Implementation Agreement OIF-CEI-04.

0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
IA Title: Common Electrical I/O (CEI) - 15
Electrical and Jitter Interoperability 16
17
agreements for 6G+ bps, 11G+ bps, 18
19
25G+ bps I/O and 56G+ bps 20
21
22
IA # OIF-CEI-04.0 23
24
25
December 29, 2017 26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Implementation Agreement created and approved 44
45
by the Optical Internetworking Forum 46
47
www.oiforum.com 48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 1

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The OIF is an international non profit organization with over 100 member companies,
2 including the world's leading carriers and vendors. Being an industry group uniting
3 representatives of the data and optical worlds, OIF's purpose is to accelerate the
4 deployment of interoperable, cost-effective and robust optical internetworks and their
5 associated technologies. Optical internetworks are data networks composed of routers
6 and data switches interconnected by optical networking elements.
7
8 With the goal of promoting worldwide compatibility of optical internetworking
9 products, the OIF actively supports and extends the work of national and international
10 standards bodies. Working relationships or formal liaisons have been established with
11 CFP-MSA, COAST, Ethernet Alliance, Fibre Channel T11, IEEE 802.1, IEEE 802.3, IETF,
12 InfiniBand, ITU-T SG13, ITU-T SG15, MEF, ONF, Rapid I/O, SAS T10, SFF Committee,
13 TMF and TMOC.
14
15 For additional information contact:
16
17 The Optical Internetworking Forum,
18
19 5177 Brandin Ct,
20
21 Fremont, CA 94538 USA
22
23
24 +1.510.492.4040 F [email protected]
25
26 www.oiforum.com
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

2 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Implementation Agreement: OIF-CEI-04.0 1

2
3
Working Group: Physical and Link Layer 4
5
6
Title: Common Electrical I/O (CEI) - Electrical and Jitter 7
Interoperability agreements for 6G+ bps, 11G+ bps and 25G+ 8
bps I/O 9
10
11
Source: Graeme Boyd Henrik Johansen Anthony Sanders
12
PMC-Sierra Intel Corporation Infineon Technologies
8555 Baxter Place Mileparken 22 Balanstr 13
Burnaby, BC, V5A 4V7 DK-2740, Skovlunde Munich 14
Canada Denmark Germany
01.0 Phone: +1-604-415-6000 Phone: +45 44 54 61 10 Phone: +49-170-6344266 15
[email protected] [email protected] [email protected] 16
17
Peter Dartnell Mike Lerer 18
Bookham Technology PLL Chair 19
Caswell Xilinx Inc. / FPGA.com
Northamptonshire, NN12 8EQ Box 636 20
UK Londonderry,NH 03053, USA 21
02.0 Phone: +44-1732-350677 Phone 1-603-548-3704
[email protected] [email protected]
22
23
Iain Robertson Klaus-Holger Otto David R. Stauffer, Ph. D. 24
Texas Instruments Alcatel-Lucent PLL Chair 25
800 Pavilion Drive Thurn-und-Taxis-Str. 10 IBM Corporation
Northampton, NN4 7YL 90411, Nuremberg 1000 River Road, MC 862J
26
UK Germany Essex Jct., VT 05452, USA 27
03.0 Phone: +44-1604-663412 Phone: +49 911 526 3594 Phone: (802) 769-6914 28
[email protected] [email protected] [email protected]
29
Klaus-Holger Otto Tom Palkert David R. Stauffer, Ph. D. 30
Alcatel-Lucent Molex PLL Chair 31
Thurn-und-Taxis-Str. 10  Kandou Bus, S.A. 32
90411, Nuremberg  
Germany   33
03.1 Phone: +49 911 526 3594 Phone: +1 (952) 200 8542 Phone: +1 (802) 316-0808 34
Klaus-Holger.Otto@alcatel- [email protected] [email protected] 35
lucent.com
Klaus-Holger Otto  David R. Stauffer, Ph. D. Edward Frlan 36
TC Chair PLL Chair TC Vice Chair 37
Nokia  Semtech
04.0 Thurn-und-Taxis-Str. 10/2 
Kandou Bus, S.A.
 38
EPFL Innovation Quartier Building I,
90411 Nuremberg 
1015 Lausanne Switzerland
 39
Germany  Canada
Phone: +49 911 526 3594 Phone: +1 (802) 316-0808 Phone: +1 613-416-1343 40
[email protected] [email protected] [email protected]
41
Dr. Mike Peng Li 42

Intel Corporation 43

 44
USA 
Phone: +1 (408)-544-8312 45
[email protected] 46
47
48
DATE: December 29, 2017 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 3

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 ABSTRACT:
2
3 This document is the CEI implementation agreement, which specifies the transmitter,
4 receiver and interconnect channel associated with 6G+ bps, 11G+ bps, 25G+ bps and
5 56G+ bps interfaces for application in high speed backplanes, chip to chip interconnect
6 and optical modules. Also included is the Jitter definition and measurement
7 methodologies associated with CEI interfaces.
8
9 Notice: This Technical Document has been created by the Optical Internetworking Forum (OIF). This
document is offered to the OIF Membership solely as a basis for agreement and is not a binding proposal
10 on the companies listed as resources above. The OIF reserves the rights to at any time to add, amend, or
11 withdraw statements contained herein. Nothing in this document is in any way binding on the OIF or
12 any of its members.
13
The user's attention is called to the possibility that implementation of the OIF implementation agreement
14 contained herein may require the use of inventions covered by the patent rights held by third parties. By
15 publication of this OIF implementation agreement, the OIF makes no representation or warranty
16 whatsoever, whether expressed or implied, that implementation of the specification will not infringe any
17 third party rights, nor does the OIF make any representation or warranty whatsoever, whether
expressed or implied, with respect to any claim that has been or may be asserted by any third party, the
18 validity of any patent rights related to any such claim, or the extent to which a license to use any such
19 rights may or may not be available or the terms hereof.
20 © 2017 Optical Internetworking Forum
21
22 This document and translations of it may be copied and furnished to others, and derivative works that
23 comment on or otherwise explain it or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction other than the following, (1) the above copyright
24 notice and this paragraph must be included on all such copies and derivative works, and (2) this
25 document itself may not be modified in any way, such as by removing the copyright notice or references
26 to the OIF, except as needed for the purpose of developing OIF Implementation Agreements.
27 By downloading, copying, or using this document in any manner, the user consents to the terms and
28 conditions of this notice. Unless the terms and conditions of this notice are breached by the user, the
29 limited permissions granted above are perpetual and will not be revoked by the OIF or its successors or
30 assigns.
31 This document and the information contained herein is provided on an "AS IS" basis and THE OIF
32 DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY
33 WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS
OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY, TITLE OR FITNESS FOR A
34 PARTICULAR PURPOSE.
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

4 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

0 Document Structure and Contents 1

2
3
0.1 Revision History 4
5
The OIF document 2003.104 was the working document used for the development of 6
the CEI-6G-SR, CEI-6G-LR, CEI-11G-SR interfaces and the jitter methodology. The 7
history of this document is detailed in the table below: 8
9
10
11
Revision Date Description 12
13
Draft 1.0. Compiled from baseline documents oif2002.605.03 (clause 0, 1),
OIF 2003.104.00 28th March 2003, OIF2002.536.06 (clause 2), oif2002.520.02 (clauses 4, 5), OIF2002.506.02 14
(clause 6) 15
Draft 2.0. Contains changes as result from comments received from Draft 1.0. 16
OIF 2003.104.01 3rd May 2003
Section added in Clause 6 relating to transparent application, derived from 17
XFP specification. Parameters added re DC coupling option, derived from
OIF2003.129 18
Draft 3.0. Updated to include approved changes from the OIF Plenary
19
OIF 2003.104.02 24th May 2003
meeting in Scottsdale, 6-8 May 2003 20
Draft 4.0. Updated to include changes as results of comment resolution from 21
OIF 2003.104.03 2nd October 2003
CEI Straw ballot (ballot#41), approved at the Ottawa meeting July 2003 22
Draft 4.1. As draft 4.0 but including changes approved at the Berlin interim/ 23
OIF 2003.104.04 17th November 2003 plenary meetings 13 - 16 October 2003. These changes are summarized in 24
OIF2003.326.03.
25
Draft 5.0. Updated to include changes as results of comment resolution from
OIF2003.104.05 10th February 2004 the second CEI Straw ballot (ballot#49), approved at the San Diego meeting 26
January 2004 27
Draft 6.0. Updated to include changes as result of comment resolution from 28
OIF2003.104.06 5th May 2004 3rd Straw ballot (ballot no 52), as approved at the Orlando Interim meeting 29
March 15th 2004.
30
Draft 7.0. As Draft 6.0, but updated to include changes approved at the
Budapest Plenary meeting. Clause 2 reconstructed and SXI-5 and TFI-5
31
OIF2003.104.07 14th July 2004
interfaces described as new clauses 4 and 5. Previous clauses 4,5,6 are 32
renumbered as clauses 6,7,8 33
OIF2003.104.08 26th August 2004
Clause 8 modified to include changes agreed at the Hawaii Plenary meeting, 34
to address discrepancies between CEI and XFP specifications.
35
Draft 9.0. Updated to include changes as result of comment resolution from 36
OIF2003.104.09 20th October 2004
4th Straw ballot (ballot no 55),
37
Draft 10.0. As draft 9.0 with specific reference to version no of State Eye
OIF2003.104.10 8th November 2004
scripts in section 2.C.5 removed. 38
39
This revision was published as OIF-CEI-01.00 in December 2004. 40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 5

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The OIF document 2003.253 was the working document used for the development of
2 the CEI-11G-MR and CEI-11G-LR interfaces. The history of this document is detailed in
3 the table below:
4
5
6
7 Revision Date Description
8 Draft 1.0. Compiled from baseline document oif2002.127.0 with changes and
OIF 2003.253.00 20th July 2003,
9 modifications from Scottsdale motions
10 Draft 1.1. adding changes and modifications from the July 2003 meeting in
11 Ottawa.
- New entries for table 1-1 moved to OIF2003.104.
12
OIF 2003.253.01 5th October 2003 - Removed figure 1-1, table 1-2 and sections 1.8 and 3.2.10.
13
- Moved appendix 3B to OIF2003.104
14 - Changed 7.2.8, 8 Taps down to 4 Taps
15 - Changed 7.1 to required BER of 1e-15
16
Draft 2.0. adding changes and modifications from the October 2003 meeting
17 OIF 2003.253.02 9th November 2003
in Berlin.
18 Draft 2.1 resolving comments from Straw ballot #50, motions and resolutions
19 as agreed in the San Diego 2004 meeting. Corrections include:
20 - DC coupling introduced with VTT = 1.2V
21 - Channel compliance, section 7.2.7 - with introduction of reference
transmitter and -receiver.
22 OIF2003.253.03 2nd February 2004
- Changes in transmit amplitude to 1200mVppd max
23 Comment resolution spread sheet, OIF2004.054.03
24 Clause 7 Editors report, OIF2004.053.01
25 PLL Meeting motions: OIF2004.076.00
26 Draft 2.2 resolving comments from straw ballot 53 and orlando interim
27 meeting, March 15th. Corrections include
28 - DC coupling editorials
29 OIF2003.253.04 3rd May 2004 - Tap weight clarification
30 - T_Y1 = 400 mVpp, T_Y2 = 600mVpp
31 - driver and receiver absolute min and max voltages
32 - Return loss alignment to 6G-LR

33 Draft 2.3 including motions from Budapest and Hawaii meetings:

34 - Changed clause no from 7 to 9
35 OIF2003.252.05 6 September 2004 - Changed values in Table 9-1 and 9-8d
- Changed reference receiver B definitions
36
- Added appendix B, the StatEye.org template.
37
38 Draft 3.0 including the motions from the Alexandria meeting, October 26-28
- Added CEI-11G-MR
39 OIF2003.253.06 6 December 2004
- Further specification of Reference Receiver B
40
- StatEye templates for -LR Ref Receiver A and B and for -MR
41
Draft 3.1 includes corrections to table 9.11 following discussions and motion
42 OIF2003.253.06
25 January 2005 from the Dallas meeting, 18-20 January 2005.
43 Source documents uploaded as OIF2005.090.00
44
45 This revision was published as OIF-CEI-02.00 in February 2005.
46
47
48
49

6 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The OIF document 2011.004 was the working document used for the development of 1
maintenance updates to OIF-CEI-02.00. The comment resolution for this update is 2
contained in 2011.121. These updates were published as part of OIF-CEI-03.00 in 3
August 2011. 4
5
The OIF document 2008.029 was the working document used for the development of 6
the CEI-28G-SR interface defined in clause 10. The history of this document is detailed 7
in the table below: 8
9
Revision Date Description 10
OIF 2008.029.03 28th July 2008, Document taken over from Beth Donnay 11
OIF 2008.029.04 Inserted text for all tbd locations according to work session results of Q2/09
12
23rd April 2009
meeting in Boston 13
OIF 2008.029.05 Finalized text proposal after continued discussion in Q2/09 meeting in Boston.
14
23rd April 2009
Text proposal sent to Straw Ballot in Boston
15
16
oif2009.129.02: Comment resolution according CEI-28-SR/25-LR Editors
OIF 2008.029.06 Report  17
23rd July 2009
Finalized text proposal after continued discussion in Q3/09 meeting in 18
Vancouver. Text proposal sent to Straw Ballot in Vancouver
19
oif2009.280.03: Comment Resolution Worksheet for CEI-28-SR  20
OIF 2008.029.07 Finalized text proposal after continued discussion in Q4/09 meeting in
15th October 2009
Lannion. Text proposal sent to Straw Ballot in Lannion and sent as liaison to 21
IEEE 802.3ba for comments 22
oif2009.408.01: Comment Resolution Worksheet for CEI-28-SR  23
OIF 2008.029.08 Finalized text proposal after continued discussion in Q2/10 meeting in Hong
21st May 2010
Kong. Text proposal sent to Straw Ballot in electronic motion after Hong Kong
24
meeting. 25
oif2010.239.01: Comment Resolution Worksheet for CEI-28-SR  26
OIF 2008.029.09
25th August 2010
Finalized text proposal after continued discussion in Q3/10 meeting in 27
Baltimore. Text proposal sent to Straw Ballot in electronic motion after
Baltimore meeting. 28
oif2010.337.02: Comment Resolution Worksheet for CEI-25/28 
29
OIF 2008.029.10
16th November 2010
Finalized text proposal after continued discussion in Q4/10 meeting in 30
Nuremberg. Text proposal sent to Straw Ballot in electronic motion after 31
Nuremberg meeting
32
oif2010.452.01: Comment Resolution Worksheet for CEI-25/28 
OIF 2008.029.11
14th February 2011 Finalized text proposal after continued discussion in Q1/11 meeting in Dallas.
33
Text proposal sent to Straw Ballot in electronic motion after Dallas meeting 34
oif2011.129.04: Comment Resolution Worksheet for CEI-25/28  35
OIF 2008.029.12
7th April 2011
Finalized text proposal after continued discussion in Q2/11 meeting in 36
Glasgow. Text proposal sent to Straw Ballot during Glasgow meeting with
option for Principal. 37
oif2011.198.01: Comment Resolution Worksheet for CEI-25/28 
38
Resolution of LSI, Qlogic Straw Ballot comments. Text proposal sent to 39
OIF 2008.029.13
3rd June 2011 another Straw Ballot in electronic motion. 40
oif2011.271.01: Comment Resolution Worksheet for CEI-25/28 41
Document sent to principal member ballot at Philadelphia meeting
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 7

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The OIF document 2008.161 was the working document used for the development of
2 the CEI-25G-LR interface defined in clause 11. The history of this document is detailed
3 in the table below:
4
5 Revision Date Description
6 OIF 2008.161.03 28th July 2008, Document taken over from Beth Donnay
7 OIF 2008.161.04 Inserted text for all tbd locations according to work session results of Q2/09
8 23rd April 2009
meeting in Boston
9
OIF 2008.161.05 Finalized text proposal after continued discussion in Q2/09 meeting in Boston.
10 23rd April 2009
Text proposal sent to Straw Ballot in Boston
11
oif2009.129.02: Comment resolution according CEI-28-SR/25-LR Editors
12 OIF 2008.161.06 Report 
23rd July 2009
13 Finalized text proposal after continued discussion in Q3/09 meeting in
Vancouver. Text proposal sent to Straw Ballot in Vancouver
14
15 OIF 2008.161.07
oif2009.281.02: Comment Resolution Worksheet for CEI-25-LR 
Finalized text proposal after continued discussion in Q4/09 meeting in
16 15th October 2009
Lannion. Text proposal sent to Straw Ballot in Lannion and sent as liaison to
17 IEEE 802.3ba for comments
18 oif2009.409.01: Comment Resolution Worksheet for CEI-25-LR 
OIF 2008.161.08 Finalized text proposal after continued discussion in Q2/10 meeting in Hong
19 21st May 2010
Kong. Text proposal sent to Straw Ballot in electronic motion after Hong Kong
20 meeting.
21 oif2010.240.01: Comment Resolution Worksheet for CEI-25-LR 
22 OIF 2008.161.09
25th August 2010
Finalized text proposal after continued discussion in Q3/10 meeting in
Baltimore. Text proposal sent to Straw Ballot in electronic motion after
23 Baltimore meeting.
24
oif2010.337.02: Comment Resolution Worksheet for CEI-25/28 
25 OIF 2008.161.10 Finalized text proposal after continued discussion in Q4/10 meeting in
16th November 2010
26 Nuremberg. Text proposal sent to Straw Ballot in electronic motion after
Nuremberg meeting
27
28 OIF 2008.161.11 oif2010.452.01: Comment Resolution Worksheet for CEI-25/28 
14th February 2011 Finalized text proposal after continued discussion in Q1/11 meeting in Dallas.
29 Text proposal sent to Straw Ballot in electronic motion after Dallas meeting
30 oif2011.129.04: Comment Resolution Worksheet for CEI-25/28 
31 OIF 2008.161.12
7th April 2011
Finalized text proposal after continued discussion in Q2/11 meeting in
Glasgow. Text proposal sent to Straw Ballot during Glasgow meeting with
32 option for Principal.
33
oif2011.198.01: Comment Resolution Worksheet for CEI-25/28 
34 Resolution of LSI, Qlogic Straw Ballot comments. Text proposal sent to
OIF 2008.161.13
35 3rd June 2011 another Straw Ballot in electronic motion.
36 oif2011.271.01: Comment Resolution Worksheet for CEI-25/28
37 Document sent to principal member ballot at Philadelphia meeting

38
39
40
41
42
43
44
45
46
47
48
49

8 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The OIF document 2010.189 was the working document used for the development of 1
the Test Methodologies for CEI-28G-SR and CEI-25G-LR defined in clause 12. The 2
history of this document is detailed in the table below: 3
4
Revision Date Description 5
Separate Clause extracted for common ‘TX Jitter and Channel Compliance 6
OIF 2010.189.00
12th May 2010 Methodologies for CEI-28G-SR and CEI-25G-LR’ in Q2/10 meeting in Hong 7
Kong.
8
OIF 2010.189.01 12th May 2010
Modifications during Hong Kong meeting 9
10
OIF 2010.189.02 Editorial changes of PLL chair, see change bars 11
21st May 2010 Text proposal sent to Straw Ballot in electronic motion after Hong Kong
meeting. 12
oif2010.241.01: Comment Resolution Worksheet for Clause 12
13
OIF 2010.189.03
25th August 2010
Finalized text proposal after continued discussion in Q3/10 meeting in 14
Baltimore. Text proposal sent to Straw Ballot in electronic motion after 15
Baltimore meeting.
16
oif2010.337.02: Comment Resolution Worksheet for CEI-25/28 
OIF 2010.189.04 Finalized text proposal after continued discussion in Q4/10 meeting in
17
16th November 2010
Nuremberg. Text proposal sent to Straw Ballot in electronic motion after 18
Nuremberg meeting 19
OIF 2010.189.05 oif2010.452.01: Comment Resolution Worksheet for CEI-25/28  20
14th February 2011 Finalized text proposal after continued discussion in Q1/11 meeting in Dallas.
Text proposal sent to Straw Ballot in electronic motion after Dallas meeting 21
oif2011.129.04: Comment Resolution Worksheet for CEI-25/28 
22
OIF 2010.189.06
7th April 2011
Finalized text proposal after continued discussion in Q2/11 meeting in 23
Glasgow. Text proposal sent to Straw Ballot during Glasgow meeting with 24
option for Principal.
25
oif2011.198.01: Comment Resolution Worksheet for CEI-25/28 
Resolution of LSI, Qlogic Straw Ballot comments. Text proposal sent to
26
OIF 2010.189.07
3rd June 2011 another Straw Ballot in electronic motion. 27
oif2011.271.01: Comment Resolution Worksheet for CEI-25/28 28
Document sent to principal member ballot at Philadelphia meeting 29
30
31
32
The combined revision including changes of above documents was published as OIF- 33
CEI-03.00 in September 2011. 34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 9

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The OIF document oif2013.329 was the working document used for the development of
2 maintenance updates to OIF-CEI-03.00. The comment resolution for this update is
3 contained in oif2014.051. These updates were published as part of OIF-CEI-04.0 in
4 October 2013.
5
6 The OIF document 2010.404 was the working document used for the development of
7 the CEI-28G-VSR interface defined in clause 13. The history of this document is
8 detailed in the table below:
9
10 Revision Date Description
11 OIF 2010.404.00 26th October 2010, Baseline text proposal
12 OIF 2010.404.01 28th October 2010, Updated baseline text proposal
13
OIF 2010.404.02 26th May 2011 Updated baseline text proposal
14
OIF 2010.404.03 14th July 2011 Updated baseline text proposal
15
16 OIF 2010.404.04 27th July 2011 oif2011.180.00: VSR change document
17 OIF 2010.404.05 2nd December 2011 oif2011.411.02: VSR editor report
18 OIF 2010.404.06 12th March 2012 oif2012.055.01: VSR editor report -05
19 OIF 2010.404.07 14th May 2012 oif2012.164.02: VSR rev6-0 editor recommendations
20 OIF 2010.404.08 18th October 2012 oif2012.253.02: CEI-28G-VSR draft 7 comment summary
21
OIF 2010.404.09 6th February 2013 oif2013.036.02: VSR editor report Jan 2013
22
23 OIF 2010.404.10 21st May 2013
oif2013.150.01: CEI-28G-VSR 9-0 comment summary with editor
recommendations
24
OIF 2010.404.11 25th July 2013 oif2013.255.02: CEI-28G-VSR rev10-0 editor report
25
26 OIF 2010.404.12 17th September 2013 oif2013.317.00: Working document for CEI VSR draft 11-1

27 OIF 2010.404.13 18th September 2013 Wrong file uploaded

28 OIF 2010.404.14 18th September 2013 oif2013.380.01: CEI-28G-VSR 14-0 comments
29 OIF 2010.404.15 30th October 2013 oif2014.068.00: CEI-VSR rev 0-15 comments
30
31 The OIF document 2013.066 was the working document used for the development of
32 the CEI-28G-MR interface defined in clause 14. The history of this document is detailed
33 in the table below:
34
35 Revision Date Description
36
OIF 2013.066.00 17th January 2013 oif2013.037.01: MR editor report Jan 2013
37
OIF 2013.066.01 21st May 2013 oif2013.162.02: MR 1-0 comment resolution spreadsheet
38
39 OIF 2013.066.02 25th July 2013 oif2013.254.03: CEI-28G-MR rev2-0 editors report
40 OIF 2013.066.03 17th September 2013 Final version
41
42 The combined revision including changes of above documents was published as OIF-
43 CEI-03.01 in February 2014.
44
45 The OIF document oif2017.059 was the working document used for the development of
46 maintenance updates to OIF-CEI-03.01 according worklist in oif2015.001.02. The
47 comment resolution for this update is contained in oif2017.321. These updates were
48 published as part of OIF-CEI-04.0 in December 2017.
49

10 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The OIF document 2014.230 was the working document used for the development of 1
the CEI-56G-VSR-PAM4 interface defined in clause 16. The history of this document is 2
detailed in the table below: 3
4
Revision Date Description 5
OIF 2014.230.00 07/21/2014 Baseline Text Proposal 6
OIF 2014.230.01 10/22/2014 oif2014.372.00 CEI-56G-VSR-PAM4 Editor's report
7
8
OIF 2014.230.02 02/06/2015 oif2014.443.02 CEI-56G-VSR-PAM4 comments resolution spreadsheet
9
OIF 2014.230.03 05/08/2015 oif2015.144.02 CEI-56G-VSR-PAM4 comment and resolution spreadsheet 10
OIF 2014.230.04 09/09/2015 11
OIF 2014.230.05 11/06/2015
oif2015.492.01CEI-56G-VSR/XSR-PAM4 comment and resolution 12
spreadsheets 13
OIF 2014.230.06 02/22/2016
oif2016.004.04 CEI-56G-VSR/XSR-PAM4 comment and resolution 14
spreadsheets
15
oif2016.128.05 Prague CEI-56G-VSR/XSR-PAM4 comment and resolution 16
OIF 2014.230.07 06/10/2016
spreadsheets
17
oif2016.252.04 Quebec City CEI-56G-VSR-PAM4 comment and resolution
OIF 2014.230.08 09/13/2016
spreadsheet 18
oif2016.402.03 Auckland CEI-56G-VSR-PAM4 comment and resolution
19
OIF 2014.230.09 11/18/2016
spreadsheet 20
oif2017.027.01 San Jose CEI-56G-VSR-PAM4 comment and resolution 21
OIF 2014.230.10 02/17/2017
spreadsheet 22
OIF 2014.230.11 06/06/2017
oif2017.148.02 Ljub. CEI-56G-VSR-PAM4 comment and resolution 23
spreadsheet 24
OIF 2014.230.12 09/25/2017
oif2017.333.02 Halifax CEI-56G-VSR-PAM4 comment and resolution 25
spreadsheet
26
OIF 2014.230.13 11/01/2017
Final version: oif2017.492.01 Shanghai CEI-56G-VSR-PAM4 comment and 27
resolution spreadsheet
28
29
The OIF document 2014.245 was the working document used for the development of 30
the CEI-56G-MR-PAM4 interface defined in clause 17. The history of this document is 31
detailed in the table below: 32
33
Revision Date Description
34
OIF 2014.245.00 07/21/2014 Baseline Text Proposal 35
OIF 2014.245.01 10/22/2014
oif2014.247.00 Method and Highlights for Creating a CEI-56G-MR Baseline 36
Proposal
37
OIF 2014.245.02 03/08/2015 oif2014.444.03 CEI-56G-MR comment and resolution spreadsheet 38
OIF 2014.245.03 05/13/2015
oif2015.198.02 Editor's initial reponses to the comments received against 39
CEI-56G-MR spec (OIF2014.245.02) ballot
40
OIF 2014.245.04 09/11/2015
oif2015.358.02 CEI-56G-MR-PAM4 comment resolution spreadsheet - OIF 41
Q3/2015 Meeting
42
oif2015.504.02 Editor's initial reponses to comments and resolutions against
OIF 2014.245.05 12/13/2015
the CEI-56G-MR-PAM4 (i.e., oif2014.245.04.pdf) spec draft
43
44
oif2016.060.02 Editor's initial reponses to comments and resolutions against
OIF 2014.245.06 04/03/2016
the CEI-56G-MR-PAM4 (i.e., oif2014.245.05.pdf) spec draft 45
OIF 2014.245.07 06/28/2016 oif2016.194.03 Editor's initial reponses to comments and resolutions
46
47
oif2016.331.02 Editor's Responses for Comment Resolution for CEI-56G-MR-
OIF 2014.245.08 09/27/2016
PAM4 (oif2014.245.07) 48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 11

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Revision Date Description

2
oif2016.446.02 Editor's initial reponses to CEI-56G-MR-PAM4 comments and
3 OIF 2014.245.09 12/12/2016
resolutions
4 oif2017.082.02 Editor's reponses to CEI-56G-MR-PAM4 comments and
5 OIF 2014.245.10 04/09/2017
resolutions (Q1/2017)
6 oif2017.233.03 Editor's reponses to CEI-56G-MR-PAM4 comments and
OIF 2014.245.11 06/27/2017
7 resolutions (Q2/2017)
8 OIF 2014.245.12 08/25/2017
oif2017.419.02 Editor's reponses to CEI-56G-MR-PAM4 comments and
9 resolutions (Q3/2017)
10 OIF 2014.245.13 11/01/2017
Final version: oif2017.563.02 Worksheet with editor's comment&resolution
pre-Q4/17 meeting for CEI-56G-MR-PAM4 (oif2014.245.12)
11
12
13 The OIF document 2014.267 was the working document used for the development of
14 the CEI-56G-USR-NRZ interface defined in clause 18. The history of this document is
15 detailed in the table below:
16
Revision Date Description
17
18 OIF 2014.267.00 07/25/2014 Baseline Text Proposal
19 OIF 2014.267.01 10/14/2014 oif2014.304.00 CEI 56G USR Editorial Comments on oif2014.267.00
20 OIF 2014.267.02 10/21/2014 oif2014.369.00 Editors report for 56G-USR
21 OIF 2014.267.03 01/22/2015 oif2015.002.03 CEI-56G-USR-NRZ comment and resolution spreadsheet
22
OIF 2014.267.04 04/22/2015 oif2015.103.02 CEI-56G-USR Version3 comment and resolution spreadsheet
23
24 OIF 2014.267.05 07/29/2015 D oif2015.254.01 CEI-56G-USR Version4 comment and resolution spreadsheet

25 OIF 2014.267.06 11/05/2015 oif2015.472.02 CEI-56G-USR-NRZ Comment Resolution

26 OIF 2014.267.07 01/27/2016 oif2016.038.03 CEI-56G-USR-NRZ Comment Resolution
27 OIF 2014.267.08 05/10/2016 oif2016.121.02 CEI-56G-USR-NRZ Comment Resolution 07
28 OIF 2014.267.09 08/10/2016 oif2016.267.02 Comment resolution for CEI-56G-USR-NRZ 08
29
OIF 2014.267.10 10/31/2016 oif2016.363.03 Comment resolution for CEI-56G-USR-NRZ 09
30
OIF 2014.267.11 01/17/2017 Final Version: oif2016.485.01 Comment resolution for CEI-56G-USR-NRZ 10
31
32
33 The OIF document 2014.268 was the working document used for the development of
34 the CEI-56G-XSR-NRZ interface defined in clause 19. The history of this document is
35 detailed in the table below:
36
Revision Date Description
37
38 OIF 2014.268.00 07/25/2014 Baseline Text Proposal
39 OIF 2014.268.01 10/14/2014 oif2014.305.00 CEI 56G XSR Editorial Comments on oif2014.268.00
40 OIF 2014.268.02 10/21/2014 oif2014.370.00 Editors report for CEI-56G-XSR-NRZ
41 OIF 2014.268.03 01/22/2015 oif2015.003.04 CEI-56G-XSR-NRZ comment and resolution spreadsheet
42
oif2015.102.02 CEI-56G-XSR-NRZ Version3 comment and resolution
43 OIF 2014.268.04 04/22/2015
spreadsheet
44 oif2015.255.01 CEI-56G-XSR-NRZ Version4 comment and resolution
45 OIF 2014.268.05 07/29/2015
spreadshee
46 OIF 2014.268.06 11/05/2015 oif2015.476.02 CEI-56G-XSR-NRZ Comment Resolution
47 OIF 2014.268.07 01/27/2016 oif2016.043.04 CEI-56G-XSR-NRZ Comment Resolution
48
OIF 2014.268.08 05/10/2016 oif2016.123.02 CEI-56G-XSR-NRZ Comment Resolution 07
49

12 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Revision Date Description 1

2
OIF 2014.268.09 08/10/2016 oif2016.269.01 Comment resolution for CEI-56G-XSR-NRZ 08
3
OIF 2014.268.10 10/31/2016 oif2016.364.02 Comment resolution for CEI-56G-XSR-NRZ 09 4
OIF 2014.268.11 01/17/2017 Final Version: oif2016.486.01 Comment resolution for CEI-56G-XSR-NRZ 10 5
6
The OIF document 2014.380 was the working document used for the development of 7
the CEI-56G-LR-PAM4 interface defined in clause 21. The history of this document is 8
detailed in the table below: 9
10
Revision Date Description 11
OIF 2014.380.00 10/18/2014 Baseline Text Proposal 12
oif2014.381.03 Method and Highlights for Creating a PAM-4 Baseline
13
OIF 2014.380.01 04/18/2015
Proposal for CEI-56G-LR 14
oif2015.505.01 Editor's initial reponses to comments and resolutions against 15
OIF 2014.380.02 04/03/2016
the CEI-56G-LR-PAM4 (i.e., oif2014.380.01.pdf) spec draft 16
OIF 2014.380.03 06/28/2016
oif2016.205.02 Editor's initial reponses to CEI-56G-LR-PAM4 comments and 17
resolutions 18
OIF 2014.380.04 09/28/2016
oif2016.334.03 Editor's Responses for Comment Resolution for CEI-56G-LR- 19
PAM4 (oif2014.380.03)
20
oif2016.447.02 Editor's initial reponses to CEI-56G-LR-PAM4 comments and 21
OIF 2014.380.05 12/12/2016
resolutions (Q4/16)
22
oif2017.083.02 Editor's reponses to CEI-56G-LR-PAM4 comments and
OIF 2014.380.06 04/09/2017
resolutions (Q1/2017) 23
oif2017.234.02 Editor's reponses to CEI-56G-LR-PAM4 comments and
24
OIF 2014.380.07 06/27/2017
resolutions (Q2/2017) 25
oif2017.420.02 Editor's reponses to CEI-56G-LR-PAM4 comments and 26
OIF 2014.380.08 08/25/2017
resolutions (Q3/2017) 27
OIF 2014.380.09 11/01/2017
Final version: oif2017.564.02 Worksheet with editor's comment&resolution 28
pre Q4/17 meeting for CEI-56G-LR-PAM4 (oif2014.380.08) 29
30
The OIF document 2014.364 was the working document used for the development of 31
the CEI-56G-LR-ENRZ interface defined in clause 22. The history of this document is 32
detailed in the table below: 33
34
Revision Date Description 35
OIF 2014.364.00 10/13/2014 Baseline Text Proposal 36
OIF 2014.364.01 01/14/2015 oif2015.046.00 CEI-56G-LR ENRZ Proposal Updates 37
OIF 2014.364.02 04/12/2015 oif2015.046.00 CEI-56G-LR ENRZ Proposal Updates
38
39
OIF 2014.364.03 07/30/2015 oif2015.341.00 CEI-56G-LR-ENRZ Proposed Updates to Baseline Text
40
OIF 2014.364.04 01/28/2016 oif2015.440.01 CEI-56G-LR-ENRZ Comment Resolution 4Q2015 41
OIF 2014.364.05 05/12/2016
Final version: oif2016.155.00 CEI-56G-LR-ENRZ Comment Resolution 42
Spreadsheet 2Q2016
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 13

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 0.2 Document Structure

2
3 The CEI document is created as a clause based document to allow for a successive
4 completion of the document as clauses are added. This reflects the split project
5 schedule where there are different schedules for completion different application
6 specifications.
7
8 The first release of the document included all clauses common for the applications
9 covered by the CEI project. These clauses were completed to cover the requirements
10 of the included applications. Further common specifications may be included as new
11 application clauses are added, resulting in an update of the common clauses. The
12 process of creating the CEI document can be explained as follows:
13
14 1. Prepare and complete all clauses necessary for the first release of the document,
15 make it the master for future documents and submit it for its approval process
16 (balloting cycles).
17 2. Follow on documents include new clauses for new functions and corrections and
18 additions to all affected clauses of the Master document. Unchanged clauses from
19 prior documents are not included, only deltas are listed (additions and deletions).
20
21 3. Once the Master document and following documents are approved it is an editorial
22 task to merge the documents.
23 4. All requirements and specifications in the application specific clauses shall be
24 referenced to the common clauses when appropriate.
25
26 5. Annexes and Appendices providing explanatory and informative text for a specific
27 application shall be included in the corresponding clause and covered by the clause
28 revision history. Information included in Annexes is normative with respect to the
29 particular clause. Information included in Appendices is informative only with
30 respect to the particular clause.
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

14 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Contents 1
2
0 Document Structure and Contents............................................................................................5 3
0.1 Revision History .....................................................................................................................5 4
0.2 Document Structure .............................................................................................................14 5
0.3 List of companies belonging to the OIF when the document is approved............................37 6
7
1 Common electrical I/O project - Introduction, definitions and formats. ..............................39
8
1.1 Introduction ..........................................................................................................................39 9
1.2 Overview ..............................................................................................................................39
10
1.3 Objectives and Requirements ..............................................................................................41
1.4 References...........................................................................................................................41
11
1.5 Abbreviations .......................................................................................................................43 12
1.6 Definitions ............................................................................................................................45 13
1.6.1 Definition of Amplitude and Swing ...............................................................................50 14
1.6.2 Definition of Skew and Relative wander ......................................................................51 15
1.6.3 Definition of Total wander ............................................................................................51 16
1.7 Table Entries and Specifications..........................................................................................52 17
1.7.1 Transmitter Electrical Output Specification..................................................................52 18
1.7.2 Receiver Electrical Input Specification.........................................................................53 19
1.7.3 Receiver input Jitter Specification................................................................................54 20
1.8 Reference Model..................................................................................................................54 21
1.A Signal Definitions .................................................................................................................56
22
1.B Examples of CEI links in Typical systems............................................................................58
23
2 Jitter and Interoperability Methodology..................................................................................59 24
2.1 Method A..............................................................................................................................59 25
2.1.1 Defined Test Patterns ..................................................................................................59 26
2.1.1.1 CID Jitter Tolerance Pattern ..................................................................................59 27
2.1.1.2 Jitter Tolerance and General Test Patterns...........................................................59 28
2.1.2 Channel Compliance ...................................................................................................59 29
2.1.3 Transmitter Compliance...............................................................................................60 30
2.1.4 Receiver Compliance...................................................................................................60
31
2.2 Method B..............................................................................................................................61
2.2.1 Defined Test Patterns ..................................................................................................61
32
2.2.2 Channel Compliance ...................................................................................................61 33
2.2.3 Transmitter Compliance...............................................................................................62 34
2.2.4 Receiver Compliance...................................................................................................62 35
2.3 Method C .............................................................................................................................63 36
2.3.1 Defined Test Patterns ..................................................................................................63 37
2.3.2 Channel Compliance ...................................................................................................63 38
2.3.3 Transmitter Compliance...............................................................................................64 39
2.3.4 Receiver Compliance...................................................................................................64 40
2.4 Method D .............................................................................................................................65 41
2.4.1 Defined Test Patterns ..................................................................................................65 42
2.4.2 Channel Compliance ...................................................................................................65
43
2.4.3 Transmitter Compliance...............................................................................................66
2.4.4 Receiver Compliance...................................................................................................66
44
2.5 Method E..............................................................................................................................67 45
2.5.1 Defined Test Patterns ..................................................................................................67 46
2.5.1.1 CID Jitter Tolerance Pattern ..................................................................................67 47
2.5.2 Channel Compliance ...................................................................................................68 48
2.5.3 Transmitter Compliance...............................................................................................68 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 15

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.5.4 Receiver Compliance ..................................................................................................70

2 2.A Masks.................................................................................................................................71
3 2.A.1 Total Wander Mask ........................................................................................................71
4 2.A.2 Relative Wander Mask ...................................................................................................71
2.A.3 Random Jitter Mask........................................................................................................72
5
2.B Pulse Response Channel Modelling ..................................................................................73
6 2.B.1 Generating a Pulse Response........................................................................................73
7 2.B.2 Basic Pulse Response Definitions..................................................................................74
8 2.B.3 Transmitter Pulse Definition ...........................................................................................75
9 2.B.4 Receiver Pulse Response ..............................................................................................76
10 2.B.5 Crosstalk Pulse Response .............................................................................................77
11 2.B.6 Decision Feedback Equalizer .........................................................................................77
12 2.B.7 Time Continuous Transverse Filter ................................................................................78
13 2.B.7.1 Annex - Time Continuous Zero-Pole Equalizer adaption .......................................78
14 2.B.8 Time Continuous Zero/Pole............................................................................................78
15 2.B.9 Degrees of Freedom.......................................................................................................79
16 2.B.9.1 Receiver Sample Point............................................................................................79
2.B.9.2 Transmit Emphasis..................................................................................................79
17
2.C Jitter Modelling ...................................................................................................................80
18 2.C.1 High Frequency Jitter vs. Wander ..................................................................................80
19 2.C.2 Total Wander vs. Relative Wander.................................................................................80
20 2.C.3 Correlated vs. Uncorrelated Jitter...................................................................................81
21 2.C.4 Jitter Distributions ...........................................................................................................82
22 2.C.4.1 Unbounded and Bounded Gaussian Distribution ....................................................82
23 2.C.4.2 Bounded Gaussian Distribution...............................................................................83
24 2.C.4.3 High Probability Jitter ..............................................................................................83
25 2.C.4.4 Total Jitter................................................................................................................83
26 2.C.4.5 Probability Distribution Function vs. Cumulative Distribution Function ...................84
27 2.C.4.6 BathTub...................................................................................................................85
28 2.C.4.7 Specification of GJ and HPJ....................................................................................87
2.C.4.8 Example of Bounded Gaussian...............................................................................87
29
2.C.5 Statistical Eye Methodology ...........................................................................................88
30 2.C.5.1 Derivation of Cursors and Calculation of PDF.........................................................88
31 2.C.5.2 Inclusion of Sampling Jitter .....................................................................................91
32 2.C.5.3 Generation of Statistical Eye ...................................................................................93
33 2.D Definition of CEI Test Patterns...........................................................................................95
34 2.D.1 PRBS31..........................................................................................................................95
35 2.D.2 Short Stress Pattern Random (SSPR) ...........................................................................95
36 2.D.3 Short Stress Pattern SDH 16 (SSPS-16) .......................................................................96
37 2.D.4 Short Stress Pattern SDH 64 (SSPS-64) .......................................................................97
38 2.D.5 Use of CEI Test Patterns................................................................................................98
39 2.D.6 Text Definitions of Patterns ............................................................................................99
40 2.E Lab Setups.........................................................................................................................106
2.E.1 High Frequency Transmit Jitter Measurement............................................................106
41
2.E.1.1 BERT Implementation ..........................................................................................106
42 2.E.1.2 Spectrum analyzer and Oscilloscope Methodology .............................................107
43 2.E.2 Total Transmit Wander Measurement ........................................................................110
44 2.E.3 Relative Transmit Wander Measurement ...................................................................111
45 2.E.4 Jitter Tolerance ...........................................................................................................112
46 2.E.4.1 Jitter Tolerance with Relative Wander Lab Setup ................................................112
47 2.E.4.2 Jitter Tolerance with no Relative Wander Lab Setup ...........................................114
48 2.E.4.3 Jitter Tolerance with Defined ISI and no Relative Wander...................................115
49 2.E.5 Jitter Transfer..............................................................................................................115

16 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.E.6 Network Analysis Measurement .................................................................................117 1

2.E.7 Eye Mask Measurement Setup...................................................................................118 2
2.F BER Adjustment Methodology ...........................................................................................120 3
2.F.1 Extrapolation of Correlated Bounded Gaussian Jitter to low BERs ............................120 4
2.F.2 Confidence Level of Errors Measurement ..................................................................120
5
2.F.3 Eye Mask Adjustment for Sampling Oscilloscopes.....................................................121
2.F.3.1 Theory ..................................................................................................................122
6
2.F.3.2 Usage ...................................................................................................................124 7
8
3 Common Electrical Specification ..........................................................................................125 9
3.1 Introduction ........................................................................................................................125 10
3.2 General requirements ........................................................................................................125 11
3.2.1 Data Patterns.............................................................................................................125 12
3.2.2 Signal Levels .............................................................................................................126 13
3.2.3 Bit Error Ratio ............................................................................................................126 14
3.2.4 Ground Differences....................................................................................................126
15
3.2.5 Cross Talk..................................................................................................................126
3.2.6 Driver Test Load ........................................................................................................127
16
3.2.7 Driver Lane-to-Lane Skew .........................................................................................127 17
3.2.8 Input Lane-to-Lane Skew...........................................................................................127 18
3.2.9 Driver Short Circuit Current .......................................................................................127 19
3.2.10 Differential Resistance and Return Loss, Driver and Receiver..................................127 20
3.2.11 Baud Rate Tolerance.................................................................................................128 21
3.2.12 Termination and DC Blocking ....................................................................................128 22
3.A Transmission Line Theory and Channel Information .........................................................130 23
3.A.1 Transmission Lines Theory.........................................................................................130 24
3.A.1.1 Impedance Matching. ...........................................................................................130 25
3.A.1.2 Impedance Definition Details................................................................................131 26
3.A.2 Density considerations................................................................................................133
27
3.A.3 Common-Mode Impedance and Return Loss .............................................................134
3.A.4 Crosstalk Considerations. ...........................................................................................134
28
3.A.5 Equation based Channel Loss by curve fit. ................................................................136 29
30
4 SxI-5, SFI-4.2, SFI-5.1 & SPI-5.1 Interfaces ...........................................................................139 31
4.1 Introduction ........................................................................................................................139 32
4.2 General Requirements .......................................................................................................139 33
4.2.1 Channel Compliance .................................................................................................139 34
4.3 Electrical Characteristics....................................................................................................140 35
4.A StatEye.org Template ........................................................................................................141 36
5 TFI-5 Interface..........................................................................................................................145 37
5.1 Introduction ........................................................................................................................145 38
5.2 General Requirements .......................................................................................................145 39
5.2.1 Channel Compliance .................................................................................................145 40
5.3 Electrical Characteristics....................................................................................................146 41
5.A StatEye.org Template ........................................................................................................147 42
6 CEI-6G-SR Short Reach Interface ..........................................................................................151
43
44
6.1 Introduction ........................................................................................................................151
45
6.2 Requirements.....................................................................................................................151
6.3 General Requirements .......................................................................................................151
46
6.3.1 Data Patterns.............................................................................................................151 47
6.3.2 Signal levels...............................................................................................................152 48
6.3.3 Signal Definitions .......................................................................................................152 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 17

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 6.3.4 Bit Error Ratio ............................................................................................................152

2 6.3.5 Ground Differences ...................................................................................................152
3 6.3.6 Cross Talk .................................................................................................................152
4 6.3.7 Channel Compliance .................................................................................................152
6.4 Electrical Characteristics....................................................................................................153
5
6.4.1 Driver Characteristics ................................................................................................153
6 6.4.1.1 Driver Test Load ..................................................................................................154
7 6.4.1.2 Driver Baud Rate .................................................................................................154
8 6.4.1.3 Driver Amplitude and Swing ................................................................................155
9 6.4.1.4 Driver Rise and Fall Times ..................................................................................155
10 6.4.1.5 Driver Resistance and Return Loss.....................................................................155
11 6.4.1.6 Driver Lane-to-Lane Skew...................................................................................155
6.4.1.7 Driver Short Circuit Current .................................................................................155
12 6.4.1.8 Driver Template and Jitter ...................................................................................155
13 6.4.1.9 Driver Training Pattern ........................................................................................156
14 6.4.2 Receiver Characteristics............................................................................................156
15 6.4.2.1 Input Baud Rate...................................................................................................157
16 6.4.2.2 Reference Input Signals ......................................................................................158
17 6.4.2.3 Input Signal Amplitude.........................................................................................158
6.4.2.4 Absolute Input Voltage ........................................................................................158
18 6.4.2.5 Input Common Mode Impedance ........................................................................158
19 6.4.2.6 Input Lane-to-Lane Skew ....................................................................................159
20 6.4.2.7 Input Resistance and Return Loss ......................................................................159
21 6.4.2.8 Input Jitter Tolerance...........................................................................................159
22 6.A Link and Jitter Budgets ......................................................................................................160
23 6.B StatEye.org Template ........................................................................................................161
24 7 CEI-6G-LR Long Reach Interface...........................................................................................165
25 7.1 Introduction ........................................................................................................................165
26 7.2 Requirements.....................................................................................................................165
27 7.3 General Requirements.......................................................................................................165
28 7.3.1 Data Patterns.............................................................................................................165
29 7.3.2 Signal levels ..............................................................................................................166
30 7.3.3 Signal Definitions .......................................................................................................166
31 7.3.4 Bit Error Ratio ............................................................................................................166
32 7.3.5 Ground Differences ...................................................................................................166
33 7.3.6 Cross Talk .................................................................................................................166
34 7.3.7 Channel Compliance .................................................................................................166
7.4 Electrical Characteristics....................................................................................................167
35
7.4.1 Driver Characteristics ................................................................................................168
36 7.4.1.1 Driver Test Load ..................................................................................................169
37 7.4.1.2 Driver Baud Rate .................................................................................................169
38 7.4.1.3 Driver Amplitude and Swing ................................................................................169
39 7.4.1.4 Driver Rise and Fall Times ..................................................................................169
40 7.4.1.5 Output Resistance and Return Loss....................................................................169
41 7.4.1.6 Driver Lane-to-Lane Skew...................................................................................169
7.4.1.7 Driver Short Circuit Current .................................................................................170
42 7.4.1.8 Driver Template and Jitter ...................................................................................170
43 7.4.1.9 Driver Training Pattern ........................................................................................170
44 7.4.2 Receiver Characteristics............................................................................................171
45 7.4.2.1 Baud Rate............................................................................................................172
46 7.4.2.2 Reference Input Signals ......................................................................................172
47 7.4.2.3 Input Signal Amplitude.........................................................................................172
7.4.2.4 Absolute Input Voltage ........................................................................................172
48 7.4.2.5 Input Common Mode Impedance ........................................................................172
49

18 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7.4.2.6 Input Lane-to-Lane Skew ....................................................................................173 1

7.4.2.7 Input Resistance and Return Loss ......................................................................173 2
7.4.2.8 Jitter Tolerance ....................................................................................................173
7.A Link and Jitter Budgets ......................................................................................................174
3
7.B StatEye.org Template ........................................................................................................175 4
5
8 CEI-11G-SR Short Reach Interface ........................................................................................179 6
8.1 Requirements.....................................................................................................................179 7
8.2 General Requirements .......................................................................................................179 8
8.2.1 Data Patterns.............................................................................................................179 9
8.2.2 Signal levels...............................................................................................................180 10
8.2.3 Signal Definitions .......................................................................................................180 11
8.2.4 Bit Error Ratio ............................................................................................................180
12
8.2.5 Ground Differences....................................................................................................180
8.2.6 Cross Talk..................................................................................................................180
13
8.2.7 Channel Compliance .................................................................................................180 14
8.3 Electrical Characteristics....................................................................................................181 15
8.3.1 Driver Characteristics ................................................................................................181 16
8.3.1.1 Driver Baud Rate .................................................................................................183 17
8.3.1.2 Driver Test Load ..................................................................................................183 18
8.3.1.3 Driver Resistance and Return Loss .....................................................................183 19
8.3.1.4 Driver Lane-to-Lane Skew ...................................................................................183 20
8.3.1.5 Driver Short Circuit Current .................................................................................183
8.3.2 Receiver Characteristics............................................................................................183
21
8.3.2.1 Input Baud Rate...................................................................................................185 22
8.3.2.2 Reference Input Signals ......................................................................................185 23
8.3.2.3 Input Resistance and Return Loss ......................................................................185 24
8.3.2.4 Input Lane-to-Lane Skew ....................................................................................186 25
8.4 Specifications for Jitter-transparent applications................................................................186 26
8.4.1 Jitter Requirements for Transparent Applications in Telecom systems.....................186 27
8.4.1.1 Sinusoidal Jitter tolerance mask for Ingress direction, CEI receiver at reference  28
point RI. ...............................................................................................................187
8.4.1.2 Sinusoidal Jitter tolerance mask for Egress direction, CEI receiver at reference  29
point RE. ..............................................................................................................188 30
8.4.1.3 Telecom Jitter transfer .........................................................................................188 31
8.4.1.4 Telecom Jitter Generation for Egress Direction...................................................189 32
8.4.2 Jitter Requirements for Transparent Applications in Datacom systems ....................190 33
8.4.2.1 Sinusoidal Jitter tolerance mask for Ingress direction, CEI Receiver at reference  34
point D .................................................................................................................190
8.4.2.2 Datacom Jitter transfer ........................................................................................191
35
8.4.3 Jitter Transparency compliance nomenclature ..........................................................191 36
8.A Informative Jitter Budget ....................................................................................................192 37
8.B StatEye.org Template ........................................................................................................193 38
8.C XFP reference points .........................................................................................................195 39
40
9 CEI-11G-LR/MR Long/Medium Reach Interface....................................................................197
41
9.1 Requirements.....................................................................................................................198 42
9.2 General Requirements .......................................................................................................198 43
9.2.1 Data Patterns.............................................................................................................198
44
9.2.2 Signal Levels .............................................................................................................198
9.2.3 Signal Definitions .......................................................................................................198
45
9.2.4 Bit Error Ratio ............................................................................................................198 46
9.2.5 Ground Differences....................................................................................................198 47
9.2.6 Cross Talk..................................................................................................................198 48
9.2.7 Channel Compliance .................................................................................................199 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 19

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9.2.7.1 CEI-11G-LR Channel Compliance ......................................................................199

2 9.2.7.2 CEI-11G-MR Channel Compliance .....................................................................200
3 9.3 Electrical Characteristics, CEI-11G-LR and CEI-11G-MR .................................................200
9.3.1 Driver Characteristics ................................................................................................200
4
9.3.1.1 Driver Baud Rate .................................................................................................202
5 9.3.1.2 Driver Amplitude and Swing ................................................................................202
6 9.3.1.3 Driver Resistance and Return Loss.....................................................................202
7 9.3.1.4 Driver Lane-to-Lane Skew...................................................................................202
8 9.3.1.5 Driver Short Circuit Current .................................................................................202
9 9.3.1.6 Driver Template and Jitter ...................................................................................202
10 9.3.2 CEI-11G-LR Receiver Characteristics .......................................................................203
9.3.2.1 Input Baud Rate...................................................................................................203
11 9.3.2.2 Absolute Input Voltage ........................................................................................204
12 9.3.2.3 Input Resistance and Return Loss ......................................................................204
13 9.3.2.4 Input Signal Amplitude.........................................................................................204
14 9.3.2.5 Input Lane-to-Lane Skew ....................................................................................204
15 9.3.3 CEI-11G-MR Receiver Characteristics ......................................................................204
16 9.3.3.1 Input Baud Rate...................................................................................................205
17 9.3.3.2 Reference Input Signals ......................................................................................206
9.3.3.3 Input Resistance and Return Loss ......................................................................206
18 9.3.3.4 Input Lane-to-Lane Skew ....................................................................................206
19 9.A Informative Jitter Budgets ..................................................................................................207
20 9.A.1 Informative Jitter Budget for Long Reach ...................................................................207
21 9.A.2 Informative Jitter Budget for Medium Reach ..............................................................207
22 9.B StatEye.org templates........................................................................................................209
23 9.B.1 StatEye.org templates for CEI-11G-LR, reference receiver A ....................................209
24 9.B.2 StatEye.org Templates for CEI-11G-LR, reference receiver B...................................211
25 9.B.3 StatEye.org templates for CEI-11G-MR reach ...........................................................213
26 10 CEI-28G-SR Short Reach Interface........................................................................................217
27 10.1 Requirements.....................................................................................................................217
28 10.2 General Requirements.......................................................................................................217
29 10.2.1 Data Patterns.............................................................................................................217
30 10.2.2 Signal levels ..............................................................................................................217
31 10.2.3 Signal Definitions .......................................................................................................217
32 10.2.4 Bit Error Ratio ............................................................................................................218
33 10.2.5 Ground Differences ...................................................................................................218
34 10.2.6 Channel Compliance .................................................................................................218
35 10.2.6.1 Reference Model .................................................................................................218
10.2.6.2 Insertion Loss ......................................................................................................219
36 10.2.6.3 Fitted insertion loss..............................................................................................221
37 10.2.6.4 Insertion loss deviation (ILD) ...............................................................................221
38 10.2.6.5 Channel differential return loss............................................................................222
39 10.2.6.6 Channel integrated crosstalk noise .....................................................................222
40 10.3 Electrical Characteristics....................................................................................................224
41 10.3.1 Transmitter Characteristics........................................................................................224
42 10.3.1.1 Transmitter Baud Rate ........................................................................................225
10.3.1.2 Transmitter Amplitude and Swing........................................................................225
43 10.3.1.3 Transmitter Resistance and Return Loss ............................................................225
44 10.3.1.4 Transmitter Lane-to-Lane Skew ..........................................................................225
45 10.3.1.5 Transmitter Short Circuit Current.........................................................................226
46 10.3.1.6 Transmitter output waveform requirements.........................................................226
47 10.3.1.6.1 Summary of requirements.............................................................................226
48 10.3.1.6.2 Process to compute coefficients ...................................................................226
10.3.1.6.3 Waveform acquisition....................................................................................228
49

20 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

10.3.1.6.4 Linear fit to the waveform measured at T......................................................228 1

10.3.1.6.5 Removal of the transfer function between the transmit function and T .........230 2
10.3.2 Receiver Characteristics............................................................................................231 3
10.3.2.1 Input Baud Rate...................................................................................................231
10.3.2.2 Reference Input Signals ......................................................................................232
4
10.3.2.3 Input Resistance and Return Loss ......................................................................232 5
10.3.2.4 Input Lane-to-Lane Skew ....................................................................................232 6
10.3.2.5 Absolute Input Voltage ........................................................................................232 7
11 CEI-25G-LR Long Reach Interface.........................................................................................233 8
9
11.1 Requirements.....................................................................................................................233
11.2 General Requirements .......................................................................................................233
10
11.2.1 Data Patterns.............................................................................................................233 11
11.2.2 Signal levels...............................................................................................................233 12
11.2.3 Signal Definitions .......................................................................................................233 13
11.2.4 Bit Error Ratio ............................................................................................................234 14
11.2.5 Ground Differences....................................................................................................234 15
11.2.6 Channel Compliance .................................................................................................234 16
11.2.6.1 Reference Model .................................................................................................234 17
11.2.6.2 Insertion Loss ......................................................................................................235 18
11.2.6.3 Fitted insertion loss..............................................................................................237
11.2.6.4 Insertion loss deviation (ILD) ...............................................................................237
19
11.2.6.5 Channel Return Loss ...........................................................................................238 20
11.2.6.6 Channel integrated crosstalk noise .....................................................................238 21
11.3 Electrical Characteristics....................................................................................................240 22
11.3.1 Transmitter Characteristics........................................................................................240 23
11.3.1.1 Transmitter Baud Rate ........................................................................................241 24
11.3.1.2 Transmitter Amplitude and Swing........................................................................241 25
11.3.1.3 Transmitter Resistance and Return Loss ............................................................241
11.3.1.4 Transmitter Lane-to-Lane Skew ..........................................................................241
26
11.3.1.5 Transmitter Short Circuit Current.........................................................................242 27
11.3.1.6 Transmitter output waveform requirements .........................................................242 28
11.3.1.6.1 Summary of requirements.............................................................................242 29
11.3.1.6.2 Process to compute coefficients ...................................................................242 30
11.3.1.6.3 Waveform acquisition....................................................................................244 31
11.3.1.6.4 Linear fit to the waveform measured at T......................................................244
11.3.1.6.5 Removal of the transfer function between the transmit function and T .........246 32
11.3.2 Receiver Characteristics............................................................................................247 33
11.3.2.1 Input Baud Rate...................................................................................................247 34
11.3.2.2 Reference Input Signals ......................................................................................248 35
11.3.2.3 Input Resistance and Return Loss ......................................................................248 36
11.3.2.4 Input Lane-to-Lane Skew ....................................................................................248 37
11.3.2.5 Absolute Input Voltage ........................................................................................248
38
12 Test Methodologies for CEI Baud Rates above 11G............................................................249 39
12.1 TX jitter measurement methodology ..................................................................................249 40
12.1.1 Data Dependent Jitter (DDJ) measurement ..............................................................250 41
12.2 Channel compliance methodology .....................................................................................251 42
12.2.1 Channel Compliance .................................................................................................251 43
12.2.1.1 Fitted insertion loss..............................................................................................251 44
12.2.1.2 Integrated crosstalk noise....................................................................................254 45
12.3 Common Mode Noise ........................................................................................................256
46
13 CEI-28G-VSR Very Short Reach Interface.............................................................................257 47
13.1 Requirements.....................................................................................................................257 48
13.2 General CEI Requirements ................................................................................................257 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 21

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 13.2.1 Data Patterns.............................................................................................................257

2 13.2.2 Signal levels ..............................................................................................................258
3 13.2.3 Signal Definitions .......................................................................................................258
4 13.2.4 Bit Error Ratio ............................................................................................................258
13.2.5 Ground Differences ...................................................................................................258
5
13.3 Electrical Characteristics....................................................................................................258
6 13.3.1 Compliance Point Specifications ...............................................................................258
7 13.3.2 Host-to-Module Electrical Specifications ...................................................................260
8 13.3.3 Module-to-Host Electrical Specifications ...................................................................262
9 13.3.4 Output Differential Voltage, pk-pk..............................................................................263
10 13.3.5 Common Mode Noise ................................................................................................263
11 13.3.6 Differential Termination Resistance Mismatch ..........................................................263
12 13.3.7 Differential Return Loss .............................................................................................264
13 13.3.8 Common to differential mode and differential to common mode conversion.............265
14 13.3.9 Common Mode Return Loss......................................................................................267
15 13.3.10 Transition Time ..........................................................................................................267
16 13.3.11 Eye Width, Eye Height and Stressed Input tests .......................................................268
13.3.11.1 Host and Module output Eye Width and Eye Height test.....................................268
17 13.3.11.1.1 Host and Module output test method ............................................................270
18 13.3.11.2 Host and Module stressed input test ...................................................................272
19 13.3.11.2.1 Host (TP4a) and Module (TP1) stressed input test method..........................274
20 13.3.11.2.1.1 Host input test signal calibration ...........................................................275
21 13.3.11.2.1.2 Module input test signal calibration .......................................................276
22 13.3.11.3 Reference receiver ..............................................................................................276
23 13.3.12 Input Differential Voltage Tolerance ..........................................................................279
24 13.4 Measurement methods ......................................................................................................279
13.4.1 Compliance Boards ...................................................................................................279
25
13.4.1.1 HCB and MCB insertion loss ...............................................................................279
26 13.4.1.2 Mated HCB and MCB S parameters ...................................................................280
27 13.A Recommended Electrical Channel.....................................................................................285
28 13.A.1 Insertion Loss ............................................................................................................285
29 13.B Informative Host Transmitter output Electrical Characteristics ..........................................286
30 13.B.1 Host Transmitter output specification point................................................................287
31 13.B.1.1 Host-to-Module transmitter output Electrical Specifications .................................287
32
14 CEI-28G-MR Medium Reach Interface ...................................................................................289
33
34 14.1 Requirements.....................................................................................................................289
14.2 General Requirements.......................................................................................................289
35
14.2.1 Data Patterns.............................................................................................................289
36 14.2.2 Signal levels ..............................................................................................................289
37 14.2.3 Signal Definitions .......................................................................................................290
38 14.2.4 Bit Error Ratio ............................................................................................................290
39 14.2.5 Ground Differences ...................................................................................................290
40 14.2.6 Channel Compliance .................................................................................................290
41 14.2.6.1 Reference Model .................................................................................................290
42 14.2.6.2 Insertion Loss ......................................................................................................291
43 14.2.6.3 Fitted insertion loss..............................................................................................293
14.2.6.4 Insertion loss deviation (ILD) ...............................................................................293
44 14.2.6.5 Channel Return Loss...........................................................................................294
45 14.2.6.6 Channel integrated crosstalk noise .....................................................................295
46 14.3 Electrical Characteristics....................................................................................................297
47 14.3.1 Transmitter Characteristics........................................................................................297
48 14.3.1.1 Transmitter Baud Rate ........................................................................................298
49 14.3.1.2 Transmitter Amplitude and Swing........................................................................298

22 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14.3.1.3 Transmitter Return Loss ......................................................................................298 1

14.3.1.4 Transmitter Lane-to-Lane Skew ..........................................................................298 2
14.3.1.5 Transmitter Short Circuit Current.........................................................................299
14.3.1.6 Transmitter output waveform requirements .........................................................299
3
14.3.1.6.1 Summary of requirements.............................................................................299 4
14.3.1.6.2 Process to compute coefficients ...................................................................299 5
14.3.1.6.3 Waveform acquisition....................................................................................301 6
14.3.1.6.4 Linear fit to the waveform measured at T......................................................301 7
14.3.1.6.5 Removal of the transfer function between the transmit function and T .........303 8
14.3.2 Receiver Characteristics............................................................................................304
9
14.3.2.1 Input Baud Rate...................................................................................................304
14.3.2.2 Reference Input Signals ......................................................................................305 10
14.3.2.3 Input Resistance and Return Loss ......................................................................305 11
14.3.2.4 Input Lane-to-Lane Skew ....................................................................................305 12
14.3.2.5 Single Ended Input Voltage .................................................................................305 13
15 Reserved to add future Clause ..............................................................................................307 14
15
16 CEI-56G-VSR-PAM4 Very Short Reach Interface..................................................................309 16
16.1 Requirements.....................................................................................................................309 17
16.2 General CEI Requirements ................................................................................................310 18
16.2.1 Data Patterns.............................................................................................................310 19
16.2.2 Transmitter equalizer function ...................................................................................310 20
16.2.3 Bit Error Ratio ............................................................................................................310
21
16.2.4 Ground Differences....................................................................................................310
16.3 Electrical Characteristics....................................................................................................310
22
16.3.1 Compliance Point Specifications ...............................................................................310 23
16.3.2 Host-to-Module Electrical Specifications ...................................................................311 24
16.3.3 Module-to-Host Electrical Specifications ...................................................................314 25
16.3.4 Output Differential Voltage, pk-pk..............................................................................315 26
16.3.5 Common Mode Noise ................................................................................................315 27
16.3.6 Differential Termination Resistance Mismatch ..........................................................316 28
16.3.7 Differential Return Loss .............................................................................................316 29
16.3.8 Common to differential mode and differential to common mode conversion .............317 30
16.3.9 Common Mode Return Loss......................................................................................318 31
16.3.10 Eye Width, Eye Height and Stressed Input tests .......................................................318 32
16.3.10.1 Host and Module output Eye Width and Eye Height test.....................................318
16.3.10.1.1 Host and Module output test method ............................................................320
33
16.3.10.2 Measured PAM4 Eye Parameter Definitions .......................................................323 34
16.3.10.3 Host and Module stressed input test ...................................................................325 35
16.3.10.3.1 Host (TP4a) and Module (TP1) stressed input test method..........................327 36
16.3.10.3.1.1 Host input test signal calibration............................................................328 37
16.3.10.3.1.2 Module input test signal calibration .......................................................329 38
16.3.10.4 Reference receiver ..............................................................................................329 39
16.3.11 Input Overload Voltage Tolerance .............................................................................332
40
16.4 Measurement methods ......................................................................................................333
16.4.1 Compliance Boards ...................................................................................................333
41
16.4.1.1 Mated HCB and MCB S-parameters ...................................................................333 42
16.A Recommended Electrical Channel.....................................................................................334 43
16.A.1 Insertion Loss .............................................................................................................334 44
16.B Informative Host Transmitter output Electrical Recommendations ....................................336 45
16.B.1 Host Transmitter output test point...............................................................................336 46
16.B.1.1 Host-to-Module transmitter output Electrical Recommendations .........................336 47
16.C General PAM4 Requirement ..............................................................................................337 48
16.C.1 Transmit Functional Requirements .............................................................................337 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 23

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 16.C.1.1 Gray Coding .........................................................................................................337

2 16.C.1.2 PAM4 Encoding ...................................................................................................337
3 16.C.1.3 Optional Precoding...............................................................................................338
4 16.C.2 Receive Functional Requirements..............................................................................338
16.C.2.1 Optional Precoding Decoder ................................................................................338
5
16.C.3 Test Patterns ..............................................................................................................338
6 16.C.3.1 Quaternary PRBS13 test pattern - QPRBS13-CEI...............................................338
7 16.C.3.2 Quaternary PRBS31 test pattern - QPRBS31-CEI...............................................339
8 16.C.4 PAM4 Signal Parameters ...........................................................................................339
9 16.C.4.1 Transition Time and Slew Time Based on the QPRBS13-CEI Test Pattern ........339
10 16.C.4.2 Eye Linearity ........................................................................................................339
11 16.C.4.3 Transmitter Linearity ............................................................................................340
12 16.C.5 PAM4 Data Patterns ...................................................................................................341
13 16.D FEC Guidance and Background Material ..........................................................................342
14 17 CEI-56G-MR-PAM4 Medium Reach Interface........................................................................345
15
17.1 Requirements.....................................................................................................................345
16 17.2 General Requirements.......................................................................................................345
17 17.2.1 Data Patterns.............................................................................................................345
18 17.2.2 Bit Error Ratio ............................................................................................................346
19 17.2.3 Ground Differences ...................................................................................................346
20 17.2.4 Channel Compliance .................................................................................................346
21 17.2.4.1 Reference Model .................................................................................................346
22 17.2.4.2 Channel Operating Margin ..................................................................................346
23 17.2.4.3 Informative Channel Insertion Loss .....................................................................348
17.2.4.4 Channel Return Loss...........................................................................................349
24 17.2.4.5 Channel AC-coupling...........................................................................................350
25 17.3 Electrical Characteristics....................................................................................................351
26 17.3.1 Transmitter Characteristics........................................................................................351
27 17.3.1.1 Transmitter Baud Rate ........................................................................................352
28 17.3.1.2 Transmitter Amplitude and Swing........................................................................352
29 17.3.1.3 Transmitter Return Loss ......................................................................................352
30 17.3.1.4 Transmitter Lane-to-Lane Skew ..........................................................................354
17.3.1.5 Transmitter Short Circuit Current.........................................................................354
31 17.3.1.6 Transmitter output waveform requirements.........................................................354
32 17.3.1.6.1 Linear fit to the measured waveform.............................................................355
33 17.3.1.6.2 Steady-state voltage and linear fit pulse peak ..............................................356
34 17.3.1.6.3 Transmitter equalizer coefficients .................................................................356
35 17.3.1.6.4 Transmitter Output Noise and Distortion.......................................................356
36 17.3.1.7 Transmitter output jitter........................................................................................357
17.3.1.7.1 J4u and JRMS Jitter......................................................................................358
37 17.3.1.7.2 Even-Odd Jitter (EOJ)...................................................................................358
38 17.3.2 Receiver Characteristics............................................................................................359
39 17.3.2.1 Input Baud Rate...................................................................................................360
40 17.3.2.2 Reference Input Signals ......................................................................................360
41 17.3.2.3 Receiver Input Return Loss .................................................................................360
42 17.3.2.4 Receiver Interference Tolerance .........................................................................361
17.3.2.5 Receiver Jitter Tolerance.....................................................................................362
43 17.3.2.6 Single Ended Input Voltage .................................................................................363
44 17.3.2.7 Input Lane-to-Lane Skew ....................................................................................363
45
46 18 CEI-56G-USR-NRZ Ultra Short Reach Interface ...................................................................365
47 18.1 Requirements.....................................................................................................................365
48 18.2 General Requirements.......................................................................................................365
49 18.2.1 Data Patterns.............................................................................................................365

24 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.2.2 Signal levels...............................................................................................................365 1

18.2.3 Signal Definitions .......................................................................................................366 2
18.2.4 Bit Error Ratio ............................................................................................................366 3
18.2.5 Ground Differences....................................................................................................366 4
18.2.6 Cross Talk..................................................................................................................366
5
18.2.7 Channel Compliance .................................................................................................366
18.2.7.1 Reference Model .................................................................................................366
6
18.2.7.2 Insertion Loss ......................................................................................................367 7
18.2.7.3 Fitted Insertion Loss ............................................................................................368 8
18.2.7.4 Insertion Loss Deviation (ILD) .............................................................................369 9
18.2.7.5 Channel Return Loss ...........................................................................................369 10
18.2.7.6 Channel integrated crosstalk noise .....................................................................370 11
18.3 Electrical Characteristics....................................................................................................371
12
18.3.1 Reference Clock ........................................................................................................371
18.3.2 Transmitter Characteristics........................................................................................372
13
18.3.2.1 Transmitter Amplitude and Swing........................................................................373 14
18.3.2.2 Transmitter Resistance and Return Loss ............................................................373 15
18.3.2.3 Transmitter Lane-to-Lane Skew ..........................................................................374 16
18.3.2.4 Transmitter Template and Jitter...........................................................................374 17
18.3.3 Receiver Characteristics............................................................................................375 18
18.3.3.1 Reference Receiver Input Signals .......................................................................375 19
18.3.3.2 Receiver Return Loss ..........................................................................................376
18.3.3.3 Input Lane-to-Lane Skew ....................................................................................376
20
21
19 CEI-56G-XSR-NRZ Extra Short Reach Interface ...................................................................381 22
19.1 Requirements.....................................................................................................................381 23
19.2 General Requirements .......................................................................................................381 24
19.2.1 Data Patterns.............................................................................................................381 25
19.2.2 Signal levels...............................................................................................................381 26
19.2.3 Signal Definitions .......................................................................................................382 27
19.2.4 Bit Error Ratio ............................................................................................................382
28
19.2.5 Ground Differences....................................................................................................382
19.2.6 Cross Talk..................................................................................................................382
29
19.2.7 Channel Compliance .................................................................................................382 30
19.2.7.1 Reference Model .................................................................................................382 31
19.2.7.2 Insertion Loss ......................................................................................................383 32
19.2.7.3 Fitted Insertion Loss ............................................................................................384 33
19.2.7.4 Insertion Loss Deviation (ILD) .............................................................................385 34
19.2.7.5 Channel Return Loss ...........................................................................................385 35
19.2.7.6 Channel integrated crosstalk noise .....................................................................386
19.3 Electrical Characteristics....................................................................................................387
36
19.3.1 Reference Clock ........................................................................................................387 37
19.3.2 Transmitter Characteristics........................................................................................388 38
19.3.2.1 Transmitter Amplitude and Swing........................................................................389 39
19.3.2.2 Transmitter Resistance and Return Loss ............................................................389 40
19.3.2.3 Transmitter Lane-to-Lane Skew ..........................................................................390 41
19.3.2.4 Transmitter Template and Jitter...........................................................................390 42
19.3.3 Receiver Characteristics............................................................................................391 43
19.3.3.1 Reference Receiver Input Signals .......................................................................391
19.3.3.2 Receiver Return Loss ..........................................................................................392
44
19.3.3.3 Input Lane-to-Lane Skew ....................................................................................392 45
46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 25

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 20 Reserved to add future Clause ..............................................................................................393

2 21 CEI-56G-LR-PAM4 Long Reach Interface .............................................................................395
3
21.1 Requirements.....................................................................................................................395
4 21.2 General Requirements.......................................................................................................395
5 21.2.1 Data Patterns.............................................................................................................395
6 21.2.2 Bit Error Ratio ............................................................................................................396
7 21.2.3 Ground Differences ...................................................................................................396
8 21.2.4 Channel Compliance .................................................................................................396
9 21.2.4.1 Reference Model .................................................................................................396
10 21.2.4.2 Channel Operating Margin ..................................................................................396
11 21.2.4.3 Informative Channel Insertion Loss .....................................................................398
21.2.4.4 Channel Return Loss...........................................................................................399
12 21.2.4.5 Channel AC-coupling...........................................................................................400
13 21.3 Electrical Characteristics....................................................................................................401
14 21.3.1 Transmitter Characteristics........................................................................................401
15 21.3.1.1 Transmitter Baud Rate ........................................................................................401
16 21.3.1.2 Transmitter Amplitude and Swing........................................................................402
17 21.3.1.3 Transmitter Return Loss ......................................................................................402
18 21.3.1.4 Transmitter Lane-to-Lane Skew ..........................................................................404
21.3.1.5 Transmitter Short Circuit Current.........................................................................404
19 21.3.1.6 Transmitter output waveform requirements.........................................................404
20 21.3.1.6.1 Linear fit to the measured waveform.............................................................405
21 21.3.1.6.2 Steady-state voltage and linear fit pulse peak ..............................................406
22 21.3.1.6.3 Transmitter equalizer coefficients .................................................................406
23 21.3.1.6.4 Transmitter Output Noise and Distortion.......................................................406
24 21.3.1.7 Transmitter output jitter........................................................................................407
21.3.1.7.1 J4u and JRMS Jitter......................................................................................408
25 21.3.1.7.2 Even-Odd Jitter (EOJ)...................................................................................408
26 21.3.2 Receiver Characteristics............................................................................................409
27 21.3.2.1 Input Baud Rate...................................................................................................410
28 21.3.2.2 Reference Input Signals ......................................................................................410
29 21.3.2.3 Input Resistance and Return Loss ......................................................................410
30 21.3.2.4 Receiver Interference Tolerance .........................................................................411
21.3.2.5 Receiver Jitter Tolerance.....................................................................................412
31 21.3.2.6 Single Ended Input Voltage .................................................................................413
32 21.3.2.7 Input Lane-to-Lane Skew ....................................................................................413
33
34 22 CEI-56G-LR-ENRZ Long Reach Interface..............................................................................415
35 22.1 Requirements.....................................................................................................................415
36 22.2 General Requirements.......................................................................................................415
37 22.2.1 Data Patterns.............................................................................................................415
22.2.2 Signal levels ..............................................................................................................416
38
22.2.3 Signal Definitions .......................................................................................................419
39 22.2.4 Bit Error Ratio ............................................................................................................420
40 22.2.5 Ground Differences ...................................................................................................420
41 22.2.6 Channel Compliance .................................................................................................420
42 22.2.6.1 Reference Model .................................................................................................420
43 22.2.6.2 Insertion Loss ......................................................................................................421
44 22.2.6.3 Fitted insertion loss..............................................................................................423
45 22.2.6.4 Insertion loss deviation (ILD) ...............................................................................423
22.2.6.5 Channel Return Loss...........................................................................................424
46 22.2.6.6 Channel integrated crosstalk noise .....................................................................425
47 22.2.6.7 Insertion Loss to Conversion Loss Ratio (ICOR).................................................426
48 22.3 Electrical Characteristics....................................................................................................427
49 22.3.1 Transmitter Characteristics........................................................................................427

26 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.3.1.1 Transmitter Baud Rate ........................................................................................428 1

22.3.1.2 Transmitter Amplitude and Swing........................................................................428 2
22.3.1.3 Transmitter Return Loss ......................................................................................428
22.3.1.4 Transmitter Quad-to-Quad Skew.........................................................................428
3
22.3.1.5 Transmitter Wire-to-Wire Skew............................................................................428 4
22.3.1.6 Transmitter Short Circuit Current.........................................................................428 5
22.3.1.7 Transmitter output waveform requirements .........................................................429 6
22.3.1.7.1 Summary of requirements.............................................................................429 7
22.3.1.7.2 Requirements for Transmitter Linearity.........................................................429 8
22.3.2 Receiver Characteristics............................................................................................430
9
22.3.2.1 Input Baud Rate...................................................................................................430
22.3.2.2 Reference Input Signals ......................................................................................431 10
22.3.2.3 Input Resistance and Return Loss ......................................................................431 11
22.3.2.4 Input Quad-to-Quad Skew ...................................................................................431 12
22.3.2.5 Absolute Input Voltage ........................................................................................431 13
22.3.2.6 Receiver Interference Tolerance Test .................................................................432 14
22.3.2.7 Receiver Imbalance Tolerance Test ....................................................................432
22.3.2.8 Receiver Skew Tolerance Test............................................................................433
15
22.A.1 PRBS9 Polynomial .....................................................................................................434 16
22.A.2 Quad Short Stress Pattern Random (SSPR-Q) ..........................................................434 17
22.A.3 Linearity Test Pattern..................................................................................................435 18
22.A.4 Clock Jitter Test Patterns............................................................................................436 19
22.A.4.1 Clock Jitter Pattern A............................................................................................436 20
22.A.4.2 Clock Jitter Pattern B............................................................................................436 21
22.A.5 Test Pattern Definitions ..............................................................................................437 22
22.B Transmitter Compliance Tests .........................................................................................440 23
22.B.1 Transmitter Output Waveform Tests...........................................................................440 24
22.B.1.1 Test Equipment Setup..........................................................................................440
25
22.B.1.2 Linear Mismatch Ratio..........................................................................................441
22.B.1.3 Waveform acquisition for Linear Fit Measurements .............................................441
26
22.B.1.4 Determining Linear Fit to Measured Waveform....................................................442 27
22.B.1.5 Transmitter Equalization Coefficients...................................................................443 28
22.B.1.6 Steady State Voltage and Linear Fit Pulse Peak .................................................445 29
22.B.1.7 Linear Fit Error .....................................................................................................445 30
22.B.1.8 Removal of the Transfer Function between the Transmitter and T ......................445 31
22.B.2 Transmitter Output Jitter .............................................................................................446 32
22.B.2.1 Test Equipment Setup..........................................................................................446 33
22.B.2.2 Deterministic and Random Clock Jitter ................................................................446 34
22.B.2.3 Even-Odd Jitter ....................................................................................................447 35
22.B.3 Transmitter Output Noise and Distortion.....................................................................448 36
22.C Receiver Compliance Tests .............................................................................................450
37
22.C.1 Receiver Interference Tolerance ................................................................................450
22.C.1.1 Test Equipment Setup..........................................................................................450
38
22.C.1.2 Test Transmitter ...................................................................................................451 39
22.C.1.3 Frequency Synthesizer and Clock Source ...........................................................451 40
22.C.1.4 Test Channel ........................................................................................................453 41
22.C.1.5 Interference Noise Source....................................................................................453 42
22.C.1.6 Test Procedure.....................................................................................................454 43
22.C.2 Receiver Imbalance Tolerance ...................................................................................454 44
22.C.2.1 Test Equipment Setup..........................................................................................454 45
22.C.2.2 Test Transmitter ...................................................................................................455 46
22.C.2.3 Bias Voltage .........................................................................................................455 47
22.C.2.4 Test Channel ........................................................................................................456 48
22.C.2.5 Interference Noise Source....................................................................................456
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 27

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.C.2.6 Test Procedure.....................................................................................................456

2 22.C.3 Receiver Skew Tolerance...........................................................................................457
3 22.C.3.1 Test Equipment Setup..........................................................................................457
4 22.C.3.2 Test Transmitter ...................................................................................................457
22.C.3.3 Test Channel ........................................................................................................457
5
22.C.3.4 Interference Noise Source ...................................................................................457
6 22.C.3.5 Test Procedure.....................................................................................................458
7 22.D Network Analysis Measurement ........................................................................................459
8 22.D.1 S Parameter Measurement with an 8-port VNA .........................................................459
9 22.D.2 S Parameter Measurement with an 4-port VNA .........................................................459
10 22.D.3 Calculation of Insertion Loss and Conversion Loss....................................................461
11 22.E Printed Circuit Board Reference Geometry .......................................................................462
12 22.E.1 Reference Geometry ..................................................................................................462
13 22.E.2 Channel Skew.............................................................................................................463
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

28 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

List of Figures 1
2
Figure 1-1 Definition of Driver Amplitude and Swing ...............................................................50 3
Figure 1-2 Skew and Relative Wander between in band Signals............................................51 4
Figure 1-3 Total Wander of a Signal........................................................................................51 5
Figure 1-4 Transmit Eye Mask.................................................................................................53 6
Figure 1-5 Receiver Input Mask...............................................................................................54 7
Figure 1-6 Reference Model ....................................................................................................55 8
Figure 1-7 Signal Diagram.......................................................................................................56 9
Figure 1-8 Some typical systems.............................................................................................58
10
Figure 2-1 CID Jitter Tolerance Pattern...................................................................................59
Figure 2-2 CID Jitter Tolerance Pattern...................................................................................67
11
Figure 2-3 Total Wander Mask ................................................................................................71 12
Figure 2-4 Relative Wander Mask ...........................................................................................71 13
Figure 2-5 Random Jitter Spectrum.........................................................................................72 14
Figure 2-6 Graphical Representation of Receiver Pulse .........................................................74 15
Figure 2-7 Transmit Pulse .......................................................................................................75 16
Figure 2-8 Transmitter FIR Filter Function...............................................................................75 17
Figure 2-9 Receiver Pulse Definition .......................................................................................76 18
Figure 2-10 Crosstalk Pulse Definition ......................................................................................77 19
Figure 2-11 Decision Feedback Equalizer.................................................................................77 20
Figure 2-12 Feed Forward Filter ................................................................................................78 21
Figure 2-13 Generation of Total and Relative Wander ..............................................................81
22
Figure 2-14 Jitter Probability Density Function ..........................................................................82
Figure 2-15 Example of Total Jitter PDF ...................................................................................84
23
Figure 2-16 Example of Total Jitter CDF ...................................................................................84 24
Figure 2-17 Bathtub Definition ...................................................................................................86 25
Figure 2-18 Example of Bounded Gaussian..............................................................................87 26
Figure 2-19 Statistics of Pulse Response Cursor ......................................................................88 27
Figure 2-20 Variation of the c0 sampling time ...........................................................................89 28
Figure 2-21 Varying the Receiver Sampling Point.....................................................................91 29
Figure 2-22 Generation of the Data Eye and Bathtub ...............................................................93 30
Figure 2-23 Statistical Eye.........................................................................................................94 31
Figure 2-24 Short Stress Pattern Random (SSPR) ...................................................................95 32
Figure 2-25 Short Stress Pattern SDH 16 (SSPS-16) ...............................................................96 33
Figure 2-26 Short Stress Pattern SDH 64 (SSPS-64) ...............................................................97
34
Figure 2-27 BERT with Golden PLL ........................................................................................106
Figure 2-28 Spectral Measurement Setup...............................................................................108
35
Figure 2-29 Single Side Band Relative Power Spectrum for Phase Modulated Signal ...........109 36
Figure 2-30 Transmit Wander Lab Setup ................................................................................110 37
Figure 2-31 Relative Wander Lab Setup .................................................................................111 38
Figure 2-32 Jitter Tolerance with Relative Wander Lab Setup ................................................112 39
Figure 2-33 Jitter Tolerance with no Relative Wander.............................................................114 40
Figure 2-34 Jitter Tolerance with Defined ISI ..........................................................................115 41
Figure 2-35 Jitter Transfer Lab Setup......................................................................................116 42
Figure 2-36 S-parameter Port definitions.................................................................................117 43
Figure 2-37 Eye Mask Measurement with Golden PLL ...........................................................119 44
Figure 2-38 Example Data Mask .............................................................................................122 45
Figure 2-39 Example Data Mask .............................................................................................123
46
Figure 2-40 Cumulative Distribution Function of Maximum Amplitude ....................................124
Figure 3-1 Driver and Input Differential Return Loss .............................................................128
47
Figure 3-2 Termination Example ...........................................................................................129 48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 29

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 3-3 Transmission Line as 2-port .................................................................................131

2 Figure 3-4 PI Network Termination........................................................................................132
3 Figure 3-5 Measurement of Zodd, Zeven ..............................................................................132
4 Figure 3-6 Minimisation of crosstalk at IC pins. .....................................................................135
Figure 3-7 Minimisation of crosstalk at connector pins..........................................................136
5
Figure 3-8 Minimisation of crosstalk over backplane.............................................................136
6 Figure 3-9 Equation based Channel Loss curves..................................................................137
7 Figure 8-1 Jitter Ingress Receiver Input Telecom Sinusoidal Jitter .......................................187
8 Figure 8-2 Jitter Egress Receiver Input Telecom Sinusoidal Jitter ........................................188
9 Figure 8-3 Jitter Ingress Receiver Input Datacom Sinusoidal Jitter.......................................190
10 Figure 8-4 Reference Model ..................................................................................................195
11 Figure 10-1 CEI-28G-SR Reference Model.............................................................................218
12 Figure 10-2 CEI-28G-SR Normative Channel Insertion Loss at 28.05 Gsym/s.......................220
13 Figure 10-3 Illustration integrated crosstalk noise limits ..........................................................223
14 Figure 11-1 CEI-25G-LR Reference Model .............................................................................234
15 Figure 11-2 CEI-25G-LR Normative Channel Insertion Loss at 25.80 Gsym/s. ......................236
16 Figure 11-3 Illustration integrated crosstalk noise limits ..........................................................239
Figure 12-1 DDJ Measurement Method ..................................................................................250
17
Figure 13-1 Measurement points using compliance boards ....................................................259
18 Figure 13-2 Host Differential Termination Resistance Mismatch measurement setup............264
19 Figure 13-3 Module Differential Termination Resistance Mismatch measurement setup .......264
20 Figure 13-4 SDD11, SDD22 for host output (TP1a), host input (TP4a), module input (TP1) 
21 and module output (TP4) (for fb = 28 GHz)..........................................................265
22 Figure 13-5 SDC11 and SCD11 for module input (TP1) and host input (TP4a) (for fb = 28 
23 GHz) .....................................................................................................................266
24 Figure 13-6 SDC22 and SCD22 for module output (TP4) and host output (TP1a) (for fb = 28 
25 GHz) .....................................................................................................................267
26 Figure 13-7 Host output test setup ..........................................................................................269
27 Figure 13-8 Module output test setup ......................................................................................269
28 Figure 13-9 TP1a and TP4 jitter and Eye Height parameters..................................................272
Figure 13-10 Host input test setup ............................................................................................273
29
Figure 13-11 Module stressed input test setup .........................................................................274
30 Figure 13-12 Host input and Module input Sinusoidal Jitter ......................................................275
31 Figure 13-13 Host output Reference receiver equalizer (CTLE) transfer function for gains of 
32 1 dB to 9 dB .........................................................................................................277
33 Figure 13-14 Module output Reference receiver equalizer (CTLE) transfer function for gains 
34 of 1 and 2 dB ........................................................................................................278
35 Figure 13-15 Reference SDD21 of HCB and MCB printed circuit board traces ........................280
36 Figure 13-16 Mated HCB-MCB SDD11, SDD22 .......................................................................281
37 Figure 13-17 Mated HCB-MCB SCD21, SCD12 .......................................................................282
38 Figure 13-18 Mated HCB-MCB SCD11, SCD22, SDC11, SDC22 ............................................283
39 Figure 13-19 Mated HCB-MCB SDD21, SDD12 .......................................................................284
40 Figure 13-20 CEI-28G-VSR full Channel Reference Model ......................................................285
Figure 13-21 Recommended minimum SDD21 of the VSR channel (for fb = 28 GHz).............286
41
Figure 14-1 CEI-28G-MR Reference Model ............................................................................290
42 Figure 14-2 CEI-28G-MR normative channel insertion loss at 28.1 Gsym/s ...........................292
43 Figure 14-3 CEI-28G-MR normative channel return loss at 28.1 Gsym/s ...............................295
44 Figure 14-4 Illustration of integrated crosstalk noise limits ......................................................296
45 Figure 16-1 Measurement points using compliance boards. ...................................................311
46 Figure 16-2 SDD11, SDD22 for host output (TP1a), host input (TP4a), module input (TP1) 
47 and module output (TP4) (for fb = 29 GHz)..........................................................316
48 Figure 16-3 SCD11 for module input (TP1) and host input (TP4a), and SDC22 for module 
49 output (TP4) and host output (TP1a) (for fb=29 GHz)..........................................317

30 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-4 Host output test setup ..........................................................................................319 1

Figure 16-5 Module output test setup ......................................................................................319 2
Figure 16-6 TP1a and TP4 Eye Width and Eye Height parameters ........................................321 3
Figure 16-7 TP1a and TP4 Eye Width, Eye Height and Eye Amplitude ..................................323 4
Figure 16-8 PAM4 Horizontal Eye Mask..................................................................................325
5
Figure 16-9 Host input test setup.............................................................................................326
Figure 16-10 Module input test setup ........................................................................................327
6
Figure 16-11 Host input and Module input Sinusoidal Jitter ......................................................328 7
Figure 16-12 Host output reference equalizer (CTLE) transfer function for 1 dB to 9 dB of  8
peaking at 25 Gsym/s...........................................................................................330 9
Figure 16-13 Module near-end output reference equalizer (CTLE) transfer function for 1 dB  10
to 2 dB of peaking at 25 Gsym/s ..........................................................................331 11
Figure 16-14 TP4 Far-end Host Channel SDD21 for fb = 29 GHz ............................................332 12
Figure 16-15 CEI-56G-VSR-PAM4 Channel Reference Model .................................................334 13
Figure 16-16 Recommended minimum SDD21 of the end-to-end channel (for fb = 29 GHz) ...335 14
Figure 16-17 Transmit signaling and mapping diagram ............................................................337 15
Figure 16-18 Receive signaling and mapping diagram .............................................................338 16
Figure 16-19 QPRBS13-CEI pattern generator .........................................................................339
17
Figure 16-20 Input to Output BER of common FEC codes........................................................342
Figure 17-1 CEI-56G-MR Reference Model ............................................................................346
18
Figure 17-2 Channel Insertion Loss Limit for 29 Gsym/s.........................................................348 19
Figure 17-3 Channel Return Loss Limit for 29 Gsym/s............................................................349 20
Figure 17-4 Transmitter differential return loss limit for 29 Gsym/s .........................................353 21
Figure 17-5 Transmitter common mode return loss limit for 29 Gsym/s. .................................354 22
Figure 17-6 Transmit equalizer functional model.....................................................................355 23
Figure 17-7 Receiver differential to common-mode return loss limit for 29 Gsym/s. ...............361 24
Figure 17-8 Receiver Jitter Tolerance Mask............................................................................363 25
Figure 18-1 CEI-56G-USR-NRZ Reference Model..................................................................366 26
Figure 18-2 CEI-56G-USR-NRZ normative channel insertion loss at 58.0 Gsym/s ................368 27
Figure 18-3 CEI-56G-USR-NRZ normative channel return loss at 58.0 Gsym/s.....................370 28
Figure 18-4 Transmitter Amplitude Test Load .........................................................................373
29
Figure 18-5 Illustration of Return Loss for T_Baud = 58.0 Gsym/s..........................................374
Figure 18-6 Forwarded Clock Architecture ..............................................................................377
30
Figure 18-7 Common Clock Architecture.................................................................................378 31
Figure 18-8 Transmitter Measurements Using a Golden Clock Multiplier ...............................379 32
Figure 19-1 CEI-56G-XSR-NRZ Reference Model..................................................................382 33
Figure 19-2 CEI-56G-XSR-NRZ normative channel insertion loss at 58.0 Gsym/s.................384 34
Figure 19-3 CEI-56G-XSR-NRZ normative channel return loss at 58.0 Gsym/s.....................386 35
Figure 19-4 Transmitter Amplitude Test Load .........................................................................389 36
Figure 19-5 Illustration of Return Loss for T_Baud = 58.0 Gsym/s..........................................390 37
Figure 21-1 CEI-56G-LR Reference Model .............................................................................396 38
Figure 21-2 Channel Insertion Loss Limit for 29Gsym/s..........................................................398 39
Figure 21-3 Channel Return Loss Limit for 29 Gsym/s............................................................399 40
Figure 21-4 Transmitter differential return loss limit for 29 Gsym/s .........................................403
41
Figure 21-5 Transmitter common mode return loss limit for 29 Gsym/s. .................................404
Figure 21-6 Transmit equalizer functional model.....................................................................405
42
Figure 21-7 Receiver differential to common-mode return loss limit for 29 Gsym/s. ...............411 43
Figure 21-8 Receiver Jitter Tolerance Mask............................................................................413 44
Figure 22-1 Definition of Driver Amplitude and Swing .............................................................416 45
Figure 22-2 Definition of Driver Amplitude and Swing .............................................................416 46
Figure 22-3 Signal Diagram.....................................................................................................419 47
Figure 22-4 CEI-56G-LR-ENRZ Reference Model ..................................................................420 48
Figure 22-5 CEI-56G-LR normative channel insertion loss at 37.5 Gsym/s ............................422 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 31

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 22-6 CEI-56G-LR normative channel return loss at 37.5 Gsym/s ................................424
2 Figure 22-7 Illustration of integrated crosstalk noise limits ......................................................425
3 Figure 22-8 PRBS9 Pattern Generation ..................................................................................434
4 Figure 22-9 Quad Short Stress Pattern Random (SSPR-Q) ...................................................435
Figure 22-10 Linearity Test Pattern Waveform..........................................................................435
5
Figure 22-11 Driver Test Load...................................................................................................440
6 Figure 22-12 Receiver Interference Tolerance Test Setup........................................................450
7 Figure 22-13 Configuration for Test Transmitter Calibration .....................................................451
8 Figure 22-14 Configuration for Pattern Generator Calibration...................................................452
9 Figure 22-15 ENRZ to Differential Test Conversion Network ....................................................452
10 Figure 22-16 Configuration for Interference Noise Source Calibration......................................454
11 Figure 22-17 Receiver Imbalance Tolerance Test Setup ..........................................................455
12 Figure 22-18 Configuration for Bias Voltage Calibration ...........................................................455
13 Figure 22-19 Bias Network ........................................................................................................456
14 Figure 22-20 Receiver Skew Tolerance Test Setup ..................................................................457
15 Figure 22-21 S-Parameter Measurement Using an 8-port VNA ................................................459
16 Figure 22-22 S-Parameter Measurement Using a 4-port VNA ..................................................460
Figure 22-23 Reference PCB Routing Geometry for ENRZ Channel........................................462
17
Figure 22-24 Length Matching Example Through a Right Angle Turn ......................................463
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

32 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

List of Tables 1
2
Table 1-1 Abbreviations...........................................................................................................43 3
Table 1-2 General Definitions (with exception of Jitter and Wander) ......................................45 4
Table 1-3 Jitter and Wander Definitions ..................................................................................47 5
Table 1-4 Transmitter Electrical Output Specification..............................................................52 6
Table 1-5 Transmitter Output Jitter Specification ...................................................................52 7
Table 1-6 Receiver Electrical Input Specification ....................................................................53 8
Table 1-7 Receiver Input Jitter Specification ..........................................................................54 9
Table 1-8 Receive Signal Summary ........................................................................................57
10
Table 1-9 Transmit Signal Summary .......................................................................................57
Table 1-10 Example specification of reference clock ................................................................57
11
Table 2-1 Use of CEI Test Patterns.........................................................................................98 12
Table 3-1 Definition of load types ..........................................................................................126 13
Table 3-2 Curve fit Coefficients .............................................................................................137 14
Table 6-1 CEI-6G-SR Transmitter Output Electrical Specifications.......................................153 15
Table 6-2 CEI-6G-SR Transmitter Output Jitter Specifications .............................................154 16
Table 6-3 CEI-6G-SR Driver Return Loss Parameters..........................................................155 17
Table 6-4 CEI-6G-SR Near-End (Tx) Template Intervals ......................................................156 18
Table 6-5 CEI-6G-SR Receiver Electrical Input Specifications .............................................156 19
Table 6-6 CEI-6G-SR Receiver Input Jitter Tolerance Specifications ...................................157 20
Table 6-7 CEI-6G-SR Input Return Loss Parameters ...........................................................159 21
Table 6-8 CEI-6G-SR Far-End (Rx) Template Intervals ........................................................159
22
Table 6-9 CEI-6G-SR Informative Loss, Skew and Jitter Budget ..........................................160
Table 6-10 CEI-6G-SR High Frequency Jitter Budget.............................................................160
23
Table 7-1 CEI-6G-LR Receiver Equalization Output Eye Mask ............................................167 24
Table 7-2 CEI-6G-LR Transmitter Output Electrical Specifications .......................................168 25
Table 7-3 CEI-6G-LR Transmitter Output Jitter Specifications..............................................168 26
Table 7-4 CEI-6G-LR Driver Return Loss Parameters ..........................................................169 27
Table 7-5 CEI-6G-LR Near-End Template Intervals..............................................................170 28
Table 7-6 CEI-6G-LR Training Pattern ..................................................................................171 29
Table 7-7 CEI-6G-LR Receiver Electrical Input Specifications..............................................171 30
Table 7-8 CEI-6G-LR Input Return Loss Parameters............................................................173 31
Table 7-9 CEI-6G-LR Informative Loss, Skew and Jitter Budget ..........................................174 32
Table 7-10 CEI-6G-LR High Frequency Jitter Budget .............................................................174 33
Table 8-1 Transmitter Electrical Output Specification............................................................182
34
Table 8-2 Transmitter Output Jitter Specification ..................................................................182
Table 8-3 Driver Return Loss Parameters .............................................................................183
35
Table 8-4 Receiver Electrical Input Specification ..................................................................184 36
Table 8-5 Receiver Input Jitter Specification ........................................................................184 37
Table 8-6 Driver Return Loss Parameters .............................................................................185 38
Table 8-7 Telecom Signal Conditioner, Egress direction.......................................................189 39
Table 8-8 Telecom Signal Conditioner, Ingress Direction .....................................................189 40
Table 8-9 Telecom Egress Jitter Generation budget .............................................................189 41
Table 8-10 Telecom Egress Driver Jitter Generation ..............................................................190 42
Table 8-11 Datacom Signal Conditioner Egress direction .......................................................191 43
Table 8-12 Datacom Signal Conditioner Ingress Direction......................................................191 44
Table 8-13 Datacom Signal Conditioner Ingress Direction......................................................191 45
Table 8-14 Informative Jitter Budget .......................................................................................192
46
Table 9-1 CEI-11G-LR Receiver Equalization Output Eye Mask ..........................................200
Table 9-2 Transmitter Output Electrical Specifications..........................................................201
47
Table 9-3 Transmitter Output Jitter Specifications.................................................................201 48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 33

