D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Background
▪ Educational Background (Mechanical Engineering)
➢ B.Eng. with first class honors from NTU (2006 – 2010)
➢ Dual Ph.Ds. from NTU and CMU (2011 – 2016)
➢ Post-Doctoral Researcher at MPI (2016 – 2017)
CMU – Carnegie Mellon University (USA), MPI – Max Planck Institute for Intelligent Systems (Germany)

▪ Contact Details
➢ Email: [email protected]
➢ Office: N3.2-01-22

2
 Teaching Plan (Part I)
LAMS and Live Short Review Lectures
Week Lecture Topics Tutorial
1 Introduction to Systems and Laplace Transformation
2 Mathematical Modelling of Dynamic Systems 1
3 Mathematical Modelling of Mechanical Systems 2
4 Response and Stability 3
5 First-Order and Second-Order Systems 4
6 Transient Response 5
7 Basic Control Actions (Quiz Week)

3
 Teaching Plan (Part II)

Live Lectures
Week Lecture Topics Tutorial
8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

4
 Assessment
Quiz (20%)
• Week 7 during lecture hours.
• Hint: Simpler than final exams – Don’t miss it!

Homework (20%)
• Released in Week 10 and two weeks to complete

Final Exam (60%)

• Restricted Open Book: 1 double-sided A4 cheatsheet (typed or hand-written)
• IMPORTANT: No formula sheet attached as appendix

5
 Suggestions

• Read the PDF lecture notes

• Watch the LAMS video and do their activities (Week 1 to 7)
• Attend the lecture sessions
• Please ask the questions at the end of the lecture
• Attempt the tutorial questions before attending the tutorials

6
 Motivation
1. A method to unify sub-systems across different disciplines

2. To regulate the output with respect to time t

For example:
Desired Electrical System Mechanical System Actual
Velocity Voltage Torque Velocity
Controller

Sensors

7
 Laplace Transformation
Converting differential equations (t) algebraic equations (s)

Some Basic Transform s-shifting

1
ℒ 1 = 1
𝑠 ℒ −1 = 𝑒 −𝑎𝑡
𝑠 𝑠+𝑎
ℒ cos 𝜔𝑡 = 2
𝑠 + 𝜔2
𝜔 𝜔
ℒ sin 𝜔𝑡 = 2 ℒ −1 = 𝑒 −𝑎𝑡 sin 𝜔𝑡
𝑠 + 𝜔2 𝑠 + 𝑎 2 + 𝜔2

𝑑𝑓 0
ℒ 𝑓(𝑡) = 𝐹(𝑠), ℒ = 𝑠𝐹 𝑠 − 𝑓(𝑡 = 0)
𝑑𝑡

8
 Physical System

R(s) + E(s) C(s)

Gc(s) P(s)
- Controller Plant

H(s)
Sensor

Software
Programming

9
Microcontroller
 Example (Modelling)
Derive equation of motion
F Using f = ma,
C1 𝐹 − 𝐶1 𝑥ሶ + 𝐾1 𝑥 = 𝑀𝑥ሷ
M 𝐹 = 𝑀𝑥ሷ + 𝐶1 𝑥ሶ + 𝐾1 𝑥
K1 Laplace Transform
C1 x
𝐹(𝑠) = 𝑀𝑠 2 𝑋 + 𝐶1 𝑠𝑋 + 𝐾1 𝑋
𝐶1 𝑥ሶ 𝐶1 𝑥ሶ
𝐹 𝑠 = 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 𝑋

K1x K1x 𝑋(𝑠) 1

K1 = = 𝐺(𝑠)
𝐹(𝑠) 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1
F
𝐶1 𝑥ሶ + 𝐾1 𝑥
M
F(s)
G(s)
𝑥ሷ X(s)
10
10
 Example (Solving I)
Ns 𝑁
1 Assume that: 𝑀 = 1 kg, 𝐶1 = 7 , 𝐾1 = 6 .
m 𝑚
F(s) 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 X(s) Solve 𝑥(𝑡) when 𝑓 𝑡 = 1 N.

1
Input in s-domain: 𝐹(𝑠) =
𝑠

Solving 𝒙(𝒕)
1 1 1
𝑋 𝑠 = 𝐹(𝑠) 𝑋(𝑠) = 2
𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 𝑠 + 7𝑠 + 6 𝑠

1
𝑋(𝑠) =
𝑠 𝑠 + 1 [𝑠 + 6]

11
11
 Example (Solving II)
Ns 𝑁
1 Assume that: 𝑀 = 1 kg, 𝐶1 = 7 , 𝐾1 = 6 .
m 𝑚
F(s) 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 X(s) Solve 𝑥(𝑡) when 𝑓 𝑡 = 1 N.

1 A B C
Partial Fraction: 𝑋(𝑠) = 𝑠 𝑠 + 1 [𝑠 + 6] → 𝑋 s = s + s + 1 + s + 6

Comparing numerator: 𝐴 𝑠 + 1 𝑠 + 6 + 𝐵𝑠 𝑠 + 6 + 𝐶𝑠(𝑠 + 1) = 1

1
Sub s = 0: 𝐴 1 6 + 0 = 1 → 𝐴 =
6
1 1/6 −1/5 1/30
Sub s = -1: 𝐵 −1 5 =1→𝐵=−
5 𝑋 s = + +
s s+1 s+6
1
Sub s = -6: 𝐶 −6 −5 = 1 → 𝐶 =
30
1 1 −𝑡 1 −6𝑡
Laplace Inverse: 𝑥 𝑡 = ℒ −1 𝑋 s = − 𝑒 + 𝑒
6 5 30

Particular Solution Homogeneous Solution

12
12
 Example (Solving II)
Ns 𝑁
1 Assume that: 𝑀 = 1 kg, 𝐶1 = 7 , 𝐾1 = 6 .
m 𝑚
F(s) 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 X(s) Solve 𝑥(𝑡) when 𝑓 𝑡 = 1 N.

1 A B C
Partial Fraction: 𝑋(𝑠) = 𝑠 𝑠 + 1 [𝑠 + 6] → 𝑋 s = s + s + 1 + s + 6

Comparing numerator: 𝐴 𝑠 + 1 𝑠 + 6 + 𝐵𝑠 𝑠 + 6 + 𝐶𝑠(𝑠 + 1) = 1

1
Sub s = 0: 𝐴 1 6 + 0 = 1 → 𝐴 =
6
1 1/6 −1/5 1/30
Sub s = -1: 𝐵 −1 5 =1→𝐵=−
5 𝑋 s = + +
s s+1 s+6
1
Sub s = -6: 𝐶 −6 −5 = 1 → 𝐶 =
30
1 1 −𝑡 1 −6𝑡
Laplace Inverse: 𝑥 𝑡 = ℒ −1 𝑋 s = − 𝑒 + 𝑒
6 5 30

Steady-state Transient Response

response 13
13
 Final Value Theorem (FVT)
Ns 𝑁
Assume that: 𝑀 = 1 kg, 𝐶1 = 7 , 𝐾1 = 6 .
m 𝑚
1
Solve 𝑥𝑠𝑠 when 𝑓 𝑡 = 1 N.
F(s) 𝑀𝑠 2 + 𝐶1 𝑠 + 𝐾1 X(s)
1
Given: 𝑋(𝑠) = 𝑠 𝑠+1 [𝑠+6]

Recomputing steady-state 𝒙 (𝒙𝒔𝒔 ) with FVT

𝑥𝑠𝑠 = lim 𝑥(𝑡) = lim 𝑠𝑋(𝑠)
𝑡→∞ 𝑠→0

1
= lim𝑠
𝑠→0 𝑠 𝑠 + 1 [𝑠 + 6]

1 1
= lim =
𝑠→0 1 [6] 6

Valid only when the system is stable

14
14
Position Control System
Physical System
F
C1

K1 Displacement
x sensor

Block Diagrams
Xd(s) + E(s) F(s) X(s)
-
Gc (s) G(s)
Controller Plant

H(s)
Sensor 15
 Summary

1. Introduction and motivation of learning control theory

2. Laplace transformation
▪ Partial fraction and Laplace Inverse
▪ Final value theorem

16
 Acknowledgement

Professor Cuong Professor Tegoeh

17
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Assistant Professor Guo Zhan Lum

1
 Teaching Plan (Part I)
LAMS and Live Short Review Lectures
Week Lecture Topics Tutorial
1 Introduction to Systems and Laplace Transformation
2 Mathematical Modelling of Dynamic Systems 1
3 Mathematical Modelling of Mechanical Systems 2
4 Response and Stability 3
5 First-Order and Second-Order Systems 4
6 Transient Response 5
7 Basic Control Actions (Quiz Week)

2
 Block Diagrams

Control Signal

Transfer Function

Objectives
• Understand the relationship between control signals and transfer functions
• Simplify the block diagrams
• Establish the relationship between the input and output control signals

3
3
 General Rules
1. Multiply

𝐺1 𝐺2 𝐺1 𝐺2

𝐴 𝐴𝐺1 𝐺2
𝐺1 𝐺2

2. Summation

4
4
 General Rules
3. Pre-Multiply (G) -> Post-Multiply (G)
AH AH
𝐻 𝐻/G
AG A
𝐺 𝐺
A AG

A A+B G(A+B) A AG G(A+B)

G G

B
BG
G
B
5
5
 General Rules
4. Post-Multiply (G) -> Pre-Multiply (G)

AGH AGH
𝐻 𝐺𝐻

𝐺 𝐺
A AG A AG

A+B/G
A AG+B
A AG AG+B G
G
B/G
B
1/G
B

6
6
 General Rules
5. Closed-loop R – CH E(G)

R(s) E(s) C(s)

G(s)
CH
H(s)
EG = C
(R – CH)G = C, RG = C(1+HG)
𝐶 𝐺
=
𝑅 1 + 𝐺𝐻
7
7
 General Rules
5. Closed-loop R – CH E(G)

R(s) E(s) C(s)

G(s)
CH
H(s)

R(s) 𝐺 C(s)
1 + 𝐺𝐻
8
8
 Example I
Aim: Establish the relationship between Y(s) and X(s)

X A AX
𝑌 = 𝐴𝑋 + 𝐵𝑋 + 𝐶𝑋
X X BX Y
B + 𝑌 = 𝐴+𝐵+𝐶 𝑋

CX 𝑌
X C =𝐴+𝐵+𝐶
𝑋

X(s) Y(s)
𝐴+𝐵+𝐶

9
9
 Example II (PYP April 2014)
Question 3B: Simplify the system and find C/R (7 marks)

R C
+
- G1 +
- G2 G3

H1

H2

R C
+
- +
- G1 G2 G3

H1
𝐻2
𝐺1

10
10
 Example II (PYP April 2014)
Question 3B: Simplify the system and find C/R (7 marks)

R C
+
- +
- G1 G2 G3

H1
𝐻2
𝐺1

R C
+
- +
- G1 G2 G3

H1
𝐻2
𝐺1

11
11
 Example II (PYP April 2014)
Question 3B: Simplify the system and find C/R (7 marks)

R C
+
- +
- G1 G2 G3

H1
𝐻2
𝐺1

R 𝐺1 𝐺2 C
+
- G3
1+𝐺1 𝐺2 𝐻1
𝐶 𝐺
G =
𝑅 1 + 𝐺𝐻
𝐻2
𝐺1
12
H 12
 Example III (PYP April 2015)
Question 3A: Simplify the system and find C/R (6 marks)

R - C
+
- 𝐺1 + G2 +
- 𝐺3 G4

𝐻1 𝐻2

1
𝐺4
R - C
+
- 𝐺1 + G2 +
- 𝐺3 G4

𝐻1 𝐻2

13
13
 Example III (PYP April 2015)
Question 3A: Simplify the system and find C/R (6 marks)
1
𝐺4
R - 𝐺3 𝐺4
+
- 𝐺1 + G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

1
𝐺4
R - C
+
- 𝐺1 + G2 +
- 𝐺3 G4

𝐻1 𝐻2

14
14
 Example III (PYP April 2015)
Question 3A: Simplify the system and find C/R (6 marks)
1
𝐺4
R - 𝐺3 𝐺4
+
- 𝐺1 + G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

1
𝐺1 𝐺4
R - 𝐺3 𝐺4
+
- + 𝐺1 G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

15
15
 Example III (PYP April 2015)
Question 3A: Simplify the system and find C/R (6 marks)
1
𝐺1 𝐺4
R - 𝐺3 𝐺4
+ +
- 𝐺1 G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

1
𝐺1 𝐺4
R - 𝐺3 𝐺4
+
- + 𝐺1 G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

16
16
 Example III (PYP April 2015)
Question 3A: Simplify the system and find C/R (6 marks)
1
𝐺1 𝐺4
R - 𝐺3 𝐺4
+ +
- 𝐺1 G2
1 + 𝐺3 𝐺4 𝐻2 C
𝐻1

H 1
𝐺1 𝐺4

R - 𝐺1 𝐺2 𝐺3 𝐺4 𝐶 𝐺
+
1 + 𝐺1 𝐺2 𝐻1 1 + 𝐺3 𝐺4 𝐻2 C =
𝑅 1 + 𝐺𝐻
G
17
17
 Summary

▪ Analysis of block diagrams

❖ Manipulate control signals and transfer functions
❖ Establish basic rules
❖ Simplify the block diagrams

▪ Concepts are demonstrated via three examples

18
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Assistant Professor Guo Zhan Lum

1
 Teaching Plan (Part I)
LAMS and Live Short Review Lectures
Week Lecture Topics Tutorial
1 Introduction to Systems and Laplace Transformation
2 Mathematical Modelling of Dynamic Systems 1
3 Mathematical Modelling of Mechanical Systems 2
4 Response and Stability 3
5 First-Order and Second-Order Systems 4
6 Transient Response 5
7 Basic Control Actions (Quiz Week)

2
 Mechanical Systems
PYP Dec 2015 Qn 2B: Find Y/X (12 marks)
𝑧

Y(s)
G(s) X(s)
General Steps
1. Separate the springs, dampers and mass
2. Analyze forces on springs and dampers
3. Use free-body diagrams to write down the equation of motion
4. Laplace Transform
3
3
 Steps 1–2
1. Separate the springs, dampers and mass
2. Analyze forces on springs and dampers

𝐾1𝑧 K1 𝐾1𝑧 𝐾2(𝑥 − 𝑧) K2 𝐾2(𝑥 − 𝑧)

𝐾0(𝑥 − 𝑦)
𝐵1𝑧ሶ 𝐵1𝑧ሶ 𝐵2𝑥ሶ 𝐵2𝑥ሶ
𝐾0(𝑥 − 𝑦)
B2 K0
B1
𝑧 𝑥
𝐾1𝑧 + 𝐵1𝑧ሶ 𝐾2 (𝑥 − 𝑧) 𝐾0(𝑥 − 𝑦)
𝑀1 𝑀2
𝐾2 (𝑥 − 𝑧)
𝑧ሷ + 𝐵2𝑥ሶ
𝑥ሷ
4
4
 Steps 3-4 (Mass 1)
3. Use free-body diagrams to write down the equation of motion
𝑧 𝑥
𝐾1𝑧 + 𝐵1𝑧ሶ 𝐾2 (𝑥 − 𝑧) 𝐾0(𝑥 − 𝑦)
𝑀1 𝑀2
𝐾2 𝑥 − 𝑧
𝑧ሷ + 𝐵2𝑥ሶ 𝑥ሷ
Using 𝑓 = 𝑚𝑎 on M1: 𝐾2 𝑥 − 𝑧 − 𝐾1𝑧 − 𝐵1𝑧ሶ = 𝑀1 𝑧ሷ

𝐾2𝑥 = 𝑀1 𝑧ሷ + 𝐵1𝑧ሶ + 𝐾1 + 𝐾2 𝑧

4. Laplace Transform: 𝐾2𝑋 = 𝑀1 𝑠 2 𝑍 + 𝐵1𝑠𝑍 + 𝐾1 + 𝐾2 𝑍

𝐾2𝑋 = (𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2 )𝑍

𝐾2
𝑍= 𝑋 −−− −(1)
𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2

5
5
 Steps 3-4 (Mass 2)
3. Use free-body diagrams to write down the equation of motion
𝑧 𝑥
𝐾1𝑧 + 𝐵1𝑧ሶ 𝐾2 (𝑥 − 𝑧) 𝐾0(𝑥 − 𝑦)
𝑀1 𝑀2
𝐾2 𝑥 − 𝑧
𝑧ሷ + 𝐵2𝑥ሶ 𝑥ሷ
𝐾2
𝑍= 𝑋 −−− −(1)
𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2

