Chapter 10:

Frequency Response Techniques

Course Instructor
Prof. Adel Abdennour
Electrical Engineering Department
Islamic University of Madinah

EE 3631 Control Theory Spring Term: 2021-2022 1

 We will cover in this chapter:
• Introduction
• Asymptotic Approximations: Bode Plots
• The Nyquist Diagram
• Stability, Gain Margin and Phase Margin via the Nyquist Diagram
• Stability, Gain Margin, and Phase Margin via Bode Plots

EE 3631 Control Theory 2

 Introduction:
Recall that for a linear system, the steady-state response will exhibit the
same frequency ω as the input, but will vary in magnitude and phase.

The output response is the product of the input phasor by the system
phasor:
Mo (ω)∠φo (ω) = Mi (ω)M(ω)∠ [φi (ω) + φ(ω)] The
magnitude frequency response is M(ω) and the phase frequency
response is is φ(ω).

EE 3631 Control Theory 3

 Example:
What is the frequency response of the following transfer function:

Substitute jω for s, we can easily obtain:

We can plot G (jω) in two ways:

• As separate magnitude and phase plots

• As a polar plot

EE 3631 Control Theory 4

 First way: Magnitude and phase plots (Bode plots):
To cover a very wide range of ω, we plot the magnitude in decibels and
the frequency on a logarithmic scale: dB(x ) = 20 log10(x ); the phase
plot we will display the phase directly on the vertical axis.

Magnitude and phase

plots for
G (s) = 1/(s + 2),

EE 3631 Control Theory 5

 Second way: Polar plot :
On a polar plot, both the magnitude and phase are combined. Beginning
with ω = ꝏ we plot the head of the vector with magnitude and phase
|G (jω)| ∠G (jω)=
ω=ꝏ ω=0

EE 3631 Control Theory 6

 Bode Plots:
The Bode plot of the function G(j) is composed of two plots:
• One with the magnitude of G(j) plotted in decibels (dB) versus
log10()
• The other with the phase of G(j) plotted in degrees versus
log10()
Feature of the Bode Plots
• Since the magnitude of G(j) is expressed in dB, products and
division factors in G(j) become additions and subtraction.
• The phase relations are also added and subtracted from
each other algebraically.
• The magnitude plot of Bode of G(j) can be approximated by
straight-line segments which allow the simple sketching of the
bode plot without detailed computation.
EE 3631 Control Theory 7
 The Bode Form:
In order to simplify the Bode plot, it is convenient to use the so-called
Bode form given by: A.( s  z1 )(s  z2 )...( s  zm )
L( s ) 
s l ( s  p1 )(s  p2 )...( s  pzn )
The bode form can be expressed as:
L ( j )  L ( s ) s  j

 j   j   j 
K .1  .1  ...1  
L( )   z1   z2   zm 
l j   j   j 
 j  1  .1  ...1  
 p1   p2   pn 

𝑧1 𝑧2 … 𝑧𝑚
where: 𝐾 =
𝑝1 𝑝2 … 𝑝𝑛

EE 3631 Control Theory 8

 Magnitude and Phase of L(j):
The magnitude of L() in dB is obtained by:
j j
L( j )  20 log L( j )  20 log K  20 log1   ...  20 log1 
z1 zm
j j
 20.l. log j  20 log 1   ...  20 log 1 
p1 pn

The phase of L() in degree is obtained by:

j j j j
L( j )  K  1   ...  1   lj  1   ...1 
z1 zm p1 pn
1  1  1  1 
L( j )  0  l.90  tan ( )  ...  tan ( )  tan ( )...  tan ( )
z1 zm p1 pn

EE 3631 Control Theory 9

 In general, L() can contain four simple type of factors which are:

•Constant gain L( j )  20 log L( j )  20 log K

j
•Zeros at origin L( j )  20 log L( j )  20 log1 
z
•Poles at origin L( j )  20 log L( j )  20 l log j

•First order lag j

1
L( j )  20 log L( j )  20 log z
j
•First order lead 1
p
•Second order term s2 + 2ξωns + ωn2

EE 3631 Control Theory 10

The magnitude and phase plots for the following simple transfer functions
have well-known approximations:

The transfer functions for most physical systems are composed of some
combination of these. Therefore, knowing these approximations provides
two benefits:

