UNIVERSIDAD DE LAS FUERZAS ARMADAS ESPE

SIGNAL AND SYSTEMS

Fausto Granda G.

Sangolqui - Ecuador
Signal and Systems

Fourier Transform

• Oppenheim, A. Willsky, and H. Nawab, Signals and Systems, 2ª edición, 1997, Prentice Hall, ISBN # 0-13-814757-4.
• Hwei P. Hsu, “Signals and Systems”, 2nd Edition, McGrawHill Schaum Outlines, ISBN: 978-0-07-163473-1, 2011.
• Chaparro Luis, “Signal and Systems using Matlab”, Elsevier, Oxford UK, ISBN 978-0-12-374716-7 , 2011
• https://see.stanford.edu/materials/lsoftaee261/book-fall-07.pdf, 10/07/2020
Signal and Systems

One of the Fundamental Secrets of the Universe:

All waveforms, no matter what you scribble or

observe in the universe, are actually just the sum of
simple sinusoids of different frequencies

The Fourier Transform decomposes a waveform -basically any real world waveform- into
sinusoids

The Fourier Transform, in essence, consists of a different method of viewing the universe
(that is, a transformation from the time domain to the frequency domain).
Fourier Transform: Introduction
Methods based on the Fourier transform are used in virtually all areas of engineering
and science and by virtually all engineers and scientist.

Fourier analysis was originally concerned with representing and analyzing periodic
phenomena, via Fourier series, and later with extending those insights to nonperiodic
phenomena, via the Fourier transform.

A discrete set of frequencies in the periodic case becomes a continuum of frequencies in

the nonperiodic case, the spectrum is born, and with it comes the most important principle
of the subject:

“ Every signal has a spectrum and is determined by its spectrum. You can analyze the signal
either in the time (or spatial) domain or in the frequency domain.”
Fourier Transform: Introduction
Whereas for periodic signals the complex exponential building blocks are harmonically
related, for aperiodic signals they are infinitesimally close in frequency, and the
representation in terms of a linear combination takes the form of an integral rather than a
sum.

The resulting spectrum of coefficients in this representation is called the Fourier

transform,

The synthesis integral itself, which uses these coefficients to represent the signal as a linear
combination of complex exponentials, is called the inverse Fourier transform.
Fourier Transform: magnitude and phase

Fourier Spectra: If x(t) is Real

Magnitude Phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Fourier Transform: magnitude and phase
Magnitude Phase
Fourier Transform: magnitude and phase
Dirichlet conditions: convergence

Fourier and Laplace connection:

Only if x(t) is absolutely integrable.

Fourier Transform: magnitude and phase
a) 𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖 𝒕 𝒄𝒐𝒏 𝒂 >𝟎

12 + 02 1
𝑥 𝑤 = =
𝑎2 + 𝑤2 𝑎2 + 𝑤 2
Fourier Transform: magnitude and phase
a) 𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖 𝒕 𝒄𝒐𝒏 𝒂 >𝟎

𝑤 𝑤
∅ 𝑤 = 𝑡𝑎𝑛−1 0 −𝑡𝑎𝑛−1 = −𝑡𝑎𝑛−1
𝑎 𝑎 𝜋
2 a=2

𝜋
−
2

a=5
Fourier Transform: magnitude and phase
a) 𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖 𝒕 𝒄𝒐𝒏 𝒂 >𝟎 FourierTransform[e^(-a*t)*unitstep(t),t,w]

12 + 02 1
𝑥 𝑤 = =
𝑎2 + 𝑤2 𝑎2 + 𝑤 2

plot (abs(1/(1+i*w))), w=-5..5 plot(phase(1/(1+i*w))), w=-5..5

Fourier Transform: magnitude and phase
a) 𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖 𝒕 𝒄𝒐𝒏 𝒂 >𝟎

w = logspace(-1,3);
X1 = 1./(2+j*w);
subplot(2,1,1)
semilogx(w,(abs(X1)));
%semilogx(w,20*log10(abs(X1)));
%magnitude in dB
title('Magnitude'); ylabel('Magnitude’);
xlabel('Log of Frequency’);grid on

subplot(2,1,2)
semilogx(w,angle(X1)*180/pi);
title('Phase'); ylabel('Phase
(°)');xlabel('Log of Frequency')
grid on
Fourier Transform: magnitude and phase
a) 𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖 𝒕 𝒄𝒐𝒏 𝒂 >𝟎
Fourier Transform: magnitude and phase

b) 𝒙 𝒕 = 𝒆−𝒂|𝒕| 𝒄𝒐𝒏 𝒂 > 𝟎

plot(abs((2*a)/(a^2 + w^2))), w=-10 to 10, a=5

Fourier Transform: magnitude and phase
Inverse Fourier Transform

La regla de l'Hopital: el
límite del cociente de una
función es igual al límite del
cociente de las derivadas.