Using 𝑓 = 𝑚𝑎 on M2: −𝐾2 𝑥 − 𝑧 − 𝐵2𝑥ሶ − 𝐾0 (𝑥 − 𝑦) = 𝑀2 𝑥ሷ

𝐾2𝑧 + 𝐾0 𝑦 = 𝑀2 𝑥ሷ + 𝐵2𝑥ሶ + 𝐾0 + 𝐾2 𝑥

4. Laplace Transform: 𝐾2𝑍 + 𝐾0 𝑌 = 𝑀2 𝑠 2 𝑋 + 𝐵2𝑠𝑋 + 𝐾0 + 𝐾2 𝑋

𝐾2𝑍 + 𝐾0 𝑌 = 𝑀2 𝑠 2 + 𝐵2𝑠 + 𝐾0 + 𝐾2 𝑋 −−− −(2)

Substitute Eq. (1) into (2)

6
6
 Steps 3-4 (Solving X/Y)
3. Use free-body diagrams to write down the equation of motion
4. Laplace Transform
𝑧 𝑥
𝐾1𝑧 + 𝐵1𝑧ሶ 𝐾2 (𝑥 − 𝑧) + 𝐵2𝑥ሶ 𝐾0(𝑥 − 𝑦)
𝑀1 𝑀2
𝐾2 𝑥 − 𝑧
𝑧ሷ + 𝐵2𝑥ሶ 𝑥ሷ
𝐾2 2+𝐵 𝑠+ 𝐾 +𝐾
𝐾2 𝑋 + 𝐾0 𝑌 = 𝑀2 𝑠 2 0 2 𝑋
𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2

𝐾22
𝐾0 𝑌 = 𝑀2 𝑠 2 + 𝐵2𝑠 + 𝐾0 + 𝐾2 − 𝑋
𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2

𝑋 𝐾0
= = 𝐺(𝑠)
𝑌 𝐾22
𝑀2 𝑠 2 + 𝐵2𝑠 + 𝐾0 + 𝐾2 −
𝑀1 𝑠 2 + 𝐵1𝑠 + 𝐾1 + 𝐾2

Y(s)
G(s) X(s) 7
7
 Linear & Rotary Systems
Springs Linear Rotary
𝜃1 𝜃2
𝐾𝐿 𝜃1 𝐾𝑅 𝜃2
Scenario 1
𝜃2 > 𝜃1
𝐾𝐿 (𝜃2 −𝜃1 ) 𝐾𝐿 (𝜃2 −𝜃1 ) 𝐾𝑅 (𝜃2 −𝜃1 ) 𝐾𝑅 (𝜃2 −𝜃1 )

𝜃1 𝜃2
𝐾𝐿 𝜃1 𝐾𝑅 𝜃2
Scenario 2
𝜃2 < 𝜃1
𝐾𝐿 (𝜃1 −𝜃2 ) 𝐾𝐿 (𝜃1 −𝜃2 ) 𝐾𝑅 (𝜃1 −𝜃2 ) 𝐾𝑅 (𝜃1 −𝜃2 )

8
 Example (PYP Aug 2019)
Question 2 (modified): Find the transfer function of the system
Given: Damping coefficient of air = C, mass = m, initial velocity = 𝑣𝑜

𝐶𝑣

𝑣 𝐶𝑣
𝑚𝑔 𝐶𝑣
𝑣ሶ
𝑚𝑔
Steps
1. Separate the springs, dampers and mass
2. Analyze forces on springs and dampers

9
9
 Example (PYP Aug 2019)
Question 2 (modified): Find the transfer function of the system
Given: Damping coefficient of air = C, mass = m, initial velocity = 𝑣𝑜

Step 3: Use free-body diagrams to write down the equation of motion

𝐶𝑣 Using 𝑓 = 𝑚𝑎: 𝑓 − 𝐶𝑣 = 𝑚𝑣ሶ

𝑓 = 𝑚𝑣ሶ + 𝐶𝑣
𝑣ሶ
Step 4: Laplace Transform: 𝐹(𝑠) = 𝑚(𝑠𝑉 − 𝑣0 ) + 𝐶𝑉
𝑚𝑔 = 𝑓 𝑉 1
𝐹 𝑠 + 𝑚𝑣0 = 𝑚𝑠 + 𝐶 𝑉 → = = 𝐺(𝑠)
𝐹 𝑠 + 𝑚𝑣0 𝑚𝑠 + 𝐶

𝑚𝑔
𝐹= 1 𝑉
𝑠
𝑚𝑠 + 𝐶
𝑚𝑣0 10
10
 Summary

Deriving the transfer function of mechanical systems

1. Separate the springs, dampers and mass

2. Analyze forces on springs and dampers
3. Use free-body diagrams to write down the equation of
motion
4. Laplace Transform

11
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Assistant Professor Guo Zhan Lum

1
 Teaching Plan (Part I)
LAMS and Live Short Review Lectures
Week Lecture Topics Tutorial
1 Introduction to Systems and Laplace Transformation
2 Mathematical Modelling of Dynamic Systems 1
3 Mathematical Modelling of Mechanical Systems 2
4 Response and Stability 3
5 First-Order and Second-Order Systems 4
6 Transient Response 5
7 Basic Control Actions (Quiz Week)

2
 Poles and Stability
𝑁 𝑠
Transfer function: 𝐺(𝑠) =
𝐷(𝑠)

G(s) Characteristics Equation (CE): 𝐷 𝑠 = 0

Y(s) X(s)
Poles: All ‘s’ that satisfy CE

Objectives
1. Understand what are poles
2. The relationship between poles and stability of a transfer function
3. Routh–Hurwitz stability criterion

3
3
 Transfer Function Poles
Differential Equation Laplace Transformation
𝑀𝑥ሷ + 𝐶 𝑥ሶ + 𝐾𝑥 = 𝑓 𝑀𝑠 2 + 𝐶𝑠 + 𝐾 𝑋(𝑠) = 𝐹(𝑠)
𝑋(𝑠) 1
Homogeneous solution: 𝑀 𝑥ሷ + 𝐶 𝑥ሶ + 𝐾𝑥 = 0 =
𝐹(𝑠) 𝑀𝑠 2 + 𝐶𝑠 + 𝐾
𝑀𝑑 2 + 𝐶𝑑 + 𝐾 = 0 Transfer Function Poles: 𝑀𝑠 2 + 𝐶𝑠 + 𝐾 = 0

1st possibility 2nd possibility 1st possibility 2nd possibility

𝑑 = −1, 2 𝑑 = ±𝑗5 𝑠 = −1, 2 𝑠 = ±𝑗5
𝑥= 𝐴e−𝑡 + 𝐵e2𝑡 𝑥 = 𝐴cos(5𝑡 + ∅) 𝑥 = 𝐴e−𝑡 + 𝐵e2𝑡 𝑥 = 𝐴cos(5𝑡 + ∅)

3rd possibility 3rd possibility

𝑑 = −2 ± 𝑗5 𝑠 = −2 ± 𝑗5
𝑥 = 𝐴e−2𝑡 cos 5𝑡 + ∅ 𝑥 = 𝐴e−2𝑡 cos 5𝑡 + ∅

Poles dictate homogeneous 4

solution/ transient response
 Stable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 + 3𝑠 + 2

Transfer Function Poles: 𝑠 2 + 3𝑠 + 2 = 0 𝑠+1 𝑠+2 =0 𝑠 = −1, −2

Im Transient Response Step Response

x(t)
𝐴𝑒 −𝑡 + 𝐵𝑒 −2𝑡

-2 -1 Re

𝑠 = −1 𝑠 = −2 Time
Pole #1 Pole #2
Transient response decays

s-plane 𝐴, 𝐵 are constants Stable system

5
Partially-Stable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 +4

Transfer Function Poles: 𝑠2 + 4 = 0 𝑠 = ±𝑗2

Im Transient Response Step Response

x(t)
2
𝐴 cos(2𝑡 + ∅)

Re
-2
𝑠 = ±𝑗2 Time
Complex Poles Transient response oscillates

s-plane 𝐴, ∅ are constants Partially-stable system

6
 Unstable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 +𝑠−2

Transfer Function Poles: 𝑠2 + 𝑠 − 2 = 0 𝑠−1 𝑠+2 =0 𝑠 = 1, −2

Im Transient Response Step Response

x(t)
𝐴𝑒 𝑡 + 𝐵𝑒 −2𝑡

-2 1 Re

𝑠=1 𝑠 = −2 Time
Pole #1 Pole #2
Transient response → ∞
s-plane 𝐴, 𝐵 are constants unstable system

7
 Stability
Three types of poles
1. Real poles (e.g. s = a) Stable systems require
Transient response 𝐴𝑒 𝑎𝑡 𝑒 𝑎𝑡 → 0, when 𝑡 → ∞
𝑎<0
2. Pure complex poles (e.g. s = ±jb)
Left hand s-plane
Transient response 𝐴 cos(𝑏𝑡 + ∅) Im
(always partially-stable)

Stable Unstable
3. Complex poles with real parts (e.g. s = a ± jb)
Re
Transient response 𝐴𝑒 𝑎𝑡 cos(𝑏𝑡 + ∅)

𝐴, ∅ are constants
s-plane 8
 Stability

Routh–Hurwitz stability criterion

This pole has a +ve real part
1st possibility
𝑠 = −1, 2

▪ A simple way to check number of poles

𝑥 = 𝐴e−𝑡 + Be2t

that have positive real parts

→ ∞, unstable

9
9
 Example
Characteristics equation: 𝑠 4 + 8𝑠 3 + 322 + 80𝑠 + 100 = 0

Determine number of roots with positive real parts:

Using Routh–Hurwitz stability criterion

s4: 1 32 100

s3: 8 80

s2: 8 × 32 − 1 × 80 8 × 100
8 8
s1:

s0:

10
10
 Example
Characteristics equation: 𝑠 4 + 8𝑠 3 + 322 + 80𝑠 + 100 = 0

Determine number of roots with positive real parts:

Using Routh–Hurwitz stability criterion

s4: 1 32 100

s3: 8 80

s2: 22 100

s1: 22 × 80 − 8 × 100
22
s0:

11
11
 Example
Characteristics equation: 𝑠 4 + 8𝑠 3 + 322 + 80𝑠 + 100 = 0

Determine number of roots with positive real parts:

Using Routh–Hurwitz stability criterion

s4: 1 32 100

s3: 8 80

s2: 22 100

s1: 43.6

s0: 43.6 × 100

= 100
43.6
There is no change in sign for the first column, -> there are no roots with
positive real parts
12
12
 Special Case I
One of the elements in the first numerical column is 0

s3: 1 5

s2: 1 5

s1: 0 𝜀
s0: 5𝜀 − 1 0
=5
𝜀 NB: 𝜀 is small positive number

Check for sign changes in first numerical column

Important: System is either partially stable or unstable

13
13
 Special Case II
Entire row of elements are 0

s5: XX XX XX

s4: 1 5 20

s3: 0 0

s2:

Auxiliary polynomial: 𝑃 𝑠 = 𝑠 4 + 5𝑠 2 + 20
d𝑃 𝑠
= 4𝑠 3 + 10𝑠
d𝑠
d𝑃 𝑠
Use the coefficients of to replace the 0 elements
d𝑠

14
14
 Special Case II
Entire row of elements are 0

s5: XX XX XX

s4: 1 5 20

s3: 4 10

s2:

Auxiliary polynomial: 𝑃 𝑠 = 𝑠 4 + 5𝑠 2 + 20
d𝑃 𝑠
= 4𝑠 3 + 10𝑠
d𝑠
d𝑃 𝑠
Use the coefficients of to replace the 0 elements
d𝑠

15
15
 Special Case II
Entire row of elements are 0

s5: XX XX XX

s4: 1 5 20

s3: 4 10

s2: 2.5 20

Continue as per normal and check for sign changes in first numerical column

Important: System is either partially stable or unstable

16
16
 PYP Dec 2017 Qn 1

R(s) + 1 C(s) 𝐶 𝐺
𝐺CL = =
- 𝑠 3 + 𝑠 2 + 2𝑠 + 23 𝑅 1+𝐺
FLTF: G(s) 1
= 3
𝑠 + 𝑠 2 + 2𝑠 + 24

Characteristic Equation (CE): 𝑠 3 + 𝑠 2 + 2𝑠 + 24 = 0

17
17
 PYP Dec 2017 Qn 1
Characteristic Equation (CE): 𝑠 3 + 𝑠 2 + 2𝑠 + 24 = 0

Using Routh–Hurwitz stability criterion

s3: 1 2

s2: 1 24

s1: 1 × 2 − 1 × 24 = −22
1
s0: 24

Two sign changes -> Two roots with positive real parts -> Two unstable poles

18
18
 Summary

▪ Understanding poles of transfer function

▪ Relationship between poles and stability
▪ Routh-Hurwitz stability criterion

19
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Assistant Professor Guo Zhan Lum

1
 Teaching Plan (Part I)
LAMS and Live Short Review Lectures
Week Lecture Topics Tutorial
1 Introduction to Systems and Laplace Transformation
2 Mathematical Modelling of Dynamic Systems 1
3 Mathematical Modelling of Mechanical Systems 2
4 Response and Stability 3
5 First-Order and Second-Order Systems 4
6 Transient Response 5
7 Basic Control Actions (Quiz Week)

2
 Transfer Function
𝑠+3
R(s) 𝑠 3 + 2𝑠 2 + 5𝑠 + 10 C(s)

System Order: Highest power of ‘s’ in the denominator (e.g., 3rd-order system)
Poles of transfer function: Dictate transient response
DC gain of transfer function: Dictate steady-state response (step inputs)

Objectives
1. First order, second order and higher order systems
2. Step, impulse, ramp and parabolic responses
3. Pole-zero cancellation

3
3
 First Order Systems
𝑘 Find: c(t) when r(t) is a step input with magnitude p
𝑝
R(s) 𝜏𝑠 + 1 C(s) 𝑟 𝑡 = 𝑝𝑢(𝑡) 𝑅 𝑠 =
𝑠

Solving: c(t) when r(t) is a step input with magnitude p

𝑝 𝑘 𝐴 𝐵
𝐶(𝑠) = = +
𝑠 𝜏𝑠+1 𝑠 𝜏𝑠 + 1

Steady-state output Transient output

𝐴 and 𝐵 are residues

4
 First Order Systems
𝑘 Find: c(t) when r(t) is a step input with magnitude p
𝑝
R(s) 𝜏𝑠 + 1 C(s) 𝑟 𝑡 = 𝑝𝑢(𝑡) 𝑅 𝑠 =
𝑠

Solving: c(t) when r(t) is a step input with magnitude p

𝑝 𝑘 𝐴 𝐵
𝐶(𝑠) = = +
𝑠 𝜏𝑠+1 𝑠 𝜏𝑠 + 1
𝑝𝑘 𝑝𝑘𝜏
= −
𝑠 𝜏𝑠 + 1
𝑡
−𝜏
𝑐 𝑡 = 𝑝𝑘 − 𝑝𝑘e

Steady-state output
Transient output
5
 First Order Systems
𝑘 Find: c(t) when r(t) is a step input with magnitude p
𝑝
R(s) 𝜏𝑠 + 1 C(s) 𝑟 𝑡 = 𝑝𝑢(𝑡) 𝑅 𝑠 =
𝑠
𝑡
−
Solution: 𝑐 𝑡 = 𝑝𝑘 − 𝑝𝑘e 𝜏

e.g., p = k = 1 Steady-state
value

2% settling time: 𝑡𝑠 = 4𝜏
2% settling
time

6
 Key Parameters
𝑘 Find: r(t) when c(t) is a step input with magnitude p

R(s) 𝜏𝑠 + 1 C(s) Solution: 𝑐 𝑡 = 𝑝𝑘 − 𝑝𝑘e

𝑡
−𝜏

𝐺(𝑠)
𝑘
DC gain of 𝐺 𝑠 = lim =𝑘
𝑠→0 𝜏𝑠+1

𝑐𝑠𝑠 = 𝑝𝑘 (product of DC gain and magnitude of step)

Im
1
Poles of 𝐺 𝑠 : 𝑠 = − 𝜏
𝑡 1 Re
−𝜏 −
→ 𝑐𝑡𝑟 𝑡 = 𝐴𝑒 𝜏

𝑐 𝑡 = 𝑐𝑠𝑠 + 𝑐𝑡𝑟
7
s-plane
 Second Order Systems
𝑘𝜔𝑛2
Find: 𝑐 𝑡 when r(t) is a step
R(s) 𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2 C(s) input with magnitude p

