Exam III: Solutions

BME 171-02, Signals and Systems

Exam III: Solutions
100 points total

0. (5 pts.) Laplace transform tables.

1. (20 pts.) A signal x(t) has Laplace transform

s+2
X(s) =
s2 + 4s + 5
Find the Laplace transforms of the following signals without computing the inverse Laplace
transform of X(s):
(a) y1 (t) = x(2t − 1)u(2t − 1)
Solution:
1 −s/2  s  1 −s/2 (s/2) + 2 s+4
Y1 (s) = e X = e = 2 e−s/2
2 2 2 (s/2)2 + 4(s/2) + 5 s + 8s + 20

(b) y2 (t) = e−3t x(t)

Solution:
(s + 3) + 2 s+5
Y2 (s) = X(s + 3) = 2
= 2
(s + 3) + 4(s + 3) + 5 s + 10s + 26

(c) y3 (t) = x(t) cos(7t)

Solution:
 
1 1 (s + 7j) + 2 (s − 7j) + 2
Y3 (s) = [X(s + 7j) + X(s − 7j)] = +
2 2 (s + 7j)2 + 4(s + 7j) + 5 (s − 7j)2 + 4(s − 7j) + 5

(d) y4 (t) = x(t) ⋆ (x(t − 1)u(t − 1))

Solution:
2
s2 + 4s + 4

s+2
Y4 (s) = e−s X 2 (s) = e−s = e−s
s2 + 4s + 5 s4 + 8s3+ 26s2 + 40s + 25
 Exam III: Solutions

2. (20 pts.) Obtain the inverse Laplace transforms of the following functions:
5
(a) X(s) = 3 (Hint: s + 1 is a factor.)
s + s2 + 9s + 9
Solution:
5 5 5
X(s) = = 2 =
s2 (s + 1) + 9(s + 1) (s + 9)(s + 1) (s + 3j)(s − 3j)(s + 1)

The poles are: p1 = −3j, p2 = 3j, p3 = −1. Partial fraction expansion:

c1 c2 c3
X(s) = + +
s + 3j s − 3j s+1
Find the residues:
 
5 5 1 1
c1 = [(s + 3j)X(s)]s=−3j = =− =− + j
(s − 3j)(s + 1) s=−3j 18 + 6j 4 12
r
5 j(π+tan−1 (−1/3))
= e
72
r
1 1 5 j(π+tan−1 (−1/3))
c2 = c̄1 = − − j = e
4 12 72
 
5 5 1
c3 = [(s + 1)X(s)]s=−1 = = =
(s + 3j)(s − 3j) s=−1 10 2
r   
5 1 1
x(t) = 2 cos 3t + π + tan−1
− u(t) + e−t u(t)
72 3 2
(Note: tan−1 (−1/3) ≈ −0.32, so the phase is ≈ 2.82)
s2 + 5s + 7
(b) X(s) =
s2 + 3s + 2
Solution:
(s2 + 3s + 2) + (2s + 5) 2s + 5
X(s) = 2
=1+ 2 ≡ 1 + Y (s).
s + 3s + 2 s + 3s + 2
Decompose the second term into partial fractions:
2s + 5 2s + 5 c1 c2
Y (s) = = = +
s2 + 3s + 2 (s + 2)(s + 1) s+2 s+1

Find the residues:

 
2s + 5
c1 = [(s + 2)Y (s)]s=−2 = = −1
s+1 s=−2
 
2s + 5
c2 = [(s + 1)Y (s)]s=−1 = =3
s+2 s=−1

x(t) = δ(t) − e−2t u(t) + 3e−t u(t)