Gibss phenomenon: As W->ꝏ , this

signal converges to x(t) everywhere,
except at the discontinuities. The signal
𝑤𝑇1 = ±𝑛π exhibits ripples near the discontinuities.
 Fourier Transform: magnitude and phase

Functions of the form given in the two

last exercises, arise frequently in
Fourier analysis and in the study of
LTI systems and are referred to as
sinc functions. A commonly used
precise form for the sine function is:

Inverse Fourier Transform

Fourier Transform: inverse relationship that exists between the time
and frequency domains,

…as W increases, X(jw) becomes broader, while the main peak of x(t) at t = 0 becomes higher and the width
of the first lobe of this signal becomes narrower.
Fourier Transform: inverse relationship that exists between the time
and frequency domains,

syms x w % symbolic variables

fun = @(x)x*0+1; % f(x)=1 constant
b1=int(fun*exp(-1i*w*x),-1/2,1/2) %
integral
fplot(b1,[-20,20]); grid on % plot
Fourier Transform: inverse relationship that exists between the time
and frequency domains,
Fourier Transform: inverse relationship that exists between the time
and frequency domains: WATCH OUT!!!
Wolfram: different response???

integral (A/B*t+A)*e^(-i*w*t) from -B to 0 + integral (-A/B*t+A)*e^(-

i*w*t) from 0 to B

What is the
correct answer?
Fourier Transform: inverse relationship that exists between the time
and frequency domains: WATCH OUT!!!
Wolfram: different response???

integral (A/B*t+A)*e^(-i*w*t) from -B to 0 + integral (-A/B*t+A)*e^(-

i*w*t) from 0 to B

syms A B x w % symbolic variables

fun = @(x)(A/B)*x+A; % f(x)=1 constant
fun1 = @(x)(-A/B)*x+A; % f(x)=1 constant
b1=int(fun*exp(-1i*w*x),-B,0)+ int(fun1*exp(-
1i*w*x),0,B); % integral
b1=expand(simplify(b1));
pretty(b1)
% f(x)=1 constant
b2=subs(b1,[A,B],[5,2]) % evaluate b1 for A=5, B=2
ezplot(b2,[-10,10]); % plot the function
Fourier Transform: exercise.

integral (A/B*t+A)*e^(-i*w*t) from -B to 0 + integral (-A/B*t+A)*e^(-

=2 i*w*t) from 0 to B

= -1 =1
Piecewise[{{2*x+2, -1≤ x < 0}, {-2*x+2, 0 ≤x <1}}]

2/1*((sinc(x/2))^2) 2/1*(((sin(x/2)/(x/2))^2)

¿Cuál
respuesta es
correcta?
2*1*((sinc(1*x/2))^2) 2*1*(((sin(1*x/2)/(1*x/2))^2)
Fourier Transform: exercise.

=5

= -2 =2
Piecewise[{{5/2*x+5, -2≤ x < 0}, {-5/2*x+5, 0 ≤x <2}}]

integral (5/2*t+5)*e^(-i*w*t) from -2 to 0 + integral

(-5/2*t+5)*e^(-i*w*t) from 0 to 2
Fourier Transform: inverse relationship that exists between the time
and frequency domains,
Find the magnitude and phase of the spectrum of x(t).

1) Find x(t)

2) Find X(w) using the definition.

3) Find X(w) in terms of sinc function

4) Plot the magnitude of X(w)

5) Plot the phase of X(w)

6) Plot x(t) in Wolfram.

7) Find the Fourier Transform X(w) in Wolfram

8) Plot the magnitude of X(w) in wolfram.

9) Compare the results of the points 4 and 8

Fourier Transform: inverse relationship that exists between the time
and frequency domains,
Find the magnitude and phase of the spectrum of x(t).