𝐺(𝑠)
2
𝑘𝜔𝑛
DC gain of 𝐺 𝑠 = lim 2 =𝑘
𝑠→0 𝑠 2 +2𝜁𝜔𝑛 𝑠+𝜔𝑛

𝑐𝑠𝑠 = 𝑝𝑘 (product of DC gain and magnitude of step)

Poles of 𝐺 𝑠 : 𝑠 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1

 > 1 (overdamped)  = 1 (critically damped)  < 1 (Underdamped)

𝑠 = 𝑎, 𝑏 𝑠 = −𝜁𝜔𝑛 𝑠 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2

(repeated pole)
8
 Overdamped Systems
𝑘𝜔𝑛2 Find: 𝑐 𝑡 when r(t) is a step
input with magnitude p
R(s) 𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2 C(s)
𝑐𝑠𝑠 = 𝑝𝑘,  > 1
𝐺(𝑠)

Poles: 𝑠 = 𝑎, 𝑏 → 𝑐𝑡𝑟 = 𝐴𝑒 𝑎𝑡 + 𝐵𝑒 𝑏𝑡

Im
Two first order poles

b a Re
Pole ‘a’ is dominant
because it decays slower
than pole ‘b’
s-plane

𝑐 𝑡 = 𝑐𝑠𝑠 + 𝑐𝑡𝑟
e.g., p = k = 1 9
 Critically Damped Systems
𝑘𝜔𝑛2 Find: 𝑐 𝑡 when r(t) is a step
input with magnitude p
R(s) 𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2 C(s)
𝑐𝑠𝑠 = 𝑝𝑘,  = 1
𝐺(𝑠)

Poles: 𝑠 = −𝜁𝜔𝑛 → 𝑐𝑡𝑟 = 𝐴𝑡𝑒 −𝜁𝜔𝑛 𝑡

Im
Repeated pole

Re
−𝜁𝜔𝑛
Critically damped systems
are faster than
overdamped systems
s-plane

𝑐 𝑡 = 𝑐𝑠𝑠 + 𝑐𝑡𝑟
e.g., p = k = 1
10
 Underdamped Systems
𝑘𝜔𝑛2 Poles
𝑠 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2 ,
R(s) 𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2 C(s) Transient
𝐴𝑒 −𝜁𝜔𝑛 𝑡 cos(𝜔𝑛 1 − 𝜁 2 𝑡 + ∅)

Step Response (p = k = 1)
Im 𝜋𝜁
−
%OS = 𝑒 1−𝜁 2
𝜁𝜔𝑛
𝜔𝑛
𝜔𝑛 1 − 𝜁2 𝛽
Re
4
𝑡𝑠 = steady-state
𝜁𝜔𝑛
cos 𝛽 = 𝜁 value

s-plane
11
𝑐 𝑡 = 𝑐𝑠𝑠 + 𝑐𝑡𝑟
 Higher Order Systems

• Higher order systems can be decomposed into multiple first and second
order systems

• Expect some deviations from first and second order systems

• Transient response will be dominated by slowest poles (dominant poles)

• Steady-state value is equal to the product of DC gain of transfer function and

magnitude of input step

12
 Output Responses
Unit Impulse Unit Step Unit Ramp Unit Parabolic

𝑡2
𝑟 𝑡 = 𝛿(𝑡) 𝑟 𝑡 = 𝑢(𝑡) 𝑟 𝑡 = 𝑡. 𝑢(𝑡) 𝑟 𝑡 = . 𝑢(𝑡)
2
r r r r
t-domain

t t t t

1 1 1
s-domain 1
𝑠 𝑠2 𝑠3
𝑡 𝑡 𝑡
d𝑐 𝑡
න 𝑐(𝑡) d𝑡 න න 𝑐(𝑡) d𝑡 d𝑡
Output d𝑡 𝑐(𝑡) 0 0 0

13
 Example (I)
𝑅(𝑠) 𝐸(𝑠) 1 𝐶(𝑠)
𝐴 Find K, A such that the unit step response:
𝑠+𝐵
2% settling time = 8 s, ess = 0.1

𝐴 2% settling time criterion

𝐶 𝐴
= 𝐺𝐶𝐿 = 𝑠 + 𝐵 =
𝑅 𝐴 𝑠+𝐵+𝐴 𝑡𝑠 = 4𝜏 = 8 → 𝜏=2
1+𝑠+𝐵

𝐴 Steady-state error criterion

𝐴+𝐵 𝑘 𝑒𝑠𝑠 = 𝑟 − 𝑐𝑠𝑠 = 0.1
= =
1 𝜏𝑠 + 1
𝐴+𝐵 𝑠+1 𝑐𝑠𝑠 = 0.9 = 𝑘

1
Time constant of 𝐺𝐶𝐿 : Comparison
𝐴+𝐵
1 𝐴
DC gain of 𝐺𝐶𝐿 :
𝐴 =2 = 0.9
𝐴+𝐵 𝐴+𝐵 𝐴+𝐵

𝐴 = 0.45, 𝐵 = 0.05 14
14
 Example (II)
0.9 Aim: Find unit impulse and ramp response
R(s) 2𝑠 + 1 C(s) −
𝑡
Given unit step response: 𝑐 𝑡 = 0.9 − 0.9e 2

Unit impulse response

𝑡
𝑑 −
𝑐 𝑡 = (0.9 − 0.9e )
2
𝑑𝑡

𝑡
−2
𝑐 𝑡 = 0.45e

Unit ramp response

𝑡 𝑡 𝑡 𝑡
−2 −2
𝑐 𝑡 = න 0.9 − 0.9e 𝑑𝑡 = 0.9𝑡 + 1.8e 0
0
𝑡
−2
= 0.9𝑡 + 1.8e − 1.8

15
15
 Pole-Zero Cancellation

𝑁 𝑠
Transfer function: 𝐺(𝑠) =
G(s) 𝐷(𝑠)

Y(s) X(s) Zero(es): All ‘s’ that satisfy 𝑁 𝑠 = 0

Pole-zero cancellation
A phenomenon in which the effects of a pole will be cancelled
when its location in the s-plane is very near a zero

16
 Example
𝑠 + 0.99 Effects of pole s = 1 is minimized due to
R(s) (𝑠 + 1)(𝑠 + 5) C(s) pole-zero cancellation

1 𝑠 + 0.99
Unit step response: 𝐶(𝑠) =
𝑠 (𝑠 + 1)(𝑠 + 5)

𝐴1 𝐵1 𝐵2
= + +
𝑠 𝑠+1 𝑠+5

0.198 0.0025 0.2005

= + −
𝑠 𝑠+1 𝑠+5

𝑐 𝑡 = 0.198 + 0.0025𝑒 −𝑡 − 0.2005𝑒 −5𝑡

Very small residue

17
 Cancelling Unwanted Pole(s)
𝑅(𝑠) 1 𝐶(𝑠)
𝐺𝑐
(𝑠 + 𝐴)(𝑠 + 𝐵)(𝑠 + 𝐷)
controller Plant

If pole 𝑠 = −𝐴 is undesirable, 𝐺𝑐 can be designed to be:

𝐺𝑐 = 𝑠 + 𝐴 ± 𝜖

where 𝜖 is a very small number

NOTE: Pole-zero cancellation can only be done on stable poles!!

18
 Summary

▪ First order, second order and higher order systems

▪ Step, impulse, ramp and parabolic responses
▪ Pole-zero cancellation

19
MA3005: Control Theory

School of Mechanical and Aerospace

Engineering

Faculty: Asst Prof Tegoeh Tjahjowidodo

Email: [email protected]
Phone: (+65)6790 4952
Office: N3-02c-68
Week 7 – Basic Control Actions
Part I – Lecture Map
2. Laplace
Transform
3. Modelling

1. Introduction to Systems

6. First-Order 4. Response and Stability

5. Routh Stability
8. Transient 7. Second-Order
Response

9. Basic
Control PART II. CONTROLLER DESIGN
Actions

Week 7 3
Learning Objectives
At the end of this week, you should be able to describe different types of controller.

+q
Target

Current position

Week 7 4
Control System
A control system can be:
• Open loop
R(s) C(s)
G(s)
• Closed-loop

Plant/
R(s) Controller
System C(s)
+-

Week 7 5
Open Loop Control System
Advantages:
• Stability is not a problem if plant is stable
• Less expensive
• Easier to control, especially if output cannot be measured

Disadvantages:
• Sensitive to disturbance and modelling errors
• Periodic calibration is required

Week 7 6
Closed-Loop Control System
Advantages:
• Reduced sensitivity to disturbance inputs
• Reduced sensitivity to changes in plant (modelling errors)
• Ability to stabilise an open loop unstable plant
• Ability to change dynamics of system to get the required system performance

Disadvantages:
• Need to use a sensor, which adds complexity and cost
• Risk of instability

Week 7 7
Block Diagram of an Industrial Control System

Automatic controller

Error detector

Reference
input Output
Amplifier Actuator Plant
Set point +-
Actuating
error signal
Sensor

Week 7 8
Simplified Feedback Block Diagram
Considering unity feedback system i.e. H(s) = 1:

R E U C
Controller Plant
+-

Week 7 9
Classification of Industrial Controllers
This includes:
• Two-position or on-off controller
• Proportional (P) controller
• Integral (I) controller
• Proportional-plus-Integral (P+I, PI) controller
• Proportional-plus-Derivative (P+D, PD) controller
• Proportional-plus-Integral-plus-Derivative (P+I+D, PID) controller

Week 7 10
On-Off Control Action
Characteristics:
• Two fixed positions, either on or off
• Relatively simple and inexpensive
• Output signal from controller u(t) remains at a minimum U2 or maximum U1 value,
depending on whether the actuating error signal e(t) is positive or negative

Block Diagram Differential gap

U1
e u e u
+- U2 +-

Week 7 11
Controller

Week 7 12
On-Off Controller Logic
R E Controller
U Plant
C
+-

Owing to the air conditioning (AC) concept, the controller will activate the plant (e.g. heat
pump) when C > R (or E < 0) and deactivate it when E > 0.
U
1
Note:
• 1 means activation of the plant.
0 E • 0 means deactivation of the plant.

Week 7 13
On-Off Controller Logic (Cont’d.)
R E Controller
U Plant
C
+-

Consider a case of a servo system (angular positioning of a motor shaft):

+q
Target

Current position

Week 7 14
Controlling Servo System

U
+q 1
Target

Current position 0 E

–1

E = target - current

Note:
• 1 means to rotate the motor counterclockwise (ccw) in full speed.
• –1 means to rotate the motor clockwise (cw) in full speed.
Week 7 15
Controlling Servo System (Cont’d.)

U
+q 1
Target
0.5
Current position 0 E
–0.5

–1

E = target - current

Note:
• 1 means to rotate the motor ccw in full speed.
• –1 means to rotate the motor cw in full speed.
Week 7 16
Controlling Servo System (Cont’d.)

U
+q 1
Target
0.5
Current position 0 E
–0.5

–1

Note:
• 1 means to rotate the motor ccw in full speed.
• –1 means to rotate the motor cw in full speed.
Week 7 17
Controlling Servo System (Cont’d.)

U
+q 1
Target
0.5
Current position 0 E
–0.5

–1

E = target - current

The output signal of the controller u(t) is

proportional to the error signal e(t).
R E U C
Note: Controller Plant

• 1 means to rotate the motor ccw in full speed. + - = Kp

• –1 means to rotate the motor cw in full speed.
Week 7 18
P and I Controllers

19
P Control Action
The output signal of the controller u(t) is proportional to the error signal e(t) i.e.:
U (s)
u(t) = Kp e(t) or  Kp
E (s)
• Kp is called the proportional gain.
• P controller is just an amplifier with an adjustable gain.

Week 7 20
Block Diagram for P Controller

E(s) U(s)
Kp
+-

e(t) u(t)

0 t 0 t

The plot of u(t) has the same shape as that of e(t) and is just scaled by the value of Kp.

Week 7 21
Effects of P Controller
A high value of Kp will result in:
• High system sensitivity and improved transient response
• Better disturbance rejection
• Lower steady-state error

An excessively high value of Kp can also lead to instability.

Week 7 22
I Control Action
Controller output u(t) is changed at a rate proportional to the actuating error signal e(t) i.e.:
du (t ) t
 K i e(t ) or u (t )  K i  e(t )dt
dt 0
t
u (t ) . K i  e(t )dt
where Ki is an adjustableorconstant
0

Transfer function for integral control:

U ( s) Ki

E ( s) s

Week 7 23
Block Diagram of an I Controller
E(s) Ki U(s)
+- s

e(t) u(t)
Area = A u(t1) = KI A

0 t1 t2 t 0 t1 t2 t

In general, even if e(∞) = 0, there can be a non-zero actuation signal, i.e., u(∞) = constant
value ≠ 0.

Week 7 24
Effects of I Controller
There can be a constant u(∞) even when e(∞) = 0.
It is able to eliminate steady-state error due to step reference and disturbance inputs.
If e(t) does not change sign, u(t) will increase and become very large. For physical controller,
the output will saturate.
If e(t) changes sign, a larger KI value results in oscillatory u(t), which may result in a more
oscillatory response.

Week 7 25
PI Control Action
Control action of a PI controller is defined by:
t
u (t )  K p e(t )  K I  e(t )dt
0

Transfer function representation is:

U s  
 K I 
 K p 1
E s  
 K p s 

It is common practice to express integral control gain, in terms of integral time Ti, as:

U ( s)  1 
 K p 1  
E (s)  Ti 
s

Week 7 26
Block Diagram of PI Controller

E(s) K p (1  Ti s ) U(s)
+- Ti s

where Kp is the proportional gain and Ti is the integral time (1/Ti is the reset rate).
Both Kp and Ti are adjustable.

Week 7 27
Effects of PI Controller
Its main effect is similar to I controller.
Both proportional and integral controllers contribute to u(t), whose larger value accelerates
the output towards the set point.
However, larger u(t) can result in the system to overreact, oscillate and may result in
instability (similar to I controller).
The I term allows for the elimination of steady-state error that may occur with P controller
alone.

Week 7 28
P, I and D Controllers

29
PD Control Action
Control action of a PD controller is defined as:
de(t )
u(t )  K p e(t )  K pTd
dt
and the transfer function is:
U ( s)
 K p 1  Td s 
E ( s)

Week 7 30
Block Diagram of PD controller

E(s) U(s)
Kp(1+Td s)
+-

Td is a constant called the derivative time. Both Kp and Td are adjustable.

Derivative control action is also called rate control.
Magnitude of controller output is proportional to rate of change of actuating error signal.

Week 7 31
Effects of PD Controller
PD controller is able to anticipate error to improves stability, i.e., to increase the damping of
the system.
It tends to amplify high frequency signals, which may lead to saturation of system.
D controller is always combined with a P or PI controller.

Week 7 32
PID Control Action
PID control action is the combination of P, I, and D control actions. It has the advantages of
each of the three individual control actions.
The combine action is given as:
Kp t de(t )
u ( t )  K p e( t ) 
Ti 0
e(t )dt  K pTd
dt
Transfer function representation:
U ( s)  1 
 K p 1   Td s 
E ( s)  Ti s 
where Kp is the proportional gain, Ti is the integral time, and Td is the derivative time.

Week 7 33
Remarks on PID controller
PID control is extremely common in process industry.
It maintains controlled variables such as temperature, pressures, and levels at constant values
despite disturbances and parameters variation.
Gains of PID controller are set by established procedure such as Ziegler-Nichols method
(Section 8-2: Tuning Rules for PID controllers).

Week 7 34
Summary
Proportional control improves speed of response and disturbance rejection.
Integral control can remove steady-state error. However, it may also lead to oscillatory
response or instability.
Derivative control adds damping to the system and tends to increase the stability of the
system.
Derivative control permits the use of a larger proportional gain, which will result in an
improvement in the steady-state accuracy.

Week 7 35
Lesson Summary (Cont’d.)
At the end of this week, you should be able to describe different types of controller.

