Contents

C HAPTER 1
7 Introduction to Control systems

1.1 Introduction 7
1.2 System categories 8
1.3 Advantages of Control Systems 9
1.4 Examples of controlled systems 10
1.4.1 Fuel cell hybrid electric vehicle 10
1.4.2 Green-house system 10

C HAPTER 2
13 Mathematical Modelling of Systems

2.1 Differential equation 13

2.2 Transfer function 15
2.2.1 Poles and zeros 16
2.2.2 Examples 17
2.2.3 Electrical Network Transfer Function 17
2.2.4 Transnational Mechanical System Transfer Functions 19
2.3 State space model 21
2.3.1 Converting from State Space to a Transfer Function 24
2.4 Converting a Transfer Function to State Space 25
2.5 Problems 29

C HAPTER 3
39 Block Diagrams Models

3.1 Block Diagram Models 39

3.1.1 Cascade Form 40
3.1.2 Parallel Form 41
3.1.3 Feedback Form 41
3.1.4 Moving of summing points 41
3.1.5 Moving of pick off points 42

1
2 CONTENTS

3.2 Signal Flow Graph (SFG) 44

3.2.1 Basic Elements of Signal Flow Graph 44
3.2.2 Flow-Graph Algebra 45
3.2.3 Mason’s gain formula 45
3.3 Problems 49

C HAPTER 4
55 Time Response Analysis

4.1 Transient-Response Specifications 56

4.2 First order system 57
4.2.1 Time constant τ 57
4.2.2 Rise Time, Tr 58
4.2.3 Settling Time, Ts 58
4.3 Second order system 59
4.3.1 Critically damped case ξ = 1 60
4.3.2 Overdamped case ξ > 1 60
4.3.3 Underdamped case 0 < ξ < 1 60
4.4 Transient system specifications for Second-Order Systems 63
4.5 Steady-state error 65
4.5.1 Type characterization of Control Systems 65
4.5.2 Error Constants 65
4.5.3 Static Position Error Constant Kp 66
4.5.4 Static velocity Error Constant Kv 66
4.5.5 Static acceleration Error Constant Ka 67
4.6 Problems 69

C HAPTER 5
81 Stability of linear system

5.1 Introduction 81
5.2 Definition 81
5.2.1 Routh’s stability criterion 83
5.2.2 Stability Design via Routh’s Stability Criterion 87
5.3 Problems 89

C HAPTER 6
93 Root Locus Technique
 CONTENTS 3

6.1 The Root Locus Concept 93

6.2 Angle and magnitude criteria 95
6.3 Construction Rules 96
6.3.1 Number of branches 96
6.3.2 Symmetry 96
6.3.3 Starting and ending points 96
6.3.4 Real axis loci 96
6.3.5 Asymptotes 96
6.3.6 Breakaway point 96
6.3.7 Imaginary axis crossing 97
6.3.8 Departure and arrival angles 97
6.4 Problems 101

C HAPTER 7
111 The Frequency-response analysis

7.1 Frequency Response 111

7.2 Frequency Response Characteristics 114
7.3 Plotting Frequency Response 114
7.3.1 Bode plot 114
7.3.2 Nyquist Plots 117
7.3.3 Relative stability: Nyquist criterion 118
7.4 Frequency-Domain Performance Specifications 122
7.4.1 Damping ratio and frequency response 122
7.4.2 Damping ratio and phase margin 123
7.4.3 Response speed and frequency response 123
7.5 Problems 124

C HAPTER 8
131 Control design

8.1 Root-locus design 131

8.1.1 PD controller 133
8.1.2 PI Controller 133
8.1.3 PID Controller 134
8.1.4 phase-Lag and phase-Lead Controllers 134
8.2 Improvement of Transient Response 136
8.2.1 PD Controller Design 136
8.2.2 Phase-Lead Controller Design 139
4 CONTENTS

8.3 Improvement of Steady State Errors 143

8.3.1 PI Controller Design 143
8.3.2 Phase-Lag Controller Design 146
8.4 Improving Steady-State Error and Transient Response 147
8.4.1 PID Controller Design 147
8.4.2 Lag-Lead Controller Design 148
 Preface
Control engineering is an exciting and a challenging field. By its very nature, control engineering is
a multi-disciplinary subject, and it has taken its place as a core course in the engineering curriculum.
Control engineering is considered as an important and integral part of space-vehicle systems,
robotic systems, modern manufacturing systems, and any industrial operations involving control of
temperature, pressure, humidity, flow, etc.
This book, intended to be a text book on control systems at the junior level, presents a control
engineering methodology that, while based on mathematical fundamentals, stresses physical system
modeling and practical control system designs with realistic system specifications.
It is written for electrical students. The reader is expected to have fulfilled the following
prerequisites: introductory courses on differential equations, Laplace transforms, vector matrix
analysis and circuit analysis.

5
 CHAPTER
1 Introduction to
Control systems

Content

1.1 Introduction 7
1.2 System categories 8
1.3 Advantages of Control Systems 9
1.4 Examples of controlled systems 10
1.4.1 Fuel cell hybrid electric vehicle 10
1.4.2 Green-house system 10

1.1 Introduction

Control systems are an integral part of modern society. From devices as simple as a toaster or a
toilet, to complex machines like space shuttles, a self-guided vehicle and power steering, control
engineering is a part of our everyday life. We are not the only creators of automatically controlled
systems; these systems also exist in nature. Within our own bodies are numerous control systems,
such as the pancreas, which regulates our blood sugar. Control system engineers are concerned
with understanding and controlling segments of their environment, often called systems, to provide
useful economic products for society. The most characteristic quality of control engineering is the
opportunity to control machines and industrial and economic processes for the benefit of society.
Control engineering is based on the foundations of feedback theory and linear system analysis,
and it integrates the concepts of network theory and communication theory. Therefore control
engineering is applicable to many engineering discipline such as aeronautical, chemical, mechanical,
environmental, civil, and electrical engineering.
A control system is an interconnection of components forming a system (plants or process)
for the purpose of obtaining a desired output with desired performance, given a specified input. A
process to be controlled can be represented by a block, as shown in Figure 4.11, where the input
represents a desired output.

7
8 Chapter 1. Introduction to Control systems

Input Output
Control System

Figure 1.1: Description of a control system

Here, are some formal definitions of terms used throughout this book:

Definition 1.1: System

A system is a combination of components that act together to perform a function not possible
with any of the individual parts.
A control system is an arrangement of physical components connected in such a manner as
to command, or regulate itself or another system.

Definition 1.2: Input/Output

The input is the stimulus, excitation or command applied to a control system, typically
from an external energy source, usually in order to produce a specified response from the
control system.
The output is the actual response obtained from a control system. It may or may not be
equal to the specified response implied by the input.

Systems can be categorized by the number of inputs and the number of outputs the system has.

SIS0 system is a system with one input and one output, also called single-input, single output, or
SISO.

MIMO system is a system with multiple inputs and multiple outputs, also called called multi-input,
multi-output, or MIMO.

1.2 System categories

Control systems are classified into two major categories or configurations: open-loop and closed-
loop systems.

• As shown in Figure 1.2, an open loop system is a system without feedback.

 1.3 Advantages of Control Systems 9

Disturbance Disturbance

Input or
Reference + +
Output
Transducer Controller + Process +

Figure 1.2: Open-loop system

• A closed-loop system is interconnected in a cycle and uses an additional measure of the

actual output to compare the actual output with the desired output response, Figure 1.3.

Disturbance Disturbance

Input or
Reference + +
Output
Transducer +
- Controller + Process +

Sensor

Figure 1.3: Closed-loop system

1.3 Advantages of Control Systems

Control systems are built for four primary reasons:

1. Power amplification

2. Remote control

3. Convenience of input form

4. Compensation for disturbances

Control systems are appropriate to compensate human disabilities and also useful in remote or
dangerous locations. For example, a remote-controlled robot arm can be used to pick up material
in a radioactive environment. Another advantage of a control system is the ability to compensate
for disturbances. Typically, we control such variables as temperature in thermal systems, position
and velocity in mechanical systems, and voltage, current, or frequency in electrical systems. The
system must be able to yield the correct output even with a disturbance.
10 Chapter 1. Introduction to Control systems

1.4 Examples of controlled systems

1.4.1 Fuel cell hybrid electric vehicle

The use of hybrid cars is becoming increasingly popular. Fuel cell hybrid electric vehicles (FCHEVs)
are electric vehicles powered by batteries and fuel cell. In conventional electric vehicles batteries
provide power to the electric motor, in the CHEVs batteries and fuel cell are connected in a parallel
system and together provide power. Figure 1.4 shows a fuel cell stack connected with a DC/ DC
converter needed to provide a regulated voltage at the output. Battery pack is connected with
auxiliary devices and with fuel cell. The DC/ AC inverter converts the direct current (DC) in
alternate current (AC) in order to fed electric motor. The motor is able to recover part of energy
that would normally be lost due to braking (regenerative braking). This recovered energy is used to
recharge batteries.

Figure 1.4: Fuel cell hybrid electric vehicle

1.4.2 Green-house system

A fruit crop experiment station needed a monitoring and control solution for its climate systems
in a number of greenhouses. The proposed system needed to monitor soil moisture content and
temperature and actively control irrigation and airflow to create a favorable environment for the
crop in order to reach predetermined results for high yield, high quality and low costs. Figure 1.5.
 1.4 Examples of controlled systems 11

Figure 1.5: Greenhouse system

 CHAPTER Mathematical
2 Modelling of
Systems
Content

2.1 Differential equation 13

2.2 Transfer function 15
2.2.1 Poles and zeros 16
2.2.2 Examples 17
2.2.3 Electrical Network Transfer Function 17
2.2.4 Transnational Mechanical System Transfer Functions 19
2.3 State space model 21
2.3.1 Converting from State Space to a Transfer Function 24
2.4 Converting a Transfer Function to State Space 25
2.5 Problems 29

Before designing control systems for many real processes, the designer must have a com-
prehensive understanding of how these processes operate. Thus, it is compulsory to analyse the
relationships between the system variables and obtain quantitative mathematical equations using
the physical laws. The mathematical representation obtained is called model.
In this chapter we provide an introduction to the concept of modelling and present some basic
material on the following specific methods commonly used in feedback and control systems:

1. differential equations.

2. transfer functions.

3. state space equations.

2.1 Differential equation

A model is a mathematical representation of physical, biological or information systems that allows
us to understand the systems behaviours. This model is defined as a set of differential equations.
In this text we are limited to the linear systems which are governed by linear ordinary differential
equations.

13
14 Chapter 2. Mathematical Modelling of Systems

Definition 2.1: Linear systems

• A system is called linear if the principle of superposition applies.

• The principle of superposition states that the response produced by the simultane-
ous application of two different forcing functions is the sum of the two individual
responses.

Definition 2.2: Linear ordinary differential equation

A system represented by a linear ordinary differential equation, of any order n, having
constant coefficients, and with input u(t) and output y(t):

dn y(t) dn−1 y(t) dy(t) dm u(t) du(t)

an n
+ a n−1 n−1
+ · · · + a 1 + a 0 y = b m m
+ · · · + b1 + b0 u (2.1)
dt dt dt dt dt
is linear time-invariant if the system is not initially energized (i.e., the initial conditions are
zero, and the input u(t) is zero for t < 0).

Example 2.1
Find the differential equation relating the capacitor voltage, vC (t) , to the input voltage, e(t)
in Figure 2.13.

L R

e(t) + vC (t) C
−
−

Figure 2.1: RLC network

Answer:
Summing the voltages around the loop, assuming zero initial conditions, yields the differen-
tial equation for this network as

di(t)
Ri(t) + L + vC (t) = e(t) (2.2)
dt
dvC (t)
Changing variables from i = C , yields
dt
d 2 vC (t) dvC (t)
LC + RC + vC (t) = e(t) (2.3)
dt 2 dt
 2.2 Transfer function 15

Example 2.2
Consider the spring-mass system shown in Figure 2.4.

F1 F2

k1 k2
m1 m2

c1 c2
y1 y2

Figure 2.2: Mechanical system

The following two equations of motion for this system can be written by using the basic laws
of dynamics

2
 F1 = m1 ddty21 + c1 ( dydt1 − dydt2 ) + k1 (y1 − y2 )
2
(2.4)
F
2 = m2 ddty22 − c1 ( dydt1 − dydt2 ) − k1 (y1 − y2 ) + c2 dydt2 + k2 y2

2.2 Transfer function

In control theory, transfer functions are commonly used to characterize the input-output relationships
of components or systems that can be described by linear, time-invariant, differential equations.
The transfer function of a system represents the relationship describing the dynamics of the system
under consideration.
First, we begin by defining the transfer function and then we derive the transfer function of a
differential equation system.

Definition 2.3: Transfer function

The transfer function of a linear system is defined as the ratio of the Laplace transform
of the output variable to the Laplace transform of the input variable, with all initial
conditions assumed to be zero.

Let us begin by writing a general nth-order, linear, time-invariant differential equation,

dn y(t) dn−1 y(t) dy(t) dm u(t) du(t)

an + an−1 + · · · + a 1 + a 0 y = b m + · · · + b1 + b0 u (2.5)
dt n dt n−1 dt dt m dt

where y(t) is the output, u(t) is the input,

16 Chapter 2. Mathematical Modelling of Systems

Taking Laplace Transforms converts the differential equation into the following algebraic equation:

an snY (s) + an−1 s(n−1)Y (s) + · · · + a1 sY (s) + a0Y (s) (2.6)

= bm smU(s) + bm−1 s(m−1)U(s) + · · · + b1 sU(s) + b0U(s)

or equivalently

Y (s) = G(s)U(s) (2.7)

Now form the ratio of the output transform, Y (s), divided by the input transform, U(s), we obtain

N(s) bm sm + bm−1 s(m−1) + · · · + b1 s + b0

G(s) = = (2.8)
D(s) an s(n) + an−1 s(n−1) + · · · + a1 s + a0

where

N(s) = bm sm + bm−1 s(m−1) + · · · + b1 s + b0 (2.9)

D(s) = an s(n) + an−1 s(n−1) + · · · + a1 s + a0 (2.10)

G(s) is the transfer function.

2.2.1 Poles and zeros

Definition 2.4: Poles and zeros

• The poles of a transfer function are any roots of the denominator of the transfer
function that are common to roots of the numerator. The poles are plotted on the
complex s-plane using a cross (×).

• The zeros of a transfer function are any roots of the numerator of the transfer function
that are common to roots of the denominator. The zeros are plotted on the complex
s-plane using a circle (◦).

For the transfer function G(s):

• The roots of N(s) = 0 are the system zeros: m zeros.

• The roots of D(s) = 0 are the system poles: n poles.

• If m < n we say that the model is strictly proper.

• If m = n we say that the model is biproper .

• If m > n we say that the model is improper.

 2.2 Transfer function 17

2.2.2 Examples

Example 2.3
Y (s)
Find the transfer function, G(s) = , corresponding to the differential equation
R(s)

d3y d2y dy dr
+ 2 + 5 + 7y = 2 + 3r (2.11)
dt 3 dt 2 dt dt
Answer:
Y (s) 2s + 3
G(s) = = (2.12)
R(s) s3 + 2s2 + 5s + 7

Example 2.4
Find the differential equation corresponding to the transfer function,

X(s) s+1
G(s) = = 2 (2.13)
U(s) s + 2s + 4

Answer:
d2x dx du
+ 2 + 4x = +u (2.14)
dt 2 dt dt

2.2.3 Electrical Network Transfer Function

By combining electrical components into circuits and deciding on the input and output you can find
the transfer function. Our guiding principles are Kirchhoff’s laws.

Component Voltage-current Current-voltage Impedance

V (s)
Z(s) =
I(s)

vR = Ri vR R
Resistor i=
R
R
i
+ −
vR

Inductance di 1R Ls
vL = L i= vL dt
L dt L
i
+ vL −

Capacitance 1R dvc 1
C vc = ic dt ic = C
C dt Cs
i
+ −
vC

Table 2.1: Voltage-current and impedance relationships for capacitors, resistors, and inductors
18 Chapter 2. Mathematical Modelling of Systems

Example 2.5
Find the transfer function relating the capacitor voltage, VC (s) , to the input voltage, E(s) in
Figure 2.13.
Answer:
Taking the Laplace transform of (2.3) and assuming zero initial conditions, rearranging
terms, and simplifying yields

(LCs2 + RCs + 1)VC (s) = E(s) (2.15)

Then, we can obtain the following transfer function

VC (s) 1
= 2
(2.16)
E(s) LCs + RCs + 1

Example 2.6
I2 (s)
Given the network of Figure 2.3, find the transfer function, .
E(s)

2Ω 2Ω
i2

e(t) + 1H 1
− F
8

Figure 2.3: Two-loop electrical network

The equivalent circuit in s domain is

2Ω 2Ω

I1 (s) I2 (s)

E(s) + s VL (s) 8
−
s
IL (s)

According to the current Kirchhoff law, we have

I1 (s) = IL (s) + I2 (s) (2.17)

E(s) −VL (s) VL (s)
= + I2 (s) (2.18)
2 s
 2.2 Transfer function 19

Then we obtain
s+2
E(s) = 2I2 (s) + VL (s) (2.19)
s
An another hand, we have
8 (2s + 8)
VL (s) = (2 + )I2 (s) = I2 (s) (2.20)
s s
Thus, we get

2s2 + (2s + 8)(s + 2)

E(s) = I2 (s) (2.21)
s2
So
I2 (s) s2
= 2 (2.22)
E(s) 4s + 12s + 16

2.2.4 Transnational Mechanical System Transfer Functions

Consider a simple system with a mass that is separated from a wall by a spring and a dashpot. The
mass could represent a car, with the spring and dashpot representing the car’s bumper. An external
force is also shown. Only horizontal motion and forces are considered.

Component Force-velocity Force-displacement Impedance

F(s)
Z(s) =
X(s)

Rt
f (t) = K f (t) = Kx(t) K
Spring 0 v(τ)dτ
K F

Viscous damper f (t) = Cv(t) dx(t) Cs

f (t) = C
c dt
F

Mass v(t) d 2 x(t) Ms2

f (t) = M f (t) = M
F dt dt 2
M

Table 2.2: Force-velocity, force-displacement, and impedance translational relationships for springs,
viscous dampers, and mass
20 Chapter 2. Mathematical Modelling of Systems

Example 2.7
X(s)
Find the transfer function, , for the system of Figure 2.4
F(s)

k c

Figure 2.4: Mass, spring, and damper system

The mechanical model of the system is given as

d 2 x(t) dx(t)
m +c + kx(t) = F (2.23)
dt 2 dt
Taking the Laplace transform of (2.23) and assuming zero initial conditions, rearranging
terms, and simplifying yields

(ms2 + cs + k)X(s) = F(s) (2.24)

Then, we can obtain the following transfer function

X(s) 1
= 2 (2.25)
F(s) ms + cs + k
 2.3 State space model 21

2.3 State space model

As systems become more complex, representing them with differential equations or transfer
functions becomes cumbersome. On the other hand, the transfer function representation is not
useful for,

• Systems with initial conditions

• Nonlinear systems

• Time varying systems and

• Multiple input multiple output (MIMO) systems

The state space concept, promoted in the 1950s, was evolved to overcome the draw backs of
transfer functions models. This representation forms the basis for the development of modern
control systems. It is considered as the most powerful and dominant technique used for modelling,
analysing, and designing a wide range of dynamic systems.
For a given dynamic system, there exists a minimum set of physical variables, called the state
variables, that fully describe the system and its response to any given set of inputs.

Definition 2.5: State model

The state of a system is a set of variables whose values, together with the input signals and
the equations describing the dynamics, will provide the future state and output of the system.

The basic terms associated with the state models, are now defined:

Definition 2.6

• State vector of a dynamic system is the column vector x whose components are the
state variables, x = [x1 , x2 , · · · , xn ].

• State space is an n-dimensional space containing the n-system state variables. The
state of a dynamic system at any instant of time t is represented by a single point in
the state space.

• State equations is a set of n simultaneous, first-order differential equations with n

variables, where the n variables to be solved are the state variables.

• Output equation is an algebraic equation that expresses the output variables of a

system as linear combinations of the state variables and the inputs.
22 Chapter 2. Mathematical Modelling of Systems

A linear system is represented in state-space form by the following equations:


 ẋ = Ax + Bu
(2.26)
 y = Cx + Du

for t ≥ t0 and initial conditions x(t0 ), where

• x : state vector

• ẋ : time derivative of the state vector

• y : output vector

• u : input or control vector

• A : system matrix

• B : input matrix

• C : output matrix

• D : feed forward matrix

This representation of a system provides complete knowledge of all variables of the system at any
t ≥ t0 .

• The first equation in 2.26 is called the state equation, and the vector x, the state vector,
contains the state variables.

