DigitalCommthr Compiled Suma
DigitalCommthr Compiled Suma
CHAPTER-4
Line Codes
In base band transmission best way is to map digits or symbols into pulse
waveform. This waveform is generally termed as Line codes.
RZ: Return to Zero [ pulse for half the duration of Tb ]
NRZ Return to Zero[ pulse for full duration of Tb ]
1 0 1 0 1 1 1 0 0
Unipolar
NRZ
Polar NRZ
NRZ-inverted
(differential
encoding)
Bipolar
encoding
Manchester
encoding
Differential
Manchester
encoding
Unipolar (NRZ)
Unipolar NRZ
Unipolar NRZ
“1” maps to +A pulse “0” maps to no pulse
• Poor timing
• Low-frequency content
• Simple
• Long strings of 1s and 0s ,synchronization problem
Polar - (NRZ)
Polar NRZ
“1” maps to +A pulse “0” to –A pulse
• Better Average Power
• simple to implement
• Long strings of 1s and 0s ,synchronization problem
• Poor timing
Bipolar Code
NRZ-
Bipolar
V
1 0 1 0 0 1 1 0 1
0
1 0 1 0 1 1 1 0 0
Manchester
Encoding
• “1” maps into A/2 first for Tb/2, and -A/2 for next Tb/2
• “0” maps into -A/2 first for Tb/2, and A/2 for Tb/2
• Every interval has transition in middle
– Timing recovery easy
• Simple to implement
• Suitable for satellite telemetry and optical communications
Differential encoding
• It starts with one initial bit .Assume 0 or 1.
• Signal transitions are used for encoding.
M-ary formats
Bandwidth can be properly utilized by employing M-ary formats. Here
grouping of bits is done to form symbols and each symbol is assigned some
level.
Example
Polar quaternary format employs four distinct symbols formed by dibits.
Gray and natural codes are employed
1. Ruggedness
2. DC Component
3. Self Synchronization.
4. Error detection
5. Bandwidth utilization
6. Matched Power Spectrum
Symbol 1=+a
A =[ Symbol 0 =−a
k
Polar
Alternate Symbol 1 takes =+a, −a
Bipolar A
k
=[ Symbol 0 =0
PSD & auto correlation function form Fourier Transform pair & hence auto
correlation function tells us something about bandwidth requirement in frequency
domain.
2 ∞
V (f) ∑ R (n)e − j2π fnT
1
Sx (f)=
T n = −∞ A
Where V(f) is Fourier Transform of basic pulse V(t). V(f) & RA(n) depends on
different line codes.
Consider unipolar form with symbol 1’s and 0’s with equal probability i.e.
P(Ak=0) = ½ and P(Ak=1) = ½
For n=0;
Probable values of Ak.Ak = 0 x 0 & a x a
=E [ Ak.Ak-0]
= E[Ak2] = 02 x P [ Ak=0] + a2 x P[Ak=1]
RA(0) = a2/2
If n ≠ 0
Ak.Ak-n will have four possibilities (adjacent bits)
0 x 0, 0 x a, a x 0, a x a with probabilities ¼ each.
E[Ak.Ak-n] = 0 x ¼ + 0 x ¼ + 0 x ¼ + a2 / 4
= a2 / 4
V(t) is rectangular pulse of unit amplitude, its Fourier Transform will be
sinc function.
2 ∞
V (f) ∑ R (n)e − j2π fnT
1
Sx (f)=
T n = −∞ A
substituting the values of V(f) and RA(n)
∞ − j2π fnT
S (f)= 1 T 2 Sinc2 (fT ) ∑ R (n)e b
X T b b n = −∞ A
b
2
∞ − j2π fnT
= T Sinc (fT ) R (0) + ∑ R (n) e b
b b A A
n = −∞
n ≠0
2
2
a
= T S in c( f T ) + a 2 ∞ e− j 2π f n Tb
b b 2 4 ∑
n= − ∞
n≠0
a 2 2 a 2 2 ∞ − j2π fn Tb
= T Sinc (fT ) + T Sinc (fT ) ∑ e
4 b b 4 b b n= −∞
2 2 ∞
S (f) = a T Sinc 2 (fT ) + a T Sinc 2 (fT ) 1 ∑ δ(f − n )
X 4 b b 4 b b T n = −∞ T
b b
∞ n)
∑ δ(f −
n= −∞ T is Dirac delta train which multiplies Sinc function
b
which
1 2
has nulls at ± , ± . . . . . . . . . .
