# Instructor's Solutions Manual For Modern Control Systems, 12th Ed Systems Edition

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 754

© 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained
from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

MODERN CONTROL SYSTEMS
SOLUTION MANUAL

Richard C. Dorf

Robert H. Bishop

University of California, Davis

Marquette University

A companion to
MODERN CONTROL SYSTEMS
TWELFTH EDITION
Richard C. Dorf
Robert H. Bishop

Prentice Hall
Upper Saddle River Boston Columbus San Francisco New York
Indianapolis London Toronto Sydney Singapore Tokyo Montreal Dubai
Madrid Hong Kong Mexico City Munich Paris Amsterdam Cape Town

Educator Home | eLearning & Assessment | Support/Contact Us | Find your rep | Exam copy bookbag

Instructor's Solutions Manual
for Modern Control Systems, 12/E
Richard C. Dorf, University of California, Davis
Robert H. Bishop, University of Texas at Austin
ISBN-10: 013602498X
ISBN-13: 9780136024989
Publisher: Prentice Hall
Format: On-line Supplement
Published: 08/16/2010

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

P R E F A C E

In each chapter, there are five problem types:
Exercises
Problems
Design Problems/Continuous Design Problem
Computer Problems
In total, there are over 1000 problems. The abundance of problems of increasing complexity gives students confidence in their problem-solving
ability as they work their way from the exercises to the design and
computer-based problems.
and the Control System Toolbox or to LabVIEW and the MathScript RT
Module. All of the computer solutions in this Solution Manual were developed and tested on an Apple MacBook Pro platform using MATLAB 7.6
Release 2008a and the Control System Toolbox Version 8.1 and LabVIEW
2009. It is not possible to verify each solution on all the available computer
platforms that are compatible with MATLAB and LabVIEW MathScript
RT Module. Please forward any incompatibilities you encounter with the
scripts to Prof. Bishop at the email address given below.
The authors and the staff at Prentice Hall would like to establish an
open line of communication with the instructors using Modern Control
Systems. We encourage you to contact Prentice Hall with comments and
suggestions for this and future editions.
Robert H. Bishop

rhbishop@marquette.edu

iii

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

T A B L E - O F - C O N T E N T S

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

iv

Introduction to Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Mathematical Models of Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
State Variable Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Feedback Control System Characteristics . . . . . . . . . . . . . . . . . . . . . . . 133
The Performance of Feedback Control Systems . . . . . . . . . . . . . . . . . 177
The Stability of Linear Feedback Systems . . . . . . . . . . . . . . . . . . . . . . 234
The Root Locus Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Frequency Response Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
Stability in the Frequency Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
The Design of Feedback Control Systems . . . . . . . . . . . . . . . . . . . . . . . 519
The Design of State Variable Feedback Systems . . . . . . . . . . . . . . . . 600
Robust Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
Digital Control Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

1

Introduction to Control Systems

There are, in general, no unique solutions to the following exercises and
problems. Other equally valid block diagrams may be submitted by the
student.

Exercises
E1.1

A microprocessor controlled laser system:
Controller

Desired
power
output

Error

-

Microprocessor

Current i(t)

Laser

Power
Sensor

power

A driver controlled cruise control system:
Controller

Process

Foot pedal
Desired
speed

Power
out

Measurement

Measured

E1.2

Process

-

Driver

Car and
Engine

Actual
auto
speed

Measurement

Visual indication of speed

E1.3

Speedometer

Although the principle of conservation of momentum explains much of
the process of fly-casting, there does not exist a comprehensive scientific
explanation of how a fly-fisher uses the small backward and forward motion of the fly rod to cast an almost weightless fly lure long distances (the
1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

2

CHAPTER 1

Introduction to Control Systems

current world-record is 236 ft). The fly lure is attached to a short invisible
leader about 15-ft long, which is in turn attached to a longer and thicker
Dacron line. The objective is cast the fly lure to a distant spot with deadeye accuracy so that the thicker part of the line touches the water first
and then the fly gently settles on the water just as an insect might.
Fly-fisher
Desired
position of
the fly

Controller

-

Wind
disturbance

Mind and
body of the
fly-fisher

Process

Rod, line,
and cast

Actual
position
of the fly

Measurement

Visual indication
of the position of
the fly

E1.4

Vision of
the fly-fisher

An autofocus camera control system:
One-way trip time for the beam

Conversion factor
(speed of light or
sound)

K1
Beam
Emitter/
Beam return

Distance to subject

Subject
Lens focusing
motor

Lens

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

3

Exercises

E1.5

Tacking a sailboat as the wind shifts:

Error

Desired
sailboat
direction

-

Controller

Actuators

Sailor

Rudder and

Wind

Process

Sailboat

Actual
sailboat
direction

Measurement
Measured sailboat direction

Gyro compass

E1.6

An automated highway control system merging two lanes of traffic:
Controller

Error

Desired
gap

-

Embedded
computer

Actuators

Brakes, gas or
steering

Process

Active
vehicle

Actual
gap

Measurement
Measured gap

E1.7

Using the speedometer, the driver calculates the difference between the
measured speed and the desired speed. The driver throotle knob or the
brakes as necessary to adjust the speed. If the current speed is not too
much over the desired speed, the driver may let friction and gravity slow
the motorcycle down.
Controller

Desired
speed

Error

-

Driver

Actuators

Throttle or
brakes

Measurement
Visual indication of speed

Speedometer

Process

Motorcycle

Actual
motorcycle
speed

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

4

CHAPTER 1

E1.8

Introduction to Control Systems

Human biofeedback control system:
Controller

Desired
body
temp

Process

Hypothalumus

-

Message to
blood vessels

Actual
body
temp

Human body

Measurement
Visual indication of
body temperature

E1.9

TV display

Body sensor

E-enabled aircraft with ground-based flight path control:
Corrections to the
flight path

Desired
Flight
Path

-

Controller

Aircraft

Gc(s)

G(s)

Flight
Path
Health
Parameters

Meteorological
data

Location
and speed

Optimal
flight path

Ground-Based Computer Network
Optimal
flight path
Meteorological
data

Desired
Flight
Path

E1.10

Specified
Flight
Trajectory

Health
Parameters

Corrections to the
flight path

Gc(s)

G(s)

Controller

Aircraft

Location
and speed

Flight
Path

Unmanned aerial vehicle used for crop monitoring in an autonomous
mode:
Trajectory
error

-

Controller

UAV

Gc(s)

G(s)

Flight
Trajectory

Sensor
Location with
respect to the ground

Map
Correlation
Algorithm

Ground
photo

Camera

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

5

Exercises

E1.11

An inverted pendulum control system using an optical encoder to measure
the angle of the pendulum and a motor producing a control torque:
Actuator

Voltage

Error

Desired
angle

-

Controller

Process

Torque

Motor

Pendulum

Angle

Measurement

Measured
angle

E1.12

In the video game, the player can serve as both the controller and the sensor. The objective of the game might be to drive a car along a prescribed
path. The player controls the car trajectory using the joystick using the
visual queues from the game displayed on the computer monitor.
Controller

Desired
game
objective

Optical
encoder

Error

-

Player

Actuator

Joystick

Measurement

Player
(eyesight, tactile, etc.)

Process

Video game

Game
objective

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

6

CHAPTER 1

Introduction to Control Systems

Problems
P1.1

Desired
temperature
set by the
driver

An automobile interior cabin temperature control system block diagram:

Error

-

Controller

Process

Thermostat and
air conditioning
unit

Automobile
cabin

Automobile
cabin temperature

Measurement
Measured temperature

P1.2

Temperature
sensor

A human operator controlled valve system:
Controller

Process

Error *

Desired
fluid
output *

-

Tank

Valve

Fluid
output

Measurement
Visual indication
of fluid output *

Meter
* = operator functions

P1.3

A chemical composition control block diagram:
Controller

Process

Error
Desired
chemical
composition

-

Mixer tube

Valve

Measurement
Measured chemical
composition

Infrared analyzer

Chemical
composition

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

7

Problems

P1.4

A nuclear reactor control block diagram:
Controller

Process

Error
Desired
power level

Reactor
and rods

Motor and
amplifier

-

Output
power level

Measurement
Measured chemical
composition

P1.5

A light seeking control system to track the sun:

Measurement

Light
source

Dual
Photocells

P1.6

Ionization chamber

Controller

Ligh
intensity

Trajectory
Planner

Desired
carriage
position

Controller

-

Motor,
carriage,
and gears

K

Photocell
carriage
position

If you assume that increasing worker’s wages results in increased prices,
then by delaying or falsifying cost-of-living data you could reduce or eliminate the pressure to increase worker’s wages, thus stabilizing prices. This
would work only if there were no other factors forcing the cost-of-living
up. Government price and wage economic guidelines would take the place
of additional “controllers” in the block diagram, as shown in the block
diagram.
Controller

Process
Market-based prices

Initial
wages

Process

Motor
inputs

Error

-

Industry

Government
price
guidelines

Controller

Wage increases

Government
wage
guidelines

Cost-of-living

K1

Prices

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

8

CHAPTER 1

P1.7

Introduction to Control Systems

Assume that the cannon fires initially at exactly 5:00 p.m.. We have a
positive feedback system. Denote by ∆t the time lost per day, and the
net time error by ET . Then the follwoing relationships hold:
∆t = 4/3 min. + 3 min. = 13/3 min.
and
ET = 12 days × 13/3 min./day .
Therefore, the net time error after 15 days is
ET = 52 min.

P1.8

The student-teacher learning process:
Process

Controller

Lectures

Error
Desired
knowledge

-

Teacher

Knowledge

Student

Measurement

Exams

Measured knowledge

P1.9

A human arm control system:
Process

Controller
u
Desired
arm
location

e

y

s
Brain

Nerve signals

z
Measurement

Visual indication of
arm location

Pressure
Eyes and
pressure
receptors

Arm &
muscles

d

Arm
location

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

9

Problems

P1.10

An aircraft flight path control system using GPS:
Controller

Desired
flight path
from air traffic
controllers

Actuators

Computer
Auto-pilot

Error

-

Process

Ailerons, elevators,
rudder, and
engine power

Flight
path

Aircraft

Measurement
Measured flight path

P1.11

The accuracy of the clock is dependent upon a constant flow from the
orifice; the flow is dependent upon the height of the water in the float
tank. The height of the water is controlled by the float. The control system
controls only the height of the water. Any errors due to enlargement of
the orifice or evaporation of the water in the lower tank is not accounted
for. The control system can be seen as:

Desired
height of
the water
in float tank

P1.12

Global Positioning
System

-

Controller

Process

Float level

Flow from
upper tank
to float tank

Actual
height

Assume that the turret and fantail are at 90◦ , if θw 6= θF -90◦ . The fantail
operates on the error signal θw - θT , and as the fantail turns, it drives the
turret to turn.

y

Wind

qW = Wind angle
qF = Fantail angle
qT = Turret angle

Controller

*

qW
qF
qT

qW

*

Turret

x

-

Process
Torque

Error

Fantail

Fantail

Gears & turret

qT

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

10

CHAPTER 1

P1.13

Introduction to Control Systems

This scheme assumes the person adjusts the hot water for temperature
control, and then adjusts the cold water for flow rate control.
Controller

Error

Desired water
temperature

Process

Hot water
system

-

Hot
water

Actual
water temperature
and flow rate
Desired water
flow rate

Cold water
system

-

Cold
water

Measurement

Measured water flow
Measured water temperature

P1.14

Human: visual
and touch

If the rewards in a specific trade is greater than the average reward, there
is a positive influx of workers, since
q(t) = f1 (c(t) − r(t)).
If an influx of workers occurs, then reward in specific trade decreases,
since
c(t) = −f2 (q(t)).
Controller

Average
rewards
r(t)

P1.15

Desired
Fuel
Pressure

Error

-

f1(c(t)-r(t))

Process
q(t)

- f2(q(t))

Total of
rewards
c(t)

A computer controlled fuel injection system:

-

Controller

Process

Electronic
Control Unit

High Pressure Fuel
Supply Pump and
Electronic Fuel
Injectors

Measurement
Measured fuel pressure

Fuel Pressure
Sensor

Fuel
Pressure

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

11

Problems

P1.16

With the onset of a fever, the body thermostat is turned up. The body
adjusts by shivering and less blood flows to the skin surface. Aspirin acts
to lowers the thermal set-point in the brain.
Controller

Desired temperature
or set-point from body
thermostat in the brain

Process

within the
body

-

Body
temperature

Body

Measurement
Measured body temperature

Internal sensor

P1.17

Hitting a baseball is arguably one of the most difficult feats in all of sports.
Given that pitchers may throw the ball at speeds of 90 mph (or higher!),
batters have only about 0.1 second to make the decision to swing—with
bat speeds aproaching 90 mph. The key to hitting a baseball a long distance is to make contact with the ball with a high bat velocity. This is
more important than the bat’s weight, which is usually around 33 ounces
(compared to Ty Cobb’s bat which was 41 ounces!). Since the pitcher can
throw a variety of pitches (fast ball, curve ball, slider, etc.), a batter must
decide if the ball is going to enter the strike zone and if possible, decide
the type of pitch. The batter uses his/her vision as the sensor in the feedback loop. A high degree of eye-hand coordination is key to success—that
is, an accurate feedback control system.

P1.18

Define the following variables: p = output pressure, fs = spring force
= Kx, fd = diaphragm force = Ap, and fv = valve force = fs - fd .
The motion of the valve is described by ÿ = fv /m where m is the valve
mass. The output pressure is proportional to the valve displacement, thus
p = cy , where c is the constant of proportionality.

Constant of
proportionality

Spring

Screw
displacement
x(t)

K

fs

-

Valve position

fv

Valve

c

y

Diaphragm area

fd

A

Output
pressure
p(t)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

12

CHAPTER 1

P1.19

Introduction to Control Systems

A control system to keep a car at a given relative position offset from a

Throttle

Position of
follower

Follower
car

Actuator

u

-

Controller

Relative
position

-

Position

Fuel
throttle
(fuel)

Video camera
& processing
algorithms

Reference
photo

Desired relative position

P1.20

A control system for a high-performance car with an adjustable wing:

Desired

-

Process

Actuator

Controller

Computer

wing

conditions

Race Car

Measurement

P1.21

K

Tire internal
strain gauges

A control system for a twin-lift helicopter system:
Measurement
Measured separation
distance

Desired separation
distance

-

Controller

Process
Separation distance

Pilot
Desired altitude

Helicopter
Altitude

Measurement
Measured altitude

Altimeter

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

13

Problems

P1.22

The desired building deflection would not necessarily be zero. Rather it
would be prescribed so that the building is allowed moderate movement
up to a point, and then active control is applied if the movement is larger
than some predetermined amount.
Process
Controller

Desired
deflection

Hydraulic
stiffeners

-

Building

Deflection

Measurement

Measured deflection

P1.23

Strain gauges
on truss structure

K

The human-like face of the robot might have micro-actuators placed at
strategic points on the interior of the malleable facial structure. Cooperative control of the micro-actuators would then enable the robot to achieve
various facial expressions.
Controller

Process

Error
Desired
actuator
position

-

Voltage

Electromechanical
actuator

Amplifier

Actuator
position

Measurement

Position
sensor

Measured position

P1.24

We might envision a sensor embedded in a “gutter” at the base of the
windshield which measures water levels—higher water levels corresponds
to higher intensity rain. This information would be used to modulate the
Process

Controller

Desired
wiper speed

and motor

Electronic
Control Unit

-

Measurement

K

Measured water level

Water depth
sensor

Wiper
speed

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

14

CHAPTER 1

Introduction to Control Systems

A feedback control system for the space traffic control:

P1.25

Controller

Error

Desired
orbit position

Control
law

-

Actuator
Jet
commands

Process
Applied
forces

Reaction
control jets

Satellite

Actual

orbit position

Measurement
Measured orbit position

Earth-based control of a microrover to point the camera:

P1.26

Microrover
Camera position
command

Transmitter

Controller

G(s)

Gc(s)

Rover
position

Camera

Camera
Position

m
Ca
ap
er

Sensor

ea

iti
os

M

Measured camera
position

on

d
re
su

d
an

m
m
co

ap
er

m
ca

on

iti
os

P1.27

Desired
Charge
Level

Control of a methanol fuel cell:

-

Controller

Recharging
System

Gc(s)

GR(s)

Methanol water
solution

G(s)
Sensor

Measured charge level

Fuel Cell

H(s)

Charge
Level

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

15

AP1.1

Control of a robotic microsurgical device:

Microsurgical
robotic manipulator

Controller
Desired
End-effector
Position

-

G(s)

Gc(s)

End-effector
Position

Sensor

H(s)

AP1.2

An advanced wind energy system viewed as a mechatronic system:
AERODYNAMIC DESIGN
STRUCTURAL DESIGN OF THE TOWER
ELECTRICAL AND POWER SYSTEMS

SENSORS
Rotor rotational sensor
Wind speed and direction sensor
ACTUATORS
Motors for manipulatiing the propeller pitch

Physical System Modeling

CONTROL SYSTEM DESIGN AND ANALYSIS
ELECTRICAL SYSTEM DESIGN AND ANALYSIS
POWER GENERATION AND STORAGE

Sensors and Actuators
WIND ENERGY
SYSTEM

Software and
Data Acquisition

CONTROLLER ALGORITHMS
DATA ACQUISTION: WIND SPEED AND DIRECTION
ROTOR ANGULAR SPEED
PROPELLOR PITCH ANGLE

AP1.3

Signals and Systems

Computers and
Logic Systems

COMPUTER EQUIPMENT FOR CONTROLLING THE SYSTEM
SAFETY MONITORING SYSTEMS

The automatic parallel parking system might use multiple ultrasound
sensors to measure distances to the parked automobiles and the curb.
The sensor measurements would be processed by an on-board computer
to determine the steering wheel, accelerator, and brake inputs to avoid
collision and to properly align the vehicle in the desired space.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

16

CHAPTER 1

Introduction to Control Systems

Even though the sensors may accurately measure the distance between
the two parked vehicles, there will be a problem if the available space is
not big enough to accommodate the parking car.
Controller

Desired
automobile
position

Error

Actuators

On-board
computer

-

Steering wheel,
accelerator, and
brake

Process

Actual
automobile
position

Automobile

Measurement

Position of automobile
relative to parked cars
and curb

Ultrasound

There are various control methods that can be considered, including placing the controller in the feedforward loop (as in Figure 1.3). The adaptive
optics block diagram below shows the controller in the feedback loop, as
an alternative control system architecture.

AP1.4

Process

Astronomical
object
Uncompensated
image

Astronomical
telescope
mirror

Compensated
image

Measurement

Wavefront
reconstructor

Wavefront
corrector

Wavefront
sensor

Actuator & controller

AP1.5

Desired
floor

Error

-

The control system might have an inner loop for controlling the acceleration and an outer loop to reach the desired floor level precisely.

Controller #2

Outer
Loop

Desired
acceleration

Error

-

Controller #1

Elevator
motor,
cables, etc.

Inner
Loop
Measured acceleration

Acceleration
Measurement

Elevator

Floor

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

17

An obstacle avoidance control system would keep the robotic vacuum
cleaner from colliding with furniture but it would not necessarily put the
vacuum cleaner on an optimal path to reach the entire floor. This would
require another sensor to measure position in the room, a digital map of
the room layout, and a control system in the outer loop.

AP1.6

Process
Desired
distance
from
obstacles

Error

-

Controller

Measured distance from obstacle

Motors,
wheels, etc.

Infrared
sensors

Robotic
vacuum
cleaner

Distance
from
obstacles

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

18

CHAPTER 1

Introduction to Control Systems

Design Problems
CDP1.1

The machine tool with the movable table in a feedback control configuration:
Controller

Error

Desired
position
x

Amplifier

-

Actuator

Process

Machine
tool with
table

Positioning
motor

Actual
position
x

Measurement

Position sensor

Measured position

DP1.1

Use the stereo system and amplifiers to cancel out the noise by emitting
signals 180◦ out of phase with the noise.
Process

Controller
Noise
signal
Desired
noise = 0

Shift phase
by 180 deg

-

Machine
tool with
table

Positioning
motor

Noise in
cabin

Measurement

Microphone

DP1.2

Desired
speed
of auto
set by
driver

1/K

An automobile cruise control system:
Controller

Desired
shaft
speed

-

Electric
motor

Process

Automobile
and engine

Valve

Measurement

Measured shaft speed

Shaft speed
sensor

Drive shaf t speed

K

Actual
speed
of auto

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

19

Design Problems

DP1.3

An automoted cow milking system:
Measurement
Cow location

Vision system

Motor and
gears

-

Desired cup
location

Process

Actuator

Controller

Location
of cup

Robot arm and
cup gripper

Cow and
milker

Milk

Measurement

Vision system

Measured cup location

DP1.4

A feedback control system for a robot welder:
Controller

Desired
position

Process

Computer and
amplifier

Error

-

Voltage

Motor and
arm

Weld
top
position

Measurement

Vision camera

Measured position

DP1.5

A control system for one wheel of a traction control system:
Antislip
controller

Engine torque

+

-

Wheel
dynamics

+

-

Wheel
speed

Sensor

+
Actual slip

1/Rw

Vehicle
dynamics

Brake torque

+

Vehicle speed

Antiskid
controller

Sensor

Measured
slip

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

20

CHAPTER 1

Introduction to Control Systems

A vibration damping system for the Hubble Space Telescope:

DP1.6

Controller
Desired
jitter = 0

Error

Computer

-

Actuators

Gyro and
reaction wheels

Process
Signal to
cancel the jitter

Spacecraft
dynamics

Jitter of
vibration

Measurement

Measurement of 0.05 Hz jitter

DP1.7

A control system for a nanorobot:
Controller

Desired
nanorobot
position

Rate gyro
sensor

Error

-

Biocomputer

Actuators

Plane surfaces
and propellers

Process

Nanorobot

Actual
nanorobot
position

Measurement

External beacons

Many concepts from underwater robotics can be applied to nanorobotics
within the bloodstream. For example, plane surfaces and propellers can
provide the required actuation with screw drives providing the propulsion. The nanorobots can use signals from beacons located outside the
skin as sensors to determine their position. The nanorobots use energy
from the chemical reaction of oxygen and glucose available in the human
body. The control system requires a bio-computer–an innovation that is
not yet available.
For further reading, see A. Cavalcanti, L. Rosen, L. C. Kretly, M. Rosenfeld, and S. Einav, “Nanorobotic Challenges n Biomedical Application,
Design, and Control,” IEEE ICECS Intl Conf. on Electronics, Circuits
and Systems, Tel-Aviv, Israel, December 2004.
DP1.8

The feedback control system might use gyros and/or accelerometers to
measure angle change and assuming the HTV was originally in the vertical
position, the feedback would retain the vertical position using commands
to motors and other actuators that produced torques and could move the
HTV forward and backward.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

21

Design Problems

Process
Desired angle
from vertical (0o)

Error

-

Controller

Measured angle from vertical

Motors,
wheels, etc.

Gyros &
accelerometers

HTV

Angle from
vertical

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

2

Mathematical Models of Systems

Exercises
E2.1

We have for the open-loop
y = r2
and for the closed-loop
e = r − y and y = e2 .
So, e = r − e2 and e2 + e − r = 0 .
16

14

12

y

10

8

open-loop

6

4

closed-loop

2

0

0

0.5

1

1.5

2
r

2.5

3

3.5

4

FIGURE E2.1
Plot of open-loop versus closed-loop.

For example, if r = 1, then e2 + e − 1 = 0 implies that e = 0.618. Thus,
y = 0.382. A plot y versus r is shown in Figure E2.1.
22

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

23

Exercises

E2.2

Define
f (T ) = R = R0 e−0.1T
and
∆R = f (T ) − f (T0 ) , ∆T = T − T0 .
Then,
∆R = f (T ) − f (T0 ) =

∂f
∂T

T =T0 =20◦

∆T + · · ·

where
∂f
∂T

T =T0 =20◦

= −0.1R0 e−0.1T0 = −135,

when R0 = 10, 000Ω. Thus, the linear approximation is computed by
considering only the first-order terms in the Taylor series expansion, and
is given by
∆R = −135∆T .
The spring constant for the equilibrium point is found graphically by
estimating the slope of a line tangent to the force versus displacement
curve at the point y = 0.5cm, see Figure E2.3. The slope of the line is
K ≈ 1.
2
1.5
Spring breaks

1
0.5
0

Force (n)

E2.3

-0.5
-1
-1.5
-2
-2.5
-3
-2

Spring compresses
-1.5

-1

-0.5

0

0.5

1

y=Displacement (cm)

FIGURE E2.3
Spring force as a function of displacement.

1.5

2

2.5

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

24

CHAPTER 2

E2.4

Mathematical Models of Systems

Since
R(s) =

1
s

we have
Y (s) =

4(s + 50)
.
s(s + 20)(s + 10)

The partial fraction expansion of Y (s) is given by
Y (s) =

A1
A2
A3
+
+
s
s + 20 s + 10

where
A1 = 1 , A2 = 0.6 and A3 = −1.6 .
Using the Laplace transform table, we find that
y(t) = 1 + 0.6e−20t − 1.6e−10t .
The final value is computed using the final value theorem:
4(s + 50)
lim y(t) = lim s
=1.
2
t→∞
s→0
s(s + 30s + 200)


E2.5



The circuit diagram is shown in Figure E2.5.
R2

v+

A
+

vin
-

FIGURE E2.5
Noninverting op-amp circuit.

With an ideal op-amp, we have
vo = A(vin − v − ),

+
v0
-

R1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

25

Exercises

where A is very large. We have the relationship
R1
vo .
R1 + R2

v− =
Therefore,

vo = A(vin −

R1
vo ),
R1 + R2

and solving for vo yields
vo =

A
1+

AR1
R1 +R2

1
Since A ≫ 1, it follows that 1 + RAR
≈
1 +R2
vo simplifies to

vo =
E2.6

vin .

AR1
R1 +R2 .

Then the expression for

R1 + R2
vin .
R1

Given
y = f (x) = ex
and the operating point xo = 1, we have the linear approximation
y = f (x) = f (xo ) +

∂f
∂x

x=xo

(x − xo ) + · · ·

where
df
dx

f (xo ) = e,

= e,
x=xo =1

and x − xo = x − 1.

Therefore, we obtain the linear approximation y = ex.
E2.7

The block diagram is shown in Figure E2.7.

R(s)

Ea(s)

+

G1(s)

G2(s)

-

H(s)
FIGURE E2.7
Block diagram model.

I(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

26

CHAPTER 2

Mathematical Models of Systems

Starting at the output we obtain
I(s) = G1 (s)G2 (s)E(s).
But E(s) = R(s) − H(s)I(s), so
I(s) = G1 (s)G2 (s) [R(s) − H(s)I(s)] .
Solving for I(s) yields the closed-loop transfer function
G1 (s)G2 (s)
I(s)
=
.
R(s)
1 + G1 (s)G2 (s)H(s)
E2.8

The block diagram is shown in Figure E2.8.
H2(s)
-

R(s)

K
-

E(s)

-

G1(s)

W(s) -

A(s)

G2(s)

Z(s)

1
s

Y(s)

H3(s)

H1(s)

FIGURE E2.8
Block diagram model.

Starting at the output we obtain
Y (s) =

1
1
Z(s) = G2 (s)A(s).
s
s

But A(s) = G1 (s) [−H2 (s)Z(s) − H3 (s)A(s) + W (s)] and Z(s) = sY (s),
so
1
Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s) + G1 (s)G2 (s)W (s).
s
Substituting W (s) = KE(s) − H1 (s)Z(s) into the above equation yields
Y (s) = −G1 (s)G2 (s)H2 (s)Y (s) − G1 (s)H3 (s)Y (s)
1
+ G1 (s)G2 (s) [KE(s) − H1 (s)Z(s)]
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

27

Exercises

and with E(s) = R(s) − Y (s) and Z(s) = sY (s) this reduces to
Y (s) = [−G1 (s)G2 (s) (H2 (s) + H1 (s)) − G1 (s)H3 (s)
1
1
− G1 (s)G2 (s)K]Y (s) + G1 (s)G2 (s)KR(s).
s
s
Solving for Y (s) yields the transfer function
Y (s) = T (s)R(s),
where
T (s) =
E2.9

KG1 (s)G2 (s)/s
.
1 + G1 (s)G2 (s) [(H2 (s) + H1 (s)] + G1 (s)H3 (s) + KG1 (s)G2 (s)/s

From Figure E2.9, we observe that
Ff (s) = G2 (s)U (s)
and
FR (s) = G3 (s)U (s) .
Then, solving for U (s) yields
U (s) =

1
Ff (s)
G2 (s)

FR (s) =

G3 (s)
U (s) .
G2 (s)

and it follows that

Again, considering the block diagram in Figure E2.9 we determine
Ff (s) = G1 (s)G2 (s)[R(s) − H2 (s)Ff (s) − H2 (s)FR (s)] .
But, from the previous result, we substitute for FR (s) resulting in
Ff (s) = G1 (s)G2 (s)R(s)−G1 (s)G2 (s)H2 (s)Ff (s)−G1 (s)H2 (s)G3 (s)Ff (s) .
Solving for Ff (s) yields
G1 (s)G2 (s)
Ff (s) =
R(s) .
1 + G1 (s)G2 (s)H2 (s) + G1 (s)G3 (s)H2 (s)




from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

28

CHAPTER 2

Mathematical Models of Systems

H2(s)
+ -

R(s)

U(s)

G2(s)

Ff (s)

U(s)

G3(s)

FR(s)

G1(s)

-

H2(s)
FIGURE E2.9
Block diagram model.

E2.10

The shock absorber block diagram is shown in Figure E2.10. The closedloop transfer function model is
T (s) =

Gc (s)Gp (s)G(s)
.
1 + H(s)Gc (s)Gp (s)G(s)

Controller

Gear Motor

Plunger and
Piston System

Gc(s)

Gp(s)

G(s)

+
R(s)
Desired piston
travel

-

Y(s)
Piston
travel

Sensor

H(s)

Piston travel
measurement

FIGURE E2.10
Shock absorber block diagram.

E2.11

Let f denote the spring force (n) and x denote the deflection (m). Then
K=

∆f
.
∆x

Computing the slope from the graph yields:
(a) xo = −0.14m → K = ∆f /∆x = 10 n / 0.04 m = 250 n/m
(b) xo = 0m → K = ∆f /∆x = 10 n / 0.05 m = 200 n/m
(c) xo = 0.35m → K = ∆f /∆x = 3n / 0.05 m = 60 n/m

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

29

Exercises

E2.12

The signal flow graph is shown in Fig. E2.12. Find Y (s) when R(s) = 0.

-K

Td(s)
1

1
K2

G(s)
Y (s)

-1
FIGURE E2.12
Signal flow graph.

The transfer function from Td (s) to Y (s) is
Y (s) =

G(s)(1 − K1 K2 )Td (s)
G(s)Td (s) − K1 K2 G(s)Td (s)
=
.
1 − (−K2 G(s))
1 + K2 G(s)

If we set
K1 K2 = 1 ,
then Y (s) = 0 for any Td (s).
E2.13

The transfer function from R(s), Td (s), and N (s) to Y (s) is
K
K
1
R(s)+ 2
Td (s)− 2
N (s)
Y (s) = 2
s + 10s + K
s + 10s + K
s + 10s + K












Therefore, we find that
Y (s)/Td (s) =
E2.14

s2

1
+ 10s + K

and

Y (s)/N (s) = −

s2

K
+ 10s + K

Since we want to compute the transfer function from R2 (s) to Y1 (s), we
can assume that R1 = 0 (application of the principle of superposition).
Then, starting at the output Y1 (s) we obtain
Y1 (s) = G3 (s) [−H1 (s)Y1 (s) + G2 (s)G8 (s)W (s) + G9 (s)W (s)] ,
or
[1 + G3 (s)H1 (s)] Y1 (s) = [G3 (s)G2 (s)G8 (s)W (s) + G3 (s)G9 (s)] W (s).
Considering the signal W (s) (see Figure E2.14), we determine that
W (s) = G5 (s) [G4 (s)R2 (s) − H2 (s)W (s)] ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

30

CHAPTER 2

Mathematical Models of Systems

H1(s)

+

G1(s)

R1(s)

+

G7(s)

R2(s)

G4(s)

-

+

G2(s)

G3(s)
+

Y1(s)

G9(s)

G8(s)

+

+

G6(s)

G5(s)

Y2(s)

W(s)

-

H2(s)
FIGURE E2.14
Block diagram model.

or
[1 + G5 (s)H2 (s)] W (s) = G5 (s)G4 (s)R2 (s).
Substituting the expression for W (s) into the above equation for Y1 (s)
yields
Y1 (s)
G2 (s)G3 (s)G4 (s)G5 (s)G8 (s) + G3 (s)G4 (s)G5 (s)G9 (s)
=
.
R2 (s)
1 + G3 (s)H1 (s) + G5 (s)H2 (s) + G3 (s)G5 (s)H1 (s)H2 (s)
E2.15

For loop 1, we have
di1
1
R1 i1 + L1
+
dt
C1

Z

(i1 − i2 )dt + R2 (i1 − i2 ) = v(t) .

And for loop 2, we have
1
C2
E2.16

Z

di2
1
i2 dt + L2
+ R2 (i2 − i1 ) +
dt
C1

Z

(i2 − i1 )dt = 0 .

The transfer function from R(s) to P (s) is
P (s)
4.2
= 3
.
2
R(s)
s + 2s + 4s + 4.2
The block diagram is shown in Figure E2.16a. The corresponding signal
flow graph is shown in Figure E2.16b for
P (s)/R(s) =

s3

+

4.2
.
+ 4s + 4.2

2s2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

31

Exercises

v1(s)

R(s)

v2(s)

7

-

q(s)

0.6
s

1
s2+2s+4

P(s)

(a)

R(s )

1

V1

7

1
s2 + 2 s + 4

0.6
s

V2

P (s)

-1

(b)
FIGURE E2.16
(a) Block diagram, (b) Signal flow graph.

E2.17

A linear approximation for f is given by
∆f =

∂f
∂x

∆x = 2kxo ∆x = k∆x
x=xo

where xo = 1/2, ∆f = f (x) − f (xo ), and ∆x = x − xo .
E2.18

The linear approximation is given by
∆y = m∆x
where
m=

∂y
∂x

.
x=xo

(a) When xo = 1, we find that yo = 2.4, and yo = 13.2 when xo = 2.
(b) The slope m is computed as follows:
m=

∂y
∂x

= 1 + 4.2x2o .
x=xo

Therefore, m = 5.2 at xo = 1, and m = 18.8 at xo = 2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

32

CHAPTER 2

E2.19

Mathematical Models of Systems

The output (with a step input) is
Y (s) =

15(s + 1)
.
s(s + 7)(s + 2)

The partial fraction expansion is
18 1
3 1
15
−
+
.
14s
7 s+7 2s+2

Y (s) =

Taking the inverse Laplace transform yields
y(t) =
E2.20

15 18 −7t 3 −2t
− e
+ e
.
14
7
2

The input-output relationship is
A(K − 1)
Vo
=
V
1 + AK
where
K=

Z1
.
Z1 + Z2

Assume A ≫ 1. Then,
Vo
K−1
Z2
=
=−
V
K
Z1
where
Z1 =

R1
R1 C 1 s + 1

and Z2 =

R2
.
R2 C 2 s + 1

Therefore,
Vo (s)
R2 (R1 C1 s + 1)
2(s + 1)
=−
=−
.
V (s)
R1 (R2 C2 s + 1)
s+2
E2.21

The equation of motion of the mass mc is
mc ẍp + (bd + bs )ẋp + kd xp = bd ẋin + kd xin .
Taking the Laplace transform with zero initial conditions yields
[mc s2 + (bd + bs )s + kd ]Xp (s) = [bd s + kd ]Xin (s) .
So, the transfer function is
bd s + kd
0.7s + 2
Xp (s)
=
= 2
.
Xin (s)
mc s2 + (bd + bs )s + kd
s + 2.8s + 2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

33

Exercises

E2.22

The rotational velocity is
ω(s) =

2(s + 4)
1
.
2
(s + 5)(s + 1) s

Expanding in a partial fraction expansion yields
ω(s) =

81
1 1
3
13 1
1
+
−
−
.
2
5 s 40 s + 5 2 (s + 1)
8 s+1

Taking the inverse Laplace transform yields
ω(t) =
E2.23

8
1
3
13
+ e−5t − te−t − e−t .
5 40
2
8

The closed-loop transfer function is
Y (s)
K1 K2
= T (s) = 2
.
R(s)
s + (K1 + K2 K3 + K1 K2 )s + K1 K2 K3

E2.24

The closed-loop tranfser function is
Y (s)
10
= T (s) = 2
.
R(s)
s + 21s + 10

E2.25

Let x = 0.6 and y = 0.8. Then, with y = ax3 , we have
0.8 = a(0.6)3 .
Solving for a yields a = 3.704. A linear approximation is
y − yo = 3ax2o (x − xo )
or y = 4x − 1.6, where yo = 0.8 and xo = 0.6.

E2.26

The equations of motion are
m1 ẍ1 + k(x1 − x2 ) = F
m2 ẍ2 + k(x2 − x1 ) = 0 .
Taking the Laplace transform (with zero initial conditions) and solving
for X2 (s) yields
X2 (s) =

(m2

s2

k
F (s) .
+ k)(m1 s2 + k) − k 2

Then, with m1 = m2 = k = 1, we have
X2 (s)/F (s) =

1
.
s2 (s2 + 2)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

34

CHAPTER 2

E2.27

Mathematical Models of Systems

The transfer function from Td (s) to Y (s) is
Y (s)/Td (s) =

E2.28

G2 (s)
.
1 + G1 G2 H(s)

The transfer function is
R2 R4 C
R2 R4
Vo (s)
=
s+
= 24s + 144 .
V (s)
R3
R1 R3

E2.29

(a) If
G(s) =

s2

1
+ 15s + 50

and

H(s) = 2s + 15 ,

then the closed-loop transfer function of Figure E2.28(a) and (b) (in
Dorf & Bishop) are equivalent.
(b) The closed-loop transfer function is
T (s) =

(a) The closed-loop transfer function is
T (s) =

G(s) 1
10
=
2
1 + G(s) s
s(s + 2s + 20)

where G(s) =

0.8
0.7
0.6
Amplitude

E2.30

1
.
s2 + 17s + 65

0.5
0.4
0.3
0.2
0.1
0

0

1

2

3
Time sec

FIGURE E2.30
Step response.

4

5

6

s2

10
.
+ 2s + 10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

35

Exercises

(b) The output Y (s) (when R(s) = 1/s) is
Y (s) =

0.5 −0.25 + 0.0573j
−0.25 − 0.0573j
−
+
.
s
s + 1 − 4.3589j
s + 1 + 4.3589j

(c) The plot of y(t) is shown in Figure E2.30. The output is given by
√
√
1
1
y(t) =
1 − e−t cos 19t − √ sin 19t
2
19




E2.31



The partial fraction expansion is
V (s) =

a
b
+
s + p1 s + p2

where p1 = 4 − 19.6j and p2 = 4 + 19.6j. Then, the residues are
a = −10.2j

b = 10.2j .

The inverse Laplace transform is
v(t) = −10.2je(−4+19.6j)t + 10.2je(−4−19.6j)t = 20.4e−4t sin 19.6t .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

36

CHAPTER 2

Mathematical Models of Systems

Problems
P2.1

The integrodifferential equations, obtained by Kirchoff’s voltage law to
each loop, are as follows:
R1 i1 +

1
C1

R3 i2 +

1
C2

Z

i1 dt + L1

d(i1 − i2 )
+ R2 (i1 − i2 ) = v(t)
dt

(loop 1)

and

P2.2

Z

i2 dt + R2 (i2 − i1 ) + L1

d(i2 − i1 )
=0
dt

(loop 2) .

The differential equations describing the system can be obtained by using
a free-body diagram analysis of each mass. For mass 1 and 2 we have
M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t)
M2 ÿ2 + k12 (y2 − y1 ) = 0 .
Using a force-current analogy, the analagous electric circuit is shown in
Figure P2.2, where Ci → Mi , L1 → 1/k1 , L12 → 1/k12 , and R → 1/b .

FIGURE P2.2
Analagous electric circuit.

P2.3

The differential equations describing the system can be obtained by using
a free-body diagram analysis of each mass. For mass 1 and 2 we have
M ẍ1 + kx1 + k(x1 − x2 ) = F (t)
M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 .
Using a force-current analogy, the analagous electric circuit is shown in
Figure P2.3, where
C→M

L → 1/k

R → 1/b .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

37

Problems

FIGURE P2.3
Analagous electric circuit.

(a) The linear approximation around vin = 0 is vo = 0vin , see Figure P2.4(a).
(b) The linear approximation around vin = 1 is vo = 2vin − 1, see Figure P2.4(b).

(a)

(b)

0.4

4

3.5

0.3

3
0.2
2.5
0.1

2
vo

vo

P2.4

0

1.5

linear approximation
1

-0.1

0.5
-0.2
0
-0.3

-0.4
-1

linear approximation

-0.5

-0.5

0
vin

0.5

FIGURE P2.4
Nonlinear functions and approximations.

1

-1
-1

0

1
vin

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

38

CHAPTER 2

P2.5

Mathematical Models of Systems

Given
Q = K(P1 − P2 )1/2 .
Let δP = P1 − P2 and δPo = operating point. Using a Taylor series
expansion of Q, we have
Q = Qo +

∂Q
∂δP

(δP − δPo ) + · · ·

δP =δPo

where
Qo = KδPo1/2

∂Q
∂δP

and

=
δP =δPo

K −1/2
.
δP
2 o

Define ∆Q = Q − Qo and ∆P = δP − δPo . Then, dropping higher-order
terms in the Taylor series expansion yields
∆Q = m∆P
where
m=
P2.6

K
1/2

2δPo

.

From P2.1 we have
R1 i1 +

1
C1

R3 i2 +

1
C2

Z

i1 dt + L1

d(i1 − i2 )
+ R2 (i1 − i2 ) = v(t)
dt

and
Z

i2 dt + R2 (i2 − i1 ) + L1

d(i2 − i1 )
=0.
dt

Taking the Laplace transform and using the fact that the initial voltage
across C2 is 10v yields
[R1 +

1
+ L1 s + R2 ]I1 (s) + [−R2 − L1 s]I2 (s) = 0
C1 s

and
[−R2 − L1 s]I1 (s) + [L1 s + R3 +

1
10
+ R2 ]I2 (s) = −
.
C2 s
s

Rewriting in matrix form we have



R1 +

1
C1 s

+ L 1 s + R2

−R2 − L1 s

−R2 − L1 s
L 1 s + R3 +

1
C2 s

+ R2




I1 (s)
I2 (s)





=

0
−10/s



 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

39

Problems

Solving for I2 yields










1
0
R2 + L 1 s
1  L1 s + R3 + C2 s + R2

=

 .
1
∆
−10/s
I2 (s)
R2 + L 1 s
R1 + C1 s + L1 s + R2

I1 (s)

or
I2 (s) =

−10(R1 + 1/C1 s + L1 s + R2 )
s∆

where
∆ = (R1 +
P2.7

1
1
+ L1 s + R2 )(L1 s + R3 +
+ R2 ) − (R2 + L1 s)2 .
C1 s
C2 s

Consider the differentiating op-amp circuit in Figure P2.7. For an ideal
op-amp, the voltage gain (as a function of frequency) is
V2 (s) = −

Z2 (s)
V1 (s),
Z1 (s)

where
Z1 =

R1
1 + R1 Cs

and Z2 = R2 are the respective circuit impedances. Therefore, we obtain
V2 (s) = −

Z



R2 (1 + R1 Cs)
V1 (s).
R1


Z

1

C

+
R1

2

R2

+

+

V1(s)

V2(s)

-

-

FIGURE P2.7
Differentiating op-amp circuit.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

40

CHAPTER 2

Let

∆=

G2 + Cs

−Cs

−G2

−Cs

G1 + 2Cs

−Cs

−G2

−Cs

Cs + G2

.

Then,
Vj =

∆ij
I1
∆

or

or

V3
∆13 I1 /∆
=
.
V1
∆11 I1 /∆

Therefore, the transfer function is
−Cs 2Cs + G1
T (s) =

∆13
V3
=
=
V1
∆11

−G2

−Cs

2Cs + G1

−Cs

−Cs

Cs + G2

Pole-zero map (x:poles and o:zeros)
3

2

o

1

Imag Axis

P2.8

Mathematical Models of Systems

0

x

x

-1

-2

-3
-8

o

-7

-6

-5

-4
Real Axis

FIGURE P2.8
Pole-zero map.

-3

-2

-1

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

41

Problems

=

C 2 R1 R2 s2 + 2CR1 s + 1
.
C 2 R1 R2 s2 + (2R1 + R2 )Cs + 1

Using R1 = 0.5, R2 = 1, and C = 0.5, we have
T (s) =

s2 + 4s + 8
(s + 2 + 2j)(s + 2 − 2j)
√
√ .
=
2
s + 8s + 8
(s + 4 + 8)(s + 4 − 8)

The pole-zero map is shown in Figure P2.8.
From P2.3 we have
M ẍ1 + kx1 + k(x1 − x2 ) = F (t)
M ẍ2 + k(x2 − x1 ) + bẋ2 = 0 .
Taking the Laplace transform of both equations and writing the result in
matrix form, it follows that


M s2 + 2k



−k

M s2 + bs + k

−k




X1 (s)
X2 (s)





=

F (s)
0



 ,

Pole zero map
0.4

0.3

0.2

0.1
Imag Axis

P2.9

0

- 0.1

-0.2

-0.3

-0.4
-0.03

FIGURE P2.9
Pole-zero map.

-0.025

-0.02

-0.015
Real Axis

-0.01

-0.005

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

42

CHAPTER 2

Mathematical Models of Systems

or










k
F (s)
1  M s2 + bs + k

=


2
∆
X2 (s)
k
M s + 2k
0
X1 (s)

where ∆ = (M s2 + bs + k)(M s2 + 2k) − k 2 . So,
G(s) =

M s2 + bs + k
X1 (s)
=
.
F (s)
∆

When b/k = 1, M = 1 , b2 /M k = 0.04, we have
G(s) =

s2 + 0.04s + 0.04
.
s4 + 0.04s3 + 0.12s2 + 0.0032s + 0.0016

The pole-zero map is shown in Figure P2.9.
P2.10

From P2.2 we have
M1 ÿ1 + k12 (y1 − y2 ) + bẏ1 + k1 y1 = F (t)
M2 ÿ2 + k12 (y2 − y1 ) = 0 .
Taking the Laplace transform of both equations and writing the result in
matrix form, it follows that


or

M1 s2 + bs + k1 + k12




M2 s2 + k12

−k12


−k12




Y1 (s)
Y2 (s)





=



F (s)



0





k12
F (s)
1  M2 s2 + k12

=


∆
Y2 (s)
k12
M1 s2 + bs + k1 + k12
0
Y1 (s)

where

2
∆ = (M2 s2 + k12 )(M1 s2 + bs + k1 + k12 ) − k12
.

So, when f (t) = a sin ωo t, we have that Y1 (s) is given by
Y1 (s) =

aM2 ωo (s2 + k12 /M2 )
.
(s2 + ωo2 )∆(s)

For motionless response (in the steady-state), set the zero of the transfer
function so that
(s2 +

k12
) = s2 + ωo2
M2

or

ωo2 =

k12
.
M2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

43

Problems

P2.11

The transfer functions from Vc (s) to Vd (s) and from Vd (s) to θ(s) are:
K1 K2
, and
(Lq s + Rq )(Lc s + Rc )
Km
.
θ(s)/Vd (s) =
2
(Js + f s)((Ld + La )s + Rd + Ra ) + K3 Km s

Vd (s)/Vc (s) =

The block diagram for θ(s)/Vc (s) is shown in Figure P2.11, where
θ(s)/Vc (s) =

K1 K2 Km
θ(s) Vd (s)
=
,
Vd (s) Vc (s)
∆(s)

where
∆(s) = s(Lc s + Rc )(Lq s + Rq )((Js + b)((Ld + La )s + Rd + Ra ) + Km K3 ) .

Vc

1
L cs+R c

Ic

K1

Vq

1
L qs+R q

Iq
K2

Vd +

1
(L d+L a)s+R d+R a

Id

Tm
Km

-

1
Js+f

w

1
s

q

Vb
K3

FIGURE P2.11
Block diagram.

P2.12

The open-loop transfer function is
Y (s)
K
=
.
R(s)
s + 20
With R(s) = 1/s, we have
Y (s) =

K
.
s(s + 20)

The partial fraction expansion is
K
Y (s) =
20



1
1
−
,
s s + 20


and the inverse Laplace transform is
y(t) =


K
1 − e−20t ,
20

As t → ∞, it follows that y(t) → K/20. So we choose K = 20 so that y(t)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

44

CHAPTER 2

Mathematical Models of Systems

approaches 1. Alternatively we can use the final value theorem to obtain
y(t)t→∞ = lim sY (s) =
s→0

K
=1.
20

It follows that choosing K = 20 leads to y(t) → 1 as t → ∞.
P2.13

The motor torque is given by
Tm (s) = (Jm s2 + bm s)θm (s) + (JL s2 + bL s)nθL (s)
= n((Jm s2 + bm s)/n2 + JL s2 + bL s)θL (s)
where
n = θL (s)/θm (s) = gear ratio .
But
Tm (s) = Km Ig (s)
and
Ig (s) =

1
Vg (s) ,
(Lg + Lf )s + Rg + Rf

and
Vg (s) = Kg If (s) =

Kg
Vf (s) .
Rf + L f s

Combining the above expressions yields
θL (s)
Kg Km
=
.
Vf (s)
n∆1 (s)∆2 (s)
where
∆1 (s) = JL s2 + bL s +

Jm s2 + bm s
n2

and
∆2 (s) = (Lg s + Lf s + Rg + Rf )(Rf + Lf s) .
P2.14

For a field-controlled dc electric motor we have
ω(s)/Vf (s) =

Km /Rf
.
Js + b

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

45

Problems

With a step input of Vf (s) = 80/s, the final value of ω(t) is
80Km
= 2.4
Rf b

ω(t)t→∞ = lim sω(s) =
s→0

or

Km
= 0.03 .
Rf b

Solving for ω(t) yields
80Km −1
1
ω(t) =
L
Rf J
s(s + b/J)




=

80Km
(1−e−(b/J)t ) = 2.4(1−e−(b/J)t ) .
Rf b

At t = 1/2, ω(t) = 1, so
ω(1/2) = 2.4(1 − e−(b/J)t ) = 1 implies

b/J = 1.08 sec .

Therefore,
ω(s)/Vf (s) =
P2.15

0.0324
.
s + 1.08

Summing the forces in the vertical direction and using Newton’s Second
Law we obtain
ẍ +

k
x=0.
m

The system has no damping and no external inputs. Taking the Laplace
transform yields
X(s) =

s2

x0 s
,
+ k/m

where we used the fact that x(0) = x0 and ẋ(0) = 0. Then taking the
inverse Laplace transform yields
x(t) = x0 cos
P2.16

s

k
t.
m

Using Cramer’s rule, we have


1 1.5

x1





or


2

4





x1
x2





=

6
11







1  4 −1.5   6 

=
∆ −2
x2
1
11

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

46

CHAPTER 2

Mathematical Models of Systems

where ∆ = 4(1) − 2(1.5) = 1 . Therefore,
x1 =

4(6) − 1.5(11)
= 7.5
1

and x2 =

−2(6) + 1(11)
= −1 .
1

The signal flow graph is shown in Figure P2.16.
11
1/4

6

1

-1/2

X2

X1
-1.5

FIGURE P2.16
Signal flow graph.

So,
x1 =
P2.17

6(1) − 1.5( 11
4 )
= 7.5
3
1− 4

and x2 =

11( 41 ) +
1−

−1
2 (6)
3
4

= −1 .

(a) For mass 1 and 2, we have
M1 ẍ1 + K1 (x1 − x2 ) + b1 (ẋ3 − ẋ1 ) = 0
M2 ẍ2 + K2 (x2 − x3 ) + b2 (ẋ3 − ẋ2 ) + K1 (x2 − x1 ) = 0 .
(b) Taking the Laplace transform yields
(M1 s2 + b1 s + K1 )X1 (s) − K1 X2 (s) = b1 sX3 (s)
−K1 X1 (s) + (M2 s2 + b2 s + K1 + K2 )X2 (s) = (b2 s + K2 )X3 (s) .
(c) Let
G1 (s) = K2 + b2 s
G2 (s) = 1/p(s)
G3 (s) = 1/q(s)
G4 (s) = sb1 ,
where
p(s) = s2 M2 + sf2 + K1 + K2
and
q(s) = s2 M1 + sf1 + K1 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

47

Problems

The signal flow graph is shown in Figure P2.17.

G4
G3

X3
G1

G2

X1

K1
K1

FIGURE P2.17
Signal flow graph.

(d) The transfer function from X3 (s) to X1 (s) is
X1 (s)
K1 G1 (s)G2 (s)G3 (s) + G4 (s)G3 (s)
=
.
X3 (s)
1 − K12 G2 (s)G3 (s)
P2.18

The signal flow graph is shown in Figure P2.18.
I1

V1

Va

Z2

Y3

Ia

Z4
V2

Y1
-Z 2

-Y 1

-Y 3

FIGURE P2.18
Signal flow graph.

The transfer function is
V2 (s)
Y 1 Z2 Y 3 Z4
=
.
V1 (s)
1 + Y 1 Z2 + Y 3 Z2 + Y 3 Z4 + Y 1 Z2 Z4 Y 3
P2.19

For a noninerting op-amp circuit, depicted in Figure P2.19a, the voltage
gain (as a function of frequency) is
Vo (s) =

Z1 (s) + Z2 (s)
Vin (s),
Z1 (s)

where Z1 (s) and Z2 (s) are the impedances of the respective circuits. In
the case of the voltage follower circuit, shown in Figure P2.19b, we have

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

48

CHAPTER 2

Mathematical Models of Systems

Z2
Z1

+

vin

v0

+

vin

(a)

v0

(b)

FIGURE P2.19
(a) Noninverting op-amp circuit. (b) Voltage follower circuit.

Z1 = ∞ (open circuit) and Z2 = 0. Therefore, the transfer function is
Vo (s)
Z1
=
= 1.
Vin (s)
Z1
P2.20

(a) Assume Rg ≫ Rs and Rs ≫ R1 . Then Rs = R1 + R2 ≈ R2 , and
vgs = vin − vo ,
where we neglect iin , since Rg ≫ Rs . At node S, we have
vo
= gm vgs = gm (vin − vo ) or
Rs

vo
gm Rs
=
.
vin
1 + gm Rs

(b) With gm Rs = 20, we have
vo
20
=
= 0.95 .
vin
21
(c) The block diagram is shown in Figure P2.20.

vin(s)

gmRs

-

FIGURE P2.20
Block diagram model.

P2.21

From the geometry we find that
∆z = k

l1 − l2
l2
(x − y) − y .
l1
l1

vo(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

49

Problems

The flow rate balance yields
A

dy
= p∆z
dt

which implies

Y (s) =

p∆Z(s)
.
As

By combining the above results it follows that
l2
p
l1 − l2
Y (s) =
k
(X(s) − Y (s)) − Y (s) .
As
l1
l1
 





Therefore, the signal flow graph is shown in Figure P2.21. Using Mason’s
-1
(l 1 - l 2)/l 1

k
X

DZ

p/As
Y

1
-l 2 / l 1

FIGURE P2.21
Signal flow graph.

gain formula we find that the transfer function is given by
Y (s)
=
X(s)
1+

k(l1 −l2 )p
l1 As
k(l1 −l2 )p
l2 p
l1 As +
l1 As

=

K1
,
s + K2 + K1

where
K1 =
P2.22

k(l1 − l2 )p
p
l1 A

and K2 =

l2 p
.
l1 A

(a) The equations of motion for the two masses are
L 2
L
M L θ¨1 + M gLθ1 + k
(θ1 − θ2 ) = f (t)
2
2
 2
L
M L2 θ¨2 + M gLθ2 + k
(θ2 − θ1 ) = 0 .
2
 

2

With θ˙1 = ω1 and θ˙2 = ω2 , we have
g
k
k
f (t)
ω˙1 = −
+
θ1 +
θ2 +
L 4M
4M
2M L


k
g
k
ω˙2 =
θ1 −
+
θ2 .
4M
L 4M




from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

50

CHAPTER 2

Mathematical Models of Systems

a

-

F (t)

w1
1/s

1/s

1/2ML

(a)

q1

b

w2

1/s

1/s

q2

a

Imag(s)
+ j

+ j

(b)

g
k
L + 4M

g
k
L + 2M

X
O

X

+ j

g
L

Re(s)
FIGURE P2.22
(a) Block diagram. (b) Pole-zero map.

(b) Define a = g/L + k/4M and b = k/4M . Then
θ1 (s)
1
s2 + a
=
.
F (s)
2M L (s2 + a)2 − b2
(c) The block diagram and pole-zero map are shown in Figure P2.22.
P2.23

The input-output ratio, Vce /Vin , is found to be
β(R − 1) + hie Rf
Vce
=
.
Vin
−βhre + hie (−hoe + Rf )

P2.24

(a) The voltage gain is given by
vo
RL β1 β2 (R1 + R2 )
.
=
vin
(R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

51

Problems

(b) The current gain is found to be
ic2
= β1 β2 .
ib1
(c) The input impedance is
vin
(R1 + R2 )(Rg + hie1 ) + R1 (R1 + R2 )(1 + β1 ) + R1 RL β1 β2
=
,
ib1
R1 + R2
and when β1 β2 is very large, we have the approximation
vin
RL R1 β1 β2
≈
.
ib1
R1 + R2
P2.25

The transfer function from R(s) and Td (s) to Y (s) is given by


Y (s) = G(s) R(s) −

1
(G(s)R(s) + Td (s)) + Td (s) + G(s)R(s)
G(s)


= G(s)R(s) .
Thus,
Y (s)/R(s) = G(s) .
Also, we have that
Y (s) = 0 .

when R(s) = 0. Therefore, the effect of the disturbance, Td (s), is eliminated.
P2.26

The equations of motion for the two mass model of the robot are
M ẍ + b(ẋ − ẏ) + k(x − y) = F (t)
mÿ + b(ẏ − ẋ) + k(y − x) = 0 .
Taking the Laplace transform and writing the result in matrix form yields



M s2 + bs + k
−(bs + k)

−(bs + k)

ms2

+ bs + k




X(s)
Y (s)





k
m

 .

=

Solving for Y (s) we find that
1
Y (s)
mM (bs +k)
=
m b
F (s)
s2 [s2 + 1 + M
ms +

]

F (s)
0



 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

52

CHAPTER 2

P2.27

Mathematical Models of Systems

The describing equation of motion is
mz̈ = mg − k

i2
.
z2

Defining
f (z, i) = g −

ki2
mz 2

z̈ = f (z, i) .
The equilibrium condition for io and zo , found by solving the equation of
motion when
ż = z̈ = 0 ,
is
ki2o
= zo2 .
mg
We linearize the equation of motion using a Taylor series approximation.
With the definitions
∆z = z − zo

and ∆i = i − io ,

˙ = ż and ∆z
¨ = z̈. Therefore,
we have ∆z
¨ = f (z, i) = f (zo , io ) + ∂f
∆z
∂z

z=zo
i=io

∆z +

∂f
∂i

z=zo
i=io

∆i + · · ·

But f (zo , io ) = 0, and neglecting higher-order terms in the expansion
yields
2
¨ = 2kio ∆z − 2kio ∆i .
∆z
mzo3
mzo2

Using the equilibrium condition which relates zo to io , we determine that
¨ = 2g ∆z − g ∆i .
∆z
zo
io
Taking the Laplace transform yields the transfer function (valid around
the equilibrium point)
∆Z(s)
−g/io
= 2
.
∆I(s)
s − 2g/zo

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

57

Problems

R(s )

+

K1
s (s+1)

Y (s)

1 +K 2s

1
0.9
0.8
0.7
<---- time to 90% = 0.39 sec

y(t)

0.6
0.5
0.4
0.3
0.2
0.1
0
0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

time(sec)

FIGURE P2.35
The equivalent block diagram and the system step response.

P2.36

(a) Given R(s) = 1/s2 , the partial fraction expansion is
Y (s) =

s2 (s

24
3
8/3
3/4
1 13/12
=
−
+
+ 2−
.
+ 2)(s + 3)(s + 4)
s+2 s+3 s+4 s
s

Therefore, using the Laplace transform table, we determine that the
ramp response is
8
3
13
y(t) = 3e−2t − e−3t + e−4t + t −
,
3
4
12

t≥0.

(b) For the ramp input, y(t) ≈ 0.21 at t = 1. second (see Figure P2.36a).

(c) Given R(s) = 1, the partial fraction expansion is
Y (s) =

24
12
24
12
=
−
+
.
(s + 2)(s + 3)(s + 4)
s+2 s+3 s+4

Therefore, using the Laplace transform table, we determine that the
impulse response is
y(t) = 12e−2t − 24e−3t + 412e−4t ,

t≥0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

53

Problems

P2.28

The signal flow graph is shown in Figure P2.28.
-d
G

B
+b

P

+c

D

+a
-m
+e

-k

M

+g
+f

+h

S

C

FIGURE P2.28
Signal flow graph.

(a) The PGBDP loop gain is equal to -abcd. This is a negative transmission since the population produces garbage which increases bacteria
and leads to diseases, thus reducing the population.
(b) The PMCP loop gain is equal to +efg. This is a positive transmission since the population leads to modernization which encourages
immigration, thus increasing the population.
(c) The PMSDP loop gain is equal to +ehkd. This is a positive transmission since the population leads to modernization and an increase
in sanitation facilities which reduces diseases, thus reducing the rate
of decreasing population.
(d) The PMSBDP loop gain is equal to +ehmcd. This is a positive
transmission by similar argument as in (3).
P2.29

Assume the motor torque is proportional to the input current
Tm = ki .
Then, the equation of motion of the beam is
J φ̈ = ki ,
where J is the moment of inertia of the beam and shaft (neglecting the
inertia of the ball). We assume that forces acting on the ball are due to
gravity and friction. Hence, the motion of the ball is described by
mẍ = mgφ − bẋ

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

54

CHAPTER 2

Mathematical Models of Systems

where m is the mass of the ball, b is the coefficient of friction, and we
have assumed small angles, so that sin φ ≈ φ. Taking the Laplace transfor
of both equations of motion and solving for X(s) yields
X(s)/I(s) =
P2.30

gk/J
.
+ b/m)

s2 (s2

Given
H(s) =

k
τs + 1

where τ = 4µs = 4 × 10−6 seconds and 0.999 ≤ k < 1.001. The step
response is
Y (s) =

k
1
k
k
· = −
.
τs + 1 s
s s + 1/τ

Taking the inverse Laplace transform yields
y(t) = k − ke−t/τ = k(1 − e−t/τ ) .
The final value is k. The time it takes to reach 98% of the final value is
t = 15.6µs independent of k.
P2.31

From the block diagram we have
Y1 (s) = G2 (s)[G1 (s)E1 (s) + G3 (s)E2 (s)]
= G2 (s)G1 (s)[R1 (s) − H1 (s)Y1 (s)] + G2 (s)G3 (s)E2 (s) .
Therefore,
Y1 (s) =

G1 (s)G2 (s)
G2 (s)G3 (s)
R1 (s) +
E2 (s) .
1 + G1 (s)G2 (s)H1 (s)
1 + G1 (s)G2 (s)H1 (s)

And, computing E2 (s) (with R2 (s) = 0) we find
G4 (s)
E2 (s) = H2 (s)Y2 (s) = H2 (s)G6 (s)
Y1 (s) + G5 (s)E2 (s)
G2 (s)


or
E2 (s) =

G4 (s)G6 (s)H2 (s)
Y1 (s) .
G2 (s)(1 − G5 (s)G6 (s)H2 (s))

Substituting E2 (s) into equation for Y1 (s) yields
Y1 (s) =

G1 (s)G2 (s)
R1 (s)
1 + G1 (s)G2 (s)H1 (s)



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

55

Problems

+

G3 (s)G4 (s)G6 (s)H2 (s)
Y1 (s) .
(1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s))

Finally, solving for Y1 (s) yields
Y1 (s) = T1 (s)R1 (s)
where
T1 (s) =


G1 (s)G2 (s)(1 − G5 (s)G6 (s)H2 (s))
(1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s)



.



.

Similarly, for Y2 (s) we obtain
Y2 (s) = T2 (s)R1 (s) .
where
T2 (s) =


P2.32

G1 (s)G4 (s)G6 (s)
(1 + G1 (s)G2 (s)H1 (s))(1 − G5 (s)G6 (s)H2 (s)) − G3 (s)G4 (s)G6 (s)H2 (s)

The signal flow graph shows three loops:
L1 = −G1 G3 G4 H2
L2 = −G2 G5 G6 H1
L3 = −H1 G8 G6 G2 G7 G4 H2 G1 .
The transfer function Y2 /R1 is found to be
Y2 (s)
G1 G8 G6 ∆1 − G2 G5 G6 ∆2
=
,
R1 (s)
1 − (L1 + L2 + L3 ) + (L1 L2 )
where for path 1
∆1 = 1
and for path 2
∆ 2 = 1 − L1 .
Since we want Y2 to be independent of R1 , we need Y2 /R1 = 0. Therefore,
we require
G1 G8 G6 − G2 G5 G6 (1 + G1 G3 G4 H2 ) = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

56

CHAPTER 2

P2.33

Mathematical Models of Systems

The closed-loop transfer function is
G3 (s)G1 (s)(G2 (s) + K5 K6 )
Y (s)
=
.
R(s)
1 − G3 (s)(H1 (s) + K6 ) + G3 (s)G1 (s)(G2 (s) + K5 K6 )(H2 (s) + K4 )

P2.34

The equations of motion are
m1 ÿ1 + b(ẏ1 − ẏ2 ) + k1 (y1 − y2 ) = 0
m2 ÿ2 + b(ẏ2 − ẏ1 ) + k1 (y2 − y1 ) + k2 y2 = k2 x
Taking the Laplace transform yields
(m1 s2 + bs + k1 )Y1 (s) − (bs + k1 )Y2 (s) = 0
(m2 s2 + bs + k1 + k2 )Y2 (s) − (bs + k1 )Y1 (s) = k2 X(s)
Therefore, after solving for Y1 (s)/X(s), we have
Y2 (s)
k2 (bs + k1 )
=
.
2
X(s)
(m1 s + bs + k1 )(m2 s2 + bs + k1 + k2 ) − (bs + k1 )2

P2.35

(a) We can redraw the block diagram as shown in Figure P2.35. Then,
T (s) =

K1 /s(s + 1)
K1
= 2
.
1 + K1 (1 + K2 s)/s(s + 1)
s + (1 + K2 K1 )s + K2

(b) The signal flow graph reveals two loops (both touching):
L1 =

−K1
s(s + 1)

and

L2 =

−K1 K2
.
s+1

Therefore,
T (s) =

K1 /s(s + 1)
K1
= 2
.
1 + K1 /s(s + 1) + K1 K2 /(s + 1)
s + (1 + K2 K1 )s + K1

(c) We want to choose K1 and K2 such that
s2 + (1 + K2 K1 )s + K1 = s2 + 20s + 100 = (s + 10)2 .
Therefore, K1 = 100 and 1 + K2 K1 = 20 or K2 = 0.19.
(d) The step response is shown in Figure P2.35.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

58

CHAPTER 2

Mathematical Models of Systems

(d) For the impulse input, y(t) ≈ 0.65 at t = 1 seconds (see Figure P2.36b).
(a) Ramp input

(b) Impulse input
0.8

2

1.8

0.7

1.6
0.6
1.4
0.5

y(t)

y(t)

1.2

1

0.8

0.4

0.3

0.6
0.2
0.4
0.1

0.2

0

0

1

2

0

3

0

1

Time (sec)

2

3

Time (sec)

FIGURE P2.36
(a) Ramp input response. (b) Impulse input response.

P2.37

The equations of motion are
m1

d2 x
= −(k1 + k2 )x + k2 y
dt2

and

m2

d2 y
= k2 (x − y) + u .
dt2

When m1 = m2 = 1 and k1 = k2 = 1, we have
d2 x
= −2x + y
dt2
P2.38

and

d2 y
=x−y+u .
dt2

The equation of motion for the system is
J

d2 θ
dθ
+ b + kθ = 0 ,
dt2
dt

where k is the rotational spring constant and b is the viscous friction
coefficient. The initial conditions are θ(0) = θo and θ̇(0) = 0. Taking the

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

59

Problems

Laplace transform yields
J(s2 θ(s) − sθo ) + b(sθ(s) − θo ) + kθ(s) = 0 .
Therefore,
θ(s) =

(s + Jb θo )
(s2

+

b
Js

+

K
J)

=

(s + 2ζωn )θo
.
+ 2ζωn s + ωn2

s2

Neglecting the mass of the rod, the moment of inertia is detemined to be
J = 2M r 2 = 0.5 kg · m2 .
Also,
s

ωn =

k
J

and ζ =

b
= 0.01 .
2Jωn

Solving for θ(t), we find that
q
θo
θ(t) = p
e−ζωn t sin(ωn 1 − ζ 2 t + φ) ,
1 − ζ2

where tan φ =

p

1 − ζ 2 /ζ). Therefore, the envelope decay is
θo
θe = p
e−ζωn t .
1 − ζ2

So, with ζωn = 2 × 10−4 , θo = 4000o and θf = 10o , the elapsed time is
computed as
t=
P2.39

1
θo
ln p
= 8.32 hours .
ζωn
1 − ζ 2 θf

When t < 0, we have the steady-state conditions
i1 (0) = 1A ,

va (0) = 2V

and

vc (0) = 5V ,

where vc (0) is associated with the 1F capacitor. After t ≥ 0, we have
2

di1
+ 2i1 + 4(i1 − i2 ) = 10e−2t
dt

and
Z

i2 dt + 10i2 + 4(i2 − i1 ) − i1 = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

60

CHAPTER 2

Mathematical Models of Systems

Taking the Laplace transform (using the initial conditions) yields
2(sI1 − i1 (0)) + 2I1 + 4I1 − 4I2 =

10
s+2

or

(s + 3)I1 (s) − 2I2 (s) =

s+7
s+2

and
1
[ I2 −vc (0)]+10I2 +4(I2 −I1 ) = I1 (s) or
s

−5sI1 (s)+(14s+1)I2 (s) = 5s .

Solving for I2 (s) yields
I2 =

5s(s2 + 6s + 13)
,
14(s + 2)∆(s)

where
∆(s) =

s+3

−2

−5s

14s + 1

= 14s2 + 33s + 3 .

Then,
Vo (s) = 10I2 (s) .
P2.40

The equations of motion are
J1 θ̈1 = K(θ2 − θ1 ) − b(θ̇1 − θ̇2 ) + T

and J2 θ̈2 = b(θ̇1 − θ̇2 ) .

Taking the Laplace transform yields
(J1 s2 + bs + K)θ1 (s) − bsθ2 (s) = Kθ2 (s) + T (s)
and
(J2 s2 + bs)θ2 (s) − bsθ1 (s) = 0 .
Solving for θ1 (s) and θ2 (s), we find that
θ1 (s) =

(Kθ2 (s) + T (s))(J2 s + b)
∆(s)

and

θ2 (s) =

b(Kθ2 (s) + T (s))
,
∆(s)

where
∆(s) = J1 J2 s3 + b(J1 + J2 )s2 + J2 Ks + bK .
P2.41

Assume that the only external torques acting on the rocket are control
torques, Tc and disturbance torques, Td , and assume small angles, θ(t).
Using the small angle approximation, we have
ḣ = V θ

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

61

Problems

J θ̈ = Tc + Td ,
where J is the moment of inertia of the rocket and V is the rocket velocity
(assumed constant). Now, suppose that the control torque is proportional
to the lateral displacement, as
Tc (s) = −KH(s) ,
where the negative sign denotes a negative feedback system. The corresponding block diagram is shown in Figure P2.41.
Td

H desired=0

K
+

Tc

+
+

-

1
Js 2

V
s

H( s)

FIGURE P2.41
Block diagram.

P2.42

(a) The equation of motion of the motor is
J

dω
= Tm − bω ,
dt

where J = 0.1, b = 0.06, and Tm is the motor input torque.
(b) Given Tm (s) = 1/s, and ω(0) = 0.7, we take the Laplace transform
of the equation of motion yielding
sω(s) − ω(0) + 0.6ω(s) = 10Tm
or
ω(s) =

0.7s + 10
.
s(s + 0.6)

Then, computing the partial fraction expansion, we find that
ω(s) =

A
B
16.67
15.97
+
=
−
.
s
s + 0.6
s
s + 0.6

The step response, determined by taking the inverse Laplace transform, is
ω(t) = 16.67 − 15.97e−0.6t ,

t≥0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

62

CHAPTER 2

P2.43

Mathematical Models of Systems

The work done by each gear is equal to that of the other, therefore
Tm θm = TL θL .
Also, the travel distance is the same for each gear, so
r1 θ m = r2 θ L .
The number of teeth on each gear is proportional to the radius, or
r1 N 2 = r2 N 1 .
So,
θm
r2
N2
=
=
,
θL
r1
N1
and
N1 θ m = N2 θ L
N1
θL =
θm = nθm ,
N2
where
n = N1 /N2 .
Finally,
Tm
θL
N1
=
=
=n.
TL
θm
N2

P2.44

The inertia of the load is
JL =

πρLr 4
.
2

Also, from the dynamics we have
T2 = JL ω̇2 + bL ω2
and
T1 = nT2 = n(JL ω̇2 + bL ω2 ) .
So,
T1 = n2 (JL ω̇1 + bL ω1 ) ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

63

Problems

since
ω2 = nω1 .
Therefore, the torque at the motor shaft is
T = T1 + Tm = n2 (JL ω̇1 + bL ω1 ) + Jm ω̇1 + bm ω1 .
P2.45

Let U (s) denote the human input and F (s) the load input. The transfer
function is
P (s) =

G(s) + KG1 (s)
Gc (s) + KG1 (s)
U (s) +
F (s) ,
∆(s)
∆(s)

where
∆ = 1 + GH(s) + G1 KBH(s) + Gc E(s) + G1 KE(s) .
P2.46

Consider the application of Newton’s law (
mv we obtain

P

F = mẍ). From the mass

mv ẍ1 = F − k1 (x1 − x2 ) − b1 (ẋ1 − ẋ2 ).
Taking the Laplace transform, and solving for X1 (s) yields
X1 (s) =

b1 s + k1
1
F (s) +
X2 (s),
∆1 (s)
∆1 (s)

where
∆1 := mv s2 + b1 s + k1 .
From the mass mt we obtain
mt ẍ2 = −k2 x2 − b2 ẋ2 + k1 (x1 − x2 ) + b1 (ẋ1 − ẋ2 ).
Taking the Laplace transform, and solving for X2 (s) yields
X2 (s) =

b1 s + k1
X1 (s),
∆2 (s)

where
∆2 := mt s2 + (b1 + b2 )s + k1 + k2 .
Substituting X2 (s) above into the relationship fpr X1 (s) yields the transfer function
∆2 (s)
X1 (s)
=
.
F (s)
∆1 (s)∆2 (s) − (b1 s + k1 )2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

64

CHAPTER 2

P2.47

Mathematical Models of Systems

Using the following relationships
h(t) =

Z

(1.6θ(t) − h(t))dt

ω(t) = θ̇(t)
J ω̇(t) = Km ia (t)
va (t) = 50vi (t) = 10ia (t) + vb (t)
θ̇ = Kvb
we find the differential equation is
d3 h
Km
+ 1+
3
dt
10JK


P2.48



Km dh
8Km
d2 h
+
=
vi .
2
dt
10JK dt
J

(a) The transfer function is
V2 (s)
(1 + sR1 C1 )(1 + sR2 C2 )
=
.
V1 (s)
R1 C 2 s
(b) When R1 = 100 kΩ, R2 = 200 kΩ, C1 = 1 µF and C2 = 0.1 µF , we
have
V2 (s)
0.2(s + 10)(s + 50)
=
.
V1 (s)
s

P2.49

(a) The closed-loop transfer function is
T (s) =

G(s)
6205
= 3
.
1 + G(s)
s + 13s2 + 1281s + 6205

(b) The poles of T (s) are s1 = −5 and s2,3 = −4 ± j35.
(c) The partial fraction expansion (with a step input) is
Y (s) = 1 −

1.0122 0.0061 + 0.0716j
0.0061 − 0.0716j
+
+
.
s+5
s + 4 + j35
s + 4 − j35

(d) The step response is shown in Figure P2.49. The real and complex
roots are close together and by looking at the poles in the s-plane we
have difficulty deciding which is dominant. However, the residue at
the real pole is much larger and thus dominates the response.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

65

Problems

1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Time (secs)

FIGURE P2.49
Step response.

P2.50

(a) The closed-loop transfer function is
T (s) =

s3

+

45s2

14000
.
+ 3100s + 14500

(b) The poles of T (s) are
s1 = −5 and

s2,3 = −20 ± j50.

(c) The partial fraction expansion (with a step input) is
Y (s) =

0.9655 1.0275 0.0310 − 0.0390j
0.0310 + 0.0390j
−
+
+
.
s
s+5
s + 20 + j50
s + 20 − j50

(d) The step response is shown in Figure P2.50. The real root dominates
the response.
(e) The final value of y(t) is
yss = lim sY (s) = 0.9655 .
s→0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 2

Mathematical Models of Systems

1
0.9
0.8
0.7

Amplitude

66

0.6
0.5
0.4
0.3
0.2
0.1
0
0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Time (secs)

FIGURE P2.50
Step response.

P2.51

Consider the free body diagram in Figure P2.51. Using Newton’s Law
and summing the forces on the two masses yields
M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t)
M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t)

k1x

M1

k1

x
. .
b1(x - y)

k2

M1
x

. .
b1(y - x) k2 y

b1
M2

M2

y

y
u(t)

FIGURE P2.51
Free body diagram.

u(t)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

67

AP2.1

The transfer function from V (s) to ω(s) has the form
ω(s)
Km
=
.
V (s)
τm s + 1
ωss = lim s
s→0



Km
5
= 5Km .
τm s + 1 s


So,
Km = 70/5 = 14 .
Also,
ω(t) = Vm Km (1 − e−t/τm )
where V (s) = Vm /s. Solving for τm yields
τm =

−t
.
ln(1 − ω(t)/ωss )

When t = 2, we have
τm =

−2
= 3.57 .
ln(1 − 30/70)

Therefore, the transfer function is
ω(s)
14
=
.
V (s)
3.57s + 1
AP2.2

The closed-loop transfer function form R1 (s) to Y2 (s) is
Y2 (s)
G1 G4 G5 (s) + G1 G2 G3 G4 G6 (s)
=
R1 (s)
∆
where
∆ = [1 + G3 G4 H2 (s)][1 + G1 G2 H3 (s)] .
If we select
G5 (s) = −G2 G3 G6 (s)
then the numerator is zero, and Y2 (s)/R1 (s) = 0. The system is now
decoupled.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

68

CHAPTER 2

AP2.3

Mathematical Models of Systems

(a) Computing the closed-loop transfer function:
G(s)Gc (s)
R(s) .
Y (s) =
1 + Gc (s)G(s)H(s)




Then, with E(s) = R(s) − Y (s) we obtain
E(s) =



1 + Gc (s)G(s)(H(s) − 1)
R(s) .
1 + Gc (s)G(s)H(s)


If we require that E(s) ≡ 0 for any input, we need 1 + Gc (s)G(s)(H(s) −
1) = 0 or
H(s) =

Gc (s)G(s) − 1
n(s)
=
.
Gc (s)G(s)
d(s)

Since we require H(s) to be a causal system, the order of the numerator
polynomial, n(s), must be less than or equal to the order of the denominator polynomial, d(s). This will be true, in general, only if both Gc (s)
and G(s) are proper rational functions (that is, the numerator and denominator polynomials have the same order). Therefore, making E ≡ 0
for any input R(s) is possible only in certain circumstances.
(b) The transfer function from Td (s) to Y (s) is
Gd (s)G(s)
Y (s) =
Td (s) .
1 + Gc (s)G(s)H(s)




With H(s) as in part (a) we have
Gd (s)
Y (s) =
Td (s) .
Gc (s)




(c) No. Since
Y (s) =



Gd (s)G(s)
Td (s) = T (s)Td (s) ,
1 + Gc (s)G(s)H(s)


the only way to have Y (s) ≡ 0 for any Td (s) is for the transfer function
T (s) ≡ 0 which is not possible in general (since G(s) 6= 0).
AP2.4

(a) With q(s) = 1/s we obtain
τ (s) =

1/Ct
s+

QS+1/R
Ct

·

1
.
s

Define
α :=

QS + 1/R
Ct

and

β := 1/Ct .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

69

Then, it follows that
τ (s) =

β
1
−β/α β/α
· =
+
.
s+α s
s+α
s

Taking the inverse Laplace transform yields
τ (t) =

β
−β −αt β
e
+ = [1 − e−αt ] .
α
α
α

1
(b) As t → ∞, τ (t) → αβ = Qs+1/R
.
(c) To increase the speed of response, you want to choose Ct , Q, S and
R such that

α :=

Qs + 1/R
Ct

is ”large.”
AP2.5

Considering the motion of each mass, we have
M3 ẍ3 + b3 ẋ3 + k3 x3 = u3 + b3 ẋ2 + k3 x2
M2 ẍ2 + (b2 + b3 )ẋ2 + (k2 + k3 )x2 = u2 + b3 ẋ3 + k3 x3 + b2 ẋ1 + k2 x1
M1 ẍ1 + (b1 + b2 )ẋ1 + (k1 + k2 )x1 = u1 + b2 ẋ2 + k2 x2
In matrix form the three equations can be written as


0
 M1

 0
M2



AP2.6

0

0











0



0

0   ẍ1   b1 + b2
−b2
0

 

 
0 
b2 + b3 −b3
  ẍ2  +  −b2

M3



ẍ3



−b3

b3

−k2
0
 k1 + k2

+
k2 + k3 −k3
 −k2
−k3

k3





  ẋ1 


  ẋ 
 2 


ẋ3















  x 1   u1 

 

 x  =  u  .
 2   2 


x3

u3

Considering the cart mass and using Newton’s Law we obtain
M ẍ = u − bẋ − F sin ϕ
where F is the reaction force between the cart and the pendulum. Considering the pendulum we obtain
m

d2 (x + L sin ϕ)
= F sin ϕ
dt2



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

70

CHAPTER 2

Mathematical Models of Systems

m

d2 (L cos ϕ)
= F cos ϕ + mg
dt2

Eliminating the reaction force F yields the two equations
(m + M )ẍ + bẋ + mLϕ̈ cos ϕ − mLϕ̇2 sin ϕ = u
mL2 ϕ̈ + mgL sin ϕ + mLẍ cos ϕ = 0
If we assume that the angle ϕ ≈ 0, then we have the linear model
(m + M )ẍ + bẋ + mLϕ̈ = u
mL2 ϕ̈ + mgLϕ = −mLẍ
AP2.7

The transfer function from the disturbance input to the output is
Y (s) =

1
Td (s) .
s + 20 + K

When Td (s) = 1, we obtain
y(t) = e−(20+K)t .
Solving for t when y(t) < 0.1 yields
t>

2.3
.
20 + K

When t = 0.05 and y(0.05) = 0.1, we find K = 26.05.
AP2.8

The closed-loop transfer function is
T (s) =

200K(0.25s + 1)
(0.25s + 1)(s + 1)(s + 8) + 200K

The final value due to a step input of R(s) = A/s is
v(t) → A

200K
.
200K + 8

We need to select K so that v(t) → 50. However, to keep the percent
overshoot to less than 10%, we need to limit the magnitude of K. Figure AP2.8a shows the percent overshoot as a function of K. Let K = 0.06
and select the magnitude of the input to be A = 83.3. The inverse Laplace
transform of the closed-loop response with R(s) = 83.3/s is
v(t) = 50 + 9.85e−9.15t − e−1.93t (59.85 cos(2.24t) + 11.27 sin(2.24t))

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

71

The result is P.O. = 9.74% and the steady-state value of the output is
approximately 50 m/s, as shown in Figure AP2.8b.

25

Percent Overshoot (%)

20

15

10

5

0

0

0.01

0.02

0.03

0.04

0.05
K

0.06

0.07

0.08

0.09

0.1

Step Response
60

System: untitled1
Peak amplitude: 54.9
Overshoot (%): 9.74
At time (sec): 1.15

50

Amplitude

40

30

20

10

0

0

0.5

1

1.5
Time (sec)

FIGURE AP2.8
(a) Percent overshoot versus the gain K. (b) Step response.

AP2.9

The transfer function is
Vo (s)
Z2 (s)
=−
,
Vi (s)
Z1 (s)

2

2.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

72

CHAPTER 2

Mathematical Models of Systems

where
Z1 (s) =

R1
R1 C 1 s + 1

and Z2 (s) =

R2 C 2 s + 1
.
C2 s

Then we can write
KI
Vo (s)
= Kp +
+ KD s
Vi (s)
s
where
KP = −



R1 C 1
+1 ,
R2 C 2


KI = −

1
,
R1 C 2

KD = −R2 C1 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

73

Design Problems

Design Problems
CDP2.1

The model of the traction drive, capstan roller, and linear slide follows
closely the armature-controlled dc motor model depicted in Figure 2.18
in Dorf and Bishop. The transfer function is
T (s) =

rKm
,
s [(Lm s + Rm )(JT s + bm ) + Kb Km ]

where
JT = Jm + r 2 (Ms + Mb ) .

Va(s)
-

1
JTs+bm

Km
Lms+Rm

1
s

q

r

X(s)

Kb

Back EMF

DP2.1

w

The closed-loop transfer function is
Y (s)
G1 (s)G2 (s)
=
.
R(s)
1 + G1 (s)H1 (s) − G2 (s)H2 (s)
When G1 H1 = G2 H2 and G1 G2 = 1, then Y (s)/R(s) = 1. Therefore,
select
G1 (s) =

DP2.2

1
G2 (s)

and H1 (s) =

G2 (s)H2 (s)
= G22 (s)H2 (s) .
G1 (s)

At the lower node we have
1 1
v
+ + G + 2i2 − 20 = 0 .
4 3




Also, we have v = 24 and i2 = Gv . So
1 1
v
+ + G + 2Gv − 20 = 0
4 3




from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

74

CHAPTER 2

Mathematical Models of Systems

and
G=
DP2.3

20 − v



1
4

+

1
3

3v



=

1
S.
12

Taking the Laplace transform of
3 1
1
y(t) = e−t − e−2t − + t
4
4 2
yields
Y (s) =

1
1
3
1
−
−
+ 2 .
s + 1 4(s + 2) 4s 2s

Similarly, taking the Laplace transform of the ramp input yields
R(s) =

1
.
s2

Therefore
G(s) =
DP2.4

Y (s)
1
=
.
R(s)
(s + 1)(s + 2)

For an ideal op-amp, at node a we have
vin − va vo − va
+
=0,
R1
R1
and at node b
vin − vb
= C v̇b ,
R2
from it follows that


1
1
+ Cs Vb =
Vin .
R2
R2


Also, for an ideal op-amp, Vb − Va = 0. Then solving for Vb in the above
equation and substituting the result into the node a equation for Va yields
Vo
=
Vin

1
R2

"

2
1
−
R
+ Cs
2

1
R2

+ Cs
2

or
Vo (s)
R2 Cs − 1
=−
.
Vin (s)
R2 Cs + 1

#

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

75

Design Problems

For vin (t) = At, we have Vin (s) = A/s2 , therefore
2
2
vo (t) = A e−βt + t −
β
β




where β = 1/R2 C.
DP2.5

The equation of motion describing the motion of the inverted pendulum
(assuming small angles) is
ϕ̈ +

g
ϕ=0.
L

Assuming a solution of the form ϕ = k cos ϕ, taking the appropriate
derivatives and substituting the result into the equation of motion yields
the relationship
ϕ̇ =

r

g
.
L

If the period is T = 2 seconds, we compute ϕ̇ = 2π/T . Then solving for L
yields L = 0.99 meters when g = 9.81 m/s2 . So, to fit the pendulum into
the grandfather clock, the dimensions are generally about 1.5 meters or
more.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

76

CHAPTER 2

Mathematical Models of Systems

Computer Problems
CP2.1

The m-file script is shown in Figure CP2.1.
pq =
1 9
P=
-5
-2
Z=
-2
value =
4

p=[1 7 10]; q=[1 2];
% Part (a)
pq=conv(p,q)
% Part (b)
P=roots(p), Z=roots(q)
% Part (c)
value=polyval(p,-1)

24

20

FIGURE CP2.1
Script for various polynomial evaluations.

The m-file script and step response is shown in Figure CP2.2.
numc = [1]; denc = [1 1]; sysc = tf(numc,denc)
numg = [1 2]; deng = [1 3]; sysg = tf(numg,deng)
% part (a)
sys_s = series(sysc,sysg);
sys_cl = feedback(sys_s,[1])
% part (b)
step(sys_cl); grid on

Transfer function:
s+2
------------s^2 + 5 s + 5

Step Response
From: U(1)
0.4

0.35

0.3

To: Y(1)

0.25

Amplitude

CP2.2

0.2

0.15

0.1

0.05

0

0

0.5

1

1.5

2

Time (sec.)

FIGURE CP2.2
Step response.

2.5

3

3.5

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

77

Computer Problems

Given
ÿ + 4ẏ + 3y = u
with y(0) = ẏ = 0 and U (s) = 1/s, we obtain (via Laplace transform)
Y (s) =

s(s2

1
1
=
.
+ 4s + 3)
s(s + 3)(s + 1)

Expanding in a partial fraction expansion yields
Y (s) =

1
1
1
−
−
.
3s 6(s + 3) 2(s + 1)

Taking the inverse Laplace transform we obtain the solution
y(t) = 0.3333 + 0.1667e−3t − 0.5e−t .
The m-file script and step response is shown in Figure CP2.3.