• The second equation in 2.26 is called the output equation. This equation is used to calculate
any other system variables.

Example 2.8
Find the state space system of the following circuit. Figure 2.5.

L R

e(t) + vC C
−

Figure 2.5: RLC network

Answer: Summing the voltages around the loop, assuming zero initial conditions, yields the
 2.3 State space model 23

differential equation for this network as

di(t)
Ri(t) + L + vC (t) = e(t) (2.27)
dt
di(t) R 1
= −vC (t) − i(t) + e(t) (2.28)
dt L L
dvC (t)
From the current equation i = C , yields
dt
dvC (t) 1
= i(t) (2.29)
dt C
Select the capacitor voltage vC (t) and the inductor current i(t) as state variables, and generates
the following pair of state equations:
 
" # 1 " #  
v̇C (t)  0 vC (t) 0
= 1 C 
+  1  e(t) (2.30)
i̇(t) R i(t)
− − L
L L
The required output equation is:
" #
h i v (t) h i
C
vC (t) = 1 0 + 0 e(t) (2.31)
i(t)

Example 2.9
Given the network of Figure 2.6, find the state space system.

5Ω
i3
5Ω 0.5H 0.5H
i1 i2

e(t) + vc 0.2F 3Ω
−
ic

Figure 2.6: Electrical network

24 Chapter 2. Mathematical Modelling of Systems

According to the Kirchhoff laws we have,

di1
e(t) = 5i1 + 0.5 + vc (2.32)
dt
di2
vc (t) = 0.5 + 3i2 (2.33)
dt
i3 + i1 = i2 + ic (2.34)

From the current equation, yields

e − vc 1 dvc
+ i1 − i2 = (2.35)
5 5 dt
Finally, we obtain

di1
 
    
 dt  −10 0 −2 i 2
 di2    1  
= 0 −6 2   i2  + 0 (2.36)
e
    
 dt  
 dv 
c 5 −5 −1 vc 1
dt  
h i  i1 
vc = 0 0 1   i2 
 (2.37)
vc

2.3.1 Converting from State Space to a Transfer Function

Using the Laplace transform, we can derive the transfer function for a state space system. Consider
a linear state space system described by

ẋ = Ax + Bu
(2.38)
y = Cx + Du

Taking Laplace transforms, yields

sX(s) − x(0) = AX(s) + BU(s) (2.39)

Y (s) = CX(s) + DU(s) (2.40)

and the state equations may be rewritten:

(sI − A)X(s) = x(0) + BU(s) (2.41)

where I is the identity matrix.

Solving for X(s) in Equation 2.41,

X(s) = (sI − A)−1 x(0) + (sI − A)−1 BU(s) (2.42)

 2.4 Converting a Transfer Function to State Space 25

Substituting X(s) in the output equation:

Y (s) = C((sI − A)−1 x(0) + (sI − A)−1 BU(s)) + DU(s) (2.43)

under the assumption that all initial values are zero, we obtain

Y (s) = C((sI − A)−1 B + D)U(s) (2.44)

Therefore the transfer function is

G(s) = C(sI − A)−1 B + D (2.45)

Example 2.10
Find the transfer function of the following system
" # " #
0 1 0
ẋ = x+ u (2.46)
−3 −4 1
h i
y = 10 0 x (2.47)

• First, find sI − A
" #
s −1
(sI − A) = (2.48)
3 s+4

• Then, we can calculate

" #
1 s + 4 1
(sI − A)−1 =
s(s + 4) + 3 −3 s

• Finally, we obtain the transfer function:

10
G(s) = (2.49)
s2 + 4s + 3

2.4 Converting a Transfer Function to State Space

State equations may be obtained from an nth order differential equation or directly from the system
model by identifying appropriate state variables. To illustrate the first method, consider a linear
plant model described by the differential equation

dn y(t) dn−1 y(t) dy(t)

n
+ a n−1 n−1
+ · · · + a1 + a0 y(t) = b0 u(t) (2.50)
dt dt dt
Where y(t) is the plant output and u(t) is the plant input. According to the choice of a set of state
variables, the state model describing this system is not unique.
26 Chapter 2. Mathematical Modelling of Systems

A useful set of state variables, referred to as phase variables, is defined as:




 x1 = y(t)

dy(t)

x = ẋ1 =

 2


 dt
d2 y(t)
x3 = ẋ2 = dt 2
(2.51)


 ..



 .

 dn−1 y(t)
xn = ẋn−1 =

dt n−1

From equation 2.50 we have

dn y(t)
ẋn = = −a0 x1 − a1 x2 − · · · − an−1 xn + b0 u(t) (2.52)
dt n
In matrix form, this looks like
      
ẋ1 0 1 0 ··· 0 x1 0
      
ẋ2   0 0 1 ··· 0  x2   0 
   
 . = .  .  +  .  u(t) (2.53)
   
 ..   .. .. .. .. ..  . .
   . . . .  .   . 
   

ẋn −a0 −a1 · · · −an−1 xn b0

The output equation is

 
x1
 
h i x2 
y= 1 0 0 ··· 0  .  (2.54)
 
 .. 
 
xn

Example 2.11

R(s) 24 Y (s)
s3 + 9s2 + 26s + 24

Find the state space representation of the given system.

1. We will find the differential equation of the system

d 3 y(t) d 2 y(t) dy(t)

3
+ 9 2
+ 26 + 24y(t) = 24r(t) (2.55)
dt dt dt

2. Let
 
x =y ẋ = x2
 1  1

 

x2 = ẏ −→ ẋ2 = x3

 

x3 = ÿ ẋ3 = −24x1 − 26x2 − 9x3 + 24r
 
 2.4 Converting a Transfer Function to State Space 27

      
ẋ 0 1 0 x 0
 1    1  



ẋ2  =  0
 x2  +  0  r




   0 1    

 ẋ3
 −24 −26 −9 x3 24
  (2.56)
i x1 



 h




 y= 1 0 0  x2 


x3


Note that the third row of the matrix A is has the same coefficients as the denominator of the
transfer function but negative and in reverse order.

In the next example, we will consider a case where the transfer function has a polynomial in s in
the numerator that is of order less than the polynomial in the denominator.

Example 2.12
Example of Transfer function with Polynomial Term in the numerator

R(s) s2 + 7s + 2 Y (s)
s3 + 9s2 + 26s + 24

In order to find the state-space representation of the given system we have to handle the
numerator and the denominator separately.

R(s) 1 X1 (s) Y (s)

s2 + 7s + 2
s3 + 9s2 + 26s + 24

We convert the first block into a state-space representation using the state variables x1 , x2
and x3 , in which x1 is the output.
 
x1 = x1

 ẋ1 = x2


x2 = ẋ1 −→ ẋ2 = x3

 

x3 = ẍ1 ẋ3 = −24x1 − 26x2 − 9x3 + r
 

For the second block we can express the output of the system as:

Y (s) = (s2 + 7s + 2)X1 (s) (2.57)

Taking the inverse Laplace transform with zero initial condition, we get:

d 2 x1 dx1
+7 + 2x1 = y(t) (2.58)
dt 2 dt
28 Chapter 2. Mathematical Modelling of Systems

In the matrix form, the following system can be obtained:

      
ẋ1 0 1 0 x 0
  1  


   
ẋ2  =  0 1  x2  + 0

0 r
       



 ẋ3
 −24 −26 −9 x3 1
  (2.59)
i x1 



 h




 y= 2 7 1  x2 


x3


Note that the row vector of the output matrix has the same coefficients as the numerator of
the transfer function but in reverse order.
 2.5 Problems 29

2.5 Problems

Exercise 2.1

You are given the LR circuit shown in Figure 2.7 where e(t) is the excitation and u(t) is the
response.

1. Determine the governing differential equation for the circuit.

U(s)
2. Determine the transfer function G(s) = for the circuit.
E(s)

e(t) + u(t) R
−

Figure 2.7: RL circuit

Solution 2.1

1. According to Khirchoff law, we have

e(t) = vl + vR (2.60)
di(t)
e(t) = L + Ri (2.61)
dt

2. Using Laplace transform, we obtain

E(s) = LsI(s) + RI(s) (2.62)

than
I(s) 1
= (2.63)
E(s) Ls + R
U(s) I(s) R
=R = (2.64)
E(s) E(s) Ls + R

Exercise 2.2

Consider the following electric circuit

30 Chapter 2. Mathematical Modelling of Systems

4Ω i1 0.5H

+ 1
e(t) − Vc (t) F 2Ω
6

i2

Figure 2.8: Electric circuit

Vc (s)
1. Determine the transfer function G(s) = for the circuit.
E(s)
I1 (s)
2. Determine the transfer function F(s) = for the circuit.
E(s)
VL (s)
3. Determine the transfer function H(s) = for the circuit.
E(s)

Solution 2.2

The equivalent circuit in s domain is

4 I1 (s) 0.5s

+ 6
E(s) − Vc (s) 2
s
Ic (s)
I2 (s)

Figure 2.9: Equivalent circuit

6
1. Let Z1 (s) = 4 + 0.5s, Zc (s) = and Z2 (s) = 2. According voltage divider rule, we
s
have
6
Zeq = Zc //Z2 (s) = (2.65)
s+3
Zeq
Vc (s) = E(s) (2.66)
Zeq + Z1 (s)

Thus
Vc (s) 12
G(s) = = 2 (2.67)
E(s) s + 11s + 36
 2.5 Problems 31

2. From the first loop, we have

E(s) −Vc (s)

I1 (s) = (2.68)
Z1 (s)
1 2(s + 3)
F(s) = (1 − G(s)) = 2 (2.69)
Z1 (s) s + 11s + 36

3. Write

VL (s) = 0.5sI1 (s) (2.70)

s(s + 3)
H(s) = 0.5sF(s) = (2.71)
s2 + 11s + 36

Exercise 2.3

Consider the following circuit

C2
R2
C1

−
R1
+
vi
vo

Figure 2.10:

Vo (s)
1. Determine the transfer function G(s) = for the circuit.
Vi (s)

Solution 2.3

1 R1 1 C2 R2 s + 1
Let Z1 (s) = R1 // = and Z2 (s) = R2 + = .
C1 s 1 + R1C1 s C2 s C2 s
32 Chapter 2. Mathematical Modelling of Systems

I2 (s) Z2 (s)
Z1 (s)
−
I1 (s)
+
vi
vo

Figure 2.11:

From the circuit we have

I1 = +I2 (2.72)
Vi (s) Vo (s)
=− (2.73)
Z1 (s) Z2 (s)

Then

I1 = +I2 (2.74)
Vo (s) Z2 (s)
=− (2.75)
Vi (s) Z1 (s)
(1 + R1C1 s)(1 +C2 R2 s)
=− (2.76)
C2 R1 s

Exercise 2.4

You are given the following differential equation where x(t) is the excitation and y(t) is the
response

dy3 (t) dy2 (t) dy(t) dx(t)

3
+ 8 2
+ 37 + 50y(t) = + 5x(t) (2.77)
dt dt dt dt
Y (s)
1. Find the transfer function G(s) =
X(s)
2. It can be shown that the transfer function G(s) may be written as
s+a
G(s) = (2.78)
(s + 2)(s2 + bs + c)

(a) Determine the constants a, b and c.

(b) Determine the poles and zeros G(s).
 2.5 Problems 33

Solution 2.4

1. Using Laplace transform, we obtain

(s3 + 8s2 + 37s + 50)Y (s) = (s + 5)X(s) (2.79)

than
Y (s) s+5
G(s) = = (2.80)
X(s) s3 + 8s2 + 37s + 50

2.
s+a s+a
G(s) = 2
= 3 2
(2.81)
(s + 2)(s + bs + c) s + (b + 2)s + (c + 2b)s + 2c

then
s+5
G(s) = (2.82)
(s + 2)(s2 + 6s + 25)

The poles are −2, −3 + 4 j and −3 − 4 j, whereas the zero is −5

Exercise 2.5

A system with input u and output y is identified by a differential equation of the form

dy2 (t) dy(t) du(t)

2
+ (a + b) + aby(t) = + cu(t) (2.83)
dt dt dt
dy(0− )
where a, b, c are positive real numbers. Assume that u(0− ) = 0, y(0− ) = 0 and = 0.
dt
1. Find the transfer function from U(s) to Y (s) and identify its poles and zeros.

2. Assume that c = b. Find the output y(t), t ≥ 0 when u is a unit step input.

3. Assume that a = b. Find the output y(t),t ≥ 0 when u is a unit step input.

Solution 2.5

1. The transfer function is obtained as

s+c s+c
G(s) = = (2.84)
(s2 + (a + b)s + ab) (s + a)(s + b)
The poles are −a, and −b, whereas the zero is −c.
1
2. For c = b, and the input is unit step, U(s) = , we have
s
1 1
1
Y (s) = = a− a (2.85)
s(s + a) s s+a
34 Chapter 2. Mathematical Modelling of Systems

The time domain solution is obtained as

1 1
y(t) = ( − e−at )u(t) (2.86)
a a

1
3. For a = b, and the input is unit step, U(s) = , we have
s
c c a−c
s+c a 2 a2 a
Y (s) = = − + (2.87)
s(s + a)2 s s + a (s + a)2

Using the Laplace transform table, we find that

1 1 a − c −at
y(t) = ( 2
− 2 e−at + te )u(t) (2.88)
a a a

Exercise 2.6

A common actuator in control systems is the DC motor. It directly provides rotary motion
and, coupled with wheels or drums and cables, can provide translational motion. The electric
equivalent circuit of the armature and the free-body diagram of the rotor are shown in the
following figure.

R L
i

ω, θ
+ −
U − E +

Figure 2.12: DC Motor

We will assume that the input of the system is the voltage source (U) applied to the motor’s
armature, while the output is the rotational speed of the shaft ω. The rotor and shaft are
assumed to be rigid. We further assume a viscous friction model, that is, the friction torque
is proportional to shaft angular velocity.
The physical parameters for our example are:

• (J) moment of inertia of the rotor 0.01kg.m2

• (b) motor viscous friction constant 0.1N.m.s

• (Ke ) electromotive force constant 0.01V /rad/sec

• (Kt ) motor torque constant 0.01N.m/A

 2.5 Problems 35

• (R) electric resistance 1Ω

• (L) electric inductance 0.5H

From the figure above, we can derive the following governing equations based on Newton’s
2nd law and Kirchhoff’s voltage law.

 L di + Ri = U − E

dt (2.89)
dω
 J
 + bω = T
dt
dθ
where T = Kt i, E = Ke ω and ω =
dt
1. Sketch the block diagram of the system.
" #
i
2. Assume the state vector is . Determine the matrix differential equation of the
ω
system.
ω(s)
3. Determine the transfer function of the system, F(s) = .
U(s)

Solution 2.6

Response:

1. System (2.89 )can be written as


di
L + Ri = U − Ke ω


dt (2.90)
dω

 J + bω = Kt i
dt
Using Laplace transform we obtain

1
 I(s) =
 .(U(s) − Ke ω(s))
Ls + R (2.91)
1

 ω(s) = .Kt I(s)
Js + b

R(s) 1 I(s) T (s) 1 ω(s)

+
- Kt
Ls + R Js + b

Ke

Figure 2.13: Block diagram representation of the DC motor

36 Chapter 2. Mathematical Modelling of Systems

2. From system 2.90, we have


di R ω U

 = − i − Ke +
dt L L L (2.92)
dω Kt b

 = i− ω
dt J J
Then, we obtain the following system matrix:
   
di R Ke " #  1 
 dt  − L − L  i
 dω  =  Kt b ω + L U
  (2.93)
− 0
dt J J
" #
h i i
ω= 0 1 (2.94)
ω

 
di " #" # " #
 dt  −2 −0.02 i 2
 dω  = + U (2.95)
1 −10 ω 0
dt
" #
h i i
ω= 0 1 (2.96)
ω

3. The transfer function is given by

ω(s)
G(s) = = C(sI − A)−1 B + D (2.97)
U(s)

By some algebraic calculation, we obtain

" #−1 " #
−1 s+2 0.02 1 s + 10 −0.02
(sI − A) = = (2.98)
−1 s + 10 (s + 2)(s + 10) + 0.02 1 s+2

2
then, G(s) =
(s + 2)(s + 10) + 0.02
a
" # " #
a b 1 d −b
a For A= , the inverse can be found using this formula: A−1 =
c d ad − bc −c a

Exercise 2.7

A schematic diagram of an automobile suspension system is shown in Figure 2.14. As the

car moves along the road, the vertical displacements at the tires act as the motion excitation
to the automobile suspension [Link] motion of this system consists of a translational
motion of the centre of mass and a rotational motion about the centre of mass.
 2.5 Problems 37

Figure 2.14: Automobile suspension system

A simplified mathematical model of the system is given by the following differential equation:

dxo2 dxo dxi

m 2
+b − + k xo − xi = 0 (2.99)
dt dt dt
where the motion xi at point P is the input to the system and the vertical motion xo of the
body is the output.

1. Obtain the transfer function of the system.

2. Obtain the state space model.

Solution 2.7

Response: The model 2.99 can be written as

dxo2 dxo dxi

m +b + kxo = + kxi (2.100)
dt 2 dt dt
Then, the transfer function is given by

bs + k
G(s) = (2.101)
ms2 + bs + k

Exercise 2.8

The linearized equations governing the motion of a BOEING’s aircraft are

β̇ (t) = −0.313β (t) + 56.7p(t) + 0.232F(t)

ṗ(t) = −0.0139β (t) − 0.426p(t) + 0.0203F(t) (2.102)
α̇(t) = 56.7p(t)

where β (t) in the aircraft angle of attack, p(t) is the pitch rate, and α(t) represents the
aircraft pitch angle. The driving force F(t) stands for the elevator deflection angle.
38 Chapter 2. Mathematical Modelling of Systems

1. Obtain the transfer function of the system.

2. Obtain the state space model.

 CHAPTER
3 Block Diagrams
Models

Content

3.1 Block Diagram Models 39

3.1.1 Cascade Form 40
3.1.2 Parallel Form 41
3.1.3 Feedback Form 41
3.1.4 Moving of summing points 41
3.1.5 Moving of pick off points 42
3.2 Signal Flow Graph (SFG) 44
3.2.1 Basic Elements of Signal Flow Graph 44
3.2.2 Flow-Graph Algebra 45
3.2.3 Mason’s gain formula 45
3.3 Problems 49

We know that a process is a combination of a set components. These components are related
between them to perform a function. To understand the behavior of such process we commonly
use a graphical model called block diagram. A block diagram consist of a set of blocks related
with arrows, where the block represents a function carried out by the system and the arrows the
represent the flux of the signals between thee blocks.
This chapter provides the definition of the block diagram and presents a method for obtaining block
diagrams for physical systems. Finally, we discuss the techniques to simplify such diagrams.

3.1 Block Diagram Models

We can use the transfer functions of a model to construct a visual representation of the dynamics
of the model. Such a representation is a block diagram 3.1. Block diagrams can be used to
describe how system components interact with each other. Unlike a schematic diagram, which
shows the physical connections, the block diagram shows the cause and effect relations between
the components, and thus helps us to understand the system’s dynamics.

39
40 Chapter 3. Block Diagrams Models

Description Symbol

1. The arrow, which is used to represent a R(s)

variable and the direction of the cause-and-
effect relation.

2. The block, which is used to represent the R(s) Y (s)

G(s)
input-output relation of a transfer function.

3. The circle, generically called a summer, R(s) E(s)

+
+
which represents addition as well as sub-
traction, depending on the sign associated X(s)
with the variable’s arrow.

4. The takeoff point, which is used to obtain X(s)

the value of a variable from its arrow, for
use in another part of the diagram. X(s)

Table 3.1: Elements of block diagram

X(s) Y (s)
G(s)

Figure 3.1: Block diagram

We will now examine some common topologies for interconnecting subsystems and derive the
single transfer function representation for each of them. These common topologies will form the
basis for reducing more complicated systems to a single block.

3.1.1 Cascade Form

Two blocks in cascade can be replaced by an equivalent block with a transfer function is the product
of the subsystem’s transfer functions.
 3.1 Block Diagram Models 41

X1 (s) X2 (s) X3 (s) X1 (s) X3 (s)

G1 (s) G2 (s) ≡ G1 (s)G2 (s)

Figure 3.2: Cascade form and equivalent transfer function

3.1.2 Parallel Form

Parallel subsystems have a common input and an output formed by the algebraic sum of the outputs
from.

G1 (s)

X(s) +
Y (s) X(s) Y (s)
G2 (s) +
+
≡ G1 (s) + G2 (s) + G3 (s)

G3 (s)

Figure 3.3: Parallel form and equivalent transfer function

3.1.3 Feedback Form

The third topology is the feedback form, which will be seen repeatedly in subsequent chapters. The
feedback system forms the basis for our study of control systems engineering.