Tb Tb
∞
As a result, Sin 2 (fT ). ∑ δ(f − n ) = δ(f)
b n =−∞ Tb
For n=0
E[AK2] = a x a P(AK = a) + (0 x 0) P[AK = 0] +
(-a x –a) P(AK = -a)
= a2/4 + 0 + a2/4 = a2/2
For n>1, 3 bits representation 000,001,010 . . . . . . 111. i.e. with each probability
of 1/8 which results in
E[AK.AK-n] = 0
a2 / 2 n=0
Therefore RA(n) = -a2 / 4 n = ±1
0 n>1
2 ∞
V (f) ∑ R (n)e − j2π fnT
1
Sx (f)=
T n = −∞ A
PSD is given by
1 2 j2π fnTb - j2π fnTb
S x (f) = T SinC 2 (fT ) R A ( −1)e + R A (0) + R A (1)e
T b b
b
b
[
Sx (f) = T SinC 2 (fT )
b 2
]
a 2 a 2 j2π fnTb
−
4
(e +e
−j2π fTb
)
a 2T
S x (f) = b SinC 2 (fT ) [1 −Cos(2 πfT b ) ]
2 b
a 2 T
[ ]
S x (f) = b SinC 2 (fT ) 2Sin 2 ( fTb)
2 b
[ ][
S x (f) = a 2 T SinC 2 (fT ) Sin 2 ( fTb)
b b
]
Tb
Transmitted Waveform Pulse Dispersion
BASEBAND TRANSMISSION:
PAM signal transmitted is given by
∞ -------------------------- (1)
x(t) = ∑ a V(t − KTb )
K = −∞ K
The output pulse μ P(t) is obtained because input signal ak .V(t) is passed
through series of systems with transfer functions HT(f), HC(f), HR(f)
The receiving filter output y(t) is sampled at ti = iTb. where ‘i’ takes intervals
i = ±1, ±2 . . . . .
∞
y (iT b ) = µ ∑a
K =−
∞
k P (iT b −KT b )
∞
y (iT b ) = µ a i P (0) + µ ∑a k P (iT b −KT b ) ---------------------(4)
K =−
∞
K=i K≠i
In equation(4) first term μai represents the output due to ith transmitted bit.
Second term represents residual effect of all other transmitted bits that are
obtained while decoding ith bit. This unwanted residual effect indicates ISI. This
is due to the fact that when pulse of short duration T b is transmitted on band
limited channel, frequency components of the pulse are differentially attenuated
due to frequency response of channel causing dispersion of pulse over the
interval greater than Tb.
In absence of ISI desired output would have y (ti) = μai
∞
1
pδ ( f ) =
Tb
∑ p( f
n =−∞
− n / Tb ) ----------------(6)
∞ ∞
∫ ∑p(mT ) δ (t − mT b ) e dt = p (t ).δ (t )
− j 2πft
pδ ( f ) = b
−∞ m =−
∞
Where m = i-k, then i=k, m=0; so
∞
∫ p(0) δ(t ) e
− j 2πft
pδ ( f ) = dt
−∞
Using property of delta function
∞
i.e ∫ δ(t ) dt
−∞
=1
Therefore pδ ( f ) = p (0) =1
Pδ(f) = 1 ------------(7)
p(0) =1 ,i.e pulse is normalized (total area in frequency domain is unity)
∞
1
Or ∑ p( f
n =−∞
− n / Tb ) = Tb =
Rb
----------- (8)
Ideal Solution
Ideal Nyquist filter that achieves best spectral efficiency and avoids ISI is
designed to have bandwidth as suggested
1 f
P(f) = rect
2B0 2B0
1 if f < B0
P(f) = 2B0
0 f > B0
Impulse response in time domain is given by
sin(2π B0 t)
P(t) =
2π B0 t
=sinc(2B t)
0
Practical solution
Raised Cosine Spectrum
• To design raised cosine filter which has transfer function consists of a flat
portion and a roll off portion which is of sinusoidal form
1 is an adjustable value between B
• Bandwidth =
B 2 o and 2Bo.
