page 1

Dynamic System Modeling and Control

by

Hugh Jack

(Draft Version 2.4, November 26, 2003)

page 1

1.

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1
1.1 1.2 1.3 1.4 BASIC TERMINOLOGY EXAMPLE SYSTEM SUMMARY PRACTICE PROBLEMS 1.1 1.3 1.3 1.3

2.

TRANSLATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1
2.1 2.2 INTRODUCTION MODELING 2.2.1 Free Body Diagrams 2.2.2 Mass and Inertia 2.2.3 Gravity and Other Fields 2.2.4 Springs 2.2.5 Damping and Drag 2.2.6 Cables And Pulleys 2.2.7 Friction 2.2.8 Contact Points And Joints SYSTEM EXAMPLES OTHER TOPICS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 2.1 2.3 2.4 2.4 2.7 2.10 2.16 2.19 2.21 2.23 2.24 2.34 2.35 2.35 2.40 2.44

2.3 2.4 2.5 2.6 2.7 2.8

3.

ANALYSIS OF DIFFERENTIAL EQUATIONS . . . . . . . . . . . . 3.1
3.1 3.2 3.3 INTRODUCTION EXPLICIT SOLUTIONS RESPONSES 3.3.1 First-order 3.3.2 Second-order 3.3.3 Other Responses RESPONSE ANALYSIS NON-LINEAR SYSTEMS 3.5.1 Non-Linear Differential Equations 3.5.2 Non-Linear Equation Terms 3.5.3 Changing Systems CASE STUDY SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 3.1 3.2 3.15 3.16 3.22 3.27 3.30 3.32 3.33 3.37 3.40 3.46 3.50 3.50 3.55 3.60

3.4 3.5

3.6 3.7 3.8 3.9 3.10

4.

NUMERICAL ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1

page 2

4.1 4.2 4.3

4.4 4.5 4.30 4.6

INTRODUCTION 4.1 THE GENERAL METHOD 4.1 4.2.1 State Variable Form 4.2 NUMERICAL INTEGRATION 4.10 4.3.1 Numerical Integration With Tools 4.10 4.3.2 Numerical Integration 4.15 4.3.3 Taylor Series 4.21 4.3.4 Runge-Kutta Integration 4.23 SYSTEM RESPONSE 4.29 4.4.1 Steady-State Response 4.30 DIFFERENTIATION AND INTEGRATION OF EXPERIMENTAL ADVANCED TOPICS 4.6.1 Switching Functions 4.6.2 Interpolating Tabular Data 4.6.3 Modeling Functions with Splines 4.6.4 Non-Linear Elements CASE STUDY SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 4.33 4.33 4.36 4.37 4.39 4.39 4.46 4.47 4.50 4.60

DATA

4.7 4.8 4.9 4.10 4.11

5.

ROTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1
5.1 5.2 INTRODUCTION MODELING 5.2.1 Inertia 5.2.2 Springs 5.2.3 Damping 5.2.4 Levers 5.2.5 Gears and Belts 5.2.6 Friction 5.2.7 Permanent Magnet Electric Motors OTHER TOPICS DESIGN CASE SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 5.1 5.2 5.3 5.7 5.12 5.14 5.15 5.19 5.22 5.23 5.23 5.28 5.28 5.35 5.44

5.3 5.4 5.5 5.6 5.7 5.8

6.

INPUT-OUTPUT EQUATIONS . . . . . . . . . . . . . . . . . . . . . . . . . 6.1
6.1 6.2 6.3 INTRODUCTION THE DIFFERENTIAL OPERATOR INPUT-OUTPUT EQUATIONS 6.1 6.1 6.4

page 3

6.4 6.5 6.6 6.7 6.8 6.9

6.3.1 Converting Input-Output Equations to State Equations 6.3.2 Integrating Input-Output Equations DESIGN CASE SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSGINMENT PROBLEMS REFERENCES

6.6 6.9 6.11 6.20 6.20 6.22 6.26 6.27

7.

ELECTRICAL SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1
7.1 7.2 INTRODUCTION MODELING 7.2.1 Resistors 7.2.2 Voltage and Current Sources 7.2.3 Capacitors 7.2.4 Inductors 7.2.5 Op-Amps IMPEDANCE EXAMPLE SYSTEMS PERMANENT MAGNET DC MOTORS INDUCTION MOTORS BRUSHLESS SERVO MOTORS FILTERS OTHER TOPICS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 7.1 7.1 7.2 7.4 7.8 7.10 7.11 7.16 7.18 7.26 7.28 7.29 7.32 7.33 7.33 7.34 7.38 7.43

7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13

8.

FEEDBACK CONTROL SYSTEMS . . . . . . . . . . . . . . . . . . . . . . 8.1
8.1 8.2 8.3 INTRODUCTION TRANSFER FUNCTIONS CONTROL SYSTEMS 8.3.1 PID Control Systems 8.3.2 Manipulating Block Diagrams 8.3.3 A Motor Control System Example 8.3.4 System Error 8.3.5 Controller Transfer Functions 8.3.6 Feedforward Controllers 8.3.7 State Equation Based Systems 8.3.8 Cascade Controllers SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS 8.1 8.1 8.3 8.5 8.7 8.12 8.17 8.21 8.21 8.22 8.24 8.24 8.24 8.31

8.4 8.5 8.6

page 4

8.7

ASSIGNMENT PROBLEMS

8.37

9.

PHASOR ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1
9.1 9.2 9.3 9.4 9.5 9.6 9.7 INTRODUCTION PHASORS FOR STEADY-STATE ANALYSIS VIBRATIONS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 9.1 9.1 9.8 9.10 9.10 9.12 9.14

10.

BODE PLOTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1
10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 INTRODUCTION BODE PLOTS SIGNAL SPECTRUMS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS LOG SCALE GRAPH PAPER 10.1 10.5 10.21 10.22 10.22 10.25 10.35 10.36

11.

ROOT LOCUS ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1
11.1 11.2 11.3 11.4 11.5 11.6 INTRODUCTION ROOT-LOCUS ANALYSIS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 11.1 11.1 11.9 11.10 11.13 11.24

12.

NONLINEAR SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1
12.1 12.2 12.3 INTRODUCTION SOURCES OF NONLINEARITY 12.2.1 Non-Linear Relationships NON-LINEAR ELEMENTS 12.3.1 Time Variant 12.3.2 Switching 12.3.3 Deadband 12.3.4 Hysteresis and Slip SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASIGNMENT PROBLEMS 12.1 12.1 12.1 12.2 12.3 12.3 12.4 12.7 12.8 12.8 12.8 12.8

12.4 12.5 12.6 12.7

13.

ANALOG INPUTS AND OUTPUTS . . . . . . . . . . . . . . . . . . . . 13.1

page 5

13.1 13.2 13.3 13.4

13.5 13.6 13.7 13.8 13.9

INTRODUCTION ANALOG INPUTS ANALOG OUTPUTS NOISE REDUCTION 13.4.1 Shielding 13.4.2 Grounding CASE STUDY SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS

13.1 13.3 13.9 13.12 13.12 13.14 13.15 13.15 13.15 13.15 13.16

14.

CONTINUOUS SENSORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.1
14.1 14.2 INTRODUCTION 14.1 INDUSTRIAL SENSORS 14.2 14.2.1 Angular Displacement 14.3 Potentiometers 14.3 14.2.2 Encoders 14.4 Tachometers 14.8 14.2.3 Linear Position 14.8 Potentiometers 14.8 Linear Variable Differential Transformers (LVDT)14.9 Moire Fringes 14.11 Accelerometers 14.12 14.2.4 Forces and Moments 14.15 Strain Gages 14.15 Piezoelectric 14.18 14.2.5 Liquids and Gases 14.20 Pressure 14.21 Venturi Valves 14.22 Coriolis Flow Meter 14.23 Magnetic Flow Meter 14.24 Ultrasonic Flow Meter 14.24 Vortex Flow Meter 14.24 Positive Displacement Meters 14.25 Pitot Tubes 14.25 14.2.6 Temperature 14.25 Resistive Temperature Detectors (RTDs) 14.26 Thermocouples 14.26 Thermistors 14.28 Other Sensors 14.30 14.2.7 Light 14.30 Light Dependant Resistors (LDR) 14.30 14.2.8 Chemical 14.31 pH 14.31

page 6

Conductivity 14.3 14.4 14.5 14.6 14.7 14.8 14.9 14.2.9 Others INPUT ISSUES SENSOR GLOSSARY SUMMARY REFERENCES PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS

14.31 14.32 14.32 14.37 14.38 14.39 14.39 14.40 14.42

15.

CONTINUOUS ACTUATORS . . . . . . . . . . . . . . . . . . . . . . . . . 15.1
15.1 15.2 INTRODUCTION ELECTRIC MOTORS 15.2.1 Basic Brushed DC Motors 15.2.2 AC Motors 15.2.3 Brushless DC Motors 15.2.4 Stepper Motors 15.2.5 Wound Field Motors HYDRAULICS OTHER SYSTEMS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 15.1 15.1 15.2 15.7 15.12 15.14 15.16 15.20 15.21 15.22 15.22 15.23 15.23

15.3 15.4 15.5 15.6 15.7 15.8

16.

