Chapter 4

Baseband Communication Systems

in the Absence of Noise

4.1 Problem Solutions

Problem 4.1

a. Split phase or bipolar RZ;

b. NRZ change or NRZ mark;

c. Split phase;

d. NRZ mark;

e. Polar RZ;

f. Unipolar RZ.

Problem 4.2
This is a matter of adapting Figures 4.2 top and bottom to the data sequence given in the
problem statement.

Problem 4.3
These are a matter of …lling in the details given at the ends of Examples 4.1 through 4.5.

Problem 4.4
This is a matter of adapting Figure 4.2, second plot, to the data sequence given in the
problem statement.

1
2CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Problem 4.5
This is a matter of adapting Figure 4.2, third - …fth plots, to the data sequence given in
the problem statement.

Problem 4.6

a. 4 kbps;

b. 2 kbps;

c. 2 kbps;

d. 4 kbps for …rst null and 2 kbps for second null.

Problem 4.7
The input pulse is

x1 (t) = Au (t) Au (t T)

The response to the …rst step of amplitude A is

y11 (t) = A [1 exp ( t=RC)] u (t)

and the response to the second step, using time invariance, is

y12 (t) = A [1 exp ( (t T ) =RC)] u (t T)

By the superposition property of a linear system, the total response is y1 (t) = y11 (t) y12 (t)
which gives (4:30) :
4.1. PROBLEM SOLUTIONS 3

Problem 4.8
Because PRC (f ) is even, it follows that
Z 1
1
pRC (t) = F [PRC (f )] = 2 PRC (f ) cos (2 f t) df
0
Z (1 )=2T Z (1+ )=2T
T 1
= 2T cos (2 f t) df + T 1 + cos f cos (2 f t) df
0 (1 )=2T 2T
1 1+ 1
sin T t sin T t sin T t
= 2T +T T
2 t 2 t 2 t
Z (1+ )=2T
T 1
+T cos f cos (2 f t) df
(1 )=2T 2T
1 1+
sin T t sin T t
= T +T
2 t 2 t
Z (1+ )=2T
T T 1 T 1
+ cos f + 2 f t + cos f 2 ft df
2 (1 )=2T 2T 2T
1 1+
sin T t sin T t
= T +T
2 t 2 t
h i (1+ )=2T h i (1+ )=2T
T 1 T 1
T sin f 2T + 2 ft T sin f 2T 2 ft
+ T
+ T
2 +2 t 2 2 t
(1 )=2T (1 )=2T
1 1+
sin T t sin T t
= T +T
2h t 2 t i h i
T 1+ 1 1+ T 1 1 1
T sin 2T 2T + T t T sin 2T 2T + T t
+
2 T= + 2 t 2 T= + 2 t
h i i h
T 1+ T 1 1 1+
1 1
T sin t 2T T sin 2T 2T TT t 2T
+
2 T= 2 t 2 T= 2 t
sin ( t=T ) cos ( t=T ) 1 1 1 t t
= + cos sin
t=T 2 t=T T =2 t 1 T =2 t + 1 T T
1 sin ( t=T )
= 1 2 cos ( t=T )
1 (T =2 t) t=T
" #
(T =2 t)2 sin ( t=T ) cos ( t=T )
= 2 cos ( t=T ) = sinc (t=T )
1 (T =2 t) t=T 1 (2 t=T )2
4CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Problem 4.9
From (4:37) the ‡at-top portion of the spectrum goes to f = 1=2T and the transition region
is 0 in frequency extent. The remaining spectrum is 0 for 1=2T < f 1=T . Setting = 0
in (4:36) clearly gives pRC (t) = sinc(t=T ).

