0% found this document useful (0 votes)
51 views19 pages

ELEN3026A - Lecture 2 - Worked Examples - Questions & Solutions

Uploaded by

Bongani Mofokeng
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
51 views19 pages

ELEN3026A - Lecture 2 - Worked Examples - Questions & Solutions

Uploaded by

Bongani Mofokeng
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

School of Electrical and Information Engineering

University of the Witwatersrand, Johannesburg


ELEN3026A – Signals and Systems IIA (PT)

Fourier Transform - Worked Examples - 2022

Question 1
Find the Fourier transform of the following common signals
(i) δ(t)
(ii) e−bt u(t) for where b > 0 is a constant, plot the magnitude and phase spectra
for b = 1
(iii) e−a|t| , where a > 0 is a constant
(iv) cos(ω0 t)
(v) u(t) − u(t − 1)

Question 2
Find the Fourier transforms of the following signals
(i) x1 (t) = x(2t), where the signal x(t) = u(t) − u(t − 1)
(ii) x(t) cos(ω0 t)
(iii) e−|t| cos(10t)

Question 3
Given the signal x(t) = e−at u(t)
(a) Compare its energy in
(i) Time domain
(ii) Frequency domain
and hence verify that Parseval’s theorem for x(t).
(b) Determine the frequency W (rad/s) so that the energy contributed by the spec-
tral components of all the frequencies below W is 95% of the signal energy
Ex .

1
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Question 4
(i) Determine the aperiodic function x(t) whose Fourier transform X(ω) is the
function shown in Figure 1

Figure 1: X(jω)

(ii) Given the following is a Fourier transform pair:


 
t  ωτ 
rect ⇔ τ sinc
τ 2π
Use the duality property to calculate the Fourier transform of
 
W Wt
x(t) = sinc
π π

Question 5
(i) Determine the frequency response function of the linear time invariant RL sys-
tem shown in Figure 2, where x(t) is the input to the system, y(t) is the output,
R > 0 is the resistance and L > 0 is the inductance.

Figure 2: A linear time invariant RL system

2
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

(ii) The impulse response of the linear time invariant RC system shown in Figure
3 is given by
1 − t
h(t) = e RC u(t)
RC
where u(t) is the unit step function, R > 0 is the resistance and C > 0 is the
capacitance.

If the system is subject to a unit step input voltage x(t) = u(t), find the output
response of the system in:

(a) The frequency domain.


(b) The time domain.

and the output is y(t).

Figure 3: A linear time invariant RC system

3
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

How to find the magnitude and phase spectrum


Given an generalized frequency response X(jω) of the form

a + jb
X(jω) =
c + jd
where a, b, c, d are constants, the magnitude and phase spectra for −∞ < ω < ∞ are
√    
a2 + b 2 −1 b −1 d
|X(jω)| = √ and ∠X(jω) = tan − tan
c2 + d 2 a c

P.T.O. for solutions

4
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Solutions
Question 1:
(i) Making use of the properties of the impulse function, i.e.
Z ∞ (
∞ t=0
δ(t) = 1, δ(t) = and δ(t − t0 )x(t) = δ(t0 )x(t0 )
−∞ 0 otherwise

One finds
Z ∞ Z ∞ Z ∞
−jωt −jωt j0
X(jω) = F [x(t)] = x(t)e dt = δ(t)e dt = e δ(t) dt = 1
−∞ −∞ −∞

The magnitude and phase spectra for −∞ < ω < ∞ are


√  
0
−1
|X(jω)| = 12 = 1 and ∠X(jω) = tan =0
1

The spectrum is plotted in Figure 4.

Figure 4: |X(jω)| and ∠X(jω)

5
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

(ii) Let x(t) = e−bt u(t)


Z ∞ Z ∞ Z ∞
−jωt −bt −jωt
F {x(t)} = x(t)e dt = e e dt = e−(b+jω)t dt
−∞ 0 0
1  −(b+jω)t ∞
= e 0
−(b + jω)
1  −(b+jω)t ∞
=− e 0
b + jω
1  −∞
− e0

=− e
b + jω
1
=− [0 − 1]
b + jω
1
=
b + jω
To obtain the magnitude and phase spectra substitute b = 1 into X(jω)
1
X(jω) =
1 + jω
the magnitude and phase spectra are

