ECE 318 - Communication Systems 1

Spring 2024
Tutorial 5

TA: Seyyed Mojtaba Pourjaafari ([email protected])

Office Hours: Friday from 4:30 pm to 6:30 pm (EIT-4116)
Textbook: Fundamentals of Communication Systems, 2nd edition, J.G. Proakis and M.
Salehi

In this tutorial we will go through the following items related to the topic of Angle
Modulation.

• Review of Phase & Frequency Modulation

• Spectral Characteristics of Angle-Modulated Signals

• Solve Some Exercise Problems

1 Review of Angle Modulation

An angle-modulated signal is generally expressed as:

u(t) = Ac cos(2πfc t + ϕ(t)) (1)

where Ac denotes the carrier amplitude, fc denotes the carrier frequency and ϕ(t) denotes a
time-varying phase.
The instantaneous frequency of cos(θ(t)) is generally defined as:

1 d
fi (t) = θ(t)
2π dt
Therefore in this case the instantaneous frequency of the angle modulated signal is given as:
1 d
fi (t) = (2πfc t + ϕ(t))
2π dt

1 d
fi (t) = fc + ϕ(t) (2)
2π dt

1
1.1 Phase Modulation (PM)
Suppose m(t) is the message signal. In a Phase Modulation (PM) system, the phase is
proportional to the message signal, i.e.,

ϕ(t) = kp m(t) (3)

where kp is the phase deviation constant.

The maximum phase deviation in a PM system is given by:

∆ϕmax = kp max[|m(t)|] (4)

The modulation index of a PM system βp for a general non-sinusoidal signal m(t) is:

βp = kp max[|m(t)|] (5)

Therefore, from Equations (4) and (5), we get the modulation index in terms of the maximum
phase deviation:

βp = ∆ϕmax (6)

For a sinusoidal message signal of the form m(t) = a cos(2πfm t), the modulation index

βp = kp a = ∆ϕmax (7)

Finally, the general representation of a Phase Modulated signal uP M (t) is:

uP M (t) = Ac cos(2πfc t + kp m(t)) (8)

1.2 Frequency Modulation (FM)

For a message signal m(t), in a Frequency Modulation (FM) system, the instantaneous
frequency deviation from the carrier frequency is proportional to the message signal, i.e.,

fi (t) − fc = kf m(t) (9)

where kf is the frequency deviation constant.

Using the value of fi (t) from Equation (2) in the above equation,

1 d
fc + ϕ(t) − fc = kf m(t)
2π dt
Therefore
d
ϕ(t) = 2πkf m(t)
dt

2
Integrating the above equation in order to find the value of ϕ(t), we get:
Z t
ϕ(t) = 2πkf m(τ )dτ (10)
0

The maximum frequency deviation in a FM system is given by:

∆fmax = kf max[|m(t)|] (11)

The modulation index of a FM system βf for a general message signal m(t) is defined as:

kf max[|m(t)|]
βf = (12)
W
where W denotes the bandwidth of the message signal m(t).
Therefore, from Equations (11) and (12), we get the modulation index in terms of the
maximum frequency deviation:
∆fmax
βf = (13)
W
For a sinusoidal message signal of the form m(t) = a cos(2πfm t), the modulation index βf
is:
kf a ∆fmax Maximum Frequency Deviation
βf = = = (14)
fm fm Frequency of the Message Signal

Finally, the general representation of a Frequency Modulated signal uF M (t) is:

Z t
uF M (t) = Ac cos(2πfc t + 2πkf m(τ )dτ ) (15)
0

2 Review: Spectral Characteristics of Angle-Modulated

Signals
Since Angle-Modulated systems are inherently non-linear in nature, it is quite difficult to
characterize their spectral properties precisely. Usually, analysis is done on the spectral
characteristics of simple modulating signals and then certain approximations are made to
extend this analysis to more complex signals. Consider the following angle-modulated signal
where the message signal is a simple sinusoid (to be precise, sine in PM and cosine in FM):

u(t) = Ac cos(2πfc t + β sin(2πfm t)) (16)

where β is the modulation index. Using Euler’s relation the above can be rewritten as:

3
 u(t) = Re(Ac ej2πfc t ejβsin(2πfm t) ) (17)

Since, sin(2πfm t) is periodic with period Tm = f1m , the same is true for the complex
exponential signal ejβsin(2πfm t) . Therefore, it can be expanded in terms of Fourier Series
expansion. We know that the Fourier Series coefficients are obtained as:

Z α+T0
1 −j2π Tn t
xn = x(t)e 0 dt (integral over one period for some arbitrary α)
T0 α

In this case, let us call the coefficients as cn , T0 = Tm and x(t) = ejβsin(2πfm t) . Therefore,
the coefficients are:

Z Tm
1 n
cn = ejβsin(2πfm t) e−j2π Tm t dt (Let α = 0)
Tm 0

1
Since Tm = fm
, Thus:

Z 1 Z 1
fm fm
jβsin(2πfm t) −j2πnfm t
cn = f m e e dt = fm ej(βsin(2πfm t)−2πnfm t) dt
0 0

Substituting u = 2πfm t in the above equation, it becomes:

Z 2π
1
cn = ej(βsin(u)−nu) du
2π 0

The above equation is a well known integral called the Bessel Function of the first
kind of order n and is denoted by Jn (β). Therefore,

Z 2π
1
Jn (β) = ej(βsin(u)−nu) du (18)
2π 0

Bessel Functions for various values of n are shown in Figure-1.

A function x(t) is expressed in terms of Fourier Series expansion as:

∞
j2π Tn t
X
x(t) = xn e 0

n=−∞

Therefore, in this case:

4
 Figure 1: Bessel Functions for various values of n.

∞
X
x(t) = ejβsin(2πfm t) = Jn (β)ej2πnfm t (19)
n=−∞

1
where T0 = Tm = fm
. Substituting the result in Equation (19) into Equation (17), we
have:

∞
X ∞
X
j2πnfm t j2πfc t
u(t) = Re(Ac Jn (β)e e ) = Re(Ac Jn (β)ej2π(fc +nfm )t ) (20)
n=−∞ n=−∞

Therefore,

∞
X
u(t) = Ac Jn (β) cos(2π(fc + nfm )t) (21)
n=−∞

This shows that even when we are modulating a simple sinusoidal signal, the angle-
modulated signal contains all the frequencies of the form fc + nfm for n = 0, ±1, ±2, ±3, ...,
and therefore, the actual bandwidth of the modulated signal is infinite. Here fc is the

5
carrier component and ±nfm are called harmonics. However, the amplitude of the frequency
components corresponding to large values of n is very small and hence in case of Angle-
modulated signals the concept of Effective Bandwidth is used, i.e., the bandwidth which
contains most of the modulated signal power. Now the question is, how many sidebands or
harmonics are important in the transmission of angle-modulated signals? A common rule is
that a sideband or harmonic is significant if its magnitude is equal to or exceeds 1% of the
unmodulated carrier.
• For β << 1, only the n = 0, ±1 terms in Jn (β) are significant. In this case the the
Effective Bandwidth is approximately:
Bc = 2fm (22)

• The bandwidth for large β is approximately:

Bc = 2nfm (23)
As β becomes large, Jn (β) diminishes rapidly and the ratio βn → 1. Therefore, when
β >> 1, n ≈ β. And thus for large β the approximate bandwidth is:
Bc = 2nfm ≈ 2βfm (24)
We know that for a FM system the maximum frequency deviation or peak fre-
quency deviation is given by:

∆fmax = ∆f = kf max[|m(t)|] (25)

If the message signal is a sinusoid of the form m(t) = a cos(2πfm t), then in this case
the peak frequency deviation becomes:

∆fmax = kf a (26)

Also, for a sinusoidal message signal of the form m(t) = a cos(2πfm t), the modulation
index β is:

kf a ∆f
β= = (27)
fm fm

And thus,
∆f = βfm (28)

Therefore, Equation (24) becomes:

Bc ≈ 2βfm = 2∆f (29)

