Name ID3

EES 351: Principles of Communications 2020/1

HW 7 — Due: October 28, 11:59 PM
Lecturer: Prapun Suksompong, Ph.D.

Instructions
(a) This assignment has 5 pages.

(b) Some MATLAB scripts are available from http://www2.siit.tu.ac.th/prapun/ees351/EES351_2020_HW_4_MATLAB.zip.

(c) (1 pt) Two choices for submission:

(i) Online submission via Google Classroom

• PDF only. Paper size should be the same as the posted file.
• Only for those who can directly work on the posted PDF file using devices with pen input.
• No scanned work, photos, or screen capture.
• Your file name should start with your 10-digit student ID: “5565242231 351 HW4.pdf”
(ii) Hardcopy submission: Work and write your answers directly on a hardcopy of the posted file (not on another blank sheet
of paper).

(d) (1 pt) Write your first name and the last three digits of your student ID in the spaces provided on the upper-right corner of this page.

(e) (8 pt) Try to solve all problems.

(f) Late submission will be heavily penalized.

78 Problem 1. Consider the two receiver

Optimum signals s1for
(t) binary
and s2 (t)
datashown in Figure 6.1. Note that V and
transmission
Tb are some positive constants. Your answers should be given in terms of them.

s1(t) s2(t)

V V

Tb
t t
0 0 Tb

−V
(a)
Figure 7.1: Signal set for Question 1
φ 1(t) φ 2(t)
(a) Find the energy in each signal.
1 Tb 1 Tb

Tb
t t
0 7-1 0 Tb

−1 Tb
(b)
EES 351 HW 7 — Due: October 28, 11:59 PM 2020/1

(b) Are they energy signals?

(c) Are they power signals?

(d) Find the (average) power in each signal.

(e) Are the two signals s1 (t) and s2 (t) orthogonal? (Two signals are orthogonal if their
inner product is 0.)

Problem 2. (Power Calculation) For each of the following signals g(t), find (i) its corre-

2
2
sponding power Pg = |g (t)| , (ii) the power Px = |x (t)| of x (t) = g (t) cos (10t), and
(iii) the power Py = |y (t)|2 of y (t) = g (t) cos (50t)



(a) g (t) = 3 cos (10t + 30◦ ).

7-2
EES 351 HW 7 — Due: October 28, 11:59 PM 2020/1

(b) g (t) = 3 cos (10t + 30◦ ) + 4 cos (10t + 120◦ ). (Hint: First, use phasor form to combine
the two components into one sinusoid.)

(c) g (t) = 3 cos (10t) + 3 cos (10t + 120◦ ) + 3 cos (10t + 240◦ )

7-3
EES 351 HW 7 — Due: October 28, 11:59 PM 2020/1

Extra Question
Here is an optional question for those who want more practice.

Problem 3. Consider a signal g(t). Recall that |G(f )|2 is called the energy spectral
density of g(t). Integrating the energy spectral density over all frequency gives the signal’s
total energy.R Furthermore, the energy contained in the frequency band I can be found from
the integral I |G(f )|2 df where the integration is over the frequencies in band I. In particular,
if the band is simply an interval of frequency from f1 to f2 , then the energy contained in
this band is given by Z f2
|G(f )|2 df. (7.1)
f1

In this problem, assume

g(t) = 1[−1 ≤ t ≤ 1].

(a) Find the (total) energy of g(t).

(b) Figure 6.2 define the main lobe of a sinc pulse. It is well-known that the main lobe
of the sinc function contains about 90% of its total energy. Check this fact by first
computing the energy contained in the frequency band occupied by the main lobe and
then compare with your answer from part (a).
Hint: Find the zeros of the main lope. This give f1 and f2 . Now, we can apply (6.1).
MATLAB or similar tools can then be used to numerically evaluate the integral.

7-4
EES 351 sinc function
HW 7 — Due: October 28, 11:59 PM 2020/1

Main lobe
(null to null)

Figure 7.2: Main lobe of a sinc pulse

1

(c) Suppose we want to include more energy by considering wider frequency band. Let
this band be the interval I = [−f0 , f0 ]. Find the minimum value of f0 that allows the
band to capture at least 99% of the total energy in g(t).

7-5
 Name ID3

EES 351: Principles of Communications 2020/1

HW 8 — Due: November 11, 11:59 PM
Lecturer: Prapun Suksompong, Ph.D.