+q
Target

Current position

Week 7 36
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Background
▪ Educational Background (Mechanical Engineering)
➢ B.Eng. with first class honors from NTU (2006 – 2010)
➢ Dual Ph.Ds. from NTU and CMU (2011 – 2016)
➢ Post-Doctoral Researcher at MPI (2016 – 2017)
CMU – Carnegie Mellon University (USA), MPI – Max Planck Institute for Intelligent Systems (Germany)

▪ Contact Details
➢ Email: [email protected]
➢ Office: N3.2-01-22

2
 Teaching Plan

Week Lecture Topics Tutorial

8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

3
 Continuous Assessment

Homework 2 (20%)
• Released in Week 10
• Two weeks to complete

4
 Motivation
1. A method to unify sub-systems across different disciplines

2. To regulate the output with respect to time t

For example:
Desired Electrical System Mechanical System Actual
Velocity Voltage Torque Velocity
Controller

Sensors

5
 Laplace Transformation
Converting differential equations (t) algebraic equations (s)

Some Basic Transform s-shifting

1
ℒ 1 = 1
𝑠 ℒ −1 = 𝑒 −𝑎𝑡
𝑠 𝑠+𝑎
ℒ cos 𝜔𝑡 = 2
𝑠 + 𝜔2
𝜔 𝜔
ℒ sin 𝜔𝑡 = 2 ℒ −1 = 𝑒 −𝑎𝑡 sin 𝜔𝑡
𝑠 + 𝜔2 𝑠 + 𝑎 2 + 𝜔2

𝑑𝑓 0
ℒ 𝑓(𝑡) = 𝐹(𝑠), ℒ = 𝑠𝐹 𝑠 − 𝑓(𝑡 = 0)
𝑑𝑡

6
 Transfer Function Poles
Differential Equation Laplace Transformation
𝑀𝑥ሷ + 𝐶 𝑥ሶ + 𝐾𝑥 = 𝑓 𝑀𝑠 2 + 𝐶𝑠 + 𝐾 𝑋(𝑠) = 𝐹(𝑠)
𝑋(𝑠) 1
Homogeneous solution: 𝑀 𝑥ሷ + 𝐶 𝑥ሶ + 𝐾𝑥 = 0 =
𝐹(𝑠) 𝑀𝑠 2 + 𝐶𝑠 + 𝐾
𝑀𝑑 2 + 𝐶𝑑 + 𝐾 = 0 Transfer Function Poles: 𝑀𝑠 2 + 𝐶𝑠 + 𝐾 = 0

1st possibility 2nd possibility 1st possibility 2nd possibility

𝑑 = −1, 2 𝑑 = ±𝑗5 𝑠 = −1, 2 𝑠 = ±𝑗5
𝑥= 𝐴e−𝑡 + 𝐵e2𝑡 𝑥 = 𝐴cos(5𝑡 + ∅) 𝑥 = 𝐴e−𝑡 + 𝐵e2𝑡 𝑥 = 𝐴cos(5𝑡 + ∅)

3rd possibility 3rd possibility

𝑑 = −2 ± 𝑗5 𝑠 = −2 ± 𝑗5
𝑥 = 𝐴e−2𝑡 cos 5𝑡 + ∅ 𝑥 = 𝐴e−2𝑡 cos 5𝑡 + ∅

Poles dictate homogeneous 7

solution/ transient response
 Stable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 + 3𝑠 + 2

Transfer Function Poles: 𝑠 2 + 3𝑠 + 2 = 0 𝑠+1 𝑠+2 =0 𝑠 = −1, −2

Im Transient Response Step Response

x(t)
𝐴𝑒 −𝑡 + 𝐵𝑒 −2𝑡

-2 -1 Re

𝑠 = −1 𝑠 = −2 Time
Pole #1 Pole #2
Transient response decays

s-plane 𝐴, 𝐵 are constants Stable system

8
Partially-Stable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 +4

Transfer Function Poles: 𝑠2 + 4 = 0 𝑠 = ±𝑗2

Im Transient Response Step Response

x(t)
2
𝐴 cos(2𝑡 + ∅)

Re
-2
𝑠 = ±𝑗2 Time
Complex Poles Transient response oscillates

s-plane 𝐴, ∅ are constants Partially-stable system

9
 Unstable Systems
Transfer function poles dictate stability
𝑋(𝑠) 1
= 2
𝐹(𝑠) 𝑠 +𝑠−2

Transfer Function Poles: 𝑠2 + 𝑠 − 2 = 0 𝑠−1 𝑠+2 =0 𝑠 = 1, −2

Im Transient Response Step Response

x(t)
𝐴𝑒 𝑡 + 𝐵𝑒 −2𝑡

-2 1 Re

𝑠=1 𝑠 = −2 Time
Pole #1 Pole #2
Transient response → ∞
s-plane 𝐴, 𝐵 are constants unstable system

10
 Stability
Three types of poles
1. Real poles (e.g. s = a) Stable systems require
Transient response 𝐴𝑒 𝑎𝑡 𝑒 𝑎𝑡 → 0, when 𝑡 → ∞
𝑎<0
2. Pure complex poles (e.g. s = ±jb)
Left hand s-plane
Transient response 𝐴 cos(𝑏𝑡 + ∅) Im
(always partially-stable)

Stable Unstable
3. Complex poles with real parts (e.g. s = a ± jb)
Re
Transient response 𝐴𝑒 𝑎𝑡 cos(𝑏𝑡 + ∅)

𝐴, ∅ are constants
s-plane 11
 Dominant Poles
Dominant poles are the
rightmost poles (slowest) Step Response

e.g., compare step response of two

systems 5s

System #1 has two poles: s = -2 ± j

2s
transient response: 𝐴𝑒 −2𝑡 cos(𝑡 + ∅)

System #2 has two poles: s = -1 ± j

transient response: 𝐴𝑒 −𝑡 cos(𝑡 + ∅)
Slower!

12
 Transient Response
E.g., Underdamped Second Order System
Poles
𝑋(𝑠) 𝜔𝑛2 𝑠 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2 ,
= 2 2, 𝜁<1 Transient
𝐹(𝑠) 𝑠 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛 𝐴𝑒 −𝜁𝜔𝑛 𝑡 cos(𝜔𝑛 1 − 𝜁 2 𝑡 + ∅)

Im Step Response
𝜁𝜔𝑛 𝜋𝜁
−
%OS = 𝑒 1−𝜁 2
𝜔𝑛
𝜔𝑛 1 − 𝜁2 𝛽
Re

4
cos 𝛽 = 𝜁 𝑡𝑠 =
𝜁𝜔𝑛 steady-state
value
s-plane
13
 Closed-Loop System

R(s) + E(s) C(s)

Gc(s) P(s)
-
Controller
Plant

Objective: Design GC(s) to achieve the desired steady-

state errors and poles for closed-loop transfer function

Specifically in this lecture To make c(t) stable and control its

1. Steady-state error
2. Transient response
14
 Physical System

R(s) + E(s) C(s)

Gc(s) P(s)
-
Controller
Plant

Software
Programming

15
Microcontroller
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

System Types

16
 Unity feedback systems
R(s) + E(s) G (s) P(s)
C(s)
c
-

Simplify R(s)+ E(s) C(s)

G(s)
- C(s)
G(s): Open-loop Transfer Function
𝐸 = 𝑅 − 𝐶,
𝐶 = 𝐺𝐸 → 𝐶 = 𝐺 𝑅 − 𝐶 → 𝐶 1 + 𝐺 = 𝐺𝑅
𝐶 𝐺
= 17
𝑅 (1 + 𝐺)
 System Type
G(s): Open-loop transfer function for unity feedback system

K (Ta s + 1)(Tb s + 1) (Tm s + 1)

G (s) =
s N (T1s + 1)(T2 s + 1) (T s + 1)
p

G(s) is Type N system, System Type = power of “s” in denominator

where N can be 0, 1, 2, etc..
steady-state error (desirable)
make closed-loop
system
N stability (undesirable)

N is a trade off between steady-state

error and stability
18
 Steady-State Error
R(s)+ E(s) C(s) 𝐶 𝐺
G(s)
- Transfer function: 𝑅 = (1 + 𝐺) ,

𝐶 𝐺
𝐸 =𝑅−𝐶 → 𝐸 =𝑅− 𝑅 → 𝐸 =𝑅 1−
𝑅 1+𝐺
𝑅
𝐸(𝑠) =
(1 + 𝐺)
Compute steady-state error ess via final value theorem:

ess = lim e(t ) = lim sE ( s )

t → s →0

sR( s )
= lim
s →0 1 + G ( s ) 19
 Step Input ess
𝑠𝑅
General formula: 𝑒𝑠𝑠 = lim
s→0 (1 + 𝐺)

1 s 1 1
For step input: 𝑅 = → ess = lim ∙ =
s s→0 1 + 𝐺 𝑠 1 + 𝐺(𝑠 = 0)
We define Kp to be static position error constant, such that

𝐾𝑝 = lim 𝐺 𝑠 = 𝐺(𝑠 = 0)
𝑠→0

Using this definition, the step input ess can be expressed as:

1
ess =
1+ K p 20
 Kp vs System Types
For type 0 systems
K (Ta s + 1)(Tb s + 1)
K p = lim =K
s →0 (T1s + 1)(T2 s + 1) 

1 1
→ 𝑒𝑠𝑠 = =
1 + 𝐾𝑝 1 + 𝐾 (Non-zero error)
For type 1 or higher systems
K (Ta s + 1)(Tb s + 1) 
K p = lim N → , for N  1
s →0 s (T s + 1)(T s + 1) 
1 2

1
𝑒𝑠𝑠 = → 𝑒𝑠𝑠 = 0
1 + 𝐾𝑝
(steady-state error eliminated) 21
 Observation (I)
For step inputs, the steady-state error ess:
1. Type 0 systems
1
ess = ( Non-zero value)
1+ K
2. Type 1 or higher systems

ess = 0 (steady-state error

eliminated)

Require at least one integrator in forward path

22
 Ramp Input ess
𝑠𝑅
General formula: 𝑒𝑠𝑠 = lim
s→0 (1 + 𝐺)

1 s 1 1
For ramp input: 𝑅 = 2 → ess = lim ∙ 2=
s s→0 1 + 𝐺 𝑠 lim𝑠𝐺
s→0
We define Kv to be static velocity error constant, such that

K v = lim sG( s )
s→0
Using this definition, the ramp input ess can be expressed as:

1
ess =
Kv
23
 Kv vs System Types
For type 0 systems
sK (Ta s + 1)(Tb s + 1) 1
K v = lim = 0 → 𝑒𝑠𝑠 = → ∞ (can’t follow input)
s →0 (T1s + 1)(T2 s + 1)  𝐾𝑣

For type 1 systems

sK (Ta s + 1)(Tb s + 1) 1
K v = lim =K → 𝑒𝑠𝑠 = (non-zero error)
s →0 s (T1s + 1)(T2 s + 1)  𝐾
For type 2 and higher systems
sK (Ta s + 1)(Tb s + 1) 
K v = lim N → , for N  2 → 𝑒𝑠𝑠 = 0
s →0 s (T s + 1)(T s + 1) 
1 2

(steady-state error eliminated)

24
 Observation (II)
For ramp inputs, the steady-state error ess:
1. Type 0 systems:

𝑒𝑠𝑠 → ∞ (can’t follow ramp input)

2. Type 1 systems
1
𝑒𝑠𝑠 = (non-zero value)
𝐾𝑣
3. Type 2 or higher systems
𝑒𝑠𝑠 = 0 (steady-state error eliminated)

25
 Unit Parabolic Input ess
𝑠𝑅
General formula: 𝑒𝑠𝑠 = lim
s→0 (1 + 𝐺)

1 s 1 1
Parabolic input: 𝑅 = 3 → ess = lim ∙ 3=
s s→0 1 + 𝐺 𝑠 lim𝑠 2 𝐺
s→0
We define Ka to be static acceleration error constant:

K a = lim s G ( s )
2
s →0

Using this definition, the parabolic input ess can be expressed as:

1
ess =
Ka 26
 Ka vs System Types
For type 0 systems:

s 2 K (Ta s + 1)(Tb s + 1)

K a = lim =0
s →0 (T1s + 1)(T2 s + 1)
1
→ 𝑒𝑠𝑠 = →∞ (can’t follow input)
𝐾𝑎
For type 1 systems
s K (Ta s + 1)(Tb s + 1)
2
K a = lim =0
s →0 s (T1s + 1)(T2 s + 1) 

1
→ 𝑒𝑠𝑠 = →∞ (can’t follow input)
𝐾𝑎
27
 Ka vs System Types
For type 2 systems:
s K (Ta s + 1)(Tb s + 1)
2
K a = lim 2 =K
s→0 s (T s + 1)(T s + 1) 
1 2
1
→ 𝑒𝑠𝑠 = (non-zero error)
𝐾
For type 3 or higher systems
s 2 K (Ta s + 1)(Tb s + 1) 
K a = lim N → , for N  3
s →0 s (T s + 1)(T s + 1) 
1 2

1
→ 𝑒𝑠𝑠 = = 0 (steady-state error eliminated)
𝐾𝑎 28
 Observation (III)
For parabolic inputs, the steady-state error ess:
1. Type 0 and 1 systems:

𝑒𝑠𝑠 → ∞ (can’t follow parabolic input)

2. Type 2 systems
1
𝑒𝑠𝑠 = (non-zero value)
𝐾𝑎
3. Type 3 or higher systems
𝑒𝑠𝑠 = 0 (steady-state error eliminated)

29
Parabolic response of Type 2

30
 Summary
Input “level”

System Steady-state Error (ess)

Types Unit Step input Unit Ramp input Unit Parabolic input
1
System “level”

Type 0 ∞ ∞
1 + 𝐾𝑝
1
Type 1 0 ∞
𝐾𝑣
1
Type 2 0 0
𝐾𝑎
Type 3 0 0 0
Input “Level” higher than System “level” 𝑒𝑠𝑠 → ∞
Input “Level” equal to System “level” 𝑒𝑠𝑠 is non-zero
Input “Level” lower than System “level” 𝑒𝑠𝑠 = 0
31
 Discussion
▪ Steady-state errors are related to Kp, Kv, and Ka

▪ Increasing system type requires an integrator in the

forward path but this will have destabilizing effect

▪ Designing a stable system with more than two

integrators in the forward path is generally difficult

32
 Acknowledgement

Professor Cuong Professor Tegoeh

33
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Teaching Plan
Week Lecture Topics Tutorial

8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

2
 Closed Loop System
Purpose: Create desired Changing the system’s
transient and steady-state pole locations
response

Input, e.g., Plant Output, e.g.,

Desired Velocity Actual Velocity
Controller

Sensors

3
 Root Locus
Using a proportional controller, i.e., Controller = K

R(s) E(s) KE(s) C(s)

K G(s)

H(s)

Root Locus: The trajectory of the closed-loop system’s poles as K

varied from zero to infinity

4
 Illustration
c(t) K= 0.01
R(s) C(s)
+- K G(s)

jw
t
K=0.01 c(t) K=1

K=1

K→ b
× × × × × s t
K=10
K = 10
c(t)

s-plane
5
t
 Motivation
Learning root locus allows you to:

1. Visualize how the closed-loop poles will move

2. Select a suitable controller

▪ e.g., P, PI, PD, PID controller

First step towards learning how to design controllers

6
 Recap: Complex Numbers
𝜎
Im 𝜎 + 𝑗𝜔 = 𝑟𝑒 𝑗𝜃
𝑟 −1 𝜔
𝜔 where 𝑟 = 𝜎2 + 𝜔2 ,𝜃= tan
𝜎

𝜃 𝜎1 + 𝑗𝜔1 𝑟1 𝑒 𝑗𝜃1
Re =
𝜎2 + 𝑗𝜔2 𝑟2 𝑒 𝑗𝜃2

complex plane
𝑟1 𝑗(𝜃 −𝜃 )
= 𝑒 1 2
𝑟2
Phase
Magnitude 7
 Recap: Complex Numbers
Im 𝑠 + 𝑧 = 𝑟𝑧 𝑒 𝑗𝜃𝑧 ,
where 𝑟𝑧 = |𝑠 + 𝑧 |
𝑠+𝑝
𝑠 𝑠 + 𝑝 = 𝑟𝑝 𝑒 𝑗𝜃𝑝 ,
𝑟𝑝 𝑟𝑧 𝑠 + 𝑧
where 𝑟𝑝 = |𝑠 + 𝑝 |
𝜃𝑝 𝜃𝑧

−𝑝 −𝑧 Re 𝑠+𝑧 𝑟𝑧 𝑗(𝜃 −𝜃 )
= 𝑒 𝑧 𝑝
𝑠+𝑝 𝑟𝑝
complex plane
Phase
Magnitude
8
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Root Locus: Sketching

9
 Characteristic Equation

R(s) C(s) C (s ) KG (s )
+- K G(s) =
R(s ) 1 + KG (s )H (s )
H(s) Closed-loop
characteristic equation

Characteristic equation (CE): 1 + KG(s)H(s) = 0

For a given K, any s that satisfied CE will be a closed-loop pole
When K varies, the location of the closed-loop pole changes:
root locus represents the trajectory of the closed-loop poles

10
 Required Conditions
CE: 1 + KG(s)H(s) = 0

rewrite KG(s)H(s) = -1 equal

𝑁 𝑠 𝑠 + 𝑧1 𝑠 + 𝑧2 … (𝑠 + 𝑧𝑚 )
𝐾
𝐷(𝑠)
=𝐾
𝑠 + 𝑝1 𝑠 + 𝑝2 … (𝑠 + 𝑝𝑛 ) 1𝑒 ±𝑗(𝜋±2𝑞𝜋)

𝑁 𝑠 𝑁 𝑠
𝐾 = 1𝑒 ±𝑗(𝜋±2𝑞𝜋) 𝐾 𝑒𝑗 ∠𝑁−∠𝐷
= 1𝑒 ±𝑗(𝜋±2𝑞𝜋)
𝐷(𝑠) 𝐷(𝑠)

Magnitude condition Phase condition

𝑁 𝑠 ∠𝑁 − ∠𝐷 = −𝜋 − 2𝑞𝜋
𝐾 =1
𝐷(𝑠) ∠𝑁 − ∠𝐷 = −(1 + 2𝑞)𝜋 11
 Starting and Ending Points
R(s) C(s)
+- K G(s) 𝑁(𝑠)
CE: 1 + 𝐾 𝐷(𝑠) = 0 𝐷(𝑠) + 𝐾𝑁(𝑠) = 0
H(s)

Smallest K (i.e., K = 0),

CE: 𝐷(𝑠) + 𝐾𝑁(𝑠) = 0 𝐷 𝑠 =0

Solution shows that closed-loop POLES = open-loop POLES!