Problem 4.10
The MATLAB program is given below. The plot of the desired frequency response functions
follows the program.
% Problem 4_10
% Computation and plotting of transmitter and receiver frequency responses
% for raised cosine signaling through lowpass Butterworth …ltered channel
%
clear all
A = char(’-’,’:’,’–’,’-.’);
clf
f3dB = input(’Enter channel …lter 3-dB frequency, Hz => ’);
n_poles = input(’Enter number of poles for Butterworth …lter repr channel => ’);
Rb = input(’Enter data rate, bits/s => ’);
beta0 = input(’Enter vector of betas (max no = 3) => ’);
Lbeta = length(beta0);
for n = 1:Lbeta
beta = beta0(n);
f1 = 0:.1:(1-beta)*Rb/2;
HR1 = sqrt(1+(f1/f3dB).^(2*n_poles));
f2 = (1-beta)*Rb/2:.1:(1+beta)*Rb/2;
HR2 = sqrt(1+(f2/f3dB).^(2*n_poles)).*cos(pi*(f2-(1-beta)*Rb/2)./(2*beta*Rb));
f3 = (1+beta)*Rb/2:.1:Rb;
HR3 = zeros(size(f3));
f = [f1 f2 f3];
HR = [HR1 HR2 HR3];
plot(f,HR,A(n,:), ’LineWidth’, 1.5),xlabel(’{nitf}, Hz’),...
ylabel(’j{nitH_R}({nitf})j or j{nitH_T}({nitf})j’)
if n == 1
hold on
end
end
title([’Bit rate = ’,num2str(Rb),’bps; channel …lter 3-dB frequency = ’,num2str(f3dB),
’Hz; no. of poles = ’,num2str(n_poles)]),...
if Lbeta == 1
legend([’nbeta = ’,num2str(beta0(1))],1)
4.1. PROBLEM SOLUTIONS 5

Bit rate = 5000 bps; channel filter 3-dB frequency = 5000 Hz; no. of poles = 1
1.4
β=1
1.2 β = 0.5

1
|HR(f )| or |HT(f )|

0.8

0.6

0.4

0.2

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
f, Hz

elseif Lbeta == 2
legend([’nbeta = ’,num2str(beta0(1))],[’nbeta = ’,num2str(beta0(2))],1)
elseif Lbeta == 3
legend([’nbeta = ’,num2str(beta0(1))],[’nbeta = ’,num2str(beta0(2))],...
[’nbeta = ’,num2str(beta0(3))],1)
end
Problem 4.11

1
a. This spectrum has the zero-ISI property. The appropriate sample rate is T = 1:5W .

b. This spectrum does not have the zero-ISI property.

c. This spectrum does not have the zero-ISI property.

1
d. This spectrum has the zero-ISI property. The appropriate sample rate is T = W.
1
e. This spectrum has the zero-ISI property. The appropriate sample rate is T = 3W .
6CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Problem 4.12

a. Consider P (f ) = T2 (T f =2) which is a triangle T =2 units high going from 2=T to

2=T . Clearly, superimposing it with its translates by 1=T; 2=T; ::: produces a
constant height of T =2 so Nyquist’s theorem is satis…ed.

b. Use the transform pair (t) $ sinc2 (f ) with duality to produce the transform pair
sinc2 (t) $ (f ). Then use scale change to get sinc2 (at) $ a 1 (f =a). Apply this
to P (f ) with a = 2=T to get

p (t) = sinc2 (2t=T )

which has zero crossings at t = T =2; T; :::

Problem 4.13
Use (4:37) and set the maximum frequency equal to 7 kHz:
1+
= 7 kHz
2T
1
But T = 9 kbps, so
2 7
1+ =
9
14 5
= 1 = = 0:556
9 9

Problem 4.14

a. The channel response matrix is

2 3
1:0 0:1 0:01
[Pc ] = 4 0:2 1:0 0:1 5
0:02 0:2 1:0

Its inverse is 2 3
1:0217 0:0163 0:0209
[Pc ] 1
= 4 0:0216 1:0423 0:0163 5
0:0626 0:0216 1:0217
The equalizer coe¢ cients are the middle column:

[ 1 0 1] = [ 0:01631 :0423 0:0216]