• The ability to plot the magnitude and phase response without a

computer
• The ability to recognize these responses from a measurement of a
system’s frequency response

EE 3631 Control Theory 11

Illustration : L() = K (constant gain)
Frequency response:

L( ) dB ,  ( ) 
20logK

0dB  (log )
0.1 1 10 100
0o
 (log )
0.1 1 10 100
-180

EE 3631 Control Theory 12

Integrating element: L() = 1/(j)l
Frequency response:

L()dB ,  () 
40 dB

20 dB

0dB

-20 dB l=1
l=2
-40 dB  (log  )
0.1 1 10 100
0o  (log  )
0.1 1 10 100
-90 l=1
-180 l=2

EE 3631 Control Theory 13

 Differentiating element: L() = (j)l
Frequency response:

l=2
L()dB ,  ()  l=3
20 dB
l=1
0 dB

-20dB

-40 dB

-60 dB  (log  )
0.1 1 10 100
270 l=3
180 l=2
90
l=1
 (log  )
0o 0.1 1 10 100

EE 3631 Control Theory 14

Simple zero:

EE 3631 Control Theory 15

 The following plot shows the true and approximate Bode plots for
G(s) = s + 3. In magnitude for high and low frequency
approximations meet at the break frequency ω = 3.

For the phase, the sloped

component
begins at one tenth the break
frequency, 0.3 and ends at 10
times the break frequency, 30.

The slope in magnitude is 20 dB / decade.

EE 3631 Control Theory 16

Simple Pole: L(s) = 1/(1 + s/p)

Frequency response:

Magnitude response Phase response

• Low frequency • Low frequency
asymptote = 0dB asymptote = 0
• Breakpoint at = p • -45 at breakpoint (=p)
• High frequency • High frequency
asymptote= -20 dB/decade asymptote= -90

EE 3631 Control Theory 17

 L()dB ,  () 
0 dB

-10 dB -20 db/decade

-20dB

-30 dB  /  p (log /  p )
0.1 1 10  /  p (log /  p )
0

-45

-90

Example: L(s) = 1/(1 + s/2)

EE 3631 Control Theory 18

 Example: 𝐾(𝑠+3)
Draw the Bode plots for 𝐺 𝑠 =
𝑠(𝑠+1)(𝑠+2)
Solution:
G(s) in bode (normalized) form is:

The effect of K is to move the magnitude curve up (increasing K) or

down (decreasing K) by the amount of 20 log K. K has no effect upon
the phase curve. Let K = 1:
𝑗𝜔
𝐺 𝑗𝜔 = 20 log 1.5 − 20 log 𝜔 − 20log( 1 + 𝑗𝜔 − 20log( 1 + + 20log( 1 + 𝑗𝜔/3
2

If we start at ω=0.1, the starting gain will be 20log(1.5)-20log(.1) at

-20dB/decade slope, the rest of the components are not active yet.
The initial gain is therefore: 20 log 1.5 +20=23.5dB

EE 3631 Control Theory 19

 𝑗𝜔
𝐺 𝑗𝜔 = 20 log 1.5 − 20 log 𝜔 − 20log( 1 + 𝑗𝜔 − 20log( 1 + + 20log( 1 + 𝑗𝜔/3
2

2 3

EE 3631 Control Theory 20

 𝑗𝜔
𝐺 𝑗𝜔 = 20 log 1.5 − 20 log 𝜔 − 20log( 1 + 𝑗𝜔 − 20log( 1 + + 20log( 1 + 𝑗𝜔/3
2
Φ 𝑗𝜔 = 0 − 90 − atan(𝜔) − atan(𝜔/2) + atan(𝜔/3)
Φ 0 =-90 Φ ꝏ = −90 − 90 − 90 + 90 = −180

EE 3631 Control Theory 21

 Example: Identification of the transfer function
L()dB Given an approximate Bode magnitude,
20 dB Find the transfer function
15 dB
10 dB -20 db/decade
5 dB
2 300  (log  )
0 dB
0.1 1 10 100 1000 10000
-5 dB
-10 dB
-15 dB

EE 3631 Control Theory 22

 Solution
a. Finding the Bode DC gain K’ (in decimal):
L()dB

20 dB

b. Find poles and zeros: 15 dB

10 dB -20 db/decade
Zero : s = -2
5 dB
Poles : s = - 300, 0 2 300  (log  )
0 dB
0.1 1 10 100 1000
c. Transfer function: -5 dB
10000

-10 dB

-15 dB

EE 3631 Control Theory 23

 Bode Plots for G(s) = s2 + 2ξωns + ωn2
Unlike the first-order frequency response approximation, the
difference between the approximation and the actual response can be
great for some values of ξ. A correction to the Bode diagrams should
be made to improve the accuracy.