MOTION CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.1
16.1 16.2 INTRODUCTION MOTION PROFILES 16.2.1 Velocity Profiles 16.2.2 Position Profiles MULTI AXIS MOTION 16.3.1 Slew Motion Interpolated Motion 16.3.2 Motion Scheduling PATH PLANNING CASE STUDIES SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 16.1 16.2 16.2 16.11 16.14 16.15 16.16 16.17 16.19 16.21 16.23 16.23 16.24 16.25

16.3

16.4 16.5 16.6 16.7 16.8 16.9

17.

LAPLACE TRANSFORMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.1
17.1 17.2 INTRODUCTION APPLYING LAPLACE TRANSFORMS 17.1 17.3

page 7

17.3 17.4 17.5 17.6

17.7

17.8 17.9 17.10 17.11 17.12 17.13

17.2.1 A Few Transform Tables MODELING TRANSFER FUNCTIONS IN THE s-DOMAIN FINDING OUTPUT EQUATIONS INVERSE TRANSFORMS AND PARTIAL FRACTIONS EXAMPLES 17.6.1 Mass-Spring-Damper Vibration 17.6.2 Circuits ADVANCED TOPICS 17.7.1 Input Functions 17.7.2 Initial and Final Value Theorems A MAP OF TECHNIQUES FOR LAPLACE ANALYSIS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS REFERENCES

17.4 17.9 17.11 17.14 17.21 17.21 17.23 17.25 17.25 17.26 17.27 17.28 17.28 17.31 17.35 17.36

18.

CONTROL SYSTEM ANALYSIS . . . . . . . . . . . . . . . . . . . . . . 18.1
18.1 18.2 INTRODUCTION 18.1 CONTROL SYSTEMS 18.1 18.2.1 PID Control Systems 18.3 18.2.2 Analysis of PID Controlled Systems With Laplace Transforms 18.2.3 Finding The System Response To An Input 18.2.4 Controller Transfer Functions ROOT-LOCUS PLOTS 18.3.1 Approximate Plotting Techniques DESIGN OF CONTINUOUS CONTROLLERS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 18.8 18.13 18.13 18.17 18.21 18.21 18.22 18.27 18.27

18.5

18.3 18.4 18.5 18.6 18.7 18.8

19.

CONVOLUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.1
19.1 19.2 19.3 19.4 19.5 19.6 19.7 19.8 19.9 19.10 INTRODUCTION UNIT IMPULSE FUNCTIONS IMPULSE RESPONSE CONVOLUTION NUMERICAL CONVOLUTION LAPLACE IMPULSE FUNCTIONS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 19.1 19.1 19.3 19.5 19.6 19.9 19.10 19.10 19.10 19.10

page 8

20.

STATE SPACE ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.1
20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 INTRODUCTION OBSERVABILITY CONTROLLABILITY SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS BIBLIOGRAPHY 20.1 20.11 20.13 20.16 20.16 20.16 20.16 20.17

21.

STATE SPACE CONTROLLERS . . . . . . . . . . . . . . . . . . . . . . . 21.1
21.1 21.2 21.3 21.4 21.5 21.6 INTRODUCTION FULL STATE FEEDBACK SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 21.1 21.1 21.5 21.5 21.6 21.6

22.

ELECTROMECHANICAL SYSTEMS . . . . . . . . . . . . . . . . . . . 22.1
22.1 22.2 22.3 22.4 22.5 22.6 22.7 INTRODUCTION MATHEMATICAL PROPERTIES 22.2.1 Induction EXAMPLE SYSTEMS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 22.1 22.1 22.1 22.9 22.16 22.16 22.16 22.16

23.

FLUID SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.1
23.1 23.2 SUMMARY MATHEMATICAL PROPERTIES 23.2.1 Resistance 23.2.2 Capacitance 23.2.3 Power Sources EXAMPLE SYSTEMS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEMS SOLUTIONS ASSIGNMENT PROBLEMS 23.1 23.1 23.2 23.4 23.6 23.8 23.10 23.10 23.10 23.10

23.3 23.4 23.5 23.6 23.7

24.

THERMAL SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.1
24.1 24.2 INTRODUCTION MATHEMATICAL PROPERTIES 24.1 24.1

page 9

24.3 24.4 24.5 24.6 24.7

24.2.1 Resistance 24.2.2 Capacitance 24.2.3 Sources EXAMPLE SYSTEMS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS

24.1 24.3 24.4 24.4 24.7 24.7 24.7 24.7

25.

OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.1
25.1 25.2 25.3 25.4 INTRODUCTION OBJECTIVES AND CONSTRAINTS SEARCHING FOR THE OPTIMUM OPTIMIZATION ALGORITHMS 25.4.1 Random Walk 25.4.2 Gradient Decent 25.4.3 Simplex SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 25.1 25.2 25.6 25.9 25.9 25.10 25.10 25.10 25.10 25.10 25.10

25.5 25.6 25.7 25.8

26.

FINITE ELEMENT ANALYSIS (FEA) . . . . . . . . . . . . . . . . . . . 26.1
26.1 26.2 26.3 26.4 26.5 26.6 26.7 26.8 INTRODUCTION FINITE ELEMENT MODELS FINITE ELEMENT MODELS SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS BIBLIOGRAPHY 26.1 26.2 26.4 26.12 26.13 26.13 26.13 26.13

27.

FUZZY LOGIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.1
27.1 27.2 27.3 27.4 27.5 27.6 27.7 INTRODUCTION COMMERCIAL CONTROLLERS REFERENCES SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS 27.1 27.7 27.7 27.7 27.8 27.8 27.8

28.

NEURAL NETWORKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28.1
28.1 SUMMARY 28.9

page 10

28.2 28.3 28.4 28.5

PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSIGNMENT PROBLEMS REFERENCES

28.10 28.10 28.10 28.10

29.

WRITING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29.1
29.1 29.2 29.3 29.4 29.5 FORGET WHAT YOU WERE TAUGHT BEFORE WHY WRITE REPORTS? THE TECHNICAL DEPTH OF THE REPORT TYPES OF REPORTS LABORATORY REPORTS An Example First Draft of a Report An Example Final Draft of a Report RESEARCH DRAFT REPORTS PROJECT REPORT OTHER REPORT TYPES 29.9.1 Executive 29.9.2 Consulting 29.9.3 Memo(randum) 29.9.4 Interim 29.9.5 Poster 29.9.6 Progress Report 29.9.7 Oral 29.9.8 Patent LAB BOOKS REPORT ELEMENTS 29.11.1 Figures 29.11.2 Graphs 29.11.3 Tables 29.11.4 Equations 29.11.5 Experimental Data 29.11.6 Result Summary 29.11.7 References 29.11.8 Acknowledgments 29.11.9 Abstracts 29.11.10 Appendices 29.11.11 Page Numbering 29.11.12 Numbers and Units 29.11.13 Engineering Drawings 29.11.14 Discussions 29.11.15 Conclusions 29.11.16 Recomendations 29.11.17 Appendices 29.11.18 Units 29.1 29.2 29.3 29.3 29.3 29.5 29.11 29.11 29.11 29.12 29.13 29.13 29.13 29.13 29.14 29.14 29.14 29.15 29.15 29.16 29.16 29.17 29.18 29.19 29.19 29.20 29.21 29.21 29.21 29.22 29.22 29.23 29.23 29.23 29.24 29.24 29.24 29.25 29.25

29.6 29.7 29.8 29.9

29.10 29.11

page 11

29.12 29.13 29.14 29.15 29.16

GENERAL WRITING ISSUES WRITERS BLOCK TECHNICAL ENGLISH EVALUATION FORMS PATENTS

29.25 29.26 29.26 29.29 29.31

30.

PROJECTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30.1
30.1 30.2 30.3 OVERVIEW 30.2.1 The Objectives and Constraints MANAGEMENT 30.3.1 Timeline - Tentative 30.3.2 Teams DELIVERABLES 30.4.1 Conceptual Design 30.4.2 EGR 345/101 Contract 30.4.3 Progress Reports 30.4.4 Design Proposal 30.4.5 The Final Report REPORT ELEMENTS 30.5.1 Gantt Charts 30.5.2 Drawings 30.5.3 Budgets and Bills of Material 30.5.4 Calculations APPENDICES 30.6.1 Appendix A - Sample System 30.6.2 Appendix B - EGR 345/101 Contract 30.6.3 Appendix C - Forms 30.1 30.1 30.2 30.3 30.3 30.4 30.5 30.5 30.5 30.6 30.6 30.7 30.8 30.8 30.9 30.9 30.10 30.10 30.10 30.18 30.19

30.4

30.5

30.6

31.

