THYB01 Proceedings of IPAC2016, Busan, Korea

PERFORMANCE OF SUPERCONDUCTING CAVITIES FOR THE

EUROPEAN XFEL
D. Reschke†, DESY, Hamburg, Germany
for the European XFEL Accelerator Consortium

Abstract used as a quality control tool as well as for further R&D.

The superconducting accelerator of the European XFEL For 8 of these 24 cavities a subsequent assembly of the
consists of the injector part and the main linac. The injec- He-tank was made. In addition 4 of the additional 16
tor includes one 1.3 GHz accelerator module and one cavities used for infrastructure set-up and commissioning
3.9 GHz third-harmonic module, while the main linac have since been fitted with a He-tank for use in the as-
consists of 100 accelerator modules, each housing eight sembly of the 102 cryomodules.
1.3 GHz TESLA-type cavities, operated at an average Both vendors must exactly follow well-defined speci-
design gradient of 23.6 MV/m. The fabrication and sur- fications for the mechanical fabrication and surface treat-
face treatment by industry as well as the vertical and ments, but no cold RF performance guarantee is required.
cryomodule RF tests of the required 808 superconducting The surface preparation at both vendors started with a
1.3 GHz cavities are analysed and presented. bulk electro-polishing (EP) followed by 800° annealing,
but for the final surface treatment two alternative recipes
INTRODUCTION have been used: EZ applied a final chemical surface re-
moval (“Flash-BCP”), while RI applied a final EP. All
The 17.5 GeV SRF linac for the European XFEL is cur-
cavities were fully equipped with their HOM antennas,
rently under construction by a consortium consisting of
pick-up probe and a High-Q input coupler antenna with a
several European institutes [1]. At the beginning of 2015
fixed coupling. All cavity transports took place horizon-
the cryomodule production and testing rate was increased
tally in a dedicated transport box [10] under UHV condi-
from an average of 1 to 1.25 eight-cavity-modules per
tions by truck. No performance degradation after
week, in order to meet the expected tunnel closure date of
transport has been observed. The procedures before and
September 30, 2016. Testing of both individual cavities
after the vertical acceptance test at 2K are described in
and cryomodules is performed in a dedicated test facility
[10]. Once received at DESY, an initial incoming inspec-
at DESY (AMTF) [2,3,4]. In early 2016 all of the 816
tion was performed (mechanical, electrical, warm RF and
series EU-XFEL TESLA-type 1.3 GHz SRF cavities have
vacuum checks). While originally intended to check for
been produced, and have each undergone at least one
transport damages, the incoming inspection proved neces-
vertical acceptance test in AMTF. As of April 15, 2016 87
sary to identify unexpected non-conformities, with the
of the 102 EU-XFEL cryomodules (101 for installation +
result that 54 cavities were sent back to the vendors be-
1 spare) have been tested at AMTF. Vertical and module
fore vertical testing.
testing is performed by a team from IFJ-PAN Krakow as
All 832 tested cavities have clearly demonstrated that
an in-kind contribution. The installation of cryomodules
the chosen scheme for mechanical production and surface
and first steps of commission for the main linac are in full
preparation was successful implemented at both vendors.
swing. The injector commissioning started successfully in
December 2015.
Copyright © 2016 CC-BY-3.0 and by the respective authors

XFEL CAVITIES AND VERTICAL AC-

CEPTANCE TEST
Production Overview
Series production of the 1.3 GHz TESLA cavities was Figure 1: 3-D model of the series XFEL cavity equipped
equally divided between E. Zanon Spa. (EZ), Italy, and for delivery to DESY.
Research Instruments GmbH (RI), Germany. Production Vertical Testing Scheme
included both mechanical fabrication and surface prepara-
tion [5] together with required extensive documentation The vertical acceptance tests of up to 10 cavities per
[6]. Details about the niobium and niobium-titanium ma- week have been made using two independent test sys-
terial used can be found in [7]. The RF measurements for tems, each consisting of an independent bath cryostat and
quality assurance during the cavity production are de- RF test stand. Each test cryostat accepts an “insert” which
scribed in [8]. 804 XFEL series cavities (401 by EZ; 403 supports up to four cavities (Fig. 2), greatly increasing the
by RI) were delivered complete with helium tank (Fig. 1), efficiency of cool-down / warm-up cycles. All 1,225 ver-
ready for vertical testing at DESY in AMTF. Each vendor tical acceptance tests of the 832 cavities are now complete
also produced additional 12 cavities without helium tank (except of potentially few non-conform returns from
for the ILC-HiGrade programme [9], which have been string assembly). Each vertical test was categorized ac-
___________________________________________
†
[email protected]
ISBN 978-3-95450-147-2 07 Accelerator Technology
3186 T07 Superconducting RF
 Proceedings of IPAC2016, Busan, Korea THYB01