Tb
0
1 f < f1
2Bo
P(f) =
1 f1≤ f < 2Bo−f 1
π f −f1
4Bo1+cos
2Bo−2f1
0 f ≥ 2Bo − f1
The frequency f1 and bandwidth Bo are related by
f
α = 1− 1
cos(2 π αBot)
P(t) = sinc(2Bot)
1 −16 α Bo 2 t 2
P(t) has two factors
• 1
second factor that decreases as 2
- helps in reducing tail of sinc pulse
i.e. fast decay t
P(t) =
(
sinc 4 B0 t )
• For α =1, 1 −16 B0 2 t 2
At t=Tb p(t)=0.5
2
Pulse width measured exactly equal to bit duration T b. Zero crossings
occur at t = ±3Tb, ±5Tb… In addition to usual crossings at t = ±Tb, ±2Tb…
Which helps in time synchronization at receiver at the expense of double the
transmission bandwidth
B = 2Bo- f1
Bo = 1 Nyquist bandwidth
2Tb
f1
But α = 1-
B0
using
f1 = B0 (1- α)
B = 2 B0 – B0(1- α)
therefore B = B0(1+ α)
Solution:
α = 0.25 ---- roll off
B = 3.5Khz ---transmission bandwidth
B = Bo(1+ α)
1 Rb
B0 = = Ans: Rb= 5600bps
2Tb 2
Example2
A source outputs data at the rate of 50,000 bits/sec. The transmitter uses
binary PAM with raised cosine pulse in shaping of optimum pulse width.
Determine the bandwidth of the transmitted waveform. Given
a.α = 0 b.α = 0.25 c. α = 0.5 d. α = 0.75 e. α = 1
Solution
B = B0(1+ α) B0=Rb/2
a. Bandwidth = 25,000(1 + 0) = 25 kHz
b. Bandwidth = 25,000(1 + 0.25) = 31.25 kHz
c. Bandwidth = 25,000(1 + 0.5) = 37.5 kHz
d. Bandwidth = 25,000(1 + 0.75) = 43.75 kHz
e. Bandwidth = 25,000(1 + 1) = 50 kHz
Example 3
A communication channel of bandwidth 75 KHz is required to transmit
binary data at a rate of 0.1Mb/s using raised cosine pulses. Determine the
roll off factor α.
Rb = 0.1Mbps
B=75Khz
α=?
B = Bo(1+ α)
B0 = Rb/2 Ans : α =0.5
Correlative coding :
So far we treated ISI as an undesirable phenomenon that produces a
degradation in system performance, but by adding ISI to the transmitted signal in
a controlled manner, it is possible to achieve a bit rate of 2Bo bits per second in a
channel of bandwidth Bo Hz. Such a scheme is correlative coding or partial-
response signaling scheme. One such example is Duo binary signaling.
Duo means transmission capacity of system is doubled.
The correlation between the pulse amplitude Ck comes from bk and previous
bk-1 digit, can be thought of as introducing ISI in controlled manner., i.e., the
interference in determining {bk} comes only from the preceding symbol {b k-1}
The symbol {bk} takes ±1 level thus Ck takes one of three possible values
-2,0,+2 . The duo binary code results in a three level output. in general, for M-
ary transmission, we get 2M-1 levels
Transfer function of Duo-binary Filter
The ideal delay element used produce delay of Tb seconds for impulse will have
transfer function e -j 2π f Tb .
Overall transfer function of the filter H(f)
−j 2 π f T b
H(f) =H c (f) +H c (f)e
− j2 π f T b
H(f) =H c (f) 1+e
j πfT −j π f T b
e b
+e −j π f T b
= 2 H c (f) e
2
− j π f Tb
= 2 Hc (f)c
As ideal channel transfer
o sπ( f Tb) e
function
1
1 f ≤
Hc ( f= ) 2 Tb
0 o t h e r w i s e
Thus overall transfer function
− j π f Tb 1
2 c oπ sf T( b) e f ≤
2 Tb
H ( =
H(f) which )has a gradual roll off to the band edge, can also be implemented by
f
practical and 0 realizable analog filtering oFig
t h e r
shows w i s e
Magnitude and phase plot of
Transfer function
Advantage of obtaining this transfer function H(f) is that practical implementation
is easy
Impulse response
Impulse response h(t) is obtained by taking inverse Fourier transformation of
H(f)
∞
h(t) = ∫ H(f)e j 2π f t df
−∞
1
2 Tb
− jπ f Tb
= ∫ 2 cos( π f T b)e [ e j 2π f t ] df
− 1
2 Tb
πt π( t − T b )
sin sin
= Tb + Tb
πt π( t − T b )
T
b Tb
πt πt
sin
sin
= Tb − T b
πt π( t − T b )
T
b Tb
πt
T 2sin
b Tb
h(t ) =
πt( Tb − 1)
Impulse response has two sinc pulses displaced by Tb sec. Hence overall
impulse response has two distinguishable values at sampling instants t = 0 and t
= Tb.