X(s) Y (s)
+
- G(s)
X(s) G(s) Y (s)
≡
1 + G(s)H(s)
H(s)

Figure 3.4: Feedback form and equivalent transfer function

The product, G(s)H(s), is called the open-loop transfer function, or loop gain.

3.1.4 Moving of summing points

Figure 3.5 shows equivalent block diagrams formed when the summing junction is moved from the
left or the right side of the block.
42 Chapter 3. Block Diagrams Models

X1 (s) X3 (s)
X1 (s) X3 (s) G(s) +
+
- G(s) -

≡
X2 (s)
X2 (s) G(s)

X1 (s) X3 (s) X1 (s) X3 (s)

G(s) +
-
+
- G(s)
≡
X2 (s) 1 X2 (s)
G(s)

Figure 3.5: Equivalent form for moving the summing junction to the left or to the right

3.1.5 Moving of pick off points

Figure 3.6 shows equivalent block diagrams formed when the pickoff point is moved from the left
or the right side of the block.

X1 (s) X2 (s)
X1 (s) X2 (s) G(s)
G(s)
≡
X1 (s) X1 (s) 1
G(s)

X1 (s) X2 (s)
X1 (s) X2 (s) G(s)
G(s)
≡
X2 (s) X2 (s)
G(s)

Figure 3.6: Equivalent form for moving the pickoff to the left or to the right

Example 3.1
Y (s)
Find the transfer function, , for the system of Figure 3.7
R(s)
 3.1 Block Diagram Models 43

H2 (s)

R(s) - Y (s)
+
- G1 (s) + G2 (s) +
+
G3 (s) G4 (s)

H1 (s)

H3 (s)

Figure 3.7: Complex block diagram

Answer: By three steps of transformation, we can obtain the transfer function of the system:
Step 1

H2 (s)
G4 (s)

R(s) - Y (s)
+
- G1 (s) + G2 (s) +
+
G3 (s) G4 (s)

H1 (s)

H3 (s)

Step 2

H2 (s)
G4 (s)

R(s) - - G3 (s)G4 (s) Y (s)

+ G1 (s) + G2 (s)
1 − G3 (s)G4 (s)H1 (s)

H3 (s)

Step 3
44 Chapter 3. Block Diagrams Models

R(s) G2 (s)G3 (s)G4 (s) Y (s)

+
- G1 (s)
1 − G3 (s)G4 (s)H1 (s) + G2 (s)G3 (s)H2 (s)

H3 (s)

Step 4

R(s) Y (s)
G(s)

G1 (s)G2 (s)G3 (s)G4 (s)

G(s) =
1 − G3 (s)G4 (s)H1 (s) + G2 (s)G3 (s)H2 (s) + G1 (s)G2 (s)G3 (s)G4 (s)H3 (s)

3.2 Signal Flow Graph (SFG)

When interconnections between blocks are complex with multi-inputs and multi-outputs, the
analysis and simplification of such graphical model becomes difficult and fastidious. As an
alternative for analyzing a complex system, the signal flow graph will be also introduced in this
chapter and the Mason’s rule will be provided to show how it can be further used to simplify the
block diagram models.
An SFG is graphical model that consists of a set of nodes which are connected by directed
branches. Each node represents a system’s variable, and the branch represents the functional
connections between two nodes

3.2.1 Basic Elements of Signal Flow Graph

• A node carry out two functions:

– Addition: the value of the variable designated by a node is equal to the sum of all
signals entering the node.
– Transmission: the value of the variable designated by a node is transmitted on every
branch leaving that node.

x a y = ax b z = by

Figure 3.8: Nodes in signal flow graph.

Three types of nodes are of particular interest:

 3.2 Signal Flow Graph (SFG) 45

1. Source nodes (independent nodes).These represent independent variables and have

only outgoing branches. In Figure 3.8 x is a source node.
2. Sink nodes (dependent nodes). These represent dependent variables and have only
incoming branches. In Figure 3.8 z is a sink node.
3. Mixed nodes (general nodes). These have both incoming and outgoing branches. In
Figure 3.8, y is a mixed node.

• A path is any connected sequence of branches whose arrows are in the same direction.

• A forward path between two nodes is one that follows the arrows of successive branches
and in which a node appears only once.

3.2.2 Flow-Graph Algebra

To simplify a signal flow graph, the following rules are useful:

• Series paths (cascade nodes): Series paths can be combined into a single path by multiplying
the transmittance.

• Forward-path gain: Is the product of gains found by traversing a path from the input node to
the output node of the signal-flow graph in the direction of signal flow.
The product of the transmittances in a series path.

• Parallel paths: Parallel paths can be combined by adding the transmittances.

• Feedback loop: A closed path that starts at a node and ends at the same node.

• Loop gain: The product of the transmittances of a feedback loop.

• Non-touching Loops: Are loops that do not have any common nodes

3.2.3 Mason’s gain formula

To find the transfer function between input and output nodes, a powerful formula was derived
by Mason (1953-1956). That formula is known as Mason’s gain formula which is given by the
following definition:

Definition 3.1: Mason’s formula

∑Nk=1 Pk ∆k
G(s) = (3.1)
∆
• Pk is the path gain for the kth path.

• ∆ is the determinant of the signal flow graph.

46 Chapter 3. Block Diagrams Models

• ∆k is the cofactor of path k.

∆ and ∆k are computed as follows:

• ∆ = 1- (sum of all loop gains) + (sum of products of the loop gains of all possible
combinations of non-touching loops taken two at a time) - (sum of products of the
loop gains of all possible combinations of non-touching loops taken three at a time)
+· · ·

• ∆k = value of ∆ for the part of the flow graph not touching the kth forward path.

Example 3.2
Find the equivalent signal flow of the following diagram. Figure 3.9.

H2 (s)

H1 (s)

R(s) - - Y (s)
+
- G1 (s) + G2 (s) +
- G3 (s)

−H2

−H1

1 G1 1 G2 G3 1
R(s) Y(s)
−1 −1

Figure 3.9: Block diagram and its signal flow graph.

• First, identify the forward-path gains. In this example there is only one:

T1 = G1 G2 G3 (3.2)

• Second, identify the loop gains. There are four, as follows:

1. L1 = −G1 (3.3)
2. L2 = −G3 (3.4)
3. L3 = −G2 G3 H1 (3.5)
4. L4 = −G1 G2 G3 H2 (3.6)
 3.2 Signal Flow Graph (SFG) 47

• Third, identify the non-touching loops taken two at a time. From Figure 3.9, we can
see that loop 1 does not touch loop 2 and loop 1 does not touch loop 3. Thus, the
combinations of non-touching loops taken two at a time are as follows:

1. G1 G3 (3.7)
2. G1 G2 G3 H1 (3.8)

• Now, we can form ∆ as,

∆ = 1 + G1 + G3 + G2 G3 H1 + G1 G2 G3 H2 + G1 G3 + G1 G2 G3 H1 (3.9)

We form ∆k by eliminating from ∆ the loop gains that touch the kth forward path.
Hence

∆1 = 1 (3.10)

• Using equation 3.1, we obtain the following transfer function:

G1 G2 G3
G= (3.11)
1 + G1 + G3 + G2 G3 H1 + G1 G2 G3 H2 + G1 G2 + G1 G2 G3 H1

Example 3.3
Find the equivalent signal flow of the following diagram. Figure 3.10.

G6

G1 G2 G3 G4 G5
R Y
−H1 −H2

Figure 3.10: Signal flow graph.

• First, identify the forward-path gains. In this example there is only one:

T1 = G1 G2 G3 G4 G5 (3.12)
T2 = G6 G4 G5 (3.13)

• Second, identify the loop gains. There are two, as follows:

1. L1 = − H1 G2 (3.14)
2. L2 = − H2 G4 (3.15)

• Third, identify the non touching loops taken two at a time. From Figure 3.10, we can
see that loop 1 does not touch loop 2.
48 Chapter 3. Block Diagrams Models

• Now, we can form ∆ and ∆k . Hence,

∆ = 1 + G2 H1 + G4 H2 + G2 G4 H1 H2 (3.16)

We form ∆k by eliminating from ∆ the loop gains that touch the kth forward path:

∆1 = 1 (3.17)
∆2 = 1 + G2 H1 (3.18)

• Using equation 3.1, we obtain the following transfer function:

G1 G2 G3 G4 G5 + G6 G4 G5 (1 + G2 H1 )
G= (3.19)
1 + G2 H1 + G4 H2 + G2 G4 H1 H2
 3.3 Problems 49

3.3 Problems

Exercise 3.1

Fig. 3.11 shows the block diagrams for a dynamic system.

R(s) - 1 1 Y (s)
+ +
-
s+a s+b

Figure 3.11: Block diagram

1. Find the transfer function Y (s)/R(s) by using block-diagram manipulations.

2. Consider a unit step reference input and assume that b = a + 1. Find all a values for
which the steady-state value of e(t) = r(t) − y(t) (i.e. ess = limt−→∞ e(t)) is zero.

Solution 3.1

By moving the peak point after the second block , we obtain the following figure:
50 Chapter 3. Block Diagrams Models

s+b

R(s) - 1 1 Y (s)
+ +
-
s+a s+b

Elimination the negative feedback loop, we get the following figure :

s+b

R(s) - 1 1 Y (s)
+
s+a s+b+1

Replacing two series blocks by one block as the following figure:

R(s) 1 Y (s)
+
-
(s + a)(s + b + 1)

s+b

Finally, elimination of negative feedback loop, results the following figure :

R(s) 1 Y (s)
(s + a)(s + b + 1) + (s + b)

Figure 3.12: Simplified block diagram

Exercise 3.2

C(s) C(s)
Obtain transfer functions and of the system shown in Figure 3.13.
R(s) D(s)
 3.3 Problems 51

G4
D(s)
R(s) E(S) +
U(S) +
C(s)
+
- G1 + G2 + G3

Figure 3.13: Block diagram

Solution 3.2

C(s) = G2 G3U(s) + G3 D(s) (3.20)

U(s) = G1 E(s) + G4 R(s) (3.21)
E(s) = R(s) − HC(s) (3.22)

Using these equations, we obtain

C(s) = G2 G3 (G1 + G4 )R(s) − G1 G2 G3 HC(s) + G3 D(s) (3.23)

Arranging, equation 3.3, yields

G2 G3 (G1 + G4 ) G3
C(s) = R(s) + D(s) (3.24)
1 + G1 G2 G3 H 1 + G1 G2 G3 H
Then, we get

C(s) G2 G3 (G1 + G4 ) C(s) G3

= , = D(s) (3.25)
R(s) 1 + G1 G2 G3 H D(s) 1 + G1 G2 G3 H

Exercise 3.3

C
Obtain transfer function of the system shown in Figure 3.14.
R
52 Chapter 3. Block Diagrams Models

H2

R - C
+
-
+
- G1 + G2 G3

H1

Figure 3.14: Block diagram

Solution 3.3

Answer: By four steps of transformation, we can obtain the transfer function of the system:
Step 1

H2

R - C
+
-
+
- G1 + G2 G3

H1
G3

R G2 G3 C
+
-
+
- G1
1 + G2 G3 H2

H1
G3

Step 3
 3.3 Problems 53

R G1 G2 G3 C
+
-
1 + G2 G3 H2 + G1 G2 H1

Step 4

R G1 G2 G3 C
1 + G2 G3 H2 + G1 G2 H1 + G1 G2 G3

Exercise 3.4

Consider the system shown in Figure 3.15. Obtain the system transfer function using the
Mason’s gain formula.

H3

−1
R 1 G1 G2 G3 G4 G5 Y

H1 H2

Figure 3.15:

Solution 3.4

Answer:

• Therea are 4 pathes:

P1 = G1 G2 G3 G4 G5
P2 = H1 G3 G4 G5
(3.26)
P3 = G1 G2 H2 G5
P4 = H1 H2 G5

• There are two loops

L1 = −G2 G3
(3.27)
L2 = H3

• L1 and L2 are touching loops, so ∆ is given as

∆ = 1 + G2 G3 − H3 (3.28)
54 Chapter 3. Block Diagrams Models

• The cofactors are given as

∆1 = 1
∆2 = 1 − H3
(3.29)
∆3 = 1
∆4 = 1 − H3

Y
• The transfer function is
R
Y G1 G2 G3 G4 G5 + H1 G3 G4 G5 (1 − H3 ) + G1 G2 H2 G5 + H1 H2 G5 (1 − H3 )
= (3.30)
R 1 + G2 G3 − H3
 CHAPTER
4 Time Response
Analysis

Content

4.1 Transient-Response Specifications 56

4.2 First order system 57
4.2.1 Time constant τ 57
4.2.2 Rise Time, Tr 58
4.2.3 Settling Time, Ts 58
4.3 Second order system 59
4.3.1 Critically damped case ξ = 1 60
4.3.2 Overdamped case ξ > 1 60
4.3.3 Underdamped case 0 < ξ < 1 60
4.4 Transient system specifications for Second-Order Systems 63
4.5 Steady-state error 65
4.5.1 Type characterization of Control Systems 65
4.5.2 Error Constants 65
4.5.3 Static Position Error Constant K p 66
4.5.4 Static velocity Error Constant Kv 66
4.5.5 Static acceleration Error Constant Ka 67
4.6 Problems 69

The first step in the analysis of a control system is describing the system by a mathematical
model. In chapter 2, we have seen how any given system can be modelled by a differential equation,
a transfer function or a state space model.

The next step consist of obtaining the system response, both transient and steady state, to a specific
input. The input can be a time varying function which may be described by known mathematical
functions.

The standard input signals used in control are : an impulse, a step, a ramp and a parabolic
input. By using these standard signals, the analysis and design of control systems are carried out,
defining certain performance measures for the system.

55
56 Chapter 4. Time Response Analysis

4.1 Transient-Response Specifications

The desired performances, characteristics of control system, are frequently specified in terms
of the transient response to a unit step-function input. The performance of a system is usually
evaluated in terms of the following qualities.

• How fast it is able to respond to the input ?

• How fast it is reaching the desired output ?

• What is the error between the desired output and the actual output, once the transients die
down and steady state is achieved?

• Does it oscillate around the desired value?

• Is the output continuously increasing with time or is it bounded?

In specifying the transient-response performances of a control system to a unit-step input, we

define the following quantities:

Definition 4.1: Response performances

1. Delay time, td : The delay time is the time required for the response to reach 50% of
the steady state value for the first time

2. Rise time, tr : the time required for the step response to rise from 10% to 90% of its
final value for critical and overdamped cases, and from 0% to 100% for underdamped
cases.

3. Peak time, t p : the time required for the response to reach the maximum or Peak value
of the response.

4. Maximum overshoot, Os : the difference between the peak value of the response and
the steady state vafue. It is usually expressed in percent of the steady state value.
yt p − y∞
Os% = 100 × (4.1)
y∞

5. Settling time, ts : the time required for the response to reach and remain within a
specified tolerance limits (usually ±2%or ±5%) around the steady state value.

6. Steady-state error ess : the error between the desired output and the actual output as
t −→ ∞ or under steady-state conditions.

These specifications are shown graphically in Figure 4.1.

Figure 4.1: Response specifications

 4.2 First order system 57

4.2 First order system

The dynamics of many systems of interest to engineers may be represented by a simple model
containing one independent energy storage element. For example, the braking of an automobile,
the discharge of an electronic camera flash, the flow of fluid from a tank, and the cooling of a cup
of coffee may all be approximated by a first-order differential equation, which may be written in a
standard form as

dy
+ ay = au(t) (4.2)
dt

The correspondent transfer function can be written as

a
G(s) = (4.3)
s+a

The first-order system can be described by the following block diagram:

U(s) a Y (s)
s+a

Figure 4.2: Block diagram of first order system

1
For a unit step input (U(s) = ) , the output Y (s) is given by:
s
a
Y (s) = (4.4)
s(s + a)

Taking the inverse transform, the step response is given by

1 1
y(t) = L−1 ( − ) = 1 − e−at (4.5)
s s+a

For a first order system, we define three transient response performance specifications.

4.2.1 Time constant τ

1
The system is defined by the single parameter τ = , the system time constant.
a
From Equation (4.5), the time constant is the time it takes for the step response to rise to 63% of its
final value.
58 Chapter 4. Time Response Analysis

4.2.2 Rise Time, Tr

Rise time is defined as the time for the waveform to go from 0.1 to 0.9 of its final value. Rise time
is found by solving equation 4.5 for the difference in time at y(t) = 0.9 and y(t) = 0.1 Hence,

2.2
Tr = (4.6)
a

4.2.3 Settling Time, Ts

Settling time is defined as the time for the response to reach, and stay within, 2%of its final value.
Letting y(t) = 0.98 in equation 4.5 and solving for time, t, we find the settling time to be

4
Ts = (4.7)
a

Figure 4.3 shows the transient response specifications, the constant time, the rise time and the
settling time.

← 63% of final value at t=τ

Tr
t
1 Ts
τ=
a

Figure 4.3: First order step response

Example 4.1
50
A system has a transfer function, G(s) = . Find the time constant, τ , settling time, Ts
s + 50
, and rise time, Tr .
 4.3 Second order system 59

Answer
1
τ= = 0.02s
50
Ts = 4τ = 0.08s
Tr = 2.2τ = 0.044s

4.3 Second order system

In this section, we shall define the second order system and the corresponding transient response
specifications. A general form of a second order system is:

d2 y dy
2
+ 2ξ ωn + ωn2 y = ωn2 u(t) (4.8)
dt dt

The correspondent transfer function can be written

ωn2
G(s) = (4.9)
s2 + 2ξ ωn s + ωn2

This form is called the standard form of the second-order system.

The second-order system can be described by the following block diagram:

U(s) ωn2 Y (s)

s2 + 2ξ ωn s + ωn2

Figure 4.4: Block diagram of second order system

The denominator polynomial of G(s), given by

D(s) = s2 + 2ξ ωn s + ωn2 (4.10)

is known as the characteristic polynomial of the system and D(s) = 0 is known as the characteristic
equation of the system.
Two important parameters widely used in characterizing the responses of second-order systems
are:

• ξ : the damping ratio of the system,

• ωn (rad/s) : the natural frequency of the system.

The damping ratio ξ represents the amount of damping in a system, whereas the natural frequency
ωn is a frequency of oscillations in an idealized system with zero damping.
60 Chapter 4. Time Response Analysis

1. ξ = 0, the roots of D(s) are imaginaries. The system is then called undamped , and
oscillations continue indefinitely,

2. If ξ > 1, the roots of D(s) are negative real and unequal. The system is then called over-
damped ,

3. If ξ = 1, the roots of D(s) are negative real and equal. The system is then called critically-
damped ,

4. If 0 < ξ < 1, the roots of D(s) are complex conjugates and lie in the left-half s plane. The
system is then called under-damped, and the transient response is oscillatory.

We shall now solve for the response of the second order system to a unit-step input.

4.3.1 Critically damped case ξ = 1

1
The poles of G(s) are equal. For a unit-step input, U(s) = and Y (s) can be written
s
ωn2
Y (s) = (4.11)
s(s + ωn )2

The output response may be found as

y(t) =1 − (1 + ωnt)e−ωnt (4.12)

4.3.2 Overdamped case ξ > 1

1
The poles of G(s) are negative real and unequal. For a unit-step input, U(s) = and Y (s) can be
s
written
ωn2
Y (s) = (4.13)
s(s + s1 )(s + s2 )
p p
where s1 = ξ ωn + ωn ξ 2 − 1 and s2 = ξ ωn − ωn ξ 2 − 1
The output response may be found as

ωn  e−s1t e−s2t 
y(t) =1 + p − (4.14)
2 ξ 2 − 1 s1 s2

4.3.3 Underdamped case 0 < ξ < 1

1
The poles of G(s) are complex. For a unit-step input, U(s) = and Y (s) can be written
s
ωn2
Y (s) = (4.15)
s(s + σd + jωd )(s + σd − jωd )
 4.3 Second order system 61

p
where σd = ξ ωn and ωd = ωn 1 − ξ 2.

The output response may be found as

e−σd −1
 p1 − ξ 2 
y(t) =1 − p sin(ωd t + φ ), φ = tan (4.16)
1−ξ2 ξ

A family of unit-step response curves y(t) with various values of ξ is shown in Figure 4.5. We
remark as ξ decreases, the roots approach the imaginary axis, and the response becomes increasingly
oscillatory.

y(t)
ξ=0

0.2

0.4
0.6
0.8
1.0

Figure 4.5: Plot of Unit-Step Response Curves with ωn = 1 and ξ = 0, 0.2, 0.4, 0.6, 0.8, 1
62 Chapter 4. Time Response Analysis

Value of ξ poles location step response

overdamped jω
ξ >1

σ
x x
t

critically damped jω
ξ =1

σ
x
t

underdamped jω
0<ξ <1 x

σ
t
x

undamped ξ = 0 jω
x

σ
t
x

unstable jω
ξ <0 t

σ
x x

Table 4.1: Second-order systems, pole plots, and step responses

 4.4 Transient system specifications for Second-Order Systems 63

Example 4.2
Consider the following transfer functions,
1 1 1
G1 (s) = , G2 (s) = , G3 (s) =
s2 + 8s + 12 s2 + 8s + 16 s2 + 8s + 20
find the value of ξ and report the kind of response expected.