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 9-4 Driver Return Loss Parameters.............................................................................202

2 Table 9-5 CEI-11G-LR Receiver Electrical Specifications.....................................................203
3 Table 9-6 CEI-11G-LR Receiver Input Jitter Specification ...................................................203
4 Table 9-7 Driver Return Loss Parameters.............................................................................204
Table 9-8 CEI-11G-MR Receiver Electrical Specifications....................................................205
5
Table 9-9 CEI-11G-MR Receiver Input Jitter Specification ..................................................205
6 Table 9-10 CEI-11G-LR Informative Jitter Budget...................................................................207
7 Table 9-11 CEI-11G-MR Informative Jitter Budget..................................................................208
8 Table 10-1 Measured Channel Parameters ............................................................................219
9 Table 10-2 Calculated Channel Parameters ...........................................................................219
10 Table 10-3 Channel Insertion Loss Frequency Range ............................................................219
11 Table 10-4 Channel fitted insertion loss characteristics ..........................................................221
12 Table 10-5 Channel integrated crosstalk aggressor parameters.............................................222
13 Table 10-6 Transmitter Electrical Output Specification. ..........................................................224
14 Table 10-7 Transmitter Output Jitter Specification ..................................................................225
15 Table 10-8 Transmitter Differential Return Loss Parameters ..................................................225
16 Table 10-9 Coefficient range and step size .............................................................................226
Table 10-10 Linear fit pulse and equalizing filter parameters....................................................227
17
Table 10-11 Transmitter output waveform requirements...........................................................227
18 Table 10-12 Receiver Electrical Input Specifications ................................................................231
19 Table 10-13 Receiver Input Jitter Specification ........................................................................231
20 Table 10-14 Receiver Differential Return Loss Parameters ......................................................232
21 Table 11-1 Measured Channel Parameters ............................................................................235
22 Table 11-2 Calculated Channel Parameters ...........................................................................235
23 Table 11-3 Channel Insertion Loss Frequency Range ............................................................235
24 Table 11-4 Channel fitted insertion loss characteristics ..........................................................237
25 Table 11-5 Channel integrated crosstalk aggressor parameters.............................................238
26 Table 11-6 Transmitter Electrical Output Specification. ..........................................................240
27 Table 11-7 Transmitter Output Jitter Specification ..................................................................241
28 Table 11-8 Transmitter Differential Return Loss Parameters ..................................................241
Table 11-9 Coefficient range and step size .............................................................................242
29
Table 11-10 Linear fit pulse and equalizing filter parameters....................................................243
30 Table 11-11 Transmitter output waveform requirements...........................................................243
31 Table 11-12 Receiver Electrical Input Specifications ................................................................247
32 Table 11-13 Receiver Input Jitter Specification ........................................................................247
33 Table 11-14 Receiver Differential Return Loss Parameters ......................................................248
34 Table 13-1 Host-to-Module Electrical Specifications at TP1a (host output) ............................260
35 Table 13-2 Host-to-Module Electrical Specifications (module input) .......................................261
36 Table 13-3 Crosstalk parameters for host output test and module stressed input test 
37 calibration at TP4 ..................................................................................................261
38 Table 13-4 Module-to-Host Electrical Specifications at TP4 (module output) .........................262
39 Table 13-5 Module-to-Host Electrical Specifications (host input) ............................................262
40 Table 13-6 Crosstalk parameters for module output test and host stressed input test 
calibration at TP1a ................................................................................................263
41
Table 13-7 Sinusoidal jitter frequency for TP4 and TP1 testing ..............................................275
42 Table 13-8 Reference equalizer coefficients for rate of 28 GBd..............................................278
43 Table 13-9 Host-to-Module Electrical Specifications at TP0a .................................................287
44 Table 14-1 Measured Channel Parameters ............................................................................291
45 Table 14-2 Calculated Channel Parameters ...........................................................................291
46 Table 14-3 Channel Insertion Loss Frequency Range ............................................................291
47 Table 14-4 Channel fitted insertion loss characteristics ..........................................................293
48 Table 14-5 Channel integrated crosstalk aggressor parameters.............................................295
49 Table 14-6 Transmitter Electrical Output Specification. ..........................................................297

34 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 14-7 Transmitter Output Jitter Specification ..................................................................298 1

Table 14-8 Transmitter Differential Return Loss Parameters ..................................................298 2
Table 14-9 Coefficient range and step size .............................................................................299 3
Table 14-10 Linear fit pulse and equalizing filter parameters ....................................................300 4
Table 14-11 Transmitter output waveform requirements...........................................................300
5
Table 14-12 Receiver Electrical Input Specifications.................................................................304
Table 14-13 Receiver Input Jitter Specification ........................................................................304
6
Table 14-14 Receiver Differential Return Loss Parameters ......................................................305 7
Table 16-1 Host-to-Module Electrical Specifications at TP1a (host output) (See Note 5) .......312 8
Table 16-2 Host-to-Module Electrical Specifications (module input) .......................................313 9
Table 16-3 Crosstalk parameters for host output test and module stressed input test  10
calibration at TP4 ..................................................................................................313 11
Table 16-4 Module-to-Host Electrical Specifications at TP4 (module output) .........................314 12
Table 16-6 Crosstalk parameters for module output test and host stressed input test  13
calibration at TP1a ................................................................................................315 14
Table 16-5 Module-to-Host Electrical Specifications (host input) ............................................315 15
Table 16-7 Transmission Line Model Parameters and Values ................................................322 16
Table 16-8 Sinusoidal jitter frequency for TP4 and TP1a testing ............................................328
17
Table 16-9 Reference equalizer coefficients for rates of 25 GBd and above ..........................331
Table 16-10 Host-to-Module Electrical Recommendations at TP0a..........................................336
18
Table 16-11 Common FEC Codes ............................................................................................343 19
Table 17-1 COM Parameter Values ........................................................................................347 20
Table 17-2 Transmitter Electrical Output Specification............................................................351 21
Table 17-3 Transmitter Output Jitter Specification .................................................................351 22
Table 17-4 Coefficient Range and Step Size ..........................................................................356 23
Table 17-5 QPRBS13-CEI Pattern Symbols Used for Jitter Measurement .............................358 24
Table 17-6 Receiver Electrical Input Specification ..................................................................359 25
Table 17-7 Receiver interference tolerance parameters (Note 3) ...........................................360 26
Table 17-8 Receiver Jitter Tolerance Parameters ...................................................................362 27
Table 18-1 Measured Channel Parameters ............................................................................367 28
Table 18-2 Calculated Channel Parameters ...........................................................................367
29
Table 18-3 Channel Insertion Loss Frequency Range ............................................................367
Table 18-4 Channel fitted insertion loss characteristics ..........................................................368
30
Table 18-5 Channel integrated crosstalk aggressor parameters.............................................370 31
Table 18-6 Reference Clock Electrical Specification. ..............................................................371 32
Table 18-7 Transmitter Electrical Output Specification............................................................372 33
Table 18-8 Transmitter Output Jitter Specification ..................................................................372 34
Table 18-9 Driver Return Loss Parameters .............................................................................373 35
Table 18-10 Receiver Electrical Input Specification ..................................................................375 36
Table 18-11 Receiver Input Return Loss Parameters ...............................................................376 37
Table 19-1 Measured Channel Parameters ............................................................................383 38
Table 19-2 Calculated Channel Parameters ...........................................................................383 39
Table 19-3 Channel Insertion Loss Frequency Range ............................................................383 40
Table 19-4 Channel fitted insertion loss characteristics ..........................................................384
41
Table 19-5 Channel integrated crosstalk aggressor parameters.............................................386
Table 19-6 Reference Clock Electrical Specification. ..............................................................387
42
Table 19-7 Transmitter Electrical Output Specification............................................................388 43
Table 19-8 Transmitter Output Jitter Specification ..................................................................388 44
Table 19-9 Driver Return Loss Parameters .............................................................................389 45
Table 19-10 Receiver Electrical Input Specification ..................................................................391 46
Table 19-11 Receiver Input Return Loss Parameters ...............................................................392 47
Table 21-1 COM Parameter Values ........................................................................................397 48
Table 21-2 Transmitter Electrical Output Specification............................................................401 49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 35

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 21-3 Transmitter Output Jitter Specification .................................................................401

2 Table 21-4 Coefficient Range and Step Size ..........................................................................406
3 Table 21-5 QPRBS13-CEI Pattern Symbols Used for Jitter Measurement.............................408
4 Table 21-6 Receiver Electrical Input Specification ..................................................................409
Table 21-7 Receiver interference tolerance parameters (Note 3) ...........................................410
5
Table 21-8 Receiver Jitter Tolerance Parameters...................................................................412
6 Table 22-1 ENRZ Signal Drive States .....................................................................................417
7 Table 22-2 ENRZ Signaling Code Map ...................................................................................418
8 Table 22-3 ENRZ Linear Combination Table ..........................................................................418
9 Table 22-4 Receive Signal Summary ......................................................................................419
10 Table 22-5 Transmit Signal Summary .....................................................................................419
11 Table 22-6 Measured Channel Parameters ............................................................................421
12 Table 22-7 Calculated Channel Parameters ...........................................................................421
13 Table 22-8 Channel Insertion Loss Frequency Range ............................................................421
14 Table 22-9 Channel fitted insertion loss characteristics ..........................................................423
15 Table 22-10 Channel integrated crosstalk aggressor parameters.............................................425
16 Table 22-11 Transmitter Electrical Output Specification. ..........................................................427
Table 22-12 Transmitter Output Jitter Specification ..................................................................427
17
Table 22-13 Transmitter Single-Ended Return Loss Parameters..............................................428
18 Table 22-14 Coefficient range and step size .............................................................................429
19 Table 22-15 Transmitter output waveform requirements...........................................................429
20 Table 22-16 Receiver Electrical Input Specifications ................................................................430
21 Table 22-17 Receiver Input Jitter Specification ........................................................................430
22 Table 22-18 Receiver Single-Ended Return Loss Parameters..................................................431
23 Table 22-19 Receiver Interference Tolerance Test Parameters ...............................................432
24 Table 22-20 Receiver Imbalance Tolerance Test Parameters ..................................................432
25 Table 22-21 Receiver Skew Tolerance Test Parameters..........................................................433
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

36 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

0.3 List of companies belonging to the OIF when the document 1

is approved 2
3
Acacia Communications ADVA Optical Networking 4
Alibaba (China) Co. Ltd. Amphenol Corp. 5
Anritsu Arista Networks 6
Barefoot Networks Broadcom Limited 7
BRPhotonics Cavium 8
China Telecom Ciena Corporation 9
Cisco Systems Coriant 10
Corning Credo Semiconductor (HK) LTD 11
Dell, Inc. Elenion Technologies, LLC 12
Epson Electronics America, Inc. Fiberhome Technologies Group 13
Finisar Corporation Foxconn Interconnect Technology, Ltd. 14
Fujikura Fujitsu 15
Furukawa Electric Japan Gigamon Inc. 16
Global Foundries Google 17
Hewlett Packard Enterprise (HPE) Hitachi 18
Huawei Technologies Co., Ltd. Infinera 19
Inphi Integrated Device Technology 20
Intel Invecas, Inc. 21
IPG Photonics Corporation Ixia 22
JCRFO Juniper Networks 23
Kandou Bus KDDI Research, Inc. 24
Keysight Technologies, Inc. Lumentum 25
MACOM Technology Solutions Maxim Integrated Inc. 26
MaxLinear Inc. Mellanox Technologies 27
Microsemi Inc. Microsoft Corporation 28
Mitsubishi Electric Corporation Molex 29
NEC Corporation NeoPhotonics 30
Nokia NTT Corporation 31
O-Net Communications (HK) Limited Oclaro 32
Orange PETRA 33
Precise-ITC, Inc. Qorvo 34
Ranovus Rianta Solutions, Inc. 35
Rockley Photonics Samsung Electronics Co. Ltd. 36
Samtec Inc. Semtech 37
SiFotonics Technologies Co., Ltd. Silab Tech Private Ltd. 38
Sino-Telecom Technology Co., Inc. Socionext Inc. 39
Spirent Communications Sumitomo Electric Industries 40
Sumitomo Osaka Cement TE Connectivity 41
Tektronix Teledyne LeCroy 42
TELUS Communications, Inc. UNH InterOperability Laboratory (UNH-IOL) 43
Verizon Viavi Solutions Deutschland GmbH 44
Xilinx Yamaichi Electronics Ltd. 45
ZTE Corporation 46
47
48
49

Optical Internetworking Forum - Clause 0 : Document Structure and Contents 37

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

38 Clause 0 : Document Structure and Contents - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Common electrical I/O project - Introduction, 1

2
definitions and formats. 3
4
5
1.1 Introduction 6
7
The development of a Next Generation Common Electrical I/O Project was proposed in 8
the OIF 2002.571.01 and approved in the Orlando Plenary meeting November 14, 9
2002. The purpose of the project is outlined in the problem statement: 10
11
A faster electrical interface is required to provide higher density and/or lower cost 12
interfaces for payloads of 10Gbps and higher, including SERDES to Framer Interface 13
(SFI), System Packet Interface (SPI), TDM-Fabric to framer Interface (TFI). 14
15
16
1.2 Overview 17
18
This Common Electrical IO Implementation Agreement includes: 19
• Electrical and jitter methodologies for new high speed interfaces and including the 20
following older OIF interfaces: SxI-5, SFI-4.2, SFI-5.1, SPI-5.1 and TFI-5. 21
22
• A CEI-6G-SR specification for: 23
Data lane(s) that support bit rates from 4.976 to 6.375Gsym/s over PCB. 24
Physical reach from 0 to 200mm and up to 1 connector. 25
26
• A CEI-6G-LR specification for 27
Data lane(s) that support bit rates from 4.976 to 6.375Gsym/s over PCB. 28
Physical reach from 0 to 1m and up to 2 connectors. 29
30
• A CEI-11G-SR specification for: 31
Data lane(s) that support bit rates from 9.95 to 11.2Gsym/s over PCB. 32
33
• A CEI-11G-LR specification for: 34
Data lane(s) that support bit rates from 9.95 to 11.2Gsym/s over PCB. 35
Physical reach from 0 to 1m with up to two connectors 36
37
• A CEI-11G-MR specification for: 38
Data lane(s) that support bit rates from 9.95 to 11.2Gsym/s over PCB. 39
Physical reach from 0 to 600mm with up to two connectors 40
41
• A CEI-28G-SR specification for: 42
Data lane(s) that support bit rates from 19.90 to 28.05Gsym/s over PCB. 43
Physical reach from 0 to 300mm with up to one connector 44
45
• A CEI-25G-LR specification for:
46
Data lane(s) that support bit rates from 19.90 to 25.80Gsym/s over PCB. 47
Physical reach from 0 to 686mm with up to two connectors 48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 39

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • A CEI-28G-VSR specification for:

2 Data lane(s) that support bit rates from 19.60 to 28.10Gsym/s over PCB.
3 Physical reach from a minimum of 100 mm of host PCB trace plus one connector
4 and a minimum of 50 mm of module PCB trace
5
6 • A CEI-28G-MR specification for:
7 Data lane(s) that support bit rates from 19.90 to 28.10Gsym/s over PCB.
8 Physical reach from 0 to 500 mm with up to one connector
9
10 • A CEI-56G-VSR-PAM4 specification for:
11 Data lane(s) that support bit rates from 18.0 to 29.0 Gsym/s over PCB.
12 Physical reach from a minimum of 125 mm of host PCB trace plus one connector
13 and a minimum of 25 mm of module PCB trace
14
15 • A CEI-56G-MR-PAM4 specification for:
16 Data lane(s) that support bit rates from 18.0 to 29.0 Gsym/s over PCB.
17 Physical reach from 0 to 500 mm with up to one connector
18
19 • A CEI-56G-USR-NRZ specification for:
20 Data lane(s) that support bit rates from 19.6 to 58.0 Gsym/s inside a package.
21 Physical reach from 0 to 10 mm on package substrate
22
23 • A CEI-56G-XSR-NRZ specification for:
24 Data lane(s) that support bit rates from 39.8 to 58.0 Gsym/s over PCB.
25 Physical reach from 0 to 50 mm of PCB trace
26
27 • A CEI-56G-LR-PAM4 specification for:
28 Data lane(s) that support bit rates from 18.0 to 29.0 Gsym/s over PCB.
29 Physical reach from 0 to 1000 mm of PCB and up to 2 connectors
30
31 • A CEI-56G-LR-ENRZ specification for:
32 Data lane(s) that support bit rates from 99.5 Gbit/s and 112.4 Gbit/s across four
33 wires over PCB.Corresponding baud rates are within the range of 33.16 Gbaud/s
34 and 37.50 Gbaud/s.
35 Physical reach from 0 to 1000 mm of PCB and up to 2 connectors
36
37
38 The Implementation Agreement defines applicable data characteristics (e.g. DC
39 balance, transition density, maximum run length), channel models and compliance
40 points/parameters supporting the physical reach and conditions. The
41 Implementation Agreement specifically excludes any pinout, management interface,
42 power-supply specification, connector or higher-level activity such as addressing or
43 error control. It does not endorse or specify any particular data protocol.
44
45
46
47
48
49

40 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1.3 Objectives and Requirements 1

2
The objectives and requirements for the CEI are given by the project definition as 3
follows: 4
5
The data path shall: 6
7
• allow single and multi-lane applications 8
• support AC coupling 9
10
• support Hot Plug 11
• achieve Bit Error Ratio of lower than 10-15 per lane but the test requirement will be 12
to verify 10-12 per lane. 13
14
The CEI Electrical Implementation Agreement and the CEI Protocol Implementation 15
Agreement are peer documents. Adherence to one does not force adherence to the 16
other. For example, a 10G SONET framer may connect directly to an optical module 17
using CEI electricals with SONET scrambled data. In this case, CEI Protocol would be 18
absent. It is also possible to use CEI Protocol without CEI Electricals. An example 19
would be to encapsulate TFI-5 frames with CEI Protocol to provide forward error 20
correction capability. 21
22
23
1.4 References 24
25
1. XFP Rev 3.1 (10 gigabit Small form factor Pluggable Module) April 25th 2003. 26
27
2. ITU Recommendation O.172 (03/01) Jitter and wander measuring equipment for 28
digital systems which are based on the synchronous digital hierarchy (SDH). 29
3. ITU G.825 (03/00) The control of jitter and wander within digital networks which are 30
based on the synchronous digital hierarchy (SDH). G.825 Erratum 1 (08/01) 31
Erratum to Recommendation ITU-T G.825 (03/00). 32
33
4. Optical Internetworking Forum "OIF-TFI5-0.1.0 TFI-5 TDM Fabric to Framer 34
Interface Implementation Agreement", September 16 2003. 35
36
5. Telecordia, GR-253-CORE, Issue 3 Sept. 2000 - "Synchronous Optical Network
37
(SONET) Transport System: Common Generic Criteria"
38
6. ITU-T, Recommendation G.707, Oct. 2000 - "Network Node Interface For The 39
Synchronous Digital Hierarchy (SDH)" 40
41
7. ITU-T, Recommendation G.707, Amendment 2, 2002 - "Network Node Interface For 42
The Synchronous Digital Hierarchy (SDH), Amendment 2" 43
8. ITU-T, Recommendation G.709, Feb. 2001 - "Network Node Interface for the 44
Optical Transport Network (OTN)" 45
46
9. Fiber Channel - Methodology for Jitter and Signal Quality Specification - MJSQ, 47
INCITs T11.2 project 1316-DT 48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 41

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 10. Optical Internetworking Forum, OIF-SFI4-02.0 - Serdes Framer Interface Level 4

2 (SFI4) Phase 2: Implementation Agreement for 10Gb/s Interface for Physical layer
3 devices.
4
11. Optical Internetworking Forum, OIF-SFI5-01.0 - Serdes Framer Interface Level 5
5
(SFI5): 40Gb/s Interface for Physical Layer devices.
6
7 12. Optical Internetworking Forum, OIF-SPI5-01.0 - System Packet Interface Level 5
8 (SPI5): OC-768 System Interface for Physical Layer devices.
9
10 13. Optical Internetworking Forum, OIF-SxI5-01.0 - System Interface Level 5 (SxI5):
11 Common Electrical Characteristics for 2.488 - 3.125Gb/s Parallel Interfaces.
12 14. Infiniband Architecture Release 1.0.a, Volume 2 — Physical Specifications,
13 Infiniband Trade Association, 2001
14
15 15. High Speed Digital Interconnection, Thomas J. Buck, Dynamic Details Inc.
16 16. Even Mode Impedance, An Introduction, App Note 157, Polar Instruments
17
18 17. Eric Bogatin, ‘Differential Impedance... finally made simple, Bogatin Enterprises,
19 2000
20 18. R.J.Weber, Introduction to Microwave Circuits, IEEE Press, 2001
21
22 19. IEEE, 802.3TM-2008, "Information Technology - Local & Metropolitan Area
23 Networks - Part 3: Carrier Sense Multiple Access with Collision Detection (CSMA/
24 CD) Access Method and Physical Layer Specifications"
25 20. Low Voltage Differential Swing (LVDS), ANSI/TIA/EIA-644-A-2001
26
27 21. ITU-T Recommendation O.150 May 1996 and corrigendum May 2002. General
28 requirements for instrumentation for performance measurements on digital
29 transmission equipment.
30 22. ITU-T Recommendation G.957 July 1999 and amendment Dec 2003. Optical
31 interfaces for equipments and systems relating to the synchronous digital hierarchy.
32
33 23. Fiber Channel - Physical Interfaces, INCITs T11.2 project 1235D
34
24. Optical Internetworking Forum, OIF 2002.507.01 - High Speed Backplane (HSB)
35
Interface Electrical Specification for 5-6.375Gbps Baud Rates over Currently
36
Existing Communications Backplanes.
37
38 25. IEEE Std 802.3™-2015, IEEE Standard for Ethernet.
39
40 26. IEEE Std, 802.3by™-2016, IEEE Standard for Ethernet, Amendment 2: Media
41 Access Control Parameters, Physical Layers, and Management Parameters for 25
42 Gb/s Operation.
43 27. Draft IEEE Std, 802.3bs™-201x, IEEE Standard for Ethernet, Amendment 2: Media
44 Access Control Parameters, Physical Layers and Management Parameters for 200
45 Gb/s and 400 Gb/s Operation.
46
47 28. ITU-T Recommendation G.975.1 February 2004. Forward error correction for high
48 bit-rate DWDM submarine systems
49

42 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1.5 Abbreviations 1
2
3
Table 1-1. Abbreviations 4
5
Abbreviation Meaning 6
BER Bit Error Ratio 7
8
BERT Bit Error Ratio Test or Tester
9
BUJ Bounded Uncorrelated Jitter 10
CBGJ Correlated Bounded Gaussian Jitter 11
12
CBHPJ Correlated Bounded High Probability Jitter
13
CEI Common Electrical I/O 14
CDF Cumulative Distribution Function 15
CDR Clock Data Recovery
16
17
CID Consecutive Identical Digits 18
CML Current Mode Logic 19
Cn Cursor number 20
21
DCD Duty Cycle Distortion 22
dB Decibel 23
DDJ Data Dependent Jitter 24
25
DFE Decision Feedback Equalizer
26
DJ Deterministic Jitter 27
DUT Device Under Test 28
29
EMI Electro-Magnetic Interference
30
erf error function 31
erfinv inverse error function 32
ESD Electro-Static Discharge
33
34
FEXT Far End Cross Talk 35
FFT Fast Fourier Transform 36
FIR Finite Impulse Response 37
38
Gbps Giga bits per second 39
GJ Gaussian Jitter 40
Gsym/s Giga symbols per second 41
42
HF High Frequency
43
HPF High Pass Filter 44
HPJ High Probability Jitter 45
46
IA Implementation Agreement
47
ISI Inter-Symbol Interference 48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 43

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 1-1. Abbreviations

2
3 Abbreviation Meaning
4 LMS Least Mean Square
5 LPF Low Pass Filter
6
7 LVDS [ 20] Low Voltage Differential Signal
8 LR Long Reach
9 mA milli-Amp
10
11 mV milli-Volt
12 NEXT Near End Cross Talk
13 NRZ Non Return to Zero
14
PCB Printed Circuit Board
15
16 PDF Probability Distribution Function
17 PECL Positive Emitter Coupled Logic
18
PJ Periodic Jitter
19
20 pp Peak to Peak
21 ppd Peak to Peak Differential (as in 300mVppd)
22 PLL Phase Locked Loop
23
24 ps pico second
25 PRBS Pseudo Random Bit Stream
26 Q Inverse error function
27
28 RJ Random Jitter
29 RV Random Variable
30 RX Receiver
31
S11 and S22 reflection coefficient
32
33 S21 transmission coefficient
34 SCC11 and SCC22 Common mode reflection coefficients
35
SCD11 and SCD22 Differential to common mode conversion coefficient
36
37 SDD11 and SDD22 Differential reflection coefficients
38 SDC11 and SDC22 Common mode to differential conversion coefficient
39 SFI SERDES - Framer Interface
40
41 SJ Sinusoidal Jitter
42 SPI System Packet Interface
43 SR Short Reach
44
45 sym/s symbols/second
46 TJ Total Jitter
47 TDM Time Division Multiplexed data
48
TFI TDM Fabric to Framer Interface
49

44 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 1-1. Abbreviations 1

2
Abbreviation Meaning 3
TX Transmitter 4
UBHPJ Uncorrelated Bounded High Probability Jitter
5
6
UI Unit Interval = 1/(baud rate) 7
UUGJ Uncorrelated Unbounded Gaussian Jitter 8
XAUI 10 Gigabit Attachment Unit Interface 9
10
11
1.6 Definitions 12
13
14
Table 1-2. General Definitions (with exception of Jitter and Wander) (Sheet 1 of 2) 15
16
Parameter Description 17
18
A parameter that reflects the quality of the serial transmission and detection
scheme. The Bit Error Ratio is calculated by counting the number of
19
Bit Error Ratio 20
erroneous bits output by a receiver and dividing by the total number of
transmitted bits over a specified transmission period. 21
Number of symbols per second, where a symbol can consist of more than
22
Baud rate one bit. 23
24
25
In this specification Channel shall mean electrical differential channel. The 26
channel is combination of electrical interconnects that together form the
signal path from reference points T to R - see Figure 1-6. The channel will 27
Channel 28
typically consist of PCB traces, via holes, component attachment pads and
connectors. A characteristic of a signal channel is the complex characteristic 29
impedance Z. 30
Common Mode Voltage Average of the Vhigh and Vlow voltage levels - see Figure 1-1 31
32
The use of this definition shall be understood as being with reference to a
Confidence level
Gaussian Distribution 33
34
The difference in the DC termination resistance with respect to ground of any 35
two signals forming a differential pair. Usually due to large process spread
the absolute termination resistance is specified relatively loose, e.g. 20% 36
Differential Termination 37
where the relative difference of resistors of the same device will be much
Resistance mismatch 38
less, e.g 5%. This parameter is used to specify the relative difference tighter
than the overall resistance for the purpose of minimizing differential signal 39
mode conversion 40
A statistical distribution (also termed “normal”) characterized by populations 41
Gaussian that are not bound in value and have well defined “tails”. The term “random” 42
in this document always refers to a Gaussian distribution. 43
Refers to a defined clock extraction unit which phase tracks the inherent 44
Golden PLL
clock present in a data signal. The phase tracking bandwidth is usually 45
defined in terms of a corner frequency and if not defined with a corner 46
frequency of baud/1667, a roll off of 20dB/dec and <0.1dB peaking
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 45

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 1-2. General Definitions (with exception of Jitter and Wander) (Sheet 2 of 2)
2
3 Parameter Description
4 An otherwise compliant channel that has been selected or altered to test
5 Stress Channel receiver or transmitter compliance (see also Stressed Signal (or) Stressed
6 Eye.)
7 Data dependent deterministic jitter caused by the time differences required
8 for the signal to arrive at the receiver threshold when starting from different
9 places in bit sequences (symbols). For example when using media that
10 attenuates the peak amplitude of the bit sequence consisting of alternating 0,
1, 0, 1... more than peak amplitude of the bit sequence consisting of 0, 0, 0,
11 0, 1, 1, 1, 1... the time required to reach the receiver threshold with the 0, 1,
12 0, 1... is less than required from the 0, 0, 0, 0, 1, 1, 1, 1... The run length of 4
13 Intersymbol Interference
produces a higher amplitude which takes more time to overcome when
14 changing bit values and therefore produces a time difference compared to
15 the run length of 1 bit sequence. When different run lengths are mixed in the
same transmission the different bit sequences (symbols) therefore interfere
16 with each other. Intersymbol Interference is expected whenever any bit
17 sequence has frequency components that are propagated at different rates
18 by the transmission media.
19 Lane A single CEI Channel
20
21 A functional connection between the Tx and Rx ports of 2 components, that
Link can be multiple or parallel CEI Lanes defined as 1:N. The definition a Link
22 does not imply duplex operation.
23
24 Defines an application where the high frequency transmit jitter of a device is
non-transparent
defined independently to the high frequency jitter present at any data input of
25 applications
the same device
26
27 The constant portion of the difference in the arrival time between the data of
Skew
any two in-band signals.
28
29 In order to test the tolerance of a receiver a stressed signal or eye is defined
30 Stressed Signal (or) which when applied to the receiver must be received with the defined Bit
Stressed Eye Error Rate. The stressed signal or eye is defined in terms of its horizontal
31 closure or jitter and amplitude normally in conjunction with an eye-mask.
32
33 Defines an application where the high frequency transmit jitter of a device is
Transparent applications dependent on the high frequency jitter present at one or more of the data
34 inputs of the same device
35
36 Symbol Unit of information conveyed by a single state transition in the medium
37 Symbol spaced Describes a time difference equal to the nominal period of the data signal
38 One nominal bit period for a given signaling speed. It is equivalent to the
39 Unit Interval shortest nominal time between signal transitions. UI is the reciprocal of
40 Symbol.
41
42
43
44
45
46
47
48
49

46 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Table 1-3. Jitter and Wander Definitions (Sheet 1 of 3) 2
3
Parameter Description 4
Jitter is deviation from the ideal timing of an event at the mean 5
amplitude of the signal population. Low frequency deviations are 6
tracked by the clock recovery circuit, and do not directly affect the 7
timing allocations within a bit interval. Jitter that is not tracked by the
Jitter
clock recovery circuit directly affects the timing allocations in a bit
8
interval. Jitter is phase variations in a signal (clock or data) after 9
filtering the phase with a single pole high pass filter with the -3 dB point 10
at the jitter corner frequency. 11
Total Jitter Convolution of all jitter components. 12
13
Jitter Jitter generation is the process whereby jitter appears at the output
Generation port in the absence of applied input jitter at the input port.
14
15
The ratio of the jitter output and jitter input for a component, device, or 16
system often expressed in dB. A negative dB jitter transfer indicates
the element removed jitter. A positive dB jitter transfer indicates the
17
Jitter Transfer 18
element added jitter. A zero dB jitter transfer indicates the element had
no effect on jitter. The ratio should be applied separately to 19
deterministic components and Gaussian (random) jitter components. 20
To enable enhancements in jitter methodology, more descriptive 21
terminology has been adopted. To enable the reader to understand the 22
Previous Terminology
mapping of previous descriptions the following terms are included for 23
clarity. 24
Data 25
Dependent Now referred to as Correlated Bounded High Probability Jitter 26
Jitter 27
Deterministic 28
Now referred to as High Probability Jitter
Jitter 29
Random Jitter Now referred to as Gaussian Jitter 30
31
Duty Cycle The term Duty Cycle Distortion was formerly used for both Duty Cycle
Distortion Distortion and Even-Odd Jitter.
32
33
An overall term that defines a jitter distribution that at the BER of 34
interest e.g. 1e-15 still shows a Gaussian distribution. Unless
Gaussian Jitter
otherwise specified Gaussian Jitter is the RMS sum of CBGJ and
35
UUGJ. 36
37
Jitter distribution that shows a true Gaussian distribution where the
observed peak to peak value has an expected value that grows as a
38
function of the measurement time. This form of jitter is assumed to 39
Jitter, 40
arise from phase noise random processes typically found in VCO
Unbounded
Gaussian
structures or clock sources. It is usually quantified as either the Root 41
Mean Square (RMS) or Sigma of the Gaussian distribution, or as the 42
expected peak value for a given measurement population. (Formally 43
defined as T_RJ)
44
Jitter distribution where the value of the jitter shows a correlation to the 45
Correlated
signal level being transmitted. The distribution is quantified, using a 46
Bounded
Gaussian approximation, as the gradient of the bathtub linearization at
Gaussian Jitter
the Bit Error Rate of interest. R_RJ = R_GJ 47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 47

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 1-3. Jitter and Wander Definitions (Sheet 2 of 3)

2
3 Parameter Description
4 Jitter distribution that at the BER of interest is approximated by a dual
5 Dirac. Unless otherwise specified High Probability Jitter is the
6 High probability Jitter convolution of UBHPJ, CBHPJ, PJ, SJ, DCD. The distribution is
7 quantified, using a dual Dirac approximation, as the offset of the
bathtub linearization at the Bit Error Rate of interest.
8
9 Uncorrelated
10 Bounded High Jitter distribution where the value of the jitter show no correlation to
Probability any signal level being transmitted. Formally defined as T_DJ.
11 Jitter.
12
13 Correlated Jitter distribution where the value of the jitter shows a strong
Bounded High correlation to the signal level being transmitted. This jitter may
14 Probability considered as being equalisable due to its correlation to the signal
15 Jitter level.
16
A sub form of HPJ that defines a jitter which has a single fundamental
17 Periodic Jitter
harmonic plus possible multiple even and odd harmonics.
18
19 Sinusoidal A sub form of HPJ that defines a jitter which has a single frequency
Jitter harmonic.
20
21 Even-Odd Jitter is measured on two repetitions of a repeating pattern
22 with an odd number of bits such as PRBS9. The deviation of the time
of each transition from an ideal clock at the signaling rate is measured.
23 Even-Odd Even-Odd Jitter is defined as the magnitude of the difference between
24 Jitter the average deviation of all even-numbered transitions and the
25 average deviation of all odd-numbered transitions but only actual
26 transitions are measured and averaged. Even-Odd Jitter is part of the
27 UBHPJ distribution and is measured at the time-averaged signal level.
28 The absolute value of the difference in the average width of a ’1’
29 Duty Cycle symbol or a ’0’ symbol and the ideal periodic time in a clock-like
30 Distortion repeating 0,1,0,1 sequence. Duty Cycle Distortion is part of the CBHPJ
distribution and is measured at the time-averaged signal level.
31
32 The peak to peak variation in the phase of a signal (clock or data) after
33 Wander filtering the phase with a single pole low pass filter with the -3db point
at the wander corner frequency. Wander does not include skew.
34
35 Correlated
Components of wander that are common across all applicable in band
36 wander
signals.
37
38 Relative Components of wander that are uncorrelated between any two in band
39 wander signals (See Figure 1-2)
40 The convolution of the correlated and uncorrelated wander. (See
41 Total wander
Figure 1-3)
42 Uncorrelated Components of wander that are not correlated across all applicable in
43 wander band signals.
44
Unit
45
46
47
48
49

48 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 1-3. Jitter and Wander Definitions (Sheet 3 of 3) 1

2
Parameter Description 3
Peak-to-Peak For any type of jitter, Peak to Peak Jitter is the full range of the jitter 4
Jitter distribution that contributes within the specified BER. 5
The root mean square value or standard deviation of jitter. See clause 6
Jitter RMS 7
2 for more information.
Refers to the standard deviation of a random variable modelled as a
8
Gaussian Distribution. When used in reference to jitter, it refers to the 9
standard deviation of the Gaussian Jitter component(s). When used in 10
Sigma reference to confidence levels of a result refers to the probability that 11
the result is correct given a Gaussian Mode, e.g. a measured result 12
with 3 sigma confidence level would imply that 99.9% of the 13
measurements are correct.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 49

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 1.6.1 Definition of Amplitude and Swing

2
3 See Figure 1-1 for an illustration of absolute driver output voltage limits and definition of
4 differential peak-to-peak amplitude.
5 Figure 1-1. Definition of Driver Amplitude and Swing
6
7
8
9
10
11
12
V high
13 True
14
15 M ax abs output

16 V CM
17
18
GND
19 Com plem ent
M in abs output

20 V low

21
22
23
24 True - Com plem ent Differential voltage
25 (peak-to-peak)

26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

50 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1.6.2 Definition of Skew and Relative wander 1

2
See Figure 1-2 for an illustration of skew and relative wander. 3
Figure 1-2. Skew and Relative Wander between in band Signals
4
5
6
Relative Wander between lanes 7
X and Y 8
Peak to peak 9
10
11
Lane Y The rising edges shown 12
are logical coincident data 13
with the transmitter 14
15
16
Lane X 17
18
19
Skew between
20
Lanes X and Y
21
22
Skew and Relative Wander 23
24
25
26
1.6.3 Definition of Total wander
27
See Figure 1-3 for an illustration of total wander in a signal 28
29
Figure 1-3. Total Wander of a Signal 30
31
32
33
Total Wander of a
Data or clock signal
34
Peak to peak 35
36
37
38
Lane Y 39
40
41
Edge of clean clock 42
that is frequency 43
locked to lane Y 44
45
Total Wander 46
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 51

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 1.7 Table Entries and Specifications

2
3 The CEI IA shall use a common tabular definition of the parameters specified. The
4 following section outlines examples of tables required for the definitions and the
5 corresponding entries. All clauses must use this structure. Additional clause specific
6 parameters are allowed.
7
8
9 1.7.1 Transmitter Electrical Output Specification
10
11 Table 1-4. Transmitter Electrical Output Specification
12
13 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
14 Baud Rate T_Baud Gsym/s
15 Output Differential Voltage T_Vdiff mVppd
16 DC Common mode Voltage T_Vcm mV
17
18 Output AC Common Mode Voltage T_VcmAC mVrms

19 Differential Resistance T_Rd 

20 Differential Termination Resistance Mismatch T_Rdm %
21 Output Rise and Fall Time (20% to 80%) T_tr, T_tf ps
22 Differential Output Return Loss T_SDD22 dB
23
Common Mode Output Return Loss T_SCC22 dB
24
25 NOTES:
26
27 Table 1-5. Transmitter Output Jitter Specification
28
29 Characteristic Symbol Condition MIN. TYP. MAX. UNIT

30 Uncorrelated Bounded High probability Jitter T_UBHPJ UIpp

31 Uncorrelated Unbounded Gaussian Jitter T_UUGJ UIpp
32 Duty cycle distortion T_DCD UIpp
33
Total Jitter T_TJ UIpp
34
Eye Mask T_X1 UI
35
36 Eye Mask T_X2 UI
37 Eye Mask T_Y1 mV
38 Eye Mask T_Y2 mV
39 NOTES:
40 1. Uncorrelated Unbounded Gaussian Jitter must be defined with respect to specified BER of 1e-15, Q=7.94
41
42
43
44
45
46
47
48
49

52 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Figure 1-4. Transmit Eye Mask 2
3
4
5
T_Y2 6
7
T_Y1
8
9
Amplitude (mV) 0 10
11
12
-T_Y1 13
-T_Y2 14
15
0.0 T_X1 T_X2 1-T_X2 1-T_X1 1.0 16
17
Time (UI) 18
19
20
21
22
1.7.2 Receiver Electrical Input Specification 23
Table 1-6. Receiver Electrical Input Specification 24
25
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
26
27
Baud Rate R_Baud Gsym/s
28
Input Differential Voltage R_Vdiff mVppd 29
DC Common mode voltage R_Vrcm mV 30
AC Common mode Voltage R_VcmAC mV 31
Differential Input Resistance R_Rdin  32
Input Resistance Mismatch R_Rm %
33
34
Differential Input Return Loss R_SDD11 dB
35
Common Mode Input Return Loss R_SCC11 dB 36
Differential to Common Mode Input
R_SCD11 dB 37
Conversion2
38
NOTES: 39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 53

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 1.7.3 Receiver input Jitter Specification

2
3 Table 1-7. Receiver Input Jitter Specification
4
5 Characteristic Symbol Condition MIN. TYP. MAX. UNIT

6 Uncorrelated Bounded High probability Jitter R_UBHPJ UIpp

7 Correlated Bounded High probability Jitter R_CBHPJ UIpp
8 Gaussian Jitter R_GJ UIpp
9
Sinusoidal Jitter R_SJ UIpp
10
Total Jitter R_TJ UIpp
11
12 Eye Mask R_X1 UI
13 Eye Mask R_Y1 mV
14 Eye Mask R_Y2 mV
15 NOTES:
16 1. Gaussian Jitter must be defined with respect to specified BER of 1e-15, Q=7.94
17
18
19 Figure 1-5. Receiver Input Mask
20
21
22
23
24 R_Y2
25
26 R_Y1
27
Amplitude (mV)

28 0

29
-R_Y1
30
31
32 -R_Y2
33
34 0.0 R_X1 0.5 1-R_X1 1.0
35
36 Time (UI)
37
38
39
40 1.8 Reference Model
41
42 The CEI common reference model is defined in Figure 1-6. In cases where
43 transmission direction matters the Ingress and Egress suffix is used, e.g. RI for
44 Receiver in the Ingress direction. In all other cases the R and T are used without a
45 suffix. Note that the RX and TX blocks include all off-chip components associated with
46 the respective function. Note also that a CEI Link does not imply a duplex connection,
47 so the reference model shown in Figure 1-6 represents 2 CEI links.
48
49

54 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 1-6. Reference Model 1

2
3
4
Component Component 5
Edge Edge 6
Egress 7
8
9
Channel 10
TX TE RE RX 11
12
13
14
Ingress 15
16
17
Channel
RX RI TI TX 18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 55

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4 1.A Appendix - Signal Definitions
5
6 Signals defined in this appendix are not referred to in this document, but relate to
7 subsequent applications of CEI Links, e.g. SFI, SPI, TFI. Possible applications for CEI
8 Links are described, but do not try to limit applications.
9
10 Whilst it is shown that CEI links can originate from a Serdes component, this is by no
11 means essential. It is likely that CEI Links will be generated and received by TX and RX
12 ports of an ASIC or FPGA component. In this case it will be necessary to have
13 multiplexing and demultiplexing functions within the ASIC or FPGA. When a Serdes
14 component is referred to, it can mean the Serializer/Deserializer is integrated within an
15 ASIC or FPGA component, as well as being a separate component. In some
16 applications, it will be necessary to also transmit control or status signals in parallel with
17 the CEI Link. Some applications will also require clocks to be transmitted with the data.
18
The signal paths or CEI Lanes are unidirectional point-to-point connections. Each CEI
19
Lane is made up of a balanced differential pair. A CEI Link can be comprised of a
20
unidirectional single lane or parallel lanes in either the transmit or receive direction. A
21
CEI Link does not imply duplex operation. See Figure 1-7 below for more information,
22
which shows 2 CEI Links, in the receive and transmit directions.
23
24 Figure 1-7. Signal Diagram
25
26 Optional
27 Reference
Clock
28
RefClk

29 Optional
30
Receive
Transmit

CEI
CEI

31
32
Connector

Connector

TxData0P
33 TxData0N
34
35 CEI Device
36
Connector

Connector

RxData0P
37 RxData0N
38
Receive

Transmit

39
CEI

CEI

40
41 Card Backplane Card
42
43
44
45
46
47
48
49

56 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

. 1
2
Table 1-8. Receive Signal Summary 3
Signal Name Direction Function 4
The Receive Data (RXDATA[n]) signals are the inputs
5
RXDATA[n..0]P/N Input to SERDES Component
to the SERDES component. 6
7
8
9
Table 1-9. Transmit Signal Summary 10
Signal Name Direction Function 11
12
The Transmit Data (TXDATA[n]) signals are the outputs
TXDATA[n..0]P/N Output of SERDES Component
of the SERDES component. 13
14
15
16
An example specification for the reference clock for a typical application is proposed in 17
Table 1-10 below. 18
19
Table 1-10. Example specification of reference clock 20
Characteristic Description
21
22
Input Buffer Internal Terminated LVDS
23
Frequency Divide by 16 (e.g. 622MHz @9.95Gsym/s) 24
Rise/fall time (20/80%) 200ps 25
Duty cycle variation <10% 26
Receiver Reference Clock frequency tolerance against data +/-100ppm 27
Phase noise -125dBc at 1MHz
28
29
It is expected that the reference clock input supports DC coupling, with AC coupling 30
31
being optional (LVDS input having center tap or self biasing).
32
One reference clock input can support multiple Rx and Tx channels. 33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 1 : Introduction, Definitions and Formats 57

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 1.B Appendix - Examples of CEI links in Typical systems

2
3
4
Figure 1-8. Some typical systems
5
6
7 System-Packet SERDES Framer SERDES Framer
8 Interface (SPI) Interface (SFI) Interface (SFI)
9 Transmit
10 Interface
11 Data (SPI) Data Data
12 Transmit
13 Link Layer
14 Device Serdes
15 Status Clock Clock Device
Phy FEC
16 Device Device And
17 Receive Data Data Data Optics
18 Link Layer
19 Device
20
21 Status Clock Clock
Receive
22 Interface
23 (SPI)
24
25
26
27
28 TDM Framer SERDES Framer SERDES Framer
29 Interface (TFI) Interface (SFI) Interface (SFI)
30
31
32
33 Data Data Data
34
35
36 Serdes
Clock Clock
37 TDM Switch Framer FEC Device
38 Device Device And
39 Data Data Data Optics
40
41
42 Clock Clock
43
44
45
46
47
48
49

58 Clause 1: Introduction, Definitions and Formats - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2 Jitter and Interoperability Methodology 1

2
3
This clause describes the requirements for interoperability testing of electrical 4
interfaces as defined within this implementation agreement. The clause is organized 5
into several methods of which the later Clauses will reference as the method for jitter or
6
interoperability testing. 7
8
9
2.1 Method A 10
11
This sub-clause defines the interoperability methodology specifically for interfaces 12
where neither transmit emphasis or receiver equalization are required for the receiver 13
eye to be open to within the BER of interest. 14
15
2.1.1 Defined Test Patterns1 16
17
The following patterns shall be used for the testing of jitter tolerance and output jitter 18
compliance. 19
20
2.1.1.1 CID Jitter Tolerance Pattern 21
22
• The pattern is inverting to exercise possible weaknesses in rise and fall time 23
symmetry 24
25
• 72 bits are defined for the Consecutive Identical Digits (CID) which aligns to [22.] 26
recommendation 27
• The length of the PRBS31 is defined as greater than or equal to 10328 28
29
• The pattern is based on transition density comparisons between various PRBS 30
patterns and a 3 sigma worst case analysis of a scrambled OC-768 frame. 31
Figure 2-1. CID Jitter Tolerance Pattern 32
33
34
Seed =

Seed =

CID PRBS31 CID inv PRBS31

35
all 1

all 1

72 x zero >=10328 bits 72 x one >=10328 bits

36
37
38
2.1.1.2 Jitter Tolerance and General Test Patterns 39
40
• The pattern is a free running PRBS31 polynomial 41
42
2.1.2 Channel Compliance 43
44
The following steps shall be made to identify which channels are to be considered 45
compliant. 46
47
48
1. All descriptions to PRBS31 imply the standard polynomial as described in [21.] 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 59

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 1. The forward channel and significant crosstalk channels shall be measured using a
2 Network analyzer for the maximum baud rate that the channel is to operate at shall
3 be used (see Appendix 2.E.6 for a suggested method)
4
5 2. An effective transmit filter as defined by the reference transmitter shall be used
6 3. An amplitude as defined by the reference transmitter shall be used
7
8 4. A transmitter jitter distribution (see Annex 2.C.4) as defined by the reference
9 transmitter shall be used
10 5. A transmitter return loss as defined by the reference transmitter shall be used
11
12 6. A sampling point as defined by the reference receiver shall be used
13 7. A receiver return loss as defined by the reference receiver shall be used
14
15 8. The opening of the eye shall be calculated using Statistical Eye Analysis methods,
16 as per Annex 2.C.5, and confirmed to be within the requirements at the required
17 BER of the Implementation Agreement, usually,
18 — Amplitude at the zero time offset sampling point
19
20 — Time jitter measured at the zero amplitude sampling point
21
22 2.1.3 Transmitter Compliance
23
24 The following steps shall be made to identify which transmitters are to be considered
25 compliant.
26
27 1. The high frequency transmit jitter shall be within that specified (see Appendix 2.E.1
28 for suggested methods)
29 2. The specified transmit eye mask shall not be violated (see Appendix 2.E.7 for a
30 suggested method), after adjusting the horizontal time positions for the measured
31 time with a confidence level of 3 sigma (see Appendix 2.F.3 for a suggested method
32 of calculating Q given a measurement population)
33
34 3. The total wander shall be within that specified (see Appendix 2.E.2 for a suggested
35 measurement method)
36 4. The relative wander shall be within that specified (see Appendix 2.E.3 for a
37 suggested measurement method)
38
39
40 2.1.4 Receiver Compliance
41
42 The following step shall be made to identify which receivers are to be considered
43 compliant.
44 1. The DUT shall be measured to have a BER1 better than specified for a stressed
45 signal (see Appendix 2.E.4.1 for a suggested method) with a confidence level of
46 three sigma (see Appendix 2.F.2 for a suggested method), given:
47
48
49 1. if the defined measurement BER is different to system required BER, adjustments to applied stressed eye TJ are necessary

60 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

— The defined sinusoidal jitter mask for relative and total wander as per Annex 1
2.A.1 and Annex 2.A.2, with a high frequency total/relative wander of 0.1UI and 2
a maximum total/relative wander as defined in the Implementation Agreement. 3
Note that in some Implementation Agreements one needs to reduce the amount 4
of High Probability Jitter by 0.1UI to account for this sinusoidal jitter. 5
6
7
2.2 Method B 8
9
This sub-clause defines the interoperability methodology specifically for interfaces 10
where transmit emphasis may be used however receiver equalization is not required 11
for the receiver eye to be open to within the BER of interest. 12
13
2.2.1 Defined Test Patterns1 14
15
The following pattern shall be used for the testing of jitter tolerance and output jitter 16
compliance. 17
18
• A free running PRBS31 polynomial 19
20
2.2.2 Channel Compliance 21
22
The following steps shall be made to identify which channels are to be considered 23
compliant. 24
25
1. The forward channel and significant crosstalk channels shall be measured using a 26
Network analyzer for the maximum baud rate that the channel is to operate at shall 27
be used (see Appendix 2.E.6 for a suggested method) 28
2. An n-tap emphasized transmitter as per Annex 2.B.3, where “n” is defined by the 29
reference transmitter shall be used 30
31
3. An effective transmit filter as defined by the reference transmitter shall be used 32
4. An amplitude as defined by the reference transmitter shall be used 33
34
5. A transmitter jitter distribution (see Annex 2.C.4) as defined by the reference 35
transmitter shall be used 36
6. A transmitter return loss as defined by the reference transmitter shall be used 37
38
7. A sampling point as defined by the reference receiver shall be used 39
40
8. A receiver return loss as defined by the reference receiver shall be used 41
9. The opening of the eye shall be calculated using Statistical Eye Analysis methods, 42
as per Annex 2.C.5, and confirmed to be within the requirements at the required 43
BER of the Implementation Agreement, usually: 44
45
— Amplitude at the zero time offset sampling point 46
— Time jitter measured at the zero amplitude sampling point 47
48
1. All descriptions to PRBS31 imply the standard polynomial as described in [21.] 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 61

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.2.3 Transmitter Compliance

2
3 The following step shall be made to identify which transmitters are to be considered
4 compliant.
5
1. It shall be verified that the measured eye is equal or better than the calculated eye
6
for the given measurement probability Q (see Appendix 2.F.3 for a suggested
7
method of calculating Q given a measurement population), given:
8
9 — A stress channel that is otherwise compliant as per 2.2.2, that requires at least
10 half the maximum transmit emphasis as specified in the relevant clause or IA,
11 with no receiver filtering or equalisation to produce an open eye.
12
13 — Using this channel the transmitter shall be then optimally adjusted and the
14 resulting eye measured (see Appendix 2.E.7 for a suggested method).
15 — Using this channel the statistical eye shall then be calculated, as per Annex
16 2.C.5, using the maximum defined transmit jitter and the actual transmitter's
17 amplitude and emphasis.
18
19 If the transmit jitter or transmit eye mask is additionally defined then the following steps
20 shall also be made to identify which transmitters are to be considered compliant:
21
22 1. The high frequency transmit jitter shall be within that specified (see Appendix 2.E.1
23 for suggested methods)
24 2. The specified transmit eye mask shall not be not violated (see Appendix 2.E.7 for a
25 suggested method), after adjusting the horizontal time positions for the measured
26 time with a confidence level of 3 sigma (see Appendix 2.F.3 for a suggested
27 method)
28
29
2.2.4 Receiver Compliance
30
31 The following step shall be made to identify which receivers are to be considered
32 compliant.
33
34 1. The DUT shall be measured to have a BER1 better than specified for a stressed
35 signal (see Appendix 2.E.4.2 for a suggested method) with a confidence level of
36 three sigma (see Appendix 2.F.2 for a suggested method), given:
37
— The defined sinusoidal jitter mask for total and relative wander as per Annex
38
2.A.1 and Annex 2.A.2, with a high frequency total/relative wander and a
39
maximum total/relative wander as defined in the Implementation Agreement
40
41 — The specified amount of High Probability Jitter and Gaussian jitter.
42
43
44
45
46
47
48
49 1. if the defined measurement BER is different to system required BER, adjustments to applied stressed eye TJ are necessary

62 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.3 Method C 1
2
This sub-clause defines the interoperability methodology specifically for interfaces 3
where transmit emphasis may be used and the receiver eye requires Linear 4
Continuous Time equalization (from channel interoperability point of view) to be open to 5
within the BER of interest. 6
7
8
2.3.1 Defined Test Patterns1 9
10
The following pattern shall be used for the testing of jitter tolerance and output jitter 11
compliance.
12
• A free running PRBS31 polynomial 13
14
2.3.2 Channel Compliance 15
16
The following steps shall be made to identify which channels are to be considered 17
compliant. 18
19
1. The forward channel and significant crosstalk channels shall be measured using a 20
Network analyzer for the maximum baud rate that the channel is to operate at shall 21
be used (see Appendix 2.E.6 for a suggested method) 22
2. An n-tap emphasized transmitter as per Annex 2.B.3, where “n” is defined by the 23
reference transmitter shall be used 24
25
3. An effective transmit filter as defined by the reference transmitter shall be used 26
27
4. An amplitude as defined by the reference transmitter shall be used 28
5. A transmitter jitter distribution (see Annex 2.C.4) as defined by the reference 29
transmitter shall be used 30
31
6. A transmitter return loss as defined by the reference transmitter shall be used 32
7. An ideal receiver filter of the form in Annex 2.B.7, using the restrictions as defined 33
by the reference receiver shall be used 34
35
8. A sampling point as defined by the reference receiver shall be used 36
9. A receiver return loss as defined by the reference receiver shall be used 37
38
10. The opening of the eye shall be calculated using Statistical Eye Analysis methods, 39
as per Annex 2.C.5, and confirmed to be within the requirements at the required 40
BER of the Implementation Agreement, usually: 41
— Amplitude at the zero time offset sampling point 42
43
— Time jitter measured at the zero amplitude sampling point 44
45
46
47
48
1. All descriptions to PRBS31 imply the standard polynomial as described in [21.] 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 63

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.3.3 Transmitter Compliance

2
3 The following step shall be made to identify which transmitters are to be considered
4 compliant.
5
1. It shall be verified that the measured eye is equal or better than the calculated eye
6
for the given measurement probability Q (see Appendix 2.F.3 for a suggested
7
method of calculating Q given a measurement population), given:
8
9 — A stress channel that is otherwise compliant as per 2.3.2, that requires at least
10 half the maximum transmit emphasis as specified in the relevant clause or IA,
11 with no receiver filtering or equalisation to produce an open eye.
12
13 — Using this channel the transmitter shall be then optimally adjusted and the
14 resulting eye measured (see Appendix 2.E.7 for a suggested method).
15 — Using this channel the statistical eye shall then be calculated, as per Annex
16 2.C.5, using the maximum defined transmit jitter and the actual transmitter's
17 amplitude and emphasis.
18
19 If the transmit jitter or transmit eye mask is additionally defined then the following steps
20 shall also be made to identify which transmitters are to be considered compliant:
21
22 1. The high frequency transmit jitter shall be within that specified (see Appendix 2.E.1
23 for suggested methods)
24 2. The specified transmit eye mask shall not be violated (see Appendix 2.E.7 for a
25 suggested method), after adjusting the horizontal time positions for the measured
26 time with a confidence level of 3 sigma (see Appendix 2.F.3 for a suggested
27 method)
28
29
2.3.4 Receiver Compliance
30
31 The following step shall be made to identify which receivers are to be considered
32 compliant.
33
34 1. The DUT shall be measured to have a BER1 better than specified for a stressed
35 signal (see Appendix 2.E.4.3 for a suggested method) with a confidence level of
36 three sigma (see Appendix 2.F.2 for a suggested method), given:
37
— The defined sinusoidal jitter mask for total and relative wander as per Annex
38
2.A.1 and Annex 2.A.2, with a high frequency total/relative wander and a
39
maximum total/relative wander as defined in the Implementation Agreement
40
41 — The specified amount of High Probability Jitter and Gaussian jitter.
42
43 — A stress channel or filter as identified by the methods of 2.3.2. If the optional
44 transmit filter of Appendix 2.E.4.3 is not included then no transmit emphasis
45 shall be enabled in the reference transmitter. If the transmitter filter of Appendix
46 2.E.4.3 is present then the standard reference transmitter (as used in channel
47 compliance) shall be used. The transmit filter characteristics (e.g. emphasis
48
49 1. if the defined measurement BER is different to system required BER, adjustments to applied stressed eye TJ are necessary

64 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

settings) shall be set in accordance with the optimised values resulting when the 1
methods of 2.3.2 are applied. 2
3
— An additive crosstalk signal of amplitude such that the resulting statistical eye, 4
given the channel, jitter and crosstalk, is as close as feasible in amplitude when
5
compared to the defined minimum amplitude for channel compliance.
6
7
8
2.4 Method D 9
10
This sub-clause defines the interoperability methodology specifically for interfaces 11
where transmit emphasis may be used and the receiver eye requires DFE equalization 12
(from channel interoperability point of view) to be open to within the BER of interest. 13
14
2.4.1 Defined Test Patterns1 15
16
The following pattern shall be used for the testing of jitter tolerance and output jitter 17
compliance. 18
• A free running PRBS31 polynomial 19
20
21
2.4.2 Channel Compliance 22
23
The following steps shall be made to identify which channels are to be considered 24
compliant. 25
1. The forward channel and significant crosstalk channels shall be measured using a 26
Network analyzer for the maximum baud rate that the channel is to operate at shall 27
be used (see Appendix 2.E.6 for a suggested method) 28
29
2. An n-tap emphasized transmitter as per Annex 2.B.3, where “n” is defined by the 30
reference transmitter shall be used 31
3. An effective transmit filter as defined by the reference transmitter shall be used 32
33
4. An amplitude as defined by the reference transmitter shall be used 34
5. A transmitter jitter distribution (see Annex 2.C.4) as defined by the reference 35
transmitter shall be used 36
37
6. A transmitter return loss as defined by the reference transmitter shall be used 38
39
7. An ideal receiver filter of the form in Annex 2.B.6, using the restrictions as defined 40
by the reference receiver shall be used 41
8. Any parameters that have degrees of freedom e.g. filter coefficients or sampling 42
point, shall be optimised against the amplitude, at the zero phase offset, as 43
generated by the Statistical Eye Output. e.g. by sweeping all degrees of freedom 44
and selecting the parameters giving the maximum amplitude. A receiver return loss, 45
as defined by the reference receiver, shall be used 46
47
48
1. All descriptions to PRBS31 imply the standard polynomial as described in [21.] 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 65

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9. The opening of the eye shall be calculated using Statistical Eye Analysis methods,
2 as per Annex 2.C.5, and confirmed to be within the requirements at the required
3 BER of the Implementation Agreement, usually:
4
5 — Amplitude at the zero time offset sampling point
6 — Time jitter measured at the zero amplitude sampling point
7
8
2.4.3 Transmitter Compliance
9
10 The following step shall be made to identify which transmitters are to be considered
11 compliant.
12
13 1. It shall be verified that the measured eye is equal or better than the calculated eye
14 for the given measurement probability Q (see Appendix 2.F.3 for a suggested
15 method of calculating Q given a measurement population), given:
16
— A stress channel that is otherwise compliant as per 2.4.2, that requires at least
17
half the maximum transmit emphasis as specified in the relevant clause or IA,
18
with no receiver filtering or equalisation to produce an open eye.
19
20 — Using this channel the transmitter shall be then optimally adjusted and the
21 resulting eye measured (see Appendix 2.E.7 for a suggested method).
22
23 — Using this channel the statistical eye shall then be calculated, as per Annex
24 2.C.5, using the maximum defined transmit jitter and the actual transmitter's
25 amplitude and emphasis.
26
If the transmit jitter or transmit eye mask is additionally defined then the following steps
27
shall also be made to identify which transmitters are to be considered compliant:
28
29 1. The high frequency transmit jitter shall be within that specified (see Appendix 2.E.1
30 for suggested methods)
31
32 2. The specified transmit eye mask shall not be violated (see Appendix 2.E.7 for a
33 suggested method), after adjusting the horizontal time positions for the measured
34 time with a confidence level of 3 sigma (see Appendix 2.F.3 for a suggested
35 method)
36
37 2.4.4 Receiver Compliance
38
39 The following step shall be made to identify which receivers are to be considered
40 compliant.
41 1. The DUT shall be measured to have a BER1 better than specified for a stressed
42 signal (see Appendix 2.E.4.3 for a suggested method) with a confidence level of
43 three sigma (see Appendix 2.F.2 for a suggested method), given:
44
45 — The defined sinusoidal jitter mask for total and relative wander as per Annex
46 2.A.1 and Annex 2.A.2, with a high frequency total/relative wander and a
47 maximum total/relative wander as defined in the Implementation Agreement
48
49 1. if the defined measurement BER is different to system required BER, adjustments to applied stressed eye TJ are necessary

66 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

— The specified amount of High Probability Jitter and Gaussian jitter. 1

2
— A stress channel or filter as identified by the methods of 2.4.2. If the optional 3
transmit filter of Appendix 2.E.4.3 is not included then no transmitter emphasis 4
shall be enabled in the reference transmitter. If the transmitter filter of Appendix 5
2.E.4.3 is present then the standard reference transmitter (as used in channel
6
compliance) shall be used. The transmit filter characteristics (e.g. emphasis 7
settings) shall be set in accordance with the optimised values resulting when the 8
methods of 2.4.2 are applied. 9
— An additive crosstalk signal of amplitude such that the resulting statistical eye, 10
given the channel, jitter and crosstalk, is as close as feasible in amplitude when 11
compared to the defined minimum amplitude for channel compliance 12
13
14
2.5 Method E 15
16
The following sub-clause defines the Interoperability methodology for interfaces where 17
a simple receiver equalization may be used to improve the margin of the link and 18
transparent applications may be used and the receiver eye is still open to within the 19
BER of interest. 20
21
22
2.5.1 Defined Test Patterns 23
The following pattern shall be used for the testing jitter tolerance and output jitter 24
25
compliance
26
• A free running PRBS31 polynomial 27
28
when used in transparent applications the additional test pattern defined in 2.5.1.1 must 29
be additionally tested. 30
31
2.5.1.1 CID Jitter Tolerance Pattern 32
33
• The pattern is inverting to exercise possible weaknesses in rise and fall time 34
symmetry 35
36
• 72 bits are defined for the Consecutive Identical Digits (CID) which aligns to [22.] 37
recommendation 38
• The length of the PRBS31 is defined as greater than or equal to 10328 39
40
• The pattern is based on transition density comparisons between various PRBS 41
patterns and a 3 sigma worst case analysis of a scrambled OC-768 frame. 42
Figure 2-2. CID Jitter Tolerance Pattern 43
44
45
Seed =

Seed =

CID PRBS31 CID inv PRBS31

46
all 1

all 1

72 x zero >=10328 bits 72 x one >=10328 bits

47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 67

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.5.2 Channel Compliance

2
3 The following steps shall be made to identify which channels are to be considered
4 compliant.
5
1. The forward channel and significant crosstalk channels shall be measured using a
6
Network analyzer for the maximum baud rate that the channel is to operate at shall
7
be used (see Appendix 2.E.6 for a suggested method)
8
9 2. An effective transmit filter as defined by the reference transmitter shall be used
10
11 3. An amplitude as defined by the reference transmitter shall be used
12 4. A transmitter jitter distribution (see Annex 2.C.4) as defined by the reference
13 transmitter shall be used
14
15 5. A transmitter return loss as defined by the reference transmitter shall be used
16 6. All defined reference receivers
17
18 7. A sampling point as defined by the reference receiver shall be used
19 8. A receiver return loss as defined by the reference receiver shall be used
20
21 9. The opening of the eye shall be calculated using Statistical Eye Analysis methods,
22 as per Annex 2.C.5, and confirmed to be within the requirements at the required
23 BER of the Implementation Agreement for both receiver types, usually:
24 — Amplitude at the zero time offset sampling point
25
26 — Time jitter measured at the zero amplitude sampling point
27 10. Any parameters that have degrees of freedom e.g. filter coefficients, shall be
28 optimised against the amplitude, at the zero phase offset, as generated by the
29 Statistical Eye Output. e.g. by sweeping all degrees of freedom and selecting the
30 parameters giving the maximum amplitude.
31
32
33 2.5.3 Transmitter Compliance
34
35 The following steps shall be made to identify whether a transmitter is considered
36 compliant.
37 1. the high frequency transmit jitter shall be within that specified (see Appendix 2.E.1
38 for suggested methods)
39
40 • for jitter transparent applications the bandwidth of any defined Golden PLL should
41 be adjusted according to the specific Implementation Agreement e.g. 8MHz for ITU
42 2. Specifically for “transparent ITU application egress transmitters” the transmit peak
43 to peak jitter and optionally rms jitter with the defined bandwidth shall be less than
44 that specified (see Appendix 2.E.1.2 for suggested methods)
45
46 3. Specifically for “transparent ingress transmitters” the defined jitter transfer mask
47 shall be less than that specified (see Appendix 2.E.5 for suggested methods)
48
49

68 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

• an applied sinusoidal jitter conforming to the defined jitter tolerance mask for this 1
line interface 2
3
4. the specified transmit eye mask is not violated (see Appendix 2.E.7 for a suggested 4
method), after adjusting the horizontal time positions for the measured time and a 5
confidence level of 3 sigma (see Appendix 2.F.3 for a suggested method of
6
calculating Q given a measurement population ) 7
5. the total wander is less than that specified (see Appendix 2.E.2 for a suggested 8
method) 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 69

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.5.4 Receiver Compliance

2
3 The following steps shall be made to identify whether a receiver is considered
4 compliant.
5
1. The DUT shall be measured to have a BER1 better than specified for a stressed
6
signal (see Appendix 2.E.4.2 for a suggested method) with a confidence level of
7
three sigma (see Appendix 2.F.2. for a suggested method) given
8
9 • for non-transparent applications, the defined sinusoidal jitter mask for relative and
10 total wander as per Annex 2.A.1 and Annex 2.A.2, with a high frequency total/
11 relative wander and a maximum total/relative wander as defined in the
12 Implementation Agreement
13
14 • for transparent application, the defined appropriate sinusoidal jitter mask for the
15 specific optical standard
16 • the high frequency jitter should be calibrated by either
17
18 — applying the maximum specified amount of receiver High Probability Jitter and
19 Gaussian jitter2 including CBHPJ
20
21 or
22 — applying the maximum specified amount of receiver High Probability Jitter and
23 Gaussian jitter3 excluding CBHPJ
24
25 — cascading with a compliance channel or filter as identified by 2.5.2.
26 — applying an additive crosstalk signal of amplitude such that the resulting
27 statistical eye, given the channel, jitter and crosstalk, is as close as feasible in
28 amplitude when compared to the defined minimum amplitude for channel
29 compliance
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46 1. if the defined measurement BER is different to system required BER, adjustments to applied stressed eye TJ are necessary
47 2. for jitter “transparent application ingress receivers” the bandwidth of any defined Golden PLL for the calibration of the HPJ and
GJ should be adjusted according to the specification Implementation Agreement e.g. 8MHz for ITU
48 3. for jitter “transparent application ingress receivers” the bandwidth of any defined Golden PLL for the calibration of the HPJ and
49 GJ should be adjusted according to the specification Implementation Agreement e.g. 8MHz for ITU

70 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.A Annex - Masks 1

2
3
2.A.1 Annex - Total Wander Mask 4
Total wander specifications should be considered as accumulated low frequency jitter. 5
6
As modern CDRs are digitally based they show a corner tracking frequency plus slew
7
limitation which has been guaranteed, therefore for jitter tolerance testing the total
8
wander needs to be spectrally defined to ensure correct operation.
9
To this end, for jitter tolerance testing, the wander is considered a sinusoidal jitter 10
source as shown below. 11
12
Figure 2-3. Total Wander Mask 13
14
Total Wander Amplitude
15
16
17
18
20dB/dec
19
SJ

High
Frequency 20
Amplitude 21
22
baud/1 667 20MHz
23
24
25
At higher frequency this jitter source is used to ensure margin in the high frequency 26
jitter tolerance of the receiver. At lower frequencies the higher SJ should then be 27
tracked by the CDR. 28
29
30
2.A.2 Annex - Relative Wander Mask 31
32
Specifically for interfaces defining relative wander, Figure 2-4 is also defined in terms of 33
a sinusoidal jitter source as shown below. 34
Figure 2-4. Relative Wander Mask 35
36
37
Relative Wander
Amplitude
38
39
40
20dB/dec
41
42
SJ

High
Frequency 43
Amplitude 44
45
baud/1 667 20MHz
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 71

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.A.3 Annex - Random Jitter Mask

2
3 To ensure that the random jitter modulation of stressed signals is above the CDR
4 bandwidth and therefore untracked, the following filter mask shall be applied where
5 necessary.
6 Figure 2-5. Random Jitter Spectrum
7
8
9
0dB
10
11
12 20dB/dec
Power(dB)

13
14
15
16
17
18 1 0MHz f baud /2

19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

72 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.B Annex - Pulse Response Channel Modelling 1

2
This annex shall describe the theoretical background for channel modelling 3
4
5
2.B.1 Annex - Generating a Pulse Response 6
7
Given the spectral transfer function as per Chapter 2.E.6 the pulse response of the 8
channel can be calculated using tools such as Matlab. 9
10
The Pulse Response of the channel is the received pulse for an ideal square wave and
11
is calculated by either
12
• convolving the pulse with the impulse response of the channel or 13
14
• multiplying the Fourier spectrum of the ideal transmitted square wave with the 15
channel response and taking the inverse Fourier transform, 16
1 17
t step = ----------- 18
f max 19
20
t = t step  n 21
22
n = [1,P] 23
24
25
tx  t  = H(0)  H(t period – t) 26
27
28
rx() = tx()  Tr() 29
30
31
rx(t) = ifft(rx()) 32
33
34
where 35
36
f max is difference between the maximum positive and minimum negative frequency 37
P is the number of equally space points in the frequency array 38
39
tx  t  is the transmit signal pulse 40
41
tx    is the transmit signal pulse in the frequency domain 42
43
Tr() is the transfer function of the channel 44
45
rx(t) is the resulting pulse response of the channel 46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 73

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.B.2 Annex - Basic Pulse Response Definitions

2
3 A receive pulse response as calculated above can be graphically represented, Figure
4 2-6.
5 Figure 2-6. Graphical Representation of Receiver Pulse
6
7 c0
8
9
10
Amplitude

11
12
13 c -1 c1 c4
14 c2
15
16 c3
Time
17 (Baud spaced intervals)
18
19 Cursors are defined as being the amplitude of the received pulse at symbol spaces
20 from the maximum signal energy at c , and extend to infinity in both negative and
21 positive time. The exact position of c 0is arbitrary and is defined specifically by the
0
22 various methodologies.
23
24 A precursor is defined as a cursor that occurs before the occurrence of the main signal
25 c , i.e. c where n<0, usually convergences to zero within a small number of bits
0 n
26
27 A post cursor is defined as a cursor that occurs after the occurrence of the main signal
28 c0, i.e. cn where n>0, and usually convergences to zero within twice the propagation
29 time of the channel.
30
31 Given a deterministic data stream travelling across the channel, the superposition of
32 the channel pulses give rise to Inter-Symbol Interference (ISI). This ISI has a maximum
33 occurring for a worst case pattern, which for a channel response where all cursors are
34 positive would be a single 1 or 0 in the middle of a long run of 0s or 1s respectively. This
35 maximum is referred to Total Distortion
36 n=
37
38
39
 =  cn
40  n = –    n  0 
41
42
43 Due to ISI an enclosure in the time domain also occurs which can be determined by
44 either running exhaustive simulations or simulations with determined worst case
45 patterns. For the case where the ISI is so large that the eye is closed, Inherent Channel
46 Jitter has no meaning.
47
48
49

74 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.B.3 Annex - Transmitter Pulse Definition 1

2
A transmitter is defined by its ability to generate a transmit pulse. A single 1 transmit 3
symbol has different amplitudes at symbol space intervals, tn, where post taps have 4
n>0, and pre-taps have n<0. 5
Figure 2-7. Transmit Pulse 6
7
8
9
10
Amplitude
t(0) 11
12
13
Time (Unit Intervals)
14
t(-1)
15
t(1)
16
17
-1UI 0UI 1UI 2UI 18
19
20
When a pulse train is transmitted the exact transmitted amplitude is therefore the 21
superposition of the pulses from the previous and to be transmitted pulses, so as in a 22
FIR filter. 23
Figure 2-8. Transmitter FIR Filter Function 24
25
+tn-1+tn+tn+1 26
an z-1 z-1 +tn-1+tn-tn+1 27
-tn-1+tn+tn+1
28
29
tn-1 tn tn+1 -tn-1+tn-tn+1 30
31
32
33
transmit +tn-1-tn+tn+1 34
+ signal
+tn-1-tn-tn+1 35
-tn-1-tn+tn+1
36
37
-tn-1-tn-tn+1
38
39
This superposition can be understood by referring to the amplitudes depicted for 40
various bit sequences in Figure 2-8. 41
42
The transmit emphasis can be defined to have certain limits of maximum transmit 43
amplitude or ratios of emphasis as defined below 44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 75

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 t1
3 P post = ----
t0
4
5 1 + P post
6 E = 20 log ----------------------
1 – P post
7
8
9
10
 t n  T_Vdiff
11
12 where
13
14 P post is the first coefficient of the transmit FIR
15
16 E is the emphasis of the transmit emphasis
17
18 T_Vdiff is the maximum transmit amplitude
19
20
21
22 2.B.4 Annex - Receiver Pulse Response
23
24 Given an emphasized transmitter the pulse response of the receiver should be
25 recalculated using the emphasized transmit pulse as opposed to a simple NRZ pulse.
26
27 the receiver pulse cursors are then defined as follows
28 Figure 2-9. Receiver Pulse Definition
29
30
r0
31
32
33
Amplitude

34
35
36 r-1 r4
37 r2
r1
38
39 r3
Time
40 (Baud spaced intervals)
41
42
43
44
45
46
47
48
49

76 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.B.5 Annex - Crosstalk Pulse Response 1

2
The crosstalk pulse response is analogous to the receiver pulse response as defined in 3
Annex 2.B.4 but using the crosstalk channel, i.e. NEXT or FEXT network analysis 4
measurement.. The transmit signal as seen in the system should be used for the 5
Figure 2-10. Crosstalk Pulse Definition 6
7
x0 8
9
10
11
Amplitude

12
13
x -2 x4
14
x2
15
16
x1 x3
17
x -1
Time
18
(Baud spaced intervals) 19
20
21
calculation of the resulting crosstalk pulse response, e.g.an emphasized transmitter 22
from above, or XAUI transmit NRZ pulse. 23
24
The Crosstalk pulse response is then defined as above, as being a set of cursors xn 25
usually oscillatory in form. The position of x0 is defined as being at the maximum 26
amplitude of the pulse response. 27
28
2.B.6 Annex - Decision Feedback Equalizer 29
30
The following filter function can be used to verify the capability of the channel to be 31
used in such an application. 32
Figure 2-11. Decision Feedback Equalizer 33
34
35
- z-1 z-1 z-1 z-1 36
37
kn kn-1 kn-2 kn-3
38
39
40
41
42
+ 43
44
45
46
The value of the coefficients are calculated directly from the channel pulse response or 47
the receiver pulse using an emphasized transmitter. 48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 77

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 kn = cn for unemphasized transmitters, or

2 n = [1,m]
3 kn = rn for emphasized transmitters
4 n = [ 1 , m]
5 This equalizer is capable of equalizing a finite number of post cursors, whose individual
6 values may be limited.
7
8 2.B.7 Annex - Time Continuous Transverse Filter
9
10 A.k.a. Feed forward Filter, Finite Input Response or Comb Structure, the Transverse
11 Filter, Figure 2-12 consists of a finite number of coefficients, k. The sum of the
12 continuous value of symbol spaced delayed samples multiplied by these coefficients
13 then gives the resulting signal.
14 Figure 2-12. Feed Forward Filter
15
16
17 rn z-1 z-1 z-1 z-1 z-1 z-1
18
19
kn-1 kn kn+1 kn+2 kn+3 kn+4
20
21
22
23
+ yn
24
25
26
27 2.B.7.1 Annex - Time Continuous Zero-Pole Equalizer adaption
28
29 The pole-zero algorithm takes the SDD21 magnitude response for the through channel
30 and inverts it to produce a desired CTE filter response curve. From a set of initial
31 conditions for 3 poles and 3 zeros, the squared differences are minimized between the
32 CTE response and the inverse channel response curve. The minimization is done
33 using a simplex method, specifically the Nelder-Mead Multidimensional Unconstrained
34 Non-Linear Minimization Method. The Nelder-Mead method provides a local
35 minimization of the square of the difference between the two curves by descending
36 along the gradient of the difference function. Once the optimization result is obtained, it
37 is compared to a specified threshold. If the threshold exceeds the target tolerance, an
38 incrementally offset seed point is generated from a 6-dimensional grid of seed points,
39 and the process is iterated until the correct curve is obtained within the target tolerance.
40
41
2.B.8 Annex - Time Continuous Zero/Pole
42
43 The Zero/Pole Filter is defined, in the frequency domain by
44
45
46 p  z + j2f 
H  f  = ---  ------------------------
47 z  p + j2f 
48
49

78 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

and consists of a single zero, z , and single pole, p . 1

2
2.B.9 Annex - Degrees of Freedom 3
4
5
2.B.9.1 Annex - Receiver Sample Point 6
7
A receiver shall be allowed to either position the centre sampling point fully 8
independently to the signal transitions or exactly in between the mean crossover of the 9
receiver signal. 10
11
2.B.9.2 Annex - Transmit Emphasis 12
13
Transmit emphasis and receiver filter coefficients must be optimised with the defined 14
resolution to give the best achievable results. Unless otherwise stated it shall be 15
assumed that the coefficients are defined using floating point variables. 16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 79

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.C Annex - Jitter Modelling

2
3 This annex describes the theoretical background of the methodology used for jitter
4 budgeting and jitter measurement. To avoid fundamental issues with the addition of
5 jitter using the dual Dirac model through a bandlimited channel, a fundamental
6 methodology call “stateye” is defined in Annex 2.C.5, which uses only convolution of
7 the jitter distribution for the calculation of the jitter at the receiver.
8
9
10 2.C.1 Annex - High Frequency Jitter vs. Wander
11
Jitter is defined as the deviation of the signal transition from an origin, usually its mean.
12
This deviation has an amplitude and an associated spectrum. High frequency jitter is
13
defined by a 1st order high pass phase filter with a corner frequency equal to the ideal
14
CDR bandwidth. The low frequency Jitter or Wander is defined by a 1st order low pass
15
phase filter with a corner frequency equal to the bandwidth.
16
17
18 2.C.2 Annex - Total Wander vs. Relative Wander
19
20 Generation of Total and Relative Wander can be achieved using a “Common” and
21 “AntiPhase” Sinusoidal Source, where the total and relative wander are then related as
22 defined below.
23
24 A total = A common + A antiphase
25 A relative = 2A antiphase
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

80 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

By adding sinusoidal frequencies of slightly differing frequencies the maximum total 1

and relative wander is achieved at various phase relationships, Figure 2-13. 2
Figure 2-13. Generation of Total and Relative Wander 3
4
Generated Jitter 5
1 6
Relative
Total
7
0.5
8
0
9
10
-0.5 11
12
-1 13
0 2 4 6 8 10 12 14 16 18
4
14
Jitter Sources
x 10 15
0.5 16
Common
Antiphase 17
18
0 19
20
21
22
-0.5 23
0 2 4 6 8 10 12 14 16 18
4 24
x 10
25
26
27
2.C.3 Annex - Correlated vs. Uncorrelated Jitter 28
29
If a correlation exists between the amplitude of the jitter and the current, past and future
30
signal level of a data channel, this type of jitter is deemed correlated. Typically this is
31
encountered when band limitation and inter-symbol interference occurs. Due to
32
amplitude to phase conversion of the ISI, a jitter is observed which has a direct
33
correlation to the data pattern being transmitted.
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 81

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.C.4 Annex - Jitter Distributions

2
3 High frequency is traditionally measured and described using probability density
4 functions, Figure 2-14 (bottom) which describe the probability of the data signal
5 crossing a decision threshold.
6 Figure 2-14. Jitter Probability Density Function
7
8
9
10
11 Decision Level
12
13
14
15
16
17
18
19
20
Sample Error :
21
Error probability is equal to
22
23 1-area under distribution
24
25
26
27
28
29
Sample Time
30
31
32 The low probability part of the jitter distribution can be described by two components,
33 mathematically described below.
34
35
2.C.4.1 Annex - Unbounded and Bounded Gaussian Distribution
36
37 We define a Unbounded Gaussian distribution function in terms of sigma as below.
38
2
39 
40 – --------
-
2
1 1 2
41 GJ( ) = ----------  ---  e
42 2 
43
44
45 For every offset  , there exists a finite and non-zero probability.
46
47
48
49

82 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.C.4.2 Annex - Bounded Gaussian Distribution 1

2
We define a Bounded Gaussian Distribution function1 in terms of sigma and a 3
maximum value as below. 4

2 5
---------2 6
1 1 2    max
7
GJ( ) = ----------  ---  e if
2     max 8
9
0 10
11
12
For random processes consisting of a finite number of random variables there exists a 13
finite non-zero probability only if    max . For example a bandlimited channel is 14
bounded but shows a Gaussian Distribution below its maximum. See Annex 2.C.4.8 for 15
an explanation concerning extrapolation. 16
17
18
2.C.4.3 Annex - High Probability Jitter 19
20
We define a dual Dirac distribution function for a High Probability jitter (W) as below
21
W W 22
( – -----) ( + -----)
2 2 23
HPJ( W) = -------------------- + --------------------- 24
2 2
25
26
2.C.4.4 Annex - Total Jitter 27
28
We define the convolution of the High Probability and Gaussian jitter as being the total 29
jitter and define it as below. 30
31
W 2 W 2 32
   – -----    + -----
 2  2 33
– ------------------------
2
– ------------------------
2
- 34
1 1 2 2
TJ( W ) = --------------  ---  e +e 35
2 2  36
37
38
39
40
41
42
43
44
45
46
47
1. Due to its bounded nature the function does not comply with the requirement that the integral of the PDF from minus infinity to
48
infinity is one. This small inaccuracy is recognized and accepted in this context. 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 83

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.C.4.5 Annex - Probability Distribution Function vs. Cumulative Distribution

2 Function
3
4 An example of the convolution of GJ (magenta), HPJ (green) to give TJ (red) can be
5 seen Figure 2-15. When integrating the probability distribution functions, same colours,
6 we obtain the cumulative distribution function or half the bathtub, Figure 2-16.
7 Figure 2-15. Example of Total Jitter PDF
8
9 Probability Distribution Function of Convolution of Basic Elements
10 0.18

11 0.16
12
13 0.14

14 0.12
15
16 0.1
Probabiliy

17 0.08
18
19 0.06

20 0.04
21
22 0.02

23 0
24 -0.1 -0.05 0 0.05 0.1 0.15
25 Time (UI)
26
27 Figure 2-16. Example of Total Jitter CDF
28
29 Cumulative Distribution Function of Convolution of Basic Elements
0
30 10

31
32 -1
10
33
34
35 10
-2

36
Probabiliy

37
38 10
-3

39
40
41 10
-4

42
43
44 -0.1 -0.05 0 0.05 0.1 0.15
45 Time (UI)

46
47
48
49

84 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.C.4.6 Annex - BathTub 1

2
Given a measured bathtub curve consisting of measured BER for various sampling 3
offsets, the defined Gaussian and High Probability Distributions can be used to 4
describe the important features of the distribution. 5
6
Initially the BER axis should be converted to Q as defined below, e.g. a BER of 10-12 is 7
a Q=7.04, and a BER of 10-15 a Q=7.94.1 8
–1 9
Q = 2  erf  2   1 – BER  – 1  10
11
12
where 13
–1 14
erf  x  is the inverse function of the error function erf  x  and 15
z 16
2
2 –t 17
erf(z) = -------  e dt
  18
19
0
20
Note: this conversion from BER to Q is only valid given a large time offset from the 21
optimal sampling point. The use of the nomenclature BER in this reference should 22
therefore be carefully used. Any accurate prediction of the BER towards the centre of 23
the eye should be done using Marcum’s Q function, and is outside the scope of this 24
document. 25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
1. It is assumed that when measuring the jitter bathtub that the left and right parts of the bathtub are independent to each other,
48
e.g. the tail of the right hand part of the bathtub and negligible effect on the left hand side of the bathtub. 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 85

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 By linearising the bathtub, Figure 2-17, we can describe the function of the left and
2 Figure 2-17. Bathtub Definition
3
4 arbitrary origin arbitrary origin
5 tauoffset plus 1UI
6
7 HPJleft
8 HPJright
9 Q=0
10
11
12
13
14

Qleft   offset  HPJ left  1  GJ left
15
16
17

Qright  HPJ right   offset  1  GJ right
18
19 Q=6
20
21 Q=7
22
23 GJ right  QBER
24 HPJleft
25
26
GJ left  QBER
HPJright
27
28
29
30
31
32 right hand linear parts of the bathtub in terms of an offset (HPJ) and gradient (1/GJ)
33
34 1
35 Q left( offset) =   offset – HPJ left   ---------------
36 GJ left
37 1
38 Q right( offset) =  HPJ left –  offset   ------------------
GJ right
39
40
41 The conversion to a linearised bathtub from a measurement should be calculated using
42 a polynomial fit algorithm for parts of the measurement made at low BERs or high Q.
43
44
45
46
47
48
49

86 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.C.4.7 Annex - Specification of GJ and HPJ 1

2
In Implementation Agreements the left and right hand terms are combined to give a 3
single definition as below. 4
5
HPJ total = 1 –  HPJ right – HPJ left  6
7
8
GJ total = GJ left  Q BER + GJ right  Q BER = 2Q BER  GJ rms 9
10
GJ left + GJ right 11
GJ rms = ---------------------------------------- 12
2
13
J total = GJ total + HPJ total 14
15
16
– 12 17
where Q BER is the Q for the BER of interest, e.g Q=7.04 for a BER = 10 18
19
2.C.4.8 Annex - Example of Bounded Gaussian 20
21
Assuming that the Cumulative Distribution Function of the jitter could be measured to 22
the probabilities shown, Figure 2-18 shows an example of when a jitter should be 23
classified as Correlated High Probability or Correlated Bounded Gaussian.. 24
Figure 2-18. Example of Bounded Gaussian 25
26
27
0 28
29
-1
30
31
-2
32
33
-3
34
-4
35
Q

36
-5 37
38
-6 39
40
-7 41
42
-8
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
43
Time Offset (arbituary) 44
45
46
The convolution of a true Unbounded Gaussian Jitter (green) with a Bounded Gaussian 47
Jitter (Red) can be seen (Magenta). It can be clearly seen and measured that at a Q of 48
-3 the Bounded Jitter is still Gaussian and the resulting convolution can be calculated 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 87

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 using RMS addition. Below a Q of -5 the Bounding effect can be seen, and if we
2 linearize the Bathtub we measure a non-zero High Probability Jitter and Gaussian
3 component.
4
5 2.C.5 Annex - Statistical Eye Methodology
6
7 The following section describes the fundamental underlying the StatEye methodology.
8 For a golden implementation please refer to the scripts on the OIF website, which are
9 published separately, and to the appropriate appendix in this document for the
10 compliance template.
11
12
13 2.C.5.1 Annex - Derivation of Cursors and Calculation of PDF
14
15 The Statistical Eye Methodology uses a channel pulse response and crosstalk pulse
16 response in conjunction with a defined sampling jitter to generate an equivalent eye
17 which represents the eye opening as seen by the receiver for a given probability of
18 occurrence.
19 Figure 2-19. Statistics of Pulse Response Cursor
20
21
22
23
24 Each possible amplitude A   dn rn
is the convolution of n
25 the data stream ndwith
26 the cursors rn d  {1,1}
27
28
29 Probability
30
Amplitude

31
32
33
34
35
36
37
38
39 All equal probable combination of cursors
40
41
42 Given a pulse response (black left), Figure 2-19, we locate c at an arbitrary point (red
0
43 arrow), and measure the symbol space cursors (blue arrows)
44
45 Given a DFE the post cursors should be adjusted by negating the measured post
46 cursors by the appropriate static coefficient of the DFE, up to the maximum number of
47 cursors specified.
48
49

88 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

According to the exact data pattern these cursors superimpose to Inter-symbol 1

Interference. Each possible combination of these cursors is calculated and from these 2
combinations a histogram is generated to form the probability density function (PDF) 3
(green). 4
Figure 2-20. Variation of the c0 sampling time 5
6
7
8
Pulse Response
Additional blue cursors
for each sampling point
9
are removed to avoid 10
confusion 11
12
13
14
15
16
17
18
19
pdf(Amplitude) 20
Signal Amplitude
21
22
23
24
By varying the reference sampling point for c0, Figure 2-20, the previous function is 25
repeated and family of conditional PDFs build up, which can be represented 26
mathematically below. 27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 89

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Given,
2
3 r n    are the cursors of the pulse response at sampling 
4
5 e b is the ideal static equalization coefficients of the b tap DFE
6
7 c    is the set of equalization cursors at sampling 
8 –1
9     = lim  x is the Dirac or delta function
10 0
11 d n b are all the possible combinations of the data stream and is either 1 or 0
12
13
p(ISI ) is the probability density function of the ISI for a given sample time
14
15
c    = r– m () ... r – 1() r 1() – e 1 ... r b() – e b r b + 1() ... r m()
16 ---- ----
17 2 2
18
19
20 d 1 1 d 1 ... d 1 m
21
22 d = d ... 1 d ... ... d ... m
23
24 d m d m d m
2 1 2  ... 2 m
25
26
27
b–1
28
29
n =  d n b  2 +1
30 b = [1,m]
31
32
33
34 1
35
36
p(ISI ) = ------
2
m    c      2d n' – 1  – ISI 
m
37 n = [1,2 ]
38
39 A similar family of PDFs are generated for the crosstalk pulse response and any other
40 aggressors in the system using the cursor set below, noting that the entire pulse
41 response is used
42
43 () ... r – 1() r 0() r 1() ... r m()
44 c    = r–m
---- ----
45 2 2
46
47
48
49

90 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.C.5.2 Annex - Inclusion of Sampling Jitter 1

2
In a real system the sampling point c0 is defined by the CDR and is jittered, for the sake 3
of standardization, by the transmitter. This jitter has a probability density function which 4
is centred at the receiver CDR sampling point and defines the probability of each of the 5
previous conditional PDFs occurring1. 6
Figure 2-21. Varying the Receiver Sampling Point 7
8
9
10
Pulse Response 11
12
13
14
15
16
17
18
Signal 19
Amplitude 20
pdf(Amplitude)
21
22
Joint 23
Distribution
pdfs 24
25
26
Zero
Signal
27
line
Amplitude 28
29
30
31
pdf(Transmit Jitter)
32
33
34
35
36
pdf(Amplitude)
37
38
39
By multiplying each of the conditional PDFs by its associated sampling jitter probability 40
and summing their results together, the joint probability density function at the given 41
receiver CDR sample point can be calculated, Figure 2-21. 42
43
44
45
46
47
48
1. Currently DCD effects are not taken into account 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 91

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Given,
2
3 p jitter( w ) is the dual Dirac probability density function of the sampling jitter in
4 the system, as defined in Annex 2.C.4.4
5
6 p crosstalk(ISI ) is the probability density function of the crosstalk
7
8 p forward(ISI ) is the probability density function of the ISI of the forward channel
9
10 a  b is the convolution operative
11
12 p average(ISI ) =
13
14 
15
16
17    pcrosstalk(ISI  +  + w)  pforward(ISI  + )   pjitter( w )  d
18 –
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

92 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.C.5.3 Annex - Generation of Statistical Eye 1

2
By varying the receiver CDR sampling point a new joint probability density function, 3
Figure 2-21 can be generated. 4
Figure 2-22. Generation of the Data Eye and Bathtub 5
6
Signal 7
Amplitude 8
9
pdf(Amplitude)
10
11
12
Joint
Distribution
13
pdfs 14
15
16
17
18
Zero 19
line
20
21
22
Signal
Amplitude 23
24
25
26
27
28
29
30
31
Zero Zero line 32
line 33
34
35
cdf(Jitter)
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 93

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 By integrating the Joint Probability Density Function to give the Cumulative Distribution
2 function, and creating a contour plot an equivalent of the receiver eye can be generated
3 which shows the exact probability of obtaining a given amplitude, Figure 2-22, this
4 equivalent eye is termed the statistical eye, Figure 2-23
5
6 By only plotting the probability against time by cutting the statistical Eye along the
7 decision threshold axis, a bathtub of the jitter can be generated,Figure 2-22.
8 Figure 2-23. Statistical Eye
9
10 Arx(tsample, Q)
11
12
13 Q=5
14 Q=6
15 Q=7
16 Q=8
17
18 teye
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

94 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.D Annex - Definition of CEI Test Patterns 1

2
3
2.D.1 Annex - PRBS31 4
The pattern is a free running PRBS31 polynomial in accorance with [21.]. The sequence is 5
generated using taps 28 and 31.
6
7
8
2.D.2 Annex - Short Stress Pattern Random (SSPR) 9
10
The SSPR pattern was chosen to have baseline wander and timing content that are at 11
least as stressful as 10,000 years of random binary. 12
• The baseline wander was assessed with a cut-off frequency of baudrate/10,000. 13
14
• The clock content was assessed with a corner frequency of baudrate/1667. 15
• The period of 10,000 years was chosen on the basis of random binary exceeding 16
the baseline wander timing content limits of the short pattern once in 10 years in a 17
network containing 1000 random streams. 18
19
The SSPR pattern is defined as: 20
21
Figure 2-24. Short Stress Pattern Random (SSPR)
22
23
PRBS28 CID PRBS28 PRBS28 PRBS28 CID PRBS28 PRBS28 24
Seed=0080080 1, 72 x 0 Seed=FFFFFFF Seed=0080080 Seed=0080080 0, 72 x 1 Seed=FFFFFFF Seed=0080080
Diff encoded Diff encoded
25
26
5437 bits 73 bits 5437 bits 5434 bits 5437 bits 73 bits 5437 bits 5434 bits
27
28
• Total length 32,762 bits 29
30
• All 228-1 PRBS28 sequences are generated using taps 25 and 28 31
• Block 1 is 5437 bits of PRBS28 seed = 0x0080080 and begins with 8 x 0, 1, 11 x 0, 32
1, 12 x 0, 1 ... 33
34
• Block 2 is 1 followed by 72 x 0 35
• Block 3 is 5437 bits of PRBS28 seed = 0xFFFFFFF and begins with 28 x 1, 25 x 0, 36
3 x 1, 22 x 0 ... 37
38
• Block 4 takes the same sequence as block 1 (omitting the last 3 bits) and encodes it 39
as follows: 40
— A zero causes a change of output 41
42
— A one causes no change of output 43
44
— The output before the first bit is assumed to have been zero
45
— This block begins 1010101001010101010110101010101011011010 ... 46
47
• Blocks 5 to 8 are the inverse of blocks 1 to 4 respectively. 48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 95

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Under some circumstances (e.g. to accommodate the restrictions of some pieces of

2 test equipment) it may be desirable to modify this short pattern to have a total length of
3 32,768 bits (215) rather than 32,762 bits. To make use of this option, the differentially
4 encoded blocks (blocks 4 and 8) should be extended by 3 bits making these blocks
5 5437 bits long.
6
7 2.D.3 Annex - Short Stress Pattern SDH 16 (SSPS-16)
8
9 The SSPS-16 pattern was chosen to have baseline wander and timing content that are
10 at least as stressful as 10,000 years of STM-16 framed random binary.
11
12 • The baseline wander was assessed with a cut-off frequency of baud/10,000.
13 • The clock content was assessed with a corner frequency of baudrate/1667.
14
15 • The period of 10,000 years was chosen on the basis of STM-16 framed random
16 binary exceeding the baseline wander and timing content limits of the short pattern
17 once in 10 years in a network containing 1000 STM-16 framed streams.
18
19 The SSPS-16 pattern is defined as:
20 Figure 2-25. Short Stress Pattern SDH 16 (SSPS-16)
21
22 A1 A2 NU PRBS28 CID PRBS28 PRBS28 A1 A2 NU PRBS28 CID PRBS28 PRBS28
23 F6 28 AA Seed 1, 72 0’s Seed Seed 09 D7 55 Seed 0, 72 1’s Seed Seed
24 0080080 FFFFFFF 0080080 0080080 FFFFFFF 0080080
Diff. enc. Diff. enc.
25 384 384 258 5095 73 5095 5092 384 384 258 5095 73 5095 5092
26 bits bits bits bits bits bits bits bits bits bits bits bits bits bits
27
28 • Total length 32,762 bits
29
30 • All 228-1 PRBS28 sequences are generated using taps 25 and 28
31
• Block 1 is A1 (11110110) repeated 48 times to give 384 bits
32
33 • Block 2 is A2 (00101000) repeated 48 times to give 384 bits
34
35 • Block 3 is the National Use bits and consists of 1010 repeated for 258 bits
36 • Block 4 takes 5095 bits of PRBS28 seed = 0x0080080 and encodes it as follows:
37
38 — A zero causes a change of output
39 — A one causes no change of output
40
41 — The output before the first bit is assumed to have been zero
42 — This block begins 1010101001010101010110101010101011011010 ...
43
44 • Block 5 is 1 followed by 72 x 0
45 • Block 6 is 5095 bits of PRBS28 seed = 0xFFFFFFF and begins 28 x 1, 25 x 0, 3 x 1,
46 22 x 0 …
47
48 • Block 7 is 5092 bits of PRBS28 seed = 0x0080080 and begins with 8 x 0, 1, 11 x 0,
49 1, 12 x 0, 1 …

96 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

• Blocks 8 to 14 are the inverse of 1 to 7 respectively. 1

2
Under some circumstances (e.g. to accommodate the restrictions of some pieces of 3
test equipment) it may be desirable to modify this short pattern to have a total length of 4
32,768 bits (215) rather than 32,762 bits. To make use of this option, the last block in 5
each half (blocks 7 and 14) should be extended by 3 bits making these blocks 5095 bits 6
long. 7
8
2.D.4 Annex - Short Stress Pattern SDH 64 (SSPS-64) 9
10
The SSPS-64 pattern was chosen to have baseline wander and timing content that are 11
at least as stressful as 10,000 years of STM-64 framed random binary. 12
13
• The baseline wander was assessed with a cut-off frequency of baud/10,000. 14
• The clock content was assessed with a corner frequency of baudrate/1667. 15
16
• The period of 10,000 years was chosen on the basis of STM-64 framed random 17
binary exceeding the baseline wander and timing content limits of the short pattern 18
once in 10 years in a network containing 1000 STM-64 framed streams. 19
20
The SSPS-64 pattern is defined as: 21
Figure 2-26. Short Stress Pattern SDH 64 (SSPS-64) 22
23
A1 A2 NU PRBS28 CID PRBS28 PRBS28 A1 A2 NU PRBS28 CID PRBS28 PRBS28 24
F6 28 AA Seed 1, 72 0’s Seed Seed 09 D7 55 Seed 0, 72 1’s Seed Seed 25
0080080
Diff. enc.
FFFFFFF 0080080 0080080
Diff. enc.
FFFFFFF 0080080 26
1536 1536 1026 4071 73 4071 4068 1536 1536 1026 4071 73 4071 4068 27
bits bits bits bits bits bits bits bits bits bits bits bits bits bits 28
29
• Total length 32,762 bits 30
31
• All 228-1 PRBS28 sequences are generated using taps 25 and 28 32
• Block 1 is A1 (11110110) repeated 192 times to give 1536 bits 33
34
• Block 2 is A2 (00101000) repeated 192 times to give 1536 bits 35
36
• Block 3 is the National Use bits and consists of 1010 repeated for 1026 bits
37
• Block 4 takes 4071 bits of PRBS28 seed = 0x0080080 and encodes it as follows: 38
39
— A zero causes a change of output 40
— A one causes no change of output 41
42
— The output before the first bit is assumed to have been zero 43
• This block begins 1010101001010101010110101010101011011010 ... 44
45
• Block 5 is 1 followed by 72 x 0 46
• Block 6 is 4071 bits of PRBS28 seed = 0xFFFFFFF and begins 28 x 1, 25 x 0, 3 x 1, 47
22 x 0 … 48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 97

