Communications PDF

Overview of Communication Topics
• Sinusoidal amplitude modulation

• Amplitude demodulation (synchronous and asynchronous)
• Double- and single-sideband AM modulation
• Pulse-amplitude modulation
• Pulse code modulation
• Frequency-division multiplexing
• Time-division multiplexing
• Narrowband frequency modulation
J. McNames Portland State University ECE 223 Communications Ver. 1.11 1

Handy Trigonometry Identities
cos(a + b) = cos(a) cos(b) − sin(a) sin(b)

sin(a + b) = sin(a) cos(b) + cos(a) sin(b)
cos(a) cos(b) = 1
2 cos(a − b) + 2 cos(a + b)
1
sin(a) sin(b) = 1
2 cos(a − b) − 2 cos(a + b)
1
sin(a) cos(b) = 1
2 sin(a − b) + 2 sin(a + b)
1

Introduction to Communication Systems
• Communications is a very active and large area of electrical
engineering
• Experienced a lot of growth through the nineties with the advent
of wireless cell phones and the internet
• Still an active area of research
• Fundamentals of signals and systems are essential to grasping
communications concepts
• The next two lectures will merely introduce some of the
fundamental concepts
• Will primarily focus on modulation and demodulation in
continuous-time
• Analogous concepts apply in discrete-time

Introduction to Amplitude Modulation
x(t) × y(t)
c(t)
y(t) = x(t) · c(t) c(t) = cos(ωc t + θc )
• Modulation: the process of embedding an information-bearing
signal into a second signal
• Demodulation: extracting the information-bearing signal from
the second signal
• Sinusoidal Amplitude modulation: a sinusoidal carrier c(t) has
its amplitude modified by the information-bearing signal x(t)

Fourier Analysis of Sinusoidal Amplitude Modulation
For convenience, we will assume θc = 0.
c(t) = cos(ωc t)
C(jω) = π [δ(ω − ωc ) + δ(ω + ωc )]
y(t) = x(t) · c(t)
Y (jω) = 1
2π [X(jω) ∗ C(jω)]
X(jω) ∗ δ(ω − ωc ) = X (j(ω − ωc ))
Y (jω) = 12 X (j(ω − ωc )) + 12 X (j(ω + ωc ))
• Thus, sinusoidal AM shifts the baseband signal x(t) so that it is

centered at ±ωc
• Thus, x(t) can be recovered only if ωc > ωx so that the replicated
spectra don’t overlap

Fourier Analysis of Sinusoidal Amplitude Modulation
X(jω)
1
ω
−ωx 0 ωx
C(jω)
π
ω
-ωc 0 ωc
1
Y (jω)
2
ω
-ωc -ωx -ωc -ωc +ωx 0 ωc -ωx ωc ωc +ωx

Example 1: Sinusoidal AM of a Random Signal
Example of Sinusoidal Amplitude Modulation
0.2
x(t)
−0.2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

1 −3
x 10
c(t)
−1
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
−3
0.2 x 10
y(t)
−0.2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Time (s)
−3
x 10

Example 1: MATLAB Code
function [] = AMTimeDomain();
close all;
N = 2000; % No. samples

fc = 50e3; % Carrier frequency
fs = 1e6; % Sample rate
k = 1:N;
t = (k-1)/fs;
xh = randn(1,N); % Random high-frequency signal

[n,wn] = ellipord(0.01,0.02,0.5,60);
[b,a] = ellip(n,0.5,60,wn);
x = filter(b,a,xh); % Lowpass filter to create baseband signal
c = cos(2*pi*fc*t);
y = x.*c;
figure;
FigureSet(1,’LTX’);
subplot(3,1,1);
h = plot(t,x,’b’);
set(h,’LineWidth’,0.2);
xlim([0 max(t)]);
ylim([-0.3 0.3]);
ylabel(’x(t)’);
title(’Example of Sinusoidal Amplitude Modulation’);
box off;
AxisLines;
subplot(3,1,2);
h = plot(t,c,’b’);
xlim([0 max(t)]);
ylim([-1.1 1.1]);

ylabel(’c(t)’);
box off;
AxisLines;
subplot(3,1,3);
h = plot(t,y,’b’,t,x,’g’,t,-x,’r’);
xlim([0 max(t)]);
ylim([-0.3 0.3]);
xlabel(’Time (s)’);
ylabel(’y(t)’);
box off;
AxisLines;
AxisSet(6);
print -depsc AMTimeDomain;