• ξ = 1.155 for G1 , which is thus over-damped

• ξ = 1 for G2 , which is thus critically damped,

• ξ = 0.894 for G3 , which is thus under-damped.

4.4 Transient system specifications for Second-Order Systems

The under-damped second-order system, a common model for physical problems, displays unique
behaviour that must be itemized. For this system with a response described by equation (4.16), we
seek to determine the transient specifications.
Figure 4.6, depicts the poles of the under-damped second-order system. Using the Pythagorean
we can verify that

ξ = cos β (4.17)

jω
p
x jωd = jωn 1−ξ2
ωn

β σ
-σd = −ωn ξ

Figure 4.6: Pole plot for an under-damped second-order system

These transient specifications for under-damped second-order system are defined as follows:
64 Chapter 4. Time Response Analysis

1. Rise time:

1 ωd π −β
tr = tan−1 (− ) = (4.18)
ωd σd ωd

2. Peak time :

π π
tp = p = (4.19)
ωn 1 − ξ 2 ωd

3. Maximum overshoot :

−ξ π −πσd
p
1 − ξ 2
Os% = 100e = 100e ωd (4.20)

4. Settling time:

4 4
Ts = = (4.21)
ωn ξ σd

Figure 4.7 shows the maximum overshoot Os and the normalized peak time, Tp ωn versus the
damping ratio, ξ .

100 8

80 7

60 6
Tp ωn
Mp

40 5

20 4

0 3
0 0.2 0.4 0.6 0.8 1
ξ

Figure 4.7: Maximum overshoot Os and normalized peak time, Tp ωn versus ξ

 4.5 Steady-state error 65

4.5 Steady-state error

The steady-state error is considered as one of the important design specification. The steady-state
error reflects the accuracy of the system. In this section, we shall investigate the deviation of the
output from the desired reference caused by the incapability of the system to track particular types
of inputs.

4.5.1 Type characterization of Control Systems

. Control systems may be classified according to their ability to follow particular types of inputs.
Consider the following open-loop transfer function G(s):
K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)
G(s) = (4.22)
sN (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)
Transfer function (4.22) involves the term sN in the denominator, representing a pole of multiplicity
N at the origin. N indicates the number of integrations in the open-loop transfer function and
characterizes the type of the system.
A system is called

• type 0, if N = 0,

• type 1, if N = 1,

• type 2, if N = 2,

4.5.2 Error Constants

Let us consider the following feedback configuration of a single-input single-output system:

R(s) E(s) Y (s)

+
- G(s)

Figure 4.8: Feedback Form

From the block diagram given we have

E(s) = R(s) − G(s)E(s) (4.23)

Then the error signal E(s) is given by

1
E(s) = R(s) (4.24)
1 + G(s)
66 Chapter 4. Time Response Analysis

The steady state error component can be obtained by using the final value theorem of the Laplace
transform as

n sR(s) o
ess = lim e(t) = lim sE(s) = lim (4.25)
t−→∞ s−→0 s−→0 1 + G(s)

Equation 4.31 shows that the steady-state error is a function of the input R(s) and the open loop
transfer function G(s). Let us consider various standard test signals and obtain the steady-state
error for these inputs.

4.5.3 Static Position Error Constant Kp

1
For a unit step input, R(s) = . The steady-state error of the system is given by
s

n s 1o
ess = lim sE(s) lim
s−→0 s−→0 1 + G(s) s
1
=  (4.26)
1 + lim G(s)
s−→0
1
=
1 + G(0)
Defining the Static Position Error Constant as

Kp = lim G(s) = G(0) (4.27)

s−→0

Thus, the steady-state error in terms of the static position error constant Kp is given by

1
ess = (4.28)
1 + Kp

4.5.4 Static velocity Error Constant Kv

1
For a ramp input, R(s) = . The steady-state error of the system is given by
s2
n s 1o
ess = lim sE(s) lim
s−→0 s−→0 1 + G(s) s2
1 (4.29)
= 
lim sG(s)
s−→0

Defining the Static Velocity Error Constant as

Kv = lim sG(s) (4.30)

s−→0
 4.5 Steady-state error 67

Thus, the steady-state error in terms of the static velocity error constant Kv is given by

1
ess = (4.31)
Kv

4.5.5 Static acceleration Error Constant Ka

1
For a unit-parabolic input , R(s) = . The steady-state error of the system is given by
s3
n s 1o
ess = lim sE(s) lim
s−→0 s−→0 1 + G(s) s3
1 (4.32)
= 2
lim s G(s)
s−→0

Defining the Static Acceleration Error Constant as

Ka = lim s2 G(s) (4.33)

s−→0

Thus, the steady-state error in terms of the static acceleration error constant Ka is given by

1
ess = (4.34)
Ka

For the special case of unity of feedback system, H(s) = 1, error constants (4.27), (4.30) and (4.33)
are given by

Kp = lim G(s) (4.35)

s−→0

Kv = lim sG(s) (4.36)

s−→0

Ka = lim s2 G(s) (4.37)

s−→0

The steady state error, is dependent on the type of the system. We obtain the steady state error for
various standard test signals for type 0, type 1 and type 2 systems.
For a type 0 system,

K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kp = lim =K (4.38)
s−→0 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)
68 Chapter 4. Time Response Analysis

sK(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kv = lim =0 (4.39)
s−→0 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

s2 K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Ka = lim =0 (4.40)
s−→0 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

For a type 1 system,

K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kp = lim =∞ (4.41)
s−→0 s(τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

sK(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kv = lim =K (4.42)
s−→0 s(τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

s2 K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Ka = lim =0 (4.43)
s−→0 s(τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

For a type 2 system,

K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kp = lim =∞ (4.44)
s−→0 s2 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

sK(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Kv = lim =∞ (4.45)
s−→0 s2 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

s2 K(τz0 s + 1)(τz1 s + 1) · · · (τzm s + 1)

Ka = lim =K (4.46)
s−→0 s2 (τ p0 s + 1)(τ p1 s + 1) · · · (τ pn s + 1)

Table 4.2 summarizes the steady-state errors for type 0, type 1, and type 2 systems.

Standard input Type 0 Type 1 Type 2

step 1 0 0
1 + kp

Ramp ∞ 1 0
kv

Parabola ∞ ∞ 1
ka
Table 4.2: Steady state errors for various inputs and type of systems
 4.6 Problems 69

4.6 Problems

Exercise 4.1

The general form of a first-order system is described by:

dy(t)
T + y(t) = Ku(t) (4.47)
dt
where T is the time constant, K is the gain, u(t) and y(t) are the input and output of the
system respectively. The unit-step response of a first-order system is shown in the following
figure. Determine the parameters K and T from this figure.

0.5

0.4

0.3

0.2

0.1

0
0 2 4 6 8 10 12 14 16 18

Figure 4.9: First order response

Solution 4.1

• The transfer function of the system is given by

K
Y (s) K
G(s) = = = T1 (4.48)
U(s) (T s + 1) s + T
Then, the step response is
K
Y (s) = (4.49)
s(T s + 1)
According to figure 4.9, the final value is given by
K
y(∞) = lim s = 0.5 (4.50)
s−→0 s(T s + 1)
Thus K = 0.5.
70 Chapter 4. Time Response Analysis

• We know that y(τ) = 0.63y(∞) = 0.315. From the figure, we read τ = 2s. Thus,
T = τ = 2s

Exercise 4.2

Consider a second order system with transfer function:

361
G(s) = (4.51)
s2 + 16s + 361
Find ξ , ωn , Ts ,Tr (10 to 90 percent), Tp and percent overshoot (PO).

Solution 4.2

We have
√ 16
ωn = 361 = 19rad/s, ξ= = 0.42
2 × 19
p
ωd = ωn 1 − ξ 2 , σd = ωn ξ

So, we calculate
4
• Ts = = 0.5s.
σd
π
• Tp = = 0.18s.
ωd
πσd
−
• e ωd = 0.23.
ωd
• β = tan−1 ( ) = 1.136,
σd
π −β
• Tr = = 0.116.
ωd

Exercise 4.3

When the second-order system

Y (s) K
= 2 (4.52)
U(s) T s + s + K

is subjected to a unit-step input, the system output responds as shown in the following figure.
Determine K and T from the response curve.
 4.6 Problems 71

Solution 4.3

• The transfer can be written as the second order prototype as

K
K T
G(s) = = (4.53)
(T s2 + s + K) s2 + T1 s + KT

Then, we identify
r
K 1
ωn = , ξ= √ (4.54)
T 2 KT
Then, we have
p
ωd = ωn 1 − ξ 2, σd = ωn ξ (4.55)

Then, the step response is

K
T
Y (s) = (4.56)
s(s2 + T1 s + KT )

According to figure ??, the peak time is given by

π π
Tp = p = = 0.55 (4.57)
ωn 1 − ξ 2 ωd

Thus,
π
ωd = = 5.712 (4.58)
0.55
From the figure, we can determine the value of overshot Os = 0.35. Then, we have
−σd π
−Ln(Os)
Os = e ωd = 0.35 ⇔ σd = ωd = 1.91 (4.59)
q π
σd
ωn = σd2 + ωd2 = 6rad/s ⇔ ξ = = 0.32 (4.60)
ωn
Thus, we compute

T = 0.264 and K = 1.56 (4.61)

Exercise 4.4

Given the circuit of Figure 4.10, where C = 10µF, find R and L to yield 15% overshoot with
a settling time of 7ms for the capacitor voltage. The input, v(t), is a unit step.
72 Chapter 4. Time Response Analysis

L R

−
v(t) C
+

Figure 4.10: RLC circuit

Solution 4.4

The transfer function of the circuit is written as

1
1
Zc Cs LC
G(s) = = = (4.62)
1
Zc + ZL + ZR R + Cs + Ls s2 + R s + 1
L LC
So we have
r
1 R
ωn = , σd = ωn ξ = (4.63)
LC 2L
On the other hand we have
−σd π
σd π
Os = e ωd = 0.15 ⇔ = 1.9 (4.64)
ωd
4
Ts = = 0.007 (4.65)
σd
We can deduce that

ωd = 946.27rad/s, σd = 571.43rad/s, ωn = 1105.4rad/s (4.66)

Then, we obtain
1
ωn2 = ⇔ L = 0.081835H (4.67)
LC
R
= 2σd ⇔ R = 93.526Ω (4.68)
L

Exercise 4.5

For each pair of second-order system specifications that follow, find the location of the
second-order pair of poles.

1. %OS = 12%; Ts = 0.6s

2. %OS = 10%; Tp = 5s
 4.6 Problems 73

3. Ts = 7s; Tp = 3s

Solution 4.5

4
1. σd = = 6.667rad/s
Ts
σd π
− σd −Ln(Os)
Os = e ωd ⇔ = = 0.675 ⇔ ωd = 9.878rad/s.
ωd π
q
ωn = σd2 + ωd2 = 11.91rad/s
σd
ξ= = 0.6
ωn
π
2. ωd = = 0.628
Tp
σd π
− σd −Ln(Os)
Os = e ωd ⇔ = = ⇔ σd = 0.46
ωd π
q
ωn = σd2 + ωd2 = 0.779rad/s
σd
ξ= = 0.59.
ωn
4
3. σd = = 0.57rad/s
Ts
π
ωd = = 1.047rad/s
Tp
q
ωn = σd2 + ωd2 = 1.19rad/s
σd
ξ= = 0.48.
ωn

Exercise 4.6

Consider the control system whose closed loop poles are given in Fig 4.11.

jω
5

θ σ
−4 −2

−5

Figure 4.11: poles location

74 Chapter 4. Time Response Analysis

1. Find ξ , ωn , %OS, Tp , Ts and Tr .

Solution 4.6

From the figure we read the pole coordinates p1 = −3 + 7.5 j. So σd = 3, ωd = 7.5.

We calculate
4
• Ts = = 1.33s.
σd
π
• Tp = = 0.418s.
ωd
πσd
−
• Os = e ωd = 0.284.
ωd π −β
• β = tan−1 ( ) = 1.19, Tr = = 0.261,
σd ωd

Exercise 4.7

For the system shown in Fig 4.12, determine the values of the gain K and the velocity
feedback constant Kh so that the maximum overshoot in the unit-step response is 0.2 and the
peak time is 1 sec.
With these values of K and Kh , obtain the rise time and settling time. Assume that J = 1kg.m2
and B = 1Nm/rad/sec.

R(s) K 1 Y (s)
+ +
- -
Js + B s

Kh

Figure 4.12: Block diagram

Solution 4.7

The closed-loop transfer function is given by

K
K
G(s) = = 2 αJ (4.69)
Js + αs + K s + J s + KJ

where α = B + KKh .
 4.6 Problems 75

Then, we have
r
K α
ωn = , σd = ωn ξ = (4.70)
J 2J
Therefore, we have also
π
Tp == 1 ⇔ ωd = π (4.71)
ωd
−σd π
σd −Ln(0.2)
Os = e ωd ⇔ = = 0.512 (4.72)
ωd π
q
Thus σd = 1.608 and ωn = σd2 + ωd2 = 3.53rad/s. Then, we obtain K = Jωn2 = 12.45,
α = 2Jσd = 3.216 and Kh = 0.178. Figure 4.13 depicts the corresponding time response to
verify the performance requirement.

1.4

1.2

0.8

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5 4

Figure 4.13: Response output

Exercise 4.8

For the system shown in Fig 4.14, determine the values of K and b so that the damping ratio
1
is 0.707 and the steady state error is ess = .
20
76 Chapter 4. Time Response Analysis

R(s) K Y (s)
+
-
s(s + b)

Figure 4.14: Block diagram

Solution 4.8

Since the type of the system is 1 and the ess ids different to 0, then the input must be a ramp.
K K
Kv = lim s = (4.73)
s−→0 s(s + b) b

So we have
1 1
ess = = , ⇔ K = 20b (4.74)
Kv 20
Te closed-loop transfer function of the system is
K √
T (s) = , ⇒ ωn = K, 2ξ ωn = b (4.75)
s2 + bs + K
We can deduce that
√ K
1.414 K = , ⇔ K = 800, b = 40 (4.76)
20

Exercise 4.9

There is a fundamental trade-off between the steady-state value of the tracking error e(t)
and the initial peaking of the control input u(t), which is the main theme of this question.
Consider the unity feedback system in the following figure 4.15.

R(s) E(s) U(s) s+3 Y (s)

+
- K P(s) =
s2 − 1

Figure 4.15: Block diagram

E(s)
1. Find the transfer function from R(s) to E(s) (i.e. S(s) = ) and the transfer function
R(s)
U(s)
from R(s) to U(s) (i.e. G(s) = ).
R(s)
2. Find the range of K for which both S(s) andG(s) are stable, i.e. they have no poles in
 4.6 Problems 77

the right half-plane or on the imaginary axis.

3. Assuming that K is chosen to ensure closed-loop stability, find es s

4. Find u(0+ ) for a unit step reference input.

5. Assume that the control input is required to satisfy |u(+ )| ≤ 17. Choose K such that
the value of |ess | for a unit step reference input is as small as possible. Find ess for this
value of K.

Solution 4.9

1. The transfer function form R(s) to E(s) is obtained as

E(s) 1
S(s) = = (4.77)
R(s) 1 + KP(s)
s2 − 1
= 2 (4.78)
s + Ks + 3K − 1
The transfer function from R(s) to U(s) is then obtained as

U(s) E(s)
G(s) = =K (4.79)
R(s) R(s)
2
K(s − 1)
= 2 (4.80)
s + Ks + 3K − 1

2. In order to have a stable feedback system, the roots of s2 + Ks + 3K − 1 should be

in the left half plane. Using the Routh criterion the system is stable if K > 0 and
1
3K − 1>0. We hence obtain the range of K for a stable system as K >
3
3. We apply the final value theorem with the assumption that K is properly chosen to
ensure closed-loop stability. In this fashion, we obtain
1 1
ess = lim sS(s) = − (4.81)
s−→0 s 3K − 1

4. By applying the initial value theorem, we find

1
u(0+ ) = lim sG(s) = K (4.82)
s−→∞ s

5. We first note that K should be chosen to ensure closed-loop stability, i.e. K > 1/3.
The bound on |u(0+ )| = |K| then implies K ≤ 17. In conclusion, we need to choose
1
K ∈]1/3, 17]. For this range of K, we have |ess | = , which clearly decreases
3K − 1
with increasing K. To obtain minimum possible |ess | , we hence choose K = 17. This
leads to ess = −0.02
78 Chapter 4. Time Response Analysis

Exercise 4.10

Consider the system shown in Fig 4.16 with PI control.

D(s)
R(s) Kp s + 10 + 10 Y (s)
+ +
- 2
s s + s + 20

Figure 4.16: Block diagram for a PI control system

1. Determine the transfer function from R(s) to Y (s) .

2. Determine the transfer function from D(s) to Y (s) .

3. Use Routh’s criterion to find the range of Kp for which the system is stable.

4. Find the expression of E(s) versus D(s) and R(s).

5. Assume D(s) = 0, Find E(s) and calculate the steady state error for a ramp reference
tracking?

6. Assume R(s) = 0, Find E(s) and calculate the steady state error with respect to
unit-step disturbance rejection?

Solution 4.10

Kp s + 10 10
Set G1 (s) = and G2 (s) = 2
s s + s + 20
1. The transfer function from R(s) to Y (s) is
Y (s) 10(10 + Kp s)
T1 (s) = = (4.83)
R(s) s(s2 + s + 20) + 10(10 + Kp s)

2. The transfer function from D(s) to Y (s) is

Y (s) 10s
T2 (s) = = (4.84)
D(s) s(s2 + s + 20) + 10(10 + Kp s)

3. The characteristic equation is s3 + s2 + (10Kp + 20)s + 100 = 0. The Routh array is

For stability we must have Kp > 8.

4. The expression of E(s) is given by

1 G2 (s)
E(s) = R(s) − D(s) (4.85)
1 + G1 (s)G2 (s) 1 + G1 (s)G2 (s)
 4.6 Problems 79

1
5. Assume D(s) = 0 and R(s) = , We find .
s2
s(s2 + s + 20)
E(s) = R(s) (4.86)
s(s2 + s + 20) + 10(10 + K p s)

Then, we obtain

ess = lim sE(s) = 0.2 (4.87)

s−→0

1
6. Assume R(s) = 0 and D(s) = , We find .
s
10s
E(s) = − D(s) (4.88)
s(s2 + s + 20) + 10(10 + K p s)

Then, we obtain

ess = lim sE(s) = 0 (4.89)

s−→0
 CHAPTER
5 Stability of
linear system
Content

5.1 Introduction 81
5.2 Definition 81
5.2.1 Routh’s stability criterion 83
5.2.2 Stability Design via Routh’s Stability Criterion 87
5.3 Problems 89

5.1 Introduction
Stability is the most important system specification. If a system is unstable, transient response and
steady-state errors are moot points. An unstable system cannot be designed for a specific transient
response or steady-state error requirement.
A linear, time-invariant system is stable if the natural response approaches zero as time
approaches infinity.
A linear, time-invariant system is unstable if the natural response grows without bound as time
approaches infinity.
A linear, time-invariant system is marginally stable if the natural response neither decays nor
grows but remains constant or oscillates as time approaches infinity.

5.2 Definition

Definition 5.1: Stability

• A continuous system is stable if every bounded input produces a bounded output.

• A continuous system is unstable if every bounded input produces a unbounded

output.

81
82 Chapter 5. Stability of linear system

We note that the impulse response of a linear time-invariant continuous system is a sum of
exponential time functions whose exponents are the roots of the system characteristic equation.

• If the real parts of the roots of the characteristic equation has negative real parts (the poles
are within left half-plane), the system is stable,

• If the characteristic equation has some roots with negative real parts and one imaginary
root of multiplicity 1, the system is marginally stable,

• If the characteristic equation has some roots with positive real parts or roots with negative
real parts and imaginary roots of multiplicity greater than 1, the system is unstable.