ENGINEERING PROBLEM SOLVING . . . . . . . . . . . . . . . . . . 31.1
31.1 31.2 31.3 BASIC RULES OF STYLE EXPECTED ELEMENTS SEPCIAL ELEMENTS 31.3.1 Graphs 31.3.2 EGR 345 Specific SCILAB TERMINOLOGY 31.1 31.1 31.2 31.2 31.2 31.2 31.3

31.4 31.5

32.

MATHEMATICAL TOOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1
32.1 INTRODUCTION 32.1.1 Constants and Other Stuff 32.1.2 Basic Operations Factorial 32.1.3 Exponents and Logarithms 32.1 32.2 32.3 32.4 32.4

page 12

32.2

32.3

32.4

32.5

32.6

32.7

32.1.4 Polynomial Expansions 32.1.5 Practice Problems FUNCTIONS 32.2.1 Discrete and Continuous Probability Distributions 32.2.2 Basic Polynomials 32.2.3 Partial Fractions 32.2.4 Summation and Series 32.2.5 Practice Problems SPATIAL RELATIONSHIPS 32.3.1 Trigonometry 32.3.2 Hyperbolic Functions Practice Problems 32.3.3 Geometry 32.3.4 Planes, Lines, etc. 32.3.5 Practice Problems COORDINATE SYSTEMS 32.4.1 Complex Numbers 32.4.2 Cylindrical Coordinates 32.4.3 Spherical Coordinates 32.4.4 Practice Problems MATRICES AND VECTORS 32.5.1 Vectors 32.5.2 Dot (Scalar) Product 32.5.3 Cross Product 32.5.4 Triple Product 32.5.5 Matrices 32.5.6 Solving Linear Equations with Matrices 32.5.7 Practice Problems CALCULUS 32.6.1 Single Variable Functions Differentiation Integration 32.6.2 Vector Calculus 32.6.3 Differential Equations First-order Differential Equations Guessing Separable Equations Homogeneous Equations and Substitution Second-order Differential Equations Linear Homogeneous Nonhomogeneous Linear Equations Higher Order Differential Equations Partial Differential Equations 32.6.4 Other Calculus Stuff 32.6.5 Practice Problems NUMERICAL METHODS

32.5 32.6 32.9 32.9 32.9 32.11 32.14 32.16 32.17 32.17 32.22 32.23 32.24 32.41 32.43 32.45 32.45 32.48 32.49 32.50 32.51 32.51 32.52 32.57 32.59 32.59 32.64 32.65 32.70 32.70 32.70 32.73 32.77 32.79 32.80 32.81 32.81 32.82 32.83 32.83 32.84 32.86 32.86 32.87 32.87 32.93

page 13

32.7.1 32.93

Approximation of Integrals and Derivatives from Sampled Data 32.94 32.94 32.95 32.95 32.96 32.96 32.99 32.102 32.102 32.103

32.8 32.9 32.10 32.11 32.12

32.7.2 Euler First-order Integration 32.7.3 Taylor Series Integration 32.7.4 Runge-Kutta Integration 32.7.5 Newton-Raphson to Find Roots LAPLACE TRANSFORMS 32.8.1 Laplace Transform Tables z-TRANSFORMS FOURIER SERIES TOPICS NOT COVERED (YET) REFERENCES/BIBLIOGRAPHY

33.

A BASIC INTRODUCTION TO ‘C’ . . . . . . . . . . . . . . . . . . . . . 33.1
33.1 33.2 33.3 33.4 33.5 33.6 WHY USE ‘C’? BACKGROUND PROGRAM PARTS HOW A ‘C’ COMPILER WORKS STRUCTURED ‘C’ CODE ARCHITECTURE OF ‘C’ PROGRAMS (TOP-DOWN) 33.6.1 How? 33.6.2 Why? CREATING TOP DOWN PROGRAMS HOW THE BEAMCAD PROGRAM WAS DESIGNED 33.8.1 Objectives: 33.8.2 Problem Definition: 33.8.3 User Interface: Screen Layout (also see figure): Input: Output: Help: Error Checking: Miscellaneous: 33.8.4 Flow Program: 33.8.5 Expand Program: 33.8.6 Testing and Debugging: 33.8.7 Documentation Users Manual: Programmers Manual: 33.8.8 Listing of BeamCAD Program. PRACTICE PROBLEMS 33.1 33.2 33.2 33.11 33.13 33.14 33.14 33.15 33.16 33.17 33.18 33.18 33.18 33.18 33.19 33.20 33.20 33.20 33.21 33.22 33.22 33.24 33.25 33.25 33.26 33.26 33.26

33.7 33.8

33.9

34.

UNITS AND CONVERSIONS . . . . . . . . . . . . . . . . . . . . . . . . . 34.1
34.1 34.2 HOW TO USE UNITS HOW TO USE SI UNITS 34.1 34.2

page 14

34.3 34.4 34.5

THE TABLE ASCII, HEX, BINARY CONVERSION G-CODES

34.2 34.6 34.8

35. 35. 36.

ATOMIC MATERIAL DATA . . . . . . . . . . . . . . . . . . . . . . . . . . 35.1 MECHANICAL MATERIAL PROPERTIES . . . . . . . . . . . . . . 35.1
35.1 FORMULA SHEET 35.4

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.1
36.1 TEXTBOOKS 36.1.1 Slotine and Li 36.1.2 VandeVegte 36.1 36.1 36.1

37.

TOPICS IN DEVELOPMENT . . . . . . . . . . . . . . . . . . . . . . . . . . 37.1
37.1 37.2 37.3 37.4 37.5 37.6 37.7 37.8 37.9 37.10 37.11 37.12 37.13 37.14 37.15 37.16 37.17 37.18 37.19 37.20 37.21 37.22 37.23 37.24 37.25 37.26 37.27 37.28 UPDATED DC MOTOR MODEL ANOTHER DC MOTOR MODEL STATE SPACE IDENTIFICATION BLOCK DIAGRAMS AND UNITS SIGNAL FLOW GRAPHS ZERO ORDER HOLD TORSIONAL DAMPERS MISC Nyquist Plot Nichols Plot BESSEL POLYNOMIALS ITAE LQR XXXXX XXXXX XXXXX XXXXX XXXXX XXXXX XXXXX XXXXX XXXXX SUMMARY PRACTICE PROBLEMS PRACTICE PROBLEM SOLUTIONS ASSGINMENT PROBLEMS REFERENCES BIBLIOGRAPHY 37.1 37.4 37.8 37.17 37.17 37.17 37.18 37.18 37.19 37.19 37.19 37.20 37.21 37.23 37.23 37.23 37.23 37.23 37.23 37.23 37.24 37.24 37.24 37.24 37.24 37.24 37.24 37.24

page 1

PREFACE
How to use the book. • read the chapters and do drill problems as you read • examine the case studies - these pull together concepts from previous chapters • problems at the ends of chapters are provided for further practice Tools that should be used include, • graphing calculator that can solve differential equations, such as a TI-85 • computer algebra software that can solve differential equations, such as Mathcad Supplemental materials at the end of this book include, • a lab guide for the course • a writing guide • a summary of math topics important for engineers • a table of generally useful engineering units • properties of common materials Acknowledgement to, Dr. Hal Larson for reviewing the calculus and numerical methods chapters Dr. Wendy Reffeor for reviewing the translation chapter Student background a basic circuits course a basic statics and mechanics of materials course math up to differential equations a general knowledge of physics computer programming, preferably in ’C’

Special notes - despite all common wisdom, inertia is presentaed as a force, this makes it easier for students attempting to learn, and keep sign conventions correct -

TO BE DONE small italicize variables and important terms fix equation numbering (auto-numbering?) fix subscripts and superscripts fix problem forms to include therefores, mark FBDs, etc. check C programs for ANSI compliance

page 2

big proofread and complete writing chapters add more drill problems and solutions verify the phase angle relationships cos vs/ sin. chapter rotation replace the rotational case with IC motor chapter non-linear systems develop chapter chapter magnetic consider adding/writing this chapter chapter fluids consider adding/writing this chapter chapter thermal consider adding/writing this chapter chapter c programming review section add problems

Buzzword topics: - Distributed systems - Fault Detection and Diagnosis - Vision and sensors - Robots and mobile machines - Vibration and control - Computational intelligence in mechatronics - Intelligent Transportation Systems - Micro mechatronics - Man Machine Interfaces - Hardware/software co-design - Intelligent manufacturing systems - Adaptive control - Architectures for signal processing and control algorithms - Communication protocols in control - Computing and communications - Control in communication networks - Decentralized control - Discrete event systems - Distributed control - Hybrid systems - Identification - Industrial automation - Intelligent control - Non-linear systems - Predictive control

Robotics .Robust control .page 3 .Real-time control .