Synchronous Sinusoidal Amplitude Demodulation
Transmitter Receiver
y(t) y(t) w(t)
x(t) × Channel × H(s) x̂(t)
c(t) c(t)
y(t) = x(t) cos(ωc t)
w(t) = y(t) cos(ωc t)
= x(t) cos2 (ωc t)
1 1
= x(t) 2 + 2 cos(2ωc t)
= 12 x(t) + 12 x(t) cos(2ωc t)
• Synchronous demodulation assumptions
– The carrier c(t) is known exactly
– ω c > ωx
• The x(t) can be extracted by multiplying y(t) by the same carrier
and lowpass filtering the signal

Fourier Analysis of Sinusoidal AM Demodulation
1 Y (jω)
2
ω
−ωc 0 ωc
C(jω)
π
ω
−ωc 0 ωc
1 W (jω)
2
ω
−2ωc 0 2ωc
2
H(jω)
ω
0
1
R(jω)

Synchronous AM Demodulation Observations
Transmitter Receiver
y(t) y(t) w(t)
x(t) × Channel × H(s) x̂(t)
c(t) c(t)
• The lowpass filter H(s) should have a passband gain of 2

• The transition band is very wide so the filter does not need to be
close to ideal (e.g. it can be low order)
• We learned how to design this type of filter in ECE 222
• We assumed the signal spectrum X(jω) was real
• The same ideas hold if X(jω) is complex
• Called synchronous demodulation because we assumed the
transmitter and receiver carrier signals c(t) were in phase

Synchronous AM Demodulation Carrier Phase Analysis
Suppose the transmitter and receiver carrier signals differ by a phase
shift:
cT (t) = cos(ωc t + θ)
cR (t) = cos(ωc t + φ)
w(t) = y(t) cR (t)
= x(t) cT (t) cR (t)
= x(t) cos(ωc t + θ) cos(ωc t + φ)
1
= x(t) 2 cos(θ − φ) + 2 cos(2ωc t + θ + φ)
1
2 x(t) cos(θ − φ) + 12 x(t) cos(2ωc t + θ + φ)

1
=

Synchronous AM Demodulation Carrier Phase Comments
w(t) = 12 x(t) cos(θ − φ) + 12 x(t) cos(2ωc t + θ + φ)
• If θ = φ then we have the same case as before and we recover

x(t) exactly after a lowpass filter with a passband gain of 2
π
• If |θ − φ| = 2, we lose the signal completely
• Otherwise, the received signal is attenuated
• The phase relationship of the oscillators must be maintained over
time
• This type of careful synchronization is difficult to maintain
• Phase-locked loops (PLL) can be used to solve this problem
• In ECE 323 you will design and build PLL’s
• The carrier frequency ωc of the transmitter and receiver must also
be the same and remain so over time

Introduction to Asynchronous AM Demodulation
• Asynchronous modulation does not require the carrier signal
c(t) be available in the receiver
• Thus, there is no need for synchronization
• Asynchronous Modulation Assumptions:
ωc ωx
x(t) > 0 for all t
• However, it does require that the baseband signal x(t) be positive

• In this case, the envelope of the modulated signal y(t) is
approximately the same as the baseband signal x(t)
• Thus, we can recover a good approximation of x(t) with an
envelope detector

Example 2: Asynchronous Amplitude Modulation
Example of Asynchronous Sinusoidal AM Modulation

0.4
0.2
x(t)
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
1 −3
x 10
c(t)
−1
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
−3
0.2 x 10
y(t)
0
−0.2
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Time (s) −3
x 10

function [] = AAMTimeDomain();
close all;

k = 1:N;
t = (k-1)/fs;
xh = rand(1,N)-0.5; % Random high-frequency signal limited to [-0.5 0.5]

[n,wn] = ellipord(0.02,0.03,0.5,60);
[b,a] = ellip(n,0.5,60,wn);
x = x + 0.2; % Convert to positive signal
c = cos(2*pi*fc*t);
y = x.*c;
figure;
subplot(3,1,1);
xlim([0 max(t)]);
ylim([-0.1 0.4]);
ylabel(’x(t)’);
title(’Example of Asynchronous Sinusoidal AM Modulation’);
box off;
AxisLines;
subplot(3,1,2);
h = plot(t,c,’r’);
xlim([0 max(t)]);

ylim([-1.1 1.1]);
ylabel(’c(t)’);
box off;
AxisLines;
subplot(3,1,3);
h = plot(t,y,’g’,t,x,’b’,t,-x,’b’);
xlim([0 max(t)]);
ylim([-0.39 0.39]);
ylabel(’y(t)’);
box off;
AxisLines;
AxisSet(8);
print -depsc AAMTimeDomain;

Diodes
I Ideal Model Real Model

+
I I
V
- V V
0 0.7 0 0.7
• Diodes can be used as a key component in envelope detectors

• Like resistors, diodes do not have memory and the relationship
between voltage and current does not depend on time
• Key idea: diodes only allow current to flow in one direction
• When they are on (V ≥ 0.7 V), they act like an ideal voltage
source
• When they are off (V < 0.7 V), they act like an open circuit

Example 3: Full Wave Rectifier
R
vs
+ vo -
Draw the equivalent circuits for vs > 0 and vs < 0 assuming an ideal
model of the diode with a threshold voltage of 0 V.