• At low frequencies:
The magnitude, M, in dB at low frequencies therefore is:

• At high frequencies:

EE 3631 Control Theory 24

EE 3631 Control Theory 25
 a magnitude correction of +20 log 2ξ can be made at the natural,
or break, frequency on the Bode asymptotic plot.
Bode Plots for G(s) =1/(s2 + 2ξωns + ωn2)
• The procedure is exactly the same as before and the plots are
exactly the inverse.
• We find that the magnitude curve breaks at the natural frequency
and decreases at a rate of -40 dB/decade.
• The phase plot is 0° at low frequencies. It begins a decrease of
-90°/decade and continues until ω =10ωn, where it levels off at
-180°.
• The normalized magnitude at the scaled natural frequency is
-20 log 2ξ, which can be used as a correction at the break
frequency on the Bode asymptotic plot.

EE 3631 Control Theory 26

 Example:
Draw the Bode plots for the system given by:
(𝑠 + 3)
𝐺(𝑠) =
(𝑠 + 2)(𝑠 2 + 2𝑠 + 25)
The second-order term is normalized by factoring out ωn2:

At low-frequency: The magnitude plot starts at 3/50, or -24.44 dB.

The correction to magnitude curve due to the underdamped term
can be made by plotting a point -20 log 2ξ above the asymptotes at
the natural frequency. Since ξ = 0.2, the correction is 7.96 dB

EE 3631 Control Theory 27

 7.96dB

EE 3631 Control Theory 28

Gain Margin and Phase Margin from Bode plot:
• The Phase-Crossover frequency (p) is the frequency
where: L(jp)=180 
• The Gain-Crossover frequency (g) is the frequency
where: L(jg)=1

• Gain margin (GM in dB)= 20 log10 (1/ L(jp)

= -20 log10  L(jp)
• Phase Margin (PM in )= (L(jg))-180 

EE 3631 Control Theory 29

EE 3631 Control Theory 30
 Gain and Phase Margins

EE 3631 Control Theory 31

 Bode Plots using Matlab: Examples
Example mtlb 1: L(s) = 10/s(s+1)(s+5)
Bode Form: Consider L(jω) =2/jω(1+jω)(1+jω/5)

GM=9.54 dB at 2.23 rad/sec

PM=25.39  at 1.227 rad/sec

p = 2.23 rad /sec

g = 1.227 rad /sec

GM >0 and PM >0

System is stable

EE 3631 Control Theory 32

Example mtlb 2: L(s) = 10(s+10)/s(s+2)(s+5)
Bode Form: Consider L(jω) =10/jω(1+jω/2)(1+jω/5)

GM=7.36 dB at 5.77 rad/sec

PM=10.64 at 3.68 rad/sec

p = 5.77 rad /sec

g = 3.68 rad /sec

GM >0 and PM >0

 System is stable

EE 3631 Control Theory 33

 Example mtlb 3:Matlab code 5 s + 0.5
---------------------
s^3 + 1.5 s^2 + 0.5 s
Bode Diagram
>> w=logspace(-2,3) Gm = Inf dB (at Inf rad/sec) , Pm = 36.2 deg (at 2.1 rad/sec)
>> num=[5 0.5] 100
>> den=[1 1.5 0.5 0]
>> sys=tf(num,den) 50

Magnitude (dB)
>> bode (sys,w)
>> margin(sys) 0
>> grid on
-50

-100
-45
GM=
PM=36.2  at 2.1 rad/sec
Phase (deg)

-90

GM = and PM >0 -135

 System is stable
-180
-3 -2 -1 0 1 2
10 10 10 10 10 10
Frequency (rad/sec)

EE 3631 Control Theory 34

EE 3631 Control Theory 35
Nyquist Plots:
R(s) + Ys)
• Critical point: G

The characteristic equation is: -

H
1  G ( s ) H ( s )  1  L( s )  0
 L( s )   1
Therefore, instead of examining [1+L(s)] with the critical point
being zero, we could simply examine L(s) with the critical point
being -1.
• Minimum phase systems:
• A transfer function having neither poles nor zeros in the
right-half s-plane is a minimum-phase transfer function
• Systems having poles and/or zeros in the right-half s-plane
are non-minimum-phase systems
36
EE 3631 Control Theory 36
 • Encircled: a point is said to be encircled by a path if it is
Inside the path
• Enclosed: a point is said to be enclosed by a path if it is
to the left of the path regardless of whether it is
encircled or not.