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • Block 7 is 4068 bits of PRBS28 seed = 0x0080080 and begins with 8 x 0, 1, 11 x 0,

2 1, 12 x 0, 1 …
3
4 • Blocks 8 to 14 are the inverse of 1 to 7 respectively.
5
Under some circumstances (e.g. to accommodate the restrictions of some pieces of
6
test equipment) it may be desirable to modify this short pattern to have a total length of
7
32,768 bits (215) rather than 32,762 bits. To make use of this option, the last block in
8
each half (blocks 7 and 14) should be extended by 3 bits making these blocks 4071 bits
9
long.
10
11
12 2.D.5 Annex - Use of CEI Test Patterns
13
14 The Test patterns required for the various electrical interfaces covered by CEI are
15 specified in Table 2-1.
16 Table 2-1. Use of CEI Test Patterns
17
18 Test Patterns
19 Electrical Requirement “Method” IA Data Mandatory Recommended
20
SFI-4.2 Scrambled PRBS31 or SSPR
21
22 Other Scrambled PRBS31 or SSPR
23 CEI Clause 4 (SxI-5) A SPI-5 Scrambled PRBS31 or SSPR
24 SFI-5.1 Partially scrambled PRBS31 or SSPR SSPR
25 SFI-5.1s Partially scrambled PRBS31 or SSPR SSPS-16
26 Scrambled PRBS31 or SSPR
27 CEI Clause 5 (TFI-5) B TFI-5
Partially scrambled PRBS31 or SSPR SSPS-16
28
29 TDM-P Scrambled PRBS31 or SSPR

30 CEI Clause 6 (CEI-6G-SR) B

CEI-P Scrambled PRBS31 or SSPR
31 Other Scrambled PRBS31 or SSPR
32 Other Partially scrambled PRBS31 or SSPR SSPS-16
33 TDM-P Scrambled PRBS31 or SSPR
34
CEI-P Scrambled PRBS31 or SSPR
35 CEI Clause 7 (CEI-6G-LR) D
Other Scrambled PRBS31 or SSPR
36
37 Other Partially scrambled PRBS31 or SSPR SSPS-16
38 TDM-P Scrambled PRBS31 or SSPR
39 CEI-P Scrambled PRBS31 or SSPR
40 CEI Clause 8 (CEI-11G-SR) E SFI5.2 Scrambled PRBS31 or SSPR
41 Other Scrambled PRBS31 or SSPR
42
Other Partially scrambled PRBS31 or SSPR SSPS-64
43
44 E for 11G-
MR
TDM-P Scrambled PRBS31 or SSPR

45 CEI-P Scrambled PRBS31 or SSPR

46 CEI Clause 9 (CEI-11G-LR/MR) C&D without Other Scrambled PRBS31 or SSPR
47 Tx emphasis
Other Partially scrambled PRBS31 or SSPR SSPS-64
for 11G-LR
48
49

98 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 2-1. Use of CEI Test Patterns 1

Test Patterns
2
3
Electrical Requirement “Method” IA Data Mandatory Recommended 4
PRBS9 for DDJ & 5
CEI Clause 10 (CEI-28G-SR) C Other Scrambled Tx emphasis, 6
otherwise SSPR
7
CEI Clause 11 (CEI-25G-LR) D Other Scrambled
PRBS9 for DDJ &
Tx emphasis,
8
otherwise SSPR 9
PRBS9 for eye, 10
CEI Clause 13 (CEI-28G-VSR) - Other Scrambled
otherwise PRBS31 11
PRBS9 for DDJ & 12
CEI Clause 14 (CEI-28G-MR) C Other Scrambled Tx emphasis, 13
otherwise SSPR
14
CEI Clause 16 (CEI-56G-VSR-PAM4) - Other Scrambled
For PAM-4 see
Appendix 16.C.3
15
16
For PAM-4 see
CEI Clause 17 (CEI-56G-MR-PAM4) - Other Scrambled
Appendix 16.C.3 17
PRBS9 for DDJ,
18
CEI Clause 18 (CEI-56G-USR-NRZ) E Other Scrambled
otherwise SSPR 19
PRBS9 for DDJ, 20
CEI Clause 19 (CEI-56G-XSR-NRZ) E Other Scrambled
otherwise SSPR 21
CEI Clause 21 (CEI-56G-LR-PAM4) - Other Scrambled
For PAM-4 see 22
Appendix 16.C.3 23
CEI Clause 22 (CEI-56G-LR-ENRZ) - Other Scrambled
For ENRZ see 24
Appendix 22.A.5
25
26
2.D.6 Annex - Text Definitions of Patterns 27
28
Below are definitions of the patterns described in Annex 2.D.2, Annex 2.D.3 and Annex 29
2.D.4 as hexadecimal digits with the most significant bit of each digit transmitted first. 30
Since these patterns are 32,762 bits long (which is not divisible by 4), the two least 31
significant bits of the last digit shown are not included in the sequence. 32
33
Short Stress Pattern Random (SSPR) 34
35
008008004804802082081249248800000C8000068800032C8001A48800C80C8068868832CB2C
9A49248480000A080005A480028808016C8480A08A085A4DA4A882081EC9248E88000FAC8007 36
2C8803E48C81CC0E88FAC7ACF2CFACB64B2C90412481248008800804C8048228820936C92408 37
0800448480260A081165A489A4880C480C86E0868B1E4B2D3EC1244C8C80628E88376FAC98C1 38
2C85EC848AA88A0DFECDA6208A09724DA42E020855E124AFAE801D2DA80F440E875647ABDE47 39
AF52C7AD5C4FAC7BE32CFA4FA4B283281269A68812492C8800048C80020E880127AC8081AC88
48CC8CA0EA8E9A7BEFA49A4928048006820803292481A60008C96004E826023A91613EE1A68C 40
9EC92E868805AB2C828F24896F600C214606D2B7630470C7B27F6FA218412B2DA48724080BE0 41
44854E260AF3F165D6C7A4B60FA810672E8935E5AC09AC8CC44C8EAE628FBDD76F2536C160C8 42
0CA668869952CB253C4920CDE0026A2E0113B5E098E1AE45FECDC6A08A3F3A4DBC6E8205F1A9
22A7CE03799BE18B544EDD1F63835E47F99AC78354CFB99F2B27566721DE55E2F2CFAF564B2D 43
5E41247AC4807ACE083ACBE49EC94C068832C32C9A4DA48482080A092485A4000A884005ECA4 44
02A898417EC5A4A88E881ECFAC8E8B2C8FAD248F2C400F64E407443C43D65DE5D64B2CB64124 45
9044800126080081648048A40820D8449261A60016C9600A082605A491628801A76C80C9C088
687C4CB2B9E292776F601CC1460FACB7672C90C5E4816EAC08A1BCC4DAC5AE20CE8DF26BAE26 46
136DF1688027A2C811AB4889CF10CC7B796AFA0B23D2A523D478C3D77BEDD6CA483609809864 47
5845B469A681724928AE0006DDE003032E01B1A5E0C3C8AE6DD8DDD031E3351BEFA9DC492E73 48
E005DECE02B28BE1726D4EAE1073BDE93EE52A0C9CC7A687AFA92BAD2E076C45E3C400000000 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 99

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 0000000003FFFFFFC000001C00000FC000071C0003FFC001C01C00FC0FC071C71C3FFFFFDC00
2 0013C00008DC0004E3C0023FDC013C13C08DC8DC4E38E3E3FFFFCFC0001B1C000C3FC006DC1C
0303CFC1B1DB1CC3F03FADC71C2C3FFFD4DC001723C00AE3DC05DFD3C2B214DD722B236E3723
3 81F8E3F8E3FFC3FFC01DC01C0F3C0FC76DC71FC03FFE1C1C00EFCFC0791B1C3B1C3FDE3FDC12
4 FC13C851C8D8ADF8E1DC23FEF3D3C096D4DC420723E523E3CCC3CFDAADDB10FC303971DB1F2F
5 F03E65071CD4D3FFA724C029E02C166E14CA51EB298DEB265E2B214AF722B1D4E373F73F8EC4
EC3F8E38DC3FFFE3DC000FD3C00714DC03FB23C1C223DCFD33D3B14AD4E3B1C73FE3FFEC0FC0
6 08C71C04EFFFC239001D3F100F4C790752FB13DC5238D3EC3FE4C8DC0C28E3C6D6FFDF061012
7 7369081E82148EA92B0FBE073724E3E8E03FCAFE1C19D0EFCD75791A6DEB1C902B3F81172C38
8 9AE4DFC4DC221E23D32EF3D4A596D7189206FDC023113C13B98DC8E75E38FDDAFFF130D0078B
6503BD04D1E55225ECFC30A8B1DB5ED3F01A84C70CEA2FF6BBB5043661D25856F409AE15444D
9 EBF6622B445737166EE8FA519AF28DD4D66E372651F8E14DE3FEB22FC0B2351C5239DFEC3F72
10 08DC4E24E3E3F03FCFC71C1B1FFFCC3E001ADCE00CC3BE06ADE4E33C2C3FADD4DC2C3723D4D8
11 E3D721FFD6E2E0161F5E0A6E5AE591C8DC91F8E381E3FFF8EFC003F91C01C31FC0FDBE1C7104
12 EFFF923900303F101B1C790C3FFB16DC023A03C13EA1DC8CBAF38E96D6FFA206102B23691723
821AE3F92CDFC304A21DB21B2F022C257134D0EF8925793C00EB0DC07B363C3A287DDEB6B932
13 B0370A7318F59EBDF596B5259231C0903BFC411E41E49EC4EC068E38C32FFFEDA5000808D004
14 84E5020A3CD125BDA5808508984AD4C5A1C72E8AFFE5ADD00C8C35068ED9D32F8174A538AD18
15 D54AAB555B55B552B52B56DB6DB4AAAAABB55555CB555513B555724B5544ABB55CB5CB513B13
B724924A4AAAAB2B55558DB55534AB5563B5B54D4B2B58DB8DB34AD4A83B6DB434AAABF3B555
16 EC4B55024BB57BABCB40C5F3BEC40C4E24EC495A924A924AAB4AAB55BB55B52CB52B6E3B6DAB
17 54AAA5B5B5522B2B569D8DB48DB4ABA4ABB5C2B5CB17DB13903A925BB44AA2CBCB51E3F3B745
18 6C4A7C4A4B334B2B803B8DD2B4D49EDB8DAC2AD4A67D6DB3F3CAA86C73541A5435E625F307DA
0C193A0C67A40C5F02EC40DBF24ECAECA9237234A92493B4AAAA4BB5552BCB556DF3B54A8C4B
19 5B444BB2BCCBC8DF03F248DB6CAA4AAA352B55136DB5722AAB449D55BCADD52F369D6FC28DCB
20 B7C493CA34AA7313B534124B63E2ABAD65D5C6CE1D15A145F1201C0D28A56CEC724A1254AB02
21 A5B59BD22B3EFE9D863A0DB1D40CA95DEC34918273AB73B445A44BCC22CBF079E3ED91C562BB
22 544DDCB5C8993B126FA492BB82AADCD3D5690E7D48BD73DA7FC47A32B4D010DB8FA3CAD58173
6D32042AE088FD74E65BC797E2F51825FF773A0AE6440F77ECEDE6221287D982C13BF3E624EC
23 67DA925F3A4AA0C42B50C4FDB7C49BAA34AEC513B724724A4AD4AB2B6DB58DAAAB34A55583B2
24 553348A5603A724CB434A83BF3B434EC4BF3924BEC5AABE24255E5AFA506278279DB13B1DA92
25 495A4AAA922B554A9DB55B4DAB52B8A5B6DD722AA9C49D54D4ADD58DB69D34AA8DE3B544854B
5CA05BB13082C921E3E2A84565D40C4E1DEC4945824A9C33AB4D7045B8C58C2D44347EDCF3D2
26 291C7E9CB5520D3B568CE4B48416BBA0E09CC0D4ED06CD92F9A0BAF9E0FC79C4DB51D48AB75D
27 A75A71A372356124934C2AAA387D551513D571727D4504B3DC78B879517D119703F3785B6C21
28 02AA783BD53134FD61239BCC295EF07C903D932BB7BA0DCA0C0C930C6C2A1C5A7D07FBFFBFFD
BFDBFEFBEFBF6DB6DBBFFFFF9BFFFFCBBFFFE69BFFF2DBBFF9BF9BFCBBCBBE69A69B2DB6DBDB
29 FFFFAFBFFFD2DBFFEBBFBFF49BDBFAFBAFBD2D92DABBEFBF09B6DB8BBFFF829BFFC69BBFE0DB
30 9BF19F8BB829C29869829A4DA69B7DF6DBF6DBFFBBFFBFD9BFDBEEBBEFB649B6DFBFBFFDDBDB
31 FECFAFBF74D2DBB2DBBF9DBF9BC8FBCBA70DA69609F6DD9B9BFCEB8BBE44829B39F69BD09BDB
32 AABBAF900992CEFBAFB46D92DE8FEFBD50F6DA828BFF1692BF85DF8BC54DC2A10DC28569C295
1D829C20E6982D82DA6BE6BF6CB2CBBF6DB69BBFFFDB9BFFEF8BBFF6C29BFBF29BBDB99B9AF8
33 AB8B2C2082DB2DB6BFDBFFCBEFBFE6B6DBF2CFFFB9B4FFD8BECFEE2B74F608F2CB9B09B68BCB
34 BFD2A69BEB86DBB484FF9EF5CFC96A44E7DC79C26C0482EF3DF6A692DBC6DFBFA0FDDBD58ECF
35 A86074D149C2DA4F82BF7CC68BB650D29FB29B99DD9B8A8CEB82114486D649F4F9BF9ACCBBCB
3569A6D61DB6F990FFECAE8FF76250FB38F28DD00991CAFBAE062D921C8BEFD072B6EAC18FE4
36 3320F3A55D891704E3E50DC03329C3E559823306A6C54CC6F10D50E86982854DA6950DF6DC29
37 DBFC298FBE29A0DB09B59FCBBE69E69B2D92DBDBEFBFAFB6DBD2DFFFABBDFFD09ADFEABB3DF4
38 09D2DABB8BBF09829B8BA69B8296DB869DFF84D8DFC5DE1DE14D10D14DA69A4DF6DB7DDBFFF6
CFBFFBF4DBFDBADFBEF93DDB6CF2CFFF49B4FFAFBECFD2DB74EBBFF2C49BF9B1FBBCBC1D9A6A
39 30EB6C4484FF19F5CF829A44C69B79D0DBF48A9FBAF219D929D28EF98B906CA28ECF649074BB
40 FEC2E9BF72A5BBB187799C2434A82FA6E16AD6E15C39E144209149ADBE4FB3FB3CDD3DD25CB2
41 CBF46DB6BA8FFFC910FFE7E68FF272D0F9E1BA8C9139117E70E6572082B11DB68C60FFD1098F
42 EA6BA0F46C958A8F7C6210B608D6AF9B19C2CBC282B6A2968FC49DD0E1DFFFFFFFFFFFFFFFFF
E0000001FFFFFF1FFFFF81FFFFC71FFFE001FFF1FF1FF81F81FC71C71E0000011FFFFF61FFFF
43 B91FFFD8E1FFEE011FF61F61FB91B91D8E38E0E0000181FFFF271FFF9E01FFC91F1FE7E181F2
44 712719E07E0291C71E9E0001591FFF46E1FFA8E11FD10161EA6F59146EA6E48E46E3F038E038
45 E001E001FF11FF1F861F81C491C701FE000F1F1FF88181FC37271E271E010E011F681F61BD71
B93A9038F11EE0086161FB49591DEFC6E0D6E0E199E1812A9127781E7E34712706807E0CD7C7
46 19596002C6D9FEB0FE9F4C8F59AD70A6B390A6CD0EA6F5A846EA7158E460460389D89E038E39
47 1E0000E11FFF8161FFC7591FE026E1F1EEE11816616275A958E271C600E0009F81FFB9C71FD8
48 8001EE37FF16077F859C37C56827611D6E39609E00D9B91F9EB8E1C9480107CF7F6C64B7BF0B
EF5B8AB6A7820FC646D8E0B8FE01A80F1F3178819454372C90A71B7EA603F7469E3B28D901D9
49

100 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1EEF0EE16688615AD349473B6FC811FEE7761F623391B8C50E3811280076797FC3A4D7E217D9 1
70D56ED0981E7ABA712509607F2BD9C798AE804A2257DE4CF16D3D485FB28F55DD90A04CEEA5 2
DD46474C88B82D73286B91594C8E46CD7038F590E00A6E81FA6E571D6E31009E046FB91D8ED8
E0E07E0181C71F2700019E0FFF2918FF99E20FCA90D8E61E9E0291591E9E46E15938E146F001 3
48E8FF4F050FAC8D28D371B91B7038E3F0E0003881FFE0371FF1E701F8120F1C77D880036E37 4
FE7E077F271C379E0027491FEE2FE1F60AF11B9A28638B494802EFCF7EA6E4B746E3EF28E036 5
9901E7DAEF126F2687EE9ED476597883B6D4361FF8A791FC264E1E2EBC110A4A366A7E47AC67
38530A10534A56D36E71FB7E201DF70DF0DB09D89FCB8E39E6800092D7FFBFB97FDBD8D7EFAE 6
1976D212D3FBD7BB3DA959D2F1C68BA800D2917F9B9E57CB89316683F45AD63A973955AAA555 7
2552556A56A54924925AAAAAA2555551A555576255546DA555DAA2551A51A576276246DB6DAD 8
AAAAA6A555539255565AA554E2525595A6A539239265A95ABE24925E5AAAA06255509DA557ED
A25422A1A5F9D06209DF9D8ED89DB52B6DAB6DAAA5AAA5522552569A56A48E2492A55AAAD252 9
556EA6A54B13925B925AA2DAA251EA51A74127637E2B6D225DAAE9A1A570E06245D49DAC1DAD 10
A665A6A3FE23916A595B0923929EA95ACC14926061AABC9C655F2D5E50CED067C12F9F362F9C 11
C2DF9D07E89DF9206D89A89AB6E46E5AB6DB625AAAADA25556A1A554906255AB9DA525DDA26A 12
19A1B907E06DB9249AADAAAE56A5576492546EAAA5DB15521A9156864B1481EB91A241DB61AE
5AAC6762565F6DA4E0EAA294D151C98F1752F5D076FF1F96BAD4989C6DAF6D5AA7EAD253216E 13
A6080B13CE2F92715F9AB5109E5B73EC62A4625DD2DDA19EE9A07C30E09371D4EA255D911A51 14
BB36276C82DB6A23EAA9196154B78C15BA14612C01DC2E6A597F792382E1A953F464966FDEA8 15
FBB81458CD21C340E8573ED004462FA8CDDF84409890CEEF6BC133E9F62060CED89CC12B6D06
2DAAF9DEA579D81241DB22AE5A89D76246DC6DADA95AA6A4925392AAA65AD553E26D5665BAD4 16
FE2C6D9A5E5ABE20625E589DA0636DA09D2AA0EDED50D282D7CEC3EC312762712B6DB52DAAAB 17
6EA555AB12552592A56A3AD249146EAAB1DB15595A9153924B165AAB90E255DBD5A51AFD2276 18
7BE9B6F0E0EABDD4D15F9D8F109DB5D3EDAB1E62A5947DD239D39E95DE5C0918616EB71C0B1A
556F96254B98DA5BDF4A22F8FB19F95897C9936832FA2830F81C31D925715BAA4512C52C72E4 19
6E54F6DB659EAAAE3C1557576154746C15D7DA611C3A3C3574177347E0643D249EF7EAAC3E21 20
56765814F6E3219EB5087C1B7E1366A2422F91AF9F9B679C9EAF1D2C17C 21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 101

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Short Stress Pattern SDH 16 (SSPS-16)

2
3 F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
F6F6F6F6F6F6F6F6F6F628282828282828282828282828282828282828282828282828282828
4 2828282828282828282828282828282828282828AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
5 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9556AAB6AB6AA56A56ADB6DB695555576AAAAB96AAAA27
6 6AAAE496AA89576AB96B96A276276E49249495555656AAAB1B6AAA6956AAC76B6A9A9656B1B7
1B6695A95076DB68695557E76AABD896AA04976AF75796818BE77D88189C49D892B524952495
7 569556AB76AB6A596A56DC76DB56A9554B6B6AA45656AD3B1B691B695749576B856B962FB627
8 207524B76895459796A3C7E76E8AD894F8949666965700771BA569A93DB71B5855A94CFADB67
9 E79550D8E6A834A86BCC4BE60FB418327418CF4818BE05D881B7E49D95D95246E46952492769
5554976AAA5796AADBE76A951896B6889765799791BE07E491B6D95495546A56AA26DB6AE455
10 56893AAB795BAA5E6D3ADF851B976F8927946954E6276A682496C7C5575ACBAB8D9C3A2B428B
11 E240381A514AD9D8E49424A956054B6B37A4567DFD3B0C741B63A81952BBD8692304E756E768
12 8B4897984597E0F3C7DB238AC576A89BB96B91327624DF4925770555B9A7AAD21CFA917AE7B4
13 FF88F46569A021B71F4795AB02E6DA640855C111FAE9CCB78F2FC5EA304BFEEE7415CC881EEF
D9DBCC44250FB3058277E7CC49D8CFB524BE7495418856A189FB6F893754695D8A276E48E494
14 95A95656DB6B1B5556694AAB0764AA66914AC074E49968695077E76869D897E72497D8B557C4
15 84ABCB5F4A0C4F04F3B62763B52492B495552456AA953B6AB69B56A5714B6DBAE45553893AA9
16 A95BAB1B6D3A69551BC76A890A96B940B76261059243C7C5508ACBA8189C3BD8928B04953867
569AE08B718B185A8868FDB9E7A45238FD396AA41A76AD19C969082D7741C13981A9DA0D9B25
17 F34175F3C1F8F389B6A3A9156EBB4EB4E346E46AC2492698555470FAAA2A27AAE2E4FA8A0967
18 B8F170F2A2C000000000000000003FFFFFFC000001C00000FC000071C0003FFC001C01C00FC0
19 FC071C71C3FFFFFDC000013C00008DC0004E3C0023FDC013C13C08DC8DC4E38E3E3FFFFCFC00
01B1C000C3FC006DC1C0303CFC1B1DB1CC3F03FADC71C2C3FFFD4DC001723C00AE3DC05DFD3C
20 2B214DD722B236E372381F8E3F8E3FFC3FFC01DC01C0F3C0FC76DC71FC03FFE1C1C00EFCFC07
21 91B1C3B1C3FDE3FDC12FC13C851C8D8ADF8E1DC23FEF3D3C096D4DC420723E523E3CCC3CFDAA
22 DDB10FC303971DB1F2FF03E65071CD4D3FFA724C029E02C166E14CA51EB298DEB265E2B214AF
23 722B1D4E373F73F8EC4EC3F8E38DC3FFFE3DC000FD3C00714DC03FB23C1C223DCFD33D3B14AD
4E3B1C73FE3FFEC0FC008C71C04EFFFC239001D3F100F4C790752FB13DC5238D3EC3FE4C8DC0
24 C28E3C6D6FFDF0610127369081E82148EA92B0FBE073724E3E8E03FCAFE1C19D0EFCD75791A6
25 DEB1C902B3F81172C389AE4DFC4DC221E23D32EF3D4A596D7189206FDC023113C13B98DC8E75
26 E38FDDAFFF130D0078B6503BD04D1E55225ECFC30A8B1DB5ED3F01A84C70CEA2FF6BBB504366
1D25856F409AE15444DEBF6622B445737166EE8FA519AF28DD4D66E372651F8E14DE3FEB22FC
27 0B2351C5239DFEC3F7208DC4E24E3E3F03FCFC71C1B1FFFCC3E001ADCE00CC3BE06ADE4E33C2
28 C3FADD4DC2C3723D4D8E3D721FFD6E2E0161F5E0A6E5AE591C8DC91F8E381E3FFF8EFC003F91
29 C01C31FC0FDBE1C7104EFFF923900303F101B1C790C3FFB16DC023A03C13EA1DC8CBAF38E96D
6FFA206102B23691723821AE3F92CDFC304A21DB21B2F022C257134D0EF8925793C00EB0DC07
30 B363C3A287DDEB6B932B0370A7318F59EBDF596B5259231C0903BFC411E41E49EC4EC068E38C
31 32FFFEDA500040040024024010410409249244000006400003440001964000D2440064064034
32 43441965964D249242400005040002D24001440400B642405045042D26D25441040F64924744
33 0007D64003964401F24640E607447D63D67967D65B2596482092409240044004026402411441
049B649204040022424013050408B2D244D244062406437043458F25969F6092264640314744
34 1BB7D64C609642F64245544506FF66D3104504B926D21701042AF09257D7400E96D407A20743
35 AB23D5EF23D7A963D6AE27D63DF1967D27D25941940934D34409249644000246400107440093
36 D64040D6442466465075474D3DF7D24D2494024003410401949240D3000464B0027413011D48
B09F70D3464F6497434402D59641479244B7B00610A303695BB1823863D93FB7D10C209596D2
37 43920405F02242A7130579F8B2EB63D25B07D4083397449AF2D604D646622647573147DEEBB7
38 929B60B064065334434CA965929E249066F00135170089DAF04C70D722FF66E350451F9D26DE
39 374102F8D49153E701BCCDF0C5AA276E8FB1C1AF23FCCD63C1AA67DCCF9593AB3390EF2AF179
67D7AB2596AF20923D62403D67041D65F24F64A603441961964D26D242410405049242D20005
40 442002F65201544C20BF62D25447440F67D647459647D69247962007B27203A21E21EB2EF2EB
41 25965B20924822400093040040B240245204106C224930D3000B64B005041302D248B14400D3
42 B64064E044343E26595CF1493BB7B00E60A307D65BB3964862F240B7560450DE626D62D71067
43 46F935D71309B6F8B44013D16408D5A444E788663DBCB57D0591E95291EA3C61EBBDF6EB6524
1B04C04C322C22DA34D340B924945700036EF001819700D8D2F061E45736EC6EE81848484848
44 4848484848484848484848484848484848484848484848484848484848484848484848484848
45 48484848484EBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBE
46 BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAB554AAA4AA4AAD4AD4A924924B5555544AAAAA34AAAAEC4AAA8DB4A
47 ABB544AA34A34AEC4EC48DB6DB5B55554D4AAAA724AAACB54AA9C4A4AB2B4D4A724724CB52B5
48 7C4924BCB55540C4AAA13B4AAFDB44A845434BF3A0C413BF3B1DB13B6A56DB56DB554B554AA4
49 4AA4AD34AD491C49254AB555A4A4AADD4D4A962724B724B545B544A3D4A34E824EC6FC56DA44

102 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

BB55D3434AE1C0C48BA93B583B5B4CCB4D47FC4722D4B2B61247253D52B5982924C0C3557938 1
CABE5ABCA19DA0CF825F3E6C5F3985BF3A0FD13BF240DB1351356DC8DCB56DB6C4B5555B44AA 2
AD434AA920C4AB573B4A4BBB44D43343720FC0DB7249355B555CAD4AAEC924A8DD554BB62AA4
3522AD0C962903D7234483B6C35CB558CEC4ACBEDB49C1D54529A2A3931E2EA5EBA0EDFE3F2D 3
75A93138DB5EDAB54FD5A4A642DD4C1016279C5F24E2BF356A213CB6E7D8C548C4BBA5BB433D 4
D340F861C126E3A9D44ABB2234A3766C4ED905B6D447D55232C2A96F182B7428C25803B85CD4 5
B2FEF24705C352A7E8C92CDFBD51F77028B19A438681D0AE7DA0088C5F519BBF088131219DDE
D78267D3EC40C19DB1398256DA0C5B55F3BD4AF3B02483B6455CB513AEC48DB8DB5B52B54D49 6
24A725554CB5AAA7C4DAACCB75A9FC58DB34BCB57C40C4BCB13B40C6DB413A5541DBDAA1A505 7
AF9D87D8624EC4E256DB6A5B5556DD4AAB5624AA4B254AD475A49228DD5563B62AB2B522A724 8
962CB55721C4ABB7AB4A35FA44ECF7D36DE1C1D57BA9A2BF3B1E213B6BA7DB563CC54B28FBA4
73A73D2BBCB81230C2DD6E381634AADF2C4A9731B4B7BE9445F1F633F2B12F9326D065F45061 9
F03863B24AE2B7548A258A58E5C8DCA9EDB6CB3D555C782AAEAEC2A8E8D82BAFB4C23874786A 10
E9FFFFFFFFFFFFFFFFFE0000001FFFFFF1FFFFF81FFFFC71FFFE001FFF1FF1FF81F81FC71C71 11
E0000011FFFFF61FFFFB91FFFD8E1FFEE011FF61F61FB91B91D8E38E0E0000181FFFF271FFF9 12
E01FFC91F1FE7E181F2712719E07E0291C71E9E0001591FFF46E1FFA8E11FD10161EA6F59146
EA6E48E46E3F038E038E001E001FF11FF1F861F81C491C701FE000F1F1FF88181FC37271E271 13
E010E011F681F61BD71B93A9038F11EE0086161FB49591DEFC6E0D6E0E199E1812A9127781E7 14
E34712706807E0CD7C719596002C6D9FEB0FE9F4C8F59AD70A6B390A6CD0EA6F5A846EA7158E 15
460460389D89E038E391E0000E11FFF8161FFC7591FE026E1F1EEE11816616275A958E271C60
0E0009F81FFB9C71FD88001EE37FF16077F859C37C56827611D6E39609E00D9B91F9EB8E1C94 16
80107CF7F6C64B7BF0BEF5B8AB6A7820FC646D8E0B8FE01A80F1F3178819454372C90A71B7EA 17
603F7469E3B28D901D91EEF0EE16688615AD349473B6FC811FEE7761F623391B8C50E3811280 18
076797FC3A4D7E217D970D56ED0981E7ABA712509607F2BD9C798AE804A2257DE4CF16D3D485
FB28F55DD90A04CEEA5DD46474C88B82D73286B91594C8E46CD7038F590E00A6E81FA6E571D6 19
E31009E046FB91D8ED8E0E07E0181C71F2700019E0FFF2918FF99E20FCA90D8E61E9E0291591 20
E9E46E15938E146F00148E8FF4F050FAC8D28D371B91B7038E3F0E0003881FFE0371FF1E701F 21
8120F1C77D880036E37FE7E077F271C379E0027491FEE2FE1F60AF11B9A28638B494802EFCF7 22
EA6E4B746E3EF28E0369901E7DAEF126F2687EE9ED476597883B6D4361FF8A791FC264E1E2EB
C110A4A366A7E47AC6738530A10534A56D36E71FB7E201DF70DF0DB09D89FCB8E39E6800092D 23
7FFDFFDFFEDFEDFF7DF7DFB6DB6DDFFFFFCDFFFFE5DFFFF34DFFF96DDFFCDFCDFE5DE5DF34D3 24
4D96DB6DEDFFFFD7DFFFE96DFFF5DFDFFA4DEDFD7DD7DE96C96D5DF7DF84DB6DC5DFFFC14DFF 25
E34DDFF06DCDF8CFC5DC14E14C34C14D26D34DBEFB6DFB6DFFDDFFDFECDFEDF75DF7DB24DB6F
DFDFFEEDEDFF67D7DFBA696DD96DDFCEDFCDE47DE5D386D34B04FB6ECDCDFE75C5DF22414D9C 26
FB4DE84DEDD55DD7C804C9677DD7DA36C96F47F7DEA87B6D4145FF8B495FC2EFC5E2A6E15086 27
E142B4E14A8EC14E10734C16C16D35F35FB65965DFB6DB4DDFFFEDCDFFF7C5DFFB614DFDF94D 28
DEDCCDCD7C55C59610416D96DB5FEDFFE5F7DFF35B6DF967FFDCDA7FEC5F67F715BA7B047965
CD84DB45E5DFE9534DF5C36DDA427FCF7AE7E4B52273EE3CE1360241779EFB53496DE36FDFD0 29
7EEDEAC767D4303A68A4E16D27C15FBE6345DB28694FD94DCCEECDC54675C108A2436B24FA7C 30
DFCD665DE59AB4D36B0EDB7CC87FF65747FBB1287D9C7946E804C8E57DD70316C90E45F7E839 31
5B7560C7F2199079D2AEC48B8271F286E01994E1F2ACC119835362A6637886A87434C142A6D3 32
4A86FB6E14EDFE14C7DF14D06D84DACFE5DF34F34D96C96DEDF7DFD7DB6DE96FFFD5DEFFE84D
6FF55D9EFA04E96D5DC5DF84C14DC5D34DC14B6DC34EFFC26C6FE2EF0EF0A68868A6D34D26FB 33
6DBEEDFFFB67DFFDFA6DFEDD6FDF7C9EEDB67967FFA4DA7FD7DF67E96DBA75DFF9624DFCD8FD 34
DE5E0ECD351875B622427F8CFAE7C14D22634DBCE86DFA454FDD790CEC94E9477CC5C8365147 35
67B2483A5DFF6174DFB952DDD8C3BCCE121A5417D370B56B70AE1CF0A21048A4D6DF27D9FD9E
6E9EE92E5965FA36DB5D47FFE4887FF3F347F939687CF0DD46489C88BF0 36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 103

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Short Stress Pattern SDH 64 (SSPS-64)

2
3 F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
4 F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
5 F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
6 F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6F6
F6F6282828282828282828282828282828282828282828282828282828282828282828282828
7 2828282828282828282828282828282828282828282828282828282828282828282828282828
8 2828282828282828282828282828282828282828282828282828282828282828282828282828
9 2828282828282828282828282828282828282828282828282828282828282828282828282828
2828282828282828282828282828282828282828282828282828282828282828282828282828
10 28282828AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
11 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
12 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
13 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA9556AAB6AB6AA56A56ADB6DB695555576AAAAB
96AAAA276AAAE496AA89576AB96B96A276276E49249495555656AAAB1B6AAA6956AAC76B6A9A
14 9656B1B71B6695A95076DB68695557E76AABD896AA04976AF75796818BE77D88189C49D892B5
15 24952495569556AB76AB6A596A56DC76DB56A9554B6B6AA45656AD3B1B691B695749576B856B
16 962FB627207524B76895459796A3C7E76E8AD894F8949666965700771BA569A93DB71B5855A9
4CFADB67E79550D8E6A834A86BCC4BE60FB418327418CF4818BE05D881B7E49D95D95246E469
17 524927695554976AAA5796AADBE76A951896B6889765799791BE07E491B6D95495546A56AA26
18 DB6AE45556893AAB795BAA5E6D3ADF851B976F8927946954E6276A682496C7C5575ACBAB8D9C
19 3A2B428BE240381A514AD9D8E49424A956054B6B37A4567DFD3B0C741B63A81952BBD8692304
E756E7688B4897984597E0F3C7DB238AC576A89BB96B91327624DF4925770555B9A7AAD21CFA
20 917AE7B4FF88F46569A021B71F4795AB02E6DA640855C111FAE9CCB78F2FC5EA304BFEEE7415
21 CC881EEFD9DBCC44250FB3058277E7CC49D8CFB524BE7495418856A189FB6F893754695D8A27
22 6E48E49495A95656DB6B1B5556694AAB0764AA66914AC074E49968695077E76869D897E72497
23 D8B557C484ABCB5F4A0C4F04F3B62763B52492B495552456AA953B6AB69B56A5714000000000
000000003FFFFFFC000001C00000FC000071C0003FFC001C01C00FC0FC071C71C3FFFFFDC000
24 013C00008DC0004E3C0023FDC013C13C08DC8DC4E38E3E3FFFFCFC0001B1C000C3FC006DC1C0
25 303CFC1B1DB1CC3F03FADC71C2C3FFFD4DC001723C00AE3DC05DFD3C2B214DD722B236E37238
26 1F8E3F8E3FFC3FFC01DC01C0F3C0FC76DC71FC03FFE1C1C00EFCFC0791B1C3B1C3FDE3FDC12F
C13C851C8D8ADF8E1DC23FEF3D3C096D4DC420723E523E3CCC3CFDAADDB10FC303971DB1F2FF
27 03E65071CD4D3FFA724C029E02C166E14CA51EB298DEB265E2B214AF722B1D4E373F73F8EC4E
28 C3F8E38DC3FFFE3DC000FD3C00714DC03FB23C1C223DCFD33D3B14AD4E3B1C73FE3FFEC0FC00
29 8C71C04EFFFC239001D3F100F4C790752FB13DC5238D3EC3FE4C8DC0C28E3C6D6FFDF0610127
369081E82148EA92B0FBE073724E3E8E03FCAFE1C19D0EFCD75791A6DEB1C902B3F81172C389
30 AE4DFC4DC221E23D32EF3D4A596D7189206FDC023113C13B98DC8E75E38FDDAFFF130D0078B6
31 503BD04D1E55225ECFC30A8B1DB5ED3F01A84C70CEA2FF6BBB5043661D25856F409AE15444DE
32 BF6622B445737166EE8FA519AF28DD4D66E372651F8E14DE3FEB22FC0B2351C5239DFEC3F720
33 8DC4E24E3E3F03FCFC71C1B1FFFCC3E001ADCE00CC3BE06ADE4E33C2C3FADD4DC2C3723D4D8E
3D721FFD6E2E0161F5E0A6E5AE591C8DC91F8E00400400240240104104092492440000064000
34 03440001964000D244006406403443441965964D249242400005040002D24001440400B64240
35 5045042D26D25441040F649247440007D64003964401F24640E607447D63D67967D65B259648
36 2092409240044004026402411441049B649204040022424013050408B2D244D2440624064370
43458F25969F60922646403147441BB7D64C609642F64245544506FF66D3104504B926D21701
37 042AF09257D7400E96D407A20743AB23D5EF23D7A963D6AE27D63DF1967D27D25941940934D3
38 4409249644000246400107440093D64040D6442466465075474D3DF7D24D2494024003410401
39 949240D3000464B0027413011D48B09F70D3464F6497434402D59641479244B7B00610A30369
5BB1823863D93FB7D10C209596D243920405F02242A7130579F8B2EB63D25B07D4083397449A
40 F2D604D646622647573147DEEBB7929B60B064065334434CA965929E249066F00135170089DA
41 F04C70D722FF66E350451F9D26DE374102F8D49153E701BCCDF0C5AA276E8FB1C1AF23FCCD63
42 C1AA67DCCF9593AB3390EF2AF17967D7AB2596AF20923D62403D67041D65F24F64A603441961
43 964D26D242410405049242D20005442002F65201544C20BF62D25447440F67D647459647D692
47962007B27203A21E21EB2EF2EB25965B20924822400093040040B240245204106848484848
44 4848484848484848484848484848484848484848484848484848484848484848484848484848
45 4848484848484848484848484848484848484848484848484848484848484848484848484848
46 4848484848484848484848484848484848484848484848484848484848484848484848484848
4848484848484848484848484848484848484848484848484848484848484848484848484848
47 48484848484848484848484848484848484848484848484848484848484848484848484EBEBE
48 BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBE
49 BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBE

104 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBE 1
BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBE 2
BEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBEBA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 3
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 4
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 5
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAB554AAA4AA4AAD4AD4A924924B5555544AAAAA34AAAAEC4
AAA8DB4AABB544AA34A34AEC4EC48DB6DB5B55554D4AAAA724AAACB54AA9C4A4AB2B4D4A7247 6
24CB52B57C4924BCB55540C4AAA13B4AAFDB44A845434BF3A0C413BF3B1DB13B6A56DB56DB55 7
4B554AA44AA4AD34AD491C49254AB555A4A4AADD4D4A962724B724B545B544A3D4A34E824EC6 8
FC56DA44BB55D3434AE1C0C48BA93B583B5B4CCB4D47FC4722D4B2B61247253D52B5982924C0
C3557938CABE5ABCA19DA0CF825F3E6C5F3985BF3A0FD13BF240DB1351356DC8DCB56DB6C4B5 9
555B44AAAD434AA920C4AB573B4A4BBB44D43343720FC0DB7249355B555CAD4AAEC924A8DD55 10
4BB62AA43522AD0C962903D7234483B6C35CB558CEC4ACBEDB49C1D54529A2A3931E2EA5EBA0 11
EDFE3F2D75A93138DB5EDAB54FD5A4A642DD4C1016279C5F24E2BF356A213CB6E7D8C548C4BB 12
A5BB433DD340F861C126E3A9D44ABB2234A3766C4ED905B6D447D55232C2A96F182B7428C258
03B85CD4B2FEF24705C352A7E8C92CDFBD51F77028B19A438681D0AE7DA0088C5F519BBF0881 13
31219DDED78267D3EC40C19DB1398256DA0C5B55F3BD4AF3B02483B6455CB513AEC48DB8DB5B 14
52B54D4924A725554CB5AAA7C4DAACCB75A9FC58DB34BCB57C40C4BCB13B40C6DB413A5541DB 15
DAA1A505AF9D87D8624EC4E256DB6A5B5556DD4AAB5624AA4B254AD475FFFFFFFFFFFFFFFFFE
0000001FFFFFF1FFFFF81FFFFC71FFFE001FFF1FF1FF81F81FC71C71E0000011FFFFF61FFFFB 16
91FFFD8E1FFEE011FF61F61FB91B91D8E38E0E0000181FFFF271FFF9E01FFC91F1FE7E181F27 17
12719E07E0291C71E9E0001591FFF46E1FFA8E11FD10161EA6F59146EA6E48E46E3F038E038E 18
001E001FF11FF1F861F81C491C701FE000F1F1FF88181FC37271E271E010E011F681F61BD71B
93A9038F11EE0086161FB49591DEFC6E0D6E0E199E1812A9127781E7E34712706807E0CD7C71 19
9596002C6D9FEB0FE9F4C8F59AD70A6B390A6CD0EA6F5A846EA7158E460460389D89E038E391 20
E0000E11FFF8161FFC7591FE026E1F1EEE11816616275A958E271C600E0009F81FFB9C71FD88 21
001EE37FF16077F859C37C56827611D6E39609E00D9B91F9EB8E1C9480107CF7F6C64B7BF0BE 22
F5B8AB6A7820FC646D8E0B8FE01A80F1F3178819454372C90A71B7EA603F7469E3B28D901D91
EEF0EE16688615AD349473B6FC811FEE7761F623391B8C50E3811280076797FC3A4D7E217D97 23
0D56ED0981E7ABA712509607F2BD9C798AE804A2257DE4CF16D3D485FB28F55DD90A04CEEA5D 24
D46474C88B82D73286B91594C8E46CD7038F590E00A6E81FA6E571D6E31009E046FB91D8ED8E 25
0E07E0181C71F2700019E0FFF2918FF99E20FCA90D8E61E9E0291591E9E46E15938E146F0014
8E8FF4F050FAC8D28D371B91B7038FFDFFDFFEDFEDFF7DF7DFB6DB6DDFFFFFCDFFFFE5DFFFF3 26
4DFFF96DDFFCDFCDFE5DE5DF34D34D96DB6DEDFFFFD7DFFFE96DFFF5DFDFFA4DEDFD7DD7DE96 27
C96D5DF7DF84DB6DC5DFFFC14DFFE34DDFF06DCDF8CFC5DC14E14C34C14D26D34DBEFB6DFB6D 28
FFDDFFDFECDFEDF75DF7DB24DB6FDFDFFEEDEDFF67D7DFBA696DD96DDFCEDFCDE47DE5D386D3
4B04FB6ECDCDFE75C5DF22414D9CFB4DE84DEDD55DD7C804C9677DD7DA36C96F47F7DEA87B6D 29
4145FF8B495FC2EFC5E2A6E15086E142B4E14A8EC14E10734C16C16D35F35FB65965DFB6DB4D 30
DFFFEDCDFFF7C5DFFB614DFDF94DDEDCCDCD7C55C59610416D96DB5FEDFFE5F7DFF35B6DF967 31
FFDCDA7FEC5F67F715BA7B047965CD84DB45E5DFE9534DF5C36DDA427FCF7AE7E4B52273EE3C 32
E1360241779EFB53496DE36FDFD07EEDEAC767D4303A68A4E16D27C15FBE6345DB28694FD94D
CCEECDC54675C108A2436B24FA7CDFCD665DE59AB4D36B0EDB7CC87FF65747FBB1287D9C7946 33
E804C8E57DD70316C90E45F7E8395B7560C7F2199079D2AEC48B8271F286E01994E1F2ACC119 34
835362A6637886A87434C142A6D34A86FB6E14EDFE14C7DF14D06D84DACFE5DF34F34D96C96D 35
EDF7DFD7DB6DE96FFFD5DEFFE84D6FF55D9EFA04E96D5DC5DF84C14DC5D34DC14B6DC34EFFC2
6C6FE2EF0EF0A68868A6D34D26FB6DBEEDFFFB67DFFDFA6DFEDD6FDF7C8 36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 105

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.E Appendix - Lab Setups

2
3 All methodology described in this Appendix is only relevant for verification of low level
4 CDR functionality, and does not cover any required tests for protocol compliance e.g.
5 deskew. The methodology is based on the assumption that either an integrated BERT
6 is present in the DUT or a loop or functionality for the attachment of external
7 equipment.
8
9
10 2.E.1 High Frequency Transmit Jitter Measurement
11
The following sub-clause describes various methods for measuring high frequency
12
jitter, which depending upon the baud rate can be applied for various levels of accuracy.
13
14
15 2.E.1.1 BERT Implementation
16
17 Referring to Figure 2-27, this sub-clause describes test methodology based on bathtub
18 extraction, which relies on equipment being available for the given baud rate.
19 Figure 2-27. BERT with Golden PLL
20
21
22 Differential
to single
23 DUT ended amp +
24 Golden PLL

25
26 Trigger
27
28
29 Signal
30 Clock BERT
Ref
31
32
33
34
35 • This same methodology can be used by equalized transmitters, by initially turning
36 the equalization off, or by performing the measurement at the output of a Stress
37 Channel .
38 • The transmitter under test shall transmit the specified data pattern, while all other
39 signals are active.
40
41 — The other channels can transmit the same pattern if they have at least a 16 bit
42 offset with the channel under test.
43 — All links within a device under test to be active in both transmit and receive
44 directions, and receive links are to use asynchronous clocks with respect to
45 transmit links (to maximum allowed ppm. offset as specified in the protocol
46 specifications).
47
48 • The data should be differentially analysed using an external differential amp or
49 differential input BERT and Golden PLL.

106 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

— Use of single ended signals will give an inaccurate measurement and should not 1
be used. 2
3
— The use of a balun will most likely degrade the signal integrity and is only 4
recommended for 3Gsym/s signalling when the balun is linear with a return loss
5
of better than -15dB until three times the baud rate.
6
• Inherent bandwidth of clock reference inputs of BERT should be verified e.g. in the 7
case of parBERTs. Additional bandwidth limitation of the BERT will lead to 8
inaccurate results. 9
10
• The use of a Golden PLL is required to eliminate inherent clock content (Wander) in 11
transmitted data signals for long measurement periods. 12
— The Golden PLL should have at maximum a bandwidth of baud rate over 1667, 13
with a maximum of 20dB/dec rolloff, until at least baud rate over 16.67, with no 14
peaking around the corner frequency. 15
16
• The output jitter for the DUT is not defined as the contributed jitter from the DUT but 17
as the total output jitter including the contributions from the reference clock. To this 18
end, the reference clock of the DUT should be verified to have a performance 19
similar to the real application. 20
• a confidence level of three sigma should be guaranteed in the measurement of BER 21
for the Bathtub as per Appendix 2.F.2.1 22
23
• The High Probability and Gaussian Jitter components should be extracted from the 24
bathtub measurement using the methodology defined in Annex 2.C.4.6. 25
• If not defined the maximum Gaussian jitter is equal to the maximum total jitter minus 26
the actual High Probability jitter. 27
28
29
2.E.1.2 Spectrum analyzer and Oscilloscope Methodology 30
31
Bandlimited2 Unbounded Gaussian Noise 32
33
Referring to Figure 2-28, bandlimited or high frequency Gaussian noise can be 34
measured at the transmitter of the DUT accurately using a high frequency 101010 35
pattern and measuring the spectral power3.
36
37
38
39
40
41
42
43
44
45
1. It is assumed due to the magnitude of jitter present at the transmitter that the left and right hand parts of the bathtub are 46
independent to each other 47
2. Normal CEI application will integrate from the defined ideal CDR bandwidth to infinity, while some CEI-11G-SR application will
integrate over a specific band
48
3. The spectral power should be measured using averaging 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 107

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 2-28. Spectral Measurement Setupa
3
4
5 Spectrum
DUT
6 Analysiser
7
8 Differential to
9 single ended
10 amplifier
11 Clock
12 Reference (a)
13
14
15 a.The clock reference is such that its power noise represents the typical power noise of the reference in the
16 system
17
18 The spectral power is calculating by integrating over the frequency band of interest and
19 converting into time jitter.
20 P(f)
100f 2 ---------
21
1 1  f1  j  f 10

22
23
 rms = ------ 2  -------------------------------------------------------------  10
2  1 + j  f  f1   1 + j  f  f2 
24 f 1  100
25
26
27 where
28
29  rms is the time jitter
30
31 P(f) is the measured spectral power for 1Hz Bandwidth
32
33 It should be noted that the measured Gaussian noise for a driver can usually be
34 considered equivalent to that derived from a full bathtub jitter distribution.
35
36 Bandlimited 60 second Total Jitter Measurements
37
38 In certain CEI-11G-SR applications total jitter measurements of 60 seconds are
39 required. The Gaussian jitter, as measured above, should be multiplied by a Q of 6.961.
40 If spurs are present in the spectrum then these must be converted to time jitter
41 separately using an inverse of the Bessel function as per Figure 2-29, which describes
42 the power spectrum for a given phase modulated signal.
43
44 where
45
46
F(P n) is the inverse spectral SSB power to time modulation (below)
47
1. Traditional measurements are performed for 60 seconds using a demodulator and performing a real time peak to peak
48 measurement of the jitter. Given this, the number of bits transmitter across the link in 60 seconds is calculated and the
49 associated three sigma confidence level, peak to peak multiplication factor, Q, for the random jitter.

108 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
 pkpk = 2Q rms +  F(Pn) 5
6
n
7
8
9
10
Pn is the relative SSB power of a spur 11
Figure 2-29. Single Side Band Relative Power Spectrum for Phase Modulated Signal 12
13
14
Phase Modulation Power Spectrum
50 15
16
17
45 18
19
40 20
21
22
35
23
24
dBc

30 25
26
27
25
28
29
20 30
31
32
15
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 33
Modulation Index (UI) 34
35
36
Uncorrelated High Probability Jitter 37
38
After measuring the Gaussian Jitter, as above, an oscilloscope measurement, as per 39
Appendix 2.E.7, of the peak to peak jitter should be performed using a 101010 pattern. 40
41
The Uncorrelated High Probability Jitter is then calculated by removing the 42
accumulated Unbounded Gaussian jitter. 43
 UBHJ =  pkpk – 2Q rms 44
45
46
using a Q calculated for a 3 sigma confidence level1 as per Appendix 2.F.3. 47
48
1. It is recommended that enough samples on the oscilloscope should be made such that Q>4 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 109

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Total High Probability Jitter

2
3 After measuring the Unbounded Gaussian Jitter, as above, an oscilloscope
4 measurement, as per Appendix 2.E.7, of the peak to peak jitter should be performed
5 using the standard pattern e.g. PRBS31.
6
7 The Total High Probability Jitter is then calculated by removing the accumulated
8 Gaussian jitter.
9
10  HPJ =  pkpk – 2Q rms
11
12 1
13 using a Q calculated for a 3 sigma confidence level as per Appendix 2.F.3.
14
15 2.E.2 Total Transmit Wander Measurement
16
17 This sub-clause describes the total transmit wander of a simple non-equalized
18 transmitter as depicted below
19 Figure 2-30. Transmit Wander Lab Setup
20
21
22 Differential
23 DUT to single
Golden PLL
1/n Scope
ended Amp
24
25
26
27
28
29
30
31
32
33 Clock Reference

34
35
36
37 • The transmitter under test shall transmit the specified data pattern, while all other
38 signals are active.
39
40 — The other channels can transmit the same pattern if they have at least a 16 bit
41 offset with the channel under test.
42 — All lanes to be active in both transmit and receive directions, and opposite ends
43 of the link, i.e. transmit to receiver, are to use asynchronous clocks (to maximum
44 allowed ppm. offset as specified in the protocol specifications).
45
46 • The transmitter can be tested single ended as high frequency jitter components are
47 filtered by the Golden PLL
48
49 1. It is recommended that enough samples on the oscilloscope should be made such that Q>4

110 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

• Temperature and Supply Voltage should be cycled with a rate slower than baud rate 1
over 166700Hz during test to exercise any delay components in the DUT. 2
3
• The inherent clock wander in signal shall be extracted using Golden PLL and 4
divided, by the 1/n block, such as to limit the measured wander to 1UI at the divided 5
frequency, and thus allowing it to be measured on an oscilloscope.
6
— The Golden PLL should have at a minimum bandwidth of baud rate over 1667, 7
with a maximum of 20dB/dec rolloff, until at least baud rate over 16.67, and is 8
suggested to have no peaking around the corner frequency. 9
10
• The peak to peak total wander of the extracted clock should be measured using a 11
scope trigger by the reference clock. The measured peak to peak wander should be 12
verified to be bounded by repeating the measurement for ever increasing periods of 13
time until the measurement is constant. 14
15
2.E.3 Relative Transmit Wander Measurement 16
17
This sub-clause describes specifically for SxI-5 interfaces, where limitations are defined 18
in terms of relative wander between data lane and clocks, whose relative wander can 19
be measured as depicted below. 20
Figure 2-31. Relative Wander Lab Setup 21
22
23
24
1/n Scope
25
Golden PLL
26
27
DUT
28
Trigger
29
30
Golden PLL
1/n 31
32
33
34
Clock Reference
35
36
37
38
• The transmitter under test shall transmit the specified data pattern, while all other 39
signals are active. 40
41
— The other channels can transmit the same pattern if they have at least a 16 bit 42
offset with the channel under test. 43
— All lanes to be active in both transmit and receive directions, and opposite ends 44
of the link, i.e. transmit to receiver, are to use asynchronous clocks (to maximum 45
allowed ppm. offset as specified in the protocol specifications). 46
47
• The transmitters can be tested single ended as high frequency jitter components 48
are filtered by the Golden PLL 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 111

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • Temperature and Supply Voltage should be cycled with a rate slower than baud rate
2 over 166700Hz during test to exercise any delay components in the DUT.
3
• The inherent clock wander in each signal shall be extracted using Golden PLL and
4
divided, by the 1/n block, such as to limit the measured wander to 1UI at the divided
5
frequency, and thus allowing it to be measured on an oscilloscope.
6
7 — The Golden PLL should have at a minimum bandwidth of baud rate over 1667,
8 with a maximum of 20dB/dec rolloff, until at least baud rate over 16.67, and is
9 suggested to have no peaking around the corner frequency.
10
11 • The peak to peak relative wander between the extracted clocks should be
12 measured using a scope trigger by one of the extracted clocks. The measured peak
13 to peak wander should be verified to be bounded by repeating the measurement for
14 ever increasing periods of time until the measurement is constant.
15
16 2.E.4 Jitter Tolerance
17
18 2.E.4.1 Jitter Tolerance with Relative Wander Lab Setup
19
20 The following sub-clause describes the required jitter tolerance methodology for
21 devices where Relative Wander is applicable e.g. SxI.5 and where no receive
22 equalization is implemented.
23
Figure 2-32. Jitter Tolerance with Relative Wander Lab Setup
24
25
Data Output
26 BERT
transmitting defined test
Voltage
Signal Filter
for defining
Controlled
27 pattern Delay Line
edge rate

28 Control Voltage
Clock
29 Reference Inverted Input
DUT

Input
30 Signal Filter
Voltage
31 Controlled
for defining
edge rate
Delay Line
32
33 Control
Voltage
Calibrated
34 Input
Test Data

35 DUT Clock
Reference
36 (100ppm offset to BERT)

37
38
39 +
White Noise Source
for generating Gaussian
40 Jitter Control
Signal Filter
Jitter

41
42
43 PRBS Generator
for generating

44 Clock
Reference
Jitter Control
Signal Filter
Uncorrelated High
Probability Jitter

45 modulated by Common SJ
Wander Source

46 Common SJ Wander Antiphase SJ Wander

47 Source, which together

with Antiphase
generates Total Wander
Source for generating
Relative Wander

48
49

112 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

General 1
2
• The transmitter under test shall transmit the specified data pattern, while all other 3
signals are active.
4
— The other channels can transmit the same pattern if they have at least a 16 bit 5
offset with the channel under test. 6
7
— All lanes to be active in both transmit and receive directions, and opposite ends 8
of the link, i.e. transmit to receiver, are to use asynchronous clocks (to maximum 9
allowed ppm. offset as specified in the protocol specifications). 10
• The Device Under Test (DUT) shall be tested using an internal BERT or loop to 11
have the defined BER performance 12
13
• The confidence level of the BER measurement should be at least three sigma as 14
per Appendix 2.F.2. 15
16
Synchronization 17
• All lanes are to be active in both transmit and receive direction. 18
19
• All reference clocks should have the maximum offset frequency, with respect to 20
each other, as defined in the implementation agreement. 21
22
Jitter 23
• The applied calibrated test signal shall have applied a calibrated amount of HF GJ 24
and HPJ 25
26
• The jitter control signal for generating High Probability Jitter should be filtered using 27
at least a first order low pass filter with a corner frequency between 1/20 - 1/10 of 28
the baud rate of the PRBS generator to ensure that high frequency components are 29
removed. The distribution of the jitter after the filter must be reasonably even, 30
symmetrical, and large spikes should be avoided. The order of the PRBS 31
polynomial may be between 7 and 11, inclusive, to allow flexibility in meeting this 32
objective. The rate of the PRBS generator should be between 1/10 - 1/3 of the data 33
rate of the DUT being tested, and their rates must be not harmonically related. The 34
upper -3 dB frequency of the filtered HPJ should be at least 1/100 of the data rate of 35
the DUT being tested to represent transmitter jitter that is above the tracking 36
frequencies of the DUT's CDR. Calibration of HPJ must be done with a golden PLL 37
in place. Once these objectives are achieved, there is no need to vary these 38
settings; any combination of settings that meets all the objectives is satisfactory. 39
• The jitter control signal for generating Unbounded Gaussian Jitter shall be filtered 40
as per Figure 2-5 using the “Jitter Control Signal Filter”. However, the upper 41
frequency of the Gaussian jitter spectrum will be, acceptably, limited by the 42
bandwidth of the voltage controlled delay line. The crest factor of the White Noise 43
generator should be better than 18dB. 44
45
• The calibrated test signal shall have a calibrated amount of Total Wander and 46
Relative Wander as compared to the used clock by using the Common SJ Wander 47
and Antiphase SJ Sources with 1% frequency offsets. (Note the use of the inverted 48
input to the uppermost delay line), as per Annex 2.C.2 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 113

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • The amplitude of the Total Wander and Relative Wander is defined by the sinusoidal
2 masks defined in Annex 2.A.1 and Annex 2.A.2 with the specified amplitudes from
3 the implementation agreement.
4
• Wander should be applied
5
6 — from a frequency equivalent to 1UI of Total Jitter up to 20MHz modulation
7 frequency
8
9 — at a maximum of 2MHz frequency steps above the corner frequency
10 — at a maximum of 200kHz frequency steps below the corner frequency.
11
12 Amplitude
13
14 • The calibrated data signals should be filtered using a single pole low pass filter with
15 a corner frequency of 0.7 times the baud rate, to define the edge rate.
16 • The amplitude of signal should be adjusted such that it just passes the defined
17 receiver data eye sensitivity.
18
19 • For testing of DC coupled receivers either a pattern generator capable of generating
20 differential signals and setting the common mode should be used or a combined AC
21 coupled signal together with a biased-T. Using this setup the common mode should
22 be varied between the defined maximum and minimum.
23
24 2.E.4.2 Jitter Tolerance with no Relative Wander Lab Setup
25
26 The following sub-clause describes the required jitter tolerance methodology for
27 devices where Relative Wander is not applicable and no receive equalization is
28 implemented.
29 Figure 2-33. Jitter Tolerance with no Relative Wander
30
31
32 BERT
Data Output
Voltage
Signal Filter
for defining
DUT
Controlled
33 Delay Line
edge rate

34
35 Clock
Reference Calibrated
36 Input Test Data

37 Wander can be
optionally applied
directly to FM
38 Clock
input
+
White Noise Source
for generating
Unbounded Gaussian
Jitter Control
39 Reference
Signal Filter Jitter

40
41
42 PRBS Generator
for generating
uncorrelated High
43 Total
Jitter Control
Signal Filter Probability Jitter
SJ Wander
44 Source

45
46 Referring to Figure 2-33, the DUT shall be tested as per the description in Appendix
47 2.E.4.1, omitting any requirements relating to relative wander and where only Total
48 Wander is applied via the SJ Source shown.
49

114 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.E.4.3 Jitter Tolerance with Defined ISI and no Relative Wander 1

2
The following sub-clause describes the required jitter tolerance methodology for 3
devices where Relative Wander is not applicable e.g. SxI.5 and where receive 4
equalization is implemented and the performance of the equalization must be verified. 5
Figure 2-34. Jitter Tolerance with Defined ISI 6
7
8
+
Voltage Signal Filter
BERT Controlled FFE to Define DUT 9
Delay Line (Optional) Edge Rate
Stress
10
Clock
Channel 11
Reference 12
Input Calibrated
Test Data 13
14
Sinusoidal Noise
Source for 15
Generating
Crosstalk 16
Wander Can
be applied to
17
FM Input White Noise Source 18
Clock Reference + Jitter Control
Signal Filter
for Generating
Unbounded Gaussian 19
Jitter
20
21
Jitter Control
PRBS Generator for
Uncorrelated High
22
Sinusoidal
Wander Signal Filter Probability Jitter 23
Source 24
25
Referring to Figure 2-34, the DUT shall be tested as per the description in Appendix 26
2.E.4.1, omitting any requirements relating to relative wander, and additionally 27
28
• The transmit jitter and amplitude shall be initially calibrated as per Appendix 2.E.1 at 29
the output of the delay line. 30
• The stress channel shall have the characteristics specified in the relevant test 31
method. 32
33
• The use of a Transmit Equalizing Filter (FFE) is optional. If it is included then its 34
characteristics should be adjusted in accordance with the relevant test method. 35
• The defined amount of uncorrelated additive noise shall be applied via a sinusoidal 36
source differentially to the signal. The frequency used shall be between 100MHz 37
and the lesser of 1/4 the data rate and 2GHz. There is no need to sweep the 38
frequency. 39
40
41
2.E.5 Jitter Transfer 42
43
This section describes how jitter transfer relevant interfaces can be tested for 44
compliance, e.g. CEI-11-SR-Transparent, SxI-5. Referring to Figure 2-35 45
• The BERT shall generate a data pattern as defined by the IA 46
47
• The jitter present before the delay line should be minimized as much as possible so 48
as to maximize any transfer bandwidth function of the DUT 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 115

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • A sinusoidal jitter should be applied following the same defined SJ mask as used for
2 jitter tolerance, with the same resolution as described in Appendix 2.E.4.
3
4 The peak to peak jitter for a 60 second period measured on the scope should be
5 compared before and after the application of the sinusoidal jitter. The ratio of the
6 difference to the jitter applied is then defined as the jitter transfer function.
7 Figure 2-35. Jitter Transfer Lab Setup
8
9
10
BERT Delay Line DUT Scope
11
12
13
14
Total
15 Wander can be SJ Wander
optionally applied Source
16 directly to FM
input
17 Clock
Reference
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

116 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2.E.6 Network Analysis Measurement 1

2
To enable accurate analysis of a channel the following methodology should be followed 3
for the measurement and calculation of the effective channel transfer function. 4
Figure 2-36. S-parameter Port definitions 5
6
Forward 7
Channel 8
Term. Term. 9
10
TxP,N RxP,N
11
12
13
Term. Term. 14
15
NEXT 16
Term. Term.
17
18
19
Term. RxP,N
20
21
Term. TxP,N 22
23
FEXT 24
25
TxP,N Term.
26
27
Term. RxP,N 28
29
Term. Term. 30
31
• Figure 2-36 shows an overview of the termination and port definitions typically used 32
when measuring the forward channel and NEXT/FEXT crosstalk aggressors 33
34
• The intermediate frequency (IF) bandwidth should be set to a maximum of 300 35
Hertz with 100 Hertz preferred. The launch power shall be specified to the highest 36
available leveled output power not to exceed 0 dBm.1 37
• Either direct differential measurements of the channel S21 and S11 should be 38
performed or multiple single ended measurements from which the differential 39
modes should be calculated.2 40
41
• Linear frequency steps of the measurements shall be no larger than 12.5MHz. 42
43
• A frequency range from no higher than 100MHz to no lower than three times the
44
fundamental frequency should be measured.
45
46
47
1. Please refer to Agilent PLTS data sheet #5989-0271EN, and Agilent TDR Users Guide #54753-97015, section 2.2
48
2. Special care must be taken when performing multiple single ended measurements if the system is tightly coupled 49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 117

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 • Extrapolation towards DC should be performed linearly on magnitude part with the

2 phase being extrapolated to zero at DC, i.e. only a real part is present at DC.
3
• The channel response of the channel should be calculated by cascading the
4
complete 4 port s-parameter matrix with a worst case transmitter and receiver. The
5
transmitter/receiver should be described as a parallel R and C, where R is the
6
defined maximum allowed DC resistance of the interface and C is increased until
7
the defined maximum Return Loss at the defined frequency is reached.
8
9 • Any defined effective transmit or receiver filters should also be cascaded with the
10 channel response
11
12 • The time resolution should be increased by resampling the impulse response in the
13 time domain
14 • If required interpolation of the frequency domain should be performed on the
15 magnitude and unwrapped phase components of the channel response
16
17 S () S 21()
1 1 Rx 11() 1
18 Tr() =  11 
19 1 Tx 22() S 12() S 22() 1 1
20
21
22
23 where
24
25
S m n is the measured 4 port differential data of the channel
26 Tx is the transmitter return loss
27 22
28 Rx is the receiver return loss
29 11
30 Tr() is the receiver return loss
31
32 converting the original frequency range to time domain, we obtain
33
34 i(t m) = ifft(Tr())
35
36
37 where
38
3
39  = [– --- f
3
baud,--- f ]
40 4 4 baud
41
42
43 2.E.7 Eye Mask Measurement Setup
44
45 The measurement of an eye mask is defined by the various Implementation
46 Agreements in terms of a polygon for the probability of the required Bit Error Rate. This
47 polygon may have to be altered given that the sample population of the scope is limited
48
49

118 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

and must be adjusted as per Appendix 2.F.3. For the measurement of the signal the 1
laboratory setup shown in Figure 2-37 should be used, including the recommendations 2
list in Appendix 2.E.1. 3
Figure 2-37. Eye Mask Measurement with Golden PLL 4
5
6
Differential 7
to single
ended amp + 8
DUT
Golden PLL 9
10
11
Trigger
12
13
Signal
14
Clock Oscilloscope 15
Ref
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 119

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.F Appendix - BER Adjustment Methodology

2
3
2.F.1 Extrapolation of Correlated Bounded Gaussian Jitter to low BERs
4
5 For IAs with BER requirements of 1 10–15 or lower, measurements to that level are
6 very time consuming (or rely on averaging multi-links), hence more practical to only
7 take measurements to Qs around 7 (BER around 1 10–12 ).
8
9 Bathtub Measurements
10
11 CBGJ can appear as either GJ or CBHPJ depending upon the Q at which it is
12 linearised.
13
14 If HPJ and GJ are measured using a bathtub there is no knowledge as to if the GJ is
15 UUGJ or CBGJ. For system budgeting it is recommended that the bathtub GJ should
16 be assumed to be all UUGJ.
17
18 If combined spectral, oscilloscope methods are used then UUGJ, UBHPJ and CBHPJ
19 can be estimated. It is not possible to estimate the CBGJ as it has already become
20 bounded and appears as CBHPJ. For system budgeting it is recommended that this
21 peak value is valid for the extrapolated Q of interest.
22
23
2.F.2 Confidence Level of Errors Measurement
24
25 Assuming that a link, with a given BER, can be modelled as a Bernoulli random
26 process, the following statistics can be assumed.
27
28 Given,
29
30 p is the probability of error
31
32 q =  1 – p  is the probability of not erroring
33
34 n is the number of bits received and measured
35
36 then,
37
38 m = np is the expected number of errors received
39
40  = npq is the sigma of the variation of the number of errors received
41
42
43
44
45
46
47
48
49

120 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

As an example process, for a 3 sigma confidential level 1

2
– 12 3
p = 10
4
12 5
n = 100  10 6
7
8
m = 100 9
10
 = 10 11
12
13
min Q=3
m =  m + Q  14
max Q = –3 15
min 70 16
m = 17
max 130
18
19
To assess the accuracy of such a measurement an equivalent process with a higher 20
BER can be calculated that would show the same limit of error for the same confidence 21
level and measured number of bits. 22
23
m = E  m  – Q 24
max
25
m = np – Q npq 26
max
m = np – Q np  1 – p  27
max 28
29
Solving the quadrative equation for p 30
31
– 12 32
p = 1.69 10
33
34
35
2.F.3 Eye Mask Adjustment for Sampling Oscilloscopes
36
In all Interoperability Agreement the data mask is defined for the bit error rate of the 37
link. Given that this bit error rate is very small, typical oscilloscope measurement will 38
not sample enough points to be able to verify compliance to these mask. 39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 121

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.F.3.1 Theory
2
3 Figure 2-38. Example Data Mask
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 0UI X1 0.5UI 1-X1 1UI
19
20
21 Given an example eye mask, Figure 2-38, the extremes of the mask, X1 are defined as
22 a linear addition of a Gaussian and High Probability jitter component.
23
HPJ
24 X1 = ----------- + Q  GJ rms
25 2
26
27 where
28
29 HPJ is the high probability jitter
30
31 GJ
rms is the gaussian distributed jitter
32
33 Q is the GJ multiplication factor
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

122 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Given a low sample population and the requirements for mask verification to achieve a 1
hit or no-hit result, X1 must be adjusted according to the sample population and the 2
confidence level that a particular peak to peak is achieved., Given a random process 3
Figure 2-39. Example Data Mask 4
5
Sample 6
Population of n 7
8
9
10
11
Maximum of 12
Population 13
14
15
16
17
18
Amplitude 19
20
21
the probability of measuring a particular maximum amplitude on an oscilloscope, 22
requires one sample to lie on the maximum and all other samples to lie below this 23
value. Referring this all to a half Gaussian distribution and a population of n, there are n 24
different ways this can occur, 25
xm n–1 26
  27
P  x m  = nQ  x m  Q  x  dx

 28
  29
0 
30
31
where 32
33
x m is the random variable of the maximum amplitude measured 34
35
x is the random variable of the underlying random jitter process 36
37
Q  x  is the Q function of the Normal probability density function 38
39
n is the sample population 40
41
P  x m  is a probability density function 42
43
The equation above is solved and the probability of attaining a given maximum 44
(normalized to the sigma) for various populations plotted, Figure 2-40. 45
46
47
48
49

Optical Internetworking Forum - Clause 2: Jitter and Interoperability Methodology 123

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 2.F.3.2 Usage
2
3 Given a known sampling population, n, calculated from the measurement time, average
4 transition density and sampling/collection
–3
frequency of the oscilloscope the three sigma
5 confidence level (i.e. 1.3 10 ) of the measured Gaussian jitter peak value can be read
6 from Figure 2-40. This value should be multiplied by 2 to give the full peak to peak
7 value of the random jitter.
8
9 The three sigma confidence level should be understood as ensuring that 99.96% of all
10 good devices do not violate the eye mask. To limit the number of bad devices that also
11 pass the eye mask it is strongly recommended that the sample population be chosen
12 as to give a Q larger than 5.
13
14 e.g. refering to the red circled intersections Figure 2-40, if we calculate that the sample
15 population for an oscilloscope was 100 i.e. n=100, then for a 3 sigma confidence this
16 equals a Q of 4.2. As the recommended Q value is 5 we should increase the sample
17 population to 10k to give a Q of 5.2.
18 Figure 2-40. Cumulative Distribution Function of Maximum Amplitude
19
20
21 10
0

22
23
24
25
26 -1
10
27
28
29 n=1
Confidence Level

30 n=10
31 -2
n=100
32 10 n=1k
33 n=10k
34 n=100k
35 n=1M
36
-3 3 sigma confidence
37 10
38 level
39
40
41
42 10
-4

43 0 1 2 3 4 5 6 7
44 Q
45
46
47
48
49

124 Clause 2 : Jitter and Interoperability Methodology - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

3 Common Electrical Specification 1

2
3
3.1 Introduction 4
5
This clause specifies electrical parameters and attributes common to all links defined in 6
clause 1. In the event of a difference between an individual clause and these general 7
requirements, the respective individual clause shall prevail. 8
9
10
3.2 General requirements 11
12
13
3.2.1 Data Patterns 14
15
This IA does not have any requirements for specific data patterns (i.e. 8B/10B, 64/66B, 16
SONET scrambling, stream cipher, raw data, etc.), however the following requirements 17
are necessary to insure proper operation. If all of these conditions are not met, then the 18
link may not work to the full distance, or meet the BER, or in fact work at all. 19
• Average transition density needs to converge to 0.5 over a long period (>109 bits), 20
but can in the extreme be between 0.45 and 0.55 over a 30,000 bit period with a 21
probability of at least one minus the BER ratio (1-10-15 with a test requirement to 22
verify 1-10-12) 23
24
• Average DC balance needs to converge to 0.5 over a long period (>109 bits), but 25
can in the extreme be between 0.45 and 0.55 over a 30,000 bit period with a 26
probability of at least one minus the BER ratio (1-10-15 with a test requirement to 27
verify 1-10-12). 28
• Probability of run lengths over 10 to be proportional to 2-N for N-like bits in a row 29
(N10). Hence, a run length of 40 bits would occur with a max probability of 2-40. 30
31
• If a fixed block coding scheme is used (e.g. 8B/10B, SONET), the raw data must be 32
scrambled before coding or the coded data must be scrambled prior to 33
transmission. This is to prevent the so called worst case patterns (e.g. CJPAT-like 34
patterns). 35
36
SONET can be viewed as a coding scheme that can create worst case patterns (via the 37
un-encoded overhead bytes). Two such cases would be the A1/A2 pattern and the Z0 38
byte that can be anything (each unscrambled byte is repeated N times in an OC-N 39
stream [N = 3, 12, 48, 192]). 40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 125

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 3.2.2 Signal Levels

2
3 The signal is a low swing differential interface. This implies that the receiver has a wide
4 common mode range (within the max. absolute input voltages). All devices must
5 support load type 0 defined in Table 3-1, SR devices can optionally support any/all of
6 the other 3 load types while LR devices can optionally support load type 1.
7 Table 3-1. Definition of load types
8
9 Characteristic Load Type 0 Load Type 1 Load Type 2 Load Type 3 Unit
10 R_Zvtt >1k <30 <30 <30 
11
12 Nominal Vtt undefined 1.2 1.0 0.8 V
13
14 This type of differential interface allows for interoperability between components
15 operating from different supply voltages and different I/O types (CML, LVDS-like,
16 PECL, etc.). Low swing differential signaling provides noise immunity and improved
17 electromagnetic interference (EMI). Differential signal swings are defined in following
18 sections and depend on several factors such as transmitter pre-equalization, receiver
19 equalization and transmission line losses.
20
21 3.2.3 Bit Error Ratio
22
23 The link will operate with a Bit Error Ratio (BER) of 10-15 (with a test requirement to
24 verify 10-12 - see Clause 2 for more information on the jitter model and how to measure
25 BER)
26
27
28 3.2.4 Ground Differences
29
30 The maximum ground difference between the driver and the receiver shall be 50mV
31 for SR links and 100mV for LR links. This will affect the absolute maximum voltages at
32 compliance point 'R'. If driver and receiver are on the same PCB with no intervening
33 connectors, then the ground difference is approximately 0 mV.
34
35 3.2.5 Cross Talk
36
37 Cross talk arises from coupling within the connectors, on the PCB, the package and the
38 die. Cross talk can be categorized as either Near-End or Far-End Cross talk (NEXT and
39 FEXT). In either of these categories, the amount of cross talk is dependent upon signal
40 amplitudes, signal spectrum, and trace/cable length. There can be many aggressor
41 channels onto one victim channel, however typically only a few are dominant.
42
43 Further consideration of Crosstalk can be found in Appendix 3.A.4.
44
45
46
47
48
49

126 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

3.2.6 Driver Test Load 1

2
All driver characteristics should be implemented and measured to a differential 3
impedance of 1% at DC with a return loss of better than 20dB from baud rate 4
divided by 1667 to 1.5 times the baud rate, unless otherwise noted. 5
6
3.2.7 Driver Lane-to-Lane Skew 7
8
While the protocol layer will control some of the lane to lane skew, the electrical level is 9
allowed up to 500ps of lane-to-lane skew caused by the driver circuitry and associated 10
routing. Hence, the total output (i.e. measured) lane-to-lane skew is to be specified in 11
the protocol standards with this 500ps taken into account. The driver lane-to-lane skew 12
is only for the Serdes TX and does not include any effects of the channel. 13
14
15
3.2.8 Input Lane-to-Lane Skew 16
17
While the protocol layer will control the maximum amount of lane to lane skew that is 18
allowed, it must allow for up to 1000ps of skew caused by the driver & receiver circuitry
19
and associated routing (that is 500ps for the driver and 500ps for the Rx). The input
20
lane-to-lane skew does not include any skew effects of the channel.
21
22
3.2.9 Driver Short Circuit Current 23
24
The max DC current into or out of the driver pins when either shorted to each other or 25
to ground shall be ±100mA when the device is fully powered up. From a hot swap point 26
of view, the ±100mA limit is only valid after 10 s 27
28
3.2.10 Differential Resistance and Return Loss, Driver and Receiver 29
30
The DC differential resistance shall be between 80 and 120. 31
32
The differential return loss shall be better than A0 from f0 to f1 and better than A0 + 33
Slope*log10(f/f1) where f is the frequency from f1 to f2. See Figure 3-1 for definitions. 34
Differential return loss is measured at compliance points T and R. If AC coupling is 35
used, then all components (internal or external) are to be included in this requirement. 36
The reference impedance for the differential return loss measurements is . 37
38
Common mode return loss measurement shall be better than -6dB between a minimum 39
frequency of 100MHz and a maximum frequency of 0.75 times the baud rate. The 40
reference impedance for the common mode return loss is 25. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 127

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 3-1. Driver and Input Differential Return Loss
3
4
5
6
7
8
9 Slope
10
11
Loss (dB)

12
13 A0
14
15
16 Acceptable
17 Region
18
19
20
21
22 f0 f1 f2
23 Frequency (Hz)
24
25
26
27
28
29 3.2.11 Baud Rate Tolerance
30
The range of operating Baud rates is defined specifically for each interface in the
31
specific clauses. Each CEI interface is required to operate asynchronously with a
32
tolerance of +/-100ppm from the nominal baud rate.
33
34
35 3.2.12 Termination and DC Blocking
36
37 Each link requires a nominal 100differential source termination at the driver and a
38 nominal 100differential load termination at the receiver. The terminations shall provide
39 both differential and common mode termination to effectively absorb differential or com-
mon mode noise and reflections. Receivers and transmitters shall support AC coupling
40 and may also optionally support DC coupling. AC Coupled receivers require a differential
41 termination >1kat DC (by blocking capacitors in or near receivers as shown in Figure
42 3-2 or by circuit means within the receiver). DC Coupled Devices shall meet additional
43 electrical parameters T_Vcm, R_Vrcm, R_Vtt, R_Zvtt. All termination components are
44 included within the Rx and TX blocks as shown in the reference model as defined in Sec-
45 tion 1.8.
46
47
48
49

128 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 3-2. Termination Example 1

50 0, 1, 2 2
50
ohm Connectors 3
ohm 4
5
6
7
8
9
10
11
AC Capacitors AC 12
Gnd (Optional) Gnd 13
14
Driver Receiver 15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 129

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4 3.A Appendix - Transmission Line Theory and Channel
5 Information
6
7 3.A.1 Transmission Lines Theory
8
9 The performance of a high frequency transmission line is strongly affected by
10 impedance matching, high frequency attenuation and noise immunity.
11
12 It is possible to design a high frequency transmission line using only a single conductor.
13 Nevertheless most high frequency signals use differential transmission lines (i.e. a pair
14 of coupled conductors carrying signals of opposite polarity). Although differential
15 signaling appears wasteful of both pins and signal traces it results in much better noise
16 immunity. Differential signals produce less conducted noise because the opposite
17 power and ground current flows cancel each other both in the line driver and in the
18 transmission line. Differential signals produce less radiated noise because over a
19 modest distance the opposite fields induced by the opposite currents cancel each
20 other. Differential signals are less susceptible to noise because most sources of noise
21 (common mode noise) tend to affect both signal lines identically, producing a variation
22 in common mode voltage but not in differential voltage.
23
24 3.A.1.1 Impedance Matching.
25
26 The AC impedance of a single conductor is determined by the trace geometry, distance
27 to the nearest AC ground plane(s) and the dielectric constant of the material between
28 the trace and the ground plane(s). If the distance between the signal trace and the
29 nearest ground plane is significantly less than the distance to other signal traces the
30 signal trace will behave as a single-ended transmission line. Its AC impedance does
31 not vary with signal polarity although it may vary with frequency due to the properties of
32 the dielectric material. This impedance is often called single ended impedance, Zse.
33
34 The AC impedance, Z of a differential transmission line is affected by the configuration
35 of the pair of conductors and the relationship between their signal polarities, in addition
36 to the trace geometry, distance to the nearest AC ground plane(s) and the dielectric
37 constant of the material between the trace and the ground plane(s). If the paired
38 conductors are close enough to interact (coupled), then the impedance for signals of
39 opposite polarity (odd mode impedance, Zodd) will be lower than the impedance for
40 signals of the same polarity (even mode impedance, Zeven).
41
42 If there is minimal coupling between the paired conductors then Zodd = Zeven = Zse.
43 Coupled transmission lines always produce Zodd < Zse < Zeven. The following
44 equations relate effective differential impedance, zdiff to common mode impedance,
45 Zcm and single ended impedance, Zse to even and odd mode impedances:
46
47 Zeven Zeven + Zodd
Zdiff = 2Zodd Zcm = --------------- Zse = -----------------------------------
48 2 2
49

130 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Most differential data signals are designed with zdiff = 100 and 25< Zcm < 50. 1
2
There is a trade-off in the choice of Zcm. With Zcm = 25 (no coupling) may reduce 3
conducted noise for transmission lines with inadequate AC or DC grounding. Zcm = 4
50 (close coupling) may reduce radiated noise (crosstalk) which is more critical in 5
backplanes. However close coupling requires careful ground construction to control 6
common mode noise. 7
8
The reader may wonder why common mode impedance is meaningful in a differential 9
transmission system. In a perfectly constructed system only odd mode (opposite 10
polarity) signals propagate. However imperfections in the transmission system cause 11
differential to common mode conversion. Once converted into common mode the 12
energy may convert back to differential mode by the same imperfections. Thus, these 13
imperfections convert some of the signal energy from opposite polarities to the same 14
polarity and back. 15
16
The two main sources of mode conversion are impedance mismatches which cause 17
part of the energy to be reflected, and differential skew which causes variations in 18
forward signal propagation delay between the individual paths of the differential pair. 19
Impedance mismatches typically occur at boundaries between transmission line 20
segments, including wire bonds, solder joints, connectors, vias and trace-to-via 21
transitions. Often ignored sources of impedance mismatches at these boundaries are 22
discontinuities within the AC ground itself as well as asymmetric coupling between the 23
individual traces and the AC ground. Differential skew can occur at these same 24
boundaries and also due to mismatched trace lengths in device packages and in PCBs. 25
26
3.A.1.2 Impedance Definition Details 27
28
Differential transmission lines consist of two conductors and a ground plane. The 29
voltage-current relationships at one end of this line can be formulated in terms of a two- 30
port as in Figure 3-3. 31
32
Figure 3-3. Transmission Line as 2-port
33
I1 34
Conductor 1 35
I2 36
V1
Conductor 2 37
V2 38
Ground Plane 39
40
41
Transmission Line
42
43
The voltage current relationships are: 44
45
V 1 = Z 11 I 1 + Z 12 I 2 V 2 = Z 21 I 1 + Z 22 I 2 46
47
If the line is infinitely long or perfectly terminated, then these four impedance values are 48
the characteristic impedance of the line. The characteristic impedance is a 2 x 2 matrix: 49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 131

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Z 11 Z 12
3 Zˆc =
4 Z 21 Z 22
5
6
7 Generally, all four of the matrix entries are complex. But, at frequencies of interest, the
8 inductance and capacitance per unit length dominate so that all four quantities are
9 approximately real, positive numbers. For engineering purposes it is common to speak
10 of the impedances as though they are resistances, with no imaginary part; keeping in
11 mind that the imaginary part exists. Since the line is passive and symmetric, we have 
12 Z11 = Z22 and Z12 = Z21 so that the line is described by just two impedance values. If
13 the line is to be perfectly terminated, then we must create a network that is equivalent
14 to Zˆc . That is, we need a 3-terminal (2 nodes + ground) network that presents the same
15 values of Z11 and Z12 as the line. A T or pi network could be used. The pi network is
16 shown in Figure 3-4, along with the impedance values in terms of Z11 and Z12.
17
18 .
19 Figure 3-4. PI Network Termination
20 

21 Z 11 – Z 12
2 2 I1
Za
22 Za = Z 11 + Z 12 Zb = ---------------------------
23 Z 12
I2 Zb Za
24
V1
25 V2
26 ZaZb
27 Zodd = ----------------------- = Z 11 – Z 12 Zeven = Za = Z 11 + Z 12
2Za + Zb
28
29
30
The odd and even mode impedances, Zodd and Zeven, are other impedance
31
definitions that are more descriptive referring to the polarity of the signal propagating
32
the differential pair. In the case of opposite signal polarity in the two lines of the signal
33
pair the odd mode impedance is used. In the case of same signal polarity the even
34
mode is used. Zodd and Zeven are measured as shown in Figure 3-5.
35
36
37
38 Figure 3-5. Measurement of Zodd, Zeven
39
40 I
41 + 1 + 1
42 I
V
I
43 - -
2 - + 2
44 V V
-
45 +
gnd gnd
46
47
48
49

132 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Zodd Zeven 1
 2
V = V1 = –V2 V = V1 = V2  3
 4
I = I1 = –I2 I = I1 = I2  5
 6
V V
Zodd = --- Zeven = ---  7
 I I 8
9
10
11
12
Odd mode impedance is the impedance measured when the two halves of the line are 13
driven by equal voltage or current sources of opposite polarity. Even mode impedance 14
is the impedance measured when the two halves of the line are driven by equal voltage 15
or current sources of the same polarity. In this specification the differential mode 16
impedance, Zdiff and the common mode impedance, Zcm are used. The relationship to 17
even and odd mode impedances is given as: 18
19
Zeven Zeven + Zodd
Zdiff = 2Zodd Zcm = --------------- Zse = ----------------------------------- 20
2 2 21
22
From the above equations we see that Zeven is always greater than Zodd by 2Z12, 23
where Z12 is a measure of the amount of coupling between the lines. This means that 24
Zeven is larger than Zodd for coupled transmission lines. 25
26
3.A.2 Density considerations 27
28
The preceding section showed that, for two idealized forms of termination, Zodd is 29
correctly terminated but Zeven is not. The first illustrated case, using a 50 ohm resistor 30
(or its equivalent) from either terminal to ground (or to AC ground), has become 31
relatively standard. Because it has ZoddT = ZevenT = 50 ohm, it provides correct 32
differential termination and is often close to providing correct common-mode 33
termination. 34
35
By increasing the conductor spacing in the transmission line we can decrease Zeven 36
(decrease Z12) and bring it closer to 50 ohm. But dense backplanes require a large 37
number of transmission lines per unit cross-sectional area of the printed circuit board. 38
This means that the two printed circuit traces comprising the differential transmission 39
line are forced close together, which increases Z12. The backplane design is therefore, 40
a compromise between the desire for high density of transmission lines and a desire for 41
correct common-mode termination. 42
43
Transmission lines act as low-pass filters due to skin effect and dielectric absorption. 44
As the density of transmission lines increases, both the series resistance per unit 45
length and the parallel conductance per unit length increase. This, in turn, results in 46
greater attenuation at a given frequency. Thus, high speed backplane design is not just 47
a compromise between density and common-mode matching. There is also a 48
compromise between density and attenuation. 49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 133

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 3.A.3 Common-Mode Impedance and Return Loss

2
3 It is demonstrated above that increasing the density of transmission lines in a
4 backplane results in higher common-mode impedance, which is known as interference
5 and for high amplitudes the receiver is likely to be disrupted.
6
7 Common-mode interference arises from several sources. Among them are:
8 1. Imperfections in driver circuits.
9
10 2. A difference in length between the two conductors of the transmission line
11 3. Imperfections in impedance matching across board boundaries connectors and
12 vias causing mode conversion, differential to Common mode
13
14 4. EMI.
15
16 The interference resulting from the driver probably has a spectrum that is the same as
17 or similar to that of the signal. EMI arising from coupling into the printed circuit traces
18 should be small, assuming that coupled stripline is used. However, connector pins may
19 be exposed. EMI may have frequency components that are well below signal
20 frequencies, which means that it won’t necessarily be attenuated to the extent that
21 signals are. But, at the same time, the lower frequencies are probably poorly coupled
22 into the backplane circuit.
23
24 Earlier, two ideal forms of termination were presented based on either one or two
25 resistors. These ideal terminating devices are helpful in examining the relationship
26 between the parameters of the transmission line versus those of the device. Real
27 devices, however, are not simple resistances. They contain parasitic components and a
28 non-ideal path from package pins to die. There may also be a need to AC-couple the
29 terminations.
30
31 The most that we can do in this situation is to make the package and the die appear as
32 close to ideal as possible over as much of the signal spectrum as possible. The extent
33 of the deviation from ideal is specified and measured as a function of frequency. The
34 preferred measures are S11 (single-ended return loss) or SDD11 (differential return loss)
35 as functions of frequency. (Sometimes S22 or SDD22 are used to indicate an output.)
36 Ideally these return losses are 0 (no reflection) over the frequency range of interest. In
37 dB this is -.
38
39 Note: Sometimes a return loss is specified as a positive number, it being understood
40 that this still refers to the log of a reflection coefficient in the range of 0 to 1.
41
42 3.A.4 Crosstalk Considerations.
43
44 This IA assumes that the dominant cross talk can come from aggressors other than the
45 transmitter associated with the receiver. Hence NEXT cancellation is not useful.
46
47 Crosstalk between CEI channels should be minimized by good design practices. This
48 includes the pin-out arrangement to the driving/receiving IC’s, connectors and
49 backplane tracking.

134 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Optimum arrangement for minimising crosstalk between channels at IC pins is 1

illustrated in Figure 3-6 below. Crosstalk between channels can be reduced by 2
grouping TX and RX pins and avoiding close proximity between individual TX and Rx 3
pins. This practice will minimize coupling of noise from TX drivers into RX inputs. 4
5
6
Figure 3-6. Minimisation of crosstalk at IC pins. 7
8
9
Diff pair 10
Rx 11
Tx
12
Rx Can increase cross talk
Tx 13
Rx
due to Tx beside an Rx, 14
Chip Tx yet is good to allow for 15
Rx
Tx
loopback debug testing 16
Rx 17
Tx 18
Rx 19
Rx 20
Rx Best for cross talk prevention 21
Rx
Rx due to separating Rx and Tx, 22
Chip but harder to design in 23
Tx
loopback debug testing 24
Tx 25
Tx
Tx 26
Tx 27
28
29
Crosstalk at connector pins can be minimized by careful optimisation of connections as 30
shown in Figure 3-7 below. 31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 135

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 3-7. Minimisation of crosstalk at connector pins
3
4
Diff pair single trace
5
6
7 Rx
Rx
8 Rx
Rx Best for crosstalk prevention
9
Tx due to separating Rx and Tx,
10 Tx
but might be harder to route
11 Tx
Tx
12
13 Poor design for crosstalk prevention
Rx
14 Tx due to Tx beside an Rx,
15 Rx
Tx Might be easier to route.
16
Rx Note quite a lot of the crosstalk is in
17 Tx
Rx the via’s, while routing and internal
18
19
Tx
parts of the connector cause the rest
20
21
22
23 Crosstalk between channels over a backplane can be minimized by careful
24 arrangement of tracking, avoiding coupling of noise into RX inputs and increasing
25 spacing “d” between channels as far as possible as shown in Figure 3-8 below.
26
27
28
29 Figure 3-8. Minimisation of crosstalk over backplane
30
31
32
33
34
35
36
37
38
39
40
41 3.A.5 Equation based Channel Loss by curve fit.
42
43 This section describes a technique with specific limitations. It does not include any
44 phase data for the SDD21, and includes no return loss information about SDD11 or
45 SDD22, neither phase nor magnitude, information that is critical for the evaluation of a
46 specific topology's performance. The above proposed statistical-eye characterization
47 includes these effects by including the full 4-port s-parameter measurements. The
48 following method is included for information only and is believed to be of relevance to
49 the overall understanding of the channel transfer loss.