Example 3: Workspace

Envelope Detectors
+ i(t) + + i(t) +
y(t) R e(t) y(t) C R r(t)
- - - -
• A diode can be used to extract the upper half of the modulated

signal
• This is called a half-wave rectifier
• Roughly speaking, when y(t) > 0, e(t) ≈ y(t)
• By connecting a capacitor in parallel with the resistor, the RC
circuit acts like a first-order lowpass filter
• This smoothes the received waveform
• A full-wave rectifier that recovers both the negative and positive
peaks has better performance

Example 4: Envelope Tracking
Example of Envelop Tracking of an Asynchronous AM Signal
0.2
y(t)
0
−0.2
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Half−wave Rectifier
−3
0.3 x 10
0.2
0.1
0
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Full−wave Rectifier
−3
0.3 x 10
0.2
0.1
0
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Time (s)
−3
x 10

function [] = EnvelopeTracking();
close all;

al = 0.95; % First-order filter parameter
k = 1:N;
t = (k-1)/fs;
rand(’state’,10);
xh = rand(1,N)-0.5; % Random high-frequency signal limited to [-0.5 0.5]
[n,wn] = ellipord(0.01,0.02,0.5,60);
[b,a] = ellip(n,0.5,60,wn);
x = x + 0.2; % Convert to positive signal
c = cos(2*pi*fc*t);
y = x.*c;
eh = y.*(y>0);
rh = filter(1-al,[1 -al],eh-mean(eh))+mean(eh);
ef = abs(y);
rf = filter(1-al,[1 -al],ef-mean(ef))+mean(ef);
figure;
subplot(3,1,1);
h = plot(t,y,’g’,t,x,’b’,t,-x,’b’);
xlim(1e-3*[0.8 1.8]);
ylim([-0.39 0.39]);
ylabel(’y(t)’);
title(’Example of Envelop Tracking of an Asynchronous AM Signal’);
box off;

AxisLines;
subplot(3,1,2);
h = plot(t,eh,’b’,t,rh,’r’);
xlim(1e-3*[0.8 1.8]);
ylim([-0.02 0.39]);
ylabel(’Half-wave Rectifier’);
box off;
AxisLines;
subplot(3,1,3);
h = plot(t,ef,’b’,t,rf,’r’);
xlim(1e-3*[0.8 1.8]);
ylim([-0.02 0.39]);
ylabel(’Full-wave Rectifier’);
box off;
AxisLines;
AxisSet(6);
print -depsc EnvelopeTracking;

Asynchronous Amplitude Modulation Terminology
x(t) + × y(t)
A c(t)
• Most baseband signals will not be positive

• We can make amplitude-limited signals, |x(t)| ≤ xmax , positive by
adding a constant A such that A > xmax
• The envelope detector then approximates x(t) + A
• x(t) can then be extracted with a highpass filter to remove A (the
DC component)
• The ratio m = xmax /A is called the modulation index
• If expressed in percentage, 100xmax /A, it is called the percent
modulation
• The spectrum of y(t) contains impulses to account for A

Spectrum of Asynchronous Amplitude Modulation
X(jω)
1
ω
−ωx 0 ωx
C(jω)
π
ω
-ωc 0 ωc
Aπ Y (jω)
1
2
ω

Asynchronous Amplitude Modulation Tradeoffs
x(t) + × y(t)
A c(t)
• In most applications, the FCC limits the transmission power

• For asynchronous AM, transmitting the carrier component requires
a portion of this power
• Thus, asynchronous AM is less efficient than synchronous AM
• However, the receiver is easier and cheaper to build
• As m → 1, more of the transmitter power is used for the baseband
signal x(t)
• As m → 0, the signal is easier to demodulate with an envelope
detector

Single Sideband AM
1
Y (jω)
2
ω
• Let us define the bandwidth of the signal as ωx , the highest

frequency component of the signal
• The signal transmitted requires twice the bandwidth, 2ωx
• Near ωc , the signal content for both negative and positive
frequencies is transmitted
• We don’t need this much information to reconstruct X(jω)
• If we know X(jω) for either positive or negative frequencies, we
can use symmetry to construct the other part
• Thus, we only need to transmit one of the sidebands

Single Sideband AM Continued
1
Y (jω)
2
ω
-ωc -ωc +ωx 0 ωc -ωx ωc
• What we have discussed so far uses double-sideband modulation