Im Im Im Im

Re Re Re Re

Encircled Encircled Not Encircled Not Encircled

Enclosed Not Enclosed Not Enclosed Enclosed

EE 3631 Control Theory 37

 Nyquist Stability Criterion
Let:
• (s) = 1+G(s)H(s)= 1+ L(s)=0
• Z= Number of zeros of 1+G(s)H(s) in the right-half s-plane
(closed-loop poles)
• P= Number of poles of 1+G(s)H(s) in the right-half s-plane
(open-loop poles)
• N= Number of clockwise encirclements of the critical point
-1+j0 by the mapping L(s)

Principle of the Argument: N= Z - P

The closed-loop system is stable if and only if Z = 0.

EE 3631 Control Theory 38

 Given a control system, with L(s) being the loop transfer function,
the application of the Nyquist criterion to the stability problem
involves the following step:
1. The L(s) plot corresponding to the Nyquist path is constructed
in the L(s)-plane for all values of ω from -∞ to +∞.

2. The value of N, the number of encirclements of the (–1,j0)

point made by the L(s) plot is observed.

3. The Nyquist criterion follows from (N=Z–P)

→ for closed-loop stability Z must be zero.
→ for open-loop stability P must be zero.

EE 3631 Control Theory 39

• For a closed-loop system to be stable (Z=0), the L(s) plot must
encircle (–1,j0) point as many times as the number of poles of L(s)
that are in the right half s–plane, and the encirclement, if any, must
be made in the clockwise direction. i.e., N = – P

• A minimum phase transfer function does not have poles or zeros

in the right half s–plane or on the jω–axis, excluding the origin
(P=0). In this case, the Nyquist criterion reduces to N = 0

∴ For a closed loop system with loop transfer function L(s) that is of
minimum phase type, the system is closed-loop stable if the L(s) plot
that corresponds to the Nyquist path does not enclose the (–1,j0)
point. If the (–1,j0) is enclosed by the Nyquist plot, the system is
unstable.
EE 3631 Control Theory 40
 In Brief:
• For Minimum phase systems, the closed-loop is
stable if and only if N = 0.
• Therefore, We only need to see whether the (–
1,j0) is enclosed or not.
• We don't care how many times it is Enclosed.
• Therefore we need to only plot L(s) for the
positive values of ω; starting from ω = ∞ and
ending at ω = 0

EE 3631 Control Theory 41

 Example 1:
Example:
K
L(s) = G(s). H(s) =
s(s + 2). (s + 10)

1) Substitute s = jω :
K
L(s) =
jω ( jω + 2)( jω + 10)

2) Substitute ω = 0 To get the zero frequency

L( j0 ) = ∞ ∠ – 90
3) Substitute ω = ∞
L(j ∞) = 0 ∠ 90

EE 3631 Control Theory 42

4) Find intersection with real axis
K K
L(j ω) = 2
=
j ω (– ω + 12 j ω + 20 ) – 12 ω2 + j ( 20 ω – ω 3 )
K[−12ω2 − j(20ω − ω3 ]
=
−12ω2 + j(20ω − ω3 ) [−12ω2 − j(20ω − ω3 )]
K[−12ω2 − jω(20ω − ω2 )]
=
144ω4 + (20ω − ω3 )2

EE 3631 Control Theory 43

 Intersect with real axis
→ Im ( L(jω)) = 0
−Kω (20 − ω2 ) −Kω (20 − ω2 )
4 3 2
=0 → 2 2 2 2
=0
144ω + (20ω − ω ) ω 144ω + (20 − ω )

−K(20 − ω2 )
→ =0
ω[144ω2 + (20 − ω2 )2 ]

→ ω = ∞ or ω =± 20 rad/ sec Since we need ω positive ,

ω = 20
Note: This is not the intersection. This is the value of ω
at which the intersection occurs!!!! We need L(jω) and
not ω!