136 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

One way to specify the channel loss is to have an average or worst case “curve” fit to 1
several real channels. This method includes effects of real vias and connectors. This 2
method typically uses the equation below: 3
4
5
6
7

Att = – 20 log  e  a 1 f + a 2 f + a 3 f
2
 8
9
10
11
12
13
Where f is frequency in Hz, a1, a2, & a3 are the curve fit coefficients and Att is in dB. 14
15
Table 3-2 gives some examples of these coefficients and Figure 3-9 plots them along 16
with the PCB model and a real 75cm backplane (with 5cm paddle cards on both ends). 17
These examples are representative for CEI-6G-LR applications but do not represent 18
specifications that a CEI link are to comply with. 19
20
Table 3-2. Curve fit Coefficients 21
a1 a2 a3
22
23
XAUI [ 19] (50cm) 6.5e-6 2.0e-10 3.3e-20
24
75cm [ 24] “Worse” 6.5e-6 3.9e-10 6.5e-20 25
75cm [ 24] “Typical” 6.0e-6 3.9e-10 3.5e-20 26
27
28
Figure 3-9. Equation based Channel Loss curves 29
30
0.00 31
-5.00 32
-10.00
33
34
35
-15.00

-20.00
36
Loss (dB)

-25.00
37
-30.00 38
-35.00 39
-40.00
40
41
42
-45.00

-50.00
43
1.00E+08

1.00E+09

1.00E+10

44
Frequency (Hz) 45
XAUI Model (50cm) 75cm Typical Case 75cm Worst Case Real 75cm Channel
46
47
48
49

Optical Internetworking Forum - Clause 3 : Common Electrical Specification 137

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

138 Clause 3 : Common Electrical Specification - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

4 SxI-5, SFI-4.2, SFI-5.1 & SPI-5.1 Interfaces 1

2
3
4.1 Introduction 4
5
This clause details the requirements for the SxI-5 electrical interface (which includes  6
the following three OIF Implementation Agreements SFI-4.2, SFI-5.1 and SPI-5.1). 7
8
9
4.2 General Requirements 10
11
This clause uses “Method A” of the Jitter and Interoperability Methodology section. 12
13
14
4.2.1 Channel Compliance 15
16
As per 2.1.2, with the following reference transmitter and reference receiver (note these 17
conditions do not specify any required implementation but rather indicate a 18
methodology for testing channel compliance), and shall meet the received eye mask as 19
specified in [13], [10], [11] or [12] as required. 20
21
Also refer to Appendix 3.A for more information on the channel characteristics. 22
23
Reference Transmitter: 24
1. No emphasis 25
26
2. A concatenated first order low pass transmit filter with 0.75 times baud rate 27
3. An amplitude equal to the defined minimum transmit amplitude in the specific 28
Implementation Agreement 29
30
4. A jitter distribution equal to the defined maximum allowed transmit jitter in the 31
specific Implementation Agreement 32
5. Worst case transmitter return loss described as a parallel RC elements, see 2.E.6. 33
34
Reference Receiver: 35
36
1. No sampling jitter 37
2. No equalisation 38
39
3. A sampling point defined at the midpoint between the average zero crossings of the 40
differential signal 41
4. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 42
43
5. A BER as per [13]. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 4: SxI-5 Interface 139

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 4.3 Electrical Characteristics

2
3 Refer to [13] for detailed information on SxI-5, [10] for detailed infromation on SFI-4.2,
4 [11] for detatiled information on SFI-5.1 and [12] for detailed information on SPI-5.1.
5
6 Note these implementation agreements require that one drop the high frequency jitter
7 tolerance number by 0.1UI for the addition of the sinusoidal jitter.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

140 Clause 4: SxI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

4.A Appendix - StatEye.org Template 1

2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 3
4
% example template for setting up a standard, i.e. equaliser 5
% jitter and return loss 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
param.version = [param.version ’_v1.0’]; 10
11
% these are internal variables and should not be changed 12
13
param.scanResolution = 0.010; 14
param.binsize = 0.0005; 15
param.points = 2^13; 16
17
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 18
19
% set the transmitter and baud rate. The tx filter has two 20
% parameters defined for the corner frequency of the poles 21
22
%param.bps = 2.488e9; % lower rate SxI-5 23
param.bps = 3.125e9; 24
param.bitResolution = 1/(4*param.bps); 25
param.txFilter = ’singlepole’; 26
param.txFilterParam = [0.75]; 27
28
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 29
30
% set the return loss up. The return loss can be turned off 31
% using the appropriate option 32
33
param.returnLoss = ’on’; 34
param.cpad = 2.25; 35
36
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 37
38
% set the transmitter emphasis up. Some example setting are 39
% included which can be uncommented 40
41
% single tap emphasis 42
param.txpre = []; 43
param.signal = 1.0; 44
param.txpost = []; 45
param.vstart = [-0.3 -0.3]; 46
param.vend = [+0.0 +0.0]; 47
param.vstep = [0.1 0.05 0.025]; 48
49

Optical Internetworking Forum - Clause 4: SxI-5 Interface 141

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2
3 % set the de-emphasis of 4-point transmit pulse
4 % the de-emphasis run if param.txpre = [] and param.txpost = []
5
6 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
7
8 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9
10 % set the data coding changing the transmit pulse spectrum
11 % the coding run if param.txpre = [] and param.txpost = []
12
13 param.datacoding = 1; % the coding is off
14
15 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16
17 % set PAM amplitude and rate
18
19 param.PAM = 2; % PAM is swithed off
20
21 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22
23 % the rxsample point does not need to be changed as it is
24 % automatically adjusted by the optimisation scripts.
25 % The number of DFE taps should be set, however, the initial
26 % conditions are irrelevant.
27
28 param.rxsample = -0.1;
29
30 % no DFE
31 param.dfe = [];
32
33 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
34
35 % sampling jitter in HPJpp and GJrms is defined here
36
37 param.txdj = 0.17;
38 param.txrj = 0.18/(2*7.04);
39
40 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
41
42 % the following options are not yet implemented and should
43 % not be changed
44
45 param.user = [0.0];
46 param.useuser = ’no’;
47 param.usesymbol = ’’;
48 param.xtAmp = 1.0;
49

142 Clause 4: SxI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1
2
param.TransmitAmplitude = 0.500; % mVppdif 3
param.MinEye = 0.175; % mVppdif 4
5
param.Q = 2*704; 6
param.maxDJ = 0.20; 7
param.maxTJ = 0.56; 8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 4: SxI-5 Interface 143

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

144 Clause 4: SxI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

5 TFI-5 Interface 1
2
3
5.1 Introduction 4
5
This clause details the requirements for the TFI-5 electrical interface. 6
7
8
5.2 General Requirements 9
10
This clause uses “Method B” of the “Jitter and Interoperability Methodology” section. 11
12
13
5.2.1 Channel Compliance 14
15
As per 2.2.2, with the following reference transmitter and reference receiver (note these 16
conditions do not specify any required implementation but rather indicate a 17
methodology for testing channel compliance), and shall meet the received eye mask as 18
specified in [4]. 19
20
Also refer to Appendix 3.A for more information on the channel characteristics. 21
22
Reference Transmitter: 23
1. A single post tap transmitter, with  3dB of emphasis and infinite precision 24
accuracy. 25
26
2. A maximum amplitude equal to the defined minimum transmit amplitude in the 27
specific Implementation Agreement 28
3. A jitter distribution equal to the defined maximum allowed transmit jitter in the 29
specific Implementation Agreement 30
31
4. At the maximum baud rate as defined by the specific Implementation Agreement 32
5. Worst case transmitter return loss described as a parallel RC elements, see 2.E.6. 33
34
6. A concatenated first order low pass transmit filter with 0.75 times baud rate. 35
36
Reference Receiver: 37
1. No sampling jitter 38
39
2. No equalisation 40
3. A sampling point defined at the midpoint between the average zero crossings of the 41
differential signal 42
43
4. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 44
5. A BER as per [4]. 45
46
47
48
49

Optical Internetworking Forum - Clause 5: TFI-5 Interface 145

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 5.3 Electrical Characteristics

2
3 Refer to [4] for detailed information on TFI-5.
4
5 Note this implementation agreement requires that one drop the high frequency jitter
6 tolerance number by 0.1UI for the addition of the sinusoidal jitter.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

146 Clause 5: TFI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

5.A Appendix - StatEye.org Template 1

2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 3
4
% example template for setting up a standard, i.e. equaliser 5
% jitter and return loss 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
param.version = [param.version ’_v1.0’]; 10
11
% these are internal variables and should not be changed 12
13
param.scanResolution = 0.010; 14
param.binsize = 0.0005; 15
param.points = 2^13; 16
17
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 18
19
% set the transmitter and baud rate. The tx filter has two 20
% parameters defined for the corner frequency of the poles 21
22
23
%param.bps = 2.488e9; % lower rate TFI-5 24
param.bps = 3.11e9; 25
param.bitResolution = 1/(4*param.bps); 26
param.txFilter = ’singlepole’; 27
param.txFilterParam = [0.75]; 28
29
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 30
31
% set the return loss up. The return loss can be turned off 32
% using the appropriate option 33
34
param.returnLoss = ’on’; 35
param.cpad = 2.25; 36
37
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 38
39
% set the transmitter emphasis up. Some example setting are 40
% included which can be uncommented 41
42
% single tap emphasis 43
param.txpre = []; 44
param.signal = 1.0; 45
param.txpost = [-0.1]; 46
param.vstart = [-0.3 -0.3]; 47
param.vend = [+0.0 +0.0]; 48
param.vstep = [0.1 0.05 0.025]; 49

Optical Internetworking Forum - Clause 5: TFI-5 Interface 147

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3
4 % set the de-emphasis of 4-point transmit pulse
5 % the de-emphasis run if param.txpre = [] and param.txpost = []
6
7 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
8
9 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
10
11 % set the data coding changing the transmit pulse spectrum
12 % the coding run if param.txpre = [] and param.txpost = []
13
14 param.datacoding = 1; % the coding is off
15
16 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17
18 % set PAM amplitude and rate
19
20 param.PAM = 2; % PAM is swithed off
21
22 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23
24 % the rxsample point does not need to be changed as it is
25 % automatically adjusted by the optimisation scripts.
26 % The number of DFE taps should be set, however, the initial
27 % conditions are irrelevant.
28
29 param.rxsample = -0.1;
30
31 % no DFE
32 param.dfe = [];
33
34 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
35
36 % sampling jitter in HPJpp and GJrms is defined here
37
38 param.txdj = 0.175;
39 param.txrj = 0.175/(2*7.04);
40
41 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42
43 % the following options are not yet implemented and should
44 % not be changed
45
46 param.user = [0.0];
47 param.useuser = ’no’;
48 param.usesymbol = ’’;
49 param.xtAmp = 1.0;

148 Clause 5: TFI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 2
3
param.TransmitAmplitude = 0.350; % mVppdif 4
param.MinEye = 0.175; % mVppdif 5
6
param.Q = 2*7.04; 7
param.maxDJ = 0.37; 8
param.maxTJ = 0.65; 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 5: TFI-5 Interface 149

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

150 Clause 5: TFI-5 Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

6 CEI-6G-SR Short Reach Interface 1

2
3
6.1 Introduction 4
5
This clause details the requirements for the CEI-6G-SR short-reach high speed 6
electrical interface between nominal baud rates of 4.976Gsym/s to 6.375Gsym/s using 7
NRZ coding (hence 1 bit per symbol at the electrical level). A compliant device must 8
meet all of the requirements listed below. The electrical interface is based on high 9
speed, low voltage logic with nominal differential impedance of 100. Connections are 10
point-to-point balanced differential pair and signalling is unidirectional. 11
12
The electrical IA is based on loss & jitter budgets and defines the characteristics 13
required to communicate between a CEI-6G-SR driver and a CEI-6G-SR receiver using 14
copper signal traces on a printed circuit board. The characteristic impedance of the 15
signal traces is nominally 100 differential. These characteristics are normative for the 16
devices and informative for the channel. Rather than specifying materials, channel 17
components, or configurations, the IA focuses on effective channel characteristics. 18
Hence a short length of poorer material should be equivalent to a longer length of 19
premium material. A ‘length’ is effectively defined in terms of its attenuation rather than 20
physical length. 21
22
Short reach CEI-6G-SR devices from different manufacturers shall be inter-operable. 23
24
25
6.2 Requirements 26
27
1. Support serial baud rate from 4.976Gsym/s to 6.375Gsym/s. 28
29
2. Capable of low bit error rate (required BER of 10-15). 30
3. Capable of driving 0 – 200mm of PCB and up to 1 connector. 31
32
4. Shall support AC coupled operation and optionally DC-coupled operation. 33
5. Shall allow multi-lanes (1:N). 34
35
6. Shall support hot plug. 36
37
38
6.3 General Requirements 39
40
This clause uses “Method B” of the Jitter and Interoperability Methodology section. 41
42
6.3.1 Data Patterns 43
44
Please refer to 3.2.1 45
46
47
48
49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 151

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 6.3.2 Signal levels

2
3 Please refer to 3.2.2 and 6.4.1.
4
5 6.3.3 Signal Definitions
6
7 Please refer to 1.A
8
9
10 6.3.4 Bit Error Ratio
11
12 Please refer to 3.2.3
13
14 6.3.5 Ground Differences
15
16 Please refer to 3.2.4
17
18 6.3.6 Cross Talk
19
20 Please refer to 3.2.5
21
22
23 6.3.7 Channel Compliance
24 As per 2.2.2, with the following reference transmitter and reference receiver (note these
25 conditions do not specify any required implementation but rather indicate a
26 methodology for testing channel compliance), and shall meet the received eye mask as
27 specified in Figure 1-5 and Table 6-8.
28
29 Also refer to Appendix 3.A for more information on the channel characteristics.
30
31 Reference Transmitter:
32
33 1. A single post tap transmitter, with  3dB of emphasis and infinite precision
34 accuracy.
35 2. A transmit amplitude of 400mVppd
36
37 3. Additional Uncorrelated Bounded High Probability Jitter of 0.15UIpp (emulating part
38 of the Tx jitter)
39 4. Additional Uncorrelated Unbounded Gaussian Jitter of 0.15UIpp (emulating part of
40 the Tx jitter)
41
42 5. A Tx edge rate filter: simple 20dB/dec low pass at 75% of baud rate, this is to
43 emulate a Tx -3dB bandwidth at 3/4 baud rate.
44
6. At the maximum baud rate that the channel is to operate at or 6.375Gsym/s which
45
ever is the lowest.
46
47 7. Worst case transmitter return loss described as a parallel RC elements, see 2.E.6.
48
49 Reference Receiver:

152 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1. No Rx equalization and the Rx bandwidth is assumed to be infinite. 1

2
2. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 3
3. A BER as per 6.3.4. 4
5
4. A sampling point defined at the midpoint between the average zero crossings of the 6
differential signal 7
8
9
6.4 Electrical Characteristics 10
11
The electrical interface is based on high speed, low voltage logic with nominal 12
differential impedance of 100. Connections are point-to-point balanced differential 13
pair and signalling is unidirectional. 14
15
6.4.1 Driver Characteristics 16
17
The key driver characteristics are summarized in Table 6-1 and Table 6-2 while the 18
following sub-clauses fully detail all the requirements. 19
20
21
22
Table 6-1. CEI-6G-SR Transmitter Output Electrical Specifications
23
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 24
Baud Rate T_Baud See 6.4.1.2 4.976 6.375 Gsym/s
25
26
Output Differential voltage 
(into floating load Rload=100)
T_Vdiff See 6.4.1.3 400 750 mVppd 27
Differential Resistance T_Rd See 6.4.1.5 80 100 120 
28
29
Recommended output rise and fall times
(20% to 80%)
T_tr, T_tf See 6.4.1.4 30 ps 30
Differential Output Return Loss 
31
(100MHz to 0.75*T_Baud)
-8 dB
32
T_SDD22 See 6.4.1.5
Differential Output Return Loss  33
(0.75*T_Baud to T_Baud) 34
Common Mode Return Loss
T_SCC22 See 6.4.1.5 -6 dB
35
(100MHz to 0.75 *T_Baud) 36
Transmitter Common Mode Noise T_Ncm
5% of
mVppd 37
T_Vdiff
38
NOTES:
1. For all Load Types: R_Rdin = 100 20. For Vcm definition, see Figure 1-1
39
2. Load Type 0 with min T_Vdiff, AC-Coupling or floating load. 40
3. For Load Types 1 through 3: R_Zvtt  30Vtt is defined for each load type as follows: Load Type 1 R_Vtt = 1.2V +5%/-8%; 41
Load Type 2 R_Vtt = 1.0V +5%/-8%; Load Type 3 R_Vtt = 0.8V +5%/-8%.
4. DC Coupling compliance is optional (Type 1 through 3). Only Transmitters that support DC coupling are required to meet this 42
parameter. It is acceptable for a Transmitter to restrict the range of T_Vdiff in order to comply with the specified T_Vcm range. 43
For a Transmitter which supports multiple T_Vdiff levels, it is acceptable for a Transmitter to claim DC Coupling Compliance if
it meets the T_Vcm ranges for at least one of its T_Vdiff setting as long as those setting(s) that are compliant are indicated. 44
5. Simple CML Transmitters designed using Vdd  1.2V may still claim DC compliance if this parameter is not met. 45
6. Simple CML Transmitters designed using Vdd  0.8V may still claim DC compliance if this parameter is not met.
46
47
48
49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 153

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 6-1. CEI-6G-SR Transmitter Output Electrical Specifications

2 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
3
4 Load Type 0
0.0 1.8 V
Note 2
5
6 Output Common Mode Voltage Load Type 1
735 1135 mV
7 See Note 1, 3, 4 T_Vcm Note 6

8 Also see 3.2.2 Load Type 2 550 1060 mV

9 Load Type 3
490 850 mV
10 Note 5
11 NOTES:
12 1. For all Load Types: R_Rdin = 100 20. For Vcm definition, see Figure 1-1
2. Load Type 0 with min T_Vdiff, AC-Coupling or floating load.
13 3. For Load Types 1 through 3: R_Zvtt  30Vtt is defined for each load type as follows: Load Type 1 R_Vtt = 1.2V +5%/-8%;
14 Load Type 2 R_Vtt = 1.0V +5%/-8%; Load Type 3 R_Vtt = 0.8V +5%/-8%.
15 4. parameter.
DC Coupling compliance is optional (Type 1 through 3). Only Transmitters that support DC coupling are required to meet this
It is acceptable for a Transmitter to restrict the range of T_Vdiff in order to comply with the specified T_Vcm range.
16 For a Transmitter which supports multiple T_Vdiff levels, it is acceptable for a Transmitter to claim DC Coupling Compliance if
17 5. itSimple
meets the T_Vcm ranges for at least one of its T_Vdiff setting as long as those setting(s) that are compliant are indicated.
CML Transmitters designed using Vdd  1.2V may still claim DC compliance if this parameter is not met.
18 6. Simple CML Transmitters designed using Vdd  0.8V may still claim DC compliance if this parameter is not met.
19
20
21
22 Table 6-2. CEI-6G-SR Transmitter Output Jitter Specifications
23 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
24
Uncorrelated High Probability Jitter T_UHPJ See 6.4.1.8 0.15 UIpp
25
26 Duty Cycle Distortion T_DCD See 6.4.1.8 0.05 UIpp

27 Total Jitter T_TJ See 6.4.1.8 0.30 UIpp

28 Eye Mask T_X1 See 6.4.1.8 0.15 UI
29 Eye Mask T_X2 See 6.4.1.8 0.40 UI
30 Eye Mask T_Y1 See 6.4.1.8 200 mV
31 Eye Mask T_Y2 See 6.4.1.8 375 mV
32
33 NOTES:
34
35 6.4.1.1 Driver Test Load
36
37 Please refer to 3.2.6
38
39 6.4.1.2 Driver Baud Rate
40
41 All devices shall work from 4.976Gsym/s to the maximum baud rate specified for the
42 device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific
43 protocols will define the operating baud rate without affecting CEI compliance.
44
45
46
47
48
49

154 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

6.4.1.3 Driver Amplitude and Swing 1

2
Driver differential output amplitude shall be between 400 to 750mVppd either with or 3
without any transmit emphasis. Absolute driver output voltage shall be between -0.1V 4
and 1.9V with respect to local ground. See Figure 1-1 for an illustration of absolute 5
driver output voltage limits and definition of differential peak-to-peak amplitude. 6
7
6.4.1.4 Driver Rise and Fall Times 8
9
The recommended minimum differential rise and fall times are 30ps as measured 10
between the 20% and 80% of the maximum measured levels; the maximum differential 11
rise and fall times are defined by the Tx eye diagram (Figure 1-4 and Table 6-4). 12
Shorter rise and fall times may result in excessive high frequency components and 13
increase EMI and cross talk. 14
15
16
17
18
6.4.1.5 Driver Resistance and Return Loss 19
As per 3.2.10, with the following parameters. 20
21
Table 6-3. CEI-6G-SR Driver Return Loss Parameters 22
23
Parameter Value Units
24
A0 -8 dB 25
f0 100 MHz 26
3 27
f1 T_Baud  --- Hz
4 28
f2 T_Baud Hz 29
Slope 16.6 dB/dec 30
31
32
33
34
6.4.1.6 Driver Lane-to-Lane Skew 35
36
Please refer to 3.2.7 37
38
6.4.1.7 Driver Short Circuit Current 39
40
Please refer to 3.2.9 41
42
6.4.1.8 Driver Template and Jitter 43
44
As per 2.2.3 for a BER as per 6.3.4, the driver shall satisfy both the near-end and far- 45
end eye template and jitter requirements as given in Figure 1-4, Table 6-4, Figure 1-5 46
and Table 6-8 either with or without any transmit emphasis. 47
48
The maximum near-end duty cycle distortion (T_DCD) shall be less than 0.05UIpp. 49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 155

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 It should be noted that it is assumed the Uncorrelated High Probability Jitter component
2 of the driver jitter is not Inter-symbol Interference (ISI). This is only assumed from a
3 receiver point of view and does not in any way put any restrictions on the real driver
4 HPJ.
5
6 Table 6-4. CEI-6G-SR Near-End (Tx) Template Intervals
7 Characteristics Symbol Near-End Value Units
8 Eye Mask T_X1 0.15 UI
9
Eye Mask T_X2 0.40 UI
10
11 Eye Mask T_Y1 200 mV

12 Eye Mask T_Y2 375 mV

13 Uncorrelated Bounded High Probability Jitter T_UBHPJ 0.15 UIpp
14 Duty Cycle Distortion T_DCD 0.05 UIpp
15 Total Jitter T_TJ 0.30 UIpp
16
17
18 6.4.1.9 Driver Training Pattern
19
20 There is no requirement at the electrical level for a training pattern, however there may
21 be a training pattern requirement(s) at the protocol level.
22
23 6.4.2 Receiver Characteristics
24
25 The key receiver characteristics are summarized in Table 6-5 and Table 6-6 while the
26 following sub-clauses fully detail all the requirements.
27
28 Table 6-5. CEI-6G-SR Receiver Electrical Input Specifications
29 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
30
Rx Baud Rate R_Baud See 6.4.2.1 4.976 6.375 Gsym/s
31
32 Input Differential voltage R_Vdiff See 6.4.2.3 125 750 mVppd

33 Differential Resistance R_Rdin See 6.4.2.7 80 100 120 

34 Bias Voltage Source Impedance  R_Zvtt See Note 1 30 
35 (load types 1 to 3)
36 Differential Input Return Loss -8 dB
(100MHz to 0.75*R_Baud)
37 R_SDD11 See 6.4.2.7
Input Return Loss
38 Differential
(0.75*R_Baud to R_Baud))
39
Common mode Input Return Loss
40 (100MHz to 0.75 *R_Baud) R_SCC11 See 6.4.2.7 -6 dB
41 NOTES:
42 1. DC Coupling compliance is optional. For Vcm definition, see Figure 1-1
43 2. Receiver is required to implement at least one of specified nominal R_Vtt values, and typically implements only one of these
values. Receiver is only required to meet R_Vrcm parameter values that correspond to R_Vtt values supported.
44 3. Input common mode voltage for AC-coupled or floating load input with min T_Vdiff,
45 4. For floating load, input resistance must be 1k.
46
47
48
49

156 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 6-5. CEI-6G-SR Receiver Electrical Input Specifications 1

Characteristic Symbol Condition MIN. TYP. MAX. UNIT
2
3
R_Vtt floating,
Note 4 Not Specified V 4
5
R_Vtt = 1.2V
Nominal
1.2 - 8% 1.2 + 5% V 6
Termination Voltage
Note 1, 2
R_Vtt
R_Vtt = 1.0V
7
Nominal
1.0 - 8% 1.0 + 5% V 8
R_Vtt = 0.8V
9
Nominal
0.8 - 8% 0.8 + 5% V 10
R_Vtt floating, 11
-0.05 1.85 V
Note 3, 4 12
R_Vtt = 1.2V R_Vtt - 13
720 mV
Input Common Mode Voltage Nominal 10 14
R_Vrcm
Note 1, 2 R_Vtt = 1.0V
535
R_Vtt +
mV
15
Nominal 125 16
R_Vtt = 0.8V
475
R_Vtt +
mV
17
Nominal 105
18
Wander divider (in Figure 2-30 & Figure 2-31) n 10 19
NOTES: 20
1. DC Coupling compliance is optional. For Vcm definition, see Figure 1-1
2. Receiver is required to implement at least one of specified nominal R_Vtt values, and typically implements only one of these
21
values. Receiver is only required to meet R_Vrcm parameter values that correspond to R_Vtt values supported. 22
3. Input common mode voltage for AC-coupled or floating load input with min T_Vdiff, 23
4. For floating load, input resistance must be 1k.
24
25
26
Table 6-6. CEI-6G-SR Receiver Input Jitter Tolerance Specifications 27
28
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
29
Bounded High Probability Jitter R_BHPJ See 6.4.2.8 0.45 UIpp 30
Sinusoidal Jitter, maximum R_SJ-max See 6.4.2.8 5 UIpp 31
Sinusoidal Jitter, High Frequency R_SJ-hf See 6.4.2.8 0.05 UIpp 32
Total Jitter (Does not include Sinusoidal Jitter) R_TJ See 6.4.2.8 0.60 UIpp
33
34
Eye Mask R_X1 See 6.4.2.8 0.30 UI
35
Eye Mask R_Y1 See 6.4.2.8 62.5 mV 36
Eye Mask R_Y2 See 6.4.2.8 375 mV 37
NOTES: 38
39
6.4.2.1 Input Baud Rate 40
41
All devices shall work from 4.976Gsym/s to the maximum baud rate specified for the 42
device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific 43
protocols will define the operating baud rate without affecting CEI compliance. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 157

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 6.4.2.2 Reference Input Signals

2
3 Reference input signals to the receiver have the characteristics determined by
4 compliant driver. The reference input signal must satisfy the transmitter near-end
5 template and jitter given in Figure 1-4 and Table 6-4, as well as the far-end eye
6 template and jitter given in Figure 1-5 and Table 6-8, with the differential load
7 impedance of 1001% at DC with a return loss of better than 20dB from baud rate
8 divided by 1667 to 1.5 times the baud rate. Note that the input signal might not meet
9 either of these templates when the actual receiver replaces this load.
10
11 6.4.2.3 Input Signal Amplitude
12
13 The receiver shall accept differential input signal amplitudes produced by compliant
14 transmitters connected without attenuation to the receiver. This may be larger than the
15 750mVppd maximum of the driver due to output/input impedances and reflections.
16
17 The minimum input amplitude is defined by the far-end driver template, the actual
18 receiver input impedance and the loss of the actual PCB. Note that the far-end driver
19 template is defined using a well controlled load impedance, however the real receiver is
20 not, which can leave the receiver input signal smaller than the minimum 125mVppd.
21
22
23 6.4.2.4 Absolute Input Voltage
24 The absolute voltage levels with respect to the receiver ground at the input of the
25 receiver are dependent on the driver implementation, the inter-ground difference,
26 whether the receiver is AC or DC coupled, and (in the case of DC coupling load types 1
27 to 3) the nominal R_Vtt supported by the receiver. The voltage levels at the input of a
28 DC coupled receiver shall be consistent with R_Vrcm and R_Vdiff values defined in
29 Table 6-5.
30
31 The voltage levels at the input of an AC coupled receiver (if AC coupling is done within
32 the receiver) or at the Tx side of the external AC coupling cap (if AC coupling is done
33 externally) shall be between -0.15 to 1.95V with respect to local ground.
34
35
36 6.4.2.5 Input Common Mode Impedance
37
38 The input common mode impedance (R_Zvtt) at the input of the receiver is dependent
39 on whether the receiver is AC or DC coupled. The value of R_Zvtt as measured at the
40 input of an AC coupled receiver is undefined. The value of R_Zvtt as measured at the
41 input of a DC coupled receiver is defined as per Table 6-5.
42
43 If AC coupling is used, it is to be considered part of the receiver for the purposes of this
44 specification unless explicitly stated otherwise. It should be noted that various methods
45 for AC coupling are allowed (for example, internal to the chip or done externally). See
46 also 3.2.12 for more information.
47
48
49

158 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

6.4.2.6 Input Lane-to-Lane Skew 1

2
Please refer to 3.2.8 3
4
6.4.2.7 Input Resistance and Return Loss 5
6
Please refer to 3.2.10 with the following parameters. 7
8
Table 6-7. CEI-6G-SR Input Return Loss Parameters 9
Parameter Value Units
10
11
A0 -8 dB
12
f0 100 MHz 13
f1
3
R_Baud  --- Hz 14
4 15
f2 R_Baud Hz 16
Slope 16.6 dB/dec 17
18
19
20
6.4.2.8 Input Jitter Tolerance 21
22
As per 2.2.4, the receiver shall tolerate at least the far-end eye template and jitter 23
requirements as given in Figure 1-5 and Table 6-8 with an additional SJ with any 24
frequency and amplitude defined by the mask of Figure 2-4 where the minimum & 25
maximum total wander amplitude are 0.05UIpp & 5UIpp respectively. This additional SJ 26
component is intended to ensure margin for wander, hence is over and above any high 27
frequency jitter from Table 6-8. 28
29
30
31
Table 6-8. CEI-6G-SR Far-End (Rx) Template Intervals 32
Far-End
33
Characteristics Symbol
Value
Units
34
Eye Mask R_X1 0.30 UI
35
36
Eye Mask R_Y1 62.5 mV
37
Eye Mask R_Y2 375 mV
38
Uncorrelated Bounded High Probability Jitter R_UBHPJ 0.15 UIpp 39
Correlated Bounded High Probability Jitter R_CBHPJ 0.30 UIpp 40
Total Jitter (Does not include Sinusoidal Jitter) R_TJ 0.60 UIpp 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 159

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 6.A Appendix - Link and Jitter Budgets

2
3 The primary intended application is as a point-to-point interface of up to approximately
4 200mm (8”) and up to one connector between integrated circuits using controlled
5 impedance traces on low-cost printed circuit boards (PCBs). Informative loss and jitter
6 budgets are presented in Table 6-9 (see also Appendix 3.A for more information) to
7 demonstrate the feasibility of legacy FR4 epoxy PCB’s. The jitter budget is given in
8 Table 6-10. The performance of an actual transceiver interconnect is highly dependent
9 on the implementation.
10
11
12
13 Table 6-9. CEI-6G-SR Informative Loss, Skew and Jitter Budget
14 Differential
Bounded High
15 Loss (dB)
Skew (ps)
Probability TJ (UIpp)
(UIpp)
16
Driver 0 15 0.15 0.30
17
18 Interconnect (with Connector) 6.6
25
0.15 0.15
19 Other 3.5 0.15 0.15
20 Total 10.1 40 0.45 0.60
21
22
23 Table 6-10. CEI-6G-SR High Frequency Jitter Budget
24 Uncorrelated Jitter Correlated Jitter Total Jitter
25 Bounded Bounded
CEI-6G-SR Unbounded High Bounded Amplitude
26 Gaussian Probability Gaussian
High Gaussian Sinusoidal High Total
Probability Probability
27
Abbreviation UUGJ UHPJ CBGJ CBHPJ GJ SJ HPJ TJ k
28
Unit UIpp UIpp UIpp UIpp UIpp UIpp UIpp UIpp mVppd
29 -0.200
30 Transmitter 0.150 0.150
See 1
0.150 -0.050 0.100 400.0

31 Channel 0.500
32 Receiver Input 0.150 0.150 0.000 0.300 0.150 0.450 0.600 0.25 125
33 Clock + Sampler 0.150 0.100 0.100 -50.0
34 Budget 0.212 0.250 0.000 0.400 0.212 0.050 0.650 0.912 0.13 75.0
35 NOTES:
1. Due to transmitter emphasis, it reduces the ISI as seen at the receiver. Thus this number is negative
36
37
38
39
40
41
42
43
44
45
46
47
48
49

160 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

6.B Appendix - StatEye.org Template 1

2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 3
4
% example template for setting up a standard, i.e. equaliser 5
% jitter and return loss 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
param.version = [param.version ’_v1.0’]; 10
11
% these are internal variables and should not be changed 12
13
param.scanResolution = 0.01; 14
param.binsize = 0.0005; 15
param.points = 2^13; 16
17
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 18
19
% set the transmitter and baud rate. The tx filter has two 20
% parameters defined for the corner frequency of the poles 21
22
param.bps = 6.375e9; 23
param.bitResolution = 1/(4*param.bps); 24
param.txFilter = ’singlepole’; 25
param.txFilterParam = [0.75]; 26
27
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 28
29
% set the return loss up. The return loss can be turned off 30
% using the appropriate option 31
32
param.returnLoss = ’on’; 33
param.cpad = 1.0; 34
35
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 36
37
% set the transmitter emphasis up. Some example setting are 38
% included which can be uncommented 39
40
% single tap emphasis 41
param.txpre = []; 42
param.signal = 1.0; 43
param.txpost = [-0.1]; 44
param.vstart = [-0.3 -0.3]; 45
param.vend = [+0.0 +0.0]; 46
param.vstep = [0.1 0.05 0.025]; 47
48
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 161

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 % set the de-emphasis of 4-point transmit pulse
3 % the de-emphasis run if param.txpre = [] and param.txpost = []
4
5 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
6
7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8
9 % set the data coding changing the transmit pulse spectrum
10 % the coding run if param.txpre = [] and param.txpost = []
11
12 param.datacoding = 1; % the coding is off
13
14 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15
16 % set PAM amplitude and rate
17
18 param.PAM = 2; % PAM is swithed off
19
20 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
21
22 % the rxsample point does not need to be changed as it is
23 % automatically adjusted by the optimisation scripts.
24 % The number of DFE taps should be set, however, the initial
25 % conditions are irrelevant.
26
27 param.rxsample = -0.1;
28
29 % no DFE
30 param.dfe = [];
31
32 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
33
34 % sampling jitter in HPJpp and GJrms is defined here
35
36 param.txdj = 0.15;
37 param.txrj = 0.15/(2*7.94);
38
39 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40
41 % the following options are not yet implemented and should
42 % not be changed
43
44 param.user = [0.0];
45 param.useuser = ’no’;
46 param.usesymbol = ’’;
47 param.xtAmp = 1.0;
48
49 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

162 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
param.TransmitAmplitude = 0.400; % mVppdif 2
param.MinEye = 0.125; % mVppdif 3
4
param.Q = 2*7.94; 5
param.maxDJ = 0.30; 6
param.maxTJ = 0.60; 7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 6: CEI-6G-SR Short Reach Interface 163

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

164 Clause 6: CEI-6G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7 CEI-6G-LR Long Reach Interface 1

2
3
7.1 Introduction 4
5
This clause details the requirements for the CEI-6G-LR long-reach high speed 6
electrical interface between nominal baud rates of 4.976Gsym/s to 6.375Gsym/s using 7
NRZ coding (hence 1 bit per symbol at the electrical level). A compliant device must 8
meet all of the requirements listed below. The electrical interface is based on high 9
speed, low voltage logic with nominal differential impedance of 100. Connections are 10
point-to-point balanced differential pair and signalling is unidirectional. 11
12
The electrical IA is based on loss & jitter budgets and defines the characteristics 13
required to communicate between a CEI-6G-LR driver and a CEI-6G-LR receiver using 14
copper signal traces on a printed circuit board. The characteristic impedance of the 15
signal traces is nominally 100 differential. These characteristics are normative for the 16
devices and informative for the channel. Rather than specifying materials, channel 17
components, or configurations, the IA focuses on effective channel characteristics. 18
Hence a short length of poorer material should be equivalent to a longer length of 19
premium material. A ‘length’ is effectively defined in terms of its attenuation rather than 20
physical length. 21
22
Long reach CEI-6G-LR devices from different manufacturers shall be inter-operable. 23
24
25
7.2 Requirements 26
27
1. Support serial baud rate from 4.976Gsym/s to 6.375Gsym/s. 28
29
2. Capable of low bit error rate (required BER of 10-15). 30
3. Capable of driving 0 – 1m of PCB (such as IEEE 802.3 XAUI/TFI-5 compliant 31
backplane) and up to 2 connector. 32
33
4. Shall support AC coupled operation and optionally DC-coupled operation. 34
5. Shall allow multi-lanes (1:N). 35
36
6. Shall support hot plug. 37
38
39
7.3 General Requirements 40
41
This clause uses “Method D” of the Jitter and Interoperability Methodology section. 42
43
7.3.1 Data Patterns 44
45
Please refer to 3.2.1 46
47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 165

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7.3.2 Signal levels

2
3 Please refer to 3.2.2 and 7.4.1.
4
5 7.3.3 Signal Definitions
6
7 Please refer to 1.A
8
9
10 7.3.4 Bit Error Ratio
11
12 Please refer to 3.2.3
13
14 7.3.5 Ground Differences
15
16 Please refer to 3.2.4
17
18 7.3.6 Cross Talk
19
20 Please refer to 3.2.5
21
22
23 7.3.7 Channel Compliance
24 As per 2.4.2, with the following reference transmitter and reference receiver (note these
25 conditions do not specify any required implementation but rather indicate a
26 methodology for testing channel compliance), and shall meet the equalized eye mask
27 as specified in Figure 1-5 and Table 7-1. However for the case of a short reach Tx
28 talking to a long reach Rx, the Rx needs to meet all requirements as given in 6.3.7 and
29 6.4.2.
30
31 Also refer to Appendix 3.A for more information on the channel characteristics.
32
33 Reference Transmitter:
34
35 1. Either a single pre or post tap transmitter, with  6dB of emphasis, with infinite
36 precision accuracy.
37 2. A transmit amplitude of 800mVppd.
38
39 3. Additional Uncorrelated Bounded High Probability Jitter of 0.15UIpp (emulating part
40 of the Tx jitter)
41 4. Additional Uncorrelated Unbounded Gaussian Jitter of 0.15UIpp (emulating part of
42 the Tx jitter)
43
44 5. A Tx edge rate filter: simple 40dB/dec low pass at 75% of baud rate, this is to
45 emulate both Rx and Tx -3dB bandwidths at 3/4 baud rate.
46
6. At the maximum baud rate that the channel is to operate at or 6.375Gsym/s which
47
ever is lowest
48
49 7. Worst case transmitter return loss described as a parallel RC elements, see 2.E.6.

166 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Reference Receiver: 1
2
1. Rx equalization: 5 tap DFE, with infinite precision accuracy and having the following 3
restriction on the coefficient values:  4
 5
Let W[N] be sum of DFE tap coefficient weights from taps N through M where
6
 7
N = 1 is previous decision (i.e. first tap) 8
M = oldest decision (i.e. last tap) 9
R_Y2 = T_Y2 = 400mV 10
Y = min(R_X1, (R_Y2 - R_Y1) / R_Y2) = 0.30 11
Z = 2/3 = 0.66667 12
 13
Then W[N]  Y * Z(N - 1) 14
 15
For the channel compliance model the number of DFE taps (M) = 5. This gives the 16
following maximum coefficient weights for the taps: 17

18
W[1]  0.3000 (sum of taps 1 to 5) 19
W[2]  0.2000 (sum of taps 2 to 5) 20
W[3]  0.1333 (sum of taps 3 to 5) 21
W[4]  0.0889 (sum of taps 4 and 5)  22
W[5] 0.0593 (tap 5) 23

24
Notes: 25
- These coefficient weights are absolute assuming a T_Vdiff of 1Vppd 26
- For a real receiver the restrictions on tap coefficients would apply for the actual 27
number of DFE taps implemented (M) 28
2. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 29
30
3. A BER as per 3.2.3. 31
32
33
Table 7-1. CEI-6G-LR Receiver Equalization Output Eye Mask 34
35
Parameter Symbol Max Units
36
Eye mask R_X1 0.3 UI 37
Eye mask R_Y1 50 mV 38
Bounded High Probability Jitter R_BHPJ 0.325 UI 39
40
41
7.4 Electrical Characteristics 42
43
The electrical interface is based on high speed, low voltage logic with nominal 44
differential impedance of 100. Connections are point-to-point balanced differential 45
pair and signalling is unidirectional. 46
47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 167

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7.4.1 Driver Characteristics

2
3 The key driver characteristics are summarized in Table 7-2 and Table 7-3 while the
4 following sub-clauses fully detail all the requirements.
5
6
7
Table 7-2. CEI-6G-LR Transmitter Output Electrical Specifications
8
9 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
10 Baud Rate T_Baud See 7.4.1.2 4.976 6.375 Gsym/s
11 Output Differential voltage  See 7.4.1.3 &
12 (into floating load Rload=100) T_Vdiff
Note 1
800 1200 mVppd

13 Differential Resistance T_Rd See 7.4.1.5 80 100 120 

14 Recommended output rise and fall times
15 (20% to 80%) T_tr, T_tf See 7.4.1.4 30 ps

16 Differential Output Return Loss 

-8 dB
17 (100MHz to 0.75*T_Baud)
T_SDD22 See 7.4.1.5
18 Differential Output Return Loss 
19 (0.75*T_Baud to T_Baud)
20 Common Mode Return Loss T_S11 See 7.4.1.5 -6 dB
21 (100MHz to 0.75 *T_Baud)
22 Transmitter Common Mode Noise T_Ncm
5% of
mVppd
T_Vdiff
23
Load Type 0
24 Output Common Mode Voltage See Note 2
100 1700 mV
25 See Notes 2, 3 & 4 T_Vcm
Load Type 1
26 See also 3.2.2 See Note 3 & 4
630 1100 mV
27 NOTES:
28 1. The Transmitter must be capable of producing a minimum T_Vdiff greater than or equal to 800 mVppd. In applications where
29 the channel is better than the worst case allowed, a Transmitter device may be provisioned to produce T_Vdiff less than this
minimum value, but greater than or equal to 400 mVppd, and is still compliant with this specification.
30 2. Load Type 0 with min T_Vdiff, AC-Coupling or floating load.
31 3. For Load Type 1: R_Zvtt  30T_Vtt & R_Vtt = 1.2V +5%/-8%
4. DC Coupling compliance is optional (Load Type 1). Only Transmitters that support DC coupling are required to meet this
32 parameter.
33
34
35
36 Table 7-3. CEI-6G-LR Transmitter Output Jitter Specifications
37 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
38
39 Uncorrelated High Probability Jitter T_UHPJ See 7.4.1.8 0.15 UIpp

40 Duty Cycle Distortion T_DCD See 7.4.1.8 0.05 UIpp

41 Total Jitter T_TJ See 7.4.1.8 0.30 UIpp
42 Eye Mask T_X1 See 7.4.1.8 0.15 UI
43 Eye Mask T_X2 See 7.4.1.8 0.50 UI
44 Eye Mask T_Y1 See 7.4.1.8 400 mV
45
46 Eye Mask T_Y2 See 7.4.1.8 600 mV

47 NOTES:
48
49

168 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7.4.1.1 Driver Test Load 1

2
Please refer to 3.2.6 3
4
7.4.1.2 Driver Baud Rate 5
6
All devices shall work from 4.976Gsym/s to the maximum baud rate specified for the 7
device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific 8
protocols will define the operating baud rate without affecting CEI compliance. 9
10
11
7.4.1.3 Driver Amplitude and Swing 12
13
Driver differential output amplitude shall be able to drive between 800 to 1200mVppd 14
either with or without any transmit emphasis. However, for the case of this transmitter 15
talking to a short reach receiver, the differential output amplitude shall be between 400 16
to 750mVppd either with or without any transmit emphasis. DC referenced logic levels
17
are not defined since the receiver must have high common mode impedance at DC.
18
However, absolute driver output voltage shall be between -0.1 V and 1.9 V with respect
19
to local ground. See Figure 1-1 for an illustration of absolute driver output voltage limits
20
and definition of differential peak-to-peak amplitude.
21
22
7.4.1.4 Driver Rise and Fall Times 23
24
The recommended minimum differential rise and fall time is 30ps as measured 25
between the 20% and 80% of the maximum measured levels; the maximum differential 26
rise and fall times are defined by the Tx eye diagram (Figure 1-4 and Table 7-5). 27
Shorter rise and falls may result in excessive high frequency components and increase 28
EMI and cross talk. 29
30
7.4.1.5 Output Resistance and Return Loss 31
32
Please refer to 3.2.10 with the following parameters. 33
34
Table 7-4. CEI-6G-LR Driver Return Loss Parameters 35
Parameter Value Units 36
37
A0 -8 dB
38
f0 100 MHz 39
f1
3
T_Baud  --- Hz 40
4 41
f2 R_Baud Hz 42
Slope 16.6 dB/dec 43
44
7.4.1.6 Driver Lane-to-Lane Skew 45
46
Please refer to 3.2.7 47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 169

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7.4.1.7 Driver Short Circuit Current

2
3 Please refer to 3.2.9
4
5 7.4.1.8 Driver Template and Jitter
6
7 As per 2.4.3 for a BER as per 7.3.4, the driver shall satisfy both the near-end eye
8 template & jitter requirements as given in Figure 1-4, Table 7-5 either with or without
9 any transmit emphasis.
10
11 The maximum near-end duty cycle distortion (T_DCD) shall be less than 0.05UIpp.
12
13 It should be noted that it is assumed the Uncorrelated High Probability Jitter component
14 of the driver jitter is not Inter-symbol Interference (ISI). This is only assumed from a
15 receiver point of view so that a receiver can’t equalize it and does not in any way put
16 any restrictions on the real driver HPJ.
17
18
19
20 Table 7-5. CEI-6G-LR Near-End Template Intervals
21 Characteristics Symbol Near-End Value Units Comments
22 Eye Mask T_X1 0.15 UI
23
Eye Mask T_X2 0.50 UI
24
25 400
For connection
to short reach Rx
26 Eye Mask T_Y1 mV
For connection
27 400
to long reach Rx
28 For connection
29 375
to short reach Rx
Eye Mask T_Y2 mV
30 For connection
600
31 to long reach Rx
32 Uncorrelated Bounded High Probability Jitter T_UBHPJ 0.15 UIpp
33 Duty Cycle Distortion T_DCD 0.05 UIpp
34
35 Total Jitter T_TJ 0.30 UIpp

36
37 7.4.1.9 Driver Training Pattern
38
39 The driver is required to repeatedly transmit a “training pattern”. This pattern may be
40 needed by the receiver to aid in its power up adaptive process. The pattern is at least
41 384 bits long and is explained in Table 7-6. However it should be noted that other data
42 (i.e. framing bits) may be present between the repeated groups of 384 bits.
43
44
45
46
47
48
49

170 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
Table 7-6. CEI-6G-LR Training Pattern 3
Pattern (in Hex) Purpose 4
00 FF 00 FF 00 FF 48 bits - f/16 square wave
5
6
24 bits - positive impulse with 12
00 80 00
leading and trailing zeros 7
55 55 55 55 55 55 48 bits - f/2 square wave
8
9
24 bits - negative impulse with 12
FF EF FF
leading and trailing ones 10
00 FF 00 FF 00 FF 48 bits - f/16 square wave
11
12
Approximation of normal randomized
At least 192 random or pseudo-random bits
data patterns (see 3.2.1) 13
14
The means to indicate to the driver when it has to send or stop the training pattern is 15
beyond the scope of this IA. 16
17
Note there may well be other training pattern(s) requirements at the protocol level. 18
19
20
7.4.2 Receiver Characteristics 21
22
The key receiver characteristics are summarized in Table 7-7 while the following sub- 23
clauses fully detail all the requirements. 24
Table 7-7. CEI-6G-LR Receiver Electrical Input Specifications 25
26
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
27
Rx Baud Rate R_Baud See 7.4.2.1 4.976 6.375 Gsym/s 28
Input Differential voltage R_Vdiff See 7.4.2.3 1200 mVppd 29
Differential Resistance R_Rdin See 7.4.2.7 80 100 120  30
Bias Voltage Source Impedance 
31
(load type 1)
R_Zvtt See Note 1 30  32
Differential Input Return Loss 33
-8 dB
(100MHz to 0.75*R_Baud) 34
R_SDD11 See 7.4.2.7
Differential Input Return Loss 35
(0.75*R_Baud to R_Baud)) 36
Common Mode Input Return Loss
R_SCC11 See 7.4.2.7 -6 dB 37
(100MHz to 0.75 *R_Baud)
38
Load Type 0
0 1800 mV 39
See Note 2
Input Common Mode Voltage
R_Vfcm 40
See Notes: 1, 2 & 3 Load Type 1 
Notes: 1 & 3
595 R_Vtt - 60 mV 41
42
Wander divider (in Figure 2-30 & Figure 2-31) n 10
43
NOTES:
1. DC Coupling compliance is optional (Load Type 1). Only receivers that support DC coupling are required to meet this
44
parameter. 45
2. Load Type 0 with min T_Vdiff, AC-Coupling or floating load. For floating load, input resistance must be 1k
3. For Load Type 1: T_Vtt & R_Vtt = 1.2V +5%/-8%.
46
47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 171

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7.4.2.1 Baud Rate

2
3 All devices shall work from 4.976Gsym/s to the maximum baud rate specified for the
4 device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific
5 protocols will define the operating baud rate without affecting CEI compliance.
6
7 7.4.2.2 Reference Input Signals
8
9 Reference input signals to the receiver have the characteristics determined by
10 compliant driver. The reference input signal must satisfy the transmitter near-end
11 template and jitter given in Figure 1-4 and Table 7-5, as well as the far-end eye jitter
12 given in Table 7-10, with the differential load impedance of 1001% at DC with a
13 return loss of better than 20dB from baud rate divided by 1667 to 1.5 times the baud
14 rate. Note that the input signal might not meet either of these requirements when the
15 actual receiver replaces this load.
16
17
18 7.4.2.3 Input Signal Amplitude
19
20 The receiver shall accept differential input signal amplitudes produced by compliant
21 transmitters connected without attenuation to the receiver. This may be larger than the
22 1200mVppd maximum of the driver due to output/input impedances and reflections.
23 The minimum input amplitude is defined by the far-end driver template, the actual
24 receiver input impedance and the loss of the actual PCB. Note that the far-end driver
25 template is defined using a well controlled load impedance, however the real receiver is
26 not, which can leave the receiver input signal smaller than expected.
27
28
29 7.4.2.4 Absolute Input Voltage
30
31 The absolute voltage levels with respect to the receiver ground at the input of the
32 receiver are dependent on the driver implementation and the inter-ground difference.
33
34 The voltage levels at the input of an AC coupled receiver (if the effective AC coupling is
35 done within the receiver) or at the Tx side of the external AC coupling cap (if AC
36 coupling is done externally) shall be between -0.2 to 2.0V with respect to local ground.
37
38 7.4.2.5 Input Common Mode Impedance
39
40 The input common mode impedance (R_Zvtt) at the input of the receiver is dependent
41 on whether the receiver is AC or DC coupled. The value of R_Zvtt as measured at the
42 input of an AC coupled receiver is undefined. The value of R_Zvtt as measured at the
43 input of a DC coupled receiver is defined as per Table 7-7.
44
45 If AC coupling is used, it is to be considered part of the receiver for the purposes of this
46 specification unless explicitly stated otherwise. It should be noted that various methods
47 for AC coupling are allowed (for example, internal to the chip or done externally). See
48 also 3.2.12 for more information.
49

172 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7.4.2.6 Input Lane-to-Lane Skew 1

2
Please refer to 3.2.8 3
4
7.4.2.7 Input Resistance and Return Loss 5
6
Please refer to 3.2.10 with the following parameters. 7
8
9
10
11
12
Table 7-8. CEI-6G-LR Input Return Loss Parameters 13
Parameter Value Units 14
15
A0 -8 dB
16
f0 100 MHz
17
f1
3
R_Baud  --- Hz 18
4
19
f2 R_Baud Hz
20
Slope 16.6 dB/dec 21
22
23
24
7.4.2.8 Jitter Tolerance 25
26
As per 2.4.4, the receiver shall tolerate at least the far-end jitter requirements as given 27
in Table 7-1 in combination with any compliant channel, as per 7.3.7, with an additional 28
SJ with any frequency and amplitude defined by the mask of Figure 2-4 where the 29
minimum & maximum total wander amplitude are 0.05UIpp & 5UIpp respectively. This 30
additional SJ component is intended to ensure margin for wander, hence is over and 31
above any high frequency jitter from Table 7-1. 32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 173

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7.A Appendix - Link and Jitter Budgets

2
3 The primarily intended application is as a point-to-point interface of up to approximately
4 1m (40”) and up to two connector between integrated circuits using controlled
5 impedance traces on low-cost printed circuit boards (PCBs). Informative loss and jitter
6 budgets are presented in Table 7-9 (see also Appendix 3.A for more information) to
7 demonstrate the feasibility of legacy FR4 epoxy PCB’s. The jitter budget is given in
8 Table 7-10. The performance of an actual transceiver interconnect is highly dependent
9 on the implementation.
10
11
12
13 Table 7-9. CEI-6G-LR Informative Loss, Skew and Jitter Budget
14 Differential
Bounded High
15 Loss (dB)
Skew (ps)
Probability TJ (UIpp)
(UIpp)
16
Driver 0 15 0.15 0.30
17
18 Interconnect (with Connector) 15.9
25
0.35 0.513
19 Other 4.5 0.10 0.262
20 Total 20.4 40 0.60 0.875
21
22
23 Table 7-10. CEI-6G-LR High Frequency Jitter Budget
24 Uncorrelated Jitter Correlated Jitter Total Jitter
25 Bounded Bounded
CEI-6G-LR Unbounded High Bounded Amplitude
26 Gaussian Probability Gaussian
High Gaussian Sinusoidal High Total
Probability Probability
27
Abbreviation UUGJ UHPJ CBGJ CBHPJ GJ SJ HPJ TJ k
28
Unit UIpp UIpp UIpp UIpp UIpp UIpp UIpp UIpp mVppd
29 Transmitter 0.150 0.150 0.150 0.150 0.300 800.0
30 Channel 0.230 0.525
31 0.0
Receiver Input 0.150 0.150 0.230 0.525 0.275 0.675 0.950 0.00
32 See 2
33 Equalizer
-0.350
See 1
34
Post Equalization 0.150 0.150 0.230 0.175 0.275 0.325 0.60 0.20 100.0
35 DFE Penalties 0.100 -0.08 -45.0
36 Clock + Sampler 0.150 0.100 0.100 -45.0
37 Budget 0.212 0.250 0.230 0.375 0.313 0.050 0.625 0.988 0.06 10.0
38 NOTES:
39 1. Due to receiver equalization, it reduces the ISI as seen inside the receiver. Thus this number is negative.
2. It is assumed that the eye is closed at the receiver, hence receiver equalization is required as indicated below.
40
41
42
43
44
45
46
47
48
49

174 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7.B Appendix - StatEye.org Template 1

2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 3
4
% example template for setting up a standard, i.e. equalizer 5
% jitter and return loss 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
param.version = [param.version ’_v1.0’]; 10
11
% these are internal variables and should not be changed 12
13
param.scanResolution = 0.01; 14
param.binsize = 0.0005; 15
param.points = 2^13; 16
17
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 18
19
% set the transmitter and baud rate. The tx filter has two 20
% parameters defined for the corner frequency of the poles 21
22
param.bps = 6.375e9; 23
param.bitResolution = 1/(4*param.bps); 24
param.txFilter = ’twopole’; 25
param.txFilterParam = [0.75 0.75]; 26
27
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 28
29
% set the return loss up. The return loss can be turned off 30
% using the appropriate option 31
32
param.returnLoss = ’on’; 33
param.cpad = 1.00; 34
35
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 36
37
% set the transmitter emphasis up. Some example setting are 38
% included which can be uncommented 39
40
% single tap emphasis 41
param.txpre = [-0.1]; 42
param.signal = 1.0; 43
param.txpost = []; 44
param.vstart = [-0.3 -0.3]; 45
param.vend = [+0.0 +0.0]; 46
param.vstep = [0.1 0.05 0.025]; 47
48
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 175

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 % set the de-emphasis of 4-point transmit pulse
3 % the de-emphasis run if param.txpre = [] and param.txpost = []
4
5 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
6
7 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8
9 % set the data coding changing the transmit pulse spectrum
10 % the coding run if param.txpre = [] and param.txpost = []
11
12 param.datacoding = 1; % the coding is off
13
14 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15
16 % set PAM amplitude and rate
17
18 param.PAM = 2; % PAM is swithed off
19
20 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
21
22 % the rxsample point does not need to be changed as it is
23 % automatically adjusted by the optimisation scripts.
24 % The number of DFE taps should be set, however, the initial
25 % conditions are irrelevant.
26
27 param.rxsample = -0.1;
28
29 param.dfe = [0.3 0.1 0.1 0.1 0.1];
30
31 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
32
33 % sampling jitter in HPJpp and GJrms is defined here
34
35 param.txdj = 0.15;
36 param.txrj = 0.15/(2*7.94);
37
38 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
39
40 % the following options are not yet implemented and should
41 % not be changed
42
43 param.user = [0.0];
44 param.useuser = ’no’;
45 param.usesymbol = ’’;
46 param.xtAmp = 1.0;
47
48 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49

176 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

param.TransmitAmplitude = 0.800; % mVppdif 1

param.MinEye = 0.100; % mVppdif 2
3
param.Q = 2*7.94; 4
param.maxDJ = 0.325; 5
param.maxTJ = 0.60; 6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 7: CEI-6G-LR Long Reach Interface 177

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

178 Clause 7: CEI-6G-LR Long Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

8 CEI-11G-SR Short Reach Interface 1

2
3
This clause details the requirements for the CEI-11G-SR short-reach high speed 4
electrical interface between nominal baud rates of 9.95 Gsym/s to 11.2 Gsym/s using 5
NRZ coding. A compliant device must meet all of the requirements listed below. The
6
electrical interface is based on high speed, low voltage logic with nominal differential 7
impedance of 100 . Connections are point-to-point balanced differential pair and 8
signaling is unidirectional. 9
The electrical IA is based on loss & jitter budgets and defines the characteristics 10
required to communicate between a CEI-11G-SR driver and a CEI-11G-SR receiver 11
using copper signal traces on a printed circuit board. The characteristic impedance of 12
the signal traces is nominally 100  differential. These characteristics are normative for 13
the devices and informative for the channel. Rather than specifying materials, channel 14
components, or configurations, the IA focuses on effective channel characteristics. 15
Hence a short length of poorer material should be equivalent to a longer length of 16
premium material. A ‘length’ is effectively defined in terms of its attenuation rather than 17
physical length. 18
19
Short reach CEI-11G-SR devices from different manufacturers shall be inter-operable. 20
21
22
23
24
8.1 Requirements 25
26
1. Support serial data rate from 9.95 Gsym/s to 11.2 Gsym/s. 27
28
2. Capable of low bit error rate (required BER1 of 10-15). 29
3. Capable of driving 0 – 200 mm of PCB and up to 1 connector. 30
31
4. Shall support AC-coupled and optionally DC-coupled operation. 32
33
5. Shall allow multi-lanes (1 to n). 34
6. Shall support hot plug. 35
36
37
8.2 General Requirements 38
39
This clause uses “Method E” of the Jitter and Interoperability Methodology section. 40
41
42
8.2.1 Data Patterns 43
Please refer to 3.2.1 44
45
46
47
48
1. If optical components are included, i.e XFP modules, the BER is constrained by the optical specification. 49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 179

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 8.2.2 Signal levels

2
3 Please refer to 3.2.2
4
5 8.2.3 Signal Definitions
6
7 Please refer to 1.A
8
9
10 8.2.4 Bit Error Ratio
11
12 Please refer to 3.2.31
13
14 8.2.5 Ground Differences
15
16 Please refer to 3.2.4
17
18 8.2.6 Cross Talk
19
20 Please refer to 3.2.5
21
22
23 8.2.7 Channel Compliance
24 As per 2.5.2, with the following reference transmitter and reference receivers (note
25 these conditions do not specify any required implementation but rather indicate a
26 methodology for testing channel compliance), and shall meet the receive eye mask as
27 specified in Figure 1-5 and Table 8-5 when:
28
29 a. Using reference receiver A and Electrical Characteristic R_X1 less R_SJ-hf in
30 Table 8-5
31
b. Using reference receiver B and Electrical Characteristic R_X1LessCBHPJ in
32
Table 8-5
33
34 Also refer to Appendix 3.A for more information on the channel characteristics.
35
36 Reference Transmitter:
37
38 1. A transmitter with no emphasis
39 2. A transmit amplitude of both 360 mVppd and 770 mVppd
40
41 3. Additional Uncorrelated Bounded High Probability Jitter of 0.15 UIpp (emulating
42 part of the Tx jitter)
43
4. Additional Uncorrelated Unbounded Gaussian Jitter of 0.15UIpp (emulating part of
44
the Tx jitter)
45
46 5. At the maximum baud rate that the channel is to operate at or 11.2Gsym/s which
47 ever is the lowest.
48
49 1. If optical components are included, i.e XFP modules, the BER is constrained by the optical specification.

180 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

6. A Tx edge rate filter: simple 20dB/dec low pass at 75% of baud rate, this is to 1
emulate a Tx -3dB bandwidth at 3/4 baud rate. 2
3
7. Worst case transmitter return loss described as a parallel RC elements, see 2.E.6. 4
Reference Receiver A: 5
6
1. No Rx equalization and the Rx bandwidth is assumed to be infinite. 7
8
2. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 9
3. A BER1 as per 3.2.3. 10
11
4. A wander divider (n in Figure 2-30 & Figure 2-31) equal to 10 12
5. A sampling point defined at the midpoint between the average zero crossings of the 13
differential signal 14
15
Reference Receiver B2: 16
17
1. A receiver with a single zero single pole filter (as per Annex 2.B.8) and the Rx 18
bandwidth is assumed to be infinite. 19
2. Worst case receiver return loss described as a parallel RC elements, see 2.E.6. 20
21
3. A BER1 as per 3.2.3. 22
4. A wander divider (n in Figure 2-30 & Figure 2-31) equal to 10 23
24
5. A sampling point defined at the midpoint between the average zero crossings of the 25
differential signal 26
27
28
8.3 Electrical Characteristics 29
30
The electrical signaling is based on high speed low voltage logic with a nominal 31
differential impedance of 100 . 32
33
All devices shall work from 9.95Gsym/s to the maximum baud rate specified for the 34
device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific 35
protocols will define the operating baud rate without affecting CEI compliance. 36
37
38
39
8.3.1 Driver Characteristics 40
41
The driver electrical specifications at compliance point T are given in tableTable 8-1. As 42
per 2.4.3, the driver shall satisfy both the near-end and far-end eye template and jitter 43
requirements as given in Figure 1-4, Table 8-2, Figure 1-5 and Table 8-5. It is assumed 44
45
46
47
1. If optical components are included, i.e XFP modules, the BER is constrained by the optical specification.
48
2. Reference receiver B allows compliance to XFP Rev. 3.1 (10 gigabit Small form factor Pluggable Module) April 25th 2003 49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 181

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 that the UBHPJ component of the driver jitter is not Inter-symbol Interference (ISI),
2 hence it cannot be equalized in the receiver. To attenuate noise and absorb even/odd
3 mode reflections, the source must provide a common mode return path.
4
5 For termination and DC-blocking information, please refer to 3.2.12
6
7 Table 8-1. Transmitter Electrical Output Specification.
8
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
9
10 Baud Rate T_Baud 9.95 11.2 Gsym/s
11 Output Differential Voltage T_Vdiff 360 770 mVppd
12 Differential Resistance T_Rd 80 100 120 
13
Differential Termination Resistance Mismatch T_Rdm 5 %
14
15 Output Rise and Fall Time (20% to 80%) T_tr, T_tf 24 ps

16 Differential Output Return Loss T_SDD22 See 8.3.1.3 dB

17 Common mode Output Return Loss T_SCC22 See 8.3.1.3 -6 dB
18 Transmitter Common Mode Noise T_Ncm 15 mVrms
19 Load Type 0
20 Note 2
0.05 3.55 V
21 Load Type 1
22 Output Common Mode Voltage
T_Vcm Note 6
735 1135 mV
23 Note 1, 3, 4
Load Type 2 550 1060 mV
24
Load Type 3
25 490 850 mV
Note 5
26
27 NOTES:
1. For Load Types 1, 2 and 3: R_Rdin = 100 ohms  20 ohms, R_Zvtt  30 ohms. For Vcm definition, see Figure 1-1
28 2. Load Type 0, AC-Coupling or floating load, R_Rdin = 100 ohms  20 ohms.Number includes ground difference
29 3. For Load Types 1 through 3: Vtt is defined for each load type as follows: Load Type 1 R_Vtt = 1.2V +5% / -8%; Load Type 2
R_Vtt = 1.0V +5% / -8%; Load Type 3 R_Vtt = 0.8V +5% / -8%.
30 4. DC Coupling compliance is optional (Type 1 through 3). Only Transmitters that support DC coupling are required to meet this
31 parameter. It is acceptable for a Transmitter to restrict the range of T_Vdiff in order to comply with the specified T_Vcm range.
For a Transmitter which supports multiple T_Vdiff levels, it is acceptable for a Transmitter to claim DC Coupling Compliance if
32 it meets the T_Vcm ranges for at least one of it's T_Vdiff setting as long as those setting(s) are that are compliant are
33 indicated
5. Simple CML Transmitters designed using Vdd  1.2V may still claim DC compliance if this parameter is not met.
34 6. Simple CML Transmitters designed using Vdd  0.8V may still claim DC compliance if this parameter is not met.
35
36 Table 8-2. Transmitter Output Jitter Specification
37
38 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
39 Uncorrelated Bounded High Probability Jitter T_UBHPJ 0.15 UIpp
40
Uncorrelated Unbounded Gaussian Jitter T_UUGJ Note 1 0.15 UIpp
41
42 Total Jitter T_TJ 0.30 UIpp

43 Eye Mask T_X1 0.15 UI

44 Eye Mask T_X2 0.4 UI
45 Eye Mask T_Y1 180 mV
46 Eye Mask T_Y2 385 mV
47
NOTES:
48 1. BER=10-15, Q=7.94
49

182 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

8.3.1.1 Driver Baud Rate 1

2
All devices shall work from 9.95Gsym/s to the maximum baud rate specified for the 3
device, with the baud rate tolerance as per 3.2.11. 4
5
8.3.1.2 Driver Test Load 6
7
Please refer to 3.2.6. 8
9
10
8.3.1.3 Driver Resistance and Return Loss 11
12
Please refer to 3.2.10 with the following parameters.. 13
Table 8-3. Driver Return Loss Parameters 14
15
Parameter Value Units
16
A0 -8 dB 17
18
f0 100 MHz
19
3 20
f1 T_Baud  --- Hz
4 21
3 22
f2 T_Baud  --- Hz
2 23
Slope 16.6 dB/dec
24
25
26
8.3.1.4 Driver Lane-to-Lane Skew 27
28
Please refer to 3.2.7 29
30
8.3.1.5 Driver Short Circuit Current 31
32
Please refer to 3.2.9 33
34
8.3.2 Receiver Characteristics 35
36
Receiver electrical specifications are given in Table 8-4 and measured at compliance 37
point R. To dampen noise sources and absorption of both even and odd mode 38
reflections, the source in addition to improve differential termination must provide a 39
common mode return path. Jitter specifications at reference R are listed in Table 8-5 40
and the compliance mask is shown in Figure 1-5. 41
42
As per 2.2.4, the receiver shall tolerate at least the far-end eye template and jitter 43
requirements as given in Figure 1-5 and Table 8-5 with an additional SJ with any 44
frequency and amplitude defined by the mask of Figure 2-4 where the maximum total 45
wander amplitude is 5UIpp. This additional SJ component is intended to ensure margin 46
for wander. 47
48
49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 183

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 For termination and DC-blocking information, please refer to 3.2.12.

2
3
4
5 Table 8-4. Receiver Electrical Input Specification
6
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
7
8 Baud Rate R_Baud 9.95 11.2 Gsym/s
9 Input Differential Voltage R_Vdiff 110 1050 mVppd
10 Differential Input Resistance R_Rdin 80 100 120 
11 Receiver Common Mode Noise R_Ncm 25 mVrms
12
Input Resistance Mismatch R_Rm 5 %
13
14 Differential Input Return Loss R_SDD11 See 8.3.2.3 dB

15 Common mode Return Loss R_SCC11 See 8.3.2.3 -6 dB

16 Differential to Common mode input conversion R_SCD11 See 8.3.2.3 -12 dB
17 R_Vtt floating, Not Specified V
18 Note 3
19 R_Vtt = 1.2V
1.2 - 8% 1.2 + 5% V
20 Termination Voltage
R_Vtt
Nominal

21 Note 1, 2 R_Vtt = 1.0V

1.0 - 8% 1.0 + 5% V
Nominal
22
23 R_Vtt = 0.8V
0.8 - 8% 0.8 + 5% V
Nominal
24
25 R_Vtt floating,
Note 3
0 3.60 V
26
R_Vtt = 1.2V R_Vtt
27 Nominal
720 mV
-10
28 Input Common Mode Voltage
R_Vrcm
Note 1, 2 R_Vtt
29 R_Vtt = 1.0V
535 mV
Nominal +125
30
31 R_Vtt = 0.8V
475
R_Vtt
mV
Nominal +105
32
33 NOTES:
1. DC Coupling compliance is optional. Only Receivers which support DC coupling are required to meet this parameter. For Vcm
34 definition, see Figure 1-1
35 2. Receiver is required to implement at least one of specified nominal R_Vtt values, and typically implements only one of these
values. Receiver is only required to meet R_Vrcm parameter values that correspond to R_Vtt values supported.
36 3. Input common mode voltage for AC-coupled or floating load input.
37
38
39
40 Table 8-5. Receiver Input Jitter Specification
41
42 Characteristic Symbol Condition MIN. TYP. MAX. UNIT

43 Uncorrelated Bounded High Probability Jitter R_UBHPJ 0.25 UIpp

44 Correlated Bounded High probability Jitter R_CBHPJ 0.20 UIpp
45 Gaussian Jitter (UUGJ + CBGJ) R_GJ Note 2 0.20 UIpp
46
NOTES:
47 1. TJ includes high frequency sinusoidal jitter. The receiver must tolerate the total deterministic and random jitter with addition of
48 the sinusoidal jitter. For transparent applications the specified jitter tolerance mask replace R_SJ.
2. BER=10-15, Q=7.94
49

184 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Characteristic Symbol Condition MIN. TYP. MAX. UNIT 1

2
Sinusoidal Jitter, Maximum R_SJ-max See 2.2.4 5 UIpp
3
Sinusoidal Jitter, High Frequency R_SJ-hf See 2.2.4 0.05 UIpp 4
Total Jitter, including R_SJ-hf R_TJ Note 1 0.70 UIpp 5
R_TJLess
Total Jitter excl. Correlated High Probability Jitter CHPJ 0.50 UIpp
6
7
Eye Mask incl. Correlated High Probability. Jitter R_X1 0.35 UI 8
Eye mask excl. Correlated High Probability Jitter
R_X1Less
0.25 9
CHPJ
10
Eye Mask R_Y1 55 mV 11
Eye Mask R_Y2 525 mV 12
NOTES: 13
1. TJ includes high frequency sinusoidal jitter. The receiver must tolerate the total deterministic and random jitter with addition of
the sinusoidal jitter. For transparent applications the specified jitter tolerance mask replace R_SJ.
14
2. BER=10-15, Q=7.94 15
16
. 17
18
19
8.3.2.1 Input Baud Rate 20
All devices shall work from 9.95Gsym/s to the maximum baud rate specified for the 21
device, with the baud rate tolerance as per 3.2.11. 22
23
24
8.3.2.2 Reference Input Signals 25
26
Reference input signals to the receiver shall have the characteristics determined by a 27
compliant driver. The reference input signal must satisfy the transmitter near-end 28
template and jitter given in Figure 1-4 and Table 8-2, as well as the far-end eye 29
template and jitter given in Figure 1-5 and Table 8-5, with the differential load 30
impedance of 1% at DC and a return loss of better than 20dB from baud rate 31
over 1667 to 1.5 times the baud rate. Note that the input signal might not meet either of 32
these templates when the actual receiver replaces this load. 33
34
8.3.2.3 Input Resistance and Return Loss 35
36
Please refer to 3.2.10 with the following parameters. 37
38
Table 8-6. Driver Return Loss Parameters 39
Parameter Value Units 40
41
A0 -8 dB
42
f0 100 MHz 43
44
3 45
f1 R_Baud  --- Hz
4
46
f2
3
R_Baud  --- Hz 47
2
48
Slope 16.6 dB/dec 49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 185

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 SCD11 relates to the conversion of Differential to Common mode and the associated
2 generation of EMI. The common mode reference impedance is 25, measurement
3 range is f0 to f1 of Table 8-6.
4
5 8.3.2.4 Input Lane-to-Lane Skew
6
7 Please refer to 3.2.8
8
9
10 8.4 Specifications for Jitter-transparent applications
11
12 The CEI interface for short reach may be used for applications where connected
13 elements are transparent to other clock domains with requirements to jitter
14 performance that in some implementations may interfere with the CEI jitter
15 requirements. Consider a situation using the CEI reference model, Figure 1-6, where
16 the Ingress Transmitter T does not filter the jitter from the adjacent clock domain with a
I
17 low frequency low pass filter and the Egress Receiver RE likewise pass the CEI
18 channel jitter unfiltered to the adjacent clock domain. In this case the requirements to
19 handle the combined jitter of the CEI interface and the adjacent clock domain is
20 evident. In the Ingress direction the unfiltered Jitter from the input to the Ingress
21 Transmitter will be superimposed to the jitter of the Transmitter, link and Receiver. In
22 the Egress direction the jitter of the Transmitter, Link and Receiver will be passed
23 beyond the Egress Receiver R into the adjacent clock domain. The following sections
E
24 specify the requirements to devices intended for use in transparent applications. The
25 requirements have an effect on the previously defined channel, transmitter, and
26 receiver compliance testing and must be carefully understood, please refer to 2.5 for
27 further details.
28
29
30 8.4.1 Jitter Requirements for Transparent Applications in Telecom systems
31
32 Telecom systems are Sonet as defined by ANSI: T1.105.03-2003 and Telcordia: GR-
33 253, SDH systems as defined by ITU-T: G.783, G.812, G.813, G.825 and OTN systems
34 as defined by ITU-T: G.8251 (for OTN jitter).
35
36 Currently there are discrepancies between Telcordia GR-253 and ITU-T G.783. This IA
37 is compliant to both with respect to jitter transfer and aligned with ITU-T G.783 with
38 respect to jitter generation
39
40
41
42
43
44
45
46
47
48
49

186 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

8.4.1.1 Sinusoidal Jitter tolerance mask for Ingress direction, CEI receiver at 1
reference point RI. 2
3
4
Figure 8-1. Jitter Ingress Receiver Input Telecom Sinusoidal Jitter 5
6
7
17 8
Sinusoidal Jitter Tolerance (UIp-p)

9
-20dB/Dec 10
11
12
1.7 13
14
15
16
0.17 17
18
0.05 19
20
0.01E-3 2E-3 20E-3 0.4 4 8 27.2 80 21
Frequency (MHz) 22
23
24
25
26
The Sinusoidal Jitter mask is aligned with the Telecom requirements for the Input Jitter 27
Tolerance at the Signal Conditioner input and a required maximum loop BW of 8MHz in 28
the case of a simple PLL based Signal Conditioner. Margins are added to the jitter 29
amplitude to allow for added jitter by the signal conditioner and the CEI interconnect. 30
This margin is not intended to alter in any way the telecom network limits as specified 31
by ANSI/ITU-A but is required to assure the limits to be met by an Ingress CEI receiver 32
that needs to tolerate the combined telecom network maximum jitter and CEI channel 33
maximum jitter. 34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 187

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 8.4.1.2 Sinusoidal Jitter tolerance mask for Egress direction, CEI receiver at
2 reference point RE.
3
4
5 Figure 8-2. Jitter Egress Receiver Input Telecom Sinusoidal Jitter
6
7
8
9 15.2
Sinusoidal Jitter Tolerance (UIp-p)

10
11
12 -20dB/Dec
13 1.7
14
15
16
17
18 0.17
19
20 0.05
21 0.01E-3 2E-3 17.9E-3 0.12 4.08 80
22 Frequency (MHz)
23
24
25
26 The Sinusoidal Jitter mask is aligned with the Telecom requirements for the Input Jitter
27 of an Ingress Signal Conditioner with additional margin for the signal transfer to the
28 Egress path in accordance with 8.4.1.3. This implies a required minimum loop BW of
29 4MHz in the case of a simple PLL based Signal Conditioner. The low frequency
30 amplitude is required for tolerance testing only and does not reflect a valid condition
31 during operation.
32
33
34
35 8.4.1.3 Telecom Jitter transfer
36
37 Jitter transfer specifications are necessary to constrain the Peaking and Bandwidth
38 transfer function of the elements in a telecom system due to the synchronous timing of
39 network elements. Measurements as per Annex 2.E.5. The following specifications
40 assume an overall transfer -3dB bandwidth (20db/dec) limited to 120kHz by circuits
41 outside the scope of this IA.
42
43
44
45
46
47
48
49

188 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
Table 8-7. Telecom Signal Conditioner, Egress direction 3
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 4
Jitter Transfer Bandwidth BW Data see 1 8 MHz
5
6
Frequency
<120kHz
0.03 dB 7
Jitter Peaking
Frequency
8
>120kHz
1 dB 9
NOTES: 10
1. PRBS 231-1, OC-192/SDH-64 Sinusoidal Jitter Tolerance Mask 11
12
13
14
Table 8-8. Telecom Signal Conditioner, Ingress Direction 15
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 16
Jitter Transfer Bandwidth BW Data, see 1 8 MHz
17
18
Frequency
<120kHz
0.03 dB 19
Jitter Peaking
Frequency
20
>120kHz
1 dB 21
NOTES: 22
1. PRBS 231-1, OC-192/SDH-64 Sinusoidal Jitter Tolerance Mask 23
24
25
26
8.4.1.4 Telecom Jitter Generation for Egress Direction 27
28
The Jitter generation measured at the Egress output of the Jitter Transparent Element 29
is the sum of the jitter at the Egress Driver Output (reference point TE in Figure 1-6), the 30
CEI channel and the Jitter Transparent Element in which the CEI receiver RE (Figure 1- 31
6) resides. The maximum allowed Jitter Generation at the output of the Jitter 32
Transparent Element is allocated in Table 8-9. 33
34
Table 8-9. Telecom Egress Jitter Generation budget 35
Measurement range 36
Lower Frequency Upper Frequency
Budget allocation 37
38
TE Egress output lower Signal conditioner max
Egress driver
measurement limit transfer bandwidth
42.5% 39
TE Egress output lower Signal conditioner max
40
Egress channel
measurement limit transfer bandwidth
7.5% 41
Egress TE, signal conditioner TE Egress output lower TE Egress output upper 42
50%
and path to Egress output measurement limit measurement limit 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 189

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Informative values for the Egress Driver is given in Table 8-10 based on current
2 telecom recommendations...
3
4 Table 8-10. Telecom Egress Driver Jitter Generation
5 TE Output Measurement
Method Value Unit
6 Specified Range Range
7 50kHz - 80MHz 50kHz - 8MHz not specified, note 1 6.5 mUIrms
8 Telcordia GR-253 50kHz - 80MHz 50kHz - 8MHz not specified, note 1 43 mUIpp
9 20kHz - 80MHz 20kHz - 8MHz 60 sec 129 mUIpp
10 ITU-T G.783
4MHz - 80MHz 4MHz - 8MHz 60 sec 43 mUIpp
11
12 NOTES:
1. The ITU-T specifications are applicable, Telcordia plans to align GR-253 those specifications when/if GR-253 is reissued
13
14
The measurement range corresponds to the transfer bandwidth as stated in Table 8-7.
15
16
17 8.4.2 Jitter Requirements for Transparent Applications in Datacom systems
18
19 Datacom systems are 10GE as defined by IEEE 802.3ae-2002and the 10GFC as
20 defined by INCITS, T11.2.
21
22 8.4.2.1 Sinusoidal Jitter tolerance mask for Ingress direction, CEI Receiver at
23 reference point D
24
25 .
26 Figure 8-3. Jitter Ingress Receiver Input Datacom Sinusoidal Jitter
27
28
29
Sinusoidal Jitter Tolerance (UIp-p)

30
31 0.2
1.13x  ------- + 0.1 , f in MHz
32  f 
33 5.76
34 -20dB/Dec
35
36
37
38
0.17
39
40 0.05
41
42 0.04 4 8 27.2 80
43 Frequency (MHz)
44
45
46
47
48
49

190 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The Sinusoidal Jitter mask is aligned with the Datacom requirements for the Input Jitter 1
Tolerance at the Signal Conditioner input and a required maximum loop BW of 8MHz in 2
the case of a simple PLL based Signal Conditioner. Margins are added to the jitter 3
amplitude to allow for added jitter by the signal conditioner and the CEI interconnect. 4
5
8.4.2.2 Datacom Jitter transfer 6
7
The jitter transparent Signal Conditioner of the Ingress and Egress directions need to 8
be specified to constrain the overall signal jitter transferred to the receive end of the 9
CEI channel and for the Egress direction further onto the transmit side of the signal 10
conditioner. 11
12
13
14
Table 8-11. Datacom Signal Conditioner Egress direction 15
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
16
17
Jitter Transfer Bandwidth BW Data see 1 8 MHz
18
Jitter Peaking
Frequency
1 dB 19
>50kHz
20
NOTES:
1. Based on IEEE 802.3ae-2002 Clause 52 Sinusoidal Jitter Tolerance Mask, figure 52-4
21
22
23
24
Table 8-12. Datacom Signal Conditioner Ingress Direction 25
26
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
27
Jitter Transfer Bandwidth BW Data, see 1 8 MHz 28
Jitter Peaking
Frequency
1 dB 29
>50kHz
30
NOTES: 31
1. Based on IEEE 802.3ae-2002 Clause 52 Sinusoidal Jitter Tolerance Mask, figure 52-4
32
33
8.4.3 Jitter Transparency compliance nomenclature 34
35
For compliance to Jitter-transparent applications transmitters and receivers shall be 36
identified as shown in table 37
38
Table 8-13. Datacom Signal Conditioner Ingress Direction
39
Characteristic Symbol 40
Telecom Receiver, Ingress CEI 11GSR - TR(I) 41
Telecom Transmitter, Ingress CEI 11GSR - TT(I)
42
43
Telecom Receiver, Egress CEI 11GSR - TR(E)
44
Telecom Transmitter, Egress CEI 11GSR - TT(E) 45
Datacom Receiver, Ingress CEI 11GSR - DR(I) 46
NOTES: 47
48
49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 191

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 8-13. Datacom Signal Conditioner Ingress Direction

2 Characteristic Symbol
3
4 Datacom Transmitter, Ingress CEI 11GSR - DT(I)

5 Datacom Receiver, Egress CEI 11GSR - DR(E)

6 Datacom Transmitter, Egress CEI 11GSR - DT(E)
7 NOTES:
8
9
10
11
12
13
14
8.A Appendix - Informative Jitter Budget
15
16 The Jitter Budget is presented in Table 8-14. Contributors in the ‘Source’ column should
17 not exceed the value of the ‘Value’ column.
18 Table 8-14. Informative Jitter Budget
19
20 Uncorrelated Jitter Correlated Jitter Total Jitter
Source Amplitude
21 Unbounded Bounded Bounded Bounded
Gaussian Sinusoidal High Prob. Total
22 Gaussian High Prob. Gaussian High Prob.

23 Abbreviation UUGJ UBHPJ CBGJ CBHPJ k

24 Unit UIpp UIpp UIpp UIpp UIpp UIpp UIpp UIpp mVppd
25 Transmitter 0.150 0.150 0.150 0.150 0.300 360
26 Channel 0,100 0,132 0.200 0,050
27
Receiver Input 0.150 0.250 0,132 0.200 0.200 0,050 0.450 0.650 0.31 110
28
29 Equalizer -0.200

30 Post Equalizer 0.150 0.250 0,132 0.000 0.200 0,050 0.250 0.450 0.31 110
31 Clock & Sampler 0.150 0.100 0.100 -50
32 Budget with
0.212 0.350 0,132 0.100 0.250 0.050 0.450 0.750 60
33 Equalizer
34 Budget without 0.212 0.350 0,132 0.300 0.250 0.050 0650 0.950 60
35 equalizer
36 Note: Values in yellow are specified values from Table 8-2 and Table 8-5
37
38
39
40
41
42
43
44
45
46
47
48
49

192 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

8.B Appendix - StatEye.org Template1 1

2
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 3
4
% example template for setting up a standard, i.e. equaliser 5
% jitter and return loss 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
param.version = [param.version ’_v1.0’]; 10
11
% these are internal variables and should not be changed 12
13
param.scanResolution = 0.01; 14
param.binsize = 0.0005; 15
param.points = 2^13; 16
17
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 18
19
% set the transmitter and baud rate. The tx filter has two 20
% parameters defined for the corner frequency of the poles 21
22
param.bps = 11.1e9; 23
param.bitResolution = 1/(3*param.bps); 24
param.txFilter = ’singlepole’; 25
param.txFilterParam = [0.75]; 26
27
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 28
29
% set the return loss up. The return loss can be turned off 30
% using the appropriate option 31
32
% param.returnLoss = ’off’; 33
param.returnLoss = ’on’; 34
param.cpad = 0.60; 35
36
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 37
38
% set the transmitter emphasis up. Some example setting are 39
% included which can be uncommented 40
41
% single tap emphasis 42
param.txpre = []; 43
param.signal = 1.0; 44
param.txpost = []; 45
param.vstart = [-0.3]; 46
param.vend = [+0.0]; 47
48
1. for Reference receiver B in 8.2.7, pls refer to XFP Rev. 3.1 (10 gigabit Small form factor Pluggable Module) April 25th 2003 49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 193

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 param.vstep = [0.1 0.05 0.025];

2
3 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4
5 % set the de-emphasis of 4-point transmit pulse
6 % the de-emphasis run if param.txpre = [] and param.txpost = []
7
8 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
9
10 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11
12 % set the data coding changing the transmit pulse spectrum
13 % the coding run if param.txpre = [] and param.txpost = []
14
15 param.datacoding = 1; % the coding is off
16
17 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
18
19 % set PAM amplitude and rate
20
21 param.PAM = 2; % PAM is swithed off
22
23 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24
25 % the rxsample point does not need to be changed as it is
26 % automatically adjusted by the optimisation scripts.
27 % The number of DFE taps should be set, however, the initial
28 % conditions are irrelevant.
29
30 param.rxsample = -0.1;
31
32 param.dfe = [];
33
34 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
35
36 % sampling jitter in HPJpp and GJrms is defined here
37
38 param.txdj = 0.15;
39 param.txrj = 0.15/(2*7.94);
40
41 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42
43 % the following options are not yet implemented and should
44 % not be changed
45
46 param.user = [0.0];
47 param.useuser = ’no’;
48 param.usesymbol = ’’;
49 param.xtAmp = 1.0;

194 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 2
3
param.TransmitAmplitude = 0.360; % mVppdif 4
param.MinEye = 0.110; % mVppdif 5
6
param.Q = 2*7.94; 7
param.maxDJ = 0.45; 8
param.maxTJ = 0.65; 9
10
11
8.C Appendix - XFP reference points 12
13
The specification of the CEI-11G-SR is compatible with the XFI interface specified for 14
the XFP (10 gigabit Small form factor Pluggable Module). However the definition of 15
reference points diverts somewhat. Where the CEI is defining the active component 16
interfaces to a generic compliant channel the XFP specifies the normative reference 17
points at the edges of the XFP connector that forms the interface between an XFP 18
module and its host board. The XFP reference points A and D at the component edge 19
are informative only for XFP but identical to the CEI RI and TE respectively. Figure 8-4 20
shows the reference points of the XFP in comparison to the CEI. Note that the XFP 21
specification does not define test points for the component edge of the components in 22
the XFP module, the signal conditioners. Also note that CEI does not define the XFP 23
reference points B, B’, C and C’ for the connector as this is considered part of the 24
channel. 25
26
Figure 8-4. Reference Model
27
28
Component Component
Connector 29
Edge Edge 30
31
D C’ C 32
33
Channel * 34
RX RI TI TX
35
36
37
38
* 39
Channel 40
TX TE RE RX
41
A 42
B’ B 43
ASIC / Serdes XFP Module 44
45
* Signal Conditioner 46
Host Board 47
48
49

Optical Internetworking Forum - Clause 8: CEI-11G-SR Short Reach Interface 195

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

196 Clause 8: CEI-11G-SR Short Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

9 CEI-11G-LR/MR Long/Medium Reach Interface 1

2
3
This clause details the requirements for the CEI-11G-LR and CEI-11G-MR high speed 4
electrical interface between nominal baud rates of 9.95 Gsym/s to 11.2 Gsym/s using 5
NRZ coding. A compliant device must meet all of the requirements listed below. The
6
electrical interface is based on high speed, low voltage logic with nominal differential 7
impedance of 100 . Connections are point-to-point balanced differential pair and 8
signaling is unidirectional. 9
The electrical IA is based on loss and jitter budgets and defines the characteristics 10
required to communicate between a CEI-11G-LR driver and a CEI-11G-LR receiver and 11
between a CEI-11G-MR driver and a CEI-11G-MR receiver, using copper signal traces 12
on a printed circuit board. The characteristic impedance of the signal traces is 13
nominally 100 differential. Rather than specifying materials, channel components or 14
configurations, the IA focuses on effective channel characteristics. Hence a short 15
length of poorer material should be equivalent to a longer length of premium material. A 16
length is effectively defined in terms of its attenuation and phase response rather than 17
its physical length. 18
19
CEI-11G-LR as well as CEI-11G-MR devices from different manufacturers shall be 20
inter-operable.The CEI-11GLR/MR channel is tested to insure compliance using the 21
statEye scripts. The transmitter is specified in terms of its ability to pre-equalize the 22
transmit signal and the receiver must work to the given BER using a compliant driver 23
and channel. 24
25
The primary focus of the CEI-11G-LR implementation agreement will be for non-legacy 26
applications, optimized for overall cost-effective system performance including total 27
power dissipation. Future clauses may address schemes otherwise optimized. 28
29
This clause also provides for a CEI-11G-MR low power option. The CEI-11G-MR option 30
is based upon the following: 31
32
• A channel compliance specification is defined in this clause for CEI-11G-MR which 33
is more stringent than that of CEI-11G-LR. 34
• CEI-11G-MR uses the same Transmitter device as is specified for CEI-11G-LR, 35
making use of certain features otherwise defined as optional. 36
37
• CEI-11G-MR uses a Receiver device that is similar to the device specified for CEI- 38
11G-SR in Clause 8, but with extended T_Vdiff range. Relevant specifications for 39
this receiver device are incorporated by reference to Clause 8. 40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 197

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9.1 Requirements
2
3 1. Support NRZ coded serial data rate from 9.95 Gsym/s to 11.2 Gsym/s.
4
5 2. Capable of low bit error rate (required BER < 10-15).
6 3. Capable of driving 0 — 1 meter (39 inches) of PCB and up to 2 connectors.
7
8 4. Capable of driving 0 — 600 mm of PCB and up to 2 connectors for low-power
9 applications.
10 5. Shall support AC-coupled and optionally DC-coupled operation.
11
12 6. Shall allow multi-lanes (1 to n).
13 7. Shall support hot plug.
14
15
16 9.2 General Requirements
17
18
19 9.2.1 Data Patterns
20
21 See 3.2.1
22
23 9.2.2 Signal Levels
24
25 See 3.2.2
26
27 9.2.3 Signal Definitions
28
29 See 1.A
30
31
32 9.2.4 Bit Error Ratio
33
34 See 3.2.3
35
36 9.2.5 Ground Differences
37
38 See 3.2.4
39
40 9.2.6 Cross Talk
41
42 See 3.2.5
43
44
45
46
47
48
49

198 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

9.2.7 Channel Compliance 1

2
9.2.7.1 CEI-11G-LR Channel Compliance 3
4
A forward channel and associated dominant crosstalk channels are deemed compliant 5
if for the specified reference transmitter and both the specified reference receivers, the 6
signal conforms to the defined eye mask and does not exceed the defined jitter using 7
the “Statistical Eye” methodology defined in 2.C 8
9
10
11
Reference Transmitter: 12
13
1. Maximum Transmit Pulse, as per 2.E.7, of T_Vdiff min. of Table 9-1 14
2. A TX edge rate filter simple 40dB/dec low pass at 75% of Baud Rate 15
16
3. Effective Driver UUGJ, UBHPJ and DCD as in Table 9-3 17
4. Equalizing Filter with 2 tap baud spaced emphasis no greater than a total of 6dB 18
with finite resolution no better than 1.5dB. 19
20
5. Worst case Transmitter return loss described as a parallel RC element, see 2.E.6 21
6. Maximum baud rate that the channel is to operate at or 11.2 Gsym/sec whichever is 22
the lowest, see 9.3.1.1 23
24
Reference Receiver A: 25
26
1. 4-tap baud spaced Non-Linear Discrete Inverse Channel Filter (DFE), with infinite 27
precision accuracy and having the following restrictions:  28
 29
Let W[N] be sum of DFE tap coefficient weights from taps N through M where 30
 31
N = 1 is previous decision (i.e. first tap) 32
M = 4 33
R_Y2 = T_Y2 = 400mV 34
Y = min(R_X1, (R_Y2 - R_Y1) / R_Y2) = 0.2625 35
Z = 2/3 = 0.66667 36
 37
Then W[N]  Y * Z(N - 1) 38
 39
For the channel compliance model the number of DFE taps (M) = 4. This gives the 40
following maximum coefficient weights for the taps: 41
 42
W[1]  0.2625 (sum of absolute value of taps 1 and2) 43
W[2]  0.1750 (sum of absolute value of taps 2, 3 and 4) 44
W[3]  0.1167 (sum of absolute value of taps 3 and 4) 45
W[4]  0.0778 (sum of absolute value of tap 4) 46
 47
Notes: 48
- Coefficient weights are absolute, assuming a T_Vdiff of 1Vppd 49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 199

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 - For a real receiver the restrictions on tap coefficients would apply for the actual
2 number of DFE taps implemented (M)
3 - LMS, Least Mean Squared Adaptation Algorithm.
4
5 2. Worst case Receiver return loss described as a parallel RC, see 2.E.6
6
7 Resulting Eye Mask of either receiver:
8 Table 9-1. CEI-11G-LR Receiver Equalization Output Eye Mask
9
10 Parameter Symbol Max Units
11 Eye mask R_X1 0.2625 UI
12 Eye mask R_Y1 50 mV
13 Correlated Bounded High Probability Jitter, pre-equalizer R_CBHPJ 0.40 UIpp
14
Correlated Bounded High Probability Jitter, post-equalizer R_CBHPJ 0.10 UIpp
15
16 Uncorrelated Bounded High Probability Jitter R_UBHPJ 0.15 UIpp

17 Uncorrelated Unbounded Gaussian Jitter R_UUGJ 0.15 UIpp

18 Quality of signal (SNR in real number) Q 7.94
19
20
21 9.2.7.2 CEI-11G-MR Channel Compliance
22
As per 2.5.2, with the following reference transmitter and reference receiver (note
23
these conditions do not specify any required implementation but rather indicate a
24
methodology for testing channel compliance), and shall meet the receive eye mask as
25
specified in Figure 1-5 and Table 9-9 when using electrical characteristic R_X1 less
26
R_SJ-hf in Table 9-9.
27
28 Reference Transmitter as defined in “Reference Transmitter” in section 9.2.7.1.
29
30 Reference Receiver as defined in “Reference Receiver A” in Section 8.2.7.
31
32
33 9.3 Electrical Characteristics, CEI-11G-LR and CEI-11G-MR
34
35 The electrical signaling is based on high speed low voltage logic with a nominal
36 differential impedance of 100 .
37
38
39 9.3.1 Driver Characteristics
40
41 For termination and DC-blocking information, please refer to 8.2.7
42
43
44
45
46
47
48
49

200 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Table 9-2. Transmitter Output Electrical Specifications 2
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
3
4
Baud Rate T_Baud 9.95 11.2 Gsym/s
5
Pre-emphasis 6
off or Tx Filter
Output Differential Voltage T_Vdiff
Applied, see
800 1200 mVppd 7
note 1 8
Differential Output Impedance T_Rd 80 100 120  9
Differential Termination Impedance Mismatch T_Rm 10 % 10
Output Rise and Fall Time (20% to 80%) T_tr, T_tf 24 ps
11
12
Differential Output Return Loss T_SDD22 See 9.3.1.3
13
Common Mode Return Loss T_SCC22 See 9.3.1.3 -6 dB
14
Transmitter Common Mode Noise T_Ncm
5% of
mVppd 15
T_Vdiff
16
Load Type 0
See Note 2
100 1700 mV 17
Output Common Mode Voltage
See Notes 2, 3 & 4
T_Vcm 18
Load Type 1
See Note 3 & 4
630 1100 mV 19
NOTES:
20
1. In applications where the channel is better than the worst case allowed, a transmitter device may be provisioned to produce 21
T_Vdiff less than this minimum value but360mVppd and be compliant with this specification. 22
2. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load.
3. For Load Type 1: R_Zvtt  30T_Vtt & R_Vtt = 1.2V +5%/-8% 23
4. DC Coupling compliance is optional (Load Type). Only Transmitters that support DC coupling are required to meet this 24
parameter.
25
26
27
28
Table 9-3. Transmitter Output Jitter Specifications 29
30
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
31
Uncorrelated Unbounded Gausian Jitter T_UUGJ
See 9.3.1.6,
0.15 UIPP 32
Note 1
33
Uncorrelated Bounded High Probability Jitter T_UBHPJ
See 9.3.1.6,
Note 1
0.15 UIPP 34
35
Duty Cycle Distortion (component of UBHPJ) T_DCD See 9.3.1.6 0.05 UIPP
36
Total Jitter T_TJ See 9.3.1.6 0.30 UIPP
37
Eye Mask T_X1 See 9.3.1.6 0.15 UI 38
Eye Mask T_X2 See 9.3.1.6 0.50 UI 39
See 9.3.1.6 40
Eye Mask T_Y1 400 mV
Note 3 41
Eye Mask T_Y2 See 9.3.1.6 600 mV 42
NOTES:
43
1. UBHPJ is composed of DCD, inter-symbol-interference (ISI), and Sinusoidal Jitter. 44
2. Except for amplitude, the CEI-11G+ long-reach driver electrical specifications of Table 9-3 are intended to be the same as for
CEI-11G+ short-reach
45
3. The minimum value for channel compliance is 300mV and not 180mV. The 180mV is to allow lower power for channels that 46
are better than the worst case channels allowed 47
48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 201

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9.3.1.1 Driver Baud Rate

2
3 All devices shall work from 9.95Gsym/s to the maximum baud rate specified for the
4 device, with the baud rate tolerance as per 3.2.12. Note that implementation of specific
5 protocols will define the operating baud rate without affecting CEI compliance.
6
7
9.3.1.2 Driver Amplitude and Swing
8
9 Driver differential output amplitude shall be able to drive between 800 to 1200mVppd
10 either with or without any transmit emphasis. However, for the case of this transmitter
11 talking to a short reach receiver, the differential output amplitude shall be between 380
12 to 770mVppd either with or without any transmit emphasis. DC referenced logic levels
13 are not defined since the receiver must have high common mode impedance at DC.
14 However, absolute driver output voltage shall be between -0.1 V and 1.9 V with respect
15 to local ground. See Figure 1-1 for an illustration of absolute driver output voltage limits
16 and definition of differential peak-to-peak amplitude.
17
18
19 9.3.1.3 Driver Resistance and Return Loss
20
21 Please refer to 3.2.10 with the following parameters.
22 Table 9-4. Driver Return Loss Parameters
23
24 Parameter Value Units
25 A0 -8 dB
26
27 f0 100 MHz

28 3
29 f1 T_Baud  --- Hz
4
30
31 f2 T_Baud Hz

32
Slope 16.6 dB/dec
33
34
35 9.3.1.4 Driver Lane-to-Lane Skew
36
37 Please refer to 3.2.7
38
39 9.3.1.5 Driver Short Circuit Current
40
41 Please refer to 3.2.9
42
43
44 9.3.1.6 Driver Template and Jitter
45
46 As per 2.2.3 for a BER as per 9.2.4, the driver shall satisfy the eye template and jitter
47 requirements as given in Figure 1-4.
48
49

202 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

9.3.2 CEI-11G-LR Receiver Characteristics 1

2
This section defines receiver characteristics for CEI-11G-LR receivers. Receiver 3
characteristics for CEI-11G-MR receivers are defined in 9.3.3. 4
5
Receiver electrical specifications are given in Table 9-5 and measured at compliance 6
point R. For termination and DC-blocking information, please refer to 3.2.12 7
8
Table 9-5. CEI-11G-LR Receiver Electrical Specifications 9
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 10
11
Baud rate R_Baud 9.95 11.2 GSym/s
12
Input Differential Voltage R_Vdiff Note 1 1200 mVppd
13
Differential Input Impedance R_Rdin 80 100 120  14
Input Impedance Mismatch R_Rm 10 % 15
Differential Input Return Loss R_SDD11 See 9.3.2.3 16
Common Mode Input Return Loss R_SCC11 Below 10 GHz -6 dB 17
Load Type 0
18
Input Common Mode Voltage See Note 3
0 1800 mV 19
See Notes: 2, 3 & 4
R_Vcm
Load Type 1  R_Vtt - 20
595 mV
See Notes 2, 4 60 21
Wander Divider n See Note 5 10 22
NOTES: 23
1. The long-reach receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the 24
combined transmitter and channel. The channel response shall include the worst case effects of the return losses at the
transmitter and receiver.
25
2. DC Coupling compliance is optional (Load Type 1). Only receivers that support DC coupling are required to meet this 26
parameter.
3. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load. For floating load, input resistance must be 1k
27
4. For Load Type 1: T_Vtt & R_Vtt = 1.2V +5%/-8%. 28
5. Used in Statistical Eye script, must be set to 10 29
30
Table 9-6. CEI-11G-LR Receiver Input Jitter Specification 31
32
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
33
Sinusoidal Jitter, Maximum R_SJ-max See 2.5.4, note 5 UIpp 34
1, 2
35
Sinusoidal Jitter, High Frequency R_SJ-hf
See 2.5.4, note
1, 2 0.05 UIpp 36
NOTES:
37
1. The Receiver shall tolerate the sum of these jitter contributions: Total Driver jitter from Table 9-2;Sinusoidal jitter as defined in 38
Table 9-6;The effects of a channel compliant to the Channel Characteristics ( 9.2.7). 39
2. The receiver must tolerate the total deterministic and random jitter with addition of the sinusoidal jitter.
40
41
9.3.2.1 Input Baud Rate 42
43
All devices shall work from 9.95 Gsym/s to the maximum baud rate specified for the 44
device, with the baud rate tolerance as per 3.2.11. Note that implementation of specific 45
protocols will define the operating baud rate without affecting CEI compliance. 46
47
48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 203

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9.3.2.2 Absolute Input Voltage

2
3 The absolute voltage levels with respect to the receiver ground at the input of the
4 receiver are dependent on the driver implementation and the inter-ground difference.
5
6 The voltage levels at the input of an AC coupled receiver (if the effective AC coupling is
7 done within the receiver) or at the Tx side of the external AC coupling cap (if AC
8 coupling is done externally) shall be between -0.2 to 2.0V with respect to local ground.
9
10 9.3.2.3 Input Resistance and Return Loss
11
12 Please refer to 3.2.10 with the following parameters.
13
14 Table 9-7. Driver Return Loss Parameters
15 Parameter Value Units
16
A0 -8 dB
17
18 f0 100 MHz
19
20 3
f1 R_Baud  --- Hz
4
21
22 f2 R_Baud Hz
23
24 Slope 16.6 dB/dec
25
26
9.3.2.4 Input Signal Amplitude
27
28 The receiver shall accept differential input signal amplitudes produced by compliant
29 transmitters connected without attenuation to the receiver. This may be larger than the
30 1200 mVppd maximum of the driver due to output/input impedances and reflections.
31
32 The minimum input amplitude is defined by the far-end driver template, the actual
33 receiver input impedance and the loss of the actual PCB. Note that the far-end driver
34 template is defined using a well controlled load impedance, however the real receiver is
35 not, which can leave the receiver input signal smaller than expected.
36
37
38 9.3.2.5 Input Lane-to-Lane Skew
39
40 Please refer to 3.2.8
41
42 9.3.3 CEI-11G-MR Receiver Characteristics
43
44 This section defines receiver characteristics for CEI-11G-MR receivers. Receiver
45 characteristics for CEI-11G-LR receivers are defined in 9.3.2.
46
47 Receiver electrical specifications are given in Table 9-8 and measured at compliance
48 point R. Jitter specifications at reference R are listed in Table 9-9 and the compliance
49 mask is shown in Figure 1-5.