• We can use single-sideband modulation by removing the upper or
lower sidebands
• Requires only half the bandwidth!
• An obvious approach: lowpass (to retain lower sidebands) or
highpass filter (to retain upper sidebands)
• Requires a nearly ideal high-frequency filter
• SSB modulation increases the cost of the transmitter
• If asynchronous modulation is used, it also increases the cost and
complexity of the receiver

Frequency-Division Multiplexing
1
X1 (jω) 1
X2 (jω) 1
X3 (jω)
ω ω ω
1 Y (jω)
2
ω
• We can transmit multiple signals using a single transmitting
antenna with frequency-division multiplexing (FDM)
• Each baseband signal is shifted to a different frequency band
• Thus, multiple baseband signals can be transmitted
simultaneously over a single wideband channel
• The different modulated signals y1 (t), y2 (t), and y3 (t) are simply
summed before sending to the antenna
• To recover a specific signal, the corresponding frequency band
usually is extracted with a bandpass filter

Time-Division Multiplexing
• The sampling theorem tells us we can represent any bandlimited
signals by its samples x(nT ) as long as ωs > 2ωx
• Thus we can convert multiple bandlimited signals into
discrete-time signals: x1 (t) → x1 [n], x2 (t) → x2 [n],
x3 (t) → x3 [n]
• Time-Division multiplexing interleaves these signals to form a
composite signal
. . . x1 [0], x2 [0], x3 [0], x1 [1], x2 [1], x3 [1], x1 [2], x2 [2], x3 [2] . . .
• A different time interval is assigned to each signal

• We could then form a continuous-time signal using bandlimited
interpolation
• If M signals are multiplexed and each signal has a bandwidth of
ωx , the multiplexed signal y(t) will require a bandwidth of M × ωx

Example 5: Time-Division Multiplexing
Example of Time−Division Multiplexing

1
x [n]
0.5
1
0
0 1 2 3 4 5 6 7 8 9 10 11
1
x [n]
0.5
2
0
0 1 2 3 4 5 6 7 8 9 10 11
1
x [n]
0.5
3
0
0 1 2 3 4 5 6 7 8 9 10 11
1
y[n]
0.5
0
0 5 10 15 20 25 30

function [] = TDMultiplexing();
close all;
k = 1:N;
x1 = rand(N,1);
x2 = rand(N,1);
x3 = rand(N,1);
y = zeros(3*N,1);
k1 = 1:3:3*N;
k2 = 2:3:3*N;
k3 = 3:3:3*N;
y(k1) = x1;
y(k2) = x2;
y(k3) = x3;
wc = pi;
T = 1; % Sample rate
t = 0:0.01:3*N+1;
n = 1:3*N;
yr = zeros(size(t)); % Reconstructed signal

for cnt = 1:length(n),
yr = yr + (wc*T/pi)*y(cnt)*sinc(wc*(t-n(cnt)*T)/pi);
end;
figure;
subplot(4,1,1);
h = stem(k,x1,’g’);
set(h(1),’MarkerSize’,2);

set(h(1),’MarkerFaceColor’,’g’);
hold off;
xlim([0 N+1]);
ylim([0 1.05]);
ylabel(’x_1[n]’);
title(’Example of Time-Division Multiplexing’);
box off;
subplot(4,1,2);
h = stem(k,x2,’b’);
set(h(1),’MarkerFaceColor’,’b’);
hold off;
xlim([0 N+1]);
ylim([0 1.05]);
box off;
subplot(4,1,3);
h = stem(k,x3,’r’);
set(h(1),’MarkerFaceColor’,’r’);
hold off;
xlim([0 N+1]);
ylim([0 1.05]);
box off;
subplot(4,1,4);
h = plot(t,yr,’k’);
hold on;
h = stem(k1,y(k1),’g’);
set(h(1),’MarkerFaceColor’,’g’);
set(h(3),’Visible’,’Off’);
h = stem(k2,y(k2),’b’);
h = stem(k3,y(k3),’r’);

set(h(1),’MarkerFaceColor’,’r’);
hold off;
xlim([0 3*N+1]);
ylim([-0.3 1.3]);
ylabel(’y[n]’);
box off;
AxisLines;
AxisSet(6);
print -depsc TDMultiplexing;

Pulse Amplitude Modulation
T
x(t) × y(t)
+∞
y(t) = x(nT ) p(t − nT )
n=−∞
p(t)
• In modern communication systems, the baseband signal x(t) is

first sampled to form x(nT ) in accord with the sampling theorem
• In a pulse-amplitude modulation (PAM) system, each sample is
multiplied by a pulse p(t)
• Time-division multiplexing can be easily combined with PAM
• Thus we could use p(t) = sinc( tw
π ) to ensure y(t) is bandlimited
to w
• We require w > 2ωx = 2π
T to satisfy the sampling theorem
• AM could then be used to shift this to any frequency band