EE 3631 Control Theory 44

 Substitute the value of ω found into the real part of L(jω) to
find the intersection:
−12 𝐾 𝜔2 −12 𝐾𝜔2
𝑅𝑒(𝐿 𝑗 𝜔) = = 2
144𝜔 + (20 − 𝜔 ) 𝜔 144𝜔 2 + (20 − 𝜔 2 )2
4 3 2

−12𝐾
=
144𝜔 2 + (20 − 𝜔 2 )2

−12𝐾 −12𝐾
𝐿 𝑗 20 = 2 2 2
= → 𝐿( 𝑗 20 ) = – 0.004167 𝐾
144𝜔 + (20 − 𝜔 ) 2880

EE 3631 Control Theory 45

 Im(L)

-0.004167K
= Re(L)
-1

=0

The system is a stable if -1 is not enclosed (It is not to that

left of the path as it goes from ω = ∞ to ω = 0 )
∴ K < 240
K=240 → System Marginally stable.
K> 240 → –1 is enclosed and the system is unstable

EE 3631 Control Theory 46

 Example:
Example2: The characteristic equation given by
Ks3 + ( 2K + 1 ) s2 + ( 2K + s )s + 1 = 0

Divide by all the terms that don't have K in them :

s3 + 2s2 + 2s s(s2 + 2s + 2)
1+K 2 = 0 → L(s) = 1 + K 2
s + 5s + 1 s + 5s + 1
Note: improper system

EE 3631 Control Theory 47

 K(jω)(−ω2 + 2jω + 2) ω[−2ω + j(2 − ω2 )
1) L ( j ω ) = 2
=K
– ω + 5jω + 1 (1 − ω2 ) + 5jω
2) L ( j0 ) = 0 ∠ 90 L(∞) = ∞ ∠ 90°

K(jω) −2ω + j(2 − ω2 ) (1 − ω2 ) − jω

3) L (jω ) =
(1−ω2 ) + 5jω [(1−ω2 ) − 5jω
ω2 5(2−ω2 ) − 2(1−ω2 ) + jω[10ω2 + (2−ω2 )(1−ω2 )]
=K
(1−ω2 )2 + 25ω2
Im L ( j ω ) = 0 → 10ω2 + (2−ω2 )(1−ω2 ) = 0
10ω2 + 2−ω2 − 2ω2 + ω4 = 0
ω4 + 7ω2 + 2 = 0
− All roots are imaginary !!!
The line doesn't intersect The real axis except at ω = 0

EE 3631 Control Theory 48

 Im(L)
ω=ꝏ
Enclosed
Area

ω=0
-1 Re(L)

The point –1 is not enclosed for any value of K

→ closed loop system is always stable.

EE 3631 Control Theory 49

General Nyquist criterion (minimum and non-minimum phase):

• The general system criterion is given by: N = Z – P

• For the closed loop system to be stable, there should be no
zeros of ∆(s) = 1 + L(s) in the right half plane.
Therefore, Z = 0 and the criterion becomes N = – P
• The above condition requires that the complete Nyquist plot
should be drawn (for all values of ω from -∞ to +∞) in order to
know how many times the –1 point is enclosed.
• In other words, seeing if -1 is enclosed is not sufficient.

EE 3631 Control Theory 50

• Here, we will present a new and simple method that we can apply to
minimum and non-minimum phase systems.
• Also, with this method, we don’t need to plot the entire Nyquist
plot. Only the positive jω-axis part of the Nyquist path is needed.
Let: Z= # of zeros of 1+L(s) that in the RHP.
P= # of poles of 1+L(s) thatThe
in the RHP. Nyquist criterion states:
simplified
P_ω= # of poles of L(s) on the jω-axis.
ϕ_11= angle traversed by the Nyquist plot ϕ11with
= [ Zrespect
– P – 0.5toPω-1] 180°

The simplified Nyquist criterion states: For closed-loop stability, Z=0

ϕ11 = [ Z – P – 0.5 Pω ] 180° ϕ11 = −( 0.5 Pω + P) 180°

EE 3631 Control Theory

For closed-loop stability, Z=0 51
Important Observations:
• Since P_ω and P are positive, if ϕ_11 is positive, then (-1,j0) is
enclosed and the closed loop is unstable.
• For non-minimum phase systems, even when (-1,j0) is not
enclosed, the angle criteria must be satisfied for the closed-loop
systems to be stable.
• With this method we need to plot the Nyquist plot only for
positive ω.
• ϕ_11 is the angle variation. Therefore ϕ_11= -270 ≠ ϕ_11= 90
• For systems with minimum phase loop transfer functions (P=0),
the equation simplifies to: ϕ_11= -P_ω× 90°