Ck = bk + bk-1
Decoding:
if
^
bk
is estimate of original sequence bk then
^ = − ^
b k Ck b k − 1
Disadvantage
Precoding
In case of duo binary coding if error occurs in a single bit it reflects as multiple
errors because the present decision depends on previous decision also. To
make each decision independent we use a precoder at the receiver before
performing duo binary operation.
The precoding operation performed on the input binary sequence {bk}
converts it into another binary sequence {ak} given by
a k = b k ⊕ a k −1
C k = a k + a k −1
± 2 vi f bk = s y m0 b o l
Ck =
0 v i f bk = s y m1 b o l
The decision rule for detecting the original input binary sequence {bk} from {ck} is
s y 0 mi fCk > b1 ov l
^b =
k s y 1 im fCk ≤ b1 ov l
Example: with start bit as 0, reference bit 1
Example
The binary data 001101001 are applied to the input of a duo binary system.
a)Construct the duo binary coder output and corresponding receiver output,
without a precoder.
b) Suppose that due to error during transmission, the level at the receiver input
produced by the second digit is reduced to zero. Construct the new receiver
output.
c) Repeat above two cases with use of precoder
without a precoder
errors errors
With a precoder (start bit 1)
Ck = ak – ak-2
Receiver
At the receiver we may extract the original sequence {bk} using the decision rule
bk = symbol 0 if |Ck| >1v
symbol 1 if |Ck| ≤1v
- j 4 π f Tb
= Hc(f) 1- e
+ j 2 π f Tb - j 2 π f Tb
- j 2 π f Tb e −e
= 2 jHc(f) e
2j
- j 2 π f Tb
= 2 jHc(f) sin (2π f Tb) e
1
1 f ≤
W h e rHce( f i) s 2 Tb
0 o t h e r w i s e
The Transfer function has zero value at origin, hence suitable for poor dc
channels
- j 2 π f Tb 1
2 sj i ( n2 π f Tb) e f≤
H (= f ) 2 Tb
0 O th e rw is e
Impulse response
∞
h(t) = ∫ H(f)e j 2π f t df
−∞
1
2 Tb
− j π f Tb
= ∫ 2 jsin ( 2π f T b)e [ e j 2π f t ] df
− 12
πt Tb π ( t − 2T b )
sin sin
Tb Tb
= −
πt π ( t − 2T b )
T Tb
b
πt ( πt )
sin sin
= Tb − Tb
πt π( t − 2T b )
T Tb
b
π( t )
2T b 2 sin
Tb
=
πt ( 2T b −has
Impulse response t ) three distinguishable levels at the sampling instants.
To eliminate error propagation modified duo binary employs Precoding option
same as previous case.
Prior to duo binary encoder precoding is done using modulo-2 adder on signals
spaced 2Tb apart
a =b ⊕ a
k k k− 2
Example
The binary data 011100101 are applied to the input of a modified duo binary
system.
a) Construct the modified duobinary coder output and corresponding receiver
output, without a precoder.
b) Suppose that due to error during transmission, the level at the receiver input
produced by the third digit is reduced to zero. Construct the new receiver output.
c) Repeat above two cases with use of precoder
Ck = ∑f
n =0
n b k -n
In base band M-ary PAM, output of the pulse generator may take on any one of
the M-possible amplitude levels with M>2 for each symbol
The blocks of n- message bits are represented by M-level waveforms with
M=2n.
Ex: M=4 has 4 levels. possible combination are 00, 10, 11, 01
M-ary PAM system is able to transmit information at a rate of log2M faster than
binary PAM for given channel bandwidth.
