EE 179

Digital and Analog Communication Systems

Homework #3 Solutions

April 28, 2014

Handout #21

1. DSB-SC modulator (Lathi & Ding 4.2-3). You are asked to design a DSB-SC modulator to generate
a modulated signal km(t) cos(c t + ), where m(t) is a signal bandlimited to B Hz. Figure P4.2-3
shows a DSB-SC modulator available in the stockroom. The carrier generator available generates
not cos c t but cos3 c t. Explain whether you would be able to generate the design using only this
equipment. You may use any kind of filter you like.

a. What kind of filter is required in Fig. P4.2.3?

b. Determine the signal spectra at points b and c, and indicate the frequency bands occupied by
these spectra.
c. What is the minimum usable value of c ?
d. Would this scheme work in the carrier generator output were sin3 c t? Explain.
e. Would this scheme work in the carrier generator output were cosn c t for any integer n 2?
Solution (10 points)
a. Using trigonometric identities or Eulers formula, we can expand cos3 as
cos3 c t =

3
4

cos c t + 14 cos 3c t .

When modulated by cos3 c t, the transmitted signal contains the term 34 m(t) cos c t, which
is the desired modulated signal with spectrum centered at fc . The other term has spectrum
centered at 3fc . A bandpass filter centered at fc allows the passage of the desired term but
suppresses the unwanted term.
b. The signal spectra at points b and c are shown in the following figure.

c. In order to avoid spectral folding overlap after modulation, the minimum usable value of fc
is B, where B is the bandwidth of the lowpass signal m(t).
d. Yes. Since sin3 c t is cos3 c t delayed by /2, if we modulate with sin3 c t = cos3 (c t /2),
then we can demodulate with cos3 (c t /2) = sin3 c t.

e. The expansion of cosn c t contains a term a1 cos c t when n is odd but not when n is even.
Therefore the system works for carrier cosn c t if and only if n is odd.
2. Audio scrambler (Lathi & Ding 4.2-8). The system shown in Fig. P4.2-8 is used for scrambling
audio signals. The output y(t) is the scrambled version of the input m(t).

a. Find the spectrum of the scrambled signal y(t).

b. Suggest a method for descrambling y(t) to obtain m(t).
Solution (10 points)
a. The spectrum Y (f ) of the scrambled signal is shown in the following figure.
ss

s ss

s ss

b. Observe that Y (f ) is the same as M(f ) with frequency spectrum inverted, that is, the high
frequencies are shifted to lower frequencies and vice versa. To get back to the original spectrum,
we must invert Y (f ), which can be done using the same scrambler.
3. AM signal (Lathi & Ding 4.3-1). In an amplitude modulation system, the message signal is given
by Fig. P4.3-1 and the carrier frequency is 1 KHz. The modulator output is
sAM (t) = 2(b + 0.5m(t)) cos c t .

a. Determine the average message power.

b. If b = 1, determine the modulation index and the modulation power efficiency.
c. Sketch the modulated signal of part (a) in the time domain.
d. If b = 0.5, repeat parts (a) and (b).
Solution (10 points)
a. In general, the sawtooth signal with period T and amplitude a is defined by its values in one
period, [T /2, T /2]:
2a
m(t) =
t , T /2 t T /2
T
Page 2 of 6

EE 179, Spring 2014

The power of this signal is

Z
Z
1 T /2 2
2 T /2 2
2
m (t) =
m (t) dt =
m (t) dt
T T /2
T 0
Z
Z
8a2 t3
8a2 T /2 2
2 T /2  2a 2
t dt = 3
t dt = 3
=
T 0
T
T 0
T 3

 T /2

8a2 T 3
a2

=
=
.

T3 3 8
3
0

Note that the power is independent of T and is proportional to the square of the amplitude.
For the sawtooth signal of this problem, a = 2 and T = 0.1, so the power is
4
22
Pm =
= .
3
3
We can rewrite sAM (t) as
sAM (t) = (2b + m(t)) cos c t .
Therefore the power of the modulated signal is
Ps = Pc + Pm = 12 (2b)2 + 12 ( 34 ) = 2b2 + 32 .
b. The peak value of m(t) is 2. If b = 1, the modulation index is mp /A = 2/2b = 1. The
modulation power efficiency is
=

Ps
2/3
1
=
= = 25% .
Pc + Ps
2 + 2/3
4

c. Shown below is the modulated output for one cycle of the input signal.
4
2
0
2
4
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

d. If b = 0.5, the modulation index is mp /A = 2/2b = 2. The modulation power efficiency is

=

2/3
4
Ps
=
= = 57% .
Pc + Ps
1/2 + 2/3
7

e. Shown below is the modulated output for one cycle of the input signal.
4
2
0
2
4

0.01

Homework #3 Solutions

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Page 3 of 6

Since > 1, the input signal cannot be recovered by envelope detection.