Step Response
0.35

0.3

0.25

Amplitude

CP2.3

n=[1]; d=[1 4 3]; sys = tf(n,d);
t=[0:0.1:5];
y = step(sys,t);
ya=0.3333+0.1667*exp(-3*t)-0.5*exp(-t);
plot(t,y,t,ya); grid;
title('Step Response');
xlabel('Time (sec)');
ylabel('Amplitude');

0.2

0.15

0.1

0.05

0

0

FIGURE CP2.3
Step response.

0.5

1

1.5

2

2.5
3
Time (sec)

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

78

CHAPTER 2

CP2.4

Mathematical Models of Systems

The mass-spring-damper system is represented by
mẍ + bẋ + kx = f .
Taking the Laplace transform (with zero initial conditions) yields the
transfer function
X(s)/F (s) =

s2

1/m
.
+ bs/m + k/m

The m-file script and step response is shown in Figure CP2.4.
m=10; k=1; b=0.5;
num=[1/m]; den=[1 b/m k/m];
sys = tf(num,den);
t=[0:0.1:150];
step(sys,t)
Step Response
From: U(1)
1.8

1.6

1.4

1
To: Y(1)

Amplitude

1.2

0.8

0.6

0.4

0.2

0

0

50

100

150

Time (sec.)

FIGURE CP2.4
Step response.

CP2.5

The spacecraft simulations are shown in Figure CP2.5. We see that as J
is decreased, the time to settle down decreases. Also, the overhoot from
10o decreases as J decreases. Thus, the performance seems to get better
(in some sense) as J decreases.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

79

Computer Problems
Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted)
18
16

Spacecraft attitude (deg)

14
12
10
8
6
4
2
0
0

10

20

30

40

50

60

70

80

90

100

Time (sec)

%Part (a)
a=1; b=8; k=10.8e+08; J=10.8e+08;
num=k*[1 a];
den=J*[1 b 0 0]; sys=tf(num,den);
sys_cl=feedback(sys,[1]);
%
% Part (b) and (c)
t=[0:0.1:100];
%
% Nominal case
f=10*pi/180; sysf=sys_cl*f ;
y=step(sysf,t);
%
% Off-nominal case 80%
J=10.8e+08*0.8; den=J*[1 b 0 0];
sys=tf(num,den); sys_cl=feedback(sys,[1]);
sysf=sys_cl*f ;
y1=step(sysf,t);
%
% Off-nominal case 50%
J=10.8e+08*0.5; den=J*[1 b 0 0];
sys=tf(num,den); sys_cl=feedback(sys,[1]);
sysf=sys_cl*f ;
y2=step(sysf,t);
%
plot(t,y*180/pi,t,y1*180/pi,'--',t,y2*180/pi,':'),grid
xlabel('Time (sec)')
ylabel('Spacecraft attitude (deg)')
title('Nominal (solid); Off-nominal 80% (dashed); Off-nominal 50% (dotted)')

FIGURE CP2.5
Step responses for the nominal and off-nominal spacecraft parameters.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

80

CHAPTER 2

CP2.6

Mathematical Models of Systems

The closed-loop transfer function is
T (s) =

4s6 + 8s5 + 4s4 + 56s3 + 112s2 + 56s
,
∆(s)
p=
7.0709
-7.0713
1.2051 + 2.0863i
1.2051 - 2.0863i
0.1219 + 1.8374i
0.1219 - 1.8374i
-2.3933
-2.3333
-0.4635 + 0.1997i
-0.4635 - 0.1997i

num1=[4]; den1=[1]; sys1 = tf(num1,den1);
num2=[1]; den2=[1 1]; sys2 = tf(num2,den2);
num3=[1 0]; den3=[1 0 2]; sys3 = tf(num3,den3);
num4=[1]; den4=[1 0 0]; sys4 = tf(num4,den4);
num5=[4 2]; den5=[1 2 1]; sys5 = tf(num5,den5);
num6=[50]; den6=[1]; sys6 = tf(num6,den6);
num7=[1 0 2]; den7=[1 0 0 14]; sys7 = tf(num7,den7);
sysa = feedback(sys4,sys6,+1);
sysb = series(sys2,sys3);
sysc = feedback(sysb,sys5);
sysd = series(sysc,sysa);
syse = feedback(sysd,sys7);
sys = series(sys1,syse)
poles
%
pzmap(sys)
%
p=pole(sys)
z=zero(sys)

z=
0
1.2051 + 2.0872i
1.2051 - 2.0872i
-2.4101
-1.0000 + 0.0000i
-1.0000 - 0.0000i

Polezero map
2.5

2

1.5

1

Imag Axis

0.5

0

-0.5

-1

-1.5

-2

-2.5
-8

-6

-4

-2

0

Real Axis

FIGURE CP2.6
Pole-zero map.

2

4

6

8

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

81

Computer Problems

where
∆(s) = s10 + 3s9 − 45s8 − 125s7 − 200s6 − 1177s5
− 2344s4 − 3485s3 − 7668s2 − 5598s − 1400 .
CP2.7

The m-file script and plot of the pendulum angle is shown in Figure CP2.7.
With the initial conditions, the Laplace transform of the linear system is
θ(s) =

s2

θ0 s
.
+ g/L

To use the step function with the m-file, we can multiply the transfer
function as follows:
θ(s) =

s2
θ0
,
2
s + g/L s

which is equivalent to the original transfer function except that we can
use the step function input with magnitude θ0 . The nonlinear response
is shown as the solid line and the linear response is shown as the dashed
line. The difference between the two responses is not great since the initial
condition of θ0 = 30◦ is not that large.

30

L=0.5; m=1; g=9.8;
theta0=30;
% Linear simulation
sys=tf([1 0 0],[1 0 g/L]);
[y,t]=step(theta0*sys,[0:0.01:10]);
% Nonlinear simulation
[t,ynl]=ode45(@pend,t,[theta0*pi/180 0]);
plot(t,ynl(:,1)*180/pi,t,y,'--');
xlabel('Time (s)')
ylabel('\theta (deg)')

20

θ (deg)

10

0

-10

function [yd]=pend(t,y)
L=0.5; g=9.8;
yd(1)=y(2);
yd(2)=-(g/L)*sin(y(1));
yd=yd';

-20

-30

0

2

4

6
Time (s)

FIGURE CP2.7
Plot of θ versus xt when θ0 = 30◦ .

8

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

82

CHAPTER 2

CP2.8

Mathematical Models of Systems

The system step responses for z = 5, 10, and 15 are shown in Figure CP2.8.

z=5 (solid), z=10 (dashed), z=15 dotted)
1.5

x(t)

1

0.5

0

0

0.5

1

1.5

2

2.5
3
Time (sec)

3.5

4

4.5

5

FIGURE CP2.8
The system response.

CP2.9

(a,b) Computing the closed-loop transfer function yields

T (s) =

G(s)
s2 + 2s + 1
= 2
.
1 + G(s)H(s)
s + 4s + 3

The poles are s = −3, −1 and the zeros are s = −1, −1.
(c) Yes, there is one pole-zero cancellation. The transfer function (after
pole-zero cancellation) is

T (s) =

s+1
.
s+3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

83

Computer Problems

Pole?Zero Map
1

0.8

0.6

0.4

Imaginary Axi s

0.2

0

?-0.2

?-0.4

?-0.6

?-0.8

?-1
?-3

?-2.5

?-2

?-1.5

?-1

?-0.5

0

Real Axi s

ng=[1 1]; dg=[1 2]; sysg = tf(ng,dg);
nh=[1]; dh=[1 1]; sysh = tf(nh,dh);
sys=feedback(sysg,sysh)
%
pzmap(sys)
%
pole(sys)
zero(sys)

>>
Transfer function:
s^2 + 2 s + 1
------------s^2 + 4 s + 3

poles

p=
-3
-1

zeros

z=
-1
-1

FIGURE CP2.9
Pole-zero map.

CP2.10

Figure CP2.10 shows the steady-state response to a unit step input and a
unit step disturbance. We see that K = 1 leads to the same steady-state
response.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 2

Mathematical Models of Systems

0.35

K=[0.1:0.1:10];
sysg=tf([1],[1 20 20]);
for i=1:length(K)
nc=K(i); dc=[1];sysc=tf(nc,dc);
syscl=feedback(sysc*sysg,1);
systd=feedback(sysg,sysc);
y1=step(syscl);
Tf1(i)=y1(end);
y2=step(systd);
Tf2(i)=y2(end);
end
plot(K,Tf1,K,Tf2,'--')
xlabel('K')

0.3

0.25

84

0.2

0.15

0.1

0.05

K=1
0

FIGURE CP2.10

0

1

2

3

4

5
K

6

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

3

State Variable Models

Exercises
E3.1

One possible set of state variables is
(a) the current iL2 through L2 ,
(b) the voltage vC2 across C2 , and
(c) the current iL1 through L1 .
We can also choose vC1 , the voltage across C1 as the third state variable,
in place of the current through L1 .

E3.2

We know that the velocity is the derivative of the position, therefore we
have
dy
=v ,
dt
and from the problem statement
dv
= −k1 v(t) − k2 y(t) + k3 i(t) .
dt
This can be written in matrix form as














0
1
y
0
d  y  

+
i .
=
dt
v
−k2 −k1
v
k3

Define u = i, and let k1 = k2 = 1. Then,

ẋ = Ax + Bu
where


A=

0

1

−1 −1



 ,



B=

0
k3



 , and



x=

y
v



 .

85

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

86

CHAPTER 3

E3.3

State Variable Models

The charactersitic roots, denoted by λ, are the solutions of det(λI − A) =
0. For this problem we have


λ

det(λI − A) = det 

−1

1 λ+2



 = λ(λ + 2) + 1 = λ2 + 2λ + 1 = 0 .

Therefore, the characteristic roots are
λ1 = −1 and λ2 = −1 .
E3.4

The system in phase variable form is
ẋ = Ax + Bu
y = Cx
where


E3.5



A=



0
0



1

0 

1 
 ,

0

−8 −6 −4







 0 



B=
 0  ,



C=

20

h

1 0 0

i

.

From the block diagram we determine that the state equations are
ẋ2 = −(f k + d)x2 + ax1 + f u
ẋ1 = −kx2 + u
and the output equation is
y = bx2 .
Therefore,
ẋ = Ax + Bu
y = Cx + Du ,
where


A=
E3.6

0

−k

a −(f k + d)



 ,



B=

1
f



 ,

C=

h

0 b

i

(a) The state transition matrix is
Φ(t) = eAt = I + At +

1 2 2
A t + ···
2!

and D = [0] .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

87

Exercises

But A2 = 0, thus A3 = A4 = · · · = 0. So,


Φ(t) = eAt = I + At = 

1 0
0 1





+

0 1
0 0





t = 

1 t
0 1



 .

(b) The state at any time t ≥ 0 is given by
x(t) = Φ(t)x(0)
and since x1 (0) = x2 (0) = 1, we determine that
x1 (t) = x1 (0) + tx2 (0) = 1 + t
x2 (t) = x2 (0) = 1 .
E3.7

The state equations are
ẋ1 = x2
x˙2 = −100x1 − 20x2 + u
or, in matrix form


ẋ = 

0

1

−100 −20





x + 

0
1



u .

So, the characteristic equation is determined to be


det(λI − A) = det 

λ

−1

100 λ + 20



 = λ2 + 20λ + 100 = (λ + 10)2 = 0 .

Thus, the roots of the characteristic equation are
λ1 = λ2 = −10 .
E3.8

The characteristic equation is


 λ −1

det(λI − A) = det 
 0 λ


0

6

0
−1
λ+3





 = λ(λ2 + 3λ + 6) = 0 .



Thus, the roots of the characteristic equation are
λ1 = 0 ,

λ2 = −1.5 + j1.9365 and λ3 = −1.5 − j1.9365 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

88

CHAPTER 3

E3.9

State Variable Models

Analyzing the block diagram yields
1
ẋ1 = −x1 + x2 + r
2
3
ẋ2 = x1 − x2 − r
2
3
y = x1 − x2 − r.
2
In state-variable form we have


ẋ = 

−1
1



1
2
− 32



x+



1

r ,

−1

The characteristic equation is


h
i
3
x + −1 r .
y= 1 −
2


5
1
s2 + s + 1 = (s + 2)(s + ) = 0 .
2
2
E3.10

(a) The characteristic equation is


det[λI − A] = det 

λ



−6

 = λ(λ+ 5)+ 6 = (λ+ 2)(λ+ 3) = 0 .

1 (λ + 5)

So, the roots are λ1 = −2 and λ2 = −3.
(b) We note that
−1

Φ(s) = [sI − A]



=

s

−6

1 s+5

−1






s+5 6
1

 .
=
(s + 2)(s + 3)
−1 s

Taking the inverse Laplace transform yields the transition matrix


E3.11

Φ(t) = 

3e−2t − 2e−3t
−e−2t

+

6e−2t − 6e−3t

e−3t

−2e−2t

+

3e−3t

A state variable representation is



 .

ẋ = Ax + Br
y = Cx
where


A=

0

1

−12 −8



 ,



B=

0
1



 ,

C=

h

12 4

i

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

89

Exercises

The equation of motion is
L

di
+ Ri + vc = vin
dt

where
vc =

1
C

Z

i dt .

Unit step response
1.8
1.6
1.4
1.2
State response

E3.12

x1: capacitor voltage
1
0.8
0.6
0.4
0.2
x2: inductor current

0
−0.2

0

0.05

0.1

0.15
0.2
Time(sec)

0.25

0.3

FIGURE E3.12
State variable time history for a unit step input.

Selecting the state variables x1 = vc and x2 = i, we have
1
x2
C
R
1
1
ẋ2 = − x2 − x1 + vin .
L
L
L
ẋ1 =

This can be written in matrix form as


ẋ = 

0

1/C

−1/L −R/L





x + 

0
1/L



 vin .

0.35

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

90

CHAPTER 3

State Variable Models

When C = 0.001F , R = 4Ω, and L = 0.1H, we have


ẋ = 

0

1000

−10

−40





x+

0
10



 vin .

The step response is shown in Figure E3.12.
E3.13

(a) Select the state variables as x1 = y and x2 = ω.
(b) The corresponding state equation is
ẋ1 = −x1 − ax2 + 2u
ẋ2 = bx1 − 4u
or, in matrix form


ẋ = 

−1 −a
b

0





x + 

2
−4





u

and x = 

x1
x2



 .

(c) The characteristic equation is


det[λI − A] = det 

λ+1 a
−b

λ



 = λ2 + λ + ab = 0 .

So, the roots are
1 1√
1 − 4ab .
λ=− ±
2 2
E3.14

Assume that the mass decay is proportional to the mass present, so that
Ṁ = −qM + Ku
where q is the constant of proportionality. Select the state variable, x, to
be the mass, M . Then, the state equation is
ẋ = −qx + Ku .

E3.15

The equations of motion are
mẍ + kx + k1 (x − q) + bẋ = 0
mq̈ + kq + bq̇ + k1 (q − x) = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

91

Exercises

In state variable form we have


0


 (k+k1 )
 −

m
ẋ = 

0


k1
m

1

0

0

b
−m

k1
m

0

0

0

1

0

1)
− (k+k
m

b
−m

where x1 = x, x2 = ẋ, x3 = q and x4 = q̇.
E3.16







x




The governing equations of motion are
m1 ẍ + k1 (x − q) + b1 (ẋ − q̇) = u(t)
m2 q̈ + k2 q + b2 q̇ + b1 (q̇ − ẋ) + k1 (q − x) = 0 .
Let x1 = x, x2 = ẋ, x3 = q and x4 = q̇. Then,


0



 − k1

ẋ =  m1
 0


k1
m2



1

0

0

b1
−m
1

k1
m1

b1
m1

0

0

1

b1
m2

2)
− (k1m+k
2

− (b1m+b2 2 )

E3.17

h

0 0 1 0

i

x.

At node 1 we have
C1 v̇1 =

va − v1 v2 − v1
+
R1
R2

C2 v˙2 =

vb − v2 v1 − v2
+
.
R3
R2

and at node 2 we have

Let
x 1 = v1
and
x 2 = v2 .



 0 


 1 




 m 

 x +  1  u(t) .
 0 








Since the output is y(t) = q(t), then
y=



0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

92

CHAPTER 3

State Variable Models

Then, in matrix form we have


 −

E3.18

ẋ = 



1
R1 C1

+

1
R2 C1

− R21C2



1
R2 C1

−



1
R3 C2

+

1
R2 C2

The governing equations of motion are






  x+

1
R1 C1

0

0

1
R3 C2




va
vb



 .

di1
+ v = va
dt
di2
L2
+ v = vb
dt
dv
.
iL = i1 + i2 = C
dt

Ri1 + L1

Let x1 = i1 , x2 = i2 , x3 = v, u1 = va and u2 = vb . Then,




ẋ = 



y=

E3.19

h

− LR1

0

0

0

1
C

1
C

0 0 1

First, compute the matrix



− L11
− L12

1
L1





x+ 0





0
i



0 

1
L2

0

0

x + [0] u .



sI − A = 

s

−1

3 s+4




u





 .

Then, Φ(s) is
Φ(s) = (sI − A)−1
where ∆(s) = s2 + 4s + 3, and
G(s) =
E3.20

h

10 0

i




s+4
∆(s)
3
− ∆(s)





1  s+4 1 
=
∆(s)
−3 s
1
∆(s)
s
∆(s)




0
1



=

s2

10
.
+ 4s + 3

The linearized equation can be derived from the observation that sin θ ≈ θ
when θ ≈ 0. In this case, the linearized equations are
θ̈ +

g
k
θ + θ̇ = 0 .
L
m

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

93

Exercises

Let x1 = θ and x2 = θ̇. Then in state variable form we have
ẋ = Ax
y = Cx
where


A=

E3.21

0

1

−g/L −k/m

The transfer function is



 ,

C=

h

1 0

i

,

G(s) = C [sI − A]−1 B + D =

and

s2



x(0) = 

θ(0)
θ̇(0)



.

−1
.
+ 2s + 1

The unit step response is
y(t) = −1 + e−t + te−t .
E3.22

The transfer function is
G(s) =

s2

s−6
.
− 7s + 6

The poles are at s1 = 1 and s2 = 6. The zero is at s = 6. So, we see
that there is a pole-zero cancellation. We can write the system in state
variable form as
√
ẋ = x − 2u
√
2
y=−
x
2
and the transfer function is
G(s) =
E3.23

1
.
s−1

The system in state variable form can be represented by
ẋ = Ax + Bu
y = Cx + Du
where




A=



0

1

0

0



0 

1 
 ,

−1 −3 −3







 0 



B=
 0  ,



1

C=

h

0 1 −1

i

,

D=

h

1

i

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

94

CHAPTER 3

State Variable Models

+
-

U(s)

+
- --

1
s

x3

3
3

FIGURE E3.23
Block diagram.

1
s

x2

1
s

x1

+

X(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

95

Problems

Problems
P3.1

The loop equation, derived from Kirchoff’s voltage law, is
di
1
R
1
=
v−
i − vc
dt
L
L
L
where
vc =

1
C

Z

i dt .

(a) Select the state variables as x1 = i and x2 = vc .
(b) The corresponding state equations are
R
1
1
v−
x1 − x2
L
L
L
1
x1 .
ẋ2 =
C

ẋ1 =

(c) Let the input u = v. Then, in matrix form, we have


ẋ = 

−R/L −1/L
1/C

0





x+

1/L
0



u .

-R/L
1/C

v
1/s

1/L

x1

1/s
x2

-1/L

FIGURE P3.1
Signal flow graph.

P3.2

Let
−2
−2R1 R2
,
a22 =
,
(R1 + R2 )C
(R1 + R2 )L
1
R2
= b12 =
, b21 = −b22 =
.
(R1 + R2 )C
(R1 + R2 )L

a11 =
b11

The corresponding block diagram is shown in Figure P3.2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

96

CHAPTER 3

State Variable Models

2/(R1+R2)C

v1

1/s

1/(R1+R2)C

x1

R2

(a)
1/(R1+R2)C

v2

1/s

x2

-

2R1R2/(R1+R2)C

a 11
v1

b11
x1

1/s

b12
(b)
b21
v2

1/s
x2

b22
a 22

FIGURE P3.2
(a) Block diagram. (b) Signal flow graph.

P3.3

Using Kirchoff’s voltage law around the outer loop, we have
L

diL
− vc + v2 − v1 = 0 .
dt

Then, using Kirchoff’s current law at the node, we determine that
C

dvc
= −iL + iR ,
dt

where iR is the current through the resistor R. Considering the right loop
we have
iR R − v2 + vc = 0

or

iR = −

vc v2
+
.
R
R

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

97

Problems

Thus,
dvc
vc
iL
v2
=−
−
+
dt
RC
C
RC

vc
v1 v2
diL
=
+
−
.
dt
dt
L
L

and

In matrix form, the state equations are



ẋ1
ẋ2





=

0

1/L

−1/C −1/RC




x1
x2





+

1/L −1/L
0

1/RC




v1
v2



 ,

where x1 = iL and x2 = vc . The signal flow graph is shown in Figure P3.3.

v1
1/L
-1/L
-1/C

1/s

v2
1/L

x2

1/s

x1
-1/RC

1/RC

FIGURE P3.3
Signal flow graph.

P3.4

(a) The block diagram model for phase variable form is shown in Figure P3.4a. The phase variable form is given by




ẋ = 



y=

h

0

1

0

0

−10 −6 −4
10 2 1

i











0 
 0 





1 
x +  0 r

x.

1



(b) The block diagram in input feedforward form is shown in Figure P3.4b.
The input feedforward form is given by




ẋ = 



y=

h











−4 1 0 
 1 





−6 0 1 
 x +  2  r(t)

−10 0 0
1 0 0

i

x.

10



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

98

CHAPTER 3

State Variable Models

1
2
x2

x3
R(s)

1
s

+
-

--

1
s

x1

1
s

+

10

+ +

Y(s)

4
6
10
(a)
1
2

R(s)

10

+

.
.
x3 1 + + x2
s -

1 +
s -

+

.
x1

1
s

Y(s)

4
6
10
(b)

FIGURE P3.4
(a)Block diagram model for phase variable form. (b) Block diagram model for input feedforward form.

P3.5

(a) The closed-loop transfer function is
T (s) =

s3

s+1
.
+ 4s2 − 11s + 1

(b) A matrix differential equation is
ẋ = Ax + Bu
y = Cx
where




A=



0

1

0

0



0 

1 
 ,

−1 11 −4











 0 



B=
 0  ,

1

The block diagram is shown in Figure P3.5.

C=

h

1 1 0

i

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

99

Problems

1
R(s)

-

x3

1
s

+
--

x2

1
s

x1

1
s

+

1

+

Y(s)

4
-11
1

FIGURE P3.5
Block diagram model.

P3.6

The node equations are
dv1
vi − v1
+ iL −
=0
dt
4000
dv2
v2
0.0005
− iL +
− i3 = 0
dt
1000
diL
0.002
+ v2 − v1 = 0 .
dt
0.00025

Define the state variables
x 1 = v1

x 2 = v2

x3 = iL .

Then,
ẋ = Ax + Bu
where


 −1

A=
0


P3.7





0 −4000 

−2
2000 
 ,

500 −500

Given K = 1, we have
KG(s) ·

0







0 
 1



B=
 0 2000 


0

0



(s + 1)2
1
=
.
s
s(s2 + 1)

We then compute the closed-loop transfer function as
T (s) =

s2 + 2s + 1
s−1 + 2s−2 + s−3
=
.
3s3 + 5s2 + 5s + 1
3 + 5s−1 + 5s−2 + s−3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

100

CHAPTER 3

State Variable Models

A state variable model is




ẋ = 



y=

P3.8

h

0

1

0

0

−1/3 −5/3 −5/3
1 2 1

The state-space equations are

i











0 
 0 





1 
x +  0 r

x.

1/3



ẋ1 = x2
ku
ẋ2 =
−g
x3
ẋ3 = u .
This is a set of nonlinear equations.
P3.9

(a) The closed-loop transfer function is
T (s) =

10s−3
10
=
,
Js3 + (b + 10J)s2 + 10bs + 10K1
1 + 10.1s−1 + s−2 + 5s−3

where K1 = 0.5, J = 1, and b = 0.1.
(b) A state-space model is




ẋ = 



ω=

h

0

1

0

0

1 0 0









0 
 0 





1 
x +  0 r

−5 −1 −10.1
i



x.

10



(c) The characteristic equation is


 s −1

det[sI − A] = det 
 0 s


5

1

0
−1
s + 10.1





 = s3 + 10.1s2 + s + 5 = 0 .



The roots of the characteristic equation are
s1 = −10.05

and s2,3 = −0.0250 ± 0.7049j .

All roots lie in the left hand-plane, therefore, the system is stable.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

101

Problems

P3.10

(a) From the signal flow diagram, we determine that a state-space model
is given by


−K1





ẋ = 
y=



K2

x + 

−K1 −K2
y1
y2



=





K1 −K2
K1



K2



1 0

r1
r2




x .

0 1

(b) The characteristic equation is
det[sI − A] = s2 + (K2 + K1 )s + 2K1 K2 = 0 .
(c) When K1 = K2 = 1, then


A=

−1

1

−1 −1



 .

The state transition matrix associated with A is


o

n

Φ = L−1 [sI − A]−1 = e−t 

P3.11

cos t

sin t

− sin t cos t

The state transition matrix is


Φ(t) = 

(2t − 1)e−t
2te−t

−2te−t

(−2t + 1)e−t



 .

So, when x1 (0) = x2 (0) = 10, we have
x(t) = Φ(t)x(0)
or
x1 (t) = 10e−t
x2 (t) = 10e−t
P3.12

(a) A state variable representation is given by




ẋ = 



0

1

0

0











0 
 0 





1 
x +  0 r

−48 −44 −12

1





 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

102

CHAPTER 3

State Variable Models

y = [40 8 0]x .
(b) The state transition matrix is
Φ(t) =



.
.
Φ1 (t)..Φ2 (t)..Φ3 (t)



,

where




e−6t − 3e−4t + 3e−2t



−6t + 12e−4t − 6e−2t
Φ1 (t) = 
 −6e


36e−6t − 48e−4t + 12e−2t


P3.13



Φ3 (t) = 













Φ2 (t) = 



1 −6t
− 41 e−4t + 18 e−2t
8e
− 43 e−6t + e−4t − 14 e−2t
9 −6t
− 4e−4t + 12 e−2t
2e

(a) The RLC circuit state variable representation is


ẋ = 

−10 −4
6

0





x+

4
0

3 −6t
− 2e−4t + 54 e−2t
4e
− 29 e−6t + 8e−4t − 25 e−2t

27e−6t − 32e−4t + 5e−2t





 .





u .

The characteristic equation is
s2 + 10s + 24 = 0 .
All roots of the characteristic equation (that is, s1 = −4 and s2 = −6)
are in the left half-plane; therefore the system is stable.
(b) The state transition matrix is


Φ(t) = 

3e−6t − 2e−4t

−3e−6t

+

3e−4t

2e−6t + 2e−4t
−2e−6t

+

3e−4t



 .

(c) Given
x1 (0) = 0.1 ,

x2 (0) = 0 and e(t) = 0 ,

we have
i(t) = x1 (t) = 0.3e−6t − 0.2e−4t
vc (t) = x2 (t) = −0.3e−6t + 0.3e−4t .








from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

103

Problems

(d) When x(0) = 0 and u(t) = E, we have
x(t) =

Z

t

Φ(t − τ )Bu(τ )dτ ,

0

where


Bu(t) = 

Integrating yields

4E
0



 .

x1 (t) = (−2e−6t + 2e−4t )E
x2 (t) = (1 + 2e−6t − 3e−4t )E .
P3.14

A state space representation is
ẋ = Ax + Br ,

y = Cx

where


P3.15





A=




0

1

0

0

0

1

0

0

0





0 

0 


 ,

1 


−50 −34 −10 −12



 ,
 0 





B=

A state variable representation is




ẋ = 



0

1

0

0



 0 


 0 



C = [50 1 0 0] .

1











0 
 0 





1 
x+ 0 r

−16 −31 −10

y = [56 14 0]x .
The block diagram is shown in Figure P3.15.

1



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

104

CHAPTER 3

State Variable Models

14
R(s)

1
s

+
-

--

x3

x2

1
s

1
s

x1

56

+

+

Y(s)

10
31
16

FIGURE P3.15
Block diagram model.

(a) The characteristic equation is

x1 - solid; x2 - dotted; x3 - dashed
0.5
0
-0.5

Step response)

P3.16

-1
-1.5
-2
-2.5
-3
-3.5

0

20

40

60

80

100

Time (s)

FIGURE P3.16
Step response of magnitude 0.285◦ .



s


det(sI − A) = det 
 0.0071

0

−1
s + 0.111
−0.07

0





−0.12 



s + 0.3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

105

Problems

= s3 + 0.411s2 + 0.032s + 0.00213 = 0 .
The roots are
s1 = −0.3343

and

s2,3 = −0.0383 ± 0.0700j .

All the poles lie in the left half-plane, therefore, the system is stable.
(b) The solution of the system to a step of magnitude 0.285◦ is given by
x1 (t) = −2.66 − 0.11e−0.33t + e−0.038t (2.77 cos 0.07t + 0.99 sin 0.07t)
x2 (t) = 0.037e−0.33t − e−0.038t (0.037 cos 0.07t + 0.23 sin 0.07t)
x3 (t) = 0.069 − 0.075e−0.33t + e−0.038t (0.006 cos 0.07t − 0.06 sin 0.07t)
P3.17

The transfer function is
G(s) = C(sI − A)−1 B =

P3.18

−4s + 12
.
s3 − 14s2 + 37s + 20

Define the state variables as
x1 = φ1 − φ2
ω1
x2 =
ωo
ω2
x3 =
.
ωo
Then, the state equations of the robot are
ẋ1 = ωo x2 − ωo x3
−J2 ωo
x1 −
ẋ2 =
J1 + J2
J1 ωo
ẋ3 =
x2 +
J1 + J2

b
x2 +
J1
b
x2 −
J2

b
Km
x3 +
i
J1
J1 ωo
b
x3
J2

or, in matrix form


0

1



ẋ = ωo 
 a−1


a

where
a=

J1
,
(J1 + J2 )

b1 =

−b1

b2 −b2

b
,
J1 ωo











−1 
 0 





b1 
x +  d i

b2 =

0

b
and
J2 ωo



d=

Km
.
J1 ωo

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

106

CHAPTER 3

P3.19

State Variable Models

The state equation is given by


0

ẋ = 

1

−2 −3



x

where x1 (0) = 1 and x2 (0) = −1. The state transition matrix is


Φ(t) = 

−e−2t + 2e−t −e−2t + e−t
2e−2t

−

2e−t

2e−2t

− e−t



 .

The system response is








x1 (t) = −e−2t + 2e−t x1 (0) + −e−2t + e−t x2 (0)








x2 (t) = 2e−2t − 2e−t x1 (0) + 2e−2t − e−t x2 (0) .
The state response is shown in Figure P3.19.

1
0.8

x1

0.6

System response

0.4
0.2
0
-0.2
-0.4

x2

-0.6
-0.8
-1

0

1

2

3
Time (s)

4

5

6

FIGURE P3.19
Response with x1 (0) = 1 and x2 (0) = −1.

P3.20

The state equation is given by


ẋ = 

− 0.693
6.7
−1

0
− 0.693
9.2



x



where x(0) = 

0.3 × 1016
7×

1016



 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

107

Problems

The state transition matrix is


Φ(t) = 

e−0.103433t
35.5786(e−0.103433t

−

0

e−0.0753261t )

The system response is

e−0.075326t



 .

x1 (t) = e−0.103433t x1 (0)
h

i

x2 (t) = 35.5786 e−0.103433t − e−0.0753261t x1 (0) + e−0.075326t x2 (0) .
The state response is shown in Figure P3.20.
7

Nucleide densities in atoms per unit volume

6

X=Xenon 135
I=Iodine 135

5

4

3

2

1

0

-1

0

10

20
30
Time (hours)

40

FIGURE P3.20
Nuclear reactor state response to initial conditions.

50

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

108

CHAPTER 3

P3.21

State Variable Models

Referring to Figure P3.21 we have
1
1
1
Y (s) = W (s) =
h1 U (s) + Q(s)
s
s
s
1
h1
U (s) + 2 [h0 U (s) − a0 Y (s) − a1 sY (s) + a1 h1 U (s)] .
=
s
s




Gathering like terms and re-arranging yields


1+

a1 a0
+ 2 Y (s) =
s
s




h1 h0 a1 h1
U (s)
+ 2 + 2
s
s
s


or
Y (s) =



h1 s + h0 + a1 h1
U (s) .
s 2 + a1 s + a0


Computing the transfer function from the state variable representation
yields
G(s) = C (sI − A)−1 B
=

h

1 0

i




s+a1
s2 +a1 s+a0
−a0
s2 +a1 s+a0

1
s2 +a1 s+a0
s
s2 +a1 s+a0




h1
h0



=

h1 s + h0 + a1 h1
.
s 2 + a1 s + a0

h1

U(s)

Q(s)

h0 +
--

1
s

+

+

1
s

W(s)

a1
a0
FIGURE P3.21
Block diagram with labeled signals.

P3.22

The governing equations are
L
C1

di
= v2
dt

dv1
1
1
+
(v1 − v) +
(v1 − v2 ) = 0
dt
R1
R2

Y(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

109

Problems

C2

dv2
1
v2
+
(v2 − v1 ) + i +
=0.
dt
R2
R3

Let u = v, x1 = i, x2 = v1 and x3 = v2 . Then,




ẋ = 



0
0
− C12

− a1



1
R1

+

1
R2

1
R2 C2





1
L
1
C1 R2

0

−



1
R2 C2

+

1
R3 C2







x+


 


0
1
R1 C1

0





u



y = [0 0 1]x .
P3.23

A state variable representation is given by




ẋ = 



0

1

0

0











0 
 0 





1 
x +  0 r

−30 −31 −10

y = [1 0 0]x .

1



Other representations include the input feedforward representation












 −10 1 0 
 0 







ẋ = 
 −31 0 1  x +  0  r


−30 0 0

y = [1 0 0]x ,

1



the physical variable representation


1
 −3


ẋ =  0 −2

0

















0 
 0 





1 x+
 0 r

0 −5

y = [1 0 0]x ,



1



and the decoupled representation


0
 −3

ẋ = 
 0 −2

0



0 
 1 





0 
x +  1 r

0 −5

1



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

110

CHAPTER 3

State Variable Models

y=
P3.24



1 1
1
x.
−
6 3
2


The matrix representation of the state equations is


ẋ = 

3 0
0 2





x + 



1 1



0 1

u1
u2





+

0
1



d .

When u1 = 0 and u2 = d = 1, we have
ẋ1 = 3x1 + u2
ẋ2 = 2x2 + 2u2
So we see that we have two independent equations for x1 and x2 . With
U2 (s) = 1/s and zero initial conditions, the solution for x1 is found to be
x1 (t) = L

−1

1
{X1 (s)} = L
s(s − 3)



1
1 1
1
−1
=L
− +
= − 1 − e3t
3s 3 s − 3
3
−1





and the solution for x2 is
x2 (t) = L−1 {X2 (s)} = L−1
P3.25



2
s(s − 2)



1
1
= L−1 − +
s s−2


= −1+e2t .

Since Φ(s) = (sI − A)−1 , we have


Φ(s) = 

s+1
−2

0
s+3

−1




=

s+3
2

0
s+1




1
∆(s)

where ∆(s) = (s + 1)(s + 3). The state transition matrix is


Φ(t) = L−1 {Φ(s)} = 
P3.26



e−t

0

e−t − e−3t e−3t

The state variable differential equation is


ẋ = 

0

1

−25 −6

y = [1 0]x .





x + 

0
25



r



 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

111

Problems

and


s+6 1

Φ(s) = (sI − A)−1 = 

−25

s



1
∆(s)



where ∆(s) = s2 + 6s + 25.
P3.27

Equating the change in angular momentum to the sum of the external
torques yields
J θ̈ − Hω cos θ = −bθ̇ − kθ
where b is the damping coefficient, k is the spring constant, and J is the
wheel moment of inertia. Defining the state variables x1 = θ and x2 = ẋ
and the input u = ω, we can write the equations of motion as
ẋ1 = x2
k
b
H
ẋ2 = − x1 − x2 + u cos x1
J
J
J
With a small angle assumption (that is, cos x1 ≈ 1) we have


ẋ = 

0

1

−k/J

−b/J

y=θ=

P3.28

h

1 0

i

x.





x + 



0

u

H/J

The governing equations of motion are
m1 y¨1 + k(y1 − y2 ) + by˙1 = u
m2 y¨2 + k(y2 − y1 ) + by˙2 = 0
y = y2 .
Let x1 = y1 , x2 = ẏ1 , x3 = y2 and x4 = ẏ2 . Then


0



 − k

ẋ =  m1
 0


k
m2

y=

h

1

0

− mb1

k
m1

0

0

0

1

− mk2

− mb2

0

0 0 1 0

i

x.

0







 0 



 1 




 m 
x +  1 u

 0 







0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

112

CHAPTER 3

P3.29

State Variable Models

The equations of motion are
I q¨1 + M gL sin q1 + k(q1 − q2 ) = 0
J q¨2 − k(q1 − q2 ) = u .
Let x1 = q1 , x2 = q̇1 , x3 = q2 , and x4 = q̇2 and linearize the equations
using small angle assumptions (i.e. sin q1 ≈ q1 ). Then, we have
x˙1 = x2
M gL
k
x˙2 = −
x1 − (x1 − x3 )
I
I
x˙3 = x4
k
1
x˙4 = (x1 − x3 ) + u .
J
J

P3.30

Using Kirchoff’s current law, we find that
C

dv c
= i2 + i3
dt

where i3 = current in R3 . Let i1 = current in R1 . Using Kirchoff’s voltage
law, we have
L

diL
= v1 − R1 i1
dt

and
R1 i1 + R2 i2 + vc = v1 .
But
i2 = i1 − iL ,
so
(R1 + R2 )i1 = v1 − vc + R2 iL .
Using Kirchoff’s voltage law once again, we calculate i3 as
i3 =

v2 − vc
.
R3

Utilizing the above equations, we can solve for diL /dt and dvc /dt, as
follows:
diL
R2
R1
R1 R2
=
v1 +
vc −
iL
dt
L(R1 + R2 )
L(R1 + R2 )
L(R1 + R2 )
vc
v1
vc
vc
R1 iL
v2
=
−
−
−
+
dt
C(R1 + R2 ) C(R1 + R2 ) CR3 C(R1 + R2 ) CR3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

113

Problems

Define the state variables x1 = vc and x2 = iL . Then, in matrix form we
have


ẋ = 

1 +R2 +R3 )
− (R
CR3 (R1 +R2 )

R1
L(R1 +R2 )

y = i2 =
P3.31

h

1
− (R1 +R
2)



1
− C(RR1 +R
2)

x+

R1 R2
− L(R
1 +R2 )
1
− (R1R+R
2)



i

x+

h

1
CR3

1
C(R1 +R2 )
R2
L(R1 +R2 )

1
(R1 +R2 )

0
i

0







v1
v2

v1
v2








A state variable representation is


ẋ = 

0



1



x + 

−3 −4

0
30



u .

The state transition matrix can be computed as follows:
n

Φ = L−1 [sI − A]−1

o




s 



 1
s+4

= L−1
 ∆(s)
−3


=

3 −t
− 21 e−3t
2e
− 32 e−t + 23 e−3t

1

1 −t
− 21 e−3t
2e
− 12 e−t + 32 e−3t




where
∆(s) = s2 + 4s + 3 = (s + 1)(s + 3) .
P3.32

A state variable representation is
ṁ1 = −k1 m1 + r
ṁ2 = k1 m1 − k2 m2
where k1 and k2 are constants of proportionality. In matrix form, we have


ẋ = Ax + Br = 

−k1

0

k1 −k2





x+

1
0



r

where x1 = m1 and x2 = m2 . Let k1 = k2 = 1 and assume that r(t) = 0
and x1 = 1 and x2 = 0. Then


x(t) = Φ(t)x(0) = 

e−t

0

te−t

e−t





 x(0) = 

e−t
te−t



 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

114

CHAPTER 3

State Variable Models

The simulation is shown in Figure P3.32.

0.4

1
0.9

0.35

0.8
0.3
0.25

x1

0.6

x2

state history, x(t)

0.7

0.5
0.4

0.2
0.15

0.3
0.1
0.2
0.05

x2

0.1
0
0

5

t=0

0
0

10

0.5

time (sec)

1

x1

FIGURE P3.32
Actual versus approximate state response.

P3.33

The system (including the feedback) is described by


ẋ = Ax = 

0

1

−1/2 −1



x .

The charactersitic equation is


det[λI − A] = det 

λ

−1

1/2 λ + 1



 = λ2 + λ +

The roots of the characteristic equation are
1
1
λ1,2 = − ± j .
2
2

1
=0.
2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

115

Problems

The system response is


x(t) = eAt x(0) = 

e−t/2 cos 2t + e−t/2 sin 2t

2e−t/2 sin 2t

−e−t/2 sin 2t

e−t/2 cos 2t − e−t/2 sin 2t



= e−t/2 

2 sin 2t
cos 2t − sin 2t





 x(0)



where x1 (0) = 0 and x2 (0) = 1.
P3.34

(a) The state space representation is


1



−6 −11 −6

 0

ẋ = 
 0











0 
 0 





1 
x +  0 r

0

y = [6 0 0] x .

1



(b) The element φ11 (t) of the state transition matrix is
φ11 (t) = e−3t − 3e−2t + 3e−t .
P3.35

The state equations are
1
8
[80θ − 50h] = −x1 + x2
50
5
θ̇ = ẋ2 = ω = x3
Km
Km Kb
Km Ka
353
25000
ω̇ = ẋ3 =
ia = −
ω+
vi = −
x3 +
vi .
J
JRa
JRa
30
3
ḣ = ẋ1 =

In state variable form, we have


8
 −1 5

ẋ = 
 0 0

P3.36



0

0
1

0 − 353
30









x+





0
0
25000
3





 vi .



Using Newton’s Law and summing the forces on the two masses yields
M1 ẍ(t) + b1 ẋ(t) + k1 x(t) = b1 ẏ(t)
M2 ÿ(t) + b1 ẏ(t) + k2 y(t) = b1 ẋ(t) + u(t)
Let
z1 = x, z2 = ẋ, z3 = y, and z4 = ẏ .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

116

CHAPTER 3

State Variable Models

Then we write the system in state variable form as


0



 − k1

ż =  M1
 0



1

0

0

b1
−M
1

0

b1
M1

0

0

1

b1
M2

k2
−M
2

b1
−M
2

0

y=
P3.37

h

1 0 0 0

i

z.







 0 


 0 







z + 
u

 0 






1
M2

From the block diagram in Figure P3.37, we obtain
ẋ1
ẋ2
ẋ3
y

= x2
= x3
= −10x1 − 4x2 − 3x3 + u
= x1 + 12x2 + 5x3

or




ẋ = 



0

1

0

0

−10 −4 −3

y = [1 12 5] x .











0 
 0 





1 
x +  0 u
1



The third-order differential equation model is
...
y +3ÿ + 4ẏ + 10y = 5ü + 12u̇ + u .

5
12

U(s)

+ - --

∫

x3

∫

x2

3
4
10

FIGURE P3.37
Block diagram with states labeled.

∫

x1

++
+

Y(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

117

AP3.1

With the state variables are defined as




 x 



z=
 ẋ  ,



i

the nonlinear equations of motion are










 ż1  

 
 ż  =  g −
 2  


ż3

z2
K (Io +z3 )2
m (Xo +z1 )2

1
L (v

− Rz3 )





 ,



where the control is the voltage v. We assume that z1 = x is measurable.
The linearized equations of motion are
ż = Az + Bv
y = Cz

where




A=



0

1

0

2K Io2
m Xo3

0

0

0

Io
− 2K
m Xo2
−R
L

The transfer function is





 ,







 0 
h
i



B=
.
 0  , and C = 1 0 0


1
L



G(s) = C(sI − A)−1 B .
With the constants
R = 23.2
L = 0.508
m = 1.75
K = 2.9 × 10−4
Io = 1.06
Xo = 4.36 × 10−3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

118

CHAPTER 3

State Variable Models

the transfer function is
G(s) =
AP3.2

s3

+

−36.38
.
+ 4493s + 205195

45.67s2

The differential equation describing the motion of y is
mÿ + bẏ + ky = bu̇ + ku .
Taking Laplace tranforms (with zero initial conditions) yields the transfer
function
Y (s)
(b/m)s + (k/m)
= 2
.
U (s)
s + (b/m)s + (k/m)
In state space form, we have


ẋ = 
y=

AP3.3

h

0

1

−k/m −b/m
k/m b/m

i





x + 

0
1

x.



u

The transfer function is
Y (s)
2s2 + 6s + 5
= 3
.
R(s)
s + 4s2 + 5s + 2
In (nearly) diagonal form, we have


1
 −1

A=
 0 −1

0



0 

0 
 ,

0 −2







 0 



B=
 1  , and



1

C=

h

1 1 1

i

.

The matrix A is not exactly diagonal due to the repeated roots in the
denominator of the transfer function.
AP3.4

The differential equations describing the motion of y and q are
mÿ + k2 ẏ + k1 (y − q) = f
−bq̇ + k1 (y − q) = f
where k1 = 2 and k2 = 1. Assume the mass m = 1. Then with the state

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

119

variables defined as z =


h

y ẏ q
0 1



ż = 
 −3 0


iT

y=

1 0 0

i





0





−1/b

0 




2 z + 


2/b 0 −2/b

h

, we have the state variable model

z



1



f



If we model a large bump at high speeds as an impulse and a small bump
at low speeds as a step, then b = 0.8 provides good performance. In both
cases, the ride settles out completely in about 10 seconds.
AP3.5

The differential equations describing the motion of x and θ are
(M + m)ẍ + M L cos θ θ̈ − M L sin θ θ̇ 2 = −kx
g sin θ + cos θẍ + Lθ̈ = 0
Assuming θ and θ̇ are small, it follows that
(M + m)ẍ + M Lθ̈ = −kx
ẍ + Lθ̈ = −gθ
Define the state variables as z =
able model is


0


 −k/m

ż = 

0



h

x ẋ θ θ̇

1

0

0

gM/m

0

0

iT

. Then, the state vari

0 

0 


k/(Lm) 0 −g(M + m)/(Lm) 0

AP3.6
AP3.7

z

1 



Computing the closed-loop system yields


A − BK = 

−1

1

−K1 −K2



 ,

The characteristic polynomial is



B=

0
1



 , and

C=

h

2 1

i

.

|sI − (A − BK)| = s2 + (K2 + 1)s + K1 + K2 = 0.
The roots are in the left-half plane whenever K2 +1 > 0 and K1 +K2 > 0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

120

CHAPTER 3

AP3.8

State Variable Models

(a) A state variable representation is given by
x˙1
x˙2
x˙3
y

= x2
= x3
= −Kx1 − 12x2 − 6x3 + Kr
= x1

or, in matrix form




ẋ = 



y=

h

0

1

0

0

1 0 0

x









0 
 0 





1 
x +  0 r

−K −12 −6
i



K



(b) The characteristic roots are found by solving
det [λI − A] = 0
or
λ3 + 6λ2 + 12λ + K = 0
When K = 8, we have characteristic roots at λ1 = −2, λ2 = −2, and
λ3 = −2, as desired.
(c) The unit step response is given by
y(t) = 1 − e−2t − 2te−2t − 2t2 e−2t .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

121

Design Problems

Design Problems
CDP3.1

The transfer model of the traction drive, capstan roller, and linear slide
was given in CDP2.1 as
rKm
X(s)
=
,
Va (s)
s [(Lm s + Rm )(JT s + bm ) + Kb Km ]
where
JT = Jm + r 2 (Ms + Mb ) .
Define x1 = x, x2 = ẋ, and x3 = ẍ. Then, a state variable representation
is
ẋ = Ax + Bva
y = Cx
where


 0

A=
 0


0

C=

DP3.1

h

1

0

0

1

b Km
− Rm bLmm+K
JT

m JT
− Lm bLmm+R
JT

1 0 0

i

.





 ,







B=



(a) The equation of motion of the spring-mass-damper is
mÿ + bẏ + ky = u
or
ÿ = −

b
k
1
ẏ − y + u .
m
m
m

Select the state variables
x1 = y

and x2 = ẏ .

Then, we have
ẋ = Ax + Bu
y = Cx

0
0
rKm
Lm J T








from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 3

State Variable Models

where


A=

0

1

−20 −9





 ,

0

B=

1



 ,

C=

h

1 0

i

.

A is the system matrix. The characteristic equation is


det[λI − A] = det 

s

−1

20 s + 9



 = s2 + 9s + 20 = 0 .

The roots of the characteristic equation are s1 = −4 and
and the transistion matrix is


Φ(t) = 

5e−4t − 4e−5t

e−4t − e−5t

−20e−4t + 20e−5t −4e−4t + 5e−5t

s2 = −5 ,



 .

(b) Assume the initial conditions are x1 (0) = 1 and x2 (0) = 2. The zeroinput response is shown in Figure DP3.1.
(c) Suppose we redesign the system by choosing b and k to quickly damp
out x2 and x1 . We can select b and k to achieve critical damping.
Critical damping: b/m=20, k/m=100

b/m=9, k/m=20
2

2

1.5
1
x1

1
0

x1
State response, x

0.5
State response, x

122

0

x2

-1

-0.5
-2
x2

-1

-3
-1.5

-2

0

0.5

1
Time(sec)

FIGURE DP3.1
Zero input state response.

1.5

2

-4

0

0.5

1
Time(sec)

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

123

Design Problems

If we desire the characteristic polynomial to be pd (s) = (s + 10)2 =
s2 + 20s + 100, then we need b = 20 and k = 100.
DP3.2

The desired transfer function is
6
Y (s)
= 2
.
U (s)
s + 7s + 10
The transfer function derived from the phase variable representation is
Y (s)
d
= 2
.
U (s)
s + bs + a
Therefore, we select d = 6, a = 10 and b = 7.
Assume the aircraft lands precisely on the centerline. The linearized equations of motion are
m3 ẍ3 + KD ẋ3 + K2 (x3 − x2 ) = 0
m2 ẍ2 + K2 (x2 − x3 ) + K1 (x2 − x1 ) = 0
2
m1 ẍ1 = − √ K2 (x1 − x2 )
2
where x1 (0) = x2 (0) = ẋ2 (0) = ẋ3 = 0 and ẋ1 (0) = 60. The system
response is shown in Figure DP3.3 where KD = 215. The aircraft settles
out at 30 m, although initially it overshoots by about 10 m at 1 second.

45
40
35
30

Amplitude

DP3.3

25
20
15
10
5
0
0

1

2

3

4

5
Time (secs)

FIGURE DP3.3
Aircraft arresting gear response.

6

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

124

CHAPTER 3

DP3.4

State Variable Models

We can model the bungi cord system as a mass-spring-damper. This is
actually an over-simplification because the bungi cord cannot “push” the
jumper down as a spring would—it can only exert a restoring force when
the cord is stretched (that is, when the jumper exceeds the length, L, of
the cord). The problem is nonlinear! When the distance of the jumper
from the platform is less than L we should model the cord spring constant
and damping as K = 0 and b = 0, respectively. Only gravity acts on the
jumper. Also, when ẋ (the jumper velocity) is negative (where we define
positive towards the ground), then we should model b = 0. A reasonable
set of equations of motion are
ẋ1 = x2
K
b
ẋ2 = − x1 − x2 + g
m
m
where x1 is the distance measured from the top of the platform and x2
is the jumper velocity. For the initial conditions we have x1 (0) = 10
and x2 (0) = 0. A reasonable set of parameters for the bungi cord are
L = 40 m, K = 40 N/m and b = 20 kg/m. The system response is
shown in Figure DP3.4 for a person with m = 100 kg. The accelerations
experienced by the jumper never exceed 1.5 g.

global MASS GRAVITY LENGTH K b
MASS=100; HEIGHT=100; GRAVITY=9.806;
LENGTH=40; SPRINGCONSTANT=40; SPRINGDAMPING=20;
x0=[10;0];
t=0; dt=0.1;
n=round(120/dt);
for i=1:n;
if x0(1) 1. For example, we can take K = 8.
(b) The transfer function from Td (s) to ω(s) is given by
−10s
ω(s)
= 2
.
Td (s)
s + 10s + 100
The error plot is shown in Figure AP4.4, where e(s) = −ω(s) (V (s) =
0.)

0.2
0.18
0.16
0.14

e(t)

0.12
0.1
0.08
0.06
0.04
0.02
0
0

1

2

3

4

5

6

7

8

9

10

Time (sec)

FIGURE AP4.4
Error plot with a ramp disturbance input.

AP4.5

(a) The transfer function from the disturbance Td (s) to the output Y (s)
is
Y (s)
−s
= 3
.
2
Td (s)
s + 4s + 4s + K
The steady-state error (when Td (s) = 1/s) is
ess = lim s
s→0

s3

+

4s2

1
s
=0.
+ 4s + K s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

158

CHAPTER 4

Feedback Control System Characteristics

(b) The closed-loop transfer function is
Y (s)
K
= 3
.
R(s)
s + 4s2 + 4s + K
The steady-state error (when R(s) = 1/s2 ) is
ess = lim s(1 − T (s))
s→0

s3 + 4s2 + 4s
4
1
=
lim
=
.
2
3
2
s→0
s
s(s + 4s + 4s + K)
K

(c) Let K = 8. Then,
Y (s)
−s
= 3
.
Td (s)
s + 4s2 + 4s + 8
The error plot is shown in Figure AP4.5, for r(t) = 0.

0.15

0.1

e(t)

0.05

0

-0.05

-0.1

0

2

4

6

8

10

12

Time (sec)

FIGURE AP4.5
Error plot with a step disturbance input and K =8.

AP4.6

(a) The transfer function is
Vo (s)
1 + RCs
=
.
V (s)
2 + RCs

14

16

18

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

159

(b) The system sensitivity is defined as
SCG =

∂G/G
.
∂C/C

Therefore, the sensitivity is determined to be
SCG =

RCs
=
(2 + RCs)(1 + RCs)
1+

1
2
RCs

(c) Let V (s) = 1/s. Then

Vo (s) =



1+

1
RCs

 .

1 + RCs 1
0.5
0.5RC
=
+
.
2 + RCs s
s
RCs + 2

Taking the inverse Laplace transform yields
vo (t) = 0.5(1 + e−2t/RC )u(t)
where u(t) is the unit step function. A plot of vo (t) versus t/RC is
shown in Figure AP4.6.

1
0.95
0.9
0.85

Vo

0.8
0.75
0.7
0.65
0.6
0.55
0.5

0

0.5

1

1.5

2

2.5
t / RC

FIGURE AP4.6
Step response.

3

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

160

CHAPTER 4

AP4.7

Feedback Control System Characteristics

(a) The transfer function from Td (s) to Y (s) is
s
Y (s)
=−
.
Td (s)
s(s + 1) + K
(b) The transfer function from N (s) to Y (s) is
Y (s)
K
=
.
N (s)
s(s + 1) + K
(c) Let Td (s) = A/s and N (s) = B/s. Then,
ess = −yss = lim s
s→0

A
K
B
s
− lim s
= −B .
s(s + 1) + K s s→0 s(s + 1) + K s

So, K has no effect on the steady-state errors. However, choosing
K = 100 will minimize the effects of the disturbance Td (s) during the
transient period.
AP4.8

(a) The closed-loop transfer function is
T (s) =

Kb
.
s + Kb + 2

(b) The sensitivity is determined to be
SbT =

∂T /T
s+2
=
.
∂b/b
s + Kb + 2

(c) The transfer function from Td (s) to Y (s) is
Y (s)
b
=
.
Td (s)
s + Kb + 2
So, choose K as large as possible, to make Y (s)/Td (s) as “small” as
possible. Thus, select
K = 50 .
This also minimizes SbT at low frequencies.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

161

Design Problems

Design Problems
The model of the traction drive, capstan roller, and linear slide was developed in CDP2.1:
θ(s)
Km
=
.
Va (s)
s [(Lm s + Rm )(JT s + bm ) + Kb Km ]
The step response for the closed-loop system (with the tachometer not in
the loop) and various values of the controller gain Ka is shown below.

% System parameters
Ms=5.693; Mb=6.96; Jm=10.91e-03; r=31.75e-03;
bm=0.268; Km=0.8379; Kb=0.838; Rm=1.36; Lm=3.6e-03; Lm=0;
% Controller gain
Ka=100;
% Motor and slide model
Jt=Jm+r^2*(Ms+Mb);
num=[Km];
den=[Lm*Jt Rm*Jt+Lm*bm Kb*Km+Rm*bm 0];
sys=tf(num,den);
%Closed-loop tf and step response
sys_cl=feedback(Ka*sys,[1]);
step(sys_cl)
1.5

Ka=2
Ka=5
Ka=10
Ka=100

1
Theta step response

CDP4.1

0.5

0

0

0.2

0.4

0.6
Time (sec)

0.8

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

162

CHAPTER 4

(a) The transfer function from the load disturbance to the output speed
is
−G(s)
−s
ω(s)
=
= 2
.
Td (s)
1 + Gc G(s)
s + 4s + K
Thus, the effect on ω(s) (of a unit step disturbance) at steady-state
is
lim ω(t) = lim s

t→∞

s→0



−s
2
s + 4s + K



1
=0.
s

We see that the load disturbance has no effect on the output at steadystate.
(b) The system response for 10 ≤ K ≤ 25 is shown in Figure DP4.1.
K=10,12,16,18,20,23,25
100.04
100.02
100

K=25

99.98
99.96

w(t)

DP4.1

Feedback Control System Characteristics

99.94

K=10

99.92
99.9
99.88
99.86
99.84
0

0.5

1

1.5

2

2.5

3

Time(sec)

FIGURE DP4.1
Speed control system response.

For example , if we select K = 16, then ωn = 4, ζ =
response due to a unit step disturbance is
−s
ω(s) = 2
s + 4s + 16

1
s

 

=

1
2,

−1
.
(s + 2)2 + 12

Hence, if we are originally at ω(t) = 100 for t < τ , we have
√
1
ω(t) = 100 − √ e−2t sin 12t
12

t≥τ .

and the

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

163

Design Problems

DP4.2

With θd = 0, we have
θ(s) =

s
G(s)
T (s) = 3
Td (s) .
KK1 d
2
s + 4s + 9s + KK1
1 + G(s) s

For Td = A/s, we have
θ(s) =

A
.
s3 + 4s2 + 9s + KK1

The system response to a unit step disturbance for various values of KK1
are shown in Figure DP4.2. From the plot we see that when KK1 is small
the response is slow but not oscillatory. On the other hand, when KK1
is large the response is fast but highly oscillatory. In fact, if KK1 > 35,
the system is unstable. Thus, we might select KK1 = 10 as a reasonable
trade-off between fast performance and stability.
Unit step response for KK1=1,5,10,15,20,25

0.12
0.1

KK1=1

0.08
0.06

KK1=5

q

0.04
0.02
0
-0.02
KK1=25

-0.04
-0.06

0

1

2

3

4

5

6

7

8

time(sec)

FIGURE DP4.2
Aircraft roll angle control system response to a disturbance.

DP4.3

(a) The closed-loop transfer function is
T (s) =

ω(s)
K
= 2
.
ωd (s)
s + 5s + KK1

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

164

CHAPTER 4

Feedback Control System Characteristics

Then,
E(s) = (1 − T (s))ωd (s) =

s2 + 5s + K(K1 − 1) 1
.
s2 + 5s + KK1
s

So, if
0.99 < K1 < 1.01 ,
then
|ess | < 0.01 .
(b) The transfer function from Td (s) to ω(s) is
ω(s) =

s2

−s
Td (s) .
+ 5s + KK1

So, with E(s) = −ω(s) and Td (s) = 2/s2 , we have
lim sE(s) =

s→0

2
.
KK1

Therefore, we select KK1 > 20 to obtain ess < 0.1.
DP4.4

The steady-state error for a step input command is zero for any K1 . The
transfer function from Td (s) to Y (s) is
Y (s)
G(s)
2
=
= 3
.
2
Td (s)
1 + KG(s)
s + 5s + 4s + 2K
Thus, the output at steady-state due to a step disturbance Td (s) = A/s
is
lim sY (s) =

s→0

A
.
K

We want to maximize K to reduce the effect of the disturbance. As we will
see in Chapter 6, we cannot select K too high or the system will become
unstable. That is why the problem statement suggests a maximum gain
of K = 10. For the design we choose
K = 10 .
DP4.5

The transfer function from V (s) to Vo (s) is
Vo (s)/V (s) =

ks
s+a

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

165

Design Problems

where
k=

R2 + R3
R2

and a =

1
.
R1 C

Computing the step response, we find that
vo (t) = ke−at = 5e−100t .
Solving for R1 , R2 , R3 and C yields
R1 C = 0.01
DP4.6

and

R2
=4.
R3

(a) The closed-loop transfer function is
θ(s) =

s2

K/J
θd (s) .
+ K/J

Since J > 0, the system is unstable when K < 0 and marginally
stable when K > 0.
(b) Since the system is marginally stable, the system response does not
have a steady-state value—it oscillates indefinitely.
(c) The closed-loop transfer function is
θ(s) =

KD s + KP
θd (s) .
Js2 + KD s + KP

The system is stable for all KD > 0 and KP > 0, given that J > 0.
(d) The tracking error E(s) = θd (s) − θ(s) is
E(s) =

Js2
.
Js2 + KD s + KP

Therefore, using the final value theorem we obtain the steady-state
value
lim sE(s) = lim s

s→0

DP4.7

s→0

Js2
1
· =0.
2
Js + KD s + KP s

(a) The closed-loop transfer function is
Y (s) =

s2

s
1
Td (s) = 2
+ Ks + 2K
s + Ks + 2K

where the disturbance is a unit step Td (s) = 1/s. Considering the
poles of the closed-loop system, we find that when K > 8 the system

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

166

CHAPTER 4

Feedback Control System Characteristics

has two real poles. In that case the disturbance step response is


1
y(t) = √
e−αt − e−βt ,
K 2 − 8K

where
α=

K−

√

K 2 − 8K
2

and β =

K+

√

K 2 − 8K
2

Bounding the maximum y(t) yields the inequality
|y(t)| = √

K2

1
e−αt − e−βt ≤ 0.05.
− 8K

We know that
e−αt − e−βt ≤ 1,
for any α and β computed as shown above where K > 8. So, if we
choose K such that
√

1
≤ 0.05.
K 2 − 8K

we will guarantee that the maximum bound of 0.05 is not exceeded.
Solving for K yields
K > 24.4.
For any K > 24.4 we know that the maximum value of the disturbance
step response will be less than 0.05. When K = 24.4 the maximum
unit step disturbance response is 0.035. Solving explicitly for K so
that the maximum is 0.05 we find that K = 16.3 (this was found
numerically since it is very difficult to obtain analytically).
(b) Since the system is type 2, we know that the steady-state value of
the disturbance step response is zero for a unit step disturbance.
DP4.8

(a) The sensitivities are
"

#

"

#

SτT1

∂T /T
−s2 (τ2 s + 1) τ1
=
=
T (s)
∂τ1 /τ1
K

SτT2

∂T /T
−s2 (τ1 s + 1) τ2
=
=
T (s)
∂τ2 /τ2
K

and

where we assume that K 6= 0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

167

Design Problems

(b) Computing the closed-loop transfer function yields
Y (s) =

s (τ1 s + 1)
Td (s)
s (τ1 s + 1) (τ2 s + 1) + K

When Td (s) = 1/s, using the final value theorem we find that
s (τ1 s + 1)
=0
s→0 s (τ1 s + 1) (τ2 s + 1) + K

lim sY (s) = lim

s→0

as long as K 6= 0. We assume here that final value theorem applies
(i.e., the system is stable, more on this in Chapter 6).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

168

CHAPTER 4

Feedback Control System Characteristics

Computer Problems
CP4.1

The step response and an m-file script which generates the step response
is shown in Figure CP4.1. The closed-loop transfer function is
T (s) =

s2

12
.
+ 2s + 22

The percent overshoot is P.O. = 50.2% and the steady-state error is
ess = 0.45.

Step Response
0.9

0.8

System: sys_cl
Peak amplitude: 0.82
Overshoot (%): 50.2
At time (sec): 0.67

0.7

Amplitude

0.6

num = [12]; den = [1 2 10];
sys = tf(num,den);
sys_cl = feedback(sys,[1])
step(sys_cl)

0.5

0.4

0.3

0.2

0.1

0

0

1

2

3
Time (sec)

4

5

6

FIGURE CP4.1
Step response.

CP4.2

The transfer function is
G(s) =

s2

4
.
+ 2s + 20

An m-file script which generates the step response is shown in Figure CP4.2.
The step response is also shown in Figure CP4.2. The step response is
generated using the step function. In the script, the transfer function
numerator is represented by num and the denominator is represented by
den. The steady-state value is yss = 0.2 and the desired value is 1.0.
ess = 0.8 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

169

Computer Problems

Step Response
0.35

0.3

0.25

Amplitude

num=[4]; den=[1 2 20];
sys = tf(num,den);
axis([0 6 0 1]);
t=[0:0.01:6];
step(sys,t)
y = step(sys,t);
yss = y(length(t))

0.2

0.15

0.1

yss =
0.20

0.05

0

0

1

2

3
Time (sec)

4

5

6

FIGURE CP4.2
Step response.

CP4.3

The step responses and the m-file script which generates the step responses is shown in Figure CP4.3.

7

K=10
K=200
K=500

6

K=[10,200,500];
t=[0:0.01:7];
for i=1:3
num=5*K(i); den=[1 15 K(i)];
sys = tf(num,den)
y(:,i)= step(sys,t);
end
plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':')
legend('K=10','K=200','K=500',-1)

5

4

3

2

1

0

FIGURE CP4.3
Step responses for K = 10, 100, 500.

0

1

2

3

4

5

6

7

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

170

CHAPTER 4

CP4.4

Feedback Control System Characteristics

(a,b) The m-file and plots are shown in Figure CP4.4.

Step response

1.5

y(t) 1

ng=1;dg=[1 1.91 0];sysg=tf(ng,dg);
K=10;
syscl=feedback(K*sysg,1);
figure(1)
subplot(211)
step(syscl)
subplot(212)
syst=feedback(sysg,K)
step(syst)

0.5

0

0

1

2

3
Time (sec)

4

5

6

Disturbance response

0.2
0.15

y(t) 0.1
0.05
0

0

1

2

3
Time (sec)

4

5

6

FIGURE CP4.4
Step response and disturbance response.

(c) The estimated steady-state tracking error due to a unit step input is
zero, and the estimated steady-state tracking error to a unit disturbance is 0.1.
(d) The estimated maximum tracking error due to a unit step input is
0.4, and the estimated maximum tracking error to a unit disturbance
is 0.14. The maximum occurs at approximately t = 1 s.
CP4.5

The step response and the m-file script which generates the step response
is shown in Figure CP4.5. The closed-loop transfer function is determined
to be
T (s) =

s2

10
.
+ 3.7s + 10

Using the m-file script, a trial-and-error search on k yields
k = 3.7 .
The percent overshoot P.O. = 10% and the steady-state value is 1, as
expected.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

171

Computer Problems
1.4

1.2

1

0.8
y(t)

k = 3.7; % Final value of k=3.7
numcg = [10]; dencg = [1 k 0];
sys_o = tf(numcg,dencg);
sys_cl = feedback(sys_o,[1])
t = [0:0.1:5];
[y,t] = step(sys_cl,t);
plot(t,y,[0 5],[1.1 1.1],'--'); grid
xlabel('Time (sec)'); ylabel('y(t)');

0.6

0.4

Transfer function:
10
---------------s^2 + 3.7 s + 10

0.2

0

0

0.5

1

1.5

2

2.5
Time (sec)

3

3.5

4

4.5

5

FIGURE CP4.5
Step response.

CP4.6

The closed-loop transfer function is
T (s) =

K
s−a+K

where K = 2. When a = 1 and R(s) = 1/s, the final value is
lim sT (s)R(s) = lim T (s) =

s→0

s→0

K
=2.
K −a

The output is within 2% of the final value at around t = 4.6 seconds. The
plot of the step responses for
a = 1, 0.5, 2, 5
is shown in Figure CP4.6. The output is unstable for
a > 2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

172

CHAPTER 4

Feedback Control System Characteristics

K=2; t=[0:0.1:5];
num=K*[1];
a=[1 0.5 2 5];
for i=1:4
den=[1 -a(i)]; sys = tf(num,den);
sys_cl = feedback(sys,[1]);
y(:,i)=step(sys_cl,t);
end
plot(t,y(:,1),t,y(:,2),':',t,y(:,3),'--',t,y(:,4),'-.')
axis([0 5 0 5]);
xlabel('Time (sec)'), ylabel('y(t)')
title('a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot)')
a=1 (solid); a=0.5 (dotted); a=2 (dashed); a=5 (dashdot)
5

4.5

4

3.5

y(t)

3

2.5

2

1.5

1

0.5

0

0

0.5

1

1.5

2

2.5
Time (sec)

3

3.5

4

4.5

5

FIGURE CP4.6
Step response for a=1, 0.5, 2, and 5.

CP4.7

The transfer function from the disturbance to the output is
T (s) =

G(s)
1
=
.
2
1 + K0 G(s)
Js + bs + k + K0

The disturbance response is shown in Figure CP4.7. The compensated
system response is significantly reduced from the uncompensated system
response. The compensated system output is about 11 times less than the
uncompensated system output. So, closed-loop feedback has the advantage of reducing the effect of unwanted disturbances on the output.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

173

Computer Problems

J=1; k=5; c=0.9;
num=[1/J]; den=[1 c/J k/J];
sys = tf(num,den);
t=[0:0.1:10];
%
yu=step(sys,t); % Part (a)
K0=50;
numk=[K0]; denk=[1]; sysk = tf(numk,denk);
sys_cl = feedback(sys,sysk);
yc=step(sys_cl,t); % Part (b)
plot(t,yu,t,yc,'--')
xlabel('Time (sec)'), ylabel('\theta')
title('Uncompensated response (solid) & Compensated response (dashed)')
Uncompensated response (solid) & Compensated response (dashed)
0.35

0.3

0.25

q

0.2

0.15

0.1

0.05

0

0

1

2

3

4

5
Time (sec)

6

7

8

9

10

FIGURE CP4.7
Disturbance responses for both the uncompensated and compensated systems.