Instructions
(a) This assignment has 11 pages.

(b) (1 pt) Two choices for submission:

(i) Online submission via Google Classroom

• PDF only. Paper size should be the same as the posted file.
• Only for those who can directly work on the posted PDF file using devices with pen input.
• No scanned work, photos, or screen capture.
• Your file name should start with your 10-digit student ID: “5565242231 351 HW4.pdf”
(ii) Hardcopy submission: Work and write your answers directly on a hardcopy of the posted file (not on another blank sheet
of paper).

(c) (1 pt) Write your first name and the last three digits of your student ID in the spaces provided on the upper-right corner of this page.

(d) (8 pt) Try to solve all problems.

(e) Late submission will be heavily penalized.

Problem 1. Consider the impulse train G(f ) shown on the right in Figure 8.1. Plot g(t).

1 1 1 1 1 1 1

𝑡 
 
 𝑓
2 2  1 6 4 2 0 2 4 6

Figure 8.1: A train of impulses in the frequency domain

8-1
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

Problem 2. Consider a signal r(t) = ∞ j4πkt

P
k=−∞ 2e . Plot r(t) and R(f ).
Hint: Don’t try to actually plot each complex-expo. func. and add them. It is quite hopeless
to determine their combination.

Problem 3. Consider an AM transmitter.

(a) Suppose the message is m(t) = 4 cos(10πt) and the transmitted signal is

xAM (t) = A cos(100πt) + m(t) cos(100πt).

Find the value of A which yields the modulation index in each part below.

(i) µ = 50%
(ii) µ = 100%
(iii) µ = 150%

(b) Suppose the message is m(t) = α cos(10πt) and the transmitted signal is

xAM (t) = 4 cos(100πt) + m(t) cos(100πt).

Find the value of α which yields the modulation index in each part below.

(i) µ = 50%
(ii) µ = 100%
(iii) µ = 150%

8-2
 th e follow ing.
(a) Sketch the spectrum of m (c).
(b) Sketch the spec trum of the DSB-SC signal m(c) cos IO,OOO.rrt.
(c) Identify the upper sideband (USB ) and the lower s ideband (LSB) spectra.

4.2-2 Repeat Prob. 4.2-1 if (i) m(t ) = sine ( lOOm); (ii) m(t) = ( l + t 2 )- 1; (iii) m(t) = e- IO it- l l _
EES 351 Observe thatHWe- 10 lt -811—
is e-Due: November
IO irl delayed by l second.11, 11:59
For the PM
last case yo u need to consider both2020/1
th e amplitude and the phase spectra.

4.2-3You are asked to design a DSB-SC modulator to generate a mod ulated signal km(l) cos (we t +8),
Extra Questions where m(t) is a signal band -limited to B Hz. Figure P4 .2-3 shows a DSB-SC modu lator ava il able
in the stoc kroo m. The ca tTi er generato r ava il able ge nerates not cos We t, but cos 3 We i . Exp lain
whether you wo uld be able to generate the desired signal using onl y thi s equip ment. You may use
Here are someany
optional
kind of filquestions
ter yo u li k~ . for those who want more practice.

Problem 4. You (a) What

are kind
asked of filter is req uireda inDSB-SC
to design Fi g. P4 .2 -3?
modulator to generate a modulated signal
(b) Determ ine the signal spectra
km(t) cos(2πfc t), where m(t) is a signal band-limited at points b and c, and to
ind icate
B Hz. th e frequ ency bands
Figure occ upied
8.2 shows a by
DSB-SC
these spectra.
modulator available in the stockroom. Note that, as usual, ωc = 2πfc . The carrier generator
(c) What is the minim um usable value of eve?
available generates not cos (2πfc t), but cos3 (2πfc t). Explain whether you would be able to
(d) Wo ul d this sc heme wo rk if the ca rri er ge nerator output were sin 3 cvc l') Explain .
generate the desired signal using only this equipment. You may use any kind of filter you
(f) Wo ul d thi s sc he me work if th e carri er ge nerator out put were cos" cvet for an y integer n ::: 2.?
like. [Lathi and Ding, 2009, Q4.2-3]
Figure P.4.2-3
M (.f )
m (t)
., I Filter
@..__ __,

-8 B
J- -
(a) (b)

Figure
4.2-4 You are asked to design a DSB-SC 8.2: Problem
modulator to generate 4a modul ated signal km(l ) cos wet with
the carrier frequency f e = 500kHz (we = 2.rr x 500, 000). The follo wing equipment is ava ilable
in the stockroom: (i) a signal generator of frequency I 00 k.H z; (ii) a ring modulator; (iii ) a bandpass
fi Iter tuned to 500 k.Hz.
(a) A real-valued signal r(t) that is even and periodic with period T can be expanded 0
using Fourier series into

r(t) = c0 + a1 cos(2πf0 t) + a2 cos(2π(2f0 )t) + a3 cos(2π(3f0 )t) + · · · (8.1)

1
where f0 = T0
. Consider the signal r(t) = cos3 (2πfc t).