12 1
|X(jω)| = √ =√
12 + ω 2 1 + ω2
 
0 ω  ω 
∠X(jω) = tan−1 − tan−1 = − tan−1
1 1 1
Plotting the spectrum in Figure 5
(iii) Let x(t) = cos(ω0 t), using the identity cos(θ) = (ejθ + e−jθ )/2
Z ∞ Z ∞  jω0 t
+ e−jω0 t −jωt

−jωt e
F {x(t)} = x(t)e dt = e dt
−∞ −∞ 2
Using the sifting property of the impulse function and the inverse Fourier trans-
form
Z ∞
−1 1 1 jω0 t
F {δ(ω − ω0 )} = δ(ω − ω0 )ejωt dω = e
2π −∞ 2π
Hence the Fourier transform
 
1 jω0 t
F e = δ(ω − ω0 )

6
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Figure 5: |X(jω)| and ∠X(jω)

1
Since 2π
is a constant, multiply both sides by 2π

F ejω0 t = 2πδ(ω − ω0 )


For −ω0 we obtain

F e−jω0 t = 2πδ(ω + ω0 )


Hence combining results


Z ∞
1 ∞  jω0 t
Z
−jωt
+ e−jω0 t e−jωt dt

F {cos(ω0 t)} = x(t)e dt = e
−∞ 2 −∞
1
= [2πδ(ω − ω0 ) + 2πδ(ω + ω0 )]
2
= πδ(ω − ω0 ) + πδ(ω + ω0 )

The spectrum is shown in Figure 6

(iv) Let x(t) = e−a|t|


(
e−at 0<t<∞
x(t) = e−at u(t) + eat u(−t) =
eat −∞ < t < 0

7
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Figure 6: X(jω)

The Fourier transform is


Z ∞ Z ∞ Z 0
−jωt −at −jωt
X(jω) = F [x(t)] = x(t)e dt = e ·e dt + eat · e−jωt dt
−∞ 0 −∞
Z ∞ Z 0
= e−(a+jω)t dt + e(a−jω)t dt
0 −∞
1  −(a+jω)t ∞ 1  (a−jω)t 0
= e 0
+ e −∞
−(a + jω) a − jω
1  −∞ 1  0
− e0 + e − e−∞
 
=− e
a + jω a − jω
1 1
=− [0 − 1] + [1 − 0]
a + jω a − jω
1 1
= +
a + jω a − jω
Finding a common denominator
a − jω a + jω
X(jω) = F [x(t)] = +
(a + jω)(a − jω) (a + jω)(a − jω)

Since j 2 = −1 we obtain
2a
X(jω) = F [x(t)] =
a2 + ω2

8
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

(v) Let x(t) = u(t) − u(t − 1). Taking the Fourier transform
Z ∞ Z 1
−jωt
X(jω) = F [x(t)] = x(t)e dt = 1 · e−jωt dt
−∞ 0
1  −jωt 1
=− e 0

1  −jω
− e−j0

=− e

1 
1 − e−jω

=

One can go further by factoring out ejω/2 ,

ejω/2 − e−jω/2
 
1  1
1 − e−jω = e−jω/2

X(jω) =
jω j2 ω/2

Using the identities sin(θ) = (ejθ − e−jθ )/j2 and the un-normalizes sinc identity
sinc(θ) = sin(θ)/θ

sin(ω/2) −jω/2 ω 
X(jω) = e = sinc e−jω/2
ω/2 2

9
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Question 2:
(i) Given that x1 (t) = x(2t), we can use the time scaling property for a = 2 together
with the result X(jω) from Question 1(v)
1 ω 
X1 (jω) = F {x1 (t)} = F {x(2t)} = X
|2| 2
 
1 ω/2 −j(ω/2)/2
= sinc e
2 2
1 ω 
= sinc e−jω/4
2 4
Compare x(t) vs x1 (t) and X(jω) vs X1 (jω) in Figure 7.