Pulse-Code Modulation
T T
Modulation/
x[n] Sign × Sign r[n]
Demodulation
x(t) r(t)
p(t)
• In practice, digital systems encode discrete-time signals with
discrete amplitudes
• Most digital signal processing (DSP) uses discrete-valued signals
• Continuous-valued signals are converted to discrete-valued signals
using analog-to-digital (ADC) converters
• Discrete-valued signals can be encoded using binary 1’s and 0’s
• These discrete signals can be transmitted over a communications
channel by transmitting
– 0: Transmit −p(t)
– 1: Transmit +p(t)

Example 6: PCM
Create a random digital (discrete-time and discrete-valued) signal
consisting of fifty 0’s and 1’s. Encode the baseband signal x(t) such
that the bandwidth is limited to 100 Hz. What is the minimum time
required to transmit the signal? Plot the discrete-time signal x[n], the
baseband encoded signal x(t), and an “eye” diagram of the
overlapping received pulses. Assume that the channel does not cause
any distortion and that the receiver and transmitter sampling times are
synchronized. Hint: recall that

W tW
sinc ⇔ PW (jω)
π π

Example 6: Workspace

Example 6: Plot of x[n] and x(t)
Example of Pulse−Code Modulation

1
0.8
0.6
x [n]
1
0.4
0.2
0
0 1 2 3 4 5 6 7 8 9 10 11
1
x(t)
−1
−2
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

Example 6: Eye Diagram
Eye Diagram
2
1.5
0.5
x(t)
−0.5
−1
−1.5
−2
−0.01 −0.008 −0.006 −0.004 −0.002 0 0.002 0.004 0.006 0.008 0.01
Time (sec)

function [] = PCMEx();
close all;

n = 1:N; % Discrete-time index
xd = (rand(N,1)>0.5); % Digital signal
wc = 2*pi*50; % Limit pulse bandwidth to 100 Hz (-50 to 50)

T = pi/wc; % Sample period
Ts = 0.0005;
t = 0:Ts:(N+1)*T;
figure;
subplot(2,1,1);
h = stem(n,xd,’b’);
hold off;
xlim([0 11]);
ylim([0 1.05]);
title(’Example of Pulse-Code Modulation’);
box off;
subplot(2,1,2);
xc = zeros(size(t)); % Modulated signal x(t)
s = -1*(xd(cnt)==0) + 1*(xd(cnt)==1);
p = s*sinc(wc*(t-n(cnt)*T)/pi);
plot(t,p,’b’);
hold on;
xc = xc + p;
end;
plot(t,xc,’g’);

plot(n*T,-1*(xd==0)+1*(xd==1),’ro’,’MarkerSize’,2,’MarkerFaceColor’,’r’);
hold off;
xlim([0 11*T]);
ylabel(’x(t)’);
box off;
AxisLines;
AxisSet(6);
print -depsc PCMSignals;
figure;
k = -round(T/Ts):round(T/Ts);
plot(k*Ts,xc(round(n(cnt)*T/Ts)+k+1));
hold on;
end;
hold off;
xlim([min(k*Ts) max(k*Ts)]);
ylabel(’x(t)’);
xlabel(’Time (sec)’);
title(’Eye Diagram’);
box off;
AxisSet(6);
AxisLines;
print -depsc PCMEyeDiagram;

function [] = PCMEx();
close all;


Ts = 0.0005;
t = 0:Ts:(N+1)*T;
figure;
subplot(2,1,1);
hold off;
xlim([0 11]);
ylim([0 1.05]);
box off;
subplot(2,1,2);
s = -1*(xd(cnt)==0) + 1*(xd(cnt)==1);
plot(t,p,’b’);
hold on;
xc = xc + p;
end;
plot(t,xc,’g’);

hold off;
xlim([0 11*T]);
ylabel(’x(t)’);
box off;
AxisLines;
AxisSet(6);
print -depsc PCMSignals;
figure;
plot(k*Ts,xc(round(n(cnt)*T/Ts)+k+1));
hold on;
end;
hold off;
ylabel(’x(t)’);
box off;
AxisSet(6);
AxisLines;
print -depsc PCMEyeDiagram;

Example 7: Noise Tolerance of PCM
Repeat the previous example, but this time assume that the channel
adds noise that is uniformly distributed between -0.5 and 0.5. Can you
still accurately receive the signal?