Largest K (i.e., 𝑲 → ∞),

CE: 𝐷(𝑠) + 𝐾𝑁(𝑠) = 0 𝑁 𝑠 =0
Solution shows that closed-loop POLES = open-loop ZEROS!

12
 Infinity Open-Loop Zeros
𝑠 + 𝑧1 𝑠 + 𝑧2 … 𝑠 + 𝑧𝑚
1+𝐾 =0
𝑠 + 𝑝1 𝑠 + 𝑝2 … (𝑠 + 𝑝𝑛 )

𝑠 + 𝑧1 𝑠 + 𝑧2 … 𝑠 + 𝑧𝑚
𝐾 = 1𝑒 −𝑗𝜋(1+2𝑞)
𝑠 + 𝑝1 𝑠 + 𝑝2 … (𝑠 + 𝑝𝑛 )

In general: 𝑚 ≤ 𝑛, some CL poles end at infinity open-loop zeros

𝑠 + 𝑧1 𝑠 + 𝑧2 … 𝑠 + 𝑧𝑚
lim 𝐾 = 1𝑒 −𝑗𝜋(1+2𝑞)
𝑠→∞ 𝑠 + 𝑝1 𝑠 + 𝑝2 … (𝑠 + 𝑝𝑛 )
𝑠𝑚 𝑚−𝑛 lim 𝐾 𝑠 𝑒 𝑗∠𝑠
𝑚−𝑛
= 1𝑒 −𝑗𝜋(1+2𝑞)
lim 𝐾 𝑛 = 1𝑒 −𝑗𝜋(1+2𝑞) , lim 𝐾 𝑠 = 1𝑒 −𝑗𝜋(1+2𝑞) , 𝑠→∞
𝑠→∞ 𝑠 𝑠→∞

Magnitude condition Phase condition

𝑚−𝑛 𝑚 − 𝑛 ∠𝑠 = −𝜋(1 + 2𝑞)
𝐾 𝑠 =1
𝜋 1 + 2𝑞
∠𝑠 = 13
𝑛−𝑚
 Physical Interpretation
Im Phase condition at s → ∞
𝜋(1 + 2𝑞)
∠𝑠 = , 𝑞 = 0, … , 𝑛 − 𝑚 − 1
𝑛−𝑚
∠𝑠 represents the asymptote angle
-p2 Required to substitute 𝑞 with
|𝑛 – 𝑚| number of integer values

Asymptote location is on real axis

-z1 -p1 Re
σ𝑛𝑖 −𝑝𝑖 − σ𝑚
𝑗 −𝑧𝑗
-p3 𝜎𝑎 =
𝑛−𝑚
Example
Infinity 𝑠 + 𝑧1
1+𝐾 =0
𝑠 + 𝑝1 𝑠 + 𝑝2 (𝑠 + 𝑝3 )
Starting point (OL POLES)

Ending point (OL ZEROS) 𝑚 = 1, 𝑛 = 3 14

 Physical Interpretation
Im Example
𝑠 + 𝑧1
1+𝐾 =0
𝑠 + 𝑝1 𝑠 + 𝑝2 (𝑠 + 𝑝3 )
Asymptote Formula
-p2
3𝜋 𝜋(1 + 2𝑞)
σ𝑛𝑖 −𝑝𝑖 − σ𝑚
𝑗 −𝑧𝑗
𝜋 2 𝜎𝑎 = , ∠𝑠 = ,
2 𝑛−𝑚 𝑛−𝑚
sa 𝑚 = 1, 𝑛 = 3
-z1 -p1 Re

-p3 σ3𝑖 −𝑝𝑖 + 𝑧1 𝜋(1 + 2𝑞)

𝜎𝑎 = , ∠𝑠 = ,
3−1 3−1
𝜋
Infinity ∠𝑠 = + 𝑞𝜋, 𝑞 = 0, 1
2
Starting point (OL POLES) 𝜋 3𝜋
→ ∠𝑠 = ,
Ending point (OL ZEROS)
2 2 15
 Locus on the Real Axis
Im Phase Condition

𝑠 + 𝑧1 𝑠 + 𝑧2 … 𝑠 + 𝑧𝑚
∠ = ±𝜋
𝑠 + 𝑝1 𝑠 + 𝑝2 … (𝑠 + 𝑝𝑛 )

𝑚 𝑛
-p3
−𝜃 ෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
… s + p3
s 𝑖=1 𝑗=1

s + p4
-p2 -z2 -p1 -z1 Re
𝜃
-p4 Angle contributions Angle contributions
from all OL zeros from all OL poles

Region 1
Ignore complex open-loop poles/zeros
Test point of s
Conjugate pairs cancel their angle
Starting points (OL POLES) contributions
16
Ending points (OL ZEROS)
 Locus on the Real Axis
Im Phase Condition
𝑚 𝑛

෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

∠ 𝑠 + 𝑧𝑖 = ∠ 𝑠 + 𝑝𝑗 = 0
… s+p2→ ∠ 𝑠 + 𝑝2 = 0 s

-p2 -z2 -p1 -z1 Re Region 1: LHS angles = 0 ≠ ±𝜋

No locus in Region 1!
Region 1

Test point of s

Starting points (OL POLES)

17
Ending points (OL ZEROS)
 Locus on the Real Axis
Im Phase Condition
𝑚 𝑛

෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

∠ 𝑠 + 𝑧1 = 𝜋 The vectors for all the and

…
that are on the left of produce:
-p2 -z2 -p1 -z1 Re ∠ 𝑠 + 𝑧2 = ∠ 𝑠 + 𝑝1 = ∠ 𝑠 + 𝑝2 = 0

Only vector ∠ 𝑠 + 𝑧1 = 𝜋
Region 2
Region 2: LHS angles = 𝜋 = RHS
Test point of s

Starting points (OL POLES) Region 2 has locus!

18
Ending points (OL ZEROS)
 Locus on the Real Axis
Im Phase Condition
𝑚 𝑛

෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

2 vectors on the right of

∠ 𝑠 + 𝑧1 = 𝜋 produce angles of 𝜋:
…
∠ 𝑠 + 𝑧1 = ∠ 𝑠 + 𝑝1 = 𝜋
-p2 -z2 -p1 -z1 Re

Net phase on LHS = 0 ≠ ±𝜋

Region 3
No locus in Region 3!
Test point of s

Starting points (OL POLES)

19
Ending points (OL ZEROS)
 Locus on the Real Axis
Im Phase Condition
𝑚 𝑛

෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

∠ 𝑠 + 𝑧1 = 𝜋 3 vectors on the right of

… produce angles of 𝜋:
-p2 -z2 -p1 -z1 Re ∠ 𝑠 + 𝑧2 = ∠ 𝑠 + 𝑧1 = ∠ 𝑠 + 𝑝1 = 𝜋

Net phase on LHS = 𝜋 = ±𝜋

Region 4

Locus in Region 4!
Test point of s

Starting points (OL POLES)

20
Ending points (OL ZEROS)
 Locus on the Real Axis
Im Phase Condition
𝑚 𝑛

෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

…
Regions with odd number of
-p2 -z2 -p1 -z1 Re and that are on the right of
will satisfy the phase condition

Starting point (OL POLES)

21
Ending point (OL ZEROS)
 Break-Out Points
Example
Im
𝑠+5
CE: 1 + 𝐾 =0
𝑠 + 1 𝑠 + 2 (𝑠 + 3)

3𝜋 Finding Asymptote
𝜋
2 2 𝜋(1 + 2𝑞)
Angle: ∠𝑠 =
3−1
-5 -3 -2 -1 -0.5 Re 𝜋 3𝜋
∠𝑠 = , , 𝑞 = 0, 1
2 2

−1 − 2 − 3 − (−5)
Location: 𝜎𝑎 = = −0.5
3−1

Starting point (OL POLES)

22
Ending point (OL ZEROS)
 Break-Out Points
Example
Im
𝑠+5
CE: 1 + 𝐾 =0
𝑠 + 1 𝑠 + 2 (𝑠 + 3)

Breakout s = -1.45 𝑠+1 𝑠+2 𝑠+3

point 𝐾=−
𝑠+5
𝑑𝐾
=0 (max point)
-5 -3 -2 -1 -0.5 Re 𝑑𝑠

→ 𝑠 = −1.45, −2.62, −6.42

rejected because
not on locus

Starting point (OL POLES)

23
Ending point (OL ZEROS)
 Break-Out Points
Example
Im
𝑠+5
CE: 1 + 𝐾 =0
𝑠 + 1 𝑠 + 2 (𝑠 + 3)

𝑠+1 𝑠+2 𝑠+3

𝐾=−
𝑠+5
𝑑𝐾
=0 (max point)
-5 -3 -2 -1 -0.5 Re 𝑑𝑠

→ 𝑠 = −1.45, −2.62, −6.42

rejected because
not on locus

Starting point (OL POLES)

24
Ending point (OL ZEROS)
 Break-In Points
Example
Im
𝑠+3
CE: 1 + 𝐾 =0
𝑠+1 𝑠+2

Asymptote
𝜋(1 + 2𝑞)
𝜋 Angle: ∠𝑠 =
2−1

∠𝑠 = 𝜋, 𝑞=0
-3 -2 -1 Re
−1 − 2 − (−3)
Location: 𝜎𝑎 = =0
2−1

Starting point (OL POLES)

25
Ending point (OL ZEROS)
 Break-In Points
Example
Im
𝑠+3
CE: 1 + 𝐾 =0
𝑠+1 𝑠+2

𝑠+1 𝑠+2
𝐾=−
𝑠+3
𝑑𝐾
= 0 (max and min point)
-3 -2 -1 Re 𝑑𝑠

→ 𝑠 = −1.58, −4.41

Break-out point Break-in point

(max point) (min point)

Starting point (OL POLES)

26
Ending point (OL ZEROS)
 Departure Angles
Example
Im
𝑠+4
CE: 1 + 𝐾 =0
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
-1+j Finding Asymptote
𝜋
𝜋(1 + 2𝑞)
2 Angle: ∠𝑠 =
3−1

-4 -1 0.5 Re 𝜋 3𝜋
3𝜋 ∠𝑠 = , , 𝑞 = 0, 1
2 2
2
−1 − 1 − 1 − (−4)
-1-j Location: 𝜎𝑎 = = 0.5
3−1

Starting point (OL POLES)

27
Ending point (OL ZEROS)
 Departure Angles
Example
Im
𝑠+4
CE: 1 + 𝐾 =0
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
-1+j ?
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
𝐾=−
𝑠+4

𝑑𝐾
-4 -1 0.5 Re =0
𝑑𝑠
No break-out points
-1-j

Starting point (OL POLES)

28
Ending point (OL ZEROS)
 Departure Angles
Example
Im
𝑠+4
s CE: 1 + 𝐾 =0
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
-1+j 𝛼
𝑚=1 𝑛=3

Phase: ෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
−1
1 𝜋 𝑖=1 𝑗=1
tan
3 2
-4 -1 0.5 Re 1 𝜋 𝜋
tan−1 − − − 𝛼 = −𝜋
3 2 2
𝜋
2 1
-1-j 𝛼= tan−1
3

Starting point (OL POLES)

29
Ending point (OL ZEROS)
 Departure Angles
Example
Im
𝑠+4
CE: 1 + 𝐾 =0
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
-1+j 1
tan−1 𝑚=1 𝑛=3
3
Phase: ෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
−1
1 𝜋 𝑖=1 𝑗=1
tan
3 2
-4 -1 0.5 Re 1 𝜋 𝜋
tan−1 − − − 𝛼 = −𝜋
3 2 2
𝜋
2 1
-1-j 𝛼= tan−1
3

Starting point (OL POLES)

30
Ending point (OL ZEROS)
Intersection of Imaginary Axis
Example
Im
𝑠+4
CE: 1 + 𝐾 =0
𝑠+1 𝑠+1+𝑗 𝑠+1−𝑗
-1+j 𝜔 = 14 𝑠 3 + 3𝑠 2 + 4 + 𝐾 𝑠 + (2 + 4𝐾) = 0
Sub s = jw
(𝑗𝜔)3 +3(𝑗𝜔)2 + 4 + 𝐾 𝑗𝜔 + (2 + 4𝐾) = 0
-4 -1 0.5 Re
−𝑗𝜔3 − 3𝜔2 + 4 + 𝐾 𝑗𝜔 + (2 + 4𝐾) = 0

(−3𝜔2 + 2 + 4𝐾 ) + 𝑗( 4 + 𝐾 𝜔 − 𝜔3 ) = 0
-1-j
−𝜔
Real parts Imaginary parts

Starting point (OL POLES) 𝐾 = 10, 𝜔 = 14

31
Ending point (OL ZEROS)
 Arrival Angles
Example
Im
𝑠+1+𝑗 𝑠+1−𝑗
CE: 1 + 𝐾 =0
(𝑠 + 1) 𝑠 + 5
-1+j 𝛼?
𝑠+1 𝑠+5
𝐾=−
𝑠 = −1.78 𝑠+1−𝑗 𝑠+1+𝑗

𝑑𝐾
-5 -1 Re =0
𝑑𝑠

→ 𝑠 = −1.78 (maximum point)

-1-j → break-out point

Starting point (OL POLES)

32
Ending point (OL ZEROS)
 Arrival Angles
Example
Im
𝑠+1+𝑗 𝑠+1−𝑗
CE: 1 + 𝐾 =0
(𝑠 + 1) 𝑠 + 5
-1+j 𝛼?
𝑚 𝑛

1 Phase: ෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
tan−1 𝜋 𝑖=1 𝑗=1
4 2
𝜋 𝜋 −1
1
-5 -1 Re + 𝛼 − − tan =𝜋
2 2 4
𝜋
1
2 𝛼=𝜋+ tan−1
4
-1-j

Starting point (OL POLES)

33
Ending point (OL ZEROS)
 Arrival Angles
Example
Im
𝑠+1+𝑗 𝑠+1−𝑗
CE: 1 + 𝐾 =0
(𝑠 + 1) 𝑠 + 5
-1+j 𝛼
𝑚 𝑛

Phase: ෍ ∠ 𝑠 + 𝑧𝑖 − ෍ ∠ 𝑠 + 𝑝𝑗 = ±𝜋
𝑖=1 𝑗=1

𝜋 𝜋 −1
1
-5 -1 Re + 𝛼 − − tan =𝜋
2 2 4
1
𝛼=𝜋+ tan−1
4
-1-j

Starting point (OL POLES)

34
Ending point (OL ZEROS)
Root Locus Summary
1. Starting and ending points of the locus
a. Starting points – open-loop POLES
b. Ending points – open-loop ZEROS

2. Asymptotes (infinite open-loop ZEROS)

a. Angles
b. Location

3. Locus on real axis

4. Break-in and break-out points

a. Break-in (minimum K)
b. Break-out (maximum K)

5. Departure and arrival angles

6. Imaginary axis intersection

35
 MATLAB Functions
Explore root locus with MATLAB:

1. Create the desired open-loop transfer function

e.g., G = tf([1 2], [5 6 7]); %creating G(s)

Coefficients of Coefficients of the

the numerator denominator

𝑠+2
The mathematical expression of G is: 𝐺 𝑠 = 5𝑠 2 + 6𝑠 + 7

2. Plot the root locus of G(s)

i.e., rlocus(G); %plotting

36
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Teaching Plan
Week Lecture Topics Tutorial

8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

2
 Root Locus
R(s)
C (s ) KG (s )
C(s)
+- K G(s)
=
R(s ) 1 + KG (s )H (s )
Controller Plant

H(s)
Sensor

Characteristic equation (CE): 1 + KG(s)H(s) = 0

Root Locus: How closed-loop poles change for 0 < K < .