Figure 7: x(t) vs x1 (t) and X(jω) vs X1 (jω)

(ii) Let y(t) = x(t) cos(ω0 t) and using the identity cos(θ) = (ejθ + e−jθ )/2
Z ∞ Z ∞
+ e−jω0 t
 jω0 t 
−jωt e
Y (jω) = F {y(t)} = y(t)e dt = x(t) dt
−∞ −∞ 2
1 ∞
Z
x(t)ejω0 t + x(t)e−jω0 t dt

=
2 −∞
Using the frequency shifting property
1
Y (jω) = [X(ω − ω0 ) + X(ω + ω0 )]
2
10
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

(iii) Let y(t) = e−|t| cos(10t), and x(t) = e−|t| . We can use the result from Question
1(iii) for a = 1, i.e.
2a 2
X(jω) = = (1)
a2 + ω2 a=1 1 + ω2

Therefore if y(t) = x(t) cos(10t), by the frequency shifting property we find


1
Y (jω) = F {y(t)} = [X(ω − 10) + X(ω + 10)]
2
Substituting Equation 1
 
1 2 2
Y (jω) = F {y(t)} = +
2 1 + (ω − 10)2 1 + (ω + 10)2
1 1
= 2
+
1 + (ω − 10) 1 + (ω + 10)2
The spectrum is shown in Figure 8 shows that x(t) in the frequency domain is
shifted from being centred at ω = 0 to |ω| = 10 by the carrier wave.

Figure 8: x(t), y(t), X(jω), Y (jω)

11
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Question 3:
(a) (i) The energy in the time domain is
Z ∞ Z ∞ Z ∞
−at 2
Ex = 2
|x(t)| dt = (e ) dt = e−2at dt
−∞ 0 0
1  −2at ∞
= e 0
−2a
1 −∞
− e0
 
=− e
2a
1
= − [0 − 1]
2a
1
=
2a

(ii) Using the tables, the Fourier transform pair of x(t) = e−at u(t) is given by
1
e−at u(t) ⇔
a + jω

The energy in the frequency domain is therefore given by (for positive


frequencies)

1 ∞ 1 ∞
Z Z
2 1
Ex = |X(ω)| dω = dω
π 0 π 0 a + ω2
2
 ∞
1 1 −1 ω
= tan
π a a 0
  
1 1 −1 ∞
  1 0
−1
= tan − tan
π a a a a
 
1 1π 1
= − (0)
π a2 a
1
=
2a
The results in (i) and (ii) are equal, verifying Parseval’s theorem for x(t).

(b) The energy is given by


Z π
1
Ex = |X(ω)|2 dω
π 0

12
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

assuming the band [0, W ] contains 95% of the signal energy, that is Ex =
0.95/(2a) one can write

0.95 1
Z W
1

1  W  
−1 ω 1 −1 W
= dω = tan = tan
2a π 0 ω 2 + a2 aπ a 0 aπ a

Solving for W
 
0.95 −1 W
(aπ) = tan
2a a
 
W
tan(0.475π) =
a
W = tan(0.475π)a
∴ W = 12.706a

13
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Question 4:
(i) From Figure 1, one can find
(
1 |ω| ≤ W
X(ω) =
0 |ω| > W

Applying the inverse Fourier transform


Z ∞ Z W
−1 1 jωt 1
x(t) = F {X(jω)} = X(jω)e dω = 1 · ejωt dω
2π −∞ 2π −W
W
1 ejωt

=
2π jt −W
1  jW t
− e−jW t

= e
j2πt
1
= [2j sin(W t)]
j2πt
sin(W t)
=
πt  
W Wt
= sinc
π π

(ii) Interchanging the role of variables t and ω in the following Fourier transform
pair:
 
t  ωτ 
rect ⇔ τ sinc
τ 2π
we obtain
 
tτ  ω
X(t) = τ sinc and x(−ω) = rect − .
2π τ
Using the duality property, we obtain
   
tτ −ω
τ sinc ⇔ 2πrect
2π τ
Substituting τ = 2W
   
2W t −ω
2W sinc ⇔ 2πrect
2π 2W

14
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

dividing by 2π
   
2W Wt −ω
sinc ⇔ rect
2π π 2W

simplifying, taking into account the fact that rectangular pulse is an even func-
tion, i.e. rect(-a) = rect(a)
 
W Wt  ω 
sinc ⇔ rect
π π 2W

This is the same result Question 4(i).