1) Find x(t)
Piecewise[{{-B/A*t, -1≤t < 0}}] +Piecewise[{{B/A*t, 0≤t ≤ 1}}]

2)Find X(w) using the definition.

integrate ((-B*t/A)*e^(-iwt)) from t= -1 to 0 + integrate ((B*t/A)*e^(-iwt)) from t= 0 to 1
Result: (2 B (w sin(w) + cos(w) - 1))/(A w^2)

3) Find X(w) in terms of sinc function

(2B/AW)*Sin(w) -(B/A)*sinc^2(w/2)

4) Plot the magnitude of X(w)

A= 1, B=2
plot (2*2/(w*1))*Sin(w) -(2/1)*sinc^2(w/2), from w=-10 to 10

5) Plot the phase of X(w)

Fourier Transform: inverse relationship that exists between the time
and frequency domains,
Find the magnitude and phase of the spectrum of x(t).

6) Plot x(t) in Wolfram.

For A=1, B=2
Piecewise[{{-2/1*t, -1≤t < 0}}] +Piecewise[{{2/1*t, 0≤t ≤ 1}}]

7) Find the Fourier Transform X(w) in Wolfram

fouriertransform( Piecewise[{{-2/1*t, -1≤t < 0}}] +Piecewise[{{2/1*t, 0≤t ≤ 1}}],x,w)
Result: plot((4 (-1+Cos[w] + w*Sin[w]))/w^2)
8) Plot the magnitude of X(w) in wolfram.
plot((4 (-1+Cos[w] + w*Sin[w]))/w^2) from w=-10 to 10

9) Compare the results of the points 4 and 8

 Fourier Transform for periodic signals.
If X(jw) is of the form of a linear combination of impulses equally spaced in
frequency, that is,

Fourier series representation

of a periodic signal

The Fourier transform of a periodic signal with Fourier series coefficients {ak} can be interpreted as a train of
impulses occurring at the harmonically related frequencies and for which the area of the impulse at the kth
harmonic frequency kwo is 2π times the kth Fourier series coefficient ak.
Fourier Transform: Properties.
Fourier Transform: Properties.
Fourier Transform: Properties.

integral (A/B*t)*e^(-i*w*t) from 0 to

B + integral (-A/B*t+2*A)*e^(-i*w*t)
from B to 2*B
 Fourier Transform: Properties.

integral (A/B*t)*e^(-i*w*t) from 0 to B + integral (-

A/B*t+2*A)*e^(-i*w*t) from B to 2*B
Fourier Transform: Properties.

Fourier pair
Fourier Transform: Properties.

Fourier pair
Bode Plots
Bode Plots
• It is often convenient to use a logarithmic scale to plot the magnitude of a
frequency response.
• Frequency responses can have a wide dynamic range covering many orders of
magnitude(ej. frequencies amplified by 1000, frequencies attenuated by 10−4 .
• Using a log scale, we can add rather than multiply the magnitudes of cascaded
Fourier transforms, which is easier to do graphically:

• A Bode plot is the combination of a magnitude plot (magnitude-> Log scale, frequency
-> Log scale) and a phase plot (radians/degrees->linear scale).
Bode Plots
• It is customary to use the decibel (dB) as the logarithmic unit. The bel (B) was
defined (after Alexander Graham Bell) as a power amplification of a system.

• The decibel is one tenth of a bel. Therefore, for a system with a power gain of
10 at frequency w , its power gain in dB is

• To measure the actual magnitude gain (not the power gain) of a system, we use
the identity:
Bode Plots
 Bode Plots
• Consider a first-order system with transfer function:

The break frequency is 2 rad/s. The Bode magnitude plot is the graph of:
 Bode Plots
• Consider a first-order system with transfer function:

• For w=10 rad/s, we get –20 dB, for w=100

rad/s, we get –40 dB, etc. The slope of the
asymptote is therefore –20 decibels per
decade of frequency (a decade is an increase
by a factor of 10.).
• (dotted line: actual magnitude.)
 Bode Plots
• Consider a first-order system with transfer function:
 Bode Plots
• Consider a first-order system with transfer function:
 Bode Plots
• Consider a first-order system with transfer function:

• Magnitude and phase ?

 Bode Plots
• Consider a first-order system with transfer function:

• Magnitude and phase ?