Example 5.1

U(s) 1 Y (s)
(s + 2)(s2 + 0.5s + 1)

2
y(t)
0.6
1

0.4
−3 −2 −1 1

0.2 −1

t
−2
10 20 30 40

Figure 5.1: poles and response for stable system

Example 5.2

U(s) 1 Y (s)
(s + 2)(s2 − s + 1.25)
 5.2 Definition 83

·108
2
y(t)
1
1
0.5

−3 −2 −1
t
10 20 30 40 −1

−0.5
−2

Figure 5.2: poles and response for unstable system

5.2.1 Routh’s stability criterion

In linear control systems, the most important problem concerns stability. That is, under what
conditions the considered system become unstable? In other words, how many poles are in the left
half-plane, in the right half-plane, and on the jω-axis.

The Routh-Hurwitz stability criterion is considered as a powerful method that provides an

answer to the question of stability by considering the characteristic equation of the system. This
method determines how many roots of the characteristic equation lie in the right half of s-plane.

Let the characteristic equation be given by the following polynomial in s domain:

an sn + an−1 sn−1 + · · · + a1 s + a0 = 0 (5.1)

The Routh-Hurwitz criterion is based on ordering the coefficients of the characteristic equation into
an array according to the following pattern:
84 Chapter 5. Stability of linear system

sn an an−2 an−4 · · ·
sn−1 an−1 an−3 an−5 · · ·
an an−2 an an−4
an−1 an−3 an−1 an−5
sn−2 − = b1 − = b2 b3
an−1 an−1
an−1 an−3 an−1 an−5
b1 b2 b1 b3
sn−3 − = c1 − = c2 c3
b1 b1
b1 b2 b1 b3
c1 c2 c1 c3
sn−4 − = d1 − = d1 d3
c1 c1
··· ··· ···
s1 j1 ···
s0 k1 ···

Table 5.1: Routh table

This pattern is continued until the rest of the c’s are all equal to zero.

Definition 5.2: Routh Criterion

• Routh’s criterion states that the number of roots of the characteristic equation with
positive real parts is equal to the number of changes of sign of the coefficients in the
first column.

• The system is stable if the elements of the first column have the same sign.

Example 5.3
Find the conditions for second order system to be stable:

D(s) = a2 s2 + a1 s + a0 (5.2)

s2 a2 a0
s1 a1 0
s0 b1 0

a2 a0
a1 0
b1 = − = a0 (5.3)
a1
Therefore, second order system is stable if all three coefficients have the same sign
 5.2 Definition 85

Example 5.4
The characteristic equation of a given system is

s4 + s3 + 3s2 + 2s + 1 = 0 (5.4)

Find the range of K in order to ensure that the system is stable?

s4 1 3 1
s3 1 2 0
s2 1 1 0
s1 1 0
s0 1 0

From the table there are no sign changes in the first column. According to Routh’s criterion,
this means that system is stable.

Theorem 5.1: Division of a Row

The coefficients of any row may be multiplied or divided by a positive number without
changing the signs of the first column.

Example 5.5
Determine the stability of the following transfer function:
s+1
T (s) = (5.5)
s6 + 3s5 + 2s4 + 9s3 + 5s2 + 12s + 20

s6 1 2 5 20
s5 0
s4 -1 1 20 0
s3 0
s2 7 20
s1 22
s0 20

There are two changes of signs in the first column and, therefore, there are two roots with
positive real parts.

Theorem 5.2: Zero Coefficient in the First Column

When the first term in a row is zero but not all the other terms are zero, the following methods
can be used:
86 Chapter 5. Stability of linear system

1
1. Substitute s = in the original equation; then solve for the roots of x with positive
x
real parts. The number of roots x with positive real parts will be the same as the
number of s roots with positive real parts.

2. Multiply the original polynomial by the factor (s + 1), which introduces an additional
negative root. Then apply the Routh criterion for the new polynomial.

Example 5.6
Determine the stability of the following transfer function:
3
T (s) = (5.6)
s4 + s3 + 2s2 + 2s + 5

s4 1 2 5
s3 1 2
s2 0 5

The zero in the first column prevents completion of the array. The following methods
overcome this problem.
1
• Letting s = and rearranging the polynomial gives
x

D(x) = 5x4 + 2x3 + 2x2 + x + 1 = 0 (5.7)

s4 5 2 1
s3 2 1
s2 -1 2
s1 5
s0 2

There are two changes of sign in the first column. Therefore, there are two roots of x in the
right half plane. The number of roots s with positive real parts is also two.

Theorem 5.3: A Zero Row

When all the coefficients of one row are zero, the procedure is as follows:

1. The auxiliary equation can be formed from the preceding row,

2. The Routh table can be completed by replacing the all-zero row with the coefficients
obtained by differentiating the auxiliary equation.
 5.2 Definition 87

3. The roots of the auxiliary equation are also roots of the original equation. These
roots occur in pairs and are the negative of each other. Therefore, these roots may be
imaginary (complex conjugates) or real (one positive and one negative), may lie in
quadruplets (two pairs of complex-conjugate roots), etc.

Example 5.7
Determine the number of right-half-plane poles in the closed-loop transfer function
5
T (s) = (5.8)
s5 + 2s4 + 24s3 + 48s3 − 25s − 50

s5 1 24 -25
s4 2 48 -50
s3 0 0

The presence of a zero row (the s3 row) indicates that there are roots that are the negatives of
each other. The next step is to form the auxiliary equation from the preceding row, which is
the s4 row.

a(s) = 2s4 + 48s2 − 50 (5.9)

Using quadratic formula, roots of a(s) are s2 = 1, −25.

Hence the roots of a(s) are s = ±1, ± j5.
Next we differentiate the polynomial with respect to s and obtain

da(s)
= 8s3 + 96s = 0 (5.10)
ds

s5 1 24 -25
s4 2 48 -50
s3 8 96
s2 24 -50
s1 112.7 0
s0 -50

One sign change in first column indicates one root in right half plane.

5.2.2 Stability Design via Routh’s Stability Criterion

Routh’s criteria can be applied to a transfer function to determine whether the system is stable or
not. However it can also be used to study the effects of parameter variation on system stability.
88 Chapter 5. Stability of linear system

In the following example, we shall consider the problem of determining the stability range of a
parameter value.

Example 5.8
The characteristic equation of a given closed-loop system is

D(s) = s4 + 6s3 + 11s2 + 6s + K = 0 (5.11)

Find the range of K in order to ensure that the system is stable?

s4 1 11 K
s3 6 6 0
s2 10 K 0
60 − 6K
s1 0
10
s0 K 0

For the system to be stable, 60 - 6K > 0, and K > 0. Thus 0 < K < 10.
 5.3 Problems 89

5.3 Problems

Exercise 5.1

Determine the stability of the closed-loop transfer function

s2 + 3
G(s) = (5.12)
s5 + s4 + 3s3 + 3s2 + 6s + 4

Solution 5.1

D(s) = s5 + s4 + 3s3 + 3s2 + 6s + 4 = 0 (5.13)

s5 1 3 6 0
s4 1 3 4 0
s3 0 2 0 0

Table 5.2: Routh table

We have a row with a zero first column Then consider the following characteristic equation
1
with s =
x

D(x) = 4x5 + 6x4 + 3x3 + 3x2 + x + 1 = 0 (5.14)

x5 4 3 1
x4 6 3 1
x3 6 2 0
x2 1 1 0
x1 -4 0 0
x0 1

Table 5.3: Routh table

Two signs change than two poles in the right half plane.

Exercise 5.2

Determine the stability of the closed-loop transfer function

s + 10
G(s) = (5.15)
s6 + s5 + 6s4 + 5s3 + 10s2 + 5s + 5
90 Chapter 5. Stability of linear system

Solution 5.2

D(s) = s6 + s5 + 6s4 + 5s3 + 10s2 + 5s + 5 = 0 (5.16)

s6 1 6 10 5
s5 1 5 5 0
s4 1 5 5 0
s3 0 0 0 0
s3 4 10 0 0
s2 5 10 0 0
s1 2 0 0 0
s0 2 0 0 0

Table 5.4: Routh table

The auxiliary polynomial is

A(s) = s4 + 5s2 + 5 = 0 (5.17)

Then
dA(s)
= 4s3 + 10s (5.18)
d(s)

The zero coefficients in row s3 should be replaced by 4 and 10. The roots of A(s) are
√ √
−5 − 5 −5 + 5
s1,2 = ± j and s3,4 = ± j (5.19)
2 2

Exercise 5.3

Write the characteristic equation and construct the Routh array for the control system shown
in Fig 5.3 Is the system stable for (a) K = 9.5, (b) K = 11, (c) K = 12?

R(s) 1 Y (s)
+
- K
(s + 1)(0.5s + 1)

3
s+3

Figure 5.3: Block diagram

 5.3 Problems 91

Solution 5.3

The closed-loop transfer function is given by

K(s + 3)
G(s) = (5.20)
(s + 3)(0.5s + 1)(s + 1) + 3K
2K(s + 3)
= 3 (5.21)
s + 6s2 + 11s + (6 + 6K)

The characteristic equation is

D(α) = s3 + 6s2 + 11s + α = 0, α = (6 + 6K) (5.22)

s3 1 11

s2 6 α
66 − α
s1 0
6
s0 α

Table 5.5: Routh table

The system is table if

66 − α > 0, and α >0 (5.23)

(5.24)

then, we obtain

0 < K < 10 (5.25)

Exercise 5.4

k1 s + k2 1
Consider a PI controller C(s) = for an unstable transfer function G(s) = 2
s s +s−1
in the feedback system in Figure 5.4.

R(s) Y (s)
+
- C(s) G(s)

Figure 5.4: Block diagram of PI control system

1. Find the range (k1 , k2 ) for which the closed-loop system is stable. Draw the range in
92 Chapter 5. Stability of linear system

(k1 , k2 ) axis.

2. For values (k1 = 3, k2 = 1) and a unit step input, compute the steady state error.

Solution 5.4

1. The closed-loop transfer function is given by

Ks + K2
G(s) = 2 (5.26)
s(s + s − 1) + K1 s + K2
The characteristic equation is

D(α) = s3 + s2 + αs + K2 = 0, α = (K1 − 1) (5.27)

s3 1 α
s2 1 K2
s1 α − K2 0
s0 K2

Table 5.6: Routh table

The system is table if K2 > 0, and K1 − K2 − 1 < 0

K2
4

K1
2 4

2. For k1 = 3 and k2 = 1 the system is stable.

The open-loop transfer function is
3s + 1
T (s) = C(s)G(s = (5.28)
s(s2 + s − 1)
Then we have
1
Kp = lim T (s) = ∞, ⇒ ess = =0 (5.29)
s−→0 1 + Kp
 CHAPTER
6 Root Locus
Technique
Content

6.1 The Root Locus Concept 93

6.2 Angle and magnitude criteria 95
6.3 Construction Rules 96
6.3.1 Number of branches 96
6.3.2 Symmetry 96
6.3.3 Starting and ending points 96
6.3.4 Real axis loci 96
6.3.5 Asymptotes 96
6.3.6 Breakaway point 96
6.3.7 Imaginary axis crossing 97
6.3.8 Departure and arrival angles 97
6.4 Problems 101

The relative stability and the transient performance of a closed-loop system is closely related to
the location of the closed-loop poles. If the system has a variable loop gain, then the location of the
closed-loop poles depends on the value of the loop gain chosen.
It is important to know how the closed-loop poles move in the s-plane as the loop gain is varied.
From the designer viewpoint, in some systems simple gain adjustments may move the closed-loop
poles to the desired locations. Then the design problem may become the selection of an appropriate
gain value.
The root locus technique is a graphical method for sketching the locus of roots in the s-plane as
a parameter is varied and has been utilized extensively in control engineering practice. The root
locus technique may be used to great advantage in conjunction with the Routh criterion.
Using the method developed by W. R. Evans, we will present the general procedure for sketching
root loci using illustrative examples.

6.1 The Root Locus Concept

The situation of interest is illustrated in Figure 6.1, where G(s) is the transfer function of the linear
system and K is the open-loop gain.

93
94 Chapter 6. Root Locus Technique

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

Figure 6.1: Unit feedback system

The standard root locus only applies to the case where G(s) is a rational function of s. That is,

N(s) Πm (s + zi )
G(s) = = ni=1 (6.1)
D(s) Π j=1 (s + p j )

The closed loop transfer function is given by

KG(s) KN(s)
T (s) = = (6.2)
1 + KG(s) D(s) + KN(s)

Note that G(s) and of T (s) have the same numerators. Since the denominator of T (s) depends on
the real K, the root locus answer to the question: what happens to the poles of T (s) as K varies? By
definition, the root locus plot is a plot of the poles of T (s) in the complex plane as the parameter K
varies.
Consider the following example with

1
G(s) = (6.3)
s2 + 2s

then

1
T (s) = (6.4)
s2 + 2s + K

The poles of T (s), denoted by s1 and s2 , are given by

√ √
s1 = −1 − 1 − K, s2 = −1 + 1 − K (6.5)

Figure 6.2 plots the poles s1 and s2 in the complex plane as K varies. The root locus gives the pole
locations for every possible closed-loop system that can be created from the open-loop plant and
any positive gain K.
 6.2 Angle and magnitude criteria 95

10

Imag Axis
0

−5

−10
−3 −2 −1 0 1
Real Axis

Figure 6.2: Root locus

6.2 Angle and magnitude criteria

From the closed loop transfer function (6.2) the characteristic equation is defined as

1 + KG(s) = 0 (6.6)

or

KG(s) = −1 (6.7)

In polar form,

KG(s) = |KG(s)|∠KG(s) = 1∠180(2q + 1), q = 0, ±1 · · · (6.8)

In order for a branch of a root-locus to pass through a particular point s0 in the complex plane, it is
necessary that s0 be a root of the characteristic equation (6.7) for some real value of K.
If s0 satisfies equation (6.7), then it necessarily obeys to:

magnitude criterion:

1
|G(s0 )| = (6.9)
K

angle criterion:

∠G(s0 ) = (2q + 1)180 (6.10)

96 Chapter 6. Root Locus Technique

6.3 Construction Rules

6.3.1 Number of branches

The number of loci, that is, the number of branches of the root locus is equal to the number of poles
of the open-loop transfer function.

6.3.2 Symmetry

The root locus is symmetrical about the real axis.

6.3.3 Starting and ending points

The root locus starts (K = 0) at the open-loop poles and terminates (K = ±∞) at the open-loop
zeros or at infinity.

6.3.4 Real axis loci

On the real axis, for K > 0 the root locus lies to the left of an odd number of finite poles and zeros.

6.3.5 Asymptotes

The root locus approaches straight lines as asymptotes as the locus approaches infinity. Further, the
equation of the asymptotes is given by the real-axis intercept, σa and angle, θa as follows:

∑ni=1 pi − ∑m
i=1 zi
σa =
n−m
180(2q + 1)
θa =
n−m

6.3.6 Breakaway point

Two root locus branches starting from two open loop poles on real axis for K = 0, meet at a point
on the real axis for some particular value of K = Kb . At this value of K, the two closed loop poles
will be equal. Such points on root locus where multiple closed loop poles occur, are known as
breakaway points. The root locus branches meet at this point and for values of K > Kb , the closed
loop poles become complex. The breakaway point can be determined by taking the derivative of the
loop sensitivity K with respect to s. Equate this derivative to zero and find the roots of the resulting
equation.
N(s)
• Express K as K = −
D(s)
 6.3 Construction Rules 97

dK
• Find the root of the equation =0
ds

6.3.7 Imaginary axis crossing

If a root locus branch crosses the imaginary axis, the cross over point can be obtained by setting up
the Routh-Hurwitz table from the closed-loop characteristic equation.

6.3.8 Departure and arrival angles

1. When there are complex poles of G(s), the root locus branches will depart from an angle θd
for a complex pole s = pc

∠(s + pc )G(s) − θd = 180◦ , for s = pc (6.11)

2. When there are complex zeros of G(s), the root locus branches will arrive at an angle θa for
a complex zero s = zc

G(s)
∠ + θa = 180◦ , for s = zc (6.12)
(s + zc )

Example 6.1
Consider a closed-loop system as the one shown in Figure 6.1, where the open-loop transfer
function:
K
G(s) = (6.13)
s(s + 4)(s + 10)

Let us draw the root-locus plot and determine the value of K so that the damping factor ξ of
a pair of dominant complex-conjugate closed-loop poles is 0.5.
Using the previous procedure we obtain:

1. The open-loop transfer function has no zeros, thus m = 0, and three poles, p1 = 0,
p2 = −4, p3 = −10, thus n = 3.

2. Locate the open-loop poles of the open-loop transfer function with symbol "x", as
shown in Figure 6.3.

3. Locus has n = 3 branches.

4. Root locus on real axis lies between [−4, 0] and ] − ∞, 10].

5. The root locus is symmetrical with respect to the horizontal real axis.

6. The locus proceed to the zeros at infinity along asymptotes centered at σa and with
98 Chapter 6. Root Locus Technique

angles θa
ΣP − ΣZ 0 − 4 − 10 −14
σa = = =
n−m 3−0 3
(2k + 1).180◦
θa = ± = ±60, 180,
3−0
Thus, there are three asymptotes. The one having the angle 180 is the negative real
axis. The asymptotes are shown in Figure 6.3.

7. Using Routh criterion, we can determine the point at which the locus crosses the
imaginary axis. Given the characteristic equation as

s3 + 14s2 + 40s + K = 0

and the Routh array:

s3 1 40

s2 14 K
(456 − K)
s1 0
14
s0 K

The crossing points on the imaginary axis that makes the s1 term in the first column
(456 − K)
equal zero is obtained from = 0, or K = 456. The auxiliary equation is
14
14s2 + 456 = 0

Then the roots of the previous equation are the crossing points on the imaginary axis

s1,2 = ± j14

8. The breakaway point on the real axis will be determined from the solution of

K = −s(s + 4)(s + 10)

then
dK
= 3s2 + 28s + 40 = 0
ds
yields

s1 = −7.57, s2 = −1.76

Since the breakaway point must lie between 0 and -4, s2 = −1.76 corresponds to the
actual breakaway point.
 6.3 Construction Rules 99

Based on the information obtained in the foregoing steps, the root locus obtained is shown in
Figure 6.3.

15

10

5
Imag Axis

−5

−10

−15
−14 −12 −10 −8 −6 −4 −2 0 2 4
Real Axis

Figure 6.3: Root locus

Example 6.2
A feedback control system has an open-loop transfer function
K
G(s) = (6.14)
s(s + 3)(s2 + 2s + 2)

sketch the root-locus.

1. The root locus is symmetrical about the real axis. The number of poles is n = 4 and
the number of zeros is m = 0.

2. The number of root locus branches equals the number of open-loop poles. Therefore
in this case, the number of root locus branches is equal to 4.
The open-loop poles are s1 = 0, s2 = −3 and s3,4 = −1 ± j,

3. Each branch of the root locus starts from the open-loop pole at K = 0, and terminates
at K = ∞, either on an open-loop zero or infinity. Since the system given does not
have any zero, all the branches terminate at infinity.

4. Root locus on real axis lies between [−3 0].

5. The angle and the point of intersection of the asymptotes for the system defined are
100 Chapter 6. Root Locus Technique

given by
0−3−1−1
σa = = −1.25
4
180(2q + 1)
θa = = 45, 135, 225, 315
4
6. The breakaway points of the root locus are determined from the roots of the equation
dK
= 0.
ds
The characteristic equation of the system is given by

s(s + 3)(s2 + 2s + 2) + K = 0

Then
dK
= −4(s3 + 3.75s2 + 4s + 1.5) = 0
ds
s = −2.3, s = −0.73 ± j0.35

A breakaway point must occur at s = −2.3 as this part of the real axis is on the root
locus and the two root locus branches starting at s =0 and s = −3 are approaching
each other.

7. The angle of departure from the open-loop pole is obtained by subtracting the sum of
the angles subtended by the zero with the complex pole from the sum of the angles
subtended by all the open loop poles and then adding the result to 180.
In this case, the angle of departure from the open loop pole −1 + j is θ = −71.6.

8. To determine the value of K at which the root locus cuts the imaginary axis.
The characteristic equation of the system is given by

s4 + 5s3 + 8s2 + 6s + K = 0

Building the Routh table, we get

s4 1 8 K

s3 5 6 0

s2 34/5 K 0

s1 6 − 25K/34 0 0

s0 K

Therefore the value of K at which the root locus cuts the imaginary axis is K = 8.16
and the crossing points on the imaginary axis s1,2 = ±1.09 j
 6.4 Problems 101

6.4 Problems

Exercise 6.1

Consider the following feedback system, where k is the feedback gain, Fig 6.4.

R(s) k Y (s)
+
-
s(s2 + 6s + 25)

Figure 6.4: Block diagram

1. Determine the open loop transfer function of the system.

2. Determine the characteristic equation of the system.

3. How many branches are there in this system’s root locus ?

4. How many branches are there in real-axis root locus?

5. How many asymptotes are there in this system’s root locus ? Find the asymptote
angles and the point (centroid) where the asymptotes intercept the real axis

6. Determine the angle of departure from the complex pole in the upper half s plane.

7. Sketch the root locus based on the information from the previous question.

Solution 6.1

1. The open loop transfer function is

k
G(s) = (6.15)
s(s2 + 6s + 25)

2. The characteristic equation is

s3 + 6s2 + 25s + k (6.16)

3. The open-loop transfer function has only n = 3 poles, p1 = 0, p2,3 = −3 ± 4 j, thus

the root locus has 3 branches.

