The Frequency Domain

Time vs Frequency

A square wave is composed of an infinite number of

sinusoids.
magnitude
y(t)
2
4
y(t) = sin t
!
0 (a)

-2
-4 -3 -2 -1 0 1 2 3 4
2
4 1
y(t) = (sin t + sin 3t)
! 3
0
(b)

-2
-4 -3 -2 -1 0 1 2 3 4
2
4 1 1
y(t) = (sin t + sin 3t + sin 5t)
! 3 5
0
(c)

-2
-4 -3 -2 -1 0 1 2 3 4
time

Time signals can be decomposed into a number of sinusoidal

signals of differing frequencies with varying magnitudes and
phases. (Fourier analysis.)

By applying sinusoidal signals of different frequencies to a

system and monitoring the change in amplitude and phase at
the system output, we can determine a lot of information
about the system.
Learning Objectives
• to understand the meaning of gain and phase applied to a
system
• to read a frequency scale
• to recognise that there are different ways of presenting
gain and phase information
• to recognise features on a frequency response plot
• to find special frequency points on a frequency response
plot
• to interpret a frequency response plot
Identification of magnitude and phase values from a
sinusoidal signal

The standard form of a sinusoidal signal is

u(t) = A sin (ωt + ϕ)
ω represents the frequency in rad/s,
A the amplitude (or strength) of the signal and
ϕ the phase shift.

Example
Let the following three signals be the main components of a
time varying signal.

(i) u1(t) = 6 sin (4t ), ω = 4 rad/s, A= 6, ϕ = 0 rads

(ii) u2(t) = 2 sin (6t - π/4), ω = 6 rad/s, A= 2, ϕ = - π/4 rads
(iii) u3(t) = 1.2 sin (10t - π/2), ω = 10 rad/s, A=1.2, ϕ = - π/2
Magnitude and phase at different frequencies →Frequency
Response

Known sinusoidal signals are injected at the input to the

process. By varying the input signal frequency, the
magnitude and phase effect of the output can be measured.
This is termed the frequency response. True only for linear
systems.

u(t) = Asin(ωt+φ1) y(t) = Bsin(ωt+φ2)

Physical
System

Input: u(t) Output: y(t)

Frequency ω rad/s ω rad/s
Magnitude A B
Phase φ1 rads φ2 rads

System Gain

Amplitude of y(t) = system GAIN * amplitude of u(t)

Amplitude of y(t) B
System Gain = =
Amplitude of u(t) A

The physical units of the gain: [Gain] =

The gain can also be expressed in terms of dB (deciBels)

GaindB = 20 log10 Gain ---- Gain = 10Gain(dB)/20

Gain < 1  dB negative dB

Gain = 1  0dB and
Gain >1  positive dB
System phase shift
u(t) = Asin(ωt+φ1) y(t) = Bsin(ωt+φ2)
Physical
System

φ2 = φ1 + System Phase shift

The phase shift is often negative, which indicates a phase lag,

that is the output is lagging behind the input signal. For
positive phase shift, we have phase lead.
Frequency and Logarithmic Frequency Scales
Radio
frequencies
Human Frequency
hearing
Frequency
Control
Engineering
Frequency Infra-red
Radar
To Zero
Frequency Frequency
(Hz)

0 102 104 106 108 10 1012 1014 1016

10 -1 10 103 105 107 109 10 1011 1013 1015

angular frequency  ω rad/s = 2 π f where f is in Hz.

To low
0.0016 0.016 0.16 1.6 16 160 Hz
Frequency

10-2 10 -1 0 10 102 103 rad/s

Logarithmic scale x=log10ω

0.848
0.903
0.954
0.477

0.602

0.778
0.301

0.699

2 3 4 5 6 7 89 20 30 40 50 60 7080 90
1
1=10
o 10=10 100=10
2
Presentation of gain and phase information
The triplets (frequency, gain, phase) indicate how a system
affects the input signal at a particular frequency. This
information can be presented in three different graphical
forms:
(i) a Bode plot is a plot og gain vs frequency and phase vs
frequency
(ii) a Nichol’s plot is a plot gain vs phase
(iii) a Nyquist plot is a plot imaginary part vs real part