204 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

For termination and DC-blocking information, please refer to 3.2.12. 1

2
Table 9-8. CEI-11G-MR Receiver Electrical Specifications 3
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
4
5
Baud rate R_Baud 9.95 11.2 GSym/s
6
Input Differential Voltage R_Vdiff Note 1 110 1200 mVppd 7
Differential Input Impedance R_Rdin See R_Rdin in Table 8-4  8
Input Impedance Mismatch R_Rm See R_Rm in Table 8-4 % 9
Differential Input Return Loss R_SDD11 See 9.3.2.3 See R_SDD11 in Table 8-4 10
Common Mode Input Return Loss R_SCC11 Below 10 GHz See R_SCC11 in Table 8-4 dB
11
12
Input Common Mode Voltage R_Vcm Note 2 See R_Vcm in Table 9-5 mV
13
Wander Divider n See Note 5 See n in Table 9-5 14
NOTES: 15
1. The medium-reach receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the
combined transmitter and channel. The channel response shall include the worst case effects of the return losses at the 16
transmitter and receiver. 17
2. DC Coupling compliance is optional (Load Type 1). Only receivers that support DC coupling are required to meet this
parameter.
18
19
20
21
Table 9-9. CEI-11G-MR Receiver Input Jitter Specification 22
23
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 24
Uncorrelated Bounded High Probability Jitter R_UBHPJ see R_UBHPJ in Table 8-5 UIpp 25
Correlated Bounded High probability Jitter R_CBHPJ see R_CBHPJ in Table 8-5 UIpp 26
Gaussian Jitter (UUGJ + CBGJ) R_GJ Note 2 see R_GJ in Table 8-5 UIpp
27
28
Sinusoidal Jitter, Maximum R_SJ-max See 2.2.4 see R_SJmax in Table 8-5 UIpp
29
Sinusoidal Jitter, High Frequency R_SJ-hf See 2.2.4 see R_SJ-hf in Table 8-5 UIpp 30
Total Jitter, including R_SJ-hf R_TJ Note 1 see R_TJ in Table 8-5 UIpp 31
Eye Mask incl. Correlated High Probability. Jitter R_X1 see R_X1 in Table 8-5 UI 32
Eye Mask R_Y1 see R_GJ in Table 8-5 mV 33
Eye Mask R_Y2 600 mV
34
35
NOTES:
1. TJ includes high frequency sinusoidal jitter. The receiver must tolerate the total deterministic and random jitter with addition of 36
the sinusoidal jitter. For transparent applications the specified jitter tolerance mask replace R_SJ. 37
2. BER=10-15, Q=7.94
38
39
. 40
41
9.3.3.1 Input Baud Rate 42
43
Refer to 8.3.2. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 205

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9.3.3.2 Reference Input Signals

2
3 Reference input signals to the receiver shall have the characteristics determined by a
4 compliant driver. The reference input signal must satisfy the transmitter near-end
5 template and jitter given in Figure 1-4 and Table 9-3, as well as the far-end eye
6 template and jitter given in Figure 1-5 and Table 9-9, with the differential
7 loadimpedence of 100 ohms +/- 1% at DC and a return loss of better than 20dB from
8 baud rate over 1667 to 1.5 times the baud rate. Note that the input signal might not
9 meet either of these templates when the actual receiver replaces this load.
10
11 9.3.3.3 Input Resistance and Return Loss
12
13 Please refer to 3.2.10 with the parameters shown in Table 8-6.
14
15
16 9.3.3.4 Input Lane-to-Lane Skew
17
18 Please refer to 3.2.8
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

206 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
9.A Appendix - Informative Jitter Budgets 4
5
9.A.1 Informative Jitter Budget for Long Reach 6
7
The following table is an informative jitter budget for long reach. It includes the specified 8
transmit jitter and an estimate of receiver jitter. A receiver may trade its ability to 9
equalize against its own internal jitter; possibly leading to different numbers than are 10
shown here. The receiver must tolerate sinusoidal jitter in addition to jitter contained in 11
this table. 12
13
Although only total jitter (TJ) and Uncorrelated Bounded High Probability Jitter (UBHPJ) 14
are normative to the specification, a realistic jitter budget must account for Uncorrelated 15
Unbounded Gaussian Jitter (UUGJ) of both the Receiver and Transmitter as well as 16
Correlated Bounded Gaussian Jitter of the Channel. A budget based entirely on 17
Uncorrelated bounded high Probability Jitter would be overly pessimistic or would 18
unfairly burden the equalization. 19
20
Table 9-10. CEI-11G-LR Informative Jitter Budget
21
Uncorrelated Jitter Correlated Jitter Total Jitter 22
Source
Unbounded Bounded Bounded Bounded
Amplitude 23
Gaussian High Prob. Gaussian High Prob.
Gaussian Sinusoidal High Prob. Total
24
Abbreviation UUGJ UBHPJ CBGJ CBHPJ k 25
Unit UIpp UIpp UIpp UIpp UIpp UIpp UIpp UIpp mVppd
26
27
Transmitter 0.150 0.150 0.150 0.150 0.300 800
28
Channel 0.230 0.400
29
Receiver Input 0.150 0.150 0.230 0.400 0.275 0.550 0.825 0
0 30
See 2 31
Equalizer
-0.300 32
See 1 33
Post Equalizer 0.150 0.150 0.230 0.100 0.275 0.250 0.525 0.25 100 34
DFE Penalties 0.100 -45 35
Clock & Sampler 0.150 0.100 0.100 -45 36
Budget 0.212 0.250 0.230 0.300 0.313 0.050 0.550 0.913 0.13 10
37
38
Note:
1. Due to receiver equalization, it reduces the ISI as seen inside the receiver. Thus this number is negative. 39
2. It is assumed that the eye is closed at the receiver, hence receiver equalization is required. 40
3. Values in yellow are specified values fromTable 9-5and Table 9-6
41
42
9.A.2 Informative Jitter Budget for Medium Reach 43
44
The following table is an informative jitter budget for medium reach. It includes the 45
specified transmit jitter and an estimate of receiver jitter. A receiver may trade its ability 46
to equalize against its own internal jitter; possibly leading to different numbers than are 47
shown here. The receiver must tolerate sinusoidal jitter in addition to jitter contained in 48
this table. 49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 207

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Although only total jitter (TJ) and Uncorrelated Bounded High Probability Jitter (UBHPJ)
2 are normative to the specification, a realistic jitter budget must account for Uncorrelated
3 Unbounded Gaussian Jitter (UUGJ) of both the Receiver and Transmitter as well as
4 Correlated Bounded Gaussian Jitter of the Channel. A budget based entirely on
5 Uncorrelated bounded high Probability Jitter would be overly pessimistic or would
6 unfairly burden the equalization.
7
8 Table 9-11. CEI-11G-MR Informative Jitter Budget
9 Uncorrelated Jitter Correlated Jitter Total Jitter
10 Source Amplitude
Unbounded Bounded Bounded Bounded
11 Gaussian High Prob. Gaussian High Prob.
Gaussian Sinusoidal High Prob. Total

12 Abbreviation UUGJ UBHPJ CBGJ CBHPJ k

13
Unit UIpp UIpp UIpp UIpp UIpp UIpp UIpp UIpp mVppd
14
Transmit equalizer -0.200
15
16 Transmitter 0.150 0.150 -0.200 0.150 -0.050 0.100 800
17 Channel 0.100 0.132 0.400 0.0
18 Receiver Input 0.150 0.250 0.132 0.200 0.200 0.050 0.450 0.700 0 110
19 Clock & Sampler 0.150 0.100 0.100 -45
20 Budget 0.212 0.350 0.132 0.300 0.250 0.050 0.650 0.950 0.13 10
21
Note:
22 1. Due to receiver equalization, it reduces the ISI as seen inside the receiver. Thus this number is negative.
23 2. Values in yellow are specified values fromTable 9-8and Table 9-9
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

208 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

9.B Appendix - StatEye.org templates 1

2
3
9.B.1 StatEye.org templates for CEI-11G-LR, reference receiver A 4
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 5
6
% example template for setting up a standard, i.e. equaliser 7
% jitter and return loss 8
9
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 10
11
param.version = [param.version ’_v1.0’]; 12
13
% these are internal variables and should not be changed 14
15
param.scanResolution = 0.01; 16
param.binsize = 0.0005; 17
param.points = 2^13; 18
19
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 20
21
% set the transmitter and baud rate. The tx filter has two 22
% parameters defined for the corner frequency of the poles 23
24
param.bps = 11.1e9; 25
param.bitResolution = 1/(3*param.bps); 26
param.txFilter = ’twopole’; 27
param.txFilterParam = [0.75 0.75]; 28
29
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 30
31
% set the return loss up. The return loss can be turned off 32
% using the appropriate option 33
34
param.returnLoss = ’on’; 35
param.cpad = 0.60; 36
37
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 38
39
% set the transmitter emphasis up. Some example setting are 40
% included which can be uncommented 41
42
% single tap emphasis 43
param.txpre = [-0.1]; 44
param.signal = 1.0; 45
param.txpost = [-0.1]; 46
param.vstart = [-0.3 -0.3]; 47
param.vend = [+0.0 +0.0]; 48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 209

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 param.vstep = [0.1 0.05 0.025];

2
3 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4
5 % set the de-emphasis of 4-point transmit pulse
6 % the de-emphasis run if param.txpre = [] and param.txpost = []
7
8 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
9
10 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11
12 % set the data coding changing the transmit pulse spectrum
13 % the coding run if param.txpre = [] and param.txpost = []
14
15 param.datacoding = 1; % the coding is off
16
17 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
18
19 % set PAM amplitude and rate
20
21 param.PAM = 2; % PAM is swithed off
22
23 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24
25 % the rxsample point does not need to be changed as it is
26 % automatically adjusted by the optimisation scripts.
27 % The number of DFE taps should be set, however, the initial
28 % conditions are irrelevant.
29
30 param.rxsample = -0.1;
31
32 param.dfe = [0.3 0.1 0.1 0.1];
33
34 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
35
36 % The CTE shall be controlled.
37
38 param.cte = 0; % CTE setting “0” = off; “1” = on;
39 param.ctethresh = 0; % max gain;
40
41 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42
43 % sampling jitter in HPJpp and GJrms is defined here
44
45 param.txdj = 0.15;
46 param.txrj = 0.15/(2*7.94);
47
48 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49

210 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

% the following options are not yet implemented and should 1

% not be changed 2
3
param.user = [0.0]; 4
param.useuser = ’no’; 5
param.usesymbol = ’’; 6
param.xtAmp = 1.0; 7
8
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 9
10
param.TransmitAmplitude = 0.800; % mVppdif 11
param.MinEye = 0.100; % mVppdif 12
13
param.Q = 2*7.94; 14
param.maxDJ = 0.275; 15
param.maxTJ = 0.525; 16
17
18
19
9.B.2 StatEye.org Templates for CEI-11G-LR, reference receiver B 20
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
21
22
% example template for setting up a standard, i.e. equaliser
23
% jitter and return loss
24
25
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26
27
param.version = [param.version ’_v1.0’];
28
29
% these are internal variables and should not be changed
30
31
param.scanResolution = 0.01;
32
param.binsize = 0.0005;
33
param.points = 2^13;
34
35
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
36
37
% set the transmitter and baud rate. The tx filter has two
38
% parameters defined for the corner frequency of the poles
39
40
param.bps = 11.1e9;
41
param.bitResolution = 1/(3*param.bps);
42
param.txFilter = ’twopole’;
43
param.txFilterParam = [0.75 0.75];
44
45
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
46
47
% set the return loss up. The return loss can be turned off
48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 211

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 % using the appropriate option

2
3 param.returnLoss = ’on’;
4 param.cpad = 0.60;
5
6 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7
8 % set the transmitter emphasis up. Some example setting are
9 % included which can be uncommented
10
11 % single tap emphasis
12 param.txpre = [-0.1];
13 param.signal = 1.0;
14 param.txpost = [-0.1];
15 param.vstart = [-0.3 -0.3];
16 param.vend = [+0.0 +0.0];
17 param.vstep = [0.1 0.05 0.025];
18
19 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20
21 % set the de-emphasis of 4-point transmit pulse
22 % the de-emphasis run if param.txpre = [] and param.txpost = []
23
24 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
25
26 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27
28 % set the data coding changing the transmit pulse spectrum
29 % the coding run if param.txpre = [] and param.txpost = []
30
31 param.datacoding = 1; % the coding is off
32
33 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
34
35 % set PAM amplitude and rate
36
37 param.PAM = 2; % PAM is swithed off
38
39 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40
41 % the rxsample point does not need to be changed as it is
42 % automatically adjusted by the optimisation scripts.
43 % The number of DFE taps should be set, however, the initial
44 % conditions are irrelevant.
45
46 param.rxsample = -0.1;
47
48 param.dfe = [];
49

212 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 1
2
% The CTE shall be controlled. 3
4
param.cte = 1; % CTE setting “0” = off; “1” = on; 5
param.ctethresh = 3; % max gain; 6
7
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 8
9
% sampling jitter in HPJpp and GJrms is defined here 10
11
param.txdj = 0.15; 12
param.txrj = 0.15/(2*7.94); 13
14
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 15
16
% the following options are not yet implemented and should 17
% not be changed 18
19
param.user = [0.0]; 20
param.useuser = ’no’; 21
param.usesymbol = ’’; 22
param.xtAmp = 1.0; 23
24
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 25
26
param.TransmitAmplitude = 0.800; % mVppdif 27
param.MinEye = 0.100; % mVppdif 28
29
param.Q = 2*7.94; 30
param.maxDJ = 0.275; 31
param.maxTJ = 0.525; 32
33
34
35
36
37
9.B.3 StatEye.org templates for CEI-11G-MR reach 38
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 39
40
% example template for setting up a standard, i.e. equaliser 41
% jitter and return loss 42
43
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 44
45
param.version = [param.version ’_v1.0’]; 46
47
% these are internal variables and should not be changed 48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 213

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 param.scanResolution = 0.01;
2 param.binsize = 0.0005;
3 param.points = 2^13;
4
5 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6
7 % set the transmitter and baud rate. The tx filter has two
8 % parameters defined for the corner frequency of the poles
9
10 param.bps = 11.1e9;
11 param.bitResolution = 1/(3*param.bps);
12 param.txFilter = ’twopole’;
13 param.txFilterParam = [0.75 0.75];
14
15 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16
17 % set the return loss up. The return loss can be turned off
18 % using the appropriate option
19
20 param.returnLoss = ’on’;
21 param.cpad = 0.60;
22
23 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
24
25 % set the transmitter emphasis up. Some example setting are
26 % included which can be uncommented
27
28 % single tap emphasis
29 param.txpre = [-0.1];
30 param.signal = 1.0;
31 param.txpost = [-0.1];
32 param.vstart = [-0.3 -0.3];
33 param.vend = [+0.0 +0.0];
34 param.vstep = [0.1 0.05 0.025];
35
36 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
37
38 % set the de-emphasis of 4-point transmit pulse
39 % the de-emphasis run if param.txpre = [] and param.txpost = []
40
41 param.txdeemphasis = [1 1 1 1]; % de-emphasis is off
42
43 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44
45 % set the data coding changing the transmit pulse spectrum
46 % the coding run if param.txpre = [] and param.txpost = []
47
48 param.datacoding = 1; % the coding is off
49 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

214 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
% set PAM amplitude and rate 2
3
param.PAM = 2; % PAM is swithed off 4
5
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 6
7
% the rxsample point does not need to be changed as it is 8
% automatically adjusted by the optimisation scripts. 9
% The number of DFE taps should be set, however, the initial 10
% conditions are irrelevant. 11
12
param.rxsample = -0.1; 13
14
param.dfe = []; 15
16
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 17
18
% The CTE shall be controlled. 19
20
param.cte = 0; % CTE setting “0” = off; “1” = on; 21
param.ctethresh = 0; % max gain; 22
23
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 24
25
% sampling jitter in HPJpp and GJrms is defined here 26
27
param.txdj = 0.15; 28
param.txrj = 0.15/(2*7.94); 29
30
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 31
32
% the following options are not yet implemented and should 33
% not be changed 34
35
param.user = [0.0]; 36
param.useuser = ’no’; 37
param.usesymbol = ’’; 38
param.xtAmp = 1.0; 39
40
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 41
42
param.TransmitAmplitude = 0.800; % mVppdif 43
param.MinEye = 0.100; % mVppdif 44
45
param.Q = 2*7.94; 46
param.maxDJ = 0.275; 47
param.maxTJ = 0.525; 48
49

Optical Internetworking Forum - Clause 9 : CEI-11G-LR/MR Long/Medium Reach Interface 215

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

216 Clause 9: CEI-11G-LR/MR Long/Medium Reach Interface - Optical Internetworking Forum

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

10 CEI-28G-SR Short Reach Interface 1

2
3
This clause details the requirements for the CEI-28G-SR short reach high speed 4
electrical interface between nominal baud rates of 19.90 Gsym/s and 28.05 Gsym/s 5
using NRZ coding. A compliant device shall meet all of the requirements listed below.
6
The electrical interface is based on high speed, low voltage logic. Connections are 7
point-to-point balanced differential pairs and signaling is unidirectional. 8
The electrical IA is based on loss and jitter budgets and defines the characteristics 9
required to communicate between a CEI-28G-SR transmitter and a CEI-28G-SR 10
receiver using copper signal traces on a printed circuit board. The characteristic 11
impedance of the signal traces is nominally 100  differential. A ‘length’ is effectively 12
defined in terms of its attenuation and phase response rather than its physical length. 13
Refer to Section 10.2.6 for channel requirements. 14
15
Short reach CEI-28G-SR devices from different manufacturers shall be interoperable. 16
17
18
10.1 Requirements 19
20
1. Support serial baud rates within the range from 19.90 Gsym/s to 28.05 Gsym/s. 21
22
2. Capable of low bit error ratio (10-15, with a test requirement to verify 10-12). 23
24
3. Capable of driving up to 300 mm of PCB and up to 1 connector. 25
4. Shall support AC-coupled operation 26
27
5. Shall allow multi-lanes (1 to n). 28
6. Shall support hot plug. 29
30
31
10.2 General Requirements 32
33
34
10.2.1 Data Patterns 35
Please refer to Section 3.2.1 36
37
38
10.2.2 Signal levels 39
40
Please refer to Section 3.2.2. All transmitter and receiver devices shall support “Load 41
Type 0”. Other load types are not supported by this clause. 42
43
10.2.3 Signal Definitions 44
45
Please refer to Section 1.A 46
47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 217
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 10.2.4 Bit Error Ratio

2
3 Please refer to Section 3.2.3
4
5 10.2.5 Ground Differences
6
7 Please refer to Section 3.2.4
8
9
10 10.2.6 Channel Compliance
11
12 A forward channel and associated dominant crosstalk channels are deemed compliant
13 if the channel characteristics conform to the requirements in this section.
14
15 10.2.6.1 Reference Model
16
17 The channel consists of PCB traces, vias, and 0 or 1 connector. The reference PCB
18 trace differential impedance is 100.
19
20 Figure 10-1 shows a diagram of test points on an example board.
21
22
23 Figure 10-1. CEI-28G-SR Reference Model
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 Note: Test points differ from definitions in Section 1.8, as DC blocking capacitors, if
41 physically located outside of the package, are part of the channel.
42
43
44
45
46
47
48
49

218 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Measured at these test points, several channel characteristics are parametrized. Port 1
definitions noted in Figure 2-33 allow proper measurement of the parameters in Table 2
10-1 used for calculation of the channel parameters found in Table 10-2. 3
4
Table 10-1. Measured Channel Parameters 5
Symbol Description 6
7
IL(f) Differential insertion loss, -SDD21 magnitude (dB)
8
RL1(f) Differential input return loss, -SDD11 magnitude (dB) 9
RL2(f) Differential output return loss, -SDD22 magnitude (dB) 10
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB)
11
12
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 13
14
Table 10-2. Calculated Channel Parameters
15
Symbol Description 16
ILfitted(f) Fitted insertion loss (dB)
17
18
ILD(f) Insertion loss deviation (dB) 19
ICN(f) Integrated crosstalk noise (mV, RMS) 20
FOMILD RMS value of the insertion loss Deviation (dB) 21
22
23
10.2.6.2 Insertion Loss 24
25
Channel insertion losses, including PCB traces and connectors, shall comply with the 26
limits specified by equations (10-1), (10-2) and plotted in Figure 10-2. Note that the vari- 27
able fb is the maximum baud rate to be supported by the channel under test (19.90 28
Gsym/s fb 28.05 Gsym/s). 29
30
Table 10-3. Channel Insertion Loss Frequency Range
31
Parameter Value Units 32
fmin 50 MHz 33
fmax fb GHz
34
35
36
37
38
fb  39
 f  28  05-
f  28  05- + 0 68 -----------------------
 0  1188 + 1  54 -----------------------   f min  f  ----  40
 fb fb 2  41
IL max =   (10-1) 42
 f  28  05- fb
 – 15  43 + 2 2 -----------------------  ----  f  f b  43
 fb 2  44
45
46
47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 219
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5  0 f min  f  1GHz 
6  
7 1  (10-2)
8 IL min = ---  f – 1   1GHz  f  175 GH z 
3 
9  55  17 5GH z  f  f b 
10  
11 
12 Note: f in (10-1) and (10-2) is in GHz.
13
14
15 Figure 10-2. CEI-28G-SR Normative Channel Insertion Loss at 28.05 Gsym/s
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

220 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

10.2.6.3 Fitted insertion loss 1

2
For fitted insertion loss definitions, please refer to Section 12.2.1.1 3
4
The channel shall meet the insertion loss requirements defined in Table 10-4. Note that 5
the variable fb is the maximum baud rate to be supported by the channel under test.  6
7
Table 10-4. Channel fitted insertion loss characteristics 8
Value 9
Parameter Units
Min. Max. 10
Minimum frequency, fILmin GHz 0.05 - 11
Maximum frequency, fILmax GHz - fb 12
Fitted Insertion loss at Nyquist dB - 15.42
Fitted insertion loss, a0 dB -1 1.5
13
Fitted insertion loss, a1 dB 0 9.533 14
Fitted insertion loss, a2 dB 0 30.855 15
Fitted insertion loss, a4 dB 0 14.162
16
17
18
10.2.6.4 Insertion loss deviation (ILD) 19
20
The insertion loss deviation ILD is the difference between the measured insertion IL 21
and the fitted insertion loss ILfitted as defined in (10-3). 22
23
24
ILD  IL  IL fitted (10-3) 25
26
The insertion loss deviation ILD shall be within the region defined by (10-4) and (10-5) 27
where fb is the maximum baud rate to be supported by the channel under test and fILmin 28
and fILmax are given in Table 10-4. 29
30
 1.0  12.0( f f b ) f ILmin  f  f b 4  31
ILD  ILDmin   (10-4) 32

  4 .0 f b 4  f  (3 / 4) f ILmax  33
34
1.0  12.0( f f b ) f ILmin  f  f b 4   35
ILD  ILDmax    (10-5) 36
 4.0 f b 4  f  (3 / 4) f ILmax  37
38
39
FOMILD, a figure of merit for the channel, is calculated as indicated below. In OIF-CEI- 40
03.0 and OIF-CEI-03.1, FOMILD was called ILDrms. 41
42
Define the weight at each frequency f using equation (10-6) below. 43
44
2 1 1 45
W  f  = sin c  f  f b  ------------------------4- -------------------------
-
8
(10-6) 46
1 +  f  ft  1 +  f  fr  47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 221
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum 20 to

2 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365 (e.g. T_tr
3 x ft = 0.2365), where T_tr is in nano seconds and ft is in GHz. In addition, fr is the -3 dB
4 reference receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum
5 baud rate to be supported by the channel.
6
7 2
8
9
 W  f   I LD  f 
FOM ILD = --------------------------------------------
(10-7)

10 N
11
12 where N is the number of frequency points, the summation is done over the frequency
13 range of ILD and FOMILD shall be less than 0.3dB for valid channels.
14
15 10.2.6.5 Channel differential return loss
16
17 Channel differential return loss shall be bounded by:
18
19
20 • RL(f) >= 12 dB for fmin < f  fb/4 (10-8)
21 • RL(f) >= 12 dB - 15 Log10(4f/fb) for fb/4 < f < fb 
22 (10-9)
23
24 Note: fmin is as defined in Table 10-3
25
26 10.2.6.6 Channel integrated crosstalk noise
27
28 Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of
29 Table 10-5, the total integrated crosstalk noise for the channel shall be less than the
30 value specified by Equation (10-10) and illustrated in Figure 10-3.
31
32 Table 10-5. Channel integrated crosstalk aggressor parameters
33 Parameter Symbol Value Units
34
35 Baud rate max. Baud Rate
fb Gsym/s
36 sup. by Channel
37 Near-end aggressor peak to peak differential output amplitude Ant 1200 mVppd
38 Far-end aggressor peak to peak differential output amplitude Aft 1200 mVppd
39
40 Near-end aggressor 20 to 80% rise and fall times Tnt 8 ps
41 Far-end aggressor 20 to 80% rise and fall times Tft 8 ps
42
43
44  x   x max = 10  mV RMS  for 3 dB  IL  5 3 dB (10-10)
45
46 = 12  4 – 0  45 IL  mV RMS  for 5  3 d B  IL  1542 dB
47 .
48 In Equation (10-10), the IL denotes the value of the channel insertion loss in dB at 1/2
49 baud rate (NRZ).

222 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 10-3. Illustration integrated crosstalk noise limits 1

2
12
3
4
Integrated crosstalk noise (mV, RMS)

10 5
6
7
8 8
9
6 10
11
12
4 13
14
15
2
16
17
0 18
3 5 7 9 11 13 15 17 19 19
20
Insertion loss at Nyquist (dB) 21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 223
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 10.3 Electrical Characteristics

2
3 The electrical signaling is based on high speed low voltage logic with a nominal
4 differential impedance of 100 .
5
6 All devices shall work within the range from 19.90 Gsym/s to 28.05 Gsym/s as specified
7 for the device, with the baud rate tolerance as per Section 3.2.11. Note that
8 implementation of specific protocols will define the operating baud rate without affecting
9 CEI compliance.
10
11
12 10.3.1 Transmitter Characteristics
13
The transmitter electrical specifications at compliance point T are given in Table 10-6.
14
The transmitter shall satisfy jitter requirements specified in Table 10-7. Jitter is
15
measured as specified in Section 2.3.3, for a BER as specified in Section 10.2.4. It is
16
assumed that the UBHPJ component of the transmitter jitter is not data-dependent jitter
17
(DDJ) from the receiver view point, hence it cannot be equalized in the receiver. To
18
attenuate noise and absorb even/odd mode reflections, the transmitter shall satisfy the
19
Common Mode Output Return Loss requirement of Table 10-6.
20
21 Link budgets in this document assume adaptive TX FIR equalization that is part of the
22 system management function. The specific implementation is outside the scope of this
23 document.
24
25 Table 10-6. Transmitter Electrical Output Specification.
26
27 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
28 Baud Rate T_Baud 19.90 28.05 Gsym/s
29
Emphasis off.
30 Output Differential Voltage T_Vdiff
See Note 4
800 1200 mVppd
31 Differential Resistance T_Rd 80 100 120 
32
33 Differential Termination Resistance
Mismatch (see Table 1-2) T_Rdm 10 %
34 Output Rise and Fall Time
Emphasis off.
35 (20% to 80%) T_tr, T_tf See Note 2 8 ps
36 Common Mode Noise T_Ncm Note 3 12 mVrms
37
Differential Output Return Loss T_SDD22 See Section 10.3.1.3 dB
38
Below 10 GHz -6
39 Common Mode Output Return Loss T_SCC22 dB
40 10 GHz to baud rate -4
41 Output Common Mode Voltage T_Vcm
Load Type 0
-100 1700 mV
See Note 1
42
43 NOTES:
1. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load.
44 2. The transmitter under test is preset such that C0 is its maximum value (C0_max) and all other coefficients are zero. The
45 20% and 80% values are of the steady state one and zero. The max value is limited by the linear fit pulse peak value in
Table 10-11.
46 3. Measurement procedure is defined in Section 12.3.
47 4. T_Vdiff is two times the steady-state value Vf as defined in Section 10.3.1.6.2. The value is given as differential p-p voltage.
48
49

224 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 10-7. Transmitter Output Jitter Specification 1

2
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 3
Uncorrelated Unbounded Gaussian Jitter T_UUGJ 0.15 UIPP 4
Uncorrelated Bounded High Probability Jitter T_UBHPJ Note 2 0.15 UIPP
5
6
Even-Odd Jitter (component of UBHPJ) T_EOJ Note 3 0.035 UIPP
7
Total Jitter T_TJ Note 1 0.28 UIPP
8
NOTES: 9
1. T_TJ includes all of the jitter components measured without any transmit equalization.
2. Measured with all possible values of transmitter equalization, excluding DDJ as defined in Section 12.1.1. 10
3. included in T_UBHPJ 11
12
13
14
10.3.1.1 Transmitter Baud Rate 15
16
All devices shall work within the range from 19.90 Gsym/s to 28.05 Gsym/s as specified 17
for the device, with the baud rate tolerance as per Section 3.2.11. Note that 18
implementation of specific protocols will define the operating baud rate without affecting 19
CEI compliance. 20
21
10.3.1.2 Transmitter Amplitude and Swing 22
23
Transmitter differential output amplitude shall be able to drive between 800 to 1200 24
mVppd with transmit emphasis disabled. The absolute transmitter output voltage shall 25
be between -0.3V and 1.9 V with respect to local ground. Transmitter differential output 26
amplitude shall additionally adhere to the requirements in Section 10.3.1.6. 27
28
29
10.3.1.3 Transmitter Resistance and Return Loss 30
31
Please refer to Section 3.2.10 with the following parameters. 32
Table 10-8. Transmitter Differential Return Loss Parameters 33
34
Parameter Value Units 35
A0 -12 dB 36
37
f0 50 MHz 38
0.1714 x 39
f1 Hz
T_Baud 40
f2 T_Baud Hz
41
42
Slope 12.0 dB/dec 43
44
10.3.1.4 Transmitter Lane-to-Lane Skew 45
46
Please refer to Section 3.2.7 47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 225
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 10.3.1.5 Transmitter Short Circuit Current

2
3 Please refer to Section 3.2.9
4
5 10.3.1.6 Transmitter output waveform requirements
6
7 The transmitter shall include an equalizer defined as:
8
9
10 H ( Z )  C 1  C 0 z  1  C 1 z  2 
11 (10-11)
12
13 10.3.1.6.1 Summary of requirements
14
15 The normalized amplitudes of the coefficients of the transmitter equalizer (computed
16 per 10.3.1.6.2) shall meet the requirements in Table 10-9.
17
18 Table 10-9. Coefficient range and step size
19
20 Normalized Amplitude Normalized
Coefficient
21 Min (%) Max (%) Step Size (%)
22
23 C-1 -10 0 1.25 to 5
24
C1 -25 0 1.25 to 5
25
26 C0 40 100 1.25 to 5
27
28 The amplitude of a coefficient can be computed by multiplying its normalized amplitude
29 by vf, which is defined in equation (10-12). "min" is defined as the minimum normalized
30 amplitude of the coefficient that must be supplied by the transmitter to be compliant.
31 "max" is defined as the maximum normalized amplitude of the coefficient that must be
32 supplied by the transmitter to be compliant.
33
34 In addition:
35
36 a) |C-1|+|C0|+|C1|, the peak output voltage shall not exceed 1200 mVppd.
37
38 b) C-1 + C0 + C1, the steady-state output voltage shall be greater than or equal to 140
39 mVppd.
40
41 10.3.1.6.2 Process to compute coefficients
42
43 The coefficients of the transmitter equalizer shall be determined from the measured
44 waveform during TX compliance test using the process described below.
45
46 1. The transmitter under test is preset such that C0 is its maximum value (C0_max)
47 and all other coefficients are zero.
48
49

226 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2. Capture at least one complete cycle of the test pattern PRBS9 at T [ T is defined 1
as the test point at the output of transmitter package] per 10.3.1.6.3. 2
3
3. Compute the linear fit to the captured waveform per 10.3.1.6.4. 4
5
4. Define tx to be the time where the rising edge of the linear fit pulse, p, from step 6
3 crosses 50% of its peak amplitude. 7
8
5. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time  9
t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi. 10
11
6. Use pi to compute the vector of coefficients, w, of a T_Nw-tap symbol-spaced 12
transversal filter that equalizes for the transfer function from the transmit function 13
to T per 10.3.1.6.5. 14
15
The parameters of the pulse fit and the equalizing filter are given in Table 10-10. 16
Table 10-10. Linear fit pulse and equalizing filter parameters 17
18
Parameter Value (UI) 19
20
Linear fit pulse length T_Np 8 21
Linear fit pulse delay T_Dp 2 22
Equalizer length T_Nw 8 23
24
Equalizer delay T_Dw 2 25
26
The differential zero to peak output voltage at T in the steady state, vf, is estimated by 27
equation (10-12). 28
29
30
1 M T _ Np (10-12) 31
vf    p(k )
M k 1 32
 33
In (10-12), p is the linear fit pulse from step 3 and M is the number of samples per 34
symbol as defined in 10.3.1.6.3. The peak value of the linear fit pulse from step 3, pmax, 35
shall satisfy the requirements of Table 10-11.The RMS value of the error between the 36
linear fit and measured waveform from step 3, e, shall satisfy the requirements of 37
Table 10-11. 38
39
Table 10-11. Transmitter output waveform requirements
40
Parameter Condition Units 41
42
Steady state output voltage, 2 x vf max mVppd 1200 43
Steady state output voltage, 2 x vf min mVppd 800 44
45
Linear fit pulse peak, pmax min - 0.80 x vf 46
RMS error, e max - 0.027 x vf 47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 227
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5 For each configuration of the transmit equalizer:
6
7 7. Configure the transmitter under test as required.
8
9 8. Capture at least one complete cycle of the test pattern PRBS9 at T.
10
11 9. Compute the linear fit to the captured waveform per 10.3.1.6.4.
12
13 10. Define tx to be the time where the rising edge of the linear fit pulse, p, from step
14 3 crosses 50% of its peak amplitude.
15 11. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time 
16 t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi.
17
18 12. Equalize the sampled pulse, pi, using the coefficient vector, w, computed in step
19 6 per 10.3.1.6.5 to yield the equalized pulse qi.
20
21 The RMS value of the error between the linear fit and measured waveform from step 9,
22  , shall satisfy the requirements of Table 10-11.
e
23
24 The normalized amplitude of coefficient C-1 is the value of qi at time t0 + (T_Dw - 1) UI.
25 The normalized amplitude of coefficient C0 is the value of qi at time t0 + T_Dw UI. 
26 The normalized amplitude of coefficient C1 is the value of qi at time t0 + (T_Dw +1) UI.
27
28
29 10.3.1.6.3 Waveform acquisition
30
The transmitter under test repetitively transmits the specified test pattern. The
31
waveform shall be captured with an effective sample rate that is M times the signaling
32
rate of the transmitter under test. The value of M shall be an integer not less than 7.
33
Averaging multiple waveform captures is recommended.
34
35 The captured waveform shall represent an integer number of repetitions of the test
36 pattern totaling N bits. Hence the length of the captured waveform should be M .N
37 samples. The waveform should be aligned such that the first M samples of waveform
38 correspond to the first bit of the test pattern, the second M samples to the second bit,
39 and so on.
40
41
42 10.3.1.6.4 Linear fit to the waveform measured at T
43
44 Given the captured waveform y(k) and corresponding aligned symbols x(n) derived
45 from the procedure defined in 10.3.1.6.2, define the M-by-N waveform matrix Y as
46 shown in (10-13).
47
48
49

228 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

 1
 y(1) y( M  1)  y( M ( N  1)  1)  2
 y(2 ) y (M  2)  y (M ( N  1)  2 ) (10-13) 3
Y  4
     
  5
 y ( M ) y (2 M )  y (MN )  6
 7
8
9
Rotate the symbols vector x by the specified pulse delay Dp to yield xr. 10
11
12

x r  x (T _ D p  1) x (T _ D p  2 )  x( N ) x (1)  x (T _ D p )  (10-14) 13
14
15
Define the matrix X to be an N-by-N matrix derived from xr as shown in (10-15). 16
17
 18
 xr (1) xr ( 2)  xr ( N )  19
 x ( N ) x (1) 20
 xr ( N  1)  (10-15)
X  21
r r
      22
  23
 xr (2) xr (3)  xr (1) 
24
25
Define the matrix X1 to be the first T_Np rows of X concatenated with a row vector of 1's 26
of length N. The M-by-(T_Np + 1) coefficient matrix, P, corresponding to the linear fit is 27
then defined by (10-16). 28
29
30
P  YX 1T ( X 1 X 1T ) 1 (10-16) 31
32
33
34
In (10-16) the superscript "T" denotes the matrix transpose operator. 35
36
 37
 e(1) e( M  1)  e(M ( N  1)  1)  
 e( 2) e( M  2) 38
 e( M ( N  1)  2)  39
E  PX 1  Y   (10-17)
      40
  41
e( M ) e(2 M )  e(MN )  42
43
44
The error waveform, e(k), is then read column-wise from the elements of E. 45
46
Define P1 to be a matrix consisting of the first T_Np columns of the matrix P as shown 47
in (10-18). 48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 229
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 
2  p(1) p( M  1)  p ( M (T _ N p  1)  1)  
3  p (2 ) p ( M  2 )  p( M (T _ N  1)  2 )
P1   
p
4 (10-18)
5      
 
6  p( M ) p(2 M )  p ( MT _ N p ) 
7
8
9
10 The linear fit pulse response, p(k), is then read column-wise from the elements of P1.
11
12 10.3.1.6.5 Removal of the transfer function between the transmit function and T
13
14 Rotate sampled pulse response pi by the specified equalizer delay T_Dw to yield pr as
15 shown in (10-19).

16
17 
pr  pi (T _ Dw  1) pi (T _ Dw  2)  pi (T _ N p ) pi (1)  pi (T _ Dw )  
18 (10-19)
19
20 Define the matrix P2 to be a T_Np-by-T_Np matrix derived from pr as shown in (10-20).
21
22 
 pr (1) pr (T _ N p )  pr (2) 
23  p (2) 
24 pr (1)  pr (3) 
P2  
r
25 (10-20)
     
26  
27  pr (T _ N p ) pr (T _ N p  1)  pr (1) 
28 
29 Define the matrix P to be the first T_N rows of P . Define a unit pulse column vector
3 w 2
30 x of length T_N . The value of element xp(T_Dp + 1) is 1 and all other elements have a
p p
31 value of 0. The vector of filter coefficients w that equalizes p is then defined by (10-21).
i
32
33 
34 w  (P3T P3 ) 1 P3T x p

35 (10-21)
36
37 Given the column vector of equalizer coefficients, w, the equalized pulse response qi is
38 determined by (10-22).
39
40 qi  P3 w 
41 
42 (10-22)
43
44
45
46
47
48
49

230 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

10.3.2 Receiver Characteristics 1

2
A compliant receiver shall operate at the specified BER with the worst case 3
combination of a compliant transmitter and a compliant channel. 4
5
Receiver electrical specifications are given in Table 10-12 and measured at compliance 6
point R. To dampen noise sources and absorption of both even and odd mode 7
reflections, the receiver shall satisfy the Common Mode Input Return Loss requirement 8
of Table 10-12. Jitter specifications at reference R are listed in Table 10-13. 9
10
Table 10-12. Receiver Electrical Input Specifications 11
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
12
13
Baud rate R_Baud 19.90 28.05 GSym/s
14
Input Differential Voltage R_Vdiff Note 1 1200 mVppd 15
Differential Input Impedance R_Rdin 80 100 120  16
Input Impedance Mismatch R_Rm 10 % 17
Differential Input Return Loss R_SDD11 See 10.3.2.3 18
Below 10 GHz -6
19
Common Mode Input Return Loss R_SCC11 dB 20
10GHz to baud rate -4
21
Input Common Mode Voltage R_Vcm
Load Type 0 
See Note 2
-200 1800 mV 22
23
NOTES:
1. The receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the combined 24
transmitter and channel. The channel response shall include the worst case effects of the return losses at the transmitter and 25
receiver.
2. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load. For floating load, input resistance shall be 1k
26
27
28
29
Table 10-13. Receiver Input Jitter Specification 30
31
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
32
Sinusoidal Jitter, Maximum R_SJ-max See Section 5 UIpp 33
2.3.4, note 1
34
Sinusoidal Jitter, High Frequency R_SJ-hf
See Section
2.3.4, note 1 0.05 UIpp 35
36
NOTES:
1. The Receiver shall tolerate the sum of these jitter contributions: Total transmitter jitter from Table 10-7;Sinusoidal jitter as 37
defined in Table 10-13;The effects of a channel compliant to the Channel Characteristics (Section 10.2.6). 38
39
10.3.2.1 Input Baud Rate 40
41
All devices shall work within the range from 19.90 Gsym/s to 28.05 Gsym/s as specified 42
for the device, with the baud rate tolerance as per Section 3.2.11. 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 10: CEI-28G-SR Short Reach Interface 231
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 10.3.2.2 Reference Input Signals

2
3 The receiver shall accept differential input signal amplitudes produced by a compliant
4 transmitter connected with the minimum attenuation specified in Figure 10-2 to the
5 receiver. This may be larger than the 1200 mVppd maximum of the transmitter due to
6 output/input impedances and reflections.
7
8 The minimum input amplitude is defined by the minimum transmitter amplitude, the
9 actual receiver input impedance and the loss of the actual PCB. Note that the minimum
10 transmitter amplitude is defined using a well controlled load impedance, however the
11 real receiver is not, which can leave the receiver input signal smaller than expected.
12 Additionally it will be determined by the environmental noise inside and outside the
13 receiver.
14
15 10.3.2.3 Input Resistance and Return Loss
16
17 Please refer to Section 3.2.10 with the following parameters.
18
19 Table 10-14. Receiver Differential Return Loss Parameters
20 Parameter Value Units
21
A0 -12 dB
22
23 f0 50 MHz

24 f1
0.1714 x
Hz
R_Baud
25
26 f2 R_Baud Hz

27 Slope 12.0 dB/dec

28
29 10.3.2.4 Input Lane-to-Lane Skew
30
31 Please refer to Section 3.2.8.
32
33
34 10.3.2.5 Absolute Input Voltage
35
36 The absolute voltage levels with respect to the receiver ground at the input of the
37 receiver are dependent on the transmitter implementation and the inter-ground
38 difference. The voltage levels at the input of an AC coupled receiver (if the effective AC
39 coupling is done within the receiver) or at the TX side of the external AC coupling cap (if
40 AC coupling is done externally) shall be between -0.3 to 2.0V with respect to local
41 ground.
42
43
44
45
46
47
48
49

232 Clause 10: CEI-28G-SR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

11 CEI-25G-LR Long Reach Interface 1

2
3
This clause details the requirements for the CEI-25G-LR long reach high speed 4
electrical interface between nominal baud rates of 19.90 Gsym/s and 25.80 Gsym/s 5
using NRZ coding. A compliant device shall meet all of the requirements listed below.
6
The electrical interface is based on high speed, low voltage logic. Connections are 7
point-to-point balanced differential pairs and signaling is unidirectional. 8
The electrical IA is based on loss and jitter budgets and defines the characteristics 9
required to communicate between a CEI-25G-LR transmitter and a CEI-25G-LR 10
receiver using copper signal traces on a printed circuit board. The characteristic 11
impedance of the signal traces is nominally 100  differential. A ‘length’ is effectively 12
defined in terms of its attenuation and phase response rather than its physical length. 13
Refer to Section 11.2.6 for transmission line guidelines to meet the channel 14
requirements. 15
16
Long reach CEI-25G-LR devices from different manufacturers shall be interoperable. 17
18
19
20
21
11.1 Requirements 22
23
1. Support serial baud rates within the range from 19.90 Gsym/s to 25.80 Gsym/s. 24
25
2. Capable of low bit error ratio (10-15, with a test requirement to verify 10-12). 26
3. Capable of driving up to 686 mm of PCB and up to 2 connectors. 27
28
4. Shall support AC-coupled operation. 29
30
5. Shall allow multi-lanes (1 to n). 31
6. Shall support hot plug. 32
33
34
11.2 General Requirements 35
36
11.2.1 Data Patterns 37
38
Please refer to Section 3.2.1 39
40
41
11.2.2 Signal levels 42
43
Please refer to Section 3.2.2. All transmitter and receiver devices shall support “Load 44
Type 0”. Other load types are not supported by this clause. 45
46
11.2.3 Signal Definitions 47
48
Please refer to Section 1.A 49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 233
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 11.2.4 Bit Error Ratio

2
3 Please refer to Section 3.2.3
4
5 11.2.5 Ground Differences
6
7 Please refer to Section 3.2.4
8
9
10 11.2.6 Channel Compliance
11
12 A forward channel and associated dominant crosstalk channels are deemed compliant
13 if the channel characteristics conform to the requirements in this section.
14
15 11.2.6.1 Reference Model
16
17 The channel consists of PCB traces, vias, and up to 2 connectors. The reference PCB
18 trace differential impedance is 100.
19
20 Figure 11-1 shows a diagram of test points on an example board.
21
22
23 Figure 11-1. CEI-25G-LR Reference Model
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Note: Test points differ from definitions in Section 1.8, as DC blocking capacitor, if
45
physically located outside of the package, is part of the channel.
46
47
48
49

234 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Measured at these test points, several channel characteristics are parametrized. Port 1
definitions as noted in Figure 2-33 allow proper measurement of the parameters in 2
Table 11-1 used for calculation of the channel parameters found in Table 11-2. 3
4
Table 11-1. Measured Channel Parameters 5
Symbol Description 6
7
IL(f) Differential insertion loss, -SDD21 magnitude (dB)
8
RL1(f) Differential input return loss, -SDD11 magnitude (dB) 9
RL2(f) Differential output return loss, -SDD22 magnitude (dB) 10
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB)
11
12
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 13
14
Table 11-2. Calculated Channel Parameters
15
Symbol Description 16
ILfitted(f) Fitted insertion loss (dB)
17
18
ILD(f) Insertion loss deviation (dB) 19
ICN(f) Integrated crosstalk noise (mV, RMS) 20
FOMILD RMS value of the insertion loss Deviation (dB) 21
22
23
11.2.6.2 Insertion Loss 24
25
Channel insertion losses, including PCB traces and connectors, shall comply with the 26
limits specified by equations (11-1), (11-2) and plotted in Figure 11-2. Note that the vari- 27
able fb is the maximum baud rate to be supported by the channel under test (19.90 28
Gsym/s fb 25.80 Gsym/s). 29
30
Table 11-3. Channel Insertion Loss Frequency Range 31
Parameter Value Units 32
fmin 50 MHz
33
34
fmax fb GHz
35
36
37
38
39
 f  258 f  258 fb  40
- + 0 96 --------------------
 1 083 + 3  35 -------------------- - f min  f  ---- 
 fb fb 2 (11-1) 41
IL max =   42
 f  258- f
– 9 25 + 2 694 --------------------  ---b-  f  f b  43
 fb 2  44
  45
46
47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 235
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2  0 f min  f  1GHz 
 
3 1  (11-2)
4 IL min =  ---  f – 1   1GHz  f  175 GH z
5 3 
 5 5  175 GH z  f  f b 
6  
7
8
9
10 Note: f in (11-1) and (11-2) is in GHz.
11
12 Figure 11-2. CEI-25G-LR Normative Channel Insertion Loss at 25.80 Gsym/s.
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

236 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

11.2.6.3 Fitted insertion loss 1

2
For fitted insertion loss definitions, please refer to Section 12.2.1.1 3
4
The channel shall meet the insertion loss requirements defined in Table 11-4. Note that 5
the variable fb is the maximum baud rate to be supported by the channel under test.  6
7
Table 11-4. Channel fitted insertion loss characteristics 8
Value 9
Parameter Units
Min. Max. 10
Minimum frequency, fILmin GHz 0.05 - 11
Maximum frequency, fILmax GHz - fb 12
Fitted Insertion loss at Nyquist dB - 25.5
Fitted insertion loss, a0 dB -1 2.0
13
Fitted insertion loss, a1 dB 0 20.317 14
Fitted insertion loss, a2 dB 0 51.6 15
Fitted insertion loss, a4 dB 0 25.294
16
17
18
11.2.6.4 Insertion loss deviation (ILD) 19
20
The insertion loss deviation ILD is the difference between the measured insertion IL 21
and the fitted insertion loss ILfitted as defined in (11-3). 22
23
24
ILD  IL  IL fitted (11-3) 25
26
The insertion loss deviation ILD shall be within the region defined by (11-4) and (11-5) 27
where fb is the maximum baud rate to be supported by the channel under test and fILmin 28
and fILmax are given in Table 11-4. 29
30
 1.0  12.0( f f b ) f ILmin  f  f b 4  31
ILD  ILDmin   (11-4) 32

  4. 0 f b 4  f  (3 / 4) f ILmax  33
34
1.0  12.0( f f b ) f ILmin  f  f b 4   35
ILD  ILDmax    (11-5) 36
 4.0 f b 4  f  (3 / 4) f ILmax  37
38
39
FOMILD, a figure of merit for the channel, is calculated as indicated below. In OIF-CEI- 40
03.0 and OIF-CEI-03.1, FOMILD was called ILDrms. 41
42
Define the weight at each frequency f using equation (11-6) below. 43
44
2 1 1 45
W  f  = sin c  f  f b  ------------------------4- -------------------------
-
8
(11-6) 46
1 +  f  ft  1 +  f  fr  47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 237
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum 20 to

2 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365 (e.g. T_tr
3 x ft = 0.2365), where T_tr is in nano seconds and ft is in GHz. In addition, fr is the -3 dB
4 reference receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum
5 baud rate to be supported by the channel.
6
7 2
8
9
 W  f   I LD  f 
FOM ILD = --------------------------------------------
(11-7)

10 N
11
12 where N is the number of frequency points, the summation is done over the frequency
13 range of ILD and FOMILD shall be less than 0.3dB for valid channels.
14
15 11.2.6.5 Channel Return Loss
16
17 Channel Return Loss shall be bounded by:
18
19
20 • RL(f) >= 12 dB for fmin < f  fb/4 
21 • RL(f) >= 12 dB - 15 Log10(4f/fb) for fb/4 < f < fb (11-8)
22 
23 (11-9)
24
25 Note: fmin is as defined in Table 11-3
26
27 11.2.6.6 Channel integrated crosstalk noise
28
29 Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of
30 Table 11-5, the total integrated crosstalk noise for the channel shall be less than the
31 value specified by Equation (11-10) and illustrated in Figure 11-3.
32
Table 11-5. Channel integrated crosstalk aggressor parameters
33
34 Parameter Symbol Value Units
35
36 Baud rate max. Baud Rate
fb Gsym/s
sup. by Channel
37
38 Near-end aggressor peak to peak differential output amplitude Ant 1200 mVppd
39 Far-end aggressor peak to peak differential output amplitude Aft 1200 mVppd
40 Near-end aggressor 20 to 80% rise and fall times Tnt 8 ps
41
42 Far-end aggressor 20 to 80% rise and fall times Tft 8 ps
43
44  x   x max = 10  mV RMS  for 3 dB  IL  5 3 dB
45 (11-10)
46 = 12  4 – 0  45 IL  mV RMS  for 5  3 d B  IL  25 5 dB
47 
48 In Equation (11-10), the IL denotes the value of the channel insertion loss in dB at 1/2 .
49 baud rate (NRZ).

238 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 11-3. Illustration integrated crosstalk noise limits 1

2
12 3
4
5
Integrated crosstalk noise (mV, RMS)

10
6
7
8 8
9
10
6 11
12
4 13
14
15
2 16
17
18
0
19
3 5 7 9 11 13 15 17 19 21 23 25 27
20
Insertion loss at Nyquist (dB) 21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 239
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 11.3 Electrical Characteristics

2
3 The electrical signaling is based on high speed low voltage logic with a nominal
4 differential impedance of 100 .
5
6 All devices shall work within the range from 19.90 Gsym/s to 25.80 Gsym/s as specified
7 for the device, with the baud rate tolerance as per Section 3.2.11. Note that
8 implementation of specific protocols will define the operating baud rate without affecting
9 CEI compliance.
10
11
12 11.3.1 Transmitter Characteristics
13
The transmitter electrical specifications at compliance point T are given in Table 11-6.
14
The transmitter shall satisfy jitter requirements specified in Table 11-7. Jitter is
15
measured as specified in Section 2.4.3, for a BER as specified in Section 11.2.4. It is
16
assumed that the UBHPJ component of the transmitter jitter is not data-dependent jitter
17
(DDJ) from the receiver view point, hence it cannot be equalized in the receiver. To
18
attenuate noise and absorb even/odd mode reflections, the transmitter shall satisfy the
19
Common Mode Output Return Loss requirement of Table 11-6.
20
21 Link budgets in this document assume adaptive TX FIR equalization that is part of the
22 system management function. The specific implementation is outside the scope of this
23 document.
24
25 Table 11-6. Transmitter Electrical Output Specification.
26
27 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
28 Baud Rate T_Baud 19.90 25.80 Gsym/s
29
Emphasis off.
30 Output Differential Voltage T_Vdiff
See Note 4.
800 1200 mVppd
31 Differential Resistance T_Rd 80 100 120 
32
33 Differential Termination Resistance
Mismatch (see Table 1-2) T_Rdm 10 %
34 Output Rise and Fall Time Emphasis off.
35 (20% to 80%) T_tr, T_tf
See Note 2.
8 ps
36 Common Mode Noise T_Ncm See Note 3. 12 mVrms
37
Differential Output Return Loss T_SDD22 See Section 11.3.1.3 dB
38
Below 10 GHz -6
39 Common Mode Output Return Loss T_SCC22 dB
40 10 GHz to baud rate -4
41 Output Common Mode Voltage T_Vcm
Load Type 0
-100 1700 mV
See Note 1
42
43 NOTES:
1. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load.
44 2. The transmitter under test is preset such that C0 is its maximum value (C0_max) and all other coefficients are zero. The 20%
45 and 80% values are of the steady state one and zero. The max value is limited by the linear fit pulse peak value in Table 11-
11.
46 3. Measurement procedure is defined in Section 12.3.
47 4. T_Vdiff is two times the steady-state value Vf as defined in Section 11.3.1.6.2. The value is given as differential p-p voltage.
48
49

240 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 11-7. Transmitter Output Jitter Specification 1

2
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 3
Uncorrelated Unbounded Gaussian Jitter T_UUGJ 0.15 UIPP 4
Uncorrelated Bounded High Probability Jitter T_UBHPJ Note 2 0.15 UIPP
5
6
Even-Odd Jitter (component of UBHPJ) T_EOJ Note 3 0.035 UIPP
7
Total Jitter T_TJ Note 1 0.28 UIPP
8
NOTES: 9
1. T_TJ includes all of the jitter components measured without any transmit equalization.
2. Measured with all possible values of transmitter equalization, excluding DDJ as defined in Section 12.1.1. 10
3. included in T_UBHPJ 11
12
13
14
11.3.1.1 Transmitter Baud Rate 15
16
All devices shall work within the range from 19.90 Gsym/s to 25.80 Gsym/s as specified 17
for the device, with the baud rate tolerance as per Section 3.2.11. Note that 18
implementation of specific protocols will define the operating baud rate without affecting 19
CEI compliance. 20
21
11.3.1.2 Transmitter Amplitude and Swing 22
23
Transmitter differential output amplitude shall be able to drive between 800 to 1200 24
mVppd with transmit emphasis disabled. The absolute transmitter output voltage shall 25
be between -0.3V and 1.9 V with respect to local ground. Transmitter differential output 26
amplitude shall additionally adhere to the requirements in Section 11.3.1.6. 27
28
29
11.3.1.3 Transmitter Resistance and Return Loss 30
31
Please refer to Section 3.2.10 with the following parameters. 32
Table 11-8. Transmitter Differential Return Loss Parameters 33
34
Parameter Value Units 35
A0 -12 dB 36
37
f0 50 MHz 38
0.1714 x 39
f1 Hz
T_Baud 40
f2 T_Baud Hz
41
42
Slope 12.0 dB/dec 43
44
11.3.1.4 Transmitter Lane-to-Lane Skew 45
46
Please refer to Section 3.2.7 47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 241
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 11.3.1.5 Transmitter Short Circuit Current

2
3 Please refer to Section 3.2.9
4
5 11.3.1.6 Transmitter output waveform requirements
6
7 The transmitter shall include an equalizer defined as:
8
9 (11-11)
10 H ( Z )  C 1  C 0 z  1  C 1 z  2
11
12
13 11.3.1.6.1 Summary of requirements
14
The normalized amplitudes of the coefficients of the transmitter equalizer (computed
15
16 per 11.3.1.6.2) shall meet the requirements in Table 11-9.
17 Table 11-9. Coefficient range and step size
18
19 Normalized Amplitude Normalized
20 Coefficient
Min (%) Max (%) Step Size (%)
21
22 C-1 -25 0 1.25 to 5
23
24 C1 -25 0 1.25 to 5
25 C0 40 100 1.25 to 5
26
27 The amplitude of a coefficient can be computed by multiplying its normalized amplitude
28 by vf, which is defined in equation (11-12). "min" is defined as the minimum normalized
29 amplitude of the coefficient that must be supplied by the transmitter to be compliant.
30 "max" is defined as the maximum normalized amplitude of the coefficient that must be
31 supplied by the transmitter to be compliant.
32
33 In addition:
34
35 a) |C-1|+|C0|+|C1|, the peak output voltage shall not exceed 1200 mVppd.
36
37 b) C-1 + C0 + C1, the steady-state output voltage shall be greater than or equal to 80
38 mVppd.
39
40
11.3.1.6.2 Process to compute coefficients
41
42 The coefficients of the transmitter equalizer shall be determined from the measured
43 waveform during TX compliance test using the process described below.
44
45 1. The transmitter under test is preset such that C0 is its maximum value (C0_max)
46 and all other coefficients are zero.
47
48
49

242 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

2. Capture at least one complete cycle of the test pattern PRBS9 at T [ T is defined 1
as the test point at the output of transmitter package] per 11.3.1.6.3. 2
3
3. Compute the linear fit to the captured waveform per 11.3.1.6.4. 4
5
4. Define tx to be the time where the rising edge of the linear fit pulse, p, from step 6
3 crosses 50% of its peak amplitude. 7
8
5. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time  9
t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi. 10
11
6. Use pi to compute the vector of coefficients, w, of a T_Nw-tap symbol-spaced 12
transversal filter that equalizes for the transfer function from the transmit function 13
to T per 11.3.1.6.5. 14
15
The parameters of the pulse fit and the equalizing filter are given in Table 11-10. 16
Table 11-10. Linear fit pulse and equalizing filter parameters 17
18
Parameter Value (UI) 19
20
Linear fit pulse length T_Np 8 21
Linear fit pulse delay T_Dp 2 22
Equalizer length T_Nw 8 23
24
Equalizer delay T_Dw 2 25
26
The differential zero to peak output voltage at T in the steady state, vf, is estimated by 27
equation (11-12). 28
29
30
1 M T _ Np (11-12) 31
vf    p(k )
M k 1 32
 33
In (11-12), p is the linear fit pulse from step 3 and M is the number of samples per 34
symbol as defined in 11.3.1.6.3. The peak value of the linear fit pulse from step 3, pmax, 35
shall satisfy the requirements of Table 11-11.The RMS value of the error between the 36
linear fit and measured waveform from step 3, e, shall satisfy the requirements of 37
Table 11-11. 38
39
Table 11-11. Transmitter output waveform requirements
40
Parameter Condition Units 41
42
Steady state output voltage, 2 x vf max mVppd 1200 43
Steady state output voltage, 2 x vf min mVppd 800 44
45
Linear fit pulse peak, pmax min - 0.80 x vf 46
RMS error, e max - 0.027 x vf 47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 243
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3 For each configuration of the transmit equalizer:
4
5 7. Configure the transmitter under test as required.
6
7 8. Capture at least one complete cycle of the test pattern PRBS9 at T.
8
9 9. Compute the linear fit to the captured waveform per 11.3.1.6.4.
10
11 10. Define tx to be the time where the rising edge of the linear fit pulse, p, from step
12 3 crosses 50% of its peak amplitude.
13
14 11. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time 
t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi.
15
16 12. Equalize the sampled pulse, pi, using the coefficient vector, w, computed in step
17 6 per 11.3.1.6.5 to yield the equalized pulse qi.
18
19 The RMS value of the error between the linear fit and measured waveform from step 9,
20  , shall satisfy the requirements of Table 11-11.
e
21
22 The normalized amplitude of coefficient C is the value of q at time t + (T_D - 1) UI.
23 The normalized amplitude of coefficient C-1is the value of q iat time t 0 + T_D wUI. 
24 The normalized amplitude of coefficient C0 is the value of qi at time t0 + (T_Dw +1) UI.
1 i 0 w
25
26
27 11.3.1.6.3 Waveform acquisition
28
29 The transmitter under test repetitively transmits the specified test pattern. The
30 waveform shall be captured with an effective sample rate that is M times the signaling
31 rate of the transmitter under test. The value of M shall be an integer not less than 7.
32 Averaging multiple waveform captures is recommended.
33
The captured waveform shall represent an integer number of repetitions of the test
34
pattern totaling N bits. Hence the length of the captured waveform should be M .N
35
samples. The waveform should be aligned such that the first M samples of waveform
36
correspond to the first bit of the test pattern, the second M samples to the second bit,
37
and so on.
38
39
40 11.3.1.6.4 Linear fit to the waveform measured at T
41
42 Given the captured waveform y(k) and corresponding aligned symbols x(n) derived
43 from the procedure defined in 11.3.1.6.2, define the M-by-N waveform matrix Y as
44 shown in (11-13).
45
46
47
48
49

244 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

 1
 y(1) y( M  1)  y( M ( N  1)  1)  2
 y (2 ) y (M  2)  y (M ( N  1)  2 ) (11-13) 3
Y  4
     
  5
 y( M ) y (2 M )  y (MN )  6
 7
8
9
Rotate the symbols vector x by the specified pulse delay Dp to yield xr. 10
11
12

xr  x(T _ Dp  1) x(T _ Dp  2)  x( N ) x(1)  x(T _ Dp )  (11-14) 13
14
15
Define the matrix X to be an N-by-N matrix derived from xr as shown in (11-15). 16
17
 18
 xr (1) xr ( 2)  xr ( N )  19
 x ( N ) x (1) 20
 xr ( N  1)  (11-15)
X  21
r r
      22
  23
 xr (2) xr (3)  xr (1) 
24
25
Define the matrix X1 to be the first T_Np rows of X concatenated with a row vector of 1's 26
of length N. The M-by-(T_Np + 1) coefficient matrix, P, corresponding to the linear fit is 27
then defined by (11-16). 28
29
30
P  YX 1T ( X 1 X 1T ) 1 (11-16) 31
32
33
34
In (11-16) the superscript "T" denotes the matrix transpose operator. 35
36
 37
 e (1) e( M  1)  e( M (N 1)  1)  
 e(2) e(M  2) 38
 e(M ( N  1)  2) 39
E  PX 1  Y   (11-17)
      40
  41
e (M ) e(2 M )  e (MN )  42
43
44
The error waveform, e(k), is then read column-wise from the elements of E. 45
46
47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 245
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Define P1 to be a matrix consisting of the first T_Np columns of the matrix P as shown
2 in (11-18).
3
4 
5  p (1 ) p ( M  1 )  p ( M ( T _ N p  1)  1)  
6  p (2 ) p ( M  2 )  p( M (T _ N  1)  2 )
P1   
p
7 (11-18)
     
8  
9  p( M ) p (2 M )  p ( MT _ N p ) 
10
11
12
13 The linear fit pulse response, p(k), is then read column-wise from the elements of P1.
14
15 11.3.1.6.5 Removal of the transfer function between the transmit function and T
16
17 Rotate sampled pulse response pi by the specified equalizer delay T_Dw to yield pr as
18 shown in (11-19).

19
20 p r  pi( T _ D w  1) p i (T _ D w  2)  pi (T _ N p ) p i (1)  pi (T _Dw ) 
21 (11-19)
22
23 Define the matrix P2 to be a T_Np-by-T_Np matrix derived from pr as shown in (11-20).
24
25 
 pr (1) pr (T _ N p )  pr (2) 
26  p (2) 
27 pr (1)  pr (3) 
P2  
r
(11-20)
28      
29  
30  pr (T _ N p ) pr (T _ N p  1)  pr (1) 
31 
32 Define the matrix P to be the first T_N rows of P . Define a unit pulse column vector
3 w 2
33 x of length T_N . The value of element xp(T_Dp + 1) is 1 and all other elements have a
p p
34 value of 0. The vector of filter coefficients w that equalizes p is then defined by (11-21).
i
35
36 
37 w  (P3T P3 ) 1 P3T x p
(11-21)
38
39 Given the column vector of equalizer coefficients, w, the equalized pulse response qi is
40 determined by (11-22).
41
42 qi  P3 w 
43 (11-22)
44
45
46
47
48
49

246 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

11.3.2 Receiver Characteristics 1

2
A compliant receiver shall operate at the specified BER with the worst case 3
combination of a compliant transmitter and a compliant channel. 4
5
Receiver electrical specifications are given in Table 11-12 and measured at compliance 6
point R. To dampen noise sources and absorption of both even and odd mode 7
reflections, the receiver shall satisfy the Common Mode Input Return Loss requirement 8
of Table 11-12. Jitter specifications at reference R are listed in Table 11-13. 9
10
Table 11-12. Receiver Electrical Input Specifications 11
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
12
13
Baud rate R_Baud 19.90 25.80 GSym/s
14
Input Differential Voltage R_Vdiff Note 1 1200 mVppd 15
Differential Input Impedance R_Rdin 80 100 120  16
Input Impedance Mismatch R_Rm 10 % 17
Differential Input Return Loss R_SDD11 See 11.3.2.3 18
Below 10 GHz -6
19
Common Mode Input Return Loss R_SCC11 dB 20
10GHz to baud rate -4
21
Input Common Mode Voltage R_Vcm
Load Type 0 
See Note 2
-200 1800 mV 22
23
NOTES:
1. The receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the combined 24
transmitter and channel. The channel response shall include the worst case effects of the return losses at the transmitter and 25
receiver.
2. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load. For floating load, input resistance shall be 1k
26
27
28
29
Table 11-13. Receiver Input Jitter Specification 30
31
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
32
Sinusoidal Jitter, Maximum R_SJ-max See Section 5 UIpp 33
2.4.4, note 1
34
Sinusoidal Jitter, High Frequency R_SJ-hf
See Section
2.4.4, note 1 0.05 UIpp 35
36
NOTES:
1. The Receiver shall tolerate the sum of these jitter contributions: Total transmitter jitter from Table 11-7;Sinusoidal jitter as 37
defined in Table 11-13;The effects of a channel compliant to the Channel Characteristics (Section 11.2.6). 38
39
11.3.2.1 Input Baud Rate 40
41
All devices shall work within the range from 19.90 Gsym/s to 25.80 Gsym/s as specified 42
for the device, with the baud rate tolerance as per Section 3.2.11. 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 11: CEI-25G-LR Short Reach Interface 247
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 11.3.2.2 Reference Input Signals

2
3 The receiver shall accept differential input signal amplitudes produced by a compliant
4 transmitter connected with the minimum attenuation specified in Figure 11-2 to the
5 receiver. This may be larger than the 1200 mVppd maximum of the transmitter due to
6 output/input impedances and reflections.
7
8 The minimum input amplitude is defined by the minimum transmitter amplitude, the
9 actual receiver input impedance and the loss of the actual PCB. Note that the minimum
10 transmitter amplitude is defined using a well controlled load impedance, however the
11 real receiver is not, which can leave the receiver input signal smaller than expected.
12 Additionally it will be determined by the environmental noise inside and outside the
13 receiver.
14
15 11.3.2.3 Input Resistance and Return Loss
16
17 Please refer to Section 3.2.10 with the following parameters.
18
19 Table 11-14. Receiver Differential Return Loss Parameters
20 Parameter Value Units
21
A0 -12 dB
22
23 f0 50 MHz

24 f1
0.1714 x
Hz
R_Baud
25
26 f2 R_Baud Hz

27 Slope 12.0 dB/dec

28
29 11.3.2.4 Input Lane-to-Lane Skew
30
31 Please refer to Section 3.2.8.
32
33
34 11.3.2.5 Absolute Input Voltage
35
36 The absolute voltage levels with respect to the receiver ground at the input of the
37 receiver are dependent on the transmitter implementation and the inter-ground
38 difference. The voltage levels at the input of an AC coupled receiver (if the effective AC
39 coupling is done within the receiver) or at the TX side of the external AC coupling cap (if
40 AC coupling is done externally) shall be between -0.3 to 2.0V with respect to local
41 ground.
42
43
44
45
46
47
48
49

248 Clause 11: CEI-25G-LR Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

12 Test Methodologies for CEI Baud Rates above 11G 1

2
3
This test methodology clause defines some common requirements needed for the 4
following variants: CEI-28G-SR, CEI-28G-MR, CEI-25G-LR, CEI-56G-USR, CEI-56G- 5
XSR-NRZ, and CEI-56G-XSR-PAM4.
6
7
8
12.1 TX jitter measurement methodology 9
10
• TX jitter measurements are performed using the Short Stress Pattern Random 11
(SSPR) defined in Appendix 2.D.2 of the "Implementation Guide for the Common 12
Electrical Interface 2.0", except for DDJ, which is measured using PRBS9. 13
• Unless otherwise specified, TX jitter parameters defined in Table 10-7 and 11-7 are 14
measured with TX FIR equalization turned-off and on. 15
16
• Jitter distributions are defined in 2.C.4, and are the basis for determining the jitter 17
parameters 18
• Jitter distributions are measured with any jitter measurement capable instrument 19
(e.g., scope, BERT) referenced to a golden PLL recovery clock timing with its corner 20
frequency set at baud rate/1667. 21
22
• T_UUGJ, T_UBHPJ, and T_TJ are derived with the method defined in 2.C.4.6 from 23
the BER CDF. T_UBHPJ is calculated as HPJtotal - DDJ. 24
• T_DCD is defined in Section 1.6, Table 1-3 25
26
• T_EOJ is defined in Section 1.6, Table 1-3 27
• The DDJ difference with TX FIR on and off is defined as:  28
diff_DDJ = T_DDJ (FIR on) -T_DDJ (FIR off) 29
30
• T_UUGJ, T_UBHPJ, T_TJ, T_DCD, and T_DDJ need to be measured with TX FIR 31
on and off 32
• diff_DDJ should be subtracted from the T_TJ measured when the FIR is on 33
34
• T_UUGJ, T_UBHPJ, T_TJ, and T_DCD measured with FIR on and off should be 35
within the limits as defined in Table 10-7 and 11-7 36
37
• The measurement instrument bandwidth should be at least 40 GHz. If the 38
measurement bandwidth affects the result, it can be corrected using post- 39
processing 40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR 249
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 12.1.1 Data Dependent Jitter (DDJ) measurement

2
3 A high-resolution oscilloscope, time interval analyzer, or other instrument with
4 equivalent capability may be used to measure DDJ. Establish a crossing level equal to
5 the average value of the entire waveform being measured.
6
7 Synchronize the instrument to the pattern repetition frequency and average the
8 waveforms or the crossing times sufficiently to remove the effects of random jitter and
9 noise in the system. The mean time of each crossing is then compared to the expected
10 time of the crossing, and a set of timing variations is determined. DDJ is the range
11 (max-min) of the timing variations. Keep track of the signs (early/late) of the variations.
12 Note, it may be convenient to align the expected time of one of the crossings with the
13 measured mean crossing. All edges of the repeating pattern that have been averaged
14 need to be included in the measurement.
15
16 The following Figure 12-1 illustrates the method. The vertical axis is in arbitrary units,
17 and the horizontal axis is plotted in UI. The waveform is AC coupled to an average
18 value of 0, therefore 0 is the appropriate crossing level. The rectangular waveform
19 shows the expected crossing times, and the other is the waveform with jitter that is
20 being measured. Only 16 UI are shown in this example. The waveforms have been
21 arbitrarily aligned with (t2 = 0) at 5 UI.
22 Figure 12-1. DDJ Measurement Method
23
24
1
25
26
27
28
29 0
30
31
32
33
34 -1

35
36 0 2 4 6 8 10 12 14 16

37
38 t1 t2 t3 t4 t5 t6 t7 t8 t9 . . .
39 Set to 0

40
41 DDJ = max(t1, t2,…. tn) – min(t1, t2,…. tn)
42
43
44
45
46
47
48
49

250 Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

12.2 Channel compliance methodology 1

2
3
12.2.1 Channel Compliance 4
A forward channel and associated dominant crosstalk channels are deemed compliant 5
if the channel characteristics conform to the requirements in the relevant clause, using 6
the methodologies described in this section. 7
8
9
12.2.1.1 Fitted insertion loss 10
11
The weighted fitted insertion loss ILfitted as a function of frequency f is defined by the 12
equation below. 13
14
f f 2 15
IL fitted  f  = a 0 + a 1 ---f- + a 2 ---- + a 4  ----   dB  (12-1) 16
fb fb  fb 
17
 18
Where fb is the maximum symbol rate to be supported by the channel under test. 19
Given the channel insertion loss measurement at N uniformly-spaced frequencies fn 20
spanning fILmin to fILmax with a maximum frequency spacing of 10MHz. The coefficients 21
of the fitted insertion loss are computed as follows. 22
23
Note: fILmin, fILmax are defined in Table 10-4/ 11-4. 24
25
Define the weighted frequency matrix F as shown below, where "mag(ILf)" is the 26
magnitude of the measured insertion loss at each frequency point [mag(ILfx) = 10^(- 27
ILfx/20)]. Note: mag(ILf) is a real number between 0 and 1. 28
29
 30
f f1 f1 2 31
mag  IL f1  mag  IL f 1   ---1- mag  IL f 1   ---- mag  IL f1    ----  32
fb fb  fb 
33
f2 f2 f2 2 (12-2) 34
F = mag  IL f2  mag  IL f 2   ---
- mag  IL f 2   ---- mag  IL f2    ----  35
fb fb  fb 
36
    37
 38
f fN fN 2
mag  IL fN  mag  IL f N   ----N mag  IL f N   ---- mag  IL fN    ----  39
fb fb  fb  40
 41
42
43
The polynomial coefficients a0, a1, a2, and a4 are determined using the Equation below. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR 251
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 
2
3
4 a0
5
6 a1 T –1 T
7 =  F F  F mag  IL f   I L f (12-3)
8 a2
9 a4
10
11
12
13
14
Where T denotes the matrix transpose operator and ILf is a column vector of the
15
measured insertion loss values, in dB, at each frequency point.
16
17 This polynomial fit process is expected to yield values for the coefficients a0, a1, a2, and
18 a4 that are greater than the minimum and less than the maximum coefficients (as
19 specified in the specific clauses). If any of the coefficients in the equation are below the
20 minimum allowed value they are forced to the minimum value and the fitting process is
21 iterated (see example below). Iteration is done by creating a newIL by subtracting all
22 coefficients below the minimum allowed value from the original IL, removing those
23 coefficients from F and recalculating the remaining coefficients. At the end of the
24 iteration, limit all coefficients to the maximum allowed, followed by a final iteration on
25 any coefficients not previously limited.
26
27 For each iteration only one additional coefficient should be forced to a value. If multiple
28 coefficients are below the minimum or above the maximum then the coefficients should
29 be forced to a value in the following order - a4 followed by a1 followed by a2 and last
30 a0.
31
32 Example iteration: If a2 needs to be set to zero, but all other coefficients are within the
33 range, then calculate newIL and solve for a0, a1 & a4 as indicated below.
34
35 f
36 newIL = IL – a 2fixed  ---- 
fb
37 (12-4)
38
39 Define the frequency matrix F as shown below
40
41
42
43
44
45
46
47
48
49

252 Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
f f1 2 2
mag  IL f1  mag  IL f1   ---1- mag  IL f1    ----  3
fb  fb 
4
f2 f2 2 5
mag  IL  mag  IL   ---
- mag  IL f2    ----  (12-5) 6
F = f 2 f 2 fb  fb 
7
   8
f fN 2 9
mag  IL f N  mag  IL fN   ----N 
mag  IL fN    ----  10
fb  fb 11
 12
The polynomial coefficient a0, a1 & a4 are determined using the Equation below. 13
14
 15
a0 16
T –1 T (12-6) 17
a1 =  F F  F mag  IL f   I L f
 18
a4 19
20
Where T denotes the matrix transpose operator and ILf is a column vector of the 21
measured insertion loss values, in dB, at each frequency point. 22
23
If after this iteration, a1 is below minimum allowed value, then another newIL is 24
calculated as indicated below. 25
26
f 27
newIL = IL – a 1fixed  ---f- + a2  ----  28
fb fixed fb 29
(12-7)
30
Define the frequency matrix F as shown below 31
32
f1 2 33
mag  IL f1  mag  IL f1    ----  34
fb 35
f2 2 36
mag  IL f2  mag  IL f2    ----  37
F = fb  38
(12-8) 39
 
40
fN 2 41
mag  IL fN  mag  IL fN    ----  42
 fb  
 43
The polynomial coefficient a0 & a4 are determined using the Equation below. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR 253
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 a 0 =  F T F  –1 F T 
3 mag  IL f   I L f (12-9)
4 a 4
5
6
7 Where T denotes the matrix transpose operator and ILf is a column vector of the
8 measured insertion loss values, in dB, at each frequency point.
9
10 If after this iteration all values are within range, the calculation is finished.
11
12 12.2.1.2 Integrated crosstalk noise
13
14 Given multi-disturber near-end crosstalk loss MDNEXTloss and multi-disturber far-end
15 crosstalk loss MDFEXTloss measured over N frequencies fx spanning 0.05 GHz to fb
16 (where fb is the maximum baud rate supported by the channel), with uniform frequency
17 step f, the RMS value of the integrated crosstalk noise x shall be calculated as
18 follows.
19
20 MDNEXTloss is determined from all individual pair-to-pair differential NEXT loss values
21 using Equation (12-10).
22
23 
24
25  all NEXTs 
26  –  NLi ( f )   10
27
MDNEXT loss( f ) = – 10  log
10  10 
 dB 
(12-10)
 i=0 
28
29
30

31
for 0.05 GHz  f  fb
32
33 where
34 MDNEXTloss(f) is the MDNEXT loss at frequency f,
35 NLi(f) is the NEXT loss at frequency f of pair combination i, in dB,
36 f is the frequency in GHz,
37 i is all pair-to-pair combinations.
38
39 MDFEXT
loss is determined from all individual pair-to-pair differential FEXT loss values
40 using Equation (12-11).
41
42 
43
44  all F EXTs 
45  –  NLi ( f )   10
46
MDFEXT loss( f ) = – 10  log
10  10 
 dB 
(12-11)
47  i=0 
48
49

254 Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

 1
 2
for 0.05 GHz  f  fb 3
4
where 5
MDFEXTloss(f) is the MDFEXT loss at frequency f, 6
NLi(f) is the FEXT loss at frequency f of pair combination i, in dB, 7
f is the frequency in GHz, 8
i is all pair-to-pair combinations. 9
10
Define the weight at each frequency fn using Equation (12-12) and Equation (12-13). 11
12
13
 1  1  (12-12)
Wn t ( f )  ( An2t 4 f b ) sinc 2 ( f f b )  4  8 
14
1  ( f f nt )  1  ( f f r )  15
 16
 1  1  17
W ft ( f )  ( A2ft 4 f b ) sinc 2 ( f f b ) 4  8
(12-13) 18
1  ( f f ft )  1  ( f f r )  19

20
21
Note that -3 dB transmit filter bandwidths fnt and fft are inversely proportional to the 20 22
to 80% rise and fall times Tnt and Tft respectively. The constant of proportionality is 23
0.2365 (e.g. Tnt fnt = 0.2365), where Tnt is in nano seconds and fnt is in GHz. In 24
addition, fr is the -3 dB reference receiver bandwidth, which should be set at 3/4 the 25
maximum baud rate specified for the device. 26
27
The near-end integrated crosstalk noise nx is calculated using Equation (12-14). 28
29
30
1/ 2 (12-14) 31
 
 nx   2f  Wnt ( f n )10 MDNEXT loss ( fn ) 10
 32
 n  33
34
35
36
The far-end integrated crosstalk noise fx is calculated using Equation (12-15). 37
38
 39
 
1/ 2
40
 fx   2f W ft ( f n )10 MDFEXT loss ( fn ) 10
 (12-15) 41
 n  42
 43
44
45
The total integrated crosstalk noise x is calculated using Equation (12-16). 46
47
48
49

Optical Internetworking Forum - Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR 255
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 (12-16)
3  x   nx2   2fx
4
5
6
7
8
9 12.3 Common Mode Noise
10
11 Common mode noise specification is to be measured using the following test
12 procedure.
13
14 The data pattern is normal traffic or a common test pattern. Connect both waveform
15 polarities through a suitable test fixture to a 50 ohm communication analysis
16 oscilloscope system. Waveforms are not triggered (free-run mode). Scope shall have a
17 minimum bandwidth (including probes) of 1.8 times the signaling rate.
18
19 No filtering except AC coupling with a high-pass 3dB low frequency not greater than
20 10MHz.
21
22 The two inputs are summed for common mode analysis. Set the horizontal scale for full
23 width to span one UI. Set up a vertical histogram with full display width. Measure the
24 rms value of the histogram. Common mode rms value (Ncm) is half the rms value of the
25 histogram.
26
27 Follow equation (12-17) below to account for instrumentation noise.
28
29
30 T _ NcmorR _ Ncm  measured _ Ncm  instrumentation _ noise
2 2 (12-17)
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

256 Clause 12: Test Methodologies for CEI-28G-SR and CEI-25G-LR - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

13 CEI-28G-VSR Very Short Reach Interface 1

2
3
This clause details the requirements for the CEI-28G-VSR very short reach high speed 4
chip-to-module electrical interface of nominal baud rates of 19.6 Gsym/s to 5
28.1 Gsym/s. A compliant host or module shall meet all of the relevant requirements
6
listed below. The electrical interface is based on high speed, low voltage logic, and 7
connections are point-to-point balanced differential pairs. 8
This clause defines the characteristics required to communicate between 9
CEI-28G-VSR drivers and CEI-28G-VSR receivers using copper signal traces on a 10
printed circuit board, a mated connector pair and copper signal traces inside an optical 11
module. These specifications are normative at the test points shown in Figure 13-1. A 12
‘length’ is effectively defined in terms of its attenuation and phase response rather than 13
its physical length. 14
15
Hosts and modules compliant to CEI-28G-VSR from different manufacturers shall be 16
interoperable. 17
18
19
13.1 Requirements 20
21
The objectives and requirements for the CEI-28G-VSR implementation agreement are 22
given by the project definition as follows: 23
24
•Support serial baud rates (fb) within the range from 19.6 Gsym/s to 28.1 Gsym/s as 25
specified for the device using NRZ coding. Note that implementation of specific 26
protocols will define the operating baud rate without affecting CEI compliance. 27
28
•Capable of driving up to a minimum of 100 mm of host PCB trace plus one connector 29
and a minimum of 50 mm of module PCB trace 30
•Capable of achieving Bit Error Ratio of 10-15 or better per lane 31
32
•Shall support AC-coupled operation. 33
•Shall allow multi-lanes (1 to n). 34
35
•Shall support hot plug. 36
•The IA will document the constraints of the chip-to-module application(s) used to 37
derive the channel model specifications 38
39
•The IA shall define a compliance test methodology including compliance boards. 40
41
42
13.2 General CEI Requirements 43
44
13.2.1 Data Patterns 45
46
See 3.2.1. 47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 257
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 13.2.2 Signal levels

2
3 The CEI-28G-VSR interface uses low swing differential signaling. It is designed to
4 operate with load type 0 from Section 3.2.2 (no other load types are supported).
5
6 This type of differential interface allows for interoperability between components
7 operating from different supply voltages and different I/O types (CML, LVDS-like,
8 PECL, etc.). Low swing differential signaling provides noise immunity and improved
9 electromagnetic interference (EMI). Differential signal swings are defined in later
10 sections and depend on several factors: such as transmitter pre-equalization, receiver
11 equalization and transmission line losses.
12
13 13.2.3 Signal Definitions
14
15 Each signal path, or CEI lane, is a point-to-point connection made up of two
16 complementary signals making a balanced differential pair. This specification allows for
17 bi-directional applications with multiple lanes in each direction.
18
19
20 13.2.4 Bit Error Ratio
21 See 3.2.3.
22
23
24 13.2.5 Ground Differences
25
26 The maximum ground difference between the host and module shall be ±50 mV. This
27 will affect the absolute maximum voltages at the compliance points.
28
29
30 13.3 Electrical Characteristics
31
32 Hosts and modules shall meet the appropriate specifications defined in Table 13-1,
33 Table 13-2, Table 13-4, and Table 13-5. Note that the direction of a given lane (host-to-
34 module or module-to-host) will determine which of the listed tables give applicable
35 specifications.
36
37 13.3.1 Compliance Point Specifications
38
39 Figure 13-1 below gives the reference model and test points associated with host-to-
40 module and module-to-host lanes.
41
42 Reference test fixtures, called compliance boards, are used to access the electrical
43 specification parameters. The output of the Host Compliance Board (HCB) provides
44 access to the host-to-module electrical signal (host electrical output) defined at TP1a.
45 Additional module electrical input specifications, for host-to-module communication,
46 are defined at TP1, the input of the Module Compliance Board (MCB). The output of the
47 Module Compliance Board (MCB) provides access to the module to host electrical
48 signal (module electrical output) defined at TP4. Additional host electrical input
49

258 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

specifications, for module-to-host communication, are defined at TP4a, the input of the 1
Host Compliance Board (HCB). Informative specifications for the host transmit function 2
(TP0a) are given in Appendix 13.B. 3
4
Figure 13-1. Measurement points using compliance boards
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 259
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 13.3.2 Host-to-Module Electrical Specifications

2
3 Each host-to-module lane shall meet the specifications of Table 13-1 and Table 13-2.
4 Definitions and methodologies can be found in Sections 13.3.4 to 13.3.11. The host
5 shall provide a recommended CTLE peaking value selected from Table 13-8 such that
6 the requirements defined in Section 13.3.11.1 are met. The method of providing this is
7 outside the scope of this document.
8
Table 13-1. Host-to-Module Electrical Specifications at TP1a (host output)
9
10
Parameter Min. Max. Units Conditions
11
12 Differential Voltage pk-pk - 900 mV
13 Common Mode Noise RMS - 17.5 mV See Section 13.3.5
14
15 Differential Termination Resistance At 1 MHz See Section
- 10 %
Mismatch 13.3.6
16
17 Differential Return Loss See Equation
- dB
18 (SDD22) (13-2)
19 Common Mode to Differential See Equation
20 Conversion (SDC22) - dB
(13-4)
21 Common Mode Return Loss (SCC22) - -2 dB From 250 MHz to 30 GHz
22
23 Transition Time, 20 to 80% 10 - ps See Section 13.3.10
24 Common Mode Voltage -0.3 2.8 V Referred to host ground
25 Eye Width at 10-15 probability (EW15)
26 0.46 - UI See Section 13.3.11
Note 1
27 Limit 1 95 - mV
28 Eye Height at 10-15
See Section 13.3.11
probability (EH15) Note 1 Limit 2 80 - mV
29
30 NOTES:
1. Open eye is generated through the use of a reference Continuous Time Linear Equalizer (CTLE)
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

260 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 13-2. Host-to-Module Electrical Specifications (module input) 1

2
Parameter Test Point Min. Max. Units Conditions 3
4
Overload Differential Voltage See Section
pk-pk
TP1a 900 - mV
13.3.12
5
6
Common Mode Voltage (Vcm) 7
TP1 -350 2850 mV
Note 1 8
9
Differential Termination At 1 MHz See
Resistance Mismatch
TP1 - 10 %
Section 13.3.6 10
11
Differential Return Loss
TP1 -
See Equation
dB
12
(SDD11) (13-2) 13
Differential to Common Mode See Equation
14
Conversion (SCD11)
TP1 -
(13-3)
dB 15
16
See
See Section 17
Stressed Input Test TP1a Section - 18
13.3.12
13.3.11.2.1
19
NOTES: 20
1. Vcm is generated by the host. Specification includes effects of ground offset voltage.
21
Table 13-3. Crosstalk parameters for host output test and 22
module stressed input test calibration at TP4 23
24
Parameter Target value Units 25
26
Crosstalk Amplitude Differential Voltage pk-pk 900 mV 27
Crosstalk Transition Time, 20 to 80% 9.5 ps 28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 261
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 13.3.3 Module-to-Host Electrical Specifications

2
3 Table 13-4. Module-to-Host Electrical Specifications at TP4 (module output)
4
5 Parameter Min. Max. Units Conditions
6 Differential Voltage, pk-pk - 900 mV
7
Common Mode Voltage (Vcm)
8 Note 1
-350 2850 mV
9
10 Common Mode Noise, RMS - 17.5 mV See Section 13.3.5
11 Differential Termination
- 10 % At 1 MHz
12 Resistance Mismatch
13 Differential Return Loss
14 - See Equation (13-2) dB
(SDD22)
15
Common Mode to Differential
16 Conversion (SDC22)
- See Equation (13-4) dB
17
18 Common Mode Return Loss From 250 MHz to
- -2 dB
(SCC22) 30 GHz
19
20 Transition Time, 20 to 80% 9.5 - ps See Section 13.3.10
21 Vertical Eye Closure (VEC) - 5.5 dB See Section 13.3.11.1.1
22 Eye Width at 10-15 probability
23 (EW15)
0.57 - UI See Section 13.3.11
24
Eye Height at 10-15 probability
25 (EH15)
228 - mV See Section 13.3.11
26
NOTES:
27 1. Vcm is generated by the host. Specification includes effects of ground offset voltage.
28
29 Table 13-5. Module-to-Host Electrical Specifications (host input)
30
31 Parameter Test Point Min. Max. Units Conditions
32
33 Overload
pk
Differential Voltage pk-
TP4 900 - mV
See Section
13.3.12
34
35 Differential Termination TP4a - 10 %
36 Resistance Mismatch
37 Differential Return Loss TP4a -
See Equation
dB
38 (SDD11) (13-2)
39 Differential to Common Mode See Equation
40 Conversion (SCD11) TP4a - dB
(13-3)
41 See Section See Section
42 Stressed Input Test TP4
13.3.11.2.1
-
13.3.12
43
44 Common Mode Voltage Note 1 TP4a -0.3 2.8 V
45 NOTES:
1. Referred to host ground. Common mode voltage is generated by host
46
47
48
49

262 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
Table 13-6. Crosstalk parameters for module output test and 3
host stressed input test calibration at TP1a
4
5
Parameter Target value Units
6
Crosstalk Amplitude differential voltage pk-pk 900 mV 7
Crosstalk transition time 20% to 80% 10 ps 8
9
10
11
12
13.3.4 Output Differential Voltage, pk-pk 13
14
The differential voltage, pk-pk, (see Section 1.6.1 for definition of differential voltage pk- 15
pk), including any transmit de-emphasis, shall meet the specifications given in Table 16
13-1 or Table 13-4 for the respective communication direction. DC referenced values 17
are not defined for the module because AC coupling is required in the module for both 18
Tx and Rx. The waveform is observed through a fourth-order Bessel-Thomson 19
response with a bandwidth of 40 GHz using a PRBS31 pattern. 20
21
13.3.5 Common Mode Noise 22
23
24
See Section 12.3 with the exception that the minimum oscilloscope BW shall be 40 25
GHz. 26
27
13.3.6 Differential Termination Resistance Mismatch 28
29
Differential Termination Resistance Mismatch is the percentage difference in low 30
frequency termination resistance with respect to ground of any two signals forming a 31
differential pair. This parameter is used to specify the difference between the two 32
resistances more tightly than each individual resistance for the purpose of minimizing 33
common mode to differential mode conversion. 34
35
Differential Termination Resistance Mismatch may be measured by applying a low- 36
frequency test signal (high enough to overcome the high-pass effects of the AC 37
coupling capacitors) to both the positive, Ip, and negative, In, terminals. The measured 38
differential impedance, Zdiff, and currents going into both (the positive, Ip, and negative, 39
In) terminals of the input are used to calculate the Differential Termination Resistance 40
Mismatch using Equation (13-1) below. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 263
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 13-2. Host Differential Termination Resistance Mismatch measurement setup
3
4
5 Test  I
p Capacitor Connector
6
frequency
Z p
7 50 ohms Z s
8
9 50 ohms
10
I
n Capacitor Z n AC 
11
12 Ground
13 Ground
14
15
16 Figure 13-3. Module Differential Termination Resistance Mismatch measurement setup
17
18
19
Test  p I Connector Capacitor
20
21 frequency
Z p
22 50 ohms Z s
23 50 ohms
24
25 n I Capacitor Z n AC 
26
27 Ground
28
Ground
29
30
31
32  I p – I n  Z diff + 100
Z = 2  -  --------------------------  100%
---------------------- (13-1)
33 mismatch  Ip + In  Z diff
34
35
36
37 13.3.7 Differential Return Loss
38
39 When measured at the respective test point the differential return loss shall not exceed
40 the limits given in Equation (13-2) (illustrated in Figure 13-4 for fb=28 GHz).
41
42 The test points are TP1a for host output, TP4a for host input, TP1 for module input and
43 TP4 for module output.
44
45
SDD11, SDD22 < -11dB for 0.05 < f < f b  7
46
47 (13-2)
f
48 SDD11, SDD22 < -6.0+9.2*log 10  2 ---- dB for f b  7  f  f b
49 fb

264 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 13-4. SDD11, SDD22 for host output (TP1a), host input (TP4a), module input (TP1) and 1
module output (TP4) (for fb = 28 GHz) 2
3
4
5
0
6
7
8
-2 9
10
SDD11, SDD22 (dB)

-4 11
12
-6 13
14
-8 15
16
17
-10 18
19
-12 20
21
-14 22
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 23
Frequency (GHz) 24
25
26
27
28
29
30
31
32
33
13.3.8 Common to differential mode and differential to common mode 34
conversion 35
36
The common to differential mode and differential to common mode conversion 37
specifications are intended to limit the amount of unwanted signal energy that is 38
allowed to be generated due to conversion of common mode voltage to differential 39
mode voltage or vice versa. 40
41
When measured at the respective input test point, common to differential mode or 42
differential to common mode conversion shall not exceed the limits given in Equation 43
(13-3) (illustrated in Figure 13-5 for fb=28 GHz). 44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 265
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 .
2 Figure 13-5. SDC11 and SCD11 for module input (TP1) and host input (TP4a)
3 (for fb = 28 GHz)
4
5
6
7 
8
9
10 
11
6&'6'&G%

12
13 
14
15
16 
17
18
19 
20
21
22 
23               
)UHTXHQF\*+]
24
25
26
27
28
29
30 SDC11 SCD11 < -22+14*  f  f b dB for 0.05  f  f b  2
31
32 (13-3)
33 SDC11 SCD11 < -18+6*  f  f b dB for f b  2  f  f b
34
35
36 When measured at the respective output test point, common to differential mode or
37 differential to common mode conversion shall not exceed the limits given in Equation
38 (13-4) (illustrated in Figure 13-6 for fb=28 GHz).
39
40
41
42
43
44
45
46
47
48
49

266 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Figure 13-6. SDC22 and SCD22 for module output (TP4) and host output 2
(TP1a) (for fb = 28 GHz) 3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
SDC22 SCD22 < -25+20*  f  f b dB for 0.05  f  f b  2 30
(13-4) 31
SDC22 SCD22 < -18+6*  f  f b dB for f b  2  f  f b 32
33
34
13.3.9 Common Mode Return Loss 35
36
The common mode output return loss specification is intended to limit the amount of 37
common mode energy that can be reflected by the host and module outputs. This has 38
an effect on EMI radiation and differential mode signals generated via common mode to 39
differential mode conversion. The common mode to differential mode conversion 40
specification for the host and module outputs is more stringent than for the inputs to 41
take into account the lack of a common mode input return loss specification. 42
43
44
13.3.10 Transition Time 45
46
Rise and fall time define the limits on the Transition Time. These limits are intended to
47
bound crosstalk as well as near-end reflections due to channel return loss. 48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 267
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Transition times (rise and fall times) are defined as the time between the 20% and 80%
2 times, or 80% and 20% times, respectively, of isolated edges.
3
4 If the test pattern is the square wave with eight ones and eight zeros, the 0% level and
5 the 100% level are the average values of the center 20% of the two time intervals of the
6 square wave.
7
9 5
8 If the test pattern is PRBS9 the pattern is generated by the polynomial x + x + 1 as
9 specified in ITU-T O.150. The binary (0,1) data sequence d(n) is given by:
10 d(n)=d(n - 9) + d(n – 5), modulo 2. 
11 The transitions within sequences of five zeros and four ones, and nine ones and five
12 zeros, respectively, are measured. These are bits 10 to 18 and 1 to 14, respectively,
13 where bits 1 to 9 are the run of nine ones. In this case, the 0% level and the 100% level
14 may be estimated by the average signal within windows from –3 UI to –2 UI and from 2
15 UI to 3 UI relative to the edge.
16
17 The waveform is observed through a fourth-order Bessel-Thomson response with a
18 bandwidth of 40 GHz.
19
20 NOTE—This definition is not the same as the rise and fall times typically reported by an
21 oscilloscope from an eye diagram, which take all the edges into account.
22
23 13.3.11 Eye Width, Eye Height and Stressed Input tests
24
25 Eye Width and Eye Height are specified in Table 13-1 (host output) and Table 13-4
26 (module output). Compliance is verified using the test setup shown in Figure 13-7
27 (host) and Figure 13-8 (module). The Eye Width and Eye Height correspond to eye
28 contours at a probability of 10-15 to be consistent with those generated by simulator and
29 oscilloscopes based on CDF/histogram data. Compliance to the input specifications
30 defined in Table 13-2 and Table 13-5 is verified using the test setup shown in Figure 13-
31 10 (host) and Figure 13-11 (module).
32
33 13.3.11.1 Host and Module output Eye Width and Eye Height test
34
35 The host output Eye Width and Eye Height are measured at TP1a of Figure 13-1 using
36 a Host Compliance Board as defined in Section 13.4.1. The test setup is shown in
37 Figure 13-7.
38
39
40
41
42
43
44
45
46
47
48
49

268 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Figure 13-7. Host output test setup 2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
The module output Eye Width and Eye Height is tested at TP4 of Figure 13-1 using a 25
Module Compliance Board as defined in Section 13.4.1. The test setup is shown in 26
Figure 13-8. 27
Figure 13-8. Module output test setup 28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 269
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 13.3.11.1.1Host and Module output test method

2
3 The signal at TP1a may be a closed eye. Therefore, a reference receiver with a
4 continuous time linear equalizer (CTLE) (see Section 13.3.11.3) is used to measure
5 Eye Width and Eye Height. Although the signal at TP4 is an open eye, the reference
6 receiver is also used to equalize the module output signal without the use of transmit
7 equalization. The measured signal after the reference receiver shall meet the
8 specifications listed in Section 13.3.2 for host to module and Section 13.3.3 for module
9 to host. All co-propagating and counter-propagating lanes are active as crosstalk
10 sources, using a PRBS31 test pattern or a valid CEI signal. Amplitude and Transition
11 Times for counter-propagating lanes are defined in Table 13-3 and Table 13-6. The
12 lanes under test are asynchronous to the lanes in the opposing direction within the
13 PPM offset defined by the protocol in use.
14
15 The test method for measuring either host or module output Eye Width and Eye Height
16 as illustrated in Figure 13-9 is as follows:
17 1) Set the host or module to PRBS9 pattern (see Section 13.3.10).
18
19 -This allows the use of a sampling oscilloscope with a pattern lock.
20 2) Capture the input signal at TP1a or TP4 with a scope triggered with a clock from a
21 reference clock recovery unit (CRU) with a first order transfer function with a 3 dB track-
22 ing bandwidth of fb/2578.
23
24 -For TP1a, the scope shall be AC coupled.
25 -The reference CRU can be a software CRU in case of a real time scope.
26
27 -Sample the signal with a minimum sampling rate of 3 (equally spaced) samples
28 per bit. Collect sufficient samples equivalent to 4 million bits in order to con-
struct normalized cumulative distribution functions (normalized CDFs) (see Fig-
29
30 ure 13-9) of the post-processed captured signals to a probability of 10-6 (without
31 extrapolation) as described below. Depending on the sampling rate, careful
interpolation using a method such as sin(x)/x or cubic spline may be needed for
32 good accuracy.
33
34 3) Apply the reference receiver as defined in Section 13.3.11.3 to equalize the captured
35 signal in step 2.
36
37 -For TP4 compliance test, the CTLE peaking in the reference receiver shall be
38 set at either 1 dB or 2 dB. Any CTLE setting that meets both the EH15 and
39 EW15 settings defined for TP4 in Table 13-4 is acceptable.
40
41 -For TP1a compliance test, the CTLE peaking in the reference receiver shall be
42 set at two or three values. These are: a) the recommended CTLE peaking value
43 provided by the host, b) the value 1 dB higher if present in Table 13-8 and c) the
44 value 1 dB lower if present in Table 13-8. The two or three results are used in
45 step 9 to determine compliance.
46
47 4) Use the differential equalized signal from step 3 to construct CDFs of the jitter at zero
48 crossing, for both the left edge (CDFL) and the right edge (CDFR) of the eye, as a
49 distance in time from the middle of the eye.