cording to well-defined “test reasons”. Depending on the  quench gradient (quench limited);
result a categorized “decision” was taken and documented  gradient at which Q0 drops below 1010 (Q0 lim-
in the cavity and cryomodule managing system [11] of the ited);
AMTF as well as in the XFEL cavity data base [12,13].  gradient at which either X-ray detector ex-
Cavities without non-conformities (see below) and with ceeds the threshold (field-emission limited).
acceptable performance usual have only one vertical ac- At the beginning of production, the criterion for ac-
ceptance test (“as received”) after which they receive the ceptance for module assembly was specified as
decision “send to string assembly”. In case of non- Eacc,us ≥ 26 MV/m. In May 2014 it was reduced to
conformities (e.g. insufficient cavity performance, RF Eacc,us ≥ 20 MV/m, in order to optimise the number of
problem, vacuum leak, mechanical deviation, etc.) the vertical tests while still maintaining an average module
cavity was retested, retreated or sent back to the vendor gradient of 23.6 MV/m [17].
eventually resulting in additional vertical tests [14]. Cavities with Eacc,us < 20 MV/m were considered for
further processing or re-treatment. The exact nature of the
handling of low-performance cavities was judged on a
case-by-case basis. As there was no vendor performance
guarantee, retreatments were in general the responsibility
of DESY. Nevertheless both vendors did agree to perform
several retreatments depending on the case.

VERTICAL TEST RESULTS

‘As received’ from Vendor
Figure 3 shows histograms and yield curves for the ver-
tical test performance for usable gradient “as received”
from the vendors. The final analysis is based on 743 ver-
tical tests (EZ: 368; RI: 375). Table 1 summarises the
average of the distributions shown in Fig. 3. The average
usable gradients for both vendors are above the required
operational gradient for XFEL. The usable gradient is
reduced from the maximum performance by 3.7 MV/m on
average, predominantly due to the Q0-value dropping
Figure 2: Test inserts for vertical testing at AMTF. below 1010. The effect can be seen in Fig. 3 as an increase
(top to bottom plot) in the numbers of cavities with per-
The vertical acceptance tests followed a standardised formance less than 30 MV/m. For both vendors ~13% of
procedure, which included the measurement of the un- the cavity tests “as received” result in a necessary re-
loaded Q-value (Q0) versus the accelerating gradient Eacc treatment due to field emission.
at 2 K, as well as the frequencies of the fundamental There is also a statistically significant difference in the
modes. For each point of the Q0(Eacc)-curve, X-rays were average performance of the two vendors (~3 MV/m for
measured inside the concrete shielding above and below the maximum and usable gradient), and gradients above
the cryostat. No general administrative gradient limit was 40 MV/m have mainly been observed with RI cavities.
applied. The average measurement error is calculated to

Copyright © 2016 CC-BY-3.0 and by the respective authors

The better performance is attributed to the use by RI of
be 3.3 % for Eacc and 6.6 % for Q0 [15]. In general the electro-polishing as the final surface preparation scheme
systematic error of the RF measurement is about ~10% as described above, but also to the fact that RI cavities
for Eacc and up to ~20 % for Q0. showed less quenches at low gradients.
In addition to the Q0(Eacc) curves many cavities had the The percentage (“yield”) of cavities above 26 MV/m
higher-order mode frequencies of the TE111, TM110 and (20 MV/m) usable gradient is 59% (83%) for EZ and 73%
TM011 modes measured [16], depending on the fabrica- (89%) for RI, with a total yield of 66% (86%).
tion process.
After a successfully completed test, selected key data Table 1: Average (±1.std.dev) of the Maximum and Usa-
were transferred to the XFEL Cavity Data Base, which ble Gradient “As received”
forms the basis of the analyses report here. Tests Maximum Eacc Usable Eacc
Definition of “Usable Gradient” and Ac- [MV/m] [MV/m]
ceptance Criteria Total 743 31.4 ± 6.8 27.7 ± 7.2
Although all cavities are tested to their maximum EZ 368 29.8 ± 6.6 26.3 ± 6.8
achievable gradient (Eacc,max), of greater importance for
accelerator operation is the “Usable Gradient” (Eacc,us), RI 375 33.0 ± 6.6 29.0 ± 7.4
which takes Q0 as well as field-emission performance into
account. It is defined [17] as the lowest value of:
07 Accelerator Technology ISBN 978-3-95450-147-2
T07 Superconducting RF 3187
 THYB01 Proceedings of IPAC2016, Busan, Korea