Example:
1
G(s) = 2
s + 0.6 s + 1

Frequency rad/s 0.4 1.0 2.0

Polar Gain (dB) 1.20 4.45 -10.2
Phase (deg) -16o -910 -158o
Cartesian x -1.11 -0.03 -0.29
y -0.32 -1.67 -0.12
The Bode plot
The Bode plot is a graphical representation of the frequency
response in the semilog Cartesian co-ordinate system. It
comprises of two plots, that is,

Bode Diagrams

i) a Magnitude plot: The

10

0
ω1 ω2
ω3
gain of the system versus
Phase (deg); Magnitude (dB)

-10

-20

-30
frequency
-40

-50
ω1 a Phase plot:The phase
ω2
-100

-150
ω3 shift induced by the
-200
10
-1
10
0 1
10 system versus frequency
Frequency (rad/sec)
Nichols Plots
Frequency is an implicit parameter on the frequency response
line.
Nichols Charts
20
Horizontal (Phase) axis
is used to represent the
0
!2 !1

-20 !3 phase and usually has units

Open-Loop Gain (dB)

-40
of degrees.
-60

-80

-100
Vertical (Magnitude) axis
represents the gain in dB
-120
-180 -160 -140 -120 -100 -80 -60 -40 -20 0

Open-Loop Phase (deg)

Nyquist plots
This plot has a Cartesian grid; there is no frequency axis, nor
any gain or phase axis. The gain and phase are plotted as
polar co-ordinates. Therefore lines of constant gain are circles
from the origin. Each point on the response curve
corresponds to a specific frequency point.
Nyquist Diagrams
2

unit circle
1.5

1
Imaginary Axis

0.5

ω3 ω1
-0.5

-1

ω2
-1.5

-2
-1 -0.5 0 0.5 1 1.5

Real Axis
The Frequency Response And System Features

A sin ωxt

??
Process: G(s)

Consider the system:

400
G(s) =
(s + 10)(s2 + 0.4s + 4)
Amplification and Attenuation
(i) An input signal at ω, is amplified if the gain is greater
than 1, (or 0 dB).
(ii) An input signal at ω, is attenuated if the gain is less than
1, (or 0 dB).
(iii) If the gain is 1, (or 0 dB), the input and output signal are
the same.
 Frequency
Frequency range of
range of attenuation Nyquist Diagrams

amplification 60
Bode Diagrams

40 40 Circle radius 1
20
inside circle shows
attenuation
0
20

Imaginary Axis
Phase (deg); Magnitude (dB)

-20

-40

-60 0
-80

0
-20
-50

-100

-150
-40
-200

-250

-300 -60
10-1 100 101 102 -30 -20 -10 0 10 20 30

Frequency (rad/sec)
Real Axis

Nichols Charts
50

0
Open-Loop Gain (dB)

-50
Range of
gain > 1
-100

Range of
-150 gain < 1

-200
-300 -250 -200 -150 -100 -50 0

Open-Loop Phase (deg)

d.c. gain
The system gain at zero frequency.
. Nyquist Diagrams
Bode Diagrams
60

40

20
40
0

d.c.gain
Phase (deg); Magnitude (dB)

-20
20 d.c.gain

Imaginary Axis
-40

-60

-80 0
0

-50
-20
-100

-150

-200 -40

-250

-300 -60
10-1 100 101 102 -30 -20 -10 0 10 20 30

Frequency (rad/sec) Real Axis

Nichols Charts
50

0
Open-Loop Gain (dB)

-50
d.c.gain

-100

-150

-200
-300 -250 -200 -150 -100 -50 0

Open-Loop Phase (deg)

Roll-off rate

The gradient of the gain plot at high frequencies is

calculated by determining the change in magnitude over one
decade of frequency. The roll-off rate is then stated, for
example, as: -20 dB/decade, or -40 dB/decade.
In the system response shown, the system is 3rd order and this
is associated with a roll-off rate of –60 dB/decade.
Bode Diagrams