4.1. PROBLEM SOLUTIONS 7

b. The output samples are given by

1
X
peq (mT ) = n pc [(m n) T ]
n= 1
[peq ] = [ 0:0213 0:0000 1:0000 0:0000 0:0635]

Problem 4.15

a. The channel response matrix is

2 3
1:0 0:1 0:01 0:001 0:001
6 0:2 1:0 0:1 0:01 0:001 7
6 7
[Pc ] = 6
6 0:02 0:2 1:0 0:1 0:01 7
7
4 0:005 0:02 0:2 1:0 0:1 5
0:003 0:005 0:02 0:2 1:0

Its inverse is
2 3
1:0218 0:1067 0:0218 0:0046 0:0018
6 0:2111 1:0438 0:1111 0:0227 0:0046 7
6 7
[Pc ] = 6
6 0:0651 0:2179 1:0451 0:1111 0:0218 7
7
4 0:0234 0:0673 0:2179 1:0438 0:1067 5
0:0101 0:0234 0:0651 0:2111 1:0218
The equalizer coe¢ cients are the middle column:

[ 2 1 0 1 2] = [0:0218 0:1111 1:0451 0:2179 0:0651]

b. The output samples are given by

1
X
peq (mT ) = n pc [(m n) T ]
n= 1
[peq ] = [0:0000 0:0000 1:0000 0:0000 0:0000]

Problem 4.16

a. The output in terms of the input is

y (t) = x (t) + x (t m)
8CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Using the superposition and time delay theorems, its Fourier transform is

Y (f ) = X (f ) [1 + exp ( j2 m f )]

The frequency response function is

Y (f )
H (f ) = = [1 + exp ( j2 m f )]
X (f )
q
jH (f )j = [1 + cos (2 m f )]2 + sin2 (2 mf )
q
= 1 + 2 cos (2 m f ) + 2

b. From the block diagram

N
X
z (t) = i y [t (i 1) ]
i=1

Using the superposition and time delay theorems, the Fourier transform of z (t) is
N
X
Z (f ) = Y (f ) i exp [ j2 (i 1) ]
i=1

The frequency response function is

N
Z (f ) X
Heq (f ) = = i exp [ j2 (i 1) ]
Y (f )
i=1

c. From part a,
1 1
=
H (f ) 1 + exp ( j2 mf )
2 3
= 1 exp ( j2 mf ) + exp ( j4 mf ) exp ( j6 mf ) + :::

Equate the above to Heq (f ) with m = : Clearly i =( )i 1

; i = 1; 2; :::; N:

Problem 4.17

a. For a timing jitter of zero

1 6
PE0 = exp ( z0 ) = 10
2
or z0 = ln 2 + 6 ln 10
= 13:12 = 11:18 dB
4.1. PROBLEM SOLUTIONS 9

b. For a timing jitter of 5%

1 1
PE5% = exp ( z) + exp [ z (1 2 0:05)]
4 4
1 1
= exp ( z) + exp ( 0:9z) = 10 6
4 4

Trial and error provides the value

z5% = 14:054 = 11:48 dB

Degradation = 11:48 11:18 = 0:3 dB

c.
1 4
PE0 = exp ( z0 ) = 10
2
or z0 = ln 2 + 4 ln 10
= 8:52 = 9:3 dB

z5% = 9:07 = 9:6 dB

Degradation = 9:6 9:3 = 0:3 dB

Problem 4.18
The problem statement is incorrect. The FFT length should have been given as NF F T =
5; 000. Since this corresponds to fs = 50 Hz, the spectral line at the bit rate corresponds
to 50 1000=5000 = 1 Hz:

Problem 4.19
The modi…ed program follows with the spectrum shown in Fig. 4.2. The spectral line at
the bit rate does not appear to be much di¤erent than that of Fig 4.16.