Non-linear • linearity and superposition • reversibility • through and across variables . Discrete • Linear vs.1.1 1.introduction . INTRODUCTION Topics: Objectives: 1.1 BASIC TERMINOLOGY • Modeling • Simulation • Lumped parameter (masses and springs) • Distributed parameters (stress in a solid) • Continuous vs.

introduction - 1.2

• Analog vs. Digital • process vs. controllers • Basic system categories below,

Static Dynamic Stochastic Deterministic

Distributed Lumped

Non-linear Linear Continuous Discrete

Figure 1.1

Model Classifications

• control system types: servo vs. regulating/process control • open loop vs. closed loop • disturbances

introduction - 1.3

• component variations • system error • analysis vs. design • mechatronics • embedded systems • realtime systems •

1.2 EXAMPLE SYSTEM
• Servo control systems • Robot

1.3 SUMMARY

1.4 PRACTICE PROBLEMS
1.

translation - 2.1

2. TRANSLATION

Topics: • Basic laws of motion • Gravity, inertia, springs, dampers, cables and pulleys, drag, friction, FBDs • System analysis techniques • Design case Objectives: • To be able to develop differential equations that describe translating systems.

2.1 INTRODUCTION
If the velocity and acceleration of a body are both zero then the body will be static. If the applied forces are balanced, and cancel each other out, the body will not accelerate. If the forces are unbalanced then the body will accelerate. If all of the forces act through the center of mass then the body will only translate. Forces that do not act through the center of mass will also cause rotation to occur. This chapter will focus only on translational systems. The equations of motion for translating bodies are shown in Figure 2.1. These state simply that velocity is the first derivative of position, and velocity is the first derivative of acceleration. Conversely the acceleration can be integrated to find velocity, and the velocity can be integrated to find position. Therefore, if we know the acceleration of a body, we can determine the velocity and position. Finally, when a force is applied to a mass, an acceleration can be found by dividing the net force by the mass.

translation - 2.2

x,v,a

F

equations of motion dv ( t ) =  ---- x ( t )  dt d- 2 da ( t ) =  ---- x ( t ) =  ---- v ( t )  dt  dt OR x ( t ) = ∫ v ( t ) dt = ∫ ∫ a ( t ) dt v(t) =

(1) (2) (3) (4)

(5) a( t ) = F( t) --------M where, x, v, a = position, velocity and acceleratio n M = mass of the bd oy F = an applied forc e

∫ a( t ) dt

Figure 2.1

Velocity and acceleration of a translating mass

An example application of these fundamental laws is shown in Figure 2.2. The initial conditions of the system are supplied (and are normally required to solve this type of problem). These are then used to find the state of the system after a period of time. The solution begins by integrating the acceleration, and using the initial velocity value for the integration constant. So at t=0 the velocity will be equal to the initial velocity. This is then integrated once more to provide the position of the object. As before, the initial position is used for the integration constant. This equation is then used to calculate the position after a period of time. Notice that the units are used throughout the calculations, this is a good practice for any engineer.

translation - 2.3

Given an initial (t=0) state of x=5m, v=2m/s, a=3ms-2, find the system state 5 seconds later assuming constant acceleration. The initial conditions for the system at time t=0 are, x 0 = 5m Note: units are very important and should nor–1 v 0 = 2ms mally be carried through all calculations. –2 a0 = 3ms The constant acceleration can be integrated to find the velocity as a function of time. Note: v ( t ) = ∫ a 0 dt = a 0 t + C = a 0 t + v 0 (6) v ( t ) = a0 t + C v0 = a0 ( 0 ) + C v0 = C Next, the velocity can be integrated to find the position as a function of time. a0 2 (7) x ( t ) = ∫ v ( t ) dt = ∫ ( a 0 t + v0 ) dt = ----- t + v 0 t + x 0 2 This can then be used to calculate the position of the mass after 5 seconds. a0 2 x ( 5 ) = ----- t + v 0 t + x0 2 –2 3ms ( 5s ) 2 + 2ms –1 ( 5s ) + 5m = ---------------2 = 37.5m + 10m + 5m = 52.5m

Figure 2.2

Sample calculation for a translating mass, with initial conditions

2.2 MODELING
When modeling translational systems it is common to break the system into parts. These parts are then described with Free Body Diagrams (FBDs). Common components that must be considered when constructing FBDs are listed below, and are discussed in following sections. • gravity and other fields - apply non-contact forces • inertia - opposes acceleration and deceleration • springs - resist deflection • dampers and drag - resist motion • friction - opposes relative motion between bodies in contact

translation - 2.4

• cables and pulleys - redirect forces • contact points/joints - transmit forces through up to 3 degrees of freedom

2.2.1 Free Body Diagrams
Free Body Diagrams (FBDs) allow us to reduce a complex mechanical system into smaller, more manageable pieces. The forces applied to the FBD can then be summed to provide an equation for the piece. These equations can then be used later to do an analysis of system behavior. These are required elements for any engineering problem involving rigid bodies. An example of FBD construction is shown in Figure 2.3. In this case there is a mass sitting atop a spring. An FBD can be drawn for the mass. In total there are two obvious forces applied to the mass, gravity pulling the mass downward, and a spring pushing the mass upwards. The FBD for the spring has two forces applied at either end. Notice that the spring force, FR1, acting on the mass, and on the spring have an equal magnitude, but opposite direction.

M = 10 kg K = 20 N/m

FBD Mass:

FBD Spring:

FR1 FR1

Mg FR2

Figure 2.3

Free body diagram example

2.2.2 Mass and Inertia
In a static system the sum of forces is zero and nothing is in motion. In a dynamic system the sum of forces is not zero and the masses accelerate. The resulting imbalance in forces acts on the mass causing it to accelerate. For the purposes of calculation we create a virtual reaction force, called the inertial force. This force is also known as D’Alembert’s (pronounced as daa-lamb-bears) force. It can be included in calculations in one of two ways. The first is to add the inertial force to the FBD and then add it into the sum of

translation - 2.5

forces, which will equal zero. The second method is known as D’Alembert’s equation where all of the forces are summed and set equal to the inertial force, as shown in Figure 2.4. The acceleration is proportional to the inertial force and inversely proportional to the mass.

∑ F = Ma ∑ F – Ma =
Figure 2.4

0

(Newton’s) (D’Alembert’s)

(11) (12)

D’Alembert’s and Newton’s equations

An application of Newton’s equation to FBDs can be seen in Figure 2.5. In the first case an inertial force is added to the FBD. This force should be in an opposite direction (left here) to the positive direction of the mass (right). When the sum of forces equation is used then the force is added in as a normal force component. In the second case Newton’s equation is used so the force is left off the FBD, but added to the final equation. In this case the sign of the inertial force is positive if the assumed positive direction of the mass matches the positive direction for the summation.

translation - 2.6

D’Alemberts’s form: d- 2 Ma = M ---- x  dt

x d- 2 Fx = F – M  ---- x = 0 ∑  dt or d- 2 + Fx = – F + M  ---- x = 0 ∑  dt +

M

F

Note: If using an inertial force then the direction of the force should be opposite to the positive motion direction for the mass. Newton’s form: x d- 2 Fx = F = M  ---- x + ∑  dt or d- 2 Fx = – F = – M  ---- x + ∑  dt

M

F

Note: If using Newton’s form the sign of the inertial force should be positive if the positive direction for the summation and the mass are the same, otherwise if they are opposite then the sign should be negative.

Figure 2.5

Free body diagram and inertial forces

An example of the application of Newton’s equation is shown in Figure 2.6. In this example there are two unbalanced forces applied to a mass. These forces are summed and set equal to the inertial force. Solving the resulting equation results in acceleration values in the ’x’ and ’y’ directions. In this example the forces and calculations are done in vector form for convenience and brevity.

translation - 2.7

5 F1 = –4 N 0 If both forces shown act through the center of mass, what is the acceleration of the ball?

–7 F2 = – 3 N 0

M=10kg

∑F

= F 1 + F 2 = Ma

5 –7 N + – 3 N = ( 10Kg )a –4 0 0 –2 – 0.2 – 0.2 1 Kgm 1m  ------------   ----------- ------ = ∴a = N = – 0.7 – 0.7 -- 10Kg – 7  2  Kg 2 s s 0 0 0 Figure 2.6 Sample acceleration calculation

2.2.3 Gravity and Other Fields
Gravity is a weak force of attraction between masses. In our situation we are in the proximity of a large mass (the earth) which produces a large force of attraction. When analyzing forces acting on rigid bodies we add this force to our FBDs. The magnitude of the force is proportional to the mass of the object, with a direction toward the center of the earth (down). The relationship between mass and force is clear in the metric system with mass having the units Kilograms (kg), and force the units Newtons (N). The relationship between these is the gravitational constant 9.81N/kg, which will vary slightly over the surface of the earth. The Imperial units for force and mass are both the pound (lb.) which often causes confusion. To reduce this confusion the unit for force is normally modified to be, lbf. An example calculation including gravitational acceleration is shown in Figure 2.7. The 5kg mass is pulled by two forces, gravity and the arbitrary force ’F’. These forces are described in vector form, with the positive ’z’ axis pointing upwards. To find the equa-

translation - 2.8

tions of motion the forces are summed. To eliminate the second derivative on the inertia term the equation is integrated twice. The result is a set of three vector equations that describe the x, y and z components of the motion. Notice that the units have been carried through these calculations.

translation - 2.9

Assume we have a mass that is acted upon by gravity and a second constant force vector. To find the position of the mass as a function of time we first define the gravity vector and position components for the system.