6
• The Effective Bandwidth of an angle-modulated signal is generally given by the Car-
son’s rule as:

Bc = 2(β + 1)W (30)

where β is the modulation index and W is the bandwidth of the message signal m(t).
Similarly, for a sinusoidal message signal with frequency fm , the effective bandwidth of
an angle-modulated signal (which contains at least 98 % of the signal power) is given
by the relation:

Bc = 2(β + 1)fm = 2(βfm + fm ) (31)

Using the result of Equation (28), Equation (31) becomes:

Bc = 2(∆f + fm ) (32)

For a very small β (Narrowband), Equation (31) becomes:

Bc ≈ 2fm (33)

And for a very large β (Wideband), Equation (31) becomes:

Bc ≈ 2fm β = 2∆f (34)

7
3 Textbook Exercise Problem 4.4
An Angle-Modulated signal has the form:

u(t) = 100 cos[2πfc t + 4sin(2000πt)] (35)

where fc = 10 MHz.

3.1 Part-1
Determine the average transmitted power.

SOLUTION:

The modulated signal is essentially a sinusoidal signal. We know from earlier in this
2
course that the power of a sinusoidal signal of the form Ac cos(2πfc t) is A2c . Here Ac = 100.
Therefore, the power content of the modulated signal is:

A2c 1002
P = = = 5000 (36)
2 2

3.2 Part-2
Determine the peak-phase deviation.

SOLUTION:

The peak-phase deviation in a PM system is given in Equation (4) as:

∆ϕmax = kp max[|m(t)|] = max[|kp m(t)|]. (37)

The general representation of a Phase Modulated signal uP M (t) is given in Equation (8) as:

uP M (t) = Ac cos(2πfc t + kp m(t))

Comparing the above equation with Equation (35), it is clear that in this case:

kp m(t) = 4 sin(2000πt) (38)

From Equations (37) and (38), we see that the peak-phase deviation is:

∆ϕmax = max[|4 sin(2000πt)|] = 4 (39)

8
3.3 Part-3
Determine the peak-frequency deviation.

SOLUTION:

The peak-frequency deviation in a FM system is given in Equation (11) as:

∆fmax = kf max[|m(t)|] = max[|kf m(t)|] (40)

Since we do not know kf , we cannot use the straight forward approach of Part-2 to find the
peak-frequency deviation, and would have to use an alternative approach.
Note from Equation (9) that in case of a FM system, the frequency deviation of the
instantaneous frequency fi (t) around the carrier frequency fc is given as:

fi (t) − fc = kf m(t) (41)

From Equations (39) and (40) it is clear that the peak-frequency deviation can also be
represented as:

∆fmax = max[|fi (t) − fc |] (42)

The instantaneous frequency fi (t) is given by Equation (2) as:

1 d
fi (t) = fc + ϕ(t) (43)
2π dt
Comparing Equations (1) and (35), it is clear that:

ϕ(t) = 4 sin(2000πt). (44)

Therefore, Equation (43) becomes:

1 d 4 d 4
fi (t) = fc + [4 sin(2000πt)] = fc + [sin(2000πt)] = fc + × 2000π [cos(2000πt)]
2π dt 2π dt 2π
In this case the Instantaneous Frequency is:

fi (t) = fc + 4000 cos(2000πt) (45)

Using this value of fi (t) in Equation (42):

∆fmax = max[|fc + 4000 cos(2000πt) − fc |] = max[|4000 cos(2000πt)|]

Therefore, the peak-frequency deviation is:

∆fmax = 4000 Hz . (46)

9
3.4 Part-4
Is this a FM or a PM signal? Explain.

The Angle Modulated signal u(t) is:

u(t) = 100 cos[2πfc t + 4 sin(2000πt)] (47)
It can be interpreted both as a PM and a FM signal.

3.4.1 Interpretation as a PM Signal

A PM signal is generally represented as:
uP M (t) = Ac cos(2πfc t + kp m(t)) (48)
Comparing Equations (47) and (48), we conclude:
kp m(t) = 4 sin(2000πt) (49)
Therefore, one possible scenario is:
• kp = 4
• m(t) = sin(2000πt)
where we have assumed that the message signal m(t) is of the form asin(2πfm t) with a = 1.