% Problem 4-19
%
nsym = 1000; nsamp = 50; lambda = 0.7;
[b,a] = butter(3,2*lambda/nsamp);
l = nsym*nsamp; % Total sequence length
y = zeros(1,l-nsamp+1); % Initalize output vector
x =2*round(rand(1,nsym))-1; % Components of x = +1 or -1
10CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Amplitude 1

-1
0 100 200 300 400 500 600
1.5
Amplitude

0.5

0
0 100 200 300 400 500 600
1
Spectrum

0.5

0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
FFT Bin

for i = 1:nsym % Loop to generate info symbols

k = (i-1)*nsamp+1;
y(k) = x(i);
end
datavector1 = conv(y,ones(1,nsamp)); % Each symbol is nsamp long
subplot(3,1,1), plot(datavector1(1,200:799),’k’, ’LineWidth’, 1.5)
axis([0 600 -1.4 1.4]), ylabel(’Amplitude’)
…ltout = …lter(b,a,datavector1);
% datavector2 = …ltout.*…ltout;
datavector2 = abs(…ltout);
subplot(3,1,2), plot(datavector2(1,200:799),’k’, ’LineWidth’, 1.5), ylabel(’Amplitude’)
y = ¤t(datavector2); yy = abs(y)/(nsym*nsamp);
subplot(3,1,3), stem(yy(1,1:2*nsym),’k.’)
xlabel(’FFT Bin’), ylabel(’Spectrum’)
% End of script …le.

Problem 4.20
4.2. COMPUTER EXERCISES 11

Expand the signal as

h i
xPSK (t) = Ac cos 2 fc t + d (t)
h i 2 h i
= Ac cos d (t) cos (2 fc t) sin d (t) sin (2 fc t)
h 2 i h 2i
= Ac cos cos (2 fc t) sin sin (2 fc t)
h i2 h i 2
= Ac cos cos (2 fc t) sin sin (2 fc t)
h 2i 2h i
= Ac cos cos (2 fc t) d (t) sin sin (2 fc t)
2 2
The …rst term is clearly carrier and the second modulation. From the …rst line, the
total power is PT = 12 A2c From the last line the power in the carrier component is PC =
1 2 2
2 Ac cos 2 . The fraction of total power in the carrier component is

PC h i
= cos2
PT 2
We want
h i h i p 2 p
cos2 = 0:1 or cos = 0:1 or = cos 1
0:1 = 0:7952
2 2

4.2 Computer Exercises

Computer Exercise 4.1

% ce4_1: Simulation of a BB pulse transmission system …ltered
% by Butterworth …lter
%
N_order = input(’Enter order of Butterworth …lter => ’);
BWT_sym = input(’Enter bandwidth X sym dur product => ’);
T_sym = 1;
BW = BWT_sym/T_sym;
sam_sym = 100;
[num den] = butter(N_order, 2*BW/sam_sym);
N_sym = 20;
ts = T_sym/sam_sym;
12CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

n_samp = N_sym*sam_sym;
t = 0:ts:(n_samp-1)*ts;
s = 3279;
rand(’state’, s);
ss = sign(rand(1, N_sym) - 0.5); % Bipolar
data = 0.5*(ss+1); % Unipolar
x1 = zeros(size(t));
for n = 1:N_sym
x1 = x1 + ss(n)*unit_pulse((t-(n-1)*T_sym-T_sym/2)/T_sym); %NRZ change
end
y1 = …lter(num, den, x1);
x2 = zeros(size(t));
sgn = 1;
for n = 1:N_sym
if n == 1
x2 = x2 + ss(1)*unit_pulse((t-(n-1)*T_sym-T_sym/2)/T_sym); %NRZ mark
else
x2 = x2 + sgn*ss(1)*unit_pulse((t-(n-1)*T_sym-T_sym/2)/T_sym);
end
if n < N_sym
if data(n+1) == 1
sgn = -sgn;
elseif data(n+1) == 0
sgn = sgn;
end
end
end
y2 = …lter(num, den, x2);
x3 = zeros(size(t));
for n = 1:N_sym
x3 = x3 + data(n)*unit_pulse(2*(t-(n-1)*T_sym-T_sym/4)/T_sym); %Unip RZ
end
y3 = …lter(num, den, x3);
x4 = zeros(size(t));
for n = 1:N_sym
x4 = x4 + ss(n)*unit_pulse(2*(t-(n-1)*T_sym-T_sym/4)/T_sym); %Polar RZ
end
y4 = …lter(num, den, x4);
x5 = zeros(size(t));
sgn = 1;
4.2. COMPUTER EXERCISES 13