4. We draw the locus on the real axis to the left of and odd number of poles plus zeros.
102 Chapter 6. Root Locus Technique

5. Draw the asymptotes centred at σa and with angles θa

ΣP − ΣZ −6
σa = = = −2 (6.17)
n−m 3−0
(2q + 1).180◦
θa = = 60(2q + 1), (6.18)
3
= ±60◦ , 180◦ (6.19)

6. The angle of departure at the complex pole p3 = −3 + 4 j

φ3 = −180 − +∠(−3 + 4 j) − ∠(8 j)

= −180 − 126.87 − 90 = −36.87

7. Using Routh-Hurwitz criterion, we can determine the point at which the locus crosses
the imaginary axis. Given the characteristic equation 6.21, we construct the Routh
array as

s3 1 25 0

s2 6 k 0

s1 25 − k/6 0

s0 k

The crossing points on the imaginary axis that makes the s1 term in the first column
k
equal zero is obtained from 25 − = 0, or k = 150.
6
The auxiliary equation is

6s2 + 150 = 0

Then the roots of the previous equation are the crossing points on the imaginary axis
s1,2 = ± j5
 6.4 Problems 103

K>0
6
j5(k=150)
4

Imag Axis
0

−2

−4
-j5(k=150
−6
−5 −4 −3 −2 −1 0 1 2
Real Axis

Figure 6.5: Root locus

Exercise 6.2

Consider the following feedback system, where k is the feedback gain, Fig 6.6.

R(s) k Y (s)
+
-
(s + 2)(s + 4)

s + 20
s + 10

Figure 6.6: Block diagram

1. Determine the open loop transfer function of the system.

2. Determine the characteristic equation of the system.

3. How many branches are there in this system’s root locus ?

4. How many branches are there in real-axis root locus?

5. How many asymptotes are there in this system’s root locus ? Find the asymptote
angles and the point (centroid) where the asymptotes intercept the real axis

6. Determine the breakaway points of the root locus.

7. Sketch the root locus based on the information from the previous question.
104 Chapter 6. Root Locus Technique

8. Find the value of K where the root locus crosses the imaginary axis.

9. Find the maximum value of K to make all poles real.

Solution 6.2

1. The open loop transfer function is

k(s + 20)
G(s) = (6.20)
(s + 2)(s + 4)(s + 10)

2. The characteristic equation is

(s + 2)(s + 4)(s + 10) + k(s + 20) = s3 + 16s2 + (k + 68)s + (20k + 80) = 0 (6.21)

3. The open-loop transfer function has one zeros z1 = −20, thus m = 1, and n = 3 poles,
p1 = −2, p2 = −4, p3 = −10, thus the root locus has n = 3 branches.

4. We draw the locus on the real axis to the left of and odd number of poles plus zeros.

5. Draw the asymptotes centered at σa and with angles θa

ΣP − ΣZ −2 − 4 − 10 + 20
σa = = =2
n−m 3−1
(2q + 1).180◦
θa = = ±90,
3−1

6. The breakaway point on the real axis will be determined from the solution of

(s + 2)(s + 4)(s + 10) (−s3 − 16s2 − 68s − 80)

K=− =
(s + 20) (s + 20)

then
dK (2s3 + 76s2 + 640s + 1280)
=−
ds (s + 20)2

yields

s1 = −8, s2 = −2.95, s3 = −27.04

7. Using Routh-Hurwitz criterion, we can determine the point at which the locus crosses
the imaginary axis. Given the characteristic equation 6.21, we construct the Routh
array as
 6.4 Problems 105

s3 1 k + 68 0

s2 16 20k + 80 0
k
s1 63 − 0
4
s0 20k + 80

The crossing points on the imaginary axis that makes the s1 term in the first column
k
equal zero is obtained from 63 − = 0, or k = 252.
4
The auxiliary equation is

16s2 + 5120 = 0

Then the roots of the previous equation are the crossing points on the imaginary axis

s1,2 = ± j17.89

K>0
30

20
j17.9(k=252)

10
Imag Axis

−10

-j17.9(k=252)
−20

−30
−20 −15 −10 −5 0 5
Real Axis

Figure 6.7: Root locus

Exercise 6.3

Consider the following feedback system, where k is the feedback gain, Fig 6.8.
106 Chapter 6. Root Locus Technique

R(s) k Y (s)
+
-
s(s + 2)(s + 4)(s + 5)

Figure 6.8: Block diagram

Sketch the root locus.

Solution 6.3

1. Number of poles = 4, number of zeros = 0.

2. Open-Loop poles at 0, -2, -4, -5.

3. Real Axis Loci: [0 − 2] , [−4, −5]

4. Asymptote Angles ±45 , ±135

0−2−4−5
5. Asymptote Intersect: = −2.75
4
6. Breakaway Point: s = {−0.7438, −4.5771}

7. The crossing points on the imaginary axis: s = j1.9069

15

10

5
Imag Axis

−5

−10

−15
−10 −8 −6 −4 −2 0 2
Real Axis

Figure 6.9: Root locus

 6.4 Problems 107

Exercise 6.4

A feedback control system has an open-loop transfer function

K(s + 2)
G(s) = (6.22)
(s + 3)(s2 + 2s + 2)

sketch the root-locus.

Solution 6.4

1. The root locus is symmetrical about the real axis. The number of poles is n = 3 and
the number of zeros is m = 1.

2. The number of root locus branches equals the number of open-loop poles. Therefore
in this case, the number of root locus branches is equal to 3.
The open-loop poles are s1 = −3 and s2,3 = −1 ± j and zero z1 = −2,

3. Each branch of the root locus starts from the open-loop pole at K = 0, and terminates
at K = ∞. The closed-loop poles start at the open-loop poles and terminate at the
open-loop zero.

4. Segments of the real axis having an odd number of real axis open-loop poles plus
zeros to their right are parts of the root locus. Therefore the root locus has one segment
at [−3 − 2]

5. The angle and the point of intersection of the asymptotes for the system defined are
given by
−2 − 3 − 1 − 1 + 2
σa = = −2.5
3−1
180(2q + 1)
θa = = 90, −90
2

6. The breakaway points of the root locus are determined from the roots of the equation
dK
= 0.
ds
The characteristic equation of the system is given by

(s + 3)(s2 + 2s + 2) + K(s + 2) = 0

Then
dK
= −(2s3 + 11s2 + 20s + 10) = 0
ds
s = −0.8, s = −2.35 ± j0.8447

No breakaway points.
108 Chapter 6. Root Locus Technique

7. The angle of departure from the open-loop pole is obtained by subtracting the sum of
the angles subtended by the zero with the complex pole from the sum of the angles
subtended by all the open loop poles and then adding the result to 180.
In this case, the angle of departure from the open loop pole −1 + j is θ = −71.6.

8. To determine the value of K at which the root locus cuts the imaginary axis.
The characteristic equation of the system is given by

s4 + 5s3 + 8s2 + 6s + K = 0

Building the Routh table, we get

Therefore the value of K at which the root locus cuts the imaginary axis is K = 8.16.

K>0
6

2
Imag Axis

−2

−4

−6
−5 −4 −3 −2 −1 0 1
Real Axis

Figure 6.10: Root locus

Exercise 6.5

The laser can be used to drill the hip socket for the appropriate insertion of an artificial hip
joint. The use of lasers for surgery requires high accurcy for position and velocity. The
following figure shows the feedback system for a DC motor manipulator for a laser. The
amplifier gain K must be adjustable so that the steady state error for a ramp input, r(t) = t,
is less than 0.1mm, while a stable response is maintained. The parameters of the model are
given as τ1 = 0.1s and τ2 = 0.2s
 6.4 Problems 109

R(s)+ 1 Y (s)
+
- K
s(τ1 s + 1)(τ2 s + 1)

Figure 6.11:

1. Find the value of K required to achieve an ess = 0.1 for a ramp input.

2. Write the range of K such that the closed-loop system is stable.

3. Sketch the root locus of the system,

4. Determine the gain K and the point at which the locus crosses the imaginary axis. Is it
required to choose this gain, why?

Solution 6.5

1. The open-loop transfer function of the system is written as

50K
G0 (s) = (6.23)
s(s + 5)(s + 10)

2. The input is a ramp, we compute

50K
Kv = lim s =K (6.24)
s−→0 s(s + 5)(s + 10)

the steady-state error is calculated

1 1
ess = = = 0.1, ⇔ K = 10 (6.25)
Kv K

3. Consider the following table

s3 1 50

s2 15 50k
10K
s1 50 − 0
3
s0 50K

Thus, according to Routh criterion, the system is stable for K < 15.
110 Chapter 6. Root Locus Technique

4. The auxiliary equation is

15s2 + 750 = 0

Then the roots of the previous equation are the crossing points on the imaginary axis

s1,2 = ± j7
 CHAPTER The Frequency-
7 response
analysis
Content

7.1 Frequency Response 111

7.2 Frequency Response Characteristics 114
7.3 Plotting Frequency Response 114
7.3.1 Bode plot 114
7.3.2 Nyquist Plots 117
7.3.3 Relative stability: Nyquist criterion 118
7.4 Frequency-Domain Performance Specifications 122
7.4.1 Damping ratio and frequency response 122
7.4.2 Damping ratio and phase margin 123
7.4.3 Response speed and frequency response 123
7.5 Problems 124

For analysis of linear systems, two methods can be conducted. The first one is based on time
domain when a time signal like step and ramp are used to excite the system and its time response is
obtained. The second one, is based on frequency domain when the system is excited by a sinusoidal
signal of variable frequency and the magnitude and phase of the steady state output from the system
is measured.
The chapter discusses of how to obtain the frequency response of a linear system by analyzing
its poles and zeros. This discussion deals how to use Bode plots to graphically display the frequency
response.
We will discuss briefly the stability using the Nyquist stability criterion and we will introduce
the notion of stability margins.

7.1 Frequency Response

This section introduces the concept of frequency response used in control system design.

Definition 7.1: Frequency response

111
112 Chapter 7. The Frequency-response analysis

The frequency response of a system is defined as the steady-state response of the system to a
sinusoidal input signal.

u(t) = Asin(ωt) y(t) = Bsin(ωt + φ )

G(s)

where the input u(t) is a sine wave with an amplitude A. This sine wave has a Laplace transform

Aω
U(s) = (7.1)
s2 + ω 2

With zero initial conditions, the Laplace transform of the output is

Aω
Y (s) = G(s) (7.2)
s2 + ω 2

By evaluating G( jω) at s = jω, the following complex number is obtained which can be represented
in the polar form

G( jω) = |G( jω)|e jφ (ω) , φ (ω) = ∠G( jω) (7.3)

The function G( jω) is called the sinusoidal transfer function.

q
|G( jω)| = Re(G( jω))2 + Im(G( jω))2 (7.4)
 Im(G( jω)) 
φ = ∠G( jω) = tan−1 (7.5)
Re(G( jω))

The steady state response that corresponds to Y (s) is written as

yss (t) = A|G( jω)|sin(ωt + φ (ω)) (7.6)

Thus, the output has the same frequency, the amplitude is multiplied by a factor |G( jω)| and the
phase is shifted by a quantity ∠G( jω). It will also be shown that the steady-state response can be
given by

Y ( jω) = B(ω)e jφ (ω) = B(ω)∠φ (ω) (7.7)

 7.1 Frequency Response 113

Example 7.1
Consider a first order system given by
K
G(s) = (7.8)
Ts+1
Substituting s = jω into 7.8, we get
K
G( jω) = (7.9)
jT ω + 1
The magnitude is
K
B = |G( jω)| = √ (7.10)
1 + T 2ω 2
while the phase angle φ is

φ = ∠G( jω) = −tan−1 (T ω) (7.11)

Example 7.2
The basic equation of motion for the vertical motion of the system is

d2 x dx
m 2
+ b + kx = Fm , (7.12)
dt dt
where force Fm is a sinusoidal function of time Fm = 400sin(104.7t).
The transfer function relating the displacement to force Fm is obtained from equation 7.12,
assuming zero initial conditions:

X(s) 1
G(s) = = 2 (7.13)
Fm (s) ms + bs + k

One obtains the frequency-response transfer function of the system by putting s = jω in

equation 7.13:
1
G( jω) = (7.14)
k − mω 2 + jbω
For the given values of the system parameters and the frequency of vibration, ω =
104.7rad/s, the frequency-response transfer function is

1
G( jω) = (7.15)
−224.052 + j83.760

Hence the magnitude of the transfer function at the frequency of interest is |G( jω)| =
4.18.10−6 m/N. The phase angle of the vertical displacement is equal to the phase angle of
83.76
the frequency-response transfer function Φ = −π − tan−1 ( ) = −2.78rad.
224.05
114 Chapter 7. The Frequency-response analysis

7.2 Frequency Response Characteristics

Definition 7.2: System Bandwidth

This represents the frequency range in which the magnitude of the closed-loop frequency
transfer function drops no more than (3 decibels) from its zero-frequency value. The system
bandwidth can be obtained from the next equality, which indicates the attenuation of , as
1
G( jωBW ) = √ |G(0)| (7.16)
2

Definition 7.3: Peak Resonance

This is obtained by finding the maximum of the function |G( jω)| with respect to frequency
. It is interesting to point out that the systems having large maximum overshoot have also
large peak resonance.

Definition 7.4: Resonant Frequency

This is the frequency at which the peak resonance occurs. It can be obtained from

d|G( jω)|
=0 (7.17)
dω

7.3 Plotting Frequency Response

Several kinds of plots are used for representing a system frequency response G( jω); two of them
are

• as a function of frequency, with separate magnitude and phase plots,

• as a polar plot, where the phasor length is the magnitude and the phasor angle is the phase.

7.3.1 Bode plot

Bode plots consist of two graphs:

1. the magnitude |G( jω)| versus the frequency ω,

2. the phase angle ∠G( jω) versus the frequency ω,

The logarithmic scales are usually used for the frequency axes. The logarithmic magnitude of
G( jω) is 20log|G( jω)|, where the base of the logarithm is 10.
In the logarithmic representation, the curves are drawn on semi-log paper, using the log scale
for frequency and the linear scale for either magnitude (but in decibels) or phase angle (in degrees).
 7.3 Plotting Frequency Response 115

Bode plot of first order system

The system transfer function is

a 1
G(s) = = s (7.18)
s+a a +1

The frequency response is

1
G( jω) = jω (7.19)
1+ a

The logarithmic magnitude is

1 ω
20log|G( jω)| = 20log q = −10log(1 + ( )2 ) (7.20)
ω 2
1+( a ) a

The asymptotic behavior for ω ≪ a and ω ≫ a can be found as




 0 ω ≪a

|G( jω)|db = −20logω ω ≫ a (7.21)



 −3db ω = a

The phase of the transfer function is given by

ω
∠G( jω) = −tan−1 (7.22)
a

For ω ≪ a and ω ≫ a we have




 0 ω ≪a

∠G( jω) = −90 ω ≫ a (7.23)



 −45 ω = a

For the magnitude, we see that for small ω the magnitude is very close to zero. For large values of
ω, we obtain an attenuation of 20dB/decade . For small and large frequencies we have straight
line approximations. These straight lines intersect at ω = a , which is also known as a corner
frequency. The actual magnitude curve is below the straight line approximations. It has the biggest
deviation from the asymptotes at the corner frequency.
For the phase diagram, it can be seen that for large values of ω, the phase contribution is −90°.
For small, ω, the phase is close to zero, and for the phase contribution is −45°.
116 Chapter 7. The Frequency-response analysis

Magnitude (dB) −10

−20

−30

−40
0
Phase (deg)

−45

−90
10−1 100 101 102 103

Figure 7.1: Bode plot for first order system

Bode plot of second order system

The system transfer function is

ωn2 1
G(s) = 2 2
= s s (7.24)
s + 2ξ ωn s + ωn ( )2 + 2ξ +1
ωn ωn
The frequency response is
1
G( jω) = (7.25)
ω ξω
1 − ( )2 + 2 j
ωn ωn
The logarithmic magnitude is
1
20log|G( jω)| = 20log r (7.26)
ω ξω 2
(1 − ( )2 )2 + (2 )
ωn ωn
ω ξω 2
= −10log[(1 − ( )2 )2 + (2 ) ] (7.27)
ωn ωn
The asymptotic behavior for ω ≪ ωn and ω ≫ ωn can be found as




 0 ω ≪ ωn
 ω
|G( jω)|db = −40log ω ≫ ωn (7.28)

 ωn

 −20log(2ξ ) ω = ωn

 7.3 Plotting Frequency Response 117

The phase of the transfer function is given by

ω
 2ξ 
ωn
∠G( jω) = −tan−1 ω (7.29)
1 − ( )2
ωn

For ω ≪ ωn and ω ≫ ωn we have




 0 ω ≪ ωn

∠G( jω) = −180 ω ≫ ωn (7.30)



 −90 ω = ω
n

We find that the magnitude curve breaks at the natural frequency and decreases at a rate of
−40dB/decade. The phase plot is 0° at low frequencies.

Bode Diagram

0
Magnitude (dB)

−50

−100
0

−45
Phase (deg)

−90

−135

−180
10−2 10−1 100 101 102

Figure 7.2: Bode plot for second order system

7.3.2 Nyquist Plots

Nyquist plot of a sinusoidal transfer function G( jω) is a plot of the magnitude of G( jω) versus
the phase angle of G( jω) on polar coordinates as ω is varied from zero to infinity. This curve is
known also as polar plot of the given transfer function. It is useful in determining the stability of
the system in frequency domain, using Nyquist stability criterion.
118 Chapter 7. The Frequency-response analysis

0.6

0.4

Imag Axis 0.2

0
φ

−0.2

−0.4

−0.6
−0.2 0 0.2 0.4 0.6 0.8 1
Real Axis

Figure 7.3: Nyquist plot

7.3.3 Relative stability: Nyquist criterion

If the G(s) curve passes through the point (−1, j0), which represents critical conditions for stability
in the Nyquist criterion, the open-loop system gain is equal to 1 and the phase angle is −180°,
which can be written mathematically as

|G( jω)| = 1 (7.31)

φ = ∠G( jω) = −180° (7.32)

If both the magnitude condition (7.31) and the phase condition (7.32) are satisfied, the closed-loop
system is said to be marginally stable.

Example 7.3
Is a closed-loop system with the following open-loop transfer function and with K=2 stable?
K
G(s) =
s(s + 1)(4s + 1)

Find the critical value of the gain K for stability.

 7.3 Plotting Frequency Response 119

The open-loop frequency response is

K
G( jω) =
jω( jω + 1)(4 jω + 1)
K
=
jω(1 − 4ω 2 + 5 jω)
K
=
−5ω + jω(1 − 4ω 2 )
2

Let the imaginary part of G( jω) be zero, or

1 − 4ω 2 = 0

1 1
from which ω = ± Substituting ω = into G( jω), we obtain
2 2
1 4K
|G( j)| =
2 5

4K 5
The critical value of the gain K is obtained by equating = 1. Hence, K =
5 4
5
The system is stable if 0 < K < . Hence, the system with K=2 is unstable.
4

Definition 7.5: Gain Margin GM

The gain margin is the increase in open-loop gain, when the phase is equal to −180°, to
make the closed-loop system marginally stable.