Key Question
Could we use the root locus technique if Controller ≠ 𝐾?
Answer: Yes

3
 Basic Controllers
R(s) C(s)
+- E Gc A G(s)
Controller Plant

H(s)
Sensor

PD Controller PI Controller
Time Domain
d𝑒 𝑡
𝑎 = 𝐾𝑝 𝑒 + 𝐾𝐷 𝑎 = 𝐾𝑝 𝑒 + 𝐾𝐼 න 𝑒 d𝑡
d𝑡 0
Laplace Domain
𝐾𝐼
𝐴 = 𝐾𝑝 𝐸 + 𝐾𝐷 𝑠𝐸 𝐴 = 𝐾𝑝 𝐸 + 𝐸
𝑠
= 𝐸 𝐾𝑝 + 𝐾𝐷 𝑠 𝐾𝐼
= 𝐸 𝐾𝑝 +
𝑠
𝐺𝑐 for PD 𝐺𝑐 for PI
4
 Root Locus for PD Controller
R(s) C(s)
+ Kp +𝐾𝐷 𝑠 G(s)
- 𝐾𝑝
PD Controller Plant PD controller: 𝐾𝐷 𝑠 + 𝐾𝑝 = 𝐾𝐷 𝑠+
𝐾𝐷
H(s)
Sensor 𝑧𝑐

For example
jw
G1(s)

R(s) C(s)
+ KD (𝑠 + 𝑧𝑐 ) G(s)
-

s
H(s)

s-plane 5
 Root Locus for PD Controller
R(s) C(s)
+ Kp +𝐾𝐷 𝑠 G(s)
- 𝐾𝑝
PD Controller Plant PD controller: 𝐾𝐷 𝑠 + 𝐾𝑝 = 𝐾𝐷 𝑠+
𝐾𝐷
H(s)
Sensor 𝑧𝑐

For example
jw
G1(s)

R(s) C(s)
+ KD (𝑠 + 𝑧𝑐 ) G(s)
-

-zc s
H(s)

KD and zc change
closed-loop poles s-plane 6
 Root Locus for PI Controller
R(s) 𝐾 C(s)
𝑧𝑐
+
-
𝐾𝑝 + 𝑠𝐼 G(s)
𝐾𝐼
𝑠+
PI Controller Plant 𝐾𝐼 𝐾𝑝
H(s)
PI controller: 𝐾𝑝 + 𝑠 = 𝐾𝑝 𝑠
Sensor

For example
jw
G1(s)
R(s) 𝑠 + 𝑧𝑐 C(s)
+ Kp G(s)
- 𝑠

s
H(s)

s-plane 7
 Root Locus for PI Controller
R(s) 𝐾 C(s)
𝑧𝑐
+
-
𝐾𝑝 + 𝑠𝐼 G(s)
𝐾𝐼
𝑠+
PI Controller Plant 𝐾𝐼 𝐾𝑝
H(s)
PI controller: 𝐾𝑝 + 𝑠 = 𝐾𝑝 𝑠
Sensor

For example
jw
G1(s)
R(s) 𝑠 + 𝑧𝑐 C(s)
+ Kp G(s)
- 𝑠

-zc s
H(s)

PI controller eliminates
steady-state error s-plane 8
 Comparison
PD controller PI controller
Increase speed + stability Eliminate Ess
Transient Response Steady-state Response

Controllers 𝐾𝐼
𝐾𝑝 𝐾𝐷 𝑠
𝑠
Properties Proportional Derivative Integral

Stability

Steady-state error Good

(Ess)
Neutral
Speed Bad
Disturbance
rejection

Noise rejection
9
Improving Transient
Response

10
 Design of PD Controllers
R(s) C(s) Recall: closed-loop poles define
+ Kp+KDs G(s)
- transient response!
PD Controller Plant
H(s) Aim: Find Kp and KD to achieve
Sensor desired closed-loop poles

General Steps
Step 1: Determine the DESIRED closed-loop poles

Step 2: Express the characteristic equation of the closed-loop transfer

function in terms of Kp and KD
Step 3: Substitute the poles into the characteristic equation

Step 4: Solve the real and imaginary parts of the characteristic equation to
obtain the required Kp and KD

11
 Example: Step 1
R(s) 1 C(s)
+ Kp+KDs Desired Transient Response
-
𝑠 2 + 2𝑠 + 2
10% overshoot,
2% settling time of 0.5 s
Step 1: Determine the DESIRED closed-loop poles
Criterion 1: 10% OS Criterion 2: 2% settling time
−
𝜋𝜁 4
𝑒 1−𝜁 2 = 0.1 = 0.5
𝜁𝜔𝑛
𝜋𝜁 4
− = ln 0.1 = 0.5
1 − 𝜁2 (0.59)𝜔𝑛

4
ln 0.1 2 𝜔𝑛 = = 13.5
𝜁= = 0.59 (0.59) (0.5)
ln 0.1 2 + 𝜋 2

DESIRED Poles: 𝑠 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2 ≈ −8 ± 11𝑗

12
 Example: Step 2
R(s) 1 C(s)
+ Kp+KDs Desired Poles
-
𝑠 2 + 2𝑠 + 2
𝑠 = −8 ± 11𝑗

Step 2: Express the characteristic equation of the closed-loop transfer function

in terms of Kp and KD
𝐾𝑝 + 𝐾𝐷 𝑠
𝑠 2 + 2𝑠 + 2 𝐾𝑝 + 𝐾𝐷 𝑠
𝐺cl = = 2
𝐾𝑝 + 𝐾𝐷 𝑠 𝑠 + 2𝑠 + 2 + 𝐾𝑝 + 𝐾𝐷 𝑠
1+ 2
𝑠 + 2𝑠 + 2

𝐾𝑝 + 𝐾𝐷 𝑠
𝐺cl =
𝑠 2 + (2 + 𝐾𝐷 )𝑠 + (2 + 𝐾𝑝 )

Characteristic equation: 𝑠 2 + 2 + 𝐾𝐷 𝑠 + 2 + 𝐾𝑝 = 0
13
 Example: Steps 3 & 4
R(s) 1 C(s)
+
-
Kp+KDs Desired Poles: 𝑠 = −8 ± 11𝑗
𝑠 2 + 2𝑠 + 2
CE: 𝑠 2 + 2 + 𝐾𝐷 𝑠 + 2 + 𝐾𝑝 = 0

Step 3: Substitute the poles into the characteristic equation

(−8 + 11𝑗)2 + 2 + 𝐾𝐷 (−8 + 11𝑗) + 2 + 𝐾𝑝 = 0

→ (−71 + 𝐾𝑝 − 8𝐾𝐷 ) + 𝑗 11𝐾𝐷 − 154 = 0

Real part: −71 + 𝐾𝑝 − 8𝐾𝐷 = 0 −−−− −(1)

Complex part: 11𝐾𝐷 − 154 = 0 −−− −(2)

Step 4: Solve the real and imaginary parts of the characteristic equation

𝐾𝐷 = 14, 𝐾𝑝 = 183
𝐾𝑝 + 𝐾𝐷 𝑠 = 183 + 14𝑠 = 14(𝑠 + 13.1) 14
 Example: Root Locus
R(s) 1 C(s)
+
-
𝐾𝑝 +𝐾𝐷 𝑠 𝐾𝑝 + 𝐾𝐷 𝑠 = 14(𝑠 + 13.1)
𝑠 2 + 2𝑠 + 2
Desired Poles: 𝑠 = −8 ± 11𝑗
PD controller Im
-8+11j
P Controller Characteristic Equation
𝐾𝑝 -1+j
1+ 2 =0
𝑠 + 2𝑠 + 2 Re
-13.1
-1- j
PD Controller Characteristic Equation
-8-11j
𝐾𝐷 (𝑠 + 13.1)
1+ 2 =0
𝑠 + 2𝑠 + 2
𝐾𝐷 = 14 S-plane
15
Drawbacks of PD Controllers

▪ Cannot be realized with passive components

▪ Sensitive to high frequency noise

PD controller: 𝐺𝑐 = 𝐾𝐷 𝑠 + 𝐾𝑝 = 𝐾𝐷 (𝑠 + 𝑧𝑐 )
lim |𝐺𝑐 𝑠 = 𝑗𝜔 |
Noise Amplification: 𝜔→∞

= lim 𝐾𝐷 𝑗𝜔 + 𝑧𝑐 →∞
𝜔→∞

Alternative controller: Lead compensator

16
 Lead Compensator
R(s) 1 C(s) 𝑠 + 𝑧𝑐
+ 𝐺𝑐 𝐺𝑐 = 𝐾
-
𝑠 2 + 2𝑠 + 2 𝑠 + 𝑝𝑐
Additional pole
lim |𝐺𝑐 𝑠 = 𝑗𝜔 |
Noise Amplification: 𝜔→∞
𝐾 𝑗𝜔 + 𝑧𝑐
= lim
𝜔→∞ (𝑗𝜔 + 𝑝𝑐 )

𝐾 𝑗𝜔
= lim =𝐾
𝜔→∞ (𝑗𝜔)

Finite Amplification → Better than PD controller

17
Designing Lead Compensator
R(s) 𝑠 + 𝑧𝑐
𝐺(𝑠) C(s) 3 unknowns
+ 𝐾
- (𝑠 + 𝑝𝑐 )
𝐾, 𝑧𝑐 , 𝑝𝑐
Lead compensator
𝑝𝑐 > 𝑧𝑐

General Steps
Step 1: Design the desired transient response with a PD controller
𝐾𝐷 𝑠 + 𝐾𝑝 = 𝐾𝐷 (𝑠 + 𝑧𝑐′ )
Step 2: Using the obtained PD controller, specify a value of zc such that:
𝑧𝑐 < 𝑧𝑐′
Step 3: Use the characteristic equation to determine

𝐾, 𝑝𝑐
18
 Example: Steps 1 and 2
R(s) 𝑠 + 𝑧𝑐 1 C(s)
+ 𝐾
- (𝑠 + 𝑝𝑐 ) 𝑠 2 + 2𝑠 + 2
Lead compensator
Desired Poles: 𝑠 = −8 ± 11𝑗

Step 1: Design the desired transient response with a PD controller

𝐾𝐷 𝑠 + 𝐾𝑝 = 14 𝑠 + 13.1 = 𝐾𝐷 (𝑠 + 𝑧𝑐′ )

Step 2: Using the obtained PD controller, specify a value of zc such that:

𝑧𝑐 < 𝑧𝑐′ → 𝑧𝑐 = 10

𝑠 + 10
Lead compensator: 𝐾
(𝑠 + 𝑝𝑐 )
19
 Example: Step 3
R(s) 𝑠 + 10 1 C(s)
+ 𝐾
- (𝑠 + 𝑝𝑐 ) 𝑠 2 + 2𝑠 + 2
Lead compensator
Desired Poles: 𝑠 = −8 ± 11𝑗

Step 3: Use the characteristic equation to determine K and pc

Characteristic Equation: 1 + 𝐺𝑐 𝐺 = 0

𝑠 + 10 1
1+𝐾 2
=0
(𝑠 + 𝑝𝑐 ) 𝑠 + 2𝑠 + 2

(𝑠 + 𝑝𝑐 )(𝑠 2 + 2𝑠 + 2) + 𝐾 𝑠 + 10 = 0
𝑠 3 + (2 + 𝑝𝑐 )𝑠 2 + (2 + 2𝑝𝑐 + 𝐾)𝑠 + (10𝐾 + 2𝑝𝑐 ) = 0

20
 Example: Step 3 (cont)
R(s) 𝑠 + 10 1 C(s)
+ 𝐾
- (𝑠 + 𝑝𝑐 ) 𝑠 2 + 2𝑠 + 2
Lead compensator
Desired Poles: 𝑠 = −8 ± 11𝑗

Step 3: Use the characteristic equation to determine K and pc

CE: 𝑠 3 + (2 + 𝑝𝑐 )𝑠 2 + (2 + 2𝑝𝑐 + 𝐾)𝑠 + (10𝐾 + 2𝑝𝑐 ) = 0

Sub 𝑠 = −8 + 11𝑗 into CE

(2𝐾 − 71𝑝𝑐 + 2262) + 𝑗(−154𝑝𝑐 + 451 + 11𝐾) = 0

Real part: 2𝐾 − 71𝑝𝑐 + 2262 = 0 𝐾 = 669
Imaginary part: −154𝑝𝑐 + 451 + 11𝐾 = 0 𝑝𝑐 = 50.7
𝑠 + 10
𝐺𝑐 = 669
(𝑠 + 50.7) 21
 Example: S-plane
R(s) 1 C(s)
+ 14(𝑠 + 13.1)
-
𝑠 2 + 2𝑠 + 2 Desired Poles: 𝑠 = −8 ± 11𝑗
PD controller

-8+11j Im
PD Controller Characteristic Equation
14(𝑠 + 13.1) 1
1+ 2 = 0,
𝑠 + 2𝑠 + 2  -1+j
(𝑠 + 13.1) (𝑠 + 13.1)
14 2 = 1𝑒 𝑗𝜋 → ∠ =𝜋 -13.1 2 Re
𝑠 + 2𝑠 + 2 2
𝑠 + 2𝑠 + 2
-1- j
2
→ ∠ 𝑠 + 13.1 − ∠ 𝑠 + 2𝑠 + 2 = 𝜋

Phase condition
𝛼 − 𝛽1 − 𝛽2 = 𝜋 S-plane
22
 Example: S-plane
R(s) 669 𝑠 + 10 1 C(s)
+
- 𝑠 + 50.7 𝑠 2 + 2𝑠 + 2 Desired Poles: 𝑠 = −8 ± 11𝑗
Lead compensator

-8+11j Im
Characteristic Equation

669 𝑠 + 10 1 1
1+ = 0,
𝑠 + 50.7 𝑠 2 + 2𝑠 + 2
3  new -1+j

-50.7 -13.1 -10 2 Re

𝑠 + 10 1
∠ =𝜋
𝑠 + 50.7 𝑠 2 + 2𝑠 + 2 -1- j

∠ 𝑠 + 10 − ∠ 𝑠 2 + 2𝑠 + 2 − ∠ 𝑠 + 50.7 = 𝜋 → 𝛼new > 𝛼

𝛼new − 𝛽1 − 𝛽2 − 𝛽3 = 𝜋 S-plane
23
Improving Steady-State
Response

24
Design of PI Controllers
R(s) 𝐾𝐼 C(s)
+ 𝐾𝑝 + G(s)
- 𝑠
PI Controller Plant

H(s)
Sensor

Aim: Upgrade system type

Implement PI controller if transient response is satisfactory
𝐾𝐼
𝐾𝑝 𝑠 +
𝐾𝐼 𝐾𝑝
Note: 𝐾𝑝 + = ~ pole-zero
𝑠 𝑠 cancellation
𝐾𝑝 𝑠 + 𝑧𝑐
= , if zc is small
𝑠
Transient response is slightly affected
25
 Example: Root Locus
R(s) 𝑠+5 C(s)
+ Gc 2
- 𝑠 + 8 𝑠 + 10
Controller Plant

Im
Characteristic Equation for Kp controller
𝑠+5
1 + 𝐾𝑝 2 = 0,
𝑠 + 8 𝑠 + 10

-10 -8 -5 Re

-11.5

26
 Example: Root Locus
R(s) 𝑠+5 C(s)
+ Gc 2
- 𝑠 + 8 𝑠 + 10
Controller Plant

Im
Characteristic Equation for Kp controller
𝑠+5
1 + 𝐾𝑝 2 = 0,
𝑠 + 8 𝑠 + 10

Characteristic Equation for PI controller

-10 -8 -5 -0.5 Re
𝑠 + 0.5 𝑠+5
1 + 𝐾𝑝 2 = 0,
𝑠 𝑠 + 8 𝑠 + 10

PI controller
-11.5 -11.25

The root locus does not change much!