 Plot the Fourier spectrum of
 1) Your voice when you say your name (matlab)
 2) Your voice when you say your name (Audacity)
 3) Compare 1 and 2

 Useful links
 https://matlabgeeks.com/tips-tutorials/how-to-do-a-fourier-transform-in-
matlab/
 https://gist.github.com/pbianche/2584364
 https://personal.utdallas.edu/~dlm/3350%20comm%20sys/FFTandMatLab-
wanjun%20huang.pdf
 https://www.csun.edu/~skatz/ece460/matlab_tut_two.pdf
Bode Plots: Stability
Bode Plot Stability: Advantages

1. It is based on the asymptotic approximation, which provides a simple method to

plot the logarithmic magnitude curve.

2. The multiplication of various magnitude appears in the transfer function can be

treated as an addition, while division can be treated as subtraction as we are
using a logarithmic scale.

3. With the help of this plot only we can directly comment on the stability of the
system without doing any calculations.

4. Bode plots provides relative stability in terms of gain margin and phase margin.

5. It also covers from low frequency to high frequency range.

 Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
Bode Plot Stability: Terminology:

1.Gain Margin: It refers to the amount of gain, which can be increased or decreased
without making the system unstable. It is usually expressed in dB.

2.Phase Margin: It refers to the phase which can be increased or decreased without
making the system unstable. It is usually expressed in phase.

3.Gain Crossover Frequency: It refers to the frequency at which magnitude curve

cuts the zero dB axis.

4.Phase Crossover Frequency: It refers to the frequency at which phase curve cuts
the negative times the 180o axis.

5.Corner Frequency-Break frequency: The frequency at which the two asymptotes

cuts or meet each other.
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

Bode Plot Stability: Terminology:

5. Resonant Frequency: The value of frequency at which the modulus of G (jω)

has a peak value is known as the resonant frequency.

6. Factors: Loop transfer factors like constant term K, Integral factors (jω), first-
order factors ( 1 + jωT)(± n) where n is an integer, second order or quadratic
factors.

7. Slope: There is a slope corresponding to each factor and slope for each factor is
expressed in the dB per decade.

8. Angle: There is an angle corresponding to each factor (degrees).

Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

Bode Stability Criteria:

Stability conditions (applicable for closed loop systems and minimum phase
systems) are given below:

For Stable System: Both the margins should be positive, or phase margin should
be greater than the gain margin.

For Marginal Stable System: Both the margins should be zero or phase margin
should be equal to the gain margin.

For Unstable System: If any of them is negative or phase margin should be less
than the gain margin.
Other values of gain/phase margins, does not permit to conclude about stable or unstable systems
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

Minimum-phase systems : a
linear, time-invariant system is
said to be minimum-phase if
the system and its inverse are
causal and stable.

Criterios de estabilidad de diagramas de bode

https://www.youtube.com/watch?v=T0rqWJfaVh8
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

STABLE
Gm, pm >0
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

The GM from Bode is different from allmargin, why ???

GainMargin = 10^(Gm/20)= 10^(8,93dB/20)=2.79 … from Gm(dB)= 20log(GainMargin)

That means : The system can get a gain of 2.7 before to instability.
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
numG = [1 5];
denG = [1 6 11 6 0];
CLOSED LOOP SYSTEM
hs=tf(numG,denG);
impulse(hs);figure()
;
step(hs)

Bode Plot analysis is useful

for closed loop systems.
numG = [1 5];
denG = [1 6 11 6 0];
hs=tf(numG,denG);
% for negative feedback
Step response show hs1=feedback(hs,1);
instability…. Why? % its equivalence hs1=feedback(hs,1,-
1);
impulse(hs1);figure();
!System are in open step(hs1)
loop!
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
Wolfram-Open LOOP
transfer function ((s+5)/(s^4+6s^3+11*s^2+6*s)) SYSTEM

Step response show

instability…. Why?

!System are in open

loop!
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
Wolfram-Cloosed
transfer function ((s+5)/(s^4+6s^3+11*s^2+6*s))
LOOP SYSTEM?
If you need closed-loop with negative
feedback, just divide your transfer function
by 1+ this transfer function:

http://en.wikipedia.org/wiki/Closed-
loop_transfer_function

transfer function
((s+5)/(s^4+6s^3+11*s^2+6*s)/(1+(s+5
)/(s^4+6s^3+11*s^2+6*s)))

! The transfer function Step response and

has changed due to impulse response
the feedback show stability!
conditions, then, the
bode plot and system
response will change
too!
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

UNSTABLE
 Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
Bode(f)

UNSTABLE
For Closed loop
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
CLOSED LOOP SYSTEM
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)

What if :
- Gain margin is negative?
- Gain margin and phase margin are
infinity?
- Gain margin is infinity and phase
margin is negative
- Etc….