15
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Question 5:
(i) Generally speaking, the output response of a system Y (ω) is given by the prod-
uct of the Fourier transform of the impulse response of the system, H(ω) =
F {h(t)} and the Fourier transform of the input X(ω) = F {x(t)}, that is

Y (ω) = H(ω)X(ω)

Thus, the frequency response, H(ω), is

Y (ω)
H(ω) =
X(ω)

The RL circuit dynamic equations are given by the ordinary differential equa-
tion:
di(t)
x(t) = Ri(t) + L
dt
the output response is the voltage across the inductor given by

di(t)
y(t) = L
dt
Taking the Fourier transform of the input and output equations (making use of
the differentiation and linearity properties), one obtains:
   
di(t) di(t)
F {x(t)} = F Ri(t) + L and F {y(t)} = F L
dt dt

Thus

X(ω) = RI(ω) + (jω)LI(ω) and Y (ω) = (jω)LI(ω)

The factorizing the input equation and solving for I(ω) in the output equation
we find
Y (ω)
X(ω) = [R + (jω)L] I(ω) and I(ω) =
(jω)L

Thus we obtain H(ω) by substituting I(ω) into X(ω). Solving for H(ω) we
obtain
Y (ω) (jω)L
H(ω) = =
X(ω) R + (jω)L

16
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

(ii) The output response of an LTI system is given by the convolution of the impulse
response, h(t) and the input x(t), that is
Z ∞
y(t) = x(t) ∗ h(t) = x(τ )h(t − τ ) dτ
−∞

Convolution in the time domain is equivalent to multiplication in the frequency


domain (Fourier transform convolution property). Thus, the output frequency
response function is

Y (ω) = F {x(t) ∗ h(t)} = H(ω)X(ω)

where
Z ∞ Z ∞  
−jωt 1 − 1 t
H(ω) = F {h(t)} = h(t)e dt = e RC u(t) e−jωt dt
−∞ RC
Z−∞ ∞
1 h − 1 t i  −jωt 
= e RC e dt
0 RC
Z ∞
1 −( RC 1
+jω )t
= e dt
0 RC
" #∞
1
1
= 1
RC  e ( RC )
− +jω t
− RC + jω
0
   ∞
1 RC 1+jωRC
−( RC )t
= − e
RC 1 + jωRC 0
1 h 1+jωRC i∞
=− e−( RC )t
1 + jωRC 0
1
e−∞ − e0
 
=−
1 + jωRC
1
=− [0 − 1]
1 + jωRC
1
=
1 + jωRC
and from the Fourier transform pair tables
Z ∞ Z ∞
−jωt
X(ω) = F {x(t)} = x(t)e dt = u(t)e−jωt dt
−∞ −∞
1
= πδ(ω) +

17
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Hence
  
1 1
Y (ω) = H(ω)X(ω) = πδ(ω) +
1 + jωRC jω
πδ(ω) 1
= +
1 + jωRC jω(1 + jωRC)

Since δ(ω) is only non-zero at ω = 0 we find the first term in Y (ω) to be

πδ(ω) 1
= πδ(ω)
1 + jωRC 1 + jωRC ω=0
1
= πδ(ω)
1
= πδ(ω)

By partial fraction expansion we find the second term in Y (ω) to be


1 A1 A2
= +
jω(1 + jωRC) 1 + jωRC jω
A1 (jω) + A2 (1 + jωRC)
=
jω(1 + jωRC)
In order for

1 = A1 (jω) + A2 (1 + jωRC)

A1 and A2 must be chosen as

A1 = −RC and A2 = 1

Therefore
1 RC
Y (ω) = πδ(ω) + −
jω 1 + jωRC

(iii) Taking the inverse Fourier transform of Y (ω) we obtain the output response in
the time domain
 
−1 −1 1 RC
y(t) = F {Y (ω)} = F πδ(ω) + −
jω 1 + jωRC
 
−1 1 1
=F πδ(ω) + − 1
jω RC
+ jω

18
School of Electrical and Information Engineering
University of the Witwatersrand, Johannesburg
ELEN3026A – Signals and Systems IIA (PT)

Using the Fourier transform pairs tables,


1 1
u(t) ⇐⇒ πδ(ω) + and e−t/(RC) u(t) ⇐⇒ 1
jω RC
+ jω

therefore the output response in the time domain is

y(t) = F −1 {Y (ω)} = u(t) − e−t/(RC) u(t)


= (1 − e−t/(RC) )u(t)

19

You might also like