Rb
R=
log2 M
M-ary PAM system requires more power which is increased by factor equal to
M2
log 2 M for same average probability of symbol error.
M-ary Modulation is well suited for the transmission of digital data over channels
that offer a limited bandwidth and high SNR
Example
An analog signal is sampled, quantised and encoded into a binary PCM
wave. The number of representation levels used is 128. A synchronizing
pulse is added at the end of each code word representing a sample of the
analog signal. The resulting PCM wave is transmitted over a channel of
bandwidth 12kHz using binary PAM system with a raised cosine spectrum.
The roll off factor is unity.
a)Find the rate (in BPS ) at which information is transmitted through the
channel.
b) Find the rate at which the analog signal is sampled. What is the
maximum possible value for the highest frequency component of the
analog signal.
Solution
Given Channel with transmission BW B=12kHz.
Number of representation levels L = 128
Roll off α = 1
a) B = Bo(1+ α),
Hence Bo =6kHz.
Bo=Rb/2 therefore Rb = 12kbps.
b) For L=128, L = 2n , n = 7
symbol duration T = Tb log2M =nTb
sampling rate fs = Rb/n = 12/7 = 1.714kHz.
And maximum frequency component of analog signal is
From LP sampling theorem w = fs/2 = 857Hz.
Eye pattern
The quality of digital transmission systems are evaluated using the bit
error rate. Degradation of quality occurs in each process modulation,
transmission, and detection. The eye pattern is experimental method that
contains all the information concerning the degradation of quality. Therefore,
careful analysis of the eye pattern is important in analyzing the degradation
mechanism.
• Eye patterns can be observed using an oscilloscope. The received wave
is applied to the vertical deflection plates of an oscilloscope and the
sawtooth wave at a rate equal to transmitted symbol rate is applied to the
horizontal deflection plates, resulting display is eye pattern as it resembles
human eye.
• The interior region of eye pattern is called eye opening
• The width of the eye opening defines the time interval over which the
received wave can be sampled without error from ISI
Example 1
Adaptive equalization
Adaptive equalization – It consists of tapped delay line filter with set of delay
elements, set of adjustable multipliers connected to the delay line taps and a
summer for adding multiplier outputs.
The output of the Adaptive equalizer is given by
M−1
y(nt) = ∑ci x(nT − iT)
i =−0
Ci is weight of the ith tap Total number of taps are M .Tap spacing is equal to
symbol duration T of transmitted signal
In a conventional FIR filter the tap weights are constant and particular designed
response is obtained. In the adaptive equaliser the Ci's are variable and are
adjusted by an algorithm
Mechanism of adaptation
Training mode
This training sequence has maximal length PN Sequence, because it has large
average power and large SNR, resulting response sequence (Impulse) is
observed by measuring the filter outputs at the sampling instants.
The difference between resulting response y(nT) and desired response d(nT)is
error signal which is used to estimate the direction in which the coefficients of
filter are to be optimized using algorithms
Methods of implementing adaptive equalizer
i) Analog
ii) Hard wired digital
iii) Programmable digital
Analog method
• The set of adjustable tap widths are stored in digital memory locations,
and the multiplications of the analog sample values by the digitized tap
weights done in analog manner.
• Set of adjustable lap weights are also stored in shift registers. Logic
circuits are used for required digital arithmetic operations.
Programmable method
Clock
Shift Shift Shift Output
Register1 Register2 Register3
S0 S1 S 2 S3 Sequence
Logic Circuit
Fig 2: Maximum-length sequence generator for n=3
A feedback shift register is said to be Linear when the feed back logic
consists of entirely mod-2-address ( Ex-or gates). In such a case, the zero state
is not permitted. The period of a PN sequence produced by a linear feedback
shift register with ‘n’ flip flops cannot exceed 2n-1. When the period is exactly 2n-
1, the PN sequence is called a ‘maximum length sequence’ or ‘m-sequence’.
Example1: Consider the linear feed back shift register as shown in fig
2 involve three flip-flops. The input so is equal to the mod-2 sum of S1 and S3. If
the initial state of the shift register is 100. Then the succession of states will be
as follows.
100,110,011,011,101,010,001,100 . . . . . .
The output sequence (output S3) is therefore. 00111010 . . . . .
