K Vasudevan Faculty of EE IIT Kanpur ([email protected].

in) 1

EE320A
Solutions for Tutorial 1
Date: 11th Aug. 2017

1. Solution: The Fourier series for any periodic signal gp (t) is given by:
∞
X
gp (t) = cn e j 2πnt/T0
n=−∞
∞
X
= a0 + 2 an cos(2πnt/T0 ) + bn sin(2πnt/T0 ) (1)
n=1

where

 a n − j bn for n > 0
cn = a0 for n = 0 (2)
a−n + j b−n for n < 0.


Note that an and bn are real-valued, since gp (t) is real-valued. To prove

the first part, we note that
∞
X
gp (−t) = cn e−j 2πnt/T0 . (3)
n=−∞

Substituting n = −m in the above equation we get

−∞
X
gp (−t) = c−m e j 2πmt/T0
m=∞
∞
X
= c−m e j 2πmt/T0 . (4)
m=−∞

Since gp (−t) = gp (t), comparing (1) and (4) we must have c−m = cm
(even function of m). Moreover from (2), cm must be purely real.
To prove the second part, we note from (1) and (4) that c−m = −cm (odd
function of m) and moreover from (2) it is clear that cm must be purely
imaginary.
To prove the third part, we note that
∞
X
gp (t ± T0 /2) = cn e j 2πn(t±T0 /2)/T0
n=−∞
X∞
= cn (−1)n e j 2πnt/T0 . (5)
n=−∞

Since it is given that gp (t ± T0 /2) = −gp (t), comparing (1) and (5) we
conclude that cn = 0 for n = 2m.
K Vasudevan Faculty of EE IIT Kanpur ([email protected]) 2

2. Solution: We know that

vi (t) = i(t)R + v1 (t). (6)

But
dv1 (t)
i(t) = C . (7)
dt
Thus
dv1 (t)
vi (t) = RC + v1 (t). (8)
dt
Taking the Fourier transform of both sides we get

Vi (f ) = RCj 2πf V1 (f ) + V1 (f )
Vi (f )
⇒ V1 (f ) = . (9)
1 + j 2πf RC
It is given that

dv1 (t)
vo (t) =
dt
⇒ Vo (f ) = j 2πf V1 (f )
j 2πf Vi (f )
= . (10)
1 + j 2πf RC
Now
3
Vi (f ) = [δ(f − 3) + δ(f + 3)] . (11)
2
Therefore
3 j 6π 3 −j 6π
Vo (f ) = δ(f − 3) + δ(f + 3) . (12)
2 1 + j 12π 2 1 − j 12π
Taking the inverse Fourier transform we get

vo (t) = c1 e j 6πt + c2 e−j 6πt . (13)

Using Parseval’s power theorem we get the power of vo (t) as

2 × 9 36π 2
P = |c1 |2 + |c2 |2 = ≈ 9/8 watts. (14)
4 1 + 144π 2

3. Solution: Let G(f ) denote the Fourier transform of g(t). Then we have

g1 (t) = g ∗ (−t) ⇀
↽ G∗ (f ) = G1 (f ) (say). (15)
K Vasudevan Faculty of EE IIT Kanpur ([email protected]) 3

Therefore
(n)
g1 (t) ⇀
↽ (j 2πf )n G1 (f )
(n)
⇒ g1 (t) ⇀
↽ (j 2πf )n G∗ (f ). (16)
Next consider g2 (t). We have
g2 (t) = g (n) (t) ⇀
↽ (j 2πf )n G(f ) = G2 (f ) (say)
⇒ g2∗ (−t) ⇀
↽ G∗2 (f ) = (−j 2πf )n G∗ (f ). (17)
Comparing (16) and (17) we see that
(n)
g2∗ (−t) = g1 (t) when n is even
(n)
g2∗ (−t) 6= g1 (t) when n is odd. (18)

g(t)

−2 −1 1 2
dg(t)
dt

1 2 t

−2 −1
−1

d2 g(t)
dt2

−1 1 t
2
−2
−1

Figure 1: Various derivatives of g(t).

4. Solution: We start from the Fourier transform relation:

Z ∞
G(f ) = g(t)e−j 2πf t dt. (19)
f =−∞

Invoking the Schwarz’s inequality we have

Z ∞
g(t)e−j 2πf t  dt
 
|G(f )| ≤
f =−∞
K Vasudevan Faculty of EE IIT Kanpur ([email protected]) 4

Z ∞
⇒ |G(f )| ≤ |g(t)| dt. (20)
f =−∞

Similarly we have
∞
dg(t) −j 2πf t
Z
j 2πf G(f ) = e dt
f =−∞ dt
Z ∞  
 dg(t) 
⇒ |j 2πf G(f )| ≤  dt  dt
  (21)
f =−∞

and
∞
d2 g(t) −j 2πf t
Z
2
(j 2πf ) G(f ) = 2
e dt
f =−∞ dt
Z ∞  2 

2
  d g(t) 
⇒ (j 2πf ) G(f ) ≤  dt2  dt (22)
   
f =−∞

The various derivatives of g(t) are shown in Figure 1. For the given pulse
Z 2
|g(t)| dt = 3
t=−2
Z 2  
 dg(t) 
 dt  dt = 2
 
t=−2
Z 2+  2 
 d g(t) 
 dt2  dt = 4 (23)
 
t=−2−

Therefore the three bounds are

|G(f )| ≤ 3
2
|G(f )| ≤
2π|f |
4
|G(f )| ≤ . (24)
(2πf )2

These bounds on |G(f )| are shown plotted in Figure 2. It can be shown

that
cos(2πf ) − cos(4πf )
G(f ) = . (25)
2π 2 f 2

5. Solution: The Fourier series representation for a periodic signal is given

by:
∞
X
gp (t) = a0 + 2 an cos(2πnt/T ) + bn sin(2πnt/T ). (26)
n=1
K Vasudevan Faculty of EE IIT Kanpur ([email protected]) 5

3.5

2.5

2 |G(f)|
2/(2*pi*|f|)
1.5 4/(2*pi*f)^2
3
1

0.5

-0.5
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

Figure 2: Various bounds for |G(f )|.

For the given problem gp (t) = gp (−t), therefore bn = 0. We have

T−
1
Z
a0 = gp (t) dt
T t=0−
Z T−
1
= [2δ(t) + δ(t − T /2)] dt
T t=0−
3
= . (27)
T
Similarly,
T−
1
Z
an = gp (t) cos(2πnt/T ) dt
T t=0−
Z T−
1
= [2δ(t) + δ(t − T /2)] cos(2πnt/T ) dt
T t=0−
1
= [2 + cos(nπ)] for n ≥ 1. (28)
T
The Fourier transform of gp (t) is given by:
∞
X
gp (t) ⇀
↽ a0 δ(f ) + an [δ(f − n/T ) + δ(f + n/T )] (29)
n=1

where an for n ≥ 0, is given by (27) and (28).

K Vasudevan Faculty of EE IIT Kanpur ([email protected]) 6

Y (f )
38.4

19.2

4
f

−46 −40 −34 −20 −6 0 6 20 34 40 46

17
23

Figure 3: Y (f ).

6. Solution: The Fourier transform of the output is

Y (f ) = X(f ) + 0.2X(f ) ⋆ X(f ). (30)

The spectrum of Y (f ) can be found out by inspection and is shown in

Figure 3.