CP4.8

The step responses for the proportional and PI controller are shown in
Figure CP4.8. The steady-state tracking error for the proportional controller is
ess = 0.33 .
Increasing the complexity of the controller from a proportional controller
to a proportional plus integral (PI) controller allows the closed-loop system to track the unit step response with zero steady-state error. The cost
is controller complexity, which translates into higher costs ($). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 174 CHAPTER 4 Feedback Control System Characteristics numg=[10]; deng=[1 10]; sysg = tf(numg,deng); t=[0:0.001:0.5]; % Part (a) numc=[2]; denc=[1]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yk=step(sys_cl,t); % Part (b) numc=[2 20]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]); yp=step(sys_cl,t); % plot(t,yk,t,yp,'--') xlabel('Time (sec)'),ylabel('y(t)') title('Proportional controller (solid) & PI controller (dashed)') Proportional controller (solid) & PI controller (dashed) 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.05 0.1 0.15 0.2 0.25 Time (sec) 0.3 0.35 0.4 0.45 0.5 FIGURE CP4.8 Step response for proportional controller and PI controller. CP4.9 (a) The closed-loop transfer function is T (s) = G(s) 10s2 + 500s R(s) = 2 R(s) . 1 + G(s)H(s) s + 200s + 5000 The step response is shown in Figure CP4.9a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 175 Computer Problems (b) The response of the system to the sinusoidal disturbance N (s) = s2 100 + 100 is shown in Figure CP4.9b. (c) In the steady-state, the magnitude of the peak response is 0.095 and the frequency is 10 rad/sec (see Figure CP4.9b). % Part (a) ng=10*[1 0]; dg=[1 100]; sysg=tf(ng,dg); nh=[5]; dh=[1 50]; sysh=tf(nh,dh); sys=feedback(sysg,sysh) figure(1) step(sys) >> Transfer function: 10 s^2 + 500 s -----------------s^2 + 200 s + 5000 Step Response % Part (b) sysn=-feedback(sysg*sysh,1) syss=tf([100],[1 0 100]); % This is the sinusoidal input figure(2) t=[0:0.001:7]; step(syss*sysn,t) 10 Amplitude 8 6 4 2 Step Response 0.1 0 0.08 0 0.02 0.04 0.06 0.08 0.1 Time (sec) 0.06 (a) Amplitude 0.04 0.02 0 −0.02 −0.04 −0.06 −0.08 −0.1 0 1 2 3 4 Time (sec) 5 6 7 (b) FIGURE CP4.9 (a) Unit step response. (b) Response to sinusoidal noise input at ω = 10 rad/sec. CP4.10 (a) The closed-loop transfer function is T (s) = Gc (s)G(s) K(s + 1) R(s) = R(s) . 2 1 + G(s)Gc (s) (s + 15)(s + s + 6.5) + K(s + 1) (b) The step responses are shown in Figure CP4.10a. 0.12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 176 CHAPTER 4 Feedback Control System Characteristics (c) The unit disturbance response of the system is shown in Figure CP4.10b. The steady-state value is 0.14. Step Response 0.7 K=5 K=10 K=50 0.6 0.2 0.18 0.16 System: syscl Final Value: 0.14 0.5 0.4 Amplitude Step response 0.14 0.3 0.12 0.1 0.08 0.2 0.06 0.04 0.1 0.02 0 0 0 1 2 3 4 Time (s) 5 6 7 0 1 2 3 Time (sec) (a) 4 5 6 (b) FIGURE CP4.10 (a) Unit step responses for K = [5, 10, 50]. (b) Disturbance unit step response. The m-file is shown in Figure CP4.11a and the step responses in Figure CP4.11b. CP4.11 5 K=10 K=12 K=15 4 3 Step response K=[10, 12, 15]; t=[0:0.1:20]; ng=[20]; dg=[1 4.5 64]; sysg=tf(ng,dg); nh=[1]; dh=[1 1]; sysh=tf(nh,dh); for i=1:length(K) sys=K(i)*sysg; syscl=feedback(sys,sysh) y(:,i)= step(syscl,t); end plot(t,y(:,1),t,y(:,2),'--',t,y(:,3),':') xlabel('Time (s)') ylabel('Step response') legend('K=10','K=12','K=15',-1) 2 1 0 −1 −2 −3 0 5 (a) FIGURE CP4.11 (a) M-file script. (b) Unit step responses for K = [10, 12, 15]. 10 Time (s) (b) 15 20 7 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 5 The Performance of Feedback Control Systems Exercises E5.1 For a zero steady-state error, when the input is a step we need one integration, or a type 1 system. A type 2 system is required for ess = 0 for a ramp input. E5.2 (a) The closed-loop transfer function is T (s) = Y (s) G(s) 240 240 = = = 2 . R(s) 1 + G(s) (s + 4)(s + 6) + 240 s + 2ζωn s + ωn2 The steady-state error is given by ess = A , 1 + Kp where R(s) = A/s and Kp = lim G(s) = s→0 240 = 10 . 24 Therefore, ess = A . 11 (b) The closed-loop system is a second-order system with natural frequency √ ωn = 264 , 177 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 178 CHAPTER 5 The Performance of Feedback Control Systems and damping ratio 10 ζ= √ = 0.31 . 2 264 The percent overshoot is thus computed to be √ 2 P.O. = 100e−πζ/ 1−ζ = 36% . E5.3 The closed-loop transfer function is G(s) K K Y (s) = = = 2 . I(s) 1 + G(s) s(s + 14) + K s + 14s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 14s + K , so equating coefficients yields ωn = 10 and K = ωn2 = 100 . We can also compute the damping ratio as ζ= 14 = 0.7 . 2ωn Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.4 (a) The closed-loop transfer function is T (s) = G(s) 2(s + 8) = 2 . 1 + G(s) s + 6s + 16 (b) We can expand Y (s) in a partial fraction expansion as 2(s + 8) A 1 s+4 Y (s) = 2 =A − 2 (s + 6s + 16) s s s + 6s + 16   . Taking the inverse Laplace transform (using the Laplace transform tables), we find √ y(t) = A[1 − 1.07e−3t sin( 7t + 1.21)] . (c) Using the closed-loop transfer function, we compute ζ = 0.75 and © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 179 Exercises ωn = 4. Thus, 8 a = = 2.67 , ζωn 3 where a = 8. From Figure 5.13(a) in Dorf & Bishop, we find (approximately) that P.O. = 4% . (d) This is a type 1 system, thus the steady-state error is zero and y(t) → A as t → ∞. E5.5 The closed-loop transfer function is Y (s) K = 2 . R(s) s + 4s + K Utilizing Table 5.6 in Dorf & Bishop, we find that the optimum coefficients are given by s2 + 1.4ωn s + ωn2 . We have s2 + 4s + K , so equating coefficients yields ωn = 2.86 and K = ωn2 = 8.16 . We can also compute the damping ratio as ζ= 4 = 0.7 . 2ωn Then, using Figure 5.8 in Dorf & Bishop, we find that P.O. ≈ 5%. E5.6 (a) The closed-loop transfer function is T (s) = Y (s) G(s) 100 = = 2 , R(s) 1 + GH(s) s + 100Ks + 100 where H(s) = 1 + Ks and G(s) = 100/s2 . The steady-state error is computed as follows: ess = lim s[R(s) − Y (s)] = lim s[1 − T (s)] s→0 s→0 " = lim 1 − s→0 1+ 100 s2 100 (1 + s2 A s2 # A = KA . Ks) s (b) From the closed-loop transfer function, T (s), we determine that ωn = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 180 CHAPTER 5 The Performance of Feedback Control Systems 10 and ζ= 100K = 5K . 2(10) We want to choose K so that the system is critically damped, or ζ = 1.0. Thus, K= 1 = 0.20 . 5 The closed-loop system has no zeros and the poles are at p s1,2 = −50K ± 10 25K 2 − 1 . The percent overshoot to a step input is P.O. = 100e √−5πK 1−25K 2 for 0 < K < 0.2 and P.O. = 0 for K > 0.2. E5.7 The closed-loop transfer function is Y (s) KG(s) K(s + 2) K(s + 2) = = = 2 . R(s) 1 + KG(s) s(s + 1) + K(s + 2) s + s(K + 1) + 2K √ √ Therefore, ωn = 2K and ζ = 2K+1 . So, 2K T (s) = a 4 = . ζωn K +1 From Figure 5.13a in Dorf & Bishop, we determine that a ≈ 1.5 ζωn when ζ = 0.707. Thus, solving for K yields 4 = 1.5 K +1 or K = 1.67. E5.8 The pole-zero map is shown in Figure E5.8. Since the dominant poles are at s = −2 ± 2.45i we have a damping ratio ζ = 0.63. We estimate the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 7.69% The step response is shown in Figure E5.8b. The actual overshoot is 8%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 181 Exercises Pole−Zero Map 2.5 2 1.5 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −2 −2.5 −25 −20 −15 −10 −5 0 Real Axis Step Response 1.4 System: sys Time (sec): 1.28 Amplitude: 1.08 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 FIGURE E5.8 (a) Pole-zero map. (b) Unit step response. E5.9 (a) The closed-loop transfer function is T (s) = s2 + √ K . 2Ks + K 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 182 CHAPTER 5 The Performance of Feedback Control Systems √ The √ damping ratio is ζ = 2/2 and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% for ζ = 0.707. The settling time is estimated via Ts = 8 4 =√ . ζωn 2K (b) The settling time is less than 1 second whenever K > 32. E5.10 The second-order closed-loop transfer function is given by T (s) = ωn2 . s2 + 2ζωn s + ωn2 From the percent overshoot specification, we determine that P.O. ≤ 5% implies ζ ≥ 0.69. From the settling time specification, we find that Ts < 4 implies ωn ζ > 1. p And finally, from the peak time specification we have Tp < 1 implies ωn 1 − ζ 2 > π. The constraints imposed on ζ and ωn by the performance specifications define the permissible area for the poles of T (s), as shown in Figure E5.10. Im(s) wn z = 0.69 1-z 2 = P desired region 46o for poles wn Re(s) 1-z 2 = - P z wn = -1 FIGURE E5.10 Permissible area for poles of T (s). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 183 Exercises E5.11 The system is a type 1. The error constants are Kp = ∞ and Kv = 1.0 . Therefore, the steady-state error to a step input is 0; the steady-state error to a ramp input is 1.0A0 , where A0 is the magnitude (slope) of the ramp input. (a) The tracking error is given by E(s) = (s + 9)(s + 2)(s + 4) R(s) = R(s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6) The steady-state tracking error (with R(s) = 1/s) is lim sE(s) = s→0 72 . 72 + 6K We require ess < 0.05, so solving for K yields K > 228. (b) The tracking error due to the disturbance is E(s) = −G(s) −(s + 9)(s + 6) Td (s) = Td (s) . 1 + Gc G(s) (s + 9)(s + 2)(s + 4) + K(s + 6) The tracking error is shown in Figure E5.12. 0 -0.01 -0.02 -0.03 Amplitude E5.12 -0.04 -0.05 -0.06 -0.07 -0.08 0 0.1 0.2 0.3 0.4 0.5 Time (secs) FIGURE E5.12 Tracking error due a step disturbance. 0.6 0.7 0.8 0.9 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 184 CHAPTER 5 E5.13 The Performance of Feedback Control Systems The system is a type 0. The error constants are Kp = 0.4 and Kv = 0. The steady-state error to a ramp input is ∞. The steady-state error to a step input is ess = E5.14 1 = 0.71. 1 + Kp (a) The tracking error is given by E(s) = [1 − T (s)] R(s) . The steady-state tracking error (with R(s) = 1/s) is ess = lim s [1 − T (s)] R(s) = lim [1 − T (s)] = 1 − T (0) . s→0 s→0 The closed-loop transfer function is T (s) = K(s + 0.1) , s(s + 0.1)(s + 2) + K(s + 3) and T (0) = 0.033. Therefore, ess = 1 − T (0) = 0.967. (b) Use Gp (s) = 30. Then, ess = lim s [1 − T (s)Gp (s)] R(s) = 1−lim T (s)Gp (s) = 1−30 T (0) = 0 . s→0 The plot of y(t) is shown in Figure E5.15. 1.4 1.2 Response using only dominate poles 1 0.8 y(t) E5.15 s→0 0.6 Actual response 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 FIGURE E5.15 Plot of y(t) with T (s) (solid line) and approximate Ta (s) (dashed line). 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 185 Exercises Using the dominant poles, the approximate closed-loop transfer function is Ta (s) = s2 50 . + 10s + 50 The actual transfer function is T (s) = E5.16 500 . (s + 10)(s2 + 10s + 50) The partial fraction expansion is y(t) = − 10(z − 1) −t 10(z − 8) −8t e + e + 1.25 . 7z 56z The plot of y(t) for z = 2, 4, 6 is shown in Figure E5.16. z=2 (solid) & z=4 (dashed) & z=6 (dotted) 1.4 1.2 1 y(t) 0.8 0.6 0.4 0.2 0 0 1 2 3 Time (sec) 4 5 6 FIGURE E5.16 Plot of y(t) for z=2, 4, 6. E5.17 The desired pole locations for the 5 different cases are shown in Figure E5.17. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 186 CHAPTER 5 The Performance of Feedback Control Systems Im(s) desired region for poles wn = 10 37 o o 53 Re(s) (a) 0.6 < z < 0.8 and wn <10 desired region for poles Im(s) wn = 10 45 o o Re(s) 60 (b) 0.5 < z < 0.707 and wn > 1 0 Im(s) desired region for poles wn = 10 o Re(s) 60 wn = 5 (c) 0.5 < z and 5 < wn <10 FIGURE E5.17 Desired pole locations. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 187 Exercises Im(s) desired region for poles wn = 10 o Re(s) 45 wn = 5 (d) 0.707 > z and 5 < wn <10 Im(s) o wn = 6 Re(s) 53 desired region for poles (e) 0.6 < z and wn < 6 FIGURE E5.17 CONTINUED: Desired pole locations. E5.18 The output is given by Y (s) = T (s)R(s) = K G(s) R(s) . 1 + G(s) When K = 1, the steady-state error is ess = 0.2 which implies that lim sY (s) = 0.8 . s→0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 188 CHAPTER 5 The Performance of Feedback Control Systems Since we want ess = 0, it follows that lim sY (s) = 1 , s→0 or 0.8K = 1 . Therefore, K = 1.25. (a) The characteristic equation is s2 = 2ζωn s + ωn2 = s2 + 3.17s + 7 = 0 , from which it follows that √ ωn = 7 = 2.65, ζ= 3.17 = 0.6 . 2ωn Therefore, we compute the percent overshoot and the estimated settling time to be √ 4 2 P.O. = 100e−πζ/ 1−ζ = 9.53% and Ts = = 2.5 s . ζωn (b) The unit step response is shown in Figure E5.19. Step Response 1.4 System: sys Peak amplitude: 1.1 Overshoot (%): 9.53 At time (sec): 1.47 1.2 System: sys Settling Time (sec): 2.25 1 Amplitude E5.19 0.8 0.6 0.4 0.2 0 0 FIGURE E5.19 Unit step response. 0.5 1 1.5 Time (sec) 2 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 189 Exercises E5.20 (a) The closed-loop transfer function is T (s) = s2 + √ K . 2Ks + K The damping ratio is √ 2 2 √ and the natural frequency is ωn = K. Therefore, we compute the percent overshoot to be √ 2 P.O. = 100e−πζ/ 1−ζ = 4.3% ζ= for ζ = 0.707. The settling time is estimated via Ts = 4 8 =√ . ζωn 2K (b) The settling time is less than 1 second whenever K > 32. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 190 CHAPTER 5 The Performance of Feedback Control Systems Problems P5.1 (a) The system error is E(s) = 1 1+ Ka Km sτm +1 R(s) where R(s) = 25o /sec . s So, lim e(t) = lim sE(s) = t→0 s→0 25 . 1 + Ka Km (b) If we desire ess ≤ 1o /sec, then 25o /s ≤ 1o /sec , 1 + Ka Km and solving for Ka Km yields Ka Km ≥ 24 . (c) The closed-loop transfer function is T (s) = Vb (s) Ka Km = . Vc (s) sτm + 1 + Ka Km The step response of the system (i.e. vc (t) = A) is −(Ka Km +1) AKa Km t τm vb (t) = 1−e 1 + Ka Km   . So, at settling time, we have 1−e −(1+Ka Km ) t τm ≥ 0.98 , where τm = 0.4. Setting t = 0.03 and solving for Ka Km yields Ka Km ≥ 52 . P5.2 (a) The settling time specification Ts = 4 < 0.6 ζωn © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 191 Problems is used to determine that ζωn > 6.67. The P.O. < 20% requirement is used to determine ζ < 0.45 which implies θ < 63o and the P.O. > 10% requirement is used to determine ζ > 0.60 which implies θ > 53o , since cos θ = ζ. The desired region for the poles is shown in Figure P5.2. Im(s) desired region for poles 53 o 63 o Re(s) s = -6.67 FIGURE P5.2 Desired region for pole placement. (b) The third root should be at least 10 times farther in the left halfplane, so |r3 | ≥ 10|ζωn | = 66.7 . (c) We select the third pole such that r3 = −66.7. Then, with ζ = 0.45 and ζωn = 6.67, we determine that ωn = 14.8. So, the closed-loop transfer function is T (s) = 66.7(219.7) , (s + 66.7)(s2 + 13.3s + 219.7) where the gain K = (66.7)(219.7) is chosen so that the steady-state © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 192 CHAPTER 5 The Performance of Feedback Control Systems tracking error due to a step input is zero. Then, T (s) = G(s) , 1 + G(s) G(s) = T (s) . 1 − T (s) or P5.3 Given the input R(s) = 1 , s3 we compute the steady-state error as ess 1 = lim s s→0 1 + G(s)   1 1 = lim 2 3 s→0 s s G(s)    = lim  s→0 Since we require that ess ≤ 0.5 cm, we determine 1 s2  K s2   = 1 . K K≥2. P5.4 (a) The closed-loop transfer function is T (s) = G(s) K ωn2 = 2 = 2 . 1 + G(s) s + 2s + K s + 2ζωn s + ωn2 Thus, ωn = √ K and √ ζ = 1/ωn = 1/ K . √ Our percent overshoot requirement√of 5% implies that ζ = 1/ 2 , which in turn implies that ωn = 2. However, the corresponding time to peak would be 4.4 Tp = √ = 3.15 . 2 Our desired Tp = 1.1—we cannot meet both specification simultaneously. (b) Let Tp = 1.1∆ and P.O. = 0.05∆, where ∆ is the relaxation factor to be determined. We have that K = ωn2 and ζωn = 1, so 1 ζ=√ . K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 193 Problems Thus, √ √ 2 P.O. = e−πζ/ 1−ζ = e−π/ K−1 . Also, Tp = √ π = 1.1∆ . K −1 Therefore, from the proceeding two equations we determine that P.O. = 0.05∆ = e−1.1∆ . Solving for ∆ yields f (∆) = ln ∆ + ln(0.05) + 1.1∆ = 0 . The plot of f (∆) versus ∆ is shown in Figure P5.4. From the plot we 2 1 0 * D=2.07 -1 f(D) -2 -3 -4 -5 -6 -7 -8 0 0.5 1 1.5 2 D FIGURE P5.4 Solving for the zeros of f. see that ∆ = 2.07 results in f (∆) = 0. Thus, P.O. = 0.05∆ = 10% Tp = 1.1∆ = 2.3 sec. 2.5 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 194 CHAPTER 5 The Performance of Feedback Control Systems So, we can meet the specifications if they are relaxed by a factor of about 2 (i.e. ∆ = 2.07). P5.5 (a) The closed-loop transfer function is T (s) = s2 K1 K2 (s + 1) . + K1 K2 s + K1 K2 A percent overshoot less than 5% implies ζ ≥ 0.69. So, choose ζ = 0.69. Then set 2ζωn = K1 K2 and ωn2 = K1 K2 . Then 2(0.69)ωn = ωn2 ; and solving for ωn yields ωn = 1.38 . Therefore K1 K2 = ωn2 = 1.9. When K1 K2 ≥ 1.9 it follows that ζ ≥ 0.69. (b) We have a type 2 system, so the steady-state tracking error to both a step and ramp input is zero. (c) For a step input, the optimum ITAE characteristic equation is s2 + 1.4ωn s + ωn2 = 0 . For a ramp input, the optimum ITAE characteristic equation is s2 + 3.2ωn s + ωn2 = 0 . Thus, K1 K2 = ωn2 = 3.2ωn . So, ωn = 3.2 and K1 K2 = 10.24. P5.6 We have a ramp input, r(t) = t. So Kv = lim sG(s) = lim s s→0 s→0  75 75(s + 1) = = 0.6 , s(s + 5)(s + 25) 125  and ess = P5.7 |R| 1 = = 1.67 . Kv 0.6 (a) The closed-loop transfer function is T (s) = Is2 K1 K2 . + K1 K2 K3 s + K1 K2 The steady-state tracking error for a ramp input is ess = lim sE(s) = lim s(1 − T (s))R(s) = lim s(1 − T (s)) s→0 s→0 s→0 1 s2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 195 Problems = lim s→0 Is + K1 K2 K3 = K3 . Is2 + K1 K2 K3 s + K1 K2 But we desire ess = 0.01 m, so K3 = 0.01. (b) For P.O. = 10%, we have ζ = 0.6. Also, 2ζωn = 0.01K1 K2 25 and ωn2 = K1 K2 . 25 Thus, solving for K1 K2 yields K1 K2 = 36 × 104 . P5.8 (a) The closed-loop transfer function is T (s) = P (s) G(s)/s 20 = = . R(s) 1 + G(s)H(s)/s s(s + 40) Therefore, the closed-loop system time constant is τ = 1/40 sec. (b) The transfer function from Td (s) to the output P (s) is P (s) −G(s) −20 = = . Td (s) 1 + G(s)H(s)/s s + 40 The response to a unit step disturbance is 1 p(t) = − (1 − e−40t ) . 2 At settling time, p(t) = 0.98pss = −0.49. Thus, solving for t(= Ts ) we determine that Ts = 0.098 sec. P5.9 We need to track at the rate ω= v 16000 = = 1.78 × 10−3 radians/sec . r 2500 The desired steady-state tracking error is ess ≤ 1 degree = 0.1754 × 10−2 rad . 10 Therefore, with ess = |ω| , Kv © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 196 CHAPTER 5 The Performance of Feedback Control Systems we compute Kv as Kv = 1.78 × 10−3 = 1.02 . 0.1754 × 10−2 This assumes that the system is type 1. (a) The armature controlled DC motorblock diagram is shown in Figure P5.10. P5.10 amplifier R(s ) + Km K 1 J s+b R a+ L as - w(s) Kb back emf FIGURE P5.10 Armature controlled DC motor block diagram. (b) The closed-loop transfer function is T (s) = ω(s) KG(s) = , R(s) 1 + KKb G(s) where G(s) = Km . (Ra + La s)(Js + b) Thus, T (s) = K , s2 + 2s + 1 + K where Ra = La = J = b = Kb = Km = 1. The steady-state tracking error is ess A = lim s(R(s) − Y (s)) = lim s (1 − T (s)) s→0 s→0 s   K A = A(1 − T (0)) = 1 − = . 1+K 1+K   (c) For a percent overshoot of 15%, we determine that ζ = 0.5. √ From our characteristic polynomial we have 2ζωn = 2 and ωn = 1 + K. Solving for ωn yields ωn = 2, thus K = 3. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 197 Problems P5.11 (a) The closed-loop transfer function is T (s) = K Y (s) = . R(s) s+K To include the initial condition, we obtain the differential equation: ẏ(t) + Ky(t) = Kr(t) . Taking the Laplace transform yields: sY (s) − y(to ) + KY (s) = K  A s  , where y(to ) = Q. Computing the inverse Laplace transform, L−1 {Y (s)} yields y(t) = A(1 − e−Kt ) + Qe−Kt . Also, the tracking error is given by e(t) = A − y(t) = e−Kt (A − Q) . Thus, the performance index, I is determined to be (for K > 0) I= Z 0 ∞ 2 −2Kt (A − Q) e 2 dt = (A − Q) = Q)2 (A − 2K  1 −2K  e−2Kt ∞ 0 . (b) The minimum I is obtained when K = ∞, which is not practical. (c) Set K at the maximum value allowable such that the process does not saturate. For example, if K = 50, then I= P5.12 (A − Q)2 . 100 The optimum ITAE transfer function for a ramp input is T (s) = 3.25ωn2 s + ωn3 . s3 + 1.75ωn s + 3.25ωn2 s + ωn3 The steady-state tracking error, ess = 0, for a ramp input. The step response is shown in Figure P5.12 for ωn = 10. The percent overshoot is P.O. = 39%, and the settling time is Ts = 0.72 s . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 198 CHAPTER 5 The Performance of Feedback Control Systems 1.4 PO=39% 1.2 Ts=0.72s 1 y(t) 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 1 time (sec) 1.2 1.4 1.6 1.8 2 FIGURE P5.12 Step input system response. The step responses for the actual system and the approximate system are shown in Figure P5.13. It can be seen that the responses are nearly identical. 1 0.9 0.8 0.7 0.6 y(t) P5.13 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE P5.13 Closed-loop system step response: Actual T(s) (solid line) and second-order approximation (dashed line). © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 199 Problems P5.14 Consider L(s) = 2(c1 s + 1) . (s + 1)(s + 2) After cancellation of like factors, we compute H(s)/L(s), H(s) s3 + 7s2 + 24s + 24 = . L(s) (s + 3)(s + 4)2(c1 s + 1) Therefore, M (s) = s3 + 7s2 + 24s + 24 , and ∆(s) = 2[c1 s3 + (7c1 + 1)s2 + (12c1 + 7)s + 12] . Then, following the procedure outlined in Section 5.10, we have M o (0) = 24, M 1 (0) = 24, M 2 (0) = 14, M 3 (0) = 6, and ∆0 (0) = 24, ∆1 (0) = (12c1 + 7)2, ∆2 (0) = 2(2 · (7c1 + 1)), ∆3 (0) = 12c1 . For q = 1: M2 = 240, and ∆2 = 4[144c21 + 25] . Then, equating ∆2 and M2 , we find c1 , c1 = 0.493 . So, L(s) = P5.15 0.986s + 2 0.986(s + 2.028) 2(0.493s + 1) = 2 = . (s + 1)(s + 2) s + 3s + 2 (s + 1)(s + 2) The closed-loop transfer function is T (s) = (s + 4)(s2 K(s + 1) . + s + 10) + K(s + 1) The percent overshoot as function of the gain, K, is shown in Figure P5.15. It can be seen that the percent overshoot decreases as the gain increases approaching a minimum around 85%. The larger the gain, the smaller the percent overshoot. For a gain K ≈ 250, we have essentially minimized the percent overshoot. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 200 CHAPTER 5 The Performance of Feedback Control Systems 160 150 Percent Overshoot (%) 140 130 120 110 100 90 80 0 50 100 150 200 250 300 350 K FIGURE P5.15 Percent overshoot versus the gain, K. P5.16 The open-loop transfer function is G(s) = 10 . (s + 1)(50Cs + 1) Define τ = 50C. Then, the closed-loop transfer function is Vo (s) 10 10/τ   = = τ +1 2 Vin (s) (s + 1)(τ s + 1) + 10 s + τ s+ 11 τ . With ωn2 = 11 τ τ +1 1 and ζ = √ = , 2τ ωn 2 we can solve for τ , yielding τ 2 − 20τ + 1 = 0 . Therefore, τ = 19.95 and 0.05. For each value of τ we determine C as follows: τ = 19.95 = 50C, implies C = 0.399F , and τ = 0.05 = 50C, implies C = 1mF . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 201 Problems P5.17 (a) The closed-loop transfer function is T (s) = Y (s) 12K = 2 . R(s) s + 12s + 12K The percent overshoot specification P.O. ≤ 10% implies ζ ≥ 0.59. From the characteristic equation we find that ωn2 = 12K Solving for K yields √ 2(0.59) 12K = 12 and ζωn = 6 . which implies that K = 8.6 . So, any gain in the interval 0 < K < 8.6 is valid. The settling time is Ts = 4/ζωn = 4/6 seconds and satisfies the requirement. Notice that Ts is not a function of K. (b) The sensitivity is T SK (s) = 1 s(s + 12) = 2 1 + G(s) s + 12s + 120 when K = 10. (c) The sensitivity at DC (s = 0) is T SK (0) = 0 . (d) In this case, s = j2π · 1 beat/sec = j2π. So, the sensitivity at s = 2πj is T |SK (j2π)| = P5.18 85.1084 = 0.77 . 110.31 We select L(s) as L(s) = 1 , αs + 1 then H(s) 6αs + 6 = 3 . L(s) s + 6s2 + 11s + 6 Therefore, M (s) = 6αs + 6 , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 202 CHAPTER 5 The Performance of Feedback Control Systems and M o (0) = 6, M 1 (0) = 6α, M 2 (0) = 0. Also, ∆(s) = s3 + 6s2 + 11s + 6 , and ∆o (0) = 6 , ∆1 (0) = 11 , ∆2 (0) = 12. So, computing M2 and ∆2 yields M2 = 36α2 , and ∆2 = 49 . Finally, equating M2 = ∆2 yields 36α2 = 49 , or α = 1.167 . Thus, L(s) = P5.19 1 0.857 = . 1.167s + 1 s + 0.857 (a) The closed-loop transfer function is T (s) = s3 + 6s2 8 . + 12s + 8 (b) The second-order approximation is L(s) = d2 s2 1 , + d1 s + 1 where d1 and d2 are to be determined using the methods of Section 5.10 in Dorf & Bishop. Given M (s) = 8d2 s2 + 8d1 s + 8 ∆(s) = s3 + 6s2 + 12s + 8 we determine that M2 M4 ∆2 ∆4 = −128d2 + 64d21 = 64d22 = 48 = 18 . Equating M2 = ∆2 and M4 = ∆4 , and solving for d1 and d2 yields d1 = 1.35 and d2 = 0.53 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 203 Problems 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.5 1 1.5 2 2.5 Time (sec) 3 3.5 4 4.5 5 FIGURE P5.19 Closed-loop system step response: Actual T (s) (solid line) and second-order approximation (dashed line). Thus, the second-order approximation is L(s) = 0.53s2 1 . + 1.35s + 1 (c) The plot of the step response for the actual system and the approximate system is shown in Figure P5.19. P5.20 The steady-state error is (s + 5)(s + 11) + K(1 − K1 ) 55 + K(1 − K1 ) = . s→0 (s + 5)(s + 11) + K 55 + K ess = lim To achieve a zero steady-state tracking error, select K1 as follows: K1 = 1 + P5.21 55 . K The closed-loop transfer function is T (s) = s+a . s2 + (2k + a)s + 2ak + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 204 CHAPTER 5 The Performance of Feedback Control Systems (a) If R(s) = 1/s, we have the tracking error E(s) = R(s) − Y (s) = [1 − T (s)]R(s) or E(s) = s2 + (2k + a − 1)s + 2ak + 1 − a 1 · . s2 + (2k + a)s + 2ak + 1 s From the final value theorem we obtain ess = lim sE(s) = s→0 2ak + 1 − a . 2ak + 1 Selecting k = (a − 1)/(2a) leads to a zero steady-state error due to a unit step input. (b) To meet the percent overshoot specification we desire ζ ≥ 0.69. From T (s) we find ωn2 = 2ak + 1 and 2ζωn = 2k + a. Therefore, solving for a and k yields a = 1.5978 and k = 0.1871 when we select ζ = 0.78. We select ζ > 0.69 to account for the zero in the closed-loop transfer function which will impact the percent overshoot. With a and k, as chosen, we have T (s) = s2 s + 1.598 + 1.972s + 1.598 and the step response yields P.O. ≈ 4%. P5.22 The closed-loop transfer function is T (s) = 2(2s + τ ) . (s + 0.2K)(2s + τ ) + 4 (a) If R(s) = 1/s, we have the unit step response Y (s) = 2(2s + τ ) 1 . (s + 0.2K)(2s + τ ) + 4 s From the final value theorem we obtain yss = lim sY (s) = s→0 2τ . 0.2Kτ + 4 Selecting K = 10− 20/τ leads to yss = 1 and a zero steady-state error due to a unit step input. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 205 Problems (b) The characteristic equation is (s + 0.2K)(2s + τ ) + 4 = 2s2 + (0.4K + τ )s + 0.2Kτ + 4 = 0 . So, with K = 10 − 20/τ , the natural frequency and damping ratio are: ωn = √ τ and ζ = τ 2 + 4τ − 8 . 4τ 3/2 The settling time and percent overshoot are found using the standard design formulas √ π −ζπ 1−ζ 2 p Tp = and P.O. = 100e ωn 1 − ζ 2 with ωn and ζ given above (as a function of τ ). Since the closed-loop system has a zero at s = −τ /2, the formulas for Tp and P.O. will only be approximate. Also, note that for the closed-loop system poles to be in the left half-plane (that √ is, all the poles have negative real parts), we require that τ > 2 3 − 2 ≈ 1.4642. As seen √ in the next chapter, this is the condition for stability. Having τ > 2 3 − 2 insures that the damping ratio ζ is positive. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 206 CHAPTER 5 The Performance of Feedback Control Systems Advanced Problems (a) The steady-state error is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − s→0 108(3) =0. 9(36) (b) Assume the complex poles are dominant. Then, we compute a = 0.75 , ζωn since a = 3, ζ = 0.67 and ωn = 6. Using Figure 5.13 in Dorf & Bishop, we estimate the settling time and percent overshoot to be P.O. = 45% and Ts = 4 = 1 second . ζωn (c) The step response is shown in Figure AP5.1. The actual settling time and percent overshoot are P.O. = 34.4% and Ts = 1.18 second . Step Response 1.4 1.2 1 Amplitude AP5.1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 0.6 FIGURE AP5.1 Closed-loop system step response. 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 207 Advanced Problems The closed-loop transfer function is T (s) = s3 + 28s2 5440(τz s + 1) . + (432 + 5440τz )s + 5440 The closed-loop step responses are shown in Figure AP5.2. The performance results are summarized in Table AP5.2. tau=0 (solid) & tau=0.05 (dashed) & tau=0.1 (dotted) & tau=0.5 (dot-dash) 1.4 1.2 1 0.8 y(t) AP5.2 0.6 0.4 0.2 0 0 0.2 0.4 0.6 0.8 Time (sec) 1 1.2 1.4 1.6 FIGURE AP5.2 Closed-loop system step responses. τz Tr Ts P.O. 0 0.16 0.89 32.7% p = −20, −4 ± 16j 0.05 0.14 0.39 4.5% p = −10.4, −8.77 ± 21.06j 0.1 0.10 0.49 0% p = −6.5, −10.74 ± 26.84j 0.5 0.04 1.05 29.2% p = −1.75, −13.12 ± 54.16j TABLE AP5.2 closed-loop poles Performance summary. As τz increases from 0 to 0.1, the P.O. decreases and the response is faster © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 208 CHAPTER 5 The Performance of Feedback Control Systems and more stable. However, as τz is increased beyond 0.1, the P.O. and Ts increase, although Tr continues to decrease. The closed-loop transfer function is T (s) = τp s3 1 . + (1 + 2τp )s2 + 2s + 1 The closed-loop step responses for τp = 0, 0.5, 2, 5 are shown in Figure AP5.3. The performance results are summarized in Table AP5.3. tau=5 (solid) & tau=2 (dotted) & tau=0.5 (dashed) & tau=0 (dot-dash) 1.5 1 y(t) AP5.3 0.5 0 0 5 10 15 20 25 Time (sec) 30 35 40 45 FIGURE AP5.3 Closed-loop system step responses. τp Tr Ts P.O. 0 4 5.8 0% 0.5 3.6 7.4 4.75% p = −2.84, −0.58 ± 0.6j 2 4.6 22.4 27.7% p = −2.14, −0.18 ± 0.45j 5 6 45.8 46% p = −2.05, −0.07 ± 0.3j TABLE AP5.3 Performance summary. closed-loop poles p = −1, −1 50 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 209 Advanced Problems As τp increases, the P.O., Tr and Ts also increase; adding the pole makes the system less stable with more overshoot. The system transfer function is Y (s) = 15 15K R(s) + Td (s) . (s + 5)(s + 7) + 15K (s + 5)(s + 7) + 15K When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The closedloop step input and disturbance responses for K = 1, 10, 100 are shown in Figure AP5.4. The performance results are summarized in Table AP5.4. Unit step input response Unit step distrubance response 1.6 0.35 1.4 0.3 1.2 0.25 1 y(t) 0.2 y(t) AP5.4 0.8 0.15 0.6 0.1 0.4 0.05 0.2 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 0 0 0.2 0.4 0.6 Time (sec) 0.8 1 FIGURE AP5.4 Closed-loop system input and disturbance responses (K =1: solid line, K =10: dotted line, and K =100:dashed line). TABLE AP5.4 K ess Ts P.O. |y/d|max 1 0.7 0.45 0% 0.3 10 0.19 0.6 17.3% 0.1 100 0.023 0.59 60.0% 0.01 Performance summary. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 210 CHAPTER 5 The Performance of Feedback Control Systems The best value of the gain is K = 10, which is compromise between (i) percent overshoot, and (ii) disturbance rejection and tracking error. The system transfer function is 50(s + α)(s + 2) R(s) s(s + 3)(s + 4) + 50(s + α)(s + 2) 50s(s + 2) Td (s) . + s(s + 3)(s + 4) + 50(s + α)(s + 2) Y (s) = Disturbance response: alpha=0 (solid) & alpha=10 (dashed) & alpha=100 (dotted) 10 9 8 7 6 5 y(t) AP5.5 4 3 2 1 0 -1 0 0.05 0.1 0.15 0.2 0.25 Time (sec) FIGURE AP5.5 Closed-loop system disturbance response. When considering the input response, we set Td (s) = 0, and similarly, when considering the disturbance response, we set R(s) = 0. The steadystate tracking error is ess = lim s(1 − T (s))R(s) = lim 1 − s→0 s→0 50(s + α)(s + 2) . s(s + 3)(s + 4) + 50(s + α)(s + 2) When α = 0, we have ess = 1 − 100 = 0.11 , 100 + 12 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 211 Advanced Problems and, for α 6= 0 ess = 0 . The closed-loop step input and disturbance responses for α = 0, 10, 100 are shown in Figure AP5.5. For disturbance rejection and steady-state tracking error the best value of the parameter is α = 100 . However, when considering both the disturbance and input response we would select the parameter α = 10 , since it offers a good compromise between input response overshoot (about 5% for α = 10) and disturbance rejection/tracking error. AP5.6 (a) The closed-loop transfer function is T (s) = KKm . KKm + s(s + Km Kb + 0.01) The steady-state tracking error for a ramp input R(s) = 1/s2 is ess = lim s(1 − T (s))R(s) s→0 s + Km Kb + 0.01 s→0 KKm + s(s + Km Kb + 0.01) Km Kb + 0.01 = . KKm = lim (b) With Km = 10 and Kb = 0.05 , we have Km Kb + 0.01 10(0.05) + 0.01 = =1. KKm 10K Solving for K yields K = 0.051 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 212 CHAPTER 5 The Performance of Feedback Control Systems (c) The plot of the step and ramp responses are shown in Figure AP5.6. The responses are acceptable. Step input response 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 Time (sec) 12 14 16 18 20 14 16 18 20 Ramp input response 20 y(t) 15 10 5 0 0 2 4 6 8 10 Time (sec) 12 FIGURE AP5.6 Closed-loop system step and ramp responses. AP5.7 The performance is summarized in Table AP5.7 and shown in graphical form in Fig. AP5.7. K Estimated Percent Overshoot Actual Percent Overshoot 1000 8.8 % 8.5 % 2000 32.1 % 30.2 % 3000 50.0 % 46.6 % 4000 64.4 % 59.4 % 5000 76.4 % 69.9 % TABLE AP5.7 Performance summary. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 213 Advanced Problems 80 Percent Overshoot (% ) 70 Actual P.O . Estimated P.O. 60 50 40 30 20 10 0 1000 2000 3000 K 4000 5000 FIGURE AP5.7 Percent overshoot versus K. The closed-loop transfer function is T (s) = 100K . s(s + 50)(s + 100) + 100K The impact of the third pole is more evident as K gets larger as the estimated and actual percent overshoot deviate in the range 0.3% at K = 1000 to 6.5% at K = 5000. AP5.8 The closed-loop transfer function is T (s) = K(s + 2) . s2 + ( 23 + K)s + 13 + 2K Comparing T (s) to a second-order system we have ωn = q 1/3 + 2K 2/3 + K ζ= p 2 1/3 + 2K For the closed-loop transfer function to have complex roots, we require K 2 − (20/3)K − (8/9) < 0. This occurs when −0.13 ≤ K ≤ 6.8. When K = 1/3, we have the minimum ζ = 0.5, as shown in Figure AP5.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 214 CHAPTER 5 The Performance of Feedback Control Systems 0.58 0.57 0.56 ζ 0.55 0.54 0.53 0.52 0.51 0.5 0 0.05 0.1 0.15 0.2 0.25 K 0.3 0.35 0.4 0.45 0.5 FIGURE AP5.8 Damping ratio, ζ, versus K. AP5.9 The closed-loop characteristic equation is s4 + 40s3 + 375s2 + KP s + KI = 0. The desired characteristic equation is √ (s + a)(s + b)(s2 + 2ωn s + ωn2 ) = 0. Expanding the desired characteristic equation and equating terms to the actual characteristic equation yields √ abωn2 = KI , ωn2 (a + b) + 2abωn = KP √ 2(a + b)ωn + ab = 375, √ 2ωn + a + b = 40 This represents 4 equations with 5 unknowns (a, b, KP , KI , and ωn ). We can choose one variable as part of the controller design. Let KI = 0.1KP . Then, solving the 4 equations for the remaining 4 variables yields a = 29.15, b = 0.1, KP = 1720, KI = 172, and ωn = 7.6. The resulting Ts = 1.1s and P.O. = 6.4%, as shown in Figure AP5.9. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 215 Advanced Problems Step Response 1.4 System: sysa Peak amplitude: 1.06 Overshoot (%): 6.39 At time (sec): 0.618 1.2 Amplitude 1 System: sysa Settling Time (sec): 1.09 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec) FIGURE AP5.9 Step response for KP = 1720 and KI = 172. 6 7 8 9 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 216 CHAPTER 5 The Performance of Feedback Control Systems Design Problems The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: 26.035 θ(s) = , Va (s) s(s + 33.142) where we neglect the motor inductance Lm . The closed-loop transfer function from the disturbance to the output is θ(s) 26.035 = 2 . Td (s) s + 33.142s + 26.035Ka For a unit step disturbance input the steady-state response is θss = 1 . Ka Therefore, we want to use the maximum Ka while keeping the percent overshoot less than 5%. The step response for the closed-loop system (with the tachometer not in the loop) and Ka = 22 is shown below. Values of Ka greater than 22 lead to overshoots greater than 5%. Step response 1.4 1.2 q(t)/A 1 0.8 0.6 0.4 0.2 0 0 0.1 0.2 0.3 0.4 0.5 Time (sec) 0.6 0.7 0.8 0.9 1 0.7 0.8 0.9 1 Unit disturbance response 0.05 0.04 0.03 q(t) CDP5.1 0.02 0.01 0 0 0.1 0.2 0.3 0.4 0.5 Time (sec) 0.6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 217 Design Problems DP5.1 (a) The closed-loop transfer function is 12.2K 12.2K φ(s) = = 3 . 2 φd (s) s(s + 2.2)(s + 7) + 12.2K s + 9.2s + 15.4s + 12.2K (b) For K = 0.7, we have the characteristic equation s3 + 9.2s2 + 15.4s + 8.54 = 0 , with roots s1 = −7.23 and s2,3 = −0.98 ± 0.46j. For K = 3, we have the characteristic equation s3 + 9.2s2 + 15.4s + 36.6 = 0 , with roots s1 = −7.83 and s2,3 = −0.68 ± 2.05j. And for K = 6, we have the characteristic equation s3 + 9.2s2 + 15.4s + 73.2 = 0 , with roots s1 = −8.4 and s2,3 = −0.4 ± 2.9j. (c) Assuming the complex conjugate pair are the dominant roots, we expect the following: (i) for K = 0.7: P.O.=0.13% and Tp = 6.8 sec (ii) for K = 3: P.O.=35.0% and Tp = 1.5 sec (iii) for K = 6: P.O.=65.2% and Tp = 1.1 sec (d),(e) We select K = 1.71 to have a P.O. = 16% and Tp = 2.18sec. All four cases (K = 0.7, 3, 6, 1.71) are shown in Figure DP5.1. In each case, the approximate transfer function is derived by neglecting the non-dominant real pole and adjusting the gain for zero steady-state error. The approximate transfer functions are 1.18 0.7908 = + 1.965s + 1.18 (s + 0.98 + 0.46j)(s + 0.98 − 0.46j) 4.67 3.299 TK=3 (s) = 2 = s + 1.37s + 4.67 (s + 0.68 + 2.05j)(s + 0.68 − 2.05j) 8.71 6.399 TK=6 (s) = 2 = s + 0.796s + 8.71 (s + 0.4 + 2.9j)(s + 0.4 − 2.9j) 2.77 1.458 TK=1.71 (s) = 2 = s + 1.679s + 2.77 (s + 0.83 + 1.43j)(s + 0.83 − 1.43j) TK=0.7 (s) = s2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 218 CHAPTER 5 The Performance of Feedback Control Systems K=0.7 K=3 1 1 phi 1.5 phi 1.5 0.5 0 0.5 0 5 time (sec) K=6 0 10 1.5 10 0 5 time (sec) 10 1 phi phi 5 time (sec) K=1.71 1.5 2 1 0.5 0.5 0 0 0 5 time (sec) 10 0 FIGURE DP5.1 Step responses (actual response:solid lines; approximate response: dotted lines). DP5.2 The closed-loop transfer function is T (s) = Kωn2 , s3 + 2ζωn s2 + ωn2 s + Kωn2 where ζ = 0.6. From the second-order system approximation, we have Tp = ωn π . 1 − ζ2 p So, with ζ = 0.6 given, we should select ωn “large” to make Tp “small.” Also, from the problem hint, let 0.2 < K/ωn < 0.4 . As a first attempt, we can select ωn = 10. See Figure DP5.8 for various values of K/ωn . Our final selection is K = 3.33 and ωn = 10 . This results in P.O. = 3.6% and Tp = 0.66 second. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 219 Design Problems 1.4 1.2 K/ωn=0.4 1 K/ωn=0.33 y(t) 0.8 K/ω =0.2 n 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 2.5 3 FIGURE DP5.2 Closed-loop system response. DP5.3 The closed-loop transfer function is T (s) = s2 K . + qs + K From the ITAE specification, we desire T (s) = ωn2 . s2 + 1.4ωn s + ωn2 But 2ζωn = 1.4ωn which implies ζ = 0.7 . Since we want Ts ≤ 0.5, we require ζωn ≥ 8. So, ωn ≥ 8 = 11.4 . 0.7 We can select ωn = 12. Then, T (s) = 144 . s2 + 16.8s + 144 Therefore, K = 144 and q = 16.8. The predicted percent overshoot is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 220 CHAPTER 5 The Performance of Feedback Control Systems P.O. = 4.5%. DP5.4 The loop transfer function is Gc (s)G(s) = 10K 10K/70 = . (s + 70)(s + 3)(s + 7) (s/70 + 1)(s + 3)(s + 7) The second-order approximation is obtained by neglecting the fastest firstorder pole. Thus, K/7 . (s + 3)(s + 7) Gc (s)G(s) ≈ The closed-loop transfer function is T (s) = s2 K/7 . + 10s + 21 + K/7 When ζ ≥ 0.52, we have less than 15% overshoot. So, we have 2ζωn = 10 and ωn = q 21 + K/7. Eliminating ωn and solving for K (with P.O. ≤ 15%) yields K ≤ 500.19 . Also, Kp = lim GGc (s) = s→0 K 7(21) and ess = 1 1 = < 0.12 K 1 + Kp 1 + 147 implies K ≥ 1078 . Therefore, we have an inconsistency. We require 1078 ≤ K to meet the steady-state requirement and K ≤ 500.18 to meet the percent overshoot requirement. It is not possible to meet both specifications. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 221 Design Problems DP5.5 The closed-loop characteristic equation is 1 + K1 G1 (s) + K2 G1 G2 (s) = 1 + 2K2 K1 − =0 s(s + 1) s(s + 1)(s + 2) or s3 + 3s2 + (2 + K1 )s + 2(K1 − K2 ) = 0 . Assuming that K1 > 0 and K2 > 0, the range of the gains for stability is 0 < K2 < K1 . DP5.6 The closed-loop transfer function is T (s) = s2 K1 . + (K1 K2 + 1)s + K1 The percent overshoot specification P.O. ≤ 2% is satisfied when ζ > 0.78. The peak time specification Tp ≤ 0.5 s is satisfied when ωn = 10 and ζ = 0.78. So, given K1 = ωn2 and K1 K2 + 1 = 2ζωn , we determine that the specifications are satisfied when K1 = 100 and K2 = 0.15 . DP5.7 The plant is G(s) = 2 s(s + 1)(s + 4) and the PD controller is Gc (s) = KD s + KP . The characteristic equation is s3 + 6s2 + (8 + 2KD )s + 2KP = 0. The desired characteristic equation is (s + a)(s2 + 2ζωn s + ωn2 ) = s3 + (2ζωn + a)s2 + (ωn2 + 2ζωn a)s + aωn2 = 0. Equating the desired characteristic equation to the actual characteristic © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 222 CHAPTER 5 The Performance of Feedback Control Systems equation yields 2ζωn + a = 6, ωn2 + 2ζωn a = 8 + 2KD , aωn2 = 2KP , where ζ = 0.69 and ωn = 3 to meet the design specifications. This represents 3 equations in 3 unknowns (a, KD , and KP ). Solving yields a = 1.86, KD = 4.35 and KP = 8.37. The step response is shown in Figure DP5.7. Step Response 1.4 System: sys_cl Peak amplitude: 1.04 Overshoot (%): 4.28 At time (sec): 1.48 1.2 Amplitude 1 System: sys_cl Settling Time (sec): 1.96 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Time (sec) 2 2.5 3 FIGURE DP5.7 Step response withKD = 4.35 and KP = 8.37. DP5.8 The closed-loop transfer function is T (s) = s2 K + 6s + 5 + K The damping ratio and natural frequency is ζ=√ 3 K +5 and ωn = √ K +5 Using the design formulas for second-order systems we have √ 4 2 P O = 100e−ζπ/ 1−ζ and Ts = . ζωn We know that the formula for Ts is approximate and that the formulas apply only to systems with ζ < 1. For K = 1 the closed-loop poles are © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 223 Design Problems both real, so there is no overshoot and the design formula for settling time does not apply. Thus we obtain the results shown in Table DP5.8. We can choose K = 10 as a good trade-off between percent overshoot, settling time, and steady-state tracking error. The disturbance response is shown in Figure DP5.8. TABLE DP5.8 Step response for K=1, 10, and 20. K % P.O. Ts , sec Ts , sec Estimated Actual ess 1 0 - 3.24 0.83 10 2.13 1.33 1.38 0.33 20 9.48 1.33 1.19 0.20 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 FIGURE DP5.8 Closed-loop system disturbance response for K = 10. 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 224 CHAPTER 5 The Performance of Feedback Control Systems Computer Problems With the impulse input we have R(s) = 1. The transfer function is Y (s) = 15 15 R(s) = . (s + 3)(s + 5) (s + 3)(s + 5) Therefore, taking the inverse Laplace transforms yields the output response: y(t) = 15 −3t 15 −5t e − e . 2 2 The impulse response and the analytic response is shown in Figure CP5.1. n=15; d=[1 8 15]; t=[0:0.1:6]; ya=(15/2)*exp(-3.*t)-(15/2)*exp(-5.*t); sys = tf(n,d) y=impulse(sys,t); plot(t,y,t,ya,'o') xlabel('Time (sec)'), ylabel('y(t)'), legend('Computer','Analytic',-1) 1.4 Computer Analytic 1.2 1 0.8 y(t) CP5.1 0.6 0.4 0.2 0 0 FIGURE CP5.1 Impulse responses. 1 2 3 Time (sec) 4 5 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 225 Computer Problems CP5.2 The ramp response is shown in Figure CP5.2. The unity feedback system is type 2, so that the steady-state tracking error is lim ess → 0. t→∞ 60 50 Amplitude 40 n=[1 10]; d=[1 15 0 0]; t=[0:0.1:50]; sys= tf(n,d); sys_cl = feedback(sys,[1]); u=t; lsim(sys_cl,u,t); 30 20 10 0 0 5 10 15 20 25 Time (sec) 30 35 40 45 50 FIGURE CP5.2 Ramp responses. CP5.3 The m-file script and the four plots are shown in Figure CP5.3. The plots can be compared to Figure 5.17 in Dorf & Bishop. wn=2, zeta=0 2 1 1 0 0 -1 -1 -2 0 5 10 15 20 wn=1, zeta=0 1 wn=2, zeta=0.1 2 -2 0 5 15 20 wn=1, zeta=0.2 1 0.5 10 0.5 0 0 -0.5 -1 0 FIGURE CP5.3 Impulse responses. 5 10 15 20 -0.5 0 5 10 15 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 226 CHAPTER 5 The Performance of Feedback Control Systems w1=2; z1=0; w2=2; z2=0.1; w3=1; z3=0; w4=1; z4=0.2; t=[0:0.1:20]; % num1=[w1^2]; den1=[1 2*z1*w1 w1^2]; sys1 = tf(num1,den1); [y1,x1]=impulse(sys1,t); % num2=[w2^2]; den2=[1 2*z2*w2 w2^2]; sys2 = tf(num2,den2); [y2,x2]=impulse(sys2,t); % num3=[w3^2]; den3=[1 2*z3*w3 w3^2]; sys3 = tf(num3,den3); [y3,x3]=impulse(sys3,t); % num4=[w4^2]; den4=[1 2*z4*w4 w4^2]; sys4 = tf(num4,den4); [y4,x4]=impulse(sys4,t); % clf subplot(221),plot(t,y1),title('wn=2, zeta=0') subplot(222),plot(t,y2),title('wn=2, zeta=0.1') subplot(223),plot(t,y3),title('wn=1, zeta=0') subplot(224),plot(t,y4),title('wn=1, zeta=0.2') FIGURE CP5.3 CONTINUED: Impulse response m-file script. CP5.4 The closed-loop system is T (s) = s2 21 . + 2s + 21 Therefore, the natural frequency is √ ωm = 21 = 4.58 and the damping ratio is computed as 2ζωn = 2 , which implies ζ = 0.218 . The percent overshoot is estimated to be √ 2 P.O. = 100e−ζπ/ 1−ζ = 50% , since ζ = 0.218. The actual overshoot is shown in Figure CP5.4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 227 Computer Problems Step Response From: U(1) 1.5 To: Y(1) 1 Amplitude numc=[21]; denc=[1 0]; sysc = tf(numc,denc); numg=[1]; deng=[1 2]; sysg = tf(numg,deng); sys_o = series(sysc,sysg); sys_cl = feedback(sys_o,[1]) step(sys_cl) 0.5 0 0 1 2 3 4 5 6 Time (sec.) FIGURE CP5.4 Impulse responses. CP5.5 For a step input, the optimum ITAE characteristic equation is s3 + 1.75ωs2 + 2.15ω 2 s + ω 3 = 0 . Examining Figure 5.30 for n=3 in Dorf & Bishop, we estimate that ωTs = 8. So, once we decide on the desired Ts we can estimate ω. For this design we let Ts =8 so that ω = 1. Computing the desired characteristic equation and the actual characteristic equation and comparing the coefficients leads to the following relationships: p = 1.75ω − 2ζωn K = (2.15ω 2 − ωn2 − 2ζωn p)/ωn2 z = (ω 3 − pωn2 )/(Kωn2 ) where ζ = 0.59 and ωn = 0.45. The controller and prefilter are Gc (s) = 6.42 s + 0.58 s + 1.22 and Gp (s) = The unit step response is shown in Figure CP5.5. 1 . 1.3s + 0.75 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 228 CHAPTER 5 The Performance of Feedback Control Systems Step Response 1.4 wn=0.45; zeta=0.59 ng=wn^2; dg=[1 2*zeta*wn wn^2]; sysg=tf(ng,dg); Ts=8; w=8/Ts; p=1.75*w-2*zeta*wn; K=(2.15*w^2-wn^2-2*zeta*wn*p)/wn^2; z=(w^3-p*wn^2)/(K*wn^2); nc=K*[1 z]; dc=[1 p]; sysc=tf(nc,dc); sys=series(sysc,sysg); syscl=feedback(sys,1); [num,den]=tfdata(syscl,'v'); sysp=tf([den(end)],num); step(syscl*sysp) System: untitled1 Peak amplitude: 1.02 Overshoot (%): 1.98 At time (sec): 4.68 1.2 1 Amplitude System: untitled1 Settling Time (sec): 7.54 0.8 0.6 0.4 0.2 0 0 2 4 6 Time (sec) 8 10 12 FIGURE CP5.5 Closed-loop system step response m-file script. CP5.6 The unit step response is shown in Figure CP5.6. The performance numbers are as follows: Mp = 1.16, Tp = 0.73, and Ts = 1.62. Step Response 1.4 System: sys_cl Peak amplitude: 1.16 Overshoot (%): 16.3 At time (sec): 0.73 1.2 1 System: sys_cl Settling Time (sec): 1.62 Amplitude numg=[25]; deng=[1 5 0]; sys = tf(numg,deng); sys_cl = feedback(sys,[1]); t=[0:0.01:2]; step(sys_cl,t); 0.8 0.6 0.4 0.2 0 0 FIGURE CP5.6 Closed-loop system step response m-file script. 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 229 Computer Problems The m-file script and the simulations are shown in Figure CP5.7. % Part (a) numc=[2]; denc=[1]; sys_c = tf(numc,denc); nums=[-10]; dens=[1 10]; sys_s = tf(nums,dens); numg=[-1 -5]; deng=[1 3.5 6 0]; sys_g = tf(numg,deng); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); f=0.5*pi/180; % Convert to rad/sec t=[0:0.1:10]; u=f*t; [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(211) plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('Constant gain C(s) = 2: theta (solid) & input (dashed)') % Part (b) numc=[2 1]; denc=[1 0]; sys_c = tf(numc,denc); [numa,dena]=series(numc,denc,nums,dens); sysa = series(sys_c,sys_s); sysb = series(sysa,sys_g); sys = feedback(sysb,[1]); [y,x]=lsim(sys,u,t);(y(length(t),1)-u(1,length(t)))*180/pi subplot(212), plot(t,y*180/pi,t,u*180/pi,'--'), grid xlabel('Time (sec)'),ylabel('theta') title('PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed)') Constant gain C(s) = 2: theta (solid) & input (dashed) 5 theta 4 3 2 1 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 10 8 9 10 PI controller C(s) = 2 + 1/s: theta (solid) & input (dashed) 6 5 4 theta CP5.7 3 2 1 0 0 1 2 3 4 5 Time (sec) 6 7 FIGURE CP5.7 Closed-loop system response to a ramp input for two controllers. For the constant gain controller, the attitude error after 10 seconds is ess = −0.3 deg. On the other hand, the PI controller has a zero steadystate error ess = 0 deg. So, we can decrease the steady-state error by © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 230 CHAPTER 5 The Performance of Feedback Control Systems using a more sophisticated controller, in this case a PI controller versus a constant gain controller. CP5.8 The closed-loop characteristic equation is s3 + 12s2 + 610s + 500 = (s + 0.8324)(s2 + 11.1676s + 600.7027) = 0 . The natural frequency and damping ratio of the complex roots are ωn = 24.5 and ζ = 0.23. From this we predict Mp = 1.48, Ts = 0.72, and Tp = 0.13. The actual response is shown in Figure CP5.8. The differences Step Response From: U(1) 1.4 1.2 1 0.8 To: Y(1) theta dot numg=[100 100]; deng=[1 2 100]; sysg = tf(numg,deng); numc=[0.1 5]; denc=[1 0]; sysc = tf(numc,denc); sys_o = series(sysg,sysc); sys_cl = feedback(sys_o,[1]) t=[0:0.01:3]; step(sys_cl,t); ylabel('theta dot') 0.6 0.4 0.2 0 0 0.5 1 1.5 2 2.5 3 Time (sec.) FIGURE CP5.8 Missile rate loop autopilot simulation. can be explained by realizing that the system is not a second-order system. The closed-loop system actually has two zeros, one real pole, and two complex-conjugate poles: T (s) = (s + 50)(s + 1) . (s + 0.8324)(s2 + 11.1676s + 600.7027) The effect of the pole at s = −0.8324 is diminished by the zero at s = −1. The third pole and the zeros affect the overall response such that the analytic formulas for second-order systems are not exact predictors of the transient response. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 231 Computer Problems CP5.9 Figure CP5.9 shows an m-file to compute the closed-loop transfer function and to simulate and plot the step response. Step Response 1 System: sys Peak amplitude: 0.979 Overshoot (%): 95.7 At time (sec): 0.533 0.9 0.8 numg=[10]; deng=[1 10]; sysg = tf(numg,deng); numh=[0.5]; denh=[10 0.5]; sysh = tf(numh,denh); sys = feedback(sysg,sysh) step(sys); Amplitude 0.7 Transfer function: 100 s + 5 --------------------10 s^2 + 100.5 s + 10 0.6 System: sys Settling Time (sec): 39.1 0.5 0.4 0.3 0.2 0.1 0 0 10 20 30 Time (sec) 40 50 60 FIGURE CP5.9 M-file to compute the transfer function and to simulate the step response. CP5.10 Figure CP5.10 shows an m-file to compute the closed-loop transfer function and to simulate and plot the ramp response. The steady-state error Linear Simulation Results 100 90 80 70 Amplitude numg=[10]; deng=[1 20 75 0]; sysg = tf(numg,deng); sys = feedback(sysg,1) t=[0:0.1:100]; u=t; % Unit ramp input lsim(sys,u,t); 60 50 40 30 20 10 0 0 10 20 30 40 50 Time (sec) FIGURE CP5.10 M-file to compute the transfer function and to simulate the ramp response. is 7.5. 60 70 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 232 CHAPTER 5 CP5.11 The Performance of Feedback Control Systems Figure CP5.11 shows an m-file to compute the closed-loop transfer function and to simulate and plot the impulse, step, and ramp responses. Notice that the closed-loop system is unstable. 0 Amplitude -10 0 2 4 6 0 2 4 6 0 2 4 6 8 10 Time (sec) Step Response 12 14 16 18 20 8 10 12 Time (sec) Linear Simulation Results 14 16 18 20 14 16 18 20 10 0 -10 Amplitude numg=[1]; deng=[1 2 0]; sysg = tf(numg,deng); numc=[0.5 2]; denc=[1 0]; sysc = tf(numc,denc); syss=series(sysg,sysc); sys = feedback(syss,1) t=[0:0.1:20]; subplot(311) impulse(sys,t); subplot(312) step(sys,t); subplot(313) u=t; % Unit ramp input lsim(sys,u,t); Amplitude Impulse Response 10 40 20 0 8 10 Time (sec) 12 FIGURE CP5.11 M-file to compute the transfer function and to simulate the ramp response. CP5.12 Figure CP5.12 shows an m-file to simulate and plot the step response for the original system and the 2nd-order approximation. For the original system, we find Ts = 2.28 and P.O. = 80.6%. For the 2nd-order approximation we find Ts = 2.16 and P.O. = 101% © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 233 Computer Problems 2.5 2nd order approximation 2 3rd order system response Step response num=77*[1 2]; den=conv([1 7],[1 4 22]); sys = tf(num,den) na=(77/7)*[1 2]; da=[1 4 22]; sysa=tf(na,da); t=[0:0.01:5]; y=step(sys,t); ya=step(sysa,t); plot(t,y,t,ya,'--') xlabel('Time (s)'), ylabel('Step response') 1.5 1 0.5 0 0 1 2 3 Time (s) FIGURE CP5.12 Step response. 4 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 6 The Stability of Linear Feedback Systems Exercises E6.1 The Routh array is s3 1 K +1 s2 K 6 s1 b 0 so 6 where b= K(K + 1) − 6 . K For stability, we require K > 0 and b > 0. Therefore, using the condition that b > 0, we obtain K2 + K − 6 > 0 , and solving for K yields K > 2 and K < −3. We select K > 2, since we also have the condition that K > 0. E6.2 The Routh array is s3 1 2 s2 10 30 s1 -1 0 so 30 The system is unstable since the first column shows two sign changes. 234 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 235 Exercises E6.3 The Routh array is s4 1 32 s3 10 37 s2 28.3 20 s1 29.9 s0 20 20 By the Routh-Hurwitz criterion, the system is stable (i.e., all the numbers in the first column are positive). E6.4 The closed-loop transfer function is T (s) = s3 + −K(s − 2) . + (4 − K)s + 2K 5s2 Therefore, the characteristic equation is s3 + 5s2 + (4 − K)s + 2K = 0 . The corresponding Routh array is given by s3 1 s2 4−K 5 2K s1 b 0 so 2K where b= 5(4 − K) − 2K 20 − 7K = . 5 5 For stability we require K > 0 and b > 0 . Thus, the range of K for stability is 0 < K < 20/7. E6.5 The closed-loop transfer function is T (s) = s3 + 10s2 K . + 27s + 18 + K When K = 20, the roots of the characteristic polynomial are s1,2 = −1.56 ± j1.76 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 236 CHAPTER 6 The Stability of Linear Feedback Systems and s3 = −6.88 . E6.6 When K = 252, the roots of the characteristic equation are on the imaginary axis. The roots are s1,2 = ±j5.2 E6.7 and s3 = −10 . (a) The closed-loop system characteristic equation is 1 + GH(s) = 1 + or K(s + 2) =0, s(s − 1) s2 + (K − 1)s + 2K = 0 . √ We have the relationships ωn = 2K and 2ζωn = K − 1, where ζ = 0.707. Thus,   1 √ 2 √ 2K = K − 1 , 2 or  2 √ 2 2 =  K −1 √ 2K 2 , and K 2 − 6K + 1 = 0 . Solving for K yields K = 5.83 and K = 0.17. However, for stability we require K > 1 (from the Routh array), so we select K = 5.83. √ (b) The two roots on the imaginary axis when K = 1 are s1,2 = ±j 2. E6.8 The closed-loop system characteristic equation is 3 + 20s2 + (100 + K)s + 20K = 0 . The corresponding Routh array is s3 1 (100 + K) s2 20 20K s1 b 0 so 20K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 237 Exercises where b= 20(100 + K) − 20K 20(100) = = 100 . 20 20 Therefore, the system is stable for all K > 0. E6.9 The characteristic equation is s3 + 2s2 + (K + 1)s + 8 = 0 , and the Routh array is given by s3 1 K +1 s2 2 8 s1 b 0 so 8 where b= 2(K + 1) − 8 =K −3 . 2 Setting b = 0, yields K − 3 = 0 or K>3. E6.10 Stable with your eyes open and (generally) unstable with your eyes closed. E6.11 The system is unstable. The poles are s1 = −5.66, s2 = −0.90 and s3,4 = 0.28 ± j0.714. E6.12 The characteristic equation is s2 + as + b = 0, so, the Routh array is s2 1 b s1 a 0 so b The system is stable if and only if a > 0 and b > 0. For a second-order system, a necessary and sufficient condition for stability is that all the coefficients have the same sign. E6.13 The characteristic equation is s2 + (KD + 2)s + 4KP = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 238 CHAPTER 6 The Stability of Linear Feedback Systems The Routh array is s2 1 4KP s1 KD + 2 0 so 4KP The system is stable if and only if KP > 0 and KD > −2. E6.14 The characteristic equation associated with the system matrix is s3 + 3s2 + 5s + 6 = 0 . The roots of the characteristic equation are s1 = −2 and s2,3 = −5±j1.66. The system is stable. E6.15 The roots of q(s) are s1 = −4, s2 = −3, s3,4 = −1 ± j2 and s5,6 = ±j0.5. The system is marginally stable. The Routh array is s6 1 31.25 67.75 s5 9 61.25 14.75 s4 24.44 66.11 15 s3 31.909 9.2273 0 s2 60 15 s1 0 0 15 so The auxillary equation is 60s2 + 15 = 0 . Solving the auxillary equation yields two roots at s1,2 = ±j0.5. After accounting for the row of zeros, the completed Routh array verifies that the system has no poles in the right half-plane. E6.16 The Routh array is s4 1 45 s3 9 87 s2 35.33 50 s1 74.26 0 so 50 50 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 239 Exercises The system is stable. The roots of q(s) are s1,2 = −3 ± j4, s3 = −2 and s4 = −1. E6.17 The characteristic equation is s3 + 7s2 + 36s + 24 = 0 . The system is stable. The roots of the characteristic equation are s1 = −0.77, s2,3 = −3.12 ± 4.64i. E6.18 The roots of q(s) are s1 = −20 and s2,3 = ±j2.24. The system is marginally stable. The Routh array is s3 1 5 s2 20 100 s1 0 0 so The auxillary equation is 20s2 + 100 = 0 . The roots are s = ±j2.24. So, the system has roots at s = ±j2.24. Completing the Routh array (after accounting for the row of zeros) verifies that no poles lie in the right half-plane. E6.19 (a) Unstable. (b) Stable. (c) Stable. E6.20 (a) The roots are s1,2 = −2 and s3 = −1. (b) The roots are s1,2,3 = −3. E6.21 The characteristic equation is (sn − 2)3 + 10(sn − 2)2 + 29(sn − 2) + K = 0 or s3n + 4s2n + sn − 26 + K = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 240 CHAPTER 6 The Stability of Linear Feedback Systems s3 1 1 s2 4 s1 30−K 4 K − 26 so K − 26 0 If K = 30, then the auxillary equation is 4s2n + 4 = 0 or sn = ±j. Therefore, s = sn − 2 implies s = −2 ± j. E6.22 This system is not stable. The output response to a step input is a ramp y(t) = kt. E6.23 The characteristic polynomial is s3 + 4s2 + ks + 8 = 0 . The Routh array is s3 1 k s2 4 8 s1 4k−8 4 so 8 So, k > 2 for stability. E6.24 The transfer function is G(s) = C(sI − A)−1 B + D   k 0 k s+k  2  s + ks + k  = [ 1 0 0 ] −k      0      −1   0   s −1  = [ 1 0 0 ]  0 s −ks  1 s+k  s2 + ks   1  0      s   0  −ks − k s2  1 1  ∆(s) where ∆(s) = s3 + ks2 + ks + k. Thus, the transfer function is G(s) = The Routh array is s3 + ks2 1 . + ks + k © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 241 Exercises s3 1 k s2 k k s1 k−1 so k For stability k > 1. E6.25 The closed-loop transfer function is T (s) = Ks + 1 . s2 (s + p) + Ks + 1 Therefore, the characteristic equation is s3 + ps2 + Ks + 1 = 0 . The Routh array is s3 1 K s2 p 1 s1 (pK − 1)/p so 1 We see that the system is stable for any value of p > 0 and pK − 1 > 0. E6.26 The closed-loop transfer function is T (s) = 10 . 2s2 + (K − 20)s + 10 − 10K Therefore, the characteristic equation is 2s2 + (K − 20)s + 10 − 10K = 0 . The Routh array is s2 2 s1 K − 20 so 10-10K 10 − 10K We see that the system is stable for any value of K > 20 and any K < 1. Therefore, the system is unstable for all K > 0 since the gain K cannot be simultaneously greater than 20 and less than 1. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 242 CHAPTER 6 The Stability of Linear Feedback Systems Problems P6.1 (a) Given s2 + 5s + 2 , we have the Routh array s2 1 2 s1 5 0 so 2 Each element in the first column is positive, thus the system is stable. (b) Given s3 + 4s2 + 8s + 4 , we have the Routh array s3 1 8 s2 4 4 s1 7 0 so 4 Each element in the first column is positive, thus the system is stable. (c) Given s3 + 2s2 − 6s + 20 , we determine by inspection that the system is unstable, since it is necessary that all coefficients have the same sign. There are two roots in the right half-plane. (d) Given s4 + s3 + 2s2 + 12s + 10 , we have the Routh array s4 1 2 10 s3 1 12 0 s2 -10 10 0 s1 13 0 so 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 243 Problems There are two sign changes in the first column, thus the system is unstable with two roots in the right half-plane. (e) Given s4 + s3 + 3s2 + 2s + K , we have the Routh array s4 1 3 K s3 1 2 0 s2 1 K s1 2−K 0 so K Examining the first column, we determine that the system is stable for 0 < K < 2. (f) Given s5 + s4 + 2s3 + s + 6 , we know the system is unstable since the coefficient of the s2 term is missing. There are two roots in the right half-plane. (g) Given s5 + s4 + 2s3 + s2 + s + K , we have the Routh array s5 1 2 1 s4 1 1 K s3 1 s2 1−K K K s1 −K 0 so K Examining the first column, we determine that for stability we need K > 0 and K < 0. Therefore the system is unstable for all K. P6.2 (a) The closed-loop characteristic polynomial is s4 + 27.88s3 + 366.4s2 + 1500s + 1500ka = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 244 CHAPTER 6 The Stability of Linear Feedback Systems s4 1 366.4 s3 27.88 1500 s2 312.6 1500ka s1 b so 1500ka 1500ka where b = 1500 − 133.78ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka > 0 for stability. Thus, 0 < ka < 11.21 . (b) With Ts = 1.5 = 4 , ζωn we determine that ζωn = 2.67 . So, shift the axis by s = so − 2.67, and (so − 2.67)4 + 27.88(so − 2.67)3 + 366.4(so − 2.67)2 + 1500(so − 2.67) + 1500ka = s4o + 17.2s3o + 185.85s2o + 63.55so − 1872.8 + 1500ka . The Routh array is s4 1 185.85 s3 17.2 63.55 s2 182.16 1500ka -1872.8 s1 b so 1500ka -1872.8 1500ka -1872.8 where b = 240.38 − 141.63ka . Examining the first column of the Routh array, we find that b > 0 and 1500ka − 1872.8 > 0. Thus, 1.25 < ka < 1.69. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 245 Problems P6.3 (a) Given G(s) = K , (s + 1)(s + 2)(0.5s + 1) and H(s) = 1 , 0.005s + 1 the closed-loop transfer function is T (s) = 0.0025s4 K(0.005s + 1) . + 0.5125s3 + 2.52s2 + 4.01s + 2 + K Therefore, the characteristic equation is 0.0025s4 + 0.5125s3 + 2.52s2 + 4.01s + (2 + K) = 0 . The Routh array is given by s4 0.0025 2.52 2+K s3 0.5125 4.01 0 s2 2.50 2+K s1 3.6 − 0.205K 0 so 2+K Examining the first column, we determine that for stability we require −2 < K < 17.6 . (b) Using K = 9, the roots of the characteristic equation are s1 = −200 , s2,3 = −0.33 ± 2.23j , and s4 = −4.35 . Assuming the complex roots are dominant, we compute the damping ratio ζ = 0.15. Therefore, we estimate the percent overshoot as √ 2 P.O. = 100e−πζ/ 1−ζ = 62% . The actual overshoot is 27%, so we see that assuming that the complex poles are dominant does not lead to accurate predictions of the system response. P6.4 (a) The closed-loop characteristic equation is 1 + GH(s) = 1 + K(s + 40) =0, s(s + 10)(s + 20) © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 246 CHAPTER 6 The Stability of Linear Feedback Systems or s3 + 30s2 + 200s + Ks + 40K = 0 . The Routh array is s3 1 200 + K s2 30 40K s1 200 − so K 3 0 40K Therefore, for stability we require 200 − K/3 > 0 and 40K > 0. So, the range of K for stability is 0 < K < 600 . (b) At K = 600, the auxilary equation is 30s2 + 40(600) = 0 or s2 + 800 = 0 . The roots of the auxiliary equation are s = ±j28.3 . (c) Let K = 600/2 = 300. Then, to the shift the axis, first define so = s + 1. Substituting s = so − 1 into the characteristic equation yields (so −1)3 +30(so −1)2 +500(so −1)+12000 = s3o +27s2o +443so +11529 . The Routh array is s3 1 443 s2 27 11529 s1 16 0 so 11529 All the elements of the first column are positive, therefore all the roots lie to left of s = −1. We repeat the procedure for s = so − 2 and obtain s3o + 24s2o + 392so + 10992 = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 247 Problems s3 1 392 s2 24 10992 s1 -66 0 so 10992 There are two sign changes in the first column indicating two roots to right of s = −2. Combining the results, we determine that there are two roots located between s = −1 and s = −2. The roots of the characteristic equation are s1 = −27.6250 and s2,3 = −1.1875 ± 20.8082j . We see that indeed the two roots s2,3 = −1.1875±20.8082j lie between -1 and -2. P6.5 (a) Given the characteristic equation, s3 + 3s2 + 4s + 2 = 0 , we compute the roots s1 = −1, and s2,3 = −1 ± j. (b) The roots of the characteristic equation s4 + 9s3 + 30s2 + 42s + 20 = 0 are s1 = −1, s2 = −2, and s3,4 = −3 ± j1. (c) The roots of the characteristic equation s3 + 19s2 + 110s + 200 = 0 are s1 = −4, s2 = −5, and s3 = −10. P6.6 (a) The characteristic equation is 1 + G(s) = 0 , or s3 + s2 + 10s + 2 = 0 . The roots are: s1 = −0.2033, and s2,3 = −0.3984 ± j3.1112. (b) The characteristic equation is s4 + 10s3 + 35s2 + 50s + 24 = 0 . The roots are s1 = −1, s2 = −2, s3 = −3, and s4 = −4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 248 CHAPTER 6 The Stability of Linear Feedback Systems (c) The characteristic equation is s3 + 11s2 + 29s + 6 = 0 . The roots are s1 = −0.2258, s2 = −3.8206 and s3 = −6.9536. P6.7 (a) The closed-loop characteristic equation is s3 + 101s2 + (100 + 10KKa )s + 100KKa = 0 . The Routh array is s3 1 100 + 10KKa s2 101 100KKa s1 b so 100KKa where b = 100 + 910 KKa > 0 . 101 Thus, examing the first column, we determine that KKa > 0 stabilizes the system. (b) The tracking error is e(s) = lim s(1 − T (s)) s→0 100 100 = . 2 s KKa We require E(s) < 1o = 0.01745. So, KKa > 100 = 5729 . 0.01745 When KKa = 5729, the roots of the characteristic polynomial are s1 = −10.15 P6.8 and s2,3 = −45.43 ± j233.25 . (a) The closed-loop characteristic equation is 1+ K =0, (0.5s + 1)(s + 1)( 14 s + 1) or s3 + 7s2 + 14s + 8(1 + K) = 0 . The Routh array is © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 249 Problems s3 1 14 s2 7 8(1 + K) s1 b so 8(1 + K) where 7(14) − 8(1 + K) . 7 b= For stability, we require b > 0 and 8(1 + K) > 0. Therefore, the range of K for stability is −1 < K < 11.25 . (b) Let K = 11.25/3 = 3.75. Then, the closed-loop transfer function is T (s) = s3 + 3.37 . + 14s + 38 7s2 The settling time to a step input is Ts ≈ 6 seconds. (c) We want Ts = 4 sec, so Ts = 4 = 4 ζωn implies ζωn = 1 . Our desired characteristic polynomial is (s + b)(s2 + 2ζωn s + ωn2 ) = s3 + (2 + b)s2 + (ωn2 + 2b)s + bωn2 where we have used the fact that ζωn = 1 and ωn and b are to be determined. Our actual characteristic polynomial is s3 + 7s2 + 14s + 8(1 + K) = 0 . Comparing the coefficients of the actual and desired characteristic polynomials, we find the following relationships: ωn2 2+b=7 + 2b = 14 bωn2 = 8(1 + K) . Solving these three equations yields b=5, ωn = 2 and K = 1.5 . The actual settling time is Ts = 4.17 sec. This is not exactly our © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 250 CHAPTER 6 The Stability of Linear Feedback Systems desired Ts since we have the contribution of the additional pole at s = −5. The closed-loop poles are s1 = −5 and s2,3 = −1 ± 1.73j . P6.9 (a) The closed-loop characteristic equation is 1 + GH(s) = 1 + 10K , (s + 100)(s + 20)2 or s3 + 140s2 + 4400s + 40000 + 10K = 0 . The Routh array is s3 1 4400 s2 140 40000 + 10K s1 b so 40000 + 10K where b= 140(4400) − (40000 + 10K) . 140 Examining the first column and requiring all the terms to be positive, we determine that the system is stable if −4000 < K < 57600 . (b) The desired characteristic polynomial is (s+b)(s2 +1.38ωn s+ωn2 ) = s3 +(1.38ωn +b)s2 +(ωn2 +1.38ωn b)s+bωn2 where we have used the fact that ζ = 0.69 to achieve a 5% overshoot, and ωn and b are to be determined. The actual characteristic polynomial is s3 + 140s2 + 4400s + 40000 + 10K = 0 . Equating the coefficients of the actual and desired characteristic polynomials, and solving for K, b, and ωn yields b = 104.2 , ωn = 25.9 So, a suitable gain is K = 3003. and K = 3003 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 251 Problems P6.10 (a) The closed-loop characteristic equation is s4 + 7s3 + 20s2 + (24 + K)s + 10K = 0 . The Routh array is s4 1 20 10K s3 7 24 + K 0 s2 116−K 7 10K s1 b so 10K where b=  116−K 7  (24 + K) − 70K  Setting b > 0 yields 116−K 7  . 2784 − 398K − K 2 > 0 , which holds when −404.88 < K < 6.876 . Examining the first column, we also find that K < 116 and K > 0 for stability. Combining all the stability regions, we determine that for stability 0 < K < 6.876 . (b) When K = 6.876, the roots are s1,2 = −3.5 ± 1.63j , P6.11 and s3,4 = ±2.1j . Given s3 + (1 + K)s2 + 10s + (5 + 15K) = 0 , the Routh array is s3 1 10 s2 1+K 5 + 15K s1 b so 5 + 15K © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 252 CHAPTER 6 The Stability of Linear Feedback Systems where b= 5 − 5K (1 + K)10 − (5 + 15K) = . 1+K 1+K Given that K > 0, we determine that the system is stable when 5−5K > 0 or 0 0, ab − c > 0 and c > 0. When a > 0 and c > 0, we know that b > 0. So, a necessary condition for stability is that all coefficients a, b, and c be positive. The necessary and sufficient conditions for stability also require that b > c/a, in addition to a > 0 and c > 0. P6.13 The characteristic equation is s3 + (p + 2ζωn )s2 + (2ζωn p + Kωn2 )s + Kωn2 z = 0. The conditions for stability (see P6.12) are p + 2ζωn > 0, 2ζωn p + Kωn2 > (Kωn2 z)/(p + 2ζωn ), and Kωn2 z > 0. Since we know that K > 0, ζ > 0, and ωn > 0, it follows that for stability z > 0, p > −2ζωn , and 2ζωn p + Kωn2 > P6.14 Kωn2 z . p + 2ζωn The system has the roots s1,2 = ±j , s3,4 = ±j , and s5,6 = −1 ± 3j , © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 253 Problems Therefore, the system is not stable since there are repeated roots on the jω-axis. P6.15 (a) Neglecting the zeros and poles, we have the characteristic equation s4 + 30s3 + 325s2 + 2500s + K = 0 . The Routh array is s4 1 325 K s3 30 2500 0 s2 241.67 K s1 b so K where b= 604166.67 − 30K . 241.67 Therefore, the system is stable for 0 < K < 20139. (b) Without neglecting the zeros and poles, the closed-loop characteristic equation is s6 + 90s5 + 5525s4 + 12400s3 + (1255000 + K)s2 + (8500000 + 30K)s + 1125K = 0 . This is stable for 0 < K < 61818 . We see that the additional poles and zero makes the system stable for a much larger gain K. P6.16 (a) The Routh array is s3 1 5 s2 5 6 s1 3.8 so 6 Examining the first column of the Routh array, we see no sign changes. So, the system is stable. (b) The roots of the system are s1 = −0.3246 and s2,3 = −2.3377 ± 3.6080j. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 254 CHAPTER 6 The Stability of Linear Feedback Systems (c) The step response is shown in Figure P6.16. Step Response 0.18 0.16 0.14 Amplitude 0.12 0.1 0.08 0.06 0.04 0.02 0 0 5 10 15 Time (sec ) FIGURE P6.16 Unit step response. P6.17 The closed-loop transfer function is T (s) = s3 + K +1 . + 3s + K + 1 3s2 The Routh array is s3 1 3 s2 3 K+1 s1 8−K 3 so K +1 So, for stability we require −1 < K < 8. P6.18 The system characteristic equation is s2 + (h − k)s + ab − kh = 0 . For stability we require h > k and ab > kh. If k > h, the system is unstable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 255 Problems P6.19 (a) The characteristic equation is s3 + 9s2 + (K − 10)s + 2K = 0 . The Routh array is s3 1 s2 K − 10 9 2K s1 7K−90 9 so 2K For stability K > 90/7 . (b) When K = 90/7, the system is marginally stable. The roots are q s1,2 = ±j 20/7 , at the jω-axis crossing. P6.20 The closed-loop characteristic equation is q(s) = s5 + s4 + 9s3 + Ks2 + 2Ks + K . The range of stability for the vertical-liftoff vehicle is 5.177 < K < 7.823 . Therefore, for K = 6, the system is stable. When K = 6 we have q(s) = s5 + s4 + 9s3 + 6s2 + 12s + 6 The Routh array is s5 1 9 12 s4 1 6 6 s3 3 6 s2 4 6 s1 1.5 so 6 All entries in the first column are positive, so the system is stable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 256 CHAPTER 6 P6.21 The Stability of Linear Feedback Systems The state transition matrix is   (k − p1 )e−p1 t − (k2 − p2 )e−p2 t e−p1 t − e−p2 t 1   2 Φ(t, 0) = −p t −p t −p t −p t p2 − p1 1 2 1 2 −k1 e + k1 e −p1 e + p2 e where p1 p2 = k1 and p1 + p2 = k2 . We assume that p1 6= p2 . In the case when p1 = p2 , the state transition matrix will change, but the factors e−p1 t and e−p2 t will remain. The eigenvalues of A are given by the solution to det |λI − A| = λ2 + k2 λ + k1 = 0 . q Therefore, the eigenvalues are λ1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the eigenvalues are in the left half-plane, and the system is stable. The transfer function is given by G(s) = C (sI − A)−1 B = − s−1 . s2 + k2 s + k1 Therefore the characteristic equation is s2 + k2 s + k1 = 0 and the poles q are s1,2 = −k2 /2 ± k22 − 4k1 . If k2 > 0 and k1 > 0, then the poles are in the left half-plane, and the system is stable. Notice that the values of λ1,2 and s1,2 are the same. Also, the eigenvalues are the same as the values of −p1 and −p2 . So, if the eigenvalues are negative, then the elements of the state transition matrix will decay exponentially. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 257 Advanced Problems Advanced Problems The Routh array is s4 1 K1 s3 20 4 s2 20K1 −4 20 K2 s1 b 0 so K2 K2 where b= 20K1 − 4 − 100K2 . 5K1 − 1 For stability, we require K2 > 0, K1 > 0.2, and b > 0. Therefore, using the condition that b > 0, we obtain K2 < 0.2K1 − 0.04 . The stability region is shown in Figure AP6.1. 0. 4 0.35 0. 3 0.25 K2 AP6.1 0. 2 0.15 0. 1 0.05 STABLE REGION 0 0. 2 0. 4 0. 6 0. 8 1 1. 2 K1 FIGURE AP6.1 Stability region. 1. 4 1. 6 1. 8 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 258 CHAPTER 6 AP6.2 The Stability of Linear Feedback Systems The Routh array is s4 1 30 s3 9 s2 310−K 9 K − 40 s1 b so K K K 0 where b= (310 − K)(K − 40) − 81K . 310 − K Therefore, using the condition that b > 0, we obtain the stability range for K: 59.07 < K < 209.94 . AP6.3 (a) The steady-state tracking error to a step input is ess = lim s(1 − T (s))R(s) = 1 − T (0) = 1 − α . s→0 We want |1 − α| < 0.05 . This yields the bounds for α 0.95 < α < 1.05 . (b) The Routh array is s3 1 α s2 1+α 1 s1 b 0 so 1 where b= α2 + α − 1 . 1+α Therefore, using the condition that b > 0, we obtain the stability range for α: α > 0.618 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 259 Advanced Problems (c) Choosing α = 1 satisfies both the steady-state tracking requirement and the stability requirement. The closed-loop transfer function is T (s) = s3 K . + (p + 1)s2 + ps + K The Routh array is s3 1 p s2 1+p K s1 b 0 so K where b= p2 + p − K . 1+p Therefore, using the condition that b > 0, we obtain the the relationship K < p2 + p . The plot of K as a function of p is shown in Figure AP6.4. 120 100 80 K AP6.4 60 40 STABLE REGION 20 0 0 1 2 3 4 5 p FIGURE AP6.4 Stability region. 6 7 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 260 CHAPTER 6 The closed-loop transfer function is T (s) = 30K1 K2 . (s + 1 + K1 K3 )(s − 10)(2s + K2 K3 − 4) + 30K1 K2 K4 The Routh array is s3 2 a s2 b c s1 d 0 so c where a = −9K2 K3 + 16 + K1 K2 K32 − 24K1 K3 , b = 2K1 K3 + K2 K3 − 22, and c = −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 and d = (ab − 2c)/b . The conditions for stability are 2K1 K3 + K2 K3 − 22 > 0 −10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 > 0 −2(−10K2 K3 + 40 − 10K1 K2 K32 + 40K1 K3 ) + (9K2 K3 +16 + K1 K2 K32 − 24K1 K3 )(2K1 K3 + K2 K3 − 22) > 0 Valid values for the various gains are: K1 = 50, K2 = 30, K3 = 1, and K4 = 0.3. The step response is shown in Figure AP6.5. Step Response 350 300 250 Amplitude AP6.5 The Stability of Linear Feedback Systems 200 150 100 50 0 FIGURE AP6.5 Stability region. 0 5 10 15 Time (sec) 20 25 30 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 261 Advanced Problems AP6.6 The characteristic equation is s3 + 7s2 + (KD + 14)s + KP + 8 = 0. For stability we require that KP > −8 and KD > KP + 8 − 14. 7 The relationship between KD and KP is shown in Figure AP6.6. 2 0 −2 STABLE REGION K D −4 −6 −8 UNSTABLE REGION −10 −12 −14 0 10 20 30 40 50 60 70 80 90 100 K P FIGURE AP6.6 Stability region. AP6.7 The characteristic equation is 0.1s4 + 2.05s3 + s2 + 8KP s + 8KI = 0. From the Routh array we find the conditions for stability are 0 < KI < 0.3125 1.2812 − p 1.6416 − 5.2531KI < KP < 1.2812 + p 1.6416 − 5.2531KI © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 262 CHAPTER 6 The Stability of Linear Feedback Systems Design Problems CDP6.1 The plant model with parameters given in Table CDP2.1 in Dorf and Bishop is given by: θ(s) 26.035 = . Va (s) s(s + 33.142) In the above transfer function we have neglected the motor inductance Lm . The closed-loop transfer function from the input to the output is 26.035Ka θ(s) = 2 . R(s) s + 33.142s + 26.035Ka The Routh array is s2 1 26.035Ka s1 33.142 0 s0 26.035Ka Stability is achieved for any 0 < Ka < ∞. DP6.1 The closed-loop characteristic polynomial is 1 1 1 s3 + s2 (5 + p + K) + s( Kp + K + 5p) + K = 0 . 5 5 5 (i) When p = 2, we have 1 3 s3 + s2 (7 + K) + s(10 + K) + K = 0 . 5 5 The Routh array is s3 1 s2 7+ s1 b so K 10 + 35 K K 5 K where b= (7 + K/5)(10 + 3K/5) − K . 7 + 51 K When −32.98 < K < −17.69, we find that b > 0. Examining the other terms in the first column of the array, we find that the system © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 263 Design Problems is stable for any K > 0. (ii) When p = 0, we have 1 1 s3 + s2 (5 + K) + s( K) + K = 0 . 5 5 The Routh array is s3 1 1 5K s2 5 + 15 K K s1 b so K where b= (5 + 15 K) 15 K − K K 2 /25 = . (5 + K/5) (5 + K/5) Again, examination of the first column reveals that any K > 0 results in a stable system. So, we just need to select any K > 0; e.g. K = 10. DP6.2 (a) The closed-loop characteristic equation is 1+ 20(Ks + 1) =0, s2 (s + 20) or s3 + 20s2 + 20Ks + 20 = 0 . The Routh array is s3 1 20K s2 20 20 s1 b so 1 where b= 20K − 1 . 1 For stability, we require K > 0.05. (b) The desired characteristic polynomial is (s2 + as + b)(s + 5) = s3 + s2 (a + 5) + s(5a + b) + 5b = 0 . Equating coefficients with the actual characteristic equation we can © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 264 CHAPTER 6 The Stability of Linear Feedback Systems solve for a, b and K, yielding b = 4, a = 15, and K= 5a + b 79 = . 20 20 (c) The remaining two poles are s1 = −14.73 and s2 = −0.27. (d) The step response is shown in Figure DP6.2. 1 0.9 0.8 0.7 y(t) 0.6 0.5 0.4 0.3 0.2 0.1 0 0 2 4 6 8 10 time (sec) 12 14 16 18 FIGURE DP6.2 Mars guided vehicle step response. DP6.3 (a) The closed-loop characteristic equation is 2τ s3 + (τ + 2)s2 + (K + 1)s + 2K = 0 . The Routh array is s3 2τ K+1 s2 τ +2 2K s1 b so 2K where b= (τ + 2)(K + 1) − 4Kτ . (τ + 2) 20 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 265 Design Problems Examining the first column of the Routh array, we determine that for stability τ > 0, K > 0 and setting b > 0 yields the relationships: (1) K < τ +2 2 when τ > 3τ − 2 3 (2) K > 0 when 0 < τ ≤ 2 . 3 The plot of τ versus K is shown in Figure DP6.3a. 5 4.5 4 3.5 tau 3 2.5 2 1.5 1 0.5 0 STABLE REGION 1 2 3 4 5 6 7 K FIGURE DP6.3 (a) The plot of τ versus K. (b) The steady-state error is ess = A , Kv where Kv = 2K . So, ess 1 = . A 2K We require that ess ≤ 0.25A, therefore K≥2. One solution is to select τ = 0.5, then we remain within the stable region. (c) The step response is shown in Figure DP6.3b. The percent overshoot is P.O. = 57%. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 266 CHAPTER 6 The Stability of Linear Feedback Systems 1.6 P.O. = 56.77 % 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 2 4 6 8 10 12 14 16 18 20 time (sec) FIGURE DP6.3 CONTINUED: (b) Closed-loop system step response. DP6.4 (a) The closed-loop characteristic polynomial is s3 + Ks2 + [(2 + m)K − 1]s + 2mK = 0 . The Routh array is s3 1 s2 2K + mK − 1 K 2mK s1 b so 2mK Examining the first column of the Routh array, we see that for stability we require m > 0, K > 0, and b > 0, where b= (2K + mK − 1)K − 2mK = (2 + m)K − (1 + 2m) > 0 , K or K> 1 + 2m . 2+m The plot of K vs m is shown in Figure DP6.4a. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 267 Design Problems 1.6 1.4 STABLE REGION 1.2 K 1 0.8 0.6 0.4 0 0.5 1 1.5 2 2.5 3 3.5 6 7 4 4.5 5 m FIGURE DP6.4 (a) The plot of K versus m. 1.8 P.O. = 64.3208 % 1.6 1.4 1.2 y(t) 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 time (sec) FIGURE DP6.4 CONTINUED: (b) Shuttle attitude control step response. (b) The steady-state error is ess 1 1 = < 0.10 , = A Kv 2mK 8 9 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 268 CHAPTER 6 The Stability of Linear Feedback Systems or mK > 5. For example, we can select m = 0.5 and K = 2. (c) See Figure DP6.4b for the step response where P.O. = 64.3%. DP6.5 The closed-loop transfer function is T (s) = s3 + K . + 20s + K 10s2 The range of K for stability is 0 < K < 200. If we let K = Km /N where Km = 200, then N = 6.25 results in a step response with P.O. = 15.7% and Ts = 1.96 seconds. DP6.6 The closed-loop system is given by  ẋ =  0 1 2 − K1 −2 − K2   x+ 0 1  r The characteristic polynomial is s2 +(2+K2 )s+K1 −2h = 0. Soithe system is stable for K1 > 2 and K2 > −2. Selecting K = 10 2 results in closed-loop eigenvalues at s = −2 ± 2j. The closed-loop step response has a settling time of 2.11 s and a percent overshoot of 4.32%. Im(s) sin-1 ζ = sin-1 0.69=43.63ο Re(s) desired region for eigenvalues ζωn = -1 FIGURE DP6.6 Performance region. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 269 Design Problems DP6.7 (a) The inner loop closed-loop transfer function is 20s Y (s) = 3 . 2 U (s) s + 10s + 20s + 20K1 The Routh array is s3 1 20 s2 ω 20K1 s1 200−20K1 10 so 20K1 For stability 0 < K1 < 10. (b) The fastest response (that is, the quickest settling time) occurs when K1 = 2.2 (c) With K1 = 2.2, the closed-loop transfer function is Y (s) 20K2 s = 3 . R(s) s + 10s2 + (20 + 20K2 )s + 44 The Routh array is s3 1 20(K2 + 1) s2 10 44 s1 200K2 +156 10 so 44 For stability, we require 200K2 + 156 > 0 . Therefore, K2 > −0.78. DP6.8 The closed-loop characteristic equation is s2 + 4KD s + 4(KP + 1) = 0. So, it is possible to find KP and KD to stabilize the system. For example, any KP > 0 and KD > 0 leads to stability. Choosing KP ≥ 9 results in a steady-state tracking error √ less than 0.1 due to a unit step input. Then, the damping ratio ζ = 2/2 is achieved by selecting √ √ 2 KP + 1 KD = . 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 270 CHAPTER 6 The Stability of Linear Feedback Systems Computer Problems CP6.1 The m-file script is shown in Figure CP6.1. ans = -0.6063 + 2.7322i -0.6063 - 2.7322i -1.7874 pa=[1 3 10 14]; roots(pa) pb=[1 8 24 32 16]; roots(pb) pc=[1 0 2 1]; roots(pc) ans = -2.0004 -2.0000 + 0.0004i -2.0000 - 0.0004i -1.9996 ans = 0.2267 + 1.4677i 0.2267 - 1.4677i -0.4534 FIGURE CP6.1 Computing the polynomial roots with the rootsfunction. CP6.2 The m-file script is shown in Figure CP6.2. K1=1;K2=2;K3=5; den=[1 2 1]; num1=K1*[1 -1 2];num2=K2*[1 -1 2];num3=K3*[1 -1 2]; sys1 = tf(num1,den); sys2 = tf(num2,den); sys3 = tf(num3,den); sys1_cl=feedback(sys1,[1]); sys2_cl=feedback(sys2,[1]); sys3_cl=feedback(sys3,[1]); p1 = pole(sys1_cl), p2 = pole(sys2_cl), p3 = pole(sys3_cl) ans = -2.5000e -01 + 1.1990e+00i -2.5000e -01 - 1.1990e+00i ans = ans = 2.5000e -01 + 1.3307e+00i 2.5000e -01 - 1.3307e+00i 0 + 1.2910e+00i 0 - 1.2910e+00i FIGURE CP6.2 K = 1 is stable;K = 2 is marginally stable; and K = 5 is unstable. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 271 Computer Problems CP6.3 The closed-loop transfer function and the roots of the characteristic equation are shown in Figure CP6.3. Transfer function: s+1 ---------------------s^3 + 4 s^2 + 7 s + 11 numg=[1 1]; deng=[1 4 6 10]; sysg = tf(numg,deng); sys = feedback(sysg,[1]) r=pole(sys) r= -2.8946 -0.5527 + 1.8694i -0.5527 - 1.8694i FIGURE CP6.3 Closed-loop transfer function and roots. CP6.4 There are no poles in the right half-plane, but the system is unstable since there are multiple poles on the jω-axis at s = ±j and s = ±j (see Figure CP6.4). Step Response From: U(1) 25 20 -2.0000 0.0000 + 1.0000i 0.0000 - 1.0000i -0.0000 + 1.0000i -0.0000 - 1.0000i 10 5 To: Y(1) ans = 15 Amplitude num=[1]; den=[1 2 2 4 1 2]; sys = tf(num,den); pole(sys) t = 0:0.1:100; step(sys,t) 0 -5 -10 -15 -20 -25 0 10 20 30 40 50 Time (sec.) FIGURE CP6.4 Unstable system step response. 60 70 80 90 100 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 272 CHAPTER 6 CP6.5 The Stability of Linear Feedback Systems The closed-loop system poles for the slow/fast pilots are shown in Figure CP6.5. The maximum allowable time delay is 0.2045seconds. At the maximum allowable time delay, the system has roots on the jω-axis at s = ±2.6j. The slow pilot destabilizes the aircraft. nume=[-10]; dene=[1 10]; syse = tf(nume,dene); numg=[-1 -6]; deng=[1 3 6 0]; sysg = tf(numg,deng); % % Fast pilot % tau=0.1; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); fast_pilot=pole(sys) % % Slow pilot % tau=0.6; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); slow_pilot = pole(sys) % % Maximum pilot time delay, tau = 0.2045 sec % tau=0.2045; tau1=2; K=1; tau2=0.5; nump=-K*[tau1*tau tau-2*tau1 -2]; denp=[tau2*tau tau+2*tau2 2]; sysp = tf(nump,denp); sysa = series(sysp,syse); sysb = series(sysa, sysg); sys = feedback(sysb,[1]); max_pilot_delay=pole(sys) closed-loop system poles fast_pilot = -19.6267 -10.7712 -3.8885 -0.1697 + 2.7880i -0.1697 - 2.7880i -0.3742 slow_pilot = -9.4526 -4.5228 + 2.2595i -4.5228 - 2.2595i 0.2793 + 2.0314i 0.2793 - 2.0314i -0.3937 max_pilot_delay = -10.0433 + 2.2684i -10.0433 - 2.2684i -4.3153 0.0001 + 2.6040i 0.0001 - 2.6040i -0.3783 FIGURE CP6.5 Closed-loop system poles for an aircraft with a pilot in-the-loop. CP6.6 The closed-loop transfer function is T (s) = s3 + 5s2 1 . + (K − 3)s + K + 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 273 Computer Problems Utilizing the Routh-Hurwitz approach, for stability we determine that K>4. When K = 4, the roots of the characteristic equation are s1 = −5 and s2,3 = ±j . The m-file script which generates a plot of the roots of the characteristic equation as a function of K is shown in Figure CP6.6. K=[0:0.1:5]; n=length(K); for i=1:n numg=[1]; deng=[1 5 K(i)-3 K(i)]; sys_o = tf(numg,deng); sys_cl = feedback(sys_o,[1]); p(:,i)=pole(sys_cl); end plot(real(p),imag(p),'x'), grid text(-0.9,0.95,'K=4 -->'); text(-0.2,1.3,'K=5'); text(0,0.2,'K=0') % From a Routh-Hurwitz analysis we find that % minimum K for stability is K=4 Kmax=4; numg=[1]; deng=[1 5 Kmax-3 Kmax]; sysg = tf(numg,deng); sys_cl = feedback(sysg,[1]); pole(sys_cl) 1.5 K=5 1 K=4 --> 0.5 K=0 0 -0.5 -1 -1.5 -6 -5 -4 -3 -2 -1 0 FIGURE CP6.6 Roots of the characteristic equation as a function of K, where 0 < K < 5. 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 274 CHAPTER 6 CP6.7 The Stability of Linear Feedback Systems The characteristic equation is p(s) = s3 + 10s2 + 14s + 12 . A=[0 1 0;0 0 1;-12 -14 -10]; b=[0;0;12]; c=[1 1 0]; d=[0]; sys = ss(A,b,c,d); % % Part (a) % p=poly(A) % % Part (b) % roots(p) % % Part (c) % step(sys) p= 1.0000 10.0000 14.0000 12.0000 ans = -8.5225 -0.7387 + 0.9286i -0.7387 - 0.9286i Step Response 1.4 1.2 Amplitude 1 0.8 0.6 0.4 0.2 0 0 1 2 3 4 5 Time (sec) 6 7 8 9 FIGURE CP6.7 Characteristic equation from the state-space representation using the poly function. The roots of the characteristic equation are s1 = −8.5225 and s2,3 = −0.7387 ± 0.9286j . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 275 Computer Problems The system is stable since all roots of the characteristic equation are in the left half-plane. The unit step response and associated m-file script are shown in Figure CP6.7. CP6.8 The characteristic equation is s3 + 10s2 + 10s + 5K1 = 0 . (a) The Routh array is s3 1 10 s2 10 5K1 s1 100−5K1 10 so 5K1 From the Routh-Hurwitz criterion, we obtain the limits 0 < K1 < 20 for stability. (b) The plot of the pole locations is 0 < K1 < 30 is shown in Figure CP6.8. As seen in Figure CP6.8, when K1 > 20, the pole locations move into the right half-plane. Root Locus 4 3 Imaginary Axi s 2 1 k=20 0 ?-1 ?-2 ?-3 ?-4 ?-12 ?-10 ?-8 ?-6 ?-4 ?-2 0 Real Axi s FIGURE CP6.8 Pole locations for 0 < K1 < 30. CP6.9 (a) The characteristic equation is s3 + 2s2 + s + k − 4 = 0 . 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 6 The Stability of Linear Feedback Systems The Routh array is s3 1 1 s2 2 s1 6−k 2 k−4 so k−4 For stability, we obtain 4 < k < 6. (b) The pole locations for 0 < k < 10 are shown in Figure CP6.9. We see that for 0 < k < 4 the system is unstable. Similarly, for 6 < k < 10, the system is unstable. Root Locus 2 k=10 pole locations when k=0 1.5 1 0.5 pole location when k=0 increasing k k=10 0 k=6 inc rea ?-1 ?-1.5 ?-2 ?-3 k=4 gk ?-0.5 sin Imaginary Axi s 276 ?-2 ?-1 Real Axi s FIGURE CP6.9 Pole locations for 0 < k < 10. 0 1 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. C H A P T E R 7 The Root Locus Method Exercises (a) For the characteristic equation 1+K s(s + 4) =0, + 2s + 2 s2 the root locus is shown in Figure E7.1. 4 3 2 1 Imag Axis E7.1 0 x o o -1 x -2 -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.1 s(s+4) Root locus for 1 + K s2 +2s+2 = 0. 277 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 278 CHAPTER 7 The Root Locus Method (b) The system characteristic equation can be written as (1 + K)s2 + (2 + 4K)s + 2 = 0 . Solving for s yields −(1 + 2K) s= ± (1 + K) p (2 + 4K)2 − 8(1 + K) . 2(1 + K) When (2 + 4K)2 − 8(1 + K) = 0 , then we have two roots at s1,2 = − (1+2K) 1+K . Solving for K yields K = 0.31. (c) When K = 0.31, the roots are s1,2 = −(1 + 0.62) = −1.24 . (1.31) (d) When K = 0.31, the characterisitc equation is s2 + 2.472s + 1.528 = (s + 1.24)2 = 0 . Thus, ωn = 1.24 and ζ = 1, the system is critically damped. The settling time is Ts ≈ 4 sec. E7.2 (a) The root locus is shown in Figure E7.2. When K = 6.5, the roots of the characteristic equation are s1,2 = −2.65 ± j1.23 and s3,4 = −0.35 ± j0.8 . The real part of the dominant root is 8 times smaller than the other two roots. (b) The dominant roots are (s + 0.35 + j0.8)(s + 0.35 − j0.8) = s2 + 0.7s + 0.7625 . From this we determine that ωn = 0.873 and ζ = 0.7 = 0.40 . 2(0.873) Thus, the settling time is 4 4 = = 11.43 sec . ζωn 0.35 √ 2 The percent overshoot is P.O. = e−πζ/ 1−ζ = 25.4%. Ts = © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 279 Exercises 4 3 * K=6.5 2 * Imag Axis 1 x * 0 x -1 x * x * -2 -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.2 Root locus for 1 + K s(s+2)(s12 +4s+5) = 0. The root locus is shown in Figure E7.3. The roots are s1 = −8.7, s2,3 = −1.3 ± j2.2 when K = 7.35 and ζ = 0.5. 4 zeta=0.5 3 2 o * <----- K=7.35 1 Imag Axis E7.3 0 x * x -1 -2 o * -3 -4 -10 -8 -6 -4 -2 Real Axis FIGURE E7.3 2 Root locus for 1 + K ss2+4s+8 = 0. (s+4) 0 2 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 280 CHAPTER 7 E7.4 The Root Locus Method The root locus is shown in Figure E7.4. 2 1.5 1 x Imag Axis 0.5 0 o -0.5 -1 x -1.5 -2 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 Real Axis FIGURE E7.4 s+1 = 0. Root locus for 1 + K s2 +4s+5 The departure angles and entry points are θd = 225o , −225o and σb = −2.4 . E7.5 (a) The root locus is in Figure E7.5. The breakaway points are σb1 = −13.0 , σb2 = −5.89 . (b) The asymptote centroid is σcent = −18 , and φasym = ±90o . (c) The gains are K1 = 1.57 and K2 = 2.14 at the breakaway points. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 281 Exercises 40 30 20 Imag Axis 10 0 -10 < asymptote -20 -30 -40 -15 -10 -5 0 5 Real Axis FIGURE E7.5 s2 +2s+10 Root locus for 1 + K (s4 +38s3 +515s 2 +2950s+6000) = 0. The system is unstable for K > 75. 10 8 System: sys Gain: 75 Pole: −0.000981 + 8.66i Damping: 0.000113 Overshoot (%): 100 Frequency (rad/sec): 8.66 6 4 Imaginary Axis E7.6 2 0 −2 −4 −6 −8 −10 −10 −8 −6 FIGURE E7.6 15K Root locus for 1 + K s(s2 +15s+75) = 0. −4 Real Axis −2 0 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 282 CHAPTER 7 E7.7 The Root Locus Method The root locus is shown in Figure E7.7. The characteristic equation has 20 15 asymptote −−−> Imaginary Axis 10 5 0 System: sys Gain: 27.3 Pole: −1.44 + 1.11i Damping: 0.792 Overshoot (%): 1.7 Frequency (rad/sec): 1.81 −5 −10 −15 −20 −25 −20 −15 −10 −5 Real Axis 0 5 10 FIGURE E7.7 s+8 = 0. Root locus for 1 + K s(s+4)(s+6)(s+9) 4 poles and 1 zero. The asymptote angles are φ = +60o , −60o , −180o centered at σcent = −3.7. When K = 27.35 then ζ = 0.8 for the complex roots. E7.8 The characteristic equation is 1+K (s + 1) =0, s2 (s + 9) or s3 + 9s2 + Ks + K = 0 . For all the roots to be equal and real, we require (s + r)3 = s3 + 3rs2 + 3r 2 s + r 3 = 0 . Equating terms and solving for K yields K = 27. All three roots are equal at s = −3, when K = 27. The root locus is shown in Figure E7.8. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 283 Exercises 8 6 4 Imag Axis 2 3 roots at s=-3 0 x o x -2 -4 -6 -8 -15 -10 -5 0 5 Real Axis FIGURE E7.8 Root locus for 1 + K s2s+1 (s+9) = 0. E7.9 The characteristic equation is 1+K 1 =0 s(s2 + 2s + 5) or s3 + 2s2 + 5s + K = 0 . (a) The system has three poles at s = 0 and −1 ± j2. The number of asymptotes is np − nz = 3 centered at σcent = −2/3, and the angles are φasymp at ±60o , 180o . (b) The angle of departure, θd , is 90o +θd +116.6o = 180o , so θd = −26.6o . (c) The Routh array is s3 1 5 s2 2 K s1 b so K where b = 5 − K/2. So, when K = 10 the roots lie on the imaginary © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 284 CHAPTER 7 The Root Locus Method axis. The auxilary equation is 2s2 + 10 = 0 √ s1,2 = ±j 5 . which implies (d) The root locus is shown in Figure E7.9. 4 3 2 x Imag Axis 1 asymptote ---> 0 x -1 -2 x -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.9 1 = 0. Root locus for 1 + K s(s2 +2s+5) E7.10 (a) The characteristic equation is 1+ K(s + 2) =0. s(s + 1) Therefore, K=− (s2 + s) , (s + 2) and dK s2 + 4s + 2 =− =0. ds (s + 2)2 Solving s2 +4s+2 = 0 yields s = −0.586 and −3.414. Thus, the system © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 285 Exercises breakaway and entry points are at s = −0.586 and s = −3.414. (b) The desired characteristic polynomial is (s + 2 + aj)(s + 2 − aj) = s2 + 4s + 4 + a2 = 0 , where a is not specified. The actual characteristic polynomial is s2 + (1 + K)s + 2K = 0 . Equating coefficients and solving for K yields√K = 3 and a = Thus, when K = 3, the roots are s1,2 = −2 ± 2j. √ 2. (c) The root locus is shown in Figure E7.10. 2 K=3, s=-2+1.414j 1.5 * 1 Imag Axis 0.5 s=-3.41 0 o x s=-0.58 x -0.5 -1 * -1.5 -2 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 Real Axis FIGURE E7.10 s+2 Root locus for 1 + K s(s+1) = 0. E7.11 The root locus is shown in Figure E7.11 for the characteristic equation 1+ K(s + 2.5) =0. (s2 + 2s + 2)(s2 + 4s + 5) From the root locus we see that we can only achieve ζ = 0.707 when K = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 286 CHAPTER 7 The Root Locus Method 5 4 3 2 Imag Axis 1 0 x x x x -2 -1 <---- zeta=0.707 & K=0 o -1 -2 -3 -4 -5 -5 -4 -3 0 1 2 3 4 5 Real Axis FIGURE E7.11 Root locus for 1 + E7.12 K(s+2.5) (s2 +2s+2)(s2 +4s+5) = 0. (a) The root locus is shown in Figure E7.12 for the characteristic equation 1+ K(s + 1) =0. s(s2 + 6s + 18) (b) The roots of the characteristic equation are (i) K = 10: s1,2 = −2.8064 ± 4.2368j and s3 = −0.3872 (ii) K = 20: s1,2 = −2.7134 ± 5.2466j and s3 = −0.5732 (c) The step response performance of the system is summarized in Table E7.12. K 10 20 Ts (sec) 9.0 5.5 0 0 4.8 2.6 P.O. Tr (sec) TABLE E7.12 System performance when K = 10 and K = 20. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 287 Exercises Root Locus 15 10 Imaginary Axis 5 0 −5 −10 −15 −3.5 FIGURE E7.12 Root locus for 1 + E7.13 −3 −2.5 K(s+1) s(s2 +6s+18) −2 −1.5 Real Axis −1 −0.5 0 0.5 = 0. (a) The characteristic equation is s(s + 1)(s + 3) + 4s + 4z = 0 . Rewriting with z as the parameter of interest yields 1+z 4 =0. s(s + 1)(s + 3) + 4s The root locus is shown in Figure E7.13a. (b) The root locations for z = 0.6 , 2.0 , and 4.0 are shown in Figure E7.13a. When z = 0.6, we have ζ = 0.76 and ωn = 2.33. Therefore, the predicted step response is P.O. = 2.4% and Ts = 2.3 sec (ζ = 0.6) . When z = 2.0, we have ζ = 0.42 and ωn = 1.79. Therefore, the predicted step response is P.O. = 23% and Ts = 5.3 sec (ζ = 2.0) . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 4 3 * z=0.6 o z=2.0 + z=4.0 + 2 x o * Imag Axis 1 0 + o x * -1 x * o -2 + -3 -4 -4 -3 -2 -1 0 1 2 3 4 Real Axis FIGURE E7.13 4 = 0. (a) Root locus for 1 + z s(s+1)(s+3)+4s Finally, when z = 4.0, we have ζ = 0.15 and ωn = 2.19. Therefore, the predicted step response is P.O. = 62% and Ts = 12 sec. (c) The actual step responses are shown in Figure E7.13b. 1.6 1.4 ___ z=0.6 - - - z=2.0 1.2 ..... z=4.0 1 y(t) 288 0.8 0.6 0.4 0.2 0 0 2 4 6 8 time (sec) 10 FIGURE E7.13 CONTINUED: (b) Step Responses for z = 0.6, 2.0, and 4.0. 12 14 16 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 289 Exercises (a) The root locus is shown in Figure E7.14 for the characteristic equation K(s + 10) =0. s(s + 5) 1+ The breakaway point is sb = −2.93; the entry point is se = −17.1. 10 8 6 K=5, s=-5+5j * 4 2 Imag Axis E7.14 0 s=-17.1 o x -10 -5 s=-2.93 x -2 -4 -6 -8 -10 -20 -15 0 5 Real Axis FIGURE E7.14 Root locus for 1 + K(s+10) s(s+5) = 0. √ (b) We desire ζ = 1/ 2 = 0.707. So, the desired characteristic polynomial is 1 s2 + 2 √ ωn s + ωn2 = 0 . 2   Comparing the desired characteristic polynomial to the actual we find the relationships √ 2ωn = 5 + K . ωn2 = 10K and Solving for K and ωn yields K = 5 and ωn = 7.07. The roots are s1,2 = −5 ± j5 when K = 5. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 290 CHAPTER 7 (a) The characteristic equation 1+K (s + 10)(s + 2) =0 s3 has the root locus in Figure E7.15. Root Locus 15 10 Imaginary Axi s E7.15 The Root Locus Method 5 K=1.67 0 ?-5 ?-10 ?-15 ?-25 ?-20 ?-15 ?-10 -5 0 5 Real Axi s FIGURE E7.15 Root locus for 1 + K(s+10)(s+2) s3 = 0. (b) The Routh array is s3 1 12K s2 K 20K s1 b so 20K when b = 12K − 20. For stability, we require all elements in the first column to be positive. Therefore, K > 1.67 . (c) When K > 3/4, we have ess = lim sE(s) = lim s s→0 s→0 1 1 s2 · 2 = lim 3 =0. s→0 s + K(s + 1)(s + 3) 1 + GH(s) s © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 291 Exercises The expansion for e−T s is e−T s = 1 − T s + (T s)2 − ... 2! If (T s) << 1, then e−T s ≈ 1 − T s = a + bs , c + ds where a, b, c and d are constants to be determined. Using long division, 40 30 K=21 20 * 10 Imag Axis E7.16 0 x x o -10 -20 -30 -40 -40 -30 -20 -10 0 10 20 30 40 50 60 Real Axis FIGURE E7.16 Root locus for 1 + K(20−s) (s+1)(20+s) = 0. we expand (a + bs)/(c + ds) and match as many coefficients as possible. In this case, we determine that a = c = (2/T ) and also that b = −d = −1. In this case, with T = 0.1, we have e−T s = 20 − s −(s − 20) = . 20 + s (s + 20) So, the characteristic equation is 1+ −K(s − 20) , (s + 1)(s + 20) and the root locus is shown in Figure E7.16. Using a Routh-Hurwitz © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 292 CHAPTER 7 The Root Locus Method analysis with the characteristic polynomial s2 + (21 − K)s + 20 + 20K = 0 , we determine that the system is stable for −1 < K < 21. (a) The root locus is in Figure E7.17a. 2 1.5 1 0.5 Imag Axis E7.17 0 x x -0.5 -1 -1.5 -2 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 Real Axis FIGURE E7.17 (a) Root locus for 1 + K s(s−1) = 0. The root locus is always in the right half-plane; the system is unstable for K > 0. (b) The characteristic equation is 1+ K(s + 2) =0, s(s − 1)(s + 20) and the root locus is shown in Figure E7.17b. The system is stable for K > 22.3 and when K = 22.3, the roots are s1,2 = ±j1.53 and s3 = −19 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 293 Exercises 10 8 6 4 Imag Axis 2 * 0 x o K=22.3 x x -2 -4 -6 -8 -10 -30 -25 -20 -15 -10 -5 0 5 10 Real Axis FIGURE E7.17 CONTINUED: (b) Root locus for 1 + = 0. The root locus is shown in Figure E7.18. 6 4 2 Imag Axis E7.18 K(s+2) s(s+20)(s−1) x + 0 x + K=8.15 x + x + -2 -4 -6 -6 -4 -2 0 2 4 6 Real Axis FIGURE E7.18 Root locus for 1 + K s(s+3)(s2 +2s+2) = 0. When K = 8.15, the roots are s1,2 = ±j1.095 and s3,4 = −2.5 ± j0.74. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 294 CHAPTER 7 E7.19 The Root Locus Method The characteristic equation is 1+ K =0, + 6s + 64) 3)(s2 s(s + and the root locus is shown in Figure E7.19. When K = 1292.5, the roots are s1,2 = ±j4.62 and s3,4 = −4.49 ± j6.36 . 15 10 x + Imag Axis 5 + 0 x K=1292.5 x + -5 + x -10 -15 -15 -10 -5 0 5 10 15 Real Axis FIGURE E7.19 Root locus for 1 + E7.20 K s(s+3)(s2 +6s+64) = 0. The characteristic equation is 1+ K(s + 1) =0, s(s − 2)(s + 6) and the root locus is shown in Figure E7.20. The system is stable for K > 16 . The maximum damping ratio of the stable complex roots is ζ = 0.25 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 295 Exercises 20 15 10 Imaginary Axis 5 ζmax = 0.25 0            0 1 2 Real Axis FIGURE E7.20 Root locus for 1 + = 0. The gain is K = 10.8 when the complex roots have ζ = 0.66. 10 5 K=10.8 Imag Axis E7.21 K(s+1) s(s−2)(s+6) + 0 x + x o + x -5 -10 -10 -5 0 Real Axis FIGURE E7.21 Root locus for 1 + Ks s3 +5s2 +10 = 0. 5 10 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 296 CHAPTER 7 E7.22 The Root Locus Method The root locus is shown in Figure E7.22. The characteristic equation is 1+ K(s2 + 18)(s + 2) =0. (s2 − 2)(s + 12) Root Locus 5 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −5 −14 FIGURE E7.22 Root locus for 1 + E7.23 −12 −10 K(s2 +18)(s+2) (s2 −2)(s+12) −8 −6 Real Axis −4 −2 0 = 0. The characteristic equation is 5s2 + as + 4 = 0 , which can rewritten as 1+ as =0. +4 5s2 The roots locus (with a as the parameter) is shown in Figure E7.23. 2 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 297 Exercises 1.5 1 x Imag Axis 0.5 0 o -0.5 x -1 -1.5 -1.5 -1 -0.5 0 0.5 1 1.5 Real Axis FIGURE E7.23 Root locus for 1 + E7.24 as 5s2 +4 = 0. The transfer function is G(s) = C(sI − A)−1 B + D  = [ 1 0 ] = s2 s −1 4 s+k 1 . + ks + 4 −1    0 1   Therefore, the characteristic equation is s2 + ks + 4 = 0 , or 1+k s2 s =0. +4 The root locus for 0 < k < ∞ is shown in Figure E7.24. The closed-loop system is stable for all 0 < k < ∞. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 298 CHAPTER 7 The Root Locus Method 2.5 2 1.5 Imaginary Axis 1 0.5 0  5   5   5   5   5   5   5 0 0.5 Real Axis FIGURE E7.24 Root locus for 1 + k s2s+4 = 0. The characteristic equation is 1+K 10 =0. s(s + 25) The root locus shown in Figure E7.25 is stable for all 0 < K < ∞. 15 10 5 Imaginary Axis E7.25 0        Real Axis FIGURE E7.25 10 Root locus for 1 + K s(s+25) = 0.   0 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 299 Exercises E7.26 The characteristic polynomial is  det  or s −1 s+K −3 s+K +2 1+K  =0 s+1 =0. s2 + 2s − 3 The root locus shown in Figure E7.26 is stable for all 0 < K < 3. Root Locus 0.8 0.6 Imaginary Axis 0.4 0.2 0 −0.2 −0.4 −0.6 −0.8 −12 −10 −8 −6 −4 Real Axis −2 0 2 FIGURE E7.26 s+1 Root locus for 1 + K s2 +2s−3 = 0. E7.27 The characteristic equation is 1+p s2 s =0. + 4s + 40 The root locus shown in Figure E7.27 is stable for all 0 < p < ∞. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 300 CHAPTER 7 The Root Locus Method 8 6 Imaginary Axis 4 2 0           0 2 Real Axis FIGURE E7.27 s = 0. Root locus for 1 + p s2 +4s+40 The characteristic equation is 1+K s(s2 s−1 =0. + 2s + 2) The system is stable for −1.33 < K < 0. 1.5 1 0.5 Imaginary Axis E7.28 0 #5 #5 ! " Real Axis FIGURE E7.28 Root locus for 1 + K s(s2s−1 = 0. +2s+2) 0 2 4 6 8 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 301 Problems Problems P7.1 Root Locus 30 20 Imaginary Axis 10 0 −10 −20 −30 −50 −40 −30 −20 −10 Real Axis 0 10 20 Root Locus 5 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −5 −7 FIGURE P7.1 (a) Root locus for 1 + −6 −5 K s(s+10)(s+8) −4 −3 −2 Real Axis = 0, and (b) 1 + −1 0 1 K (s2 +2s+2)(s+1) 2 = 0. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 302 CHAPTER 7 The Root Locus Method Root Locus 40 30 Imaginary Axis 20 10 0 −10 −20 −30 −40 −12 −10 −8 −6 −4 Real Axis −2 0 2 Root Locus 4 3 Imaginary Axis 2 1 0 −1 −2 −3 −4 −4 −3.5 −3 −2.5 FIGURE P7.1 CONTINUED: (c) Root locus for 1 + P7.2 −2 −1.5 Real Axis K(s+5) s(s+2)(s+7) −1 −0.5 = 0, and (d)1 + 0 0.5 K(s2 +4s+8) s2 (s+7) = 0. The root locus is shown in Figure P7.2 for the characteristic equation 1+ 10Kv (s + 10) =0. s(s + 1)(s + 100) The damping ratio is ζ = 0.6 when Kv = 0.8, 135 and 648. The roots of the characteristic equation are: (a) Kv = 0.8 : s1 = −99.9, s2,3 = −0.54 ± j0.71 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 303 Problems (b) Kv = 135 : s1 = −85.9, s2,3 = −7.5 ± j10 (c) Kv = 648 : s1 = −11.7, s2,3 = −44.6 ± j59.5 30 20 Imag Axis 10 0 x o xx -10 -20 -30 -100 -80 -60 -40 -20 Real Axis FIGURE P7.2 Root locus for 1 + P7.3 10Kv (s+10) s(s+1)(s+100) = 0. (a) The breakaway point is s = −0.88 at K = 4.06. (b) The characteristic equation can be written as s(s + 2)(s + 5) + K = 0 . The Routh array is s3 1 10 s2 7 K s1 b 0 so K where b= 70 − K . 7 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 304 CHAPTER 7 The Root Locus Method √ When K = 70, the system has roots on jω-axis at s = ±j 10. (c) When K = 6, the roots are s1,2 = −0.83 ± j0.66, s3 = −5.34. (d) The characteristic equation 1+ K =0 s(s + 2)(s + 5) has the root locus shown in Figure P7.3. 10 8 6 4 Imag Axis 2 0 -2 -4 -6 -8 -10 -10 -8 FIGURE P7.3 Root locus for 1 + P7.4 -6 -4 K s(s+2)(s+5) -2 0 Real Axis 2 4 6 8 10 = 0. The characteristic equation for the large antenna is 1 + G1 G(s) = 1 + 100ka =0, (0.1s + 1)(s2 + 14.4s + 100) or 1+ 1000ka =0. (s + 10)(s2 + 14.4s + 100) The root locus is shown in Figure P7.4. Using Routh’s criteria, we find that the system is stable for −1 < ka < 4.83 . © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 305 Problems 20 * 15 <-- K=4.827 10 x Imag Axis 5 0 x -5 x -10 -15 * -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis FIGURE P7.4 Root locus for 1 + 1000ka (s2 +14.14s+100)(s+10) = 0. When ka = 4.83, we have s1,2 = ±j15.53. P7.5 (a) The characteristic equation for hands-off control is 1+ 25K2 (s + 0.03)(s + 1) =0. (s + 0.4)(s2 − 0.36s + 0.16)(s + 9) The root locus is shown in Figure P7.5a. The damping ratio is ζ = 0.707 when K2 = 1.6 or K2 = 0.74. (b) The transfer function from Td (s) to Y (s) is Y (s) = G2 (s)Td (s) , 1 + G2 (s)Gf (s) where Gf (s) = K2 (s + 1) . s+9 Using the final value theorem, we determine that yss = lim s s→0 G2 (s) 1 11.7   = 3.8 , = 1 + G2 (s)Gf (s) s 1 + 11.7 K92 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method 5 4 <-- K=1.6 3 2 <-- K=0.74 Imag Axis 1 x 0 x o x o x -1 -2 -3 -4 -5 -10 -8 -6 -4 -2 0 2 4 Real Axis FIGURE P7.5 (a) Root locus for 1 + 25K2 (s+0.03)(s+1) (s+9)(s2 −0.36s+0.16)(s+0.4) = 0. 20 15 10 5 Imag Axis 306 x 0 x o x xo x -5 -10 -15 -20 -20 -15 -10 -5 0 5 10 15 20 Real Axis FIGURE P7.5 CONTINUED: (b) Root locus for 1 + 25K1 (s+0.03)(s+9) (s+0.045)(s2 +12s+1)(s+1.33)(s2 +7.66s+29.78) = 0. where we have selected K2 = 1.6. For K2 = 0.74, we find that yss = 5.96. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 307 Problems (c) The closed-loop characteristic equation with the pilot loop added is 1+ 25K1 (s + 0.03)(s + 9) =0. (s + 0.045)(s + 1.33)(s2 + 7.66s + 29.78)(s2 + 12s + 1) The root locus is shown in Figure P7.5b. (d) Using K1 = 2, we determine that ess = 0.44 . (a) The characteristic equation is 1+ K(s + 0.20)(s2 + 4s + 6.25) =0. (s + 0.9)(s − 0.6)(s − 0.1)(s + 4) The root locus is shown in Figure P7.6. 4 zeta=0.5 zeta*wn=-1/3 3 2 o K=4 --> 1 Imag Axis P7.6 0 x x o x x -1 o -2 -3 -4 -6 -5 -4 -3 -2 -1 0 1 2 Real Axis FIGURE P7.6 Root locus for 1 + K(s+0.2)(s2 +4s+6.25) (s+0.9)(s−0.6)(s−0.1)(s+4) = 0. (b) For Ts < 12 sec, we require ζωn > 1/3. Also, we want ζ > 0.5. So, we seek roots for a stable system with ζωn > 1/3 and ζ > 0.5. This occurs when K > 4. © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 308 CHAPTER 7 (a) The characteristic equation for the speed control system is 1+ K =0, (s + 4)2 (s + δ) where K= 0.004 R and δ= 0.75 = 0.0001875 . 4000 The root locus is shown in Figure P7.7. At ζ = 0.6, we have K = 19.1, 6 4 2 <-- K=19.1 Imag Axis P7.7 The Root Locus Method 0 x x -2 -4 -6 -6 -5 -4 -3 -2 -1 0 1 2 3 Real Axis FIGURE P7.7 Root locus for 1 + K (s+4)2 (s+1.875e−04) = 0. therefore R = 0.00021 . When K = 19.1 the roots are s1,2 = −1.1 ± j1.43 and s3 = −5.80 . (b) The steady-state error is lim s∆ω(s) = lim s s→0 s→0 (0.25s + 1)2 ∆L(s) (0.25s + 1)2 (Js + b) + 1/R 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 309 Problems = 1 ∆L ≈ ∆LR , b + 1/R when R < 0.1. (a) The characteristic equation for the speed control system with the hydroturbine is 1+ K(−s + 1) =0, (s + 4)(s + 2)(s + δ) where K= 0.002 R and δ= 0.75 = 0.0001875 . 4000 The root locus is shown in Figure P7.8. At ζ = 0.6, we have K = 2.85, 2 1.5 1 K=2.85 --> 0.5 Imag Axis P7.8 0 x x x o 0 1 -0.5 -1 -1.5 -2 -6 -5 -4 -3 -2 -1 2 Real Axis FIGURE P7.8 Root locus for 1 + K(−s+1) (s+4)(s+2)(s+δ) = 0. therefore R = 0.0007 . When K = 2.85 the roots are −0.45 ± j0.60, and -5.1. 3 4 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 310 CHAPTER 7 The Root Locus Method (b) The steady-state error is lim s∆ω(s) = lim s s→0 s→0 = (0.25s + 1)(0.5s + 1) ∆L(s) (0.25s + 1)(0.5s + 1)(Js + f ) + (−s + 1)/R 1 ∆L ≈ ∆LR , f + 1/R when R < 0.1. The characteristic equation is 1+K (s + 0.5)(s + 0.1)(s2 + 2s + 289) =0 s(s + 30)2 (s − 0.4)(s + 0.8)(s2 + 1.45s + 361) where K = K1 K2 . The root locus is shown in Figure P7.9. When K = 4000 , the roots are s1,2 = −0.82 ± j19.4 50 40 30 20 10 Imag Axis P7.9 0 -10 -20 -30 -40 -50 -35 FIGURE P7.9 Root locus for 1 + -30 -25 -20 -15 Real Axis -10 K(s+0.5)(s+0.1)(s2 +2s+289) s(s+30)2 (s−0.4)(s+0.8)(s2 +1.45s+361) -5 = 0. 0 5 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 311 Problems s3 s4 s5 s6 s7 (a) The characteristic equation is 1+ K1 K2 (s + 2)2 =0. (s + 10)(s + 100)(s2 + 1.5s + 6.25) The root locus is shown in Figure P7.10. 10 8 6 4 x 2 Imag Axis P7.10 = −39.8 = −14.9 = −5.0 = −0.38 = −0.14 . 0 x x -2 o x -4 -6 -8 -10 -120 -100 -80 -60 -40 Real Axis FIGURE P7.10 Root locus for 1 + K1 K2 (s+2)2 (s+10)(s+100)(s2 +1.5s+6.25) = 0. (b) The gain K1 K2 = 1620 when ζ = 0.707. Therefore, K2 = 81000 , -20 0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 312 CHAPTER 7 The Root Locus Method since K1 = 0.02 at medium weight cruise condition. (c) At lightweight cruise condition K1 = 0.2 . Using K2 = 81000, we find the roots are s1,2 = −54 ± j119 s3,4 = −2 ± j0.6 . The roots s3,4 become negligible and the roots at s1,2 become highly oscillatory. Hence, in this case ζ = 0.41 . (a) The closed-loop characteristic equation is 1+ 20Ka (s2 + s + 0.02) =0, s(s + 1)2 (s2 + 2s + 0.8) where K2 = 10 . Then, the root locus is shown in Figure P7.11a. 3 2 1 Ka=0.035 --> Imag Axis P7.11 0 x xo x ox -1 -2 -3 -3 -2 -1 0 1 2 Real Axis FIGURE P7.11 20s2 +20s+0.4 (a) Root locus for 1 + Ka s(s+1) 2 (s2 +2s+0.8) = 0, where K2 = 10. 3 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 313 Problems (b) When Ka < 0.035 , all the roots have a damping greater than or equal to 0.60. (c) Select Ka = 0.035 . Then, the characteristic equation with K2 as the parameter is 1 + K2 0.07(s2 + s) =0. s5 + 4s4 + 5.8s3 + 3.6s2 + 0.8s + 0.014 The root locus is shown in Figure P7.11b. 3 Ka=0.035 2 Imag Axis 1 0 x x o x x xo -1 -2 -3 -3 -2 -1 0 1 2 3 Real Axis FIGURE P7.11 0.07s(s+1) CONTINUED: (b) Root locus for 1+K2 s(s+1)2 (s2 +2s+0.8)+0.014 = 0, where Ka = 0.035. P7.12 (a) The closed-loop transfer function is T (s) = 1.8s2 (s Ka Km (s + 25)(s + 15) . + 2) + Ka Km (s + 25)(s + 15) + 1.6Km s(s + 2) So, with E(s) = R(s) − Y (s), we have E(s) = (1 − T (s))R(s) and ess = lim sE(s) = 1 − T (0) = 0 . s→0 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. CHAPTER 7 The Root Locus Method Therefore, when the system is stable, it has zero steady-state error. (b) The characteristic equation is s3 + (3.6 + Ka )s2 + (3.2 + 40Ka )s + 375Ka . The Routh array is s3 1 3.2 + 40Ka s2 3.6 + Ka 375Ka s1 b so 375K Solving for b > 0 leads to 0 < Ka < 0.05 or Ka > 5.64 for stability. (c) The characteristic equation can be written as 1+ Ka (s + 25)(s + 15) =0. s(s + 2)(s + 1.6) The root locus is shown in Figure P7.12. (d) When K > 40 , 40 30 20 10 Imag Axis 314 0 o o xx x -10 -20 -30 -40 -70 -60 -50 -40 -30 -20 Real Axis FIGURE P7.12 (s+25)(s+15) Root locus for 1 + Ka s(s+2)(s+1.6) = 0, where Km = 1.8. -10 0 10 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 315 Problems the roots are s1 = −123 and s2,3 = −15.6 ± j31.2 . From the step response we find P.O. = 5% Tp = 0.67 sec Ts = 0.25 sec . (a) The characteristic equation is 1+ s(s + 3)(s2 K =0. + 4s + 7.84) The root locus is shown in Figure P7.13. The breakaway point is s = −1.09 at K = 9.72. (b) When K = 13.5, the roots are s1,2 = −0.84 ± j0.84 s3,4 = −2.66 ± j1.55 . 6 4 2 x + Imag Axis P7.13 + 0 x x + + -2 x -4 -6 -6 -4 -2 0 Real Axis FIGURE P7.13 Root locus for 1 + K s(s+3)(s2 +4s+7.84) = 0. 2 4 6 © 2011 Pearson Education, Inc., Upper Saddle River, NJ. All rights reserved. This publication is protected by Copyright and written permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458. 316 CHAPTER 7 The Root Locus Method (c) The roots s = −0.84 ± j0.84 are dominant roots. (d) For the dominant roots, we determine that ζ = 0.7 and ωn = 1.19. Therefore, the settling time is Ts = sec . The characteristic equation is 1+ K(s + 2.5)(s + 3.2) =0. + 1)(s + 10)(s + 30) s2 (s The root locus is shown in Figure P7.14. When K = 559.3, the roots are s1 = −30.75 s2 = −8.48 s3 = −1.78 s4,5 = ±j3.11 . s3 = −2.21 s4,5 = ±j10.23 . When K = 4321, the roots are s1 = −34.45 s2 = −4.35 The crossover points are s = ±j3.11 and s = ±j10.23 . 25 20 15 10 Imaginary Axis P7.14 4 = 4.8 ζωn 5 0$&
$)'$)&
$('$(&
$%&$%'