(i) Is it periodic?

(ii) Is it even?

(iii) Expand r(t) = cos3 (2πfc t) into a linear combination of cos(2π(nfc )t) as in (8.1)
above.

8-3
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

(b) What kind of filter is required in Figure 8.2?

(c) Determine the signal spectra at points (b) and (c) in Figure 8.2, and indicate the
frequency bands occupied by these spectra.

8-4
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

(d) What is the minimum usable value of fc ?

(e) Would this scheme work if the carrier generator output were cos2 (2πfc t)? Explain.

8-5
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

Problem 5. Consider the basic DSB-SC transceiver with time-delay channel presented in
class. Recall that the input of the receiver is
√
x (t − τ ) = m (t − τ ) 2 cos (ωc (t − τ ))
F
where m(t) −
)−
−*− M (f ) is bandlimited to B, i.e., |M (f )| = 0 for |f | > B. We also assume
−1
F
that fc  B.
√ √

(a) Suppose that, at the receiver, we multiply by 2 cos ((ωc t) − θ) instead of 2 cos (ωc t)
as illustrated in Figure 8.3. Assume

1, |f | ≤ B
HLP (f ) =
0, otherwise.

Find m̂(t) (the output of the LPF).

v t 
x t     H LP  f  m̂  t 

2 cos  ct    

Figure 8.3: Receiver for Problem 5a

8-6
 HWR

EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

HWR

Figure 8.4: Receiver for Problem 5b

(b) Use√the same assumptions as part (a). However, at the receiver, instead of multiplying
by 2 cos ((ωc t) − θ), we pass x(t − τ ) through a half-wave rectifier (HWR) as shown
in Figure 8.4b.
Make an extra assumption that m(t) ≥ 0 for all time t and that the half-wave rectifier
input-output relation is described by a function f (·):

x, x ≥ 0,
f (x) =
0, x < 0.

Find m̂(t) (the output of the LPF).

8-7
 b. (5 pt) Sketch the spectrum of the DSB-SC signal m  t  cos  5,000 t  . No
explanation is needed.

EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

7. (8 pt) Suppose m  t  
 M  f  is bandlimited to W, i.e., M  f   0 for
f W
F
Problem 6 .(M2011Q7).
Consider the Suppose −−
followingmDSB-SC
(t) )−*−transceiver.
M (f ) is bandlimited to W , i.e., |M (f )| = 0
−1 F
for |f | > W . Consider a DSB-SC transceiver shown in Figure 8.5.

Transmitter (modulator)

x t 
m t   Delayed by 

cos ct 

x t   

v t 
FWR H LP  f  y t 

Receiver (demodulator)

Figure 8.5: A DSB-SC transceiver

Also assume that f c W and that H LP  f   1, f W
1, |f | ≤ W
Also assume that fc  W and that HLP (f ) = 0, otherwise.
0, otherwise.
MakeMake an extra
an extra assumption
assumption that  t ≥ 00 for
thatmm(t) for all
alltime
timet and thatthat
t and the full-wave
the full-wave rectifier
(FWR)rectifier (FWR) input-output
input-output relation isby
relation is described described
a function F W R (·): f FWR   :
by affunction
, x x 0,
xx,
f  
R (x) = 
fF WFWR x  ≥ 0,
 x−x,
, x x 0.< 0.

(a) (Questions
Recall that theonhalf
start the -wave rectifier input-output relation is described by a function
nextpage.)
x, x ≥ 0,
fHW R (·) : fHW R (x) = We have seen in Problem 5b that when the
0, x < 0.

8-8
Page 5 of 12
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

receiver uses half -wave rectifier,

v (t) = x (t − τ ) × gHW R (t − τ )

where gHW R (t) = 1 [cos (ωc t) ≥ 0].