3.4.2 Interpretation as a FM Signal

A FM signal is generally represented as:
Z t
uF M (t) = Ac cos(2πfc t + 2πkf m(τ )dτ ) (50)
0

Comparing Equations (47) and (50), we conclude:

Z t
2πkf m(τ )dτ = 4 sin(2000πt)
0

Taking the derivative of the above equation with respect to t:

Z t
d d
[2πkf m(τ )dτ ] = [4 sin(2000πt)]
dt 0 dt
2πkf m(t) = 4 × 2000π cos(2000πt) ⇐⇒ kf m(t) = 4000 cos(2000πt). (51)
Therefore, one possible scenario is:
• kf = 4000
• m(t) = cos(2000πt)
where we have assumed that the message signal m(t) is of the form acos(2πfm t) with a = 1.

10
4 Additional Exercise Problem
Sketch the PM and FM waves produced by the sawtooth wave shown below.

m(t)

t
0 T0 2T0 3T0

SOLUTION:

Starting with the PM case,

ϕ(t) = Ac cos(2πfc t + kp m(t))

Consider the first interval, from 0 to T0 . The message m(t) is just a straight line:

At
m(t) = , 0 ≤ t < T0
T0
 
kp At
ϕ(t) = Ac cos 2πfc t +
T0
  
kp A
= Ac cos 2πfc + t , 0 ≤ t < T0
T0
kp A
So on this interval, we can just draw a simple cosine with a frequency of 2πfc + T0
kp A
[rad/s] or fc + 2πT 0
[Hz]:

11
 ϕ(t)

Ac

0 t
T0 2T0 3T0

−Ac
1
kp A
fc + 2πT
0

In the following interval, from T0 to 2T0 , m(t) is still a straight line, but it gets shifted
downwards by A:
At
m(t) = − A, T0 ≤ t < 2T0
T0
 
kp At
ϕ(t) = Ac cos 2πfc t + − kp A
T0
  
kp A
= Ac cos 2πfc + t − kp A , T0 ≤ t < 2T0
T0
  !!
kp A kp A
= Ac cos 2πfc + t− , T0 ≤ t < 2T0
T0 2πfc + kTp0A

So at T0 , the phase will suddenly jump by −kp A radians. Likewise, it will jump another
−kp A radians at 2T0 , and again at 3T0 , and so on.
In other words, once we reach t = T0 , we “pause” the cosine wave, then “wait” for
kp A kp A
kp A seconds worth of time, and then “restart” the cosine wave at time t = T0 + kp A
2πfc + T0
2πfc + T0
kp A
second with magnitude of 1. In the time between t = T0 and t = T0 + kp A , which
2πfc + T0
we initially “skipped”, just “back-track” the cosine wave as usual from magnitude of 1 at
t = T0 + kp Akp A to t = T0 .
2πfc + T0
An example in drawn in the following figure where we assumed that kp A = π radians.

12
 ϕ(t) kp A [rad] = kp A
[s]
kp A
2πfc + T0

Ac

0 t
T0 2T0 3T0

−Ac
1
kp A
fc + 2πT
0

The FM case is a bit easier. Note that the sudden jumps in the PM case occur if and
only if there is a sudden jump in m(t); that is, since m(t) is not continuous, then the phase
won’t be continuous either.
On the other hand, in FM, the phase is proportional to the integral of m(t) instead.
While m(t) itself is not continuous, the integral of m(t) is continuous. So we will not see the
same sudden jumps in ϕ(t) in the FM case.
It’s easier to draw ϕ(t) in the FM case by looking at the instantaneous frequency:

fi (t) = fc + kf m(t)

The instantaneous frequency will increase linearly between 0 ≤ t < T0 , from fc to fc +kf A.
Immediately after t = T0 , the instantaneous frequency will go back down to fc and start
increasing linearly again as before:

13
 ϕ(t)

frequency increases linearly until T0

Ac

0 t
T0 2T0 3T0

−Ac

14