The gain margin GM is defined as

1
GM = (7.33)
|Re(G( jω p ))|

where ω p is the frequency for which the phase angle of the open-loop transfer function is −180°.
When the gain margin has been expressed in decibels, it is calculated as

n 1 o
GM = 20log (7.34)
|Re(G( jω p ))|

Definition 7.6: Phase Margin ΦM

The phase margin is the amount of phase shift required at unity gain to make the closed-loop
system marginally stable.
120 Chapter 7. The Frequency-response analysis

The gain margin ΦM is defined as

ΦM = 180 + ∠G( jωg ) (7.35)

where ωg is the frequency for which the magnitude of the open-loop transfer function is unity
|G( jωg )| = 1.
The following figure, shows the phase margin and gain margin of stable and unstable systems
using Bode and Nyquist plots

Figure 7.4: Phase and gain gins of stable and unstable systems (Bode diagram)

Figure 7.5: Phase and gain gins of stable and unstable systems (Nyquist diagram)

Example 7.4
Determine the open-loop gain necessary for the closed-loop system shown in the following
Figure to be stable with the gain margin MG ≥ 1.2 and the phase margin ΦM ≥ 45
 7.3 Plotting Frequency Response 121

R(s) K Y (s)
+
-
(s + 1)(2s + 1)(3s + 1)

The sinusoidal transfer function is found to be

K
G( jω) =
( jω + 1)(2 jω + 1)(3 jω + 1)
K
=
(1 − 11ω ) + 6 jω(ω 2 − 1))
2

Let the imaginary part of G( jω) be zero, or

1 − ω2 = 0

from which ω p = ±1rad/s Substituting ω p = 1 into G( jω), we obtain

K
|G( j)| =
10
The condition for the gain margin, MG ≥ 1.2 , yields
10
≥ 1.2, K ≤ 8.333
K
Now, to satisfy the phase margin requirement, the phase angle of the open-loop transfer
function for the frequency at which the magnitude is 1 should be

∠G( jωg ) = −180 + 45 = −135°

From the expression for the open-loop transfer function, the phase angle of G( jω) for jωg is
 6ωg (1 − ω 2 ) 
g
∠G( jωg ) = −tan−1 = −135°
1 − 11ωg2
Taking the tangents of both sides of the previous equation yields a cubic equation for for
jωg is

6ωg3 + 11ωg2 − 6ωg − 1 = 0

which has the solution ωg = 0.54776rad/s.

The magnitude of |G( jωg )|| for this system is
K
|G( jωg )| = q =1
(1 − 11ωg2 )2 + 36ωg2 (1 − ωg2 )2

The solution for K that satisfies the phase condition: K = 3.27.

This value is smaller than the value of K obtained from the gain condition: Therefore the
open-loop gain must be equal to 3.27 (or less) to meet both the gain and the phase angle
conditions.
122 Chapter 7. The Frequency-response analysis

7.4 Frequency-Domain Performance Specifications

7.4.1 Damping ratio and frequency response

For a second order system the magnitude of the frequency response is given by the following
equation

ωn2
M = |G( jω)| = p (7.36)
(ωn2 − ω 2 )2 + 4ξ 2 ω 2 ωn2

For 0 ≤ ξ ≤ 0.707, the maximum value of M occurs at the frequency ωr

p
ωr = ωn 1 − 2ξ 2 (7.37)

Frequency ωr is the resonant frequency. At this frequency, the maximum value of M is given by

1
Mp = p (7.38)
2ξ 1 − ξ 2

Equation (7.38) shows that the maximum magnitude on the frequency response curve is directly
related to the damping ratio and, hence, the percent overshoot. Also notice from equation (7.37)
that the peak frequency, ωr , is not the natural frequency. However, for low values of damping ratio,
we can assume that the peak occurs at the natural frequency.

ω
0
100

−5

−10

−15

−20

−25
logωr logωBW

Figure 7.6: band width of system

 7.4 Frequency-Domain Performance Specifications 123

7.4.2 Damping ratio and phase margin

For an under-damped system, the phase margin is related to the damping ratio ξ .
The phase margin for this system is given by

2ξ
ΦM = tan−1 qp (7.39)
1 + 4ξ 4 − 2ξ 2

Notice that the phase margin ΦM is a function only of the damping ratio ξ . For 0 ≤ ξ ≤ 0.6, the
phase margin ΦM and the damping ratio ξ are related approximately by a straight line as follows:

ΦM = 100° × ξ (7.40)

where the phase margin is measured in degrees.

7.4.3 Response speed and frequency response

The bandwidth of the second order frequency response, is defined as the frequency, ωBW , at which
the magnitude response curve is 3dB down from its value at zero frequency. It is given by

q p
ωBW = ωn 1 − 2ξ 2 + 4ξ 4 − 4ξ 2 + 2 (7.41)

The speed of the time response (as measured by settling time, peak time, and rise time) is related to
the bandwidth of the frequency response.
4
To relate ωBW to settling time, we substitute ωn = into (7.41) and obtain
Ts ξ

q
4 p
ωBW Ts = 1 − 2ξ 2 + 4ξ 4 − 4ξ 2 + 2 (7.42)
Ts ξ

π
Similarly, to relate ωBW to peak time, we substitute ωn = p into (7.41) and obtain
Tp 1 − ξ 2

q
π p
ωBW Tp = p 1 − 2ξ 2 + 4ξ 4 − 4ξ 2 + 2 (7.43)
1−ξ2
124 Chapter 7. The Frequency-response analysis

7.5 Problems

Exercise 7.1

1
Consider a system with a transfer function G(s) = , determine exactly the gain
s(s + 2)2
margin, GM .

Solution 7.1

The open-loop frequency response is

1
G( jω) =
jω( jω + 2)2
1
=
jω(4 − ω 2 + 4 jω)
1
=
−4ω 2 + jω(4 − ω 2 )

Let the imaginary part of G( jω) be zero, or

ω2 − 4 = 0

from which ω p = 2.
Substituting ω p = 2 into G( jω), we obtain

1
ReG( jω p ) = −
16

Hence, GM = 16.

Im
Re
−0.2 −0.1

−0.1

−0.2
 7.5 Problems 125

Exercise 7.2

10
Consider a system with a transfer function G(s) = , determine exactly the phase
s2 + s + 9
margin, ΦM .

Solution 7.2

The open-loop frequency response is

10
G( jω) =
9 − ω 2 + jω

The magnitude of |G( jωg )| for this system is

100
|G( jωg )|2 = =1
(9 − ω 2 )2 + ω 2

We obtain

ωg4 − 17ωg2 − 19 = 0

So ωg2 = 18.05 and ωg = 4.25 rad/s

Now, to satisfy the phase margin requirement, the phase angle of the open-loop transfer
function for the frequency at which the magnitude is 1 should be

∠G( jω) = −∠(9 − ω 2 + jω) = −∠( − 9.06 + j4.25) = −155

Hence, ΦM = 180 − 155 = 25.

1
Im

Re
−1 1 2

−1

−2
126 Chapter 7. The Frequency-response analysis

Exercise 7.3

A feedback control system is shown in Figure 7.7.

R(s) 1 Y (s)
+
- K
s(s + 7)

Figure 7.7: Block diagram

The specification for the closed-loop system requires that the overshoot to a step input be
less than 10%.

1. Determine the corresponding specification Mp in the frequency domain for the closed-
loop transfer function

2. Determine the resonant frequency ωr

3. Determine the phase margin and the bandwidth of the closed-loop system.

Solution 7.3

1. The closed loop transfer function is

K
G(s) =
s2 + 7s + K

For Os = 10%, we calculate as ξ = 0.59, so 2ξ ωn = 7. Hence, K = ωn2 = 35.12.

The maximum value of M is given by
1
Mp = p = 1.05 (7.44)
2ξ 1 − ξ 2

2. For a second order system, the resonant frequency ωr is given by

p
ωr = ωn 1 − 2ξ 2 = 0.55ωn = 3.27rad/s (7.45)

3. the phase margin is given by

ΦM = 100° × ξ = 59° (7.46)

The bandwidth is
q p
ωBW = ωn 1 − 2ξ 2 + 4ξ 4 − 4ξ 2 + 2 = 6.8rad/s (7.47)
 7.5 Problems 127

Exercise 7.4

Consider the following feedback system, Figure 7.8

R(s) 1 Y (s)
+ K
-
(s + 2)2 (s2 + 5s + 4)

Figure 7.8: Block diagram

1. Create the Bode plots for the system, and determine the gain margin that results for
ΦM = 65. What damping ratio would you expect for this ΦM = 65?

2. Find the gain margin.

Solution 7.4

0
Magnitude (dB)

−20

−40
0

−90
Phase (deg)

−180

−270

−360
10−1 100 101

Figure 7.9: Bode graph for K = 30

1. The low-frequency gain of G(s) is found by setting s to zero. Thus, the Bode magni-
tude plot starts at K = 16. For convenience, let K = 16 so that the log- magnitude plot
starts at 0 dB. Figure 7.9 sketch the Bode graph.
128 Chapter 7. The Frequency-response analysis

For ΦM = 65, we calculate ∠G( jω = 65 − 180 = −115. Accordingly, we see that at

a frequency of 1rad/s, the phase plot is −115 and the magnitude plot is −5.5dB.
If the gain is raised by 5.5dB, the magnitude will be zero. We can find the gain value
such that 20logK ′ = 5.5. So K ′ = 1.88 and K = 16K ′ = 30.

2. We see at a frequency of 2rad/s, when the phase plot is −180, the gain margin is
8.5dB or 2.66

Exercise 7.5

Consider the following feedback system, Figure 7.10

R(s) K Y (s)
+ K
-
s(s + 20)(s + 85)

Figure 7.10: Block diagram

1. Find the value of K so that the magnitude plot is 0dB at ω = 1rad/s.

2. Find the adjusted gain to yield a 15% overshoot in the transient response for a step
input.

Solution 7.5

1. At ω = 1rad/s, we have
K
|G( j1)| = | | = 1, ⇒ K = 1702 (7.48)
j( j + 20)( j + 85)

2. For Os = 15%, we calculate ξ = 0.5169.

Using this damping ratio, we can obtain the desired phase margin
2ξ
ΦM = tan−1 qp = 53° (7.49)
1 + 4ξ 4 − 2ξ 2

The phase which yields this phase margin is φ = −180 + 53 = −127°.

From the bode plot, we see that at a frequency of 11.2rad/s, the phase plot is −127
and the magnitude plot is 22.2dB.
If the gain is raised by 22.4dB, the magnitude will be zero. We can find the gain value
such that 20logK ′ = 22. So K ′ = 12.93 and K = 1702K ′ = 22007.
 7.5 Problems 129

0
Magnitude (dB)

−100

−200
−90
Phase (deg)

−180

−270
100 101 102 103 104

Figure 7.11: Bode graph for K = 1702

 CHAPTER
8 Control design

Content

8.1 Root-locus design 131

8.1.1 PD controller 133
8.1.2 PI Controller 133
8.1.3 PID Controller 134
8.1.4 phase-Lag and phase-Lead Controllers 134
8.2 Improvement of Transient Response 136
8.2.1 PD Controller Design 136
8.2.2 Phase-Lead Controller Design 139
8.3 Improvement of Steady State Errors 143
8.3.1 PI Controller Design 143
8.3.2 Phase-Lag Controller Design 146
8.4 Improving Steady-State Error and Transient Response 147
8.4.1 PID Controller Design 147
8.4.2 Lag-Lead Controller Design 148

This chapter consists the central issue of the design of controllers. Based on the root locus
technique as well as the frequency domain, we present several methods used in control engineer to
design controllers or compensator that able to achieve the specified system performances.
First, the root locus technique is used to choose the proper loop gain to meet a transient response
specification. Then the frequency response is used to determine the controller parameters that
guarantee the control system requirements.

8.1 Root-locus design

The root-locus can be used to determine the value of the loop gain, which results in a satisfactory
closed-loop behavior. This is called the proportional compensator or proportional controller
and provides gradual response to deviations from the set point. There are practical limits as to
how large the gain can be made. In fact, very high gains lead to instabilities. If the root-locus plot
is such that the desired performance cannot be achieved by the adjustment of the gain, then it is
necessary to reshape the root-loci by adding the additional controller or compensator Gc (s) to the

131
132 Chapter 8. Control design

open-loop transfer function.

This compensator must be chosen so that the root-locus will pass through the proper region of the
s-plane. In many cases, the speed of response and/or the damping of the uncompensated system
must be increased in order to satisfy the specifications. This requires moving the dominant branches
of the root locus to the left.
The purpose of reshaping the root locus generally falls into one of the following categories.

• Category A

– System is stable; Transient response is satisfactory; Steady-state error is too large.

– Here the gain must be increased to reduce the steady-state error without appreciably
reducing system stability

• Category B

– System is stable; Transient response is unsatisfactory; Steady-state response is satisfac-

tory.
– Root locus must be reshaped so that it is moved farther to the left, away from the
imaginary axis.

• Category C

– System is stable; Transient response is unsatisfactory; Steady-state error is unsatisfac-

tory.
– Root locus must be moved to the left and gain must be increased.

• Category D

– System is unstable for all values of gain.

– Root locus must be reshaped so that part of each branch falls in the LHP, thereby
making the system stable.

Consider the following feedback system:

R(s) Y (s)
+
- K Ḡc (s) G(s)

Figure 8.1: Feedback control system

The open-loop transfer function of the system is given as

G0 (s) = K Ḡc (s)G(s) = Gc (s)G(s) (8.1)

 8.1 Root-locus design 133

From (8.1), we can see that by adding a compensator, the system dynamics is modified by adding
adding zeros and poles to the plant.

• Effects of the Addition of Poles: Generally the additions of a pole to the open-loop transfer
function pulls the root locus to the right. This tends to lower the system’s relative stability
and slow down the settling of the response.

Particularly, the addition of integral control adds a pole at the origin makes the system less
stable.

• Effects of the Addition of Zeros: Generally the addition of a zero to the open-loop transfer
function pulls the root locus to the left. This tends to make the system more stable, and
speed up the settling of the response.

Physically, a zero adds derivative control to the system, introducing anticipation into the
system, speeding up transient response. However, steady-state errors can get worse.

8.1.1 PD controller

PD stands for a proportional and derivative controller. The output signal of this controller is equal
to the sum of two signals:

• a signal obtained by multiplying the input signal by a constant gain,

• a signal obtained by differentiating and multiplying the input signal

The transfer function for a PD controller is given by

Gc (s) = KP + KD s
(8.2)
= K(s + z)

KP
where K = KD , z = .
KD
This controller is used to improve the system transient response.

8.1.2 PI Controller

Similarly to the PD controller, the PI controller produces as its output a weighted sum of the input
signal and its integral. The transfer function of a PI controller is

KI KP s + KI
Gc (s) = KP + =
s s (8.3)
(s + z)
=K
s
134 Chapter 8. Control design

KI
where K = KP , z = .
KP
This controller is used to improve the system response steady state errors since it increases the
control system type by one.

8.1.3 PID Controller

The PID controller is a combination of PD and PI controllers. The transfer of PID controller
function is given by

KI KD s2 + KP s + KI
Gc (s) = KP + + KD s =
s s (8.4)
K(s + z1 )(s + z2 )
=
s

8.1.4 phase-Lag and phase-Lead Controllers

The phase-lag and phase-lead controllers have a transfer function in the following form

s+z
Gc (s) = K (8.5)
s+ p

• If z > p, the controller is known as the phase-lag controller. It belongs to the same class as
the PI controller. It introduces a negative phase into the feedback loop, which justifies its
name. It has a zero and pole with the pole being closer to the imaginary axis.

∠(Gc (s)) = ∠(s + z) − ∠(s + p) = θz − θ p < 0 (8.6)

The phase-lag controller is used to improve steady state errors.

• If z < p, the controller is known as the phase-lead controller. The phase-lead controller
contributes to the feedback loop with a positive phase.

∠(Gc (s)) = ∠(s + z) − ∠(s + p) = θz − θ p > 0 (8.7)

The phase-lead controller is used to improve the system response transient behaviour .

The phase-lag-lead controller is obtained as a combination of phase-lead and phase-lag

controllers. Its transfer function is given by

(s + z1 )(s + z2 )
Gc (s) = K , p2 > z2 > z1 > p1 , z1 z2 = p1 p2 (8.8)
(s + p1 )(s + p2 )
 8.1 Root-locus design 135

As PID controller, phase-lead-lag controller is used to improve simultaneously both the system
transient response and steady state errors.
The control engineer’s task is to adjust the controller gains to arrive at an acceptable degree of
error reduction simultaneously with acceptable dynamic response.
136 Chapter 8. Control design

8.2 Improvement of Transient Response

In this section we show how to improve the transient response by using either the PD or phase-lead
controllers. Typically, the objective is to design a controller such that the output response has a
desirable percent overshoot and a shorter settling time than the uncompensated system.
The common feature of PD and phase lead controllers is introducing a positive phase shift, and
both of them can be implemented in a similar manner.

8.2.1 PD Controller Design

The transient response of a system can be selected by choosing an appropriate closed-loop pole
location on the s-plane.

• If this point is on the root locus, we just adjust the gain to meet the transient response
specification.

• If the closed-loop pole location is not on the root locus, then the root locus must be reshaped
such as the new root locus goes through the selected closed-loop pole location.

To accomplish the previous objective, poles and zeros can be added in the forward path to produce
a new open-loop function whose root locus goes through the design point on the s-plane.
The PD controller is given by

Gc (s) = K(s + zc ) (8.9)

using this controller, the compensated system open-loop transfer function will have one additional
zero which introduces a positive phase shift.

Design Algorithm
• From the performance specifications, determine a pair of complex conjugate poles in
s-plane that produces the desired transient performance.

• Find the required phase contribution of the PD controller.

• Determine the values of zc using a graphical method.

• Determine the value of K using the magnitude condition

Method

• Let the desired closed-loop pole location be s0 = −σd + jωd = |s0 |∠θ0 .

• The angle deficit at the closed-loop pole location

φ = −180 − ∠G(s0 ).

• For a PD design, we require the zero angle to satisfy θz = φ .

 8.2 Improvement of Transient Response 137

• Using the graph in figure 8.2, the location of the zero can be calculated from

ωd
zc = σd + (8.10)
tan(θz )

• The controller gain is obtained from the formula

1
K= (8.11)
|G0 (s0 )|

jω

s0

ωd

θ0
θz σ
zc
σd

Figure 8.2: Determination of the zero of a PD controller

Example 8.1
Consider the following control system, Figure 8.3

R(s) (s + 8) Y (s)
+
- K(s + zc )
(s + 1)(s + 3)(s + 5)

Figure 8.3: Block diagram

Design a PD controller so that the desired damping ratio is 0.707 and the settling time is
1.6s.
Using these specifications, we calculate
4
q
σd
σd = = 2.5, ωn = = 3.53, ωd = ωn2 − σd2 = 2.5 (8.12)
Ts ξ
so, the desired dominant poles are given by s0 = −2.5 ± 2.5 j
138 Chapter 8. Control design

The root locus of this system is represented in Figure 8.4. It is evident from this figure that
the desired dominant poles do not belong to the original root locus.

15

10

−5

−10

−15
−9 −8 −7 −6 −5 −4 −3 −2 −1 0 1

Figure 8.4: Root locus for uncompensated system

The deficit angle is given as

(s + 8)
φ = 180 − ∠G(s0 ) = 180 − ∠ = 40 (8.13)
(s + 1)(s + 3)(s + 5) s0 =−2.5+2.5 j

Hence, the contribution of the compensating zero should be 40.

ωd
zc = σd + , ⇒ zc = 5.48
tan(40)

The gain K can be determined from the magnitude condition. So we obtain

K = 1.11

Therefore the controller transfer function is

Gc (s) = 1.11(s + 5.48)

 8.2 Improvement of Transient Response 139

−2

−4

−6

−8
−14 −12 −10 −8 −6 −4 −2 0 2 4

Figure 8.5: Root locus for compensated system

0.8

0.6

0.4

0.2

0
0 0.5 1 1.5 2 2.5 3 3.5

Figure 8.6: Time responses

8.2.2 Phase-Lead Controller Design

As PD controller, the phase-lead controller introduces a phase shift such that the desired dominant
poles, having the specified transient response performance, belong to the root locus. The general
form of the phase-lead controller is given by
s + zc
Gc (s) = K , pc > zc > 0 (8.14)
s + pc
140 Chapter 8. Control design

which can be also written sa

Ts+1
Gc (s) = Kα , 0<α <1 (8.15)
αT s + 1

where K , zc , pc or α and T to be determined from the design specifications.

The procedures for designing a lead compensator is stated as follows:

Design Algorithm
• From the performance specifications, determine a pair of complex conjugate poles in
s-plane that produces the desired transient performance.

• Find the required phase contribution of the phase-lead controller.

• Determine the values of zc , pc (α and T ) using a graphical method.