40

20

0
Phase (deg); Magnitude (dB)

-20

-40 slope of line gives

-60 roll-off rate
-80

-50

-100

-150

-200

-250

-300
-1 0 1 2
10 10 10 10

Frequency (rad/sec)
Phase graphs
The shape of a phase curve is important when we
discuss the phase lag introduced by a system.
no phase lag, , the output signal follows the input signal.
As the frequency of the input signal increases the output
signal starts to lag behind the input signal.
The combination of amplification and phase lags of –180o can
cause closed loop system instability.
Bode Diagrams

40

20

0
Phase (deg); Magnitude (dB)

-20

-40

-60

-80

-50 more gradual

-100

-150
rapid phase phase change
-200 change
-250

-300
-1 0 1 2
10 10 10 10

Frequency (rad/sec)
Problems
(a) What is the d.c.gain?
(b) State the frequency range where amplification > 10 dB.
(c) At what frequency do we find 10 dB of attenuation?
(d) What is the roll-off rate?
(e) What is the gain when the phase reaches -1350?
Bode Diagrams

20

10

0
(a) d.c.gain = 20 dB
(b) Ampl >10 dB:
Phase (deg); Magnitude (dB)

-10

-20

-30
below 5 rad/s
-40 (c)Atten < 10dB over 20
0
rad/s.
-50
(d) Roll-off rate is –40
dB/decade
-100

(e)Gain at phase = -135o

-150

-200
10
-1 0
10
1
10
2
10
is 3 dB
Frequency (rad/sec)
Special Frequency Points

Gain Crossover frequency= ωgco = frequency at which the

gain crosses 0 dB
Phase Crossover frequency = ωpco = frequency at which the
phase crosses ‘-180o’
.
Bode Diagrams

20

0
Phase (deg); Magnitude (dB)

-20

-40
gain crossover point

-60 ωgco
-80

-60

-120
ωpco
-180

-240 phase crossover point

-300
-1 0 1 2
10 10 10 10

Frequency (rad/sec)
Nichols Charts Nyquist Diagrams
40 1.5
gain crossover point
20

0 1

-20 phase crossover point

Open-Loop Gain (dB)

unit circle
-40 0.5
Imaginary Axis

phase crossover point

-60

-80 0
o
-100 -180
-120 -0.5
gain =1
-140
-1
-160 gain crossover point
-180
-1.5
-200 -1 -0.5 0 0.5 1 1.5 2
-300 -270 -240 -210 -180 -150 -120 -90 -60 -30 0
Real Axis
Open-Loop Phase (deg)
Bandwidth
Bandwidth is defined as the range of frequencies over which
the gain is greater than G/ 2 = 0.707 G ( -3dB).
Open loop system bandwidth
.
Gain (dB) Gain (dB)

GdB
GdB -3 dB w (rad/s) w (rad/s)

wbw

(a) (b)

Closed loop system bandwidth

Closed loop system bandwidth = ωbw = the frequency range

over which the gain is 3 dB down on the zero frequency gain
.
Interpretation of frequency response plot
Position
Mass
Force

Phase
Gain
degrees
dB
6dB 0
freq(rad/s) freq(rad/s)
0 dB
ωgco -180

Low frequency range: gain is 0 dB with zero phase shift.

Middle Frequency range: input is amplified, with a maximum
gain of 2. The phase plot shows a rapid decrease in
phase for this system. Hence a large phase lag at high
frequencies.
High Frequency range: a clear increasing attenuation. At
very high frequency, the amplitude of the output signals
will be very small. The phase response for this example
shows the phase lag levelling off to -180o .
Problem: Rotating disk system
Gain Phase
dB degrees
freq(rad/s)
freq(rad/s) 0
0 dB -90
ωgco
-180

(a) (b)

Low frequency range: infinite d.c. gain or ‘high gain at low

frequencies' .
Middle Frequency range: The phase lag increases in this
region. This corresponds to the disk system not
responding fast enough to the changes in the input
signal.
High Frequency range: In this range we see the input signal
is severely attenuated; the disk cannot physically
follow a high frequency sinusoidal input. The output
signal is 180o out of phase with the input.
Performance Specification: Gain and Phase Margins
Stability determination
the open loop transfer function GOL(s) = G(s)K(s)H(s)

R(s) + Y(s)
K(s) G(s)
-
H(s)

Procedure for checking if the closed loop system is stable:

Test 1:
From the Bode plot of the open loop transfer function:
1. Find the gain crossover, ωgco.
2. Determine the phase at that point, φ(ωgco).
3. Is the phase below –180o? The closed loop system is
unstable.
Test 2:
From the Bode plot of the open loop transfer function:
1. Find the phase crossover, ωpco.
2. Determine the gain at that point, |G(ωpco)|.
3. Is the gain greater than one? If so, the closed loop system
is unstable.