(i) The receiver in this question uses full-wave rectifier. Its v(t) can be described in
a similar manner; that is

v (t) = x (t − τ ) × gF W R (t − τ ) .

Find gF W R (t). Hint: gF W R (t) = c1 × gHW R (t) + c2 for some constants c1 and c2 .
Find these constants.

8-9
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

(ii) Recall that the Fourier series expansion of gHW R (t) is given by
 
1 2 1 1 1
gHW R (t) = + cos ωc t − cos 3ωc t + cos 5ωc t − cos 7ωc t + . . . .
2 π 3 5 7

Find the Fourier series expansion of gF W R (t).

(b) Find y(t) (the output of the LPF).

Problem 7. Would the scheme in Problem 4 work if the carrier generator output were
cosn ωc t for any integer n ≥ 2?

8-10
EES 351 HW 8 — Due: November 11, 11:59 PM 2020/1

Problem 8 (M2011Q5). In this question, you are provided with a partial proof of an im-
portant result in the study of Fourier transform. Your task is to figure out the quantities/-
expressions inside the boxes labeled a,b,c, and d.
P∞
We start with a function g(t). Then, we define x (t) = g (t − `T ). It is a sum that
`=−∞
involves g(t). What you will see next is our attempt to find another expression for x(t) in
terms of a sum that involves G(f ).
∞
P
To do this, we first write x(t) as x (t) = g (t) ∗ δ (t − `T ). Then, by the convolution-
`=−∞
in-time property, we know that X(f ) is given by
∞
X  
X (f ) = G (f ) × a δ f+ b
`=−∞

R∞
We can get x(t) back from X(f ) by the inverse Fourier transform formula: x (t) = X (f ) ej2πf t df .
−∞
After plugging in the expression for X(f ) from above, we get
Z∞ ∞
X  
x (t) = ej2πf t G (f ) a δ f+ b df
−∞ `=−∞

Z∞ ∞
X  
= a ej2πf t G (f ) δ f + b df.
−∞ `=−∞

By interchanging the order of summation and integration, we have

∞ Z∞
X  
x (t) = a ej2πf t G (f ) δ f + b df.
`=−∞−∞

We can now evaluate the integral via the sifting property of the delta function and get
∞  
c
X
x (t) = a e G d .
`=−∞

8-11
 Name ID3

EES 351: Principles of Communications 2020/1

HW 9 — Due: November 25, 11:59 PM
Lecturer: Prapun Suksompong, Ph.D.

Instructions
(a) This assignment has 9 pages.

(b) (1 pt) Two choices for submission:

(i) Online submission via Google Classroom

• PDF only. Paper size should be the same as the posted file.
• Only for those who can directly work on the posted PDF file using devices with pen input.
• No scanned work, photos, or screen capture.
• Your file name should start with your 10-digit student ID: “5565242231 351 HW9.pdf”
(ii) Hardcopy submission: Work and write your answers directly on a hardcopy of the posted file (not on another blank sheet
of paper).

(c) (1 pt) Write your first name and the last three digits of your student ID in the spaces provided on the upper-right corner of this page.

(d) (8 pt) Try to solve all problems.

(e) Late submission will be heavily penalized.

√
Problem 1. Find the instantaneous frequency of the signal g(t) = 3 2 cos (12t3 + t2 )

(a) at time t = 0

(b) at time t = 2

9-1
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

Problem 2. Recall that, in QAM system, the transmitted signal is of the form
√ √
xQAM (t) = m1 (t) 2 cos (2πfc t) + m2 (t) 2 sin (2πfc t) .

We want to express xQAM in the form

√
xQAM (t) = 2E(t) cos(2πfc t + φ(t)),

where E(t) ≥ 0 and φ(t) ∈ (−180◦ , 180◦ ]. (This shows that QAM can be expressed as a
combination of amplitude modulation and phase modulation.)
Consider m1 (t) and m2 (t) plotted in Figure 9.1.
Draw the corresponding E(t) and φ(t).

m1(t)
1
t
-1
1
t
m2(t)
-1

Figure 9.1: m1 (t) and m2 (t) for Problem 2

9-2
 EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

Problem 3. Consider the message m(t) along with the carrier signal cos(2πfc t + φ) plotted
in Figure 9.2. Note that m(1.5) = 40.