 Non-conformity during string assembly (e.g. during

Maximum power coupler assembly), in which case the cavity
Gradient was shipped back to DESY and in general a HPR ap-
plied.
Performance “Send to string assembly”
The average usable gradient of the last vertical test be-
fore transport to the string assembly facility at CEA Sac-
lay is about 30 MV/m (Fig.4). The percentage (“yield”) of
cavities above 26 MV/m (20 MV/m) usable gradient is
79% (97%). As the cavities have been assigned to cry-
omodules by their performance, three cryomodules will
have a gradient below 20 MV/m.

Usable
Gradient

Figure 3: Comparison of performance distribution and Figure 4: Performance distribution and yield for usable
yield for maximum gradient (top) and usable gradient gradient “Send to String assembly”.
(bottom) “As received” from RI (red) and EZ (green).
CRYOMODULE TEST RESULTS
The trend in average usable gradient over the produc- The string and module assembly at CEA Saclay is de-
tion until August 2015 is described in [17]. For the final scribed in [21, 22]. As of April 27, 2016 93 cryomodules
months of production both Q-value and usable gradient have been assembled, of which 87 have been RF tested
remained constant compared to the previous production. [3,4,24] at the AMTF (Fig. 5). This includes the pre-series
Impact of “Retreatment” modules XM-1 and XM-2, which are equipped with EU-
XFEL series cavities.

 Non-conformities after delivery from vendor. About

Three categories for retreatments have been identified:

90 cavities showed a mechanical, vacuum, electrical

or other non-conformity, which required a retreat-
Copyright © 2016 CC-BY-3.0 and by the respective authors

ment at DESY or the vendor before the first vertical

test. These do not have an “as received” test.
 Performance. As described above, most cavities with
usable gradients below 20 MV/m underwent re-
treatment [18, 19] with a goal of increasing their per-
formance. Often a high-resolution optical inspection
was performed before the retreatment in order to lo-
calize the limiting defect [20]. Approximately 18% of
all cavities have been retreated and retested due to in-
sufficient performance. In general, high-pressure ul-
tra-pure water rinsing (HPR) is applied as a first re- Figure 5: Cryomodule test-stand installation at AMTF.
treatment. This is particular effective since most low-
performance cavities are dominated by field emis- Comparability of “Usable and Operational
sion, which is likely associated with a removable sur- Gradient” in vertical vs. cryomodule test
face emitter (e.g. particles). The average usable gra-
dient increased from 19 MV/m to 26 MV/m, while A direct comparison between the VT usable gradient as
the Q0(4MV/m) increased from 2.1·1010 to 2.4·1010. defined above and the operational gradient in the cry-
omodule test (CT) is difficult: First, no individual cavity
Q0 performance data is available for the CT; second, the
ISBN 978-3-95450-147-2 07 Accelerator Technology
3188 T07 Superconducting RF
 Proceedings of IPAC2016, Busan, Korea THYB01