We have more resources (Matlab): pzmap, isstable, Nyquist,

allmargin, ltiview etc.
Bode Plots: Stability Criteria - (Gain Margin and Phase Margin)
 %Alternative representation of the TF
Stability Criteria s=tf(‘s’);
hs=(15*s^2+10*s+30)/(s^2+6*s+11);

root_den=roots(hs.den{1})
Poles(hs)

numG = [15 20 30]; % Numerator

denG = [1 6 11]; % Denominator
hs=tf(numG,denG); % TF

roots(denG) % find poles of G(s)

pzmap(hs) % plot the poles and zeros

figure();
impulse(hs) % simulate impulse response
Stable
figure(); % simulate step response
step(hs)

% Test of stability. True for stable filter

isstable(hs1)
Stability

%Alternative representation of the TF

% s = tf('s’);
% hs = 15*s^2+20*s+30/(s^2+6*s+11)
numG = [15 20 30]; % Numerator
denG = [1 6 11]; % Denominator
hs=tf(numG,denG); % TF

roots(denG) % find poles of G(s)

pzmap(hs) % plot the poles and zeros Stable

hs1=feedback(hs,1);
figure();
impulse(hs1) % simulate impulse response

figure(); % simulate step response

step(hs1)
Stable
% Test of stability. True for stable filter
isstable(hs1)
Stability: Open Loop vs closed loop

%Alternative representation of the TF

% s = tf('s’);
% hs = 15*s^2+20*s+30/(s^2+6*s+11)
numG = [15 20 30]; % Numerator
denG = [1 6 11]; % Denominator
hs=tf(numG,denG); % TF

roots(denG) % find poles of G(s)

pzmap(hs) % plot the poles and zeros

hs1=feedback(hs,1);

figure();
impulse(hs1) % simulate impulse response
figure(); % simulate step response
step(hs1)
% Test of stability. True for stable filter
isstable(hs1)
 Stability OPEN LOOP SYSTEM

numG = [5 0 0]; % Numerator

denG = [1 0 2 3]; % Denominator
hs=tf(numG,denG); % TF

roots(denG) % find poles of G(s)

Bode(hs1)
 Stability CLOOSED LOOP

numG = [5 0 0]; % Numerator

denG = [1 0 2 3]; % Denominator
hs=tf(numG,denG); % TF

hs1=feedback(hs,1)

ltiview({'bode';'pzmap'},hs)

ltiview({'step';'impulse'},hs1)

isstable(hs)

isstable(hs1)
Stability CLOOSED LOOP- Interactive Control Designer
G = tf([5 0 0],[1 0 2 3]);
controlSystemDesigner('bode',G);
https://www.mathworks.com/help/c
ontrol/ug/bode-diagram-design.html

https://www.mathworks.com/videos/
understanding-bode-plots-why-use-
them-1-of-4-76194.html

BODE PLOT FAILS

https://www.powerelectronics.com/t
echnologies/power-electronics-
systems/article/21861083/five-things-
every-engineer-should-know-about-
bode-plots

https://www.powerelectronics.com/t
echnologies/power-electronics-
systems/article/21860614/when-
bode-plots-fail-us
Bibliografía
• Oppenheim, A. Willsky, and H. Nawab, Signals and Systems, 2ª edición, 1997, Prentice Hall,
ISBN # 0-13-814757-4.

• Hwei P. Hsu, SCHAUM'S OUTLINES OF Theory and Problems of Signals and Systems”, 1995,
McGraw Hill

• Magnitude and Phase,

https://www.comm.utoronto.ca/~dkundur/course_info/signals/notes/Kundur_FourierMagPh
ase.pdf, fecha de acceso, 15 de Diciembre del 2019.

• Bode Plot, Gain Margin and Phase Margin (Plus

Diagram),https://www.electrical4u.com/bode-plot-gain-margin-phase-margin/, fecha de
acceso, 15 de Diciembre del 2019.

• Virtual oscilloscope and spectrogram: academo.org/physics/sound/

• Spectrum analysis in Simulink: https://www.youtube.com/watch?v=80h9Y5rf5fc