F

M = 5Kg Fg g

0 0 N- = g = 0 -----0 Kg – 9.81 –9.81

m ---2 s

x(t) X( t ) = y ( t ) z( t )

fx F = f y fz Fg = Mg

Next, sum the forces and set them equal to inertial resistance. d- 2 F = Mg + F = M  ---- X ( t ) ∑  dt fx 0 x ( t) m+ d- 2  ---- 5Kg 0 ---fy = 5Kg   y ( t ) dt 2 – 9.81 s z( t ) fz fx 0 x(t ) –1 d- 2 m ---- + 0.2Kg =  ---- y ( t ) fy 0  dt 2 s – 9.81 z( t) fz Integrate twice to find the position components. Note: When an engineer solves a problem they vx fx 0 will always be looking 0 x ( t) dm + 0.2Kg– 1 at the equations and  ---- fy t + vy =   y ( t) 0 ---dt 2 unknowns. In this case 0 –9.81 s z( t ) fz there are three equavz tions, and there are 9 0 vx constants/givens fx, fy, fx x0 0 0 x(t) fz, vx0, vy0, xz0, x0, y0 –1 2 1 m -fy t + vy t + y0 = y ( t ) 0 ---- + 0.2Kg and z0. There are 4 2 2 0 s variables/unknowns x, – 9.81 z(t) fz z0 vz y, z and t. Therefore 2 0.1f x t + v x t + x 0 0 with 3 equations and 4 0 unknowns only one x(t ) 2 value (4-3) is required m 0.1f y t + v y t + y 0 y(t ) = to find all of the 0 z(t) unknown values.  – 9.81 + 0.1f  t 2 + v t + z -----------z z0 0  2 Figure 2.7 Gravity vector calculations

translation - 2.10

Like gravity, magnetic and electrostatic fields can also apply forces to objects. Magnetic forces are commonly found in motors and other electrical actuators. Electrostatic forces are less common, but may need to be considered for highly charged systems.

Given, 3 F1 = 4 N 0 0 Ng = – 9.81 -----Kg 0 M = 2Kg

FBD: F1 M

Find the acceleration.

g

ans. 1.5 a = –2.91 m --s 0

Figure 2.8

Drill problem: find the acceleration of the FBD

2.2.4 Springs
Springs are typically constructed with elastic materials such as metals and plastics, that will provide an opposing force when deformed. The properties of the spring are determined by the Young’s modulus (E) of the material and the geometry of the spring. A primitive spring is shown in Figure 2.9. In this case the spring is a solid member. The relationship between force and displacement is determined by the basic mechanics of materials relationship. In practice springs are more complex, but the parameters (E, A and L) are combined into a more convenient form. This form is known as Hooke’s Law.

translation - 2.11

Lδ =  ------  F  EA EA F =  ------  δ  L F = Kδ

F

δ
(Hooke’s law)

L

Figure 2.9

A solid member as a spring

Hooke’s law does have some limitations that engineers must consider. The basic equation is linear, but as a spring is deformed the material approaches plastic deformation, and the modulus of elasticity will change. In addition the geometry of the object changes, also changing the effective stiffness. Springs are normally assumed to be massless. This allows the inertial effects to be ignored, such as a force propagation delay. In applications with fast rates of change the spring mass may become significant, and they will no longer act as an ideal device. The cases for tension and compression are shown in Figure 2.10. In the case of compression the spring length has been made shorter than its’ normal length. This requires that a compression force be applied. For tension, both the displacement from neutral and the required force reverse direction. It is advisable when solving problems to assume a spring is either in tension or compression, and then select the displacement and force directions accordingly.

translation - 2.12

∆ x = deformed length

ASIDE: a spring has a natural or undeformed length. When at this length it is neither in tension or compression Fc

compression as positive:

∆x
Ks Fc = K s ∆ x

Fc

∆x
Ks Fc = – Ks ∆ x

Fc

∆x
Ks Fc = – Ks ∆ x

tension as positive:

Ft

Ks

∆x
Ft Ks

Ft = Ks ∆ x

∆x
Ft Ks

Ft = – Ks ∆ x

∆x

Ft = – Ks ∆ x

NOTE: the symbols for springs, resistors and inductors are quite often the same or similar. You will need to remember this when dealing with complex systems - and especially in this book where we deal with both types of components.

Figure 2.10

Sign conventions for spring forces and displacements

Previous examples have shown springs with displacements at one end only. In Figure 2.11, springs are shown that have movement at both ends. In these cases the sign of the force applied to the spring is selected with reference to the assumed compression or tension. The primary difference is that care is required to correctly construct the expressions for the tension or compression forces. In all cases the forces on the springs must be assumed and drawn as either tensile or compressive. In the first example the displacement and forces are tensile. The displacement at the left is tensile, so it will be positive, but on the right hand side the displacement is compressive so it is negative. In the second exam-

translation - 2.13

ple the force and both displacements are shown as tensile, so the terms are both positive. In the third example the force is shown as compressive, while the displacements are both shown as tensile, so both terms are negative.

F

∆ x1

Ks

∆ x2

F

F = K s ( ∆ x1 – ∆ x2 )

F

∆ x1

Ks

∆ x2

F

F = Ks ( ∆ x 1 + ∆ x2 )

F

∆ x1

Ks

∆ x2

F

F = K s ( – ∆ x1 – ∆ x2 )

Aside: it is useful to assume that the spring is either in tension or compression, and then make all decisions based on that assumption.

Figure 2.11

Examples of forces when both sides of a spring can move

Sometimes the true length of a spring is important, and the deformation alone is insufficient. In these cases the deformation can be defined as a deformed and undeformed length, as shown in Figure 2.12.

∆ x = l 1 – l0
where,

∆ x = deformation l 0 = the length when undeforme d l 1 = the length when deforme d
Using the actual spring length

Figure 2.12

translation - 2.14

In addition to providing forces, springs may be used as energy storage devices. Figure 2.13 shows the equation for energy stored in a spring.

EP

K ( ∆x ) = ----------------2

2

Figure 2.13

Energy stored in a spring

Given, N K s = 10 --m Ks Ks F

∆ x = 0.1m ∆x
Find F assuming the springs are normally 4m long when unloaded Aside: it can help to draw a FBD of the pin. 10m

ans. F=2N

Figure 2.14

Drill problem: Deformation of a two spring system

translation - 2.15

Draw the FBDs and the equations of motion for the masses F Ks M1 x1 M2 x2

ans. ·· x 1 ( M1 ) + x 1 ( Ks ) + x 2 ( – Ks ) = F ·· x1 ( Ks ) + x2 ( –M 2 ) + x2 ( –Ks ) = F

Figure 2.15

Drill problem: Draw the FBDs for the masses

translation - 2.16

2.2.5 Damping and Drag
A damper is a component that resists motion. The resistive force is relative to the rate of displacement. As mentioned before, springs store energy in a system but dampers dissipate energy. Dampers and springs are often used to compliment each other in designs. Damping can occur naturally in a system, or can be added by design. The physical damper pictured in Figure 2.16 uses a cylinder that contains a fluid. There is a moving rod and piston that can slide within the cylinder. As the piston moves, fluid is forced through a small orifice. When moved slowly the fluid moves easily, but when moved quickly the pressure required to force the fluid through the orifice rises. This rise in pressure results in a higher force of resistance. In ideal terms any motion would result in an opposing force. In reality there is also a break-away force that needs to be applied before motion begins. Other manufacturing variations could also lead to other small differences between dampers. Normally these cause negligible effects.

orifice motion fluid piston fluid

Figure 2.16

A physical damper

The basic equation for an ideal damper in compression is shown in Figure 2.17. In this case the force and displacement are both compressive. The force is calculated by multiplying the damping coefficient by the velocity, or first derivative of position. Aside from the use of the first derivative of position, the analysis of dampers in systems is similar to that of springs.

17 An ideal damper Damping can also occur when there is relative motion between two objects. Normally these forces are small.18 Viscous damping between two bodies with relative motion A damping force is proportional to the first derivative of position (velocity). except at high velocities. such as oils.18 two objects are shown with viscous friction (damping) between them. Aerodynamic drag is proportional to the velocity squared. Figure 2. The equation for drag is shown in . have a significant viscosity. When these materials are put in shear they resist the motion. oil) then there will be a damping effect. It was constructed using a small piston inside a larger pot filled with oil.17 x Kd F dF = K d  ---- x  dt (15) Aside: The symbol shown is typically used for dampers. the greater the resistance to flow. x1 · · Fd = Kd ( x1 – x2 ) Kd · · F d = Kd ( x1 – x2 ) x2 Aside: Fluids. It is based on an old damper design called a dashpot. In the example in Figure 2.translation . If the objects are lubricated with a viscous fluid (e. The higher the shear rate..g. When the system is broken into free body diagrams the forces are shown to be a function of the relative velocities between the blocks.2. Figure 2.