REMARK: need to perform both tests to determine if the

system is stable.
Example
Determine from the following Bode plot) of an open loop
system if the closed loop system would be stable.

Bode Diagrams

50
gain crossover point
ωgco =3 rad/s
0

φ(ωgco)=-210
Phase (deg); Magnitude (dB)

-50

-100

0
-30
-60
-90
-120
phase crossover point ωpco =2 rad/s
-150
-180
-210
|G(ωpco)|=15 dB
-240
-270
-1 0 1 2
10 10 10 10

Frequency (rad/sec)
ωpco ωgco
Frequency domain specifications

Phase Margin (PM) Procedure:

1. Find the gain crossover, ωgco.
2. Determine the phase at that point, φ(ωgco).
3. Calculate the phase margin:
PM = φ(ωgco) – (-180O).
If the PM is negative the system is unstable.

Gain Margin (GM) Procedure.

1. Find the phase crossover, ωpco.
2. Determine the gain at that point, G(ωpco).
3. Calculate the gain margin:
GM = 0dB - G(ωpco).
If the GM is negative the system is unstable.

REMARK: Only GM>0 + PM>0 indicate a stable system.

Example
Bode Diagrams

20
gain crossover point Gain margin procedure:
0 gain margin at !pco 1. ωpco is at 1.8 rad/s
-20
4. G(ωpco) is about –12dB
-40 5. GM = 0 dB – (-12dB) = 12dB.
-60

0
-30 phase margin at !gco
Phase margin procedure:
-60
-90
phase crossover point
1. ωgco is at 0.8 rad/s
2. φ(ωgco) is about –116o.
-120
-150
-180
-210
-240
-270
3. PM = -116 – (-180o) =
64o.
-1 0 1
10 10 10
!gco
Frequency (rad/sec) !pco
Example
Form the Nichols plot

Nichols Charts
20

gain phase margin

10
margin
Open-Loop Gain (dB)

-10

gain crossover point

-20

-30
phase crossover point

-40
-240 -210 -180 -150 -120 -90 -60 -30 0

Open-Loop Phase (deg)

From Nyquist plot

Nyquist Diagrams
Nyquist Diagrams 1
1 phase crossover point
gain crossover point 0.5
0.5

0
0
Imaginary Axis
Imaginary Axis

-0.5
-0.5

-1
-1 1/gain margin
Phase margin (angle) -1.5
-1.5

-2
unit circle -2

-2.5 -2.5
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 -2.5 -2 -1.5 -1 -0.5 0 0.5 1

Real Axis Real Axis

Link to the time domain
A general rule-of thumb which links the PM to a time domain
specification for systems with dominant second order
characteristicis, is

the damping ratio , ζ , = PΜdeg/100

where PMdeg represents the phase margin in degrees.

What we have learnt
 Time-varying signals were related to a frequency domain
decomposition using the amplitude, phase and frequency
of component sinusoids.
 Physical systems can alter the magnitude and phase of the
sinusoidal components of an input signal.
 Frequency response plots can be generated from gain,
phase and frequency information.
 Frequency response plots can be interpreted in terms of the
low, middle and high frequency ranges.
 Some important features of a frequency response plot are :
 ranges of amplification and attenuation, d.c.gain, high
frequency roll-off
 The main frequency points are gain crossover, phase
crossover, and bandwidth.
 The frequency response plots can be related to the
behaviour of the physical system.
 how the magnitude and phase of the open loop process can
determine the stability of the closed loop system
 Test procedures which determine, from the values of gain
and phase at the phase and gain crossovers, whether the
closed loop system is stable
 Procedures for determining the gain margin and phase
margin of a system. These provide an indication of how
stable the closed loop system is and are used as frequency
domain specifications on the open loop system.