50
m(t)

0
-50
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10
cos(2  fc t + )

1
0
-1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10
100
xAM(t)

0
-100
Figure 9.2: The message and the carrier signals for Problem 3.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10
(a) 1Find the carrier frequency fc from the plot. (Hint: It is an integer.)
xFM(t)

0
-1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
(b) Sketch the following signals. Make sure thatt (the unspecified parameter(s) are selected
x 10
-3

1such that) the important “features” of the graphs can be seen clearly.
xPM(t)

0
(i) xAM (t) = (A + m (t)) cos (2πfc t + φ) whose modulation index µ = 100%.
-1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10

Rt
 
(ii) xFM (t) = A cos 2πfc t + φ + 2πkf m (τ )dτ . Assume m(t) = 0 for t < 0.
−∞

9-3
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

π
(iii) xPM (t) = A cos (2πfc t + φ + kp m (t)) with kp = mp
.

Problem 4. Consider the FM transmitted signal

 
Zt
xFM (t) = A cos 2πfc t + φ + 2πkf m (τ )dτ  ,
−∞

where fc = 5 [kHz], A = 1, and kf = 75. The message m(t) is shown in Figure 9.3.
40

20
m(t)

-20

-40
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10
1
cos(2  fc t + )

0.5 Figure 9.3: The message m(t) for Problem 4

0
The magnitude spectrum |XFM (f )| is plotted in Figure 9.4.
-0.5
(a) Find the
-1 values of f1 , f2 , and f3 .
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10
1

0.5
xFM(t)

-0.5
(b) Find the
-1
width W in Figure 9.4.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
t -3
x 10

9-4
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

𝑋FM 𝑓

f
f1 f2 f3
BW W

Figure 9.4: The magnitude spectrum |XFM (f )| for Problem 4

1

(c) Find the bandwidth denoted by BW in Figure 9.4.

Extra Questions
Here are some optional questions for those who want more practice.

Problem 5. Recall that, in QAM system, the transmitted signal is of the form
√ √
xQAM (t) = m1 (t) 2 cos (2πfc t) + m2 (t) 2 sin (2πfc t) .

In class, we have shown that

n √ o
LPF xQAM (t) 2 cos (2πfc t) = m1 (t) .

Give a similar proof to show that

n √ o
LPF xQAM (t) 2 sin (2πfc t) = m2 (t) .

9-5
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

Problem 6. In quadrature amplitude modulation (QAM ) or quadrature multiplexing , two

baseband signals m1 (t) and m2 (t) are transmitted simultaneously via the following QAM signal:
√ √
xQAM (t) = m1 (t) 2 cos (ωc t) + m2 (t) 2 sin (ωc t) .

An error in the phase or the frequency of the carrier at the demodulator in QAM will result in loss
and interference between the two channels (cochannel interference).
In this problem, show that
n √ o
LPF xQAM (t) 2 cos ((ωc + ∆ω) t + δ) = m1 (t) cos ((∆ω) t + δ) − m2 (t) sin ((∆ω) t + δ)
n √ o
LPF xQAM (t) 2 sin ((ωc + ∆ω) t + δ) = m1 (t) sin ((∆ω) t + δ) + m2 (t) cos ((∆ω) t + δ) .

9-6
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

Problem 7. In QAM system, the transmitted signal is of the form

√ √
xQAM (t) = m1 (t) 2 cos (2πfc t) + m2 (t) 2 sin (2πfc t) .

We want to express xQAM in the form

√
xQAM (t) = 2E(t) cos(2πfc t + φ(t)),

where E(t) ≥ 0 and φ(t) ∈ (−180◦ , 180◦ ].

In each part below, we consider different examples of the messages m1 (t) and m2 (t).
(a) Suppose m1 (t) = cos(2πBt) and m2 (t) = sin(2πBt) where 0 < B  fc . Find E(t) and φ(t).
Hint: cos α cos β + sin α sin β = cos(α − β)

(b) Suppose m1 (t) = cos(2πBt) and m2 (t) = 2 sin(2πBt). Let fc = 5 and B = 2. Use MATLAB to
plot the corresponding E(t) and φ(t) from t = 0 to t = 5 [sec]. (Hint: the function angle or
atan2 will be helpful here.)

9-7
The plot for part (a):

The plot for part (b):

EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

Problem 8. Consider a complex-valued signal x(t) whose Fourier transform is X(f ).

(a) Find and simplify the Fourier transform of x∗ (t).