geometry of the x-ray monitors in the CT are significantly “Usable / Operational Gradient” in CT vs. VT
different; and third, the individual cavity measurements in In Fig. 7 the average operational gradients for all cry-
the CT are limited to 31 MV/m by the RF power system. omodules tested so far are presented and compared to the
Only cavities observed quench limits in both tests can be respective average vertical test results. For a fairer com-
strictly compared (see [21] for details). parison all vertical test gradients above 31 MV/m are
Table 2: Averages (±1.std.dev) of VT and CT measured clipped before averaging. Table 2 shows the mean opera-
performance (maximum and usable/operational) of all tional gradients over all cryomodules with the CT gradi-
cavities assembled into cryomodules. IMPORTANT: For ent meeting the VT gradients within 4%.
comparison the VT gradients are clipped to the CT limit Except for XM26 the order of assembly is in agreement
of 31 MV/m before averaging. with ascending cryomodule numbering. An average gra-
dient loss can be observed in about the first third of as-
Tests Maximum Eacc Operational sembled cryomodules, which then improved significantly
[MV/m] Eacc for more recently assembled modules. This is due to im-
[MV/m] provements in the cleanroom procedures and additional
VT 695 30.3 ± 1.8 28.7 ± 2.9 operator training, which are described in detail in [22,
]. XM20, XM33, XM45, XM58 and XM68 show the
CT 695 28.7 ± 3.9 27.6 ± 4.5 lowest performance. As XM33, XM58 and XM68 are
equipped
23 with cavities showing VT gradients of
“Maximum gradient” in CT vs. VT MV/m, 23 MV/m and 21MV/m, respectively, no high-
Figure 6 shows the maximum achieved CT gradients er gradients can be expected in CT. The strong degrada-
for all individual tested cavities in comparison to their VT tion of XM45 can be correlated with an accidental loss of
22
test results. The horizontal dashed red line indicates the electricity in the cleanroom during string assembly.
RF power limit in the CT (31 MV/m). In an ideal case all
results should scatter around a line with a slope = 1. A
number of cavities clearly show a reduced performance in
the CT after a good to excellent behaviour in the VT
(lower right section of the plot). The third column of
Table 2 gives the means for the maximum gradients for
the VT and CT respectively. The average systematic RF
measurement error in VT and CT is discussed in [15].
More details and possible correlations of the performance
to non-conformities during the module assembly process Figure 7: Average cryomodule operational gradients
are given in [22]. (orange) compared to the respective average vertical test
results (blue). IMPORTANT: For comparison the VT
gradients are clipped to the CT limit of 31 MV/m before
averaging.
Quality Factor at Operational Gradient in CT
vs. VT

Copyright © 2016 CC-BY-3.0 and by the respective authors

The dynamic cryogenic heat load at 2K of a module is
dominated by the Q0-values (i.e. their surface resistance)
of the cavities at their operational gradient. Figure 8
shows the CT effective average Q0-values at (20-
23.6) MV/m calculated from the cryo losses in compari-
son to the expected average Q0-values from the vertical
tests. As a main result all cryomodules except of XM34
and XM70 meet the EU-XFEL design goal of ≥ 1·1010.
The mean Q0-values for CT and VT are equal at
1.4×1010, despite the large scatter in both. Exceptions
where there is a significantly higher CT Q0-value were
either caused by an enhanced measurement error at low
heat loads due to a poor “signal to background” ratio, or
may have been caused by a dependence of the RF losses
on the cooldown procedure.
Figure 6: Individual CT – VT comparison for the maxi-
mum gradient. The horizontal red dashed line indicates
the power limit in the CT (note some early tested were
allowed to exceed this value).
07 Accelerator Technology ISBN 978-3-95450-147-2
T07 Superconducting RF 3189
 THYB01 Proceedings of IPAC2016, Busan, Korea

effect of the less accurate tunnel RF infrastructure, and

further calibration measurements are planned.
A spare 3.9 GHz module is already under fabrication
with plans for cw tests [30].

CW R&D ON EUROPEAN XFEL

CRYOMODULES
The design operation of the European XFEL is a short
pulse mode with a maximum RF pulse length of 1400 µs
Figure 8: Average cryomodule Q0-values measured by (including rise time) and a repetition rate of 10 Hz (a duty
cryogenic heat load measurement at an operational gra- factor of 1.4%). Reducing the average gradients of the
dient of (20-23) MV/m. cryomodules, the technology has the potential for much
larger duty factors up to 100% (limited by the tolerable
heat load) which will make the facility even more attrac-
INJECTOR CRYOMODULE OPERATION tive for users. A series cryomodule XM4 has been tested
The EU-XFEL injector and its commissioning progress extensively in the separate cryomodule test bench
are described in detail in [25]. Since the end of 2015 both (CMTB) at different operation temperatures and after two
the 1.3 GHz cryomodule and the 3.9 GHz third-harmonic different cooldown procedures. A summary of the excel-
system (Fig. 9) are in operation at 2 K. lent results is given in Table 3 and more details can be
found in [31]. In all conducted tests no cavity quench was
observed for long-pulse and cw operation. The maximum
demonstrated heat load was 71 W in cw mode at about
15 MV/m and stable operation. The cw activities at DESY
will be continued and expanded in the near future.
Table 3: Demonstrated maximum gradients and Q0-values
at maximum achieved gradient for the applied operation
Figure 9: 3.9 GHz third-harmonic cryomodule (red) and modes on XM4.
1.3 GHz cryomodule in the injector. Mode Short pulse Long pulse CW
DF=1.4% DF=20% DF=100%
1.3 GHz Injector Cryomodule
Max Eacc 31.8 19 15
The cryomodule XM29 in the injector is routinely op-
[MV/m]
erated at 2 K accelerating a beam up to a maximum ener-
gy of 160 MeV. The cavities have been operated up to an Qo-value@ - 2.0·1010 (2 K) 2.3·1010 (2 K)
average gradient of 21.3 MV/m without any performance max Eacc 3.5·1010 (1.8 K)
limitation. As no higher energy is necessary for a stable
injector operation it was not attempted to reach the aver-
age usable gradient of 27 MV/m achieved in the AMTF SUMMARY
module test. The accelerator cavity production and treatment at both
vendors has been successfully finished. The 1.3 GHz
3.9 GHz Third-Harmonic System
Copyright © 2016 CC-BY-3.0 and by the respective authors