EE 3631 Control Theory 52

xample 1:
Example:
𝑠 2 −𝑠+1
Let 𝐿(𝑠) = poles@ 0,1,5
𝑠(𝑠 2 −6𝑠+5)

𝑃𝜔 = 1 𝑎𝑛𝑑 𝑃 = 2 , 𝜙11 = −(0.5𝑃𝜔 + 𝑃)180° = −450°

For stability we should expect: Φ11=-450o Im(L)

-1 ω=ꝏ Re(L)

Note that for this non-minimum Enclosed

phase system, even though -1 is Area
not enclosed, the system is ϕ_11 =-90°
unstable because:
ϕ_11 =-90° And not -450. ω=0

EE 3631 Control Theory 53

 If we substitute 𝜙11 = -90 into [Equ1]
we get:

𝜙11 = [ 𝑍 – 𝑃 − 0.5 𝑃𝜔 ]180°

90
−90 = 𝑍 − 2 − 0.5 (1) 180 → − = 𝑍 − 2.5
180
→ 𝑍=2
which indicates that the closed–loop system have two poles with
RHP.

EE 3631 Control Theory 54

 Example 2:
Example:
𝐾 (𝑠−1)
𝐿(𝑠) = non-minimum phase (a zero @ s=1 ).
𝑠 (𝑠+1)

For this L(s), we have Pw=1 and p=0

The requirement for closed loop stability is:
𝜙11 = −( 0.5𝑃𝜔 + 𝑃 ) 180°
= −( 0.5(1) + 0 )180 = − 90°

Thus, the phasor drawn from (−1, 𝑗0) to the Nyquist plot of 𝐿(𝑗𝜔)
should traverse -90 as ω varies from ∞ to 0.

EE 3631 Control Theory 55

 K(jω − 1) K(jω − 1) K(jω − 1) K(jω − 1)(−ω2 − jω)
𝐿(𝑗ω) = = 2 𝐿(𝑗ω) = =
jω(jω + 1) −ω + jω −ω2 + jω (−ω2 + jω)(−ω2 − jω)

@ ω= ∞ : 𝐿(𝑗∞) = 0 ∠ − 90 K −jω3 + ω2 + ω2 + jω K 2ω2 + j(ω − ω3 )

= 4 2
=
@ ω= 0 : 𝐿(𝑗0) = ∞ ∠ 90 ω +ω ω4 +ω2
K[2ω + j(1 − ω2 )] 2K 1−ω
= = 2 + jK
ω(ω2 + 1) (ω + 1) ω(ω2 + 1)

To find the intersect of L(jω) with the real axis :

1 − ω2
𝐼𝑚(𝐿) = 0 → 2
= 0 → ω = ∞ or ω = ±1
ω(ω + 1)

2K
For ω = 1 L(j1) = => L(j1) = K
(1)2 +1

EE 3631 Control Theory 56

 Case1: 𝐾 > 0
We first note that −1 is enclosed which indicates that the
closed-loop is unstable.

Also, since ∅ is positive : We can alternatively conclude that the

system is unstable.
ω=0
90 = (Z − 0.5Pω − P)180°
Enclosed
= (Z − 0.5)180° Area
→ Z=1 ϕ_11 =+90°
(one unstable pole).
ω=ꝏ K

-1

EE 3631 Control Theory 57

Case2: −𝟏 < 𝑲 < 0
For negative K, we rotate the plot 180° around the origin. The
net rotating of ∅ is −90 which agrees with the stability
requirement. Hence, the closed-loop system is stable.
Remember that the system is non-minimum phase and first
checking the enclosed of −1 is NOT enough.