18 Figure 2.19 in vector and scalar forms.20 Drill problem: Find the required forces on the damper .1 m --dt s F d---.translation .3 m --dt s F k d = 0. the magnitude of the drag force coefficient ’D’ is approximated theoretically and/or measured experimentally. resulting in the velocities given below. v F F = –D v v Figure 2.2. F = – 0.1 Ns ----m ans. The drag force increases as the square of velocity. surface properties. Normally. object size and object geometry.02N Figure 2.x2 = – 0. The drag coefficient is a function of material type.19 Aerodynamic drag The force is acting on the cylinder. What is the applied force? d---.x 1 = 0.

6 Cables And Pulleys Cables are useful when transmitting tensile forces or displacements.22 Friction of a belt over a fixed drum .translation . and will not transmit force. A cable by itself can be represented as a force vector.21 Tension in a cable over a massless frictionless pulley If we have a pulley that is fixed and cannot rotate.= e k T1 T2 ∆θ Figure 2. T1 for a massless frictionless pulley T1 = T2 T2 Figure 2. When used in combination with pulleys.2. If this is the case then the tension in the cable on both sides of the pulley are equal. In this case we can use the coefficient of friction to determine the relative ratio of forces between the sides of the pulley. as shown in Figure 2. if the force becomes compressive. a cable can redirect a force vector or multiply a force.2. as shown in Figure 2.21. the cable becomes limp. T1 T2 µ ( ∆θ ) ----. Typically we assume that a pulley is massless and frictionless (in the rotation chapter we will assume they are not). the cable must slide over the surface of the pulley. And.19 2.22. The centerline of the cable becomes the centerline for the force.

Fup = 9.2.2  -.20 Given. the theory also applies to belts over drums.translation .81N Fdown = -------------e π 0.  2 N 9.2 Find F to start the mass moving up and down.  2 Figure 2.2  -. . M F ans.23 Drill problem: Friction forces for belts on drums Although the discussion in this section has focused on cables and pulleys.81e π 0. and then the force required to maintain a low velocity motion. µs = 0.35 M = 1Kg µ k = 0.

2. Figure 2. Beneath the slip force the object will stay in place. surface texture and surface shape. (Note: If the object begins to travel much faster then the kinetic friction force will decrease.21 2. . and the resulting kinetic friction force will be relatively constant. After that the object will begin to slide and a constant friction force will result. When the slip force is exceeded the object will begin to move. but it always opposes the applied force or motion.24 shows the classic model for (dry Coulomb) friction.2. The coefficient of friction is a function of the materials. and the touching surfaces are dry.translation .) It is common to forget that friction forces are bidirectional. In cases where there is no lubricant. dry coulomb friction may result. The friction force is a function of the coefficient of friction and the normal force across the contact surfaces. where a lubricant would provide a damping effect between two moving objects.7 Friction Viscous friction was discussed before. The force on the horizontal axis is the force applied to the friction surfaces while the vertical axis is the resulting friction force. In this case the surfaces will stick in place until a maximum force is overcome.

2. the friction forces direction opposes applied forces. After that the friction is approximately constant. Figure 2.24 Dry friction Many systems use kinetic friction to dissipate energy from a system as heat. In addition.translation . Fs F applied Fg F F k. . Fs F k = µk N Fs ≤ µ s N N Note: When solving problems with friction remember that the friction force will always equal the applied force (not the maximum force) until slip occurs. sound and vibration. and motion.22 Fresult Fs Fk Block begins to slip and the applied force exceeds the resultant and acceleration begins.

translation . These connections can be rigid or moving.23 Find the acceleration of the block for both angles indicated.2.2.2 10 kg θ θ1 = 5 ° θ 2 = 35 ° ans. In solid connections all forces and moments are transmitted and the two . µ s = 0.3 µk = 0.8 Contact Points And Joints A system is built by connecting components together.25 Drill problem: find the accelerations 2. Figure 2.

3 SYSTEM EXAMPLES An orderly approach to system analysis can simplify the process of analyzing large systems.2. In any direction there is a degree of freedom. In moving connections there is at least one degree of freedom. y and z. FAy A M1 B M1g M2 FBx FBy F Ax FBy F Bx M2g Note: Don’t forget that forces on connected FBDs should have equal magnitudes. For joint B the force components with equal magnitudes but opposite directions are added to both FBDs. 2. there are up to three degrees of freedom. but opposite directions. Assign letters/numbers to designate force components (if not already done) this will allow you to refer to components in your calculations. The list of steps below is based on general observations of good problem solving techniques.translation . This should include the . At joint A the forces are written as two components in the x and y directions. Consider the example in Figure 2. When constructing FBDs for a system we must break all of the components into individual rigid bodies. 1. a force or moment cannot be transmitted. Define positions and directions for any moving masses.24 pieces act as a single rigid body.26 FBDs for systems with connected members 2. Where the mechanism has been broken the contact forces must be added to both of the separated pieces.26. x. Figure 2. If we limit this to translation only.

6. the final value can be checked for errors using unit analysis. .(next chapter) Develop a final equation that relates input (forcing functions) to outputs (results).translation . and add forces (inertia is optional). Draw free body diagrams for each component. ----.25 selection of reference points.g. = N  m  m  m 2 s s ∴N + ( N ) + ( N ) + ( – N ) = N The units match. Write equations for each component by summing forces. 5.(next chapter) Combine the equations by eliminating unwanted variables.( Kg ) + --.. The FBD for the mass shows the applied force and the reaction forces from the spring and damper. This method involves replacing variables with their unit equivalents.27. Note: When deriving differential equations. This equation is then rearranged to a second-order non-homogeneous differential equation. 3. so there are no obvious problems.2. x 2 ( M 2 ) + x 2 ( B ) + x 2 ( K s2 ) + x 1 ( – K s2 ) = F N m m Ns N . + m  --- + m  – --. ·· · e. 4. it is opposed by a spring and damper on the right. When the forces are summed the inertia is on the right side of the equation in Newton’s form. so no others are required. The basic problem definition already contains all of the needed definitions. On the left is a force. coefficient F Ks Kd M units Kgm N = ----------2 s N --m Ns ----m Kg Consider the cart in Figure 2. All the units should match.∴ ---.

x + ------. and everything else on the other.2.26 Given the system diagram. eqn. · ·· + ∑ Fx = – F – Kd x – K s x = M 1 x This equation can be rearranged to a second-order non-homogeneous diff. F M1 x Kd Ks The FBD for the cart is F ·· M1 x M1 · Kdx Ks x The forces for the cart are in a single direction and can be summed as. Kd Ks F·· x + ------. It is useful to have all of the ’output’ variables for the system on the left hand side. Figure 2.translation .x = -------· M1 M1 M1 Aside: later on we will solve the differential equations.27 A simple translational system example . or use other methods to determine how the system will behave.

.2. x1 x2 K d1 F M1 K s1 M2 K s2 K d2 Figure 2.27 Develop the equation relating the input force to the motion (in terms of x) of the lefthand cart for the problem below. Care is also taken to ensure that all forces between bodies are equal in magnitude. and x-axis force components are irrelevant to the forces in the y-axis.translation .29. These are added to the FBDs.28 Drill problem: Find the differential equations A simplified model of an elevator (M1) and a passenger (M2) are shown in Figure 2. The wall forces are ignored because they are statically indeterminate. In this example many of the required variables need to be defined. but opposite in direction.

.translation .30.28 M1 M2 Kd K S2 K S1 Assign required quantities and draw the FBDs M2g · · FD = Kd ( y2 – y1 ) ·· M1 y 1 M1 g y2 FS2 = K s2 ( y 1 – y 2 ) FD FS2 ·· M2y2 y1 FS1 = – K s1 y 1 Figure 2.2.29 A multi-body translating system (an elevator with a passenger) The forces on the FBDs are summed and the equations are expanded in Figure 2.

and substitute relationships. sum the forces in vector form.( y 2 – y 1 ) = M2 a y dt Figure 2.81 ) + K d ---.( y 2 – y 1 ) = M1 a y dt d– K S2 ( y 2 – y 1 ) + M2 ( – 9.translation .81 ) – K d ---. ∑ FM ∑ FM 1 = F S1 + FS2 + M1 g + F D = M1 a 1 = – FS2 + M2 g – FD = M2 a 2 2 At this point the equations are expanded dK S1 ( – y 1 ) + K S2 ( y 2 – y 1 ) + M1 ( – 9.30 Equations for the elevator .2.29 Now.

2.translation . In the parallel spring combination the overall stiffness of the spring would increase.32. .30 A Write the differential equations for the system below. Figure 2. When two springs are combined in this manner they can be replaced with a single equivalent spring. Assume it is in motion B µ k1 M1 K S1 M2 µk2 ans. In the series spring combination the overall stiffness would decrease.31 Drill problem: A more complex translational systems Consider the springs shown in Figure 2.