(b) Find and simplify the Fourier transform of Re {x(t)}.

• Hint: x(t) + x∗ (t) = ?

F
−−
Problem 9. Consider a (complex-valued) baseband signal xb (t) )−−*
−
− Xb (f ) which is band-limited
F −1
to B, i.e., |Xb (f )| = 0 for |f | > B. We also assume that fc  B.

(a) The passband signal xp (t) is given by

√ n o
xp (t) = 2 Re ej2πfc t xb (t) .

Find and simplify the Fourier transform of xp (t).

9-8
EES 351 HW 9 — Due: November 25, 11:59 PM 2020/1

(b) Find and simplify    

 
√ √
 n o 

j2πfc t −j2πfc t
LPF 2  2 Re e
 xb (t)  e
 .

 | {z } 

 
xp (t)

Assume that the frequency response of the LPF is given by


1, |f | ≤ B
HLP (f ) =
0, otherwise.

9-9
 Solution

Parts a and b consider cosine with no phase shift. Therefore, we will appply the folding technique.

Parts c and d consider complex expo. Therefore, we will appply the tunneling technique.
Derivation of the "folding" and the "tunneling" techniques
 First, we need to look at the freq. domain representation of the given signal to find the maximum non-zero
frequency B.

multiplication in time domain

convolution in freq. domain

 1

0.5

−0.5

−1
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Seconds

1.5
Magnitude

0.5

0
−500 −400 −300 −200 −100 0 100 200 300 400 500
Frequency [Hz]
Problem 4: Aliasing and periodic square wave
(a)

-1

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Seconds

0.4
(b) sgn(cos𝜔0 𝑡) = 2 × 1[cos𝜔0 𝑡 ≥ 0] − 1. So, a = 2 and b = -1.
0.3
(c)
Magnitude

0.2
(c.i) From
0.1
1 2 1 1 1
1[cos𝜔0 𝑡 ≥ 0] = 2 + 𝜋 (cos𝜔0 𝑡 − 3 cos3𝜔0 𝑡 + 5 cos5𝜔0 𝑡 − 7 cos7𝜔0 𝑡 + ⋯ ),
0
-500 -400 -300 -200 -100 0 100 200 300 400 500
and Frequency [Hz]

𝑔(𝑡) = sgn(cos𝜔0 𝑡) = 2 × 1[cos𝜔0 𝑡 ≥ 0] − 1

we have
4 1 1 1
𝑔(𝑡) = sgn(cos𝜔0 𝑡) = 𝜋 (cos𝜔0 𝑡 − 3 cos3𝜔0 𝑡 + 5 cos5𝜔0 𝑡 − 7 cos7𝜔0 𝑡 + ⋯ ).

where 𝜔0 = 2𝜋𝑓0 .

(c.ii)

(c.ii.i) Theoretically, 𝐺(𝑓) should have spikes (impulses) at all the odd-integer
multiples of f0 Hz. The center spikes (at f0) should be the largest among them.
Here, 𝑓0 = 10. Therefore, the largest spike occurs at ±10 Hz, followed by smaller
spikes at all the odd-integer multiples (i.e., at ±30, ±50, ±70, etc).

(c.ii.ii) The sinc function is simply the Fourier transform of the rectangular windows.
Because the area of the rectangular window is 1×2 = 2, its Fourier transform
(which is a sinc function) has its peak value of 2. This is further scaled by a factor
 of ½ from the cosine. Therefore, each “impulse” (“sinc”) that we see should have
its height being the same as the coefficient of corresponding cosine. For
4
example, at f0, the coefficient of the cosine is 𝜋. Therefore, we expect the
4
height of the “impulse” at f0 to be 𝜋  1.2732.

The theoretical values for the height of other impulses is shown by the pink
circles in the plot. We see that our predicted values match the plot quite well.

sign(cos(2: f t )) where f = 10 using sampling int erval TS = 0.001

1

0.5

-0.5

-1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Seconds

1.5
Magnitude

0.5

0
-500 -400 -300 -200 -100 0 100 200 300 400 500
Frequency [Hz]

(d) Here, we also shows the time-domain plots as well. In the time domain, the switching
between the values -1 and 1 should be faster as we increase f0. All the plots here are
adjusted so that they show 10 periods of the “original signal” in the time domain. (This is
done so that the distorted shape (if any) of the waveform in the time domain is visible.)