27
Drawbacks of PI Controllers

Cannot be realized with passive components

Alternative controller: Lag compensator

28
 Lag Compensator ~ pole/zero
cancellation
R(s) C(s) 𝑠 + 𝑧𝑐
+
-
𝐺𝑐 𝐺 𝐺𝑐 = 𝐾
𝑠 + 𝑝𝑐
𝑝𝑐 ≪ 𝑧𝑐 ≪ 1
Objective: Enhance position error constant Kp
𝑠 + 𝑧𝑐
𝐾𝑝 = lim 𝐺𝑐 𝐺 = lim 𝐾 𝐺(𝑠)
𝑠→0 𝑠→0 𝑠 + 𝑝𝑐
𝑧𝑐
= 𝐾 𝐺(0) 𝑧
𝐾 𝑐 factor
𝑝𝑐 𝑝𝑐

1
𝑒𝑠𝑠 =
1 + 𝐾𝑝

Error reduces but not fully eliminated

29
 Example: Root Locus
R(s) 𝑠 + 0.2 𝑠+2 C(s)
+ 𝐾
- 𝑠 + 0.02 𝑠 2 + 2𝑠 + 5
Controller Plant

Extra pole Kp 10 folds

1
𝑒𝑠𝑠 =
1 + 𝐾𝑝
Uncompensated
Lag compensator

The root locus does not change much!

30
Improving Transient and
Steady-State Response

31
 PID Controllers
R(s) E(s) 𝐾𝐼 A(s) C(s)
+- 𝐾𝑝 + 𝐾𝐷 𝑠 +
𝑠
G(s) Time Domain
𝑡
PID Controller d𝑒
𝑎 = 𝐾𝑝 𝑒 + 𝐾𝐷 + 𝐾𝐼 න 𝑒 d𝑡
d𝑡 0
H(s)

S-plane Domain
𝐾𝐼 𝐾𝐷 𝑠 2 + 𝐾𝑝 𝑠 + 𝐾𝐼 𝐾𝐷 𝑠 + 𝑧1 𝑠 + 𝑧2
PID: 𝐾𝑝 + 𝐾𝐷 𝑠 + = =
𝑠 𝑠 𝑠
𝑠 + 𝑧2
= 𝐾𝐷 𝑠 + 𝑧1
𝑠
PD controller PI controller

Step 1: Use the PD controller to tune the transient response

Step 2: Use the PI controller to eliminate steady-state error

32
 Lead-Lag Compensators
R(s) 𝐾1 𝑠 + 𝑧1 𝐾2 𝑠 + 𝑧2 C(s)
+ (𝑠 + 𝑝1 ) (𝑠 + 𝑝2 ) G(s)
-
Lead-lag compensator

H(s)

𝐾1 𝑠 + 𝑧1 𝐾2 𝑠 + 𝑧2
Lead-lag compensator:
(𝑠 + 𝑝1 ) (𝑠 + 𝑝2 )

Lead compensator Lag compensator

𝑧1 < 𝑝1 𝑝2 < 𝑧2 ≪ 1

Step 1: Design the lead compensator to tune the transient response

Step 2: Use the lag compensator to reduce steady-state error

33
 Summary

Active Passive
Transient Response PD Lead
Steady-state Response PI Lag
Transient and steady-state PID Lead-Lag

34
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Teaching Plan
Week Lecture Topics Tutorial

8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

2
 Harmonic Inputs
Time Domain
𝑅(𝑠) 𝐶(𝑠)
𝐺(𝑠) Input: 𝑟 = 𝑟0 cos 𝜔𝑡
Transfer
Function Steady-state: 𝑐 = 𝑟0 |𝐺 𝑠 = 𝑗𝜔 | cos 𝜔𝑡 + ∠𝐺 𝑠 = 𝑗𝜔

Magnitude changed by a Phase shift by a factor of

factor of |𝐺 𝑠 = 𝑗𝜔 | ∠𝐺 𝑠 = 𝑗𝜔
3
𝜋
2 3
Input 3
Signals

1
Output 0.5
0
-1
-2
-3 3
T time
 Example
𝑅(𝑠) 2 𝐶(𝑠)

𝑠+2
Transfer Function

Input: 𝑟 𝑡 = 𝑟0 cos 10𝑡

2 2 2 −𝑗(tan −1 5)
𝐺 𝑠 = 10𝑗 = = −1 5) = 𝑒
10𝑗 + 2 104𝑒 𝑗(tan 104

𝑐𝑠𝑠 𝑡 = 𝑟0 𝐺 𝑠 = 𝑗𝜔 cos(10𝑡 + ∠𝐺 𝑠 = 𝑗𝜔 )

2
− tan−1 5
104 4
 Bode Plot
𝑅(𝑠) 𝐶(𝑠)
𝐺(𝑠) Frequency response of the
Transfer
transfer function
Function

Generates Two Log-Scale Plots

20 lg 𝐺 (dB)
1. Magnitude plot
One decade
20log10(|𝐺(𝑠 = 𝑗)|) against lg 𝜔

x-axis is lg 𝜔
∠𝐺 (degrees)

2. Phase plot
∠𝐺(𝑠 = 𝑗) against lg 𝜔

rad
𝜔( ) 5
𝑠
 Motivation
1. Together with Fourier series analysis, it can compute the output of periodic inputs
▪ Example, 𝑟 = 2 sin 3𝑡 + 6 cos 100𝑡 + 𝜋

Output: 𝑐 = 2 𝐺 𝑠 = 3𝑗 sin(3𝑡 + ∠𝐺 𝑠 = 3𝑗 ) + 6 𝐺 𝑠 = 100𝑗 cos 100𝑡 + 𝜋 + ∠𝐺 𝑠 = 100𝑗

2. Analysis:
▪ Resonance
▪ Filtering

3. Designing closed-loop system:

▪ Poles
▪ Steady-state response

4. Reverse engineering:
▪ Identify the transfer function of the system

6
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Sketching

7
Superposition Principle (I)
Objective: Plot two log-scale plots
𝑅(𝑠) 𝐶(𝑠) ς𝑛𝑖=1(𝑠 + 𝑧𝑖 )
𝐺(𝑠) 𝐺 𝑠 =𝐾 𝑚
ς𝑗=1(𝑠 + 𝑝𝑗 )
Transfer Function

Frequency Response
𝑗tan −1 (𝜔 )

ς𝑛𝑖=1(𝑗𝜔 + 𝑧𝑖 ) ς𝑛𝑖=1 𝜔2 + 𝑧𝑖2 𝑒 𝑧𝑖

𝐺 𝑠 = 𝑗𝜔 = 𝐾 𝑚 =𝐾 𝜔
ς𝑗=1(𝑗𝜔 + 𝑝𝑗 ) ς𝑚
𝑗tan−1 ( )
𝑝𝑗
𝑗=1 𝜔2 + 𝑝𝑗2 𝑒

ς𝑛𝑖=1 𝜔 2 + 𝑧𝑖2 −1 𝜔 −1 𝜔
𝑗 σ𝑛 𝑚
𝑖=1 tan (𝑧 )−σ𝑗=1 tan (𝑝 )
=𝐾 𝑒 𝑖 𝑗

ς𝑚
𝑗=1 𝜔 2 + 𝑝𝑗2
∠𝐺(𝑠 = 𝑗)
|𝐺(𝑠 = 𝑗)| 8
Superposition Principle (II)
Transfer function ∠𝑮 (Phase)
𝑛 𝑚
ς𝑛𝑖=1(𝑠 + 𝑧𝑖 ) 𝜔 𝜔
𝐺 𝑠 =𝐾 𝑚 ෍ tan−1 − ෍ tan −1
𝑧𝑖 𝑝𝑗
ς𝑗=1(𝑠 + 𝑝𝑗 ) 𝑖=1 𝑗=1
Linear combination of angles

𝑮 (Magnitude)
ς𝑛𝑖=1 ς𝑛𝑖=1 𝜔 2 + 𝑧𝑖2
𝜔2 + 𝑧𝑖2 Log-scale
𝐾 20 lg 𝐾
ς𝑚 ς𝑚 𝜔 2 + 𝑝𝑗2
𝑗=1 𝜔 2 + 𝑝𝑗2 𝑗=1

𝑛 𝑚

20 lg |𝐺| = 20 lg 𝐾 + ෍ 20 lg 𝜔 2 + 𝑧𝑖2 − ෍ 20 lg 𝜔 2 + 𝑝𝑗2

𝑖=1 𝑗=1

Linear combination of magnitudes 9

 Principal Components
1. Constants (𝐾)
2. Differentiator (𝑠)
1
3. Integrator
𝑠

𝑎
4. First order systems
𝑠+𝑎

2
𝜔𝑛
5. Second order systems 2
𝑠+2𝜁𝜔𝑛 𝑠+𝜔𝑛

10
 20 lg 𝐺 (dB)
Plotting Constants
𝑅(𝑠) 𝐶(𝑠)
20 lg 𝐾 𝐾
Transfer Function

0.01 0.1 1 10 100 𝜔(rad/s)

𝐺 𝑠 =𝐾

𝐺 𝑠 = 𝑗𝜔 = 𝐾𝑒 𝑗0
∠𝐺 (rad)

Magnitude
20 lg |𝐺 𝑗𝜔 | = 20 lg 𝐾
0.01 0.1 1 10 100 𝜔(rad/s)
Phase
∠𝐺 𝑗𝜔 = 0
11
 20 lg 𝐺 (dB)
Plotting Differentiator
+20 dB/dec 𝑅(𝑠) 𝐶(𝑠)
𝑠
Transfer Function

0.01 0.1 1 10 100 𝜔(rad/s) 𝐺 𝑠 =𝑠

−40 𝐺 𝑠 = 𝑗𝜔 = 𝑗𝜔
𝜋
𝑗2
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = 𝜔𝑒
𝜋
2 Magnitude
20 lg |𝐺 𝑗𝜔 | = 20 lg 𝜔
0.01 0.1 1 10 100 𝜔(rad/s)
Phase
𝜋
∠𝐺 𝑗𝜔 =
2
12
 20 lg 𝐺 (dB)
Plotting Integrator
40 𝑅(𝑠) 1 𝐶(𝑠)
𝑠
𝜔(rad/s) Transfer Function

0.01 0.1 1 10 100 1 1

𝐺 𝑠 = , 𝐺 𝑠 = 𝑗𝜔 =
𝑠 𝑗𝜔
−20 dB/dec
1 𝑗(−𝜋)
𝐺 𝑠 = 𝑗𝜔 = 𝑒 2
∠𝐺 (rad) 𝜔
Magnitude
𝜔(rad/s)
20 lg |𝐺 𝑗𝜔 | = −20 lg 𝜔
0.01 0.1 1 10 100
𝜋 Phase
−
2 𝜋
∠𝐺 𝑗𝜔 = −
2
13
 Plotting First Order Pole (I)
20 lg 𝐺 (dB)

𝑅(𝑠) 𝑎
𝐶(𝑠)
𝑠+𝑎
Transfer Function

0.01 0.1 1 𝑎 10 100 𝜔(rad/s) 𝑎

𝐺 𝑠 = 𝑗𝜔 =
𝑗𝜔 + 𝑎
When  << a
𝑎
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = = 1𝑒 𝑗0
𝑎
Magnitude

0.01 0.1 0.1𝑎 1 𝑎 10 10𝑎 100 𝜔(rad/s)

20 lg |𝐺 𝑗𝜔 | = 0
Phase
∠𝐺 𝑗𝜔 = 0
14
 Plotting First Order Pole (II)
20 lg 𝐺 (dB)

𝑅(𝑠) 𝑎
𝐶(𝑠)
𝑠+𝑎
Transfer Function
𝑎
0.01 0.1 1 𝑎 10 100 𝜔(rad/s) 𝐺 𝑠 = 𝑗𝜔 =
𝑗𝜔 + 𝑎

−20 dB/dec When  >> a

𝑎 𝑎 𝑗(−𝜋)
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = = 𝑒 2
𝑗𝜔 𝜔
Magnitude
𝑎
0.1 0.1𝑎 1 𝑎 10 10𝑎 100 𝜔(rad/s) 20 lg |𝐺 𝑗𝜔 | = 20 lg
𝜋 0.01 𝜔
−
4 𝜋 = 20 lg 𝑎 − 20 lg 𝜔
𝜋 − rad/dec
− 4
2 Phase
𝜋
∠𝐺 𝑗𝜔 = − 15
2
 First Order Pole: Example
𝑅(𝑠) 10 𝐶(𝑠) 𝑎 = 10
Corner Frequencies
𝑠 + 10
Transfer Function 𝜔 = 1, 10, 100
20 lg 𝐺 (dB)

−20 dB/dec

𝜔(rad/s)
∠𝐺 (degrees)

−45°/dec

16

𝜔(rad/s)
 Plotting First Order Zero (I)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝑠 + 𝑎 𝐶(𝑠)
𝑎
Transfer Function
𝜔(rad/s) 𝑗𝜔 + 𝑎
𝐺 𝑠 = 𝑗𝜔 =
0.01 0.1 1 𝑎 10 100 𝑎
When  << a
𝑎
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = = 1𝑒 𝑗0
𝑎
Magnitude
20 lg |𝐺 𝑗𝜔 | = 0
0.01 0.1 0.1𝑎 1 𝑎 10 10𝑎 100 𝜔(rad/s)
Phase
∠𝐺 𝑗𝜔 = 0
17
 Plotting First Order Zero (II)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝑠 + 𝑎 𝐶(𝑠)
+20 dB/dec 𝑎
Transfer Function
𝜔(rad/s) 𝑗𝜔 + 𝑎
𝐺 𝑠 = 𝑗𝜔 =
0.01 0.1 1 𝑎 10 100 𝑎
When  >> a
𝑗𝜔 𝜔 𝑗(𝜋)
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = = 𝑒 2
𝜋 𝑎 𝑎
𝜋
2 + rad/dec
𝜋 4 Magnitude
4
𝜔
20 lg |𝐺 𝑗𝜔 | = 20 lg
𝑎
0.01 0.1 0.1𝑎 1 𝑎 10 10𝑎 100 𝜔(rad/s)
= 20 lg 𝜔 − 20 lg 𝑎
Phase
𝜋
∠𝐺 𝑗𝜔 = 18
2
 First Order Zero: Example
𝑅(𝑠) 𝑠 + 10 𝐶(𝑠) 𝑎 = 10
Corner Frequencies
10
Transfer Function 𝜔 = 1, 10, 100
20 lg 𝐺 (dB)

+20 dB/dec

𝜔(rad/s)
∠𝐺 (degrees)

+45°/dec

19
𝜔(rad/s)
 Plotting Second Order Pole (I)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝜔𝑛2 𝐶(𝑠)
𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2
Transfer Function
𝜔(rad/s)
𝜔𝑛2
0.01 0.1 1 𝜔𝑛 10 100 𝐺 𝑠 = 𝑗𝜔 =
−𝜔 2 + 𝑗2𝜁𝜔𝑛 𝜔 + 𝜔𝑛2
When  << n
𝜔𝑛2
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = 2 = 1𝑒 𝑗0
𝜔𝑛
Magnitude
0.01 0.1 0.1𝜔𝑛 1 𝜔𝑛 10 10𝜔𝑛 100 𝜔(rad/s) 20 lg |𝐺 𝑗𝜔 | = 0
Phase
∠𝐺 𝑗𝜔 = 0
20
 Plotting Second Order Pole (II)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝜔𝑛2 𝐶(𝑠)
𝑠 2 + 2𝜁𝜔𝑛 𝑠 + 𝜔𝑛2
Transfer Function
𝜔(rad/s)
𝜔𝑛2
0.01 0.1 1 𝜔𝑛 10 100 𝐺 𝑠 = 𝑗𝜔 =
−𝜔 2 + 𝑗2𝜁𝜔𝑛 𝜔 + 𝜔𝑛2
−40 dB/dec When  >> n
𝜔𝑛2 𝜔𝑛2 𝑗(−𝜋)
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = 2
= 2𝑒
−𝜔 𝜔
Magnitude 2
𝜔𝑛
𝜋 0.01 0.1 0.1𝜔𝑛 1 𝜔𝑛 10 10𝜔𝑛 100 𝜔(rad/s) 20 lg |𝐺 𝑗𝜔 | = 20 lg 2
−
𝜔
2 𝜋
− rad/dec
2 = 40 lg 𝜔𝑛 − 40 lg 𝜔
−𝜋
Phase 21
∠𝐺 𝑗𝜔 = −𝜋
 Second Order Pole: Example
𝑅(𝑠) 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
𝑠 2 + 2𝜁𝑠 + 1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)