Example 7: Plot of x[n] and x(t)
Example of Pulse−Code Modulation

1
x [n]
0.5
1
0
0 1 2 3 4 5 6 7 8 9 10 11
2
x(t)
−2
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
4
2
r(t)
−2
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1

Example 7: Eye Diagram
Eye Diagram
3
1
x(t)
−1
−2
−3
−0.01 −0.008 −0.006 −0.004 −0.002 0 0.002 0.004 0.006 0.008 0.01
Time (sec)

function [] = PCMNoisEx();
close all;


Ts = 0.0002;
t = 0:Ts:(N+1)*T;
nt = n*(T/Ts);
figure;
subplot(3,1,1);
hold off;
xlim([0 11]);
ylim([0 1.05]);
box off;
subplot(3,1,2);
s = -1*(xd(cnt)==0) + 1*(xd(cnt)==1);
plot(t,p,’b’);
hold on;
xc = xc + p;
end;

plot(t,xc,’g’);
hold off;
xlim([0 11*T]);
ylabel(’x(t)’);
box off;
AxisLines;
subplot(3,1,3);
r = xc + (rand(size(xc))-0.5); % Add noise to the received signal
plot(t,r,’b’);
hold on;
plot(n*T,r(1+n*round(T/Ts)),’ro’,’MarkerSize’,2,’MarkerFaceColor’,’r’);
plot(t,xc,’g’);
hold off;
xlim([0 11*T]);
ylabel(’r(t)’);
box off;
AxisLines;
AxisSet(6);
print -depsc PCMNoiseSignals;
figure;
plot(k*Ts,r(n(cnt)*round(T/Ts)+k+1));
hold on;
end;
hold off;
ylabel(’x(t)’);
box off;
AxisSet(6);
AxisLines;
print -depsc PCMNoiseEyeDiagram;

Example 8: Communication System
Design a system (transmitter and receiver) that transmits a stereo
audio signal through a channel in the frequency band of 1.2 MHz
±40 kHz. Discuss the design tradeoffs of different approaches to this
problem and sketch the spectrum of signals at each stage of the
process.

Example 8: Workspace 1

Example 8: Workspace 2

Sinusoidal Angle Modulation
c(t) = A cos (ωc t + θc ) c(t) = A cos (θ(t))
• So far we have discussed different types of amplitude modulation

• Angle Modulation alters the angle of the carrier signal rather
than the amplitude
• Define the instantaneous angle of the carrier signal c(t) as θ(t)
• There are two forms of angle modulation
– Phase modulation (PM): θ(t) = ωc t + θ0 + kp x(t)
dθ(t)
– Frequency modulation (FM): dt = ωc + kf x(t)
• Note that for FM, θ(t) = (ωc + kf x(t)) t

Angle Modulation Versus Amplitude Modulation
Angle Modulation (say FM) versus Amplitude Modulation (AM)
+ One advantage of FM is that the amplitude of the signal
transmitted can always be at maximum power
+ FM is also less sensitive to many common types of noise than AM
- However, FM generally requires greater bandwidth than AM

Example 9: Angle Modulation
Create a random signal bandlimited to ±1 Hz and amplitude limited
to one (e.g. |x(t)| ≤ 1). Modulate the signal use PM and FM with a
carrier frequency of 3 Hz. Use kp = 3 and kf = 4π. Plot the baseband
signal, the carrier signal, and the modulated signals.

Example 9: Signal Plot
Example of Sinusoidal AM Modulation

1
x(t)
−1
0 1 2 3 4 5 6 7 8 9
1
c(t)
−1
0 1 2 3 4 5 6 7 8 9
1
PM
−1
0 1 2 3 4 5 6 7 8 9
1 Time (s)
FM
−1
0 1 2 3 4 5 6 7 8 9
Time (s)

%function [] = AngleModulation();
close all;

fc = 3; % Carrier frequency (Hz)
fs = 50; % Sample rate (Hz)
fx = 1; % Bandlimit of baseband signal
kp = 3; % PM scaling coefficient
kf = 2*pi*2; % FM scaling coefficient
wc = 2*pi*fc;
k = 1:N;
t = (k-1)/fs;
xh = randn(1,N); % Random high-frequency signal

[n,wn] = ellipord(0.95*fx/(fs/2),fx/(fs/2),0.5,60);
[b,a] = ellip(n,0.5,60,wn);
x = x/max(abs(x)); % Scale so maximum amplitude is 1
c = cos(wc*t); % Carrier signal

yp = cos(wc*t + kp*x);
theta = cumsum(wc + kf*x)/fs; % Approximate integral of angle
yf = cos(theta);
figure;
subplot(4,1,1);
xlim([0 max(t)]);
ylabel(’x(t)’);
title(’Example of Sinusoidal AM Modulation’);

box off;
AxisLines;
subplot(4,1,2);
xlim([0 max(t)]);
ylabel(’c(t)’);
box off;
AxisLines;
subplot(4,1,3);
h = plot(t,yp,’b’);
xlim([0 max(t)]);
ylabel(’PM’);
box off;
AxisLines;
subplot(4,1,4);
h = plot(t,yf,’b’);
xlim([0 max(t)]);
ylabel(’FM’);
box off;
AxisLines;
AxisSet(6);
print -depsc AngleModulation;