$(&$('

$)& Real Axis FIGURE P7.14 (s+2.5)(s+3.5) Root locus for 1 + K s2 (s+1)(s+10)(s+30) = 0.$)'

\$&

0

5

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

317

Problems

Therefore, the system is stable for
559.3 < K < 4321 .
The characteristic equation is
1+

K(s2 + 30s + 625)
.
s(s + 20)(s2 + 20s + 200)(s2 + 60s + 3400)

The root locus is shown in Figure P7.15. When K = 30000, the roots are
s1 = −18.5

s2 = −1.69

s3,4 = −9.8±j8.9

s5,6 = −30.1±j49.9.

The real root near the origin dominates, and the step response is overdamped.

100
80
60
x

40
20

Imag Axis

P7.15

o
x

0

x

x
x

-20

o

-40
x

-60
-80
-100
-100

-80

-60

-40

-20

0

20

40

Real Axis

FIGURE P7.15
s2 +30s+625
Root locus for 1 + K s(s+20)(s2 +20s+200)(s
2 +60s+3400) = 0.

60

80

100

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

318

CHAPTER 7

(a) Let τ = 0. Then, first reduce the motor and rolls to an equivalent
G(s) as follows:
G(s) =

1

0.25
s(s+1)
0.25
+ s(s+1)

=

0.25
0.25
=
.
s(s + 1) + 0.25
(s + 0.5)2

The loop transfer function is then
L(s) =

2(s + 0.5)Ka (0.25)
0.5Ka
=
.
2
2
s(s + 1) (s + 0.5)
s(s + 1)2 (s + 0.5)

The characteristic equation is
1 + Ka

0.5
=0.
s(s + 1)2 (s + 0.5)

The root locus is shown in Figure P7.16.

2
1.5
1
0.5

Imag Axis

P7.16

The Root Locus Method

+
+

0

x

x

x
+

+

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

1

1.5

Real Axis

FIGURE P7.16
Root locus for 1 +

0.5Ka
s(s+1)2 (s+0.5)

= 0.

(b) When K = 0.123, the roots of the characteristic equation are
s1,2 = −1.1 ± j0.27
s3,4 = −0.15 ± j0.15 .

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

319

Problems

The roots at s = −0.15 ± j0.15 have a damping ratio of ζ = 0.707.

(c) When τ becomes nonnegligible, the root locus will have an additional
pole, and the root locus will change accordingly.
The characteristic equation is
2
(M1 s2 + bs + k1 + k12 )(M2 s2 + k12 ) − k12
=0.
2 is negligible
If we let M1 = k1 = b = 1, and assume k12 < 1 so that k12
and k1 + k12 ≈ k1 , then the characteristic equation is

(s2 + s + 1)(M2 s2 + k12 ) = 0

or

1+

k
=0,
s2

where
k=

k12
.
M2

The root locus is shown in Figure P7.17. All the roots lie on the jω axis.
If we select
s

k12
= ωo ,
M2

then we cancel the vibration.

3

2

root locus -->

1

Imag Axis

P7.17

0

x

-1

-2

-3
-3

-2.5

-2

-1.5

-1
Real Axis

FIGURE P7.17
Root locus for 1 +

k
s2

= 0.

-0.5

0

0.5

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

320

CHAPTER 7

The characteristic equation is
βs3 + (1 + 2β)s2 + (2 + 4α)s + 4 = 0 .
When β = 0 we have
1+

s2

4αs
=0.
+ 2s + 4

The root locus for β = 0 is shown in Figure P7.18.

3
<-- zeta=0.6

beta=0

2
x
*

1

Imag Axis

P7.18

The Root Locus Method

0

o

-1
*
x

-2

-3
-3

-2

-1

0

1

Real Axis

FIGURE P7.18
4s
= 0, where β = 0.
Root locus for 1 + α s2 +2s+4

For α = 0.3, the poles are
s = −1.6 ± j1.2 .
Then, we have
1+

β(s + 2)s2
=0.
s2 + (2 + 4α)s + 4

When β = 0.121
s1,2 = −1.51 ± j1.51
s3 = −7.24 .

2

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

321

Problems

Thus,
ζ = 0.707

and ζωn = 1.5 .

So, the performance specs are met. Also,
Gc (s) =
P7.19

0.3s + 1
2.48(s + 3.33)
=
.
0.121s + 1
(s + 8.26)

The characteristic equation is
1+

Ka (s2 + 4s + 100)
=0.
s(s + 2)(s + 6)

The root locus is shown in Figure P7.19.
10
8
6

Imaginary Axis

4
2
0
−2
−4
−6
−8
−10
−10

−8

−6

−4
Real Axis

−2

0

FIGURE P7.19
s2 +4s+100
Root locus for 1 + Ka s(s+2)(s+6)
= 0.

When Ka = 0.094;, the roots are
s1,2 = −0.85 ± j0.85
s3 = −6.38 .
Thus, the complex roots have a damping ratio of ζ = 0.707.
P7.20

The characteristic equation is
s3 + (2 +

1 2
2
4
)s + ( + K)s + = 0 ,
β
β
β

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

322

CHAPTER 7

The Root Locus Method

where
K=

4α
β

α = 0.3

β = 0.121 .

The root sensitivity to changes in K is found to be
r1 ∼
SK
=

∆r1
= 1.186 −149.75o .
∆K/K

The root sensitivity to changes in the pole at s = −2 is found to be
r1 ∼ ∆r1
S∆
= 1.656 −137o ,
=
∆/2

(a) Let the pole be (s + 4 + ∆) and neglect ∆2 terms. Then, the characteristic equation is
1+∆

2s2 + (8 + 2δ)s + 8δ
=0
s3 + (8 + δ)s2 + (16 + 8δ)s + 16δ + K

where δ = 0.000788 and K = 19.1.
3

2
x

1

Imag Axis

P7.21

where the pole is s + 2 + ∆.

0

x

o

o

-1
x

-2

-3
-10

-8

-6

-4

-2

0

2

Real Axis

FIGURE P7.21
2s2 +(8+2δ)s+8δ
Root locus for 1 + ∆ s3 +(8+δ)s2 +(16+8δ)s+16δ+K = 0, (δ = 0.000788 and K = 19.1).

The root sensitivity is determined to be
r1 ∼ ∆r1
S∆
= 3.3146 −132o .
=
∆/4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

323

Problems

(b) Let R = Ro + ∆R, where R = 0.00021. Then,
r1 ∼ ∆r1
SR
= 1.31 6 −107o .
=
∆R/R

P7.22

The characteristic equation is
s3 + 2s2 + s + K ,
where K = 0.24 for ζ = 0.707. The root sensitivity to changes in the pole
at s = −1 is found to be
r1 ∼ ∆r1
S∆
= 0.956 −126o ,
=
∆

where the pole is s + 1 + ∆.
P7.23

The characteristic equation is
s3 + 5s2 + (6 + K)s + K ,
where K = 6.3 for ζ = 0.707. The root sensitivity to changes in the pole
at s = −2 is found to be
r1 ∼ ∆r1
S∆
= 1.256 −169.4o ,
=
∆/2

where the pole is s + 2 + ∆. The root sensitivity to changes in the zero at
s = −1 is found to be
r1 ∼ ∆r1
S∆
= 0.556 34.4o ,
=
∆

where the zero is s + 1 + ∆.
P7.24

The root locus for each of the four cases shown is shown in Figure P7.24.
The four open-loop transfer functions are
(a) KF (s) =

s+8
+
+ 296s3 + 1170s2 + 1575s
1
s2 + 6s + 6.75
KF (s) = 6
(d) KF (s) = 3
5
4
s + 2s + s
s + 5s2 + 4s

(b) KF (s) =
(c)

s2 + 7s + 8.25
s3 + 6s2 + 5s
s5

30s4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

324

CHAPTER 7

The Root Locus Method

(a)

5

(b)

10

0

ox

Imag Axis

Imag Axis

5
ox x

0

x

ox x x

x

-5
-5
-10

-5
Real Axis

-10
-20

0

(c)

2

5

-10
Real Axis

0

(d)

0

x

Imag Axis

Imag Axis

1
x

0

ox

ox x

-1
-2
-2

0

-5
-10

2

Real Axis

-5

0

Real Axis

FIGURE P7.24
Root locus for the four cases.

P7.25

The characteristic equation is
1 + KGc (s)G(s) = 0 ,
therefore,
KGc (s)G(s) = −1 .
Squaring both sides yields K 2 G2c (s)G2 (s) = 1 and
1 − K 2 G2c (s)G2 (s) = 0 .
The root locus with 0 < K 2 < ∞ is shown in Figure P7.25. The value of
K 2 for which the locus crosses the imaginary axis is
K 2 = 2/3 ,
p

therefore K = 2/3 = 0.8165 corresponds to the jω-axis crossing (at
s = 0). You can check that 1 + KGc (s)G(s) = 0 for K = 0.8165 and
s = 0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

325

Problems

3

2

Imag Axis

1

0

-1

-2

-3
-3

-2

-1

0
Real Axis

1

2

3

FIGURE P7.25
Root locus for the equation 1 − K 2 G2c (s)G2 (s) = 0.

P7.26

(a) The characteristic equation is
1+

K(s + 2)2
=0.
s(s2 + 1)(s + 8)

The root locus is shown in Figure P7.26.
(b) Using Routh’s criteria, we determine that
K > 14
for stability.
(c) From the Routh array, we determine that for K = 14, we have two
purely imaginary poles at
√
s = ±j 8 .
(d) When K > 50, the real part of the complex roots is approximately
equal to the real part of the two real roots and therefore the complex
roots are not dominant roots.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

326

CHAPTER 7

The Root Locus Method

15

10

Imag Axis

5

0

-5

-10

-15
-15

-10

-5

0
Real Axis

5

10

15

FIGURE P7.26
(s+2)2
Root locus for 1 + K s(s2 +1)(s+8) = 0.

P7.27

The characteristic equation is
1+

K(s2 + 0.1)
=0.
s(s2 + 2)

The root locus is shown in Figure P7.27a. The locus enters the axis at
s = −1.26
and leaves the axis at
s = −0.36 .
Define
p(s) = K =

−(s3 + 2s)
.
s2 + 0.1

Then, a plot of p(s) vs s is shown in Figure P7.27b, where it can be seen
that p(s) has two inflection points at
s = −1.28

and s = −0.36 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

327

Problems

Root Locus
2
1.5

Imaginary Axis

1
0.5
0
−0.5
−1
−1.5
−2
−3

−2.5

−2

−1.5
−1
Real Axis

−0.5

0

0.5

3.5

3

2.5

p(s)

2

1.5

1

0.5

0
−2

−1.8

−1.6

−1.4

−1.2

−1
s

−0.8

−0.6

−0.4

−0.2

0

FIGURE P7.27
s2 +0.1
s3 +2s
(a) Root locus for 1 + K s(s
2 +2) = 0. (b) Plot of p(s) = − s2 +0.1 versus s.

P7.28

The characteristic equation is
1 + L(s) = 1 +

K(s2 + 12s + 20)
=0.
s3 + 10s2 + 25s

The root locus is shown in Figure P7.28. The breakaway point is s = −5.0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

328

CHAPTER 7

The Root Locus Method

6

4

Imag Axis

2

0

-2

-4

-6
-20

-15

-10

-5

0

Real Axis

FIGURE P7.28
(s2 +12s+20)
Root locus for 1 + K s3 +10s2 +25s = 0.

and the entry point is s = −15.6. When K = 2, the roots are
s1 = −1.07
s2,3 = −5.46 ± j2.75 .
When K = 2, the roots are
s1 = −1.07
s2,3 = −4.36 ± j1.68 .
The predicted step response when K = 2 is Ts = 9 sec and P O ≈ 0%.
P7.29

The characteristic equation is
1+K

s2 + 10s + 30
=0.
s2 (s + 10)

The root locus is shown in Figure P7.29. When ζ = 0.707, the necessary
gain is K = 16. The corresponding roots are s1 = −18.87 and s2,3 =
−3.56 ± j3.56.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

329

Problems

Root Locus
4
System: sys
Gain: 16
Pole: −3.56 + 3.57i
Damping: 0.707
Overshoot (%): 4.34

3

Imaginary Axis

2
1
0
−1
−2
−3
−4
−20

−15

−10

−5

0

5

Real Axis

FIGURE P7.29
2
= 0.
Root locus for 1 + K ss+10s+30
2 (s+10)

P7.30

The transfer function is
Z(s) =

LCRs2 + Ls
Rs2 + s
=
.
LCs2 + CRs + 1
s2 + Rs + 1

So,
R
r1 = − +
2

!1

R2
−1
4

2

.

Thus, the nominal r1o = − 21 . Simultaneously,
R
r2 = − −
2

!1

R2
−1
4

2

.

Thus, the nominal r2o = −2. We see that there is a difference by a factor
of 4. Also,
ri
SR

∂r1
=
∂R

Ro

Ro Ro2
· Ro = −
+
2
4

!− 1

Ro2
−1
4

2

=

5
,
6

where Ro = 2.5. And
r2
SR

∂r2
=
∂R

Ro Ro2
Ro = −
−
2
4
Ro

!− 1

Ro2
−1
4

2

=

−10
.
3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

330

CHAPTER 7

The Root Locus Method

r2
r1
So, the magnitude of |SR
| = 4|SR
|.

P7.31

The characteristic equation is
1+K

s(s +

0.16)(s2

s+4
=0.
+ 14.6s + 148.999)

The root locus is shown in Figure P7.31. When K = 1350, the roots are

20
15
10

(+) K=326 -->

x

*

+

<-- K=1350 (*)

Imag Axis

5
+
*

0

o
*

xx
+

-5
+
x

-10

*

-15
-20
-20

-15

-10

-5

0

5

10

Real Axis

FIGURE P7.31
s+4
Root locus for 1 + K s(s+0.16)(s2 +14.6s+148.999)
= 0.

s1,2 = ±j9.6
s3,4 = −7.4 ± j1.9 .
When K = 326, the roots are
s1,2 = −6.5 ± j8.7
P7.32

s3,4 = −0.9 ± j3.2 .

The characteristic equation is
1+

K(s + 1)(s + 5)
=0.
s(s + 1.5)(s + 2)

15

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

331

Problems

4
3

*

2

*

Imag Axis

1
0

o

x

x

o**

x

-1
-2

*

-3

*

-4
-10

-8

-6

-4

-2

0

2

4

Real Axis

FIGURE P7.32
(s+1)(s+5)
Root locus for 1 + K s(s+1.5)(s+2) = 0.

K

TABLE P7.32

ζ

Ts (sec)

P.O. (%)

1.57

0.707

0.98

1.4

3.48

0.707

1.1

5.8

2.35

0.69

1.3

4.0

Step Response Results for K = 1.57, K = 3.48, and K = 2.35.

(a) The breakaway point is s = −1.73; the entry point is s = −8.62.

(b) The damping ratio ζ = 0.707 when K = 1.57 and again when K =
3.46.
(c) The minimum damping ratio ζ = 0.69 is achieved when K = 2.35.
(d) The results are summarized in Table P7.32.The best choice of gain is
K = 1.57.
P7.33

(a) The root locus for the V-22 is shown in Figure P7.33a. The system is
stable when 0 < K < 0.48 and K > 136.5.
(b) The unit step input response (for K = 280) is shown in Figure P7.33b.
The step response has a P.O. = 90% and Ts ≈ 50 sec.

(c) The plot of y(t) for a unit step disturbance is shown in Figure P7.33b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

2
1.5
1

Imag Axis

0.5
0

x

o

o

-1

-0.5

xxx

-0.5
-1
-1.5
-2
-3

-2.5

-2

-1.5

0

0.5

1

Real Axis

(i) Unit step input response
2
y(t) w/o prefilter .... (dotted line)
y(t) with prefilter ____ (solid line)

y(t)

1.5
1
0.5
0

4

0
x 10

10

20

10

20

-3

30

40
50
Time (sec)
(ii) Unit step disturbance response

60

70

80

60

70

80

3
y(t)

332

2
1
0
-1

0

30

40
Time (sec)

50

FIGURE P7.33
s2 +1.5s+0.5
(a) Root locus for 1 + K s(20s+1)(10s+1)(0.5s+1)
= 0. (b) (i) Unit step input response with
and without prefilter; (ii) Unit step disturbance response.

The response to the disturbance is oscillatory, but the maximum value
of oscillation is about 0.003; so it is negligible.
(d) The effect of adding a prefilter can be seen in Figure P7.33b. With

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

333

Problems

the prefilter we find P O = 7% and Ts ≈ 40 sec.
P7.34

The characteristic equation is
1+

K(s + 2)
=0.
(s + 1)(s + 2.5)(s + 4)(s + 10)

The root locus is shown in Figure P7.34a. The roots, predicted and actual percent overshoot for K = 400, 500, and 600 are summarized in
Table P7.34. The actual unit step input responses are shown in Figure P7.34b.

roots

ζ

predicted P.O. (%)

actual P.O. (%)

400

-13.5,-1.00 ± 5.71j,-1.98

0.173

57.6

51.6

500

-14.0,-0.75 ± 6.24j,-1.98

0.120

68.4

61.2

600

-14.4,-0.53 ± 6.71j,-1.98

0.079

77.9

69.6

TABLE P7.34

Summary for K = 400, 500, 600.

Root Locus
20

15

10

Imaginary Axis

K

5

0

−5

−10

−15

−20
−30

−25

−20

−15

−10
Real Axis

−5

FIGURE P7.34
s+2
(a) Root locus for 1 + K (s+1)(s+2.5)(s+4)(s+10)
= 0.

0

5

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

334

CHAPTER 7

The Root Locus Method

1.6

K=400 .... (dotted line)

1.4

K=500 −−− (dashed line)
K=600 ___ (solid line)

1.2

y(t)

1

0.8

0.6

0.4

0.2

0

0

2

4

6

8

10
Time (sec)

12

14

16

18

20

FIGURE P7.34
CONTINUED (b) Unit step input responses for K = 400, 500, 600.

(a) The root locus is shown in Figure P7.35 for the characteristic equation
1+

K(s + 1)2
=0.
s(s2 + 1)

3

K=4.52

2

*

1

Imag Axis

P7.35

x

0

o

-1

x

x

-2

-3
-5

*

*

-4

-3

-2
Real Axis

FIGURE P7.35
(s+1)2
Root locus for 1 + K s(s2 +1) = 0.

-1

0

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

335

Problems

(b) When K = 4.52, the roots are
s1 = −0.58
s2,3 = −1.96 ± j1.96 .
The complex roots have ζ = 0.707.
(c) The entry point is s = −3.38 when K = 7.41.

(d) The predicted P.O. = 4.5%
The characteristic equation is
1+

K(s + 1)(s + 2)(s + 3)
=0.
s3 (s − 1)

(a) The root locus is shown in Figure P7.36.

8
6
4
2

Imag Axis

P7.36

(ζ = 0.707) and the actual P.O. = 17%.

0

o

o

o

x

x

-2
-4
-6
-8
-10

-8

-6

-4

-2

Real Axis

FIGURE P7.36
(s+1)(s+2)(s+3)
Root locus for 1 + K
= 0.
s3 (s−1)

(b) When K = 2.96, the roots are
s1,2 = ±j4.08
s3,4 = −0.98 ± j0.33 .

0

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

336

CHAPTER 7

The Root Locus Method

(c) When K = 20, the roots are
s1 = −1.46
s2 = −1.07
s3,4 = −8.23 ± j2.99 .
When K = 100, the roots are
s1
s2
s3
s4

= −92.65
= −3.51
= −1.82
= −1.01 .

(d) When K = 20, the damping ratio is ζ = 0.94. Therefore, the predicted
P.O. = 0.02%. The actual overshoot is P.O. = 23%.
P7.37

Since we know that ess = 0 for a step input, we know that a = 0 or b = 0.
Select a = 0. Also, ωn = 2π/T = 20 rad/sec. The desired characteristic
polynomial is
(s + r1 )(s + j20)(s − j20) = s3 + r1 s2 + 400s + 400r1 = 0 .
The actual characteristic polynomial is
1+

2K
=0,
s(s + b)(s + 40)

s3 + (40 + b)s2 + 40bs + 2K = 0 .

or

Comparing the coefficients in the desired and actual characteristic polynomials, we determine that b = 10, r1 = 50, and K = 10000.
P7.38

(a) The characteristic equation is
1+

K(s + 1)
=0.
s(s − 3)

√
The system is stable for K > 3. When K = 3, the roots are s = ±j 3.

(b) The root locus is shown in Figure P7.38a.
(c) When K = 10 , the roots are
s1 = −2
s2 = −5 .

Since both roots are real and stable, we expect that there will be
zero overshoot. The actual response has a 40% overshoot, as seen in
Figure P7.38b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

337

Problems

6

4

Imag Axis

2

0

o

x

x

-2

-4

-6
-6

-4

-2

0

2

4

6

2

2.5

3

Real Axis
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5
Time (secs)

FIGURE P7.38
s+1
(a) Root locus for 1 + K s(s−3)
= 0. (b) Unit step response.

P7.39

The loop transfer function is
Gc (s)G(s) =

22K
.
(s + 1)(s2 + 8s + 22)

When K = 0.529, the closed-loop poles are s1,2 = −3.34 ± 1.83j and
s3 = −2.32 and have the maximum damping ζ = 0.877. The root locus is
shown in Figure P7.39a. The step response is shown in Figure P7.39b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

Root Locus
10
8
6

Imaginary Axis

4
2
0
−2
−4
−6
−8
−10
−14

−12

−10

−8

−6
Real Axis

−4

−2

0

2

Step Response
0.35

0.3

0.25

Amplitude

338

0.2

0.15

0.1

0.05

0

0

FIGURE P7.39
(a) Root locus for

0.5

1

22K
(s+1)(s2 +8s+22)

1.5
2
Time (sec)

2.5

= 0. (b) Unit step response.

3

3.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

339

The characteristic equation is
1+K

s+6
=0.
s(s + 4)(s2 + 4s + 8)

The root locus is shown in Figure AP7.1. The gain at maximum ζ is

10

5

Imag Axis

AP7.1

x
+

0

o

x+

+

x
+

x

-5

-10
-10

-5

0

5

10

Real Axis

FIGURE AP7.1
s(s+4)
Root locus for 1 + K s2 +2s+2 = 0.

K = 3.7 .
The roots at K = 3.7 are
s1 = −3.6424

s2,3 = −1.3395 ± +1.3553j

s4 = −1.6786 .

Using Figure 5.13 in Dorf & Bishop, the predicted percent overshoot and
settling time are
P.O. = 5%

and Ts = 3 sec ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

340

CHAPTER 7

The Root Locus Method

since ζ = 0.7 and
a
6
=
= 4.5 .
ωn ζ
1.9(0.7)
The actual percent overshoot and settling time are P.O. = 1% and Ts =
2.8 sec.
The characteristic equation is
1+K

(s + 1)(s + 4)
=0.
s(s − 1)(s + 5)(s + 10)

The root locus is shown in Figure AP7.2a. The selected gain is K = 43.7.
15

Imaginary Axis

10

5

0

−5

−10

−15
−12

−10

−8

−6
−4
Real Axis

−2

0

2

1.5
System: syscl
Peak amplitude: 1.48
Overshoot (%): 48.3
At time (sec): 0.857
System: syscl
Settling Time (sec): 2.31

1
Amplitude

AP7.2

0.5

0

0

0.5

1

1.5

2
2.5
Time (sec)

3

3.5

4

4.5

FIGURE AP7.2
(s+1)(s+4)
(a) Root locus for 1 + K s(s−1)(s+5)(s+10) = 0; (b) Step response for K = 43.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

341

The actual percent overshoot (see Figure AP7.2b) is P.O. = 48.3%.
AP7.3

The characteristic equation (with p as the parameter) is
1+p

s3

s(s + 1)
=0.
+ s2 + 10

The root locus is shown in Figure AP7.3.

5
4
3
2

x

Imag Axis

1
+

0

x

o

o
+

-1
x

-2
-3
-4
-5
-5

-4

-3

-2

-1

0

1

2

3

4

5

Real Axis

FIGURE AP7.3
s(s+1)
Root locus for 1 + p s3 +s2 +10 = 0.

When p = 21 the dominant roots have a damping ratio of ζ = 0.707.
AP7.4

The characteristic equation (with α as the parameter) is
1+α

s(s + 1)
=0.
+ s2 + 1

s3

The root locus is shown in Figure AP7.4a. The steady-state error is
1
=1−α .
s→0 1 + G(s)

ess = lim sE(s) = lim
s→0

To meet the steady-state error specification, we require
0.9 < α < 1.1 .
The step responses for α = 0.9, 1 and 1.1 are shown in Figure AP7.4b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

342

CHAPTER 7

The Root Locus Method

3

2

Imag Axis

1

x

0

x

o

o

x

-1

-2

-3
-3

-2

-1

0

1

2

3

Real Axis
alpha=0.9 (solid); alpha=1.0 (dashed); alpha=1.1 (dotted)
1.8
1.6
1.4

Amplitude

1.2
1
0.8
0.6
0.4
0.2
0
0

5

10

15

20

25

30

35

40

45

50

Time (sec)

FIGURE AP7.4
s(s+1)
(a) Root locus for 1 + p s3 +s2 +10 = 0. (b) Step responses for α = 0.9, 1 and 1.1.

AP7.5

The root locus is shown in Figure AP7.5. When K = 20.45, ζ = 0.707. The
r1 ∼
root sensitivity is SK
= ∆r1 /(∆K/20.45) = 3.156 87.76o . When K = 88,
the complex roots lie on the jω-axis—a 330% increase in the gain.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

343

5
4
3
2

Imag Axis

1
+

0

+x

x

x
+

-1
-2
-3
-4
-5
-15

-10

-5

0

5

Real Axis

FIGURE AP7.5
Root locus for 1 + K s3 +10s21+7s−18 = 0.

A gain of K = 13 provides an acceptable response of Ts < 1 and P.O. <
7.5%. The root locus is shown in Figure AP7.6.

Root Locus
2.5
2
1.5
1
Imaginary Axis

AP7.6

0.5
0
−0.5
−1
−1.5
−2
−2.5
−3

−2.5

−2

FIGURE AP7.6
s2 +3s+6
Root locus for 1 + K s3 +2s
2 +3s+1 = 0.

−1.5
−1
Real Axis

−0.5

0

0.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

344

CHAPTER 7

AP7.7

The Root Locus Method

The root locus for the positive feedback system is shown in Figure AP7.7.
15

10

Imag Axis

5

0

x

x

-5

-10

-15
-15

-10

-5

0

5

10

15

Real Axis

FIGURE AP7.7
−1
Root locus for 1 + K s2 +12s+32
= 0.

The root locus is shown in Figure AP7.8a. When k = 0.448, all the roots

30

20

10

Imag Axis

AP7.8

x

0

x

o
x

-10

-20

-30
-30

-20

-10

0
Real Axis

FIGURE AP7.8
(a) Root locus for 1 + k s3 +19s120s
2 +34s+120 = 0.

10

20

30

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

345

of the characteristic equation are real—the step response is shown in
Figure AP7.8b.
1
0.9
0.8
0.7

Amplitude

0.6
0.5
0.4
0.3
0.2
0.1
0
0

0.5

1

1.5

2

2.5

3

Time (secs)

FIGURE AP7.8
CONTINUED (b) Step response with k = 0.448.

The root locus for each controller is shown in Figure AP7.9.

AP7.9

(a)

(b)
5
Imaginary Axis

Imaginary Axis

5

0

−5
−15

−10

−5
Real Axis

0

0

−5
−15

5

−10

(c)

0

5

0

5

(d)

15

5

10

Imaginary Axis

Imaginary Axis

−5
Real Axis

5
0
−5

0

−10
−15
−15

−10

−5
Real Axis

FIGURE AP7.9
Root locus for the various controllers.

0

5

−5
−15

−10

−5
Real Axis

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

346

CHAPTER 7

AP7.10

The Root Locus Method

The characteristic equation (with K as the parameter) is
1+K

s2 + 7s + 20
=0.
s(s2 + 7s + 10)

The root locus is shown in Figure AP7.10. The steady-state value of the
10
8
6
4

Imag Axis

2
0
-2
-4
-6
-8
-10
-10

-8

-6

-4

-2

0
Real Axis

2

4

6

8

10

FIGURE AP7.10
s2 +7s+20
Root locus for 1 + K s(s
2 +7s+10) = 0.

step response for any K is 0.5. With K = 15 the closed-loop transfer
function is
T (s) =

10s + 150
.
s3 + 22s2 + 115s + 300

The step response has the following characteristics:
P.O. = 4.8%
AP7.11

and

Ts = 2 seconds .

The root locus is shown in Figure AP7.11a. A suitable gain is
K = 500.
The step response is shown in Figure AP7.11b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

347

Root Locus
80
60

Imaginary Axis

40
20
0
−20
−40
−60
−80
−100

−80

−60

−40
−20
Real Axis

0

20

40

FIGURE AP7.11
(s+2)2
(a) Root locus for 1 + K s(s+10)(s+20)(s2 +3s+3.5) = 0.

Step Response
1.4
System: sys_cl
Peak amplitude: 1.09
Overshoot (%): 9.01
At time (sec): 0.945

1.2

Amplitude

1
System: sys_cl
Settling Time (sec): 2.39

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2
Time (sec)

2.5

3

3.5

4

FIGURE AP7.11
CONTINUED: (b) Step response with K = 500.

AP7.12

The root locus is shown in Figure AP7.12a. The PI controller can be
written as
Gc (s) =

Kp s + KI
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method
8

6

4

Imag Axis

2

0

-2

-4

-6

-8
-7

-6

-5

-4

-3

-2

-1

0

1

2

Real Axis

Step Response
From: U(1)
1.4

1.2

1

0.8
To: Y(1)

Amplitude

348

0.6

0.4

0.2

0

0

5

10

15

Time (sec.)

FIGURE AP7.12
(s+0.2)
(a) Root locus for 1 + Kp s(s2 +7s+10) = 0. (b) Step response with Kp = 5.54.

and setting KI = 0.2Kp , the characteristic equation can be written as
1 + Kp

(s + 0.2)
=0
+ 7s + 10)

s(s2

A suitable gain is Kp = 5.55. The step response is shown in Figure AP7.12b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

349

AP7.13

The characteristic equation is
1 + K1 K2

1
= 0.
(s + 5)(s − 1)

The root locus is shown in Figure AP7.12a. The fastest expected settling

Root Locus
4

3

Imaginary Axis

2

1

0

−1

−2

−3

−4
−6

−5

−4

−3

−2
Real Axis

−1

0

1

2

FIGURE AP7.13
1
Root locus for 1 + K1 K2 (s+5)(s−1)
= 0.

time is Ts = 4/ωn ζ = 2 seconds since maximum |ωn ζ| = 2.
AP7.14

The root locus of the uncompensated transfer function is shown in Figure AP7.14a. It can be seen that the system is unstable for Ku = 131.25
with a period of Tu = 0.72, as illustrated in FigureAP7.14b. Using the
Ziegler-Nichols design formulas yields KP = 0.6Ku = 78.75, KI = 1.2Ku /Tu =
218.75, and KD = 0.6Ku Tu = 7.0875 where

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

Root Locus
30
System: sysg
Gain: 131
Pole: 0.0153 + 8.66i
Damping: −0.00176
Overshoot (%): 101

Imaginary Axis

20

10

0

−10

−20

−30
−40

−30

−20

−10
Real Axis

0

10

20

8

10

12

FIGURE AP7.14
10
= 0.
(a) Root locus for 1 + Ku s(s+10)(s+7.5)

Step Response
2
1.8
1.6
1.4
Amplitude

350

1.2
1
0.8
0.6
0.4
0.2
0

0

2

4

6
Time (sec)

FIGURE AP7.14
CONTIUED: (b) Step response at the ultimate gain Ku = 131.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

351

Step Response
1.6
System: sys_cl
Peak amplitude: 1.6
Overshoot (%): 59.5
At time (sec): 0.445

1.4

Amplitude

1.2
1
System: sys_cl
Settling Time (sec): 2.1

0.8
0.6
0.4
0.2
0

0

0.5

1

1.5
2
Time (sec)

2.5

3

3.5

FIGURE AP7.14
CONTINUED: (c) Step response with the Ziegler-Nichols tuned PID controller.

Step Response

−3

14

x 10

12
10

Amplitude

8
6
4
2
0
−2
−4

0

0.5

1

1.5
2
Time (sec)

2.5

3

3.5

FIGURE AP7.14
CONTINUED: (d) Disturbance response with the Ziegler-Nichols tuned PID controller.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

352

CHAPTER 7

The Root Locus Method

Design Problems
CDP7.1

The closed-loop transfer function from the input to the output is
26.035Ka
θ(s)
= 2
,
R(s)
s + (33.1415 + 26.035Ka K1 )s + 26.035Ka
where we consider for the first time the tachometer feedback (see Figure
CDP4.1 in Dorf and Bishop). The characteristic equation is
1 + K1

26.035Ka s
=0.
s2 + 33.1415s + 26.035Ka

The root locus is shown below. In accordance with the discussion in Chap30

20

Imag Axis

10

0

-10

-20

-30
-30

-20

-10

0
Real Axis

10

20

30

ter 5, we continue to use Ka = 22. This allows us to meet the overshoot
specification (P.O. < 5%) without the tachometer feedback and to provides good steady-state tracking errors to a step input. To meet the design
specifications of both P.O. and Ts we want the closed-loop poles to the
left of −ζω = −4/0.3 = −13.33 and ζ > 0.69. A reasonable selection is
K1 = 0.012. This places the closed-loop poles at s = −20 ± j13.
DP7.1

(a) The characteristic equation is
1+

(s2

18K(s + 0.015)(s + 0.45)
=0.
+ 1.2s + 12)(s2 + 0.01s + 0.0025)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

353

Design Problems

Since we want a negative feedback system, we have Gc (s) = −K.
When ωn > 2 and ζ = 0.15, the gain K = 0.12. The root locus is
shown in Figure DP7.1a.

6

4
x

Imag Axis

2

o

oxx

-0.5

0

0

-2
x

-4

-6
-4

-3.5

-3

-2.5

-2

-1.5

-1

0.5

1

Real Axis

FIGURE DP7.1
18(s+0.015)(s+0.45)
(a) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0.

(b) The unit step response is shown in Figure DP7.1b. The percent overshoot is
P.O. = 100% .

(c) The characteristic equation with the anticipatory controller is
1+

18K(s + 2)(s + 0.015)(s + 0.45)
=0.
(s2 + 1.2s + 12)(s2 + 0.01s + 0.002s)

The root locus is shown in Figure DP7.1c. If we select
K = 9.2/18 ,
then the complex roots have a damping ζ = 0.90. The roots are at
s1 = −0.253
s2 = −0.019
s3,4 = −5.07 ± j2.50 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

0.7

0.6

Amplitude

0.5

0.4

0.3

0.2

0.1

0
0

20

40

60

80

100

120

140

160

180

200

Time (secs)

FIGURE DP7.1
CONTINUED: (b) Unit step response for gain controller.

6

4
x

2

Imag Axis

354

0

xx
o o

o

-2
x

-4

-6
-6

-4

-2

0

2

4

6

Real Axis

FIGURE DP7.1
18(s+2)(s+0.015)(s+0.45)
CONTINUED: (c) Root locus for 1 + K (s2 +1.2s+12)(s2 +0.01s+0.0025) = 0.

(d) The unit step response for the system with the anticipatory controller
is shown in Figure DP7.1d.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

355

Design Problems

1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

20

40

60

80

100

120

140

160

180

200

Time (secs)

FIGURE DP7.1
CONTINUED: (d) Unit step response for anticipatory controller.

DP7.2

The characteristic equation is
1+

10K(s + 1)
=0.
s(s2 + 4.5s + 9)

(a) The root locus is shown in Figure DP7.2a. When K = 0.435, we have
ζ = 0.6 and the roots are
s1 = −0.368
s2,3 = −2.1 ± j2.75 .
(b) The response to a step input is shown in Figure DP7.2b. The performance results are
P.O. = 0%
Tss = 10 sec
ess = 0 .
(c) We have ζ = 0.41 when K = 1.51. The step response is shown in
Figure DP7.2b. The performance results to the step input are
P.O. = 0%
Ts = 4 sec
ess = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

356

CHAPTER 7

The Root Locus Method

5
4
3
2

x

Imag Axis

1
0

o

x

-1

0

-1
-2

x

-3
-4
-5
-5

-4

-3

-2

1

2

3

4

5

Real Axis

FIGURE DP7.2
10(s+1)
(a) Root locus for 1 + K s(s2 +4.5s+9) = 0.

1
0.9
K=0.435 ____ (solid line)

0.8

K=1.510 ---- (dashed line)

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

2

4

6

8

10

12

Time (sec)

FIGURE DP7.2
CONTINUED: (b) Unit step responses for K = 0.425, 1.51.

DP7.3

The characteristic equation is
1+

K(s2 + 6.5s + 12)
=0.
s(s + 1)(s + 2)

14

16

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

357

Design Problems

(a) The root locus is shown in Figure DP7.3.

6

4

2

Imag Axis

o

0

x

x

x

-2

-1

0

o

-2

-4

-6
-6

-5

-4

-3

1

Real Axis

FIGURE DP7.3
s2 +6.5s+12
= 0.
Root locus for 1 + K s(s+1)(s+2)

When K = 41, the roots are s1 = −37.12 and s2,3 = −3.44 ± j1.19 .

(b) The percent overshoot is P.O. ≈ 1% when ζ = 0.82 at K = 0.062.
(c) Select K > 300.
DP7.4

The characteristic equation is
1+K

10(0.01s + 1)
=0.
s(s2 + 10s + 10K1 )

If we choose K1 = 2.5, then the root locus will start at s = 0, −5 and
-5. This is shown in Figure DP7.4. The root locus then has a nice shape
so that we can select K to place the complex poles where desired and
the one real root will be farther in the left half-plane; thus the notion
of dominant poles will be valid. So, if we desire a P.O. < 5%, we want
ζ > 0.69. This occurs when K ≈ 3. Thus, our design is
K1 = 2.5

and K = 3 .

The unit step response is shown in Figure DP7.4. The settling time is less
than 3.5 sec and the P O < 4%. The response to a unit step disturbance
is also shown in Figure DP7.4. The steady-state error magnitude to the
disturbance is 0.33.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

358

CHAPTER 7

The Root Locus Method
4

3

2
K=3 -->

Imag Axis

1

0

-1

-2

-3

-4
-20

-15

-10

-5
Real Axis

0

5

10

1.4

Input step response
Disturbance step response
1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2

2.5
Time (sec)

3

3.5

4

4.5

5

FIGURE DP7.4
10(0.01s+1)
(a) Root locus for 1 + K s(s2 +10s+25) = 0. (b) System response to step input and disturbance.

DP7.5

The characteristic equation is
1+K

s+1
=0.
s(s − 0.1)(s2 + 10s + 41)

The root locus is shown in Figure DP7.5a.
The system is stable for 5 < K < 300. The step response with K =

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

359

Design Problems

Root Locus
10
8
System: sysgc
Gain: 90.5
Pole: −1.42 + 2.24i
Damping: 0.536
Overshoot (%): 13.6

6

Imaginary Axis

4
2
0
−2
−4
−6
−8
−10
−12

−10

−8

−6
−4
Real Axis

−2

0

2

Step Response
1.6
System: sys_cl
Peak amplitude: 1.57
Overshoot (%): 57
At time (sec): 1.24

1.4

Amplitude

1.2

System: sys_cl
Time (sec): 3.39
Amplitude: 0.98

1
0.8
0.6
0.4
0.2
0

0

0.5

1

1.5

2
2.5
Time (sec)

3

3.5

4

4.5

FIGURE DP7.5
s+1
(a) Root locus for 1 + K s(s−0.1)(s
2 +10s+41) = 0. (b) Step response with K = 875.

90.5 is shown in Figure DP7.5b. We choose K = 90.5 to minimize the
settling time. The damping of the dominant poles is ζ = 0.54, so that the
estimated percent overshoot is P.O. = 13%. The actual percent overshoot
and settling time are P.O. = 57% and Ts = 3.4 seconds. The match
between the actual and predicted percent overshoot can be improved by
selecting a much higher gain K, but then the step response becomes
overy oscillatory and the settling time increases too much for a typical
high-performance aircraft.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

360

CHAPTER 7

The characteristic equation is
1+K

s+2
=0.
s(s + 10)(s − 1)

The maximum damping is ζ = 0.46 at K = 55. The root locus is shown in
Figure DP7.6a; the step response is shown in Figure DP7.6b. The percent
overshoot and settling time are P.O. = 61.3% and Ts = 2 seconds.

20
15
10

Imag Axis

5
+

0

x

+

o

x x

+

-5
-10
-15
-20
-20

-15

-10

-5

0

5

10

15

20

Real Axis
1.8
1.6
1.4
1.2

Amplitude

DP7.6

The Root Locus Method

1
0.8
0.6
0.4
0.2
0
0

0.5

1

1.5

2

2.5

3

Time (secs)

FIGURE DP7.6
s+2
(a) Root locus for 1 + K s(s+10)(s−1)
= 0. (b) Step response with K = 55.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

361

Design Problems

DP7.7

The loop transfer function is
Gc (s)G(s) =

KP s + KI
.
s(s + 1)(0.5s + 1)

One possible set of PI controller gains are KP = 0.82 and KI = 0.9. The
step response is shown in Figure DP7.7.

Step Response
1.4
System: syscl
Peak amplitude: 1.05
Overshoot (%): 4.59
At time (sec): 3.57

1.2

Amplitude

1

System: syscl
Settling Time (sec): 4.94

0.8

0.6

0.4

0.2

0

0

1

2

3

4
Time (sec)

5

6

7

8

FIGURE DP7.7
Step response for with PI controller Gc (s) = (0.82s + 0.9)/s.

DP7.8

The closed-loop transfer function is
T (s) =

Vo (s)
G(s)
=
.
V (s)
1 + KG(s)

The dc gain is
T (0) =

G(0)
1
≈
.
1 + KG(0)
K

The root locus is shown in Figure DP7.8. The maximum value of K for
stability is
K = 0.062 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

362

CHAPTER 7

The Root Locus Method

x10 7
2
1.5
1

+

Imag Axis

0.5
0+

x

x

-0.5
-1

+

-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

1

1.5

Real Axis

2
x10 7

FIGURE DP7.8
3.142K1 ×1017
Root locus for 1 + K (s+3142)(s+10
7 )2 = 0.

Therefore, the minimum dc gain is about 1/0.062=16. Selecting
K = 0.05

and R1 = 10 K

yields
R2 = 19R1 = 190 K .
DP7.9

The closed-loop transfer function (with Gp (s) = 1 and K = 1) is
T (s) =

2s3 + 6s2 + 14s + 10
.
s4 + 6s3 + 13s2 + 26s + 6

So, if we select Gp (s) = 1/T (0) = 0.6, the step response (with K = 1)
will have a zero steady-state tracking error. The root locus is shown in
Figure DP7.9a. The step responses for K = 1, 1.5 and 2.85 are shown in
Figure DP7.9b. For K = 1, we have P.O. = 0%, Tr = 7.8 and Ts = 13.9;
for K = 1.5, we have P.O. = 0%, Tr = 5.4 and Ts = 9.6; and for K = 2.85,
we have P.O. = 5.2%, Tr = 0.5 and Ts = 7.3. The best gain selection is
K = 2.85.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

363

Design Problems

8
6
4

Imag Axis

2

x

0

x

o

-2

x

x

-4
-6
-8
-8

-6

-4

-2

0

2

4

6

8

Real Axis
K=1 (solid); K=1.5 (dashed); K=2.85 (dotted)
1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

18

20

Time (sec)

FIGURE DP7.9
6(s+1)
(a) Root locus for 1 + K s(s+4)(s2 +2s+5) = 0. (b) Step responses with K = 1, 1.5, 2.85.

DP7.10

A suitable selection of the various parameters is
ζ = 0.5

and q = 3/5 .

With q = 3/5, the open-loop zeros are real and equal. Then, it follows
that
λ=

2q
=3.
1−q

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

364

CHAPTER 7

The Root Locus Method

The root locus is shown in Figure DP7.10. A reasonable choice of gain is
K = 30 .
The resulting step response is extremely fast with no overshoot. The
closed-loop transfer function is approximately given by
T (s) ≈

1923
.
s + 1923

6

4
x

Imag Axis

2

0

o

x

-2
x

-4

-6
-6

-4

-2

0

2

4

6

Real Axis

FIGURE DP7.10
4s2 +4s+1
Root locus for 1 + K 0.0625s
3 +0.25s2 +s = 0.

DP7.11

The characteristic equation (with K as the parameter) is
1+K

10(s2 + 10)
=0.
s3 + 20s

The root locus is shown in Figure DP7.11a. To maximize the closed-loop
system damping we choose
K = 0.513.
The step response is shown in Figure DP7.11b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

365

Design Problems
5

4

3

2

Imag Axis

1

0

-1

-2

-3

-4

-5
-2

-1.5

-1

-0.5
Real Axis

0

0.5

1

Step Response
From: U(1)
1.4

1.2

0.8
To: Y(1)

Amplitude

1

0.6

0.4

0.2

0

0

1

2

3

4

5

Time (sec.)

FIGURE DP7.11
10(s2 +10)
(a) Root locus for 1 + K s3 +20s = 0. (b) Step response with K = 0.513.

DP7.12

The characteristic equation is
1+K

s + 1.5
=0.
(s + 1)(s + 2)(s + 4)(s + 10)

The root locus is shown in Figure DP7.12a.

6

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

10
8
6
4

Imag Axis

2
0

x

x

x o x

-2
-4
-6
-8
-10
-15

-10

-5

0

5

Real Axis
K=100 (solid); K=300 (dashed); K=600 (dotted)
1.6
1.4
1.2
1

Amplitude

366

0.8
0.6
0.4
0.2
0
0

1

2

3

4

5

6

7

8

9

10

Time (sec)

FIGURE DP7.12
s+1.5
(a) Root locus for 1 + K (s+1)(s+2)(s+4)(s+10)
= 0. (b) Step response with K =
100, 300, 600.

The closed-loop system roots are:
K = 100 : s1 = −11.38
K = 300 : s1 = −12.94
K = 600 : s1 = −14.44

s2,3 = −2.09 ± 3.10j
s2,3 = −1.29 ± 5.10j
s2,3 = −0.53 ± 6.72j

The step responses are shown in Figure DP7.12b.

s4 = −1.45
s4 = −1.48
s4 = −1.49

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

367

Design Problems

DP7.13

The closed-loop transfer function is
T (s) =

s3

Ka
.
+ Ka K2 s + Ka

s2

+

A suitable choice of gains is
Ka = 0.52

and K2 = 3 .

The step response is shown in Figure DP7.13.

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0

0

2

4

6

8

10

12

14

16

18

20

Time (secs)

FIGURE DP7.13
Step response with Ka = 0.52 and K2 = 3.

DP7.14

The characteristic equation is
s2 + 10KD s + 10(KP + 1) = 0 .
In the Evans form we have
1 + KD

10(s + τ )
=0.
s2 + 10

The root locus is shown in Figure DP7.14 for τ = 6. As τ → 0,√the
dominant closed-loop pole approaches s = 0 as KD√→ ∞. As τ → 10,
the dominant closed-loop pole approaches s = − 10 as KD → ∞. A
viable controller is KP = 72 and KD = 12 when τ = 6.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

Root Locus
8

6

4

Imaginary Axis

368

2

0

−2

−4

−6

−8
−25

FIGURE DP7.14
Root locus when τ = 6.

−20

−15

−10
Real Axis

−5

0

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

369

Computer Problems

Computer Problems
The root locus for parts (a)-(d) are shown in Figures CP7.1a - CP7.1d.

num=[30]; den=[1 14 43 30]; rlocus(sys)
30

Imaginary Axis

20

10

0

−10

−20

−30
−40

−30

−20

−10
Real Axis

0

10

20

num=[1 20]; den=[1 4 20]; rlocus(sys)
20
15
10
Imaginary Axis

CP7.1

5
0
−5
−10
−15
−20
−70

−60

−50

−40

−30
Real Axis

−20

−10

0

10

FIGURE CP7.1
s+20
(a) Root locus for 1 + k s3 +14s230
= 0. (b) Root locus for 1 + k s2 +4s+20
= 0.
+43s+30

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method

num=[1 1 2]; den=[1 6 10 0]; rlocus(sys)
1.5

Imaginary Axis

1

0.5

0

−0.5

−1

−1.5
−6

−5

−4

−3
−2
Real Axis

−1

0

1

num=[1 4 6 10 6 4]; den=[1 4 4 1 1 10 1]; rlocus(sys)
1.5

1

Imaginary Axis

370

0.5

0

−0.5

−1

−1.5
−7

−6

−5

−4

−3
Real Axis

−2

−1

0

1

FIGURE CP7.1
2
+s+2
CONTINUED: (c) Root locus for 1 + k s(ss2 +6s+10)
= 0. (d) Root locus for 1 +
5

4

3

2

+4s +6s +10s +6s+4
k s6s+4s
5 +4s4 +s3 +s2 +10s+1 = 0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

371

Computer Problems

CP7.2

The maximum value of the gain for stability is k = 0.791. The m-file
script and root locus is shown in Figure CP7.2.

Select a point in the graphics window
num=[1 -2 2]; den=[1 3 2 0]; sys = tf(num,den);

selected_point =

rlocus(sys)
rlocfind(sys)

-0.0025 + 0.6550i
ans =
0.8008
1

0.8

0.6

0.4

Imag Axis

0.2

0

−0.2

−0.4

−0.6

−0.8

−1
−3

−2.5

−2

−1.5

−1

−0.5
Real Axis

0

0.5

1

1.5

2

FIGURE CP7.2
Using the rlocfind function.

The value of k = 0.8008 selected by the rlocfind function is not exact
since you cannot select the jω-axis crossing precisely. The actual value is
determined using Routh-Hurwitz analysis.
CP7.3

The partial fraction expansion of Y (s) is
Y (s) =

s(s2

s+6
0.1667 1.6667 1.5
=
−
+
.
+ 5s + 4)
s+4
s+1
s

The m-file script and output is shown in Figure CP7.3.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

372

CHAPTER 7

The Root Locus Method
r=
0.1667
-1.6667
1.5000
p=

num=[1 6]; den=[1 5 4 0];
[r,p,k]=residue(num,den)

-4
-1
0
k=
[]

FIGURE CP7.3
Using the residue function.

The characteristic equation is
1+p

s−1
=0.
s2 + 5s + 10

The root locus is shown in Figure CP7.4. The closed-loop system is stable
for
0 < p < 10 .

n*+,-. /.01 23n,-. 4 .501 6789*:;:<:=
2.5
2
1.5
1
Imaginary Axis

CP7.4

0.5
0
−0.5
−1
−1.5
−2
−2.5
−8

−7

−6

FIGURE CP7.4
s−1
Root locus for 1 + p s2 +5s+10
= 0.

−5

−4

−3
−2
Real Axis

−1

0

1

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

373

Computer Problems

CP7.5

The characteristic equation is
1+k

s+1
=0.
s2

The root locus is shown in Figure CP7.5. For k = 2 we obtain s1,2 =
−1 ± j, that is, we have ζ = 0.707.
num=[1 1]; den=[1 0 0]; sys = tf(num,den);
hold off, clf
rlocus(sys);
hold on
plot([0 -2],[0 2*tan(acos(0.707))],'--')
plot([0 -2],[0 -2*tan(acos(0.707))],'--')
plot([-1 -1],[1 -1],'*')
2.5

2

1.5

1

Imag Axis

0.5

0

-0.5

-1

-1.5

-2

-2.5
-3

-2.5

-2

-1.5

-1
Real Axis

-0.5

0

0.5

1

FIGURE CP7.5
Root locus for 1 + k s+1
= 0.
s2

CP7.6

We choose a controller with two real poles and two real zeros selected to
meet the steady-state specification. The characteristic equation is
1+K

(s + 5.5)(s + 0.01)
10
=0.
(s + 6.5)(s + 0.0001) s3 + 15s2 + 50s

The m-file and root locus is shown in Figure CP7.4a. From the root locus we can select the value of the gain K that results in an estimated
P.O. ≤ 5% and a ζωn ≤ −2 to meet the settling time specification. We

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

374

CHAPTER 7

The Root Locus Method

select K = 8.58. The step response is shown in Figure CP7.6b showing
the percent overshoot and settling time specifications are satisfied. The
velocity constant is kv = 145.2 which implies a steady-state error to a
ramp input of ess = 1/kv = 0.0069.
25
20
15

System: untitled1
Gain: 8.58
Pole:
Damping: 0.807
Overshoot (%): 1.37

10
Imaginary Axis

ng=10; dg=conv([1 10 0],[1 5]); s
ysg=tf(ng,dg);
nh=conv([1 0.01],[1 5.5]);
dh=conv([1 6.5],[1 0.0001]);
sysh=tf(nh,dh);
figure(1)
rlocus(sysg*sysh)
K=8.58; sysh=tf(K*nh,dh);
sys=series(sysg,sysh);syscl=feedback(sys,1)
figure(2)
subplot(121)
step(syscl);
Kv=10*8.58*0.01*5.5/10/6.5/0.0001/5
systd=feedback(sysg,sysh);
subplot(122)
step(systd)

5
0
−5
−10
−15
−20
−25
−40

−35

−30

−25

Step Response

−15
−10
Real Axis

−5

0

5

10

Disturbance Response

1.4

0.14

System: syscl
Peak
1.2amplitude: 1.02
Overshoot (%): 1.77
At time (sec): 2.14

0.12

1

0.1
System: syscl
Settling Time (sec): 1.51

0.8

y(t)/Q

Amplitude

−20

0.08

0.6

0.06

0.4

0.04

0.2

0.02

0

0

1

2
3
Time (sec)

4

5

0

0

200
400
Time (sec)

600

FIGURE CP7.6
(a) Root locus. (b) Step response and disturbance response.

CP7.7

The m-file script to generate the root locus for each controller in parts
(a)-(c) is shown in Figure CP7.7. The performance region is indicated on
each root locus in Figures CP7.7b - CP7.7d. For part (a), the controller
gain is found to be Gc (s) = 11.3920. The integral controller in part (b)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

375

Computer Problems

numg=[1]; deng=[1 5 6]; sysg = tf(numg,deng);
t=[0:0.1:15];
%
% Part (a)
%
ÈSelect a point in the graphics window
sys1 = sysg;
rlocus(sys1), grid
selected_point =
hold on
plot([-0.4 -0.4],[-6 6],'--',...
-2.5030 + 3.3380i
[0 -6*tan(36.2*pi/180)],[0 6],'--',...
[0 -6*tan(36.2*pi/180)],[0 -6],'--')
ans =
hold off
[kp,poles] = rlocfind(sys1)
11.3920
%
% Part (b)
%
numc=[1]; denc=[1 0]; sysc = tf(numc,denc);
sys2 = series(sysc,sysg);
Select a point in the graphics window
figure
rlocus(sys2), grid
selected_point =
hold on
plot([-0.4 -0.4],[-6 6],'--',...
-0.6690 + 0.8210i
[0 -6*tan(36.2*pi/180)],[0 6],'--',...
[0 -6*tan(36.2*pi/180)],[0 -6],'--')
ans =
hold off
[ki,poles] = rlocfind(sys2)
4.0930
%
% Part (c)
%
Plot performance region boundaries on graph.
figure
numc=[1 1]; denc=[1 0]; sysc = tf(numc,denc);
Select a point in the graphics window
sys3 = series(sysc,sysg);
rlocus(sys3), grid
selected_point =
hold on
plot([-0.4 -0.4],[-6 6],'--',...
-2.0695+ 2.7387i
[0 -6*tan(36.2*pi/180)],[0 6],'--',...
[0 -6*tan(36.2*pi/180)],[0 -6],'--')
ans =
hold off
[kpi,poles] = rlocfind(sys3)
9.2516
%
% Part (d)
%
figure
sys1_o = kp*sys1; sys1_cl = feedback(sys1_o,[1]);
sys2_o = ki*sys2; sys2_cl = feedback(sys2_o,[1]);
sys3_o = kpi*sys3; sys3_cl = feedback(sys3_o,[1]);
[y1,t]=step(sys1_cl,t);
[y2,t]=step(sys2_cl,t);
[y3,t]=step(sys3_cl,t);
plot(t,y1,t,y2,'--',t,y3,':'),grid
xlabel('time [sec]'),ylabel('y(t)')
title('Gc(s): proportional (solid), integral (dashed) & PI (dotted)')

FIGURE CP7.7
(a) Script to generate the root locus for each controller.

is determined to be
Gc (s) =

4.093
.
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 7

The Root Locus Method
6

4

Imag Axis

2

0

-2

-4

-6
-4

-3

-2

-1
Real Axis

0

1

2

FIGURE CP7.7
CONTINUED: (b) Root locus for proportional controller with selected K = 11.3920.

The proportional integral (PI) controller in part (c) is

6

4

2

Imag Axis

376

0

-2

-4

-6
-4

-3

-2

-1
Real Axis

0

1

2

FIGURE CP7.7
CONTINUED: (c) Root locus for integral controller with selected K = 4.0930.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

377

Computer Problems

Gc (s) =

9.2516(s + 1)
.
s

The proportional controller is stable for all K > 0 but has a significant
6

4

Imag Axis

2

0

-2

-4

-6
-4

-3

-2

-1
Real Axis

0

1

2

FIGURE CP7.7
CONTINUED: (d) Root locus for PI controller with selected K = 9.2516.

but is stable only for K < 30. The PI controller has zero steady-state
error and is stable for all K > 0. Additionally, the PI controller has a
fast transient response. The step responses for each controller is shown in
Figure CP7.7e.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

378

CHAPTER 7

The Root Locus Method
Gc(s): proportional (solid), integral (dashed) & PI (dotted)

1.4

1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

5

10

15

time [sec]

FIGURE CP7.7
CONTINUED: (e) Step responses for each controller.

CP7.8

The loop transfer function can be written as
Gc (s)G(s) =

K1 + K2 s
s+5
= K̄2 2
2
Js
s

where
K̄2 = K2 /J .
The parameter of interest for the root locus is K̄2 . The root locus is shown
in Figure CP7.8. The selected value of
K̄2 = 7.1075 .
Therefore,
K2
= 7.1075
J

and

K1
= 35.5375 .
J

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

379

Computer Problems
num=[1 5]; den=[1 0 0]; sys=tf(num,den); rlocus(sys); rlocfind(sys)
10
8
6
+

4

Imag Axis

2
0

o

x

-2
-4
+

-6
-8
-10
-10

-8

-6

-4

-2

0

2

4

6

8

10

Real Axis

FIGURE CP7.8
Root locus to determine K̄2 .

The value of K that results in a damping ratio of ζ = 0.707 is K = 5.2.
The root locus is shown in Figure CP7.9.
Root Locus
5
4
3
2
Imaginary Axis

CP7.9

s = -0.68 + 0.68j

1
0
−1

s = -6.63
s = -0.68 - 0.68j

−2
−3
−4
−5
−10

−5

0
Real Axis

FIGURE CP7.9
Root locus for 1 + K s3 +8s21+10s+1 = 0.

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

380

CHAPTER 7

(a) The characteristic equation is
s3 + (2 + k)s2 + 5s + 1 = 0 .
(b) The Routh array is
s3

1

5

s2

2+k

1

s1

5k+9
2+k

so

1

Root Locus
2
1.5
1
Imaginary Axi s

CP7.10

The Root Locus Method

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

?-2

?-1.5

?-1

?-0.5

Real Axi s

FIGURE CP7.10
2
Root locus for 1 + k s3 +2ss2 +5s+1 = 0.

For stability we require
2 + k > 0 or

k > −2

and
5k + 9 > 0 or

k > −9/5 .

Therefore, the stability region is defined by
k > −1.8 .

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

381

Computer Problems

(c) Rearranging the characteristic equation yields
1+k

s2
s3 + 2s2 + 5s + 1 = 0 .

The root locus is shown in Figure CP7.10. We see that the system is
stable for all k > 0.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

8

Frequency Response Methods

Exercises
E8.1

Given the loop transfer function
L(s) =

4
,
(s + 2)2

we determine that
|L(jω)| =

4
4 + ω2

and φ(ω) = −2 tan−1 ω/2 .

The frequency response is shown in Figure E8.1.

Bode Diagram

Magnitude (dB)

0
−20
−40
−60

Phase (deg)

−80
0
−45
−90
−135
−180
−2
10

FIGURE E8.1
Frequency response for L(s) =

382

−1

10

0

10

4
.
(s+2)2

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

383

Exercises

The magnitude and phase angle for ω = 0, 0.5, 1, 2, 4, ∞ are summarized
in Table E8.1.