−40 dB/dec
∠𝐺 (degrees)

−90°/dec

22
 Second Order Pole: Example
𝑅(𝑠) 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
𝑠 2 + 2𝜁𝑠 + 1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)

𝜁 = 0.01
𝜁 = 0.1
𝜁 = 0.3

𝜔(rad/s)
∠𝐺 (degrees)

𝜁 = 0.01

𝜁 = 0.1
𝜁 = 0.3

23
𝜔(rad/s)
 Plotting Second Order Zero (I)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝑠 2 + 2𝜁𝜔 𝑠 + 𝜔2 𝐶(𝑠)
𝑛 𝑛
𝜔𝑛2
Transfer Function
−𝜔2 + 𝑗2𝜁𝜔𝑛 𝜔 + 𝜔𝑛2
0.01 0.1 1 𝜔𝑛 10 100 𝐺 𝑠 = 𝑗𝜔 =
𝜔𝑛2
𝜔(rad/s)
When  << n
𝜔𝑛2
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = 2 = 1𝑒 𝑗0
𝜔𝑛
Magnitude
0.01 0.1 0.1𝜔𝑛 1 𝜔𝑛 10 10𝜔𝑛 100 20 lg |𝐺 𝑗𝜔 | = 0
𝜔(rad/s)
Phase
∠𝐺 𝑗𝜔 = 0
24
 Plotting Second Order Zero (II)
20 lg 𝐺 (dB)
𝑅(𝑠) 𝑠 2 + 2𝜁𝜔 𝑠 + 𝜔2 𝐶(𝑠)
𝑛 𝑛
+40 dB/dec 𝜔𝑛2
Transfer Function
−𝜔2 + 𝑗2𝜁𝜔𝑛 𝜔 + 𝜔𝑛2
0.01 0.1 1 𝜔𝑛 10 100 𝐺 𝑠 = 𝑗𝜔 =
𝜔𝑛2
𝜔(rad/s)
When  >> n
−𝜔2 𝜔2 𝑗𝜋
∠𝐺 (rad) 𝐺 𝑠 = 𝑗𝜔 = 2 = 2 𝑒
𝜋 𝜔𝑛 𝜔𝑛
𝜋 𝜋
2
+
2
rad/dec Magnitude
0.01 0.1 0.1𝜔𝑛 1 𝜔𝑛 10 10𝜔𝑛 100
𝜔2
20 lg |𝐺 𝑗𝜔 | = 20 lg 2
𝜔(rad/s)
𝜔𝑛
= 40 lg 𝜔 − 40 lg 𝜔𝑛
Phase
25
∠𝐺 𝑗𝜔 = 𝜋
 Second Order Zero: Example
𝑅(𝑠) 𝑠 + 2𝜁𝑠 + 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)
∠𝐺 (degrees)

26
 Second Order Zero: Example
𝑅(𝑠) 𝑠 + 2𝜁𝑠 + 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)

𝜁 = 0.3 𝜁 = 0.1

𝜁 = 0.01
∠𝐺 (degrees)

𝜁 = 0.1

𝜁 = 0.3
27
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Example

28
 Complex Function: Asymptotes
20 lg 𝐺 (dB)
+20 dB/dec

+20 dB/dec
−20 dB/dec
20
Generate Bode plot:
0.01 0.1 1 10 100 1000
−20 𝜔(rad/s)
1000𝑠
−40 𝐺=
𝑠 2 + 𝑠 + 100
−40 dB/dec
∠𝐺 (rad) 1000 100
𝐺= (𝑠) 2
100 𝑠 + 𝑠 + 100
𝜋
2 𝜔(rad/s)
0.01 0.1 1 10 100 1000
π
−𝜋 − rad/dec
2

29
 Complex Function: Sketch
20 lg 𝐺 (dB)

+20 dB/dec
−20 dB/dec

Generate Bode plot:

0.01 0.1 1 10 100 1000
−20 𝜔(rad/s)
1000𝑠
𝐺=
𝑠 2 + 𝑠 + 100
∠𝐺 (rad) 1000 100
𝐺= (𝑠) 2
𝜋
100 𝑠 + 𝑠 + 100
2 𝜔(rad/s)
𝜋 0.01 0.1 1 10 100 1000
−
2

30
 Complex Function: MATLAB
20 lg 𝐺 (dB)

𝜔(rad/s)
∠𝐺 (deg)

𝜔(rad/s)

31
 Complex Function: Reading
𝑅(𝑠) 𝐶(𝑠)
𝐺(𝑠)
Transfer Function

Compute the output if the input is: 𝑟 𝑡 = 2 cos 𝑡 + 5 cos 10𝑡

𝑐 𝑡 = 2|𝐺 𝑠 = 𝑗 | cos(𝑡 + ∠𝐺 𝑠 = 𝑗 )

+5|𝐺 𝑠 = 10𝑗 | cos(10𝑡 + ∠𝐺 𝑠 = 10𝑗 )

32
 Complex Function: MATLAB
20
20 lg 𝐺 (dB)

𝐺 = 1020 = 10 60 dB
60
20 dB 𝐺 = 1020 = 1000

𝜔(rad/s)

𝜋 0 rad
∠𝐺 (deg)

rad
2

𝜔(rad/s)

33
 Complex Function: Reading
𝑅(𝑠) 𝐶(𝑠)
𝐺(𝑠)
Transfer Function

Compute the output if the input is: 𝑟 𝑡 = 2 cos 𝑡 + 5 cos 10𝑡

𝑐 𝑡 = 2|𝐺 𝑠 = 𝑗 | cos(𝑡 + ∠𝐺 𝑠 = 𝑗 )
𝜋
10
2
+5|𝐺 𝑠 = 10𝑗 | cos(10𝑡 + ∠𝐺 𝑠 = 10𝑗 )

1000 0

𝜋
𝑐 𝑡 = 20 cos(𝑡 + ) + 5000 cos(10𝑡) 34
2
 Summary
1. Bode plot is the frequency response of the system
a) Read the bode plot

2. Sketching techniques
a) Superposition principle
b) Sketch the principal components
c) Combine all the components together

35
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

MA3005: Control Theory

Part II: Controller Design

Assistant Professor Guo Zhan Lum

1
 Teaching Plan
Week Lecture Topics Tutorial

8 Recap & Introduction to controller design/System Types 6

9 Root Locus: Introduction and Sketching 7

10 Root Locus: Controller Design 8
11 Bode Plot: Introduction and Sketching 9
12 Bode Plot: Analysis 10
13 Revision 11

2
 Bode Plot Recap
𝑅(𝑠) 𝐶(𝑠)
𝐺(𝑠) Frequency response of the
Transfer
transfer function
Function

Generates Two Log-Scale Plots

20 lg 𝐺 (dB)
1. Magnitude plot
One decade
20log10(|𝐺(𝑠 = 𝑗)|) against lg 𝜔

x-axis is lg 𝜔
∠𝐺 (degrees)

2. Phase plot
∠𝐺(𝑠 = 𝑗) against lg 𝜔

rad
𝜔( ) 3
𝑠
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Second Order

Systems

4
 Second Order System
𝑅(𝑠) 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
𝑠 2 + 2𝜁𝑠 + 1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)

𝜁 = 0.01
𝜁 = 0.1
𝜁 = 0.3

𝜔(rad/s)
∠𝐺 (degrees)

𝜁 = 0.01

𝜁 = 0.1
𝜁 = 0.3

5
𝜔(rad/s)
 Second Order System
𝑅(𝑠) 1 𝐶(𝑠) 𝜔𝑛 = 1
Corner Frequencies
𝑠 2 + 2𝜁𝑠 + 1
Transfer Function 𝜔 = 0.1, 1, 10
20 lg 𝐺 (dB)

Resonance frequency 𝜔𝑟
𝜔𝑟 = 𝜔𝑛 1 − 2𝜁 2

𝜔(rad/s)
∠𝐺 (degrees)

𝜁 = 0.01

𝜁 = 0.1
𝜁 = 0.3

6
𝜔(rad/s)
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Filters

7
𝑋 dB
Low-Pass Filter
20 lg |𝐺| (dB)

𝑋 − 3 dB line
Pass band No pass band

𝜔c = 20

𝜔 (rad/s)

Input Signal Output Signal

𝑟 𝑡 = cos(𝜔𝑡) → 𝑐 𝑡 = |𝐺(𝑠 = 𝑗𝜔)|cos(𝜔𝑡 + ∠𝐺)

Relationship of 𝐺 𝑠 = 𝑗𝜔 and cut-off frequency (𝜔c )

≪ 𝐺 𝑠 = 𝑗𝜔low , 𝜔 ≥ 𝜔𝑐
𝐺 𝑠 = 𝑗𝜔 =ቊ
~ 𝐺 𝑠 = 𝑗𝜔low , 𝜔 < 𝜔𝑐 8
 High-Pass Filter
𝑋 dB
20 lg |𝐺| (dB)

𝑋 − 3 dB line

No pass band Pass band

𝜔c = 7

𝜔 (rad/s)

Input Signal Output Signal

𝑟 𝑡 = cos(𝜔𝑡) → 𝑐 𝑡 = |𝐺(𝑠 = 𝑗𝜔)|cos(𝜔𝑡 + ∠𝐺)

Relationship of 𝐺 𝑠 = 𝑗𝜔 and cut-off frequency (𝜔c )

~ 𝐺 𝑠 = 𝑗𝜔high , 𝜔 ≥ 𝜔𝑐
𝐺 𝑠 = 𝑗𝜔 =ቐ 9
≪ 𝐺 𝑠 = 𝑗𝜔high , 𝜔 < 𝜔𝑐
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Transfer Function

Approximation

10
 Reverse Engineering
AIM: Use the Bode plots to approximate the transfer function
63 dB
20 lg |𝐺| (dB)

𝜔=3

–60 dB/dec
No slope
𝜔 = 400
No differentiator
2nd order
and integrator –40 dB/dec
pole

0 degree 𝜔 (rad/s)
1st order zero
∠𝐺 (degree)

𝜔 = 0.3
180 deg
270 deg 𝜔 = 4000

𝜔 (rad/s)
Estimated transfer function
302 3 𝑠 + 400
𝐺 𝑠 = 2 𝐾 11
𝑠 + 2 0.01 30 𝑠 + 302 𝑠+3 400
 Reverse Engineering (cont..)
Estimated transfer function DC gain from plot
900𝐾 3 𝑠 + 400
𝐺 𝑠 = 63.5 dB
𝑠 2 + 0.6𝑠 + 900 𝑠 + 3 400

63.5
20 lg |lim 𝐺 𝑠 | = 63.5 |lim 𝐺 𝑠 | = 10 20
𝑠→0 𝑠→0

900𝐾 3 400 63.5 63.5

= 10 20 → 𝐾= 10 20 = 1500
900 3 400

Estimated Actual
10125(𝑠 + 400) 15000(𝑠 + 450)
𝐺 𝑠 = 𝐺 𝑠 =
(𝑠 2 + 0.6𝑠 + 900)(𝑠 + 3) (𝑠 2 + 𝑠 + 900)(𝑠 + 5)

12
 Example
Use the Bode plots to approximate the transfer function
𝜔𝑟 = 11.5,  ≈ 0.3
20 lg |𝐺| (dB)

16.5 dB -20 dB/dec

-60 dB/dec
An integrator

𝜔 (rad/s)
∠𝐺 (degree)

-90 deg

-270 deg

𝜔 (rad/s)
Estimated transfer function
1
𝐺 𝑠 = 𝐺 𝐾 13
𝑠 1
 Example (cont..)
Estimated transfer function r  Low freq gain
1
𝐺 𝑠 = 𝐺 𝐾 11.5 rad ~0.3 16.5 dB
𝑠 1
Resonance Frequency Analysis
12.72
𝜔𝑟 = 𝜔𝑛 1 − 2𝜁 2 = 11.5 → 𝜔𝑛 = 12.7 → 𝐺1 = 2
𝑠 + 2 0.3 12.7 𝑠 + 12.72
Low Frequency Gain Analysis
16.5
20 lg |lim 𝐺 𝑠 | = 16.5 → |lim 𝐺 𝑠 | = 10 20
𝑠→𝑗 𝑠→𝑗
𝐾 161 16.5
→ = 10 20
𝑗 −1 + 7.6𝑗 + 161 → 𝐾 = 6.6
Estimated Actual
1065 1000
𝐺 𝑠 = 2 𝐺 𝑠 = 2
𝑠 + 7.6𝑠 + 161 𝑠 𝑠 + 5𝑠 + 150 𝑠

14
 D(s)

R(s) +
C(s)
+
-
Controller + Plant

Bode Plot: Unity Feedback

15
 Closed-Loop Response

R(s) + C(s)
Gc(s) P(s)
- Controller Plant
G (open-loop)

Bode plot of G (open-loop transfer function) also reveals the

response of the unity feedback system

16
 Gain and Phase Margins
20 lg |𝐺| (dB)

GM = +18dB Gain Margin (Stable GOL)

𝜔gm 1.Occurs at the frequency
𝜔pm when ∠𝐺 = −𝜋 rad (𝜔gm )
𝜔 (rad/s)

2.Gain Margin (GM)

∠𝐺 (degree)

𝛾 = +70° = 0 − 𝐺 𝑠 = 𝑗𝜔gm
dB

𝜔 (rad/s)
Phase Margin (Stable GOL)
Stability
1. Occurs at the frequency
Unity feedback is stable if and only if: when |𝐺| = 0 dB (𝜔pm )
𝛾 > 0 & GM > 0
2. Phase Margin (𝛾)
Closed-loop 𝜻𝐜𝐥 = ∠𝐺 𝑠 = 𝑗𝜔pm − (−𝜋)
17
𝛾 ↑ implies that 𝜁cl ↑
 Bandwidth
𝑋 dB
20 lg |𝐺| (dB)

X-3 dB

𝜔BW = 16 rad/s
Bandwidth
𝜔 (rad/s)
The frequency when 𝐺 = 𝑋 − 3 dB.
∠𝐺 (degree)

We denote it as 𝜔BW

𝜔 (rad/s)

Closed-loop 𝝎𝒏,𝐜𝐥 Closed-loop Poles

𝜔BW ↑ implies that 𝜔𝑛,cl ↑ 2
𝑠1,2 = −𝜁cl 𝜔𝑛,cl ± 𝑗𝜔𝑛,cl 1 − 𝜁Cl

18
 Steady-State Errors (Type 0)
20 lg |𝐺| (dB)

20 lg |𝐺OL 𝑠 = 0.1𝑗 | ~ 20 lg |𝐺OL 𝑠 = 0𝑗 | = 20 lg 𝐾𝑝

𝜔 (rad/s)

Type 0 Systems
1
Steady-state error ess (step inputs): 1 + 𝐾𝑝

Static position error constant: 𝐾𝑝 = lim 𝐺OL

𝑠→0
14
e.g., 14 dB = 20 lg Kp → 𝐾𝑝 = 1020 =5
1 1
𝑒𝑠𝑠 = = 19
1+5 6
 Steady-State Errors (Type 1)
−20 dB/dec
20 lg |𝐺| (dB)

0 dB line

ω𝑣 = 𝐾𝑣

𝜔 (rad/s)

Type 1 Systems
1
Steady-state error ess (ramp inputs): 𝐾𝑣

e.g., 𝐾𝑣 = 𝜔𝑣 = 21
1
𝑒𝑠𝑠 =
21
20
 Steady-State Errors (Type 2)
−40 dB/dec
20 lg |𝐺| (dB)

0 dB line
ω𝑎 = 𝐾𝑎

𝜔 (rad/s)

Type 2 Systems
1
Steady-state error ess (parabolic inputs): 𝐾𝑎

e.g., 𝜔𝑎 = 10.8 → 𝐾𝑎 = 116

1
𝑒𝑠𝑠 =
116 21
 Summary
1. Low-pass and high-pass filters

2. Reverse engineering

3. Unity feedback systems (based on OL Bode plots)

a) Stability
b) Transient response
c) Steady-state errors

22