2
 Exam III: Solutions

3. (15 pts.) For each of the signals below, determine the poles of its Laplace transform without
actually computing the Laplace transform.
h i
(a) x1 (t) = 3e−2t − te−2t + 8e−t/3 u(t)
Solution: p1 = p2 = −2, p3 = −1/3
(b) x2 (t) = 4e−t cos(7t + π/4)u(t) + 9u(t)
Solution: p1 = −1 + 7j, p2 = −1 − 7j, p3 = 0
(c) x3 (t) = 2 cos(3t − π/16)u(t) + 3 sin(4t)u(t)
Solution: p1 = 3j, p2 = −3j, p3 = 4j, p3 = −4j
 −3t 
(d) x4 (t) = e cos(4t) + 2te −3t cos(4t + π/18) u(t)
Solution: p1 = p2 = −3 + 4j; p3 = p4 = −3 − 4j
 2t 2t

(e) x5 (t) = e + 3te + e −t sin(6t) u(t)
Solution: p1 = p2 = 2, p3 = −1 + 6j, p4 = −1 − 6j

3
 Exam III: Solutions

4. (20 pts.) Consider the following system:

+
X(s) _
H1(s) Y(s)

H2(s)

where
s b
H1 (s) = and H2 (s) = .
(s + 1)(s + a) s
(a) Determine a and b such that the overall transfer function is
s
H(s) =
(s + 4)(s + 5)
Solution:
s
H1 (s) (s+1)(s+a) s s
H(s) = = s b
= = 2
1 + H1 (s)H2 (s) 1 + (s+1)(s+a) s
(s + 1)(s + a) + b s + (a + 1)s + (a + b)

Since (s + 4)(s + 5) = s2 + 9s + 20, we must have a + 1 = 9 and a + b = 20, or

a = 8, b = 12

(b) Determine the output y(t) of the system with the above transfer function to the unit-step
input x(t) = u(t).
Solution: X(s) = 1/s, so
s 1 1
Y (s) = H(s)X(s) = =
(s + 4)(s + 5) s (s + 4)(s + 5)

Decompose into partial fractions:

c1 c2
Y (s) = +
s+4 s+5
Find the residues:
 
1
c1 = [(s + 4)Y (s)]s=−4 = =1
s + 5 s=−4
 
1
c2 = [(s + 5)Y (s)]s=−5 = = −1
s + 4 s=−5

x(t) = (e−4t − e−5t )u(t)

4
 Exam III: Solutions

5. (20 pts.) Consider the signal

∞
X
x(t) = 2−k/2 cos(40πkt)
k=0

(a) We want to design a lowpass filter that would remove no more than 10% of the signal
energy. What should be the cutoff frequency of this filter?
∞
X
Solution: X(ω) = 2−k/2 π [δ(ω + 40πk) + δ(ω − 40πk)] By Parseval’s theorem,
k=0

1 ∞
Z
Signal energy = |X(ω)|2 dω
2π −∞
∞ ∞
1
Z
2
X
−k 2
= 2 π [δ(ω + 40πk) + δ(ω − 40πk)] dω
2π −∞
k=0
∞
π X
= 2−k · 2
2
k=0
= 2π.

If we use a lowpass filter to remove all frequencies higher than 40πk0 , the filtered signal will be
k0
X k0
X
x̃(t) = 2−k/2 cos(40πkt), and X̃(ω) = 2−k/2 π [δ(ω + 40πk) + δ(ω − 40πk)].
k=0 k=0

∞
1
Z
Filtered signal energy = |X̃(ω)|2 dω
2π −∞
k0 Z ∞
1 X 2
= 2−k π 2 [δ(ω + 40πk) + δ(ω − 40πk)] dω
2π −∞
k=0
k0
X
= π 2−k
k=0
1 − 2−(k0 +1)
= π
1 − 1/2
 
= 2π 1 − 2−(k0 +1) .

We want
Filtered signal energy
= 1 − 2−(k0 +1) ≥ 0.9,
Signal energy
which is achieved when k0 ≥ 4. Thus, the cutoff frequency should be 40πk0 = 160π rad/s.
(b) Let x̃(t) be the corresponding filtered version of x(t). What is the bandwidth of x̃(t)?
What is the Nyquist rate?
Solution: the highest frequency component in x̃(t) is ω = 160π rad/s. Thus, the bandwidth is 160π
rad/s. The Nyquist rate is 2 × bandwidth = 320π rad/s.