• Determine the value of K using the magnitude condition

Method

1. Let s0 be the desired location for the dominant closed-loop pole:

s0 = −σd + jωd = |s0 |∠θ0

2. Find the required phase contribution of the phase-lead controller,

φ = −180 − ∠G(s0 )

3. For a lead design, we require the angle at the zero θz and the angle at the pole θ p to satisfy
θz − θ p = φ .

4. Using the graph in figure 8.7, we calculate the angles θz and θ p as

θ0 + φ
θz =
2 (8.16)
θ0 − φ
θp =
2
 8.2 Improvement of Transient Response 141

jω

s0
φ φ
2 2

ωd

θz θ0
θp σ
zc
σd

Figure 8.7: Determination of the pole and zero of a lead compensator

5. The location of the zero and pole can be calculated from

ωd
zc = σd +
tan(θz )
ωd (8.17)
pc = σd +
tan(θ p )

6. The controller gain is obtained from the formula

1
K= (8.18)
|G0 (s0 )|

Example 8.2
Consider the previous example to design a lead controller
√
The desired dominant poles are given by s0 = −2.5 ± 2.5 j = 2.5 2∠135
The deficit angle is calculated as

(s + 8)
φ = 180 − ∠G(s0 ) = 180 − ∠ = 40 (8.19)
(s + 1)(s + 3)(s + 5) s0 =−2.5+2.5 j

From (8.16), we determine

175
θz = = 87.5
2 (8.20)
95
θp = = 47.5
2
142 Chapter 8. Control design

The location of the zero and pole can be calculated from

2.5
zc = 2.5 + = 2.61
tan(87.5)
(8.21)
2.5
pc = 2.5 + = 4.8
tan(47.5)

The gain K can be determined from the magnitude condition. So we obtain

K = 5.9

Therefore the controller transfer function is

s + 2.61
Gc (s) = 5.9
s + 4.8
The root locus of the controlled open-loop system is given in the following figure

20

10

−10

−20

−9 −8 −7 −6 −5 −4 −3 −2 −1 0 1

Figure 8.8: Root locus for Example 8.2

The step response is depicted in the following figure

 8.3 Improvement of Steady State Errors 143

Step Response
0.7

0.6 System: T0
Settling time (seconds): 1.66

0.5
Amplitude

0.4

0.3 System: T
Settling time (seconds): 2.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5 3 3.5
Time (seconds)

Figure 8.9: Root locus for Example 8.2

8.3 Improvement of Steady State Errors

8.3.1 PI Controller Design

We will discuss in this section how to improve the steady-state error of a feedback control system
without appreciably affecting the transient response.
We know that the steady state errors can be improved by increasing the type of feedback control
system. To achieve this goal we add in cascade with the system a PI controller.
The PI controller is given by

KI s + zc KI
Gc (s) = KP + = KP , zc = (8.22)
s s KP

If KI ≪ KP , zc is a stable zero near the origin. Very often it is implemented as

s + zc
Gc (s) = (8.23)
s
144 Chapter 8. Control design

Design Algorithm
• From the performance specifications, determine a pair of complex conjugate poles in
s-plane that produces the desired transient performance.
σd
• Add a PI controller with a zero at zc =
10
• Tune the gain of the system to move the closed-loop pole closer to s0 .

• Check the time response and modify the design until it is acceptable.

Example 8.3
Consider the following control system, Figure 8.10.

R(s) K(s + 1) Y (s)

+
- Gc (s)
(s + 5)(s2 + 3s + 3)

Figure 8.10: Block diagram

Let the choice of the static gain K = 45 produce a pair of dominant poles on the root locus,
which guarantees the desired transient specifications.
The corresponding position constant and the steady state unit step error are given by
45 1 1
Kp = lim G(s) = = 3 ⇒ ess = = = 0.25 (8.24)
s−→0 15 1 + Kp 4

The root locus for the uncompensated system is shown in Figure 8.11 .
Using a PI controller with zc = 0.1, we obtain the following transfer function

45(s + 0.1)(s + 1)
G0 (s) = (8.25)
s(s + 5)(s2 + 3s + 3)

and we obtain the root locus shown in Figure 8.12.

From these figures we see that the compensated system’s closed-loop poles and gain are
approximately the same as the uncompensated system’s closed-loop poles and gain, so the
transient response of both systems is approximatively the same. However, figure 8.11 shows
that, the compensated system, with its pole at the origin, is a Type 1 system; unlike the
uncompensated system, it will respond to a step input with zero error.
 8.3 Improvement of Steady State Errors 145

8
-3.46 + 6.6j, (k=45)
6

−2

−4

−6

−8
−5 −4 −3 −2 −1 0 1

Figure 8.11: Root locus for uncompensated system

K>0
8
-3.42 + 6.58j, (k=45)
6

−2

−4

−6

−8
−3.5 −3 −2.5 −2 −1.5 −1 −0.5 0
b

Figure 8.12: Root locus for compensated system

146 Chapter 8. Control design

8.3.2 Phase-Lag Controller Design

A lag compensator is typically used to improve the steady-state response for a system whose
transient response is acceptable with the appropriate gain adjustment. The general form of a lag
compensator is

Ts+1
Gc (s) = β , β >1 (8.26)
βTs+1

Even the lag compensator reduces the steady-state error by the factor β it introduces a negative
angle that shifts the root locus to the right. To minimize the latter harmful effect, the lag pole and
zero are located close to the imaginary axis so that their angle contribution is small.

Design Algorithm
• Improve the transient response using proportional control and obtain the corresponding
gain K and closed-loop pole location s0 = −σd + jωd

• Calculate the steady-state error using K and obtain the required improvement in the steady-
state error β .
1 σd
• Select the compensator zero using zc = =
T 10
• Adjust the compensator gain to return the closed-loop pole closer to the desired location.

Example 8.4
Using a phase lag controller with zc = 0.1 and p = 0.01, we obtain the following transfer
function
45(s + 0.1)(s + 1)
G0 (s) = (8.27)
(s + 0.01)(s + 5)(s2 + 3s + 3)
z
Since = 10, the position constant is increased ten times, that is
p
1 1
Kp = 10 × 3 ⇒ ess = = = 0.032 (8.28)
1 + Kp 31
 8.4 Improving Steady-State Error and Transient Response 147

0.8

0.6

0.4
uncontrolledd
0.2 PI controlled
Lag controlled
0
0 10 20 30 40 50

Figure 8.13: Time responses

8.4 Improving Steady-State Error and Transient Response

In order to improve the steady-state error and the transient response independently, we can combine
the design techniques covered in the previous sections.

• If we design a PD controller followed by a PI controller, the resulting compensator is called

a proportional-plus-integral-plus-derivative (PID) controller.

• If we first design a lead compensator and then design a lag compensator, the resulting
compensator is called a lag-lead compensator.

8.4.1 PID Controller Design

The PID controller is accepted as one of the most commonly used controller for industrial plants.
The transfer function of PID controller is
KI (s + zc2 )
Gc (s) = Gc (s) = KP + + KD s = Kc (s + zc1 ) (8.29)
s s
which indicates that the transfer function of a PID controller is the product of transfer functions of
PD and PI controllers.
The following table describes the effect of a PID controller in terms of its three components.

Response Rise time Overshoot Settling time Steady state error

KP Decrease Increase No effect Decrease
KI Decrease Increase Increase Eliminate
KD No effect Decrease Decrease No effect

Table 8.1: Effects of KP , KI , KD in a PID controller.

 Design Algorithm
• Evaluate the performance of the uncompensated system to determine how much improve-
ment in transient response is required.

• Design the PD controller to meet the transient response specifications.

• Design the PI controller to yield the required steady-state error.

8.4.2 Lag-Lead Controller Design

we can improve both transient response and the steady-state error by using a lead compensator and
a lag compensator rather than the ideal PID. This controller is called a lag-lead controller.

Design Algorithm
• Evaluate the performance of the uncompensated system to determine how much improve-
ment in transient response is required.

• Design the lead controller to meet the transient response specifications.

• Design the lag controller to yield the required steady-state error.

• Check that the compensated system has the desired specifications.

8.5 Problems

Exercise 8.1

For the control system shown in Figure ?? design a PI compensator for the following
specifications:

• Zero steady-state error due to a step input.

• A pair of dominant closed-loop poles with a time constant of 0.25 seconds and a
damping ratio of 0.8.

R(s) 1 Y (s)
+
- Gc (s)
(s + 3)(s + 7)

Figure 8.14: Block diagram

 Solution 8.1

1. The closed-loop system should have the following performance specifications:

• time constant Ts = 0.25s,

• damping ratio ξ = 0.8

From these requirements we calculate σd and ωd as follow:

1 σd 4
• We have Ts = 0.25 = , then, σd = 4rad/s and ωn = = =5
σd ξ 0.8
q √
• We can deduce ωd = ωn2 − σd2 = 25 − 16 = 3rad/s

Thus, s0 = σd + jωd = −4 + 3 j is the desired closed-loop pole.

2. The PI controller is given by

Gc (s) = Kp + KI s (8.30)
(s + z) KI
= Kp , z= (8.31)
s Kp

The new open loop transfer function is

(s + z)
G0 (s) = Kp (8.32)
s(s + 3)(s + 7)

The position of controller zero for the desired location of s0 is obtained as follows:
The deficit angle is given as

1
∠G(s0 ) = ∠ = 116.5 (8.33)
s(s + 3)(s + 7) s0 =−4+3 j

Hence, the contribution of the compensating zero should be 116.5.

Using Equation 8.10, we get
3
zc = 4 + , ⇒ zc = 2.5
tan(116.5)

The gain Kp can be determined from the magnitude condition. So, we get

Kp = 20, KI = [Link] = 50

Therefore the controller transfer function is

50
Gc (s) = 20 +
s
 Exercise 8.2

Consider the following control system, Figure ??

R(s) s+2 Y (s)

+ Gc (s)
-
s(s + 1)2

Figure 8.15: Block diagram

1. Suppose Gc (s) = K with K > 0. determine whether or not it is possible to choose

a proper K such that the dominant poles of the closed-loop transfer function are
√
s1,2 = −1 ± j 3. Justify your answer.

2. Design a lead compensator of the form

s+1
Gc (s) = K (8.34)
s+ p
for some constant p so that the closed-loop transfer function of the compensated
√
system can have dominant poles at s1,2 = −1 ± j 3

3. what is the steady state velocity error constant Kv of the closed-loop system when
tracking a unit ramp input?

4. Design a lag-lead compensator Gc (s) so that the closed loop has dominant poles at
√
approximately s1,2 = −1 ± j 3, and the steady state velocity error constant of the
closed-loop system when tracking a unit ramp input is 30.

Solution 8.2
√
1. The desired closed-loop pole is s0 = −1 + j 3.
Using the angle criterion we get

s+2
∠G(s0 ) = ∠ √
= 120 ̸= 180 (8.35)
s(s + 1)2 s0 =−1+ j 3

So, the angle criterion is not satisfied and it is impossible to find K > 0 which achieve
the desired closed-loop of G(s) at the desired pole (see figure ??).
 4

−2

−4

−3 −2.5 −2 −1.5 −1 −0.5 0 0.5 1

Figure 8.16: Root locus for uncompensated system

2. The deficit angle

φ = 180 − ∠G(s0 ) = 60
= ∠Gc (s0 ) = ∠(s0 + 1) − ∠(s0 + p)

Then the angle θ p = ∠(s0 + p) = ∠(s0 + 1) − 60 = 30

Hence, we have
√
3
p = 1+ , ⇒ p=4
tan(30)

The new open loop transfer function is

(s + 2)
G0 (s) = K (8.36)
s(s + 4)(s + 1)

The gain K can be determined from the magnitude condition

1
K= =6
|G0 (s0 )|

3. The static velocity error constant is

s(s + 2) 1
Kv = lim sG0 (s) = 6 = 3, ⇒ ess =
s−→0 s(s + 4)(s + 1) 3

σd
4. In order to improve the speed constant, we need a lag compensator with zc = = 0.1
10
s + 0.1
GLag (s) = , zc > pc (8.37)
s + pc
 Using the lag-lead controller, the open loop transfer function will be

6(s + 0.1)(s + 2)
G1 (s) = K (8.38)
(s + pc )s(s + 4)(s + 1)

The static velocity error constant will be

s(s + 0.1)(s + 2) 0.3

Kv = lim sG1 (s) = 6 = = 30
s−→0 s(s + 4)(s + 1)(s + pc ) pc

Thus pc = 0.01 and the lag compensator model is

s + 0.1
GL (s) = (8.39)
s + 0.01
The lag-lead compensator is finally defined as

(s + 0.1) (s + 1)
GLL (s) =6 (8.40)
(s + 0.01) (s + 4)

Exercise 8.3

Consider the following control scheme for an unstable system. Our goal is to design a
√
controller such that the undamped natural frequency is ωn = 2 rad/s and damping ratio is
√
ξ = 2.

R(s) 1 Y (s)
+
- K(s + z)
s2 − 1

Figure 8.17: Block diagram

1. Find the desired pole location.

2. Design the PD controller that achieve the required goal.

Solution 8.3

1. From
q the closed-loop specifications we compute: σd = ξ ωn = 1rad/s and ωd =
ωn2 − σd2 = 1rad/s
Thus, s0 = −σd + jωd = −1 + 1 j is the desired closed-loop pole.

2. The PD controller is given by

Gc (s) = K(s + z) (8.41)

 3. The deficit angle is given as

1
φ = 180 − ∠G(s0 ) = 180 − ∠ = 63.43 (8.42)
s2 − 1 s0 =−1+1 j

Hence, the contribution of the compensating zero should be 40.

1
zc = 1 + , ⇒ zc = 1.5
tan(63.43)

The gain K can be determined from the magnitude condition. So we obtain

1
K= =2
|G(s0 )Gc (s0 )|

Exercise 8.4

For a unity feedback system with transfer function

K
G(s) = (8.43)
s(s/2 + 1)(s/6 + 1)

design a phase-lag controller such that the closed-loop system has

• velocity error constant Kv = 20,

• phase margin near 45°

Solution 8.4

The velocity error constant is

Kv = lim sG(s) = K (8.44)

s−→0

Thus, to meet the design specification Kv = 20 we must choose K = 20.

According to Bode plot shown in Figure ??, the phase margin of the open-loop system is
−20.7° at 5.29rad/s.
 Magnitude (dB) 50

−50

−100

−150
−90
Phase (deg)

−180

−270
10−1 100 101 102 103

Figure 8.18: Bode plot

′
As shown in Figure ??, we choose ωc = 1.12rad/s to be that frequency at which the phase
is −180 + 45 + 5 = 130°.
We calculate
′
ω ′
z = c = 0.11, α = G(ωc ) = 15.22 (8.45)
10
so lag controller has transfer function
s + 0.011
Gc = (8.46)
15.22(s + 0.0074)
 200

100
Magnitude (dB)

−100

−200
−90
Phase (deg)

−180

−270
10−4 10−3 10−2 10−1 100 101 102 103

Figure 8.19: Bode plot

Exercise 8.5

For a unity feedback system with transfer function

K
G(s) = (8.47)
s(s + 40)(s + 100)

design a lead compensator to yield a 20% overshoot and Kv = 50 with a settling time of 0.2
s.

Solution 8.5

The velocity error constant is

K
Kv = lim sG(s) = (8.48)
s−→0 4000
Thus, to meet the design specification Kv = 50 we must choose K = 200000.
Given the overshoot and the settling time specifications, we calculate the damping ratio

−Ln(Os)
ξ=p = 0.456 (8.49)
π 2 + Ln2 (Os)
The desired phase margin of the compensated system is

−1 2ξ
Φdes
M = tan q p = 48.1° (8.50)
−2ξ 2 + 1 + 4ξ 4

From the Bode diagram the phase margin of the uncompensated system is ΦM = 29.1° at
ω = 35.3rad/s

50
Magnitude (dB)

−50

−100

−150
−90
Phase (deg)

−180

−270
100 101 102 103 104

Figure 8.20: Bode plot

The necessary phase lead Φm = Φdes

M − ΦM + 10° = 29°, and α =

1 + sin29
α= = 2.89 (8.51)
1 − sin29
1 1
Setting the phase-margin frequency at ωmax , |G(s)| = = √ . Thus we look for the
|Gc (s)| α √
frequency at which the uncompensated system reaches 20log|G(s)| = 20log α = −4.61dB.
From the Bode diagram, this is at ωmax = 48.5rad/s. So, the parameters of the phase-lead
compensator are given
√ p
p = ωm α = 28.52, z = = 82.45 (8.52)
α
The controller is given by
s + 28.52
C(s) = 2.89 (8.53)
s + 82.45
 1.4

1.2

0.8

0.6

0.4

0.2

0
0 5 · 10−2 0.1 0.15 0.2 0.25 0.3

Figure 8.21: Step response

0.5

0.4

0.3

0.2

0.1

0
0 5 · 10−2 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

Figure 8.22: Ramp response

Exercise 8.6

We wish the dominant complex conjugate poles of the closed-loop system to be located at
√
−4 ± j2 3. The deficiency angle of the uncompensated system at the upper dominant pole
is 30 deg. We use a compensator in the forward path whose transfer function is
s+4
Gc (s) = (8.54)
s + pc
Determine the pole pc

Solution 8.6
√
The dominant pole is s0 = −4 + j2 3. Then θ0 = ∠s √0 = 139.11 deg.
θ0 − 30 2 3
Then, θ p = = 54.55 deg, and pc = 4 + = 7. Thus, the pole is loacted at
2 tan(54.55)
s p = −7
159
1.1 Description of a control system . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 Open-loop system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Closed-loop system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Fuel cell hybrid electric vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Greenhouse system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.1 RLC network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2 Mechanical system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Two-loop electrical network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Mass, spring, and damper system . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5 RLC network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 Electrical network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.7 RL circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.8 Electric circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.9 Equivalent circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.12 DC Motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.13 Block diagram representation of the DC motor . . . . . . . . . . . . . . . . . . . 35
2.14 Automobile suspension system . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.1 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3.2 Cascade form and equivalent transfer function . . . . . . . . . . . . . . . . . . . 41
3.3 Parallel form and equivalent transfer function . . . . . . . . . . . . . . . . . . . 41
3.4 Feedback form and equivalent transfer function . . . . . . . . . . . . . . . . . . 41
3.5 Equivalent form for moving the summing junction to the left or to the right . . . 42
3.6 Equivalent form for moving the pickoff to the left or to the right . . . . . . . . . 42
3.7 Complex block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.8 Nodes in signal flow graph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.9 Block diagram and its signal flow graph. . . . . . . . . . . . . . . . . . . . . . . 46
3.10 Signal flow graph. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.11 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.12 Simplified block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.13 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.14 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

161
4.1 Response specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.2 Block diagram of first order system . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.3 First order step response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.4 Block diagram of second order system . . . . . . . . . . . . . . . . . . . . . . . 59
4.5 Plot of Unit-Step Response Curves with ωn = 1 and ξ = 0, 0.2, 0.4, 0.6, 0.8, 1 . . 61
4.6 Pole plot for an under-damped second-order system . . . . . . . . . . . . . . . . 63
4.7 Maximum overshoot Os and normalized peak time, Tp ωn versus ξ . . . . . . . . 64
4.8 Feedback Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.9 First order response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.10 RLC circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.11 poles location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.12 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.13 Response output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.14 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.15 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.16 Block diagram for a PI control system . . . . . . . . . . . . . . . . . . . . . . . 78

5.1 poles and response for stable system . . . . . . . . . . . . . . . . . . . . . . . . 82

5.2 poles and response for unstable system . . . . . . . . . . . . . . . . . . . . . . . 83
5.3 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.4 Block diagram of PI control system . . . . . . . . . . . . . . . . . . . . . . . . 91

6.1 Unit feedback system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

6.2 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.3 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.4 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
6.5 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.6 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.7 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.8 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
6.9 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
6.10 Root locus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
6.11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

7.1 Bode plot for first order system . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

7.2 Bode plot for second order system . . . . . . . . . . . . . . . . . . . . . . . . . 117
7.3 Nyquist plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.4 Phase and gain gins of stable and unstable systems (Bode diagram) . . . . . . . . 120
7.5 Phase and gain gins of stable and unstable systems (Nyquist diagram) . . . . . . 120
7.6 band width of system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7.7 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
7.8 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.9 Bode graph for K = 30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.10 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.11 Bode graph for K = 1702 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

8.1 Feedback control system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

8.2 Determination of the zero of a PD controller . . . . . . . . . . . . . . . . . . . . 137
8.3 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
8.4 Root locus for uncompensated system . . . . . . . . . . . . . . . . . . . . . . . 138
8.5 Root locus for compensated system . . . . . . . . . . . . . . . . . . . . . . . . 139
8.6 Time responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.7 Determination of the pole and zero of a lead compensator . . . . . . . . . . . . . 141
8.8 Root locus for Example 8.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
8.9 Root locus for Example 8.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
8.10 Block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
8.11 Root locus for uncompensated system . . . . . . . . . . . . . . . . . . . . . . . 145
8.12 Root locus for compensated system . . . . . . . . . . . . . . . . . . . . . . . . 145
8.13 Time responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
2.1 Voltage-current and impedance relationships for capacitors, resistors, and inductors 17
2.2 Force-velocity, force-displacement, and impedance translational relationships for
springs, viscous dampers, and mass . . . . . . . . . . . . . . . . . . . . . . . . 19

3.1 Elements of block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.1 Second-order systems, pole plots, and step responses . . . . . . . . . . . . . . . 62

4.2 Steady state errors for various inputs and type of systems . . . . . . . . . . . . . 68

5.1 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

5.2 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.3 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.4 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.5 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
5.6 Routh table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

8.1 Effects of KP , KI , KD in a PID controller. . . . . . . . . . . . . . . . . . . . . . 147

165