270 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

-The middle of the eye is defined to be UI/2 away from the mean zero crossing 1
points of the equalized signal from step 3. 2
3
Calculate the Eye Width (EW6, see Figure 13-9) as the difference in time between 4
CDFR and CDFL with a value of 10-6. CDFL and CDFR are calculated as the 5
cumulative sum of histograms of the zero crossing samples at the left and right edges 6
of the eye normalized by the total number of sampled bits (e.g., the number of sampled 7
bits is 4 million as specified in step 2). For a pattern with 50% transition density the 8
maximum value for the CDFL and CDFR will be 0.5. CDFL and CDFR are equivalent to 9
bathtub curves where the bit error ratio (BER) is plotted versus sampling time. 10
11
5) Apply Dual-Dirac and tail fitting technique (See Agilent white paper: 5989-3206EN) 12
separately to CDFL and CDFR to estimate random jitter. Calculate the best linear fit in 13
Q-scale over the range of probabilities of 10-4 to 10-6 of the CDFL and CDFR to yield 14
RJL and RJR respectively. 15
16
-RJL is the RMS value of the jitter estimated from CDFL. 17
18
-RJR is the RMS value of the jitter estimated from CDFR. 19
20
-Eye Width (EW15) at 10-15 probability is equal to (EW6-3.19*(RJL+RJR)). 21
22
6) Use the differential equalized signal from step 3 to construct the CDFs of the signal 23
voltage in the central 5% of the horizontal eye, for both logic one (CDF1) and logic zero 24
(CDF0), as a distance in voltage from the zero crossing. 25
Calculate the Eye Height (EH6, see Figure 13-9) as the difference in voltage between 26
CDF1 and CDF0 with a value of 10-6. CDF0 and CDF1 are calculated as the cumulative 27
sum of histograms of the voltage samples at the top and bottom of the eye normalized 28
by the total number of sampled bits (e.g., the number of sampled bits is 4 million as 29
specified in step 2). For a pattern with a well balanced number of ones and zeros the 30
maximum value for CDF0 and CDF1 will be 0.5. 31
32
7) Apply dual-Dirac and tail fitting techniques to CDF1 and CDF0 to estimate 33
noise in the central 5% of the eye. Calculate the best linear fit in Q-scale over the range 34
of probabilities of 10-4 to 10-6 of the CDF1 and CDF0 to yield RN1 and RN0 35
respectively. 36
37
-RN1 is the RMS value of the noise estimated above from CDF1. 38
39
-RN0 is the RMS value of the noise estimated above from CDF0. 40
41
-Eye Height (EH15) at 10-15 probability equals (EH6-3.19*(RN0+RN1)). 42
43
8) At TP4 calculate vertical eye closure (VEC) as 20*log10 (AV/EH15): 44
45
- AV is the Eye Amplitude of the equalized waveform. Eye Amplitude is defined 46
as the mean value of logic one minus the mean value of logic zero in the central 47
5% of the eye. 48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 271
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 9) At TP1a, passing is defined as passing both the EW15 and EH15 Limit 1 specified in
2 Table 13-1 for at least one of the equalizer settings, and passing EH15 Limit 2 specified
3 in Table 13-1 at all of the two or three settings. At TP4, passing is defined as a single
4 equalizer setting that meets the EH15, EW15 and VEC specifications given in Table 13-
5 4.
6
7
8 Figure 13-9. TP1a and TP4 jitter and Eye Height parameters
9
10
11
12
13
14 CDF1
15 UI/2 RN1
Zero 
16 crossing UI
AV EH6
17 Center of 
18 the eye RN0
19
CDF0
20
21
22
23
24 CDFL Middle of  CDFR
the eye
25 10‐6
26 probability 10‐4
27 probability
RJL RJR
28 ‐4
29 10 probability
30 EW6
31 10‐6 probability
32
33
34
35
36
37 13.3.11.2 Host and Module stressed input test
38
39 The ability of the host input to tolerate the Eye Width and Eye Height specified in Table
40 13-4 and the sinusoidal jitter specified in Table 13-7 is tested using a stressed input
41 test. The test signal is applied at TP4a of Figure 13-1, and calibrated at TP4, using a
42 Host Compliance Board and Module Compliance Board specified in Section 13.4.1 The
43 test setup is shown in Figure 13-10. The UBHPJ block is used to create non-
44 compensable DJ in addition to sinusoidal jitter.
45
46
47
48
49

272 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Figure 13-10. Host input test setup 2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The ability of the module input to tolerate the Eye Width and Eye Height specified in 31
Table 13-1 and the sinusoidal jitter specified in Table 13-7 is tested using a stressed 32
input test. The test signal is applied at TP1 of Figure 13-1, and calibrated at TP1a using 33
a Host Compliance Board and Module Compliance Board specified in Section 13.4.1. 34
The test setup is shown in Figure 13-11. The module stressed input test represents the 35
worst case high loss VSR channel. It should be noted that modules are also expected 36
to operate at the BER specified in Section 3.2.3 when presented with lower loss 37
channels that require different CTLE settings as long as the signal complies with the 38
specifications in Table 13-1 and the recommended CTLE peaking value supplied by the 39
host is within 1 dB of the optimal value for the signal (see Section 13.3.11.2.1). 40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 273
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 13-11. Module stressed input test setup
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31 13.3.11.2.1Host (TP4a) and Module (TP1) stressed input test method
32
33 The host and module input shall tolerate a peak-to-peak sinusoidal jitter with the
34 frequency and amplitude defined by the mask of Figure 13-12 and Table 13-7. This
35 sinusoidal jitter shall be part of the jitter applied in the stressed input test.
36
37 The reference CRU and reference receiver as defined in Section 13.3.11.3 are used to
38 calibrate the stressed input test signal at TP4 (per Table 13-4) or TP1a (per Table 13-1)
39 using a PRBS9 pattern. The pattern is changed to PRBS31 for the stressed input test.
40
41 The crosstalk source is asynchronous to the main pattern generator. The amplitude and
42 rise/fall time of the crosstalk source are given in Table 13-3 and Table 13-6. The
43 crosstalk signal is to be calibrated at TP4 or TP1a using a PRBS9 pattern, then
44 changing the pattern to PRBS31 for the test. For multi-lane implementations additional
45 lanes shall be active with an asynchronous PRBS31 pattern using the above calibration
46 methods.
47
48
49

274 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The host under test shall meet the BER specified in Section 3.2.3. The module under 1
test shall meet the BER specified in Section 3.2.3 when provided with each of three 2
recommended CTLE values. These are: a) the optimal value found in Section 3
13.3.11.2.1.2, b) the value 1 dB higher if present in Table 13-8 and c) the value 1 dB 4
lower if present in Table 13-8. (e.g. If the optimal value found in Section 13.3.11.2.1.2 is 5
9dB then the module must meet the specified BER in Section 3.2.3 when provided with 6
recommended CTLE values of 8,9 and 10 dB). 7
8
9
10
Table 13-7. Sinusoidal jitter frequency for TP4 and TP1 testing
11
Sinusoidal jitter, 12
Frequency Range (Hz) Peak to peak 13
(UI) 14
15
f<fb/257800 Not Specified 16
fb/257800 < f < fb/2578 5*fb/(257800*f) 17
18
fb/2578 < f < 10xLB (Note 1) 0.05 19
NOTES: 20
1. LB = Receiver Loop Bandwidth 21
22
Figure 13-12. Host input and Module input Sinusoidal Jitter
23
24
Sinusoidal 25
jitter
26
amplitude
27
28
5 UIpp 29
30
31
32
33
34
35
36
0.05 UIpp 37
38
39
40
fb/257800 fb/2578 10x Loop 41
Bandwidth
42
43
44
45
13.3.11.2.1.1Host input test signal calibration 46
47
The host input is tested at TP4a of Figure 13-1 using a Host Compliance Board as 48
defined in Section 13.4.1. The host input test setup is shown in Figure 13-10. 49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 275
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 UBHPJ, UUGJ and sinusoidal jitter are added to a clean test pattern until the jitter
2 (except for DCD) at the output of the pattern generator approximates the informative
3 transmit specification (as defined in Appendix 13.B).
4
5 With the crosstalk generator calibrated to meet the specifications in Table 13-3, the Eye
6 Height and Eye Width at TP4 are measured using the reference receiver defined in
7 Section 13.3.11.3 with the optimal peaking value from Figure 13-14 and the
8 methodology defined in Section 13.3.11.1. The optimal peaking value is defined as the
9 setting that results in the maximum value of EW15*EH15.
10
11 The UUGJ and pattern generator amplitude are adjusted to give the minimum Eye
12 Height and Eye Width specified for the module output in Table 13-4.
13
14 A host input test signal should have a VEC in the range of 4.5 to 5.5 dB with a target
15 value of 5.0 dB.
16
17 13.3.11.2.1.2Module input test signal calibration
18
19 The module input is tested at TP1 of Figure 13-1 using a Module Compliance Board as
20 defined in Section 13.4.1. The module input test setup is shown in Figure 13-11.
21
22 UBHPJ, UUGJ and sinusoidal jitter are added to a clean test pattern until the jitter
23 (except for DCD) at the output of the pattern generator approximates the informative
24 transmit specification (as defined in Appendix 13.B).
25
26 The frequency-dependent attenuator is intended to represent the host channel, and
27 may be implemented with PCB traces. It should be adjusted to result in a loss of 10.25
28 dB at Nyquist from the output of the pattern generator to TP1a. The crosstalk generator
29 is calibrated to meet the specifications in Table 13-3. The Eye Height and Eye Width at
30 TP1a are measured using the reference receiver (defined in Section 13.3.11.3) with the
31 optimal peaking value and the methodology defined in Section 13.3.11.1. The optimal
32 peaking value is defined as the setting that results in the maximum value of
33 EW15*EH15.
34
35 The UUGJ and pattern generator amplitude are adjusted to give the minimum Eye
36 Height and Eye Width specified in Table 13-1.
37
38 13.3.11.3 Reference receiver
39
40 The waveform is observed through a fourth-order Bessel-Thomson response with a
41 bandwidth of 40 GHz concatenated with a Continuous Time Linear Equalizer (CTLE).
42 The filters may be implemented in software; however, the signal is not averaged. The
43 CTLE shall be implemented based on Equation (13-5) where G is the gain and Z1, P1
44 and P2 are the CTLE zero and poles coefficients. Figure 13-13 shows the frequency
45 response of the reference equalizer used for host output testing for baud rates between
46 25 and 28.1 GBd with values for Z1, P1 and P2 listed in Table 13-8. Figure 13-14
47 shows the frequency response of the reference equalizer used for module output
48 testing for baud rates between 25 and 28.1 GBd with values for Z1, P1 and P2 listed in
49 Table 13-8. Note that the peaking is centered at 14 GHz for all baud rates between 25

276 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

and 28.1 GBd. For baud rates below 25 GBd the values of Z1, P1 and P2 should be 1
multiplied by fb/28. Note that this results in peaking at fb/2. Note that the peaking value 2
approximates the difference between the low frequency gain (1 MHz) and the high 3
frequency gain at Nyquist in dB. 4
(13-5) 5
 G   P1   P2   S + Z1  6
Hs = -------------------------------- ------------------------------------------
Z1  S + P1   S + P2  7
8
S = j2f 9
10
Figure 13-13. Host output Reference receiver equalizer (CTLE) transfer function for gains of 11
1 dB to 9 dB 12
13
14
1 15
16
0 17
18
1 dB 19
-1
20
2 dB 21
-2
22
3 dB 23
CTLE gain (dB)

-3
24
4 dB 25
-4
26
5 dB 27
-5
28
6 dB 29
-6
30
7 dB 31
-7
32
8 dB 33
-8
34
9 dB 35
-9
36
-10 37
0.1 1 10 40 38
Frequency(GHz)
Frequency (GHz) 39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 277
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 13-14. Module output Reference receiver equalizer (CTLE) transfer function for
3 gains of 1 and 2 dB
4
5
6 1
7
8 0
9 1 dB
10 -1
11 2 dB
12 -2
13
CTLE gain (dB)

14 -3
15
16 -4
17
18 -5
19
20 -6
21
22 -7
23
24 -8
25
26 -9
27
28
-10
0.1 1 10 40
29 Frequency (GHz)
(GHz)
30
31
32
33 Table 13-8. Reference equalizer coefficients for rate of 28 GBd.
34
35 Peaking (dB) G P1/ 2 (GHz) P2/ 2 (GHz) Z1/ 2 (GHz)
36 1 0.891 18.6 14.1 8.31
37
2 0.794 18.6 14.1 7.10
38
39 3 0.708 15.6 14.1 5.68
40 4 0.631 15.6 14.1 4.98
41
5 0.562 15.6 14.1 4.35
42
43 6 0.501 15.6 14.1 3.82
44 7 0.447 15.6 14.1 3.43
45 8 0.398 15.6 14.1 3.00
46
47 9 0.355 15.6 14.1 2.67
48
49

278 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

13.3.12 Input Differential Voltage Tolerance 1

2
The input voltage tolerance tests the acceptance of differential input pk-pk amplitudes 3
produced by the extremes of operation from the transmitter (e.g. host output for host-to- 4
module communication or module output for module-to-host communication). 5
6
The input voltage tolerance maximum value is produced by a compliant transmitter (per 7
Table 13-1) connected with the minimum attenuation to the receiver. This may be larger 8
than the maximum of the driver due to output/input impedances and reflections. 9
10
The input voltage tolerance value is defined by the minimum driver amplitude, the actual 11
receiver input impedance, and the loss of the actual PCB. Note that the minimum driver 12
amplitude is defined using a well controlled load impedance; however the real receiver 13
is not, which can leave the receiver input signal smaller than expected. Additionally it will 14
be determined by the environmental noise inside and outside the receiver.
15
16
13.4 Measurement methods 17
18
19
13.4.1 Compliance Boards 20
21
Use of compliance boards for testing is assumed for the parameters defined in Table 22
13-1 through Table 13-6. Figure 13-1 shows the test setup for making S parameter 23
measurements of the mated compliance boards. If compliance boards do not meet the 24
specified S parameters test results should be corrected for the difference. The 25
requirements in this section are not connector specifications for an implemented 26
design. 27
28
13.4.1.1 HCB and MCB insertion loss 29
30
The reference differential insertion loss of the HCB printed circuit board trace follows 31
Equation (13-6) for 50 MHz < f < 28.1 GHz. The reference differential insertion loss of 32
the MCB printed circuit board trace follows Equation (13-7) for 50 MHz < f < 28.1 GHz. 33
(f is measured in GHz) Both the HCB and MCB equations are illustrated in Figure 13- 34
15, below. 35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 279
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 HCB SDD21 = 2.00  0.001 – 0.096  f  – 0.046  f   dB (13-6)

2
3
4
5 MCB SDD21 =  1.25   0.001 – 0.096 f – 0.046  f  dB (13-7)
6
7
Figure 13-15. Reference SDD21 of HCB and MCB printed circuit board traces
8
9
10
11
12
13 0
14
15 -0.5
16
17
18 -1
19
20
SDD21 (dB)

-1.5
21
22
23 -2
24
25 -2.5
26
27
28
-3
29
30 -3.5
31
32 -4
33 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
34 Frequency (GHz)
35
36
37
38
39
40
41 13.4.1.2 Mated HCB and MCB S parameters
42
43 The specifications given for the mated HCB and MCB shall be verified in both
44 directions (exception being differential insertion loss can be in either direction).
45
46
47
48
49

280 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The differential return loss of the mated HCB and MCB pair shall follow Equation (13- 1
8), illustrated in Figure 13-16. 2
3
4
Mated HCB-MCB SDD11, SDD22 < -20 + f dB ___for f < 4 GHz (13-8) 5
6
Mated HCB-MCB SDD11, SDD22 = -18 + f/2 dB _.for 4 GHz < f < 28.1 GHz 7
8
9
Figure 13-16. Mated HCB-MCB SDD11, SDD22 10
11
12
13
14
0 15
16
-2 17
18
-4 19
SDD11, SDD22 (dB)

-6 20
21
-8 22
23
-10 24
25
-12 26
-14 27
28
-16 29
30
-18 31
32
-20 33
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 34
Frequency (GHz) 35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 281
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The differential to common mode conversion for a mated HCB and MCB pair is given in
2 Equation (13-9) and shown in Figure 13-17, below.
3
4
5 Figure 13-17. Mated HCB-MCB SCD21, SCD12
6
7
8
9
10 0
11
12
13 -5
14
SCD21, SCD12 (dB)

15 -10
16
17
18 -15
19
20 -20
21
22
23 -25
24
25 -30
26
27
28 -35
29 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
30 Frequency (GHz)
31
32
33
34
35
36
37
38 Mated HCB-MCB SCD21, SCD12  -35+1.07f dB for f < 14 GHz
39
40 (13-9)
41 Mated HCB-MCB SCD21, SCD12  -20 dB for 14 GHz < f < 28.1 GHz
42
43
44
45
46
47
48
49

282 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The differential to common mode return loss for a mated HCB and MCB pair is given in 1
Equation (13-10) and shown in Figure 13-18, below. 2
3
4
Figure 13-18. Mated HCB-MCB SCD11, SCD22, SDC11, SDC22 5
6
7
8
9
0 10
SCD11, SCD22, SDC11, SDC22 (dB)

11
12
-5 13
14
-10 15
16
17
-15 18
19
20
-20 21
22
23
-25 24
25
-30 26
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 27
Frequency (GHz) 28
29
30
HCB-MCB SCD11,SCD22 and SDC11,SDC22  -30+(5/7)f dB for f < 14 GHz (13-10) 31
32
HCB-MCB SCD11,SCD22 and SDC11,SDC22  -25+(5/14)f dB for 14 GHz < f < 28.1 GHz 33
34
35
36
The maximum common mode return loss for a mated HCB and MCB pair shall be 3dB. 37
38
The maximum differential insertion loss for a mated HCB and MCB pair is given in 39
Equation (13-11). The minimum differential insertion loss for a mated HCB and MCB is 40
given in Equation (13-12). Both equations are shown in Figure 13-19, below. 41
42
43
Mated HCB-MCB SDD21, SDD12> -0.12-0.475 f - 0.221*f dB for f<14 GHz (13-11) 44
45
Mated HCB-MCB SDD21, SDD12 >4.25-0.66*f dB for 14 GHz<f<28.1 GHz 46
47
48
Mated HCB-MCB SDD21, SDD12 <-0.08 f - 0.2*f dB for f<28.1 GHz (13-12) 49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 283
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 13-19. Mated HCB-MCB SDD21, SDD12
3
4
5
6
7 0
8 -1
9 -2
10 -3
11
SDD21, SDD12 (dB)

-4
12
-5
13
14 -6
15 -7
16 -8
17 -9
18 -10
19 -11
20 -12
21
-13
22
23 -14
24 -15
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
25
Frequency (GHz)
26
27
28
29
30
31 The FOMILD (as calculated using the method defined in Section 10.2.6.4 and the curve
32 fit method defined in Chapter 12 with fILmax of 21 GHz and fILmin of 50 MHz) for the
33 mated HCB and MCB pair is 0.1 dB.
34
35 The Integrated Crosstalk Noise (ICN) as calculated using the method defined in
36 Chapter 12 with the aggressor amplitudes and rise/fall times as listed in Table 13-3
37 shall be less than 3.9 mV. MDNEXT shall be less than 1.35 mV RMS. MDFEXT shall
38 be less than 3.6 mV RMS.
39
40
41
42
43
44
45
46
47
48
49

284 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

13.A Appendix - Recommended Electrical Channel 1

2
The channel consists of Host PCB trace, Module PCB trace, vias, AC coupling 3
capacitor and one connector, not in this order. The recommended PCB trace differential 4
impedance is 100 ± 10  . This full channel model is shown in Figure 13-20 below. 5
Note that in practice the channel is not measurable as appropriate test points are not 6
accessible. 7
8
Figure 13-20. CEI-28G-VSR full Channel Reference Model 9
10
Connector 11
Up to 1.2 dB 12
Host PCB up to 7.3 dB Module PCB + Cap up to 1.5 dB 13
14
Receiver
15
Transmitter
16
Host IC Module IC 17
Transmitter 18
Receiver
19
20
21
22
13.A.1 Insertion Loss 23
24
Host insertion loss and module insertion loss are recommended limits only. Achieving 25
these recommended limits does not signify compliance nor guarantee successful 26
communication between two devices. Equation (13-13) (illustrated in Figure 13-21) 27
represents the highest recommended insertion loss of the full channel. 28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 285
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 13-21. Recommended minimum SDD21 of the VSR channel (for fb = 28 GHz)
3
4
5
6
7 0
8
9
10 -5
11
12
13
SDD21 (dB)

14 -10
15
16
17 -15
18
19
20 -20
21
22
23
24 -25
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
25 Frequency (GHz)
26
27
28
29
30
31
32
f 2
33
H  f  = 0.3144 – 8.1
f
---
- – 2.38 ---
- – 13.56  ---f-
34 fb fb f  (13-13)
35 b
36
37
38
39 13.B Appendix - Informative Host Transmitter output Electrical
40 Characteristics
41
42 Informative host Tx output recommendations are defined in Table 13-9.
43
44
45
46
47
48
49

286 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

13.B.1 Host Transmitter output specification point 1

2
Figure 13-1 gives the reference model and test points associated with host-to-module 3
and module-to-host CEI-28G-VSR lanes. The informative host transmitter output 4
electrical characteristics are defined to be measured at TP0a. TP0a is defined to be 5
separated from TP0, the ball of the package performing the host-to-module transmit 6
function, by 1 dB of attenuation at 14 GHz. 7
8
13.B.1.1 Host-to-Module transmitter output Electrical Specifications 9
10
It is recommended that each host-to-module lane meet the limits of Table 13-9. 11
12
Note: A 2 tap FIR filter may be advantageous in meeting the TP1a requirements. 13
14
Table 13-9. Host-to-Module Electrical Specifications at TP0a 15
16
Parameter Symbol Min. Max. Units Conditions 17
PRBS31 pattern. 18
Differential Voltage, pk-pk T_Vdiff 600 - mV Emphasis off. 19
Note 1 20
21
Common Mode Voltage T_Vcm -300 2800 mV Note 2
22
Differential resistance T_Rd 80 120 ohms 23
Differential Termination 24
T_Rdm - 10 % at 1 MHz
Resistance Mismatch 25
See 26
10.3.1.3 27
Differential Return Loss T_SDD22 - dB
(CEI-28G- 28
SR) 29
Transition Time: 20 to 80% T_tr, T_tf 8 - ps Emphasis off. 30
Common Mode Noise, RMS T_Ncm - 12 mV See 12.3
31
32
Uncorrelated Unbounded 33
T_UUGJ - 0.15 UI
Gaussian jitter
34
Uncorrelated Bounded high
T_UBHPJ - 0.15 UI Note 4 35
probability jitter including DCD 36
Duty Cycle Distortion T_DCD - 0.035 UI 37
Total Jitter T_TJ - 0.28 UI Note 3 38
39
NOTES:
1. Max voltage is limited by specifications at TP1a. Minimum voltage can be lower for low loss channels. 40
2. Load type 0 with min. T_Vdiff, AC-Coupling or floating load. 41
3. T_TJ includes all of the jitter components measured without any transmit equalization. A 1 dB CTLE can be used to achieve
this specification. (See Section 13.3.11.3). For jitter test parameters see 12.1 except use a CRU tracking BW = fb/2578.
42
4. Measured with any value of transmitter equalization See Section 12.1. 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface 287
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

288 Optical Internetworking Forum - Clause 13: CEI-28G-VSR Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14 CEI-28G-MR Medium Reach Interface 1

2
3
This clause details the requirements for the CEI-28G-MR medium reach high speed 4
electrical interface between nominal baud rates of 19.90 Gsym/s and 28.1 Gsym/s 5
using NRZ coding. A compliant device shall meet all of the requirements listed below.
6
The electrical interface is based on high speed, low voltage logic. Connections are 7
point-to-point balanced differential pairs and signaling is unidirectional. 8
The electrical IA is based on loss and jitter budgets and defines the characteristics 9
required to communicate between a CEI-28G-MR transmitter and a CEI-28G-MR 10
receiver using copper signal traces on a printed circuit board. The characteristic 11
impedance of the signal traces is nominally 100  differential. A ‘length’ is effectively 12
defined in terms of its attenuation and phase response rather than its physical length. 13
Refer to Section 14.2.6 for transmission line guidelines to meet the channel 14
requirements. 15
16
Medium reach CEI-28G-MR devices from different manufacturers shall be 17
interoperable. 18
19
20
21
22
14.1 Requirements 23
24
1. Support serial baud rates within the range from 19.90 Gsym/s to 28.1 Gsym/s. 25
26
2. Capable of low bit error ratio (10-15, with a test requirement to verify 10-12). 27
3. Capable of driving up to 500 mm of PCB and up to 1 connector. 28
29
4. Shall support AC-coupled operation. 30
31
5. Shall allow multi-lanes (1 to n). 32
6. Shall support hot plug. 33
34
35
14.2 General Requirements 36
37
14.2.1 Data Patterns 38
39
Please refer to Section 3.2.1 40
41
42
14.2.2 Signal levels 43
44
Please refer to Section 3.2.2. All transmitter and receiver devices shall support “Load 45
Type 0”. Other load types are not supported by this clause. 46
47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 289
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 14.2.3 Signal Definitions

2
3 Please refer to Section 1.A
4
5 14.2.4 Bit Error Ratio
6
7 Please refer to Section 3.2.3
8
9
10 14.2.5 Ground Differences
11
12 Please refer to Section 3.2.4
13
14 14.2.6 Channel Compliance
15
16 A forward channel and associated dominant crosstalk channels are deemed compliant
17 if the channel characteristics conform to the requirements in this section.
18
19 14.2.6.1 Reference Model
20
21 The channel consists of PCB traces, vias, and 1 connector. The reference PCB trace
22 differential impedance is 100.
23
24 Figure 14-1 shows a diagram of test points on an example board.
25
26
27
28 Figure 14-1. CEI-28G-MR Reference Model
29
30
31
32 Test Point ‘R’ Test Point ‘T’
Mezzanine
33 component edge or Board to Board component edge
34 Connector
35
Daughtercard
36
37 AC coupling
38 Mothercard
39
40
41
42
43
44
45 Note: Test points differ from definitions in Section 1.8, as DC blocking capacitor, if
46 physically located outside of the package, is part of the channel.
47
48
49

290 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Measured at these test points, several channel characteristics are parametrized. Port 1
definitions as noted in Figure 14-1 allow proper measurement of the parameters in 2
Table 14-1 used for calculation of the channel parameters found in Table 14-2. 3
4
Table 14-1. Measured Channel Parameters 5
Symbol Description 6
7
IL(f) Differential insertion loss, -SDD21 magnitude (dB)
8
RL1(f) Differential input return loss, -SDD11 magnitude (dB) 9
RL2(f) Differential output return loss, -SDD22 magnitude (dB) 10
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB)
11
12
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 13
14
Table 14-2. Calculated Channel Parameters
15
Symbol Description 16
ILfitted(f) Fitted insertion loss (dB)
17
18
ILD(f) Insertion loss deviation (dB) 19
ICN(f) Integrated crosstalk noise (mVRMS) 20
FOMILD A figure of merit of the insertion loss deviation (dB) 21
22
23
14.2.6.2 Insertion Loss 24
25
Channel insertion losses, including PCB traces and connectors, shall comply with the 26
limits specified by equations (14-1), (14-2) and plotted in Figure 14-2. Note that the vari- 27
able fb is the maximum baud rate to be supported by the channel under test (19.90 GHz 28
fb 28.1 GHz). 29
30
Table 14-3. Channel Insertion Loss Frequency Range 31
Parameter Value Units 32
fmin 50 MHz
33
34
fmax fb GHz
35
36
37
38
fb  39
 f  28.1 f  28.1
- +0.698 ------------------  f min  f  ----  40
 1.083  + 2.436 -----------------
fb fb 2
 (14-1) 41
IL max =   42
 f  28.1 fb
– 17.851 + 2 694 ------------------  ----  f  f b  43
 fb 2 
  44
45
46
47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 291
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2  0 f min  f  1GHz 
 
3 1  (14-2)
4 IL min =  ---  f – 1   1GHz  f  175 GH z
5 3 
 5 5  175 GH z  f  f b 
6  
7
8
9
10 Note: f in equations (14-1) and (14-2) is in GHz.
11
12
13 Figure 14-2. CEI-28G-MR normative channel insertion loss at 28.1 Gsym/s
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

292 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14.2.6.3 Fitted insertion loss 1

2
For fitted insertion loss definitions, please refer to Section 12.2.1.1 3
4
The channel shall meet the insertion loss requirements defined in Table 14-4. Note that 5
the variable fb is the maximum baud rate to be supported by the channel under test. 6
7
 8
9
Table 14-4. Channel fitted insertion loss characteristics 10
Value 11
Parameter Units 12
Min. Max.
Minimum frequency, fILmin GHz 0.05 - 13
Maximum frequency, fILmax GHz - fb 14
Fitted Insertion loss at Nyquist dB - 20 15
Fitted insertion loss, a0 dB -1 2 16
Fitted insertion loss, a1 dB 0 14.914 17
Fitted insertion loss, a2 dB 0 41.228 18
Fitted insertion loss, a4 dB 0 19.728 19
20
14.2.6.4 Insertion loss deviation (ILD) 21
22
The insertion loss deviation ILD is the difference between the measured insertion IL 23
and the fitted insertion loss ILfitted as defined in equation (14-3). 24
25
26
ILD  IL  IL fitted (14-3) 27
28
The insertion loss deviation ILD shall be within the region defined by equations (14-4) 29
and (14-5) where fb is the maximum baud rate to be supported by the channel under 30
test and fILmin and fILmax are given in Table 14-4. 31
32
33
 1.0  12.0( f f b ) f ILmin  f  f b 4  (14-4) 34
ILD  ILDmin   
 4. 0 f b 4  f  (3 / 4) f ILmax  35

36
1.0  12.0( f f b ) f ILmin  f  f b 4   37
ILD  ILDmax    (14-5) 38
 4.0 f b 4  f  (3 / 4) f ILmax  39
40
41
FOMILD is a figure of merit of the channel's insertion loss deviation (dB) from fILmin to 42
(3/4)*fILmax. In OIF-CEI-0.30 and OIF-CEI-03.1, this was called ILDRMS. FOMILD is 43
calculated as indicated below. 44
45
Define the weight at each frequency f using equation (14-6) below. 46
47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 293
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 1 1
2 W  f  = sinc  f  f b  ------------------------4- -------------------------8- (14-6)
3 1 +  f  ft  1 +  f  fr 
4
5
6 Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum 20 to
7 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365 (i.e. T_tr
8 x ft = 0.2365, T_tr is in ns when ft is in GHz). In addition, fr is the -3 dB reference
9 receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum baud rate
10 to be supported by the channel.
11
12 2
13
FOM =
 W  f   I LD  f 
--------------------------------------------
(14-7)
14 ILD
N
15
16
FOMILD is calculated using equation (14-7) where N is the number of frequency points.
17
The summation is done over the frequency range of ILD with f in GHz. FOMILD shall be
18
less than 0.3 dB for valid channels.
19
20
21 14.2.6.5 Channel Return Loss
22
23 Channel Return Loss shall be bounded by equation (14-8) as shown in Figure 14-3.
24
25 • RL(f) >= 12 dB for fmin < f  fb/4
26 
27 • RL(f) >= 12 dB - 15 Log10(4f/fb) for fb/4 < f < fb (14-8)
28
29
30 Note: f
min is as defined in Table 14-3
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

294 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 14-3. CEI-28G-MR normative channel return loss at 28.1 Gsym/s 1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
14.2.6.6 Channel integrated crosstalk noise 25
26
Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of 27
Table 14-5, the total integrated crosstalk noise for the channel shall be less than the 28
value specified by equation (14-9) and illustrated in Figure 14-4. 29
30
Table 14-5. Channel integrated crosstalk aggressor parameters 31
Parameter Symbol Value Units 32
33
Baud rate fb
max. Baud Rate
Gsym/s 34
sup. by Channel 35
Near-end aggressor peak to peak differential output amplitude Ant 1200 mVppd 36
Far-end aggressor peak to peak differential output amplitude Aft 1200 mVppd 37
38
Near-end aggressor 20 to 80% rise and fall times Tnt 8 ps 39
Far-end aggressor 20 to 80% rise and fall times Tft 8 ps 40
41
 x   x max = 10  mV RMS  for 3 dB  IL  5 3 dB 42
(14-9) 43
= 12  4 – 0  45 IL  mV RMS  for 5  3 d B  IL  20dB 44
 45
In equation (14-9), the IL denotes the value of the channel insertion loss in dB at 1/2 . 46
baud rate (NRZ). 47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 295
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 14-4. Illustration of integrated crosstalk noise limits

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

296 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14.3 Electrical Characteristics 1

2
The electrical signaling is based on high speed low voltage logic with a nominal 3
differential impedance of 100 . 4
5
All devices shall work within the range from 19.90 Gsym/s to 28.1 Gsym/s as specified 6
for the device, with the baud rate tolerance as per Section 3.2.11. Note that 7
implementation of specific protocols will define the operating baud rate without affecting 8
CEI compliance. 9
10
11
14.3.1 Transmitter Characteristics 12
The transmitter electrical specifications at compliance point T (see Figure 14-1) are 13
given in Table 14-6. The transmitter shall satisfy jitter requirements specified in Table 14
14-7. Jitter is measured as specified in Section 2.3.3, for a BER as specified in Section 15
14.2.4. It is assumed that the UBHPJ component of the transmitter jitter is not data- 16
dependent jitter (DDJ) from the receiver view point, hence it cannot be equalized in the 17
receiver. To attenuate noise and absorb even/odd mode reflections, the transmitter 18
shall satisfy the Common Mode Output Return Loss requirement of Table 14-6. 19
20
Link budgets in this document assume optimized TX FIR equalization that is part of the 21
system management function. The specific implementation is outside the scope of this 22
document. 23
24
Table 14-6. Transmitter Electrical Output Specification. 25
26
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 27
Baud Rate T_Baud 19.90 28.1 Gsym/s
28
29
Emphasis off.
Output Differential Voltage T_Vdiff
See Note 4.
800 1200 mVppd 30
Single Ended Transmitter Output
31
Voltage
T_Vse -0.3 1.9 V
32
Differential Resistance T_Rd 80 100 120  33
Differential Termination Resistance
34
Mismatch (see Table 1-2)
T_Rdm 10 % 35
Output Rise and Fall Time Emphasis off. 36
T_tr, T_tf 8 ps
(20% to 80%) See Note 2. 37
Common Mode Noise T_Ncm See Note 3. 12 mVRMS 38
Differential Output Return Loss T_SDD22 See Section 14.3.1.3 dB 39
Below 10 GHz -6
40
Common Mode Output Return Loss T_SCC22 dB 41
10 GHz to baud rate -4
42
Output Common Mode Voltage T_Vcm
Load Type 0
See Note 1
-100 1700 mV 43
44
NOTES:
1. Load Type 0 with min. T_Vdiff, AC-Coupling or floating load. 45
2. The transmitter under test is preset such that C0 is its maximum value (C0_max) and all other coefficients are zero. The 20% 46
and 80% values are of the steady state one and zero. The max value is limited by the linear fit pulse peak value in Table 14-
11.
47
3. Measurement procedure is defined in Section 12.3. 48
4. T_Vdiff is two times the steady-state value Vf as defined in Section 14.3.1.6.2. The value is given as differential p-p voltage. 49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 297
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3 Table 14-7. Transmitter Output Jitter Specification
4
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
5
6 Uncorrelated Unbounded Gaussian Jitter T_UUGJ Note 4 0.15 UIPP
7 Uncorrelated Bounded High Probability Jitter T_UBHPJ Note 2 0.15 UIPP
8 Even-Odd Jitter (component of UBHPJ) T_EOJ Note 3 0.035 UIPP
9 Total Jitter T_TJ Note 1 0.28 UIPP
10
NOTES:
11 1. T_TJ includes all of the jitter components measured without any transmit equalization.
12 2. Measured with all possible values of transmitter equalization, excluding DDJ as defined in Section 12.1.
3. Included in T_UBHPJ
13 4. Measured with all possible values of transmitter equalization
14
15
16
17
14.3.1.1 Transmitter Baud Rate
18
19 All devices shall work within the range from 19.90 Gsym/s to 28.1 Gsym/s as specified
20 for the device, with the baud rate tolerance as per Section 3.2.11. Note that
21 implementation of specific protocols will define the operating baud rate without affecting
22 CEI compliance.
23
24
25 14.3.1.2 Transmitter Amplitude and Swing
26
27 Transmitter differential output amplitude shall be able to drive between 800 to 1200
28 mVppd with transmit emphasis disabled. The single-ended transmitter output voltage
29 shall be between -0.3V and 1.9 V with respect to local ground. Transmitter differential
30 output amplitude shall additionally adhere to the requirements in Section 14.3.1.6.
31
32 14.3.1.3 Transmitter Return Loss
33
34 Please refer to Section 3.2.10 with the following parameters.
35
36 Table 14-8. Transmitter Differential Return Loss Parameters
37 Parameter Value Units
38
A0 -12 dB
39
40 f0 50 MHz

41 f1
0.1714 x
Hz
T_Baud
42
43 f2 T_Baud Hz

44 Slope 12.0 dB/dec

45
46 14.3.1.4 Transmitter Lane-to-Lane Skew
47
48 Please refer to Section 3.2.7
49

298 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14.3.1.5 Transmitter Short Circuit Current 1

2
Please refer to Section 3.2.9 3
4
14.3.1.6 Transmitter output waveform requirements 5
6
The transmitter shall include an equalizer defined as: 7
8
–1 –2 (14-10) 9
H  Z  = C –1 + C 0 z + C1 z 10
11
12
14.3.1.6.1 Summary of requirements 13
14
The normalized amplitudes of the coefficients of the transmitter equalizer (computed 15
per 14.3.1.6.2) shall meet the requirements in Table 14-9. 16
Table 14-9. Coefficient range and step size 17
18
Normalized Amplitude
Normalized Step 19
Coefficient
Min (%) Max (%)
Size (%) 20
21
C-1 -20 0 1.25 to 5 22
C1 -25 0 1.25 to 5 23
24
C0 40 100 1.25 to 5
25
26
The amplitude of a coefficient can be computed by multiplying its normalized amplitude 27
by vf, which is defined in equation (14-11). "min" is defined as the minimum normalized 28
amplitude of the coefficient that must be supplied by the transmitter to be compliant. 29
"max" is defined as the maximum normalized amplitude of the coefficient that must be 30
supplied by the transmitter to be compliant. 31
32
In addition:
33
a) |C-1|+|C0|+|C1|, the peak to peak output voltage shall not exceed 1200 mVppd. 34
35
b) C-1 + C0 + C1, the steady-state output voltage shall be greater than or equal to 80 36
mVppd. 37
38
39
14.3.1.6.2 Process to compute coefficients 40
41
The coefficients of the transmitter equalizer are defined by a calculation based on the 42
transmitter output waveform as described below. 43
44
1. The transmitter under test is preset such that C0 is its maximum value (C0_max) 45
and all other coefficients are zero. 46
47
2. Capture at least one complete cycle of the test pattern PRBS9 at T [ T is defined
48
as the test point at the output of transmitter package] per Section 14.3.1.6.3.
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 299
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 3. Compute the linear fit to the captured waveform per Section 14.3.1.6.4.
2
3 4. Define tx to be the time where the rising edge of the linear fit pulse, p, from step
4 3 crosses 50% of its peak amplitude.
5
6 5. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time 
7 t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi.
8
9 6. Use pi to compute the vector of coefficients, w, of a T_Nw-tap symbol-spaced
10 transversal filter that equalizes for the transfer function from the transmit function
11 to T per Section 14.3.1.6.5.
12
13 The parameters of the pulse fit and the equalizing filter are given in Table 14-10.
14 Table 14-10. Linear fit pulse and equalizing filter parameters
15
16 Parameter Value (UI)
17
18 Linear fit pulse length T_Np 8
19 Linear fit pulse delay T_Dp 2
20
21 Equalizer length T_Nw 8
22 Equalizer delay T_Dw 2
23
24 The differential zero to peak output voltage at T in the steady state, vf, is estimated by
25 equation (14-11).
26
27
28 1 M T _ Np (14-11)
29 vf    p(k )
M k 1
30 
31 In (14-11), p is the linear fit pulse from step 3 and M is the number of samples per
32 symbol as defined in 14.3.1.6.3. The peak value of the linear fit pulse from step 3, pmax,
33 shall satisfy the requirements of Table 14-11.The RMS value of the error between the
34 linear fit and measured waveform from step 3, e, shall satisfy the requirements of
35 Table 14-11.
36
Table 14-11. Transmitter output waveform requirements
37
38 Parameter Condition Units
39
40 Steady state output voltage, 2 x vf max mVppd 1200
41 Steady state output voltage, 2 x vf min mVppd 800
42
43 Linear fit pulse peak, pmax min - 0.80 x vf
44 RMS error, e max - 0.027 x vf
45
46
47
48 For each configuration of the transmit equalizer:
49

300 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

7. Configure the transmitter under test as required. 1

2
8. Capture at least one complete cycle of the test pattern PRBS9 at T. 3
4
9. Compute the linear fit to the captured waveform per Section 14.3.1.6.4. 5
6
10. Define tx to be the time where the rising edge of the linear fit pulse, p, from step 7
3 crosses 50% of its peak amplitude. 8
9
11. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time  10
t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi. 11
12
12. Equalize the sampled pulse, pi, using the coefficient vector, w, computed in step 13
6 per Section 14.3.1.6.5 to yield the equalized pulse qi. 14
15
The RMS value of the error between the linear fit and measured waveform from step 9, 16
e, shall satisfy the requirements of Table 14-11. 17
The normalized amplitude of coefficient C-1 is the value of qi at time t0 + (T_Dw - 1) UI. 18
The normalized amplitude of coefficient C0 is the value of qi at time t0 + T_Dw UI.  19
The normalized amplitude of coefficient C1 is the value of qi at time t0 + (T_Dw +1) UI. 20
21
22
14.3.1.6.3 Waveform acquisition 23
24
The transmitter under test repetitively transmits the specified test pattern. The 25
waveform shall be captured with an effective sample rate that is M times the signaling 26
rate of the transmitter under test. The value of M shall be an integer not less than 7. 27
Averaging multiple waveform captures is recommended. 28
29
The captured waveform shall represent an integer number of repetitions of the test 30
pattern totaling N bits. Hence the length of the captured waveform should be M .N 31
samples. The waveform should be aligned such that the first M samples of waveform 32
correspond to the first bit of the test pattern, the second M samples to the second bit, 33
and so on. 34
35
14.3.1.6.4 Linear fit to the waveform measured at T 36
37
Given the captured waveform y(k) and corresponding aligned symbols x(n) derived 38
from the procedure defined in Section 14.3.1.6.2, define the M-by-N waveform matrix Y 39
as shown in equation (14-12). 40
41
 42
 y(1) y( M  1)  y( M ( N  1)  1)  43
 y(2 ) y (M  2)  y (M ( N  1)  2 ) (14-12) 44
Y  45
     
  46
 y ( M ) y (2 M )  y (MN )  47
 48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 301
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Rotate the symbols vector x by the specified pulse delay Dp to yield xr.
2
3
4 
xr  x(T _ Dp  1) x(T _ Dp  2)  x( N ) x(1)  x(T _ Dp )  (14-13)
5
6
7
8 Define the matrix X to be an N-by-N matrix derived from xr as shown in equation (14-
9 14).
10
11 
 xr (1) xr ( 2)  xr ( N ) 
12
13
 x ( N ) x (1)  xr ( N  1)  (14-14)
X 
r r
14      
15  
16  xr (2) xr (3)  xr (1) 
17
18 Define the matrix X1 to be the first T_Np rows of X concatenated with a row vector of 1's
19 of length N. The M-by-(T_Np + 1) coefficient matrix, P, corresponding to the linear fit is
20 then defined by equation (14-15).
21
22
23 P  YX 1T ( X 1 X 1T ) 1 (14-15)
24
25
26
27 In equation (14-15) the superscript "T" denotes the matrix transpose operator.
28
29 
30  e (1) e( M  1)  e( M (N 1)  1)  
31  e(2) e(M  2)  e(M ( N  1)  2)
32 E  PX 1  Y    (14-16)
     
33  
34 e (M ) e(2 M )  e (MN ) 
35
36
37
38 The error waveform, e(k), is then read column-wise from the elements of E.
39
40
41
Define P1 to be a matrix consisting of the first T_Np columns of the matrix P as shown
42
in equation (14-17).
43
44
45
46
47
48
49

302 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

 1
 p(1) p( M  1)  p ( M (T _ N p  1)  1)   2
 p (2) p ( M  2)  p( M (T _ N p  1)  2) 3
P1    (14-17) 4
      5
 
 p( M ) p(2 M )  p ( MT _ N p )  6
7
8
9
The linear fit pulse response, p(k), is then read column-wise from the elements of P1. 10
11
14.3.1.6.5 Removal of the transfer function between the transmit function and T 12
13
Rotate sampled pulse response pi by the specified equalizer delay T_Dw to yield pr as 14
shown in equation (14-18). 15
 16

pr  pi (T _ Dw  1) pi (T _ Dw  2)  pi (T _ N p ) pi (1)  pi (T _ Dw )   17
(14-18) 18
19
Define the matrix P2 to be a T_Np-by-T_Np matrix derived from pr as shown in equation 20
(14-19). 21
22
 23
 pr (1) pr (T _ N p )  pr (2) 
 p (2)  24
p (1)  p (3) 25
P2   
r r r
(14-19) 26
     
  27
 pr (T _ N p ) pr (T _ N p  1)  pr (1)  28
 29
Define the matrix P3 to be the first T_Nw rows of P2. Define a unit pulse column vector 30
xp of length T_Np. The value of element xp(T_Dp + 1) is 1 and all other elements have a 31
value of 0. The vector of filter coefficients w that equalizes pi is then defined by (14-20). 32
33
 34
w  (P3T P3 ) 1 P3T x p 35

(14-20) 36
37
Given the column vector of equalizer coefficients, w, the equalized pulse response qi is 38
determined by equation (14-21). 39
40
qi  P3 w  41
 42
(14-21) 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 303
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 14.3.2 Receiver Characteristics

2
3 A compliant receiver shall operate at the specified BER with the worst case
4 combination of a compliant transmitter and a compliant channel.
5
6 Receiver electrical specifications are given in Table 14-12 and measured at compliance
7 point R. To dampen noise sources and absorption of both even and odd mode
8 reflections, the receiver shall satisfy the Common Mode Input Return Loss requirement
9 of Table 14-12. Jitter specifications at reference R (see Figure 14-1) are listed in Table
10 14-13.
11
12 Table 14-12. Receiver Electrical Input Specifications
13 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
14
15 Baud rate R_Baud 19.90 28.1 GSym/s

16 Input Differential Voltage R_Vdiff Note 1 1200 mVppd

17 Single Ended Input Voltage Vse
See 14.3.2.5
-300 2000 mV
18 See Note 2
19 Differential Input Impedance R_Rdin 80 100 120 
20 Input Impedance Mismatch R_Rm 10 %
21 Differential Input Return Loss R_SDD11 See 14.3.2.3
22
Below 10 GHz -6
23 Common Mode Input Return Loss R_SCC11 dB
10GHz to baud rate -4
24
25 Input Common Mode Voltage R_Vcm
Load Type 0 
-150 1750 mV
See Note 2
26
27 NOTES:
1. The receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the combined
28 transmitter and channel. The channel response shall include the worst case effects of the return losses at the transmitter and
29 2. receiver.
Load Type 0 with min. T_Vdiff, AC-Coupling or floating load. For floating load, input resistance shall be 1k. Only applies if
30 AC-coupling capacitor is integrated in receiver
31
32
33
Table 14-13. Receiver Input Jitter Specification
34
35 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
36
37 Sinusoidal Jitter, Maximum R_SJ-max
See Section
2.3.4, Note 1
5 UIpp
38
See Section
39 Sinusoidal Jitter, High Frequency R_SJ-hf
2.3.4, Note 1
0.05 UIpp
40 NOTES:
41 1. The Receiver shall tolerate the sum of these jitter contributions: Total transmitter jitter from Table 14-7;Sinusoidal jitter as
defined in Table 14-13;The effects of a channel compliant to the Channel Characteristics (Section 14.2.6).
42
43
44 14.3.2.1 Input Baud Rate
45
46 All devices shall work within the range from 19.90 Gsym/s to 28.1 Gsym/s as specified
47 for the device, with the baud rate tolerance as per Section 3.2.11.
48
49

304 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

14.3.2.2 Reference Input Signals 1

2
The receiver shall accept differential input signal amplitudes produced by a compliant 3
transmitter connected with the minimum attenuation specified in Figure 14-2 to the 4
receiver. This may be larger than the 1200 mVppd maximum of the transmitter due to 5
output/input impedances and reflections. 6
7
The minimum input amplitude is defined by the minimum transmitter amplitude, the 8
actual receiver input impedance and the loss of the actual PCB. Note that the minimum 9
transmitter amplitude is defined using a well controlled load impedance, however the 10
real receiver is not, which can leave the receiver input signal smaller than expected. 11
Additionally it will be determined by the environmental noise inside and outside the 12
receiver. 13
14
14.3.2.3 Input Resistance and Return Loss 15
16
Please refer to Section 3.2.10 with the following parameters. 17
18
Table 14-14. Receiver Differential Return Loss Parameters 19
Parameter Value Units 20
21
A0 -12 dB
22
f0 50 MHz 23
f1
0.1714 x
Hz 24
R_Baud
25
f2 R_Baud Hz 26
Slope 12.0 dB/dec 27
28
14.3.2.4 Input Lane-to-Lane Skew 29
30
Please refer to Section 3.2.8. 31
32
33
14.3.2.5 Single Ended Input Voltage 34
35
The single ended voltage levels with respect to the receiver ground at the input of the 36
receiver are dependent on the transmitter implementation and the inter-ground 37
difference. The voltage levels at the input of an AC coupled receiver (if the effective AC 38
coupling is done within the receiver) or at the TX side of the external AC coupling cap (if 39
AC coupling is done externally) will be between -0.3 to 2.0V with respect to local 40
ground. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface 305
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

306 Optical Internetworking Forum - Clause 14: CEI-28G-MR Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

15 Reserved to add future Clause 1

2
3
This page is intentionally blank 4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 15: Future Clause 307

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 This page is intentionally blank

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

308 Optical Internetworking Forum - Clause 15: Future Clause

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

16 CEI-56G-VSR-PAM4 Very Short Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-VSR-PAM4 very short reach high 4
speed chip-to-module electrical interface of nominal baud rates of 18.0 Gsym/s to 29.0 5
Gsym/s. A compliant host or module shall meet all of the relevant requirements listed
6
below. The electrical interface is based on high speed, low voltage logic, and 7
connections are point-to-point balanced differential pairs. 8
This clause defines the characteristics required to communicate between CEI-56G- 9
VSR-PAM4 drivers and CEI-56G-VSR-PAM4 receivers using copper signal traces on a 10
printed circuit board, a mated connector pair and copper signal traces inside an optical 11
module. These specifications are normative at the test points shown in Figure 16-1. A 12
‘length’ is effectively defined in terms of its attenuation and phase response rather than 13
its physical length. 14
15
Hosts and modules compliant to CEI-56G-VSR-PAM4 from different manufacturers 16
shall be interoperable. 17
18
19
16.1 Requirements 20
21
The objectives and requirements for the CEI-56G-VSR-PAM4 implementation 22
agreement are given by the project definition as follows: 23
24
• Support serial baud rates (fb) within the range from 18.0 Gsym/s to 29.0 Gsym/s 25
as specified for the device using PAM4 coding. A CEI implementation complies to 26
the specifications of this clause over the range of baud rates stated by the 27
implementer within this range. 28
• Capable of driving up to a minimum of 125 mm of host PCB trace plus one 29
connector and a minimum of 25 mm of module PCB trace. 30
• Capable of achieving a raw Bit Error Ratio (BER) of 10-6 or better per lane. FEC is 31
assumed to be used to achieve a corrected BER of 10-15 or better per lane. The 32
baud rate includes the overhead required for FEC. The definition of FEC is 33
outside the scope of this IA (see Appendix 16.D). 34
35
• Module electrical input to be self-adaptive and autonomous. 36
37
• Shall support AC-coupled operation. 38
• Shall allow multi-lanes (1 to n). 39
40
• Shall support hot plug. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 309
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.2 General CEI Requirements

2
3
4 16.2.1 Data Patterns
5 See Appendix 16.C.5.
6
7
8 16.2.2 Transmitter equalizer function
9
10 Both host and module Tx FIRs are likely to be required to meet TP1a and far-end TP4
11 eye requirements respectively for high loss channels (see Figure 16-1).
12
13 16.2.3 Bit Error Ratio
14
15 A raw Bit Error Ratio (BER) better than or equal to 10-6 is required on each lane. A
16 compliant host or module, when receiving from a compliant module or host, shall
17 deliver the specified raw BER to the subsequent FEC decoder. Error bursts with length
18 more than 15 PAM4 symbols delivered to the PAM4 decoder shall occur with a
19 probability of less than 1 in 1020 PAM4 symbols. See also Appendix 16.D.
20
21 16.2.4 Ground Differences
22
23 The maximum ground difference between the host and module shall be ±50 mV. The
24 common mode voltage limits are set taking this difference into account.
25
26
27 16.3 Electrical Characteristics
28
29 Hosts and modules shall meet the applicable specifications defined in Table 16-1, Table
30 16-2, Table 16-3, Table 16-4, Table 16-5 and Table 16-6 as applicable. The direction of
31 a signal (host-to-module or module-to-host) determines which table is applicable.
32
33 AC coupling is required in the module for both Tx and Rx.
34
35
36 16.3.1 Compliance Point Specifications
37
38 Figure 16-1 below gives the reference model and test points associated with host-to-
39 module and module-to-host lanes.
40
41 Reference test fixtures, called compliance boards, are used to access the electrical
42 specification parameters. The output of the Host Compliance Board (HCB) provides
43 access to the host-to-module electrical signal (host electrical output) defined at TP1a.
44 Additional module electrical input specifications, for host-to-module communication,
45 are defined at TP1, the input of the Module Compliance Board (MCB). The output of the
46 Module Compliance Board (MCB) provides access to the module to host electrical
47 signal (module electrical output) defined at TP4. Additional host electrical input
48
49

310 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

specifications, for module-to-host communication, are defined at TP4a, the input of the 1
Host Compliance Board (HCB). Informative specifications for the host transmit function 2
(TP0a) are given in Appendix 16.B. 3
4
Figure 16-1. Measurement points using compliance boards. 5
HCB PCB 6
TP0
Host PCB Trace Trace 7
8
Host‐to‐Module 
TP0a TP1a
9
Transmit function
10
11
Module‐to‐Host 
Receive function
TP4a 12
Host Channel 13
14
Module PCB Trace 
15
MCB PCB
Trace including AC coupling 16
17
Host‐to‐Module  18
TP1
Receive function 19
20
Module‐to‐Host  21
TP4
Transmit function 22
Module Channel
23
24
HCB PCB
MCB PCB
Trace
25
Trace
26
27
TP1 TP1a 28
29
TP4a
30
TP4 31
Mated MCB/HCB 32
33
34
16.3.2 Host-to-Module Electrical Specifications 35
36
Each host-to-module lane shall meet the specifications of Table 16-1 and Table 16-2. 37
Definitions and methodologies can be found in Sections 16.3.4 to 16.3.11. 38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 311
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 Table 16-1. Host-to-Module Electrical Specifications at TP1a (host output) (See Note 5)
2
3 Parameter Min. Max. Units Conditions
4 Differential Voltage pk-pk - 880 mV See Note 1
5
6 Referred to host ground
Common Mode Voltage (Vcm) -0.3 2.8 V
7 See Note 2
8 Common Mode Noise RMS - 17.5 mV See 16.3.5
9 Differential Termination
10 Resistance Mismatch
- 10.0 % At 1 MHz. See 16.3.6
11
Differential Return Loss
12 - Equation (16-1) dB See Note 3
13 (SDD22)
14 Common Mode to Differential
Equation (16-3)
15 Mode Conversion - dB See Note 3
16 (SDC22)
17 Common Mode Return Loss 250 MHz to fb GHz
18 - -2 dB
(SCC22) See Note 3
19
Transition Time 12.0 ps See 16.C.4.1
20
21 Eye Width at 10-6 probability
0.20 - UI
See 16.3.10
22 (EW6) See Note 4
23 Eye Height at 10-6 probability See 16.3.10
24 32 - mV
(EH6) See Note 4
25
26 See Equation (16-14)
Eye Linearity 0.85 - -
27 See Note 4
28 NOTES:
29 1. The differential voltage measured using a QPRBS13-CEI pattern will be less than than the in-service
30 differential voltage due to host loss and length of the QPRB13-CEI pattern.
2. Vcm is defined in Table 1-2 of Section 1.6
31 3. S-parameter specifications based on a differential reference impedance of 100  and a common
32 mode reference impedance of 25 
33 4. Open eye is generated through the use of a reference Continuous Time Linear Equalizer (CTLE)
34 applicable to all three PAM4 eyes (See Section 16.3.10.4)
35 5. The addition of a maximum VEC parameter is being considered as a maintenance action.
36
37
38
39
40
41
42
43
44
45
46
47
48
49

312 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 16-2. Host-to-Module Electrical Specifications (module input) 1

2
Parameter Test Point Min. Max. Units Conditions 3
Overload Differential Voltage
4
pk-pk
TP1a 900 - mV See 16.3.11 5
6
Common Mode Voltage (Vcm) TP1 -350 2850 mV See Note 1, 2 7
Differential Termination At 1 MHz. 8
TP1 - 10 % 9
Resistance Mismatch See 16.3.6
10
Differential Return Loss
TP1 - Equation (16-1) dB See Note 3 11
(SDD11) 12
Differential Mode to Common 13
Mode Conversion TP1 - Equation (16-2) dB See Note 3 14
(SCD11) 15
Stressed Input Test TP1a See Section 16.3.10.3 16
17
NOTES: 18
1. Vcm is defined in Table 1-2 of Section 1.6 19
2. Vcm is generated by the host. Specification includes effects of ground offset voltage
3. S-parameter specifications based on a differential reference impedance of 100  and a common 20
mode reference impedance of 25  21
22
23
Table 16-3. Crosstalk parameters for host output test and module stressed input test 24
calibration at TP4 25
26
Parameter Target value Units Conditions 27
Crosstalk Amplitude Differential Voltage pk-pk 900 mV 28
Crosstalk Slew Time (between -270 mV and +270 mV) 9.5 ps See Note 1
29
30
NOTES: 31
1. See Section 16.C.4.1 Transition Time and Slew Time
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 313
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.3.3 Module-to-Host Electrical Specifications

2
3 Each module-to-host lane shall meet the specifications of Table 16-4 and Table 16-5.
4 Definitions and methodologies can be found in Sections 16.3.4 to 16.3.11.
5
Table 16-4. Module-to-Host Electrical Specifications at TP4 (module output)
6
7
Parameter Min. Max. Units Conditions
8
9 Differential Voltage, pk-pk - 900 mV
10 Common Mode Voltage (Vcm) -350 2850 mV See Note 1, 2
11
12 Common Mode Noise, RMS - 17.5 mV See 16.3.5
13 Differential Termination - 10 % At 1 MHz. See 16.3.6
14 Resistance Mismatch
15 Differential Return Loss
16 (SDD22) - Equation (16-1) dB See Note 3
17
Common Mode to Differential
18 Mode Conversion
- Equation (16-3) dB See Note 3
19 (SDC22)
20
21 Common Mode Return Loss From 250 MHz to
- -2 dB fb GHz
22 (SCC22)
23 See Note 3
24 Transition Time 9.5 ps See 16.C.4.1
25 Near-end Eye Width at 10-6 See 16.3.10
26 probability (EW6) 0.265 - UI
See Note 4
27
-6 See 16.3.10
28 Near-end Eye Height at 10 70 - mV
29 probability (EH6) See Note 4
30 Far-end Eye Width at 10-6 See 16.3.10
0.20 - UI
31 probability (EW6) See Note 4
32
Far-end Eye Height at 10-6 See 16.3.10
33 probability (EH6)
30 - mV
See Note 4
34
35 Near-end Eye Linearity See Equation (16-14)
0.85 - -
36 See Note 4
37 NOTES:
38 1. Vcm is defined in Table 1-2 of Section 1.6
39 2. Vcm is generated by the host. Specification includes effects of ground offset voltage.
40 3. S-parameter specifications based on a differential reference impedance of 100  and a common
mode reference impedance of 25 
41 4. Open eye is generated through the use of a reference Continuous Time Linear Equalizer (CTLE)
42 applicable to all three PAM4 eyes (See Section 16.3.10.4)
43
44
45
46
47
48
49

314 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 16-5. Module-to-Host Electrical Specifications (host input) 1

2
Parameter Test Point Min. Max. Units Conditions 3
Overload Differential Voltage pk-pk TP4 900 - mV See 16.3.11
4
5
Differential Termination Resistance At 1 MHz. See 6
TP4a - 10 %
Mismatch 16.3.6
7
Differential Return Loss
TP4a - Equation (16-1) dB See Note 1
8
(SDD11) 9
Differential Mode to Common Mode 10
Conversion TP4a - Equation (16-2) dB See Note 1 11
(SCD11) 12
13
Stressed Input Test TP4 See 16.3.10.3
14
Common Mode Voltage (Vcm) TP4a -0.3 2.8 V See Note 2, 3 15
NOTES: 16
1. S-parameter specifications based on a differential reference impedance of 100  and a common 17
mode reference impedance of 25  18
2. Vcm is defined in Table 1-2 of Section 1.6 19
3. Referred to host ground. Common mode voltage is generated by host
20
21
22
Table 16-6. Crosstalk parameters for module output test and host stressed input test 23
calibration at TP1a 24
25
Parameter Target value Units Conditions 26
27
Crosstalk Amplitude differential voltage pk-pk 900 mV
28
Crosstalk Slew Time (between -270 mV and +270 mV) 19 ps See Note 1 29
NOTES: 30
1. See Section 16.C.4.1 Transition Time and Slew Time 31
32
33
16.3.4 Output Differential Voltage, pk-pk 34
The differential voltage, pk-pk, (see Section 1.6.1 for definition of differential voltage pk- 35
pk) shall meet the specifications given in Table 16-1 or Table 16-4 for the respective 36
communication direction. Host and module Tx FIRs are likely to be required to meet 37
TP1a and far-end TP4 eye requirements respectively for high loss channels. AC 38
coupling is required in the module for both Tx and Rx. The waveform is observed 39
through a fourth-order Bessel-Thomson response with a 3-dB bandwidth of 40 GHz 40
using a QPRBS13-CEI pattern (see Appendix 16.C.3.1). 41
42
43
16.3.5 Common Mode Noise 44
45
See Section 12.3. The oscilloscope bandwidth shall be 40 GHz. 46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 315
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.3.6 Differential Termination Resistance Mismatch

2
3 See Section 13.3.6.
4
5 16.3.7 Differential Return Loss
6
7 When measured at the respective test point the differential return loss shall not exceed
8 the limits given in Equation (16-1) (illustrated in Figure 16-2 for f =29.0 GHz).
b
9
10 All return loss measurements require that the relevant input or output be active. For
11 transmitter return loss measurements a QPRBS13-CEI pattern shall be used.
12
13 Figure 16-2. SDD11, SDD22 for host output (TP1a), host input (TP4a), module input (TP1) and
14 module output (TP4) (for fb = 29 GHz)
15
16 0
17
18 -1
19 -2
20 -3
21 -4
SDD11, SDD22 (dB)

22
23 -5
24 -6
25 -7
26 -8
27
28 -9
29 -10
30 -11
31 -12
32
33 -13
34 -14
35
11
0
1
2
3
4
5
6
7
8
9
10

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

36
37 Frequency (GHz)
38
39 The test points are TP1a for host output, TP4a for host input, TP1 for module input and
40 TP4 for module output.
41
42
43 SDD11, SDD22 < -11dB for 0.05 < f < f b  7.5
44
45 (16-1)
46  15 f 
SDD11, SDD22 < -6.0+9.2*log 10 ------ ---- dB for f b/7.5 < f < f b
47  7 f b
48
49

316 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

16.3.8 Common to differential mode and differential to common mode 1

conversion 2
3
The common to differential mode and differential to common mode conversion 4
specifications are intended to limit the amount of unwanted signal energy that is 5
allowed to be generated due to conversion of common mode voltage to differential 6
mode voltage or vice versa. 7
8
When measured at the respective input test point, differential to common mode 9
conversion shall not exceed the limits given in Equation (16-2) (illustrated in Figure 16- 10
3 for fb=29 GHz). 11
12
Figure 16-3. SCD11 for module input (TP1) and host input (TP4a), and SDC22 for module output
(TP4) and host output (TP1a) (for fb=29 GHz) 13
14
-6 15
-8 16
-10
17
18
-12 19
-14 20
-16 21
22
SDC22 (dB)
SCD11 (dB)

-18
23
-20 24
-22 25
-24 26
27
-26
28
-28 29
-30 30
31
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 32
Frequency (GHz) 33
34
35
SCD11 < -22+14* f  f b dB for 0.05  f  f b  2 36
37
(16-2)
38
SCD11 < -18+6*  f  f b dB for fb  2  f  f b
 39
When measured at the respective output test point, common to differential mode 40
conversion shall not exceed the limits given in Equation (16-3) (illustrated in Figure 16- 41
3 for fb=29 GHz). 42
43
44
SDC22 < -25+20* f  f b dB for 0.05  f  f b  2 45
46
(16-3) 47
SDC22 < -18+6* f  f b  dB for fb  2  f  f b
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 317
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.3.9 Common Mode Return Loss

2
3 The common mode output return loss specification is intended to limit the amount of
4 common mode energy that can be reflected by the host and module outputs. This has
5 an effect on EMI radiation and differential mode signals generated via common mode to
6 differential mode conversion. The common mode to differential mode conversion
7 specification for the host and module outputs is more stringent than for the inputs to
8 take into account the lack of a common mode input return loss specification.
9
10 16.3.10 Eye Width, Eye Height and Stressed Input tests
11
12 Eye Width and Eye Height are specified in Table 16-1 (host output) and Table 16-4
13 (module output). Compliance is verified using the test setup shown in Figure 16-4
14 (host) and Figure 16-5 (module). The Eye Width and Eye Height correspond to eye
15 contours at a probability of 10-6 to be consistent with those generated by simulator and
16 oscilloscopes based on CDF/histogram data. The 10-6 eye contours are constructed
17 from each individual eye’s contributing symbols. For example, the EH6/EW6
18 measurement could include a total of 32 million captured PAM4 symbols, at a rate of
19 one sample per symbol, with 16 million samples per individual eye (with the two middle
20 levels (i.e. +1/3 and -1/3) used for the outer eyes and the middle eye). The 10-6 vertical
21 eye points would have 8 samples at the top of each of the sub-eyes and 8 samples at
22 the bottom of each of the sub-eyes. The EW6 calculations would be based on 16
23 million edges for the middle eye and 12 million edges for the upper and lower eyes. The
24 10-6 horizontal eye points would have 8 samples at the left edge of the middle eye and
25 8 samples at the right edge for the middle eye, and 6 samples at the left edge and 6
26 samples at the right edge for the upper and lower eyes. Compliance to the input
27 specifications defined in Table 16-2 and Table 16-5 is verified using the test setup
28 shown in Figure 16-9 (host) and Figure 16-10 (module).
29
30
31 16.3.10.1 Host and Module output Eye Width and Eye Height test
32
The host output Eye Width and Eye Height are measured at TP1a of Figure 16-1 using
33
a Host Compliance Board as defined in Section 16.4.1. The test setup is shown in
34
Figure 16-4.
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

318 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-4. Host output test setup 1

2
3
Insertion  4
Interface TP1 5
VNA 6
Terminations
7
MCB 8
Crosstalk  9
calibration 10
or TP1a 11
CTLE
TP4 12
Scope DC 
block HCB 13
Reference  Crosstalk  14
CRU 15
generator 16
TP4a
17
18
19
Arrow Showing signal flow Host under test 20
Arrow Showing mechanical insertion 21
Electrical signals shown single ended for clarity 22
23
24
25
The module output Eye Width and Eye Height is tested at TP4 of Figure 16-1 using a 26
Module Compliance Board as defined in Section 16.4.1. The test setup is shown in 27
Figure 16-5. 28
29
Figure 16-5. Module output test setup
30
Insertion  31
Interface 32
TP1a TP1 33
Crosstalk  Crosstalk 
generator 34
calibration 35
HCB MCB 36
TP4 37
Terminations 38
or
DC  TP4a 39
block 40
VNA 41
42
Module under test
HCB 43
Emulated 
Host  44
Channel
Scope 45
Arrow Showing signal flow 46
Arrow Showing mechanical insertion Reference  47
Electrical signals shown single ended for clarity CTLE CRU 48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 319
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.3.10.1.1 Host and Module output test method

2
3 The signal at TP1a may be a closed eye. Therefore, a reference receiver with a
4 continuous time linear equalizer (CTLE) (see Section 16.3.10.4) is used to measure
5 Eye Width and Eye Height. The reference receiver is also used to equalize the module
6 output signal. The measured signal after the reference receiver shall meet the
7 specifications listed in Section 16.3.2 for host to module and Section 16.3.3 for module
8 to host. All co-propagating and counter-propagating lanes are active as crosstalk
9 sources, using a QPRBS13-CEI test pattern as defined in Appendix 16.C.3.1, or a
10 QPRBS31-CEI test pattern as defined in Appendix 16.C.3.2, or a valid CEI signal. The
11 QPRBS13-CEI or QPRBS31-CEI patterns on all pairs of aggressor lanes should be
12 asynchronous to each other or sufficiently delayed relative to each other to remove
13 correlation between all pairs of aggressor lanes and between the lane under test and all
14 aggressor lanes. Amplitude and Slew Times for counter-propagating lanes are defined
15 in Table 16-3 and Table 16-6. It is recognized that practical implementations may have
16 longer slew times than the target in Table 16-3.The lanes under test are asynchronous
17 to the lanes in the opposing direction within the PPM offset defined by the protocol in
18 use.
19
20 Note: Co- and counter-propagating crosstalk generators based on valid NRZ CEI
21 signals could be used but may overstress the system as crosstalk from such
22 generators may be greater than from PAM4 generators.
23
24 The test method for measuring either host or module output Eye Width and Eye Height
25 as illustrated in Figure 16-6 is as follows:
26 1) Set the host or module to QPRBS13-CEI pattern (see Appendix 16.C.3.1).
27
-This allows the use of a sampling oscilloscope with a pattern lock.
28
29 2) Capture the differential signal at TP1a or TP4 with a scope triggered with a clock
30 from a reference clock recovery unit (CRU) with a first order transfer function with a
31 3 dB tracking bandwidth of fb/6640.
32
33 -For TP1a, the scope shall be AC coupled.
34 -The reference CRU can be a software CRU in case of a real time scope.
35
36 -Sample the signal with a minimum of 3 samples per symbol or equivalent. Col-
lect sufficient samples in order to construct normalized cumulative distribution
37 functions (normalized CDFs) (see Figure 16-6) of the post-processed captured
38
39 signals to a probability of 10-6 (without extrapolation) as described below.
Depending on the sampling rate, careful interpolation using a method such as
40 sin(x)/x or cubic spline may be needed for good accuracy.
41
42
43
44
45
46
47
48
49

320 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
Figure 16-6. TP1a and TP4 Eye Width and Eye Height parameters 3
4
5
6
7
8
10‐6 9
probability
10
+1/3 level CDF‐high 11
histogram
12
13
EH6 (Vmid) 14
15
‐1/3 level CDF‐low
16
histogram 17
18
19
CDFL CDFR 20
EW6 (Hmid)
21
10‐6  22
probability 23
24
25
3) Apply the reference receiver as defined in Section 16.3.10.4 to equalize the 26
captured signal in step 2. 27
28
-For TP4 near-end compliance test, the CTLE peaking in the reference receiver 29
shall be set at 1 dB, 1.5 dB or 2 dB. Any CTLE setting that meets both the EH6 30
and EW6 settings defined for TP4 in Table 16-4 is acceptable. EW6 is the 31
minimum value from the set of Hlow, Hmid and Hupp measurements and 32
likewise EH6 is the minimum value from the set of Vlow, Vmid and Vupp 33
measurements (see Section 16.3.10.2 for these eye parameter definitions). 34
35
-For TP4 far-end compliance test, the signal measured at TP4 is first convolved 36
with an emulated loss channel (~7 dB loss at fb/2) that represents the worst 37
case channel loss. The loss channel is a host trace having 38
Zp = (151 x 29.0/fb) mm as defined below using Equation (16-4) to Equation 39
(16-9) with the transmission line parameters listed in Table 16-7. The response 40
of this channel is illustrated in Figure 16-14. The CTLE peaking in the reference 41
receiver is then set to one of the seventeen CTLE values in Table 16-9. Any 42
CTLE setting that meets both the EH6 and EW6 requirements defined for far- 43
end TP4 in Table 16-4 is acceptable. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 321
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1
2   1 – exp  –   f 2Z p  
3 S 11  f  = S 22  f  = -----------------------------------------------------
2
- (16-4)
4 1 –  exp  –   f 2Z p 
5 2
6  1 –   exp  –   f Z p 
7 S 21  f  = S 12  f  = -----------------------------------------------------
2 (16-5)
8 1 –  exp  –   f 2Z p 
9
10
11 where,
12
13
14   f  =  0 +  1 f +  2  f f
15 (16-6)
16
17
18 1 = a1  1 + j 
19 (16-7)
20
21
22
 2  f  = a 2  1 – j  2    ln f  + j2 (16-8)
23
24
25
26 Z c – 2R 0
 = ---------------------- (16-9)
27 Z c + 2R 0
28
29
30
31 with f in units of GHz in equations (16-4), (16-6), (16-6) and (16-8).
32 Table 16-7. Transmission Line Model Parameters and Values
33
34
Parameter Value Units
35
36 Zp 151 x 29.0/fb mm
37 0 0 1/mm
38 -4
a1 4.114 x 10 ns1/2/mm
39
40 a2 2.547 x 10-4 ns/mm
41  6.191 x 10-3 ns/mm
42
Zc 109.8 
43
44 R0 50.0 
45
46
47
48
49

322 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

-For TP1a compliance test the CTLE peaking in the reference receiver shall be 1
set to one of the seventeen CTLE values in Table 16-9. Any CTLE setting that 2
meets both the EH6 and EW6 requirements defined for TP1a in Table 16-1 is 3
acceptable. EH6 and EW6 have the same meaning as in the previous 4
paragraph. 5
6
4) At TP1a, passing is defined as at least one equalizer setting that meets the EH6, 7
EW6 and linearity specifications defined in Table 16-1 for the lower, upper and middle 8
eyes. At TP4, passing is defined as at least one equalizer setting that meets the near- 9
end EH6, EW6 and linearity specifications given in Table 16-4 and at least one 10
potentially different equalizer setting that meets the far-end EH6 and EW6 11
specifications given in Table 16-4. 12
13
16.3.10.2 Measured PAM4 Eye Parameter Definitions 14
15
Figure 16-7. TP1a and TP4 Eye Width, Eye Height and Eye Amplitude 16
17
18
19
+1 mean level 20
21
Hupp
Vupp 22
AVupp 23
24
25
time center of
+1/3 mean level 26
middle eye, Tmid 27
Vmid AVmid 28
Hmid
29
30
1 UI 31
-1/3 mean level 32
33
Vlow
34
Hlow AVlow 35
36
37
-1 mean level 38
39
40
41
42
All relevant PAM4 eye parameters are determined from 10-6 contours generated from 43
oscilloscope CDF/histogram data except for Tmid which is determined from a 10-3 44
contour of the middle eye. 45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 323
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 The approach described in this clause is based upon first locating the midpoint (Tmid)
2 of the middle eye’s maximum horizontal eye opening and using that information to
3 determine its vertical eye opening (Vmid) and the vertical eye opening of the upper and
4 lower eyes, Vupp and Vlow, respectively based on the same time window. The
5 midpoints of Vmid, Vupp and Vlow are used to determine Hmid, Hupp and Hlow
6 respectively.
7
8 1. Tmid - the midpoint of the maximum horizontal eye opening of the 10-3 inner eye
9 contour of the middle eye
10
11 2. Vmid - the 10-6 inner Eye Height of the middle eye determined from voltage CDFs in
12 a +/- 0.025 UI time window centered on Tmid
13
14 3. Vupp - the 10-6 inner Eye Height of the upper eye determined from voltage CDFs in a
15 +/- 0.025 UI time window centered on Tmid
16
17 4. Vlow - the 10-6 inner Eye Height of the lower eye determined from voltage CDFs in a
18 +/- 0.025 UI time window centered on Tmid
19
20 5. AVmid, the Eye Amplitude of the middle eye, is the difference of the mean levels of
21 the +1/3 level and -1/3 level voltage histograms in a +/- 0.025 UI time window centered
22 on Tmid
23
24 6. AVupp, the Eye Amplitude of the upper eye, is the difference of the mean levels of
25 the +1 level and +1/3 level voltage histograms in a +/- 0.025 UI time window centered
26 on Tmid
27 7. AVlow, the Eye Amplitude of the lower eye, is the difference of the mean levels of the
28 -1/3 level and -1 level voltage histograms in a +/- 0.025 UI time window centered on
29 Tmid
30
31 8. Hmid - the 10-6 inner Eye Width determined from CDFs of eye edges half way
32 between the 10-6 points of the voltage CDFs of the middle eye (Vmid/2)
33
34 9. Hupp - the 10-6 inner Eye Width determined from CDFs of eye edges half way
35 between the 10-6 points of the voltage CDFs of the upper eye (Vupp/2)
36
37 10. Hlow - the 10-6 inner Eye Width determined from CDFs of eye edges half way
38 between the 10-6 points of the voltage CDFs of the lower eye (Vlow/2)
39
40 11. Create an Eye Width mask centered on Tmid having a width of EW6 from the
41 relevant table (Table 16-1 or Table 16-4, as the case may be) which extends above and
42 below the waveform for the upper and lower PAM4 eyes as shown in Figure 16-8.
43
44 12. The Hmid, Hupp and Hlow eye edges shall be outside this Eye Width mask
45
46
47
48
49

324 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-8. PAM4 Horizontal Eye Mask 1

Eye Mask
2
3
4
Upper Eye
LHS meas. point
Upper Eye
RHS meas. point
5
Hupp
Vupp
AVupp
6
7
8
9
time center of
middle eye, Tmid

10
Vmid AVmid
Hmid

11
12
Lower Eye
LHS meas. point
Lower Eye
RHS meas. point 13
14
Vlow
Hlow AVlow

15
16
17
18
19
16.3.10.3 Host and Module stressed input test
20
The ability of the host input to tolerate the Eye Width, Eye Height and Eye Linearity 21
specified in Table 16-4 and the sinusoidal jitter specified in Table 16-8 is tested using a 22
stressed input test. The test signal is applied at TP4a of Figure 16-1, and calibrated at 23
TP4, using a Host Compliance Board and Module Compliance Board specified in 24
Section 16.4.1 The test setup is shown in Figure 16-9. The UBHPJ block is used to 25
create non-compensable DJ in addition to sinusoidal jitter. 26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 325
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 Figure 16-9. Host input test setup

2
Insertion 
3 Interface
4 Sinusoidal TP1 Crosstalk 
5 Jitter
generator
6 MCB Stressed signal 
7 Pattern  UBHPJ TP4 calibration
8 generator
9
UUGJ
10
CTLE Scope
11 Termination and  TP1a
12 crosstalk calibration Reference 
13 HCB CRU
14
15 or
16 TP4a
17 VNA
18
19
20 Arrow Showing signal flow Host under test
Arrow Showing mechanical insertion
21 Electrical signals shown single ended for clarity
22 UUGJ = Uncorrelated Unbounded Gaussian Jitter
23
UBHPJ = Uncorrelated Bounded High Probability 
24
Jitter
25
26
27
28 The ability of the module input to tolerate the Eye Width, Eye Height and Eye Linearity
29 specified in Table 16-1 and the sinusoidal jitter specified in Table 16-8 is tested using a
30 stressed input test. The test signal is applied at TP1 of Figure 16-1, and calibrated at
31 TP1a using a Host Compliance Board and Module Compliance Board specified in
32 Section 16.4.1. The test setup is shown in Figure 16-10. The module stressed input test
33 represents the worst case high loss host. Modules are also expected to operate at the
34 BER specified in Section 16.1 when presented with lower loss channels that require
35 different CTLE settings as long as the signal complies with the specifications in Table
36 16-1. In contrast with CEI-28G-VSR the module input shall tolerate these various
37 channels in an autonomous manner with no recommended equalization settings being
38 provided by the host.
39
40
41
42
43
44
45
46
47
48
49

326 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-10. Module input test setup 1

Insertion  2
Sinusoidal Interface 3
TP4a Crosstalk  4
Jitter 5
generator
UBHPJ
HCB 6
Pattern 
generator TP1a Stressed signal  7
UUGJ calibration 8
9
Termination and  TP4 10
crosstalk calibration CTLE Scope 11
MCB 12
Frequency  Reference  13
dependent  CRU 14
or TP1 15
attenuator
VNA 16
17
18
Arrow Showing signal flow
Arrow Showing mechanical insertion 19
Module  20
Electrical signals shown single ended for clarity
under test 21
UUGJ = Uncorrelated Unbounded Gaussian Jitter
UBHPJ = Uncorrelated Bounded High Probability  22
Jitter 23
24
25
26
27
16.3.10.3.1 Host (TP4a) and Module (TP1) stressed input test method 28
29
The host and module input shall tolerate sinusoidal jitter with the frequency and 30
amplitude defined by the mask of Figure 16-11 and Table 16-8. This sinusoidal jitter is 31
part of the jitter applied in the stressed input test. The sinusoidal jitter is calibrated at 32
10x the reference CRU’s bandwidth and must be tested at fCRU/100, fCRU/3, fCRU, 33
3fCRU, and 10fCRU, where fCRU is the jitter corner frequency given by fb/6640, with 34
sinusoidal jitter of 5 UI, 0.15 UI, 0.05 UI, 0.05 UI and 0.05 UI respectively. 35
36
The reference CRU and reference receiver as defined in Section 16.3.10.4 are used to 37
calibrate the stressed input test signal at TP4 (per Table 16-4) or TP1a (per Table 16-1) 38
using a QPRBS13-CEI pattern. The pattern is changed to QPRBS31-CEI for the 39
stressed input test. 40
41
The crosstalk source is asynchronous to the main pattern generator. The amplitude and 42
slew time of the crosstalk source are given in Table 16-3 and Table 16-6. The crosstalk 43
signal is calibrated at TP4 or TP1a using a QPRBS13-CEI pattern, then the pattern is 44
changed to QPRBS31-CEI for the test. For multi-lane implementations additional lanes 45
shall be active with either uncorrelated QPRBS31-CEI or QPRBS13-CEI patterns or 46
valid CEI signals, using the above calibration methods. 47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 327
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 Figure 16-11. Host input and Module input Sinusoidal Jitter

2
3
4
5 UIpp X
5
6 Sinusoidal
7 jitter
8 amplitude
9
10 X
11
12 0.05 UIpp X X X
13
14
15
16 fb/664000 fb/6640 10fCRU
17
18
19
20 Table 16-8. Sinusoidal jitter frequency for TP4 and TP1a testing
21 Sinusoidal jitter, peak-to-
22 Frequency Range
peak (UI)
23
24 f<fb/664000 Not specified
25 f /664000 < f < f /6640 5fb/(664000f)
b b
26
27 fb/6640 < f < 10fCRU 0.05
28
29 16.3.10.3.1.1 Host input test signal calibration
30
31 The host input is tested at TP4a of Figure 16-1 using a Host Compliance Board as
32 defined in Section 16.4.1. The host input test setup is shown in Figure 16-9.
33
34 UBHPJ, UUGJ and sinusoidal jitter are added to a clean test pattern until the jitter
35 (except for EOJ) at the output of the pattern generator approximates the informative
36 transmit recommendations given in Appendix 16.B.
37
38 With the crosstalk generator calibrated to meet the specifications in Table 16-6, the Far-
39 end Eye Height and Far-end Eye Width at TP4 are measured using the reference
40 receiver defined in Section 16.3.10.4 with the optimal CTLE peaking value from Figure
41 16-12 and the methodology defined in Section 16.3.10.1. The optimal CTLE peaking
42 value is defined as the setting that results in the maximum value of EW6*EH6.
43
44 The UUGJ and pattern generator amplitude are adjusted to give the minimum far-end
45 Eye Height and far-end Eye Width and minimum Eye Linearity specified for the module
46 output in Table 16-4. The upper and lower PAM4 eyes are adjusted to be smaller than
47 the middle eye and set to the minimum Eye Height and Eye Width requirement. This
48
49

328 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

calibrated signal must also equal or exceed the Near-end Eye Height and Near-end 1
Eye Width in Table 16-4 measured using the reference receiver defined in Section 2
16.3.10.4 with one of the CTLE peaking values from Figure 16-13. 3
4
16.3.10.3.1.2 Module input test signal calibration 5
6
The module input is tested at TP1 of Figure 16-1 using a Module Compliance Board as 7
defined in Section 16.4.1. The module input test setup is shown in Figure 16-10. 8
9
Sinusoidal jitter compliant to Table 16-8 and UUGJ are added to a clean test pattern 10
until the jitter (except for EOJ) at the output of the pattern generator approximates the 11
informative transmit recommendations given in Appendix 16.B. Note that the sinusoidal 12
jitter provides the UBHPJ component including the effects of host transmitter crosstalk. 13
14
The frequency-dependent attenuator is intended to represent the host channel, and 15
may be implemented with PCB traces. It should be chosen to result in a loss of 12.2 dB 16
at Nyquist to TP1a from the output of a pattern generator having a nominal rise/fall time 17
of 10.5 ps. (Cable or PCB loss may be added between the pattern generator and the 18
test channel to produce the 10.5 ps rise/fall time, in which case the output of the cable 19
or PCB is counted as the output of the pattern generator).The complete path from the 20
output of the pattern generator to TP1a should also meet the return loss specifications 21
given for the mated HCB and MCB (see sections 16.4.1 and 13.4.1.2). The crosstalk 22
generator is calibrated to meet the specifications in Table 16-3. The Eye Height and 23
Eye Width at TP1a are measured using the reference receiver (defined in Section 24
16.3.10.4) with the optimal CTLE peaking value and the methodology defined in 25
Section 16.3.10.1. The optimal CTLE peaking value is defined as the setting that 26
results in the maximum value of EW6*EH6. The pattern generator output is adjusted so 27
that this optimal CTLE peaking value is greater than or equal to 7 dB. 28
29
The UUGJ and pattern generator amplitude are adjusted to give the minimum Eye 30
Height and Eye Width and minimum Eye Linearity specified in Table 16-1. 31
32
The upper and lower PAM4 eyes are adjusted to be smaller than the middle eye and 33
set to the minimum Eye Height and Eye Width requirement. 34
35
16.3.10.4 Reference receiver 36
37
The waveform is observed through a fourth-order Bessel-Thomson response with a 38
bandwidth of 40 GHz concatenated with a Continuous Time Linear Equalizer (CTLE). 39
The filters may be implemented in software; however, the signal is not averaged. The 40
CTLE shall be implemented based on Equation (16-10) where G is the low-frequency 41
gain and ZLF, Z1, PLF, P1 and P2 are the CTLE zero and poles coefficients. Figure 16- 42
12 shows the frequency response of the reference equalizer used for module far-end 43
and host output testing for baud rates between 25 and 29.0 GBd with values for ZLF, 44
Z1, PLF, P1 and P2 listed in Table 16-9. Figure 16-13 shows the frequency response of 45
the reference equalizer used for module near-end output testing for baud rates 46
between 25 and 29.0 GBd with values for ZLF, Z1, PLF, P1 and P2 listed in Table 16-9. 47
The CTLE is the same for baud rates between 25 and 29.0 GBd. For baud rates below 48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 329
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 25 GBd the values of ZLF, Z1, PLF, P1 and P2 should be multiplied by fb/28. The CTLE
2 peaking value is the approximate difference between the low-frequency response
3 (1 MHz) and the maximum high-frequency response in dB.
4 (16-10)
5  G   P1   P2   PLF   jf + Z1   jf + ZLF 
H  f  = ------------------------------------------------- -------------------------------------------- --------------------------
6  Z1   ZLF   jf + P1   jf + P2   jf + PLF 
7
8
9 Figure 16-12. Host output reference equalizer (CTLE) transfer function for 1 dB to 9 dB of peaking
10 at 25 Gsym/s
11
12
13 1
14
15 0
16 1 dB curve
-1
17
CTLE Response (dB)

18 -2
19
20 -3
21 -4
22
23 -5
24 -6
25
26 -7
27
28 -8
29 9 dB curve
-9
30
31 -10
32 1E8 1E9 1E10 5E10
33
34 Frequency (Hz)
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

330 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-13. Module near-end output reference equalizer (CTLE) transfer function for 1 dB to 2 1
dB of peaking at 25 Gsym/s 2
3
4
0.5 5
6
0.0 7
8
9
CTLE Response (dB)

-0.5
10
1 dB curve 11
-1.0
12
13
-1.5 14
2 dB curve 15
-2.0 16
17
-2.5 18
19
-3.0 20
21
22
-3.5
23
1E8 1E9 1E10 5E10 24
25
Frequency (Hz) 26
27
Table 16-9. Reference equalizer coefficients for rates of 25 GBd and above 28
29
Peaking
G P1 (GHz) P2 (GHz) Z1 (GHz) PLF (GHz) ZLF (GHz) 30
(dB) 31
1 0.891251 26.5625 14.1 9.463748 1.2 1.2 32
33
1.5 0.841395 26.5625 14.1 9.248465 1.2 1.15
34
2 0.794328 26.5625 14.1 9.069645 1.2 1.1 35
2.5 0.749894 26.5625 14.1 8.640319 1.2 1.075 36
37
3 0.707946 26.5625 14.1 8.255665 1.2 1.05
38
3.5 0.668344 26.5625 14.1 7.906766 1.2 1.025 39
4 0.630957 26.5625 14.1 7.587650 1.2 1 40
4.5 0.595662 26.5625 14.1 7.076858 1.2 1
41
42
5 0.562341 26.5625 14.1 6.614781 1.2 1 43
5.5 0.530884 26.5625 14.1 6.193091 1.2 1 44
6 0.501187 26.5625 14.1 5.805801 1.2 1 45
46
6.5 0.473151 26.5625 14.1 5.448395 1.2 1 47
7 0.446684 26.5625 14.1 5.117337 1.2 1 48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 331
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 Peaking
2 G P1 (GHz) P2 (GHz) Z1 (GHz) PLF (GHz) ZLF (GHz)
(dB)
3
4 7.5 0.421697 26.5625 14.1 4.809777 1.2 1
5 8 0.398107 26.5625 14.1 4.523367 1.2 1
6 8.5 0.375837 26.5625 14.1 4.256129 1.2 1
7
8 9 0.354813 26.5625 14.1 4.006377 1.2 1
9
10 The TP4 far-end module output test uses the emulated host channel specified in
11 Section 16.3.10.1.1 with insertion loss represented by SDD21 as shown in Figure 16-
12 14.
13 Figure 16-14. TP4 Far-end Host Channel SDD21 for fb = 29 GHz
14
15 0
16 -1
17 -2
18 -3
19 -4
20 -5
21
SDD21 (dB)

-6
22
-7
23
24 -8
25 -9
26 -10
27 -11
28 -12
29 -13
30 -14
31
0

10

12

14

16

18

20

22

24

26

28
29
32
33 Frequency (GHz)
34
35
36 16.3.11 Input Overload Voltage Tolerance
37
38 The input voltage tolerance tests the acceptance of differential input pk-pk amplitudes
39 produced by the extremes of operation from the host output (for host-to-module
40 communication) or module output (for module-to-host communication).
41
42 The maximum voltage at an IC input can be larger than the maximum at the
43 compliance point due to output/input impedances and reflections.
44
45 The input overload voltage tolerance specification is to be met for any valid CEI pattern.
46 The differential voltage specified in Table 16-1 is somewhat smaller than the module
47 input voltage tolerance due to host loss and the QPRBS13-CEI’s pattern length.
48
49

332 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

16.4 Measurement methods 1

2
3
16.4.1 Compliance Boards 4
Use of compliance boards for testing is assumed for the parameters defined in Table 5
16-1 through Table 16-6. The test results for test and calibration at TP1a should be 6
corrected for any deviations between the Host Compliance Board’s loss and the Host 7
Compliance Board’s reference loss given in Section 13.4.1.1. The test results for test 8
and calibration at TP4 should be corrected for any deviations between the mated 9
compliance board loss minus the Host Compliance Board’s loss and the mated 10
compliance board reference loss (given in Section 16.4.1.1) minus the Host 11
Compliance Board’s reference loss given in Section 13.4.1.1. 12
13
Figure 16-1 shows the test setup for making S parameter measurements of the mated 14
compliance boards.The requirements in this section are not connector specifications for 15
an implemented design. The compliance boards are as defined in Section 13.4 with the 16
exceptions that their specifications are extended from 28.1 GHz to 29.0 GHz, and other 17
exceptions in Section 16.4.1.1. 18
19
20
16.4.1.1 Mated HCB and MCB S-parameters 21
22
The reference mated MCB-HCB loss is given in Equation (16-11). 23
24
2 25
SDD21 ,SDD12 =  – 0.475  f – 0.1204f – 0.002f (16-11) 26
27
28
for 0.05 GHz < f < 29.0 GHz, where f is frequency in GHz, loss in dB. 29
30
The FOMILD (as calculated using the method defined in Section 10.2.6.4 and the curve 31
fit method defined in Clause 12 with fILmax of 21.75 GHz and fILmin of 50 MHz) for the 32
mated HCB and MCB pair is 0.1 dB. 33
34
Note: FOMILD is called ILDrms in OIF-CEI-03.0 clauses 10 & 11 and OIF-CEI-03.1 35
clauses 10, 11, 13 & 14. 36
37
The Integrated Crosstalk Noise (ICN) as calculated using the method defined in Clause 38
12 with the aggressor amplitudes and rise/fall times as listed in Table 16-3 shall be less 39
than 3.9 mV RMS. MDNEXT shall be less than 1.35 mV RMS. MDFEXT shall be less 40
than 3.6 mV RMS. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 333
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.A Appendix - Recommended Electrical Channel

2
3 The channel consists of Host PCB trace, Module PCB trace, vias, AC coupling
4 capacitor and one connector, not in this order. The recommended PCB trace differential
5 impedance is 100 ± 10  . This full channel model is shown in Figure 16-15 below.
6 Note that in practice the channel is not measurable as appropriate test points are not
7 accessible.
8
9 Figure 16-15. CEI-56G-VSR-PAM4 Channel Reference Model
10
conn. module PCB
11 up to + cap up to
12 host PCB up to 7.3 dB 1.2 dB 1.5 dB
13
14
Connector AC
15 Coupling Module IC
16 Line Card Host IC
Cap
17
18
19
10.0 dB channel loss at f = fb/2
20
21
22
23 16.A.1 Insertion Loss
24
25 Host insertion loss and module insertion loss are recommended limits only. Achieving
26 these recommended limits does not signify compliance nor guarantee successful
27 communication between two devices. Equation (16-12) (illustrated in Figure 16-16)
28 represents the highest recommended insertion loss (see SDD21) of the end-to-end
29 channel.
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

334 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-16. Recommended minimum SDD21 of the end-to-end channel (for fb = 29 GHz) 1
2
0 3
4
-2
5
-4 6
VSR full channel SDD21 (dB)

7
-6
8
-8 9
10
-10 11
-12 12
13
-14 14
-16 15
16
-18 17
-20 18
19
-22 20
-24 21
22
0

10

12

14

16

18

20

22

24

26

28
29
23
24
Frequency (GHz) 25
26
27
28
f f f 2
H  f  = a 0 + a 1 ---- + a 2 ---- + a 4  ---- 29
fb fb f  (16-12) 30
b
31
32
33
where a0 = -0.1, a1 = -7.51, a2 = -2.38, a4 = -13.56
34
In addition it is recommended that the VSR channel have an FOMILD less than or equal 35
to 0.3 dB. 36
37
Note: FOMILD is called ILDrms in OIF-CEI-03.0 clauses 10 & 11 and OIF-CEI-03.1 38
clauses 10, 11, 13 & 14. 39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 335
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.B Appendix - Informative Host Transmitter output Electrical

2 Recommendations
3
4 Informative host Tx output recommendations are defined in Table 16-10.
5
6
7 16.B.1 Host Transmitter output test point
8
9 Figure 16-1 gives the reference model and test points associated with host-to-module
10 and module-to-host CEI-56G-VSR-PAM4 lanes. The informative host transmitter output
11 electrical recommendations are defined to be measured at TP0a. TP0a is defined to be
12 separated from TP0, the ball of the package performing the host-to-module transmit
13 function, by 1 dB of PCB attenuation at 14 GHz.
14
15 16.B.1.1 Host-to-Module transmitter output Electrical Recommendations
16
17 It is recommended that each host-to-module lane meet the limits of Table 16-10.
18
19 Table 16-10. Host-to-Module Electrical Recommendations at TP0a
20
21 Parameter Symbol Min. Max. Units Conditions
22 Baud Rate 18.0 29.0 GBd
23 Differential Voltage, pk-pk T_Vdiff 750 - mV See Note 1
24
25 DC Common Mode Voltage T_Vcm -0.3 2.8 V See Note 2
26 Differential Termination
T_Rdm - 10 % at 1 MHz
27 Resistance Mismatch
28 Equation
29 Differential Return Loss T_SDD22 -
(17-4)
dB at TP0
30 Transition Time: 20% to 80% T_tr, T_tf 7.5 - ps With emphasis off
31
32 Common-mode return loss -6 +
T_SCC22 - dB
3*f/fb
33
34 Common Mode Noise, RMS T_Ncm - 12 mV
35 Uncorrelated Unbounded
0.01 UIRMS
36 Gaussian Jitter (UUGJ)
37 Uncorrelated Bounded High
38 Probability Jitter (UBHPJ) 0.05 UI
39 Even-Odd Jitter (EOJ) 0.019 UI
40
41 Signal-to-noise-and-distortion 31 - dB
See Section 17.3.1.6.4
ratio for definition
42
43 NOTES:
44 1. Max voltage is limited by specifications at TP1a. Minimum voltage can be lower for low loss
channels.
45 2. Load type 0 with min. T_Vdiff, AC-Coupling or floating load.
46
47
48 It is unlikely that all TP1a requirements of Table 16-1 can be met simultaneously with
49 maximum channel insertion loss and the worst-case limits of Table 16-10.