ω

0

0.5

1

2

4

∞

|L(jω)|

1

0.94

0.80

0.50

0.20

0

φ (deg)

0

-28.07

-53.13

-90

–126.87

-180

TABLE E8.1

E8.2

Magnitude and phase for L(s) =

4
.
(s+2)2

The transfer function is
G(s) =

5000
.
(s + 70)(s + 500)

The frequency response plot is shown in Figure E8.2. The phase angle is
computed from
φ = − tan−1

ω
ω
− tan−1
.
70
500

The phase angles for ω = 10, 100 and 700 are summarized in Table E8.2.

ω

TABLE E8.2

10

200

700

|G(jω)|

-16.99

-27.17

-41.66

φ (deg)

-9.28

-92.51

-138.75

Magnitude and phase for G(s) =

5000
.
(s+70)(s+500)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

384

CHAPTER 8

Frequency Response Methods

Bode Diagram

Magnitude (dB)

0
−20
−40
−60
−80

Phase (deg)

−100
0
−45
−90
−135
−180
0
10

1

10

FIGURE E8.2
Frequency response for G(s) =

E8.3

2

3

10

4

10

10

5000
(s+70)(s+500) .

The loop transfer function is
L(s) =

300(s + 100)
.
s(s + 10)(s + 40)

The phase angle is computed via
φ(ω) = −90o − tan−1

ω
ω
ω
− tan−1
+ tan−1
.
10
40
100

At ω = 28.3, we determine that
φ = −90o − 70.5o − 35.3o + 15.8o = 180o .
Computing the magnitude yields
1

|L(jω)| =

ω 2 2
300(100)(1 + ( 100
) )
1

1

ω 2 2
ω 2 2
ω10(1 + ( 10
) ) 40(1 + ( 40
) )

when ω = 28.3. We can also rewrite L(s) as
L(s) =

s
75( 100
+ 1)
.
s
s
s( 10 + 1)( 40
+ 1)

= 0.75 ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

385

Exercises

Then, the magnitude in dB is
ω 2
ω
) ) − 10 log 10 (1 + ( )2 )
100
10
ω
− 10 log 10 (1 + ( )2 ) − 20 log10 ω = −2.5 dB ,
40

20 log 10 |L| = 20 log 10 (75) + 10 log 10 (1 + (

at ω = 28.3.
E8.4

The transfer function is
G(s) =

Ks
.
(s + a)(s + 10)2

Note that φ = 0o at ω = 3, and that
φ = +90o − tan−1

ω
ω
− 2 tan−1
.
a
10

Substituting ω = 3 and solving for a yields
a=2.
Similarly, from the magnitude relationship we determine that
K = 400 .
E8.5

The lower portion for ω < 2 is
20 log

K
= 0 dB ,
ω

at ω = 8. Therefore,
20 log

K
= 0 dB
8

which occurs when
K=8.
We have a zero at ω = 2 and another zero at ω = 4. The zero at ω = 4
yields
a = 0.25 .
We also have a pole at ω = 8, and a second pole at ω = 24. The pole at
ω = 24 yields
b = 1/24 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

386

CHAPTER 8

Frequency Response Methods

Therefore,
G(s) =
E8.6

8(1 + s/2)(1 + s/4)
.
s(1 + s/8)(1 + s/24)(1 + s/36)

The loop transfer function is
L(s) =

10
.
s(s/5 + 1)(s/100 + 1)

The Bode diagram is shown in Figure E8.6. When 20 log 10 |L(jω)| = 0 dB,
we have
Bode Diagram

Magnitude (dB)

50
0
−50
−100

Phase (deg)

−150
−90
−135
−180
−225
−270
−1
10

0

10

FIGURE E8.6
Bode Diagram for L(s) =

E8.7

1

2

10
10

3

10

4

10

10
.
s(s/5+1)(s/100+1)

The transfer function is
T (s) =

4
.
(s2 + s + 1)(s2 + 0.4s + 4)

(a) The frequency response magnitude is shown in Figure E8.7.
The frequency response has two resonant peaks at

and ωr2 = 1.9 rad/sec .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

387

Exercises

10

Gain dB

5
0
-5
-10
10-1

100

101

Amplitude

1.5
1
0.5
0
0

2

4

FIGURE E8.7
(a) Bode Diagram for T (s) =

6

8

10
12
Time (secs)

4
(s2 +s+1)(s2 +0.4s+4) .

14

16

18

20

(b) Unit step response.

(b) The percent overshoot is
P.O. = 35% ,
and the settling time is
Ts ≈ 16 sec .
(c) The step response is shown in Figure E8.7.
E8.8

(a) The break frequencies are
(b) The slope of the asymptotic plot at low frequencies is 0 dB/decade.
And at high frequencies the slope of the asymptotic plot is -20 dB/decade.
(c) The Bode plot is shown in Figure E8.8.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

388

CHAPTER 8

Frequency Response Methods

Bode Diagram
20

Magnitude (dB)

10
0
−10
−20
−30
180

Phase (deg)

135
90
45
0
−45
−90
−2
10

−1

10

FIGURE E8.8
Bode Diagram for Gc (s)G(s) =

1

2

10

3

10

100(s−1)
.
s2 +25s+100

The Bode diagram for G(s) is shown in Figure E8.9.
40

Gain dB

20
0
-20
-40
10-1

100

101

102

103

102

103

50

Phase deg

E8.9

0

10
10

0
-50
10-1

100

101

FIGURE E8.9
Bode Diagram for G(s) =

(s/5+1)(s/20+1)
.
(s+1)(s/80+1)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

389

Exercises

E8.10

The frequency response has two peaks; the first peak at f ≈ 1.8 and the
second peak at f ≈ 3.1. One possible G(jω) is
1

G(jω) =



(jωτ + 1) 1 +



2ζ1
ωn1



jω +



jω
ωn1

2  

1+



2ζ2
ωn2



jω +



jω
ωn2

2  ,

where
τ=

1
,
2π(0.2)

ωn1 = 2π(1.8 × 103 )

ζ1 = 0.15;

ζ2 = 0.15;

ωn2 = 2π(3.1 × 103 ) .

The damping ratios are estimated using Figure 8.10 in Dorf & Bishop.
E8.11

The Bode plot is shown in Figure E8.11. The frequency when 20 log10 |GC G(ω)| =
0 is ω = 9.9 rad/sec.
Bode Diagram
20

Magnitude (dB)

0
−20
−40
−60
−80
−100
−120
0

Phase (deg)

−45
−90
−135
−180
−225
−270
−1
10

0

10

FIGURE E8.11
Bode Diagram for Gc (s)G(s) =

E8.12

1

10

2

10

1000
.
(s2 +10s+100)(s+2)

(a) The transfer function is
G(s) = C(sI − A)−1 B + D =
(b) The Bode plot is shown in Figure E8.12.

−5(s − 1)
.
s2 + 3s + 2

3

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

390

CHAPTER 8

Frequency Response Methods
Bode Diagram

Phase (deg)

Magnitude (dB)

10
0
-10
-20

270
180
90
-2
10

10

FIGURE E8.12
Bode Diagram for G(s) =

0

10

10

1

10

2

−5(s−‘1)
.
s2 +3s+2

The closed-loop transfer function is
T (s) =

s3

+

100
.
+ 20s + 110

11s2

The Bode plot of T (s) is shown in Figure E8.13, where ωB = 4.9 rad/sec.

Bode Diagram

Magnitude (dB)

50

Phase (deg)

E8.13

-1

0

-3 dB

-50
-100

-45
-90
-135
-180
-225
-270
-1
10

FIGURE E8.13
Bode Diagram for T (s) =

ωb=4.9

10

0

1

10

100
s3 +11s2 +20s+110 .

10

2

10

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

391

Exercises

E8.14

The loop transfer function is
L(s) =

20
.
(s2 + 1.4s + 1)(s + 10)

The Bode plot of L(s) is shown in Figure E8.14. The frequency when
20 log 10 |L(ω)| = 0 is ω = 1.32 rad/sec.
Bode Diagram

Magnitude (dB)

50

0

−50

−100

−150
0

Phase (deg)

−45
−90
−135
−180
−225
−270
−2
10

FIGURE E8.14
Bode Diagram for L(s) =

E8.15

−1

0

10

1

10
10

2

10

3

10

20
.
(s2 +1.4s+1)(s+10)

The closed-loop transfer function is
T (s) =

3s + 5
.
s2 + s + K + 6

The bandwidth as a function of K is shown in Figure E8.15. The bandwidth as a function of K is:
(a) K = 1 and ωb = 7.0 rad/sec.
(b) K = 2 and ωb = 7.9 rad/sec.
(c) K = 10 and ωb = 14.7 rad/sec.
The bandwidth increases as K increases.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 8

Frequency Response Methods

24
22
20
18

392

16
14
12
10
8
6

0

FIGURE E8.15
Bandwith of T (s) =

2

4

3s+5
s2 +s+K+6 .

6

8

10
K

12

14

16

18

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

393

Problems

Problems
(a) The transfer function is
1
,
(1 + 0.25s)(1 + 3s)

Gc (s)G(s) =
and

1
.
(1 − 0.75ω 2 ) + j3.25ω

Gc (jω)G(jω) =

The polar plot is shown in Figure P8.1a. A summary of the magnitude and phase angles for ω = 0, 0.5, 1, 2, 5 and ∞ can be found in
Table P8.1a.
Nyquist Diagram
0.8
0.6
0.4
Imaginary Axis

P8.1

0.2
0
−0.2
−0.4
−0.6
−0.8
−1

−0.8

−0.6

FIGURE P8.1
(a) Polar plot for Gc (s)G(s) =

−0.4

−0.2

0
Real Axis

0.2

0.4

0.6

0.8

1

1
.
(1+0.25s)(1+3s)

ω

0

0.5

1

2

5

∞

|Gc (jω)G(jω)| (dB)

1.00

0.55

0.31

0.15

0.04

0

φ (deg)

0

-63.4

-85.6

-107.1

-137.51

-180

TABLE P8.1

(a) Magnitudes and phase angles for Gc (s)G(s) =

1
.
(1+0.25s)(1+3s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

394

CHAPTER 8

Frequency Response Methods

(b) The transfer function is
Gc (s)G(s) =

5(s2 + 1.4s + 1)
(s − 1)2

and
5 (1 − ω 2 ) + 1.4jω
Gc (jω)G(jω) =
.
(1 − ω 2 ) − 2jω


The polar plot is shown in Figure P8.1b. A summary of the magnitude
and phase angles for ω = 0, 0.25, 0.5, 1, 2, 8, 16 and ∞ can be found in
Table P8.1b.
Nyquist Diagram
5
4
3

Imaginary Axis

2
1
0
−1
−2
−3
−4
−5
−4

−3

−2

FIGURE P8.1
CONTINUED: (b) Polar plot for Gc (s)G(s) =

−1

0

1
Real Axis

2

3

4

5

6

5(s2 +1.4s+1)
.
(s−1)2

ω

0

0.25

0.5

1

2

8

16

∞

|Gc (jω)G(jω)| (dB)

5.00

4.71

4.10

3.50

4.10

4.92

4.98

5.00

φ (deg)

0

48.5

96.1

180.0

-96.2

-24.3

-12.2

0

TABLE P8.1

CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) =

5(s2 +1.4s+1)
.
(s−1)2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

395

Problems

(c) The transfer function is
Gc (s)G(s) =

(s2

(s − 8
.
+ 6s + 8)

The polar plot is shown in Figure P8.1c. A summary of the magnitude
and phase angles for
ω = 0, 1, 2, 3, 4, 5, 6, ∞
can be found in Table P8.1c.

Nyquist Diagram
0.8
0.6

Imaginary Axis

0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1

−0.8

−0.6

FIGURE P8.1
CONTINUED: (c) Polar plot for Gc (s)G(s) =

−0.4
−0.2
Real Axis

0

0.2

0.4

s−8
.
s2 +6s+8

ω

0

1

2

3

4

5

6

∞

|Gc (jω)G(jω)| (dB)

1.00

0.87

0.65

0.47

0.35

0.27

0.22

0.00

φ (deg)

180.0

132.3

94.4

66.3

45.0

28.5

15.3

-90.0

TABLE P8.1

CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) =

s−8
.
s2 +6s+8

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

396

CHAPTER 8

Frequency Response Methods

(d) The transfer function is
Gc (s)G(s) =

20(s + 8)
.
s(s + 2)(s + 4)

The polar plot is shown in Figure P8.1d. A summary of the magnitude
and phase angles for
ω = 1, 0.1, 0.8, 1.6, 3.2, 12.8, ∞
can be found in Table P8.1d.
Nyquist Diagram
20
15

Imaginary Axis

10
5
0
−5
−10
−15
−20
−20

−15

−10

−5

0

5

Real Axis

FIGURE P8.1
CONTINUED: (d) Polar plot for Gc (s)G(s) =

20(s+8)
.
s(s+2)(s+4)

ω

0

0.1

0.8

1.6

3.2

12.8

∞

|Gc (jω)G(jω)| (dB)

∞

199.70

22.87

9.24

2.79

0.14

0.00

φ (deg)

0

-93.6

-117.4

-139.1

-164.8

174.3

180.0

TABLE P8.1

CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) =

20(s+8)
.
s(s+2)(s+4)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

397

Problems

P8.2

(a) The Bode plot is shown in Figure P8.2a. A summary of the magnitude
and phase angles for
ω = 0.25, 0.5, 1, 2, 4, 8, 16
can be found in Table P8.2a.

Bode Diagram

Magnitude (dB)

0
−20
−40
−60

Phase (deg)

−80
0
−45
−90
−135
−180
−2
10

FIGURE P8.2
(a) Bode plot for Gc (s)G(s) =

−1

10

0

1

10

2

10

10

1
.
(1+0.25s)(1+3s)

ω

0.25

0.5

1.0

2.0

4.0

8.0

16.0

|Gc (jω)G(jω)| (dB)

-1.95

-5.19

-10.26

-16.65

-24.62

-34.60

-45.93

φ (deg)

-40.5

-63.4

-85.6

-107.1

-130.2

-151.0

-164.8

TABLE P8.2

(a) Magnitudes and phase angles for Gc (s)G(s) =

1
.
(1+0.25s)(1+3s)

(b) The transfer function is
Gc (s)G(s) =

5(s2 + 1.4s + 1)
(s − 1)2

The Bode plot is shown in Figure P8.2b. A summary of the magnitude

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

398

CHAPTER 8

Frequency Response Methods

and phase angles for
ω = 0.25, 0.5, 1, 2, 4, 8, 16
can be found in Table P8.2b.

Bode Diagram

Magnitude (dB)

14
13
12
11

Phase (deg)

10
0
−90
−180
−270
−360
−1
10

0

1

10

FIGURE P8.2
CONTINUED: (b) Bode plot for Gc (s)G(s) =

10

5(s2 +1.4s+1)
.
(s−1)2

ω

0.25

0.5

1.0

2.0

4.0

8.0

16.0

|Gc (jω)G(jω)| (dB)

13.46

12.26

10.88

12.26

13.46

13.84

13.95

φ (deg)

48.5

96.2

180.0

-96.2

-48.5

-24.3

-12.2

TABLE P8.2

CONTINUED: (b) Magnitudes and phase angles for Gc (s)G(s) =

5(s2 +1.4s+1)
.
(s−1)2

(c) The transfer function is
Gc (s)G(s) =

(s2

(s − 8)
.
+ 6s + 8)

The Bode plot is shown in Figure P8.2c. A summary of the magnitude
and phase angles for ω = 0.6, 1, 2, 3, 4, 5, 6, ∞ can be found in
Table P8.2c.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

399

Problems

Bode Diagram
0

Magnitude (dB)

−10
−20
−30
−40
−50
−60
180
Phase (deg)

135
90
45
0
−45
−90
−2
10

−1

0

10

1

2

10
10

FIGURE P8.2
CONTINUED: (c) Bode plot for Gc (s)G(s) =

3

10

10

s−8
s2 +6s+8 .

ω

0.6

1

2

3

4

5

6

∞

|Gc (jω)G(jω)| (dB)

-0.45

-1.17

-3.72

-6.49

-9.03

-11.26

-13.18

-120.00

φ (deg)

150.5

132.3

94.4

66.3

45.0

28.5

15.3

-90.0

TABLE P8.2

CONTINUED: (c) Magnitudes and phase angles for Gc (s)G(s) =

s−8
s2 +6s+8 .

(d) A summary of the magnitude and phase angles for
ω = 0.2, 0.8, 3.2, 6.4, 12.8, 25.6, 51.2
can be found in Table P8.2d. The Bode plot is shown in Figure P8.2d.

ω

0.2

0.8

3.2

6.4

12.8

25.6

51.2

|Gc (jω)G(jω)| (dB)

39.95

27.19

8.90

-3.98

-17.35

-30.0355

-42.28

φ (deg)

-97.1

-117.4

-164.8

178.0

174.2

176.0

177.8

TABLE P8.2

CONTINUED: (d) Magnitudes and phase angles for Gc (s)G(s) =

20(s+8)
.
s(s+2)(s+4)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

400

CHAPTER 8

Frequency Response Methods

Bode Diagram
60

Magnitude (dB)

40
20
0
−20
−40

Phase (deg)

−60
−90

−135

−180

−225
−1
10

0

1

10

2

10

10

FIGURE P8.2
CONTINUED: (d) Bode plot for Gc (s)G(s) =

P8.3

20(s+8)
.
s(s+2)(s+4)

(a) The bridged-T network we found has zeros at
s = ±jωn
and poles at
s=−

q
ωn
± ωn 1/Q2 − 1 .
Q

The frequency response is shown in Figure P8.3 for Q = 10.
(b) For the twin-T network, we evaluate the magnitude at
ω = 1.1ωn
or 10% from the center frequency (see Example 8.4 in Dorf & Bishop).
This yields
|G| ≈ 2.1 ×



0.1
3.9



× 1.1 = 0.05 .

Similarly, for the bridged-T network
|G| = 2.1 ×



0.1
2.1



× 0.14 = 0.707 .

The bridged-T network possesses a narrower band than the twin-T
network.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

401

Problems

0

Gain dB

-10
-20
-30
-40
10-1

100

101

w/wn

Phase deg

100
50
0
-50
-100
10-1

100

101

w/wn

FIGURE P8.3
Bode plot for G(s) =

2
s2 +ωn
2 ,
s2 +(2ωn /Q)s+ωn

where ζ = 1/Q = 0.1.

The transfer function is

P8.4

1
s
30000(2s + 1)
=
.
s(s + 10)(s + 20)(s2 + 15s + 150)

G(s) = Gc G1 H(s)

 

A summary of the magnitude and phase angles can be found in Table P8.4.
The Bode plot is shown in Figure P8.4.

ω

1

3

5

8

10

15

24

|G(jω| dB

6.95

5.78

5.08

3.38

1.59

-5.01

-17.56

−40.89o

−52.39o

−77.28o

−118.41o

−145.99o

−203.52o

−258.57o

φ(deg)
TABLE P8.4

Magnitudes and phase angles for GH(s) =

30000(2s+1)
.
s(s+10)(s+20)(s2 +15s+150)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

402

CHAPTER 8

Frequency Response Methods
Bode Diagram

Magnitude (dB)

50
0
-50
-100

Phase (deg)

-150
0
-90
-180
-270
-360
-2
10

FIGURE P8.4
Bode plot for GH(s) =

-1

0

1

10
10

10

2

10

30000(2s+1)
.
s(s+10)(s+20)(s2 +15s+150)

The Bode plot is shown in Figure P8.5.

Bode Diagram
50

Magnitude (dB)

0
−50
−100
−150
−200
−250
0
Phase (deg)

P8.5

10

−90
−180
−270
−360
−2
10

FIGURE P8.5
Bode plot for G(s) =

−1

10

0

10

1

2

10
10

10
.
(s/4+1)(s+1)(s/20+1)(s/80+1)

3

10

4

10

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

403

Problems

(a) The transfer function is
GH(s) =

3.98(1 + s/1)
.
s(1 + s/10)2

We have a zero at ω = 1 and two poles at ω = 10.0. The low frequency
approximation is K/s and at ω = 1 we have
K
20 log
ω




= 12dB .

Therefore,
K = 3.98
at ω = 1 (an approximation). The phase plot is shown in Figure P8.6a.

(a)
-40
-60

Phase deg

-80
-100
-120
-140
-160
-180
-2
10

10

-1

10

0

10

1

10

2

(b)
100

50
Phase deg

P8.6

0

-50

-100
-1
10

FIGURE P8.6
Phase plots for (a) G(s) =

10

0

3.98(s/1+1)
.
s(s/10+1)2

1

10

(b) G(s) =

10

2

10

3

s
.
(s/10+1)(s/50+1)

(b) The transfer function is
GH(s) =

s
.
(1 + s/10)(1 + s/50)

The poles are located by noting that the slope is ±20 dB/dec. The

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

404

CHAPTER 8

Frequency Response Methods

low frequency approximation is Ks, so
20 log Kω = 0dB .
At ω = 1 we determine that
K=1.
The phase plot is shown in Figure P8.6b.
The loop transfer function is
L(s) =

Kv
.
s(s/π + 1)2

(a) Set Kv = 2π. The Bode plot is shown in Figure P8.7a.

Gain dB

40

20

0

-20
10-1

100

101

-80
-100

Phase deg

P8.7

-120
-140
-160
-180
10-1

100

101

FIGURE P8.7
(a) Bode plot for L(s) =

Kv
,
s(s/π+1)

where Kv = 2π.

(b) The logarithmic magnitude versus the phase angle is shown in Figure P8.7b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

405

Problems

40

30

Gain dB

20

10

0

-10

-20
-170

-160

-150

-140

-130

-120

-110

-100

-90

Phase deg

FIGURE P8.7
CONTINUED: (b) Log-magnitude-phase curve for L(jω).

P8.8

The transfer function is
T (s) =

s2

K
.
+ 10s + K

(a) When P.O. = 15%, we determine that ζ = 0.517 by solving
√
2
15 = 100e−πζ/ 1−ζ .
So, 2ζωn = 10 implies that ωn = 9.67, hence K = ωn2 = 93.53. Also,
q

Mpω = (2ζ 1 − ζ 2 )−1 = 1.13 .
(b) For second-order systems we have
q

ωr = ωn 1 − 2ζ 2 = 6.59
when ζ = 0.517 and ωn = 9.67.
(c) We estimate ωB to be
ωB ≈ (−1.19ζ + 1.85)ωn = 11.94 rad/s .
P8.9

The log-magnitude phase curves are shown in Figure P8.9.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

406

CHAPTER 8

Frequency Response Methods

(a)

(b)

0

40

-5

30

-10
20

-20

10

Gain dB

Gain dB

-15

-25

0

-30
-10
-35
-20

-40
-45
-200

-150

-100

-50

-30
-180

0

-160

Phase deg

-120

-100

Phase deg

FIGURE P8.9
Log-magnitude-phase curve for (a) Gc (s)G(s) =
1+0.5s
.
s2

P8.10

-140

1
(1+0.5s)(1+2s)

and (b) Gc (s)G(s) =

The governing equations of motion are
F (s) = Kf If (s) and If (s) =

Vf (s)
.
Rf + L f s

Without loss of generality we can let Kf = 1.0. Also, we have
F (s) = (M s2 + bs + K)Y (s) .
Therefore, the transfer function is
GH(s) =

KKf
50K
=
.
2
(Rf + Lf s)(M s + bs + K)
(s + 0.5)(s2 + 2s + 4)

This is a type 0 system, therefore Kp = 25K.
(a) If we allow a 1% error , we have ess = |R|/(1 + Kp ) = 0.01|R|. Thus
Kp = 25K = 99. Select
K=4.
(b) The Bode plot is shown in Figure P8.10a.
(c) The log-magnitude phase curve is shown in Figure P8.10b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

407

Problems

Gain dB

40

20

0
-20
10-2

10-1

100

101

100

101

Phase deg

0

-100

-200
-300
10-2

10-1

FIGURE P8.10
(a) Bode plot for GH(s) =

200
.
(s2 +2s+4)(s+0.5)

40

30

Gain dB

20

10

0

-10

-20
-300

-250

-200

-150

-100

-50

0

Phase deg

FIGURE P8.10
CONTINUED: (b) Log-magnitude-phase curve for GH(s) =

200
.
(s2 +2s+4)(s+0.5)

(d) The closed-loop transfer function Bode plot is shown in Figure P8.10c.
We determine from the plot that Mpω = 1.6, ωr = 4.4 and ωB = 6.8.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

408

CHAPTER 8

Frequency Response Methods

5

Gain dB

0
-5
-10
-15
10-1

100

101

Phase deg

100
0

-100
-200
10-1

100

101

FIGURE P8.10
CONTINUED: (c) Bode plot for closed-loop T (s) = Y (s)/R(s).

The Bode plot is shown in Figure P8.11.

Gain dB

200
100
0

-100
10-4

10-3

10-2

10-1

100

101

100

101

100

Phase deg

P8.11

0

-100
-200
10-4

10-3

10-2

10-1

FIGURE P8.11
Bode plot for G(s) =

0.164(s+0.2)(−s+0.32)
.
s2 (s+0.25)(s−0.009)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

409

Problems

P8.12

The three transfer functions are
G1 (s) = 10

G2 (s) =

1
s(s/0.6 + 1)

G3 (s) = 3s .

(a) When G3 (s) is out of the loop, the characteristic equation is
10
=0
s(s/0.6 + 1)
√
or s2 + 0.6s + 6 = 0. Thus, ζ = 0.6/(2 6) = 0.12.
(b) With G3 (s), the characteristic equation is
1 + G1 G2 (s) = 1 +

1 + G1 G2 (s) + G2 G3 (s) = 1 +

1.85
6
+
=0,
s(s + 0.6) s(s + 0.6)

or
s2 + 2.4s + 6 = 0 .
√
Thus, ζ = 2.4/(2 6) = 0.49.
P8.13

By inspection of the frequency response, we determine
L(s) = Gc (s)G(s)H(s) =

K
.
s(s/100 + 1)(s/1000 + 1)2

For small ω, we have 20 log K/ω = 40 dB at ω = 10. So, K = 1000.
P8.14

The data we have are R1 = R2 = 1000Ω, c1 = 10−7 farad and c2 = 10−6
1

V2 (s)
C1 s
=
,
V1 (s)
R1 + C11 s
and
Vo (s)
KR2
=
.
V2 (s)
R2 + C12 s
So
Vo (s)
KR2 C2 s
109 s
=
=
.
V1 (s)
(R1 C1 s + 1)(R2 C2 s + 1)
(s + 107 )(s + 1000)
(a) The Bode plot is shown in Figure P8.14.
(b) The mid-band gain is = 40 dB.
(c) The -3 dB points are (rad/sec): ωlow ≈ 7

and ωhigh ≈ 1.5 × 109 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

410

CHAPTER 8

Frequency Response Methods

Gain dB

40

20

0

-20
100

101

102

103

104

105

106

107

108

109

1010

107

108

109

1010

Phase deg

100

0

-100

-200
100

101

102

103

104

105

106

FIGURE P8.14
Bode plot for G(s) =

P8.15

109 s
.
(s+107 )(s+103 )

The data are plotted in Figure P8.15, denoted by an asterisk (*).
50
*
*

0

*

*

*

*

*

*

*
*
*

-50
-100
10-1

100

101

102

-50
-100 *

*

*

*

*

*

*
*

-150

*

-200
-250
-300
10-1

FIGURE P8.15
Bode plot for G(s) =

*

100

809.7
;
s(s2 +6.35s+161.3)

101

*

102

tabular data is indicated by an asterick (*).

The low frequency slope is -20 dB/dec and the initial low frequency φ is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

411

Problems

−90o , so we have an integrator of the form K/s. The initial phase is −90o
and the final phase −270o , so we have a minimum phase G(s). Now, |G|
is 0.97 at ω = 8 and ω = 10 indicating two complex poles. We postulate
a transfer function of the form
G(s) =

K
s



s2
2
ωn

+

2ζs
ωn

 .

+1

The phase angle φ = −180o at ω = ωn . Then, from Figure 8.10 in Dorf &
Bishop, we determine that ωn = 12.7. At ω = 8, ωωn = 0.63 and φ, due to
the complex poles is −30o (subtract −90o due to the integrator). Again,
from Figure 8.10 in Dorf & Bishop, we estimate ζ = 0.25. To determine
K, note that when ωωn ≤ 0.1, the effect of the complex poles on magnitude
is negligible, so at ω = 1 we have
K
|G| = 5.02 ∼
.
=
j1
So K = 5.02. Therefore,
G(s) =

s

s2
161.3

+

0.5s
12.7

 =

+1

809.7
.
s(s2 + 6.35s + 161.3)

(a) The unit step input response is shown in Figure P8.16. The step

Step Response
1.4

1.2

1
Amplitude

P8.16

5.02


0.8

0.6

0.4

0.2

0

0

0.1

0.2

0.3

FIGURE P8.16
Unit step input response for T (s) =

0.4
0.5
0.6
Time (sec)

60.2
s2 +12.1s+60.2 .

0.7

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

412

CHAPTER 8

Frequency Response Methods

response is given by
y(t) = 1 − e−6.05t (cos 4.85t + 1.25 sin 4.85t) .
(b) The system bandwidth is ωB = 4.95 rad/sec.
P8.17

The transfer function is
Gc (s)G(s) =

4(0.5s + 1)
.
s(2s + 1)(s2 /64 + s/20 + 1)

(a) The Bode plot is shown in Figure P8.17.

Gain dB

50

0

-50
-100
10-1

100

101

102

101

102

-50

Phase deg

-100
-150
-200
-250
-300
10-1

100

FIGURE P8.17
Bode plot for Gs (s)G(s) =

4(0.5s+1)
.
s(2s+1)(s2 /64+s/20+1)

(b) When the magnitude is 0 dB, we have
and when φ = −180o , we have
P8.18

The transfer function is
Gc (s)G(s) =

12(s + 0.5)
0.2(2s + 1)
=
.
(s + 3)(s + 10)
(s/3 + 1)(s/10 + 1)

The Bode plot is shown in Figure P8.18. Near 0 dB, the frequency is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

413

Problems

0

Gain dB

-5
-10
-15
-20
10-1

100

101

102

101

102

50

Phase deg

0
-50
-100
-150
-200
10-1

100

FIGURE P8.18
Bode plot for Gc (s)G(s) =

P8.19

12(s+0.5)
s2 +13s+30 .

Examining the frequency response, we postulate a second-order transfer
function
θ(s)
ωn2
= 2
.
I(s)
s + 2ζωn s + ωn2
From the data we see that φ = −90o at ω = 2. Using Figure 8.10 in Dorf
& Bishop, we determine that ωn = ω = 2. We also estimate ζ = 0.4 from
Figure 8.10. Thus,
θ(s)
4
= 2
.
I(s)
s + 1.6s + 4

P8.20

The transfer function is
Gc (s)G(s) =

823(s + 9.8)
.
+ 22s + 471

s2

The Bode plot is shown in Figure P8.20. The maximum value of
20 log10 |Gc (jω)G(jω)| = 32.3 dB
occurs at ω = 20.6 rad/sec and the corresponding phase is φ = −19.6o .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

414

CHAPTER 8

Frequency Response Methods

Bode Diagram

Magnitude (dB)

35
30
25
20

Phase (deg)

15
45

0

−45

−90
−1
10

0

1

10

10

2

10

FIGURE P8.20
Bode plot for Gc (s)G(s) =

The Bode plot is shown in Figure P8.21. The gain is 24 dB when φ =
−180o
40

Gain dB

20
0
-20
-40
10-1

100

101

102

101

102

0

Phase deg

P8.21

832(s+9.8)
s2 +22s+471 .

-100
-200
-300
10-1

100

FIGURE P8.21
Bode plot for Gc (s)G(s) =

−200s2
.
s3 +14s2 +44s+40

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

415

Problems

P8.22

The transfer function is
G(s) =

P8.23

10000(s + 1)(s + 80)
.
s(s + 300)(s + 9000)

The transfer function is
G(s) =

100(s + 20)(s + 8000)
.
(s + 1)(s + 80)(s + 500)

The system is type 0 and the steady-state error to a unit step input is
ess =

1
= 0.0025
1 + Kp

since
Kp = lim G(s) = 400 .
s→0

P8.24

(a) From the Bode plot we see that
20 log10 Mpω = 12
or Mpω = 3.981. For a second-order system we know that
Mpω = (2ζ

q

1 − ζ 2 )−1 .

Solving for ζ (with Mpω = 3.981) yields ζ = 0.12. Also, from the Bode
plot,
So,
ωr
ωn = p
= 0.91 .
1 − 2ζ 2

Therefore, the second-order approximate transfer function is
T (s) =

ωn2
0.83
= 2
.
s2 + 2ζωn s + ωn2
s + 0.22s + 0.83

(b) The predicted overshoot and settling time are P.O. = 68% and Ts =
37 sec.
P8.25

The transfer function is
G(s) =

100(s + 10)
.
s2 (s + 100)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

416

CHAPTER 8

P8.26

Frequency Response Methods

The transfer function is
T (s) =

Vo (s)
1 + R1 /R2
=
.
V (s)
1 + RCs

Substituting R = 10kΩ, C = 1µF , R1 = 9kΩ, and R2 = 1kΩ yields
T (s) =

10
.
1 + 0.01s

The frequency response is shown in Figure P8.26.
Bode Diagrams

20

15

Phase (deg); Magnitude (dB)

10

5

0
0
-20
-40
-60
-80
-100
1
10

10

2

10

FIGURE P8.26
Bode plot for T (s) =

P8.27

1+R1 /R2
1+RCs

The frequency response is shown in Figure P8.27.

TABLE P8.27

K

0.75

1

10

|L(jω)|jω=0 , dB

3.52

12.04

26.02

8.3

14.0

33.4

3.5

8.7

22.9

System performance as K varies.

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

417

Problems

Bode Diagram
10

Magnitude (dB)

0

K increases

−10
−20
−30
−40

K decreases

−50
−60
0

Phase (deg)

Phase plot remains unchanged as K varies
−45

−90

−135
−1
10

FIGURE P8.27
Bode plot for K = 1

10

0

1

10

10

2

10

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

418

CHAPTER 8

Frequency Response Methods

AP8.1

The spring-mass-damper system is described by
mẍ + bẋ + kx = p .
Taking the Laplace transform (with zero initial conditions) yields
X(s)
1
=
.
2
P (s)
ms + bs + k
From Figure AP8.1(b) in Dorf & Bishop, we determine that
20 log

1
X(j0)
= 20 log
= −26dB .
P (j0)
k

Solving for k yields
k = 19.96 N/m .
Also, ωn2 = k/m implies m = k/ωn2 , where ωn = corner frequency = 3.2
m = 1.949 kg .
Comparing Figure AP8.1(b) in Dorf & Bishop to the known standard
Bode plot of a second-order system, we estimate ζ ≈ 0.32. Therefore,
b = 2mζωn = 2(1.949)(0.32)(3.2) = 3.992 N − s/m .
AP8.2

The closed-loop transfer function is
T (s) =

Y (s)
Kb
=
.
R(s)
s + 1 + 0.2Kb

WIth K = 5, we have
T (s) =

5b
.
s+1+b

The sensitivity is
SbT =

∂T b
s+1
=
.
∂b T
s+1+b

With the nominal value of b = 4, we have
SbT =

s+1
.
s+5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

419

The sensitivity plot is shown in Figure AP8.2.

0

-2

20*log(mag) (dB)

-4

-6

-8

-10

-12

-14
10-1

100

101

102

FIGURE AP8.2
Bode plot for SbT (s) =

AP8.3

s+1
s+5 .

The equation of motion is
mẍ + bẋ + Kx = bṙ + Kr .
Taking Laplace transforms yields
X(s)
bs + K
=
.
2
R(s)
ms + bs + K
Then, given the various system parameters m = 1 kg, b = 4 Ns/m, K = 18
N/m, we obtain the transfer function:
X(s)
4s + 18
= 2
.
R(s)
s + 4s + 18
√
p
Also, ωn = corner frequency = K/m = 18 = 4.243 rad/s and
ζ = damping ratio =

b/m
4
=
= 0.471 .
2ωn
2(4.243)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

420

CHAPTER 8

Frequency Response Methods

The Bode plot is shown in Figure AP8.3.
10

Gain dB

0
-10
-20
-30
10-1

100

101

102

101

102

Phase deg

0
-50
-100
-150
-200
10-1

100

FIGURE AP8.3
Bode plot for G(s) =

The Bode plot is shown in Figure AP8.4.

Bode Diagram

Magnitude (dB)

−20
−40
−60
−80
−100
−120
0
−45
Phase (deg)

AP8.4

4s+18
.
s2 +4s+18

−90
−135
−180
−225
−270
−1
10

0

10

FIGURE AP8.4
Bode plot for L(s) =

1
.
(0.4s+1)(s2 +3.9s+15)

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

421

The closed-loop transfer function with unity feedback is given by
T (s) =

10(s + 1)
Gc (s)G(s)
= 2
.
1 + Gc (s)G(s)
s + 9s + 10

(a) Solving for Gc (s)G(s) yields
Gc (s)G(s) =

10(s + 1)
.
s(s − 1)

(b) A summary of the plot data (see Figure AP8.5) is presented in Table AP8.5.
(c) The open-loop system is unstable; the closed-loop system is stable.

40

30

20

20 log|GcG(j ω)|, dB

AP8.5

10

0

−10

−20

−30

−40
100

120

140

160

180
200
Phase, degrees

220

240

260

280

FIGURE AP8.5
Log-magnitude-phase curve for Gc G(jω).

ω

1

10

50

110

500

20 log |Gc G|

40

4.85

-13.33

-20.61

-33.94

phase (deg)

101.42

250.17

267.53

268.93

269.77

TABLE AP8.5

Summary of magnitude and phase for ω = 1, 10, 50, 110, 500.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

422

CHAPTER 8

The transfer function is given by
T (s) =

1/m
.
+ (b/m)s + (k/m)

s2

Selecting k = 1 and b = 2 results
in the Bode plot magnitude always
√
less than 0 dB. Choosing b = 2/2 leads to a peak response with a sinusoidal input at ω = 0.66 rad/s. Figure AP8.6a shows the Bode plot and
Figure AP8.6b shows the response to a sinusiodal input with frequency
ω = 1 rad/s is less than 1 in the steady-state, as desired.
Bode Diagram
10
System: sys
Peak gain (dB): 6.3

0

Magnitude (dB)

−10

−20

−30

−40

−50
−2
10

−1

0

10

1

10

10

Impulse Response
1

0.5

Amplitude

AP8.6

Frequency Response Methods

0

−0.5

−1

−1.5

0

100

200

300

400
Time (sec)

500

600

700

800

FIGURE AP8.6
(a) Bode plot for b/m = 1 and k/m = 1. (b) Response to a sinusiodal input.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

423

The transfer function is
G(s) =

Vo (s)
1 + R2 C 2 s
=
.
Vi (s)
1 + R1 C 1 s

Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields
G(s) =

0.01s + 1
.
0.001s + 1

The frequency response is shown in Figure AP8.7.

Bode Diagram

Magnitude (dB)

20
15
10
5
0
60
Phase (deg)

AP8.7

30

0
0
10

FIGURE AP8.7
Bode plot for G(s) =

1

10

0.01s+1
0.001s+1

2

3

10
10

4

10

5

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

424

CHAPTER 8

Frequency Response Methods

Design Problems
CDP8.1

With the PI controller in the loop, the closed-loop transfer function from
the input to the output is
26.035K(s + 2)
θ(s)
= 2
,
R(s)
s + (33.1415 + 26.035K)s + 52.07K
where we switch off the tachometer feedback (see Figure CDP4.1 in Dorf
and Bishop). The Bode plot is shown below for K = 40. From the step
response we determine that P.O. = 0 and Ts = 0.19. With K = 40, the
closed-loop poles are both real roots with values of s1 = −1072.6 and
s2 = −1.9.
60

Gain dB

40

20

0 -1
10

10

0

10

1

10

2

Phase deg

0

-30

-60

-90
10

-1

10

0

10

1

10

2

DP8.1

The loop transfer function is
L(s) = Gc (s)G(s) =

K(s + 2)
.
s2 (s + 12)

(a,b) Let K = 1. The Bode plot of the loop transfer function and the
closed-loop transfer functions are shown in Figure DP8.1a and Figure DP8.1b, respectively.
(c) Let K = 50. The Bode plot of the loop transfer function and the
closed-loop transfer functions are shown in Figure DP8.1c and Figure DP8.1d, respectively.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

425

Design Problems
50

Gain dB

0

-50

-100 -1
10

10

0

10

1

10

2

Phase deg

-120

-140

-160

-180 -1
10

10

0

10

1

10

2

FIGURE DP8.1
(a) Bode plot for the loop transfer function Gc (s)G(s) =

(s+2)
.
s2 (s+12)

50

Gain dB

0

-50

-100 -2
10

10

-1

0

10

10

1

10

2

Phase deg

0

-90

-180
10

-2

10

-1

0

10

FIGURE DP8.1
CONTINUED: (b) Bode plot for the closed-loop T (s) =

10

1

(s+2)
.
s3 +12s2 +s+2

10

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 8

Frequency Response Methods
100

Gain dB

50

0

-50 -1
10

10

0

10

1

10

2

Phase deg

-120

-140

-160

-180 -1
10

10

0

10

1

10

2

FIGURE DP8.1
CONTINUED: (c) Bode plot for the loop transfer function Gc (s)G(s) =

50(s+2)
.
s2 (s+12)

20

Gain dB

0
-20
-40
-60 -1
10

10

0

10

1

10

2

0

Phase deg

426

-90

-180
10

-1

10

0

10

1

10

FIGURE DP8.1
CONTINUED: (d) Bode plot for the closed-loop T (s) =

50(s+2)
.
s3 +12s2 +50s+100

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

427

Design Problems

(d) The peak value of Mp ≤ 2 occurs for 14 ≤ K ≤ 350. The maximum
bandwidth is achieved for the largest gain K. Thus, we select K = 350
and the corresponding bandwidth is ωB = 29 rad/sec.
(e) The system is type 2—the steady-state error is zero for a ramp input.
The open-loop transfer function is
20(s + 1)
.
s(s + 4)(s2 + 2s + 8)

Gc (s)G(s) =

(a) The phase angle is φ = −180o when ω = 3.54 rad/sec. The magnitude
is 0 dB when ω = 0.87 rad/sec.
(b) The closed-loop transfer function is
T (s) =

s4

+

6s3

20(s + 1)
.
+ 16s2 + 52s + 20

The closed-loop Bode plot is shown in Figure DP8.2.
Bode Diagram
Gm = 6.71 dB (at 3.54 rad/sec) , Pm = 105 deg (at 0.869 rad/sec)
20

Magnitude (dB)

0
−20
−40
−60
−80
−100
−45
−90
Phase (deg)

DP8.2

−135
−180
−225
−270
−1
10

0

1

10

10

2

10

FIGURE DP8.2
Bode plot for closed-loop T (s) =

20(s+1)
.
s4 +6s3 +16s2 +52s+20

(c) When K = 22, we have
Mpω = 4.84dB ,

ωr = 3.11 ,

and ωB = 3.78 rad/sec .

When K = 25, we have
Mpω = 7.18 dB ,

and ωB = 3.94 rad/sec .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

428

CHAPTER 8

Frequency Response Methods

(d) Select K = 22.
DP8.3

The closed-loop transfer function is
T (s) =

s3

+

7s2

K(s + 5)
.
+ 12s + 10 + 5K

When K = 4.2, we have 10 log 10 Mpω = 3 dB. The system bandwidth is
ωb = 3.7178 rad/sec. The steady-state tracking error to a unit step input
is
ess = lim sE(s) = lim 1 − T (s) .
s→0

s→0

So,
ess = 1 −

5K
= 0.322 ,
10 + 5K

when K = 4.2. Since the system is unstable when K > 14.8, the steadystate error does not exist after K = 14.8. The Bode plot is shown in
Figure DP8.3.
20

Gain dB

0
-20
-40
-60
-80
10-1

100

101

102

101

102

Phase deg

0
-50
-100
-150
-200
10-1

100

FIGURE DP8.3
Bode plot for closed-loop T (s) =

DP8.4

K(s+5)
,
s3 +7s2 +12s+10+5K

where K = 4.2.

We have a second-order loop transfer function
Gc (s)G(s) =

K
.
(0.3s + 1)(0.6s + 1)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

429

Design Problems

With Mpω = 1.5, we determine that
q

Mpω = (2ζ 1 − ζ 2 )−1

or

ζ = 0.3568 .

Now the characteristic equation is
s2 + 5s + 5.56(1 + K) = 0 .
So, solving 2ζωn = 5 yields ωn = 7. Therefore,
K = 0.18ωn2 − 1 = 7.82 .
The closed-loop transfer function is
T (s) =

K
5.56(K + 1)
.
2
K + 1 s + 5s + 5.56(K + 1)

So, the overall gain of the standard second-order system will be attenuated
by the factor K/(K + 1). To compensate, we amplify the gain by a small
factor. Thus we choose K = 10. The bandwidth is ωb = 11.25 rad/sec and
the peak magnitude is Mpω =1.5.
DP8.5

From the Bode plot of G(s) we find that there exists two pnoles, at approximately ω = 1 rad/sec and ω = 10 rad/sec. Then, by examining the
Bode plot we estimate
G(s) =

10
.
(s + 1)(s + 10)

We use a scale factor of 10 because at low frequency the Bode plot has
magnitude 0 dB (or a DC gain of 1). With G(s) as above, we can utilize
the controller
Gc (s) =

500
s + 20

yielding a crossover
and a magnitude of at least 25 dB for ω < 0.1 rad/sec. Figure DP8.5
shows the compensator Bode plot of Gc (s)G(s).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

430

CHAPTER 8

Frequency Response Methods
Bode Diagram

Phase (deg)

Magnitude (dB)

50

-50
-100
-150
0
-45
-90
-135
-180
-225
10

-2

10

-1

FIGURE DP8.5
Bode Diagram for G(s)Gc (s) =

DP8.6

ωc=12.9

25 dB

0

0

1

10
10

10

2

10

3

5000
.
(s+1)(s+10)(s+20)

Let K = −1 to meet the steady-state tracking error requirement and
p = 2ζ, where ζ = 0.69 to obtain a 5% overshoot. The system is given by
ẋ = Ax + Bu
where


A=

0

1

−1 −1.38



 ,



B=

The characteristic polynomial is

−1
0



 , and

C=



0 1



.

s2 + 1.38s + 1 = 0 .
The associated damping ratio is ζ = 0.69 and the natural frequency is
ωn = 1 rad/s. Using the approximation
ωb = (−1.19ζ + 1.85)ωn
we obtain ωb ≈ 1.028 rad/s. The Bode plot is shown in Figure DP8.6.
The bandwidth is ωb = 1.023 rad/s.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

431

Design Problems

Bode Diagram
20

Magnitude (dB)

0
−20
−40
−60
−80
0

Phase (deg)

−45

−90

−135

−180
−2
10

−1

0

10

10

1

10

2

10

FIGURE DP8.6
Bode diagram for K = −1 and p = 1.38.

DP8.7

A viable controler is
Gc (s) = KP +

3.33
KI
+ KD s = 5.5 +
+ 3.5s.
s
s

The loop transfer function is
Gc (s)G(s) =

10.5s2 + 16.5s + 10
s2 (s2 + 4s + 5)

and computing Ka yields
Ka = lim s2 Gc (s)G(s) =
s→0

10
= 2,
5

as desired. The phase margin is P.O. = 44.35◦ and the bandwidth is
ωb = 4.5 rad/sec. The step response is shown in Figure DP8.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 8

Frequency Response Methods

Step Response
1.4
System: sys_cl
Peak amplitude: 1.32
Overshoot (%): 32.1
At time (sec): 1.11

1.2

System: sys_cl
Settling Time (sec): 3.93

1
Amplitude

432

0.8

0.6

0.4

0.2

0

0

1

2

3
4
Time (sec)

FIGURE DP8.7
Step response for KP = 5.5, KI = 3.33, and KD = 3.5.

5

6

7

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

433

Computer Problems

Computer Problems
CP8.1

The m-file script and Bode plot are shown in Figure CP8.1. The script
automatically computes Mpω and ωr .
num=[25]; den=[1 1 25];
sys = tf(num,den);
w=logspace(0,1,400);
[mag,phase]=bode(sys,w);
[y,l]=max(mag);
mp=20*log10(y), wr=w(l)
bode(sys,w);
mp =
14.0228
wr =
4.9458

Bode Diagrams
From: U(1)
15

10

5

Phase (deg); Magnitude (dB)

0

-5

- 10
0

To: Y(1)

- 50

- 100

- 150

- 200
0
10

10

1

FIGURE CP8.1
Generating a Bode plot with the bode function.

CP8.2

The m-file script to generate the Bode plots is shown in Figure CP8.2a.
The Bode plots are presented in Figures CP8.2b-CP8.2e. The transfer
functions are
(a) : G(s) =

1000
;
(s + 10)(s + 100)

(b) : G(s) =

s + 100
;
(s + 2)(s + 25)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 8

Frequency Response Methods

(c) : G(s) =

s2

100
;
+ 2s + 50

(d) : G(s) =

s−6
.
(s + 3)(s2 + 12s + 50)

% Part (a)
num=[1000]; den=conv([1 10],[1 100]); sys1=tf(num,den);
sys = tf(num,den);
figure(1), bode(sys1), grid
% Part (b)
num=[1 100]; den=conv([1 2],[1 25]); sys2=tf(num,den);
sys = tf(num,den);
figure(2), bode(sys2), grid
% Part (c)
num=[100]; den=[1 2 50]; sys3=tf(num,den);
sys = tf(num,den);
figure(3), bode(sys3), grid
% Part (d)
num=[1 -6]; den=conv([1 3],[1 12 50]); sys4=tf(num,den);
sys = tf(sys);
figure(4), bode(sys4), grid

FIGURE CP8.2
(a) Script to generate the four Bode plots.

Bode Diagram

Magnitude (dB)

0
−20
−40
−60
−80
−100
0
Phase (deg)

434

−45
−90
−135
−180
−1
10

0

10

1

2

10
10

FIGURE CP8.2
CONTINUED: (b) Bode plot for G(s) =

1000
.
(s+10)(s+100)

3

10

4

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

435

Computer Problems

Bode Diagram

Magnitude (dB)

20
0
−20
−40
−60

Phase (deg)

−80
0

−45

−90

−135
−1
10

0

10

1

2

10
10

FIGURE CP8.2
CONTINUED: (c) Bode plot for G(s) =

3

4

10

10

s+100
.
(s+2)(s+25)
Bode Diagram

20

Magnitude (dB)

10
0
−10
−20
−30

Phase (deg)

−40
0
−45
−90
−135
−180
0
10

1

10

FIGURE CP8.2
CONTINUED: (d) Bode plot for G(s) =

100
s2 +2s+50 .

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

436

CHAPTER 8

Frequency Response Methods

Bode Diagram
−20

Magnitude (dB)

−30
−40
−50
−60
−70

Phase (deg)

−80
180
90
0
−90
−180
−1
10

0

1

10

2

10

10

FIGURE CP8.2
CONTINUED: (e) Bode plot for G(s) =

The Bode plots are shown in Figure CP8.3(a-d) with the transfer functions
listed in the caption. The crossover frequency for (a) is 17 rad/sec.

Bode Diagram
20

Magnitude (dB)

0
−20
−40
−60
−80
−100
0
Phase (deg)

CP8.3

s−6
.
(s+3)(s2 +12s+50)

−45
−90
−135
−180
−1
10

FIGURE CP8.3
(a) Bode plot for G(s) =

0

10

1

2

10
10

2000
.
(s+10)(s+100)

3

10

4

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

437

Computer Problems

The crossover frequency for (b) is 0.99 rad/sec.
Bode Diagram
20

Magnitude (dB)

0
−20
−40
−60
−80
0

Phase (deg)

−45
−90
−135
−180
−225
−270
−2
10

−1

10

0

10

FIGURE CP8.3
CONTINUED: (b) Bode plot for G(s) =

1

2

10

10

100
.
(s+1)(s2 +10s+2)

The crossover frequency for (c) is 70.7 rad/sec.
Bode Diagram
40

Magnitude (dB)

30
20
10
0
−10
−20

Phase (deg)

−30
0

−45

−90

−135
−2
10

−1

10

0

1

10
10

FIGURE CP8.3
CONTINUED: (c) Bode plot for G(s) =

50(s+100)
.
(s+1)(s+50)

The crossover frequency for (d) is 3.1 rad/sec.

2

10

3

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

438

CHAPTER 8

Frequency Response Methods

Bode Diagram
20

Magnitude (dB)

10
0
−10
−20
−30

Phase (deg)

−40
0

−45

−90
−1
10

0

1

10

FIGURE CP8.3
CONTINUED: (d) Bode plot for G(s) =

3

10

4

10

100(s2 +14s+50)
.
(s+1)(s+2)(s+500)

The m-file script and Bode plot are shown in Figure CP8.4a and b. The
bandwidth is ωb = 10 rad/sec.

10

Magnitude (dB)

0
−10
−20
−30
−40
−50
0
Phase (deg)

CP8.4

2

10
10

−45
−90
−135
−180
−1
10

0

10

FIGURE CP8.4
(a) Bode plot for T (s) =

54
s2 +6s+54 .

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

439

Computer Problems
numg=[54]; deng=[1 6 0];
sys_o = tf(numg,deng);
sys_cl = feedback(sys_o,[1])
wb=bandwidth(sys_cl)
bode(sys_cl), grid
title(titlename)

FIGURE CP8.4
CONTINUED: (b) M-file script to obtain the closed-loop Bode plot.

The Bode plot of the closed-loop system is shown in Figure CP8.5. The
closed-loop transfer function is
T (s) =

s2

100
.
+ 6s + 100

(a) From the Bode plot we determine that
Mpω ≈ 5 dB and

Bode Diagrams
From: U(1)
20

0

- 20

- 40

Phase (deg); Magnitude (dB)

- 60

- 80
0

- 50

To: Y(1)

CP8.5

- 100

- 150

- 200
-1
10

10

0

10

1

FIGURE CP8.5
Closed-loop system Bode plot.

10

2

10

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

440

CHAPTER 8

Frequency Response Methods

(b) From Equations (8.36) and (8.37) in Dorf & Bishop, we find that
ζ ≈ 0.28

and

ωr /ωn ≈ 0.92

which implies that
ωn = ωr /0.92 = 9.8 rad/sec .
(c) From T (s) we find that

and ζ = 0.3 .

The actual values and the estimated values compare very well.
The open-loop and closed-loop Bode plots are shown in Figure CP8.6a
and b. The open-loop and closed-loop transfers functions are
Gc (s)G(s) =

25
s3 + 3s2 + 27s + 25

and
T (s) =

Gc (s)G(s)
25
= 3
.
1 + Gc (s)G(s)
s + 3s2 + 27s + 50

Loop transfer function; bode(syso)

Magnitude (dB)

0
−20
−40
−60
−80
−100
0
−45
Phase (deg)

CP8.6

−90
−135
−180
−225
−270
−2
10

−1

10

0

10

FIGURE CP8.6
(a) Open-loop system Bode plot for Gc (s)G(s) =

1

10

25
.
s3 +3s2 +27s+25

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

441

Computer Problems

Closed−loop system; bode(syscl)

Magnitude (dB)

0
−20
−40
−60
−80
−100
0
Phase (deg)

−45
−90
−135
−180
−225
−270
−1
10

0

10

1

2

10

10

FIGURE CP8.6
CONTINUED: (b) Closed-loop system Bode plot T (s) =

CP8.7

25
.
s3 +3s2 +27s+50

The m-file script and plot of ωb versus p are shown in Figure CP8.7a and
b.
p=[0:0.001:1];
w=logspace(-1,1,1000);
n=length(p);
for i=1:n
num=[1]; den=[1 2*p(i) 0];
sys = tf(num,den);
sys_cl = feedback(sys,[1]);
[mag,phase,w]=bode(sys_cl,w);
a=find(mag<0.707); wb(i)=w(a(1));
end
plot(p,wb)

FIGURE CP8.7
(a) M-file script to generate plot of ωb versus p.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

442

CHAPTER 8

Frequency Response Methods

1.6
1.5
1.4

1.3
1.2
1.1
1
0.9
0.8
0.7
0.6

0

0.1

0.2

0.3

0.4

0.5
p

0.6

0.7

0.8

0.9

1

FIGURE CP8.7
CONTINUED: (b) Plot of ωb versus p.

CP8.8

The transfer function from Td (s) to θ(s) is
θ(s)/Td (s) =

s3

+

10s2

−0.01(s + 10)
.
+ (0.01K − 10.791)s − 107.91 + 0.05K

Using the final value theorem and Td (s) = 1/s, we determine that
lim sθ(s) =

s→0

−0.1
.
−107.91 + 0.05K

The design specifications require that
|ess | < 0.1o .
So, solving for K yields
K > 3300 .
We can select
K = 3300
as the initial value of K for the design. The m-file script is shown in
Figure CP8.8a. For the design shown, the final selection for the gain is
K = 6000. The disturbance response is shown in Figure CP8.8b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

443

Computer Problems

Mb=100; Ms=10; L=1; g=9.81; a=5; b=10;
%
K=6000; % Final design value of K
%
numg=[-1/Mb/L]; deng=[1 0 -(Mb+Ms)*g/Mb/L];
sysg = tf(numg,deng);
numc=-K*[1 a]; denc=[1 b];
sysc = tf(numc,denc);
%
% Part (a)
%
sys = feedback(sysg,sysc);
w=logspace(0,1,400);
bode(sys,w)
[mag,phase]=bode(sys,w);
[M,l]=max(mag);
MpDb=20*log10(M)-20*log10(mag(1)) % Mpw in decibels
wr=w(l) % Mpw and peak frequency
%
% Part (b)
%
% From Eqs. (8.35) and (8.37)
Mpw=10^(MpDb/20);zeta=sqrt((1-sqrt(1-(1/Mpw^2)))/2);
wn=wr/sqrt(1-2*zeta^2);
ts=4/zeta/wn
po=100*exp(-zeta*pi/sqrt(1-zeta^2))
%
% Part (c)
%
t=[0:0.1:10];
[y,x]=step(sys,t);
plot(t,y*180/pi)
xlabel('time [sec]')
ylabel('theta [deg]')
grid

MpDb =
4.0003
wr =
4.7226

meets specs

ts =
2.23
po =
32.75

0
-0.005
-0.01

theta [deg]

-0.015
-0.02
-0.025
-0.03
-0.035
-0.04

0

1

2

3

4

5

6

7

time [sec]

FIGURE CP8.8
(a) Design script. (b) Disturbance response - meets all specs!

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

444

CHAPTER 8

A viable filter is
G(s) = 0.7

(s + 1000)(s + 1)
.
(s + 100)(s + 10)

The Bode plot is shown in Figure CP8.9
Bode Diagram

Magnitude (dB)

20
15
10
5
0
-5
90
Phase (deg)

CP8.9

Frequency Response Methods

45
0
-45
-90
-2
10

10

0

10

2

FIGURE CP8.9
(s+1000)(s+1)
Bode plot for G(s) = 0.7 (s+100)(s+10)

.

10

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

9

Stability in the Frequency Domain

Exercises
The Bode plot for the transfer function Gc (s)G(s) is shown in Figure E9.1,
where
Gc (s)G(s) =

2(1 + s/10)
.
s(1 + 5s)(1 + s/9 + s2 /81)

The gain and phase margins are
P.M. = 17.5o .

G.M. = 26.2 dB and

Bode Diagram
Gm = 26.2 dB (at 2.99 rad/sec) , Pm = 17.5 deg (at 0.618 rad/sec)

Magnitude (dB)

50
0
−50
−100
−150
−90
−135
Phase (deg)

E9.1

−180
−225
−270
−315
−2
10

FIGURE E9.1
Bode Diagram for Gc (s)G(s) =

−1

10

0

10

1

10

2

10

2(1+s/10)
.
s(1+5s)(1+s/9+s2 /81)

445

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

446

CHAPTER 9

E9.2

Stability in the Frequency Domain

The loop transfer function is
Gc (s)G(s) =

10.5(1 + s/5)
.
s(1 + s/2)(1 + s/10)

The Bode plot is shown in Figure E9.2. The phase margin is
P.M. = 40.4o
Bode Diagram
Gm = Inf dB (at Inf rad/sec) , Pm = 40.4 deg (at 4.96 rad/sec)

Magnitude (dB)

50

0

−50

Phase (deg)

−100
−90

−135

−180
−1
10

0

10

FIGURE E9.2
Bode Diagram for Gc (s)G(s) =

1

10

2

10

3

10

10.5(1+s/5)
.
s(1+s/2)(1+s/10)

E9.3

The phase margin P.M. ≈ 75o at 200 kHz. We estimate the −180o phase
angle at 2 MHz, so the gain margin is G.M. ≈ 25 dB.

E9.4

The loop transfer function is
Gc (s)G(s) =

100
.
s(s + 10)

The Nichols diagram is shown in Figure E9.4. When the gain is raised by
4.6 dB, Mpω = 3 and the resonant frequency is ωR = 11 rad/sec.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

447

Exercises

40

0
0.25

30
0.5
1

20
3
6

10

Gain dB

-1
-3
-6

0
-10

K=171 ------

-12

------ K=100

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-40
0

-50

Phase (deg)

FIGURE E9.4
Nichols Diagram for Gc (s)G(s) =

K)
,
s(s+10)

where K = 100 and K = 171.

E9.5

(a) The G.M. ≈ 5 dB and the P.M. ≈ 10o . (b) Lower the gain by 10 dB
to obtain P.M. ≈ 60o .

E9.6

The Bode plot of the closed-loop transfer function is shown in Figure E9.6.
The value of Mpω = 3 dB. The phase margin is P.M. = 40o when K = 50.

5
0
-5
-10

Gain dB

-15
-20
-25
-30
-35
-40
-45
10-1

100

101

FIGURE E9.6
Closed-loop Bode Diagram for T (s) =

50(s+100)
s3 +50s2 +450s+5000 .

102

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

448

CHAPTER 9

E9.7

Stability in the Frequency Domain

The Nyquist plot is shown in Figure E9.7 for K = 5; the plot is a circle
with diameter= K/5. For K > 5, we have P = 1 and N = −1 (ccw as
Nyquist Diagram
0.5
0.4
0.3

Imaginary Axis

0.2
0.1
0
−0.1
−0.2
−0.3
−0.4
−0.5
−1

−0.8

−0.6

FIGURE E9.7
Nyquist Diagram for Gc (s)G(s) =

−0.4
−0.2
Real Axis

K
s−5 ,

0

0.2

0.4

where K = 5.

shown). So Z = N + P = −1 + 1 = 0 and the system is stable for K > 5.
(a) When K = 4, the G.M. = 3.5 dB. This is illustrated in Figure E9.8.
Bode Diagram
Gm = 3.52 dB (at 1.41 rad/sec) , Pm = 11.4 deg (at 1.14 rad/sec)

Magnitude (dB)

50

0

−50

−100

−150
−90

−135
Phase (deg)

E9.8

−180

−225

−270
−1
10

0

1

10

10

FIGURE E9.8
Bode Diagram for Gc (s)G(s) =

K
,
s(s+1)(s+2)

where K = 4.

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

449

Exercises

(b) The new gain should be K = 1 for a gain margin G.M. = 16 dB.
E9.9

For K = 5, the phase margin P.M. = 5o as shown in Figure E9.9.
Bode Diagram
Gm = 1.58 dB (at 1.41 rad/sec) , Pm = 5.02 deg (at 1.29 rad/sec)
100
Magnitude (dB)

50
0
-50
-100
-150
-90

Phase (deg)

-135
-180
-225
-270
-2
10

10

FIGURE E9.9
Bode Diagram for Gc (s)G(s) =

0

10

K
,
s(s+1)(s+2)

10

1

10

2

where K = 5.

The Bode plot is shown in Figure E9.10a. The closed-loop frequency
100

Gain dB

50
0
GM=12.35 dB

-50
-100
10-2

10-1

100

101

102

0

Phase deg

E9.10

-1

-100
PM=23.14 deg
-200
-300
10-2

10-1

FIGURE E9.10
(a) Bode Diagram for Gc (s)G(s) =

100

101

326s+1304
s4 +14.76s3 +151.3s2 +23.84s .

102

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

450

CHAPTER 9

Stability in the Frequency Domain

10
0
-10

Gain dB

-20
-30
-40
-50
-60
-70
10-1

100

101

102

FIGURE E9.10
CONTINUED: (b) Closed-loop frequency response: ωB = 6 rad/sec.

response is shown in Figure E9.10b. The bandwidth is ωB = 6 rad/sec.
The Bode plot is shown in Figure E9.11. The system is stable.

Bode Diagram
Gm = 3.91 dB (at 3.74 rad/sec) , Pm = 14.4 deg (at 2.76 rad/sec)

Magnitude (dB)

100

50

0

−50

−100
−90

−135
Phase (deg)

E9.11

−180

−225

−270
−2
10

−1

10

FIGURE E9.11
Bode Diagram for Gc (s)G(s) =

0

10

10(1+0.4s)
.
s(1+2s)(1+0.24s+0.04s2 )

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

451

Exercises

E9.12

We select the gain K = 10 to meet the 10% steady-state tracking error
specification for a ramp input. The Bode plot and Nichols chart are shown
in Figures E9.12a and E9.12b, respectively.

50

Gain dB

0
GM=14.82 dB
-50
-100
-150
10-1

100

101

102

103

102

103

Phase deg

0
-100
-200

PM=31.79 deg

-300
10-1

100

101

40

0
0.25

30
0.5
1

20
10

Gain dB

-1

3
6
8

-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)

FIGURE E9.12
(a) Bode Diagram for Gc (s)G(s) =
10
.
s(0.02s+1)(0.2s+1)

E9.13

10
.
s(0.02s+1)(0.2s+1)

(b) Nichols chart for Gc (s)G(s) =

(a) The Nichols diagram is shown in Figure E9.13a and Mpω = 7.97 dB.
(b) The closed-loop Bode plot is shown in Figure E9.13b. The bandwidth
ωB = 18.65 rad/sec and the resonant frequency is ωr = 11.69 rad/sec.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

452

CHAPTER 9

Stability in the Frequency Domain

40

0
0.25

30
0.5
1

20

Gain dB

10
0

-1

3
6

-3

8

-6
-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)
10

Gain dB

0
-10
-20
-30
-40
10-1

100

101

102

101

102

Phase deg

0
-50
-100
-150
-200
10-1

100

FIGURE E9.13
(a) Nichols Diagram for Gc (s)G(s) =
150
.
s2 +5s+150

150
.
s(s+5)

(b) Closed-loop Bode Diagram for T (s) =

(c) From Mpω = 8 dB, we estimate ζ = 0.2, so the expected P.O. = 52%.
E9.14

(a) The peak resonance Mpω = 6 dB.
(b) The resonant frequency is ωr = ω2 = 3 rad/sec.
(c) The bandwidth is ωB = ω4 = 10 rad/sec.
(d) The phase margin is P.M. = 30o .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

453

Exercises

E9.15

The loop transfer function is
Gc (s)G(s) =

100
,
s(s + 20)

and the closed-loop transfer function is
T (s) =

100
.
s2 + 20s + 100

The magnitude plot for the closed-loop system is shown in Figure E9.15.
With bandwidth defined as frequency at which the magnitude is reduced

Bode Diagram
0

−1

Magnitude (dB)

−2

−3

−4

−5

−6

−7
−1
10

0

10

FIGURE E9.15
Magnitude plot for the closed-loop T (s) =

1

10

100
.
s2 +20s+100

-3 dB from the dc value, we determine that ωB = 6.4 rad/sec.
E9.16

The transfer function of the approximation is
G(jω) =

1 − jω/10
,
1 + jω/10

and the magnitude is
|G(jω)| =

1 − jω/10
=1,
1 + jω/10

which is equivalent to the actual time delay magnitude. The phase ap-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

454

CHAPTER 9

Stability in the Frequency Domain

proximation is
φ = − tan−1 ω/10 + tan−1 (−ω/10) = −2 tan−1 ω/10
and the actual phase is
φ = −0.2ω .
The phase plots are shown in Figure E9.16. The approximation is accurate

Actual _______ & Approximation −−−−−−−
0

−20

Phase deg

−40

−60

−80

−100

−120
−2
10

−1

0

10

1

10

10

FIGURE E9.16
Phase plots for time delay actual vs approximation.

E9.17

(a,b) The phase angle for P.M. = 30 is
φ = −90o + tan−1

ω
2ω
− tan−1
= −150o .
2
15 − ω 2

Solving for ω yields ω = 4.7. Then, at ω = 4.7, we have K = 10.82
when
1

|Gc G(jω)| =

K(ω 2 + 4) 2
1

ω((2ω 2 )2 + (15 − ω 2 )2 ) 2

The Bode plot is shown in Figure E9.17.

=1.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

455

Exercises
Bode Diagrams
50

Phase (deg); Magnitude (dB)

0

- 50

- 100
- 50

- 100

- 150

- 200

- 250
-1
10

10

0

10

1

10

2

FIGURE E9.17
Bode Diagram for Gc (s)G(s) =

K(s+2)
,
s3 +2s2 +15s

where K = 10.82.

(c) The steady-state error for a ramp is
ess =

A
A
= 10K = 0.60A ,
Kv
15

where R(s) = A/s2 .
E9.18

(a) The gain crossover is at ωc = 486 Hz, and the phase margin P.O. =
36.2o . So, ζ ≈ 0.36. Then, the expected percent overshoot to a step
input is
√ 2
P.O. = 100e−ζπ/ 1−ζ = 30% , where ζ = 0.36 .
(b) The estimated bandwidth is ωB ≈ 2π(600).
(c) Approximate

ωn ≈ ωr = 2π(480) .
Then,
Ts =

4
4
=
≈ 4 ms .
ζωn
(0.36)2π(480)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

456

CHAPTER 9

E9.19

Stability in the Frequency Domain

The Bode plot is shown in Figure E9.19 for K = 16.75. The phase and
gain margins are P M = 50.0o and GM = 2.72 dB.
Gm=2.7233 dB (at 20.618
r
Pm=50 deg
r
10

0

- 10

Phase (deg); Mag
nitude (dB)

- 20

- 30

- 40
0

- 100

- 200

- 300

- 400

- 500
0
10

1

2

10

3

10

10

FIGURE E9.19
−0.1s
Bode Diagram for Gc (s)G(s) = K es+10 , where K = 16.75.

The system response for both drivers is shown in Figure E9.20.
T=1 sec (solid line) & T=1.5 sec (dashed line)
1
0
-1

Automobile velocity change

E9.20

-2
-3
-4
-5
-6
-7
-8

0

1

2

3

4

5

6

7

Time (sec)

FIGURE E9.20
Change in automobile velocity due to braking for two drivers.

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

457

Exercises

E9.21

The Bode plot is shown in Figure E9.21.
50

Gain dB

0
-50

GM=12.04 dB

-100
-150
10-1

100

101

102

103

102

103

Phase deg

0
-100
-200

PM=16.85 deg

-300
10-1

100

101

FIGURE E9.21
Bode Diagram for Gc (s)G(s) =

1300
.
s(s+2)(s+50)

E9.22

When K = 10, the P.M. = 36.9o ; the system is stable. Decreasing the
gain to K = 4 results in a P.M. = 60o .

E9.23

The Nichols chart is shown in Figure E9.23.

40

0
0.25

30
0.5
1

20
10

Gain dB

-1

3
6
8

-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

Phase (deg)

FIGURE E9.23
Nichols chart for Gc (s)G(s) =

438
.
s(s+2)(s+50)

-100

-50

-40
0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

458

CHAPTER 9

Stability in the Frequency Domain

The actual values are
Mpω = 1.6598 (4.4 dB) ωr = 2.4228 rad/sec ωB = 4.5834 rad/sec .
E9.24

Using the Nyquist criterion, we have
P = 1 and N = 0
which implies
Z = N +P = 1 .
Therefore, the system has one root in the right half-plane.

E9.25

The Bode plot is shown in Figure E9.25.

Gain dB

50

0

-50
-100
10-1

100

101

102

101

102

Phase deg

0
-100
-200
-300
10-1

100

FIGURE E9.25
Bode plot for Gc (s)G(s) =

E9.26

11.7
.
s(0.05s+1)(0.1s+1)

The Nichols chart for
Gc (s)G(s) =

11.7
s(0.05s + 1)(0.1s + 1)

is shown in Figure E9.26, where we find that
Mpω = 6.76 dB ωr = 8.96 rad/sec ωB = 13.73 rad/sec .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

459

Exercises

40

0
0.25

30
0.5
1

20
3
6
8

10

Gain dB

-1
-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)

FIGURE E9.26
Nichols chart for Gc (s)G(s) =

The Bode plot for G(s) with K = 122.62 is shown in Figure E9.27.

K=122.63

50

0

Phase (deg); Magnitude (dB)

E9.27

11.7
.
s(0.05s+1)(0.1s+1)

- 50

- 100
- 50

- 100

- 150

- 200

- 250

- 300
-1
10

10

0

10

FIGURE E9.27
Bode plot for Gc (s)G(s) =

K
,
s(s+6)2

with K = 122.62.

1

10

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

460

CHAPTER 9

Stability in the Frequency Domain

The phase margin is
P.M. = 40.0o
and the gain margin is
G.M. = 10.94 dB .
E9.28

The phase margin is P.M. = 28o . The estimated damping is
ζ=

P.M.
= 0.28 .
100

The estimated percent overshoot is
√
2
P.O. = 100e−πζ/ 1−ζ = 40% .
The actual overshoot is P.O. = 44.43%.
E9.29

The F (s)-plane contour is shown in Figure E9.29, where
F (s) = 1 + Gc (s)G(s) =

s+3
.
s+2

F(s)-plane
0.6
*

0.4

0.2

*

Im

*

0

*

*

*

-0.2

*

-0.4
*

-0.6

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

Re

FIGURE E9.29
F (s)-plane contour, where F (s) = 1 + Gc (s)G(s) =

E9.30

The Bode plot is shown in Figure E9.30.

s+3
s+2 .

1.8

1.9

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

461

Exercises
Bode Diagram

Magnitude (dB)

50
0

-50

Phase (deg)

-100
0
-45
-90
-135
-180
10

-2

10

0

10

2

10

4

FIGURE E9.30
Bode plot for G(s) = C [sI − A]−1 B + D =

The Bode plot is shown in Figure E9.31. The phase margin is P.M. = 50.6
deg.
Bode Diagram
Gm = Inf , Pm = 50.6 deg (at 0.341 rad/sec)

Magnitude (dB)

80
60
40
20
0
-20
-40
-90
Phase (deg)

E9.31

1000
.
s2 +100s+10

-120

-150
-3
10

10

-2

10

-1

FIGURE E9.31
Bode plot for L(s) = G(s)H(s) =

2s+1
10s2 +s .

10

0

10

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

462

CHAPTER 9

E9.32

Stability in the Frequency Domain

The Bode plot is shown in Figure E9.32. The phase margin is P.M. = 29◦ .
Bode Diagram
Gm = Inf dB (at Inf rad/sec) , Pm = 29 deg (at 3.1 rad/sec)

Magnitude (dB)

20
0
−20
−40
−60

Phase (deg)

−80
0
−45
−90
−135
−180
−1
10

0

1

10

2

10

10

FIGURE E9.32
Bode plot for G(s) = C [sI − A]−1 B + D =

The Bode plot is shown in Figure E9.33. The phase margin is P.M. =
17.7◦ and the gain margin is G.M. = 5.45 dB.
Bode Diagram
Gm = 5.45 dB (at 5.68 rad/sec) , Pm = 17.7 deg (at 4.24 rad/sec)

Magnitude (dB)

50

0

−50

−100

−150
0
−45
Phase (deg)

E9.33

6.4
s2 +s+4 .

−90
−135
−180
−225
−270
−1
10

FIGURE E9.33
Bode plot for L(s) =

0

10

1

10

200
.
(s2 +2.83s+4)(s+10)

2

10

3

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

463

Problems

Problems
P9.1

(a) The loop transfer function is
Gc (s)G(s) =

1
.
(1 + 0.5s)(1 + 2s)

P = 0, N = 0; therefore Z = N +P = 0. The system is stable. (Note:
See P8.1 for the polar plots.)
(b) The loop transfer function is
1 + 0.5s
.
s2
P = 0, N = 0, therefore Z = N + P = 0. The system is stable.
(c) The loop transfer function is
s2

s+4
.
+ 5s + 25

P = 0, N = 0, Z = N + P = 0. Therefore, the system is stable.
(d) The loop transfer function is
30(s + 8)
.
s(s + 2)(s + 4)
P = 0, N = 2 therefore Z = P + N = 2. Therefore, the system has
two roots in the right half-plane, and is unstable.
P9.2

(a) The loop transfer function is
Gc (s)G(s) =

s(s2

K
,
+ s + 6)

and
Gc (jω)G(jω) =

K
K[−ω 2 − jω(6 − ω 2 )]
==
.
jω(−ω 2 + jω + 6)
[(6 − ω 2 )2 ω 2 + ω 4 ]

To determine the real axis crossing, we let
Im{Gc (jω)G(jω)} = 0 = −Kω(6 − ω 2 )
or
ω=

√

6.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

464

CHAPTER 9

Stability in the Frequency Domain

Then,
Re{Gc (jω)G(jω)}ω=√ 6 =

−Kω 2
ω4

√
ω= 6

=

−K
.
6

So, −K/6 > −1 for stability. Thus K < 6 for a stable system.

(b) The loop transfer function is

Gc (s)G(s) =

K(s + 1)
.
s2 (s + 6)

The polar plot never encircles the -1 point, so the system is stable for
all gains K (See Figure 10 in Table 9.6 in Dorf & Bishop).
P9.3

(a,b) The suitable contours are shown in Figure P9.3.
jw

jw

Gs

Gs

q =cos z
q

r

r approaches
infinity

r

s

-s 1

r approaches
infinity
s

(b)

(a)

FIGURE P9.3
Suitable contours Γs for (a) and (b).

(c) Rewrite the characteristic equation as
1+

96
=0.
s(s2 + 11s + 56)

In this case, −σ1 = −1. Therefore, we have one pole inside the contour
at s = 0, so P = 1. The polar plot yields N = −1, so Z = N + P = 0.
Therefore, all three roots have real parts less than -1. In fact, the
roots are s1 = −3, and s2,3 = −4 ± j4.
P9.4

(a) P = 0, N = 2, therefore Z = 2. The system has two roots in the right
hand s-plane.
(b) In this case, N = +1 − 1 = 0, so Z = 0. Therefore the system is
stable.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

465

Problems

P9.5

(a) The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

K
.
(s + 1)(3s + 1)(0.4s + 1)

ess =

|R|
.
1+K

We require ess = 0.1|R|, so K > 9.
(b) Use K = 9. The Nyquist plot is shown in Figure P9.5. We determine
that P = 0 and N = 0. Therefore, Z = 0 and the system is stable.

8
6
4

Imag Axis

2
0
-2
-4
-6
-8
-2

0

2

4

6

8

10

Real Axis

FIGURE P9.5
Nyquist Diagram for L(s) = Gc (s)G(s)H(s) =

9
.
(s+1)(3s+1)(0.4s+1)

(c) The phase and gain margins are P.M. = 18o and G.M. = 5 dB.
P9.6

The rotational velocity transfer function is
ω(s)
= G(s) = 
R(s)
1+

K
s
3.7(2π)



s
68(2π)+1

 .

At low frequency, we have the magnitude near 35 dB, so 20 log K = 35 dB

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

466

CHAPTER 9

Stability in the Frequency Domain

and K = 56. Since the frequency response plot is for rotational velocity
ω(s), and we are interested in position control, we add an integrator. The
characteristic equation is
1
56(23)(427)
1 + G(s) = 1 +
=0.
s
s(s + 23)(s + 427)
The roots are
s1 = −430

and s2,3 = −10 ± j35 .

Thus, ωn = 36 and ζ = 0.28. The time constant of the closed-loop system
is
τ=
The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

10K1 s(s + 7)
.
(s + 3)(s2 + 0.36)

(a) The Bode plot is shown in Figure P9.7 for K1 = 2.

Gain dB

100
50
0
-50
10-1

100

101

102

101

102

100
50

Phase deg

P9.7

1
= 99.6 msec .
ζωn

0
-50
-100
-150
10-1

100

FIGURE P9.7
Bode Diagram for Gc (s)G(s)H(s) =

10K1 s(s+7)
,
(s+3)(s2 +0.36)

where K1 = 2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

467

Problems

(b) The phase margin P.M. = 80o and the gain margin G.M. = ∞, since
φ never crosses = −180o .
(c) The transfer function from Td (s) to θ(s) is
θ(s) =

G(s)
Td (s) .
1 + Gc (s)G(s)H(s)

Then, for a step disturbance θ(∞) = lims→0 sθ(s) = G(0) = 10/0.36 =
27.8, since H(0) = 0.
(d) The system is so highly damped, there is very little resonant peak.
(e) The estimated ζ = P.M./100 = 0.80. The actual ζ = 0.97.
(a) The loop transfer function is
Gc (s)G(s)H(s) =



s2
ω12

+

(0.02s + 1)

2ζ1 s
ω1



s2
ω22



+1

+

2ζ2 s
ω2

 ,

+1

where ω1 = 20π = 62.8, ω2 = 14π = 43.9, ζ1 = 0.05 and ζ2 = 0.05.
The Bode plot is shown in Figure P9.8a. The phase margin is P.M. =
−9o . Therefore, the system is unstable.

Gain dB

20

0

-20
-40
100

101

102

103

102

103

0

Phase deg

P9.8

-50
-100
-150
-200
100

101

FIGURE P9.8
(a) Bode Diagram for Gc (s)G(s)H(s) =
and ω2 = 14π.

s2 /ω12 +(0.1/ω1 )s+1
,
(0.02s+1)(s2 /ω22 +(0.1/ω2 )s+1)

where ω1 = 20π

(b) In this case ζ2 = 0.25, with all other parameters the same as before.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

468

CHAPTER 9

Stability in the Frequency Domain

10

Gain dB

0
-10
-20
-30
-40
100

101

102

103

102

103

Phase deg

0
-50
-100
-150
-200
100

101

FIGURE P9.8
CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) =
where ω1 = 20π and ω2 = 14π.

s2 /ω12 +(0.1/ω1 )s+1
,
(0.02s+1)(s2 /ω22 +(0.5/ω2 )s+1)

The Bode plot is shown in Figure P9.8b. The phase margin is P.M. =
86o . Therefore, the system is now stable.
P9.9

(a) The Bode plot is shown in Figure P9.9a
The phase margin is P.M. = 83o and the gain margin is G.M. = ∞.
(b) With the compensator, the loop transfer function is
Gc (s)G(s)H(s) = K1

0.30(s + 0.05)(s2 + 1600)(s + 0.5)
,
s(s2 + 0.05s + 16)(s + 70)

where
K2 /K1 = 0.5 .
Let K1 = 1. The Bode plot is shown in Figure P9.9b. The phase
margin is P.M. = 80o and the gain margin is G.M. = ∞, essentially
the same as in (a). But the system in (b) is a type one, so that ess = 0
to a step input or disturbance. We cannot achieve a G.M. = 10 dB
by increasing or decreasing K1 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

469

Problems

40

Gain dB

20
0
-20
-40
-60
10-3

10-2

10-1

100

101

102

103

101

102

103

102

103

102

103

Phase deg

0
-100
-200
-300
10-3

10-2

10-1

100

FIGURE P9.9
(a) Bode Diagram for Gc (s)G(s)H(s) =

0.3(s+0.05)(s2 +1600)
.
(s+70)(s2 +0.05s+16)

40

Gain dB

20
0
-20
-40
-60
10-3

10-2

10-1

10-2

10-1

100
101

Phase deg

0
-100
-200
-300
10-3

100

101

FIGURE P9.9
CONTINUED: (b) Bode Diagram for Gs (s)G(s)H(s) =
K1 = 1.

0.15K1 (s+0.05)(s2 +1600)(s+0.5)
,
(s+70)(s2 +0.05s+16)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

470

CHAPTER 9

The equations of motion are
F (s) = 3I(s)

and I(s) =

Eo (s)
Eo (s)
=
.
R + Ls
0.1 + 0.2s

So,
F (s) =

30
Eo (s) .
(2s + 1)

The actuator without the spring (see Table 2.7, Number 9 in Dorf &
Bishop) is modeled via
X(s)
1
Ka
=
=
.
2
Y (s)
M s + Bs
τa s 2 + s
With the spring, we have
Ka
X(s)
=
2
Y (s)
τa s + s + Ks

or

GA (s) =

0.4s2

1
.
+ s + 1.5

Then, the loop transfer function is
L(s) =

30K1
.
(2s + 1)(0.4s2 + s + 1.5)

(a) The Bode plot for K1 = 0.2 in Fig. P9.10 shows the P.M. = 30o .
20

Gain dB

0
-20
-40
-60
10-2

10-1

100

101

100

101

0

Phase deg

P9.10

Stability in the Frequency Domain

-100
-200
-300
10-2

10-1

FIGURE P9.10
Bode Diagram for L(s) =

30K1
,
(2s+1)(0.4s2 +s+1.5)

where K1 = 0.2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

471

Problems

(b) For K1 = 0.2, we determine that Mpω = 7.8 dB, ωr = 1.9 rad/sec,
(c) The estimated percent overshoot is P.O. = 51% and the estimated
settling time is Ts = 10 sec. This is based on ζ = 0.21 and ωn ≈ ωr =
The loop transfer function is
Gc (s)G(s) =

5(K1 s + K2 )e−1.5s
.
s(5s + 1)

(a) Let K1 = K2 = 1. Then
Gc (s)G(s) =

5(s + 1) −1.5s
e
.
s(5s + 1)

The Bode plot is shown in Figure P9.11a. The phase margin is P.M. =
−48o . The system is unstable.
(b) Let K1 = 0.1 and K2 = 0.04. Then, the loop transfer function is
Gc (s)G(s) =

5(0.1s + 0.04)e−1.5s
.
s(5s + 1)

The Bode plot shown in Figure P9.11b shows P.M. = 45o . Thus, the
system is stable.
60

Gain dB

40
20
0
-20
10-2

10-1

100

101

100

101

0
-200

Phase deg

P9.11

-400
-600
-800

-1000
10-2

10-1

FIGURE P9.11
(a) Bode Diagram for Gc (s)G(s) =

5(s+1)e−sT
s(5s+1)

, where T = 1.5.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

472

CHAPTER 9

Stability in the Frequency Domain

40

Gain dB

20
0
-20
-40
10-2

10-1

100

101

100

101

0

Phase deg

-200
-400
-600
-800
-1000
10-2

10-1

FIGURE P9.11
CONTINUED: (b) Bode Diagram for Gc (s)G(s) =

5(0.1s+0.04)e−sT
s(5s+1)

, where T = 1.5.

(c) When K2 = 0.1394, the phase margin is P.M. = 0o and G.M. = 0
dB. So, for stability we require K2 ≤ 0.1394 when K1 = 0.
(a) The Bode plot is shown in Figure P9.12.

Bode Diagram
Gm = 12 dB (at 3.46 rad/sec) , Pm = 67.6 deg (at 1.53 rad/sec)
20

Magnitude (dB)

0
−20
−40
−60
−80
−100
0
−45
Phase (deg)

P9.12

−90
−135
−180
−225
−270
−1
10

0

10

FIGURE P9.12
Bode Diagram for Gc (s)G(s) =

2
.
(0.5s+1)3

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

473

Problems

The loop transfer function (without the time delay) is
Gc (s)G(s) =

2
.
(0.5s + 1)3

The phase margin is P.M. = 67.6o .
(b) With the delay, the loop transfer function is
Gc (s)G(s)H(s) =

2e−0.5s
.
(0.5s + 1)3

The phase margin is now P.M. = 23.7o . So the 0.5 sec time delay has
reduced the phase margin by 43.9◦ .
The loop transfer function is
Gc (s)G(s) =

Ka (Ks + 1) −1.2s
e
.
s

(a) Let Ka = K = 1. Without the time delay, the system has infinite
phase and gain margin. However, with the time delay, the system has
a negative gain margin, hence it is unstable.
(b) A plot of phase margin versus Ka is shown in Figure P9.13.

100
80
60

Phase margin deg

P9.13

40
20
0
-20
-40
-60
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Ka

FIGURE P9.13
Phase margin as a function of Ka for Gc (s)G(s) =

Ka (s+1)e−1.2s
.
s

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

474

CHAPTER 9

Stability in the Frequency Domain

Let K = 1, and find Ka for a stable system. Then,
Gc (s)G(s) =

Ka (s + 1)e−1.2s
.
s

If Ka = 0.8, then the phase margin is P.M. = 50o .
The loop transfer function is
Gc (s)G(s) =

Ke−0.2s
.
s(0.1s + 1)

(a) The Nichols diagram is shown in Figure P9.14 for K = 2.5.

40

0
0.25

30
0.5
1

20

-1

2
3
6

10

Gain dB

P9.14

-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)

FIGURE P9.14
Nichols diagram for Gc (s)G(s) =

Ke−0.2s
,
s(0.1s+1)

for K = 2.5.

It can be seen that
Mpω = 2.0 dB .
The phase and gain margins are P.M. = 48.5o and G.M. = 7.77 dB.
(b) We determine that ζ = 0.43 (based on Mpω = 2 dB) and ζ = 0.48
(based on the phase margin P.M. = 48.5o ).
(c) The bandwidth is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

475

Problems

(a) The ship transfer function is
G(s) =

−0.164(s + 0.2)(s − 0.32)
.
s2 (s + 0.25)(s − 0.009)

The closed-loop system is unstable; the roots are
s1 = −0.5467
s2,3 = 0.2503 ± 0.1893j
s4 = −0.1949
Therefore the ship will not track the straight track.
(b) The system cannot be stabilized by lowering the gain; this is verified
in the root locus in Figure P9.15, where it is seen that the locus has
a branch in the right half-plane for all K > 0.
(c) Yes, the system can be stabilized.
(d) When the switch is closed, we have a derivative feedback, which adds
90o phase lead. This is not enough to stabilize the system. Additional

0.6

0.4

0.2

Imag Axis

P9.15

0

x

o

xx

-0.2

0

o

-0.2

-0.4

-0.6
-0.6

-0.4

0.2

Real Axis

FIGURE P9.15
−0.164(s+0.2)(s−0.32)
Root locus for 1 + GH(s) = 1 + K s2 (s+0.25)(s−0.009) = 0.