Relationship of Angle and Frequency Modulation
dθ(t)
θ(t) = ωc t + θ0 + kp x(t) = ωc + kf x(t)
dt
• These two forms are easily related.

dx(t)
• PM with x(t) is equivalent to FM with dt
t
• FM with x(t) is equivalent to PM with 0
x(τ ) dτ
• For c(t), instantaneous frequency is defined as
dθ(t)
ωi (t) ≡
dt
• Frequency modulation: ωi (t) = ωc + kf x(t)
• Phase modulation: ωi (t) = ωc + kp dx(t)
dt

Frequency Modulation
• Consider a sinusoidal baseband signal x(t) = A cos(ωx t)
• This models a bandlimited signal limited to ±ωx
ωi (t) = ωc + kf A cos(ωx t)
• Then
• The instantaneous frequency varies between ωc + kf A and
ωc − kf A
• The modulated signal is then of the form
t
ω
y(t) = cos ωc t + kf x(τ ) dτ = cos ωc t + sin(ωx t) + θ0
−∞ ωx
• Define the following variables
ω
ω ≡ kf A m≡
ωx
• m is called the modulation index

Narrowband Frequency Modulation
y(t) = cos(ωc t + m sin(ωm t))

= cos(ωc t) cos(m sin(ωm t)) − sin(ωc t) sin(m sin(ωm t))
When m is small (m π2 ), this is called narrowband FM modulation
and we may use the following approximations.
cos(m sin(ωm t)) ≈ 1 sin(m sin(ωm t)) ≈ m sin(ωm t)
Thus
y(t) = cos(ωc t) − m sin(ωc t) sin(ωm t)
= cos(ωc t) − m
2 cos(ωc t − ωm t) +
m
2 cos(ωc t + ωm t)

Narrowband Frequency Modulation Continued
A
Y (jω)
2
ω
Y (jω)
π
-ωc +ωx mπ/2 ωc -ωx

ω
-ωc -ωx -ωc ωc ωc +ωx
−mπ/2
• Like AM, spectrum contains sidebands

• Unlike AM, sidebands are out of phase by 180◦
• Bandwidth is twice that of x(t) like double-sideband AM
• Carrier frequency is present and strong

Example 10: Angle Modulation
Create a sinusoidal baseband signal with a fundamental frequency of
1 Hz and a carrier sinusoidal signal at 12 Hz. Plot these signals and
the modulated signals after applying amplitude modulation and
frequency modulation. Use a scaling factor kf = 1 and a modulation
index of m = 0.5. Solve for the baseband signal amplitude A.

Example 10: Signal Plot
Example of Sinusoidal AM and FM Modulation

5
x(t)
−5
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
1
c(t)
−1
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
5
AM
−5
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
1 Time (s)
FM
−1
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
Time (s)

Example 10: Relevant MATLAB Code
%function [] = AMFM();
close all;
fx = 1; % Signal frequency (Hz)

fc = 15; % Carrier frequency (Hz)
fs = 100; % Sampling frequency
kf = 1; % FM scaling coefficient
m = 0.5; % Modulation index
wx = 2*pi*fx; % Signal frequency (rad/s)

wc = 2*pi*fc; % Carrier frequency (rad/s)
A = m*wx/kf; % Modulating amplitude
t = -0.75:0.001:0.75;
x = A*cos(wx*t); % Baseband signal

c = cos(wc*t); % Carrier signal
ya = x.*c; % Amplitude modulated signal
yf = cos(wc*t + m*sin(wx*t));
figure;
subplot(4,1,1);
xlim([min(t) max(t)]);
ylabel(’x(t)’);
title(’Example of Sinusoidal AM and FM Modulation’);
box off;
AxisLines;
subplot(4,1,2);

ylabel(’c(t)’);
box off;
AxisLines;
subplot(4,1,3);
h = plot(t,ya,’b’,t,x,’r’,t,-x,’g’);
ylabel(’AM’);
box off;
AxisLines;
subplot(4,1,4);
h = plot(t,yf,’b’);
ylabel(’FM’);
box off;
AxisLines;
AxisSet(6);
print -depsc AMFM;