336 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

16.C Appendix - General PAM4 Requirement 1

2
This annex contains normative requirements for PAM4 signaling and includes 3
information relevant to all of the CEI-56G PAM4 clauses. 4
5
6
16.C.1 Transmit Functional Requirements 7
8
In the transmit direction, the role of the TX is to map the signal from the higher layer 9
protocol to a PAM4 encoded signal to be passed to the physical channel for transfer. 10
The mapping process shown in Figure 16-17 includes application of Gray coding 11
followed by PAM4 encoding.
12
Figure 16-17. Transmit signaling and mapping diagram 13
14
15
16
Higher 17
PAM4
Layer Gray Coder 18
Encoder 19
Protocol
20
21
22
23
16.C.1.1 Gray Coding
24
The TX signaling process shall map consecutive pairs of bits to one of four Gray-coded 25
symbols as specified in this subclause. Each pair of bits, (A, B) of digital input bits are 26
converted to a Gray-coded symbol with one of the four Gray-coded levels as follows: 27
(0, 0) maps to 0, (0, 1) maps to 1, (1, 1) maps to 2, and (1, 0) maps to 3. The “A” bit is 28
the first bit received. 29
30
31
16.C.1.2 PAM4 Encoding 32
33
The TX transmit process shall encode each Gray coded output symbol to one of four 34
PAM4 levels as specified in this subclause. 35
36
Mapping from the Gray coder output to a PAM4 encoded symbol is as follows: 37
38
0 maps to –1, 39
40
1 maps to –1/3,
41
2 maps to +1/3, and 42
43
3 maps to +1. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 337
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.C.1.3 Optional Precoding

2
3 Certain systems susceptible to DFE burst errors, such as those based on MR and LR
4 PAM4 interfaces, may take advantage of a simple per lane precoding block
5 implementing a 1/(1+D) modulo 4 function following the Gray coder, where D is defined
6 as 1 UI of delay at the baud rate.
7
8 16.C.2 Receive Functional Requirements
9
10 The receive process shall recover the data generated by the transmit process. In the
11 receive direction, the role of the RX is to unmap the PAM4 encoded signal to the higher
12 layer protocol. The mapping processes shown in Figure 16-18 includes application of
13 PAM4 decoding followed by inverse Gray coding.
14
15 Figure 16-18. Receive signaling and mapping diagram
16
17
18 Higher
19 PAM4 Inverse
Layer
20 Decoder Gray Coder
21 Protocol
22
23
24
25 16.C.2.1 Optional Precoding Decoder
26
27 Systems that have the optional precoding block are required to have a precoding
28 decoder block located before the inverse Gray coder which implements a (1+D) modulo
29 4 function, where D is defined as 1 UI of delay at the baud rate. Such systems should
30 be able to bypass this decoder function in order to be compatible with non-precoder
31 based systems.
32
33
16.C.3 Test Patterns
34
35
36 16.C.3.1 Quaternary PRBS13 test pattern - QPRBS13-CEI
37
38 The QPRBS13-CEI test pattern is the 4-level pattern created by encoding a repeating
39 PRBS13 pattern as defined in this paragraph using the Gray coding and PAM4
40 encoding described in this Annex. Each cycle of QPRBS13-CEI is 8191 symbols long.
41 The QPRBS13-CEI pattern generator produces the same result as the implementation
42 shown in Figure 16-19, which implements the generator polynomial shown in Equation
43 (16-13). If multiple test patterns are required at one time, they shall not be aligned (for
44 example, by use of different seeds or delays).
45
46 G(x) = 1 + x + x2 + x12 + x13 (16-13)
47
48
49

338 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 16-19. QPRBS13-CEI pattern generator 1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
16.C.3.2 Quaternary PRBS31 test pattern - QPRBS31-CEI 20
21
The QPRBS31-CEI test pattern is the 4-level pattern created by encoding the PRBS31 22
test pattern defined in [21] using the Gray Coding and PAM4 encoding described in this 23
Annex. 24
25
16.C.4 PAM4 Signal Parameters 26
27
16.C.4.1 Transition Time and Slew Time Based on the QPRBS13-CEI Test 28
Pattern 29
30
Transition times (rise and fall times) are defined as the time between the 20% and 80% 31
times, or 80% and 20% times, respectively, of isolated -1 to +1 or +1 to -1 PAM4 edges. 32
Slew times are defined as the time interval between the times at defined voltages. 33
Using the QPRBS13-CEI test pattern the transitions within sequences of three -1s 34
followed by three +1s, and three +1s followed by three -1s, respectively, are measured. 35
These are PAM4 symbols 1820 to 1825 and 2086 to 2091, respectively, where symbols 36
1 to 7 are the run of seven +1s. In this case, the 0% level and 100% level may be 37
estimated as the average signal within windows from -1.5 UI to -1 UI and from 1.5 UI to 38
2 UI relative to the edge. 39
40
The waveform is observed through a fourth-order, low-pass Bessel-Thomson filter 41
response having a 3-dB bandwidth of 40 GHz. 42
43
44
16.C.4.2 Eye Linearity 45
46
Eye Linearity is a metric based upon PAM4 eye parameters as determined from 47
contours generated from oscilloscope CDF/histogram data as described in Section 48
16.3.10. Eye Linearity is calculated according to Equation (16-14). 49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 339
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1
2
min  AVupp AVmid AVlow 
3 Eye Linearity = ------------------------------------------------------------------------------- (16-14)
4 max  AVupp AVmid AVlow 
5
6
7
8 16.C.4.3 Transmitter Linearity
9
10 Transmitter linearity is defined as a function of the mean signal level transmitted for
11 each PAM4 symbol. Given the PAM4 symbols -1, -1/3, +1/3, and +1, the mean signal
12 level for each symbol are V-1, V-1/3, V+1/3, and V+1 respectively. The calculation of the
13 mean signal levels is defined below. The mid-range level Vmid is defined by Equation
14 (16-15). The mean signal levels are then normalized and the signal offset is adjusted so
15 that V-1 corresponds to -1, V-1/3 to -ES1, V+1/3 to ES2, and V+1 to +1. ES1 is defined by
16 Equation (16-16) and ES2 is defined by Equation (16-17). The level separation
17 mismatch ratio RLM is defined by Equation (16-18).
18
19
20 Vmid = (V-1 + V+1) / 2 (16-15)
21
22
23 ES1 = (V-1/3 - Vmid) / (V-1 - Vmid) (16-16)
24
25
26 ES2 = (V+1/3 - Vmid) / (V+1 - Vmid) (16-17)
27
28
29 RLM = min((3ES1), (3ES2), (2 - 3ES1), (2 - 3ES2)) (16-18)
30
The mean signal levels described above are measured from a waveform captured
31
while the transmitter is transmitting the QPRBS13-CEI test pattern. The waveform
32
consists of M samples per unit interval and is aligned such that the first M samples of
33
the waveform correspond to the first PAM4 symbol of the test pattern, the second M
34
samples to the second PAM4 symbol, and so on. This allows each sample of the
35
waveform to be associated with a specific PAM4 symbol in the test pattern.
36
37 Denote the number of PAM4 symbols in the test pattern as N. Reduce the captured
38 waveform to N samples by choosing the central sample from each unit interval. The
39 central sample is defined as the mth sample in each unit interval where m is the integer
40 closest to M / 2.
41
42 For each PAM4 symbol x, Vx is the mean value of the waveform samples that
43 correspond to that symbol.
44
45 Note. Transmitter Linearity is not used in this clause but is used by other PAM4 clauses.
46
47
48
49

340 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

16.C.5 PAM4 Data Patterns 1

2
The PAM4 Implementation Agreements do not have any requirements for specific data 3
patterns, data coding, or scrambling. However the following requirements are 4
necessary to insure proper operation of a CEI PAM4 interface. If all of these conditions 5
are not met, then the link may not work to the full distance, or meet the specified BER, 6
or in fact work at all. 7
• A transition occurs for the middle PAM4 eye if the signal level changes from either - 8
1 or -1/3 to either +1/3 or +1, or vice versa. The average transition density of middle 9
eye transitions needs to converge to 0.5 over a long period (> 109 symbols), but can 10
in the extreme be between 0.45 and 0.55 over a 30,000 symbol period with a 11
probability of at least one minus the BER ratio (1 - 10-6). 12
13
• A transition occurs for the upper PAM4 eye if the signal level changes from +1 to 14
any other signal level, or from any other signal level to +1. A transition occurs for the 15
lower PAM4 eye if the signal level changes from -1 to any other signal level, or from 16
any other signal level to -1. The average transition density of the upper or lower eye 17
transitions needs to converge to 0.375 over a long period (> 109 symbols), but can 18
in the extreme be between 0.3375 and 0.4125 over a 30,000 symbol period with a 19
probability of at least one minus the BER ratio (1 - 10-6). 20
21
• The weighted average DC balance is calculated by assigning a weights of { +1, 22
+2/3, +1/3, 0 } to the PAM4 signal levels { +1, +1/3, -1/3, and -1 } respectively. The 23
weighted average DC balance needs to converge to 0.5 over a long period (> 109 24
symbols), but can in the extreme be between 0.45 and 0.55 over a 30,000 symbol 25
period with a probability of at least one minus the BER ratio (1 - 10-6). 26
• The probability of symbol run lengths over 10 is to be proportional to 4-N for N-like 27
symbols in a row (N>10). Hence, a run length of 40 symbols would occur with a 28
maximum probability of 4-40. 29
30
• If a fixed block coding scheme is used, the raw data must be scrambled before 31
coding or the coded data must be scrambled prior to transmission to prevent the 32
occurrence of worst case patterns (e.g. CJPAT-like patterns). 33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 341
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 16.D Appendix - FEC Guidance and Background Material

2
3 This annex provides general guidance with regards to system FEC coding schemes
4 and receiver error propagation behavior in order to ensure that end-to-end system
5 performance may be sufficient to enable corrected BER less than 10-15 for the CEI-56G
6 PAM4 interfaces.
7
8 In cases where the raw BER of the electrical link is expected to be greater than the
9 targeted BER, it is necessary to ensure that a suitable FEC algorithm be chosen which
10 is able to provide the necessary coding gain to achieve the targeted end-to-end
11 corrected BER. Figure 16-20 shows a graph of the input to output BER for some
12 common FEC code examples in the form of Additive White Gaussian Noise (AWGN).
13
14 Figure 16-20. Input to Output BER of common FEC codes
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38 A different way of quantifying the correction capabilities of above FEC codes is shown
39 in Table 16-11 below. The expected performance of each FEC code in the presence of
40 AWGN is quantified by the random error coding gain. However, depending upon
41 various system behaviors a receiver may not only experience random errors but may
42 be required to mitigate a combination of random and burst errors. Therefore it is
43 important that the specified pre-FEC BER of each CEI-56G-PAM4 clause include the
44 potential bursts that are forwarded to the FEC correction block. In addition a maximum
45 PAM4 symbol burst error -20
length that a CEI-56G-PAM4 receiver shall expect with a
46 probability less than 10 is defined for each clause to allow tailoring of the FEC
47 solution with respect to randomization of bursts by means of symbol interleaving or
48 similar methods.
49

342 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Random error distribution for a FEC code can only be assumed, if the burst size does 1
not exceed the code symbol size, “m”. One significant source of burst errors are the 2
presence of DFE blocks in any system which can multiply a single error into a burst of 3
errors and is highly dependent on DFE tap weights. As an example a 10-bit burst would 4
not exceed the symbol size of either the RS(528,514) or RS(544,514) codes and 5
therefore it could not affect more than two symbols but could affect three FEC symbols 6
of the the RS(255,239) whose symbol is made up of 8-bits. The BCHxBCH(1020,988) 7
is unable to address any significant burst errors. 8
9
Table 16-11. Common FEC Codes 10
11
Random Input BER 12
Symbol Correctable Rate
Error for 10-15
Code Name size, m symbols Overhead
Coding corrected
Conditions 13
(bits) per block, t (%)
Gain (dB) BER 14
15
G.709 RS8(255,239) [8] 8 8 6.7 5.4 810-5 16
IEEE 100GBASE-KR4
10 7 2.7 4.9 210-5
17
RS(528,514) [25] 18
IEEE 100GBASE-KP4 19
10 15 5.8 6.1 210-4 See Note 1
RS(544,514) [25] 20
N/A due to 21
ITU G.975.1 (I.9) [28] 22
1 concate- 3.2 9 410-3
BCHxBCH(1020,988)
nated code 23
NOTES: 24
1. IEEE KP4 is presently under consideration by several standards bodies for electrical PAM4 25
interfaces. 26
27
One technique to mitigate burst errors is the use of code word interleaving in order to 28
randomize received errors. For example, a system based on RS (255,239) coding and 29
2x interleaving would be able to successfully decode a 10-bit burst over two code 30
blocks. However interleaving multiplies the FEC latency proportional to the depth of 31
interleaving and also increases the implementation complexity. 32
33
Another technique to mitigate burst errors due to Decision Feedback Equalizers is to 34
use the pre-coding described in Section 16.C.1.3. This converts typical error bursts 35
caused by DFEs into one symbol error at the beginning of the burst and one symbol 36
error at the end. However, pre-coding converts single symbol errors into two symbol 37
errors, doubling the error rate for random errors. 38
39
A selected FEC implementation is considered sufficient for the application, if it can 40
correct the raw BER of a dedicated CEI-56G-PAM4 clause to the desired system target 41
BER, assuming a normal, uniform random error distribution (according Figure 16-20), 42
plus implements the required burst handling to effectively randomize the maximum 43
error bursts specified in the relevant clause. 44
45
Another important consideration is the worst-case error statistics for end-to-end 46
systems, specifically where all of the error correction is performed at a far-end 47
termination point and the system is made up of several errored links. In such cases the 48
performance of each of the links must be understood in order to determine the overall 49

Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface 343
 Draft 13 of oif2014.230 Common Electrical I/O (CEI)

1 expected raw BER and the maximum possible burst error length. Raw BERs and bursts
2 lengths of the individual concatenated links may add up over the complete end-to-end
3 path and the required FEC implementation has to be selected accordingly.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

344 Optical Internetworking Forum - Clause 16: CEI-56G-VSR-PAM4 Very Short Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

17 CEI-56G-MR-PAM4 Medium Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-MR-PAM4 medium reach high speed 4
electrical interface between nominal baud rates of 18.0 Gsym/s and 29.0 Gsym/s using 5
PAM4 coding. A compliant device shall meet all of the requirements listed below. The 6
electrical interface is based on high speed, low voltage logic. Connections are point-to- 7
point balanced differential pairs and signaling is unidirectional. 8
9
The electrical IA is based on loss and jitter budgets and defines the characteristics required 10
to communicate between a CEI-56G-MR-PAM4 transmitter and a CEI-56G-MR-PAM4 11
12
receiver using copper signal traces on a printed circuit board. The characteristic impedance
13
of the signal traces is nominally 100 Ω differential. The signal trace or channel between a 14
transmitter and a receiver shall meet the channel operating margin (COM), a method and a 15
threshold quantity used for channel compliance. 16
17
CEI-56G-MR-PAM4 assumes using forward error correction (FEC) to achieve the bit error 18
ratio (BER) target. The FEC guidances are described in Appendix 16.D. 19
20
Medium reach CEI-56G-MR-PAM4 devices from different manufacturers shall be 21
interoperable. 22
23
24
17.1 Requirements 25
26
1. Support serial baud rates (fb) within the range from 18.0 Gsym/s to 29.0 Gsym/s as 27
specified for the device using PAM4 coding. A CEI implementation complies to the 28
specifications of this clause over the range of baud rates stated for the 29
implementation within this range. 30
2. Capable of achieving a raw Bit Error Ratio (BER) of 10-6 or better per lane. FEC is 31
assumed to be used in the system to achieve corrected BER of 10-15or better per 32
lane. The baud rate includes the overhead required for FEC. The definition of FEC 33
is outside the scope of this IA (see Appendix 16.D). 34
35
3. Capable of driving up to 500 mm of PCB and up to 1 connector. 36
4. Shall support AC-coupled operation. 37
38
5. Shall allow multi-lanes (1 to n). 39
40
6. Shall support hot plug.
41
42
17.2 General Requirements 43
44
45
17.2.1 Data Patterns 46
47
See Appendix 16.C.5. 48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 345
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 17.2.2 Bit Error Ratio

2
3 A raw Bit Error Ratio (BER) better than or equal to 10-6 is required on each lane. A
4 compliant receiver, when receiving from a compliant transmitter over a compliant
5 channel, shall deliver the specified raw BER to the subsequent FEC decoder. Error
6 bursts with length more than 94 PAM4 symbols delivered to the PAM4 decoder shall
7 occur with a probability of less than 1 in 1020 PAM4 symbols. See Appendix 16.D.
8
9 17.2.3 Ground Differences
10
11 Please refer to Section 3.2.4.
12
13
17.2.4 Channel Compliance
14
15 A forward channel and associated dominant crosstalk channels are deemed compliant
16 if the channel characteristics conform to the requirements defined in this section.
17
18
19 17.2.4.1 Reference Model
20
Figure 17-1. CEI-56G-MR Reference Model
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35 Note: Test points differ from definitions in Section 1.8, as a DC blocking capacitor, if
36 physically located outside of the package, is part of the channel. The mezzanine
37 connector represents any board to board connector.
38
39 The channel is defined between test point T and test point R.
40
41 17.2.4.2 Channel Operating Margin
42
43 The Channel Operating Margin (COM) of the channel is computed using the procedure
44 in Annex 93A of IEEE Std 802.3 [25] as modified by IEEE Std 802.3by [26] and IEEE
45 Std 802.3bs [27], with the Test 1 and Test 2 values in Table 17-1. Test 1 and Test 2
46 differ in the value of the device package model transmission line length zp. Moreover,
47 using Tr = 0.345 UI, and β=2 for Ht(f) in Equation (93A–19), COM shall be greater than
48
49

346 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

or equal to 3.0 dB for each test. This minimum value allocates margin for practical 1
limitations on the receiver implementation, the largest step size allowed for transmitter 2
equalizer coefficients. 3
4
Table 17-1. COM Parameter Values 5
Parameter Symbol Value Units 6
Signaling rate fb 18.0 - 29.0 Gsym/s
7
8
Maximum start frequency fmin 0.05 GHz 9
Maximum frequency step Δf 0.01 GHz 10
11
Device package model    12
160 fF 
Single-ended device capacitance Cd 
mm 
13
Transmission line length, Test 1 zp  12
Transmission line length, Test 2 zp  30 mm  14
Transmission line characteristic impedance Zc 95 Ω 15
Single-ended package capacitance at Cp 110 fF 16
package-to-board interface 17
Single-ended reference resistance R0 50 Ω 18
19
Single-ended termination resistance Rd 50 Ω 20
Receiver 3 dB bandwidth fr 0.75 × fb GHz 21
Transmitter equalizer, minimum cursor c(0) 0.60 — 22
23
Transmitter equalizer, pre-cursor coefficient c(–1)   24
Minimum value –0.15 —
Maximum value 0 — 25
Step size 0.05 — 26
27
Transmitter equalizer, post-cursor coefficient c(1)   28
Minimum value  –0.25 —
0 —
29
Maximum value 
Step size 0.05 — 30
31
Continuous time filter, DC gain gDC   32
Minimum value –15 dB 
0 dB  33
Maximum value
Step size 1 dB 34
35
Continuous time filter, DC gain2 gDC2   36
Minimum value –4 dB 
0 dB 
37
Maximum value
Step size 1 dB 38
39
Continuous time filter, scaled zero frequency fz fb /2.5 GHz 40
41
42
Continuous time filter, pole frequencies fp1 fb /2.5 GHz 43
fp2 fb GHz
44
45
Continuous time filter, low frequency fLF fb /40 GHz
46
pole/scaled zero
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 347
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 17-1. COM Parameter Values

2
  
3 Transmitter
Victim
differential peak output voltage
Av  0.41 V 
4 Far-end aggressor Afe  0.41 V 
5 Near-end aggressor Ane 0.60 V
6
7 Number of signal levels L 4 —

8 Level separation mismatch ratio RLM 0.95 —

9 Transmitter signal-to-noise ratio SNRTX 32.5 dB
10 Number of samples per unit interval M 32 —
11
12 Decision feedback equalizer (DFE) length Nb 10 UI
13 Normalized DFE coefficient magnitude limit  bmax(1) 0.5 —
14 for n = 2 to Nb bmax(2-Nb) 0.2
15 Random jitter, RMS σRJ 0.01 UI
16 Dual-Dirac jitter, peak ADD 0.02 UI
17
18 One-sided noise spectral density η0 2.6 × 10-8 V2/GHz
19 Target detector error ratio DER0 10-6 —
20 Channel operating margin, min COM 3.0 dB
21
22
23 17.2.4.3 Informative Channel Insertion Loss
24 Figure 17-2. Channel Insertion Loss Limit for 29.0 Gsym/s
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

348 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
 2
 29
 29 + 0.676 f-------------  3
IL max = 1.083 + 2.398 f-------------  f min  f  f b (17-1)
fb fb 4
5
6
7
8
9
10
 11
0 f min  1GHz 
(17-2) 12
IL min = 1
---  f – 1  1GHz  f  17.5GHz 13
3 14
5.5 17.5GHz  f  f b 15
16
17
Channel insertion loss is an informative recommendation. 18
19
The channel must comply with the normative specification in Section 17.2.4.2. 20
21
17.2.4.4 Channel Return Loss 22
23
Figure 17-3. Channel Return Loss Limit for 29.0 Gsym/s 24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
 
44
45
Channel Return Loss shall be bounded by Equation (17-3) as shown in Figure 17-3. 46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 349
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3 fb
12 f min  f  ----
4 4
RL max =
5 fb (17-3)
4f
6 12 – 15 log  -----   ----  f  f b
10 f b 4
7
8
9
10 17.2.4.5 Channel AC-coupling
11
12 The transmitter shall be AC-coupled to the receiver. The impact of a DC-blocking
13 capacitor implemented in the channel between the package balls of the transmitter and
14 receiver (i.e., between compliance points T and R) is accounted for within the channel
15 specifications. Common-mode specifications are defined as if the DC-blocking
16 capacitor is implemented in the channel between compliance points T and R. Should
17 the capacitor not be implemented between compliance point T and compliance point R,
18 it is the responsibility of implementers to consider any necessary modifications to
19 common-mode and channel specifications required for interoperability as well as any
20 impact on the verification of transmitter and receiver compliance. In particular the
21 common-mode specifications for the transmitter in Table 17-2 may not be appropriate.
22
23 The low-frequency 3 dB cutoff of the AC-coupling shall be less than 100 kHz.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

350 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

17.3 Electrical Characteristics 1

2
The electrical signaling is based on high speed low voltage logic with a nominal differential 3
impedance of 100 Ω. 4
5
6
17.3.1 Transmitter Characteristics 7
8
The transmitter electrical requirements at compliance point T (see Figure 17-1) are 9
specified in Table 17-2, and the jitter requirements are specified in Table 17-3. 10
Table 17-2. Transmitter Electrical Output Specification 11
12
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 13
14
Baud Rate T_Baud 18.0 29.0 Gsym/s
15
Output Differential Voltage T_Vdiff See Note 1, 2. 1200 mVppd 16
DC Common mode Voltage T_Vcm See Note 2. 0 1.9 V 17
Output AC Common Mode Voltage T_VcmAC See Note 1, 2. 30 mVrms 18
Single-ended Transmitter Output Voltage T_Vse See Note 1, 2. -0.3 1.9 V 19
Differential Output Return Loss T_SDD22 Equation (17-4) dB
20
21
Common Mode Output Return Loss T_SCC22 Equation (17-5) dB
22
Level Separation Mismatch Ratio T_RLM 0.95 -
23
Steady-state Voltage T_Vf
See Note 1, 2,
0.4 0.6 V 24
Linear Fit Pulse Peak T_Pk 3, 4 0.83 ×
V
25
T_Vf 26
Signal-to-Noise-and-Distortion-Ratio T_SNDR 31 dB 27
NOTES: 28
1. Signals are specified as measured through a fourth-order Bessel-Thomson low-pass response with 40 GHz 3 dB bandwidth.
2. Measured as described in Section 17.3.1.2. T_Vdiff min is set by the steady-state voltage T_Vf min.
29
3. Measured as described in Section 17.3.1.6. 30
4. T_RLM is defined in Appendix 16.C.4.3. 31
32
Table 17-3. Transmitter Output Jitter Specification 33
34
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
35
Uncorrelated Jitter (time interval from 0.005% to
T_J4u 0.118 UI 36
99.995% of the probability distribution)
37
Uncorrelated jitter RMS (standard deviation of
the probability distribution)
T_JRMS
See Note 1
0.023 UIrms 38
Even-Odd Jitter T_EOJ 0.019 UIpp
39
40
NOTES:
1. Measured as described in Section 17.3.1.7. 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 351
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 17.3.1.1 Transmitter Baud Rate

2
3 All devices shall work within the range from 18.0 Gsym/s to 29.0 Gsym/s as specified
4 for the device, with the baud rate tolerance as per Section 3.2.11. A CEI
5 implementation complies to the specifications of this clause over the range of baud
6 rates stated for the implementation within this range.
7
8
9 17.3.1.2 Transmitter Amplitude and Swing
10
11 The differential output voltage T_Vdiff is defined to be True minus Complement. The
12 common-mode output voltage T_Vcm is defined to be one half of the sum of True and
13 Complement. These definitions are illustrated in Section 1.6.1.
14
15 For a QPRBS13-CEI test pattern (Appendix 16.C.3.1), the peak-to-peak value of the
16 differential output voltage (T_Vdiff) shall be less than or equal to the limit given in Table
17
17-2 regardless of the transmit equalizer setting.
18
19
20 The DC common-mode output voltage (T_Vcm) shall be within the limits in Table 17-2
21 with respect to local ground.
22
23 The AC common-mode output voltage (T_VcmAC) shall be less than or equal to the
24 limit given in Table 17-2 with respect to local ground. Common-mode output voltage
25 requirements shall be met regardless of the transmit equalizer setting.
26
27 The single-ended transmitter output voltage (T_Vse) shall be within the limits in Table
28 17-2 with respect to local ground.
29
30 The transmitter shall be capable of providing a differential steady state output
31 amplitude (2xT_V ) between 800 and 1200 mVppd with transmit emphasis disabled.
f
32
33 Transmitter differential output amplitude shall additionally adhere to the requirements in
34 Section 17.3.1.6.
35
36 Power-down behavior is beyond the scope of CEI IA.
37
38
39 17.3.1.3 Transmitter Return Loss
40
41 The differential output return loss, in dB, of the transmitter shall meet Equation (17-4),
42 where f is the frequency in GHz. The differential return loss limit RLd(f) is shown in
43 Figure 17-4. This output impedance requirement applies to all valid output levels. The
44 reference impedance for differential return loss measurements shall be 100 Ω.
45
46
47
48
49

352 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

f  29 1
 12.05 – 0.4112  -------------  0.05  f  0.5f b  
 fb  2
RL d  f   RLmax  f  =    dB  (17-4) 3
f  29
 7.175 – 0.075  -------------  0.5f b  f  f b   4
  fb   5
6
Figure 17-4. Transmitter differential return loss limit for 29.0 Gsym/s 7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
  24
25
26
The common-mode output return loss, in dB, of the transmitter shall meet Equation (17- 27
5), where f is the frequency in GHz. The common-mode return loss limit RLC(f) is shown 28
in Figure 17-5. This output impedance requirement applies to all valid output levels. 29
The reference impedance for common-mode return loss measurements shall be 25 Ω. 30
31
32
f  29 33
 9.05 – 0.4112  -------------  0.05  f  0.5f b   34
 fb 
RL c  f   RLmax  f   =    dB  (17-5) 35
f  29
 4.175 – 0.075  -------------  0.5f b  f  f b   36
  fb   37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 353
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 17-5. Transmitter common mode return loss limit for 29.0 Gsym/s.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20  
21
22
23 17.3.1.4 Transmitter Lane-to-Lane Skew
24
25 Please refer to Section 3.2.7.
26
27 17.3.1.5 Transmitter Short Circuit Current
28
29 Please refer to Section 3.2.9.
30
31
32 17.3.1.6 Transmitter output waveform requirements
33
34 The transmitter function includes programmable equalization to compensate for the
35 frequency-dependent loss of the channel and facilitate data recovery at the receiver.
36 The functional model for the transmit equalizer is the three tap transversal filter shown
37 in Figure 17-6.
38
39
40
41
42
43
44
45
46
47
48
49

354 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 17-6. Transmit equalizer functional model 1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Link budgets in this document assume optimized TX FIR equalization that is part of the 21
system management function. The specific implementation is outside the scope of this 22
document. 23
24
17.3.1.6.1 Linear fit to the measured waveform 25
26
The following test procedure defines linear fit pulse response, linear fit error (e(k), see 27
Section 11.3.1.6.4), and normalized transmitter coefficient values. 28
29
For each configuration of the transmit equalizer, capture at least one complete cycle of 30
the QPRBS13-CEI test pattern (Appendix 16.C.3.1) at the TX package ball (see Figure 31
17-1). 32
33
Compute the linear fit pulse response p(k) from the captured waveform per Section 34
11.3.1.6.2 using Np = 12 and Dp = 2. For aligned symbol values x(n) use -1, -ES1, ES2, 35
and 1 to represent symbol values of -1, -1/3, 1/3, and 1, respectively, and where ES1 36
and ES2 are the effective symbol levels determined in Appendix 16.C.4.3. 37
38
Define r(k) to be the linear fit pulse response when transmit equalizer coefficients have 39
been set to the “preset” values (see Section 11.3.1.6.1). 40
41
For each configuration of the transmit equalizer, compute the normalized transmit 42
equalizer coefficients, c(i), according to Section 11.3.1.6.2 - Section 11.3.1.6.5. 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 355
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 17.3.1.6.2 Steady-state voltage and linear fit pulse peak

2
3 The linear fit pulse, p(k), is determined according to the linear fit procedure in Section
4 11.3.1.6.2 - Section 11.3.1.6.5, as modified by Section 17.3.1.6.1. The steady-state
5 voltage T_Vf is defined to be the sum of the linear fit pulse p(k) divided by M, as shown
6 in Equation (11-2).
7
8 The steady-state voltage, T_Vf, shall satisfy the requirements in Table 17-2.
9
10 The linear fit pulse peak, T_Pk, is the highest value of p(k). It shall satisfy the
11 requirement in Table 17-2.
12
13 17.3.1.6.3 Transmitter equalizer coefficients
14
15 Table 17-4. Coefficient Range and Step Size
16 Normalized Amplitude Normalized Step
17 Coefficients
Min (%) Max (%) Size (%)
18
c(-1) -15 0 0.5 to 5
19
20 c(1) -25 0 0.5 to 5
21 c(0) 60 100 0.5 to 5
22
23 The normalized amplitudes of the coefficients of the transmitter equalizer (computed
24 per Section 17.3.1.6.1) shall meet the requirements in Table 17-4. "min" is defined as
25 the minimum normalized amplitude of the coefficient that must be supplied by the
26 transmitter to be compliant. "max" is defined as the maximum normalized amplitude of
27 the coefficient that must be supplied by the transmitter to be compliant.
28
29 The amplitude of a coefficient can be computed by multiplying its normalized amplitude
30 by T_Vf, which is defined in Section 17.3.1.6.2.
31
32 The peak-to-peak output voltage is approximated by
33
34 ( |c(-1)| + |c(0)| + |c(1)| ) * 2 * T_Vf (17-6)
35
36 and should not exceed the limit for T_Vdiff given in Table 17-2.
37
38 17.3.1.6.4 Transmitter Output Noise and Distortion
39
40 Signal-to-noise-and-distortion ratio (SNDR) is measured at the transmitter output using
41 the following method, with the transmitter on the lane under test transmitting
42 QPRBS13-CEI and transmitters on lanes not under test enabled and transmitting
43 QPRBS31-CEI test pattern as defined in Appendix 16.C.3.2, or a valid CEI signal, or
44 transmitting the same pattern with a slightly different Baud rate on each lane so that
45 lane to lane signals are asynchronous. These transmitters shall have identical transmit
46 equalizer settings to the transmitter under test.
47
48
49

356 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Compute the linear fit to the captured waveform and the linear fit pulse response, p(k), 1
and error, e(k), according to Section 17.3.1.6.1. Denote the standard deviation of e(k) 2
as σe. 3
4
With the QPRBS13-CEI pattern and the same configuration of the transmit equalizer, 5
measure the RMS deviation from the mean voltage at a fixed point in a run of at least 6 6
consecutive identical PAM4 symbols. The RMS deviation is measured for a run of each 7
of the four PAM4 symbol levels.The average of the four measurements is denoted as 8
σn. 9
10
SNDR is defined by Equation (17-7) where pmax is the maximum value of p(k). 11
12
13
2
 pmax  14
-  dB 
SNDR = 10 log  ---------------------- (17-7) 15
10   2 +  2 
e n 16
17
18
SNDR shall be greater than 31 dB for any allowable transmit equalizer setting. 19
20
17.3.1.7 Transmitter output jitter 21
22
Jitter measurements in this sub-clause are performed with transmitters on physical 23
lanes not under test enabled and transmitting QPRBS31-CEI test pattern as defined in 24
Appendix 16.C.3.2, or a valid CEI signal, or transmitting the same pattern with a slightly 25
different Baud rate on each lane so that lane to lane signals are asynchronous. These 26
transmitters shall have identical transmit equalizer settings to the transmitter under test. 27
28
J4u, JRMS, and EOJ are defined by measurements of 12 specific transitions in a 29
QPRBS13-CEI pattern in order to exclude correlated jitter. The 12 transitions represent 30
all possible combinations of four identical symbols followed by two different identical 31
symbols as shown in Table 17-5. The sequences are located by the symbol indices 32
given in the table where symbols 1 to 7 are the run of seven +1s. 33
34
The threshold used to define each transition is given in Table 17-5 where V-1, V-1/3, 35
V1/3, and V1 are as defined in Appendix 16.C.4.3. 36
37
The jitter is measured with a clock from a clock recovery unit (CRU) (i.e., a first order 38
golden PLL, with corner frequency at fb/6640, and a 20 dB/decade slope, see Section 39
1.6) as the trigger or reference clock. 40
41
J4u, JRMS, and EOJ specifications shall be met regardless of the transmit equalization 42
setting. 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 357
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 17-5. QPRBS13-CEI Pattern Symbols Used for Jitter Measurement

2
3 Label  Description Gray Coded PAM4  Index of 
First 
Index  Index 
Transition  Transition 
Index of 
Last  Threshold Level
4 Symbols Symbol Begins Ends Symbol
5 Reference 
6 REF for symbol  3333333 1 ‐‐ ‐‐ 7 ‐‐
7 index 
8 R03 0 to 3 rise  10000 330  1830 1834 1835 1837
(V‐1+V1)/2
9 F30  3 to 0 fall 23333 001 1269 1273 1274 1276
10
R12 1 to 2 rise 0111111 2222221 3638 3644 3645 3651
11 (V‐1/3+V1/3)/2
12 F21 2 to 1 fall 022222 113 1198 1203 1204 1206
13 R01 0 to 1 rise 100000 113  6835 6840 6841 6843
(V‐1+V‐1/3)/2
14 F10 1 to 0 fall 21111 003  2992 2996 2997 2999
15 R23 2 to 3 rise 32222 330  6824 6828 6829 6831
16 (V1/3+V1)/2
17 F32 3 to 2 fall 033333 2222223  7734 7739 7740 7746
18 R02 0 to 2 rise 10000 223  3266 3270 3271 3273
(V‐1+V1/3)/2
19 F20 2 to 0 fall 122222 0000002  7282 7287 7288 7294
20 R13 1 to 3 rise 011111 331  133 138 139 141
21 (V‐1/3+V1)/2
22 F31 3 to 1 fall 23333 112  7905 7909 7910 7912

23
24 17.3.1.7.1 J4u and JRMS Jitter
25
26 For each transition i, 1 ..i ..12, of the transitions specified in Table 17-5, obtain a set Si =
27 {ti(1), ti(2), ...} of transition times modulo the period of the pattern. The 12 sets should
28 be of equal size and the size of all sets should be chosen to enable calculation of J4u
29 (as defined below) with sufficient accuracy.
30
31 Calculate the average of each set Si, Tavgi, and subtract it from all elements of that set,
32 to create a set S0i = {ti(1) - Tavgi, ti(2) - Tavgi, ...}.
33
34 Combine the sets S0i, i=1 to 12, to create an estimated probability distribution fJ(t).
35 -4
36 J4u is defined as the time interval that includes all but 10 of fJ(t), from the 0.005th to
37 the 99.995th percentile of fJ(t).
38
39 JRMS is defined as the standard deviation of fJ(t).
40
41 17.3.1.7.2 Even-Odd Jitter (EOJ)
42
43 For one of the 12 specific transitions in QPRBS13-CEI in Table 17-5:
44
45 a) Trigger once in 3 repeats of the QPRBS13-CEI test pattern.
46
47 Obtain the mean time (T3) for this transition in the first QPRBS13-CEI.
48
49 Obtain the mean time (T4) for the same transition in the second QPRBS13-CEI.

358 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

b) The difference between the two means (T4 – T3), is the estimated period of the 1
repeating pattern. 2
3
For each of the 12 specific transitions in QPRBS13-CEI in Table 17-5: 4
5
1) Trigger once in 2 repeats of the QPRBS13-CEI test pattern. 6
7
Obtain the mean time (T1) for the specific transition in the first QPRBS13-CEI. 8
9
Obtain the mean time (T2) for the same transition in the second QPRBS13-CEI. 10
11
2) Calculate EOJ for this transition as |(T2 – T1) – (T4 – T3)|. 12
13
EOJ is the maximum of the 12 measurements. 14
NOTE 1—Both of (T2 - T1) and (T4 - T3) are about 8191 UI, which is much larger than 15
the EOJ value. Hence, each of T1 through T4 should have high precision. 16
17
18
17.3.2 Receiver Characteristics 19
20
A compliant receiver shall autonomously operate at the specified BER with the worst 21
case combination of a compliant transmitter and a compliant channel. The receiver also 22
shall not cause error propagation that violates the error burst length requirement as 23
defined in Section 17.2.1. Further receiver electrical requirements at compliance point 24
R (see Figure 17-1) are specified in Table 17-6, with the receiver interference tolerance 25
parameters specified in Table 17-7. Lanes not under test should be enabled and 26
transmitting or receiving asynchronous or uncorrelated signals. 27
28
Table 17-6. Receiver Electrical Input Specification
29
30
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
31
Baud Rate R_Baud 18.0 29.0 Gsym/s 32
Differential Input Return Loss R_SDD11 Equation (17-4) dB 33
Differential to Common Mode Input
34
Conversion
R_SCD11 Equation (17-8) dB 35
Interference Tolerance Table 17-7
36
37
Jitter Tolerance Table 17-8
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 359
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 17-7. Receiver interference tolerance parameters (Note 3)

2
3 Parameter Test 1 values Test 2 values Units
4 Min Max Min Max
5
6 Pre-FEC Bit Error Ratio (BER) 10–6 10–6
7 COM, including effects of broadband noise 3 3 dB
8
9 Insertion loss at Nyquist, Note 1 10 20 dB
10 RSS_DFE4, Note 2 0.025 - 0.05 -
11 NOTES:
12 1. Measured between TX and RX package balls (see Figure 17-1).
Definition can be found in Annex 93A of IEEE Std 802.3 [25] as modified by IEEE Std 802.3by [26] and IEEE
13 2. P802.3bs [27].
14 3. See Section 17.3.2.4
15
16
17 17.3.2.1 Input Baud Rate
18
19 All devices shall work within the range from 18.0 Gsym/s to 29.0 Gsym/s as specified
20 for the device, with the baud rate tolerance as per Section 3.2.11. A CEI
21 implementation complies to the specifications of this clause over the range of baud
22 rates stated for the implementation within this range.
23
24 17.3.2.2 Reference Input Signals
25
26 The receiver shall accept differential input signal amplitudes produced by a compliant
27 transmitter connected with the minimum attenuation specified in Figure 17-2 to the
28 receiver. This may be larger than the 1200 mVppd maximum of the transmitter due to
29 output/input impedances and reflections.
30
31 The minimum input amplitude is defined by the minimum transmitter amplitude, the
32 actual receiver input impedance and the loss of the actual PCB. Note that the minimum
33 transmitter amplitude is defined using a well controlled load impedance, however the
34 real receiver is not, which can leave the receiver input signal smaller than expected.
35
36 17.3.2.3 Receiver Input Return Loss
37
38 The differential input return loss, in dB, of the receiver shall meet Equation (17-4). The
39 reference impedance for differential return loss measurements shall be 100 Ω.
40
41 The differential to common-mode return loss, in dB, of the receiver shall meet Equation
42 (17-8). The differential to common-mode return loss limit RLdc(f) is shown in Figure 17-
43 7.
44
45  f  29
46  25 – 0.6897  -------------  0.05  f  0.5f b  
RL dc  f   RLmax  f  =  fb    dB 
47 (17-8)
 15  0.5f  f  f  
48 b b
49

360 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 17-7. Receiver differential to common-mode return loss limit for 29.0 Gsym/s. 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
  18
19
20
17.3.2.4 Receiver Interference Tolerance 21
22
The receiver interference tolerance test is based on the test defined in Annex 23
120D.3.2.1 of IEEE P802.3bs [27]. 24
25
The receiver on each lane shall meet the pre-FEC BER requirement with channels 26
matching the Channel Operating Margin (COM) and loss parameters for Test 1 and 27
Test 2 in Table 17-7. 28
29
The test channel should be created using printed circuit boards with short 30
interconnecting cables. 31
32
The following considerations apply to the interference tolerance test. The transmitter 33
package is omitted in the COM calculation. The test transmitter's measured SNDR 34
should be used for SNRTX in the COM calculation. The transmitter output levels are set 35
such that RLM is equal to 0.95. The test transmitter meets the specifications in Section 36
17.3.1. The test transmitter is constrained such that for any transmitter equalizer setting 37
the differential peak-to-peak voltage is less than 800 mV, and the normalized 38
amplitudes of the coefficients of the transmitter equalizer c(-1), c(0) and c(1) are 39
between the minimum and maximum limits given in Table 17-4. 40
41
The lower frequency bound for the noise spectral density constraints, fNSD1, is 1 GHz.
42
The differential return-loss of the test channel at TP5 (as defined in Annex 93A of Std. 43
IEEE802.3 [25]) shall meet the requirements of Equation (17-4), and be 3.0 dB better
44
than the requirements of Equation (17-4) for all frequencies less than fb/2. The test
45
transmitter's jitter parameters J4u and JRMS are measured. ADD and σRJ are calculated
46
from the measured values of J4u and JRMS using Equation (17-9), and Equation (17-
47
10), respectively and used for COM parameters. Other COM parameters are set
48
according to the values in Table 17-1. The broadband noise is added and adjusted to
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 361
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 achieve the COM value in Table 17-7. The test pattern to be used is QPRBS31-CEI
2 defined in Appendix 16.C.3.2. A test system with a fourth-order Bessel-Thomson low-
3 pass response with 40 GHz 3 dB bandwidth is to be used for measurement of the
4 signal applied by the pattern generator and for measurements of the broadband noise.
5
6
7 J 4u J 4u 2 
A DD =   ------
-  2
+ Q4  Q4 + 1   J RMS – 2  ------
- 2
  Q4 + 1  (17-9)
8  2    2  
9
10
11
12 J 4u
13  RJ =  ------- – A DD   Q4  (17-10)
14 2
15
16 17.3.2.5 Receiver Jitter Tolerance
17
18 Receiver jitter tolerance shall meet the conditions and parameters defined in Table 17-
19 8. This sinusoidal jitter is part of the jitter applied in the stressed input test. The
20 sinusoidal jitter is calibrated at 10x the reference CRU’s bandwidth and must be tested
21 at fCRU/100, fCRU/3, fCRU, 3fCRU, and 10fCRU, where fCRU is the jitter corner frequency
22 given by fb/6640, with sinusoidal jitter of 5 UI, 0.15 UI, 0.05 UI, 0.05 UI and 0.05 UI
23 respectively. For this test the channel used is as for the receiver interference tolerance
24 described in Section 17.3.2.4. Note that the values measured for J4u and JRMS include
25 the effects of this added sinusoidal jitter and noise is added to obtain a COM of 3 dB
26 with these measured jitter values as for the interference tolerance test. The receiver bit
27 error ratio (BER) shall meet the requirements of Section 17.2.2 for each pair of jitter
28 frequency and peak-to-peak amplitude values listed above and shown in Figure 17-8.
29
30 Table 17-8. Receiver Jitter Tolerance Parameters
31
32 Sinusoidal jitter,
33 Frequency Range peak-to-peak
34 (UI)
35
36 f < fb/664000 Not Specified
37 fb/664000 < f < fb/6640 5*fb/(664000*f)
38
fb/6640 < f < 10fCRU 0.05
39
40
41
42
43
44
45
46
47
48
49

362 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 17-8. Receiver Jitter Tolerance Mask 1

2
3
5 UIpp X 4
5
Sinusoidal
6
jitter
amplitude
7
8
X
9
10
11
0.05 UIpp X X X 12
13
14
15
fb/664000 fb/6640 10fCRU 16
17
18
17.3.2.6 Single Ended Input Voltage 19
20
The single ended voltage levels with respect to the receiver ground at the input of the 21
receiver are dependent on the transmitter implementation and the inter-ground 22
difference. The voltage levels at the input of an AC coupled receiver (if the effective AC 23
coupling is done within the receiver) or at the TX side of the external AC coupling cap (if 24
AC coupling is done externally) will be between -0.35V and 1.95V with respect to local 25
ground. 26
27
17.3.2.7 Input Lane-to-Lane Skew 28
29
Refer to Section 3.2.8. 30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface 363
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

364 Optical Internetworking Forum - Clause 17: CEI-56G-MR-PAM4 Medium Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18 CEI-56G-USR-NRZ Ultra Short Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-USR-NRZ ultra short-reach high 4
speed electrical interface between nominal baud rates of 19.6 Gsym/s to 58.0 Gsym/s 5
using NRZ coding. A compliant device must meet all of the requirements listed below.
6
The electrical interface is based on high speed, low voltage DC-coupled logic with 7
nominal differential impedance of 100  and a channel with a nominal differential 8
impedance of 92.5 . Connections are point-to-point balanced differential pair and 9
signaling is unidirectional. 10
The electrical IA is based on loss & jitter budgets and defines the characteristics 11
required to communicate between a CEI-56G-USR-NRZ driver and a  12
CEI-56G-USR-NRZ receiver using signal traces on a package substrate. The 13
characteristic impedance of the signal traces is nominally 92.5  differential. A ‘length’ 14
is effectively defined in terms of its attenuation rather than physical length. 15
16
Ultra short reach CEI-56G-USR-NRZ devices from different manufacturers shall be 17
interoperable. 18
19
20
21
22
18.1 Requirements 23
24
1. Support serial data rate from 19.6 Gsym/s to 58.0 Gsym/s. A CEI implementation 25
complies to the specifications of this clause over the range of baud rates stated by 26
the implementer within this range. 27
28
2. Capable of low bit error rate (10-15, with a test requirement to verify 10-12). 29
3. Capable of driving 0 – 10 mm of package substrate 30
31
4. Shall support DC-coupled operation. 32
33
5. Shall allow multi-lanes (1 to n). 34
35
36
18.2 General Requirements 37
38
18.2.1 Data Patterns 39
40
Please refer to Section 3.2.1. 41
42
18.2.2 Signal levels 43
44
The signal is a low swing DC coupled differential interface. 45
46
47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 365
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 18.2.3 Signal Definitions

2
3 Please refer to Appendix 1.A.
4
5 A single clock signal is shared between the transmitting and receiving devices in the
6 ingress direction and another single clock signal is shared between the transmitting
7 and receiving devices in the egress direction, avoiding the need for a clock recovery
8 circuit at the receiver.
9
10 18.2.4 Bit Error Ratio
11
12 Please refer to Section 3.2.3.
13
14
15 18.2.5 Ground Differences
16
As the driver and receiver are on the package substrate (with no intervening
17
connectors), the ground difference is approximately 0 mV.
18
19
20 18.2.6 Cross Talk
21
22 Please refer to Section 3.2.5.
23
24 18.2.7 Channel Compliance
25
26 A forward channel and associated dominant crosstalk channels are deemed compliant
27 if the channel characteristics conform to the requirements in this section.
28
29
18.2.7.1 Reference Model
30
31 The channel consists of a package substrate trace and any required vias. The
32 reference package substrate trace differential impedance is 92.5 and all channel
33 specifications are referenced to this impedance
34
35 Figure 18-1 shows a diagram of the intended application.
36
37 Figure 18-1. CEI-56G-USR-NRZ Reference Model
38
39
40
41
42
43
44
45
46
47
48
49

366 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Several channel characteristics are parametrized according to Table 18-1 at these test 1
points and are used for calculation of the channel parameters found in Table 18-2. 2
3
4
5
Table 18-1. Measured Channel Parameters
6
Symbol Description 7
IL(f) Differential insertion loss, -SDD21 magnitude (dB) 8
RL1(f) Differential input return loss, -SDD11 magnitude (dB)
9
10
RL2(f) Differential output return loss, -SDD22 magnitude (dB)
11
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB) 12
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 13
14
15
16
Table 18-2. Calculated Channel Parameters
17
Symbol Description 18
ILfitted(f) Fitted insertion loss (dB) 19
ILD(f) Insertion loss deviation (dB)
20
21
ICN(f) Integrated crosstalk noise (mVRMS)
22
FOMILD Channel Figure of merit - Weighted insertion loss deviation (dB) 23
24
18.2.7.2 Insertion Loss 25
26
Channel insertion losses, including package substrate traces and vias, shall comply 27
with the limits specified by Equation (18-1) and plotted in Figure 18-2. Note that the 28
variable fb is the maximum baud rate to be supported by the channel under test  29
(19.6 GHz < fb < 58.0 GHz) and the measurement is to be made relative to the nominal 30
impedance of 92.5 . 31
32
Table 18-3. Channel Insertion Loss Frequency Range 33
Parameter Value Units 34
35
fmin 50 MHz
36
fmax fb GHz
37
38
39
 fb  40
f  580-
f  580- + 0 01 --------------------
 0095 + 03 --------------------  f min  f  ----  41
 fb fb 2 42
IL max =   (18-1)
 f  580 fb  43
 – 3798 + 02 --------------------- ----  f  f b  44
 fb 2  45
46
Note: f in Equation (18-1) is in GHz. 47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 367
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 18-2. CEI-56G-USR-NRZ normative channel insertion loss at 58.0 Gsym/s

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 18.2.7.3 Fitted Insertion Loss
30
31 For fitted insertion loss definitions, please refer to Section 12.2.1.1.
32
33 The channel shall meet the insertion loss requirements defined in Table 18-4 and also
34 meet the IL mask in Equation (18-1). Note that the variable fb is the maximum baud rate
35 to be supported by the channel under test.
36 Table 18-4. Channel fitted insertion loss characteristics
37
38 Parameter Units Min Value Max Value
39 Min frequency, fILmin GHz 0.05 -
40 Max frequency, fILmax GHz - fb
41
Fitted insertion loss at Nyquist dB - 2
42
Fitted insertion loss, a0 dB -1 2
43
44 Fitted insertion loss, a1 dB 0 -
45 Fitted insertion loss, a2 dB 0 -
46 Fitted insertion loss, a4 dB 0 -
47
48
49

368 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.2.7.4 Insertion Loss Deviation (ILD) 1

2
The insertion loss deviation ILD is the difference between the measured insertion loss 3
IL and the fitted insertion loss ILfitted as defined in Equation (18-2) where fILmin and 4
fILmax are given in Table 18-4. 5
6
ILD = IL – IL fitted f ILmin  f   3  4   f ILmax (18-2) 7
8
The insertion loss deviation ILD shall be within the region defined by Equation (18-3). 9
10
– 10  I LD  10 (18-3) 11
12
13
A Figure Of Merit (FOMILD) for the channel is the weighted insertion loss deviation from 14
fILmin to (3/4)*fILmax. FOMILD is calculated as indicated below. Note FOMILD is called 15
ILDRMS in OIF-CEI-03.0 clauses 10 & 11 and OIF-CEI-03.1 clauses 10, 11, 13 & 14. 16
17
Define the weight at each frequency f using Equation (18-4) below. 18
2 1 1 19
W  f  = sin c  f  f b  ------------------------4- -------------------------
-
8
(18-4) 20
1 +  f  ft  1 +  f  fr  21
22
Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum 20% 23
to 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365 (i.e. 24
min(T_tr, T_tf) x ft = 0.2365, T_tr is in ns when ft is in GHz). In addition, fr is the -3 dB 25
reference receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum 26
baud rate to be supported by the channel. 27
28
FOMILD is calculated using Equation (18-5) where N is the number of frequency points. 29
The summation is done over the frequency range of ILD with f in GHz. FOMILD shall be 30
less than 0.2 dB for compliant channels. 31
2
32
33
FOM ILD =
 W  f   I LD  f
--------------------------------------------- (18-5) 34
N 35
36
18.2.7.5 Channel Return Loss 37
38
Channel Return Loss shall be bounded by Equation (18-6) as shown in Figure 18-3 39
relative to the nominal impedance of 92.5 . 40
fb 41
RL  f   12 dB f min  f  ---- 42
4
(18-6) 43
 4f  fb 44
RL  f   12 dB – 15 log  -----  ----  f  f b
10 f b 4 45
46
Note: fmin is as defined in Table 18-3 47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 369
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 18-3. CEI-56G-USR-NRZ normative channel return loss at 58.0 Gsym/s

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 18.2.7.6 Channel integrated crosstalk noise
23
24 Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of
25 Table 18-5, the total integrated crosstalk noise for the channel shall be less than 
26 1 mVRMS.
27
Table 18-5. Channel integrated crosstalk aggressor parameters
28
29 Parameter Symbol Value Units
30 Baud rate fb max. Baud Rate supported by channel Gsym/s
31
Near-end aggressor peak to peak differential output amplitude Ant 250 mVppd
32
33 Far-end aggressor peak to peak differential output amplitude Aft 250 mVppd

34 Near-end aggressor 20% to 80% rise and fall times Tnt 4 ps

35 Far-end aggressor 20% to 80% rise and fall times Tft 4 ps
36
37
38
39
40
41
42
43
44
45
46
47
48
49

370 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.3 Electrical Characteristics 1

2
The electrical interface is based on high speed, low voltage DC-coupled logic with 3
nominal differential impedance of 100  and a channel with a nominal differential 4
impedance of 92.5 . 5
6
All devices shall work within the range 19.6 Gsym/s to 58.0 Gsym/s as specified for the 7
device, with all ingress lanes synchronous to a common reference frequency having a 8
stability of  100ppm from nominal and all egress lanes synchronous to a common 9
reference frequency having a stability of  100ppm from nominal. The reference clocks 10
of the ingress and egress directions are not necessarily synchronous to each other. 11
Note that implementation of specific protocols will define the operating baud rate 12
without affecting CEI compliance. 13
14
15
18.3.1 Reference Clock 16
Both ends of the link are to have a common clock frequency, which is set to be 1/64th of 17
the baud rate. This clock could come from the same source or could be “forwarded” 18
from the driver side to the receiver side. For details of applicable clock architecture 19
options please refer to Annex 18.A. In a forwarded clock architecture the clock path 20
needs to meet the same requirements as the data path. The electrical specifications at 21
RC are given in Table 18-6. 22
23
Table 18-6. Reference Clock Electrical Specification. 24
25
Characteristic Symbol Condition MIN. NOM. MAX. UNIT 26
Reference Clock Rate Ref_Freq fb/64 GHz
27
28
Input Differential Voltage Ref_Vdiff 240 900 mVppd
29
Input Single Ended Voltage Ref_Vse 0.1 1.2 V 30
Differential Termination Resistance Mismatch Ref_Rdm 5 % 31
Input Clock Rise and Fall Time (20% to 80%) Ref_tr, Ref_tf 4 200 ps 32
at fb/128  f < fb/4 33
Differential Input Return Loss Ref_SDD11 -6 dB
Note 1 34
Input Clock Duty Cycle Ref_DC 40 60 % 35
High Frequency Uncorrelated Unbounded 36
Ref_UUGJ_hf Note 2 0.009 UI rms
Gaussian Jitter 37
@ 1kHz offset -70 38
@10kHz offset -93 39
Reference Clock Single Side Band Phase Noise Ref_PN @100kHz offset -113 dBc/Hz
40
41
@1MHz offset -133
42
10MHz offset -143 43
NOTES: 44
1. Return loss is referenced to 100 Ohm
2. UUGJ measured using the methodology defined in Appendix 2.E.1 with a golden PLL cut-off frequency of fb/500, observing 45
with a bandwidth of 43GHz 46
47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 371
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 18.3.2 Transmitter Characteristics

2
3 The transmitter electrical specifications at compliance point T (see Figure 18-1) are
4 given in Table 18-7. The transmitter shall satisfy jitter requirements specified in Table
5 18-8. Jitter is measured relative to the reference clock as timing source, using a golden
6 clock multiplier as detailed in Annex 18.A.4, for a BER as specified in Section 3.2.3. To
7 attenuate noise and absorb even/odd mode reflections, the transmitter shall satisfy the
8 Common Mode Output Return Loss requirement of Table 18-7. The waveform is
9 observed through a fourth-order Bessel-Thomson response with a bandwidth of 
10 43 GHz using a PRBS31 pattern.
11
12 The link budget in this document assumes no Tx emphasis.
13
Table 18-7. Transmitter Electrical Output Specification.
14
15 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
16
17 Baud Rate T_Baud 19.6 58.0 Gsym/s

18 Output Differential Voltage T_Vdiff 100 250 mVppd

19 Single Ended Transmitter Output Voltage Note 1 T_Vse 0.1 1.2 V
20 Differential Termination Resistance Mismatch T_Rdm 5 %
21
Output Rise and Fall Time (20% to 80%) T_tr, T_tf 4 ps
22
See 18.3.2.2
23 Differential Output Return Loss T_SDD22
Note 2
dB
24 Common mode Output Return Loss T_SCC22 500MHz to fb -6 dB
25
26 Transmitter Common Mode Noise T_Ncm 15 mVrms

27 NOTES:
1. DC Coupling compliance is mandatory.
28 2. Return loss is referenced to 100 Ohm
29
30 Table 18-8. Transmitter Output Jitter Specification
31
32 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
33 Uncorrelated Unbounded Gaussian Jitter T_UUGJ 0.24 UIpp
34 Uncorrelated Bounded High Probability Jitter T_UBHPJ 0.25 UIpp
35 Even Odd jitter (component of UBHPJ) T_EOJ Note 2 0.035 UIpp
36
37 Total Jitter T_TJ Note 1 0.44 UI

38 Eye Mask T_X1 See 18.3.2.4 0.22 UI

39 Eye Mask T_X2 See 18.3.2.4 0.4 UI

40 Eye Mask T_Y1 See 18.3.2.4 50 mV
41 Eye Mask T_Y2 See 18.3.2.4 125 mV
42 NOTES:
43 1. T_TJ includes all of the jitter components measured without any transmit equalization.
44 2. Included in T_UBHPJ. Even-odd jitter is defined in Table 1-3.
45
46
47
48
49

372 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.3.2.1 Transmitter Amplitude and Swing 1

2
Transmitter differential output amplitude shall be able to drive between 100 and 3
250 mVppd. Further the single-ended voltage must be between 0.1 and 1.2 V. Figure 4
18-4 shows the transmitter test load configuration. 5
6
Figure 18-4. Transmitter Amplitude Test Load
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
18.3.2.2 Transmitter Resistance and Return Loss 24
25
Please refer to Section 3.2.10 with the following parameters in Table 18-9 and as 26
illustrated in Figure 18-5. 27
28
Table 18-9. Driver Return Loss Parameters 29
Parameter Value Units
30
31
A0 -12 dB 32
f0 100 MHz
33
34
f1 0.1714 x T_Baud Hz 35
f2 T_Baud Hz 36
37
Slope 12.0 dB/dec 38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 373
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 18-5. Illustration of Return Loss for T_Baud = 58.0 Gsym/s

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 18.3.2.3 Transmitter Lane-to-Lane Skew
22
23 Please refer to Section 3.2.7.
24
25 18.3.2.4 Transmitter Template and Jitter
26
27 When provided with the reference clock meeting the specifications in Section 18.3.1 as
28 a timing source, using a golden clock multiplier as detailed in Annex 18.A.4, for a BER
29 as per Section 3.2.3, the transmitter shall satisfy the eye template and jitter
30 requirements as given in Table 18-8 and Figure 1-4. The measurement of jitter and eye
31 diagram is to be relative to the "forwarded" clock if the transmitter provides one. If it
32 does not provide a "forwarded" clock then the measurement is to be relative to the
33 reference clock provided to it.
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

374 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.3.3 Receiver Characteristics 1

2
A compliant receiver shall operate at the specified BER with the worst case 3
combination of a compliant transmitter and a compliant channel. 4
5
Receiver electrical specifications are given in Table 18-10 and measured at compliance 6
point R. To dampen noise sources and absorption of both even and odd mode 7
reflections, the receiver shall satisfy the Common Mode Input Return Loss requirement 8
of Table 18-10. 9
10
The link budget in this document assumes no Rx equalization. 11
Table 18-10. Receiver Electrical Input Specification
12
13
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
14
15
Baud Rate R_Baud 19.6 58.0 Gsym/s
16
Input Differential Voltage R_Vdiff Note 1 250 mVppd 17
Receiver Common Mode Noise Tolerance R_Ncm 25 mVrms 18
Differential Termination Resistance Mismatch R_Rm 5 % 19
See 18.3.3.2 20
Differential Input Return Loss R_SDD11 dB
Note 2 21
200 MHz to 10
-6 dB
22
Common mode Return Loss R_SCC11 GHz
23
10 GHz to fb -4 dB 24
Rx Input Single Ended Voltage Range R_VinSE 0.1 1.2 V 25
NOTES: 26
1. The receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the combined 27
transmitter and channel. The channel response shall include the worst case effects of the return losses at the transmitter and
receiver 28
2. Return loss is referenced to 100 Ohm 29
30
31
32
When provided with the reference clock meeting the specifications in Section 18.3.1 as 33
a timing source, the receiver shall tolerate the sum of these jitter contributions and 34
meet the BER as per Section 3.2.3: the total transmitter jitter from Table 18-8 and the 35
effects of a channel compliant to the Channel Characteristics (Section 18.2.7). 36
37
18.3.3.1 Reference Receiver Input Signals 38
39
The receiver shall accept differential input signal amplitudes produced by a compliant 40
transmitter connected directly to the receiver. This may be larger than the 250 mVppd 41
maximum of the transmitter due to output/input impedances and reflections. 42
43
The minimum input amplitude is defined by the minimum transmitter amplitude, the 44
actual receiver input impedance and the loss of the actual package substrate. Note that 45
the minimum transmitter amplitude is defined using a well controlled load impedance; 46
however the real receiver's impedance may differ, which can leave the receiver input 47
signal smaller than expected. Additionally the real receiver may be affected by 48
environmental noise. 49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 375
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 18.3.3.2 Receiver Return Loss

2
3 Please refer to Section 3.2.10 with the following parameters in Table 18-11 and as
4 illustrated in Figure 18-5.
5
Table 18-11. Receiver Input Return Loss Parameters
6
7 Parameter Value Units
8 A0 -12 dB
9 f0 100 MHz
10
f1 0.1714 x R_Baud Hz
11
12 f2 R_Baud Hz

13 Slope 12.0 dB/dec

14
15 18.3.3.3 Input Lane-to-Lane Skew
16
17 Please refer to Section 3.2.8.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

376 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

18.A Annex - Clock Architecture Options 1

2
This annex clarifies the applicability of reference clock specifications for clauses that 3
allow forwarded clock and common clock architectures. The electrical requirements 4
specified in these clauses support several different architectures for clock distribution 5
and data recovery. Devices shall be interoperable regardless of the clock distribution 6
method used by the system or the data recovery method used by the receiver device. 7
The intent is to allow low power receiver architectures which do not use Clock Data 8
Recovery (CDR) circuits to sample data. 9
10
11
18.A.1 Forwarded Clock Architecture 12
13
Receiver devices designed to support forwarded clock architectures use a clock from 14
the transmitter and a clock multiplier circuit to synchronously sample data in the 15
receiver. The device interconnections for a forwarded clock architecture are shown in 16
Figure 18-6. 17
Figure 18-6. Forwarded Clock Architecture 18
19
20
Clock 21
Generator
22
23
24
TxRefClk_N
TxRefClk_P

25
26
27
28
T R 29
TxData0_P RxData0_P 30
TxData0_N RxData0_N 31
: :
TxDatan_P RxDatan_P
32
TxDatan_N RxDatan_N 33
34
RefClkOut_P RxRefClk_P 35
RefClkOut_N RxRefClk_N 36
TC RC 37
38
39
40
41
The signals at reference point RC that are connected to the RxRefClk input shall meet 42
the requirements defined for the Reference Clock in the applicable clause. 43
44
The TxData signals in the figure at reference point T shall meet the requirements in the 45
Transmitter Characteristics section of the applicable clause. The transmitter shall 46
satisfy the jitter requirements defined in the Transmitter Characteristics section of the 47
48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 377
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 applicable clause when the signals at reference point T are measured using the signal
2 at reference point RC as a timing source. The signals at RC are connected through a
3 Golden Clock Multiplier to make these measurements as described in Annex 18.A.4.
4
5 Requirements for the TxRefClk input signal from the clock generator are determined by
6 the device manufacturer. Device interoperability does not require the imposition of any
7 specific requirements on this signal.
8
9
18.A.2 Common Clock Architecture
10
11 Devices connected in a common clock architecture configuration drive the same clock
12 generator signal to both the Transmitter and the Receiver devices. The Receiver uses
13 this clock and a clock multiplier circuit to synchronously sample data in the receiver.
14 The device interconnections for a common clock architecture are shown in Figure 18-7.
15
16 Figure 18-7. Common Clock Architecture
17
18
19 Clock
20 Generator
21 TC
22
TxRefClk_N
TxRefClk_P

23
24
25
26
27 T R
28 TxData0_P RxData0_P
29 TxData0_N RxData0_N
30 : :
31 TxDatan_P RxDatan_P
32 TxData n_N RxDatan_N
33 RefClkOut_P RxRefClk_P
34 RefClkOut_N RxRefClk_N
35
36 RC
37
38
39
40 The signals at reference point RC that are connected to the RxRefClk input shall meet
41 the requirements defined for the Reference Clock in the applicable clause. Since
42 transmitter and receiver share a common clock source, it is implied that the signal at TC
43 meets the similar requirements as the signal at RC, although this is not required for
44 interoperability.
45
46 The TxData signals in the figure at reference point T shall meet the requirements in the
47 Transmitter Characteristics section of the applicable clause. The transmitter shall
48 satisfy the jitter requirements defined in the Transmitter Characteristics section of the
49

378 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

applicable clause when the signals at reference point T are measured using the signal 1
at reference point RC as a timing source. The signals at RC are connected through a 2
Golden Clock Multiplier to make these measurements as described in Annex 18.A.4. 3
4
5
18.A.3 Receiver Clock Data Recovery (CDR) Circuits
6
While the intent is to support receiver architectures that do not use CDR circuits, there 7
is no restriction that prevents receiver implementations from using CDR circuits to 8
sample data. Such receivers are connected using the Common Clock Architecture 9
shown in Figure 18-7. To ensure interoperability with receiver devices that do not use 10
CDR circuits, the reference clock, transmitter, and receiver requirements are applied in 11
the same manner as was described in Annex 18.A.2. 12
13
14
18.A.4 Golden Clock Multiplier 15
16
Clauses which allow a forwarded clock architecture specify that transmitter jitter is 17
measured relative to the reference clock as a timing source. This section specifies the 18
relevant measurement methodology. 19
20
Transmitter jitter for applicable clauses is measured as described in Figure 18-8. The 21
electrical characteristics of the reference clock at RC are specified by the applicable 22
clause. A Golden Clock Multiplier is used to multiply this clock to produce a full-rate 23
sample clock, and this sample clock is used to sample the transmitter data output at T. 24
Figure 18-8. Transmitter Measurements Using a Golden Clock Multiplier 25
26
27
28
29
30
T 31
TxDataN_P Scope or
TxDataN_N
32
BERT
33
34
Trigger
DUT 35
36
37
38
RefClkOut_P 39
RefClkOut_N 40
TC RC Golden Clock Multiplier
41
42
43
44
45
The Golden Clock Multiplier shown in the figure is a PLL or DLL that shall have a cutoff 46
frequency equal to the frequency of the reference clock divided by 6, a 20 dB/decade 47
rolloff, and has no peaking around the corner frequency. 48
49

Optical Internetworking Forum - Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface 379
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

380 Clause 18: CEI-56G-USR-NRZ Ultra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

19 CEI-56G-XSR-NRZ Extra Short Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-XSR-NRZ extra short reach high 4
speed electrical interface between nominal baud rates of 39.8 Gsym/s to 58.0 Gsym/s 5
using NRZ coding. A compliant device must meet all of the requirements listed below.
6
The electrical interface is based on high speed, low voltage DC-coupled logic with 7
nominal differential impedance of 100  and a channel with a nominal differential 8
impedance of 92.5 . Connections are point-to-point balanced differential pair and 9
signalling is unidirectional. 10
The electrical IA is based on loss & jitter budgets and defines the characteristics 11
required to communicate between a CEI-56G-XSR-NRZ driver and a 12
CEI-56G-XSR-NRZ receiver using copper signal traces on a printed circuit board. The 13
characteristic impedance of the signal traces is nominally 92.5  differential. A ‘length’ 14
is effectively defined in terms of its attenuation rather than physical length. 15
16
Extra short reach CEI-56G-XSR-NRZ devices from different manufacturers shall be 17
interoperable. 18
19
20
21
22
19.1 Requirements 23
24
1. Support serial data rate from 39.8 Gsym/s to 58.0 Gsym/s. A CEI implementation 25
complies to the specifications of this clause over the range of baud rates stated by 26
the implementer within this range. 27
28
2. Capable of low bit error rate (10-15, with a test requirement to verify 10-12). 29
3. Capable of driving 0 – 50 mm of PCB trace. 30
31
4. Shall support DC-coupled operation. 32
33
5. Shall allow multi-lanes (1 to n). 34
35
36
19.2 General Requirements 37
38
19.2.1 Data Patterns 39
40
Please refer to Section 3.2.1. 41
42
19.2.2 Signal levels 43
44
The signal is a low swing DC coupled differential interface. 45
46
47
48
49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 381
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 19.2.3 Signal Definitions

2
3 Please refer to Appendix 1.A.
4
5 A single clock signal is shared between the transmitting and receiving devices in the
6 ingress direction and another single clock signal is shared between the transmitting
7 and receiving devices in the egress direction, avoiding the need for a clock recovery
8 circuit at the receiver.
9
10 19.2.4 Bit Error Ratio
11
12 Please refer to Section 3.2.3.
13
14
15 19.2.5 Ground Differences
16
As the driver and receiver are on the same PCB (with no intervening connectors), the
17
ground difference is approximately 0 mV.
18
19
20 19.2.6 Cross Talk
21
22 Please refer to Section 3.2.5.
23
24 19.2.7 Channel Compliance
25
26 A forward channel and associated dominant crosstalk channels are deemed compliant
27 if the channel characteristics conform to the requirements in this section.
28
29
19.2.7.1 Reference Model
30
31 The channel consists of PCB traces and any required vias. The reference PCB trace
32 differential impedance is 92.5 and all channel specifications are referenced to this
33 impedance
34
35 Figure 19-1 shows a diagram of the intended application.
36
37 Figure 19-1. CEI-56G-XSR-NRZ Reference Model
38
39
40
41
42
43
44
45
46
47
48
49

382 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Several channel characteristics are parametrized according to Table 19-1 at these test 1
points and are used for calculation of the channel parameters found in Table 19-2. 2
3
Table 19-1. Measured Channel Parameters
4
Symbol Description 5
IL(f) Differential insertion loss, -SDD21 magnitude (dB)
6
7
RL1(f) Differential input return loss, -SDD11 magnitude (dB)
8
RL2(f) Differential output return loss, -SDD22 magnitude (dB)
9
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB) 10
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 11
12
Table 19-2. Calculated Channel Parameters 13
Symbol Description 14
ILfitted(f) Fitted insertion loss (dB)
15
16
ILD(f) Insertion loss deviation (dB)
17
ICN(f) Integrated crosstalk noise (mVRMS) 18
FOMILD Channel Figure of merit - Weighted insertion loss deviation (dB) 19
20
19.2.7.2 Insertion Loss 21
22
Channel insertion losses, including PCB traces and connectors, shall comply with the 23
limits specified by Equation (19-1), Equation (19-2) and plotted in Figure 19-2. Note 24
that the variable fb is the maximum baud rate to be supported by the channel under test  25
(39.8 GHz < fb < 58.0 GHz) and the measurement is to be made relative to the nominal 26
impedance of 92.5 . 27
28
Table 19-3. Channel Insertion Loss Frequency Range 29
Parameter Value Units
30
31
fmin 50 MHz
32
fmax fb GHz 33
34
35
 fb  36
f 
 0 01 + 1 0 -------------------- 58 0 f  58 0
 - f  f  ---- 
 - + 009 --------------------
   fb fb min 2 37
IL max =   (19-1) 38
 f  580 f  39
 – 3595 + 04 -------------------- - ---b-  f  f b 
 f b 2  40
41
42
43
  44
IL min =  0 009 f + 0 015 f f  f  f  (19-2) 45
   min b
46
47
48
Note: f in Equation (19-1) and Equation (19-2) is in GHz. 49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 383
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 Figure 19-2. CEI-56G-XSR-NRZ normative channel insertion loss at 58.0 Gsym/s
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33 19.2.7.3 Fitted Insertion Loss
34
35 For fitted insertion loss definitions, please refer to Section 12.2.1.1
36
37 The channel shall meet the insertion loss requirements defined in Table 19-4 and also
38 meet the IL mask in Equation (19-1) and Equation (19-2). Note that the variable fb is the
39 maximum baud rate to be supported by the channel under test.
40
41 Table 19-4. Channel fitted insertion loss characteristics
42 Parameter Units Min Value Max Value
43 Min frequency, fILmin GHz 0.05 -
44
Max frequency, fILmax GHz - fb
45
46 Fitted insertion loss at Nyquist dB - 8

47 Fitted insertion loss, a0 dB -1 2

48
49

384 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Table 19-4. Channel fitted insertion loss characteristics 1

Parameter Units Min Value Max Value
2
3
Fitted insertion loss, a1 dB 0 5
4
Fitted insertion loss, a2 dB 0 16 5
Fitted insertion loss, a4 dB 0 7.3 6
7
19.2.7.4 Insertion Loss Deviation (ILD) 8
9
The insertion loss deviation ILD is the difference between the measured insertion loss 10
IL and the fitted insertion loss ILfitted as defined in Equation (19-3) where fILmin and 11
fILmax are given in Table 19-4. 12
13
ILD = IL – IL fitted f ILmin  f   3  4   f ILmax (19-3) 14
15
16
The insertion loss deviation ILD shall be within the region defined by Equation (19-4). 17
18
– 15  I LD  15 (19-4) 19
20
A Figure Of Merit (FOMILD) for the channel is the weighted insertion loss deviation from 21
fILmin to (3/4)*fILmax. FOMILD is calculated as indicated below. Note FOMILD is called 22
ILDRMS in OIF-CEI-03.0 clauses 10 & 11 and OIF-CEI-03.1 clauses 10, 11, 13 & 14. 23
24
Define the weight at each frequency f using Equation (19-5) below. 25
26
2 1 1 27
W  f  = sin c  f  f b  ------------------------4- -------------------------
-
8
(19-5)
1 +  f  ft  1 +  f  fr  28
29
30
Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum 20%
31
to 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365 (i.e.
32
min(T_tr, T_tf) x ft = 0.2365, T_tr is in ns when ft is in GHz). In addition, fr is the -3 dB
33
reference receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum
34
baud rate to be supported by the channel.
35
FOMILD is calculated using Equation (19-6) where N is the number of frequency points. 36
The summation is done over the frequency range of ILD with f in GHz. FOMILD shall be 37
less than 0.2 dB for compliant channels. 38
39
2 40
FOM ILD =
 W  f   I LD  f
--------------------------------------------- (19-6)
41
N 42
43
44
19.2.7.5 Channel Return Loss
45
Channel Return Loss shall be bounded by Equation (19-7) as shown in Figure 19-3 46
relative to the nominal impedance of 92.5 . 47
48
49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 385
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 fb
2 RL  f   12 dB f min  f  ----
4
3 (19-7)
4  4f  fb
RL  f   12 dB – 15 log  -----  ----  f  f b
5 10 f b 4
6
7 Note: fmin is as defined in Table 19-3
8
9 Figure 19-3. CEI-56G-XSR-NRZ normative channel return loss at 58.0 Gsym/s
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
19.2.7.6 Channel integrated crosstalk noise
32
33 Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of
34 Table 19-5, the total integrated crosstalk noise for the channel shall be less than 
35 2 mV
36 RMS.
37 Table 19-5. Channel integrated crosstalk aggressor parameters
38 Parameter Symbol Value Units
39
40 Baud rate fb max. Baud Rate supported by channel Gsym/s

41 Near-end aggressor peak to peak differential output amplitude Ant 400 mVppd
42 Far-end aggressor peak to peak differential output amplitude Aft 400 mVppd
43 Near-end aggressor 20% to 80% rise and fall times Tnt 4 ps
44 Far-end aggressor 20% to 80% rise and fall times Tft 4 ps
45
46
47
48
49

386 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

19.3 Electrical Characteristics 1

2
The electrical interface is based on high speed, low voltage DC-coupled logic with 3
nominal differential impedance of 100  and a channel with a nominal differential 4
impedance of 92.5 . 5
6
All devices shall work within the range 39.8 Gsym/s to 58.0 Gsym/s as specified for the 7
device, with all ingress lanes synchronous to a common reference frequency having a 8
stability of  100ppm from nominal and all egress lanes synchronous to a common 9
reference frequency having a stability of  100ppm from nominal. The reference clocks 10
of the ingress and egress directions are not necessarily synchronous to each other. 11
Note that implementation of specific protocols will define the operating baud rate 12
without affecting CEI compliance. 13
14
15
19.3.1 Reference Clock 16
Both ends of the link are to have a common clock frequency, which is set to be 1/64th of 17
the baud rate. This clock could come from the same source or could be “forwarded” 18
from the driver side to the receiver side. For details of applicable clock architecture 19
options please refer to Annex 18.A.In a forwarded clock architecture the clock path 20
needs to meet the same requirements as the data path. The electrical specifications at 21
RC are given in Table 19-6. 22
23
Table 19-6. Reference Clock Electrical Specification. 24
25
Characteristic Symbol Condition MIN. NOM. MAX. UNIT 26
Reference Clock Rate Ref_Freq fb/64 GHz
27
28
Input Differential Voltage Ref_Vdiff 240 900 mVppd
29
Input Single Ended Voltage Ref_Vse 0.1 1.2 V 30
Differential Termination Resistance Mismatch Ref_Rdm 5 % 31
Input Clock Rise and Fall Time (20% to 80%) Ref_tr, Ref_tf 4 200 ps 32
at fb/128  f < fb/4 33
Differential Input Return Loss Ref_SDD11 -6 dB
Note 1 34
Input Clock Duty Cycle Ref_DC 40 60 % 35
High Frequency Uncorrelated Unbounded 36
Ref_UUGJ_hf Note 2 0.009 UI rms
Gaussian Jitter 37
@ 1kHz offset -70 38
@10kHz offset -93 39
Reference Clock Single Side Band Phase
Ref_PN @100kHz offset -113 dBc/Hz
40
Noise 41
@1MHz offset -133
42
10MHz offset -143 43
NOTES: 44
1. Return loss is referenced to 100 Ohm
2. UUGJ measured using the methodology defined in Appendix 2.E.1 with a golden PLL cut-off frequency of fb/500, observing 45
with a bandwidth of 43GHz 46
47
48
49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 387
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 19.3.2 Transmitter Characteristics

2
3 The transmitter electrical specifications at compliance point T (see Figure 19-1) are
4 given in Table 19-7. The transmitter shall satisfy jitter requirements specified in Table
5 19-8. Jitter is measured relative to the reference clock as timing source, using a golden
6 clock multiplier as detailed in Annex 18.A.4, for a BER as specified in Section 3.2.3. It is
7 assumed that the UBHPJ component of the transmitter jitter is not data-dependent jitter
8 (DDJ) from the receiver view point, hence it cannot be equalized in the receiver. To
9 attenuate noise and absorb even/odd mode reflections, the transmitter shall satisfy the
10 Common Mode Output Return Loss requirement of Table 19-7. The waveform is
11 observed through a fourth-order Bessel-Thomson response with a bandwidth of 
12 43 GHz using a PRBS31 pattern.
13
14 The link budget in this document assumes no Tx emphasis.
15
Table 19-7. Transmitter Electrical Output Specification.
16
17 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
18
19 Baud Rate T_Baud 39.8 58.0 Gsym/s

20 Output Differential Voltage T_Vdiff 250 400 mVppd

21 Single Ended Transmitter Output Voltage Note 1 T_Vse 0.1 1.2 V
22 Differential Termination Resistance Mismatch T_Rdm 5 %
23
Output Rise and Fall Time (20% to 80%) T_tr, T_tf 4 ps
24
See 19.3.2.2
25 Differential Output Return Loss T_SDD22
Note 2
dB
26 Common mode Output Return Loss T_SCC22 200MHz to fb/2 -6 dB
27
Transmitter Common Mode Noise T_Ncm 15 mVrms
28
29 NOTES:
1. DC Coupling compliance is mandatory.
30 2. Return loss is referenced to 100 Ohm
31
32 Table 19-8. Transmitter Output Jitter Specification
33
34 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
35 Uncorrelated Unbounded Gaussian Jitter T_UUGJ 0.15 UIpp
36 Uncorrelated Bounded High Probability Jitter T_UBHPJ 0.15 UIpp
37 Even Odd jitter (component of UBHPJ) T_EOJ Note 2 0.035 UIpp
38
39 Total Jitter T_TJ Note 1 0.28 UI

40 Eye Mask T_X1 See 19.3.2.4 0.14 UI

41 Eye Mask T_X2 See 19.3.2.4 0.4 UI

42 Eye Mask T_Y1 See 19.3.2.4 125 mV
43 Eye Mask T_Y2 See 19.3.2.4 200 mV
44 NOTES:
45 1. T_TJ includes all of the jitter components.
46 2. Included in T_UBHPJ. Even-odd jitter is defined in Table 1-3.
47
48
49

388 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

19.3.2.1 Transmitter Amplitude and Swing 1

2
Transmitter differential output amplitude shall be able to drive between 250 and 3
400 mVppd. Further the single-ended voltage must be between 0.1 and 1.2 V. Figure 4
19-4 shows the transmitter test load configuration. 5
6
Figure 19-4. Transmitter Amplitude Test Load
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
19.3.2.2 Transmitter Resistance and Return Loss
24
Please refer to Section 3.2.10 with the following parameters in Table 19-9 and as 25
illustrated in Figure 19-5. 26
27
Table 19-9. Driver Return Loss Parameters 28
Parameter Value Units
29
30
A0 -12 dB 31
f0 100 MHz
32
33
f1 0.1714 x T_Baud Hz 34
f2 T_Baud Hz 35
36
Slope 12.0 dB/dec 37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 389
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 19-5. Illustration of Return Loss for T_Baud = 58.0 Gsym/s

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 19.3.2.3 Transmitter Lane-to-Lane Skew
22
23 Please refer to Section 3.2.7.
24
25 19.3.2.4 Transmitter Template and Jitter
26
27 When provided with the reference clock meeting the specifications in Section 19.3.1 as
28 a timing source, using a golden clock multiplier as detailed in Annex 18.A.4, for a BER
29 as per Section 3.2.3, the transmitter shall satisfy the eye template and jitter
30 requirements as given in Table 19-8 and Figure 1-4. The measurement of jitter and eye
31 diagram is to be relative to the "forwarded" clock if the transmitter provides one. If it
32 does not provide a "forwarded" clock then the measurement is to be relative to the
33 reference clock provided to it.
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

390 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

19.3.3 Receiver Characteristics 1

2
A compliant receiver shall operate at the specified BER with the worst case 3
combination of a compliant transmitter and a compliant channel. 4
5
Receiver electrical specifications are given in Table 19-10 and measured at compliance 6
point R. To dampen noise sources and absorption of both even and odd mode 7
reflections, the receiver shall satisfy the Common Mode Input Return Loss requirement 8
of Table 19-10. 9
10
Table 19-10. Receiver Electrical Input Specification
11
Characteristic Symbol Condition MIN. TYP. MAX. UNIT
12
13
Baud Rate R_Baud 39.8 58.0 Gsym/s 14
Input Differential Voltage R_Vdiff Note 1 400 mVppd 15
Receiver Common Mode Noise Tolerance R_Ncm 25 mVrms 16
Differential Termination Resistance Mismatch R_Rm 5 % 17
See 19.3.3.2
18
Differential Input Return Loss R_SDD11
Note 2
dB
19
200 MHz to10
-6 dB
20
Common mode Return Loss R_SCC11 GHz 21
10 GHz to fb -4 dB 22
Rx Input Single Ended Voltage Range R_VinSE 0.1 1.2 V 23
NOTES: 24
1. The receiver shall have a differential input voltage range sufficient to accept a signal produced at point R by the combined 25
transmitter and channel. The channel response shall include the worst case effects of the return losses at the transmitter and
receiver 26
2. Return loss is referenced to 100 Ohm 27
28
29
30
When provided with the reference clock meeting the specifications in Section 19.3.1 as 31
a timing source, the receiver shall tolerate the sum of these jitter contributions and 32
meet the BER as per Section 3.2.3: the total transmitter jitter from Table 19-8 and the 33
effects of a channel compliant to the Channel Characteristics (Section 19.2.7). 34
35
19.3.3.1 Reference Receiver Input Signals 36
37
The receiver shall accept differential input signal amplitudes produced by a compliant 38
transmitter connected with the minimum attenuation specified in Equation (19-2) to the 39
receiver. This may be larger than the 400 mVppd maximum of the transmitter due to 40
output/input impedances and reflections. 41
42
The minimum input amplitude is defined by the minimum transmitter amplitude, the 43
actual receiver input impedance and the loss of the actual PCB. Note that the minimum 44
transmitter amplitude is defined using a well controlled load impedance; however the 45
real receiver's impedance may differ, which can leave the receiver input signal smaller 46
than expected. Additionally the real receiver may be affected by environmental noise. 47
48
49

Optical Internetworking Forum - Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface 391
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 19.3.3.2 Receiver Return Loss

2
3 Please refer to Section 3.2.10 with the following parameters in Table 19-11 and as
4 illustrated in Figure 19-5.
5
Table 19-11. Receiver Input Return Loss Parameters
6
7 Parameter Value Units
8 A0 -12 dB
9 f0 100 MHz
10
f1 0.1714 x R_Baud Hz
11
12 f2 R_Baud Hz

13 Slope 12.0 dB/dec

14
15 19.3.3.3 Input Lane-to-Lane Skew
16
17 Please refer to Section 3.2.8.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

392 Clause 19: CEI-56G-XSR-NRZ Extra Short Reach Interface - Optical Internetworking Forum
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

20 Reserved to add future Clause 1

2
3
This page is intentionally blank 4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 20: Future Clause 393

 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 This page is intentionally blank

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

394 Optical Internetworking Forum - Clause 20: Future Clause

Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

21 CEI-56G-LR-PAM4 Long Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-LR-PAM4 long reach high speed 4
electrical interface between nominal baud rates of 18.0 Gsym/s and 29.0 Gsym/s using 5
PAM4 coding. A compliant device shall meet all of the requirements listed below. The 6
electrical interface is based on high speed, low voltage logic. Connections are point-to- 7
point balanced differential pairs and signaling is unidirectional. 8
9
The electrical IA is based on loss and jitter budgets and defines the characteristics required 10
to communicate between a CEI-56G-LR-PAM4 transmitter and a CEI-56G-LR-PAM4 11
12
receiver using copper signal traces on a printed circuit board. The characteristic impedance
13
of the signal traces is nominally 100 Ω differential. The signal trace or channel between a 14
transmitter and a receiver shall meet the channel operating margin (COM), a method and a 15
threshold quantity used for channel compliance. 16
17
CEI-56G-LR-PAM4 assumes using forward error correction (FEC) to achieve the bit error 18
ratio (BER) target. The FEC guidances are described in Appendix 16.D. 19
20
Long reach CEI-56G-LR-PAM4 devices from different manufacturers shall be 21
interoperable. 22
23
24
21.1 Requirements 25
26
1. Support serial baud rates (fb) within the range from 18.0 Gsym/s to 29.0 Gsym/s as 27
specified for the device using PAM4 coding. A CEI implementation complies to the 28
specifications of this clause over the range of baud rates stated for the 29
implementation within this range. 30
2. Capable of achieving a raw Bit Error Ratio (BER) of 10-4 or better per lane. FEC is 31
assumed to be used in the system to achieve corrected BER of 10-15 or better per 32
lane. The baud rate includes the overhead required for FEC. The definition of FEC 33
is outside the scope of this IA (see Appendix 16.D). 34
35
3. Capable of driving up to 1000 mm of PCB and up to 2 connectors. 36
4. Shall support AC-coupled operation. 37
38
5. Shall allow multi-lanes (1 to n). 39
40
6. Shall support hot plug.
41
42
21.2 General Requirements 43
44
45
21.2.1 Data Patterns 46
47
See Appendix 16.C.5. 48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 395
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 21.2.2 Bit Error Ratio

2
3 A raw Bit Error Ratio (BER) better than or equal to 10-4 is required on each lane. A
4 compliant receiver, when receiving from a compliant transmitter over a compliant
5 channel, shall deliver the specified raw BER to the subsequent FEC decoder. Error
6 bursts with length more than 126 PAM4 symbols delivered to the PAM4 decoder shall
7 occur with a probability of less than 1 in 1020 PAM4 symbols. See Appendix 16.D.
8
9 21.2.3 Ground Differences
10
11 Please refer to Section 3.2.4.
12
13
21.2.4 Channel Compliance
14
15 A forward channel and associated dominant crosstalk channels are deemed compliant
16 if the channel characteristics conform to the requirements defined in this section.
17
18
19 21.2.4.1 Reference Model
20
Figure 21-1.CEI-56G-LR Reference Model
21
22
23
24
25
26
27
28
29
30 AC Coupling
31
32
33
34
35
36
37
38
39
Note: Test points differ from definitions in Section 1.8, as a DC blocking capacitor, if
40
physically located outside of the package, is part of the channel.
41
42 The channel is defined between test point T and test point R.
43
44
45 21.2.4.2 Channel Operating Margin
46
47 The Channel Operating Margin (COM) of the channel is computed using the procedure
48 in Annex 93A of IEEE Std 802.3 [25] as modified by IEEE Std 802.3by [26] and IEEE
49 Std 802.3bs [27], with the Test 1 and Test 2 values in Table 21-1. Test 1 and Test 2

396 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

differ in the value of the device package model transmission line length zp. Moreover, 1
using Tr = 0.345 UI, and β=2 for Ht(f) in Equation (93A–19), COM shall be greater than 2
or equal to 3.0 dB for each test. This minimum value allocates margin for practical 3
4
Table 21-1. COM Parameter Values
5
Parameter Symbol Value Units 6
Signaling rate fb 18.0 - 29.0 Gsym/s 7
8
Maximum start frequency fmin 0.05 GHz 9
Maximum frequency step Δf 0.01 GHz 10
11
Device package model    12
Single-ended device capacitance Cd  160 fF  13
Transmission line length, Test 1 mm 
Transmission line length, Test 2
zp  12
mm 
14
zp  30
Transmission line characteristic impedance Zc Ω 15
95
Single-ended package capacitance at package-to- Cp 110  fF 16
board interface 
  17
Single-ended reference resistance R0 50
18
Ω
19
Single-ended termination resistance Rd 50 Ω 20
Receiver 3 dB bandwidth fr 0.75 × fb GHz 21
22
Transmitter equalizer, minimum cursor coefficient c(0) 0.60 —
23
Transmitter equalizer, 2nd pre-cursor coefficient c(–2)   24
Minimum value 0 — 25
Maximum value 0.10 — 26
Step size 0.025 — 27
Transmitter equalizer, 1st pre-cursor coefficient c(–1)   28
Minimum value –0.28 — 29
Maximum value 0 — 30
Step size 0.025 — 31
Transmitter equalizer, post-cursor coefficient c(1)  
32
Minimum value  –0.28 — 33
Maximum value  0 — 34
Step size 0.025 — 35
Continuous time filter, DC gain gDC   36
Minimum value –20 dB  37
Maximum value 0 dB  38
Step size 1 dB 39
Continuous time filter, DC gain2 gDC2   40
Minimum value –6 dB  41
Maximum value 0 dB  42
Step size 1 dB 43
Continuous time filter, scaled zero frequency fz fb /2.5 GHz 44
45
46
Continuous time filter, pole frequencies fp1 fb /2.5 GHz 47
fp2 fb GHz 48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 397
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 21-1. COM Parameter Values

2
Continuous time filter, low frequency pole/scaled fLF fb /40 GHz
3 zero
4
5 Transmitter differential peak output voltage   
Victim Av  0.41
6 Far-end aggressor V 
Afe  0.41
7 Near-end aggressor V 
Ane 0.60 V
8
9 Number of signal levels L 4 —
10 Level separation mismatch ratio RLM 0.95 —
11 Transmitter signal-to-noise ratio SNRTX 32.5 dB
12
13 Number of samples per unit interval M 32 —

14 Decision feedback equalizer (DFE) length Nb 12 UI

15 Normalized DFE coefficient magnitude limit  bmax(1) 0.7 —
16 for n = 2 to N b bmax(2-Nb) 0.2
17 Random jitter, RMS σRJ 0.01 UI
18
19 Dual-Dirac jitter, peak ADD 0.02 UI

20 One-sided noise spectral density η0 -8

1.64 × 10 V2/GHz
21 Target detector error ratio DER0 10-4 —
22 Channel operating margin, min COM 3.0 dB
23
24 limitations on the receiver implementation, the largest step size allowed for transmitter
25 equalizer coefficients.
26
27
28 21.2.4.3 Informative Channel Insertion Loss
29
30 Figure 21-2.Channel Insertion Loss Limit for 29.0Gsym/s
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

398 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
f  29 f  29 3
IL max = 1.083 + 3.631 ------------- + 1.041 -------------  f min  f  f b (21-1) 4
fb fb
5
6
7
8
 9
0 f min  1GHz  10
 11
IL min = 1
---  f – 1  1GHz  f  17.5GHz (21-2)
3 12
13
5.5 17.5GHz  f  f b 14
15
16
Channel insertion loss is an informative recommendation. 17
18
The channel must comply with the normative specification in Section 21.2.4.2. 19
20
21.2.4.4 Channel Return Loss 21
22
Figure 21-3.Channel Return Loss Limit for 29.0 Gsym/s 23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
  43
44
45
Channel Return Loss shall be bounded by Equation (21-3) as shown in Figure 21-3. 46
47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 399
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2 fb
3 12 f min  f  ----
4 4
RL max = (21-3)
5 fb
6  4f 
12 – 15 log 10 -----  ----  f  f b
7 f  4
b
8
9
10 21.2.4.5 Channel AC-coupling
11
12 The transmitter shall be AC-coupled to the receiver. The impact of a DC-blocking
13 capacitor implemented in the channel between the package balls of the transmitter and
14 receiver (i.e., between compliance points T and R) is accounted for within the channel
15 specifications. Common-mode specifications are defined as if the DC-blocking
16 capacitor is implemented in the channel between compliance points T and R. Should
17 the capacitor not be implemented between compliance point T and compliance point R,
18 it is the responsibility of implementers to consider any necessary modifications to
19 common-mode and channel specifications required for interoperability as well as any
20 impact on the verification of transmitter and receiver compliance. In particular the
21 common-mode specifications for the transmitter in Table 21-2 may not be appropriate.
22
23 The low-frequency 3 dB cutoff of the AC-coupling shall be less than 100 kHz.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

400 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

21.3 Electrical Characteristics 1

2
The electrical signaling is based on high speed low voltage logic with a nominal differential 3
impedance of 100 Ω. 4
5
21.3.1 Transmitter Characteristics 6
7
The transmitter electrical requirements at compliance point T (see Figure 21-1) are 8
specified in Table 21-2, and the jitter requirements are specified in Table 21-3. 9
10
Table 21-2. Transmitter Electrical Output Specification 11
12
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 13
Baud Rate T_Baud 18.0 29.0 Gsym/s 14
.
15
Output Differential Voltage T_Vdiff
See Note 1, 2.
1200 mVppd 16
DC Common mode Voltage T_Vcm See Note 2. 0 1.9 V 17
Output AC Common Mode Voltage T_VcmAC See Note 1, 2. 30 mVrms
18
19
Single-ended Transmitter Output Voltage T_Vse See Note 1, 2. -0.3 1.9 V
20
Differential Output Return Loss T_SDD22 Equation (21-4) dB 21
Common Mode Output Return Loss T_SCC22 Equation (21-5) dB 22
Level Separation Mismatch Ratio T_RLM 0.95 - 23
Steady-state Voltage T_Vf 0.4 0.6 V 24
See Note 1, 2,
3. 4 0.83 ×
25
Linear Fit Pulse Peak T_Pk
T_Vf
V 26
Signal-to-Noise-and-Distortion-Ratio T_SNDR 31 dB 27
NOTES:
28
1. Signals are specified as measured through a fourth-order Bessel-Thomson low-pass response with 40 GHz 3 dB bandwidth. 29
2. Measured as described in Section 21.3.1.2. T_Vdiff min is set by the steady-state voltage T_Vf min. 30
3. Measured as described in Section 21.3.1.6
4. T_RLM is defined in Appendix 16.C.4.3. 31
32
Table 21-3. Transmitter Output Jitter Specification 33
34
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 35
Uncorrelated Jitter (time interval from 0.005% to 36
T_J4u 0.118 UI
99.9955% of the probability distribution) 37
Uncorrelated jitter RMS (standard deviation of
T_JRMS
See Note 1
0.023 UIrms 38
the probability distribution)
39
Even-Odd Jitter T_EOJ 0.019 UIpp 40
NOTES: 41
1. Measured as described in Section 21.3.1.7.
42
43
21.3.1.1 Transmitter Baud Rate 44
45
All devices shall work within the range from 18.0 Gsym/s to 29.0 Gsym/s as specified 46
for the device, with the baud rate tolerance as per Section 3.2.11. A CEI 47
implementation complies to the specifications of this clause over the range of baud 48
rates stated for the implementation within this range. 49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 401
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 21.3.1.2 Transmitter Amplitude and Swing

2
3 The differential output voltage T_Vdiff is defined to be True minus Complement. The
4 common-mode output voltage T_Vcm is defined to be one half of the sum of True and
5
Complement. These definitions are illustrated in Section 1.6.1.
6
7
8 For a QPRBS13-CEI test pattern (Appendix 16.C.3.1), the peak-to-peak value of the
9 differential output voltage (T_Vdiff) shall be less than or equal to the limit given in Table
10 21-2 regardless of the transmit equalizer setting.
11
12 The DC common-mode output voltage (T_Vcm) shall be within the limits in Table 21-2
13 with respect to local ground.
14
15 The AC common-mode output voltage (T_VcmAC) shall be less than or equal to the
16 limit given in Table 21-2 with respect to local ground. Common-mode output voltage
17
requirements shall be met regardless of the transmit equalizer setting.
18
19 The single-ended transmitter output voltage (T_Vse) shall be within the limits in Table
20 21-2 with respect to local ground.
21
22
23 The transmitter shall be capable of providing a differential steady state output
24 amplitude (2xT_Vf) between 800 and 1200 mVppd with transmit emphasis disabled.
25
26 Transmitter differential output amplitude shall additionally adhere to the requirements in
27 Section 21.3.1.6.
28
29 Power-down behavior is beyond the scope of CEI IA.
30
31 21.3.1.3 Transmitter Return Loss
32
33 The differential output return loss, in dB, of the transmitter shall meet Equation (21-4),
34 where f is the frequency in GHz. The differential return loss limit RLd(f) is shown in
35 Figure 21-4. This output impedance requirement applies to all valid output levels. The
36 reference impedance for differential return loss measurements shall be 100 Ω.
37
38 f  29
39  12.05 – 0.4112  -------------  0.05  f  0.5f b  
 fb 
40 RL d  f   RLmin  f  =    dB  (21-4)
 f  29 
41  7.175 – 0.075 -------------  0.5f b  f  f b  
  fb  
42
43
44
45
46
47
48
49

402 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 21-4.Transmitter differential return loss limit for 29.0 Gsym/s 1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
  20
21
22
The common-mode output return loss, in dB, of the transmitter shall meet Equation (21-
23
5), where f is the frequency in GHz. The common-mode return loss limit RLC(f) is shown
24
in Figure 21-5. This output impedance requirement applies to all valid output levels.
25
The reference impedance for common-mode return loss measurements shall be 25 Ω.
26
 27
28
f  29
 9.05 – 0.4112  -------------  0.05  f  0.5f b   29
 fb  30
RL c  f   RLmin  f   =    dB  (21-5)
f  29 31
 4.175 – 0.075  -------------  0.5f b  f  f b  
  fb   32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 403
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 21-5.Transmitter common mode return loss limit for 29.0 Gsym/s.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19  
20
21
22
23 21.3.1.4 Transmitter Lane-to-Lane Skew
24
Please refer to Section 3.2.7.
25
26
27 21.3.1.5 Transmitter Short Circuit Current
28
29 Please refer to Section 3.2.9.
30
31 21.3.1.6 Transmitter output waveform requirements
32
33 The transmitter function includes programmable equalization to compensate for the
34 frequency-dependent loss of the channel and facilitate data recovery at the receiver.
35 The functional model for the transmit equalizer is the four tap transversal filter shown in
36 Figure 21-6.
37
38
39
40
41
42
43
44
45
46
47
48
49

404 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 21-6.Transmit equalizer functional model 1

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Link budgets in this document assume optimized TX FIR equalization that is part of the 25
system management function. The specific implementation is outside the scope of this 26
document. 27
28
29
21.3.1.6.1 Linear fit to the measured waveform 30
31
The following test procedure defines linear fit pulse response, linear fit error (e(k), see 32
Section 11.3.1.6.4), and normalized transmitter coefficient values. 33
34
For each configuration of the transmit equalizer, capture at least one complete cycle of 35
the QPRBS13-CEI test pattern (Appendix 16.C.3.1) at the TX package ball (see Figure 36
21-1). 37
38
Compute the linear fit pulse response p(k) from the captured waveform per Section 39
11.3.1.6.2 using Np = 15 and Dp = 3. For aligned symbol values x(n) use -1, -ES1, ES2, 40
and 1 to represent symbol values of -1, -1/3, 1/3, and 1, respectively, and where ES1 41
and ES2 are the effective symbol levels determined in Appendix 16.C.4.3. 42
43
Define r(k) to be the linear fit pulse response when transmit equalizer coefficients have
44
been set to the “preset” values (see Section 11.3.1.6.1).
45
For each configuration of the transmit equalizer, compute the normalized transmit 46
equalizer coefficients, c(i), according to Section 11.3.1.6.2 - Section 11.3.1.6.5. 47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 405
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 21.3.1.6.2 Steady-state voltage and linear fit pulse peak

2
3 The linear fit pulse, p(k), is determined according to the linear fit procedure in Section
4 11.3.1.6.2 - Section 11.3.1.6.5, as modified by Section 21.3.1.6.1. The steady-state
5 voltage T_Vf is defined to be the sum of the linear fit pulse p(k) divided by M, as shown
6 in Equation (11-12).
7
8 The steady-state voltage, T_Vf, shall satisfy the requirements in Table 21-2.
9
10 The linear fit pulse peak, T_Pk, is the highest value of p(k). It shall satisfy the
11 requirement in Table 21-2.
12
13 21.3.1.6.3 Transmitter equalizer coefficients
14
15 Table 21-4. Coefficient Range and Step Size
16 Normalized Amplitude
17 Coefficients Normalized Step Size (%)
Min (%) Max (%)
18
19 c(-2) 0 10 0.5 to 2.5
20 c(-1) -28 0 0.5 to 2.5
21 c(1) -28 0 0.5 to 2.5
22 c(0) 60 100 0.5 to 2.5
23
24
The normalized amplitudes of the coefficients of the transmitter equalizer (computed
25
per Section 21.3.1.6.1) shall meet the requirements in Table 21-4. "min" is defined as
26
the minimum normalized amplitude of the coefficient that must be supplied by the
27
transmitter to be compliant. "max" is defined as the maximum normalized amplitude of
28
the coefficient that must be supplied by the transmitter to be compliant.
29
30 The amplitude of a coefficient can be computed by multiplying its normalized amplitude
31 by T_Vf, which is defined in Section 21.3.1.6.2.
32
33 The peak-to-peak output voltage is approximated by
34
35 ( |c(-2)| + |c(-1)| + |c(0)| + |c(1)| ) * 2 * T_Vf (21-6)
36
37 and should not exceed the limit for T_Vdiff given in Table 21-2.
38
39 21.3.1.6.4 Transmitter Output Noise and Distortion
40
41 Signal-to-noise-and-distortion ratio (SNDR) is measured at the transmitter output using
42 the following method, with the transmitter on the lane under test transmitting
43 QPRBS13-CEI and transmitters on lanes not under test enabled and transmitting
44 QPRBS31-CEI test pattern as defined in Appendix 16.C.3.2, or a valid CEI signal, or
45 transmitting the same pattern with a slightly different Baud rate on each lane so that
46 lane to lane signals are asynchronous. These transmitters shall have identical transmit
47 equalizer settings to the transmitter under test.
48
49