The forces for both of these are then summed. The first step is to draw a FBD for the mass at the bottom. and the equivalent spring coefficient is found.31 Parallel K S1 K S2 Series K S1 M K S2 M Figure 2.translation . After this. The next process is to combine the two equations to eliminate the height variable created for point P. the equation is rearranged into Hooke’s law. . and for a point between the two springs.2. P.32 Springs in parallel and series (kinematically) Figure 2.33 shows the calculations required to find a spring coefficient equivalent to the two springs in series.

33 Calculation of an equivalent spring coefficient for springs in series . + ∑ Fy = K S1 y 2 – K S2 ( y 1 – y 2 ) = 0 (1) K S2 ( y 1 – y 2 ) K S2 ( y 1 – y 2 ) M + Fg Next.  K S1 + K S2 Finally. rearrange the equations to eliminate y2 and simplify.= y 1 --------------------------K S2 K S1 + K S2 K S2   Fg = y 1  1 – --------------------------- K S2 K S1 + K S2   K S1 + K S2 – K S2 Fg = y 1  -------------------------------------------- K  K S1 + K S2  S2  K S1 K S2  F g = y 1  --------------------------. (1) becomes K S2 ( y 1 – y 2 ) – F g = 0 Fg y 1 – y 2 = --------K S2 Fg --------y2 = y1 – K S2 (2) becomes K S1 y 2 – K S2 ( y 1 – y 2 ) = 0 y1 K S1 y 2 K S1 P K S2 y2 P ∑ Fy = K S2 ( y 1 – y 2 ) – F g = 0 (2) M (3) (4) y 2 ( K S1 + K S2 ) = y 1 K S2 Fg   sub (3) into (4)  y – --------. ( K + K ) = y K S2 1 S2  1 K S2 S1 Fg K S2 y 1 – --------.2. consider the basic spring equation to find the equivalent spring coefficient. draw FBDs for P and M and sum the forces assuming the system is static.translation . K S1 K S2 K equiv = --------------------------· K S1 + K S2 Figure 2.32 First.

translation . . When an object has no mass. The only factor that changes is displacement.2. Figure 2.34 Drill problem: Find an equivalent spring for the springs in parallel Consider the drill problem.33 K S1 K S2 F ans. the force applied to one side of the spring will also be applied to the other.

or energy transfer rate is the force applied multiplied by the velocity. ans.2. These equations show that the total energy in the system is the sum of kinetic and potential energy.35 Drill problem: Prove that the force on both sides is equal 2.36.translation . In addition.4 OTHER TOPICS Designing a system in terms of energy content can allow insights not easily obtained by the methods already discussed.34 Show that a force applied to one side of a massless spring is the reaction force at the other side. Consider the equations in Figure 2. Potential energy in translating systems is a force magnitude multiplied by a distance (that force was applied over). Figure 2. the power. . Kinetic energy is half the product of mass times velocity squared.

what is the spring constant? . • The equations can be integrated for dynamic cases.36 Energy and power equations for translating masses 2.35 E = E P + EK 2 E K = Mv ---------2 E P = Fd = Mgd dP = Fv = ---.6 PRACTICE PROBLEMS 1. • Equations for translation and rotation can be written for FBDs.translation .2. 2.E dt (7) (8) (9) (10) Figure 2. or solved algebraically for static cases. If a spring has a deflection of 6 cm when exposed to a static load of 200N.5 SUMMARY • FBDs are useful for reducing complex systems to simpler parts.

Write a differential equation for the mass pictured below. a) Kd1 Kd2 b) Kd1 Kd2 3. Write the differential equations for the translating system below. F Ks M1 M2 . Find the effective damping coefficients for the pairs below from basic principles. x M B F 4.36 2.2.translation .

37 5. Write the differential equations for the system given below. Kd M1 B1 B2 Ks M2 F 7.translation . Write the differential equations for the system below. Write the differential equations for the system below.2. x1 K d1 K s1 M1 K s2 K d2 M2 F x2 K d3 K s3 . x1 K d1 K s1 M1 K s2 M2 B F x2 8. Write the differential equations for the system below. x1 K d1 K s1 M1 K s2 M2 B F x2 6.

Write the differential equations for the system below.2. F M1 x1 K s1 K d1 M2 x2 K s2 K d2 . K s2 M2 K d1 K s1 M1 x2 F µ s.translation . K s2 M2 K d1 K s1 M1 B x1 x2 F 10. µ k x1 11. Write the differential equations for the system below. Write the differential equations for the system below.38 9.

M2.39 12. The suspended mass. Write the differential equations for the system below. The cable runs over a massless.translation . is between a spring and a cable and there is viscous damping between the mass and the floor. x1 Ks M1 B R M2 x2 Kd . In this system the upper mass. frictionless pulley. x K s1 M B R K s2 13. Write the differential equations for the system below. M1. is between the cable and a damper.2. Assume that the pulley is massless and frictionless and that the system begins undeflected.

F ·· · B x + x  ---- = – --- M M 4. Ks –Ks ·· Fx 1 + x 1  -----.2. + x 1  -------- = 0  M -  M - 2 2 . µ k M2 x2 K s2 2.3N/cm 2. Write the differential equations for the system below. Ks = 33.translation . a) K eq = K d1 + K d2 b) K d1 K d2 K eq = ----------------------K d1 + K d2 3. + x 2  -------- = ----- M -  M - M1 1 1 Ks –Ks ·· x 2 + x 2  -----. K s1 M1 x1 µ s.7 PRACTICE PROBLEM SOLUTIONS 1.40 14.

 + x 2  -----.translation . + x 2  -------. = 0  M -  M -  M - 1 1 1 Ks ·· · ·· · B Fx 1 + x 1 ( B ) + x 2 + x 2  -----. = 0  M  M  M  1 1 1 Ks –Ks –F ·· · B2 x 2 + x 2  -----. = ----- M -  M - M2 2 2 . = -----M  M  M  M2 2 2 2 7. = -----M  M   M  M2 2 2 2 6. + x 2  -----. + x 1  -------.41 5. = 0 M   M   M  1 1 1 K s2 – K s2 ·· · B Fx 2 + x 2  -----. + x 2  ---------. Ks –Ks ·· · Kd + B 1 x 1 + x 1  ------------------ + x 1  -----. + x 1  ---------. + x 2  ---------. K s1 + K s2 – K s2 ·· · K d1 x 1 + x 1  -------- + x 1  ---------------------. K s1 – K s2 ·· · K d1 x 1 + x 1  -------- + x 1  ------.2. + x 2  ------.

 + x 2  -----. + x 2  ---------. + x 2  ----------. = -----M  M  M  M2 2 2 2 . + x 2  ---------------------. FBDs: · – K d1 x 1 –K s1 x 1 For M1: + ·· M1x1 M1 · · K d2 ( x 1 – x 2 ) K s2 ( x 1 – x 2 ) M2 · K d3 x 2 K s3 x 2 ·· M2 x 2 F · · · ·· ∑ F = – Kd1 x 1 – K s1 x1 – Kd2 ( x1 – x2 ) – Ks2 ( x1 – x2 ) = M 1 x1 ·· · · x 1 ( M 1 ) + x 1 ( K d1 + K d2 ) + x 1 ( K s1 + K s2 ) + x 2 ( – K d2 ) + x 2 ( – K s2 ) = 0 K s1 + K s2 –K s2 ·· · K d1 + K d2 · – K d2 x 1 + x 1  ----------------------.2. = F  M   M   M   M  2 2 2 2 9. + x 1  -----.translation . + x 2  ------. ·· · · x 1 ( M 1 ) + x 1 ( K d1 + B ) + x 1 ( K s1 ) + x 2 ( – B ) = 0 K s1 ·· · K d1 + B · –B x 1 + x 1  ------------------ + x 1  ------. + x 1  ---------------------. + x 1  ----------- + x 1  ---------.42 8. = 0  M  M  M  1 1 1 ·· · · x 2 ( M 2 ) + x 2 ( B ) + x 2 ( K s2 ) + x 1 ( – B ) = F K s2 ·· · B · –B Fx 2 + x 2  -----. = 0  M   M   M   M  1 1 1 1 For M2: + ∑F · · · · = K d2 ( x 1 – x 2 ) + K s2 ( x 1 – x 2 ) + F – K d3 x 2 – K s3 x 2 = M2 x 2 ·· · · x 2 ( M2 ) + x 2 ( K d2 + K d3 ) + x 2 ( K s2 + K s3 ) + x 1 ( – K d2 ) + x 1 ( – K s2 ) = F K s2 + K s3 – K s2 ·· · K d2 + K d3 · – K d2 x 2 + x 2  ----------------------.