0.4

0.6

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

476

CHAPTER 9

P9.16

Stability in the Frequency Domain

The loop transfer function is
Gc (s)G(s) =

K
.
(s/10 + 1)(s2 + s + 2)

When K = 3.2, the phase margin is P.M. ≈ 30o . The Bode plot is shown
in Figure P9.16.

Gm=10.88 dB, (w= 3.464) Pm=29.91 deg. (w=2.083)
50

Gain dB

0

-50

-100 -1
10

10

0

10

1

10

2

0

Phase deg

-90
-180
-270
-360 -1
10

10

0

10

1

10

2

FIGURE P9.16
Bode plot for Gc (s)G(s) =

P9.17

K
,
(s/10+1)(s2 +s+2)

where K = 3.2.

(a) We require ess ≤ 0.05A, and we have
ess =

A
< 0.05A
1 + Kp

or Kp > 19. But
20K1
s→0 (0.5s + 1)

Kp = lim G1 (s)G2 (s)G3 (s)G4 (s) = lim
s→0

So, Kp = 0.2K1 > 19, or K1 > 95.



0.1
1 + 4s

2

= 0.2K1 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

477

Problems

(b) Given
1
s+1
G1 (s) = K1 (1 + ) = K1
s
s




,

we require 1.05 < MPt < 1.30, or 0.70 > ζ > 0.36, or 70o > P.M. >
36o . Then,
G1 (s)G2 (s)G3 (s)G4 (s) =

0.2K1 (s + 1)
.
s(0.5s + 1)(4s + 1)2

When K1 = 0.8, the P.M. = 40o . The Bode plot is shown in Figure P9.17a.

Gain dB

50

0

-50

-100
10-2

10-1

100

101

100

101

Phase deg

0
-100
-200
-300
10-2

10-1

FIGURE P9.17
(a) Bode plot for G1 (s)G2 (s)G3 (s)G4 (s) =
P.M. = 40o .

0.2K1 (s+1)
,
s(0.5s+1)(4s+1)2

where K1 = 0.8 and

(c) For part (a), we had
G1 (s)G2 (s)G3 (s)G4 (s) =

2.375
.
(s + 2)(s + 0.25)2

The characteristic equation is
s3 + 2.5s2 + 1.06s + 2.50 = (s + 2.48)(s2 + 0.02s + 1.013) .
The dominant complex roots are lightly damped since ζ = 0.01 and
ζωn = 0.01.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 9

Stability in the Frequency Domain

Thus,
Ts =

4
= 400 sec .
ζωn

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

(0.2)(0.8)(s + 1)
.
s(0.5s + 1)(4s + 1)2

The characteristic equation is
8s4 + 20s3 + 8.5s2 + 1.16s + 0.16 = 0 .
The roots are s1 = −2, s2 = −0.4 and s3,4 = −0.05 ± j0.15. Thus
ζ = 0.16 and ζωn = 0.05. So,
Ts =

4
4
=
= 75 sec .
ζωn
0.05

(d) Let U (s) be a unit step disturbance and R(s) = 0. Then
Y (s)
G3 (s)G4 (s)
=
=
U (s)
1 + G1 (s)G2 (s)G3 (s)G4 (s)
1+

2
0.1
1+4s
20K1 (s+1)
s(0.5s+1)(4s+1)2


The disturbance response is shown in Figure P9.17b.
x10 -3
6
5
4
3

Amplitude

478

2
1
0
-1
-2

0

10

20

30

40

50

60

70

Time (secs)

FIGURE P9.17
CONTINUED: (b) System response to a unit disturbance U (s).

80

90

100

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

479

Problems

P9.18

The transfer function is
Gc (s)G(s)H(s) =

5.3(s2 + 0.8s + 0.32)e−T s
.
s3

The Bode plot is shown in Figure P9.18.
80

Gain dB

60
40
20
0
-20
10-1

100

101

Phase deg

300

T=0 solid ___ & T=0.1 dashed ---- & T=0.2674 dotted ....

250
200
150
100
10-1

100

101

FIGURE P9.18
K(s2 +0.8s+0.32)e−sT
, where T = 0 (solid line),
Bode diagram for Gc (s)G(s)H(s) =
s3
T = 0.1 (dashed line), and T = 0.2674 (dotted line).

The following results are verified in the figure.
(a) The phase margin is P.M. = 81o at ω = 5.3 when T = 0.
(b) For T = 0.1, the added phase is φ = −T ω (in radians). The phase
margin is P.M. = 51o at ω = 5.3 when T = 0.1.
(c) The system is borderline stable when T = 0.2674 sec. The phase
margin is P.M. = 0o at ω = 5.3.
P9.19

The transfer function is
Gc (s)G(s) =

0.5
.
s(1 + 2s)(4 + s)

(a) The Nichols diagram is shown in Figure P9.19. The gain margin is
G.M. = 31.4 dB.
(b) The phase margin is P.M. = 75o and Mpω = 0 dB. The bandwidth is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

480

CHAPTER 9

Stability in the Frequency Domain

Phase margin = 180-105=75o

Nichols Chart
40
0.25 dB
0.5 dB
1 dB
3 dB
6 dB

20
0

0 dB sys
System:
Gain (dB):
Phase (deg):

AM IW
APR IW
ARN IW

ARN
AVN

Open

AMN

AMN IW

ATN

ATN IW

APNN

APNN IW

APRN

APRN IW

APVN

APVN IW

APMN
ALMN

ALPQ

ARS0

ARRQ
APTN
APLQ
O>?@ABCC> DEFG? HI?JK

FIGURE P9.19
Nichols diagram for Gc (s)G(s) =

AUN

Gain margin = 31.4 dB

APMN IW
0
AVQ

0.5
.
s(2s+1)(s+4)

(a) Let K = 100. The Bode plot is shown in Figure P9.20a. The loop
transfer function is
Gc (s)G(s) =

K(s2 + 1.5s + 0.5)
.
s(20s + 1)(10s + 1)(0.5s + 1)

Gain dB

100
50
0
-50
10-3

10-2

10-1

100

101

100

101

0

Phase deg

P9.20

AVN IW

System: sys
Gain (dB):
4
Phase (deg):

-100
-200
-300
10-3

10-2

10-1

FIGURE P9.20
(a) Bode diagram for Gc (s)G(s) =

K(s2 +1.5s+0.5)
,
s(20s+1)(10s+1)(0.5s+1)

where K = 100.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

481

Problems

(b) The phase margin is P.M. = −3.5o and the gain margin is G.M. = 2.7
dB.
(c) You must decrease K below 100 to achieve a P.M. = 40o . For K = 0.1,
the phase margin P.M. = 37.9o .
(d) The step response is shown in Figure P9.20b for K = 0.1.

1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

50

100

150

200

250

300

350

400

Time (secs)

FIGURE P9.20
CONTINUED: (b) Unit step response K = 0.1.

P9.21

The loop transfer function is
Gc (s)G(s) =

K
.
s(s + 1)(s + 4)

(a) The Bode plot is shown in Figure P9.21 for K = 4.
(b) The gain margin is
G.M. = 14 dB .
(c) When K = 5, the gain margin is
G.M. = 12 dB .
(d) We require Kv > 3, but Kv = K4 . So, we need K > 12. This gain can
be utilized since K < 20 is required for stability.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

482

CHAPTER 9

Stability in the Frequency Domain

50

Gain dB

0
-50
-100
-150
10-1

100

101

102

101

102

Phase deg

0
-100
-200
-300
10-1

100

FIGURE P9.21
Bode diagram for Gc (s)G(s) =

P9.22

K
,
s(s+1)(s+4)

where K = 4.

(a) The resonant frequency ωr = 5.2 rad/sec is point 6 on the Nichol’s
chart.
(b) The bandwidth is between points 8 and 9. We estimate the bandwidth
to be ωB = 7.5 rad/sec.
(c) The phase margin P.M. = 30o .
(d) The gain margin G.M. = 8 dB.
(e) Since we have P.M. = 30o , then we estimate ζ = 0.3. We can also
approximate
ωn ≈ ωr = 5.2 .ap9.1
Thus,
Ts =

P9.23

4
4
=
= 2.5sec .
ζωn
1.56

The phase margin is P.M. = 60 deg when K = 266. The gain margin is
G.M. = 17.2 dB .
The Bode plot is shown in Figure P9.23.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

483

Problems
Bode Diagram
Gm = 17.2 dB (at 9.8 rad/sec) , Pm = 60 deg (at 2.58 rad/sec)

Magnitude (dB)

50
0
−50
−100

Phase (deg)

−150
−90
−135
−180
−225
−270
−1
10

0

10

FIGURE P9.23
Bode diagram for Gc (s)G(s) =

P9.24

1

2

10

K
,
s(s+8)(s+12)

10

3

10

where K = 266.

When K = 14.1, then P.M. = 45 deg, G.M. = ∞ dB and ωB = 29.3
100

Phase (deg); Magnitude (dB)

50

0

- 50
- 80

- 100

- 120

- 140

- 160

- 180
-1
10

10

0

10

1

10

2

FIGURE P9.24
Bode diagram for G(s) =

P9.25

K(s+20)
,
s2

where K = 14.1.

The phase margin is P.M. = 60 deg when K = 2.61 and T = 0.2 second.
The Bode plot is shown in Figure P9.25.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

484

CHAPTER 9

Stability in the Frequency Domain
40

Gain dB

20
0
-20
-40 -1
10

10

0

10

1

10

2

Phase deg

0

-500

-1000

-1500 -1
10

10

0

10

1

10

2

FIGURE P9.25
Bode diagram for Gc (s)G(s) =

P9.26

Ke−0.2s
,
s

where K = 2.61.

The loop transfer function is
Gc (s)G(s) =

K
.
s(0.25s + 1)(0.1s + 1)

The Bode plot is shown in Figure P9.26a for K = 10. The Nichols chart
is shown in Figure P9.26b. The phase and gain margins are
P.M. = 9o

and

G.M. = 3 dB .

The system bandwidth is ωB = 8 rad/sec. From the P.M. = 9o , we
estimate ζ = 0.09. Therefore, the predicted overshoot is
√

−πζ/

P.O. = 100e

1−ζ 2

= 75% , where ζ = 0.09 .

The resonant peak occurs at ωr = 5.5 rad/sec. If we estimate ωn ≈ ωr =
5.5 rad/sec, then the settling time is
Ts =

4
= 8 sec .
ζωn

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

485

Problems

Gain dB

50

0

-50
-100
10-1

100

101

102

101

102

Phase deg

0
-100
-200
-300
10-1

100

40

0
0.25

30
0.5
1

20
10

Gain dB

-1

2
3

-3

16

0

-6
-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)

FIGURE P9.26
(a) Bode diagram for Gc (s)G(s) =
for Gc (s)G(s) =

P9.27

K
,
s(0.25s+1)(0.1s+1)

K
,
s(0.25s+1)(0.1s+1)

where K = 10. (b) Nichols chart

where K = 10.

The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

4K
.
(s2 + 2s + 4)(s + 1)

The plot of the phase margin versus the gain K is shown in Figure P9.27.
As the gain increases towards Kmax = 3.5, the phase margin decreases

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

486

CHAPTER 9

Stability in the Frequency Domain

towards zero.

180

160

140

Phase margin (deg)

120

100

80

60

40

20

0

1

1.5

2

2.5

3

3.5

K

FIGURE P9.27
Phase margin versus the gain K.

P9.28

The loop transfer function is
Gc (s)G(s) =

KP
.
s(s + 1)

When KP = 1.414, we have P.M. ≈ 45◦ . Using the approximation that
ζ ≈ P.M./100 we estimate that ζ = 0.45. Then using the design formula
√ 2
P.O. = 100e−πζ/ 1−ζ = 20.5% .
The actual overshoot is 23.4%. The step input response is shown in Figure P9.28. The actual damping ratio is ζ = 0.42. This shows that the
approximation ζ ≈ P.M./100 is quite applicable and useful in predicting
the percent overshoot.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

487

Problems

Step Response
1.4

1.2

System: syscl
Peak amplitude: 1.23
Overshoot (%): 23.3
At time (sec): 2.97

Amplitude

1

0.8

0.6

0.4

0.2

0

0

5

10
Time (sec)

FIGURE P9.28
Step response showing a 23.3% overshoot.

15

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

488

CHAPTER 9

Stability in the Frequency Domain

The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

236607.5(s + 10)(s + 5)
.
s(s + 2)(s2 + 100s + ωn2 )(s + 1)

(a) The Bode plot for
ωn2 = 15267
is shown in Figure AP9.1a.
150

Gain dB

100
50
0
-50
-100
10-3

10-2

10-1

10-2

10-1

100
101

102

103

102

103

0

Phase deg

AP9.1

-100
-200
-300
10-3

100

101

FIGURE AP9.1
(a) Bode Diagram for L(s) =

236607.5(s+10)(s+5)
2 )(s+1) ,
s(s+2)(s2 +100s+ωn

2
where ωn
= 15267.

The phase and gain margins are
P.M. = 48.6o

and

G.M. = 15.5 dB .

(b) The Bode plot for ωn2 = 9500 is shown in Figure AP9.1b. The gain
and phase margins are
P.M. = 48.5o

and

G.M. = 10.9 dB .

Reducing the natural frequency by 38% has the effect of reducing the
gain margin by 30%.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

489

150

Gain dB

100
50
0
-50
-100
10-3

10-2

10-1

100

101

102

103

101

102

103

Phase deg

0
-100
-200
-300
10-3

10-2

10-1

100

FIGURE AP9.1
CONTINUED: (b) Bode Diagram for L(s) =

2
where ωn
= 9500.

(a) The Bode plot with T = 0.05 sec is shown in Figure AP9.2a. The
phase margin is P.M. = 47.7o and the gain margin is G.M. = 11.2
dB.
40

Gain dB

20
0
-20
-40
100

101

102

-100

Phase deg

AP9.2

236607.5(s+10)(s+5)
2 )(s+1) ,
s(s+2)(s2 +100s+ωn

-200

-300
-400
100

101

FIGURE AP9.2
(s+5)
(a) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.05s.

102

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

490

CHAPTER 9

Stability in the Frequency Domain

(b) The Bode plot with T = 0.1 sec is shown in Figure AP9.2b. The

40

Gain dB

20
0
-20
-40
100

101

102

Phase deg

0
-200
-400
-600
-800
100

101

102

FIGURE AP9.2
(s+5)
CONTINUED: (b) Bode Diagram for Gc (s)G(s)H(s) = 8 s(s+2) e−sT , where T = 0.1s.

phase margin is P.M. = 22.1o and the gain margin is G.M. = 4.18
dB. A 100% increase in time delay T leads to a 50% decrease in phase
and gain margins.
(c) The damping ratio ζ ≈ P.M./100 and

√
2
P.O. ≈ 100e−πζ/ 1−ζ .

So, for T = 0.05 sec, ζ ≈ 0.47 and P.O. ≈ 18.7%. Also, for T = 0.1
sec, ζ ≈ 0.22 and P.O. ≈ 49.2%.
AP9.3

The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

66K(1 + 0.1s)
.
(1 + 0.01s)(1 + 0.01s)(1 + 1.5s)(1 + 0.2s)

(a) When K = 1, the gain and phase margins are
G.M. = 18.4 dB and P.M. = 55o .
(b) When K = 1.5, the gain and phase margins are
G.M. = 14.9 dB and P.M. = 47.8o .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

491

(c,d) The bandwidth and settling time with K = 1 are ωB = 233.6
rad/sec and Ts = 0.4 second. When K = 1.5, we determine that
ωB = 294.20 rad/sec and Ts = 0.33 second.
AP9.4

The loop transfer function is
L(s) = Gc (s)G(s) =

K(s + 40)
.
s(s + 15)(s + 10)

The gain K = 28.8 satisfies the specifications. The actual gain and phase
margins are
G.M. = 18.8 dB and P.M. = 45o .
The system bandwidth is ωB = 10.3 rad/sec. The step response is shown
in Figure AP9.4.

Step Response

System: sys_cl
Peak amplitude: 1.23
1.4
Overshoot (%): 23.4
At time (sec): 0.476

1.2

Amplitude

1
System: sys_cl
Settling Time (sec): 1.1

0.8

0.6

0.4

0.2

0

0

0.2

0.4

0.6

0.8
1
1.2
Time (sec)

1.4

1.6

1.8

2

FIGURE AP9.4
Closed-loop system step response.

AP9.5

The loop transfer function is
L(s) = Gc (s)G(s) = K

s + 0.4
.
s4 + 9s3 + 18s2

The Bode plot for K = 1 is shown in Figure AP9.5. From the phase
response, we determine that the maximum P.M. ≈ 41o . From the magnitude response (for K = 1), we find that the gain needs to be raised to
K = 14 to achieve maximum phase margin at ω = 0.826 rad/sec. The

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

492

CHAPTER 9

Stability in the Frequency Domain

gain and phase margin with K = 14 are
G.M. = 19.3 dB and P.M. = 40.9o .
Also, the overshoot is P.O. = 38.3%.
Bode Diagram
Gm = 42.3 dB (at 3.79 rad/sec) , Pm = 16.7 deg (at 0.154 rad/sec)

Magnitude (dB)

50

0

−50

−100

Phase (deg)

−150
−135
System: sys
Phase (deg): −139

−180

−225

−270
−2
10

−1

10

0

10

1

10

2

10

FIGURE AP9.5
s+0.4
Bode plot for L(s) = K s4 +9s
3 +18s2 with K = 1.

AP9.6

With D > 2m, the gain can be increased up to K = 100, while still
retaining stability.

AP9.7

The loop transfer function is
L(s) = Gc (s)G(s) =

K(s + 4)
.
s2

We select
√
K=2 2
for P.M. = 45o . The system bandwidth is
The disturbance response is shown in Figure AP9.7. The maximum output
due to a disturbance is y(t) = 0.11.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

493

0.12

0.1

Amplitude

0.08

0.06

0.04

0.02

0
0

0.5

1

1.5

2

2.5

3

3.5

4

Time (secs)

FIGURE AP9.7
Closed-loop system disturbance response.

A reasonable choice for the gain is K = 2680. The phase margin is P.M. =
42.8◦ and the percent overshoot is P.O. = 18.9%. The Nichols chart is
shown in Figure AP9.8.

Nichols Chart
60
40
0.25 dB
0.5 dB
1 dB
3 dB
6 dB

20
Open−Loop Gain (dB)

AP9.8

0 dB
−1 dB

−20

−3 dB
−6 dB
−12 dB
−20 dB

−40

−40 dB

−60

−60 dB

−80

−80 dB

−100

−100 dB

0

−120
−360

FIGURE AP9.8
Nichols chart.

−315

−270

−225
−180
−135
Open−Loop Phase (deg)

−90

−120 dB
−45
0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

494

CHAPTER 9

AP9.9

Stability in the Frequency Domain

The loop transfer function is
L(s) = Gc (s)G(s) =

Kp (s + 0.2)
2
s (s2 + 7s + 10)

.

At the maximum phase margin, Kp = 4.9 for P.M. = 48.6o . The Bode
diagram is shown in Figure AP9.9.
Bode Diagrams
100

50

0

Phase (deg); Magnitude (dB)

-50

-100

-150
-100

-150

-200

-250

-300
-3
10

10

-2

10

-1

10

0

10

1

10

2

FIGURE AP9.9
Phase and gain margin.

AP9.10

The closed-loop transfer function is
T (s) =

s2

K
.
+ 3s + 1

We require K = 1 a zero steady-state tracking error to a unit step. The
step response is shown in Figure AP9.10. Computing T (jω) = 0.707 it
follows that
(jω)2

1
= 0.707
+ 3jω + 1

or ω 4 + 7ω 2 − 1 = 0 .

Solving for ω yields ω = 0.37 rad/s. This is the bandwidth of the system.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

495

Step Response
1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

0

2

4

6

8
Time (sec)

10

12

14

16

FIGURE AP9.10
Unit step response.

The phase margin versus time delay is shown in Figure AP9.11a.

80
Time Delay=1

70

PM=58.5285

Time Delay=3.0455

60

Phase Margin (deg)

AP9.11

PM=0.001

50
40
30
20
10
0
−10
0.5

1

FIGURE AP9.11
Phase margin versus time delay.

1.5

2
Time Delay (s)

2.5

3

3.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 9

Stability in the Frequency Domain

The maximum time delay is T = 3.04 s for stability. The step response is
shown in Figure AP9.11b. The percent overshoot is P.O. = 7.6%.

1.4

1.2

1

Amplitude

496

0.8

0.6

0.4

0.2

0

0

FIGURE AP9.11
Unit step response.

1

2

3

4

5
6
Time (sec)

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

497

Design Problems

Design Problems
CDP9.1

The plant model with parameters given in Table CDP2.1 in Dorf and
Bishop is given by:
26.035
θ(s)
=
,
Va (s)
s(s + 33.142)
where we neglect the motor inductance Lm and where we switch off the
tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). The closedloop system characteristic equation is
1+

26.035Ka
=0.
s(s + 33.142)

The phase margin is P.M. = 70.4◦ when Ka = 16. The step response with
K = 16 is shown below.
1.2

1

Amplitude

0.8

0.6

0.4

0.2

0

DP9.1

0

0.05

0.1

0.15
Time (secs)

0.2

0.25

0.3

(a) The gain and phase margins are G.M. = 7 dB and P.M. = 60o .
(b) The resonant peak and frequency are Mpω = 2 dB and ωr = 5 rad/sec.
(c) We have ωB = 20 rad/sec. From Mpω = 2 dB we estimate ζ = 0.45
(Figure 8.11 in Dorf & Bishop). Also, ωr /ωn = 0.8, so ωn = 6.25.
Thus, Ts = 1.4.
(d) We need P.O. = 30o or ζ = 0.3 or P.M. ≈ 30o . So, we need to raise
the gain by 10 dB or K = 3.2.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

498

CHAPTER 9

The loop transfer function is
L(s) = Gc (s)G(s) =

K(s + 0.5)
.
+ 7.5s + 9)

s2 (s2

When K = 6.25, we have the maximum phase margin. The phase margin maximum is P.M. = 23o . The plot of P.M. versus K is shown in
Figure DP9.2a.

24
22
20
18
Phase Margin deg

DP9.2

Stability in the Frequency Domain

16
14
12
10
8
6
4

0

1

2

3

4

5

6

7

8

9

K

FIGURE DP9.2
(a) Phase margin versus K for L(s) =

K(s+0.5)
.
s2 (s2 +7.5s+9)

The predicted damping is ζ = 0.23. It then follows that the predicted
percent overshoot is
√
2
P.O. = 100e−πζ/ 1−ζ = 48% .
The actual overshoot is 65%. The step input response is shown in Figure DP9.2b. The resonant peak occurs at ωr = 0.75 rad/sec. Approximating ωn ≈ ωr = 0.75 rad/sec, we can estimate the settling time as
Ts =

4
= 23 sec .
ζωn

The actual settling time is 20 sec.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

499

Design Problems

1.8
1.6
1.4

Amplitude

1.2
1
0.8
0.6
0.4
0.2
0
0

5

10

15

20

25

30

Time (secs)

FIGURE DP9.2
CONTINUE: (b) Closed-loop unit step response.

We want to select the gain K as large as possible to reduce the steady-state
error, but we want a minimum phase margin of P.M. = 45o to achieve
good dynamic response. A suitable gain is K = 4.2, see Figure DP9.3.
20

Gain dB

0

-20

-40 -2
10

10

-1

10

0

10

1

0
-100
Phase deg

DP9.3

-200
-300
10

-2

10

-1

10

FIGURE DP9.3
Bode plot for G(s) =

Ke−10s
40s+1 .

0

10

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

500

CHAPTER 9

DP9.4

Stability in the Frequency Domain

We are given the loop transfer function
L(s) = Gc (s)G(s) =

K
s(s + 1)(s + 4)

which can be written as
Gc (s)G(s) =

Kv
.
s(s + 1)(0.25s + 1)

The performance results are summarized in Table DP9.4.

Kv

TABLE DP9.4

G.M.

P.M.

ωB

P.O.

Ts

(dB)

(deg)

(%)

(sec)

0.40

21.9

64.2

0.62

4.4

9.8

0.75

16.5

49.0

1.09

19.0

10.1

Summary for Kv = 0.40 and Kv = 0.75.

When Kv = 0.40, we have
ess
1
=
= 2.5 ,
A
0.40
or 2 1/2 times the magnitude of the ramp. This system would be acceptable for step inputs, but unacceptable for ramp inputs.
DP9.5

(a) With a time delay of T = 0.8 second, we determine that the proportional controller
Gc (s) = K = 7
provides a suitable response with
P.O. = 8.3 % ess = 12.5 %

Ts = 4.38 sec .

(b) A suitable proportional, integral controller is
Gc (s) = K1 + K2 /s = 6 + 0.6/s .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

501

Design Problems

The response to a unit step is
P.O. = 5.14 %

ess = 0 % Ts = 6.37 sec .

The Nichols chart is shown in Figure DP9.5.

40

0
0.25

30
0.5
1

20
3
6
8

10

Gain dB

-1
-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

-50

-40
0

Phase (deg)

FIGURE DP9.5
Nichols chart for Gc (s)G(s) =

DP9.6

(K1 s+K2 )e−0.8s
,
s(10s+1)

where K1 = 6 and K2 = 0.6.

With K = 170, at the two extreme values of b, we have
b = 80
b = 300

P.M. = 91.62o
P.M. = 75.23o

G.M. = 13.66 dB
G.M. = 25.67 dB .

Since reducing the value of K only increases the P.M. and G.M., a value
of
K = 170
is suitable to meet P.M. = 40o and G.M. = 8 dB for the range of b.
DP9.7

A suitable gain is
K = 0.22 .
This results in P.M. = 60.17o and G.M. = 13.39 dB. The step reponse is
shown in Figure DP9.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

502

CHAPTER 9

Stability in the Frequency Domain

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

18

20

Time (secs)

FIGURE DP9.7
Lunar vehicle step response.

A gain of K = 315000 will satisfy the P.O. specification, while giving the
fastest response. The step response is shown in Figure DP9.8.

1.2

1

0.8

Amplitude

DP9.8

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

0.5
Time (secs)

FIGURE DP9.8
Steel rolling mill step response.

0.6

0.7

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

503

Design Problems

The closed-loop transfer function is
Ts (2) =

Gc (s)G2 (s)
G1 (s)
To (s) +
T2d (s) .
1 + Gc (s)G2 (s)
1 + Gc (s)G2 (s)

where
1
(10s + 1)(50s + 1)

G1 (s) =
and
G2 (s) =

0.01
.
(10s + 1)(50s + 1)

The steady-state error (with Gc (s) = 500) to a unit step 2A (and after
the system has settled out subsequent to a step of magnitude A) is
ess = 2(0.167) = 0.33 .
The step response is shown in Figure DP9.9.
Gc=500 (solid); Gc=1/s (dashed); Gc=600+6/s (dotted)
2.5

2

1.5

T2/A

DP9.9

1

0.5

0
0

200

400

600

800

1000

Time (sec)

FIGURE DP9.9
Two tank temperature control step response.

A suitable integral controller is
Gc (s) =

1
.
s

1200

1400

1600

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

504

CHAPTER 9

Stability in the Frequency Domain

In this case, the steady-state tracking error is zero , since the system is a
type 1. The system response is shown in Figure DP9.9. With the integral
controller, the settling time is about Ts = 438 seconds and the P.O. = 7%.
A suitable PI controller is
Gc (s) = 600 +

6
.
s

With the PI controller, the settling time is about Ts = 150 seconds and
the P.O. = 10%.
DP9.10

The system is given by
ẋ = Ax + Br
y = Cx
where


A=

0

1

2 − K1 3 − K2





 ,

B=

0
1

The associated transfer function is
T (s) =

s2



 , and

C=



1 0



.

1
.
+ (K2 − 3)s + K1 − 2

The characteristic polynomial is
s2 + (K2 − 3)s + K1 − 2 = 0 .
If we select K1 = 3, then we have a zero-steady error to a unit step
response R(s) = 1/s, since
s2 + (K2 − 3)s
=0.
s→0 s2 + (K2 − 3)s + K1 − 2

lim s [1 − T (s)] R(s) = lim

s→0

Let
K=



3 4.3



.

The step response is shown in Figure DP9.10a. The bandwidth is ωb =
1.08 rad/s, as seen in Figure DP9.10b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

505

Design Problems

Step Response
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

1

2

3

4
5
Time (sec)

6

7

8

9

Bode Diagram
5

0
System: sys
Magnitude (dB): −3

−5

Magnitude (dB)

−10

−15

−20

−25

−30

−35

−40
−1
10

0

10

1

10

FIGURE DP9.10
Step response with K = [3 4.3] and closed-loop Bode plot.

DP9.11

A time domain step response specification P.O. > 10% requires the dominant poles to have a damping ration of ζ = 0.6. This time domain specification can be transformed to a frequency response specification using
the approximation P.M. ≈ 100ζ = 60◦ . To keep the problem tractable,
we consider the controller with the form
Gc (s) = KP +

KI
1
= KP + ,
s
s

where we let KI = 1. The plot of the P.M. as a function of KP is shown

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 9

Stability in the Frequency Domain

in Figure DP9.11a. If we select KP = 0.07 we expect a phase margin
of approximately 60◦ , hence a percent overshoot P.O. ≤ 10%. The step
response is shown in Figure DP9.11b. The actual phase margin is P.M. =
60.2◦ , the percent overshoot is P.O. = 5.9% and the settling time is
Ts = 3.4 sec.

85

Phase Margin (deg)

80

75

70

65

60

55

0

0.05

0.1

0.15

0.2

0.25
KP

0.3

0.35

0.4

0.45

0.5

1.4

1.2

1

Amplitude

506

0.8

0.6

0.4

0.2

0

0

1

2

3
Time (sec)

4

5

6

FIGURE DP9.11
(a) Phase margin versus controller gain KP and KI = 1. (b) Step response with KP = 0.07
and KI = 1.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

507

Computer Problems

Computer Problems
The m-file script to generate the Bode plot (from which the gain and
phase margin can be determined) is shown in Figure CP9.1. The transfer
function is
G(s) =

s2

141
.
+ 2s + 12

The gain margin is
G.M. = ∞
and the phase margin is
P.M. = 10o .

num=141; den=[1 2 12];
sys = tf(num,den);
margin(sys);
Bode Diagram
Gm = Inf dB (at Inf rad/sec) , Pm = 10 deg (at 12.3 rad/sec)

Magnitude (dB)

40
20
0
−20
−40
0
Phase (deg)

CP9.1

−45
−90
−135
−180
−1
10

0

10

FIGURE CP9.1
Gain and phase margin with the margin function.

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

508

CHAPTER 9

The Nyquist plots are shown in Figures CP9.2a-c.

num=[5]; den=[1 5]; sys=tf(num,den); nyquist(sys)
0.5
0.4
0.3

Imaginary Axis

0.2
0.1
0
−0.1
−0.2
−0.3
−0.4
−0.5
−1

−0.8

−0.6

FIGURE CP9.2
(a) Nyquist plot for G(s) =

−0.4

−0.2

0
Real Axis

0.2

0.4

0.6

0.8

1

5
s+5 .

num=[50]; den=[1 10 25]; sys=tf(num,den); nyquist(sys)
1.5

1

Imaginary Axis

CP9.2

Stability in the Frequency Domain

0.5

0

−0.5

−1

−1.5
−1

−0.5

0

0.5
Real Axis

FIGURE CP9.2
CONTINUED: (b) Nyquist plot for G(s) =

50
s2 +10s+25 .

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

509

Computer Problems

num=[15]; den=[1 3 3 1]; sys=tf(num,den); nyquist(sys)
15

10

Imaginary Axis

5

0

−5

−10

−15
−5

0

5
Real Axis

FIGURE CP9.2
CONTINUED: (c) Nyquist plot for G(s) =

CP9.3

10

15

15
s3 +3s2 +3s+1 .

The m-file script to generate the Nichols chart for part (a) is shown in
Figure CP9.3a. The Nichols charts for (b) and (c) are similiarly generated;
all plots are in Figure CP9.3a-c.

Nichols Chart
40
0 dB
30

0.25 dB

num = [1]; den = [1 0.2];
sys = tf(num,den);
nichols(sys)
ngrid

Open−Loop Gain (dB)

0.5 dB
20

1 dB

−1 dB

3 dB

10

−3 dB

6 dB
0

−6 dB

−10

−12 dB

−20 dB
−20
−360

FIGURE CP9.3
(a) M-file script and Nichols chart for G(s) =

−315

1
s+0.1 .

−270

−225
−180
−135
Open−Loop Phase (deg)

−90

−45

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 9

Stability in the Frequency Domain

The gain and phase margin for each transfer function are as follows:
(a) G.M. = ∞ and P.M. = 102o
(b) G.M. = ∞ and P.M. = ∞

(c) G.M. = 20 dB and P.M. = ∞

Nichols Chart
40
0 dB
30

0.25 dB
0.5 dB

Open−Loop Gain (dB)

20

1 dB

−1 dB

3 dB
6 dB

10

−3 dB

0

−6 dB

−10

−12 dB

−20

−20 dB

−30
−40 dB

−40
−50
−60
−360

−315

−270

−225
−180
−135
Open−Loop Phase (deg)

FIGURE CP9.3
CONTINUED: (b) Nichols chart for G(s) =

−90

−45

−60 dB
0

1
s2 +2s+1 .

Nichols Chart
40
0 dB
0.25 dB
0.5 dB
1 dB

20

−1 dB

3 dB
6 dB

−3 dB
−6 dB

0
Open−Loop Gain (dB)

510

−12 dB
−20

−20 dB

−40

−40 dB

−60

−60 dB

−80

−80 dB

−100
−360

−315

−270

−225
−180
−135
Open−Loop Phase (deg)

FIGURE CP9.3
CONTINUED: (c) Nichols chart for G(s) =

−90

24
s3 +9s2 +26s+24 .

−45

−100 dB
0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

511

Computer Problems

CP9.4

To obtain a phase margin P.M. = 40◦ we select
K = 15
when T = 0.2 second. The variation in the phase margin for 0 ≤ T ≤ 0.3
is shown in Figure CP9.4.
T=[0:0.01:0.3]; K=15;
num=K;den=[1 12]; sys = tf(num,den);
% w=logspace(-2,1,400);
for i=1:length(T)
[mag,phase,w]=bode(sys);
ph(1:length(phase))=phase(1,1,:); ph=ph';
ph2=ph-w*T(i)*180/pi;
[Gm,Pm,Wcg,Wcp]=margin(mag,ph2,w);
clear ph ph2
PMo(i)=Pm;
end
plot(T,PMo), grid
xlabel('Time delay (sec)')
ylabel('Phase margin (deg)')
K=15
160
140

Phase margin (deg)

120
100
80
60
40
20
0
−20

0

0.05

0.1

0.15
0.2
Time delay (sec)

0.25

0.3

FIGURE CP9.4
Variation in the phase margin for 0 ≤ T ≤ 0.3 with K = 15.

CP9.5

The loop transfer function is
L(s) = Gc (s)G(s) =

K(s + 50)
.
s(s + 20)(s + 10)

0.35

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

512

CHAPTER 9

Stability in the Frequency Domain

The plot of system bandwidth versus the gain K is shown in Figure CP9.7.

K=[0.1:1:50];
w=logspace(-2,3,2000);
den=[1 30 200 0];
for i=1:length(K)
num=K(i)*[1 50]; sys = tf(num,den);
sys_cl = feedback(sys,[1]);
[mag,phase,w]=bode(sys_cl,w);
L=find(mag<0.707); wb(i)=w(L(1));
end
plot(K,wb), grid
xlabel('Gain K')

15

10

5

0

0

5

10

15

20

25
Gain K

30

35

40

45

50

FIGURE CP9.5
Variation in the system bandwidth for 0 ≤ K ≤ 50.

CP9.6

The m-file script and Bode plot are shown in Figure CP9.6. The gain and
phase margin and ωc are determined to be G.M. = 2.23, P.M. = 26o and
ωc = 12.6 rad/sec. So, the maximum value of bo is found to be
bomax = 2.13bo = 1.11 .
In this problem, there is also a minimum value of bo . Using the Routh-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

513

Computer Problems
gm =
2.2238

numg = -0.5*[1 0 -2500]; deng = [1 47 850 -3000];
sysg = tf(numg,deng);
numc = 10*[1 3]; denc = [1 0]; sysc = tf(numc,denc);
sys_o = series(sysc,sysg);
bode(sys_o)
[mag,phase,w] = bode(sys_o);
[gm,pm,wg,wc] = margin(mag,phase,w)

pm =
26.3187
wg =
26.1155
wc =
12.6487

Gain dB

50

0

-50
10-1

100

101

102

103

102

103

Phase deg

250
200
150
100
50
10-1

100

101

FIGURE CP9.6
Using the margin function to compute stability margins.

Hurwitz method, we determine that (for stability) the range of bo is
0.14 < bo < 1.11 .
CP9.7

The m-file script is shown in Figure CP9.7a. Since we do not have a value
for J, we write the loop transfer function as
Gc (s)G(s) =

K̄1 + K̄2 s
s2

where K̄1 = K1 /J and K̄2 = K2 /J. We work with K̄1 and K̄2 , then we
can always compute K1 and K2 whenever J is specified. A PD controller
which meets the specs is given by
Gc (s) = 0.04 + 0.3s .
The step response is shown in Figure CP9.7b. The Bode plot is shown in

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

514

CHAPTER 9

Stability in the Frequency Domain

%
% Part (a)
%
numc = [0.3 0.04]; denc = [1]; sysc = tf(numc,denc);
numg = [1]; deng = [1 0 0]; sysg = tf(numg,deng);
sys_o = series(sysc,sysg);
sys_cl = feedback(sys_o,[1]);
step(sys_cl), pause
%
% Part (b)
%
w = logspace(-1,1,400);
[mag,phase] = bode(sys_o,w);
[gm,pm,w1,w2] = margin(mag,phase,w);
margin(mag,phase,w), pause
%
% Part (c)
%
T = [1:0.1:5];
for i = 1:length(T)
[numd,dend] = pade(T(i),2); sysd = tf(numd,dend);
sys_o1 = series(sysd,sys_o);
sys_cl1 = feedback(sys_o1,sysd);
p(:,i) = pole(sys_cl1);
end
plot(real(p),imag(p),'*');grid
xlabel('Real Axis');
ylabel('Imag Axis')

FIGURE CP9.7
Script to assist in all three parts of the problem.

Figure CP9.7c. The phase margin is P.M. = 67.7o at ω = 0.32 rad/sec.
The loop transfer function is
Gc (s)G(s)H(s) =

K̄1 + K̄2 s −2T s
e
s2

where T is the one-way time delay. If the phase lag introduced by the
delay is greater than 67.7o at ω = 0.32 rad/sec, then the system will
become unstable. So, since the phase lag due to the time delay T̃ is
φ(ω) = ω T̃
we have 67.7o π/180 = 0.32(2T ) where T̃ = 2T . Solving for T yields
T = 1.82 seconds. This is the maximum allowable one-way time delay.
Executing the third part of the m-file script in Figure CP9.7a generates
the plot illustrating the movement of the closed-loop system roots as the
time delay is varied. The plot is shown in Figure CP9.7d. Examining
the root locations, we find that when T = 1.9, the closed-loop roots

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

515

Computer Problems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

5

10

15

20

25

30

35

40

Time (secs)

FIGURE CP9.7
CONTINUED: (b) Step response without time delays meets specs.

Gain dB

20

0
-20

-40
10-1

100

101

Phase deg

0
-100
-200
-300
10-1

100

101

FIGURE CP9.7
CONTINUED: (c) System Bode plot shows P.M. = 67.7o .

are s1 = −4.56, s2,3 = −0.94 ± 2.02j, s4 = −0.19, and s5,6 = ±0.32j.
Therefore, the system is marginally stable when T = 1.9, and is unstable
as the time delay increases.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

516

CHAPTER 9

Stability in the Frequency Domain

4
*
*

3

*

*

*

*

*

2

Imag Axis

1
0

*

*

*

*

*

**
**
**
**
***
***
****
***
***
*
***********
**
******
**
**********

* * * * * * * * * * * * * * **********************

-1
-2
-3

*

*

*

*

*

*

*
***
***
****
*
*
**
***
**
**
**

*
*

-4
-8

-7

-6

-5

-4

-3

-2

-1

0

1

Real Axis

FIGURE CP9.7
CONTINUED: (d) Closed-loop root locations as the time delay varies.

CP9.8

The Nyquist plot and associated m-file code are shown in Figure CP9.8.

Nyquist Diagram
150

a=[0 1;-1 -15]; b=[0;30]; c=[8 0]; d=[0];
sys=ss(a,b,c,d);
nyquist(sys)

Imaginary Axis

100

50

0

−50

−100

−150
−50

FIGURE CP9.8
Using the Nyquist function to obtain a Nyquist plot.

0

50

100
Real Axis

150

200

250

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

517

Computer Problems

CP9.9

The Nichols chart is shown in Figure CP9.9. The phase and gain margins
are 37.1 degrees and ∞ dB, respectively.

a=[0 1;-1 -10]; b=[0;22]; c=[10 0]; d=[0];
sys=ss(a,b,c,d);
nichols(sys)
ngrid
Nichols Chart
60

Open-Loop Gain (dB)

40
0.25 dB
0.5 dB
1 dB
3 dB
6 dB

20

0 dB
?-1 dB
?-3 dB
?-6 dB

0

?-12 dB
-20

?-20 dB

-40

?-40 dB

-60
-360

?-60 dB
-315

-270

-225

-180

-135

-90

-45

0

Open-Loop Phase (deg)

FIGURE CP9.9
The Nichols chart for the system in CP9.8.

CP9.10

(a) The Nyquist plot is shown in Figure CP9.10. The phase margin is
P.M. = 18o .
(b) When the time delay is T = 0.05 seconds, the phase margin is P.M. =
9o .
(c) When the time delay is T = 0.1 seconds, the system is marginally
stable. So, for T > 0.1 seconds, the system is unstable.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 9

Stability in the Frequency Domain

Nyquist Diagram
200
150
100
Imaginary Axi s

518

50

-1 point

0
-50
-100
-150
-200
-10

-8

-6

-4
Real Axi s

FIGURE CP9.10
Nyquist plot for G(s)H(s) =

10
.
s(s+1)

-2

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

1 0

The Design of Feedback Control
Systems

Exercises
E10.1

From the design specifications, we determine that our desired ζ = 0.69
and ωn = 5.79. The characteristic equation is
1 + Gc (s)G(s) = 1 +

K(s + a)
=0,
s(s + 2)

or
s2 + (2 + K)s + Ka = 0 .
Our desired characteristic polynomial is
s2 + 2ζωn s + ωn2 = s2 + 8s + 33.6 = 0 .
Thus, K + 2 = 8, or
K=6
and Ka = 33.6, so a = 5.6. The actual percent overshoot and settling
time will be different from the predicted values due to the presence of the
closed-loop system zero at s = −a. In fact, the actual percent overshoot
and settling time are P.O. = 12.6% and Ts = 0.87s, respectively.
E10.2

The characteristic equation is
400
1
1 + Gc (s)G(s) = 1 +
K1 +
s(s + 40)
s




=1+

400(K1 s + 1)
=0,
s2 (s + 40)

or
1 + K1

400s
=0.
s3 + 40s2 + 400
519

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

520

CHAPTER 10

The Design of Feedback Control Systems

We desire ζ = 0.45 for an overshoot of 20%. The root locus is shown
in Figure E10.2. We select a point slightly inside the performance region
(defined by ζ = 0.45 ) to account for the zero. Thus,
K1 = 0.5
and the closed-loop poles are
s1 = −35
s2,3 = −2.7 ± j2 .
The actual
P.O. = 20.7% .

50
40
30

Imag Axis

20
10
0

x

*
*

*

x
o
x

-10
-20
-30
-40
-50
-50

-40

-30

-20

-10

0

Real Axis

FIGURE E10.2
400s
Root locus for 1 + K1 s3 +40s
2 +400 = 0.

E10.3

The step response is shown in Figure E10.3 for τ = 1 and K = 0.5. It can
be seen that the
P.O. = 4% ,
so this is a valid solution.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

521

Exercises

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0

-0.2

1

0

2

3

4

5

6

7

8

Time (secs)

FIGURE E10.3
Step response for K = 0.5 and τ = 1.

The Bode plot is shown in Figure E10.4. The phase and gain margins
are marked on the plot, where it can be seen that P.M. = 75.4o and
G.M. = 28.6 dB.
Bode Diagram
Gm = 28.6 dB (at 11.8 rad/sec) , Pm = 75.4 deg (at 0.247 rad/sec)

Magnitude (dB)

150
100
50
0
-50
-100
-150
-45
Phase (deg)

E10.4

-90
-135
-180
-225
-270
-4
10

10

-2

10

0

FIGURE E10.4
Bode plot for Gc (s)G(s) =

100(s+0.15)(s+0.7)
.
s(s+5)(s+10)(s+0.015)(s+7)

10

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

522

CHAPTER 10

E10.5

The Design of Feedback Control Systems

We require that Kv ≥ 2.7, ζ = 0.5 and ωn = 3 for the dominant roots.
We want to place a zero to left of the pole at -2, so the complex roots will
dominate. Set the zero at s = −2.2. Then for the desired roots find the
location of pole p in compensator
Gc (s) =

K1 (s + 2.2)
(s + p)

to satisfy 180o phase at the desired roots. This yields p = 16.4. Using
root locus methods, we find that KK1 = 165.7, so with K1 = 7.53, we
determine that K = 22, and
Gc (s) =

7.46(s + 2.2)
.
(s + 16.4)

Then
Kv = 2.78 .
E10.6

The closed-loop transfer function is
T (s) =

326(s + 4)
Gc (s)G(s)
= 4
.
3
1 + Gc (s)G(s)
s + 14.76s + 151.3s2 + 349.8s + 1304

The roots are
s1,2 = −0.87 ± j3.2
s3,4 = −6.5 ± j8.7 .
Assuming s1,2 dominates, then we expect overshoot
P.O. = 43%

and Ts = 4.6 sec .

The discrepencies with the actual P.O. and Ts are due to the poles s3,4
and the zero at s = −4.
E10.7

The loop transfer function is
L(s) =

Ke−0.6s
.
s(s + 20)

A plot of P.M. as a function of K is shown in Figure E10.7. It can be
seen that
P.M. = 40o
when K = 26.93.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

523

Exercises

phase margin versus K (PM=40º, K=26.93)
90
80

Phase Margin deg

70
60
50
40
30
20
10
0

0

5

10

15
K

20

25

30

FIGURE E10.7
Plot of phase margin versus K.

E10.8

The open-loop transfer function is
G(s) =

2257
806071.4
=
,
s(0.0028s + 1)
s(s + 357.14)

and the compensator is
Gc (s) =

K1 (s + z)
,
s

where z = K2 /K1 . The characteristic equation is
s3 + 357.14s2 + K1 s + K2 = 0 .
Using Routh-Hurwitz methods, the system is stable for
0 < K2 < 357.14 K1
or K2 /K1 < 357.14. Select the zero z at s = −10, then using root locus
methods we determine that K1 = 0.08 and K2 = 0.8. The roots of the
characteristic equation are
s1 = −10.6

and s2,3 = −175 ± j175 ,

and ζ = 0.707, as desired. The step response is shown in Figure E10.8.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

524

CHAPTER 10

The Design of Feedback Control Systems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Time (secs)

FIGURE E10.8
Step response with K1 = 0.08 and K2 = 0.8.

E10.9

The loop transfer function is
L(s) = Gc (s)G(s) =

K1 (s + K2 /K1 )
,
s(s + 1)

and
Kv = lim sGc (s)G(s) = K2 .
s→0

Select K2 = 5. The characteristic equation is
s2 + (K1 + 1) + K2 = 0 ,
and we want
s2 + 2ζωn s + ωn2 = 0 .
√
√
Equating coefficients yields ωn = K2 = 5. Also, since we want P.O. =
5%, we require ζ = 0.69. Thus,
2ζωn = K1 + 1

implies K1 = 2.08 .
√
The step response with K1 = 2.08 and K2 = 5 yields a P.O. > 5%. This

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

525

Exercises

is due to the zero at
s = −1.08 .
So, we raise the gain K1 = 3 and then the P.O. = 5%. The step response
is shown in Figure E10.9.

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

2.5

3

Time (secs)

FIGURE E10.9
Step response with K1 = 3 and K2 = 5.

E10.10

The loop transfer function is
L(s) = Gc (s)G(s) =

(KP s + KI )
.
s(s + 1)(s + 2)

Let KI = 2. Then, the plot of the phase margin as a function of KP is
shown in Figure E10.10, where it can be seen that
P.M. = 71.6o
is the maximum achievable phase margin. When KP = 1.54 and KI = 2
we have P.M. = 60o , as desired, and P.O. = 9% and Tp = 3.4 sec.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

526

CHAPTER 10

The Design of Feedback Control Systems

75

Phase Margin (deg)

70

65

60

55

50

1

2

3

4

5

6

7

8

9

10

KP

FIGURE E10.10
Phase margin versus KP with KI = 2.

The Nichols diagram and the closed-loop Bode plot are shown in Figures E10.11a and E10.11b, respectively.

40

0
0.25

30
0.5
1

20

-1

2.3
10

Gain dB

E10.11

-3
-6

0

-12

-10

-20

-20
-30
-40

-350

-300

-250

-200

-150

-100

Phase (deg)

FIGURE E10.11
(a) Nichols diagram for Gc (s)G(s) =

1350(1+0.25s)
.
s(s+2)(s+30)(1+0.025s)

-50

-40
0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

527

Exercises

Bode Diagram
20

Magnitude (dB)

0
−20
−40
−60
−80
−100
0

Phase (deg)

−45
−90
−135
−180
−225
−270
0
10

1

2

10

3

10

10

FIGURE E10.11
CONTINUED: (b) Closed-loop Bode plot.

E10.12

The loop transfer function is
L(s) = Gc (s)G(s) =



KK1 s +

1
2

s2 (s + 5)



.

When KK1 = 5.12, the roots are
s1,2 = −0.58 ± j0.58
s3 = −3.84 .
The complex poles have ζ = 0.707 and the predicted settling time is
Ts = 4/0.58 = 6.89 sec .
The actual settling time is Ts = 6.22 s.
E10.13

For the cascade compensator, we have
T1 (s) =

Gc (s)G(s)
8.1(s + 1)
=
,
1 + Gc (s)G(s)
(s + r1 )(s + r̂1 )(s + r2 )

where r1 = −1 + j2 and r2 = −1.67. For the feedback compensator, we

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

528

CHAPTER 10

The Design of Feedback Control Systems

have
T2 (s) =

8.1(s + 3.6)
G(s)
=
,
1 + Gc (s)G(s)
(s + r1 )(s + r̂1 )(s + r2 )

where
G(s) =

8.1
s2

and
Gc (s) =

s+1
.
s + 3.6

The response of the two systems differ due to different value of the zero of
T1 and T2 , however, both systems have the same characteristic equation.
E10.14

The Bode plot (with the lag network) is shown in Figure E10.14; the
phase margin is P.M. = 46o .
Bode Diagram
Gm = 21.9 dB (at 1.84 rad/sec) , Pm = 46.4 deg (at 0.344 rad/sec)
100

Magnitude (dB)

50
0
−50
−100
−150
−90

Phase (deg)

−135

−180

−225

−270
−4
10

−3

10

FIGURE E10.14
Bode plot for Gc (s)G(s) =

E10.15

−2

10

−1

0

10

5(7.5s+1)
s(s+1)(0.25s+1)(110s+1)

10

1

10

2

10

= 0.

At the desired crossover frequency ωc = 10 rad/sec, we have
20 log |Gc (j10)G(j10)| = −8.1 dB and
6

Gc (j10)G(j10) = −169o .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

529

Exercises

Therefore, the phase margin is P.M. = 11o . So,
φ = 30o − 11o = 19o

and

M = 8.1 dB .

Since φ > 0 and M > 0, a lead compensator is required.
E10.16

At the desired crossover frequency ωc = 2 rad/sec, we have
20 log |Gc (j2)G(j2)| = 17 dB and
6

Gc (j2)G(j2) = −134o .

Therefore, the phase margin is P.M. = 46o . So,
φ = 30o − 46o = −16o
M = −17 dB .
Since φ < 0 and M < 0, a lag compensator is required.
E10.17

Using a prefilter
Gp (s) =

KI
KP s + KI

the closed-loop transfer function is
T (s) =

s2

KI
.
+ (KP + 1)s + KI

The required coefficients for a deadbeat system are α = 1.82 and Ts =
4.82. Therefore,
KI = ωn2
KP = αωn − 1 .
Since we desired a settling time less than 2 seconds, we determine that
ωn = Ts /2 = 4.82/2 = 2.41 .
Then, the gains are
KP = 3.39
KI = 5.81 .
The step response (with the prefilter) is shown in Figure E10.17. The
percent overshoot is P.O. = 0.098% and the settling time is Ts = 1.99
seconds.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

530

CHAPTER 10

The Design of Feedback Control Systems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (secs)

FIGURE E10.17
Step response for the deadbeat system.

E10.18

Consider the PI controller
Gc (s) = Kp +

Kp s + KI
30s + 300
KI
=
=
s
s
s

and the prefilter
Gp (s) = 10 .
Then, the closed-loop system is
T (s) =

s2

300s + 3000
.
+ 280s + 3000

The percent overshoot is P.O. = 9.2% and the settling time Ts = 0.16
seconds. The steady-state tracking error to a unit step is zero, as desired.
E10.19

Consider the PID controller
Gc (s) = 29

s2 + 10s + 100
.
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

531

Exercises

The closed-loop transfer function is
29(s2 + 10s + 100)
.
s3 + 24s2 + 290s + 2900

T (s) =

The settling time to a unit step is Ts = 0.94 seconds.
E10.20

Consider the PD controller
Gc (s) = KD s + Kp = 3s + 1 .
The loop transfer function is
L(s) = Gc (s)G(s) =

3s + 1
.
s(s − 2)

The Bode plot is shown in Figure E10.20. The phase margin is P.M. =
40.4◦ . This is a situation where decreasing the gain leads to instability. The
Bode plot shows a negative gain margin indicating that the system gain
can be decreased up to -3.5 dB before the closed-loop becomes unstable.
Bode Diagram
Gm = −3.52 dB (at 0.816 rad/sec) , Pm = 40.4 deg (at 2.28 rad/sec)

Magnitude (dB)

40
20
0
−20

Phase (deg)

−40
−90
−135
−180
−225
−270
−2
10

−1

10

0

10

FIGURE E10.20
Bode plot for the loop transfer function L(s) =

E10.21

3s+1
.
s(s−2)

The transfer function from Td (s) to Y (s) is
T (s) =

s2

1
.
+ 4.4s + K

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

The tracking error is E(s) = R(s) − Y (s). When R(s) = 0, then E(s) =
−Y (s). The final value of the output to a unit step disturbance is ess =
1/K. If we want the tracking error to be less than 0.1, then we require
K > 10. When K = 10, we have the disturbance response shown in
Figure E10.21.

Step Response
0.12

0.1

0.08
Amplitude

532

0.06

0.04

0.02

0

0

0.5

FIGURE E10.21
Disturbance response for K = 10.

1

1.5
Time (sec)

2

2.5

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

533

Problems

Problems
P10.1

(a) The loop transfer function is
L(s) = Gc (s)G(s)H(s) =

(1 + ατ s)K1 K2
.
α(1 + τ s)(Js2 )

We desire ζ = 0.6, Ts ≤ 2.5 or ζωn ≥ 1.6. The uncompensated closedloop system is
T (s) =

K
,
s2 + K

where K = K1 K2 /J and K = ωn2 . We can select K = 20, and then
ζωn > 1.6. First, plot the Bode diagram for
G(s)H(s) =

20
s2

where K1 K2 /αJ = 20. The phase margin of the uncompensated system is 0o . We need to add phase at ωc . After several iterations, we
choose to add 40o phase at ωc , so
sin 40o =

α−1
= 0.64 .
α+1

Therefore, α = 4.6. Then,
10 log α = 10 log 4.6 = 6.63dB .
We determine the frequency where magnitude is -6.63 dB to be ωm =
√
p = ωn α = 14.1 and z = p/α = 3.07 .
The compensated loop transfer function (see Figure P10.1a) is
20
Gc (s)G(s)H(s) = 2
s

s
3.07
s
14.1



+1
 .
+1

(b) Since we desire ζωn ≥ 1.6, we place the compensator zero at z = 1.6.
Then, we place the compensator pole far in the left half-plane; in this
case, we selected p = 20. Thus, the compensator is
Gc (s) =

s + 1.6
.
s + 20

The root locus is shown in Figure P10.1b. To satisfy the ζ = 0.6
requirement, we find K = 250, and the compensated loop transfer

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

Gain dB

100

50

0
-50
10-1

100

101

102

101

102

Phase deg

-140
-150
-160
-170
-180
10-1

100

FIGURE P10.1
(a) Compensated Bode plot for Gc (s)G(s)H(s) =

20(s/3.07+1)
.
s2 (s/14.1+1)

function is
20
250(s + 1.6)
Gc (s)G(s)H(s) = 2
= 2
s (s + 20)
s

s
1.6 + 1
s
20 + 1



.

20
15
*

10
5

Imag Axis

534

0

x

*o

x

-5
-10
*

-15
-20
-25

-20

-15

-10

-5

0

Real Axis

FIGURE P10.1
CONTINUED: (b) Root locus for Gc (s)G(s)H(s) = 1 + K s2s+1.6
.
(s+20)

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

535

Problems

P10.2

The transfer function of the system is
G(s) =

s3

1.0e + 14
,
+ 2000s2 + 1e + 11s

where we use the system parameters given in P7.11 with the following
modifications: τ1 = τ1 = 0 and K1 = 1. Also we have scaled the transfer
function so that the time units are seconds. The parameters in P7.11 are
given for time in milliseconds. A suitable compensator is
Gc (s) =

s + 500
.
s+1

The closed-loop system response is shown in Figure P10.2. The percent
overshoot is P.O. ≈ 20% and the time to settle is Ts < 0.01 second.
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

0.005

0.01

0.015

0.02

0.025
0.03
Time (secs)

0.035

0.04

0.045

0.05

FIGURE P10.2
Step response.

P10.3

The loop transfer function is
Gc (s)G(s) =

16(s + 1)
K(s + z)
.
s(s2 + 2s + 16) (s + p)

We desire dominant roots with Ts < 5 sec and P.O. < 5%, so use ζ = 0.69
and ζωn = 0.8. One solution is to select z = 1.1 (i.e. to the left of the

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

536

CHAPTER 10

The Design of Feedback Control Systems

existing zero at s = −1) and determine the pole p and gain K for dominant
roots with ζ = 0.69. After iteration, we can select p = 100, so that the
root locus has the form shown in Figure P10.3. Then, we select K = 320,
200

150

100

Imag Axis

50

0

-50

-100

-150

-200
-200

-150

-100

-50

0
Real Axis

50

100

150

200

FIGURE P10.3
16(s+1)(s+1.1)
Root locus for 1 + K s(s2 +2s+16)(s+100) = 0.

so that ζ = 0.69. The final compensator is
Gc (s) =

320(s + 1.1)
.
s + 100

The design specifications are satisfied with this compensator.
P10.4

The uncompensated loop transfer function is
G(s) =

1
1
s2 ( 40
s

+ 1)

=

40
.
+ 40)

s2 (s

We desire 10% < P.O. < 20%, so 0.58 < ζ < 0.65, and Ts < 2 implies
ζωn < 2. We will utilize a PD compensator Ka (s + a). We select a = 2,
to obtain the root locus shown in Figure P10.4. Then with Ka = 23.5, we
have the desired root location, and
Gc (s) = 23.5(s + 2) .
The design specifications are satisfied with the PD compensator.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

537

Problems

30
+
*

20

Imag Axis

10

0

+
*o

x

x

-10

-20
+
*

-30
-50

-40

-30

-20

-10

0

Real Axis

FIGURE P10.4
40(s+2)
Root locus for 1 + Ka s2 (s+40) = 0.

P10.5

We desire P.O. < 10% and Ts < 1.5 sec. The compensator is a PI-type,
given by
Gc (s) = K2 +

K2 s + K3
K2 (s + a)
K3
=
=
s
s
s

where a = K3 /K2 . So, ess = 0 for a step input and
G(s) =

3.75Ka
25Ka
=
.
(s + 0.15)(0.15s + 1)
(s + 0.15)(s + 6.67)

The loop transfer function is
Gc (s)G(s) =

25Ka K2 (s + a)
.
s(s + 0.15)(s + 6.67)

Using root locus methods, we select a = 0.2 (after several iterations) and
determine Ka K2 to yield ζ = 0.65. This results in Ka K2 = 1. The root
locus is shown in Figure P10.5. The design specifications are met. The
actual percent overshoot and settling time are P.O. = 7.4% and Ts = 1.3
s. The controller is
Gc (s) = 1 +

0.2
.
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

538

CHAPTER 10

The Design of Feedback Control Systems

20
15
10
+
*

Imag Axis

5
0

x

x+
x
*o

-5

+
*

-10
-15
-20
-20

-15

-10

-5

0

5

10

15

20

Real Axis

FIGURE P10.5
25(s+0.2)
Root locus for 1 + Ka K2 s(s+0.15)(s+6.67) = 0.

As in P10.5, using root locus we find that placing z = 15 and p = 30 yields
a root locus shape (see Figure P10.6) where the loop transfer function is

60

40

20

Imag Axis

P10.6

+
*

0

x

+
*

o

x

x
+
*

-20

-40

-60
-60

-40

-20

0
Real Axis

FIGURE P10.6
25(s+15)
Root locus for 1 + Ka (s+0.15)(s+6.67)(s+30) = 0.

20

40

60

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

539

Problems

Gc (s)G(s) =

25Ka (s + z)
.
(s + p)(s + 0.15)(s + 6.67)

and where z, p and Ka are the parameters to be determined. Properly
choosing the parameter values allows us to increase ζωn of the dominant
roots (compared to the PI compensator of P10.5). Then, with Ka = 3.7,
the dominant roots have ζ = 0.65. The design specifications are met with
the compensator.
P10.7

The plant transfer function is
G(s) =

e−50s
.
(40s + 1)2

ess =

A
< 0.1A .
1 + Kp

Therefore, Kp > 9. Insert an amplifier with the compensator with a dc
gain = 9, as follows
Gc (s)G(s) =

9e−50s (s + 2)
.
(40s + 1)2 (s + p)

The system is unstable without compensation, and it is very difficult to
compensate such a time delay system with a lead compensator. Consider
a lag network
Gc (s) =

s+z
s+p

where z > p. Let
z = 10p.
Then, a plot of the P.M. versus p is shown in Figure P10.8a. Suitable
system performance can be obtained with P.M. > 45o , so choose
p = 0.0001.
The Bode plot of the compensated and uncompensated systems is shown
in Figure P10.7c, where we have selected z = 0.001 and p = 0.0001. The
compensated system has
P.M. = 62o

and

Ts = 9 minutes .

The step response is shown in Figure P10.7b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

540

CHAPTER 10

The Design of Feedback Control Systems
(a)

Phase Margin (deg)

150
100
50
0
-50
0

0.5

1

1.5
p

2.5

3
x10 -3

(b)

1.5

Amplitude

2

1
0.5
0
0

100

200

300

400

500

600

700

800

Time (secs)

FIGURE P10.7
(a) Phase margin versus p. (b) Step response with p = 0.0001 and z = 0.001.

20

Gain dB

10
0
-10
-20
-30
10-4

10-3

10-2

10-1

10-2

10-1

0

Phase deg

-100
-200
-300
-400
-500
10-4

10-3

FIGURE P10.7
CONTINUED: (c) Bode plot for the compensated system (solid lines) and the uncompensated system (dashed line).

P10.8

The transfer function is
G(s) =

5000
.
s(s + 10)2

To meet the steady-state accuracy, we need Kv > 40. The uncompensated

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

541

Problems

Kv = 50, so the steady-state accuracy can be met.
(a) Using the Bode method, we need P.M. = 70% (to meet P.O. < 5%
specification). Let
Gc (s) =

bs + 1
.
as + 1

The plot of P.M. versus b is shown in Figure P10.8a, where we set
a = 50b. Choosing b = 20 should satisfy the P.O. specification. The
Bode plot is shown in Figure P10.8c. Thus,
(a)

Phase Margin (deg)

80
70
60
50
40
30
0

5

10

15
b

25

30

(b)

1.5

Amplitude

20

1
0.5
0
0

5

10

15

20

25

30

35

40

45

50

Time (secs)

FIGURE P10.8
(a) Phase margin versus b; (b) Step response for lag compensator designed with Bode where
a = 1000 and b = 20.

Gc (s)G(s) =

5000(20s + 1)
.
s(s + 10)2 (1000s + 1)

The step response is shown in Figure P10.8b.
(b) We require that ζ = 0.7 to meet the P.O. specifications. Let
Gc (s) =

K(bs + 1)
.
(as + 1)

Using root locus methods, we fix a and b, and then determine K for
ζ = 0.7. Let a = 50b and select b = 10 (other values will work). The
root locus is shown in Figure P10.8d. We find K = 2.5 when ζ = 0.7.
Now, Kv = 125, so the steady-state accuracy requirement is satisfied
for the step response as shown in Figure P10.8e.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

150

Gain dB

100
50
0
-50
-100
10-4

10-3

10-2

10-3

10-2

10-1
100

101

102

101

102

Phase deg

0
-100
-200
-300
10-4

10-1

100

FIGURE P10.8
CONTINUED: (c) Bode plot for the compensated system with Gc (s) =

20s+1
1000s+1 .

20
15
10
5

Imag Axis

542

*

0

*

x

*ox
*

-5
-10
-15
-20
-20

-15

-10

-5

0

5

Real Axis

FIGURE P10.8
5000(10s+1)
CONTINUED: (d) Root locus for 1 + K s(s+10)2 (500s+1) .

10

15

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

543

Problems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

18

20

Time (secs)

FIGURE P10.8
CONTINUED: (e) Step response for lag compensator designed with root locus methods,
where K = 2.5.

We desire a small response for a disturbance at 6 rad/sec. The Bode plot
of Gc (s)G(s) is shown in Figure P10.9a where we consider a compensator

Gain dB

0
-50
-100
-150
10-1

100

101

102

103

100

101

102

103

0

Phase deg

P10.9

-100
-200
-300
10-1

FIGURE P10.9
(a) Bode plot for the compensated system with Gc (s) =

10(s2 +4s+10)
.
s2 +36

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

544

CHAPTER 10

The Design of Feedback Control Systems

of the form
Gc (s) =

K(s2 + as + b)
.
s2 + 36

Notice that the magnitude is large at ω = 6, as desired. We select
a = 4,

b = 10

and K = 10 .

The response to a sinusoidal disturbance at 6 rad/sec is shown in Figure P10.9b. Notice that the effect of the disturbance is virtually eliminated

0.02
0.015
0.01

Amplitude

0.005
0
-0.005
-0.01
-0.015
-0.02

0

10

20

30

40

50

60

70

80

90

100

Time (secs)

FIGURE P10.9
CONTINUED: (b) Disturbance response for a sinusoidal disturbance at 6 rad/sec.

P10.10

The step response with Gc (s) = 1 is shown in Figure P10.10. A suitable
lag compensator is
Gc (s) =

s + 0.05
.
s + 0.005

The step response of the compensated system is also shown in Figure P10.10.
The settling time of the compensated system is
Ts = 28 seconds .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

545

Problems

Compensated system (solid) & Uncompensated system (dashed)
30

25

Amplitude

20

15

Input (dotted line)

10

5

0
0

5

10

15

20

25

30

Time (sec)

FIGURE P10.10
Step response of uncompensated and compensated systems.

The root locus is shown in Figure P10.11 where a suitable lead-lag com-

300

200

100

Imag Axis

P10.11

+

0

x

+

oo
+x
+

-100

-200

-300
-300

-200

-100

0
Real Axis

FIGURE P10.11
160(s+17)(s+10)
Root locus for 1 + K s2 (s+170)(s+1) = 0.

100

200

300

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

546

CHAPTER 10

The Design of Feedback Control Systems

pensator is
Gc (s) = K

s + 10 s + 17
.
s + 1 s + 170

The selected gain is K = 57, so that the damping of the complex roots
is about ζ = 0.7. For this particular design, the closed-loop system zeros
will affect the system response and the percent overshoot specification
may not be satisfied. Some design iteration may be necessary or aprefilter
can be utilized. A suitable prefilter is
Gp (s) =

17
.
s + 17

The acceleration constant is Ka = 9120.
We choose K = 10. This yields a velocity constant Kv = 20K = 200, as
desired. A suitable two-stage lead compensaator is
Gc (s) =

(0.05s + 1)(0.05s + 1)
.
(0.0008s + 1)(0.0008s + 1)

The Bode plot is shown in Figure P10.12. The phase margin is P.M. =
75.06o .
Phase margin=75.06 deg

100

Gain dB

50
0
-50
-100
10-1

100

101

102

103

104

103

104

0

Phase deg

P10.12

-100
-200
-300
10-1

100

101

102

FIGURE P10.12
200(0.05s+1)2
Bode plot for s(0.1s+1)(0.05s+1)(0.0008s+1)2 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

547

Problems

P10.13

(a) When
Gc (s) = K = 0.288 ,
the phase margin is P.M. = 49.3o and the bandwidth is ωB = 0.95
(b) A suitable lag compensator is
Gc (s) =

25s + 1
.
113.6s + 1

The compensated system phase margin is P.M. = 52.21o and Kv = 2,
as desired.
P10.14

Gc (s) =

1.155s + 1
.
0.032s + 1

The compensated system phase margin is P.M. = 50o and Kv = 2, as
desired. The settling time is Ts = 3.82 seconds.
One possible solution is
Gc (s) = K

(s + 12)(s + 15)
,
(s + 120)(s + 150)

where K = 900. The disturbance response is shown in Figure P10.15.
Step Response
0.1

0.08

Amplitude

P10.15

0.06

0.04

0.02

0

0

0.1

0.2

0.3
Time (sec )

FIGURE P10.15
Compensated system disturbance response.

0.4

0.5

0.6

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

548

CHAPTER 10

The PI controller is given by
K(s + b)
,
s

Gc (s) =

where K and b are to be determined. To meet the design specifications,
we need
ζ = 0.6

and ωn = 6.67 rad/sec .

The closed-loop transfer function is
T (s) =

K(s + b)
.
s2 + Ks + bK

Solving for the gains yields K = 2ζωn = 8 and b = ωn2 /K = 5.55. A
suitable prefilter is
Gp (s) =

5.55
.
s + 5.55

The step response, with and without the prefilter, is shown in Figure P10.16.

Without prefilter (solid) & with prefilter (dashed)
1.4

1.2

1

Amplitude

P10.16

The Design of Feedback Control Systems

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

Time (sec)

FIGURE P10.16
Compensated system response with and without a prefilter.

2.5

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

549

Problems

P10.17

The plant transfer function is
G(s) =

K
.
s(s + 10)(s + 50)

We desire ζωn > 10 to meet Ts < 0.4 sec and ζ = 0.65 to meet P.O. <
7.5%. Try a pole at s = −120. The root locus is shown in Figure P10.17.
The gain K = 6000 for ζ = 0.65. Thus,
Gc (s)G(s) =

6000(s/15 + 1)
s(s + 10)(s + 50)(s/120 + 1)

and Kv =

6000
= 12 .
500

200
150
100

Imag Axis

50
0

*x

x

*

*
ox x
*

-50
-100
-150
-200
-200

-150

-100

-50

0

50

100

150

200

Real Axis

FIGURE P10.17
s/15+1
Root locus for 1 + K s(s+10)(s+50)(s/120+1) .

P10.18

(a) The loop transfer function is
L(s) =

K1 e−2T s
0.25s + 1

where T = 1.28. The phase angle is
φ = −2.56ω − tan 0.25ω .
So, ω = 1.12 rad/sec when φ = −180o . However, the break frequency

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

is 4 rad/sec. Therefore, you cannot achieve P.M. = 30o and have the
system be stable for K1 < 1. The steady-state error is
ess =

A
A
=
1 + Kp
1 + K1

since K1 = Kp .
(b) Set K1 = 20, then Kp = 20 and this yields a 5% steady-state error.
Without compensation, the system is now unstable. Let
Gc (s) =

s/b + 1
s/a + 1

where b = 5 and a = 0.01. Then, the system is stable with
P.M. = 63o .
The system response is shown in Figure P10.18.

1.2

1

0.8

Amplitude

550

0.6

0.4

0.2

0

-0.2

0

2

4

6

8

10
Time (secs)

FIGURE P10.18
Unit step response with Gc (s) =

20(s/5+1)
s/0.01+1 .

12

14

16

18

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

551

Problems

P10.19

(a) The open-loop transfer function is
G(s) =

Ke−sT
,
(s + 1)(s + 3)

where T = 0.5 sec. We desire P.O. < 30%, thus ζ > 0.36. We will
design for ζ = 0.4, which implies P.M. = 40o . Then
φ = − tan−1 ω − tan−1

ω
− 0.5ω(57.3o ) .
3

At ωc = 1.75, the phase margin is P.M. = 40o , and solving
|G(jω)| =

K
[(3 −

ω 2 )2

1

+ (4ω)2 ] 2

=1

at ω = 1.75 yields K = 7. Then ess = 0.3.
(b) We want ess < 0.12, so use ess = 0.10 as the goal. Then
Gc (s)G(s) =

Ke−0.5s (s + 2)
,
(s + 1)(s + 3)(s + b)

and
ess =
where Kp =

2K
3b .

1
1 + Kp

If b = 0.1 then Kp = 6.7K and
ess =

1
.
1 + 6.7K

So, we need 6.7K = 9, or K = 1.35. We need a lag compensator (i.e.
b < 2) to meet ess < 12% and have stability.
P10.20

We desire Kv = 20, P.M. = 45o and ωB > 4 rad/sec. Thus, we set K = 20,
and
G(s) =

s

s
2

20

+1

s
6

 .

+1

Then, the Bode plot yields P.M. = −21o uncompensated at ωc = 5.2
account for the shift of the crossover to a higher frequency with the phase
Gc (s) =



1 + ατ s
1 + τs

2

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

552

CHAPTER 10

The Design of Feedback Control Systems

One solution is to use
α = 10
τ = 1/67 .
Then
Gc (s) =

100(s + 6.7)2
.
(s + 67)2

The compensator has two zeros at ω = 6.7, two poles at ω = 67 yielding
P.M. = 47o , ωc = 7.3 and ωB = 12 rad/sec.
P10.21

We desire Kv = 20, P.M. = 45o and ωB ≥ 2. The lag compensator is
Gc (jω) =

1 + jωτ
1 + jωατ

where α > 1. From the Bode plot, φ = −135o at ω ∼
= 1.3. So, at ω = 1.3,
we need to lower the magnitude by 22 dB to cause ω = 1.3 to be ωc′ , the
new crossover frequency. Thus, solving
22 = 20 log α
yields α = 14. We select the zero one decade below ωc′ or
Therefore,

1
τ

= 0.13.

1
0.13
=
= 0.0093 .
ατ
14
Then, the lag compensator is given by
Gc (s) =

s
1 + 0.13
s + 0.13
.
=
s
1 + 0.0093
14(s + 0.0093)

The new crossover is ωc′ = 1.3, and ωB = 2.14 rad/sec.
P10.22

We desire P.M. = 45o , Kv = 20 and 2 ≤ ωB ≤ 10. The lead-lag compensator is
Gc (s) =

s
1 + sb 1 + 10a
·
.
s
1 + 10b
1 + as

Since ωB ∼
= 1.5ωc , we design for a new crossover frequency ωc′ so that
1.4 < ωc′ < 7 .
Try for ωc′ = 4. The phase φ = −190o at ω = 4, so we need to add phase
lead of 55o plus phase to account for lag part of network at ωc′ . Use α = 10
and bracket ω = 4 with the lead network. Put the zero at ω = 0.8 = b

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

553

Problems

and the pole at ω = 8. For the lag compensator, put the zero at a lower
frequency than ωc′ /10. So try a zero at ω = 0.2 = 10a and a pole at
ω = 0.02 = a. Then, the lead-lag compensator is
s
s
1 + 0.8
1 + 0.2

 .
Gc (s) =
s
1 + 8s 1 + 0.02





The compensated Bode plot yields

P.M. = 50o

1
1
=
= 0.05 .
1 + Kp
1 + K/25

ess =

So, we need K/25 ≥ 19 or K ≥ 475. One possible solution is
Gc (s) =

4s + 1
12s + 1

and

K = 475 .

The compensated Bode plot is shown in Figure P10.23. The phase margin
is P.M. = 46o .
Bode Diagram
Gm = Inf dB (at Inf rad/sec) , Pm = 46 deg (at 11.5 rad/sec)

Magnitude (dB)

40
20
0
−20
−40
0
Phase (deg)

P10.23

and ωB = 6.2 rad/sec .

−45
−90
−135
−180
−3
10

FIGURE P10.23
Bode plot for Gc (s)G(s) =

−2

10

−1

0

10
10

475(4s+1)
.
(s+5)2 (12s+1)

1

10

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

554

CHAPTER 10

P10.24

The Design of Feedback Control Systems

The arm-rotating dynamics are represented by
G(s) =

80
s



s2
4900

+

s
70

 .

+1

We desire Kv = 20, and P.O. < 10%. One possible solution is the lead-lag
compensator
Gc (s) =

(s + 50)(s + 0.48)
.
4(s + 400)(s + 0.06)

With this compensator, we have
P.O. = 9.5%
P10.25

and

Kv = 20 .

Neglect the pole of the airgap feedback loop at s = 200. The characteristic
equation is
1 + K̄

(s + 20)(s + c)
=0,
s3

where
K
K1 + K2
K2 b
c=
.
K1 + K2

K̄ =

Choose
c = 10
to attain the root locus structure shown in Figure P10.25. The gain
K̄ = 38.87
insures the damping ratio of ζ = 0.5. Then, solving for K1 and b yields
K1 =

K
− K2
38.87

and
b=

0.1K
.
38.87K2

For given values of K and K2 (unspecified in the problem), we can compute K1 and b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

555

Problems

40
30
+

20

Imag Axis

10
0

o

o+

x

-10

0

-10
-20
+

-30
-40
-40

-30

-20

10

20

30

40

Real Axis

FIGURE P10.25
(s+20)(s+10)
= 0.
Root locus for 1 + K̄
s3

P10.26

The loop transfer function is
Gc (s)G(s) =

0.15K(10as + 1)
,
s(s + 1)(5s + 1)(as + 1)

where K and a are to be selected to meet the design specifications. Suitable values are
K = 6.25

and a = 0.15 .

Then, the phase margin is P.M. = 30.79o and the bandwidth is ωB =
Gc (s) = 6.25
P10.27

1.5s + 1
.
0.15s + 1

(a) Let Gc (s) = K = 11. Then the phase margin is P.M. = 50o and the
performance summary is shown in Table P10.27.
(b) Let
Gc (s) =

K(s + 12)
,
(s + 20)

where K = 32. Then, the phase margin is P.M. = 50o and the performance summary is given in Table P10.27.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

556

CHAPTER 10

compensator

The Design of Feedback Control Systems

P.M.

P.O.

Tp

Ts

Mpω

ωB

Gc (s) = K = 11

50o

18%

0.34 sec

0.78 sec

1.5 dB

32(s+12)
s+20

50o

18%

0.20 sec

0.47 sec

1.5 dB

TABLE P10.27

Performance Summary.

P10.28

The loop transfer function is
Gc (s)G(s) =

K(as + 1)
,
s(s + 10)(s + 14)(10as + 1)

where K and a are to be selected to meet the design specifications, and
we have set α = 10. The root locus is shown in Figure P10.28a. To satisfy

30

20

10

Imag Axis

Gc (s) =

*

0

*

x

x

*ox
*

-10

-20

-30
-30

-20

-10

0
Real Axis

FIGURE P10.28
1400(s+1)
(a) Root locus for 1 + K s(s+10)(s+14)(10s+1) = 0.

10

20

30

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

557

Problems

the steady-state tracking error we must select
K > 1400 .
Suitable values for the lag compensator are
K = 4060

and a = 1 .

Then, the percent overshoot is P.O. = 31% and the settling time is Ts =
2.34 sec. The lag compensator is
Gc (s) =

s+1
.
10s + 1

The step response is shown in Figure P10.28b.

1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

Time (secs)

FIGURE P10.28
CONTINUED: (b) Step response.

P10.29

The plant transfer function is
G(s) =

10e−0.05s
.
s2 (s + 10)

Gc (s) =

16(s + 0.7)
(s + 9)

2.5

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

558

CHAPTER 10

The Design of Feedback Control Systems

provides Mpω = 3.4 dB and ωr = 1.39 rad/sec. The step response is shown
in Figure P10.29. The overshoot is P.O. = 37% and Ts = 3.5 sec.

1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

1

2

3

4

5

6

Time (secs)

FIGURE P10.29
Unit step response with Gc (s) =

P10.30

16(s+0.7)
.
s+9

The vehicle is represented by
G(s) =

K
K
≈
.
s(0.04s + 1)(0.001s + 1)
s(0.04s + 1)

For a ramp input, we want
ess
1
= 0.01 =
.
A
Kv
So, let
G(s) =

100
.
s(0.04s + 1)

The uncompensated P.M. = 28o at ωc = 47 rad/sec. We need to add 17o .
Gc (s) =

1 + 0.021s
.
1 + 0.01s

The phase margin is P.M. = 45o .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

559

Problems

Gc (s) =

1 + 0.04s
.
1 + 0.005s

The phase margin is P.M. = 65o .
For Case 1, we have
P.O. = 25% ,

Ts = 0.13 sec and Tp = 0.05 sec .

For Case 2, we have
P.O. = 4% ,
P10.31

Ts = 0.04 sec and Tp = 0.03 sec .

As in P10.30, the plant is given by
G(s) =

100
.
s(0.04s + 1)

The uncompensated P.M. = 28o . We need P.M. = 50o . The phase lag
compensator
Gc (s) =

1 + 0.5s
1 + 2.5s

results in P.M. = 50o . The P.O. = 21%, Ts = 0.72 sec and Tp = 0.17 sec.
P10.32

(a) To obtain Kv = 100, we have
Gc (s)G(s) =

43.33(s + 500)
.
s(s + 0.0325)(s2 + 2.57s + 6667)

With K = 43.33, we have
P.M. = 1.2o ,

Mpω = 26 dB ,

The Bode plot is shown in Figure P10.32.
(b) Let
Gc (s) =

0.35s + 1
,
0.001s + 1

and K = 43.33 (as before). Then,
P.M. = 36o ,

Mpω = 5.4 dB ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

560

CHAPTER 10

The Design of Feedback Control Systems

100

Gain dB

50
0
-50
-100
-150
10-2

10-1

100

101

102

103

102

103

-100

Phase deg

-150
-200
-250
-300
-350
10-2

10-1

100

101

FIGURE P10.32
Bode plot with Gc (s) = K = 43.33.

The step response is shown in Figure P10.33, where
Gc (s) =

10(s + 0.71)(s + 0.02)
.
(s + 0.0017)(s + 10)

1.2

1

0.8

Amplitude

P10.33

0.6

0.4

0.2

0
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Time (secs)

FIGURE P10.33
Step response with the lead-lag compensator Gc (s) =

10(s+0.71)(s+0.02)
.
(s+0.0017)(s+10)

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

561

Problems

Then,
Kv = 80
and P.O. = 17%, Ts = 1.8 sec, and ζ = 0.54.
The process model is
G(s) =

s2 (s

1
,
+ 10)

and we consider the lead compensator
Gc (s) = K

1 + sατ
,
1 + sτ

where α = 100, τ = 0.4 and K = 0.5. Then, P.M. = 46.4o . The step
response is shown in Figure P10.34. The system performance is
P.O. = 22.7%
Ts = 5.2 sec
Tp = 1.72 sec .

1.4

1.2

1

Amplitude

P10.34

0.8

0.6

0.4

0.2

0

0

1

2

3

4

5
Time (secs)

6

7

FIGURE P10.34
40s+1
Step response with the lead compensator Gc (s) = 0.5 0.4s+1
.

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

562

CHAPTER 10

P10.35

The Design of Feedback Control Systems

The phase margin is shown in Figure P10.35. As the time delay increases,
the phase margin decreases. The system is unstable when T > 2.1843 s.
140

120

100

Phase margin (deg)

80

60

40

Stability boundary

20

0

−20

0

0.5

1

1.5

2

Time delay (s)

FIGURE P10.35
Step response with Gc (s)G(s) =

2.5

where 0 ≤ T ≤ 2.5.

One possible solution is the integral controller Gc (s) = 2/s. The step
response is shown in Figure P10.36. The steady-state tracking error to a

1.6
1.4
1.2
1

Amplitude

P10.36

2s+0.54 −T s
e
,
s(s+1.76)

T=2.1843 s

0.8
0.6
0.4
0.2
0
-0.2

0

1

2

3

4

5

6

Time (secs)

FIGURE P10.36
Step response with the integral controller Gc (s) = 2/s.

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

563

Problems

step input is zero since the system is type-1. The phase margin is P.M. =
32.8◦ and the bandwidth is ωB = 4.3 rad/s .
P10.37

One possible solution is
Gc (s) =

1600(s + 1)
.
25s + 1

The overshoot to a unit step is P.O. = 4.75% and the steady-state error
to a step input is ess = 1%. The system bandwidth is ωB = 9.7 rad/sec.
P10.38

Gc (s) =

2.88(s + 2.04)
.
s + 5.88

The Bode plot is shown in Figure P10.38. The phase margin is P.M. =
30.4o at ωc = 9.95 rad/sec and the bandwidth is ωB = 17.43 rad/sec.
60

Gain dB

40
20
0
-20
-40
10-1

100

101

102

101

102

Phase deg

0
-100
-200
-300
10-1

100

FIGURE P10.38
Bode plot for Gc (s)G(s) =

P10.39

115.29(s+2.04)
.
s(s+2)(s+5.88)

The lag compensator is
Gc (s) =

1 + 1.48s
.
1 + 11.08s

The Bode plot is shown in Figure P10.39. The steady-state error specification is satisfied since Kv = 20. The phase margin is P.M. = 28.85o at

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

564

CHAPTER 10

The Design of Feedback Control Systems

Gain dB

100

50

0
-50
10-3

10-2

10-1

100

101

100

101

Phase deg

0
-100
-200
-300
10-3

10-2

10-1

FIGURE P10.39
Bode plot for Gc (s)G(s) =

40(1+1.48s)
.
s(s+2)(1+11.08s)

ωc = 2 rad/sec and the bandwidth is ωB = 3.57 rad/sec.
P10.40

The lag compensator is
Gc (s) =

2.5(1 + 1.64s)
.
1 + 30.5s

The Bode plot is shown in Figure P10.40. The steady-state error specification is satisfied since
Kv = 50 .
The phase margin is
P.M. = 28.93o
at ωc = 1.98 rad/sec and the bandwidth is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

565

Problems

Gain dB

100

50

0

-50
10-3

10-2

10-1

100

101

100

101

Phase deg

0
-100
-200
-300
10-3

10-2

10-1

FIGURE P10.40
Bode plot for Gc (s)G(s) =

P10.41

100(1+1.64s)
.
s(s+2)(1+30.5s)

We use Table 10.2 in Dorf & Bishop to determine the required coefficients
α = 1.9

and β = 2.2 .

Also,
ωn Tr = 4.32

implies

ωn = 4.32,

since we require
Tr = 1 second.
The characteristic equation is
s3 + 8.21s2 + 41.06s + 80.62 = s3 + (1 + p)s2 + (K + p)s + Kz = 0 .
Equating coefficients and solving yields
p = 7.21
P10.42

K = 33.85

z = 2.38 .

From Example 10.4 in Dorf & Bishop, we have the closed-loop transfer
function
T (s) =

(s2

96.5(s + 4)
.
+ 8s + 80)(s + 4.83)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

566

CHAPTER 10

The Design of Feedback Control Systems

A suitable prefilter is
Gp (s) =

4
.
s+4

The step response (with and without the prefilter) is shown in Figure P10.42.
With prefilter (solid) & without prefilter (dashed)
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time (sec)

FIGURE P10.42
Step response with and without the prefilter.

P10.43

Let
K = 100.
The Bode plot is shown in Figure P10.43a and the response to a simusoidal
noise input with
is shown in Figure P10.43b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

567

Problems

Bode Diagram
60

40

Magnitude (dB)

20

0

−20
System: sysg
Magnitude (dB): −40.1

−40

−60

−80
−1
10

0

10

1

2

10

3

10

10

0.07
0.06
0.05

Amplitude

0.04
0.03
0.02
0.01
0
−0.01

−0.02

0

1

2

3

4
Time (sec)

5

6

7

8

FIGURE P10.43
(a) Bode magnitude plot. (b) Response to a noise input.

P10.44

For 0.129 < K ≤ 69.87, the system is unstable. The percent overshoot is
shown in Figure P10.44 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

Percent Overshoot

100

50

0

−50

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

85

90

95

100

K
160
Percent Overshoot

568

140

120

100
65

70

75

80
K

FIGURE P10.44
Percent overshoot.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

569

AP10.1

(a) With
Gc (s) = K ,
the closed-loop transfer function is
T (s) =

s3

+

K
.
+ 4s + K

5s2

When K = 2.05, the characteristic equation is
s3 + 5s2 + 4s + 2.05 = 0
with poles at s = −4.1563 and s = −0.4219 ± j0.5615. Therefore
ζ = 0.6, and the predicted overshoot is
√

P.O. = 100e−π0.6/

1−0.62

= 9.5% < 13% .

