ECN-311: Principles of Digital Communication

Department of Electronics and Communication Engineering

Indian Institute of Technology Roorkee

Tutorial 4 Solutions
By
Prof. Ekant Sharma

Problem 1. For a message signal given by

m(t) = Am cos (2πfm t) ,

where fm = 0.25Hz, sampling interval Ts = 1s, and pulse duration T = 0.45s

i) Plot the spectrum of the PAM wave modulated by m(t).

ii) Plot the spectrum of the reconstruction filter output, assuming an ideal reconstruction
filter.

iii) Identify the term that captures the aperture effect.

Solution: PAM signal:

∞
X ∞
X
s(t) = m(nTs )h(t − nTs ) = ( m(nTs )δ(t − nTs )) ∗ h(t) = sδ (t) ∗ h(t)
n=−∞ n=−∞

=⇒ S(f ) = Sδ (f )H(f )
∞
X
where, Sδ (f ) = fs M (f − nfs ),
n=−∞

Am
M (f ) = (δ(f − 0.25) + δ(f + 0.25)) , and
2
H(f ) = T sinc(f T ), T = 0.45s.
∞
A m fs P
=⇒ S(f ) = 2
(H(0.2 + n)δ(f − 0.2 − n) + H(−0.2 − n)δ(f + 0.2 − n))
n=−∞

Problem 2. Given a message signal m(t) with the spectrum given in the figure below.
The signal m(t) is sampled at 1kHz using flat-top pulses of duration 0.1 ms

i) Sketch the spectrum of the PAM signal.

ii) Determine the equalization required to counter the Aperture effect.

1
 Figure 1: Spectrum of m(t)

Solution: Spectrum of the PAM signal:

∞
X
S(f ) = fs H(f ) M (f − kfs )
k=−∞

∞
X
Mδ (f ) = fs M (f − kfs )
k=−∞

Figure 2: plot of H(f ), Mδ (f ), and S(f )

Problem 3. An analog signal g(t) is multiplied by a periodic train of rectangular pulses

c(t), where the pulse repitition frequency is fs , and the duration of each pulse is T (such
that fs T >> 1.

i) Find the spectrum of the sampled signal.

ii) Specify the conditions required to recover the original signal g(t) from it’s samples.

2
Solution: Pulse train: c(t) = p(t) ∗ d(t) where p(t) is a single pulse with duration T , and
∞
X
d(t) = δ(t − kTs ).
k=−∞

We need to find the spectrum of gc (t) = g(t)c(t). We have

Gc (f ) = G(f ) ∗ C(f ).
∞
X
Gc (f ) = ak G(f − kfs ).
k=−∞

where, ak = fs T sinc(kfs T ).

Problem 4. A PAM telemetry system involves the multiplexing of four input signals:
si (t), i = 1, 2, 3, 4. Two of the signals s1 (t) and s2 (t) have bandwidths of 80Hz each, whereas
the remaining two signals s3 (t) and s4 (t) have bandwidths of 1kHz each. The signals s3 (t)
and s4 (t) are each sampled at the rate of 2400 samples per second. This sampling rate is
divided by 2R (i.e., an integer power of 2) in order to derive the sampling rate for s1 (t) and
s2 (t)
i) Find the maximum value of R.
ii) Using the value of R found in part (a), design a multiplexing system that first multi-
plexes s1 (t) and s2 (t) into a new sequence s5 (t), and then multiplexes s3 (t),s4 (t) and
s5 (t).

Solution: (i) Maximum value of R:

2400
> 160 Hz
2R
leads to
R ≤ 3.
(ii) One of the solutions:
The multiplexing should take place in the following order: every
1 1
= s
2400 × 3 7200
there is a transmission as follows:
• Time 0 to 1
2400
s: Transmit s3 , s4 , s1 .
• Time 1
2400
s to 2
2400
s: Transmit s3 , s4 , s2 .
• For the next 6 intervals of length 1
2400
s reaching time 1
300
s: Transmit s3 , s4 , X, where
X is null.
• Repeat the same procedure every 1
300
s.

3
 Figure 3: Impulse response of the first-order hold

Problem 5. (Sample-and-hold First order hold): The impulse response of a first order
hold is given in the figure below:
i) Calculate and plot the frequency response of the first-order hold filter. Compare it
with the sample-and-hold (zero-order hold) filter.
ii) What is the equalization filter that needs to be cascaded with the first-order hold filter
for perfect reconstruction. Compare with the Equalization filter required with the
zero-order hold (take T /Ts ≤ 0.1).

Solution:
1
h(t) = (A(t − T ) + T A′ (t − T )) .
T
where,
A(t) = p(t) ∗ p(t).
p(t) is a pulse of duration T . This simplifies to

2
T −1 T sinc(f T )e−jπf T
H(f ) = .
1 + j2πf T

Figure 4: |H(f )| for T = 1