for n = 1:N_sym
x5 = x5 + sgn*data(n)*unit_pulse(2*(t-(n-1)*T_sym-T_sym/4)/T_sym); %Bip
RZ
if data(n) > 0
sgn = -sgn;
end
end
y5 = …lter(num, den, x5);
x6 = zeros(size(t)); %Split phase
for n = 1:N_sym
x6 = x6 + ss(n)*(unit_pulse(2*(t-(n-1)*T_sym-T_sym/4)/T_sym) ...
-unit_pulse(2*(t-(n-1)*T_sym-3*T_sym/4)/T_sym));
end
y6 = …lter(num, den, x6);
…gure(1)
subplot(6,1,1), plot(t, x1, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’NRZ change’), title(’Baseband binary data formats; un…ltered’)
subplot(6,1,2), plot(t, x2, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’NRZ mark’)
subplot(6,1,3), plot(t, x3, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Unipolar RZ’)
subplot(6,1,4), plot(t, x4, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Polar RZ’)
subplot(6,1,5), plot(t, x5, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Bipolar RZ’)
subplot(6,1,6), plot(t, x6, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Split phase’), xlabel(’Time, seconds’)
…gure(2)
subplot(6,1,1), plot(t, y1, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’NRZ change’), ...
title([’Butterworth …lter; order = ’, num2str(N_order), ’; ...
BW = ’, num2str(BWT_sym), ’/T_b_i_t’])
subplot(6,1,2), plot(t, y2, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’NRZ mark’)
subplot(6,1,3), plot(t, y3, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Unipolar RZ’)
subplot(6,1,4), plot(t, y4, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Polar RZ’)
subplot(6,1,5), plot(t, y5, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Bipolar RZ’)
14CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

subplot(6,1,6), plot(t, y6, ’LineWidth’, 1.5), axis([min(t) max(t) -1.2 1.2]), ...
ylabel(’Split phase’), xlabel(’Time, seconds’)
% End of script …le

function y = unit_step(t)
% Function for generating the unit step
%
y = zeros(size(t));
I = …nd(t >= 0);
y(I) = ones(size(I));

function y = unit_pulse(t)
% Unit rectangular pulse function
%
y = unit_step(t+0.5) - unit_step(t-0.5);

A typical run follows. Plots appesr in Figures 4.3 and 4.4.

>> ce4_1
Enter order of Butterworth …lter => 2
Enter bandwidth X sym dur product => 1

Computer Exercise 4.2

% ce4_2: Computation and plotting of transmitter and receiver frequency responses
% for raised cosine signaling through lowpass Butterworth …ltered channel
%
clf
clear all
A = char(’-’,’:’,’–’,’-.’,’-o’,’-x’);
clf
f3dB = input(’Enter channel …lter 3-dB frequency, Hz => ’);
n_poles = input(’Enter number of poles for Butterworth …lter representing channel =>
’);
Rb = input(’Enter data rate, bits/s => ’);
beta0 = input(’Enter vector of betas => ’);
delf = 100;
Lbeta = length(beta0);
for n = 1:Lbeta
4.2. COMPUTER EXERCISES 15

Baseband binary data formats; unfiltered

Split phaseBipolar RZ Polar RZUnipolar RZNRZ markNRZ change

1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
Time, seconds
16CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Butterworth filter; order = 2; BW = 1/Tbit

