HW - 2 solutions (draft)

Q1)

Given x1 (t) = x1 (t + T1 ) and x2 (t) = x2 (t + T2 ) x(t) = x1 (t) + x2 (t)

Let the period of x(t) be T x(t) = x(t + T )
x(t + T ) = x1 (t + mT1 ) + x2 (t + nT2 )
mT1 = nT2 = T where m, n are least integers.
The period of x(t) is the LCM of T1 and T2
T1
The condition for periodic is the ratio T2 must be a rational number.

Q2)

Formula :
Z ∞
x(t)δ(t − t0 )dt = x(t0 ) (1)
−∞

(a) Let t − τ = p ⇒ τ = t − p;

Z ∞ Z ∞ Z ∞
δ(t − p)x(p)dp = δ(p − t)x(p)dp = x(p)δ(p − t)dp = x(t)
−∞ −∞ −∞

(b) x(t)
(c) t = 0; e0 = 1
(d)
∞ ∞ ∞
3 1 3 1 3π −1
Z Z Z
δ(2t − 3)sin(πt)dt = δ(2(t − ))sin(πt)dt = δ(t − )sin(πt)dt = sin( ) =
−∞ −∞ 2 −∞ 2 2 2 2 2

(e) t = −3; e3
(f ) t = 1; 13 + 4 = 5
(g) t = 3; x(2 − 3) = x(−1)
(h) t = 3; e(3−1) cos( (3−5)π
2 ) = −e2

1
Q3)

When a > 0,
Z∞ Z∞   Z∞ Z∞
t 1 1 1
f (t)δ(at)dt = f δ(t) dt = f (0)δ(t) dt = f (t)δ(t) dt
a a a a
−∞ −∞ −∞ −∞

When a < 0, using b=-a

Z∞ −∞  Z∞ Z∞
−t −1

1 1
Z
f (t)δ(−bt)dt = f δ(t) dt = f (0)δ(t) dt = f (t)δ(t) dt
b b b |a|
−∞ ∞ −∞ −∞

Q4)
dx(t)
(a) y(t) = dt

It is linear,time-invariant, causal, unstable and non-ivertible.

Unstable — for any discontinuous signal, derivative at points of discontinuity becomes un-
bounded.
Non-invertible because for different constants it gives zero.
R 3t
(b) y(t) = −∞ x(τ ) dτ
It is linear, time-variant, non-causal, unstable and non-invertible.
R 3t
Time-variant, let x(t) be delayed by t0 , then y1 (t) = −∞ x(τ − to) dτ
R 3(t−t )
Now delay y(t) by t0 , then y(t − t0 ) = −∞ 0 x(τ ) dτ
y1 (t) 6= y(t − t0 )
Non-causal, upper limit of the integration is 3t. For t > 0, output depends on the future inputs.
Unstable, lower limit is −∞, for bounded input, output will be undefined.
(c) y(t) = x(t/2)
Linear, time-variant, non-causal, stable,invertible.
Non-causal, for t < 0, output depends on future input values.
(d) 
x(t) − x(t − 100) t ≥ 0

y(t) =
0

otherwise

Linear, time-variant, causal, stable and Non-invertible.

Time variant because output is always 0 for negative values of time irrespective of the input.
Therefore different shifts of the input to the negative time axis will result in non-proportional out-
puts.

2
 Non-invertible because of many to one mapping. Say, for a DC input the output of the system
will be 0 and no inverse system can predict the input.
dy
(e) dt + 3ty(t) = t2 x(t)
Linear, time-variant, causal, un-stable and non-invertible.
dy
Unstable as the output keeps increasing with time. [Say, the input signal is such that dt is
small. Then, y(t)=tx(t)]
Non-invertible because at t=0, the differential of output is 0. Therefore, the output is inde-
pendent of the input and therefore an inverse system cannot predicted the original input from the
output of this system.
(f ) y(t) = ∞
P
n=−∞ x(t)δ(t − nT )

Linear, time-variant, non-causal, stable and non-invertible.