Summary
• Modulation is the process of embedding one signal in another with
desirable properties for communication
• Most forms of modulation are nonlinear
• Sinusoidal AM is relatively simple and inexpensive
• Synchronous AM is more efficient than asynchronous AM, but is
also more expensive
• FM is more tolerant of noise than FM, but requires more
bandwidth and cost
• Filters and frequency analysis using the Fourier transform have a
crucial role in communication systems
• Frequency- (FDM) and time-division (TDM) multiplexing can be
used to merge multiple bandlimited signals into a single composite
signal with a larger bandwidth

Communications PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Communications PDF

Uploaded by

Copyright:

Available Formats

Overview of Communication Topics

• Sinusoidal amplitude modulation

J. McNames Portland State University ECE 223 Communications Ver. 1.11 1

cos(a + b) = cos(a) cos(b) − sin(a) sin(b)

J. McNames Portland State University ECE 223 Communications Ver. 1.11 2

J. McNames Portland State University ECE 223 Communications Ver. 1.11 3

J. McNames Portland State University ECE 223 Communications Ver. 1.11 4

• Thus, sinusoidal AM shifts the baseband signal x(t) so that it is

J. McNames Portland State University ECE 223 Communications Ver. 1.11 5

J. McNames Portland State University ECE 223 Communications Ver. 1.11 6

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

J. McNames Portland State University ECE 223 Communications Ver. 1.11 7

N = 2000; % No. samples

xh = randn(1,N); % Random high-frequency signal

J. McNames Portland State University ECE 223 Communications Ver. 1.11 8

J. McNames Portland State University ECE 223 Communications Ver. 1.11 9

J. McNames Portland State University ECE 223 Communications Ver. 1.11 10

J. McNames Portland State University ECE 223 Communications Ver. 1.11 11

• The lowpass ﬁlter H(s) should have a passband gain of 2

J. McNames Portland State University ECE 223 Communications Ver. 1.11 12

2 x(t) cos(θ − φ) + 12 x(t) cos(2ωc t + θ + φ)

J. McNames Portland State University ECE 223 Communications Ver. 1.11 13

• If θ = φ then we have the same case as before and we recover

J. McNames Portland State University ECE 223 Communications Ver. 1.11 14

• However, it does require that the baseband signal x(t) be positive

J. McNames Portland State University ECE 223 Communications Ver. 1.11 15

Example of Asynchronous Sinusoidal AM Modulation

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

J. McNames Portland State University ECE 223 Communications Ver. 1.11 16

N = 2000; % No. samples

xh = rand(1,N)-0.5; % Random high-frequency signal limited to [-0.5 0.5]

J. McNames Portland State University ECE 223 Communications Ver. 1.11 17

J. McNames Portland State University ECE 223 Communications Ver. 1.11 18

I Ideal Model Real Model

• Diodes can be used as a key component in envelope detectors

J. McNames Portland State University ECE 223 Communications Ver. 1.11 19

J. McNames Portland State University ECE 223 Communications Ver. 1.11 20

J. McNames Portland State University ECE 223 Communications Ver. 1.11 21

y(t) R e(t) y(t) C R r(t)

• A diode can be used to extract the upper half of the modulated

J. McNames Portland State University ECE 223 Communications Ver. 1.11 22

J. McNames Portland State University ECE 223 Communications Ver. 1.11 23

N = 2000; % No. samples

J. McNames Portland State University ECE 223 Communications Ver. 1.11 24

J. McNames Portland State University ECE 223 Communications Ver. 1.11 25

• Most baseband signals will not be positive

J. McNames Portland State University ECE 223 Communications Ver. 1.11 26

J. McNames Portland State University ECE 223 Communications Ver. 1.11 27

• In most applications, the FCC limits the transmission power

J. McNames Portland State University ECE 223 Communications Ver. 1.11 28

• Let us deﬁne the bandwidth of the signal as ωx , the highest

J. McNames Portland State University ECE 223 Communications Ver. 1.11 29

• What we have discussed so far uses double-sideband modulation

J. McNames Portland State University ECE 223 Communications Ver. 1.11 30

J. McNames Portland State University ECE 223 Communications Ver. 1.11 31

• A diﬀerent time interval is assigned to each signal

J. McNames Portland State University ECE 223 Communications Ver. 1.11 32

Example of Time−Division Multiplexing

J. McNames Portland State University ECE 223 Communications Ver. 1.11 33

N = 10; % No. samples

yr = zeros(size(t)); % Reconstructed signal

J. McNames Portland State University ECE 223 Communications Ver. 1.11 34

wc = 2pi50; % Limit pulse bandwidth to 100 Hz (-50 to 50)

wc = 2pi50; % Limit pulse bandwidth to 100 Hz (-50 to 50)

wc = 2pi50; % Limit pulse bandwidth to 100 Hz (-50 to 50)

wx = 2pifx; % Signal frequency (rad/s)

x = Acos(wxt); % Baseband signal