Split phaseBipolar RZ Polar RZUnipolar RZNRZ markNRZ change

1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
1
0
-1
0 2 4 6 8 10 12 14 16 18
Time, seconds
4.2. COMPUTER EXERCISES 17

beta = beta0(n);
f1 = 0:delf:(1-beta)*Rb/2;
HR1 = sqrt(1+(f1/f3dB).^(2*n_poles));
f2 = (1-beta)*Rb/2:delf:(1+beta)*Rb/2;
HR2 = sqrt(1+(f2/f3dB).^(2*n_poles)).*cos(pi*(f2-(1-beta)*Rb/2)./(2*beta*Rb));
f3 = (1+beta)*Rb/2:delf:Rb;
HR3 = zeros(size(f3));
f = [f1 f2 f3];
HR = [HR1 HR2 HR3];
plot(f,HR,A(n,:), ’LineWidth’, 1.5)
if n == 1
hold on
xlabel(’{nitf}, Hz’),ylabel(’j{nitH_R}({nitf})j or j{nitH_T}({nitf})j’)
title([’Bit rate = ’,num2str(Rb),’bps; ch …lter 3-dB freq = ’,num2str(f3dB),’Hz;
no. of poles = ’,...
num2str(n_poles) ’; nbeta = ’, num2str(beta0)])
end
end

A typical run follows with a plot given in Fig. 4.5.

>> ce4_2
Enter channel …lter 3-dB frequency, Hz => 2000
Enter number of poles for Butterworth …lter representing channel => 2
Enter data rate, bits/s => 5000
Enter vector of betas => [0 .5 1]

Computer Exercise 4.3

% ce4.3: Design of an N-tap zero-forcing equalizer; input and
% output plotted
%
pc = [-0.03 0.05 -0.1 0.02 -0.05 0.2 1 0.3 -0.07 0.03 -0.2 0.05 -0.1];
n_tap = input(’Number of taps => ’);
nn = ‡oor(n_tap/2)+1;
n_samp = length(pc);
n_center = ‡oor(n_samp/2)+1;
if n_tap > n_center
disp(’Not enough samples for speci…ed equalizer length’)
end
18CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Bit rate = 5000 bps; ch filter 3-dB freq = 2000 Hz; no. of poles = 2; β = 0 0.5 1
2

1.8

1.6

1.4
|HR(f )| or |HT(f )|

1.2

0.8

0.6

0.4

0.2

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
f, Hz
4.2. COMPUTER EXERCISES 19

Pc = [];
for k = 1:n_tap
Pc(:, k) = pc(n_center-k+1:n_center+n_tap-k)’;
end
Pcinv = inv(Pc);
A = Pcinv(:,nn);
disp(’The equalizer tap vector is:’)
disp(A)
nT = [-(n_center-1):n_center-1];
peq = [];
n_eq = n_samp-n_tap+1;
nTc = [-‡oor(n_eq/2):‡oor(n_eq/2)];
for n = 1:n_eq
Nd = n-1;
pcd = pc(Nd+1:n_tap+Nd)’;
peq(n) = A’*‡ipud(pcd);
end
subplot(2,1,1), stem(nT, pc), axis([min(nT) max(nT) -0.5 1.5]), ...
xlabel(’{nitn}’), ylabel(’{nitp_c}({nitn})’)
title([’Input sequence and zero-forced equalizer output for ’, num2str(n_tap), ’-tap
equalizer’])
subplot(2,1,2), stem(nTc, peq), axis([min(nT) max(nT) -0.5 1.5]), ...
xlabel(’{nitn}’), ylabel(’{nitp}_e_q({nitn})’)
% End of script …le

A typical run follows with a plot given in Fig. 4.6.

>> ce4_3
Number of taps => 5
The equalizer tap vector is:
0.1581
-0.3160
1.2068
-0.4405
0.2577

Computer Exercise 4.4: See Problem 4.18 solution

20CHAPTER 4. BASEBAND COMMUNICATION SYSTEMS IN THE ABSENCE OF NOISE

Input sequence and zero-forced equalizer output for 5-tap equalizer

1.5

1
pc (n)

0.5

-0.5
-6 -4 -2 0 2 4 6
n

1.5

1
peq(n)

0.5

-0.5
-6 -4 -2 0 2 4 6
n