Which repeats itself with period 23–1 = 7 (n=3)
Maximal length codes are commonly used PN codes
In binary shift register, the maximum length sequence is
N = 2m-1
0 0 0 1 0 0 1 1 0 1 0 1 1 1 1
3 1 2 2 1 1 1 4
1 N
R c (k) = ∑ c c
N n =1 n n -k
Where N is length or period of the sequence and
k is the lag of the autocorrelation
1 i f k= l N
R
Wherec ( k = )
l is any 1 integer. 1
≠
R c (k) =
we can also − state kAutocorrelation
l N function as N
1
Rc (k) = (7 − 8)
15
1
R c (k) =−
15
Yields PN autocorrelation as
7 127
8 255
9 511
10 1023
11 2047
12 4095
13 8191
17 131071
19 524287
A Notion of Spread Spectrum:
An important attribute of Spread Spectrum modulation is that it can
provide protection against externally generated interfacing signals with finite
power. Protection against jamming (interfacing) waveforms is provided by
purposely making the information – bearing signal occupy a BW far in excess of
the minimum BW necessary to transmit it. This has the effect of making the
transmitted signal a noise like appearance so as to blend into the background.
Therefore Spread Spectrum is a method of ‘camouflaging’ the information –
bearing signal.
V
b(t) m(t). . r(t) z(t)
Tb Decisio
∫dt
0
n
Device
+1
0
-1
a) Data Signal b(t)
+1
0
-1
To recover the original message signal b(t), the received signal r(t) is
applied to a demodulator that consists of a multiplier followed by an integrator
and a decision device. The multiplier is supplied with a locally generated PN
sequence that is exact replica of that used in the transmitter. The multiplier
output is given by
Z(t) = r(t).c(t)
= [b(t) * c(t) + n(t)] c(t)
= c2(t).b(t) + c(t).n(t)
The data signal b(t) is multiplied twice by the PN signal c(t), where as
unwanted signal n(t) is multiplied only once. But c2(t) = 1, hence the above
equation reduces to
Z(t) = b(t) + c(t).n(t)
Now the data component b(t) is narrowband, where as the spurious
component c(t)n(t) is wide band. Hence by applying the multiplier output to a
base band (low pass) filter most of the power in the spurious component c(t)n(t)
is filtered out. Thus the effect of the interference n(t) is thus significantly reduced
at the receiver output.
The integration is carried out for the bit interval 0 ≤ t ≤ Tb to provide the
sample value V. Finally, a decision is made by the receiver.
If V > Threshold Value ‘0’, say binary symbol ‘1’
If V < Threshold Value ‘0’, say binary symbol ‘0’
Direct – Sequence Spread Spectrum with coherent binary Phase shift Keying:-
Binary data
Binary PSK
b(t) m(t) x(t)
Modulator
c(t)
PN Code
Generator
Carrier
a) Transmitter
Say 1
Tb Decision
y(t) Coherent v v
if v > 0 Detector ∫dt
0
Device
Received Say 0
Signal c(t) if v < 0
Local PN
generator
Local Carrier
b) Receiver
2
c o πs fc(t 2) k cT≤ t ≤ ( k+ 1 T)c
φk ( t =) Tc
2
~ 0 s i nπ f( t2o) kt hT≤ et ≤ (r k+w1 Ti) s e
φ ( t = ) Tc c c c k =0,1,...... .......... .., N −1
k
where 0 o th e r w is e
T is chip duration,
c
N is number of chips per bit.
where
Tb
jk = ∫ j(t) φ
0
k (t) dt k = 0,1,...... ........N − 1
~ T ~ b
N −1 N −1
~
1 1
2 2
=
Tb
∑
k =0
j +
k Tb ∑
k =0
j
k
∑ jk
k =0
= ∑j
k =0
k
N−1
2 2
J=
Tb
∑ jk
k =0
Tb
2
v=
Tb ∫u(t) cos(2 π f t)dt
0
c
= v s + v cj
Tb
2
vs =
Tb ∫ s(t) cos(2 π f t)dt
0
c
Tb
2
v cj =
Tb ∫ c(t) j(t) cos(2 π f t)dt
0
c
Consider despread BPSK signal s(t)
2E b
s(t) = ± cos(2π fc t) 0 ≤ t ≤ Tb
Tb
Tc N−1 Tb
v cj =
Tb
∑c ∫ j(t) φ
k =0
k k (t) dt
0
Tc N−1
=
Tb
∑c
k =0
k
jk
E[Ck jk jk ] = jk P(C k
=1) −jkP(C k
=−1)
1 1
= jk − jk
2 2
=0
and Variance
[
Var Vc j j =] 1 N−1 2
∑ jk
N k =0
=
JTc
2
2Tb
(SNR) 0 = (SNR) I
Tc
where Tb
PG =
Tc
• .3db term on right side accounts for gain in SNR due to coherent
detection.