4. Equalization in AM modulation. Consider the DSB-SC system shown in the figure below, where
s(t) = m(t) cos 2fc t with fc B.
Channel
n(t)
LPF
m(t)

DSBSC
Mod

s(t)

x(t)
h(t)

eq

y(t)

(t)

z(t)

cos (2 fc t)

cos (2 fc t)

Assume that the noise n(t) has power spectral density Sn (f ) = 0.1 mW/Hz. The PSD Sm (f ) of
m(t) in mW/Hz is
(
10 10|f |/B |f | B
Sm (f ) =
0
|f | > B.
The frequency response of the channel H(f )

10
H(f ) = 0.5

is
|f fc | B/2
B/2 < |f fc | < B
otherwise

a. What is the power of m(t)?

b. Sketch the PSD of the modulated signal s(t).

c. Find the equalizer Heq (f ) such that in the absence of noise (i.e., for n(t) = 0), z(t) = m(t).
d. Find the PSD and power of z(t) due to noise only (i.e., for m(t) = 0), and due to signal only
(i.e., for n(t) = 0), using Heq (f ) is the equalizer you found in part (c).
e. Find the SNR of the receiver output with the equalizer of part (c).
Solution (25 points)
a. The power spectral density of m(t) is shown in the following figure.

The total power in mW is the area beneath the triangle:

Z
Pm =
Sm (f ) df = 10B .

b. By the modulation theorem, the PSD of the modulated signal is

Ss (f ) = 21 Sm (f + fc ) + Sm (f fc ) .
Page 4 of 6

EE 179, Spring 2014

c. The required equalizer is Heq (f ) = 2/H(f ). The factor of 2 is needed because cos 2fc t in the
demodulator reduces the signal amplitude by a factor of 2.

0.2 |f fc | B/2
Heq (f ) = 4
B/2 < |f fc | < B

0
otherwise

Figure 3: is due
and to the noise, so the output PSD is
d. If the signal is not present, all of the output


Szn (f ) = 0.1 14 |Heq (f + fc )|2 + |Heq (f fc )|2 |HLPF (f )|2 .

The PSD of z(t) due to noise only and due to signal only are shown in the following figure.

Thus the total noise power is

Z
Pn =
Szn (f ) df = 0.8 B + 0.002 B = 0.802B mW .

e. From part (d),

SNR =

Pm
10B
=
= 12.47 = 11.0 dB .
Pn
0.802B

5. Frequency and phase offset in QAM system. In a QAM system (Fig. 4.19 in the textbook), the
locally generated carrier has a frequency error amd a phase offset ; that is, the receiver carrier
is cos((c + )t + ) or sin((c + )t + ).
Homework #3 Solutions

Page 5 of 6

a. Show that the output of the upper receiver branch is

m1 (t) cos( t + ) m2 (t) sin( t + ) ,
instead of m1 (t), and the output of the lower receiver branch is
m1 (t) sin( t + ) + m2 (t) cos( t + ) ,
instead of m2 (t).
b. Find the signal-to-noise-plus-interference power ratio (SINR) on each branch, where the noiseplus-interference power equals the power associated with the unwanted signal (signal m2 (t) in
the upper branch and signal m1 (t) in the lower branch) added to the random noise power.
Assume that the noise n(t) signal added to the QAM signal (QAM (t) in Fig. 4.19) has flat
PSD Sn (f ) = N0 /2.
Solution (10 points)
a. When the carrier at the demodulator has frequency and phase offset,
x1 (t) = 2(m1 (t) cos c t + m2 (t) sin c t) cos((c + )t + )
= 2m1 (t) cos c t cos((c + )t + ) + 2m2 (t) sin c t cos((c + )t + )


= m1 (t) cos((2c + )t + ) + cos( t + ) +


m2 (t) sin((2c + )t + ) sin( t + )

Similarly,

x2 (t) = 2(m1 (t) cos c t + m2 (t) sin c t) sin((c + )t + )

= 2m1 (t) cos c t sin((c + )t + ) + 2m2 (t) sin c t sin((c + )t + )


= m1 (t) sin((2c + )t + ) + sin( t + ) +


m2 (t) cos( t + ) cos((2c + )t + ) .

After x1 (t) and x2 (t) are passed through a low-pass filter, the outputs are
m1 (t) = m1 (t) cos( t + ) m2 (t) sin( t + )
m2 (t) = m1 (t) sin( t + ) + m2 (t) cos( t + )

b. The desired signal for the upper branch is m1 (t) cos( t + ). The noise after demodulation
is n1 (t) = 2n(t) cos((c + )t + ), which is passed through a LPF to get n1 (t) with PSD
(
N0 |f | B
Sn1 (f ) =
0
|f | > B
The interference power is the power of the undesired signal m2 (t) sin( t + ). Therefore the
signal-to-noise-plus-interference power ratio for the upper branch is
h m21 (t) cos2 ( t + )i
.
2N0 B + h m22 (t) sin2 ( t + )i
R T /2
where hg(t)i represents the time-averaged value hg(t)i = limT T1 T /2 g(t) dt.
SINR1 =

Similarly, for the lower branch,

SINR2 =

Page 6 of 6

h m22 (t) cos2 ( t + )i

.
2N0 B + h m21 (t) sin2 ( t + )i
EE 179, Spring 2014