The actual overshoot is P.O. = 9.3% and Ts = 8.7 seconds.
(b) When
Gc (s) =

82.3(s + 1.114)
s + 11.46

the closed-loop transfer function is
82.3(s + 1.114)
+
+ 61.3s2 + 128.14s + 91.6822
82.3(s + 1.114)
=
.
(s + 1.196)(s + 12.26)(s + 1.5 ± j2)

T (s) =

s4

16.46s3

Therefore ζ = 0.6 and the predicted overshoot is P.O. = 9.5% < 13%.
The actual overshoot is P.O. = 12% and Ts = 2.5 seconds.
AP10.2

The lag network is given by
Gc =

K(s + a1 )
.
s + a2

The closed-loop transfer function is
T (s) = K

s4

+ (5 + a2

)s3

s + a1
.
+ (4 + 5a2 )s2 + (4a2 + K)s + Ka1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

570

CHAPTER 10

The Design of Feedback Control Systems

Computing the steady-state tracking error yields
s4 + (5 + a2 )s3 + (4 + 5a2 )s2 + 4a2 s
s→0 s5 + (5 + a2 )s4 + (4 + 5a2 )s3 + (4a2 + K)s2 + Ka1 s
4a2
=
< 0.125 .
a1 K

ess = lim

If we select K = 2.05 (as in AP10.1), then
a1 > 15.61a2 .
So, take a2 = a1 /16. The lag compensator can now be written as
Gc (s) = 2.05

s + a1
.
s + a1 /16

Select a1 = 0.018. Then, the closed-loop transfer function is
T (s) =

s4

+

5.0011s3

2.05(s + 0.018)
.
+ 4.0056s2 + 2.0545s + 0.0369

The performance results are P.O. = 13% and Ts = 29.6 seconds for a step
input, and ess = 0.12 for a ramp input.
AP10.3

The plant transfer function is
G(s) =

1
s(s + 1)(s + 4)

and the PI controller is given by
Gc (s) =

Kp s + KI
.
s

The closed-loop transfer function is
T (s) =

s4

+

5s3

Kp s + KI
.
+ 4s2 + Kp s + KI

For a unit ramp, the steady-state tracking error is
s4 + 5s3 + 4s2
=0.
s→0 s5 + 5s4 + 4s3 + Kp s2 + KI s

ess = lim

Any KI > 0 and Kp > 0 (such that the system is stable) is suitable and
will track a ramp with zero steady-state error. Since we want P.O. <
13%, the damping of the dominant roots should be ζ ≈ 0.6. One suitable

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

571

solution is to place the zero at s = −0.01 and select the PI controller
2.05(s + 0.01)
.
s

Gc (s) =

Therefore, Kp = 2.05 and KI = 0.0205. The closed-loop transfer function
is
T (s) =

s4

+

5s3

2.05(s + 0.01)
.
+ 4s2 + 2.05s + 0.0205

The performance results are P.O. = 11.5% and Ts = 9.8 seconds for a
step input, and ess = 0 for a unit ramp.
AP10.4

The closed-loop transfer function is
T (s) =

10K1
.
s2 + 10(1 + K1 K2 )s + 10K1

From the performance specifications, we determine that the natural frequency and damping of the dominant poles should be ωn = 5.79 and
ζ = 0.69. So,
s2 + 8(1 + K1 K2 )s + 8K1 = s2 + 2ζωn s + ωn2 = s2 + 7.99s + 33.52 .
Solving for the gains yields K1 = 4.19 and K2 = 0. The closed-loop
transfer function is
T (s) =

s2

33.52
.
+ 8s + 33.52

The performance results are P.O. = 5% and Ts = 1 second.
AP10.5

(a) From the overshoot specification P.O. = 10%. The plant transfer
function is
G(s) =

1
.
s(s + 1)(s + 10)

Let Gp = 1. A suitable compensator is
Gc = K

s + 0.5
.
s + 10

Using root locus methods, we determine that K = 45 yields P.O. ≈
10%. The closed-loop poles are s1,2 = −2.5 ± j5.1, s3 = −15.48, and
s4 = −0.45.
(b) The closed-loop transfer function is
T (s) =

s4

+

21s3

450(s + 0.5)
.
+ 120s2 + 550s + 225

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

572

CHAPTER 10

The Design of Feedback Control Systems

The step response is shown in Figure AP10.5. The overshoot and
settling time are P.O. = 9.5% and Ts = 3.4 seconds.
(c) A suitable prefilter is
Gp (s) =

0.5
.
s + 0.5

The closed-loop transfer function is
T (s) =

s4

+

21s3

225
.
+ 120s2 + 550s + 225

The step response is shown in Figure AP10.5. The overshoot and
settling time are P.O. = 0% and Ts = 6.85 seconds.

1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

1

2

3

4
Time (sec)

5

6

7

8

FIGURE AP10.5
Step response with prefilter (dashed line) and without prefilter (solid line).

AP10.6

From Example 10.12 in Dorf & Bishop, we have the relationship
ωn Ts = 4.04 .
Thereore, minimizing Ts implies maximizing ωn . Using Table 10.2 in Dorf
& Bishop, we equate the desired and actual characteristic polynomials
q(s) = s3 + 1.9ωn s2 + 2.2ωn2 s + ωn3 = s3 + (1 + p)s2 + (K + p)s + Kz .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

573

Comparing coefficients yields
(1 + p) = 1.9ωn ,

K + p = 2.2



1+p
1.9

2

Kz = ωn3 .

,

So, from the first relationship we see that maximizing ωn implies maximizing p. Solving for p while maintaining K < 52
K=

2.2 2
(p + 2p + 1) − p < 52
3.61

we determine that
−9.3643 < p < 9.005 .
The largest p = 9. Therefore, K = 51.94 and z = 2.81. The step response
is shown in Figure AP10.6. The settling time is Ts = 0.77 second.

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Time (secs)

FIGURE AP10.6
Step response with minimum settling time.

AP10.7

Let Gp = 1. The closed-loop transfer function is
T (s) =

K(s + 3)
.
s4 + 38s3 + 296s2 + (K + 448)s + 3K

When K = 311, the characteristic equation
s4 + 38s3 + 296s2 + 759s + 933 = 0

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

574

CHAPTER 10

The Design of Feedback Control Systems

√
has poles at s = −1.619 ± j1.617 (ζ = 1/ 2), s = −6.25, and s = −28.51.
(a) When Gp (s) = 1 and K = 311, the closed-loop transfer function is
T (s) =

311(s + 3)
.
s4 + 38s3 + 296s2 + 759s + 933

The step input performance is P.O. = 6.5%, Ts = 2.5 seconds, and
Tr = 1.6 seconds. With the prefilter
Gp (s) =

3
s+3

and K = 311, the closed-loop transfer function is
T (s) =

s4

+

38s3

933
.
+ 296s2 + 759s + 933

In this case, the step response is P.O. = 3.9%, Ts = 2.8 seconds, and
Tr = 1.3 seconds.
(b) Now, consider the prefilter
Gp (s) =

1.8
s + 1.8

and K = 311. The closed-loop transfer function is
(s) =

s5

+

39.8s4

559.8(s + 3)
.
+ 364.4s3 + 1291.8s2 + 2299.2s + 1679.4

The step input response is P.O. = 0.7%, Ts = 2.14 seconds and
Tr = 1.3 seconds.
AP10.8

The plant transfer function is
G(s) =

250
.
s(s + 2)(s + 40)(s + 45)

The performance specifications are P.O. < 20%, Tr < 0.5 second, Ts < 1.2
seconds and Kv ≥ 10. A suitable lead compensator is
Gc = 1483.7

s + 3.5
.
s + 33.75

The closed-loop transfer function is
T (s) =

250(1483.7)(s + 35)
s(s + 2)(s + 40)(s + 45)(s + 33.75) + 250(1483.7)(s + 3.5)

The actual step input performance (see Figure AP10.8) is P.O. = 18%,
Ts = 0.88 second, Tr = 0.18 second, and Kv = 10.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

575

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time (secs)

FIGURE AP10.8

The frequency response is shown in Figure AP10.9.

Bode Diagrams
Gm=12.4 dB (Wcg=20.9); Pm=42.0 deg. (Wcp=9.0)
150
100
50
0
Phase (deg); Magnitude (dB)

AP10.9

-50
-100
-150
-50
-100
-150
-200
-250
-300
-3
10

10

-2

10

-1

10

0

FIGURE AP10.9
Bode plot with Gc (s) =

(s+2.5)(s+0.9871)
(s+36.54)(s+0.0675)

10

1

10

2

10

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

576

CHAPTER 10

The Design of Feedback Control Systems

One lead-lag compensator that satisfies the specifications is
Gc (s) =

(s + 2.5)(s + 0.9871)
.
(s + 36.54)(s + 0.0675)

The gain and phase margins are Gm = 12.35 dB and P m = 41.8◦ , respectively. The velocity error constant is Kv = 100. Therefore, all specifications are satisfied.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

577

Design Problems

Design Problems
The plant model with parameters given in Table CDP2.1 in Dorf and
Bishop is given by:
26.035
θ(s)
=
,
Va (s)
s(s + 33.142)
where we neglect the motor inductance Lm and where we switch off the
tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a PD
controller the closed-loop system characteristic equation is
s2 + (33.142 + 26.035KD )s + 26.035Kp = 0 .
Using Table 10.2 in Dorf and Bishop we determine that for a second-order
system with a deadbeat response we have α = 1.82 and ωn Ts = 4.82. Since
we desire Ts < 0.25 seconds, we choose ωn = 19.28. Equating the actual
characteristic equation with the desired characteristic equation we obtain
s2 + ωn αs + ωn2 = s2 + (33.142 + 26.035KD )s + 26.035Kp .
Solving for Kp and KD yields the PD controller:
Gc (s) = 14.28 + 0.075s .
The step response is shown below. The settling time is Ts = 0.24 second.
1
0.9
0.8
0.7
0.6
Amplitude

CDP10.1

0.5
0.4
0.3
0.2
0.1
0

0

0.05

0.1

0.15
Time (secs)

0.2

0.25

0.3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

578

CHAPTER 10

The plant is given as
G(s) =

20
.
s (s + 2)

50(s + 1)
.
s + 20

Similarly, a suitable lag compensator is
Gclag (s) =

s + 0.1
.
s + 0.022

The loop transfer function with the lead-lag compensator is
Gc (s)G(s) =

1000(s + 1)(s + 0.1)
.
s (s + 2) (s + 0.022)(s + 20)

The step response and ramp response are shown in Figure DP10.1. The
velocity constant is Kv = 50, so the steady-state error specification is
satisfied.

Step response

1.5

1

0.5

0

0

0.1

0.2

0.3

0.4

0.5
0.6
Time (sec)

0.7

0.8

0.9

1

0

0.1

0.2

0.3

0.4

0.5
0.6
Time (sec)

0.7

0.8

0.9

1

1
Ramp response

DP10.1

The Design of Feedback Control Systems

0.8
0.6
0.4
0.2
0

FIGURE DP10.1
Step response and ramp response.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

579

Design Problems

(a) When Gc (s) = K, we require K > 20 to meet the steady-state tracking specification of less than 5%.
(b) The system is unstable for K > 20.
(c) A single stage lead compensator is
Gc1 (s) =

1 + 0.49s
.
1 + 0.0035s

With this compensator, the bandwidth is
and the phase margin is P.M. = 28.57o .
(d) A two stage lead compensator is
Gc2 (s) =

(1 + 0.0185s)(1 + 0.49s)
.
(1 + 0.00263s)(1 + 0.0035s)

With the two stage compensator, the bandwidth is
and the phase margin is P.M. = 56.79o . The step response for the
two compensators is shown in Figure DP10.2.
Single stage (solid) & two stage (dashed)
1.4

1.2

1

Amplitude

DP10.2

0.8

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

0.5

0.6

Time (sec)

FIGURE DP10.2
Step response for one- and two-stage lead compensators.

0.7

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

580

CHAPTER 10

The mast flight system is modeled as
6
.
s(s + 1.5)(s + 3.9)

G(s) =

Consider the proportional controller
Gc (s) = K = 0.85 .
The system step response is shown in Figure DP10.3. The percent overshoot is
P.O. = 15.9%,
the rise time is
Tr = 3.63 seconds,
and the phase margin is
P.M. = 52o .

Step Response
1.4

System: syscl
Peak amplitude: 1.16
Overshoot (%): 15.9
At time (sec): 3.63

1.2

1
Amplitude

DP10.3

The Design of Feedback Control Systems

0.8

0.6

0.4

0.2

0

0

2

4

FIGURE DP10.3
Step response for the mast flight system.

6
Time (sec)

8

10

12

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

581

Design Problems

DP10.4

One possible compensator is
Gc (s) = 5682

s + 12.6
.
s + 87.3

The step response is shown in Figure DP10.4. The performance results

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Time (sec)

FIGURE DP10.4
Step response for the high speed train system.

are
P.O. = 4.44%
DP10.5

Ts = 0.36 sec Kv = 14.1 .

The design specifications are Kv > 200; Ts < 12 ms and percent overshoot
P.O. < 10%. The step response is shown in Figure DP10.5. A suitable
compensator is
Gc (s) = K

s + 403
,
s + 2336

where
K = 1.9476e + 13.
Then,
P.O. = 9.5%

Ts = 10 ms

Kv = 560 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

582

CHAPTER 10

The Design of Feedback Control Systems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

18

20

Time (ms)

FIGURE DP10.5
Step response for the tape transport system.

A solution to the problem is the PI controller
Gc (s) =

4.21s + 1.2
.
s

The step response is shown in Figure DP10.6.
1.2

1

0.8

Amplitude

DP10.6

0.6

0.4

0.2

0
0

1

2

3
Time (sec)

FIGURE DP10.6
Step response for the engine control system.

4

5

6

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

583

Design Problems

The performance results are
P.O. = 8.8%

and Ts = 2.14 .

The system is a type-1, so the steady-state error for a step input is zero,
as desired.
The jet aircraft roll angle motion is represented by the transfer function
G(s) =

10
.
(s + 10)(s2 + 2s + 20)

A good controls solution is obtained with a PID controller
Gc (s) =

10s2 + 20s + 150
.
s

The system is type-1, so the steady-state tracking error is zero for a step
input. The performance results are
P.O. = 9.13%

and

Ts = 1.56 .

Step Response
1.4

1.2

1
Amplitude

DP10.7

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5
2
Time (sec)

FIGURE DP10.7
Step response for the jet aircraft roll control system.

2.5

3

3.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

584

CHAPTER 10

DP10.8

The Design of Feedback Control Systems

One good solution is obtained with the following PI controller
27.35(s + 2)
.
s

Gc (s) =

The system is type-1, so the steady-state tracking error is zero for a step
input. The step response is shown in Figure DP10.8.
Step Response
From: U(1)
1.4

1.2

0.8
To: Y(1)

Amplitude

1

0.6

0.4

0.2

0

0

0.2

0.4

0.6

0.8

1

1.2

Time (sec.)

FIGURE DP10.8
Step response for the windmill radiometer.

DP10.9

Consider the PID controller
Gc (s) = Kp + KD s +

KI
1.554s2 + 1.08s + 1
=
s
s

Gc (s) = K



s+a
s+b



s+c
s+d



= 6.04

(s + 10)(s + 2)
.
(s + 1)(s + 5)

Both are stabilizing in the presence of a T = 0.1 second time delay. For the
PID controller the phase margin is P.M. = 40o . For the lead-lag controller
the phase margin is P.M. = 45o . We find (for these particular designs)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

585

Design Problems

that the lead-lag controller is more able to remain stable in the process
of increasing time delay. For a time-delay of T = 0.2 seconds, the lead-lag
compensator has a phase margin of P.M. = 22o , while the PID controller
is unstable.
DP10.10

One solution is
Gc (s) =

50(s + 0.01)
.
s+2

The Bode magnitude is shown in Figure DP10.10. You want high gain at

Bode Diagram
80

60
System: sys
Magnitude (dB): 26.9

40

Magnitude (dB)

20

0

−20
System: sys
Magnitude (dB): −26.9

−40

−60

−80

−100
−4
10

−3

10

−2

10

−1

10

0

10

1

10

2

10

FIGURE DP10.10
Step response for the windmill radiometer.

low frequency to improve disturbance rejection and decrease sensitivity to
plant changes and low gain at high frequency to attenuate measurement
noise.
DP10.11

One solution is the PD controller
Gc (s) = 0008(s + 10) .
The step response is shown in Figure DP10.11. The closed-loop transfer

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 10

The Design of Feedback Control Systems

function is
T (s) =

s2

4
,
+ 3.4s + 4

where we use the prefilter
Gp (s) =

4
.
0.36s + 3.6

Step Response
1.4
System: sys_cl
Peak amplitude: 1.01
Overshoot (%): 0.637
At time (sec): 2.97

1.2

1
Amplitude

586

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2

2.5
Time (sec)

3

FIGURE DP10.11
Step response for the polymerase chain reaction system.

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

587

Computer Problems

Computer Problems
The m-file script and step response is shown in Figure CP10.1. The phase
margin and percent overshoot are
P.M. = 50o
P.O. ≈ 18% ,
respectively.

nnumc=[110]; denc=[1 0]; sysc = tf(numc,denc);
numg=[1]; deng=[1 10]; sysg = tf(numg,deng);
syss = series(sysc,sysg);
[Gm,Pm]=margin(syss);
Pm
%
sys_cl = feedback(syss,1);
[y,t]=step(sys_cl);
step(sys_cl); grid
S=stepinfo(y,t);
PO=S.Overshoot

Pm =
49.9158
PO =
17.5724

Step Response
1.4

1.2

1
Amplitude

CP10.1

0.8

0.6

0.4

0.2

0

0

0.2

0.4

0.6
Time (sec)

0.8

FIGURE CP10.1
Phase margin and step response for the closed-loop system.

1

1.2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

588

CHAPTER 10

CP10.2

The Design of Feedback Control Systems

Using a proportional controller the closed-loop characteristic equation is
1+K

s2

24.2
.
+ 8s + 24.2

A simple m-file script which computes the P.M. as a function of the gain
K yields the proportional controller gain K = 6. Checking the phase
margin of the system reveals that P.M. ≈ 40◦ , as desired.
n=24.2; d=[1 8 24.2]; sys = tf(n,d);
K=6;
margin(K*sys), grid
Bode Diagram
Gm = Inf dB (at Inf rad/sec) , Pm = 39.9 deg (at 11.6 rad/sec)

Magnitude (dB)

20
0
−20
−40
−60

Phase (deg)

−80
0
−45
−90
−135
−180
−1
10

0

10

1

10

2

10

3

10

FIGURE CP10.2
Bode plot with a proportional controller K = 6 in the loop.

CP10.3

The uncompensated system is type-1. To realize a zero steady-state error
to a ramp input we need to increase the system type by one. One controller
that does this is the PI controller:
Gc (s) =

KP s + KD
.
s

The step response is shown in Figure CP10.3 where it can be seen in the
tracking error plot that the settling time is Ts < 5 seconds. The actual
settling time is
Ts = 3.6 seconds .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

589

Computer Problems

KP=20; KD=10;
nc=[KP KD]; dc=[1 0]; sysc = tf(nc,dc);
n=1; d=[1 2 0]; sys = tf(n,d);
sys_o = series(sysc,sys);
sys_cl = feedback(sys_o,[1]);
t=[0:0.001:10];
sys1 = tf([1],[1 0]); sys_cl1 = series(sys_cl,sys1);
subplot(121)
y=step(sys_cl1,t);
plot(t,y,t,t,'--'), grid
xlabel(' Time (sec)'), ylabel('Ramp response')
e=y-t'; L=find(abs(e)>0.02);
Ts=t(L(length(L)))
subplot(122)
plot(t,e,[0 10],[0.02 0.02],':',[0 10],[-0.02 -0.02],':')
xlabel(' Time (sec)'), ylabel(' Track ing error')
grid
10

0.1

9
0.05
8
0

6

Tracking error

Ramp response

7

5
4
3

-0.05

-0.1

-0.15

2
-0.2
1
0

0

5
Time (sec)

10

FIGURE CP10.3
Ramp response with a PI controller Gc (s) =

CP10.4

-0.25

0

20s+10
s

5
Time (sec)

10

in the loop.

From the percent overshoot spec we determine that P.O. < 10% implies
ζ > 0.6. So, we target a phase margin P.M. = 100ζ = 60o . The m-file
script which generates the uncompensated Bode plot is shown in Figure CP10.4a.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

590

CHAPTER 10

The Design of Feedback Control Systems

numg = 100*conv([1 1],[1 0.01]);
deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101]));
sysg = tf(numg,deng)
w=logspace(-1,2,200);
[mag,phase,w]=bode(sysg,w);
[Gm,Pm,Wcg,Wcp]=margin(mag,phase,w);
%
Phi=60-Pm
Pm
Phi=(60-Pm)*pi/180;
alpha=(1+sin(Phi))/(1-sin(Phi))
M=-10*log10(alpha)*ones(length(w),1);
[mag,phase,w]=bode(sysg,w);
for i = 1:length(w),
magdB(i) = 20*log10(mag(1,1,i));
end
semilogx(w,magdB,w,M), grid
title('Uncompensated Bode Plot')
hold on
semilogx([.56072 5.6072 56.072 560.72],[20 0 -20 -40],'--')

È
Phi =
56.2111
Pm =
3.7889
alpha =
10.8408

Uncompensated Bode Plot
60

40

mag [dB]

20

0

-20

-40

-60

-80
10-1

100

101

102

FIGURE CP10.4
(a) Uncompensated Bode plot.

We assume that K = 1 and raise the gain at a later step to meet settling
time requirement. The uncompensated phase margin is P.M. = 3.7o , so
that the lead compensator needs to add φ = 56.2o . The script also calculates α = 10.84. Following the design procedure outlined in Dorf &
Bishop, we locate the compensator zero at ω = 2 rad/sec (see dashed line
in Figure CP10.4a). Then, p = αz implies p = 21.68. After several iter-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

591

Computer Problems

ations, we converge on K = 4 as a “good” value. The lead compensator
is
Gc (s) = 4

s+2
.
s + 22

The step response is shown in Figure CP10.4b. The compensated Bode is
shown in Figure CP10.4c.

K=4;
numg = 100*conv([1 1],[1 0.01]);
deng = conv([1 10],conv([1 2 2],[1 0.02 0.0101]));
sysg = tf(numg,deng)
numc=K*[1 2]; denc=[1 22]; sysc = tf(numc,denc);
sys_o = series(sysc,sysg);
sys_cl = feedback(sys_o,[1]);
t=[0:0.01:5];
f=10*pi/180;
[y,t,x]=step(f*sys_cl,t);
plot(t,y*180/pi), grid
xlabel(' Time (sec)')
ylabel('Attitude rate (deg/sec)'), pause
w=logspace(-1,2,200);
[mag,phase,w]=bode(sys_o,w);
[Gm,Pm,Wcg,Wcp]=margin(mag,phase,w);
bode(sys_o)
title(['Gain Margin = ',num2str(Gm),' Phase Margin = ',num2str(Pm)])

12

Attitude rate (deg/sec)

10

8

6

4

2

0
0

0.5

1

1.5

2

2.5
Time (sec)

FIGURE CP10.4
CONTINUED: (b) Step response.

3

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

592

CHAPTER 10

The Design of Feedback Control Systems
Gain Margin = 14.96 Phase Margin = 60.49

Gain dB

50
0
-50
-100
10-3

10-2

10-1
100

101

102

10-2

10-1

101

102

Phase deg

100
0
-100
-200
-300
10-3

100

FIGURE CP10.4
CONTINUED: (c) Bode plot with lead compensator.

CP10.5

The closed-loop transfer function is
θ(s)/θd (s) =

s2

K̄1 + K̄2 s
+ K̄2 s + K̄1

where K̄1 = K1 /J and K̄2 = K2 /J. A percent overshoot P.O. ≤ 20%
requires ζ > 0.45. Select as the initial damping
ζ = 0.7

(initial selection) .

For a second-order system with ζ = 0.7, we find that ω/ωn ≈ 0.9 when
|θ(s)/θd (s)| = 0.7. So, we select
ωn = ωB /0.9
as a starting choice. Therefore, since ωB = 10, we have
ωn = 11 .
The m-file script is shown in Figure CP10.5a. After several iterations, we
find a set of “good” values for
ζ = 0.8

and ωn = 4.5

(final selection) .

The step response and closed-loop Bode plot are shown in Figures CP10.5b
and CP10.5c.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

593

Computer Problems

% Par t (a)
wn=4.5; zeta=0.8; K2=2*zeta*wn; K1=wn^2;
% Par t (b)
num=[K2 K1]; den=[1 0 0]; sys = tf(num,den);
sys_cl = feedback(sys,[1]);
f=10*pi/180; % set-up for 10 deg step input
t=[0:.05:3];
[y,t,x]=step(f*sys_cl,t);
plot(t,y*180/pi), xlabel(' time [sec]'), ylabel(' theta [deg]'),
grid, pause
% Par t (c)
w=logspace(-1,2,400);
[mag,phase,w]=bode(sys_cl,w);
for i = 1:length(w),
magdB(i) = 20*log10(mag(1,1,i));
end
semilogx(w,magdB,[w(1) w(length(w))],[-3 -3]), grid
ylabel('Gain dB')

FIGURE CP10.5
(a) Script to generate the step response and the closed-loop Bode plot.

12

10

theta [deg]

8

6

4

2

0
0

0.5

1

1.5
time [sec]

FIGURE CP10.5
CONTINUED: (b) Step response.

2

2.5

3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

594

CHAPTER 10

The Design of Feedback Control Systems

5

0

Gain dB

-5

-10

-15

-20

-25
10-1

100

101

102

FIGURE CP10.5
CONTINUED: (c) Closed-loop Bode plot.

CP10.6

The settling time and phase margin specifications require that the dominant closed-loop poles have natural frequency and damping of ζ ≥ 0.45
and ωn ≥ 1.78. The uncompensated roots locus is shown in Figure CP10.6a.

10
+

K=10

8
6

x

4

Imag Axis

numg=[1 10]; deng=[1 2 20];
sysg = tf(numg,deng);
axis([-15,1,-10,10]);
rlocus(sysg); hold on
%
zeta=0.45; wn=1.7778;
x=[-10:0.1:-zeta*wn];
y=-(sqr t(1-zeta^2)/zeta)*x;
xc=[-10:0.1:-zeta*wn];
c=sqr t(wn^2-xc.^2);
plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':')
rlocfind(sysg),
hold off

2
0

o

-2
-4

x

-6
-8
+

-10

-14

-12

-10

-8
Real Axis

FIGURE CP10.6
(a) Uncompensated root locus.

-6

-4

-2

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

595

Computer Problems

From the final value theorem, we determine that
lim = sE(s) ≤ 0.1A

s→0

implies

A
= 0.1A .
1 + GGc (s)

Therefore, the compensated Kpcomp ≥ 9. With the compensator
Gc (s) = K

s+z
s+p

we find that
Kpcomp =

Kz
Kpuncomp .
p

But Kpuncomp = 0.5 and (from the uncompensated root locus) a gain of
K = 10 results in roots of the characteristic equation in the desired region.
Solving for
z
1 Kpcomp
=
≈2.
p
K Kpuncomp
Select z = 0.5 to minimize changing the root locus. Then, p = 0.25, and
the compensator is
Gc (s) = 10

s + 0.5
.
s + 0.25

The compensated root locus is shown in Figure CP10.6b and the step response is shown in Figure CP10.6c. The phase margin of the compensated

FIGURE CP10.6
CONTINUED: (b) Compensated root locus.

10
+

8
6
x

4

Imag Axis

numg=[1 10]; deng=[1 2 20];
sysg = tf(numg,deng);
numc=[1 0.5]; denc=[1 0.25];
sysc = tf(numc,denc);
sys_o = series(sysc,sysg);
axis([-15,1,-10,10]);
rlocus(sys_o); hold on
%
zeta=0.45; wn=1.7778;
x=[-10:0.1:-zeta*wn];
y=-(sqr t(1-zeta^2)/zeta)*x;
xc=[-10:0.1:-zeta*wn];
c=sqr t(wn^2-xc.^2);
plot(x,y,':',x,-y,':',xc,c,':',xc,- c,':')
rlocfind(sys_o)
hold off

2
0

o

o+ x

-2
-4

x

-6
-8
+

-10

-14

-12

-10

-8
Real Axis

-6

-4

-2

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

596

CHAPTER 10

The Design of Feedback Control Systems

system is P.M. = 62.3o and the settling time Ts < 5 seconds.
numg=[1 10]; deng=[1 2 20];
sysg = tf(numg,deng);
numgc=10*[1 0.5]; dengc=[1 0.25];
sysc = tf(numgc,dengc);
sys_o = series(sysc,sysg);
sys_cl = feedback(sys_o,[1]);
t=[0:0.1:5]; step(sys_cl,t)
[mag,phase,w]=bode(sys_o);
[gm,pm,w1,w2]=margin(mag,phase,w); pm

>>
pm =
62.3201

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (secs)

FIGURE CP10.6
CONTINUED: (c) Step response and phase margin verification.

CP10.7

Both design specifications can be satisfied with an integral controller
Gc (s) = K1 +

K2
10
=
.
s
s

The simulation results and m-file script are shown in Figures CP10.7a
and b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

597

Computer Problems
Unit Step Response

Phi dot

1.5

1

0.5
0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.7

0.8

0.9

1

Time (sec)
Unit Ramp Response

Tracking error

0

-0.05

-0.1
-0.15

0

0.1

0.2

0.3

0.4

0.5

0.6

Time (sec)

FIGURE CP10.7
(a) Simulation results.

K1=0; K2=10;
numc=[K1 K2]; denc=[1 0]; sysc = tf(numc,denc);
numg=[23]; deng=[1 23]; sysg = tf(numg,deng);
sys_o = series(sysc,sysg);
sys_cl = feedback(sys_o,[1]);
t=[0:0.01:1];
ys=step(sys_cl,t);
subplot(211)
plot(t,ys), xlabel(' Time (sec)'), ylabel('Phi dot')
title('Unit Step Response'), grid
u=t;
yr=lsim(sys_cl,u,t);
subplot(212)
plot(t,yr-u','--')
xlabel(' Time (sec)'), ylabel(' Track ing error')
title('Unit Ramp Response'), grid

FIGURE CP10.7
CONTINUED: (b) M-file design script.

CP10.8

From Example 10.3, we have that the loop transfer function is
Gc (s)G(s) =

8.1(s + z)
,
s2 (s + 3.6)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

598

CHAPTER 10

The Design of Feedback Control Systems

where z = 1. We want to determine a value of z so the the percent
overshoot is reduced from 46% to less than 32%. A valid design is
Gc (s)G(s) =

8.1(s + 0.45)
.
s2 (s + 3.6)

The m-file script and step response are shown in Figure CP10.8. The
percent overshoot is P.O.=27.7 %.
Step Response
From: U(1)
1.4

1.2

1

0.8
To: Y(1)

Amplitude

K1 = 8.1;
numg = [K1]; deng = [1 0 0];
sysg = tf(numg,deng);
numc = [1 0.45]; denc = [1 3.6];
sysc = tf(numc,denc);
sys_o = series(sysc,sysg);
sys_cl = feedback(sys_o,[1]);
step(sys_cl)
y=step(sys_cl);
po=100*(max(y)-1)

0.6

0.4

0.2

0

0

1.6

3.2

4.8

6.4

8

Time (sec.)

FIGURE CP10.8
Response of system with new lead compensator design.

CP10.9

From AP10.10, we have the transfer function is
Vo (s)
Vi (s)
1 + R2 C 2 s
=
.
1 + R1 C 1 s

T (s) =

Substituting C1 = 0.1 µF ,C2 = 1 mF , R1 = 10 kΩ, and R2 = 10 Ω yields
T (s) =

1 + 0.01s
.
1 + 0.001s

The frequency response is shown in Figure CP10.9.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

599

Computer Problems
Bode Diagrams

20

15

c1=0.0000001; c2=0.001;
r1=10000; r2=10;
n=[c2*r2 1]; d=[c1*r1 1];
sys=tf(n,d)
bode(sys)

Phase (deg); Magnitude (dB)

10

5

0
60
50
40
30
20
10
0
1
10

10

2

10

3

10

4

FIGURE CP10.9
Op-amp circuit frequency response.

CP10.10

The plot of K versus phase margin is shown in Figure CP10.10. The value
of K that maximizes the phase margin is K = 4.15.

60

FIGURE CP10.10
Plot of K versus phase margin.

55

50

45

40

P.M.

K=[0.1:0.01:10];
T=0.2;
for i=1:length(K)
ng=K(i)*[1 0.2]; dg=[1 6 0 0]; sysg=tf(ng,dg);
[gm,pm]=margin(sysg*sysp);
PM(i)=pm;
end
plot(K,PM), grid
[P,n]=max(PM); K(n)
xlabel('K'), ylabel('P.M.')

35

30

25

20

15

0

1

2

3

4

5
K

6

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

1 1

The Design of State Variable
Feedback Systems

Exercises
E11.1

The system is given by
ẋ = Ax + Bu
u = Kx
where


A=

0

1

−1 0






B=

1 0
0 1




and



K=

−k

0

0 −2k



 .

Then, with u = Kx, we have


ẋ = 

−k

1

−1 −2k



x .

The characteristic equation is


det[sI − A] = det 

s+k

−1

1

s + 2k



 = s2 + 3ks + 2k 2 + 1

= s2 + 2ζωn s + ωn2 = 0 .

Solving for k where ωn2 = 2k 2 + 1 and ζ = 1 (critical damping) yields
k = 2.
600

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

601

Exercises

E11.2

Let
u = −k1 x1 − k2 x2 + r .
Then,


ẋ = 

0



1



x+

9 − k1 −k2

0
1



r ,

and
det(sI − A) = s2 + k2 s + k1 − 9 = 0 .
We want ζ = 1, so the desired characteristic equation is
pd (s) = (s + co )2 ,
where co is to be determined to meet Ts = 4 and where k2 = 2co and
k1 = c2o + 9. Solving for the state response of x1 (t) to a unit step input
we find
x1 (t) = 1 − e−co t − co te−co t .
When t ≥ Ts = 4 sec we want x1 (t) ≥ 0.98. Solving for co at t = Ts yields
co = 1.459,
E11.3

k1 = 11.13,

and

k2 = 2.92 .

The controllability matrix is
Pc =

h

B AB

i



=

0

1

1 −3



 ,

and det Pc 6= 0, therefore the system is controllable. The observability
matrix is


Po = 

C
CA





=

0

2

0 −6



 ,

and det Po = 0; therefore the system is unobservable.
E11.4

The controllability matrix is
Pc =

h

B AB

i



=

0

0

2 −4



 ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

602

CHAPTER 11

The Design of State Variable Feedback Systems

and the det Pc = 0; therefore the system is uncontrollable. The observability matrix is


C

Po = 

CA





=



1 0

 ,

−10 0

and det Po = 0; therefore the system is also unobservable.
E11.5

The controllability matrix is
Pc =

h

B AB

i



=



1 −2
−2

 ,

3

and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is


Po = 

C
CA





=

1 0
0 1



 ,

and det Po = 1 6= 0; therefore the system is observable.
E11.6

The controllability matrix is
Pc =

h

B AB

i



=

0

1

1 −2



 ,

and det Pc 6= 0; therefore the system is controllable. The observability
matrix is


Po = 

C
CA





=

1 0
0 1



 ,

and det Po 6= 0; therefore the system is observable.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

603

Exercises

E11.7

The block diagram is shown in Fig. E11.7.

2
U(s)

12

1
s

1
s

+
- -

-

+

2

Y(s)

5
3
FIGURE E11.7
The block diagram for E11.7.

E11.8

The block diagram is shown in Fig. E11.8.

10
8

U(s)

4

1
s

+
- - -

+

-

1
s

1
3
9
FIGURE E11.8
The block diagram for E11.8.

1
s

2

+

++
Y(s)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

604

CHAPTER 11

E11.9

The Design of State Variable Feedback Systems

The controllability matrix is
Pc =

h

B AB

i



=

k1



k1 − k2

 ,

k2 −k1 + k2

and det Pc = −k12 + k22 . So, the condition for complete controllability is
k12 6= k22 .
E11.10

A matrix differential equation representation is




ẋ = 



0

1

0

0

−10 −6 −3

y = [−3 4
E11.11











0 
 0 





1 
x +  0 u
1



2]x + [0]u .

The system is given by
ẋ = Ax + Bu
y = Cx + Du
where




A=













0 
 0 
h
i






,
B
=
,
C
=
1 
1 2 0 , and D = [1] .
 0 

0 1
0 0

−2 0 −7

The controllability matrix is

Pc =

h

1



2

B AB A B

i



 0

=
 0




0

1 

1 −7 
 ,

1 −7

49



and det Pc = −1 6= 0; therefore the system is controllable. The observability matrix is




 C 



Po = 
 CA  =



CA2








1 2
0 1



0 

2 
 ,

−4 0 −13



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

605

Exercises

and det Po = −29 6= 0; therefore the system is observable.
The transfer function is
G(s) =

s2

6
.
+ 5s + 6

The response of the system to a unit step is
y(t) = 1 − 3e−2t + 2e−3t .
The step response is shown in Figure E11.12

1
0.9
0.8
0.7
Step Response

E11.12

0.6
0.5
0.4
0.3
0.2
0.1
0

0

FIGURE E11.12
Unit step response.

0.5

1

1.5

2

2.5
Time (s)

3

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

606

CHAPTER 11

The Design of State Variable Feedback Systems

Problems
P11.1

Consider the system
ẋ = x + u
u = −kx .
So,
ẋ = x − kx = (1 − k)x
and
x(t) = e(1−k)t x(0) .
The system is stable if k > 1. Computing the value of J (assuming k > 1)
yields
J=

Z

∞

e2(1−k)t x2 (0)dt =

0

1
.
k−1

Thus, J is minimum when k → ∞. This is not physically realizable. Select
k = 35. Then, the value of the performance index J is
J=

1
.
34

The system is not stable without feedback.
P11.2

(a) The performance index is given
J=

∞

Z

(x2 + λu2 )dt .

0

The system is
ẋ = x + u
u = −kx .
So,
J=

Z

0

∞

2

2 2

(x + λk x )dt =

Z

∞

2

2

2

(1 + λk )x dt = (1 + λk )

0

Z

∞

x2 dt .

0

Carrying out the integration (assuming k > 1) yields
J = (1 + λk 2 )

1
.
k−1

We want to determine k that minimizes J. Taking the partial of J

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

607

Problems

with respect to k and setting the result to zero yields
∂J
λk 2 − 2λk − 1
=
=0,
∂k
(k − 1)2
or
λk 2 − 2λk − 1 = 0 .
Solving for k yields
k =1+

r

1
,
λ

1+

where we reject the solution k = 1 −

q

1 + λ1 , since we require k > 1.

(b) For λ = 2, we determine that k = 2.2 and Jmin = 8.9.
P11.3

The system is given by


ẋ = 

1 0
−1 2





x + 

1
1



u

u = −k(x1 + x2 ) = −k[1 1]x .

Then, with feedback applied, the system is


ẋ = 

(1 − k)

−k

−(1 + k) (2 − k)



x .

Solving
HT P + PH = −I
yields
2p11 (1 − k) − 2p12 (k + 1) = −1
p12 (3 − 2k) − p11 k − p22 (k + 1) = 0
−2kp12 + 2p22 (2 − k) = −1 .
Solving for p11 , p12 and p22 yields
−(2k 2 − 6k + 7)
4(4k 2 − 8k + 3)
2k 2 − 2k − 1
=
4(4k 2 − 8k + 3)

p11 =
p12

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

608

CHAPTER 11

The Design of State Variable Feedback Systems

p22 =

−(2k 2 − 6k + 3)
.
4(4k 2 − 8k + 3)

The performance index is computed to be
J = xT (0)Px(0) = p11 + 2p12 + p22 =

1
,
2k − 1

when x(0) = [1 1]T . So as k → ∞, J → 0. The system is unstable without
feedback.
The performance index is
J = xT (0)Px(0) = p11 − 2p12 + p22 .
From Example 11.12 in Dorf and Bishop, we determine that
J=

2k 2 + 1
.
2k 2

So, when k → ∞, the performance index J → 1. The plot of J versus k
is shown in Figure P11.4.

60

50

40

J

P11.4

30

20

10

0

0

1

2

3

FIGURE P11.4
The performance index J versus k.

4

5
K

6

7

8

9

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

609

Problems

P11.5

The system is given by


ẋ = 

0 1
0 0

u = −kx .





x + 

0 0
1 1



u

The performance index is
J=

Z

∞

T

T

(x x + u u)dt =

0

Z

∞

(1 + k 2 )(xT x)dt .

0

First, we solve
HT P + PH = −(1 + k 2 )I ,
yielding,
(1 + k 2 )
2k
3
k + k2 + k + 1
=
2k 2
3
2k + k 2 + 2k + 1
=
.
2k

p12 =
p22
p11

The performance index is then given by
J = p11 + 2p12 + p22 =

2k 4 + 4k 3 + 3k 2 + 4k + 1
.
2k 2

Taking the partial of J with respect to k, setting the result to zero and
solving for k yields
∂J
2k 4 + 2k 3 − 2k − 1
=
=0
∂k
k3
or
2k 4 + 2k 3 − 2k − 1 = 0 .
Solving for k yields k = 0.90. The plot of J versus k is shown in Figure P11.5. The value of the performance index is
J = 6.95
when k = 0.90.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

610

CHAPTER 11

The Design of State Variable Feedback Systems

8.8
8.6
8.4
8.2

J

8
7.8
7.6
7.4
7.2
7
6.8
0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

k

FIGURE P11.5
The performance index J versus k.

P11.6

(a) For P11.3, we have
J=

1
.
2k − 1

So, as k → ∞, then J → 0. But k = ∞ is not a practical solution, so
select k = 10. Then, J = 1/19, and


ẋ = 

−9 −10
−11

−8



 x = Ax .

The closed-loop system roots are determined by solving
det[sI − A] = s2 + 17s − 38 = 0 ,
which yields s = −19 and s2 = 2. The system is unstable. The original
system was unstable, and it remains unstable with the feedback. In
general,


ẋ = 

(1 − k)

−k

−(1 + k) (2 − k)



 x = Ax

and det[sI − A] = s2 + s(2k − 3) + (2 − 4k) = 0. A Routh-Hurwitz
analysis reveals that the system is unstable for all k.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

611

Problems

(b) For P11.4, we have


ẋ = 

0

1

−k −k



 x = Ax ,

and
det[sI − A] = s2 + ks + k = 0 .
The performance index was found to be
J =1+

4k + 1
.
2k 2

As k → ∞, we have J → 0. But k = ∞ is not a practical choice for
k. Select k = 10. Then,
det[sI − A] = s2 + 10s + 10 = (s + 1.13)(s + 8.87) .
The closed-loop system is stable.
(c) In P11.5, we found that k = 0.90 for Jmin . We are given


ẋ = 

0

1

−k −k



x

and
det[sI − A] = s2 + ks + k = s2 + 0.9s + 0.9
= (s + 0.45 + j0.835)(s + 0.45 − j0.835) .
P11.7

The closed-loop system is


ẋ = 

0

1

−k1 −k2



 x = Hx ,

and
det[sI − H] = s2 + k2 s + k1 = s2 + 2ζωn s + ωn2 = 0 .
We desire ωn = 2, so set k1 = 4. With xT (0) = [1, 0], we have J = p11 ,
and solving
HT P + PH = −I

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

612

CHAPTER 11

The Design of State Variable Feedback Systems

yields



0

−4

1 −k2






p11 p12



p11 p12



−1

0

0

−1

+

p12 p22

p12 p22

=

and
p11 =




0

1

−4 −k2






 ,

20
k 2 + 20
k2
+
= 2
.
8
8k2
8k2

Select
k2 =
for Jmin , where Jmin =

√

5
2 .

√

20

Then

det[sI − H] = s2 +

√

20s + 4 = 0 ,

and ωn = 2 and ζ = 1.12. The system is overdamped.
P11.8

From Example 11.11 in Dorf and Bishop, we have


So,

P=

k22 +2
2k2
1
2

1
2
1
k2

J = xT (0)Px(0) =



 .

k22 + 2
2k2

when xT (0) = [1 0]. Taking the partial of J with respect to k2 and setting
the result to zero yields
∂J
k2 + 2
=1− 2 2 =0 .
∂k2
2k2
Solving for the optimum value of k2 yields
√
k2 = 2 .
P11.9

Let x1 = φ and x2 = ω. We have that
ω=

dφ
.
dt

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

613

Problems

The state equations are
x˙1 = x2
x˙2 = Ku .
Select a feedback such that
u = −x1 − K1 x2 + r
when r(t) is the reference input. Then,


ẋ = 

and

0

1

−K −KK1





x+

0
K



r ,

det[sI − A] = s2 + K1 Ks + K .
so that the overshoot is 4%. Since Ts = 1 = ζω4n , we
√
require ζωn = 4 or ωn = 4 2. Then, s2 + 8s + 32 = s2 + K1 Ks + K, or
8
K = 32 and K1 = 32
= 41 .
We desire ζ =

P11.10

√1 ,
2

The system with feedback is given by


ẋ = Ax = 

−10 −25
1

0



x ,

where x1 (0) = 1, and x2 (0) = −1. The characteristic equation is


det[sI − A] = det 

s + 10 25
−1

s



 = s(s + 10) + 25 = s2 + 10s + 25 = 0 .

The roots are s1,2 = −5. The solution is


x(t) = 

φ11 φ12
φ21 φ22





 x(0) = 

φ11 − φ12
φ21 − φ22




since x1 (0) = 1 and x2 (0) = −1. We compute the elements of the state
transition matrix as follows:
φ22 (t) = (1 + 5t)e−5t

and

φ21 (t) = te−5t ,

therefore
x2 (t) = −(1 + 4t)e−5t .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

614

CHAPTER 11

The Design of State Variable Feedback Systems

Similarly,
φ11 (t) = (1 − 5t)e−5t

φ12 = −25e−5t .

and

Therefore,
x1 (t) = (1 + 20t)e−5t .
P11.11

Let
u = −k1 x1 − k2 x2 + αr
where r(t) is the command input. A state variable representation of the
plant is


−5 −2

h

0 1

ẋ = 
y=

2

0
i





x+

x+

h

0

i

0.5
0

u.



u

The closed-loop transfer function is
T (s) =

α
.
s2 + (k1 /2 + 5)s + 4 + k2

To meet the performance specifications we need ωn = 4.8 and ζ = 0.826.
Therefore, the desired characteristic polynomial is
q(s) = s2 + 2(0.826)4.8s + 23 = s2 + 8s + 23 .
Equating coefficients and solving for k1 and k2 yields k2 = 19 and k1 = 6.
Select α = 23 to obtain zero steady-state error to a step input.
P11.12

A state variable representation of the dc motor is







ẋ = 






−3 −2 −0.75 0 0





1












0 
 0 





0 
x +  0 u






0 
 0 




3

0

0 0

0

2

0 0

0

0

1 0

0

0

0 2 0

y = [0 0 0 0 2.75]x .

0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

615

Problems

The controllability matrix is


1 −3



 0


Pc = 
 0


 0


0

3

4.5 −18

3 −9



13.5 


18 



−18 


9

0

6 −18

0

0

6

0

0

0



12

and the det Pc 6= 0, so the system is controllable. The observability matrix
is


0



 0


Po = 
 0


 0


33

0

0

0

2.75

0

0

5.5

0

0

5.5

0

0

11

0

0

0

0

0

0

0

and the det Po 6= 0, so the system is observable.
P11.13








 ,






To meet the Kv = 35 specification, we need K = 2450. A state variable
representation is


ẋ = 

0

1

0 −70





x + 

0
2450



u

y = [1 0]x .
Let
u = −k1 x1 − k2 x2 .
Then, the closed-loop characteristic equation is
q(s) = s2 + (2450k2 + 70)s + 2450k1 = 0 .
The desired characteristic polynomial is
s2 + 72.73s + 2644.63 = 0
where we select ζ = 0.707 and ωn = 51.42 to meet the performance
specifications. Equating coefficients and solving for the gains yields k1 =

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

616

CHAPTER 11

The Design of State Variable Feedback Systems

1.08 and k2 = 0.0011.
P11.14

Let
u = −k1 x1 − k2 x2 − k3 r
where r(t) is the command input. Then, the closed-loop system in state
variable form is


ẋ = 

−10 − k1 −k2
1

0





x+

1
0



r

y = [0 1]x .
To meet the performance specifications, we want the closed-loop characteristic polynomial to be
q(s) = s2 + 8s + 45.96 = 0
where ζ = 0.59 and ωn = 6.78. The actual characteristic polynomial is
det(sI − A) = s2 + (10 + k1 )s + k2 = 0 .
Equating coefficients and solving for the gains yields k2 = 45.96 and
k1 = −2. Select k3 = k2 = 45.96 to obtain a zero steady-state error to a
step input. This results in a settling time of Ts = 0.87 s and a percent
overshoot of P.O. = 10%.
P11.15

The transfer function is
G(s) = C(sI − A)−1 B =

1
.
s+1

The system is not controllable and not observable.
P11.16

Let
u = −Kx .
Then, Ackermann’s formula is
K = [0, 0, ..., 1]P −1
c q(A)
where q(s) is the desired characteristic polynomial, which in this case is
q(s) = s2 + 2s + 10 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

617

Problems

A state-space representation of the limb motion dynamics is


−4

ẋ = 

0

1 −1





x + 



1

u .

0

The controllability matrix is


Pc = [B AB] = 

1 −4
0

1




and



P−1
c =

1 4
0 1



 .

Also, we have


q(A) = A2 + 2A + 10I = 

18 0
−3 9



 .

Using Ackermann’s formula, we have
K = [−3

9] .

P11.17

The system is either uncontrollable or unobservable if a = 5 or a = 8.
Both of these values correspond to system real poles. So, if a takes on
either value, a pole-zero cancellation occurs in the transfer function.

P11.18

A matrix differential equation representation is


ẋ = 

0

1

−1 −2
y = [1





x+

0
1



u

0]x .

Let u(t) = −k1 x1 − k2 x2 . Then, the closed-loop characteristic equation is
q(s) = s2 + (2 + k2 )s + 1 + k1 = 0 .
We desire the characteristic equation
√
s2 + 2 2s + 2 = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

618

CHAPTER 11

The Design of State Variable Feedback Systems

Equating
coefficients and solving for the gains yields k1 = 1 and k2 =
√
2 2 − 2 = 0.828.
P11.19

A state space representation is


ẋ = 

0



1



0

x+

3 −2
y = [3

1]x .



1

1



r

The controllability matrix is
0

Pc = 



 ,

1 −2

and det Pc 6= 0, so the system is controllable. The observability matrix is


Po = 

3 1
3 1



 ,

and the det Po = 0, so the system is not observable.
P11.20

The characteristic equation associated with A is
s2 (s2 + 0.2s + 0.0015) = 0 .
There are two roots at the origin, so the system is unstable. The system
can be stabilized with
δ = −k1 x1 − k3 x3 = 20x1 − 10x3 .

P11.21

(a) Let x1 = i1 , x2 = i2 and u = v. Then, the state equation is


ẋ = 

−(R1 +R3 )
L1
R3
L2

R3
L1
−(R3 +R2 )
L2



x + 

Also,
y = vo ,
but
y = [R3



− R3 ]x .

1
L1

0



u .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

619

Problems

(b) The observability matrix is


Po = 

C
CA





=

R3


3
− RL1 R
− R32
1

and
det Po =



1
L1

+

1
L2



R2 R3
L2

−R3

+ R32



1
L1

+

1
L2

R2 R1
−
R32 .
L2
L1


So, when
R2
R1
=
,
L1
L2
det Po = 0 and the system is not observable.
(c) Let
a=

R1 + R3
,
L1

b=

R3 + R2
.
L2

and

Then


det[sI − A] = det 
"

(s + a)
3
−R
L2

3
−R
L1

(s + b)




#

R32
= (s + a)(s + b) +
= (s + r)2
L1 L2
= s2 + (a + b)s + ab +

R32
.
L1 L2

The system has two equal roots when
R32
(a + b) − 4 ab +
L1 L2
2

!

or


R1 + R3 R3 + R2
+
L1
L2

2

−4

(R1 + R3 )(R3 + R2 ) + R32
=0.
L1 L2



 

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

620

CHAPTER 11

(a) Without state feedback the state differential equation is given by


ẋ = 

y=

h

−0.4 −1
1
0 1

0
i

x.





x+

1
0



u

The step response is shown in Figure P11.22a.
(a) Without state feedback

2

x2

1.5
1
0.5
0
0

2

4

6

8

10

12

14

16

18

20

1.4

1.6

1.8

2

Time (sec)
(b) With state feedback

1.5

1

x2

P11.22

The Design of State Variable Feedback Systems

0.5
0
0

0.2

0.4

0.6

0.8

1

1.2

Time (sec)

FIGURE P11.22
Step response (a) without state feedback, and (b) with state feedback.

(b) Consider state feedback
u = −K(ax2 + bx1 ) + cr
where r is the reference input and K, a, b and c are to be determined.
Then, the state differential equation is


−0.4 − Kb −1 − Ka

h

0 1

ẋ = 

y=

1
i

0
x,





x+

c
0



r

and det(sI − A) = s2 + (0.4 + Kb)s + (1 + Ka) = 0. Our specifications
4
are P.O. = 5% and Ts = 1.35 sec. So, ζ = 0.69 and ωn = ζ1.35
= 4.3.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

621

Problems

Solving for K, a and b yields
Ka = ωn2 − 1
and
Kb = 2ζωn − 0.4 .
Select K = 1. Then, a = 17.49 and b = 5.53. Select c = 1 + Ka to
achieve a zero steady-state tracking error.
(c) The step response is shown in Figure P11.22b for the system with
state feedback.
P11.23

Using the internal model design method for step inputs, we have








 0 1 0 
 0

 e

ė

 =  0 0 1 
+ 0




 z

ż
0 0 0
1




where we choose





w ,



w = −K1 e − K2 z .
To place the poles at s = −10 and s = −2±j we use Ackermann’s formula
to compute
K1 = 50
K2 = [45

14] .

The compensator has the form shown in Figure 11.14 in Dorf and Bishop.
P11.24

Using the internal model design method for ramp inputs, we have





 0 1 0 0 
 e
 ė  

 
0 0 1 0 


 ë  = 


 
 ė

  0 0 0 1 


 z
ż
0 0 0 0


where we choose









 0 


 0 





+
w
 
  0 




1

w = −K1 e − K2 ė − K3 z .
To place the poles at s = −20 and s = −2 ± 2j we can use Ackermann’s
formula. We also need an additional pole (must be a stable pole); select

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

622

CHAPTER 11

The Design of State Variable Feedback Systems

s = −20 as the fourth pole. Then,
K1 = 3200
K2 = 1920
K3 = [568 44] .
The compensator has the form shown in Figure 11.16 in Dorf and Bishop.
P11.25

The observability matrix is


Po = 

C
CA





=

1

−4

21 −36



 ,

and det Po = 48 6= 0; therefore the system is completely observable. The
desired poles of the observer are s1,2 = −1. This implies that the desired
characteristic polynomial is
pd (s) = s2 + 2s + 1 .
The actual characteristic polynomial is
det |λI − (A − LC)| = det

λ − 1 + L1

−4 − 4L1

5 + L2

λ − 10 − 4L2

= λ2 + (L1 − 4L2 − 11)λ + 10L1 + 8L2 + 30 = 0 .
Solving for L1 and L2 yields


L=

L1
L2





=

−0.25
−3.3125



 .

Checking we find that det(λI − (A − LC)) = s2 + 2s + 1. The response of
the estimation error is shown in Figure P11.25, where e(0) = [ 1 1 ]T .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

623

Problems

Response to Initial Conditions
2.5
To: Out(1)

2
1.5
1

Amplitude

0.5
0
1
To: Out(2)

0.5
0
-0.5
?-1
-1.5

0

1

2

3

4

5

6

Time (sec )

FIGURE P11.25
Estimation error response to an initial condition.

P11.26

The observability matrix is






 C   2

 
 
Po = 
 CA  =  0

 

CA

2

−4
2

32

20



0 

−4 
 .

14



The det Po = 728 6= 0, hence the system is observable. The gain matrix




 0.14 



L=
 −0.93 



0.79

results in the observer poles at s1,2 = −1 ± j and s3 = −5, as desired.
P11.27

The observability matrix is


Po = 

C
CA





=

1

0

1 0



 .

The det Po = 0, hence the system is not completely observable. So, we
cannot find an observer gain matrix that places the observer poles at the
desired locations.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

624

CHAPTER 11

P11.28

The Design of State Variable Feedback Systems

Selecting K = 16 yields a zero steady-state error to a unit step input.
The step response is shown in Figure P11.28.

Step Response
1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

0

0.5

1

1.5
Time (sec)

FIGURE P11.28
Estimation error response to an initial condition.

P11.29

The system transfer function is
Y (s) =

2
U (s) .
s+3

The associated state variable model is
ẋ = −3x + 2u
y=x.

2

2.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

625

AP11.1

The closed loop system in state-space form is given by












 ẋ1  

 
 ẋ  = 
 2  

ẋ3

0

1

0

0

−1

2



−2KK1 −2KK2 −4 − 2KK3












  x1   0 

 

 x  +  0 u
 2  



x3

2K



x1 
h
i



y= 1 0 0 
 x2  .


The closed-loop transfer function is
T (s) =

s3

+ (2KK3 +

5)s2

x3



4K
.
+ (4KK2 + 2KK3 + 4)s + 4KK1

Setting the steady-state error to zero, we determine that
ess = 1 − T (0) = 1 −

1
.
K1

Solving for K1 yields
K1 = 0.5 .
Choosing
K2 = 0.5

and K3 = 1.5

results in a percent overshoot of P.O. = 2.82%.
AP11.2

A state variable representation is given by
ẋ = Ax + Bu
where




 −3 −1 −1 


A=
0
0 
 4
 ,



0

1

0





 3 



B=
 0  .



0



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

626

CHAPTER 11

The Design of State Variable Feedback Systems

Let
u = −Kx .
Then, with
K=

h

4.00 24.33 39.67

i

,



b1

the closed-loop system poles are s = −4, −5, and −6.
AP11.3

Given


A=

0

1

−1 −2



,

and B = 

b2



 ,

we compute the determinant of the controllability matrix as
det Pc = det[B AB] = − (b1 + b2 ) .
The system is controllable if and only if the determinant is non-zero. So,
for the system to be controllable, we require that b2 6= −b1 .
AP11.4

Consider the state variable feedback law
u = −Kx .
Using Ackermann’s formula, we determine that
K = [−14.2045 − 17.0455 − 94.0045 − 31.0455]
results in the closed-loop system characteristic roots at s = −2±j, s = −5
and s = −5.

AP11.5

The closed-loop transfer function for the system is
T (s) =

s3

+ (9 + 2K3

)s2

2Kp
.
+ (26 + 2K2 + 10K3 )s + (26 + 6K2 + 12K3 )

Setting the steady-state error for a step input to zero yields
ess = 1 −

2Kp
=0.
26 + 6K2 + 12K3

Solving for Kp in terms of K2 and K3 yields
Kp = 13 + 3K2 + 12K3 .
Now, choosing
K2 = 5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

627

K3 = 2
results in the closed-loop characteristic roots at
s1 = −4 s2 = −4 s3 = −5 .
Also, the prefilter gain is
Kp = 52 .
AP11.6

(a) A state variable representation is given by


A=
C=

h

0

1

−1 −2
1 0

i

.



 ,



B=

0
1



 ,

Since the determinant of the controllability matrix det[B AB] 6= 0,
the system is controllable.
(b) The state variable representation is
ẋ = Ax + Bu ,
or



ẋ1
ẋ2





=

0

1

−1 −2




x1
x2





+

The determinant of the controllability matrix

1
−1



u .

det Pc = det[B AB] = 0 .
Therefore, the system is uncontrollable.
AP11.7

The closed-loop transfer function is
T (s) =

s3

+ (10 + 60K3

)s2

120
.
+ (16 + 120(K3 + K2 ))s + 120

The state feedback gains
K2 = 0.283

and

K3 = 0.15

place the poles at the desired locations. The plot of the roll output for a
unit step disturbance is shown in Figure AP11.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

628

CHAPTER 11

The Design of State Variable Feedback Systems

0.35

0.3

Amplitude

0.25

0.2

0.15

0.1

0.05

0
0

0.5

1

1.5

2

2.5

3

3.5

4

Time (secs)

FIGURE AP11.7
Roll angle response to a step disturbance.

AP11.8

The state equations are (using the parameters of P3.36 in Dorf and
Bishop)
8
1
[80θ − 50h] = −x1 + x2
50
5
θ̇ = ẋ2 = ω = x3
Km
Km Kb
Km Ka
353
25000
ω̇ = ẋ3 =
ia = −
ω+
vi = −
x3 +
vi .
J
JRa
JRa
30
3
ḣ = ẋ1 =

In state variable form we have (without feedback)


8
 −1 5

ẋ = 
 0 0



0
1

0 − 353
30

0









x + 







0
0
25000
3



 vi .



(a) In this case we have vi = −kh + ar = −kx1 + ar, where k and a are
the parameters to be determined and r is the reference input. With
the feedback of h(t) we have




ẋ = 



−1

8
5

0

0

0

1

353
− 25000
3 k 0 − 30









x+





0
0
a 25000
3





r .



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

629

Since we only have one parameter to adjust, namely k, we will probably not be able to simultaneously meet both design specifications, In
fact with
k = 0.00056
we obtain the percent overshoot P.O. = 9.89%. The settling time criterion cannot simultaneously be met—the best that can be obtained
is Ts ≈ 7.5 seconds. In this case, we choose a = 0.00056 to make the
steady-state value of h(t) = 1.
(b) In this case we have vi = −k1 h − k2 θ + ar = −k1 x1 − k2 x2 + ar,
where k1 , k2 , and a are the parameters to be determined and r is the
reference input. Since we have two parameter to adjust, namely k1
and k2 we will probably be able to simultaneously meet both design
specifications. In fact with
k1 = 0.00056

and k2 = 0.001

we obtain the percent overshoot P.O. = 4.35%. The settling time
criterion is easily met— Ts ≈ 5 seconds. In this case, we choose a =
0.0012 to make the steady-state value of h(t) = 1.
AP11.9

(a) The state vector differential equation is


 0 1

 −2 0

ẋ = 
 0 0







 0 


 0 


u ,
x + 



1 
 0 




0 0 

1 0 

0



1 0 −1 0

1

where x1 = z, x2 = ż, x3 = y and x4 = ẏ.
(b) The characteristic equation is

s4 + 3s2 + 1 = (s + j0.618)(s − j0.618)(s + j1.618)(s − j1.618) = 0 .
So, the system is oscillatory.
(c) Let u = −kx4 . Then characteristic equation is
s4 + ks3 + 3s2 + 2ks + 1 = 0
which is stable if k > 0.
(d) Rewrite the characteristic equation as
1+

ks(s2 + 2)
=0.
s4 + 3s2 + 1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

630

CHAPTER 11

The Design of State Variable Feedback Systems

The root locus is shown in Figure AP11.9. A reasonable solution for
k is k = 1.35.

3

2
x
o

1

Imag Axis

x

0

o

x

-1
o
x

-2

-3
-3

-2

-1

0

1

2

3

Real Axis

FIGURE AP11.9
s(s2 +2)
Root locus for 1 + k s4 +3s2 +1 = 0.

AP11.10

The state differential equation is
ÿ = ky + αu
where k and α depend on the system parameters, such as mass and length.
The transfer function is
y
α
= 2
u
s −k
which is unstable at the top of the arc. Since we can only use ẏ for
feedback, we have
ẏ
sα
= 2
.
u
s −k
Let
Gc (s) =

K1 s + K2
.
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

631

Then
GGc (s) =

α(K1 s + K2 )
(s2 − k)

and the closed-loop characteristic equation is
αK1 s + αK2 + s2 − k = 0
or
s2 + αK1 s + αK2 − k = 0 .
Select αK2 − k > 0 and αK1 > 0 for stability.
AP11.11

The state-space representation of the plant is
ẋ = Ax + Bu
y = Cx
where


A=

0

1

−2 −3



 ,



B=

0
1



 , and

C=

h





1 0

i

.

With the intermediate variables defined as
z = ẋ

and w = u̇

we have

where



1
0
 0

ė

= 0
0
1


ż
0 −2 −3






 0 




e

+ 0 w






z
1




e=y−r .
To meet the design specifications, we require the closed-loop poles to lie
to the left of the line in the complex plane defined by s = −0.8. We choose
K2 = [10 3]
and
Gc (s) =

8
.
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

632

CHAPTER 11

The Design of State Variable Feedback Systems

This places the closed-loop poles at s = −2, −2 and −2. The closed-loop
transfer function with the internal model controller is
T (s) =

8
.
s3 + 6s2 + 12s + 8

The step response is shown on Figure AP11.11.

1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

2

4

6

8

10

12

14

Time (secs)

FIGURE AP11.11
Internal model controller step response.

AP11.12

The state-space representation of the plant is
ẋ = Ax + Bu
y = Cx
where


A=

0

1

−2 −3



 ,



B=

0
1



 , and

With the intermediate variables defined as
z = ẍ

and w = ü

C=

h

1 0

i

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

633

we have




 0 1

 0 0




 ė 


 ë  = 

 

 

ż

where e = y − r.



0



0   
 e  
 
1
0 

+

ė

 

0 0
0
1   



z
0 0 −2 −3



0 

0 


w

0 

1



6

5

Amplitude

4

3

2

1

0
0

1

2

3

4

5

6

Time (secs)

FIGURE AP11.12
Internal model controller ramp response.

To meet the design specifications, we require the closed-loop poles to lie
to the left of the line in the complex plane defined by s = −0.67. We
choose

w = −[K1 K2







 e 
 e 







K3 ] 
 ė  = −[16 32 22 5]  ė  .





z

Then,



Gc (s) =

K1 + K2 s
16 + 32s
=
.
2
s
s2

z

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

634

CHAPTER 11

The Design of State Variable Feedback Systems

The closed-loop transfer function with the internal model controller is
T (s) =

s4

+

8s3

32s + 16
.
+ 24s2 + 32s + 16

This places the closed-loop poles at s = −2, −2, −2 and −2. The ramp
response is shown in Figure AP11.12.
AP11.13

The controllability matrix is


−5 −3



4

−3

22

44

Pc = 

1




18

and the observability matrix is

Po = 



 .

Computing the determinants yields
det Pc = −87 6= 0

and

det P0 = 242 6= 0 ,

hence the system is controllable and observable. The controller gain matrix
K=

h

3.02 6.11

i

places the closed-loop poles at the desired locations. Similarly, the observer gain matrix


L=

2.38
−1.16




places the observer poles at the desired locations.
AP11.14

The controllability matrix is


 0

Pc = 
 0


0
4

4 −12

4





−12 



24

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

635

and the observability matrix is


2



Po = 
 −16


−9

2

29

41





−4 −15 
 .


120

Computing the determinants yields
det Pc = −64 6= 0 and

det P0 = 10870 6= 0 ,

hence the system is controllable and observable. The controller gain matrix
K=

h

−0.5 1.25 0.5

and the observer gain matrix



i



 57.43 



L=
 −16.11 



−104.43

yields the desired closed-loop system poles and observer poles, respectively.
AP11.15

The state-variable representation of the system is


ẋ = 

0

1

−7 −2





x+

0
1

y = [ 1 4 ]x + [0]u .



u

The observability matrix is


P0 = 

1

4

−28 −7



 ,

and det P0 = 105 6= 0, hence the system is observable. The observer gain
matrix


L=

−7.18
6.29




places the observer poles at s1,2 = −10 ± 10, as desired.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

636

CHAPTER 11

The Design of State Variable Feedback Systems

Design Problems
CDP11.1

A state variable representation is


0

h

1 0

ẋ = 
y=

1

0 −33.14
i

x





0

x+

0.827



 va

where x1 = x and x2 = ẋ. Note that we are neglecting the motor inductance and assuming that the position x(t) is the output. Assume that we
have available for feedback the angle θ and angle rate θ̇ (see CDP4.1), so
that
va = −

k1
k2
x1 − x2 + au
r
r

where u(t) is the reference input (that is, the desired position x(t)), the
gains k1 and k2 and the scaling parameter a are to be determined. Recall
that
x = rθ = 0.03175θ .
With the feedback in the loop we have


ẋ = 
y=

h

0

1

−26.03k1 −33.14 − 26.03k2
1 0

i

x





x+

0
0.827a

Choosing k1 = 50, k2 = 1 and a = 1574.1 results in
P.O. = 1.1%

and

Ts = 0.11 second .

The closed-loop poles are s1,2 = −29.59 ± 20.65j.
DP11.1

The governing differential equation is
ÿ − 2000y = −20i .
In state variable form, the system is described by


ẋ = 

0

1

2000 0





x + 

0
−20



i .



u

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

637

Design Problems

Consider the state feedback
i = −k1 x1 − k2 x2 + βr
where r(t) is the reference input and k1 , k2 and β are to be determined.
Then, the closed-loop system is


ẋ = 

0

1

2000 − 20k1 −20k2





x+

0
−20β



r .

The characteristic equation is
s2 + 20k2 s − 2000 + 20k1 = 0 .
For stability, let 20k1 − 2000 > 0. Select k1 = 125. Then, ωn = 22.36
k2 =

2ζωn
.
20

Let ζ = 0.59 to meet 10% overshoot specification. Thus,
k2 =

2(0.59)(22.36)
= 1.32 .
20

The closed-loop transfer function is
T (s) =

s2

−20β
.
+ 26.4s + 500

s2

500
.
+ 26.4s + 500

Choose β = −25 so that
T (s) =
The feedback law is
i = 125x1 + 1.32x2 − 25r .
DP11.2

The automobile engine control system (see DP10.8 in Dorf and Bishop)
is modeled as
G(s) =

2e−sT
.
(0.21s + 1)(4s + 1)

In this case, we will assume the delay is negligible. Therefore, T = 0. A

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 11

The Design of State Variable Feedback Systems

state variable representation of the system is


ẋ = 

0

1

−1.19 −5.01

y = [1 0]x .





x + 

0
1.19



r

Let
r(t) = −k1 x1 − k2 x2 + k3 u
where u(t) is the command input. Using ITAE methods, our desired characteristic polynomial is
q(s) = s2 + 1.4ωn s + ωn2 = 0 .
Select ωn = 11.315 to obtain a settling time Ts < 0.5 seconds. The characteristic polynomial of the closed-loop system is
s2 + (5.01 + 1.19k2 )s + (1.19 + 1.19k1 ) = 0 .
Equating coefficients and solving for the gains yields
k1 = 106.59

and k2 = 9.235 .

Select k3 = 107.59 to yield a zero steady-state error to a step input.

1.2

1

0.8

Amplitude

638

0.6

0.4

0.2

0
0

0.1

0.2

0.3

0.4

Time (secs)

FIGURE DP11.2
The step response of the engine control system.

0.5

0.6

0.7

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

639

Design Problems

DP11.3

The compensator is
x̂˙ = [A − BK − LC] x̂ + Ly + Mr
ũ = −Kx̂
where


A − BK − LC = 
N = 363.64 ,

K=

h

−28.7

1

−365.19 −20

344.55 15.82

i





 ,

M=


, and L = 

0
200



 ,

28.7
165.19



 .

We selected the desired eigenvalues of A − BK at p = −10 ± 10j and the
desired eigenvalues of A − LC at q = −20 ± 10j. For initial conditions we
let x(0) = [1 1] and x̂(0) = [0 0].
1.5

Actual x1

x1

1

Estimated x1

0.5

0

0

0.1

0.2

0.3

0.4

0.5
Time (s)

0.6

0.7

0.8

0.9

1

0.4

0.5
Time (s)

0.6

0.7

0.8

0.9

1

6

x2

4

Estimated x2

2
0

−2

Actual x2
0

0.1

0.2

0.3

FIGURE DP11.3
The step response showing the actual and estimated states.

DP11.4

The design specifications are
(a) Percent overshoot < 20%
(b) Ts < 1.5s, and

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

640

CHAPTER 11

The Design of State Variable Feedback Systems

(c) steady-state error less than 20% of the input magnitude.
The state differential equation is
ẋ = Ax + Bu
y = Cx
where












1
0   0
 0

 
 
A=
 0 −σ1 −α1  =  0
g −α2 −σ2












 0   0

 
 
B=
 n  =  6.27

g

The transfer function is

9.8








9.8

1
−0.415
−1.43

and C =



0

h



−0.0111 
 ,


−0.0198

1 0 0

i

.

ns + nσ2 − α1 g
θ(s)
= 3
δ(s)
s + (σ1 + σ2 )s2 + (σ1 σ2 − α1 α2 )s + α1 g
6.27s + 0.0154
= 3
.
s + 0.435s2 − 0.0077 + 0.109
Let u = −K1 x1 − K2 x2 − K3 x3 . Then the closed-loop system matrix is


0


A − BK = 
 −nK1


1
−σ1 − nK2

g − gK1 −α2 − gK2

0





−α1 − nK3 
 ,


−σ2 − gK3

where K = [K1 K2 K3 ]. From the design specifications, we have the
desired roots at
s3 +a2 s2 +a1 s+ao = s3 +36s2 +225s+1350 = (s+30)(s+3+j6)(s+3−j6) = 0 .
The actual characteristic equation is
s3 + (gK3 + K2 n + σ1 + σ2 )s2 + (−α1 α2 − α1 gK2 + K1 n − α2 nK3
+ gK3 σ1 + K2 nσ2 + σ1 σ2 )s
+ α1 g − α1 gK1 + gK3 n + σ2 nK1 = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

641

Design Problems

Comparing coefficients yields







0

n

n
−α1 g + σ2 n



g









a2 − σ 1 − σ 2
  K1  

 

 
nσ2 − α1 g −α2 n + gσ1 
  K2  =  a1 + α1 α2 − σ1 σ2
0



gn

where

K3

a0 − α1 g

a2 = 36
a1 = 225
a0 = 1350 .
The solution for K is
K = [53.11 − 28.64 21.96] .
DP11.5

The controllability and observability matrices are


Pc = 



P0 = 

0.05

−0.04

0.001 −0.001
1

0

−0.8 0.02

Computing the determinants yields
det Pc = −1.002e − 05 6= 0





 and

 , respectively.

and Po = 0.02 6= 0 ,

hence the system is controllable and observable. The feedback gain matrix
K = [ 3820 −179620 ]
yields the desired closed-loop poles. The observer gain matrix


L=

120
180000




yields the desired observer poles. The integrated system is shown in Figure DP11.5.








from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

642

CHAPTER 11

The Design of State Variable Feedback Systems

A=

-0.8 0.02
-0.02 0

B=

0.05
0.001

System Model

u

.
x=Ax+Bu

y

C

Observer

Control Law

-K

x

^
x

.
^
^
~
x=Ax+Bu+Ly

~
^
y=y-Cx

+

C= 1 0

C

K= 3820 -179620
L=

120
180000

FIGURE DP11.5
Integrated controller and observer.

DP11.6

(a) The characteristic equation associated with the system matrix is
q(s) = s2 + (12 + K2 )s + (36 + K1 ) = 0 ,
where we have assumed state feedback of the form
u = −K1 x1 − K2 x2 .
The deadbeat control characteristic equation is
s2 + αωn s + ωn2 = 0 ,
where α = 1.82 and we use ωn = 9.64 to meet the settling time
specification. Then, equating coefficients and solving for the gains
yields
K1 = 56.93

and K2 = 5.54 .

(b) Since the closed-loop poles are located at s1,2 = −8.77 ± 4, we can
select the observer poles to be about ten times farther in the left-half
plane, or
s1,2 = −88, −88 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

643

Design Problems

Then the observer gains are


L=



164

 .

5740

(c) The block diagram is shown in Figure DP11.6.

0 1
-36 -12

A=

B=

0
1

System Model

u

y

C

Observer

Control Law

-K

x

.
x=Ax+Bu

^
x

.
^
^
~
x=Ax+Bu+Ly

~
^
y=y-Cx

+

C= 1 0

C

K= 56.93 5.54
L=

164
5740

FIGURE DP11.6
Block diagram for integrated controller and observer.

DP11.7

The compensator is
x̂˙ = [A − LC] x̂ + Ly + Bu
u = −Kx̂
where


−60


A − LC = 
 −1095


1
0



0 

1 
 ,

−3750 −5 −10



from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

644

CHAPTER 11

The Design of State Variable Feedback Systems

N = 4000 ,





 60 
h
i



K = 3998 595 30 , and L = 
 1095  .




3748

We selected the desired eigenvalues of A − BK at p1,2 = −10 ± 10j,
p3 = −20 and the desired eigenvalues of A − LC at q1,2 = −20 ± 10j,
q3 = −30. For initial conditions we let x(0) = [1 1 1] and x̂(0) = [0 0 0].
The transfer function from r to y is
T (s) =

4000s3 + 2.8e05s2 + 6.8e06s + 6e07
.
s6 + 110s5 + 5100s4 + 1.29e05s3 + 1.9e06s2 + 1.58e07s + 6e07

1.5
1
0.5
0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

20
10
0
−10
200
100
0
−100
−200

FIGURE DP11.7
The step response showing the actual and estimated states.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

645

Computer Problems

Computer Problems
CP11.1

The controllability and observablity matrices have nonzero determinants,
as shown in Figure CP11.1. Therefore, the system is observable and controllable.
>>

A=[-6 2 0;4 0 7;-10 1 11]; b=[5;0;1]; c=[ 1 2 1]; d=[0];
sys = ss(A,b,c,d);
Co=ctrb(sys); dt_Co=det(Co)
Ob=obsv(sys); dt_Ob=det(Ob)

dt_Co =
-84933
dt_Ob =
-3.6030e+03

FIGURE CP11.1
Determining controllability and observability.

CP11.2

The system is controllable since the determinant of the controllability
matrix is nonzero , as shown in Figure CP11.2.
a=[0 1;-6 -5]; b=[0;6]; c=[1 0]; d=[0];
sys_ss = ss(a,b,c,d);
Pc=ctrb(sys_ss);
dt_Pc=det(Pc)
Ob=obsv(sys_ss);
dt_Ob=det(Ob)
sys_tf=tf(sys_ss)

dt_Pc =
-36
dt_Ob =
1

Transfer function:
6
------------s^2 + 5 s + 6

FIGURE CP11.2
M-file script to determine controllability and to compute equivalent transfer function model.

CP11.3

The gain matrix (computed as shown in Figure CP11.3) is K =

a=[0 1;-1 -2]; b=[1;1]; c=[1 -1]; d=[0];
p=[-1;-2];
K=acker(a,b,p)
K=
0.5000

0.5000

FIGURE CP11.3
M-file script to place the closed-loop system poles using state feedback.

h

i

0.5 0.5 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

646

CHAPTER 11

CP11.4

The Design of State Variable Feedback Systems

The constant velocity guided missile is not controllable since the controllablity matrix, Co , has a zero determinant, as shown in Figure CP11.4.
Using the tf function (see Figure CP11.4), we determine that the transfer
function is
G(s) =

s5

5s
.
+ 0.5s4 + 0.1s3

Cancelling common terms in the transfer function yields
G(s) =

s4

5
.
+ 0.5s3 + 0.1s2

Then, using the ss function, we determine a state-space representation of
G(s). As shown in Figure CP11.4, the state-space representation is
ẋ = Ax + Bu
y = Cx

A=[0 1 0 0 0;-0.1 -0.5 0 0 0;0.5 0 0 0 0;0 0 10 0 0;0.5 1 0 0 0];
b=[0;1;0;0;0];
c=[0 0 0 1 0];
d=[0];
sys_ss = ss(A,b,c,d);
Transfer function:
% Part (a)
dt_Co =
5s
Co=ctrb(sys_ss); dt_Co=det(Co)
0
----------------------% Part (b)
s^5 + 0.5 s^4 + 0.1 s^3
sys_tf = tf(sys_ss)
sys_new = minreal(sys_tf );
sys_new_ss=ss(sys_new)
a=
% Part (c)
x1
x2
x3
x4
Co_new=ctrb(sys_new_ss); dt_Co_new=det(Co_new)
x1 -0.50000 -0.10000
0
0
% Part (d)
x2
1.00000
0
0
0
evalues=eig(sys_new_ss)
x3
0
1.00000
0
0
dt_Co_new =
x4
0
0
2.00000
0
32
b=
u1
evalues =
x1
2.00000
0
x2
0
0
x3
0
-0.2500 + 0.1936i
x4
0
-0.2500 - 0.1936i
c=
x1
x2
x3
x4
y1
0
0
0
1.25000
d=
u1
y1
0
Continuous-time system.

FIGURE CP11.4
Analysis of the constant velocity guided missile state-space model.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

647

Computer Problems

where




 −0.5 −0.1 0 0 


 1
0
0 0 



A=





0 



0

1

0

0

0

2 0





 2 


 0 



B=


 0 





and C =

h

0 0 0 1.25

i

0

The reduced system is controllable but not stable, since there are two poles
at the origin. Systems that are not controllable have too many states.
After eliminating unnecessary states, a controllable system of minimal
complexity (i.e. states) is obtained. In this case, the number of states is
reduced from five to four.
CP11.5

The eigenvalues of A are
e1 = −2.0727
e2 = −0.2354
e3,4 = 0.2761 ± 0.2593j
The system is unstable since there are two eigenvalues in the right halfplane, see Figure CP11.5. The characteristic polynomial is

A = [-0.0389 0.0271 0.0188 -0.4555; 0.0482 -1.0100 0.0019 -4.0208;
>>
0.1024 0.3681 -0.7070 1.4200; 0 0 1 0];
evalues =
b1 = [0.4422;3.5446;-6.0214;0];
0.2761 + 0.2593i
b2 = [0.1291;-7.5922;4.4900;0];
0.2761 - 0.2593i
% Part (a)
-0.2354
evalues = eig(A)
-2.0727
%part (b)
p=
p = poly(A)
1.0000 1.7559 -0.6431 0.0618 0.0700
r = roots(p)
% Part (c)
dt1 =
Co1 = ctrb(A,b1); dt1 = det(Co1)
-1.8451e+03
Co2 = ctrb(A,b2); dt2 = det(Co2)

r=
-2.0727
0.2761 + 0.2593i
0.2761 - 0.2593i
-0.2354

dt2 =
-90.6354

FIGURE CP11.5
Analysis of the VTOL aircraft model.

p(s) = s4 + 1.7559s3 − 0.6431s2 + 0.0618s + 0.0700 .
The roots of the characteristic equation are the same as the eigenvalues.
Also, the system is controllable from either u1 or u2 . If the aircraft should
lose the control of the vertical motion through u1 , then the control u2 can

.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

648

CHAPTER 11

The Design of State Variable Feedback Systems

be used to control both vertical and horizontal motion, and vice versa.
CP11.6

The m-file script to analyze the translunar halo orbit problem is shown
in Figure CP11.6. The translunar equilibrium point is not a stable point

A=[0 0 0 1 0 0;0 0 0 0 1 0; 0 0 0 0 0 1;7.3809 0 0 0
0 -2.1904 0 -2 0 0; 0 0 -3.1904 0 0 0];
c=[0 1 0 0 0 0];d=[0];
b1=[0;0;0;1;0;0];
b2=[0;0;0;0;1;0];
b3=[0;0;0;0;0;1];
sys_ss_1 = ss(A,b1,c,d);
sys_ss_2 = ss(A,b2,c,d);
dt1 =
sys_ss_3 = ss(A,b3,c,d);
0
% Part (a)
evalues=eig(A)
dt2 =
% Part (b)
0
Cb1=ctrb(sys_ss_1); dt1=det(Cb1)
dt3 =
% Part (c)
0
Cb2=ctrb(sys_ss_2); dt2=det(Cb2)
% Part (d)
Cb3=ctrb(sys_ss_3); dt3=det(Cb3)
% Part (e)
sys_tf = tf(sys_ss_2);
sys_tf=minreal(sys_tf )
% Part (f )
sys_ss=ss(sys_tf );
Co=ctrb(sys_ss); dt_Co=det(Co)
if dt_Co ~= 0
disp('System is completelly Controllable')
else
disp('System in uncontrollable')
end
% Part (g)
P = [-1+i; -1-i;-10;-10];
[A,B]=ssdata(sys_ss);
K = acker(A,B,P)
dt_Co =

2 0;

evalues =
2.1587
-2.1587
0 + 1.8626i
0 - 1.8626i
0 + 1.7862i
0 - 1.7862i
Transfer function:
s^2 - 7.381
---------------------------s^4 - 1.19 s^2 - 16.17

a=
x1
x2
x3
x4

x1
x2
x3
x4
0
0.59525
0
2.02089
2.00000
0
0
0
0
2.00000
0
0
0
0
2.00000
0

x1
x2
x3
x4

u1
1.00000
0
0
0

b=

c=
y1

x1
0

y1

u1
0

x2
0.50000

x3
0

x4
-0.92261

d=

64
System is completelly Controllable

FIGURE CP11.6
Analysis of the translunar satellite halo orbit.

as evidenced by the eigenvalues of A in the right half-plane; the system is
not completely controllable from any ui individually. The transfer function

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

649

Computer Problems

from u2 to η is
T (s) =

s6

s4 − 4.191s2 − 23.55
.
+ 2s4 − 19.97s2 − 51.58

A careful analysis reveals that T (s) can be reduced by eliminating common factors. The common factors are s2 + 3.1834. The reduced transfer
function is
T (s) =

s2 − 7.3815
.
s4 − 1.1837s2 − 16.2030

Using state feedback
u2 = −Kx
the gain matrix K which places the desired poles (using Ackermann’s
formula) is
K=
CP11.7

h

22 71.56 60 27.02

i

.

The m-file script to determine the initial state is shown in Figure CP11.7a.
Given three data points at t = 0, 2, 4, we construct the three equations
A=[0 1 0;0 0 1;-2 -4 -6];
b=[0;0;0]; c=[1 0 0]; d=[0];
sys=ss(A,b,c,d);
%
% Part (b)
v1=c*expm(0*A); v2=c*expm(2*A); v3=c*expm(4*A);
V=[v1;v2;v3]; Vi=inv( V );
n=[1;-0.0256;-0.2522];
x0=Vi*n
%
% Part (c)
t=[0:0.1:4]; u=0.0*t;
[y,x]=lsim(sys,u,t,x0');
plot(t,y,[0 2 4],[1;-0.0256;-0.2522],'*'), grid
xlabel('Time (sec)'), ylabel('y(t)')
title('Data points denoted by *')

FIGURE CP11.7
(a) Script to determine the initial state from three observations.

y(0) = 1 = Ce0A x0
y(2) = −0.0256 = Ce2A x0

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 11

The Design of State Variable Feedback Systems

y(4) = −0.2522 = Ce4A x0
or, in matrix form


Ce0A



 Ce2A



Ce4A







1






 x0 =  −0.0256  .






−0.2522

The problem is solvable if the matrix




0A
 Ce



 Ce2A 




Ce4A




is invertible. In this case, the inverse does exist and the solution is




x0 = 



1
−1
1.9998





 .



The simulation is shown in Figure CP11.7b.

Data points denoted by *
1*

0.8

0.6

0.4

y(t)

650

0.2

0

*

-0.2
*

-0.4

0

0.5

1

1.5

2

2.5

3

Time (sec)

FIGURE CP11.7
CONTINUED: (b) System simulation using computed initial state.

3.5

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

651

Computer Problems

CP11.8

Suppose we are given


A=

0



1



−1 0



B=



0



1

and the feedback
u = −Kx = −[K1 K2 ]x .
Solving HT P + PH = −I for P yields
K22 + K12 + 3K1 + 2
2(K1 + 1)K2
1
=
2(K1 + 1)
K1 + 2
=
2(K1 + 1)K2

p11 =
p12
p22

Then, with xo T = [1, 0] we find that
J = xo T Pxo = p11 .
Computing the partial of J with respect to K2 yields
1
∂J
1
K1 + 2
=
−
∂K2
2 K1 + 1
K22




.

Setting
∂J
=0
∂K2
and solving for K2 , we find that
K2 =

q

(K1 + 2)(K1 + 1) .

For a given value of K1 , the value of K2 that minimizes J can be computed
via the above equation. With K2 given as above, we can compute J to be
J=

s

K1 + 2
.
K1 + 1

A plot of J versus K1 (with K2 equal to the minimizing value) is shown
in Figure CP11.8. As K1 increases, the performance index J decreases.
However, we see that the rate of decrease slows considerably after K1 >
20. Also, K2 increases as K1 increases. We want to keep both gains as

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

652

CHAPTER 11

The Design of State Variable Feedback Systems

small as possible, while still having a small J. A reasonable selection is
K1 = 20

and K2 = 21.5 .

Performance index J versus K1

1.5
1.4

J

1.3
1.2
1.1
1
0

5

10

15

20

25
K1

30

35

40

45

50

30

35

40

45

50

K2 versus K1

60

K2

40
20
0
0

5

10

15

20

25
K1

FIGURE CP11.8
Performance index as a function of K1 and K2 .

CP11.9

In this problem, A = −1 and B = 1. Computing Q yields
Q = (1 + λ(−k)2 ) = 1 + λk 2 .
Define
H = A − Bk = −1 − k .
Solving
H T P + P H = −Q
yields
p=

1 + λk 2
.
2(k + 1)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

653

Computer Problems

0.5
0.49
0.48

J/x0^2

0.47
0.46
0.45
0.44
0.43
0.42
0.41

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

6

7

8

9

10

k
2.5

2

k min

1.5

1

0.5

0
0

1

2

3

4

5
lambda

FIGURE CP11.9
Plot of J/x20 versus k and the minimizing k versus λ.