cryomodule assembly at CEA Saclay and subsequent

The third-harmonic system at 3.9 GHz of the EU- testing at DESY is close to being finished and is highly
XFEL injector section is a joint INFN-DESY contribution successful. The installation of cryomodules into the EU-
and consists of a single module housing eight SRF cavi- XFEL is in full swing. Cold commissioning of the injec-
ties and a quadrupole magnet package. Ten cavities have tor started at the end of 2015 with the successful opera-
been vertically tested and qualified for module assembly, tion of both the 1.3 GHz cryomodule as well as the
reaching at least 18 MV/m with an average gradient of 3.9 GHz third-harmonic system. In preparation for possi-
20.8 MV/m and Q0 above 109 at this field level [26,27]. In ble future cw operation, one series 1.3 GHz cryomodule
contrast to the 1.3 GHz cryomodules, only a system test has been tested in cw and long-pulse mode with excellent
of a single “cavity package” [28] consisting of a horizon- results.
tal cavity equipped with power coupler, tuner and wave
guide tuners was successfully performed before the mod- ACKNOWLEDGMENT
ule assembly [29] and its installation into the injector.
The European XFEL cavity and cryomodule fabrica-
In routine operation the 3.9 GHz module runs with
tion, testing and installation is a collaborative effort of
moderate acceleration voltages up to 30 MV. Voltages up
several European institutes and their industrial partners.
to 45 MV have been achieved without cavity limitation.
The author likes to thank the complete team of all in-
At present about 15% lower gradient is observed in the
volved institutes for their work and support.
injector operation as compared to the VT; this may be an