Non-invertible, because output will be discrete sequence, from discrete output, continous time
input cannot be reconstructed.
(g) y(t) = x(2t − 4), when input is x(t)
Linear, time-variant, non-causal, stable and invertible.
Non-causal, Output depends on the future inputs.
Eg: For t = 5, y(5) = x(2 × 5 − 4) = x(6)

Q5)
x2 (t)
Given y(t) = dx/d(t)

Homogeneity (scalar rule) means that as the strength of input signal is increased (scaled), then the
strength of output signal will be also increased (scaled) with same amount.
For the input Kx(t), output should be Ky(t)
Let the output of given system is yo (t) for input Kx(t)
x (t)2
yo (t) = K dx/d(t)
yo (t) = Ky(t), Therefore the homogeneity rule is satisfied.
Additivity denotes that the output of system can be computed as sum of the responses resulting
from each input signal acting alone.
For the input x1 (t) + x2 (t) output should be y1 (t) + y2 (t)
Let the output of the system be ya (t) for the input x1 (t) + x2 (t)
(x1 (t)+x2 (t))2
ya (t) = d(x1 +x2 )/d(t)

x21 (t) x22 (t)

y1 (t) + y2 (t) = dx1 /d(t) + dx2 /d(t)

3
 Here ya (t) is not equal to y1 (t) + y2 (t).
So the system does not satisy the additivity property.

Q6)

y (t)
x (t) 1
1
2
1

0 1 2 t 0 1 2 t

(a) (b)

x (t) y (t)
2 2
1 2

0 2 4 t 0 2 4 t

-1 -2
(c) (d)

Figure 1: Sketches represent : (a) x1 (t), (b) y1 (t), (c) x2 (t) and (d) y2 (t).

The solution follows from property of LTI systems, i.e. :

If: x1 (t) → y1 (t)

then x2 (t) → y2 (t)

x1 (t) − x1 (t − 2) → y1 (t) − y1 (t − 2)

Q7)

Figure 2: Q7) FM modulation

4
a

y(t) =cos(8πt) for t <-2 (2)

y(t) =cos(10πt + 4π) = cos(10πt) for -2 <t <1 (3)

y(t) =cos(8πt + 6π) = cos(8πt) for 1 <t (4)

y(t)|t=−2 =1 (5)

y(t)|t=1 =1 (6)

(7)

0.5
y(t)

−0.5

−1
−2 −1 0 1
ω = 8π ω = 10π ω = 8π
t→

Using the above equations the time domain FM modulated waveform can be drawn.

b Non-linear: output amplitude does not scale with input amplitude.

Time-variant: If the input is shifted by T ,
Zt
 

m(t) → y(t) = Acos ωc t + ω∆ m(τ )dτ 

−∞
t+T
 
Z
m(t + T ) → y 0 (t) = Acos ωc t + ω∆ m(τ )dτ 
−∞

But
t+T
 
Z
y(t + T ) = Acos ωc t + ωc T + ω∆ m(τ )dτ 
−∞

=⇒ y(t) 6= y 0 (t) ∴ the system is time-variant

Has Memory: Integration operation on the message cause the output at the present instant to
depend on past inputs of the message signal.
Causal: The output at present instant does not depend on future values of input.

5
Q8)

R1=1Ω R2=2Ω
t=0

+ L=4Η i2 C=0.5F +
vs i1 iL vc
− -

Figure 3: Q8) Input vs and output i1

Using KVL in the two loops,

d(i1 − i2 )
L = vs − i1 R1 (8)
Zdt
1
i2 R2 + i2 dt = vs − i1 R1 (9)
c

Solving for i2 in terms of i1 from the equations 8 and 9 yields:

D2 L
i2 = 1 i1 (10)
(D2 L + DR2 + C)
 
d
where D is the differential operator
dt
Substituting (10) in (8) gives
     
2 R2 1 2 1 R1 R2 R1
D +D + vs = D (R1 + R2 ) + D + + i1 (11)
L LC C L LC

Using the values of L, C, R1 and R2 in 11,

d2 vs dvs d2 i1 di1
2 2
+ + v s = 6 2
+5 + i1 (12)
dt dt dt dt