The performance index is
J = x20 p which implies

J/x20 =

1 + λk 2
.
2(k + 1)

The plot of J/x20 versus k is shown in Figure CP11.9. The minimum value
is achieved when k = 0.41. To arrive at this result analytically, take the

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

654

CHAPTER 11

The Design of State Variable Feedback Systems

partial of J/x20 with respect to k, set the result to zero and solve for k:
∂J/x20
= 0 when k 2 + 2K − 1/λ = 0 .
∂k
p

Solving for k yields k = −1 ± 1 + 1/λ. So, when λ = 1, k = 0.41. The
plot of kmin versus λ is shown in Figure CP11.9.
CP11.10

The m-file is shown in Figure CP11.10.
A=[0 1;-18.7 -10.4]; B=[10.1; 24.6]; C=[1 0]; D=[0];
% Controller Gains
p=[-2;-2 ];
K=acker(A,B,p)

>>
K=
-0.3081 -0.1337

% Observer Gains
q=[-20+4*j;-20-4*j];
L = acker(A',C',q); L=L

L=
29.6000
89.4600

FIGURE CP11.10
Using the acker function to compute the controller gains and the observer gains.

CP11.11

The m-file is shown in Figure CP11.11(a). The compensator can be repA=[0 1 0;0 0 1;-4.3 -1.7 -6.7]; B=[0;0;0.35]; C=[0 1 0]; D=[0];
% Controller Gains
p=[-1.4+1.4*j;-1.4-1.4*j;-2];
K=acker(A,B,p)
% Observer Gains
q=[-18+5*j;-18-5*j;-20];
L = acker(A',C',q); L=L'

>>
K=
10.1143 22.3429 -5.4286

L=
1.0e+003 *

% Simulation of closed-loop system with the observer
Ac=[A -B*K;L*C A-B*K-L*C];
Bc=[zeros(6,1)];
Cc=eye(6);
Dc=zeros(6,1);
sys=ss(Ac,Bc,Cc,Dc);
x0=[1;0;0;0.5;0.1;0.1]; t=[0:0.001:3.5];
[y,t]=initial(sys,x0,t);
subplot(311)
plot(t,y(:,1),t,y(:,4),'--'), grid
subplot(312)
plot(t,y(:,2),t,y(:,5),'--'), grid
subplot(313)
plot(t,y(:,3),t,y(:,6),'--'), grid

FIGURE CP11.11
(a) M-file.

-1.6223
0.0493
0.7370

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

655

Computer Problems

resented as
x̂˙ = (A − BK − LC)x̂ + Ly

and

u = −Kx̂ .

Since y = Cx, we can write
x̂˙ = (A − BK − LC)x̂ + LCx .
Similarly, with
ẋ = Ax + Bu

and u = −Kx̂

we obtain
ẋ = Ax − BKx̂ .
In matrix form, we have



ẋ
x̂˙





=

−BK

A

LC A − BK − LC




x
x̂



 ,

with initial conditions
h

x(0)T

x̂(0)T

iT

=

h

1 0 0 0.5 0.1 0.1

iT

.

The response of the system is shown in Figure CP11.11(b).
5

Estimated state (dashed line)

x1 0
True state 5 0
(solid line) 1

1

2

3

4

0

1

2

3

4

0

1

3

4

x2 0
1
2

x3

0
2

2
Time (sec)

FIGURE CP11.11
CONTINUED: (b) Response of system to an initial condition.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

656

CHAPTER 11

CP11.12

The Design of State Variable Feedback Systems

The Simulink block diagram is shown in Figure CP11.12.

FIGURE CP11.12

CP11.13

The m-file to design the compensator is shown in Figure CP11.13(a). The
Simulink simulation is shown in Figure CP11.13(b). The output shown
on the x-y graph depicts the state x of the system. The initial conditions
selected for the simulation are




 1 


 0 



x(0) = 


 0 





0





 0.5 


 0.1 



and x̂(0) = 

 .
 0.1 





0.1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

657

Computer Problems

A=[0 1 0 0;0 0 1 0;0 0 0 1;-2 -5 -1 -13]; B=[0;0;0;1]; C=[1 0 0 0]; D=[0];
>>
K=

% Controller Gains
p=[-1.4+1.4*j;-1.4-1.4*j;-2+j;-2-j];
K=acker(A,B,p)

17.6000 24.6800 19.1200 -6.2000

% Observer Gains
q=[-18+5*j;-18-5*j;-20;-20];
L = acker(A',C',q); L=L'

L=

% Simulation of closed-loop system with the observer
Ac=[A -B*K;L*C A-B*K-L*C];
Bc=[zeros(8,1)];
Cc=eye(8);
Dc=zeros(8,1);
sys=ss(Ac,Bc,Cc,Dc);
x0=[1;0;0;0;0.5;0.1;0.1;0.1]; t=[0:0.001:10];
[y,t]=initial(sys,x0,t);
subplot(311)
100
plot(t,y(:,1),t,y(:,4),'--'), grid
subplot(312)
0
plot(t,y(:,2),t,y(:,5),'--'), grid
subplot(313)
?100
0
2
plot(t,y(:,3),t,y(:,6),'--'), grid
2

63
1369
10495
1479

4

6

8

10

0
?2
10

0

2

4

6

8

10

0

2

4

6

8

10

0
?10

FIGURE CP11.13
(a) M-file to design the compensator, including the observer.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

658

CHAPTER 11

The Design of State Variable Feedback Systems

.
^x=[A-BK-LC]x+Ly
^
^
u=-Kx

FIGURE CP11.13

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

1 2

Robust Control Systems

Exercises
E12.1

The plant transfer function is
G(s) =

3
.
s+3

Try a PI controller, given by
Gc = K1 +

K2
.
s

The ITAE characteristic equation is
s2 + 1.4ωn s + ωn2 ,
where ωn = 30. Then
K1 = 13

and

K2 = 300 .

Without a prefilter, the closed-loop system is
Y (s)
39s + 900
= 2
,
R(s)
s + 42s + 900
and with a prefilter, the closed-loop system is
Y (s)
900
= 2
,
R(s)
s + 42s + 900
where
Gp (s) =

23.07
.
s + 23.07

The step response, with and without the prefilter, is shown in Figure E12.1.

659

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

660

CHAPTER 12

Robust Control Systems

1.4
Without prefilter
With prefilter
1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

0.05

0.1

0.15

0.2

0.25
Time (sec)

0.3

0.35

0.4

0.45

0.5

FIGURE E12.1
Step response: (a) w/o prefilter (solid line), and (b) w/prefilter (dashed line).

The disturbance response is shown in Figure E12.2.

0.05

0.04

0.03

y(t)

E12.2

0.02

0.01

0

−0.01

0

0.05

0.1

0.15

0.2

FIGURE E12.2
Disturbance response for system in E12.1.

0.25
Time (sec)

0.3

0.35

0.4

0.45

0.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

661

Exercises

E12.3

The closed-loop transfer function is
T (s) =

s2

25
,
+ bs + 25

and the sensitivity function is
S(s) =

s2 + bs
,
s2 + bs + 25

where b = 8, nominally. The sensitivity of T to changes in b is determined
to be
SbT =

∂T b
−bs
= 2
.
∂b T
s + bs + 25

The plot of T (s) and S(s) is shown in Figure E12.3, where b = 8.

10

0

Gain dB

−10

20log|T|

−20
20log|S|
−30

−40

−50

−60
−1
10

0

1

10

10

FIGURE E12.3
Plot of T (s) and the sensitivity function S(s).

E12.4

The plant transfer function is
G(s) =

1
,
(s + 20)(s + 36)

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

662

CHAPTER 12

Robust Control Systems

and the PID controller is given by
K3 (s + a)(s + b)
.
s

Gc (s) =

Let a=20, b=500, and K3 = 200. Then, the closed-loop system is
T (s) =

200s2 + 4000s + 100000
.
s3 + 256s2 + 4720s + 100000

The closed-loop poles are s1 =-237.93 and s2,3 = −9.04 ± j18.5 and the
zeros are s1,2 = −10±j20. Therefore, there is an approximate cancellation
of the complex poles and zeros and the approximate system is
T̂ (s) =

238
.
s + 238

The actual response and approximation are shown in Figure E12.4.

1.4

1.2
actual
1

approximation

y(t)

0.8

0.6

0.4

0.2

0

0

0.05

0.1

0.15

0.2

0.25
0.3
Time (sec)

0.35

0.4

0.45

FIGURE E12.4
Step response for closed-loop actual and approximate transfer functions.

E12.5

The loop transfer function is
L(s) = Gc (s)G(s) =

10KD (s + KP /KD )
.
s(s + 3)(s + 10)

0.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

663

Exercises

Select KP /KD = 10. Then
L(s) = Gc (s)G(s) =

10KD
,
s(s + 3)

and the closed-loop transfer function is
T (s) =

10KD
.
s2 + 3s + 10KD

Let ζ = 0.69, which implies P.O. < 5%. Also, 2ζωn = 3, so ωn = 2.17.
Thus,
10KD = ωn2 = 4.72 .
Thus, the controller is Gc (s) = 0.47(s + 10). The settling time is Ts = 2.8
s and the percent overshoot is P.O. = 4.6%. As K increases, the percent overshoot increases from 0% to 16% and the settling time generally
decreases from 3.8 sec to 2.6 sec.
E12.6

The loop transfer function with the PID controller is
Gc (s)Gs(s) =

KD s2 + KP s + KI
1
.
s
(s + 5)2

The ITAE step response requires
s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = s3 + (10 + KD )s2 + (25 + KP )s + KI .
For n = 3 we estimate the normalized settling time to be
ωn Ts ≈ 8 seconds.
Thus, ωn ≈ 6, and
KD = 0.5,

KP = 52.4,

and

KI = 216.

The step response is shown in Figure E12.6. The transfer function from
the disturbance to the output is
Y (s)
G(s)
s
=
= 3
.
2
Td (s)
1 + Gc (s)G(s)
s + 10.5s + 77.4s + 216
The disturbance response is shown in Figure E12.6. The system is effective
in reducing the effects of the disturbance, and the maximum output is
reduced by 1/100 for a step disturbance.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

664

CHAPTER 12

Robust Control Systems

−3

(a) Step response

1.2

10

1

8

0.8

6

x 10

(b) Disturbance response

y(t)

12

y(t)

1.4

0.6

4

0.4

2

0.2

0

0

0

0.5

1
TIme (s)

1.5

2

−2

0

0.5

1
TIme (s)

1.5

2

FIGURE E12.6
(a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.

E12.7

The plant transfer function is
G(s) =

1
,
(s + 4)2

and the PID controller is
Gc (s) =

K1 s + K2 + K3 s2
.
s

Using the ITAE criteria and selecting ωn = 10 yields
K3 = 9.5

K2 = 1000

and K1 = 199 .

The step response is shown in Figure E12.7. The disturbance response is
also shown in Figure E12.7. The maximum y(t) = 0.0041, so the system
is effective in rejecting the step disturbance.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

665

Exercises
x10 -3

(a) step response
1.4

(b) disturbance

4.5
4

1.2
3.5
1

3
2.5

y(t)

y(t)

0.8

0.6

2
1.5
1

0.4

0.5
0.2
0
0
0

0.5

1

1.5

-0.5

2

0

0.5

Time (sec)

1

1.5

2

Time (sec)

FIGURE E12.7
(a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.

The maximum ωn = 60. Then K1 = 3600 and K2 = 80. The maximum
control input is max |u(t)| ≈ 80. The plot of the step response and the
control input u(t) is shown in Figure E12.8.
(a) step response

(b) control input u(t)

1.2

90
80

1

70
60

0.8

u(t)

50

y(t)

E12.8

0.6

40
30

0.4

20
10

0.2

0
0
0

0.05

0.1

0.15

0.2

-10
0

Time (sec)

FIGURE E12.8
Step response w/o prefilter; and (b) control input u(t).

0.05

0.1
Time (sec)

0.15

0.2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

666

CHAPTER 12

E12.9

Robust Control Systems

One possible PD controller is
Gc (s) = 27.6s + 8.25s .
When
K=1,
the system roots are
s1,2 = −3.2 ± j4.3
s3 = −9.5 .
The step response is shown in Figure E12.9 for K = 0.5, 1, and 1.5.
K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted)
1.4

1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

0.2

0.4

0.6

0.8

1
1.2
Time (sec)

1.4

1.6

1.8

2

FIGURE E12.9
Step response for K = 0.5, 1, and 1.5.

E12.10

One possible PI controller is
Gc (s) =

2.2s + 22
.
s

When K = 1, the system roots are s1,2 = −1.31 ± j1.31, and s3 = −6.37.
The step response is shown in Figure E12.10.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

667

Exercises

K=1 (solid); K=0.5 (dashed); and K=1.5 (dotted)
1.4

1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2

2.5
3
Time (sec)

3.5

4

4.5

5

FIGURE E12.10
Step response for K = 0.5, 1, and 1.5.

The plot is shown in Figure E12.11.

100
90
80
70
60
P.O. (%)

E12.11

50
40
30
20
10
0

0

0.5

1

1.5

2
Time (sec)

2.5

3

FIGURE E12.11
Percent overshoot as a function of k in the interval 0.1 ≤ k ≤ 4.

3.5

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

668

CHAPTER 12

The controllability matrix is


Pc = 

and

c1

c2

c2 −ac1 − bc2




det Pc = c22 + [bc1 ]c2 + ac21 .
For controllability we require det Pc 6= 0, hence
c22 + [bc1 ]c2 + ac21 6= 0
implies
c2
b q
6= − ± (b/2)2 − a
c1
2
where (b/2)2 − a ≥ 0. For real-valued c1 and c2 , if (b/2)2 − a < 0, all real
values of c1 and c2 are valid. Valid values of the constants are c1 = 0,
c2 = 10, a = 10, and b = 3. The step response is shown in Figure E12.12.

Step Response
1.4

1.2

1

Amplitude

E12.12

Robust Control Systems

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2
Time (sec)

2.5

FIGURE E12.12
Step response with c1 = 0, c2 = 10, a = 10, and b = 3.

3

3.5

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

669

Problems

Problems
The closed-loop transfer function is
T (s) =

4(s + 2)
s2 + 4s + 8

and the sensitivity function is
S(s) =

s2
.
s2 + 4s + 8

The plot of 20 log |T | and 20 log |S| is shown in Figure P12.1. The bandwidth is
Then
T
|SK
|ωB = 0.98
T ω
|SK | B = 0.78
2

T ω
|SK
| B
4

= 0.30 .

10

0

-10
20log|T|
20log|S|
-20
Gain dB

P12.1

-30

-40

-50

-60 -1
10

0

10

FIGURE P12.1
Plot of T (s) and the sensitivity function S(s).

10

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

670

CHAPTER 12

(a) The loop transfer function is given by
Gc (s)G(s) =

K
.
s(0.02s + 1)(0.002s + 1)

When
K = 100 ,
the peak magnitude is
Mpω = 1.84 .
(b) The plot of 20 log |T | and 20 log |S| is shown in Figure P12.2a.
20
20log|S|
0

-20

Gain dB

P12.2

Robust Control Systems

20log|T|

-40

-60

-80

-100

-120
101

102

103

FIGURE P12.2
(a) Plot of T (s) and the sensitivity function S(s).

(c) The bandwidth is
and
T
|SK
|ωB = 1.47
T ω
|SK | B = 0.39
4

T ω
|SK
| B = 1.62 .
2

104

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

671

Problems

(c) The disturbance response is shown in Figure P12.2b.
x10 -7
8
7
6

Amplitude

5
4
3
2
1
0
0

0.05

0.1

0.15

0.2

0.25

0.3

Time (secs)

FIGURE P12.2
CONTINUED: (b) Disturbance response for K = 100.

P12.3

(a) The loop transfer function is
L(s) = Gc (s)G(s) =

K(s − 4)(s − 1)
.
(s + 0.02)(s + 2)2

The characteristic equation is
1 + Gc (s)G(s) = 1 + K

(s − 4)(s − 1)
=0
(s + 0.02)(s + 2)2

or
s3 + (4.02 + K)s2 + (4.08 − 5K)s + 0.08 + 4K = 0 .
Using Routh-Hurwitz we find that the system is stable for
−4.6987 < K < 0.6947 .
ess =

1
.
1 + 50K

Select K = 0.18 to obtain a steady-state error to a unit step of 0.1.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

672

CHAPTER 12

Robust Control Systems

(c,d) The plots of y(t) for
K = 0.18
K = 0.21
K = 0.15

(nominal)
(+15%)
(−15%)

are shown in Figure P12.3.

K=0.18 (solid) & K=0.21 (dashed) & K=0.15 (dotted)
1.2

1

0.8

y(t)

0.6

0.4

0.2

0

−0.2

0

5

10

15

20

25

Time (sec)

FIGURE P12.3
Step input response for K = 0.18, K = 0.21 and K = 0.15.

P12.4

(a) The plant is given by
G=

s

1
 .
+1

s
25

We desire P.O. < 10% and Ts < 100 ms. Using a PD controller
Gc (s) = 100 + 2.2s ,
we determine that P.O. = 7%, Ts < 100 ms and ess =
input. The plot of y(t) is shown in Figure P12.4.

A
100

for a ramp

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

673

Problems

(b) The sensitivity is
r
|SK
| = 27.95
1

when K1 = 1.
(c) The plot of y(t) when K1 = 2 (the compensator Gc (s) is unchanged)
is shown in Figure P12.4.
(d) The disturbance response is shown in Figure P12.4.
(b) disturbance
0.012

1

0.01

0.8

0.008

y(t)

y(t)

(a) step response
1.2

0.6

0.006

0.4

0.004

0.2

0.002

0
0

0.05

0.1

0.15

0
0

0.2

0.05

Time (sec)

0.1

0.15

0.2

Time (sec)

FIGURE P12.4
(a) Step response: K1 = 1 (solid line) and K1 = 2 (dashed line); and (b) disturbance
response.

P12.5

(a) The plant is given by
G(s) =

1
s(s + p)

where p = 2, nominally. One solution is
Gc (s) =

18.7(s + 2.9)
.
(s + 5.4)

Then,
T (s) =

18.7(s + 2.9)
√
√
.
(s + 3.41)(s + 2 + 2 3j)(s + 2 − 2 3j)

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

674

CHAPTER 12

Robust Control Systems

(b,d) The step responses are shown in Figure P12.5 for p = 2 and p = 1.
(c,d) The disturbance responses are shown in Figure P12.5 for p = 2 and
p = 1.
(a) step response

(b) disturbance

1.6

0

1.4

-0.02

1.2

-0.04

1

y(t)

y(t)

-0.06
0.8

-0.08
0.6
-0.1

0.4

-0.12

0.2
0
0

-0.14

5

0

Time (sec)

5
Time (sec)

FIGURE P12.5
(a) Step response: p = 2 (solid line) and p = 1 (dashed line); and (b) disturbance response:
p = 2 (solid line) and p = 1 (dashed line).

P12.6

(a) The plant is given by
G(s) =

s(s2

1
,
+ 4s + 5)

and the PID controller is
Gc (s) =

K(s + z)2
.
s

When
z = 1.25
and
K=4,
all roots are
s = −1 ± j1.22 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

675

Problems

Then, the closed-loop transfer function is
T (s) =

4(s + 1.25)2
.
s4 + 4s3 + 9s2 + 10s + 6.25

(b,c) The step responses with and without a prefilter are shown in Figure P12.6.
(d) The disturbance response is shown in Figure P12.6.
(a) step response

(b) disturbance

1.6

0.02

1.4
0
1.2
-0.02

y(t)

y(t)

1
0.8

-0.04

0.6
-0.06
0.4
-0.08
0.2
0
0

5

10

-0.1

0

Time (sec)

5

10

Time (sec)

FIGURE P12.6
(a) Step response: w/o prefilter (solid line) and w/prefilter (dashed line); and (b) disturbance response.

P12.7

(a) The loop transfer function is
Gc (s)G(s) =

10Ka (5s + 500 + 0.0475s2 )
.
s3

When
Ka = 374.5 ,
the phase margin is
P.M. = 40o .
(b) The root locus is shown in Figure P12.7a.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

676

CHAPTER 12

Robust Control Systems

150
*

100

o

Imag Axis

50
0

x

*

-50
o

-100
*

-150
-150

-100

-50

0

50

100

150

Real Axis

FIGURE P12.7
10(0.0475s2 +5s+500)
= 0.
(a) Root locus for 1 + Ka
s3

When
Ka = 374.5 ,
the roots are
s1 = −139.8
s2,3 = −19.1 ± j114.2 .
(c) The transfer function from Td (s) to Y (s) is
Y (s)
−s
= 3
.
2
Td (s)
s + 182s + 19150s + 1915000
The maximum is
max |y(t)| = 0.0000389 .
(d) The step responses, with and without a prefilter, are shown in Figure P12.7b.
P12.8

The polynomial under investigation is
s3 + 3s2 + 3s + 4 = 0 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

677

Problems

1.6
1.4
1.2

y(t)

1
0.8
0.6
0.4
0.2
0
0

0.05

0.1

0.15

0.2

0.25

0.3

Time (sec)

FIGURE P12.7
CONTINUED: (b) Step response: w/o prefilter (solid line) and w/prefilter (dashed line).

From the uncertainty bounds on the coefficients, we define
α0 = 4
α1 = 1
α2 = 2

β0 = 5
β1 = 4
β2 = 4

Then, we must examine the four polynomials:
s3 + 2s2 + 4s + 5 = 0
s3 + 4s2 + s + 4 = 0
s3 + 4s2 + 4s + 4 = 0
s3 + 2s2 + s + 5 = 0
The fourth polynomial is not stable—therefore, the system is not stable
for the uncertain parameters.
P12.9

One possible PID controller is
Gc (s) =

0.058s2 + 2.17s + 16.95
.
s

A first-order Pade approximation was used in the design to account for
the delay system. The step input response is shown in Figure P12.9. A
prefilter should also be used with the PID controller. A suitable prefilter

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

678

CHAPTER 12

Robust Control Systems

is
Gp (s) =

K2
.
K3 s2 + K1 s + K2

1.2

1

y(t)

0.8

0.6

0.4

0.2

0
0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Time (sec)

FIGURE P12.9
Step response with the PID controller and prefilter.

P12.10

The PID controller is given by
Gc (s) =

KD s2 + KP s + KI
.
s

Using the ITAE method, we desire the characteristic polynomial to be
q(s) = s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3 = 0 ,
where we select ωn = 4 to obtain a peak time of Tp = 1 second. Here we
use the approximation for ITAE third-order systems that ωn Tp ≈ 4 from
Figure 5.30(c) in Dorf and Bishop. The actual characteristic equation is
s3 + 25KD s2 + 25KP s + 25KI = 0 .
Equating coefficients and solving for the gains yields
KP = 1.376 ,

KD = 0.28 ,

and

KI = 2.56 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

679

Problems

The step response is shown in Figure P12.10, with the prefilter
Gp (s) =

KD

s2

KI
.
+ KP s + KI

Step Response
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5
Time (sec)

2

2.5

3

FIGURE P12.10
Step response with the PID controller and prefilter.

P12.11

We will design for the case where K = 1 and p = 1. The design plant is
G(s) =

1
.
s(s + 1)(s + 4)

The nominal plant is given by
G(s) =

2.5
,
s(s + 2)(s + 4)

and the PID controller is
Gc (s) =

KD s2 + KP s + KI
.
s

Using the ITAE method, we desire the characteristic polynomial to be
q(s) = s4 + 2.1ωn s3 + 3.4ωn2 s2 + 2.7ωn3 s + ωn4 = 0 ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

680

CHAPTER 12

Robust Control Systems

where we select ωn = 2.38 to obtain a peak time around Tp = 3 seconds.
The actual characteristic equation (with the worst-case plant) is
s4 + 5s3 + (4 + KD )s2 + KP s + KI = 0 .
Equating coefficients and solving for the gains yields KP = 36.40, KI =
32.08, and KD = 15.26. The step response is shown in Figure P12.11,
with the prefilter
Gp (s) =

KD

s2

KI
.
+ KP s + KI

1.4
Worst−case plant
Nominal plant
1.2

1

y(t)

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2

2.5
Time (sec)

3

3.5

4

4.5

5

FIGURE P12.11
Step response with the prefilter: nominal plant (dashed line) & worst-case plant (solid line).

P12.12

The transfer function is
G(s) = C(sI − A)−1 B =

h

The sensitivity is
G
SK
=

2 0

i




s

−3

5 s+K




∂G K
−Ks
.
= 2
∂K G
s + Ks + 5

0
1



=

s2

−6
.
+ Ks + 5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

681

AP12.1

Let Gp (s) = 1. A viable PID controller is
Gc (s) = KP +

1000s2 + 3000s + 100
KI
+ KD s =
.
s
s

The loop transfer function is
Gc (s)G(s) =

1000s2 + 3000s + 100)
.
s(50s2 + 1)

We can check that Kv = 100, as desired. The step response is shown in
Figure AP12.1.
Step Response
1.4
System: syscl
Peak amplitude: 1.1
Overshoot (%): 9.5
At time (sec): 0.234

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

0.2

0.4

0.6
Time (sec)

0.8

1

1.2

FIGURE AP12.1
Step response with PID controller.

AP12.2

For all three controllers, choose K = 1 as the design value. Also, use
as the nominal points a = 2 and b = 5 for each design. ITAE methods
were employed in all designs, although this did not work well for the PI
controller.
(a) PI controller: Let
Gp (s) = 1 .
Not all specifications could be met simultaneously with a PI con-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

682

CHAPTER 12

Robust Control Systems

troller. The best over-all results are achieved when using a = 3 and
b = 4.5 as the design values. An acceptable PI controller is
Gc (s) = 1.2 +

3.96
.
s

Controller

P.O.

Ts

Tp

|u(t)|max

PI

0%

2.29s

n.a.

4.43

PD

4.6%

1.72s

1.26s

12.25

PID

1.97%

0.65s

0.47s

37.25

TABLE AP12.2

PI, PD, and PID controller performance summary.

The final design is based on root locus methods since the ITAE methods did not produce an effective controller. The closed-loop transfer
function is
T (s) =

1.2s + 3.96
.
s3 + 3s2 + 5.7s + 3.96

(b) PD controller: Let
Gp (s) =

12.25
.
7.25 + 2.9s

The closed-loop transfer function is
T (s) =

7.25 + 2.9s
,
s2 + 4.9s + 12.25

where the PD controller (based on ITAE methods) is
Gc (s) = 7.25 + 2.9s .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

683

(c) PID controller: Let
Gp (s) =

15.5s2

1000
.
+ 210s + 1000

The closed-loop transfer function is
T (s) =

15.5s2 + 210s + 1000
.
s3 + 17.5s2 + 215s + 1000

And the PID controller (based on ITAE methods) is
Gc (s) =

15.5s2 + 210s + 1000
.
s

The performance of each controller is summarized in Table AP12.2.
AP12.3

(a) The PID controller is
Gc (s) =



KD s2 +

KP
KD s

+

KI
KD

s



.

Since we want P.O. < 4% and Ts < 1s, we choose the dominant
closed-loop poles to have ωn = 6 and ζ = 0.8. Therefore, we place the
zeros at
s2 +

KP
KI
s+
= s2 + 10s + 36 .
KD
KD

Solving for the constants yields,
KP
= 10 ,
KI

KI
= 36 .
KD

Then, using root locus methods, we choose KD = 91 to place the
roots near the zeros. The PID controller gains are computed to be
KP = 910, KI = 3276 and KD = 91.
(b) The loop transfer function is
Gc (s)G(s) =

KD s2 + KP s + KI
.
s2 (s2 + 5s + 4)

The closed-loop system characteristic equation is
s3 + 5s2 + 4s + KD s2 + KP s + KI = 0 .
Solving for the PID gains yields KP = 73.4, KI = 216 and KD = 5.5.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

684

CHAPTER 12

Robust Control Systems

Therefore, the controller is
Gc (s) =

5.5(s2 + 13.35s + 39.3)
.
s

Using the prefilter
Gp (s) =

s2

39.3
,
+ 13.35s + 39.3

we obtain the closed-loop transfer function
T (s) =

s3

+

216
.
+ 77.4s + 216

10.5s2

The percent overshoot is P.O. ≈ 3.5% and the settling time is Ts ≈
1.67 sec.
The PID controller is
Gc (s) =



KD s2 +

KP
KI
KD s + KD

s



.

The bounds 1 ≤ a ≤ 2 and 4 ≤ b ≤ 12 imply that 2 ≤ ωn ≤ 3.46
and 0.5 ≤ ζωn ≤ 1. One solution is to place the PID controller zeros at
1. 4
1. 2
1
Amplitude

AP12.4

0. 8
0. 6
0. 4
0. 2
0

0

0. 5

1

1. 5
Time (sec)

2

2. 5

3

FIGURE AP12.4
Family of step response with PID controller with nominal case (a, b) = (1.5, 9) denoted by
the solid line.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

685

√
s = −1 ± j 8 (i.e. ζωn = 1 and ωn = 3). So,
s2 +

KI
KP
s+
= s2 + 2ζωn s + ωn2 = s2 + 4s + 9 .
KD
KD

The nominal case for design is chosen to be a = 1.5 and b = 9. Using
root locus, we select KD = 2.1 to place the closed-loop characteristic
roots near the zeros. Then, the PID controller gains are computed to be
KP = 8.4, KI = 18.9, and KD = 2.1. The plot of the response to a step
input is shown in Figure AP12.4. The off-nominal cases shown in the
simulations are (a, b) = (1.2, 4), (1.4, 6), (1.6, 10), and (1.8, 12).
AP12.5

To obtain a phase margin of P.M. = 49.77o , select K = 1.5, b = 36 and
choose Gp (s) = 1. The PID controller is
Gc (s) =

1.5(s2 + 20s + 36)
.
s

When K1 = 0.75, the phase margin is reduced to P.M. = 45.45o ; and
when K1 = 1.25, the phase margin is increased to P.M. = 52.75o .
AP12.6

With the settling time Ts = 1 and percent overshoot P.O. < 10% specifications, we target for dominant closed-loop poles with ωn = 10. Here we
estimate ωn Ts ≈ 10 associated with the ITAE performance. The closedloop transfer function is
T (s) = Gp (s)

1.5(KD s2 + KP s + KI )
,
(1 + 1.5KD )s2 + 1.5KP s + 1.5KI

where we have neglected τ . Using the ITAE method, the desired characteristic polynomial is
s2 +

√

2ωn s + ωn2 = s2 +

1.5Kp
1.5KI
s+
.
1 + 1.5KD
1 + 1.5KD

Let KD = 0.25. Then solving for the remaining PID gains yields KP =
12.96 and KI = 91.67. The pre-filter is
Gp (s) =

0.375s2

137.5
.
+ 19.45s + 137.5

Then the closed-loop transfer function (with τ = 0.001) is
T (s) =

0.001s3

137.5
.
+ 1.375s2 + 19.45s + 137.5

The transfer function from the disturbance to the output is
Y (s)/Td (s) =

0.001s3

1.5s
.
+ 1.375s2 + 19.45s + 137.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

686

CHAPTER 12

Robust Control Systems

The step input response and disturbance response are shown in Figure AP12.6.
(a)

(b)
0.05

1.4

1.2

0.04

1

Amplitude

Amplitude

0.03
0.8

0.02

0.6
0.01
0.4

0

0.2

0

0

0.2

0.4
0.6
Time (sec)

0.8

1

−0.01

0

0.2

0.4
0.6
Time (sec)

0.8

1

FIGURE AP12.6
(a) Input response; (b) Disturbance response.

AP12.7

The PI controller is given by
Gc (s) =

KP s + KI
.
s

We will also use the prefilter
Gp (s) =

KI
.
KP s + KI

Using the ITAE method, we determine that
√
KP = 2ωn and KI = ωn2 .
Let ωn = 2.2. Then KP = 3.11 and KI = 4.8. The step response and
control u(t) are shown in Figure AP12.7.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

687

(a)

(b)

1.2

1.2

1

1

0.8

Amplitude

Amplitude

0.8

0.6

0.6

0.4

0.4
0.2
0.2

0

0
0

5

-0.2

0

5

Time (secs)

Time (secs)

FIGURE AP12.7
(a) Input response; (b) Control history u(t).

AP12.8

(a) A suitable PD controller is given by
Gc (s) = 0.6 + 0.4s .
The percent overshoot is P.O. = 18.8% and the peak time is Tp = 2.4
sec.
(b) A suitable PI controller is given by
Gc (s) = 0.15 +

0.01
.
s

The percent overshoot is P.O. = 23.7% and the peak time is Tp = 7.8
sec.
(c) A suitable PID controller is given by
Gc (s) = 0.6 +

0.01
+ 0.4s .
s

The percent overshoot is P.O. = 19.9% and the peak time is Tp = 2.5
sec.
(d) The PD or PID controllers are the best choices.
AP12.9

A robust PID controller designed with ITAE methods will be a suitable
controller. From the settling time specification we select ωn = 10, where

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

688

CHAPTER 12

Robust Control Systems

we have used ζ = 0.8. The worst case is
a = 1 and

K=2.

The desired closed-loop transfer function is
T (s) =

ωn3
s3 + 1.75ωn s2 + 2.15ωn2 s + ωn3

and the actual characteristic equation is
q(s) = s3 + (2a + KKD )s2 + (a2 + KKP )s + KKI .
Equating like terms, we find that
KP = 107

KD = 7.75 .

We use as the design plant
G(s) =

s+2
.
s(s + 3)

1.4

1.2

1

Amplitude

AP12.10

KI = 500

0.8

0.6

0.4

0.2

0
0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Time(sec)

FIGURE AP12.10
Family of step responses with the design plant (p, q, r) = (3, 0, 2) denoted by the solid line.

Select
p1 = 2

and z1 = 3

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

689

to cancel a design plant pole and zero. Then, choose p2 = 0 to have
zero steady-state error to a unit step. The remaining variables K and
z2 are selected based on ITAE methods, where ωn = 100. A suitable
compensator is
Gc (s) =

141.42(s + 3)(s + 70.71)
.
s(s + 2)

A plot of the step responses for various values of p, q and r is shown in
Figure AP12.10.
A suitable compensator is
Gc (s) =

1000(s + 1.8)(s + 3.5)(s + 5.5)
.
s(s + 600)

1.4

1.2

1

Amplitude

AP12.11

0.8

0.6

0.4

0.2

0
0

1

2

3

4

5

6

7

8

9

10

Time (sec)

FIGURE AP12.11
Step responses with nominal plant (solid line) and off-nominal plant with all poles reduced
by 50% (dashed line).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

690

CHAPTER 12

Robust Control Systems

Design Problems
The plant model with parameters given in Table CDP2.1 in Dorf and
Bishop is given by:
θ(s)
26.035
=
,
Va (s)
s(s + 33.142)
where we neglect the motor inductance Lm and where we switch off the
tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). With a
PID controller ,the closed-loop system characteristic equation is
s3 + (33.142 + 26.035KD )s2 + 26.035KP s + 26.035KI = 0 .
A suitable PID controller is
Gc (s) = 50 + s +

0.1
.
s

This PID controller places the closed-loop system poles to the left of
the −ζωn line necessary to meet the settling time requirement. The step
response is shown below. The settling time is Ts = 0.12 second. In the
steady-state the error due to a step disturbance is zero.

1.2

1

0.8
Amplitude

CDP12.1

0.6

0.4

0.2

0

0

0.05

0.1

0.15

0.2

0.25
Time (secs)

0.3

0.35

0.4

0.45

0.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

691

Design Problems

The closed-loop transfer function is
Y (s)
Km Gc (s)
.
= 2
R(s)
s + (2 + Km K1 )s + Gc (s)Km
(a) When Gc = K, we have
T (s) =

s2

15K
,
+ (2 + 15K1 )s + 15K

where Km = 15. Using ITAE criteria and ωn = 10, we determine that
K1 = 0.81 and K = 6.67. For the disturbance, we have
−1
Y (s)
= 2
.
TL (s)
s + 14.14s + 100
The input and disturbance responses are shown in Figure DP12.1,
without prefilters.
(a) Step response

(b) Disturbance response

1.2

0

1

-0.002

0.8

-0.004

y(t)

y(t)

DP12.1

0.6

-0.006

0.4

-0.008

0.2

-0.01

0
0

0.5

1

Time (sec)

-0.012

0

0.5

1

Time (sec)

FIGURE DP12.1
(a) Step response: Gc (s) = K (solid line) and Gc (s) = KP + KD s (dashed line); and (b)
disturbance response (same for both compensators).

(b) When Gc = KP + KD s, we have
Y (s)
15(KP + KD s)
= 2
.
R(s)
s + (2 + 15K1 + 15KD )s + 15KP
For ωn = 10 and with the ITAE criteria, we determine that (with

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

692

CHAPTER 12

Robust Control Systems

KD = 0.1)
Y (s)
15(6.67 + 0.1s)
= 2
.
R(s)
s + 14.14s + 100
The nominal plant is given by
1
.
s(s + 5)

G(s) =
The closed-loop transfer function is
T (s) =

K(KD s2 + KP s + KI )
.
s3 + (5 + KKD )s2 + KKP s + KKI

Let
KP = 450 ,

KI = 750 ,

and KD = 150 .

A family of responses is shown in Figure DP12.2 a for various values of
K. The percent overshoot for 0.1 ≤ K ≤ 2 is shown in Figure DP12.2b.

1.4

1.2

1
Step response

DP12.2

0.8

0.6

0.4

0.2

0

0

0.5

1

1.5

2
Time (s)

FIGURE DP12.2
(a) Family of step responses for various values of K.

2.5

3

3.5

4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

693

Design Problems

9
8

Percent overshoot

7
6
5
4
3
2
1
0

0

0.2

0.4

0.6

0.8

1
K

1.2

1.4

1.6

1.8

FIGURE DP12.2
CONTINUED: (b) Percent overshoot for various values of K.

DP12.3

(a) The dexterous hand model is given by
G(s) =

Km
,
s(s + 5)(s + 10)

where Km = 1, nominally. The PID controller is
Gc (s) =

KD (s2 + 6s + 18)
.
s

The root locus is shown in Figure DP12.3a. If we select
KD = 90 ,
the roots are
s1,2 = −5.47 ± j6.6
s3,4 = −2.03 ± j4.23 .
Thus, all roots have
ζωn > 4/3

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 12

Robust Control Systems

to meet the design specification
Ts < 3 sec .
(b) The step responses for
Km = 1
and
Km = 1/2
are shown in Figure DP12.3b. When
K = 1/2 ,
an off-nominal value, the settling time specification is no longer satisfied.
20
15
10
*

5

Imag Axis

694

o

0

x

*

x

x
o
*

-5
*

-10
-15
-20
-20

-15

-10

-5

0
Real Axis

FIGURE DP12.3
s2 +6s+18
(a) Root locus for 1 + KD s2 (s+5)(s+10)
= 0.

5

10

15

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

695

Design Problems

1.6
1.4
1.2

y(t)

1
0.8
0.6
0.4
0.2
0
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Time (sec)

FIGURE DP12.3
CONTINUED: (b) Step response (without prefilters): PID with K3 = 90 and Km = 1
(solid line) and PID with K3 = 90 and Km = 0.5 (dashed line).

DP12.4

The nominal plant is
G(s) =

s(s2

17640
,
+ 59.4s + 1764)

and the PID controller is
Gc (s) =

KI (τ1 s + 1)(τ2 s + 1)
.
s

(a) Using ITAE methods, we determine that ωn = 28.29, KI = 36.28,
τ1 + τ2 = 0.0954 and τ1 τ2 = 0.00149. So,
Gc (s) =

36.28(0.00149s2 + 0.0954s + 1)
.
s

(b) The step response for the nominal plant and the PID controller is
shown in Figure DP12.4a, with and without a prefilter.
(c) The disturbance response is shown in Figure DP12.4b.
(d) The off-nominal plant is
G(s) =

s(s2

16000
.
+ 40s + 1600)

The step response for the off-nominal plant is shown in Figure DP12.4a.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 12

Robust Control Systems

(b) off-nominal plant
1.6

1.4

1.4

1.2

1.2

1

1

y(t)

y(t)

(a) nominal plant
1.6

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0
0

0.5

0
0

1

0.5

Time (sec)

1

Time (sec)

FIGURE DP12.4
(a) Step response for (i) nominal plant: w/o prefilter (solid line) and w/prefilter (dashed
line); and (ii) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line).

disturbance response
0.3

0.25

0.2

0.15

y(t)

696

0.1

0.05

0

-0.05

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Time (sec)

FIGURE DP12.4
CONTINUED: (b) Disturbance response for the nominal plant.

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

697

Design Problems

DP12.5

One possible solution is
Gc (s) = 0.08

(0.01s + 1)(0.99s + 1)
.
s

The phase margin with this controller is
P.M. = 45.5o .
The step response is shown in Figure DP12.5 for the nominal plant (with
and without a prefilter); the step response for the off-nominal plant is also
shown in Figure DP12.5. The prefilter is
Gp (s) =

13.97s2

1411
.
+ 1411s + 1411

(a) nominal plant

(b) off nominal plant

1. 2

1. 2

1

1

0. 8

0. 8
y(t)

1. 4

y(t)

1. 4

0. 6

0. 6

0. 4

0. 4

0. 2

0. 2

0

0

10
Time (sec)

20

0

0

10
Time (sec)

20

FIGURE DP12.5
(a) Step response for nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line);
and (b) for off-nominal plant: w/o prefilter (solid line) and w/prefilter (dashed line).

DP12.6

Using ITAE methods, three controllers are designed for the nominal plant:
(i) PID controller:
Gc (s) =

0.225s2 + 0.535s + 34.3
s

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

698

CHAPTER 12

Robust Control Systems

(ii) PI controller:
Gc (s) =

0.9s + 22.5
s

(iii) PD controller:
Gc (s) = 0.9s + 22.5
The step responses for each controller is shown in Figure DP12.6. The
responses for the PID and PI controller are the same since the gains were
selected to obtain the same ITAE characteristic equation. An appropriate
prefilter is used in all cases.
(b) off-nominal plant
1.2

1

1

0.8

0.8

y(t)

y(t)

(a) nominal plant
1.2

0.6

0.6

0.4

0.4

0.2

0.2

0
0

0.5

1

1.5

Time (sec)

2

0
0

0.5

1

1.5

2

Time (sec)

FIGURE DP12.6
(a) Step response for nominal plant: PID (solid line); PI (dashed line); and PD (dotted
line); (b) for off-nominal plant: PID (solid line); PI (dashed line); and PD (dotted line).

DP12.7

The loop transfer function is
G(s) =

Ka Km
K
=
(0.5s + 1)(τf s + 1)s(s + 1)
s(s + 2)(s + 1)

since τf is negligible. A suitable PID controller is
Gc (s) =

300(s2 + 2.236s + 2.5)
KKD (s2 + as + b)
=
.
s
s

The step response is shown in Figure DP12.7. The percent overshoot is

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

699

Design Problems

P.O. = 4.6% and the settling time is Ts = 3.74 seconds.

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Time (secs)

FIGURE DP12.7
Step response for the elevator position control.

DP12.8

The system transfer function is
Y (s) =



G(s)Gc (s)Gp (s)
R(s) .
1 + G(s)Gc (s)


We are given
G(s) = e−sT

where T = 1 second .

Using a second-order Pade approximation yields
G(s) ≈

s2 − 6s + 12
.
s2 + 6s + 12

Three controllers that meet the specifications are
0.5
(Integral controller)
s
0.04s + 0.4
Gc2 (s) =
(PI controller)
s
0.01s2 + 0.04s + 0.4
Gc3 (s) =
(PID controller) .
s
Gc1 (s) =

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

700

CHAPTER 12

Robust Control Systems

In all cases, the steady-state error is zero.
Integral

PI

PID

P.O.(%)

4.05

0

0

Ts (sec)

6.03

6.12

6.02

Tp (sec)

4.75

N/A N/A

|V (t)|max (volts)

1.04

1

1

The prefilter Gp (s) = 1 is used in all designs. To compute the voltage,
the transfer function is
V (s) =

DP12.9

Gp (s)Gc (s)
R(s) .
1 + Gc (s)G(s)

The space robot transfer function is
G(s) =

1
.
s(s + 10)

(a) Consider Gc (s) = K. Then
T (s) =

Gc (s)G(s)
K
= 2
.
1 + Gc (s)G(s)
s + 10s + K

We determine that K = 50.73 for ζ = 0.702. Thus, we expect P.O. <
4.5%. So,
Gc (s) = 50.73 .
(b) Consider the PD controller
Gc (s) = KP + KD s .
Then
T (s) =

KP + KD s
.
s2 + (10 + KD )s + KP

Using the ITAE method, we compute
KP = 100

and KD = 4 .

Thus,
Gc (s) = 4s + 100 ,

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

701

Design Problems

and the prefilter is
Gp (s) =

100
.
4s + 100

(c) Consider the PI controller
Gc (s) = KP +

KP s + KI
KI
=
.
s
s

Then,
T (s) =

s3

KP s + KI
.
+ 10s2 + KP s + KI

Using the ITAE method, we have
ωn = 5.7

KP = 70.2

and KI = 186.59 .

Thus,
Gc (s) = 70.2 + 186.59/s ,
and the prefilter is
Gp (s) =

186.59
.
70.2s + 186.59

(d) Consider the PID controller
Gc (s) =

KD s2 + KP s + KI
.
s

Then,
T (s) =

KD s2 + KP s + KI
.
s3 + 10s2 + KD s2 + KP s + KI

Using the ITAE method with ωn = 10, we have
KD = 7.5

KP = 215

and KI = 1000 .

Thus,
Gc (s) =

7.5s2 + 215s + 1000
,
s

and the prefilter is
Gp (s) =

7.5s2

1000
.
+ 215s + 1000

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

702

CHAPTER 12

Robust Control Systems

A summary of the performance is given in Table DP12.9.

Gc (s)

P.O.

tp

ts

yss

max|y(t)|

K

4.5%

0.62 s

0.84 s

0

0.026

PD

5.2%

0.39 s

0.56s

0

0.010

PI

1.98%

0.81 s

1.32s

0

0.013

PID

1.98%

0.46 s

0.75 s

0

0.004

TABLE DP12.9

A summary of performance to a disturbance input.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

703

Computer Problems

Computer Problems
The closed-loop transfer function is
T (s) =

s2

8K
,
+ 2s + 8K

T , is
and the sensitivity function, SK

S(s) =

s2 + s
.
s2 + 2s + 8K

The plot of T (s) and S(s) is shown in Figure CP12.1, where K = 10.
nt=[80]; dt=[1 2 80]; syst = tf(nt,dt);
ns=[1 2 0];ds=[1 2 80]; syss = tf(ns,ds);
w=logspace(-1,2,400);
[magt,phaset]=bode(syst,w);magtdB(1,:) = 20*log10(magt(1,1,:));
[mags,phases]=bode(syss,w); magsdB(1,:) = 20*log10(mags(1,1,:));
semilogx(w,magtdB,w,magsdB,'--')
legend('20log|T|','20log|S|')
ylabel('Gain dB')
grid
20
20log|T|
20log|S|
10

0

−10
Gain dB

CP12.1

−20

−30

−40

−50

−60
−1
10

0

1

10

10

FIGURE CP12.1
Plot of T (s) and the sensitivity function S(s).

2

10

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

704

CHAPTER 12

CP12.2

Robust Control Systems

A reasonable value of the gain K = 4. The family of step responses is
shown in Figure CP12.2.

p=[0.5:0.5:20]; K=4;
t=[0:0.01:1];
for i=1:length(p)
n=[K*p(i)]; d=[1 p(i)]; sys = tf(n,d);
sys_cl = feedback(sys,[1]);
y=step(sys_cl,t); Y(:,i)=y;
[y2,t2]=step(sys_cl);
S=stepinfo(y2,t2); Ts(i)=S.SettlingTime;
end
plot(t,Y) , xlabel('Time (sec)'), ylabel('Step response')

0.9
0.8
0.7

Step response

0.6
0.5
0.4
0.3
0.2
0.1
0

0

0.1

0.2

0.3

0.4

0.5
0.6
Time (sec)

0.7

FIGURE CP12.2
Family of step responses for 0.5 < p < 20.

CP12.3

The closed-loop characteristic equation is
1 + KD

s2 + as + b
=0
Js3

where
a = KP /KD
b = KI /KD .

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

705

Computer Problems

We select
a = 1 and

b=2

to move the root locus into the left hand-plane (see Figure CP12.3a).
Then, we choose
KD = 71
from the root locus using the rlocfind function. The closed-loop Bode
plot in Figure CP12.3b verifies that the bandwidth
Also, the phase margin is
P.M. = 45.7o ,
which meets the design specification. The plot of phase margin versus J
is shown in Figure CP12.3c. We see that as J increases, the phase margin
decreases.
J=25; a=1; b=2;
ng=[1];dg=[J 0 0]; sysg=tf(ng,dg);
nc=[1 a b]; dc=[1 0]; sysc=tf(nc,dc);
sys=series(sysc,sysg); rlocus(sys)
Root Locus
2
1.5

Imaginary Axis

1
0.5
0
−0.5
−1
−1.5
−2
−1.2

−1

−0.8

−0.6

FIGURE CP12.3
2
+s+2
(a) Root locus for 1 + KD s 10s
= 0.
3

−0.4
Real Axis

−0.2

0

0.2

0.4

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 12

Robust Control Systems

J=25; a=1; b=2; KD=71; KP=a*KD; KI=b*KD;
ng=[1]; dg=[J 0 0]; sysg=tf(ng,dg);
nc=[KD KP KI]; dc=[1 0]; sysc = tf(nc,dc);
sys=series(sysc,sysg); sys_cl = feedback(sys,[1]);
bode(sys_cl);
[GM,PM]=margin(sys); PM

PM =
45.7093

20

Magnitude (dB)

10
0
−10
−20
−30

XY
10

0

1

10

2

10

10

FIGURE CP12.3
CONTINUED: (b) Closed-loop Bode plot with ωB < 5 rad/sec.

Ji=[10:1:40];
for i=1:length(Ji)
numc=[KD KP KI]; denc=[Ji(i) 0 0 0]; sysc = tf(numc,denc);
[gm,pm]=margin(sysc);
Pm(i)=pm;
end
plot(Ji,Pm), grid
xlabel('J'), ylabel('Phase Margin (deg)')

90
80
70
60
Phase Margin (deg)

706

50
40
30
20
10
0
−10

0

5

10

15

FIGURE CP12.3
CONTINUED: (c) Phase margin versus J.

20
J

25

30

35

40

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

707

Computer Problems

The closed-loop characteristic equation is

CP12.4

1+K

s2

1
=0
+ bs + a

where a = 8 and the nominal value of b = 4. The root locus is shown in
Figure CP12.4a.

clf, hold off
a=8; b=4;
num=[1]; den=[1 b a]; sys = tf(num,den);
rlocus(sys), hold on
zeta=0.59; wn=1.35;
x=[-10:0.1:-zeta*wn]; y=-(sqrt(1-zeta^2)/zeta)*x;
xc=[-10:0.1:-zeta*wn];c=sqrt(wn^2-xc.^2);
plot(x,y,':',x,-y,':',xc,c,':',xc,-c,':')
rlocfind(sys)

ÈSelect a point in the graphics window
selected_point =
-2.0165 + 2.5426i

K

ans =
2.4659

4
3
+

2

x

Imag Axis

1
0
-1
-2

x
+

-3
-4
-4

-3

-2

-1

0

1

2

3

4

Real Axis

FIGURE CP12.4
1
(a) Root locus for 1 + K s2 +4s+8
.

The performance region is specified by
ζ = 0.59

and

ωn = 1.35 ,

which derives from the design specifications
Ts < 5 sec and P.O. < 10% .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

708

CHAPTER 12

Robust Control Systems

Using an m-file, the value of
K = 2.5
is selected with the rlocfind function. The step responses for b = 0, 1, 4
and b = 40 are shown in Figure CP12.4b. When b = 0, the system is
marginally stable; b = 1 results in a stable system with unsatisfactory
performance. The nominal case b = 4 is stable and all performance specs
are satisfied. When b = 40, the system is heavily damped: the percent
overshoot specification is satisfied, but the settling time is too long.

0.5

b=0

0.45
0.4

b=1

Amplitude

0.35
0.3
0.25
0.2
0.15

b=4

0.1
0.05
0
0

b=40
1

2

3

4

5

6

7

8

9

10

Time (secs)

FIGURE CP12.4
CONTINUED: (b) Step responses for b = 0, 1, 4 and 40.

CP12.5

(a) An acceptable lead compensator (designed with root locus methods)
is
Gc (s) = K

s+a
s + 0.3
=5
.
s+b
s+2

The compensated root locus is shown in Figure CP12.5a, where
K=5
is selected to place the closed-loop poles in the performance region.
(b) The step responses for ζ = 0, 0.005, 0.1 and 1 are shown in Figure CP12.5b.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

709

Computer Problems

4
3
+

2
+

x
o

Imag Axis

1
0

+o

x

x

-1
+

-2

o
x

+

-3
-4
-4

-3

-2

-1

0

1

2

3

4

Real Axis

FIGURE CP12.5
(a) Compensated root locus.

(c) You would like the actual structural damping to be greater than the
design value, if it must be different at all.
zeta=0,0.005 (solid); zeta=0.1 (dashed); zeta=1 (dotted)
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

Time (sec)

FIGURE CP12.5
CONTINUED: (b) Step responses for ζ = 0, 0.005, 0.1 and 1.

16

18

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

710

CHAPTER 12

CP12.6

Robust Control Systems

The m-file script which computes the phase margin as a function of the
time delay (using the pade function) is shown in Figure CP12.6. The
maximum time delay (for stability) is td = 4.3 seconds.
K=5;
numg=K*[1]; deng=[1 10 2]; sysg = tf(numg,deng);
time delay vector
td=[0:0.1:5];
for i=1:length(td)
sys = series(sysg,sysd);
[mag,phase,w]=bode(sys);
[gm,pm,w1,w2]=margin(mag,phase,w);
pmv(i)=pm;
end
plot(td,pmv), grid
xlabel('time delay [sec]')
ylabel('phase margin [deg]')

120

100

phase margin [deg]

80

60

40

20

0

-20

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

time delay [sec]

FIGURE CP12.6
Phase margin versus time delay.

CP12.7

The m-file script is shown in Figure CP12.7a. The steady-state error
(shown in Figure CP12.7b) is zero when
a = 0.5
and increases rapidly as a increases past a = 0.5. The maximum initial
undershoot is shown in Figure CP12.7c. As a increases, the initial undershoot increases linearly. The gain margin is shown in Figure CP12.7d. It

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

711

Computer Problems

can be seen that as a increases, the gain margin decreases very rapidly.
a=[0.01:0.01:0.99];
t=[0:0.1:30];
for i=1:length(a)
num=a(i)*[1 -0.5]; den=[1 2 1]; sys_o = tf(num,den);
[mag,phase,w]=bode(sys_o);
[gm,pm,w1,w2]=margin(mag,phase,w);
gain margin
gmv(i)=gm;
sys_cl = feedback(sys_o,[1]);
[y,x]=step(-sys_cl,t);
negative unit step input
yf(i)=1-y(length(t));
ym(i)=-min(y)*100;
max initial undershoot
end
figure(1), plot(a,gmv), grid, xlabel('a'), ylabel('gm')
figure(2), plot(a,yf ), grid, xlabel('a'), ylabel('steady-state error')
figure(3), plot(a,ym), grid, xlabel('a'), ylabel('maximum initial undershoot [%]')

FIGURE CP12.7
Script to generate all the plots.

1
0.9
0.8

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

0

0.1

0.2

0.3

0.4

0.5
a

FIGURE CP12.7

0.6

0.7

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 12

Robust Control Systems

25

maximum initial undershoot [%]

20

15

10

5

0

0

0.1

0.2

0.3

0.4

0.5
a

0.6

0.7

0.8

0.9

1

0.6

0.7

0.8

0.9

1

FIGURE CP12.7
CONTINUED: (c) Maximum initial undershoot.

250

200

150
gm

712

100

50

0

0

0.1

0.2

FIGURE CP12.7
CONTINUED: (d) Gain margin.

0.3

0.4

0.5
a

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

713

Computer Problems

CP12.8

The plant (balloon and canister dynamics plus motor) is given by
G(s) =

1
,
(s + 2)(s + 4)(s + 10)

and the PID controller is
Gc (s) =

KD (s2 + as + b)
.
s

Let a = 6. Then using the root locus methods, we determine that with
KD = 12.5

and

b = 10

we have the roots
s1 = −8.4
s2 = −4.7
s3,4 = −1.43 ± j1.05 .
Thus, ζ = 0.8. The plot of y(t) is shown in Figure CP12.8. The percent
overshoot is less that 3%, as desired.

1.4

1.2

With prefilter
1

y(t)

0.8

Without prefilter
0.6

0.4

0.2

0

0

FIGURE CP12.8
Simulation of the GRID device.

0.5

1

1.5

2

2.5
Time (sec)

3

3.5

4

4.5

5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

C H A P T E R

1 3

Digital Control Systems

Exercises
E13.1

(a) Elevation contours on a map are discrete signals. (b) Temperature in
a room is a continuous signal. (c) A digital clock display is a discrete
signal. (d) The score of a basketball game is a discrete signal. (e) The
output of a loudspeaker is a continuous signal.

E13.2

(a) Using long-division we determine that
Y (z) = z −1 + 3z −2 + 7z −3 + 15z −4 + · · ·
Therefore, with
Y (z) =

∞
X

y(kT )z −k

k=0

we have
y(0) = 0

y(T ) = 1

y(2T ) = 3 y(3T ) = 7 y(4T ) = 15 .

(b) The exact solution is
y(kT ) = ek ln 2 − 1 .
E13.3

For the system response
y(kT ) = kT
where k ≥ 0, we have
Y (z) =

E13.4

The partial fraction expansion of Y (s) is
Y (s) =

714

Tz
.
(z − 1)2

5
0.25 0.0625 0.3125
=
+
−
.
s(s + 2)(s + 10)
s
s + 10
s+2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

715

Exercises

Then, using Table 13.1 in Dorf and Bishop, we determine that
z
z
z
+ 0.0625
− 0.3125
−10T
z −1
z −e
z − e−2T
z
z
z
= 0.25
+ 0.0625
− 0.3125
,
z −1
z − 0.135
z − 0.670

Y (z) = 0.25

where T = 0.1.
E13.5

The Space Shuttle and robot arm control block diagram is shown in Figure E13.5. The human operator uses information from the computer generated data display and visual sensory data from the TV monitor and by
looking out the window. He/she commands the robot arm via a joystick
command to the computer.

data
display

measurement
digital

analog

A/D

joint angle &
rate sensors

digital
human
operator
joystick
command

ref.

+

Computer

digital

analog

D/A

-

Robot arm
& motors/gears

tip
position

measurement

TV monitor
& window view

FIGURE E13.5
The Space Shuttle/robot arm control block diagram.

E13.6

From Section 10.8 in Dorf and Bishop, we find that the design resulted
in the compensator
Gc (s) =

6.66s + 1
s + 0.15
= 0.1
.
66.6s + 1
s + 0.015

Using the relationships
A = e−aT ,
we compute

B = e−bT ,

and C

1−A
a
=K ,
1−B
b

A = e−0.15(0.001) = 0.99985 , B = e−0.015(0.001) = 0.999985 , and C = 0.1 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

716

CHAPTER 13

Digital Control Systems

Therefore,
D(z) = C
E13.7

z − 0.99985
z−A
= 0.1
.
z−B
z − 0.999985

Using long-division, we determine that
Y (z) = 1 + 3.5z −1 + 5.75z −2 + 6.875z −3 + · · ·
Therefore, with
Y (z) =

∞
X

y(kT )z −k

k=0

we have
y(0) = 1 y(T ) = 3.5
E13.8

y(2T ) = 5.75

y(3T ) = 6.875 .

The closed-loop system with
T (z) =

z2

z
+ 0.2z − 1.0

is unstable since one of the poles of the transfer function (z = −1.1 and
z = 0.90) lies outside the unit circle in the z-plane.
E13.9

(a) Using long-division we determine that
Y (z) = z −1 + z −2 + z −3 + z −4 + · · ·
Therefore, with
Y (z) =

∞
X

y(kT )z −k

k=0

we have
y(0) = 0 y(T ) = 1 y(2T ) = 1

y(3T ) = 1 y(4T ) = 1 .

(b) The exact solution is
y(kT ) = 1 − δ(k)
where δ(k) = 1 when k = 0 and δ(k) = 0 when k 6= 0.
E13.10

We compute T /τ = 1.25.
(a) Using Figure 13.19 in Dorf and Bishop, we determine that Kτ = 0.8
which implies K = 100.
(b) Using Figure 13.21 in Dorf and Bishop, we determine that ess = 0.75.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

717

Exercises

(c) Using Figure 13.20 in Dorf and Bishop, we determine that Kτ = 0.7
which implies K = 88.
(a) The transfer function (including the zero-order hold) is
Go (s)Gp (s) =

100(1 − e−sT )
.
s(s2 + 100)

Expanding into partial fractions yields
s
1
− 2
G(z) = (1 − z )Z
s s + 100


z
z(z − cos 10T )
−1
= (1 − z )
−
.
z − 1 z 2 − 2 cos 10T z + 1




−1

When T = 0.05 we ha,ve
G(z) =

0.1224(z + 1)
.
− 1.7552z + 1

z2

(b) The system is marginally stable since the system poles, z = −0.8776±
0.4794j, are on the unit circle.
(c) The impulse response and sinusoidal input response are shown in
Figure E13.11.

Amplitude

0.5

0

-0.5

2

0

4

6

8

10

12

14

16

No. of Samples
40

Amplitude

E13.11

20
0
-20
-40
0

10

20

30

40

50

60

70

No. of Samples

FIGURE E13.11
Impulse and sinusoidal (natural frequency) input response.

80

90

100

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

718

CHAPTER 13

E13.12

Digital Control Systems

The partial fraction expansion of X(s) is
X(s) =

s2

s+1
1
2
=
−
.
+ 5s + 6
s+3 s+2

Then, with T = 1, we have
X(z) =
E13.13

z
2z
z
2z
−
=
−
.
−3
−2
z−e
z−e
z − 0.0498 z − 1353

The root locus is shown in Figure E13.13. For stability: 2.2 < K < 5.8.
Root Locus
2

1.5

Imaginary Axis

1

0.5
K=5.8
0
K=2.2
−0.5

−1

−1.5

−2
−2

−1.5

−1

−0.5

0
Real Axis

0.5

1

1.5

2

FIGURE E13.13
Root locus with unit circle (dashed curve).

E13.14

Given Gp (s), we determine that (with K = 5)
G(z) =

5(1 − e−1 )z
.
z(z − e−1 )

The closed-loop characteristic equation is
z 2 + 1.792z + 0.368 = 0
and the system is unstable, since there is a pole at z = −1.55. The
system is stable for
0 < K < 4.32 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

719

Exercises

E13.15

The transfer function G(z) is
G(z) =

z2

The sampling time is T = 1 s.
E13.16

0.1289z + 0.02624
.
− 0.3862z + 0.006738

The transfer function G(z) is
G(z) =

0.2759z + 0.1982
.
− 1.368z + 0.3679

z2

The sampling time is T = 0.5 s.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

720

CHAPTER 13

Digital Control Systems

Problems
P13.1

The plot of the input to the sampler and the output r ∗ (t) is shown in
Figure P13.1.

1
0.8
0.6
0.4

r(t), r*(t)

0.2
0
-0.2
-0.4
-0.6
-0.8
-1
0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Time (sec)

FIGURE P13.1
Plot of r(t) = sin(ωt) and r ∗ (t).

The plot of the input and the output is shown in Figure P13.2.

1
0.8
0.6
0.4
0.2

r(t)

P13.2

0
-0.2
-0.4
-0.6
-0.8
-1
0

0.2

0.4

0.6

0.8

1

1.2

Time (sec)

FIGURE P13.2
Plot of r(t) = sin(ωt) and output of sample and hold.

1.4

1.6

1.8

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

721

Problems

P13.3

The transfer function
Y (z)/R∗ (z) = G(z) =

z
.
z − e−T

The ramp input is represented by
R(z) =

Tz
.
(z − 1)2

The output Y (z) = G(z)R(z) is obtained by long division:
h

i

Y (z) = T z −1 + T (2 + e−T )z −2 − T (1 + 2e−T ) − (2 + e−T )2 z −3
h

+ T e−T + (1 + 2e−T )(2 + e−T )


− (2 + e−T ) (1 + 2e−T ) − (2 + e−T )2
P13.4

i

z −4 + · · ·

The transfer function
Y (s)/R∗ (s) =

1 − e−sT
.
s(s + 2)

The partial fraction expansion (with T = 1) yields
G(z) = (1 − z −1 )Z
=
P13.5



0.5
0.5
−
s
s+2

0.4323
.
z − 0.1353



= (1 − z −1 )



0.5z
0.5z
−
z − 1 z − 0.1353



The step input is
R(z) =

z
.
z−1

Also,
T (z) =

G(z)
0.6321
=
.
1 + G(z)
z + 0.2643

So,
Y (z) = T (z)R(z) =

0.6321
z
0.6321z
= 2
.
z + 0.2643 z − 1
z − 0.7357z − 0.2643

Using long-division we determine that
Y (z) = 0.6321z −1 + 0.4650z −2 + 0.5092z −3 + 0.4975z −4 + 0.5006z −5 + · · ·

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

722

CHAPTER 13

Digital Control Systems

Therefore, with
Y (z) =

∞
X

y(kT )z −k

k=0

we have y(0) = 0, y(T ) = 0.6321, y(2T ) = 0.4650, y(3T ) = 0.5092,
y(4T ) = 0.4975, and y(5T ) = 0.5006.
P13.6

Using the final value theorem (see Table 13.1 in Dorf and Bishop), we
determine that (for a step input)
Yss = lim (z − 1)Y (z) = lim (z − 1)
z→1

z→1

0.6321
z
0.6321
=
= 0.5 .
z + 0.2643 z − 1
1.2643

And using the initial value theorem, we compute
Yo = lim Y (z) = lim
z→∞

z→∞

0.6321
z
=0.
z + 0.2643 z − 1

P13.7

Using Figures 13.19 and 13.21 in Dorf and Bishop, we determine that
the performance specifications are satisfied when Kτ = 0.5 and Tτ = 2.
Computing K and T (with τ = 0.5) yields K = 1 and T = 1.

P13.8

We can select K = 1 and r = 0.2. The step responses for the compensated
and uncompensated systems are shown in Figure P13.8.

1.2

1
Uncompensated

Amplitude

0.8

0.6

Compensated

0.4

0.2

0
0

10

20

30
40
No. of Samples

FIGURE P13.8
Plot of compensated and uncompensated systems.

50

60

70

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

723

Problems

P13.9

Consider the compensator
Gc (s) = K

s+a
.
s+b

Then, using Bode methods we can select a = 1, b = 4, and K = 1. The
compensated system phase margin is P.M. = 50o and the gain margin is
G.M. = 15dB. The crossover frequency is ωc = 2.15 rad/sec. Utilizing the
Gc (s)-to-D(z) method and selecting T = 0.01 second, we determine
D(z) = C
We use the relationships
A = e−aT ,

z − 0.99
z−A
=
.
z−B
z − 0.96

B = e−bT ,

and C

1−A
a
=K ,
1−B
b

to compute A = e−0.01 = 0.99, B = e−0.04 = 0.96, and C = 1.
P13.10

(a) The transfer function G(z)D(z) is
G(z)D(z) = K

0.0037z + 0.0026
.
− 1.368z + 0.3679

z2

(b) The closed-loop system characteristic equation is
1+K

0.0037z + 0.0026
=0.
− 1.368z + 0.3679

z2

(c) Using root locus methods, the maximum value of K is found to be
Kmax = 239.
(d) Using Figure 13.19 in Dorf and Bishop for T /τ = 1 and a maximum
overshoot of 0.3, we find that K = 75. (e) The closed-loop transfer
function (with K = 75) is
T (z) =

0.2759z + 0.1982
.
− 1.092z + 0.5661

z2

The step response is shown in Figure P13.10.
(f) The closed-loop poles with K = 119.5 are z = 0.4641 ± 0.6843j. The
overshoot is 0.55.
(g) The step response is shown in Figure P13.10 (for K = 119.5).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

724

CHAPTER 13

Digital Control Systems

K=75

Amplitude

1.5

1

0.5

0

2

0

4

6

8

10

12

14

16

No. of Samples
K=119.5

Amplitude

2
1.5
1
0.5
0

0

2

4

6

8

10

12

14

16

18

20

No. of Samples

FIGURE P13.10
Step response for K = 75 and K = 119.5.

P13.11

(a) Consider the compensator
Gc (s) = K

s+a
.
s+b

Then, using Bode methods we can select a = 0.7, b = 0.1, and K =
150. The compensated system overshoot and steady-state tracking
error (for a ramp input) are P.O. = 30% and ess < 0.01.
(b) Utilizing the Gc (s)-to-D(z) method (with T = 0.1 second), we determine
D(z) = C
We use the relationships
A = e−aT ,

z−A
z − 0.9324
= 155.3
.
z−B
z − 0.99

B = e−bT ,

to compute
A = e−0.007 = 0.9324 ,

and C

1−A
a
=K ,
1−B
b

B = e−0.01 = 0.99 ,

and C = 155.3 .

(c) The step response for the continuous system with Gc (s) in part(a)
and for the discrete system with D(z) in part (b) is shown in Figure P13.11a.
(d) Utilizing the Gc (s)-to-D(z) method (with T = 0.01 second), we de-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

725

Problems

T=0.1 sec
1.8
1.6
1.4

Amplitude

1.2
1
0.8
0.6
0.4
0.2
0
0

2

4

6

8

10

12

14

16

18

20

No. of Samples

FIGURE P13.11
(a) Step response for continuous and discrete systems (T=0.1s) in Parts (a) and (b).

termine
D(z) = C
We use the relationships

to compute

z − 0.993
z−A
= 150
.
z−B
z − 0.999

A = e−aT
B = e−bT
1−A
a
C
=K
1−B
b
A = e−0.07 = 0.993
B = e−0.001 = 0.999
C = 150 .

The step response for the continuous system with Gc (s) in
and for the discrete system with D(z) in part (d) is shown
ure P13.11b.
(e) The ramp response for the continuous system with Gc (s) in
and for the discrete system with D(z) in part (b) is shown
ure P13.11c.

part(a)
in Figpart(a)
in Fig-

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

CHAPTER 13

Digital Control Systems
T=0.01 sec
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

20

40

60

80

100

120

140

160

180

200

No. of Samples

FIGURE P13.11
CONTINUED: (b) Step response for continuous and discrete systems (T=0.01s) in Parts
(a) and (d).

T=0.1 sec
2
1.8
1.6
1.4

Amplitude

726

1.2
1
0.8

Ramp input (dashed line)

0.6
0.4
0.2
0

0

2

4

6

8

10

12

14

16

18

20

No. of Samples

FIGURE P13.11
CONTINUED: (c) Ramp response for continuous and discrete systems (T=0.1s) in Parts
(a) and (b).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

727

Problems

P13.12

The root locus is shown in Figure P13.12. For stability: 0 < K < 2.
2

1.5
Unit circle (dashed line)

1

Imaginary Axis

0.5

0

−0.5

−1

−1.5

−2
−2

−1.5

−1

−0.5

0
Real Axis

0.5

1

1.5

2

FIGURE P13.12
z+0.5
= 0.
Root locus for 1 + K z(z−1)

The root locus is shown in Figure P13.13. When K = 0.027, the characteristic equation has two equal roots: z1,2 = 0.7247 and z3 = 0.2593.

2
Unit circle (dashed line)

1.5
1
0.5

Imag Axis

P13.13

0

o

o

xx

x

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

Real Axis

FIGURE P13.13
z 2 +1.1206z−0.0364
Root locus for 1 + K z 3 −1.7358z
2 +0.8711z−0.1353 = 0.

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

728

CHAPTER 13

P13.14

Digital Control Systems

The root locus is shown in Figure P13.14. When K = 9.5655 × 10−5 , the
two real roots break away from the real axis at z = 0.99. For stability:
K < 9.7 × 10−5 .
2
Unit circle (dashed line)

1.5
1

Imag Axis

0.5
x

0

o

o

x

x
x

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Real Axis

FIGURE P13.14
z 3 +10.3614z 2 +9.758z+0.8353
Root locus for 1 + K z 4 −3.7123z
3 +5.1644z 2 −3.195z+0.7408 = 0.

P13.15

Given
Gp (s) =

20
s−5

and the sample and hold (T=0.1s) as shown in Figure 13.18 in Dorf and
Bishop, we determine that
G(z) =

2.595
.
z − 1.649

Then, with R(z) = z/(z − 1), we have
Y (z) =

2.595z
.
(z − 1)(z + 0.9462)

Therefore, Y (z) = 2.59z −1 + 0.14z −2 + 2.46z −3 + 0.26z −4 + · · ·.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

729

Problems

P13.16

Given Gp (s) and the sample and hold (T=1s) as shown in Figure 13.18
in Dorf and Bishop, we determine that
0.22775z + 0.088984
.
− 1.0498z + 0.049787

G(z) =

z2

Then, with R(z) = z/(z − 1), we have
Y (z) =

0.22775z + 0.088984
z
.
− 0.82203z + 0.13877 z − 1

z2

The plot of y(kT ) is shown in Figure P13.16.

1
0.9
0.8
0.7

y(kT)

0.6
0.5
0.4
0.3
0.2
0.1
0

1

2

3

4

5

6

7

kT

FIGURE P13.16
Plot of y(kT ) for a step input.

P13.17

The root locus is shown in Figure P13.17 for
1+K

0.39532z + 0.30819
=0.
− 1.4724z + 0.47237

z2

The limiting value of the gain for stability is K = 1.71.

8

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

730

CHAPTER 13

Digital Control Systems

Root Locus
2
1.5

Imaginary Axis

1
0.5
0
−0.5
−1
Unit circle (dashed line)

−1.5
−2
−5

−4

−3

−2
−1
Real Axis

0

1

2

FIGURE P13.17
= 0.
Root locus for 1 + K z 20.39532z+0.30819
−1.4724z+0.47237

P13.18

The plot of the step responses for 0 ≤ T ≤ 1.2 is shown in Figure P13.18.
The overshoot and settling time summary is given in Table P13.18.

T

P.O.

Ts

0

0.2

0.4

0.6

0.8

1.0

1.2

16.3%

20.6%

25.6%

31.3%

36.9%

40.0%

51.0%

8.1

8.4

8.8

11.4

14.4

16.0

19.2

TABLE P13.18

Performance summary.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

731

Problems
1.6

1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0
0

10

20

FIGURE P13.18
Step responses for 0 ≤ T ≤ 1.2.

30

40
50
60
No. of Samples

70

80

90

100

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

732

CHAPTER 13

Digital Control Systems

AP13.1

Given the sample and hold with Gp (s), we determine that
G(z) =

10.5K(z − 0.9048)
.
(z − 1)2

The root locus is shown in Figure AP13.1. For stability: 0 < K < 0.2.
2
Unit circle (dashed line)

1.5
1

Imag Axis

0.5
0

ox

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Real Axis

FIGURE AP13.1
10.5(z−0.9048)
= 0 with unit circle (dashed line).
Root locus for 1 + K
(z−1)2

The root locus is shown in Figure AP13.2a. The loop transfer function is

2
Unit circle (dashed line)

1.5
1
0.5

Imag Axis

AP13.2

0

o

x

x

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0
Real Axis

FIGURE AP13.2
0.0379z
= 0.
(a) Root locus for 1 + K (z−1)(z−0.368)

0.5

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

733

G(z)D(z) = K

0.0379z
.
(z − 1)(z − 0.368)

For stability: Kmax = 72. We select K = 8.2. The step response is shown
in Figure AP13.2b.
1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

1

2

3

4

5

6

7

8

9

10

No. of Samples

FIGURE AP13.2
CONTINUED: (b) Step response with K = 8.2.

The root locus is shown in Figure AP13.3a.

The maximum gain for

Root Locus
2
Unit circle (dashed line)

1.5

System: sysz
Gain: 5.99
Pole: 0.736 + 0.257i
Damping: 0.596
Overshoot (%): 9.74

1
Imaginary Axis

AP13.3

0.5
0
−0.5
−1
−1.5
−2
−3

−2.5

−2

−1.5

−1

−0.5
Real Axis

FIGURE AP13.3
(a) Root locus for 1 + K z0.07441z+0.06095
2 −1.474z+0.6098 = 0.

0

0.5

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

734

CHAPTER 13

Digital Control Systems

stability is Kmax = 44.3. We select K = 6. The step response is shown in
Figure AP13.3b.

Step Response
1.4

1.2

Amplitude

1

0.8

0.6

0.4

0.2

0

0

0.2

0.4

0.6
Time (sec)

0.8

1

1.2

FIGURE AP13.3
CONTINUED: (b) Step response with K = 6.

AP13.4

The loop transfer function is
G(z) =

10(1 − e−T )
,
z − e−T

and the closed-loop transfer function is
T (z) =

10(1 − e−T )
.
z − (11e−T − 10)

For stability, we require
|11e−T − 10| < 1 .
Solving for T yields
0 < T < 0.2 .
Selecting T = 0.1s provides a stable system with rapid response; the
settling time is Ts = 0.2s. The step response is shown in Figure AP13.4.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

735

1
0.9
0.8

Amplitude

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

0

0.5

1

1.5

2

2.5

3

3.5

4

No. of Samples

FIGURE AP13.4
Step response with T = 0.1s.

The maximum gain for stability is Kmax = 63.15.
Root Locus
2

Unit circle (dashed line)

1.5

1

Imaginary Axis

AP13.5

0.5

System: sysz
Gain: 63.2
Pole: 0.725 − 0.686i
Damping: 0.00308
Overshoot (%): 99

0

−0.5

−1

−1.5

−2
−3

−2.5

−2

−1.5

−1

−0.5
Real Axis

FIGURE AP13.5
Root locus for 1 + K 0.004535z+0.004104
z 2 −1.741z+0.7408 = 0.

0

0.5

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

736

CHAPTER 13

Digital Control Systems

Design Problems
The plant model with parameters given in Table CDP2.1 in Dorf and
Bishop is given by:
26.035
,
s(s + 33.142)

Gp (s) =

where we neglect the motor inductance Lm and where we switch off the
tachometer feedback (see Figure CDP4.1 in Dorf and Bishop). Letting




G(z) = Z G≀ (∫ )G√ (∫ )
we obtain
G(z) =

1.2875e − 05(z + 0.989)
.
(z − 1)(z − 0.9674)

A suitable controller is
D(z) =

20(z − 0.5)
.
z + 0.25

The step response is shown below. The settling time is under 250 samples.
With each sample being 1 ms this means that Ts < 250 ms, as desired.
Also, the percent overshoot is P.O. < 5%.

1.2

1

0.8
Amplitude

CDP13.1

0.6

0.4

0.2

0
0

50

100

150
No. of Samples

200

250

300

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

737

Design Problems

(a) Given the sample and hold with Gp (s), we determine that
KG(z) = K

0.1228
.
z − 0.8465

The root locus is shown in Figure DP13.1a. For stablity: 0 ≤ K < 15.

Unit circle (dashed line)

1.5

1

0.5

Imag Axis

DP13.1

0

x

-0.5

-1

-1.5
-1.5

-1

-0.5

0

0.5

1

1.5

Real Axis

FIGURE DP13.1
0.1228
(a) Root locus for 1 + K z−0.8465
= 0 with unit circle (dashed line).

(b) A suitable compensator is
Gc (s) =

15(s + 0.5)
.
s+5

Utilizing the Gc (s)-to-D(z) method (with T = 0.5 second), we determine
D(z) = C

z−A
z − 0.7788
= 6.22
.
z−B
z − 0.0821

We use the relationships
A = e−aT ,

B = e−bT ,

and C

1−A
a
=K ,
1−B
b

to compute
A = e−0.5(0.5) = 0.7788 ,

B = e−0.5(5) = 0.0821 ,

and C = 6.22 .

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

738

CHAPTER 13

Digital Control Systems

(c) The step response is shown in Figure DP13.1b.

0.8
0.7
0.6

Amplitude

0.5
0.4
0.3
0.2
0.1
0

0

2

4

6

8

10

12

14

16

18

20

No. of Samples

FIGURE DP13.1
CONTINUED: (b) Closed-loop system step response.

DP13.2

With the sample and hold (T=10ms), we have
G(z) =

0.00044579z + 0.00044453
.
z 2 − 1.9136z + 0.99154

A suitable compensator is
D(z) = K

z − 0.75
,
z + 0.5

√
where K is determined so that ζ of the system is 1/ 2. The root locus is
shown in Figure DP13.2. We choose K = 1400.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

739

Design Problems

Root Locus
1.5
Curve of constant zeta=0.707 (dashed line)
1

Imaginary Axis

0.5

0

−0.5

−1

−1.5
−5

−4

−3

−2
−1
Real Axis

0

1

2

FIGURE DP13.2
0.00044579z+0.00044453
Root locus for 1 + K z−0.75
= 0.
z+0.5
z 2 −1.9136z+0.99154

The root locus is shown in Figure DP13.3a.

2
Curve of constant zeta=0.707 (dashed line)
1.5
1
0.5

Imag Axis

DP13.3

0

o

x

x

0.5

1

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

1.5

2

Real Axis

FIGURE DP13.3
z+1
(a) Root locus for 1 + K (z−1)(z−0.5)
= 0.

The gain for ζ = 0.707 is K = 0.0627. The step response is shown in
Figure DP13.3b. The settling time is Ts = 14T = 1.4s and P.O. = 5%.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

740

CHAPTER 13

Digital Control Systems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

18

No. of Samples

FIGURE DP13.3
CONTINUED: (b) Step response with K = 0.0627.

With the sample and hold (T=1s), we have
G(z) =

0.484(z + 0.9672)
.
(z − 1)(z − 0.9048)

2
Curve of constant zeta=0.5 (dashed line)
1.5
1
0.5

Imag Axis

DP13.4

0

o

x

ox x

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

Real Axis

FIGURE DP13.4
0.484(z+0.9672)
(a) Root locus for 1 + K z−0.88
z+0.5 (z−1)(z−0.9048) = 0.

1

1.5

2

20

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

741

Design Problems

1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

2

4

6

8

10

12

14

16

No. of Samples

FIGURE DP13.4
CONTINUED: (b) Step response for K = 12.5.

A suitable compensator is
D(z) = K

z − 0.88
,
z + 0.5

where K is determined so that ζ of the system is 0.5. The root locus
is shown in Figure DP13.4a. We choose K = 12.5. The step response
is shown in Figure DP13.4b. Also, Kv = 1, so the steady-state error
specification is satisfied.
DP13.5

Select T = 1 second. With the sample and hold, we have
G(z) =

0.2838z + 0.1485
.
− 1.135z + 0.1353

z2

The root locus is shown in Figure DP13.5. To meet the percent overshoot
specification, we choose K so that ζ of the system is 0.7. This results in
K = 1. The step response has an overshoot of P.O. = 4.6%. Also, from
Figure 13.21 in Dorf and Bishop, we determine that the steady-state error
to a ramp input is ess = 2 (since T /τ = 2, and Kτ = 0.3).

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

742

CHAPTER 13

Digital Control Systems

2
Curve of constant zeta=0.7 (dashed line)
1.5
1

Imag Axis

0.5
0

o

x

x

-0.5
-1
-1.5
-2
-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Real Axis

FIGURE DP13.5
Root locus for 1 + K z 20.2838z+0.1485
−1.135z+0.1353 = 0.

DP13.6

With the sample and hold at T = 1 , we have
G(z) =

z2

Consider the digital controller

0.298z + 0.296
.
− 1.98z + 0.9802

Dz) = K

z − 0.9
.
z + 0.6

The root locus is shown in Figure DP13.6. To meet the percent overshoot
specification, we choose K so that ζ of the system is greater than 0.52.
We select K = 2. The step response has an overshoot of P.O. = 11.9%
and the settling time is Ts = 17.8s.

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

743

Design Problems

Root Locus
1
0.8
0.6

Imaginary Axis

0.4
0.2
0
−0.2
−0.4
−0.6
−0.8
−1
−5

−4

−3

−2
−1
Real Axis

0

1

2

FIGURE DP13.6
0.298z+0.296
Root locus for 1 + K z−0.9
z+0.6 z 2 −1.98z+0.9802 = 0.

Step Response
1.4

System: syscl
Peak amplitude: 1.12
Overshoot (%): 11.9
At time (sec): 2

1.2

System: syscl
Settling Time (sec): 17.8

Amplitude

1

0.8

0.6

0.4

0.2

0

0

5

10

15
20
Time (sec)

FIGURE DP13.6
CONTINUED: (b) Step response for K = 2.

25

30

35

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

744

CHAPTER 13

Digital Control Systems

Computer Problems
CP13.1

The m-file script and unit step response are shown in Figure CP13.1.
num=[0.2145 0.1609]; den=[1 -0.75 0.125];
sysd = tf(num,den,1);
step(sysd,0:1:50)
1.2

1

Amplitude

0.8

0.6

0.4

0.2

0
0

5

10

15

20

25

30

35

40

45

50

No. of Samples

FIGURE CP13.1
Step response.

CP13.2

The m-file script utilizing the c2d function is shown in Figure CP13.2.

% Part (a)
num = [1]; den = [1 0]; T = 1;
sys = tf(num,den);
sys_d = c2d(sys,T,'zoh')
%
% Part (b)
num = [1 0]; den = [1 0 2]; T = 1;
sys = tf(num,den);
sys_d=c2d(sys,T,'zoh')

FIGURE CP13.2
Script utilizing the c2d function for (a) and (b).

Transfer function:
1
----z-1
Transfer function:
0.6985 z - 0.6985
-----------------z^2 - 0.3119 z + 1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

745

Computer Problems

% Part (c)
num = [1 4]; den = [1 3]; T = 1;
sys = tf(num,den);
sys_d = c2d(sys,T,'zoh')
%
% Part (d)
num = [1]; den = [1 8 0]; T = 1;
sys = tf(num,den);
sys_d = c2d(sys,T,'zoh')

Transfer function:
z + 0.267
----------z - 0.04979
Transfer function:
0.1094 z + 0.01558
------------------z^2 - z + 0.0003355

FIGURE CP13.2
CONTINUED: Script utilizing the c2d function for (c) and (d).

The continuous system transfer function (with T = 0.1 sec) is
T (s) =

s2

13.37s + 563.1
.
+ 6.931s + 567.2

The step response using the dstep function is shown in Figure CP13.3a.
The contrinuous system step response is shown in Figure CP13.3b.

1.8
1.6
1.4
1.2

Amplitude

CP13.3

1
0.8
0.6
0.4
0.2
0

0

2

4

6

8

No. of Samples

FIGURE CP13.3
(a) Unit step response using the dstep function.

10

12

14

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

746

CHAPTER 13

Digital Control Systems

1.8
*

1.6
1.4
*

1.2

*

1

*

*

*

*

*

*

*

*
*

0.8

*

*

0.6
0.4
0.2
0*
0

0.2

0.4

0.6

0.8

1

1.2

1.4

FIGURE CP13.3
CONTINUED: (b) Continuous system step response (* denote sampled-data step response).

The root locus in shown in Figure CP13.4. For stability: 0 < K < 2.45.
Root Locus
2
1.5
1
Imaginary Axis

CP13.4

0.5
0
−0.5
−1
−1.5
−2
−2

−1.5

−1

FIGURE CP13.4
z
Root locus for 1 + K z 2 −z+0.45
= 0.

−0.5

0
Real Axis

0.5

1

1.5

2

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

747

Computer Problems

CP13.5

The root locus in shown in Figure CP13.5. For stability: 0 < K < ∞.

1
0.8
0.6

Imag Axis

0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1

-0.8

-0.6

-0.4

-0.2

0
0.2
Real Axis

0.4

0.6

0.8

1

FIGURE CP13.5
(z−0.2)(z+1)
Root locus for 1 + K (z−0.08)(z−1) = 0

The root locus is shown in Figure CP13.6.

Root Locus
1.5
1

Imaginary Axis

CP13.6

0.5
0
0.5
1
1.5
1.5

1

0.5

0
Real Axis

FIGURE CP13.6
Root locus.

0.5

1

1.5

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

748

CHAPTER 13

Digital Control Systems

We determine the range of K for stability is 0.4 < K < 1.06.
% Part (a)
num=[1 4 4.25 ]; den=[1 -0.1 -1.5];
sys = tf(num,den);
rlocus(sys), hold on
xc=[-1:0.1:1];c=sqrt(1-xc.^2);
plot(xc,c,':',xc,-c,':')
hold off
%
% Part (b)
rlocfind(sys)
rlocfind(sys)

ÈSelect a point in the graphics window
selected_point =
-0.8278 + 0.5202i
ans =
0.7444

Kmax

Select a point in the graphics window
selected_point =
-0.9745 - 0.0072i
ans =
0.3481

Kmin

FIGURE CP13.6
CONTINUED: Using the rlocus and rlocfind functions.

Using root locus methods, we determine that an acceptable compensator
is
Gc (s) = 11.7

s+6
.
s + 20

With a zero-order hold and T = 0.02 sec, we find that

1.2

1
*

*

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* *

*
*
*

0.8

Amplitude

CP13.7

*
*
*

0.6
*
*

0.4
*
*

0.2
*
*

0* *
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Time (sec)

FIGURE CP13.7
System step response (* denotes sampled-data response).

0.8

0.9

1

from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying,
recording, or likewise. For information regarding permission(s), write to: Rights and Permissions Department, Pearson Education, Inc., Upper Saddle River, NJ 07458.

749

Computer Problems

D(z) =

11.7z − 10.54
.
z − 0.6703

The closed-loop step response is shown in Figure CP13.7.



Source Exif Data:
File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
Linearized                      : Yes
Page Mode                       : UseOutlines
XMP Toolkit                     : 3.1-701
Create Date                     : 2010:08:05 08:48:05-05:00
Modify Date                     : 2010:10:25 08:50:47+02:00
Producer                        : GPL Ghostscript 8.70
Format                          : application/pdf
Title                           : Instructor's Solutions Manual for for Modern Control Systems, 12th Ed
Creator                         : Richard C. Dorf, Robert H. Bishop
Subject                         : 9780136024989
Document ID                     : uuid:107e56d4-5005-604d-a266-e210e0c618a9