-1 K ω=ꝏ

Enclosed
ϕ_11 =-90° Area

ω=0
EE 3631 Control Theory 58
 Case3: 𝐊 < −𝟏
In this case, we see that the point –1 is enclosed (alternatively,
𝜙11 = +270 which differs from the required −90 ).
Therefore, the closed loop system is unstable.
𝜙11 = (𝑍 − 0.5𝑃𝜔 − 𝑃)180°
270 3
270 = (𝑍 − 0.5)180 → = Z − 0.5 → + 0.5 = Z → Z = 2
180 2
There are two closed loop poles in RHP! K ω=ꝏ
-1

Enclosed
ϕ_11 =+270° Area

ω=0

EE 3631 Control Theory 59

 Relative Stability: (Gain Margin and Phase Margin)

• Absolute stability: to find out whether the system is stable or not

• Relative stability: to find out how stable a system is.

• In the time domain, the relative stability is measured by

parameters such as Maximum overshoot and damping ratio.

• In frequency domain, the relative stability is measured by how

close the Nyquist plot of L(jω) is to the (-1,0) point.

EE 3631 Control Theory 60

Gain Margin:

Def: The phase crossover on the L(j𝜔) plot is a point at

which theplot intersects the negative real axis.

Def: The phase crossover frequency 𝜔𝑃 is the frequency of

the phase crossover, on which ∠𝐿(𝜔𝑃 ) = 180°

1 ωp ω=ꝏ
𝐆𝐚𝐢𝐧 𝐌𝐚𝐫𝐠𝐢𝐧 = 𝐺 = 20 log10 𝑑𝐵
𝐿(𝑗𝜔𝑃 ) │L(jωp)│

= −20 log10 𝐿(𝑗𝜔𝑃 )

ω=0
EE 3631 Control Theory 61
Notes:
 The L(jω) plot doesn’t intersect the negative real axis
→ L(jωP ) = 0 → GM = ∞ dB

 The L(jω) plot intersects the negative real axis between the
0 and − 1 point 0 < L(jωP ) < 1 → GM > 0 𝑑𝐵

 The L(jω) plot passes through the (−1, j0) point

𝐿(𝑗𝜔𝑃 ) = 1 → 𝐺𝑀 < 0 𝑑𝐵
 The L(jω) plot intersects the negative real axis after the point
(−1, j0) → it encloses the −1 point 𝐿(𝑗𝜔𝑃 ) > 1 → 𝐺𝑀 < 0 𝑑𝐵

EE 3631 Control Theory 62

 • The Gain Margin : is the amount of gain in dB that can be added
to the loop before the closed loop system becomes unstable.
• Therefore, in case1 above, the gain can be increased to any
value and the closed-loop system will always be stable.
• In case 2 above, the gain can be increased to a limited value
after which the closed-loop system will be unstable
• In case 3 the gain can not be increased and the closed loop
system is marginally stable.
• In case 4, the closed loop system is unstable and the loop gain
must be reduced to achieve stability.

** The above rules apply only to minimum phase systems !! **

EE 3631 Control Theory 63

 Phase Margin :

• Curve A and curve B have the same

gain margin.
• However, curve A is more stable.
• Any slight clockwise rotation will
make curve B enclose the -1 point.
• → gain margin alone is not enough
to indicate the relative stability.

If we add Phase Margin, we can get a better idea about

the relative stability of the closed-loop system

EE 3631 Control Theory 64

 Def : The Gain crossover is a point on the 𝐿(𝑗𝜔) plot at which the
magnitude of L(j𝜔) is 1

Def : Gain crossover frequency (𝜔𝑔 ) is the frequency of 𝐿(𝑗𝜔) of the

gain crossover, or where L(j𝜔𝑔 ) = 1

hase argin ( )= ∠ j − 180

EE 3631 Control Theory 65

 Phase Margin: is defined as the angle in degrees through which
the L(jω) plot must be rotated around the origin so that the gain
crossover passes through -1
Unit
circle

-1 1

Phase
Margin

EE 3631 Control Theory 66

 Example:
2500
𝐿(𝑠) =
𝑠(𝑠 + 5)(𝑠 + 50)

Gain crossover 𝜔𝑔 = 6.22 𝑟𝑎𝑑/𝑠𝑒𝑐

Phase crossover 𝜔𝑃 = 15.88 𝑟𝑎𝑑/𝑠𝑒𝑐
𝐿(𝑗𝜔𝑃 ) = 0.182
1
𝐺𝑀 = 20 log10 = 14.82 𝑑𝐵
0.182

∠𝐿 𝑗𝜔𝑔 = 211.72°
𝑃𝑀 = 211.72 − 180° = 31.72°

EE 3631 Control Theory 67