ISBN 978-3-95450-147-2 07 Accelerator Technology

3190 T07 Superconducting RF
 Proceedings of IPAC2016, Busan, Korea THYB01

REFERENCES DESY”, presented at IPAC’16, Busan, Korea, May 2016,

paper WEPMB007, this conference
[1] “XFEL: The European X-Ray Free-Electron Laser - Tech-
nical Design Report”, DESY, Hamburg, Germany, [16] A. Sulimov et al., “Practical Aspects of HOM Damping
DESY_06-097, 2006, doi:10.3204/DESY_06-097 Changes for TM011 of the E-XFEL Cavities”, in Proc.
2015, Whistler, Canada, Sep.2015, paper THPB068
[2] J. Swierblewski, “Large Scale Testing of SRF Cavities and
Modules”, in Proc. Linac14, Geneva, Switzerland, Sep. [17] D. Reschke, “Recent Progress with EU-XFEL”, in Proc.
2014, paper TUIOC01 2015, Whistler, Canada, Sep.2015, paper MOAA02
[3] J. Swierblewski et al., “Improvements of the mechanical, [18] A. Matheisen et al., “Experiences on Retreatment of XFEL
vacuum and cryogenic procedures for European XFEL Series Cavities at DESY”, in Proc. 2015, Whistler, Cana-
Cryomodule Testing”, in Proc. SRF2015, Whistler, Cana- da, Sep.2015, paper MOPB075
da, Sep.2015, paper TUPB115 [19] N.J. Walker et al., “Update and status of vertical test results
[4] M. Wiencek et al., “Improvements of the RF test procedure of the XFEL series cavities”, in Proc. 2015, Whistler, Can-
for European XFEL cryomodules”, in Proc. SRF2015, ada, Sep.2015, paper MOPB086
Whistler, Canada, Sep.2015, paper TUPB118 [20] A. Navitski et al,. “Characterisation of surface defects on
[5] W. Singer et al., “The Challenge and Realization of the E-XFEL series and ILC-HiGrade cavities”, in Proc. 2015,
Cavity Production and Treatment in Industry for the Euro- Whistler, Canada, Sep.2015, paper MOPB072
pean XFEL”, in Proc SRF2013, Paris, France, Sep. 2013, [21] C. Madec et al., “The Challenge to Assemble 100 Cry-
paper MOIOA03 omodules for the European XFEL”, in Proc. SRF2013,
[6] J. Iversen et al., “Release processes and Documentation Paris, France, Sep. 2013, paper THIOA02
Methods during Series Treatment of SRF Cavities for the [22] O. Napoly, “Module performance in XFEL cryomodule
European XFEL by using Engineering data Management mass production”, in Proc. 2015, Whistler, Canada,
System”, in Proc. SRF2015, Whistler, Canada, Sep.2015, Sep.2015, paper FRAA02
paper THPB032 [23] S. Berry et al., “Cleanliness and vacuum acceptance tests
[7] W. Singer et al., “Superconducting cavity material for the for the UHV cavity string of the XFEL linac”, in Proc.
European XFEL”, Supercond. Sci. Technol. 28 085014, 2015, Whistler, Canada, Sep.2015, paper MOPB118
2015 [24] M. Wiencek et al., “Update and status of the test results of
[8] A. Sulimov, “RF measurements for Quality Assurance dur- the XFEL series accelerator modules”, in Proc. 2015,
ing SC Cavity Mass Production”, in Proc. SRF2015, Whis- Whistler, Canada, Sep.2015, paper MOPB080
tler, Canada, Sep.2015, paper WEBA02 [25] F. Brinker, W. Decking, “Commissioning of the European
[9] A. Navitski et al., “ILC-HiGrade Cavities as a Tool of XFEL Injector”, presented at IPAC’16, Busan, Korea, May
Quality Control for European XFEL”, in Proc. SRF2013, 2016, paper TUOCA03, this conference
Paris, France, Sep. 2013, paper MOP043 [26] P. Pierini et al., “Fabrication of the 3.9 GHz SRF Structures
[10] D. Reschke, “Infrastructure, Methods and Test Results for for the European XFEL”, in Proc. 2015, Whistler, Canada,
the Testing of 800 Series Cavities for the European Sep.2015, paper THPB035
XFEL”, in Proc. SRF2013, Paris, France, Sep. 2013, paper [27] D. Sertore et al., “Vertical Tests of XFEL 3rd Harmonic
THIOA01 Cavities”, in Proc. 2015, Whistler, Canada, Sep.2015, pa-
[11] M. Wiencek at al., “Cavities and Cryomodules managing per MOPB077
system at AMTF”, in Proc. 2015, Whistler, Canada, [28] C. Maiano et al., “Horizontal RF Test of a Fully Equipped
Sep.2015, paper TUPB117 3.9 GHz Cavity for the European XFEL in the DESY
[12] S. Yasar et al., “European XFEL database structure and AMTF”, in Proc. 2015, Whistler, Canada, Sep.2015, paper
data loading system for cavities and modules”, in Proc. MOPB076

Copyright © 2016 CC-BY-3.0 and by the respective authors

2015, Whistler, Canada, Sep.2015, paper THPB038 [29] M. Schmökel at al., “Assembly of a 3.9GHz String for the
[13] S. Yasar et al., “XFEL Database User Interface” presented E-XFEL at DESY”, in Proc. 2015, Whistler, Canada,
at IPAC’16, Busan, Korea, May 2016, paper TUPOW001, Sep.2015, paper TUPB0105
this conference [30] L. Monaco et al., “Update on Third Harmonic XFEL Activ-
[14] J. Schaffran et al., “Analysis of test rate for European ities at INFN LASA”, presented at IPAC’16, Busan, Ko-
XFEL series cavities”, in Proc. 2015, Whistler, Canada, rea, May 2016, paper TUPOW005, this conference
Sep.2015, paper MOPB079 [31] J. Sekutowicz, “Performance of LG and FG dressed cavi-
[15] Y. Yamamoto, W.-D. Moeller, D. Reschke, “Error Estima- ties in XFEL cryomodules”, TTC Meeting SLAC, Dec.
tion in Cavity Performance for the European XFEL at 2015, Stanford, USA

07 Accelerator Technology ISBN 978-3-95450-147-2

T07 Superconducting RF 3191