406 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Compute the linear fit to the captured waveform and the linear fit pulse response, p(k), 1
and error, e(k), according to Section 21.3.1.6.1. Denote the standard deviation of e(k) 2
as σe. 3
4
With the QPRBS13-CEI pattern and the same configuration of the transmit equalizer, 5
measure the RMS deviation from the mean voltage at a fixed point in a run of at least 6 6
consecutive identical PAM4 symbols. The RMS deviation is measured for a run of each 7
of the four PAM4 symbol levels.The average of the four measurements is denoted as 8
σn. 9
10
SNDR is defined by Equation (21-7) where pmax is the maximum value of p(k). 11
12
13
2
 pmax  14
-  dB 
SNDR = 10 log  ---------------------- (21-7)
10   +  
2 2 15
e n 16
 17
SNDR shall be greater than 31 dB for any allowable transmit equalizer setting. 18
19
21.3.1.7 Transmitter output jitter 20
21
Jitter measurements in this sub-clause are performed with transmitters on physical 22
lanes not under test enabled and transmitting QPRBS31-CEI test pattern as defined in 23
Appendix 16.C.3.2, or a valid CEI signal, or transmitting the same pattern with a slightly 24
different Baud rate on each lane so that lane to lane signals are asynchronous. These 25
transmitters shall have identical transmit equalizer settings to the transmitter under test. 26
27
J4u, JRMS, and EOJ are defined by measurements of 12 specific transitions in a 28
QPRBS13-CEI pattern in order to exclude correlated jitter. The 12 transitions represent 29
all possible combinations of four identical symbols followed by two different identical 30
symbols as shown in Table 21-5. The sequences are located by the symbol indices 31
given in the table where symbols 1 to 7 are the run of seven +1s. 32
33
The threshold used to define each transition is given in Table 21-5 where V-1, V-1/3, 34
V1/3, and V1 are as defined in Appendix 16.C.4.3. 35
36
The jitter is measured with a clock from a clock recovery unit (CRU) (i.e., a first order 37
golden PLL, with corner frequency at fb/6640, and a 20 dB/decade slope, see Section 38
1.6) as the trigger or reference clock. 39
40
J4u, JRMS, and EOJ specifications shall be met regardless of the transmit equalization 41
setting. 42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 407
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 21-5. QPRBS13-CEI Pattern Symbols Used for Jitter Measurement

2
3 Index of  Index  Index  Index of 
Gray Coded PAM4 
Label  Description First  Transition  Transition  Last  Threshold Level
4 Symbols
Symbol Begins Ends Symbol
5
6 Reference 
7 REF for symbol  3333333 1 ‐‐ ‐‐ 7 ‐‐
8 index 
9 R03 0 to 3 rise  10000 330  1830 1834 1835 1837
10 (V‐1+V1)/2
F30  3 to 0 fall 23333 001 1269 1273 1274 1276
11
12 R12 1 to 2 rise 0111111 2222221 3638 3644 3645 3651
(V‐1/3+V1/3)/2
13 F21 2 to 1 fall 022222 113 1198 1203 1204 1206
14
15 R01 0 to 1 rise 100000 113  6835 6840 6841 6843
(V‐1+V‐1/3)/2
16 F10 1 to 0 fall 21111 003  2992 2996 2997 2999
17 R23 2 to 3 rise 32222 330  6824 6828 6829 6831
18 (V1/3+V1)/2
19 F32 3 to 2 fall 033333 2222223  7734 7739 7740 7746
20 R02 0 to 2 rise 10000 223  3266 3270 3271 3273
21 (V‐1+V1/3)/2
F20 2 to 0 fall 122222 0000002  7282 7287 7288 7294
22
23 R13 1 to 3 rise 011111 331  133 138 139 141
(V‐1/3+V1)/2
24 F31 3 to 1 fall 23333 112  7905 7909 7910 7912
25
26
27 21.3.1.7.1 J4u and JRMS Jitter
28
29 For each transition i, 1 ..i ..12, of the transitions specified in Table 21-5, obtain a set Si =
30 {ti(1), ti(2), ...} of transition times modulo the period of the pattern. The 12 sets should
31 be of equal size and the size of all sets should be chosen to enable calculation of J4u
32 (as defined below) with sufficient accuracy.
33
34 Calculate the average of each set Si, Tavgi, and subtract it from all elements of that set,
35 to create a set S0i = {ti(1) - Tavgi, ti(2) - Tavgi, ...}.
36
Combine the sets S0i, i=1 to 12, to create an estimated probability distribution fJ(t).
37
38 J is defined as the time interval that includes all but 10-4 of f (t), from the 0.005th to
4u
39 the J
99.995th percentile of fJ(t).
40
41 J
RMS is defined as the standard deviation of fJ(t).
42
43
44 21.3.1.7.2 Even-Odd Jitter (EOJ)
45
46 For one of the 12 specific transitions in QPRBS13-CEI in Table 21-5:
47
48 a) Trigger once in 3 repeats of the QPRBS13-CEI test pattern.
49

408 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Obtain the mean time (T3) for this transition in the first QPRBS13-CEI. 1
2
Obtain the mean time (T4) for the same transition in the second QPRBS13-CEI. 3
4
b) The difference between the two means (T4 – T3), is the estimated period of the 5
repeating pattern. 6
7
For each of the 12 specific transitions in QPRBS13-CEI in Table 21-5: 8
9
1) Trigger once in 2 repeats of the QPRBS13-CEI test pattern. 10
11
Obtain the mean time (T1) for the specific transition in the first QPRBS13-CEI. 12
13
Obtain the mean time (T2) for the same transition in the second QPRBS13-CEI. 14
2) Calculate EOJ for this transition as |(T2 – T1) – (T4 – T3)|. 15
16
EOJ is the maximum of the 12 measurements. 17
18
NOTE: Both of (T2 - T1) and (T4 - T3) are about 8191 UI, which is much larger than the 19
EOJ value. Hence, each of T1 through T4 should have high precision. 20
21
22
21.3.2 Receiver Characteristics 23
24
A compliant receiver shall autonomously operate at the specified BER with the worst 25
case combination of a compliant transmitter and a compliant channel. The receiver also 26
shall not cause error propagation that violates the error burst length requirement as 27
defined in Section 21.2.2. Further receiver electrical requirements at compliance point 28
R (see Figure 21-1) are specified in Table 21-6, with the receiver interference tolerance 29
parameters specified in Table 21-7. Lanes not under test should be enabled and 30
transmitting or receiving asynchronous or uncorrelated signals. 31
Table 21-6. Receiver Electrical Input Specification 32
33
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 34
35
Baud Rate R_Baud 18.0 29.0 Gsym/s
36
Differential Input Return Loss R_SDD11 Equation (21-4) dB 37
Differential to Common Mode Input
R_SCD11 Equation (21-8) dB
38
Conversion 39
Interference Tolerance Table 21-7 40
Jitter Tolerance Table 21-8 41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 409
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Table 21-7. Receiver interference tolerance parameters (Note 3)

2
3 Parameter Test 1 values Test 2 values Units
4
5 Min Max Min Max
6 Pre-FEC Bit Error Ratio (BER) 10–4 10–4
7
8 COM, including effects of broadband noise 3 3 dB
9 Insertion loss at Nyquist, Note 1 15 30 dB
10 RSS_DFE4. Note 2 0.05 0.05
- -
11
12 NOTES:
1. Measured between TX and RX package balls (see Figure 21-1).
13 2. Definition can be found in Annex 93A of IEEE Std 802.3 [25] as modified by IEEE Std 802.3by [26] and IEEE
14 P802.3bs [27].
15 3. See Section 21.3.2.4.
16
17 21.3.2.1 Input Baud Rate
18
19 All devices shall work within the range from 18.0 Gsym/s to 29.0 Gsym/s as specified
20 for the device, with the baud rate tolerance as per Section 3.2.11. A CEI
21 implementation complies to the specifications of this clause over the range of baud
22 rates stated for the implementation within this range.
23
24 21.3.2.2 Reference Input Signals
25
26 The receiver shall accept differential input signal amplitudes produced by a compliant
27 transmitter connected with the minimum attenuation specified in Figure 21-2 to the
28 receiver. This may be larger than the 1200 mVppd maximum of the transmitter due to
29 output/input impedances and reflections.
30
31 The minimum input amplitude is defined by the minimum transmitter amplitude, the
32 actual receiver input impedance and the loss of the actual PCB. Note that the minimum
33 transmitter amplitude is defined using a well controlled load impedance, however the
34 real receiver is not, which can leave the receiver input signal smaller than expected.
35
36
37 21.3.2.3 Input Resistance and Return Loss
38
39 The differential input return loss, in dB, of the receiver shall meet Equation (21-4). The
40 reference impedance for differential return loss measurements shall be 100 Ω.
41
The differential to common-mode return loss, in dB, of the receiver shall meet Equation
42
(21-8). The differential to common-mode return loss limit RLdc(f) is shown in Figure 21-
43
7.
44
45
46  f  29  0.05  f  0.5f b  
 25 – 0.6897  -------------
fb 
47 RL dc  f   RLmin  f  =    dB  (21-8)
48  15  0.5f b  f  f b  
49

410 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
Figure 21-7.Receiver differential to common-mode return loss limit for 29.0 Gsym/s. 3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
  20
21
22
23
21.3.2.4 Receiver Interference Tolerance 24
25
The receiver interference tolerance test is based on the test defined in Annex
26
120D.3.2.1 of IEEE P802.3bs [27].
27
The receiver on each lane shall meet the pre-FEC BER requirement with channels 28
matching the Channel Operating Margin (COM) and loss parameters for Test 1 and 29
Test 2 in Table 21-7. 30
31
The test channel should be created using printed circuit boards with short 32
interconnecting cables. 33
34
The following considerations apply to the interference tolerance test. The transmitter 35
package is omitted in the COM calculation. The Test transmitter's measured SNDR 36
should be used for SNRTX in the COM calculation. The transmitter output levels are set 37
such that RLM is equal to 0.95. The test transmitter meets the specifications in Section 38
21.3.1. The test transmitter is constrained such that for any transmitter equalizer setting 39
the differential peak-to-peak voltage is less than 800 mV, and the normalized 40
amplitudes of the coefficients of the transmitter equalizer c(-2), c(-1), c(0) and c(1) are 41
between the minimum and maximum limits given in Table 21-4. 42
43
The lower frequency bound for the noise spectral density constraints, fNSD1, is 1 GHz. 44
The differential return-loss of the test channel at TP5 (as defined in Annex 93A of Std. 45
IEEE802.3 [25]) shall meet the requirements of Equation (21-4), and be 3 dB better 46
than the requirements of Equation (21-4) for all frequencies less than fb/2. The test 47
transmitter's jitter parameters J4u and JRMS are measured. ADD and σRJ are calculated 48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 411
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 from the measured values of J4u and JRMS using Equation (21-9), and Equation (21-
2 10), respectively and used for COM parameters. Other COM parameters are set
3 according to the values in Table 21-1. The broadband noise is added and adjusted to
4 achieve the COM value in Table 21-7. The test pattern to be used is QPRBS31-CEI
5 defined in Appendix 16.C.3.2. A test system with a fourth-order Bessel-Thomson low-
6 pass response with 40 GHz 3 dB bandwidth is to be used for measurement of the
7 signal applied by the pattern generator and for measurements of the broadband noise.
8
9
J 4u J 4u 2
10 A DD =   ------- + Q4   Q4 + 1   J RMS 2 –  -------     Q4 + 1 
2 2
11 2 2 (21-9)
12
13
14
15 J 4u
16  RJ =  ------- – A DD   Q4 
17 2 (21-10)
18
19
20
21 21.3.2.5 Receiver Jitter Tolerance
22
23 Receiver jitter tolerance shall meet the conditions and parameters defined in Table 21-
24 8. This sinusoidal jitter is part of the jitter applied in the stressed input test. The
25 sinusoidal jitter is calibrated at 10x the reference CRU’s bandwidth and must be tested
26 at fCRU/100, fCRU/3, fCRU, 3fCRU, and 10fCRU, where fCRU is the jitter corner frequency
27 given by fb/6640, with sinusoidal jitter of 5 UI, 0.15 UI, 0.05 UI, 0.05 UI and 0.05 UI
28 respectively. For this test the channel used is as for the receiver interference tolerance
29 described in Section 21.3.2.4. Note that the values measured for J4u and JRMS include
30 the effects of this added sinusoidal jitter and noise is added to obtain a COM of 3 dB
31 with these measured jitter values as for the interference tolerance test.The receiver bit
32 error ratio (BER) shall meet the requirements of Section 21.2.2 for each pair of jitter
33 frequency and peak-to-peak amplitude values listed above and shown in Figure 21-8.
34
Table 21-8. Receiver Jitter Tolerance Parameters
35
36
37 Sinusoidal jitter,
38 Frequency Range peak-to-peak
39 (UI)
40 f < fb/664000 Not Specified
41
42 fb/664000 < f < fb/6640 5*fb/(664000*f)
43 fb/6640 < f < 10fCRU 0.05
44
45
46
47
48
49

412 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Figure 21-8. Receiver Jitter Tolerance Mask 1

2
3
5 UIpp X 4
Sinusoidal 5
jitter 6
amplitude 7
8
X 9
10
0.05 UIpp X X X 11
12
13
14
fb/664000 fb/6640 10fCRU 15
16
17
18
21.3.2.6 Single Ended Input Voltage 19
20
The single ended voltage levels with respect to the receiver ground at the input of the 21
receiver are dependent on the transmitter implementation and the inter-ground 22
difference. The voltage levels at the input of an AC coupled receiver (if the effective AC 23
coupling is done within the receiver) or at the TX side of the external AC coupling cap (if 24
AC coupling is done externally) will be between -0.35V and 1.95V with respect to local 25
ground. 26
27
21.3.2.7 Input Lane-to-Lane Skew 28
29
Refer to Section 3.2.8. 30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface 413
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

414 Optical Internetworking Forum - Clause 21: CEI-56G-LR-PAM4 Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22 CEI-56G-LR-ENRZ Long Reach Interface 1

2
3
This clause details the requirements for the CEI-56G-LR-ENRZ long reach high speed 4
electrical interface between nominal data rates of 99.5 Gbit/s and 112.4 Gbit/s using 5
ENRZ coding across four wires. This throughput, normalized per wire, is equivalent to
6
other CEI-56G variants which define signaling over two wires. 7
A compliant device shall meet all of the requirements listed below. The electrical 8
interface is based on high speed, low voltage logic. Connections are point-to-point 9
balanced quad wire groups and signaling is unidirectional. 10
11
The electrical IA is based on loss and jitter budgets and defines the characteristics 12
required to communicate between a CEI-56G-LR-ENRZ transmitter and a CEI-56G- 13
LR-ENRZ receiver using copper signal traces on a printed circuit board. The 14
characteristic impedance of the signal traces is nominally 50  between any wire of the 15
quad channel to AC ground. A ‘length’ is effectively defined in terms of its attenuation 16
and phase response rather than its physical length. Refer to Section 22.2.6 for the 17
channel requirements. 18
19
Long reach CEI-56G-LR-ENRZ devices from different manufacturers shall be 20
interoperable. 21
22
23
22.1 Requirements 24
25
1. Support data rates within the range of 99.5 Gbit/s and 112.4 Gbit/s across four 26
wires. Corresponding baud rates are within the range of 33.16 Gbaud/s and 37.50 27
Gbaud/s. 28
29
2. Capable of low bit error ratio (10-15, with a test requirement to verify 10-12). 30
3. Capable of driving up to 1000 mm of PCB and up to 2 connectors. 31
32
4. Shall support AC-coupled operation. 33
34
5. Shall allow multiple CEI lanes (1 to n). 35
6. Shall support hot plug. 36
37
38
22.2 General Requirements 39
40
22.2.1 Data Patterns 41
42
Please refer to Section 3.2.1 43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 415
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.2.2 Signal levels

2
3 The signal is a low swing Ensemble NRZ (ENRZ) interface using four wires, designated
4 W3, W2, W1, and W0. Low swing ENRZ signaling provides noise immunity and
5 improved electromagnetic interference (EMI) similar to the characteristics of differential
6 NRZ signaling. See Figure 22-1 for an illustration of absolute driver output voltage limits
7 and voltage levels for driver output states V+1, V+1/3, V-1/3, and V-1. See Figure 22-2 for
8 a definition of differential peak-to-peak amplitude and differential voltages for bit values
9 of 1 and 0 (Vdiff-1 and Vdiff-0).
10
Figure 22-1. Definition of Driver Amplitude and Swing
11
12
13
14 V+1
15
Max absolute output
16
17
18 V+1/3
19
20 VCM
21
22 V-1/3
23
24
25
V-1
26
Min absolute output
27 GND
28
29
30 Figure 22-2. Definition of Driver Amplitude and Swing
31
32
33 Vdiff-1
34
35
36 True - Complement =
Differential Voltage
(peak-to-peak)
37 (W3+W2)-(W1+W0) or
38 (W3+W1)-(W2+W0) or
39 (W3+W0)-(W2+W1)
40
41
42
43 Vdiff-0
44
45
46
47
48
49

416 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Given absolute voltages VW3, VW2, VW1, and VW0 referenced to ground on the 1
respective wires W3, W2, W1, and W0, the DC common mode voltage of the ENRZ 2
interface is defined as the average on the voltage on the four wires. 3
4
The single-ended voltage on wire i with respect to VCM is defined as: 5
6
7
V W3 + V W2 + V W1 + V W0 (22-1) 8
V CM = ----------------------------------------------------------- 9
4
10
11
The single-ended voltage on wire i with respect to VCM is defined as: 12
13
14
V SE(i) = V i – V CM  i = W3 W2 W1 W0  (22-2) 15
The maximum and minimum single-ended voltages on any of the four wires are: 16
17
18
V SE-max = max  V SE(W3) ,V SE(W2) ,V SE(W1) ,V SE  W0   19
(22-3)
V SE-min = min  V SE(W3) ,V SE(W2) ,V SE(W1) ,V SE  W0   20
21
22
The maximum peak-to-peak swing of VSE is defined as: 23
24
25
V SEpp = V SE-max – V SE-min (22-4) 26
27
As illustrated in Figure 22-1, the output driver may assume any one of four output 28
states, where the nominal drive voltages of these states are defined in Table 22-1. Each 29
of wires W3, W2, W1, and W0 can be driving different states, but only combinations of 30
driver states which maintain a constant VCM are allowed. 31
32
Table 22-1. ENRZ Signal Drive States 33
Nomenclature for Signal State Nom. VSE Nom. Vi 34
35
+1 + (VSEpp / 2) + (VSEpp / 2) + VCM
36
+(1/3) + (VSEpp / 6) + (VSEpp / 6) + VCM
37
-(1/3) - (VSEpp / 6) - (VSEpp / 6) + VCM 38
-1 - (VSEpp / 2) - (VSEpp / 2) + VCM 39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 417
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The four wire ENRZ interface carries data over three differential subchannels
2 designated as D2, D1, and D0. Each ENRZ symbol conveys one data bit on each of the
3 three subchannels. The corresponding code map is described in Table 22-2.
4
5 Table 22-2. ENRZ Signaling Code Map
6 Data Value Wire States Data Value Wire States
7 (Subchannels D2, D1, D0) (W3, W2, W1, W0) (Subchannels D2, D1, D0) (W3, W2, W1, W0)
8 000 (-1, +1/3, +1/3, +1/3) 100 (-1/3, -1/3, -1/3, +1)
9 001 (-1/3, +1, -1/3, -1/3) 101 (+1/3, +1/3, -1, +1/3)
10 010 (-1/3, -1/3, +1, -1/3) 110 (+1/3, -1, +1/3, +1/3)
11
011 (+1/3, +1/3, +1/3, -1) 111 (+1, -1/3, -1/3, -1/3)
12
13
14 Each subchannel is decoded by a comparator which implements linear combination
15 Equation (22-5), where the weights of each wire for decoder ci is defined in Table 22-3.
16
17 m–1
18 (22-5)
19
ci =   weight of wire j   signal level on wire j 
20 j=0
21
22
23
24
Table 22-3. ENRZ Linear Combination Table
25
26 Contribution of each wire to comparator decode
27 Comparator w3 w2 w1 w0
28
29 d0 +1/2 +1/2 -1/2 -1/2

30 d1 +1/2 -1/2 +1/2 -1/2

31 d2 +1/2 -1/2 -1/2 +1/2
32
33
34
35 If ci is greater than 0, then di = 1; if ci is less than 0, then di = 0. The differential
36 amplitude at the comparator in the absence of any gain in the receiver is:
37
38
39 (22-6)
40 V ppd = 0.33  V SEpp 
41
42
43
44
45
46
47
48
49

418 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.2.3 Signal Definitions 1

2
The signal paths of CEI channels are unidirectional point-to-point connections. Each 3
quad channel for CEI-56G-LR-ENRZ is made up of four balanced electrical 4
connections. Figure 22-3 illustrates two quad channels comprising a full duplex CEI 5
lane. However the CEI specification does not preclude use for unidirectional protocol 6
applications. 7
Figure 22-3. Signal Diagram 8
9
Reference Clock 10
(optional) 11
Optional 12
CEI Quad CEI Quad
13
Transmit Receive 14
W3 15

Connector
Connector
TxData0

W2 16
W1 17
W0
18
CEI-56G-LR-ENRZ 19
Device
20

Connector
Connector

W3
RxData0

W2 21
W1 22
W0 23
CEI Quad
Receive
CEI Quad
Transmit
24
25
Card Backplane Card 26
27
28
29
30
. 31
32
Table 22-4. Receive Signal Summary 33
Signal Name Direction Function 34
35
The Receive Data (RXDATA[n]) signals are the inputs
RXDATA[n..0] Input to SERDES Component
to the SERDES component. 36
37
38
39
Table 22-5. Transmit Signal Summary 40
Signal Name Direction Function
41
42
The Transmit Data (TXDATA[n]) signals are the outputs
TXDATA[n..0] Output of SERDES Component
of the SERDES component.
43
44
Please refer to Appendix 1.A for a description of the reference clock in Figure 22-3. 45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 419
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.2.4 Bit Error Ratio

2
3 Please refer to Section 3.2.3. Forward Error Correction (FEC) may be used to extend
4 BER beyond these limits.
5
6 22.2.5 Ground Differences
7
8 Please refer to the maximum ground difference for LR links in Section 3.2.4.
9
10
11 22.2.6 Channel Compliance
12
13 A forward channel and associated dominant crosstalk channels are deemed compliant
14 if the channel characteristics conform to the requirements in this section.
15
16 22.2.6.1 Reference Model
17
18 The channel consists of PCB traces, vias, and up to 2 connectors. Traces are routed as
19 two differential pairs (stripline). The reference PCB trace impedance is 50 . (single-
20 ended), with loose coupling between the differential pairs. The S-parameters of each
21 ENRZ subchannel must meet the requirements described in Section 22.2.6.2, Section
22 22.2.6.3, Section 22.2.6.4, Section 22.2.6.5, and Section 22.2.6.6. Appendix 22.D.3
23 provides guidance for meeting these requirements.
24
25 Figure 22-4 shows a diagram of test points on an example board. The DC blocking
26 capacitor, if physically located outside of the package, is part of the channel.
27 Figure 22-4. CEI-56G-LR-ENRZ Reference Model
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

420 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Eight port S-Parameter measurements of the channel are used to determine channel 1
compliance. These measurements are obtained using one of the methods described in 2
Appendix 22.D. Measured at these test points, several channel characteristics are 3
parameterized for each subchannel D2, D1, and D0. Port definitions as noted in Figure 4
22-4 allow proper measurement of the parameters in Table 22-6 used for calculation of 5
the channel parameters found in Table 22-7. 6
7
Table 22-6. Measured Channel Parameters 8
Symbol Description 9
10
IL(f) Differential insertion loss, -SDD21 magnitude (dB)
11
RL1(f) Differential input return loss, -SDD11 magnitude (dB) 12
RL2(f) Differential output return loss, -SDD22 magnitude (dB) 13
NEXTm(f) Differential near-end crosstalk loss (mth aggressor), -SDD21 magnitude (dB)
14
15
FEXTn(f) Differential far-end crosstalk loss (nth aggressor), -SDD21 magnitude (dB) 16
17
Table 22-7. Calculated Channel Parameters
18
Symbol Description 19
ILfitted(f) Fitted insertion loss (dB)
20
21
ILD(f) Insertion loss deviation (dB) 22
ICN(f) Integrated crosstalk noise (mVRMS) 23
A figure of merit fo the channel that is calculated based on weighted insertion 24
FOMILD loss deviation. 25
26
27
22.2.6.2 Insertion Loss 28
29
Insertion losses for each subchannel of the ENRZ channel, including PCB traces and 30
connectors, shall comply with the limits specified by Equation (22-7), Equation (22-8) 31
and plotted in Figure 22-5. Note that the variable fb is the maximum baud rate to be sup-
32
ported by the channel under test (33.16 Gbaud/s fb 37.50 Gbaud/s). 33
Table 22-8. Channel Insertion Loss Frequency Range 34
35
Parameter Value Units 36
fmin 50 MHz 37
fmax 37.5 GHz 38
39
40
41
42
 f  37.5 f  37.5 fb  43
- + 0.96 ------------------ 
1.083 + 3.35 ----------------- f min  f  ----  44
 fb fb 2
(22-7) 45
IL max =  
 f  37.5 f 46
– 16.8673 + 2 691 ------------------  ---b-  f  f b  47
 fb 2 
  48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 421
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3 
4 0 f min  f  1GHz 
 
5 1  (22-8)
6 IL min =  ---  f – 1   1GHz  f  175 GH z
7 3 
 5 5  175 GH z  f  f b 
8  
9
10
11 Note: f in Equation (22-7) and Equation (22-8) is in GHz.
12
13
14
15 Figure 22-5. CEI-56G-LR normative channel insertion loss at 37.5 Gsym/s
16
17
18
19
CEI‐56G‐LR Channel Loss at 37.5 Gbaud/s
10
20 0
21 ‐10
22
Insertion Loss in dB

B ‐20
23 d
 
n
 is ‐30
24 s
o ‐40
25 L 
n ‐50
26 o
it IL max
re ‐60
27 s IL min
28 In ‐70
29 ‐80
30 ‐90
31 ‐100
32 0 5 10 15 20 25 30 35
33 Frequency in GHz
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

422 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.2.6.3 Fitted insertion loss 1

2
For fitted insertion loss definitions, please refer to Section 12.2.1.1. 3
4
Each subchannel of the ENRZ channel shall meet the insertion loss requirements 5
defined in Table 22-9. Note that the variable fb is the maximum baud rate to be 6
supported by the channel under test.  7
8
Table 22-9. Channel fitted insertion loss characteristics 9
Value 10
Parameter Units 11
Min. Max.
Minimum frequency, fILmin GHz 0.05 -
12
Maximum frequency, fILmax GHz - fb 13
Fitted Insertion loss at Nyquist dB - 33.59 14
Fitted insertion loss, a0 dB -1 2.0 15
Fitted insertion loss, a1 dB 0 24.494 16
Fitted insertion loss, a2 dB 0 75.000 17
Fitted insertion loss, a4 dB 0 53.437 18
19
22.2.6.4 Insertion loss deviation (ILD) 20
21
The insertion loss deviation ILD is the difference between the measured insertion IL 22
and the fitted insertion loss ILfitted as defined in Equation (22-9). 23
24
25
ILD  IL  IL fitted (22-9) 26
27
The insertion loss deviation ILD shall be within the region defined by Equation (22-10) 28
and Equation (22-11) where fb is the maximum baud rate to be supported by the 29
subchannel under test and fILmin and fILmax are given in Table 22-9. 30
31
32
 1.0  12.0( f f b ) f ILmin  f  f b 4  (22-10) 33
ILD  ILDmin   
 4. 0 f b 4  f  (3 / 4) f ILmax  34
 35
1.0  12.0( f f b ) f ILmin  f  f b 4  36

ILD  ILDmax    (22-11) 37
 4.0 f b 4  f  (3 / 4) f ILmax  38
39
40
FOMILD is a figure of merit for the subchannel, and is calculated as indicated below. 41
42
Define the weight at each frequency f using Equation (22-12). 43
44
45
2 1 1 46
W  f  = sinc  f  f b  ------------------------4- -------------------------
-
8
(22-12)
1 +  f  ft  1 +  f  fr  47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 423
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Note that -3 dB transmit filter bandwidth ft is inversely proportional to the minimum

2 20% to 80% rise and fall times T_tr and T_tf. The constant of proportionality is 0.2365
3 (i.e. T_tr x ft = 0.2365), where T_tr is in ns when ft is in GHz). In addition, fr is the -3 dB
4 reference receiver bandwidth, which should be set at (3/4)fb, where fb is the maximum
5 baud rate to be supported by the channel.
6
7 2
8
9
 W  f   I LD  f
FOM ILD = ---------------------------------------------
(22-13)

10 N
11
12 FOMILD is calculated using Equation (22-13) where N is the number of frequency
13 points. The summation is done over the frequency range of ILD with f in GHz. FOMILD
14 shall be less than 0.3dB for each subchannel of valid ENRZ channels.
15
16 22.2.6.5 Channel Return Loss
17
18 The Return Loss for each subchannel of the ENRZ channel shall be bounded by
19 Equation (22-14) as shown in Figure 22-6.
20
21
22 • RL(f) >= 12 dB for fmin < f  fb/4 
23 • RL(f) >= 12 dB - 15 Log10(4f/fb) for fb/4 < f < fb (22-14)
24
25
26
27 Note: fmin is as defined in Table 22-8
28 Figure 22-6. CEI-56G-LR normative channel return loss at 37.5 Gsym/s
29
30
31 CEI‐56G‐LR Channel Return Loss
0
32
33
SDD11, SDD22 (dB)

‐2
34 )
35 B ‐4
d
( 
36 2
2 ‐6
37 D
D
38 S ,
1 ‐8
39 1
D
40 D
S ‐10
41 ‐12
42
43 ‐14
44 0 5 10 15 20 25 30 35
45 Frequency (GHz)
46
47
48
49

424 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.2.6.6 Channel integrated crosstalk noise 1

2
Using the Integrated crosstalk noise method of Section 12.2.1.2 and the parameters of 3
Table 22-10, the total integrated crosstalk noise from neighboring aggressors into the 4
ENRZ channel as measured on any subchannel of the ENRZ channel shall be less 5
than the value specified by Equation (22-15) and illustrated in Figure 22-7. 6
7
Table 22-10. Channel integrated crosstalk aggressor parameters
8
Parameter Symbol Value Units 9
10
max. Baud Rate 11
Baud rate fb Gsym/s
sup. by Channel
12
Near-end aggressor peak to peak single-ended output amplitude Ant 600 mVpp 13
Far-end aggressor peak to peak single-ended output amplitude Aft 600 mVpp 14
15
Near-end aggressor 20% to 80% rise and fall times Tnt 5.5 ps
16
Far-end aggressor 20% to 80% rise and fall times Tft 5.5 ps 17
18
 19
 10 (mV,RMS) 3 dB  IL  6.0 dB  20
 x   x max =  ·  (22-15) 21
 11.8 – 0.30 IL  mV,RMS  6 .0 dB  IL  33.9 dB  22
23
24
In Equation (22-15), the IL denotes the value of the channel insertion loss in dB at 25
0.5 x baud rate (ENRZ). 26
Figure 22-7. Illustration of integrated crosstalk noise limits 27
28
29
Integrated crosstalk noise limits 30
31
Integrated Crosstalk Noise (mV, rms)

)S 12
M 32
R
, 10
V 33
m
 (e 34
si 8 35
o
 n 36
kl 6
at 37
ss 38
o
rc 4
  39
d
et 40
ar 2
ge 41
t
In 0 42
3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 43
Insertion Loss at Nyquist (dB) 44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 425
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.2.6.7 Insertion Loss to Conversion Loss Ratio (ICOR)

2
3 The Insertion Loss to Conversion Loss Ratio (ICOR) of a subchannel, also called the
4 ENRZ Mode Conversion, accounts for crosstalk of one subchannel into another
5 subchannel of the same ENRZ channel.
6
7 The Insertion Loss (IL) and Conversion Loss (CL) of a subchannel can be calculated
8 from the S-Parameters of the channel as described in Annex 22.D.3. ICOR is calculated
9 as follows:
10
11
12 ICoR Di = IL Di – CL Di (22-16)
13
14 where i = 0, 1, or 2 is the subchannel designation.
15
16 Subchannels must meet the following requirement:
17
18 ICoR Di  20dB  f = 0 to fb/2  (22-17)
19
20 for each subchannel i = 0 to 2.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

426 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.3 Electrical Characteristics 1

2
The electrical signaling is based on high speed low voltage logic with a nominal 3
impedance of 50  between any signals of the quad channel and AC ground. 4
5
All devices shall work within the range from 33.16 Gsym/s to 37.50 GBaud/s as 6
specified for the device, with the baud rate tolerance as per Section 3.2.11. 7
Implementation of specific protocols will define the operating baud rate without 8
affecting CEI compliance. 9
10
11
22.3.1 Transmitter Characteristics 12
Transmitter electrical specifications at compliance point T (see Figure 22-4) are given 13
in Table 22-11. The transmitter shall satisfy jitter requirements specified in Table 22-12. 14
Jitter is measured as specified in Annex 22.B.2, for a BER as specified in Section 15
22.2.4. 16
17
Link budgets in this document assume optimized TX FIR equalization as part of the 18
system management function. Specific implementations are outside the scope of this 19
document. 20
21
Table 22-11. Transmitter Electrical Output Specification. 22
23
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 24
Baud Rate T_Baud 33.16 37.50 Gsym/s
25
26
Output Max. Peak-to-Peak Voltage
Swing (Single-Ended)
T_VSEpp See Note 1 600 mVpp 27
Output Common Mode Voltage T_Vcm See Note 2 -100 1700 mV
28
29
Common Mode Noise T_Ncm See Note 3 15 mVRMS
30
Transmitter Skew T_Skew See Note 4 0.06 UI
31
Single-Ended Output Return Loss T_S22 See Section 22.3.1.3 dB 32
NOTES: 33
1. See Equation (22-4) for definition of VSEpp. These limits shall apply for all equalizer settings.
2. See Equation (22-1) for definition of VCM. Load is defined in Annex 22.B.1.1.
34
3. Tested using the test equipment setup defined in Annex 22.B.2.1 using the procedure defined in Section 12.3. 35
4. Time between zero crossings of any two wires of the multi-wire interface for the same baud symbol transition, measured with
skew compensation disabled.
36
37
38
Table 22-12. Transmitter Output Jitter Specification 39
40
Characteristic Symbol Condition MIN. TYP. MAX. UNIT 41
Clock Random Jitter, RMS T_CRJrms See Annex 22.B.2.2 0.005 UI
42
43
Clock Deterministic Jitter T_CDJ See Annex 22.B.2.2 0.05 UI
44
Even-Odd Jitter T_EOJ See Annex 22.B.2.3 0.03 UI 45
Signal to Noise Distortion Ratio T_SNDR See Annex 22.B.3 19 dB 46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 427
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.3.1.1 Transmitter Baud Rate

2
3 All devices shall work within the range from 33.16 Gsym/s to 37.50 Gsym/s as specified
4 for the device, with the baud rate tolerance as per Section 3.2.11. Note that
5 implementation of specific protocols will define the operating baud rate without affecting
6 CEI compliance.
7
8
22.3.1.2 Transmitter Amplitude and Swing
9
10 Transmitter single-ended peak-to-peak voltage swing as defined in Equation (22-3)
11 shall not exceed 600 mVpp for any transmitter coefficient configuration. The single-
12 ended transmitter output voltage shall be between -0.3V and 1.9 V with respect to local
13 ground. Transmitter output amplitude shall additionally adhere to the requirements in
14 Section 22.3.1.7.
15
16
17 22.3.1.3 Transmitter Return Loss
18
19 The single-ended return loss shall be better than A0 from f0 to f1 and better than
20 A0 + slope*log10(f/f1) where f is the frequency from f1 to f2. Please refer to Figure 3-1
21 in Section 3.2.10 using the following parameters:
22 Table 22-13. Transmitter Single-Ended Return Loss Parameters
23
24 Parameter Value Units
25 A0 -12 dB
26 f0 50 MHz
27 0.1714 x
28 f1
T_Baud
Hz

29 f2 T_Baud Hz
30 Slope 12.0 dB/dec
31
32
33 22.3.1.4 Transmitter Quad-to-Quad Skew
34
35 Please refer to Section 3.2.7. For an ENRZ interface, this requirement restricts the
36 skew between any two quad wiring groups.
37
38 22.3.1.5 Transmitter Wire-to-Wire Skew
39
40 The transmitter implements active skew compensation on each wire with a range of at
41 least 0.40 T_Baud, and a step size no greater than 1 ps. Link budgets in this document
42 assume optimized transmit skew compensation as part of the system management
43 function. Specific implementations are outside the scope of this document.
44
45
46 22.3.1.6 Transmitter Short Circuit Current
47
Please refer to Section 3.2.9.
48
49

428 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.3.1.7 Transmitter output waveform requirements 1

2
The transmitter shall include an equalizer defined as: 3
4
(22-18) 5
H ( Z )  C 1  C 0 z  1  C 1 z  2 6
7
22.3.1.7.1 Summary of requirements 8
9
The normalized amplitudes of the coefficients of the transmitter equalizer (computed 10
per Annex 22.B.1.5) shall meet the requirements in Table 22-14. 11
12
Table 22-14. Coefficient range and step size 13
Normalized Amplitude
14
Coefficient Normalized Step 15
Size (%) 16
Min (%) Max (%)
17
C-1 -20 0 1.25 to 5
18
C1 -25 0 1.25 to 5 19
C0 40 100 1.25 to 5 20
21
The amplitude of a coefficient can be computed by multiplying its normalized amplitude 22
by vf, which is defined in equation Equation (22-29). The "min" is defined as the 23
minimum normalized amplitude of the coefficient that must be supplied by the 24
transmitter to be compliant. The "max" is defined as the maximum normalized 25
amplitude of the coefficient that must be supplied by the transmitter to be compliant. 26
27
The peak output voltage shall not exceed limits specified in Section 22.3.1.2 for any 28
values of |C-1|+|C0|+|C1|. 29
30
31
22.3.1.7.2 Requirements for Transmitter Linearity 32
The Linear Mismatch Ratio (RLM) is calculated as described in Annex 22.B.1.2 and 33
must meet the requirements specified in Table 22-15. The steady state output voltage 34
(vf) and linear fit pulse peak values (p(k)) are calculated as described in Annex 22.B.1.6 35
and must meet the requirements specified in Table 22-15. The normalized RMS linear 36
fit error (e(k)) is calculated as described in Annex 22.B.1.7 and must meet the 37
requirements specified in Table 22-15. 38
39
Table 22-15. Transmitter output waveform requirements 40
41
Parameter Condition Units Value 42
Linear Mismatch Ratio, RLM min - 0.92 43
Steady state output voltage, vf max mVpp 600
44
45
Steady state output voltage, vf min mVpp 400 46
Linear fit pulse peak, pmax min - 0.85 x vf 47
Normalized RMS Linear Fit Error, e max - 0.025 48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 429
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.3.2 Receiver Characteristics

2
3 A compliant receiver shall operate at the specified BER when the input signal at
4 compliance point R meets the electrical specifications given in Table 22-16. Jitter
5 specifications at reference R are listed in Table 22-17.
6
7 Table 22-16. Receiver Electrical Input Specifications
8 Characteristic Symbol Condition MIN. TYP. MAX. UNIT
9
10 Baud rate R_Baud 33.16 37.50 GSym/s

11 Single-ended input voltage R_VSEpp See Note 1 600 mVpp

12 Input Impedance R_Rdin See Note 4 40 50 60 
13 Input Impedance Mismatch R_Rm 10 %
14 Single-Ended Input Return Loss R_S11 See Section 22.3.2.3
15
Input Common Mode Voltage R_Vcm See Note 2 and 3 -150 1750 mV
16
17 NOTES:
1. See Equation (22-4) for definition of VSEpp. This specification applies to each wire of the quad channel.
18 2. Min. T_Vdiff, AC-Coupling or floating load. For floating load, input resistance shall be be 1k
19 3. See Equation (22-1) for definition of VCM.
4. Impedance is between any input of the quad receiver to AC ground.
20
21
22
23 Table 22-17. Receiver Input Jitter Specification
24
25 Characteristic Symbol Condition MIN. TYP. MAX. UNIT

26 Sinusoidal Jitter, Maximum R_SJ-max See Note 1 5 UIpp

27 Sinusoidal Jitter, High Frequency R_SJ-hf See Note 1 0.05 UIpp
28 NOTES:
29 1. The Receiver shall tolerate the sum of these jitter contributions: Total transmitter jitter from Table 22-12;Sinusoidal jitter as
30 defined in Table 22-17;The effects of a channel compliant to the Channel Characteristics (Section 22.2.6).

31
32
33
34 22.3.2.1 Input Baud Rate
35
36 All devices shall work within the range from 33.16 Gsym/s to 37.50 Gsym/s as specified
37 for the device, with the baud rate tolerance as per Section 3.2.11.
38
39
40
41
42
43
44
45
46
47
48
49

430 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.3.2.2 Reference Input Signals 1

2
The receiver shall accept input signal amplitudes produced by a compliant transmitter 3
connected with the minimum attenuation specified in Figure 22-5 to the receiver. This 4
may be larger than the 600 mVpp maximum value of T_VSEpp for the transmitter due to 5
output/input impedances and reflections. 6
7
The minimum input amplitude is defined by the minimum transmitter amplitude, the 8
actual receiver input impedance and the loss of the actual PCB. The receiver 9
interference tolerance test defined in Section 22.3.2.6 tests receiver compliance under 10
minimum amplitude conditions. 11
12
22.3.2.3 Input Resistance and Return Loss 13
14
The single-ended return loss shall be better than A0 from f0 to f1 and better than 15
A0 + slope*log10(f/f1) where f is the frequency from f1 to f2. Please refer to Figure 3-1 16
in Section 3.2.10 using the following parameters. 17
18
Table 22-18. Receiver Single-Ended Return Loss Parameters 19
Parameter Value Units 20
21
A0 -12 dB
22
f0 50 MHz 23
f1
0.1714 x
Hz 24
R_Baud
25
f2 R_Baud Hz 26
Slope 12.0 dB/dec 27
28
22.3.2.4 Input Quad-to-Quad Skew 29
30
Please refer to Section 3.2.8. For an ENRZ interface, this requirement restricts the 31
skew through receivers for any two quad groups. 32
33
34
22.3.2.5 Absolute Input Voltage 35
36
The absolute voltage levels with respect to the receiver ground at the input of the 37
receiver are dependent on the transmitter implementation and the inter-ground 38
difference. The voltage levels at the input of an AC coupled receiver (if the effective AC 39
coupling is done within the receiver) or at the TX side of the external AC coupling cap (if 40
AC coupling is done externally) will be between -0.3 to 2.0V with respect to local 41
ground. 42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 431
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.3.2.6 Receiver Interference Tolerance Test

2
3 The receiver interference tolerance shall consist of four separate tests as described in
4 Annex 22.C.1 with the parameters specified in Table 22-19. The receiver shall satisfy
5 the requirements for interface tolerance specified in Annex 22.C.1 for all tests.
6
7 Table 22-19. Receiver Interference Tolerance Test Parameters
8
Units /
9 Parameter Test 1 Values Test 2 Values Test 3 Values Test 4 Values
Notes
10
10-12 10-12 10-12 10-12
11 Target BER (Test Requirement)
12 mTC (min.) 1.0 0.5 1.0 0.5 See Note 1
13 Amplitude of broadband noise (min. RMS) 5.2 12 5.2 12 mV
14 Applied Sinusoidal Jitter amplitude (min.) R_SJ-max R_SJ-max R_SJ-hf R_SJ-hf See Note 2
15 Applied Sinusoidal Jitter frequency (R_Baud / 106) ±10% (R_Baud / 100) ±10%
16 f See Table 22-9
17 ILmin
18 fILmax See Table 22-9 See Note 3

19 fmin See Table 22-8

20 NOTES:
21 1. The mTC parameter is defined in Annex 22.C.1.4, Equation (22-45).
2. SJ values are defined in Table 22-17.
22 3. Also see Section 22.2.6 for definitions of these parameters.
23
24
22.3.2.7 Receiver Imbalance Tolerance Test
25
26 The receiver imbalance tolerance shall consist of a total of sixteen separate tests as
27 described in Annex 22.C.2. Test parameters are specified in Table 22-20 for four tests
28 on wire i. These tests are repeated for each of the wires W3, W2, W1, and W0 of the
29 quad channel, for a total of sixteen separate tests. The receiver shall satisfy the
30 requirements for imbalance tolerance specified in Annex 22.C.2 for all tests.
31
32 Table 22-20. Receiver Imbalance Tolerance Test Parameters
33
34 Parameter
Test 1 Test 2 Test 3 Test 4
Units / Notes
35 Values Values Values Values
36 Target BER (Test Requirement) 10-12 10-12 10-12 10-12
37 mTC (min.) 1.0 0.5 1.0 0.5 See Note 1
38 Amplitude of broadband noise (min. RMS) 5.2 12 5.2 12 mV
39
fILmin See Table 22-9
40
41 fILmax See Table 22-9 See Note 2

42 fmin See Table 22-8

43 Vbias on wire i Vcm + (0.15 × vf) Vcm - (0.15 × vf)
See Note 3, 4
44 Vbias on other wires Vcm
45 NOTES:
46 1. The mTC parameter is defined in Annex 22.C.1.4, Equation (22-45).
47 2. Also see Section 22.2.6 for definitions of these parameters.
3. Repeat the four tests for each of wires i = W3, W2, W1, and W0 of the ENRZ lane.
48 4. Parameter vf is calculated as described in Annex 22.B.1.6.
49

432 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
22.3.2.8 Receiver Skew Tolerance Test 3
4
The receiver skew tolerance shall consist of a total of eight separate tests as described 5
in Annex 22.C.3. Test parameters are specified in Table 22-21 for each test. The 6
receiver shall satisfy the requirements for imbalance tolerance specified in Annex 7
22.C.3 for all tests. 8
9
Table 22-21. Receiver Skew Tolerance Test Parameters 10
11
Parameter Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8
Units / 12
Notes 13
Target BER (Test Requirement) 10-12 10-12 10-12 10-12 10-12 10-12 10-12 10-12 14
mTC (min.) 1.0 0.5 1.0 0.5 1.0 0.5 1.0 0.5 See Note 1 15
Amplitude of broadband noise (min. RMS) 5.2 12 5.2 12 5.2 12 5.2 12 mV
16
17
fILmin See Table 22-9
18
fILmax See Table 22-9 See Note 2
19
fmin See Table 22-8 20
Preskew on wire A
Nominal + 0.15
Nominal value. Nominal value. Nominal value. 21
× (1 / R_Baud)
22
Preskew on wire B Nominal value.
Nominal + 0.15
Nominal value. Nominal value. 23
× (1 / R_Baud)
See Note 3 24
Preskew on wire C Nominal value. Nominal value.
Nominal + 0.15
× (1 / R_Baud)
Nominal value. 25
26
Nominal + 0.15
Preskew on wire D Nominal value. Nominal value. Nominal value.
× (1 / R_Baud) 27
NOTES:
28
1. The mTC parameter is defined in Annex 22.C.1.4, Equation (22-45). 29
2. Also see Section 22.2.6 for definitions of these parameters.
3. “Nominal value” represents any driver preskew setting such that the driver outputs are approximately aligned with minimal
30
skew between wires. A setting of “Nominal + N ps” indicates the additional skew on this wire must be at least N ps. 31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 433
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.A Annex - Test Patterns

2
3 The following patterns shall be used for the testing of transmitter and receiver
4 compliance, including jitter tolerance and output jitter compliance.
5
6
7 22.A.1 PRBS9 Polynomial
8
9 This pattern is a free running PRBS9 generator defined by the serial polynomial:
10
x9 + x5 + 1
11
12 Three bits of the polynomial are sent on ENRZ subchannels D2, D1, and D0 to
13 generate each baud symbol, as shown in Figure 22-8:
14
15 Figure 22-8. PRBS9 Pattern Generation
16
17
18 +
19
20
+
21 +
22
23
24 S1 S2 S3 S4 S5 S6 S7 S8 S9
25
26
27
28
29 Data to Data to Data to

30 D0 D1 D2

31
32
33 22.A.2 Quad Short Stress Pattern Random (SSPR-Q)
34
35 The SSPR-Q pattern is similar to the SSPR pattern defined in Annex 2.D.2. SSPR-Q
36 differs from SSPR in that the pattern contains an extra pad bits such that the pattern
37 length is divisible by 3.
38
SSPR patterns were chosen to have baseline wander and timing content that are at
39
least as stressful as 10,000 years of random binary. The pattern is described in this
40
section and the pattern bit sequence is defined in Annex 22.A.5.
41
42 • The baseline wander was assessed with a cut-off frequency of baudrate/10,000.
43
44 • The clock content was assessed with a corner frequency of baudrate/1667.
45 • The period of 10,000 years was chosen on the basis of random binary exceeding
46 the baseline wander timing content limits of the short pattern once in 10 years in a
47 network containing 1000 random streams.
48
49

434 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The SSPR-Q pattern is defined as described in Figure 22-9: 1

Figure 22-9. Quad Short Stress Pattern Random (SSPR-Q) 2
3
4
PRBS28
Seed=0080080
CID
1, 72 x0
PRBS28
Seed=FFFFFFF
PRBS28
Seed=0080080
PRBS28
Seed=0080080
CID
0, 72 x1
PRBS28
Seed=FFFFFFF
PRBS28
Seed=0080080
5
Diff. Encoded Diff. Encoded 6
7
5437 bits 73 bits 5437 bits 5469 bits 5437 bits 73 bits 5437 bits 5469 bits 8
9
10
• Total length 32,832 bits 11
• All 228-1 PRBS28 sequences are generated using taps 25 and 28 12
13
• Block 1 is 5437 bits of PRBS28 seed = 0x0080080 and begins with 8 x 0, 1, 11 x 0, 14
1, 12 x 0, 1, and so forth. 15
• Block 2 is 1 followed by 72 x 0 16
17
• Block 3 is 5437 bits of PRBS28 seed = 0xFFFFFFF and begins with 28 x 1, 25 x 0, 18
3 x 1, 22 x 0, and so forth. 19
• Block 4 takes the same sequence as block 1 (extended by 32 bits) and encodes it 20
as follows: 21
22
— A zero causes a change of output 23
— A one causes no change of output 24
25
— The output before the first bit is assumed to have been zero 26
— This block begins 1010101001010101010110101010101011011010 ... 27
28
• Blocks 5 to 8 are the inverse of blocks 1 to 4 respectively. 29
30
22.A.3 Linearity Test Pattern 31
32
The Linearity Test Pattern drives the waveform shown in Figure 22-10 on the driver leg 33
(W3, W2, W1, or W0) that is being tested. 34
35
Figure 22-10. Linearity Test Pattern Waveform 36
37
Vhigh
V+1 38
39
40
V+1/3 41
VCM 42
V-1/3
43
44
V-1
45
Vlow
46
GND 47
16 UI
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 435
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 This pattern is described as follows:

2
3 • The time between each transition of the waveform in Figure 22-10 is 16 UI.
4 • The data driven on subchannels (1, 2, 3) to produce these levels depends on the
5 driver leg that is being tested as described below. The corresponding pattern bit
6 sequences are defined in Annex 22.A.5.
7
8 — Driver W3: (0, 1, 1), (1, 1, 1), (0, 0, 0), (1, 0, 0), (0, 1, 1), (1, 0, 0), (0, 1, 1), (1, 0,
9 0), (0, 0, 0), (1, 1, 1), and repeat.
10 — Driver W2: (1, 1, 0), (0, 1, 0), (1, 0, 1), (0, 0, 1), (1, 1, 0), (0, 0, 1), (1, 1, 0), (0, 0,
11 1), (1, 0, 1), (0, 1, 0), and repeat.
12
13 — Driver W1: (1, 0, 1), (0, 0, 1), (1, 1, 0), (0, 1, 0), (1, 0, 1), (0, 1, 0), (1, 0, 1), (0, 1,
14 0), (1, 1, 0), (0, 0, 1), and repeat.
15 — Driver W0: (0, 0, 0), (1, 0, 0), (0, 1, 1), (1, 1, 1), (0, 0, 0), (1, 1, 1), (0, 0, 0), (1, 1,
16 1), (0, 1, 1), (1, 0, 0), and repeat.
17
18 • Signal voltage levels are measured at the points of the waveform labeled V-1, V-1/3,
19 V+1/3, and V+1 in Figure 22-10. Each measurement is performed 8 UI from the prior
20 transition edge, with a tolerance of ±1 UI.
21
22 22.A.4 Clock Jitter Test Patterns
23
24 Test patterns defined in this section are used to ensure transmitter jitter specifications
25 are met.
26
27 22.A.4.1 Clock Jitter Pattern A
28
29 Clock Jitter Pattern A drives alternating (-1, +1) states on the output driver.
30
31 The data driven on subchannels (D2, D1, D0) to produce these levels depends on the
32 driver leg being tested as described below. The corresponding pattern bit sequences
33 are defined in Annex 2.A.
34
35 — Driver W3: (0, 0, 0), (1, 1, 1), and repeat.
36 — Driver W2: (1, 1, 0), (0, 0, 1), and repeat.
37
38 — Driver W1: (1, 0, 1), (0, 1, 0), and repeat.
39 — Driver W0: (0, 1, 1), (1, 0, 0), and repeat.
40
41
42 22.A.4.2 Clock Jitter Pattern B
43
44 Clock Jitter Pattern B drives alternating (-1, +1) states on the output driver, repeated 15
45 times, and then drives alternating (+1, -1) states on the output driver, repeated 16
46 times, and so forth.
47
48
49

436 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The data driven on subchannels (D2, D1, D0) to produce these levels depends on the 1
driver leg being tested as described below. The corresponding pattern bit sequences 2
are defined in Annex 22.A.5. 3
4
— Driver W3: Repeat (0, 0, 0), (1, 1, 1) 15 times; repeat (1, 1, 1), (0, 0, 0) 16 times; 5
and repeat pattern.
6
— Driver W2: Repeat (1, 1, 0), (0, 0, 1) 15 times; repeat (0, 0, 1), (1, 1, 0) 16 times; 7
and repeat pattern. 8
9
— Driver W1: Repeat (1, 0, 1), (0, 1, 0) 15 times; repeat (0, 1, 0), (1, 0, 1) 16 times; 10
and repeat pattern. 11
— Driver W0: Repeat (0, 1, 1), (1, 0, 0) 15 times; repeat (1, 0, 0), (0, 1, 1) 16 times; 12
and repeat pattern. 13
14
15
22.A.5 Test Pattern Definitions 16
CEI Short Stress Pattern Random (SSPR-Q) 17
18
Below is the definition of the test pattern described in Section 22.A.2 as hexadecimal 19
digits with the most significant bit of each digit transmitted first. Blocks #1 to #4 of the 20
pattern are shown below. Blocks #5 to #8 are the inverse of blocks #1 to #4. 21
22
008008004804802082081249248800000C8000068800032C8001A48800C80C8068868832 23
CB2C9A49248480000A080005A480028808016C8480A08A085A4DA4A882081EC9248E8800 24
0FAC80072C8803E48C81CC0E88FAC7ACF2CFACB64B2C90412481248008800804C8048228
820936C924080800448480260A081165A489A4880C480C86E0868B1E4B2D3EC1244C8C80 25
628E88376FAC98C12C85EC848AA88A0DFECDA6208A09724DA42E020855E124AFAE801D2D 26
A80F440E875647ABDE47AF52C7AD5C4FAC7BE32CFA4FA4B283281269A68812492C880004 27
8C80020E880127AC8081AC8848CC8CA0EA8E9A7BEFA49A4928048006820803292481A600 28
08C96004E826023A91613EE1A68C9EC92E868805AB2C828F24896F600C214606D2B76304
70C7B27F6FA218412B2DA48724080BE044854E260AF3F165D6C7A4B60FA810672E8935E5 29
AC09AC8CC44C8EAE628FBDD76F2536C160C80CA668869952CB253C4920CDE0026A2E0113 30
B5E098E1AE45FECDC6A08A3F3A4DBC6E8205F1A922A7CE03799BE18B544EDD1F63835E47 31
F99AC78354CFB99F2B27566721DE55E2F2CFAF564B2D5E41247AC4807ACE083ACBE49EC9
4C068832C32C9A4DA48482080A092485A4000A884005ECA402A898417EC5A4A88E881ECF 32
AC8E8B2C8FAD248F2C400F64E407443C43D65DE5D64B2CB6412490448001260800816480 33
48A40820D8449261A60016C9600A082605A491628801A76C80C9C088687C4CB2B9E29277 34
6F601CC1460FACB7672C90C5E4816EAC08A1BCC4DAC5AE20CE8DF26BAE26136DF1688027
A2C811AB4889CF10CC7B796AFA0B23D2A523D478C3D77BEDD6CA4836098098645845B469 35
A681724928AE0006DDE003032E01B1A5E0C3C8AE6DD8DDD031E3351BEFA9DC492E73E005 36
DECE02B28BE1726D4EAE1073BDE93EE52A0C9CC7A687AFA92BAD2E076C45E3C400000000 37
0000000003FFFFFFC000001C00000FC000071C0003FFC001C01C00FC0FC071C71C3FFFFF 38
DC000013C00008DC0004E3C0023FDC013C13C08DC8DC4E38E3E3FFFFCFC0001B1C000C3F
C006DC1C0303CFC1B1DB1CC3F03FADC71C2C3FFFD4DC001723C00AE3DC05DFD3C2B214DD 39
722B236E372381F8E3F8E3FFC3FFC01DC01C0F3C0FC76DC71FC03FFE1C1C00EFCFC0791B 40
1C3B1C3FDE3FDC12FC13C851C8D8ADF8E1DC23FEF3D3C096D4DC420723E523E3CCC3CFDA 41
ADDB10FC303971DB1F2FF03E65071CD4D3FFA724C029E02C166E14CA51EB298DEB265E2B
214AF722B1D4E373F73F8EC4EC3F8E38DC3FFFE3DC000FD3C00714DC03FB23C1C223DCFD 42
33D3B14AD4E3B1C73FE3FFEC0FC008C71C04EFFFC239001D3F100F4C790752FB13DC5238 43
D3EC3FE4C8DC0C28E3C6D6FFDF0610127369081E82148EA92B0FBE073724E3E8E03FCAFE 44
1C19D0EFCD75791A6DEB1C902B3F81172C389AE4DFC4DC221E23D32EF3D4A596D7189206
FDC023113C13B98DC8E75E38FDDAFFF130D0078B6503BD04D1E55225ECFC30A8B1DB5ED3 45
F01A84C70CEA2FF6BBB5043661D25856F409AE15444DEBF6622B445737166EE8FA519AF2 46
8DD4D66E372651F8E14DE3FEB22FC0B2351C5239DFEC3F7208DC4E24E3E3F03FCFC71C1B 47
1FFFCC3E001ADCE00CC3BE06ADE4E33C2C3FADD4DC2C3723D4D8E3D721FFD6E2E0161F5E 48
0A6E5AE591C8DC91F8E381E3FFF8EFC003F91C01C31FC0FDBE1C7104EFFF923900303F10
1B1C790C3FFB16DC023A03C13EA1DC8CBAF38E96D6FFA206102B23691723821AE3F92CDF 49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 437
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 C304A21DB21B2F022C257134D0EF8925793C00EB0DC07B363C3A287DDEB6B932B0370A73
2 18F59EBDF596B5259231C0903BFC411E41E49EC4EC068E38C32FFFEDA5000808D00484E5
020A3CD125BDA5808508984AD4C5A1C72E8AFFE5ADD00C8C35068ED9D32F8174A538AD18
3 D54AAB555B55B552B52B56DB6DB4AAAAABB55555CB555513B555724B5544ABB55CB5CB51
4 3B13B724924A4AAAAB2B55558DB55534AB5563B5B54D4B2B58DB8DB34AD4A83B6DB434AA
5 ABF3B555EC4B55024BB57BABCB40C5F3BEC40C4E24EC495A924A924AAB4AAB55BB55B52C
B52B6E3B6DAB54AAA5B5B5522B2B569D8DB48DB4ABA4ABB5C2B5CB17DB13903A925BB44A
6 A2CBCB51E3F3B7456C4A7C4A4B334B2B803B8DD2B4D49EDB8DAC2AD4A67D6DB3F3CAA86C
7 73541A5435E625F307DA0C193A0C67A40C5F02EC40DBF24ECAECA9237234A92493B4AAAA
8 4BB5552BCB556DF3B54A8C4B5B444BB2BCCBC8DF03F248DB6CAA4AAA352B55136DB5722A
AB449D55BCADD52F369D6FC28DCBB7C493CA34AA7313B534124B63E2ABAD65D5C6CE1D15
9 A145F1201C0D28A56CEC724A1254AB02A5B59BD22B3EFE9D863A0DB1D40CA95DEC349182
10 73AB73B445A44BCC22CBF079E3ED91C562BB544DDCB5C8993B126FA492BB82AADCD3D569
11 0E7D48BD73DA7FC47A32B4D010DB8FA3CAD581736D32042AE088FD74E65BC797E2F51825
12 FF773A0AE6440F77ECEDE6221287D982C13BF3E624EC67DA925F3A4AA0C42B50C4FDB7C4
9BAA34AEC513B724724A4AD4AB2B6DB58DAAAB34A55583B2553348A5603A724CB434A83B
13 F3B434EC4BF3924BEC5AABE24255E5AFA506278279DB13B1DA92495A4AAA922B554A9DB5
14 5B4DAB52B8A5B6DD722AA9C49D54D4ADD58DB69D34AA8DE3B544854B5CA05BB13082C921
15 E3E2A84565D40C4E1DEC4945824A9C33AB4D7045B8C58C2D44347EDCF3D2291C7E9CB552
0D3B568CE4B48416BBA0E09CC0D4ED06CD92F9A0BAF9E0FC79C4DB51D48AB75DA75A71A3
16 72356124934C2AAA387D551513D571727D4504B3DC78B879517D119703F3785B6C2102AA
17 783BD53134FD61239BCC295EF07C903D932BB7BA0DCA0C0C930C6C2A1C5A7D054233F85F
18
19
20 Linearity Test Pattern
21
22 Below is the definition of the test pattern described in Section 22.A.3. Every 3 bits are
23 striped across subchannels 1, 2, and 3 respectively, and symbols are transmitted in the
24 order from most significant bit to least significant bit. The pattern depends upon the
25 driver leg being tested.
26 Driver W3:
27
28 000000000000 924924924924 6DB6DB6DB6DB FFFFFFFFFFFF 000000000000 
29 FFFFFFFFFFFF 000000000000 FFFFFFFFFFFF 6DB6DB6DB6DB 924924924924
30
31 Driver W2:
32 DB6DB6DB6DB6 492492492492 B6DB6DB6DB6D 249249249249 DB6DB6DB6DB6
33 249249249249 DB6DB6DB6DB6 249249249249 B6DB6DB6DB6D 492492492492
34
35 Driver W1:
36
B6DB6DB6DB6D 249249249249 DB6DB6DB6DB6 492492492492 B6DB6DB6DB6D 
37 492492492492 B6DB6DB6DB6D 492492492492 DB6DB6DB6DB6 249249249249
38
39 Driver W0:
40
41 6DB6DB6DB6DB FFFFFFFFFFFF 000000000000 924924924924 6DB6DB6DB6DB 
924924924924 6DB6DB6DB6DB 924924924924 000000000000 FFFFFFFFFFFF
42
43
44
45
46
47
48
49

438 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Clock Jitter Pattern A and B 1

2
Below is the definition of the test patterns described in Section 22.A.4. Every 3 bits are 3
striped across subchannels D2, D1, and D0 respectively, and symbols are transmitted 4
in the order from most significant bit to least significant bit. The pattern depends upon 5
the driver leg being tested. 6
7
Since pattern A is 6 bits long (which is not divisible by 4), the two least significant bits of 8
the last digit shown are not included in the sequence. 9
10
Since pattern B is 186 bits long (which is not divisible by 4), the two least significant bits 11
of the last digit shown are not included in the sequence. 12
13
Driver W3: 14
Clock Jitter Pattern A: 1C 15
16
Clock Jitter Pattern B: 1C71C71C71C71C71C71C71F8E38E38E38E38E38E38E38E0 17
18
Driver W2: 19
Clock Jitter Pattern A: C4 20
21
Clock Jitter Pattern B: C71C71C71C71C71C71C71C4E38E38E38E38E38E38E38E38 22
23
Driver W1: 24
Clock Jitter Pattern A: A8 25
26
Clock Jitter Pattern B: AAAAAAAAAAAAAAAAAAAAAA9555555555555555555555554 27
28
Driver W0: 29
Clock Jitter Pattern A: 70 30
31
Clock Jitter Pattern B: 71C71C71C71C71C71C71C7238E38E38E38E38E38E38E38C 32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 439
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.B Annex - Transmitter Compliance Tests

2
3 Test procedures described in this annex are used to test transmitter compliance.
4
5
6 22.B.1 Transmitter Output Waveform Tests
7
8 This section describes measurements to ensure the transmitter output waveform meets
9 specifications.
10
11 22.B.1.1 Test Equipment Setup
12
13 Test equipment used to measure the transmitter output for waveform tests is connected
14 to test point T (either TE or TI) as defined in Section 22.2.6.1 using the termination
15 network described in Figure 22-11. The driver being tested is connected to the test
16 equipment, and the other drivers of the CEI quad are terminated as shown.
17
18 DC resistance values should have a tolerance of 1%.
19
20 The return loss of the network should be better than 20 dB from baud rate divided by
21 1667 to 1.5 times the baud rate unless otherwise noted.
22 Figure 22-11. Driver Test Load
23
24
25 Rc Coax connection
26 to test equipment.
27 W0
Ra Rb
28 Rc
29
30 W1 Rd
Ra Rb
31
32
33
34 Rc
35
36 W2
Ra Rb Rd
37
38 Ra = 3.60 
39 Rb = 46.4 
Rc = 46.4 
40 Rc Rd > 2.0 k
41
42 W3
Ra Rb Rd
43
44
45
46
47
48
49

440 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.B.1.2 Linear Mismatch Ratio 1

2
Transmitter linearity is measured using the waveform described in Annex 22.A.3 and 3
the test load described in Annex 22.B.1.1. Signal voltage levels measured at the points 4
labeled V-1, V-1/3, V+1/3, and V+1 in Figure 22-10 are used in the following calculations. 5
6
The minimum delta between voltage levels for signal states on the wire (Smin) is 7
calculated as: 8
9
min   V +1 – V +1/3   V +1/3 – V –1/3   V –1/3 – V –1   10
S min = ----------------------------------------------------------------------------------------------------------------------- (22-19) 11
2 12
13
14
and the linear mismatch ratio (RLM) is calculated as: 15
16
6  S min 17
R LM = ---------------------------- (22-20)
 V +1 – V –1  18
19
20
Specifications for RLM must be met on each of the W3, W2, W1, and W0 transmitters. 21
22
The voltage levels measured in this section are further used to calculate the average 23
signal voltage (Vavg) and the effective voltage levels for the V-1/3 and V+1/3 wire states 24
(EV-1/3 and EV+1/3) as follows: 25
26
V +1 + V +1/3 + V -1/3 + V -1 27
V avg = ---------------------------------------------------------- (22-21) 28
4 29
V -1/3 – V avg 30
EV -1/3 = ---------------------------- (22-22) 31
V -1 – V avg 32
V +1/3 – V avg 33
EV +1/3 = ----------------------------- 34
V +1 – V avg 35
36
The EV-1/3 and EV+1/3 values are used in Annex 22.B.1.4. 37
38
22.B.1.3 Waveform acquisition for Linear Fit Measurements 39
40
The transmitter under test repetitively transmits the PRBS9 pattern defined in Annex 41
22.A.1. The waveform shall be captured with an effective sample rate that is M times 42
the signaling rate of the transmitter under test. The value of M shall be an integer not 43
less than 7. Averaging multiple waveform captures is recommended. 44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 441
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The captured waveform shall represent an integer number of repetitions of the test
2 pattern totaling N bits. Hence the length of the captured waveform should be M x N
3 samples. The waveform should be aligned such that the first M samples of waveform
4 correspond to the first bit of the test pattern, the second M samples to the second bit,
5 and so on.
6
7 For the PRBS9 pattern defined in Annex 22.A.1: N = 511.
8
9 22.B.1.4 Determining Linear Fit to Measured Waveform
10
11 Unless otherwise specified, waveforms for this procedure are measured as described
12 in Annex 22.B.1.3 with the test setup described in Annex 22.B.1.1.
13
14 Given the captured waveform y(k) and corresponding aligned symbols x(n), define the
15 M-by-N waveform matrix Y as shown in Equation (22-23).
16
17
18 y1 yM + 1   yMN – 1 + 1 (22-23)
19
Y = y2 yM + 2   yMN – 1 + 2
20    
21
22 y  M  y  2M   y  MN 
23
24 Rotate the symbols vector x by the specified pulse delay T_Dp to yield xr.
25
26
27 x r = x  T_D p + 1  x  T_D p + 2   x  N  x  1   x  T_D p  (22-24)
28
29
30
Define the matrix X to be an N-by-N matrix derived from xr as shown in Equation (22-
31
25).
32
33
34 xr  1  xr  2   xr  N – 1  xr  N  (22-25)
35
36 x  N  x  1   x  N – 2  x  N – 1 
X = r r r r
37     
38 xr  2  xr  3   xr  N  xr  1 
39
40
41
42 Define the matrix X1 to be the first T_Np rows of X concatenated with a row vector of 1's
43 of length N. The M-by-(T_Np + 1) coefficient matrix, P, corresponding to the linear fit is
44 then defined by Equation (22-26).
45
46 T T –1
47 P = YX 1  X 1 X 1  (22-26)
48
49 In Equation (22-26) the superscript "T" denotes the matrix transpose operator.

442 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

Define the error matrix E as shown Equation (22-27). 1

2
3
e1 eM + 1  eMN – 1 + 1 (22-27) 4
5
E = PX 1 – Y = e  2  e  M + 2   eMN – 1 + 2
T

    6
7
e  M  e  2M   e  MN   8
9
10
The error waveform, e(k), is then read column-wise from the elements of E. 11
12
Define P1 to be a matrix consisting of the first T_Np columns of the matrix P as shown 13
in Equation (22-28). 14
15
(22-28) 16
p  1  p  M + 1   p  M  T_N p – 1  + 1  17
p  2  p  M + 2   p  M  T_N p – 1  + 2  18
P1 = 19
     20
p  M  p  2M   p  R  T_N p   21
22
23
24
The linear fit pulse response, p(k), is then read column-wise from the elements of P1. 25
26
22.B.1.5 Transmitter Equalization Coefficients 27
28
The coefficients of the transmitter equalizer shall be determined from the measured 29
waveform during the transmitter compliance test using the process described below. 30
31
1. The transmitter under test is preset such that C0 is its maximum value (C0_max) and 32
all other coefficients are zero. 33
2. Capture at least one complete cycle of the PRBS9 pattern as described in Annex 34
22.B.1.3. 35
36
3. Compute the linear fit to the captured waveform (p(k)) per Annex 22.B.1.4 with 37
parameter values T_Dp = 2 and T_Np = 8. 38
39
4. Define tx to be the time where the rising edge of the linear fit pulse, p, from step 3 40
crosses 50% of its peak amplitude. 41
5. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time  42
t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi. 43
44
6. Use pi to compute the vector of coefficients, w, of a T_Nw-tap symbol-spaced 45
transversal filter that equalizes for the transfer function from the transmit function to 46
T per Annex 22.B.1.8. 47
48
For each configuration of the transmit equalizer: 49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 443
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 7. Configure the transmitter under test as required.

2
3 8. Capture at least one complete cycle of the PRBS9 pattern as described in Annex
4 22.B.1.3.
5 9. Compute the linear fit to the captured waveform (p(k)) per Annex 22.B.1.4 with
6 parameter values T_Dp = 2 and T_Np = 8.
7
8 10. Define tx to be the time where the rising edge of the linear fit pulse, p, from step 3
9 crosses 50% of its peak amplitude.
10 11. Sample the linear fit pulse, p, at symbol-spaced intervals relative to the time 
11 t0 = tx + 0.5 UI, interpolating as necessary to yield the sampled pulse pi.
12
13 12. Equalize the sampled pulse, pi, using the coefficient vector, w, computed in step 6
14 per Annex 22.B.1.8 to yield the equalized pulse qi.
15
16 The normalized amplitude of coefficient C-1 is the value of qi at time t0 + (T_Dw - 1) UI.
17 The normalized amplitude of coefficient C0 is the value of qi at time t0 + T_Dw UI. 
18 The normalized amplitude of coefficient C1 is the value of qi at time t0 + (T_Dw +1) UI.
19
20 Specifications for transmitter equalization coefficients shall be met for all for each of the
21 W3, W2, W1, and W0 transmitters.
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

444 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.B.1.6 Steady State Voltage and Linear Fit Pulse Peak 1

2
For each configuration of driver equalization coefficients, capture at least one complete 3
cycle of the PRBS9 pattern as described in Annex 22.B.1.3. 4
5
Compute the linear fit pulse response p(k) using the procedure in Annex 22.B.1.4 with 6
parameter values T_Dp = 2 and T_Np = 8. The aligned symbol values x(n) used to 7
construct xr should be values of either +1, EV+1/3, EV-1/3, or -1. The values of EV+1/3 8
and EV-1/3 were calculated in Annex 22.B.1.2. 9
10
The steady-state voltage vf is defined as: 11
12
M  T_Np 13
1 (22-29) 14
v f = -----
M  pk
15
k=1 16
Specifications for vf and specifications for p(k) values relative to vf shall be met for all 17
transmitter configurations and for each of the W3, W2, W1, and W0 transmitters. 18
19
20
22.B.1.7 Linear Fit Error 21
22
For each configuration of driver equalization coefficients, capture at least one complete 23
cycle of the PRBS9 pattern as described in Annex 22.B.1.3. 24
25
Compute the linear fit pulse response (p(k)) and the error between the linear fit and the 26
measured waveform (e(k)) using the procedure in Annex 22.B.1.4 with parameter 27
values T_Dp = 2 and T_Np = 8. The aligned symbol values x(n) used to construct xr 28
should be values of either +1, EV+1/3, EV-1/3, or -1. The values of EV+1/3 and EV-1/3 29
were calculated in Annex 22.B.1.2. 30
31
Specifications for vf and specifications for the RMS value of e(k) normalized to the peak 32
value of p(k), designated e, shall be met for all transmitter equalizer settings and for 33
each of the W3, W2, W1, and W0 transmitters. 34
35
22.B.1.8 Removal of the Transfer Function between the Transmitter and T 36
37
Rotate sampled pulse response pi by the specified equalizer delay T_Dw to yield pr as 38
shown in Equation (22-30). 39
40
41
p r = p i  T_D w + 1  p i  T_D w + 2   p i  N  p i  1   p i  T_D w  (22-30) 42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 445
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Define the matrix P2 to be a T_Np-by-T_Np matrix derived from pr as shown in Equation

2 (22-31).
3
4
5 pr  1  p r  T_N p   p r  3  p r  2  (22-31)
6
7 pr  2  pr  1   pr  4  pr  3 
X =
8     
9 p r  T_N p  p r  T_N p – 1   p r  2  p r  1 
10
11 
12 
13 Define the matrix P3 to be the first T_Nw rows of P2. Define a unit pulse column vector
14 xp of length T_Np. The value of element xp(T_Dp + 1) is 1 and all other elements have a
15 value of 0. The vector of filter coefficients w that equalizes pi is then defined by
16 Equation (22-32).
17
18 T –1 T
w =  P3 P3  P3 xp (22-32)
19
20 Given the column vector of equalizer coefficients, w, the equalized pulse response q is
i
21 determined by Equation (22-33).
22
23
24 qi = P3 w
(22-33)
25
26
27 22.B.2 Transmitter Output Jitter
28
29 This section describes measurements to ensure transmitter jitter requirements are met.
30
31 22.B.2.1 Test Equipment Setup
32
33 Test equipment used to measure the transmitter output for jitter tests is connected to
34 test point T (either TE or TI) as defined in Section 22.2.6.1 using the test load described
35 in Annex 22.B.1.1.
36
37 22.B.2.2 Deterministic and Random Clock Jitter
38
39 Clock random jitter (CRJrms) and clock deterministic jitter (CDJ) are measured as
40 specified in this section.
41
42 Both CRJrms and CDJ are measured with the driver transmitting tthe Clock Jitter
43 Pattern A test pattern defined in Annex 22.A.4.1. Note that the test pattern varies based
44 on which transmitter (W3, W2, W1, or W0) is being tested. The pattern length shall be
45 at least 107 symbols. Using appropriate test equipment, the zero crossings TZC(i) are
46 captured.
47
48
49

446 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The average pulse width Tavg is calculated as follows: 1

2
T ZC  N  – T ZC  1  3
T avg = ------------------------------------------ (22-34) 4
N–1 5
6
7
The jitter series (j) is calculated as follows: 8
9
  j  = T ZC  j  –   j – 1   T avg  – T ZC  1  j = 2 3  N 10
(22-35) 11
Filter the jitter samples with the effect of a single-pole high-pass filter (20 dB per 12
decade low-frequency response) with -3 dB gain at 1.6 MHz and a peak gain of 3 dB at 13
6 MHz. The filtered jitter samples are designated HPF(j). 14
15
Create a CDF as a function of HPF(j). From this CDF, determine J5 as the difference 16
between HPF at the (1 - 0.5 x 10-5) and (0.5 x 10-5) probabilities respectively. 17
Determine J6 as the difference between HPF at the (1 - 0.5 x 10-6) and (0.5 x 10-6) 18
probabilities respectively. 19
20
Using J5 and J6, calculate CRJrms and CDJ as follows: 21
22
23
CRJ rms 1.0538 – 1.0538 J 6 (22-36) 24
= 25
CDJ – 9.3098 10.3098 J 5 26
27
Specified limits for CRJrms and CDJ shall be met for all driver outputs (W3, W2, W1, 28
W0). 29
30
Specifications for CRJrms and CDJ shall be met for all transmitter equalizer settings and 31
for each of the W3, W2, W1, and W0 transmitters. 32
33
22.B.2.3 Even-Odd Jitter 34
35
Even-Odd Jitter (EOJ) is measured as specified in this section. 36
37
EOJ is measured with the driver transmitting the Clock Jitter Pattern B test pattern 38
defined in Annex 22.A.4.2. Note that the test pattern varies based on which transmitter 39
(W3, W2, W1, or W0) is being tested. 40
41
Using appropriate test equipment, the zero crossings TZC(i) are captured for each of 42
the 60 transitions. Averaging of the vertical waveform or of each zero-crossing time 43
may be used to mitigate effects of uncorrelated noise and jitter. The zero-crossing times 44
are denoted as TZC(i), i = 1, 2, ... 60, where i = 1 corresponds to the V+1 to V-1 transition 45
that follows the two consecutive V+1 wire states. 46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 447
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 A set of 40 pulse widths is calculated using the following equation:

2
3
4  T ZC  j + 10  – T ZC  j + 9  1  j  20 (22-37)
T  j  = 
5  T ZC  j + 19  – T ZC  j + 18  21  j  40
6
7
8 EOJ is then calculated as follows:
9
10 20 20
11 (22-38)
12  T  2j  –  T  2j – 1 
13 j=1 j=1
EOJ = -----------------------------------------------------------------------
-
14 40
15
16 Specifications for EOJ shall be met for all transmitter equalizer settings and for each of
17 the W3, W2, W1, and W0 transmitters.
18
19
20 22.B.3 Transmitter Output Noise and Distortion
21
22 The signal-to-noise-and-distortion ratio (SNDR) is measured as described in this
23 section.
24
25 For each configuration of driver equalization coefficients, capture at least one complete
26 cycle of the PRBS9 pattern as described in Annex 22.B.1.3.
27
Compute the linear fit pulse response (p(k)) and the error between the linear fit and the
28
measured waveform (e(k)) using the procedure in Annex 22.B.1.4 with parameter
29
values T_Dp = 2 and T_Np = 8. The aligned symbol values x(n) used to construct xr
30
should be values of either +1, EV+1/3, EV-1/3, or -1. The values of EV+1/3 and EV-1/3
31
were calculated in Annex 22.B.1.2.
32
33 Note that waveform averaging should not be used when calculating p(k) and e(k) for
34 Determine the value of S
min using Equation (22-16) in Annex 22.B.1.2.
35
36 Let k be the index to the peak of the linear fit pulse response p(k) where
37 p(k )p= max(p(k)). Error variance is for each m in the range m = k - M/4 to k + M/4
p p p
38 using the following equation:
39
40 N–1
41 2 1 2 (22-39)
42  m = -----  e  mod N  m – 1 + nM  + 1 
M
43 n=0
44
45 where M and N were defined in Annex 22.B.1.2. If M/4 is not an integer, then the value
46 is rounded up to the next largest integer. The modN(x) function returns x modulo N.
47
48
49

448 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

The SNDR is then calculated as follows: 1

2
3
S min (22-40) 4
SNDR = -----------------------
max   m  5
6
7
Specifications for SNDR shall be met for all transmitter equalizer settings and for each 8
of the W3, W2, W1, and W0 transmitters. 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 449
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.C Annex - Receiver Compliance Tests

2
3 Test procedures described in this section are used to test receiver compliance.
4
5
6 22.C.1 Receiver Interference Tolerance
7
8 This section defines the test setup and test procedures for testing interference
9 tolerance of the receiver. The purpose of this test is to ensure interoperability of the
10 receiver with any transmitter operating within specified limits.
11
12 22.C.1.1 Test Equipment Setup
13
14 The receiver interference tolerance test is performed using the test setup shown in
15 Figure 22-12 or its equivalent.
16 Figure 22-12. Receiver Interference Tolerance Test Setup
17
18
19
20
21
Frequency
22 Synthesizer
23
24 Modulation Port
25
26 Clock
27 Source
28
29
30 Test Receiver
Test
31
32
Transmitter
(ENRZ)
Channel + Under
Test
T R
33
34
Channel
35 Noise Source
36
37
38
39
40
41
42 Points T and R in Figure 22-12 are equivalent to the compliance points described in
43 Section 22.2.6.1.
44
45
46
47
48
49

450 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.C.1.2 Test Transmitter 1

2
The test transmitter meets the transmitter specifications of the Section 22.3.1. 3
4
The test transmitter is configured as follows for receiver interference tolerance tests: 5
• The VSEpp signal swing on any wire at point T shall not exceed the minimum value 6
of the steady state output voltage (vf) as defined in Table 22-15 of Section 7
22.3.1.7.2 when transmitting any of the test patterns defined in Annex 22.A.4.1 8
regardless of equalization settings. 9
10
• Transmit equalization shall be set to optimal values to maximize eye opening at 11
sample points within the receiver. These settings may be determined through 12
simulations using the Test Channel, or by measuring eye metrics and updating 13
transmit equalization coefficients through a transmitter control method. The details 14
of how the optimal transmit coefficient values are determined is outside the scope of 15
this specification. 16
17
The test transmitter may be calibrated such that the signal at T meets the above 18
requirements using the test equipment setup shown in Figure 22-13. The driver test 19
load in this figure is implemented by the circuit shown in Figure 22-11 or its equivalent 20
Calibration is performed such that the above conditions are met on each of the ENRZ 21
wires W3, W2, W1, and W0. The scope in this figure measures the single-ended signal 22
referenced to GND. 23
24
Figure 22-13. Configuration for Test Transmitter Calibration
25
26
Test
27
Driver
Transmitter Test Load Scope 28
(ENRZ) T 29
30
31
32
22.C.1.3 Frequency Synthesizer and Clock Source 33
34
The purpose of the frequency synthesizer and clock source are to inject jitter on the 35
clock used by the test transmitter. This jitter is presumed to transfer onto the output 36
signals at point T. 37
38
The frequency synthesizer and clock source are configured as follows for receiver 39
interference tolerance tests: 40
41
• The clock frequency of the clock source shall offset such that the baud rate of the 42
test transmitter is offset by ±100 ppm relative to the baud rate at which the receiver 43
is nominally set to operate. 44
• The frequency synthesizer shall modulate the clock source to generate sinusoidal 45
jitter at point T with the amplitude and frequency specified by Section 22.3.2.6. 46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 451
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The frequency synthesizer and clock source may be calibrated such that the signal at T
2 meets the above requirements using the test equipment setup shown in Figure 22-14.
3 The ENRZ to differential NRZ conversion shown in this figure can be implemented by
4 the resistor network shown in Figure 22-15 or its equivalent. Calibration is performed
5 such that the above conditions are met on each of the ENRZ subchannels D2, D1, and
6 D0. The scope in this figure measures the differential signal for the subchannel.
7
8 Figure 22-14. Configuration for Pattern Generator Calibration
9
10
Frequency
11 Synthesizer
12
13 Modulation Port
14
15 Clock
16 Source
17
18
19 Test ENRZ to
20 Transmitter Diff. NRZ Scope
21 (ENRZ) T Conversion
22
23
24
25 Figure 22-15. ENRZ to Differential Test Conversion Network
26
27 Rb
W3
28 P0
Rb
29 W2 D0 subchannel:
30 Ra
Rb (W3+W2)-(W1+W0)
31 W1
N0
32 Rb
W0
33
34 Rb
35
36 Rb P1
37 Ra D1 subchannel:
38 Rb (W3+W1)-(W2+W0)
39 N1
Rb
40
41
42 Rb
43 P2
Rb
44
45 Ra D2 subchannel:
Ra = 226 ohms Rb (W3+W0)-(W2+W1)
46
47 Rb = 110 ohms Rb N2
48
49

452 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.C.1.4 Test Channel 1

2
The test channel is a four wire ENRZ channel where traces are routed as two 3
differential pairs (stripline). The reference PCB trace differential impedance is 100 . 4
for each differential pair, with loose coupling between the differential pairs. The test 5
channel should be constructed such that it accurately represents the insertion loss and 6
group delay characteristics of differential traces on an FR-4 printed circuit board. 7
8
The test channel is specified with respect to the insertion loss and return loss. Assume 9
the insertion loss of the test channel (ILTC) is measured at N uniformly-spaced 10
frequencies fn spanning fILmin to fILmax with a maximum frequency spacing of 10 MHz. 11
The following equations calculate transmission magnitude parameters mTC and bTC: 12
13
1
m X = ----  IL max  f n  14
N (22-41) 15
n 16
1 17
m Y = ----  IL TC  f n  (22-42)
N 18
n
19
1 20
m XY = ----  IL max  f n IL TC  f n  (22-43)
N 21
n
1 22
m XX = ----  IL max  f n IL max  f n  (22-44) 23
N 24
n
25
m XY – m X m Y 26
m TC = ------------------------------- (22-45)
m XX – m X m X 27
28
29
b TC = m Y – m TC m X (22-46) 30
31
32
The values of fILmin to fILmax are specified by Section 22.3.2.6. ILmax is defined in 33
Section 22.2.6.2. The test channel shall meet the following requirements: 34
• The value of mTC shall be greater than the value specified by Table 22-19. 35
36
• The channel return loss at test points T and R shall be greater than or equal to 37
20 dB from fmin to fILmax, where fmin and fILmax are specified by Table 22-19. 38
39
22.C.1.5 Interference Noise Source 40
41
The interference noise source emulates crosstalk and other noise which cannot be 42
equalized by the receiver. It is sufficient in the test setup to connect the interference 43
noise source to one wire of the ENRZ quad at a time since these distortions will be 44
seen by all subchannels. If the noise source is connected to more than one wire of the 45
ENRZ quad, then the noise for each wire must be uncorrelated to the other wires. If the 46
noise source is connected to one wire at a time, then the test should be repeated for 47
each of the four possible injection points. 48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 453
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 The interference noise source generates white Gaussian noise with an adjustable
2 amplitude. The power spectral density of the interference noise source shall be flat
3 (within a tolerance of ± 3 dB) across the frequency range from fILmin to 0.5 times the
4 baud rate.
5
6 The amplitude of the interference noise source may be calibrated using the test setup
7 shown in Figure 22-16. The filter shown in the figure shall have a cutoff frequency that
8 is greater than or equal to 0.5 times the baud rate, and a roll-off of 40 dB per decade or
9 less.
10
Figure 22-16. Configuration for Interference Noise Source Calibration
11
12
13
Interference
14 Noise Source Filter Scope
15
16 R
17
18
19 22.C.1.6 Test Procedure
20
21 The receiver interference tolerance test is performed as follows:
22 • The test transmitter is configured as described in Annex 22.C.1.2.
23
24 • The frequency synthesizer and clock source are configured as described in Annex
25 22.C.1.3.
26 • The test transmitter drives the SSPR-Q test pattern defined in Annex 22.A.2.
27
28 • The amplitude of the noise injected by the interference noise source shall be greater
29 than or equal to the amplitude specified by Table 22-19.
30 • The BER measured at the output of the receiver shall be less than the target BER
31 specified by Table 22-19.
32
33
34 22.C.2 Receiver Imbalance Tolerance
35
36 This section defines the test setup and test procedures for testing imbalance tolerance
37 of the receiver. The purpose of this test is to ensure interoperability of the receiver with
38 any transmitter and channel operating within specified limits.
39
40 22.C.2.1 Test Equipment Setup
41
42 The receiver imbalance tolerance test is performed using the test setup shown in
43 Figure 22-17 or its equivalent. Points T and R in Figure 22-17 are equivalent to the
44 compliance points described in Section 22.2.6.1.
45
46
47
48
49

454 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
Figure 22-17. Receiver Imbalance Tolerance Test Setup
2
3
Test Receiver 4
Test 5
Transmitter
(ENRZ)
+ Channel + Under
Test 6
T R
7
8
Vbias Channel 9
Noise Source
10
11
12
13
22.C.2.2 Test Transmitter 14
15
Refer to Annex 22.C.1.2 for a definition of the test transmitter used in this test. 16
17
22.C.2.3 Bias Voltage 18
19
The bias voltage in Figure 22-17 is intended to shift the Vcm voltage on one wire of the 20
ENRZ quad relative to other wires. This will alter the switching points for the differential 21
receivers in the receiver under test and induce skew between subchannels. The bias 22
voltages that are to injected are specified by Section 22.3.2.7. 23
24
Figure 22-18. Configuration for Bias Voltage Calibration
25
26
Test Driver 27
Transmitter
(ENRZ)
+ Test Load Scope 28
T 29
30
Vbias 31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 455
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 Figure 22-19. Bias Network

2 Vdd
3
4
5 Ra1
6 A To Test
R Channel
7 a2
8
9
10
11 Rb1
12 B To Test
13 R b2 Channel
14
15
16
17
Rc1
18 C To Test
19 Rc2 Channel
20
21
22
23
Rd1
24 D To Test
25 Rd2 Channel
26
27
28
29
30 Bias voltages may be injected using the resistor network shown in Figure 22-19 or
31 equivalent circuitry. The bias voltage may be calibrated using the test setup shown in
32 Figure 22-18. The driver test load in this test setup is defined in Annex 22.C.1.1.
33
34
35 22.C.2.4 Test Channel
36
37 Refer to Annex 22.C.1.4 for a definition of the test channel used in this test.
38
39 22.C.2.5 Interference Noise Source
40
41 Refer to Annex 22.C.1.5 for a definition of the interference noise source used in this
42 test.
43
44 22.C.2.6 Test Procedure
45
46 The receiver imbalance tolerance test is performed as follows:
47
48 • The test transmitter is configured as described in Annex 22.C.2.2.
49

456 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

• The bias voltages are configured as described in Annex 22.C.2.3. Bias voltage 1
values are specified by Table 22-20. 2
3
• The test transmitter drives the SSPR-Q test pattern defined in Annex 22.A.2. 4
• The amplitude of the noise injected by the interference noise source shall be greater 5
than or equal to the amplitude specified by Table 22-20. 6
7
• The BER measured at the output of the receiver shall be less than the target BER 8
specified by Table 22-20. 9
10
22.C.3 Receiver Skew Tolerance 11
12
This section defines the test setup and test procedures for testing skew tolerance of the 13
receiver. The purpose of this test is to ensure interoperability of the receiver with any 14
transmitter and channel operating within specified limits. 15
16
22.C.3.1 Test Equipment Setup 17
18
The receiver skew tolerance test is performed using the test setup shown in Figure 22- 19
20 or its equivalent. Points T and R in Figure 22-20 are equivalent to the compliance 20
points described in Section 22.2.6.1. 21
22
Figure 22-20. Receiver Skew Tolerance Test Setup 23
24
25
Test Receiver 26
Test
Transmitter Channel + Under
Test
27
(ENRZ) T R
w/ adjustable
28
preskew 29
Channel 30
Noise Source 31
32
33
22.C.3.2 Test Transmitter 34
35
The test transmitter used for this test must meet all of the requirements described in 36
Annex 22.C.1.2, and must additionally permit provisioning of preskew on individual 37
drivers of the ENRZ quad. The preskew settings used for the test are specified by 38
Section 22.3.2.8. 39
40
22.C.3.3 Test Channel 41
42
Refer to Annex 22.C.1.4 for a definition of the test channel used in this test. 43
44
45
22.C.3.4 Interference Noise Source 46
47
Refer to Annex 22.C.1.5 for a definition of the interference noise source used in this 48
test. 49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 457
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.C.3.5 Test Procedure

2
3 The receiver skew tolerance test is performed as follows:
4 • The test transmitter is configured as described in Annex 22.C.3.2.
5
6 • Transmitter preskew is configured as specified by Table 22-21.
7
8 • The test transmitter drives the SSPR-Q test pattern defined in Annex 22.A.2.
9 • The amplitude of the noise injected by the interference noise source shall be greater
10 than or equal to the amplitude specified by Table 22-21.
11
12 • The BER measured at the output of the receiver shall be less than the target BER
13 specified by Table 22-21.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

458 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.D Appendix - Network Analysis Measurement 1

2
This appendix describes methods of measuring 8-port S-Parameters necessary for 3
channel analysis. 4
5
6
22.D.1 Appendix - S Parameter Measurement with an 8-port VNA 7
8
An 8-port Vector Network Analyzer (VNA) can be used to measure the channel 9
characteristics of an ENRZ Channel using the connections shown in Figure 22-21. 10
Figure 22-21. S-Parameter Measurement Using an 8-port VNA 11
12
ENRZ Channel 13
1 5
14
2 6 15
3 7 16
4 8 17
18
19
8-Port Vector 20
Network Analyzer 21
22
23
The single-ended response between each pair of connection points is expressed as Sji, 24
where j is the output port and i is the input port. The resulting responses are elements 25
of an 8 by 8 matrix. 26
27
28
22.D.2 Appendix - S Parameter Measurement with an 4-port VNA 29
30
The Sji responses measured in Appendix 22.D.1 can alternatively be measured using a
31
4-port VNA. Responses are measured using the 6 setup configurations shown in
32
Figure 22-22. These responses are used to fill in the 8 port S-parameter matrix as
33
follows:
34
 35
S 11 S 12 S 13 S 14 S 15 S 16 S 17 S 18 36
Measurement #1 Setup
Measurement #2 Setup
37
S 21 S 22 S 23 S 24 S 25 S 26 S 27 S 28 Measurement #3 Setup 38
S 31 S 32 S 33 S 34 S 35 S 36 S 37 S 38 Measurement #4 Setup
(22-47) 39
Measurement #5 Setup
Measurement #6 Setup 40
S 41 S 42 S 43 S 44 S 45 S 46 S 47 S 48 41
S =
S 51 S 52 S 53 S 54 S 55 S 56 S 57 S 58 42
43
S 61 S 62 S 63 S 64 S 65 S 66 S 67 S 68 44
S 71 S 72 S 73 S 74 S 75 S 76 S 77 S 78 45
46
S 81 S 82 S 83 S 84 S 85 S 86 S 87 S 88 47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 459
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3 Figure 22-22. S-Parameter Measurement Using a 4-port VNA
4
5 1
ENRZ Channel 5
6 2 6
Term 3 7 Term
7 Term
4 8 Term
8
9 4-Port Vector
Network Analyzer
10 Measurement #1 Setup
11
12 ENRZ Channel 5 Term
Term 1
13 Term 2 6
Term
3 7
14 4 8
15
16 4-Port Vector
Network Analyzer
17 Measurement #2 Setup
18
19 ENRZ Channel
1 5 Term
20 2 6
Term
21 Term 3
4
7
8
22 Term

23 4-Port Vector
24 Network Analyzer

25 Measurement #3 Setup

26
ENRZ Channel
27 Term 1
5
6
Term 2
28 3 7 Term
4
29 8 Term

30 4-Port Vector
31 Network Analyzer

32 Measurement #4 Setup
33
34 1
ENRZ Channel 5 Term
2 6 Term
35 3 7 Term
36 4 8 Term
37
4-Port Vector
38 Network Analyzer
39 Measurement #5 Setup
40
41 ENRZ Channel 5
Term 1
42 Term 2
3
6
7
43 Term
Term
4 8
44
45 4-Port Vector
Network Analyzer
46
Measurement #6 Setup
47
48
49

460 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.D.3 Calculation of Insertion Loss and Conversion Loss 1

2
Given the 8-port S-Parameters of a channel, the Insertion Loss and Conversion Loss 3
terms for each subchannel can be calculated using the following equations: 4
5
6
1 1 1
7
1 –1 –1
S 11 S 12  S 18 8
–1 1 –1
S 21 S 22  S 28 – 1
9
–1 1 = SDD21
0 0 0 0 1 1 –1 –1 D0 FEXT D1  D0 FEXT D2  D0 10
    0 0 0
11
S 81 S 82  S 88 0 0 0
0 0 0
12
0 0 0
13
14
1 1 1 15
S 11 S 12  S 18
1 –1 –1 16
–1 1 –1 17
S S  S 28 – 1 –1 1 = FEXT 18
0 0 0 0 1 – 1 1 – 1 21 22 D0  D1 SDD21 D1 FEXT D2  D1 (22-48)
    0 0 0 19
S 81 S 82  S 88 0 0 0 20
0 0 0 21
0 0 0 22
1 1 1
23
1 –1 –1 24
S 11 S 12  S 18
–1 1 –1 25
S S  S 28 – 1 –1 1 = FEXT 26
0 0 0 0 1 – 1 – 1 1 21 22 D0  D2 FEXT D1  D2 SDD21 D2 27
    0 0 0
S 81 S 82  S 88 0 0 0 28
0 0 0 29
0 0 0 30
31
32
33
The insertion loss for each subchannel is: 34
35
IL D0 = – SDD21 D0 36
(22-49) 37
IL D1 = – SDD21 D1 38
IL D2 = – SDD21 D2 39
40
41
and the conversion loss for each subchannel is: 42
43
44
CL D0 = FEXT D1  D0 + FEXT D2  D0 45
(22-50)
CL D1 = FEXT D0  D1 + FEXT D2  D1 46
47
CL D2 = FEXT D0  D2 + FEXT D1  D2 48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 461
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1 22.E Appendix - Printed Circuit Board Reference Geometry

2
3 Channels must be designed such that the S-Parameters for each subchannel meet the
4 requirements described in Clause 22.2.6, including requirements for insertion loss,
5 insertion loss deviation (ILD), and return loss. This appendix describes an example
6 reference geometry for printed circuit board routing which can meet these
7 requirements, however there are many variations of this reference geometry that can
8 also meet these requirements.
9
10
11 22.E.1 Appendix - Reference Geometry
12
13 Stripline configurations are recommended for PCB routing of ENRZ channels. While
14 microstrip configurations are possible, stripline configurations exhibit better signal
15 integrity performance.
16
A PCB routing reference geometry for an ENRZ channel using a stripline configuration
17
is shown in Figure 22-23. The differential impedance (Zdiff) of each subchannel is
18
100Each differential pair is routed as a loosely coupled stripline with 5/10/5 mil
19
spacing. Spacing between differential pairs should be at least 20 mil to ensure minimal
20
mutual inductance and capacitance between the wires of the two pairs.
21
22 These recommendations are based upon FR4 PCB material with the following
23 characteristics:r = 3.7, and tanD = 0.019.
24
25 Figure 22-23. Reference PCB Routing Geometry for ENRZ Channel
26
27
28
29 ~ 6 mil 5 mil
30
31
32 ~ 6 mil 10 mil 20 mil
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

462 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface
Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

22.E.2 Appendix - Channel Skew 1

2
The transmitter is expected to actively compensate for skew on the interface, including 3
channel skew, as defined in Section 22.3.1.5, and the receiver is expected to tolerate 4
skew consistent with the test limits specified in Section 22.3.2.8. Skew in the channel 5
must not be greater than the compensation range of the transmitter, and additionally 6
should be minimized to avoid degradation of noise immunity or EMI. Excessive skew 7
will also degrade ILD and return loss characteristics of the channel such that channel
compliance may be impacted. Degradations are minimal if the channel meets the 8
following guidelines: 9
10
• Maximum end-to-end channel skew across four ENRZ wires is less than 8 ps. 11
• Maximum routing length difference between wires of any single PCB segment is 12
less than 5 mil. 13
14
If necessary, zigzag routing may be employed to minimize length mismatch of PCB 15
traces. An example of zigzag routing to match lengths through a right angle turn is 16
shown in Figure 22-24. 17
18
Figure 22-24. Length Matching Example Through a Right Angle Turn 19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The glass weave effect introduces additional skew which cannot be ignored for these 37
skew ranges. Possible mitigation approaches include: using denser weave material; 38
using glass material with a lower Dk value; using a wider trace; or employing angled 39
routing at ~10 degree or zigzag routing. Because the skew mismatch due to glass 40
weave may vary widely based on the PCB material characteristics and manufacturing 41
process, specified guidelines cannot be provided in this appendix. 42
43
Note that all of the channel skew guidelines in this section are informative, and 44
compliance is optional as long as the channel requirements in Section 22.2.6 are met. 45
46
47
48
49

Optical Internetworking Forum - Clause 22: CEI-56G-LR-ENRZ Long Reach Interface 463
 Implementation Agreement OIF-CEI-04.0 Common Electrical I/O (CEI)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

464 Optical Internetworking Forum- Clause 22: CEI-56G-LR-ENRZ Long Reach Interface