(d.i) From the plots, as we increase f0 from 10 to 20 Hz, the locations of spikes changes from
all the odd-integer multiples of 10 Hz to all the odd-integer multiples of 20 Hz. In
particular, we see the spikes at 20, 60, 100, 140, 180, 220, 260, 300, 340,
380, 420, 460. Note that plotspect (by the way that it is coded) only plots from [-
fs/2,fs/2). So, we see a spike at -500 but not 500. Of course, the Fourier transform of the
sampled waveform is periodic and hence when we replicate the spectrum every f s, we
will have a spike at 500. Note that, in theory, we should also see spikes at 540, 580,
620, 660, and so on. However, because the sampling rate is 1000 [Sa/s], these high
 frequency spikes will suffer from aliasing and “fold back”1 into our viewing window [-
fs/2,fs/2). However, they fall back to the frequencies that already have spikes (for
example, 540 will fold back to 460, and 580 will fold back to 420) and therefore the
aliasing effect is not easily noticeable in the frequency domain.

sign(cos(2: f t )) where f = 20 using sampling int erval TS = 0.001 sign(cos(2: f t )) where f = 40 using sampling int erval TS = 0.001
1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.05 0.1 0.15 0.2 0.25
Seconds Seconds

1.5 1.5
Magnitude

Magnitude
1 1

0.5 0.5

0 0
-500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500
Frequency [Hz] Frequency [Hz]

(d.ii) When f0 = 40 Hz, we start to see the aliasing effect in the frequency domain. Instead of
seeing spikes only at 40, 120, 200, 280, 360, 440, the spikes at higher
frequencies (such as 520, 600, and so on) fold back to lower frequencies (such as
480, 400, and so on). The plot in the time domain still looks quite OK with small
visible distortion.

(d.iii) At high fundamental frequency f0 = 100 Hz, we see stronger effect of aliasing. In the
time domain, the waveform does not look quite “rectangular”.
sign(cos(2: f t )) where f = 100 using sampling int erval TS = 0.001 sign(cos(2: f t )) where f = 100 using sampling int erval TS = 0.0001
1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Seconds Seconds

1.5 1.5
Magnitude

Magnitude

1 1

0.5 0.5

0 0
-500 -400 -300 -200 -100 0 100 200 300 400 500 -5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
Frequency [Hz] Frequency [Hz]

In the frequency domain, we only see the spikes at 100, 300, and 500. These are at
the correct locations. However, there are too few of them to reconstruct a square
waveform. The rest of the spikes are beyond our viewing window. We can’t see them

1
Because the squarewave is real and even, the Fourier transform is also real and even. Therefore, the “folding
effect” is “equivalent” to the “tunneling effect”.
 directly because they fold back to the frequencies that are already occupied by the
lower frequencies. Note also that the predicted height (pink circles) at 300 Hz is quite
different from the plotspect value. This is because the content from the folded-back
higher-frequencies is being combined into the spikes.
Our problem can be mitigated by reducing the sampling interval to TS = 1/1e4 instead of
TS = 1/1e3 as shown by the plot on the right above.
(d.iv) Finally, at the highest frequency f0 = 300 Hz, if we still use T = 1/1e3, the waveform
will be heavily distorted in the time domain. This is shown in the left plot below. We
have large spikes at 300 as expected. However, the next pair which should occur at
900 is out of the viewing window and therefore folds back to 100.
sign(cos(2: f t )) where f = 300 using sampling int erval TS = 0.001 sign(cos(2: f t )) where f = 300 using sampling int erval TS = 0.0001
1 1

0.5 0.5

0 0

-0.5 -0.5

-1 -1
0 0.005 0.01 0.015 0.02 0.025 0.03 0 0.005 0.01 0.015 0.02 0.025 0.03
Seconds Seconds

1.5 1.5
Magnitude

Magnitude

1 1

0.5 0.5

0 0
-500 -400 -300 -200 -100 0 100 200 300 400 500 -5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
Frequency [Hz] Frequency [Hz]

Again, the aliasing effect can be mitigated by reducing the sampling time to T = 1/1e4
instead of T = 1/1e3. Now, more spikes show up at their expected places. Note that we
can still see a lot of small spikes scattered across the frequency domain. These are again
the spikes from higher frequency which fold back to our viewing window.
Q5 Sinc Reconstruction of Sinc: MATLAB Simulation
Wednesday, November 16, 2016 8:34 PM