• . Last term accounts for gain in SNR by use of spread spectrum.
PG is called Processing Gain
Wc
PG =
Rb
Probability of error
Decision rule is, if detector output exceeds a threshold of zero volts; received bit
is symbol 1 else decision is favored for zero.
1 Eb
Pe ≅ erfc
2 JTc
Antijam Characteristics
Consider error probability of BPSK
1 Eb
Pe = erfc
2 N0
Eb Tb P
=
N0
Tc J
J PG
or =
P Eb / N0
Eb
N
Where 0 is minimum bit energy to noise ratio needed to support a
prescribed average probability of error.
Example1
A pseudo random sequence is generated using a feed back shift register of
length m=4. The chip rate is 107 chips per second. Find the following
a) PN sequence length
b) Chip duration of PN sequence
c) PN sequence period
Solution
a) Length of PN sequence N = 2m-1
= 24-1 =15
b) Chip duration Tc = 1/chip rate =1/107 = 0.1µsec
c) PN sequence period T = NTc
= 15 x 0.1µsec = 1.5µsec
Example2
A direct sequence spread binary phase shift keying system uses a
feedback shift register of length 19 for the generation of PN sequence.
Calculate the processing gain of the system.
Solution
Given length of shift register = m =19
Therefore length of PN sequence N = 2m - 1
= 219 - 1
Processing gain PG = Tb/Tc =N
in db =10log10N = 10 log10 (219)
= 57db
Example3
A Spread spectrum communication system has the following parameters.
Information bit duration Tb = 1.024 msecs and PN chip duration of 1µsecs.
The average probability of error of system is not to exceed 10-5. calculate
a) Length of shift register b) Processing gain c) jamming margin
Solution
Processing gain PG =N= Tb/Tc =1024 corresponding length of shift register m =
10
In case of coherent BPSK For Probability of error 10-5.
[Referring to error function table]
Eb/N0 =10.8
Therefore jamming margin
Eb
(jamm ing mar gin) dB = (Pr ocessi ng gain) dB − 10log 10
N0 min
Eb
(jamm ing mar gin) dB = 10log 10 PG dB − 10log 10
N
0 min
= 30.10 – 10.33
= 19.8 db
Frequency – Hop Spread Spectrum:
In a frequency – hop Spread – Spectrum technique, the spectrum of data
modulated carrier is widened by changing the carrier frequency in a pseudo –
random manner. The type of spread – spectrum in which the carrier hops
randomly form one frequency to another is called Frequency – Hop (FH)
Spread Spectrum.
Since frequency hopping does not covers the entire spread spectrum
instantaneously. We are led to consider the rate at which the hop occurs.
Depending upon this we have two types of frequency hop.
1. Slow frequency hopping:- In which the symbol rate Rs of the MFSK signal
is an integer multiple of the hop rate Rh. That is several symbols are
transmitted on each frequency hop.
2. Fast – Frequency hopping:- In which the hop rate Rh is an integral multiple
of the MFSK symbol rate Rs. That is the carrier frequency will hoop
several times during the transmission of one symbol.
A common modulation format for frequency hopping system is that of
M- ary frequency – shift – keying (MFSK).
Fig. illustrates the variation of the frequency of a slow FH/MFSK signal with time
for one complete period of the PN sequence. The period of the PN sequence is
24-1 = 15. The FH/MFSK signal has the following parameters:
Number of bits per MFSK symbol K = 2.
Number of MFSK tones M = 2K = 4
Length of PN segment per hop k=3
Total number of frequency hops 2k = 8
Fig. illustrates the variation of the transmitted frequency of a fast FH/MFSK signal
with time. The signal has the following parameters:
Number of bits per MFSK symbol K = 2.
Number of MFSK tones M = 2K = 4
Length of PN segment per hop k=3
Total number of frequency hops 2k = 8