 + x 2  ---------------------. + x  ------.  x·1 – x·2  if if · · x1 = x2 · · x1 ≠ x2 11. = --------------M  M2 2 where. if ( x ≥ 0 ) if ( x < 0 ) T = 0 T = – K s2 x K s1 T·· · Bx + x  -----.2. + x 2  ----------. = -----M  M   M   M  M1 1 1 1 1 ·· · · x 2 ( M 2 ) + x 2 ( K d1 + K d2 ) + x 2 ( K s1 + K s2 ) + x 1 ( – K d1 ) + x 1 ( – K s1 ) = 0 K s1 + K s2 –K s1 ·· · K d1 + K d2 · – K d1 x 2 + x 2  ----------------------. (assuming no gravity ·· · · x 1 ( M 1 ) + x 1 ( K d1 ) + x 1 ( K s1 ) + x 2 ( – K d1 ) + x 2 ( – K s1 ) = F K s1 –K s1 F ·· · K d1 · – K d1 x 1 + x 1  -------- + x 1  ------.translation . FF ≤ µs M2 g FF · ·  x1 – x2  = µ k M 2 g  -----------------. = -----M   M - M1 1 1 . = 0  M   M   M   M  2 2 2 2 12.43 10. + x 2  ---------. ·· · x 1 ( M 1 ) + x 1 ( K d1 ) + x 1 ( K s1 ) = F F FF K s1 ·· · K d1 - - x 1 + x 1  -------- + x 1  ------. + x 1  ---------. + x 1  ----------. = -----M M M1 1 1 ·· x 2 ( M2 ) + x 2 ( K s2 ) = F – FF F – FF K s2 ·· x 2 + x 2  ------.

2.44 13. if ( ( x 2 – x 1 ) < 0 ) · if ( x 1 = 0 ) · if ( x 1 ≠ 0 ) T – FF K s1 ·· x 1 + x 1  ------.8 ASSIGNMENT PROBLEMS . = ------ M M2 2 T = 0 F F ≤ M 1 gµ s · x1 FF = M 1 gµk ------·x1 2. = --------------- M M1 1 K s2 ·· –T x 2 + x 2  ------. = ------------------M  – M2 2 For M2: + ∑F For T: if T <= 0 then T = 0N if T > 0 x1 = x2 14. = ------ M  M - M1 1 1 · ·· = T + Kd x 2 – M2 g = – M2 x 2 ·· · x 2 ( – M 2 ) + x 2 ( – Kd ) = T – M2 g T – M2 g ·· · Kd x 2 + x 2  -----. + x 1  -----. . FBDs: Ks x1 · Bx 1 + M1 T ·· M 1 x1 M2 M2g T ·· M 2 x2 · Kd x 2 For M1: ∑F · ·· = – K s x 1 – Bx1 + T = M 1 x 1 ·· · x 1 ( M1 ) + x 1 ( B ) + x 1 ( Ks ) = T Ks ·· · B Tx 1 + x 1  -----.translation .

. and then solve the differential equation.1 shows an abstract description of a system. The output response of the system would be the actual height of the elevator. This output response can then be used as a guide to modify design choices (parameters). In practice. elastic lift cable. The basic concept is that the system changes the inputs to outputs. Inputs to the system would be the mass of human riders and desired elevator height. • To recognize first and second-order equation forms. Say. for example.3. mass of the car. Using assumed initial values and input functions the differential equation could be solved to get an explicit equation for elevator height. the system model could be developed using differential equations for the motor. Figure 3. It is also possible to estimate the system response without solving the differential equation as will be discussed later. These equations can be used to analyze the behavior of the system and make design decisions. 3.1 INTRODUCTION In the previous chapter we derived differential equations of motion for translating systems. The most basic method is to select a standard input type (a forcing function) and initial conditions. etc. A basic test would involve assuming that the elevator starts at the ground floor and must travel to the top floor. that the system to be analyzed is an elevator. For analysis.differential equations .1 3. many of the assumptions and tests are mandated by law or by groups such as Underwriters Laboratories (UL). Canadian Standards Association (CSA) and the European Commission (CE). ANALYSIS OF DIFFERENTIAL EQUATIONS Topics: • First and second-order homogeneous differential equations • Non-homogeneous differential equations • First and second-order responses • Non-linear system elements • Design case Objectives: • To develop explicit equations that describe a system response.

The basic types of differential equations are shown in Figure 3. On the left hand side is the integration variable ’x’. These equations correspond to systems without inputs. 8rad/s at 3 seconds. such as wave height that is continually oscillating at 1Hz. then the equation is homogeneous. 4rad/s at 2 seconds. Non-homogeneous differential equations will also be reviewed. Each of these equations is linear because each of the terms on the left hand side is simply multiplied by a linear coefficient.differential equations . such as very rapidly changing a desired speed from 0Hz to 50Hz.2 EXPLICIT SOLUTIONS Solving a differential equation results in an explicit solution. If the right hand side is zero. This section will review techniques used to integrate first and second-order homogenous differential equations. This equation provides the general response as a function of time. • step . the system can be analyzed to determine its behavior.a continuously increasing input. but it can also be used to find frequencies and other characteristics of interest. and outputs are on the right. • parabolic . such as a motor speed that is 2Hz at 1 second.an exponentially increasing input. The most fundamental technique is to integrate the differential equation(s) for the system. etc. • ramp . • sinusoidal .a cyclic input that varies continuously.3. After the system has been modeled. also called unforced systems.a sudden change of input.2. 3. Figure 3. . an input type has been chosen.1 A system with and input and output response There are several standard input types used to test a system. These are listed below in order of relative popularity with brief explanations. Each of these equations is linear. such as a motor speed that increases constantly at 10Hz per minute.2 inputs forcing function system differential equations outputs response function Note: By convention inputs are on the left. and the initial conditions have been selected.

but approach zero as time goes to infinity.differential equations .2 Standard equation forms A general solution for a first-order homogeneous differential equation is given in Figure 3. .3.3 · Ax + Bx = 0 · Ax + Bx = Cf ( t ) ·· · Ax + Bx + Cx = 0 ·· · Ax + Bx + Cx = Df ( t ) first-order homogeneous first-order non-homogeneous second-order homogeneous second-order non-homogeneous Figure 3. Substitution leads to finding the value of the coefficient ’Y’. Following this. The solution begins with the solution of the homogeneous equation where a general form is ’guessed’.3. Notice that the final equation will begin at the initial displacement. the initial conditions for the equation are used to find the value of the coefficient ’X’. The e-to-the-x behavior is characteristic for a firstorder response.

differential equations . –Bt -A x h = Xe Note: The general form below is useful for finding almost all homogeneous equations x h ( t ) = Xe – Yt Next.3. x ( t ) = x0 e B -t – -A Figure 3. – Yt – Yt · x = Xe x = – YX e A ( –YX e – Yt ) + B ( Xe –Yt ) = 0 initial condition A ( –Y ) + B = 0 Y = B -A Therefore the general form is. x ( 0 ) = x0 · and Ax + Bx = 0 Guess a solution form and solve. use the initial conditions to find the remaining unknowns. –B t -A x h = Xe –B --0 A x 0 = Xe x0 = X Therefore the final equation is.4 Given the general form of a first-order homogeneous equation.3 General solution of a first-order homogeneous equation .

It is not shown. two identical real roots. the initial conditions need to be used to find the remaining coefficient values.differential equations .5. x ( t ) = 3e – 2t Figure 3. . x(0 ) = 3 · x + 2x = 0 ans.4 Drill Problem: First order homogeneous differential equation The general solution to a second-order homogeneous equation is shown in Figure 3. The complex result is the most notable because it results in sinusoidal oscillations. The solution begins with a guess of the homogeneous solution. The three cases result in three different forms of solutions. and the solution of a quadratic equation.3. as shown.5 Solve the following diffferential equation given the initial condition. There are three possible cases that result from the solution of the quadratic equation: different but real roots. or two complex roots. but after the homogeneous solution has been found.

5 As mentioned above. the general form is. the general form is. as proven in Figure 3. R1 t R2t If the values for Y are both real. and identical. If the values for Y are both real.6 Given.3. Y = σ ± ωj σt x h = X 3 e cos ( ω t + X 4 ) The initial conditions are then used to find the values of X3 and X4. R 2 xh = X 1 e + X2 e Note: The initial conditions are then used to find the values for X 1 and X2. ·· · Ax + Bx + Cx = 0 x ( 0) = x0 and · x ( 0 ) = v0 Guess a general equation form and substitute it into the differential equation. R 1 R1t R1 t xh = X1 e + X 2 te The initial conditions are then used to find the values for X1 and X2.6. Y = R1. two identical real roots. . Yt Yt 2 Yt · ·· x h = Xe x h = YXe x h = Y Xe 2 Yt Yt Yt A ( Y Xe ) + B ( YXe ) + C ( Xe ) = 0 2 A( Y ) + B ( Y ) + C = 0 2 2 – B ± ( B ) – 4 ( AC Y = ---------------------------------------------------) = – B ± B – 4AC -----------------------------------------2A 2A Note: There are three possible outcomes of finding the roots of the equations: two different real roots. Solution of a second-order homogeneous equation Figure 3. This form is very useful for analyzing the frequency response of a system. but different. as will be seen in a later chapter. the general form is.differential equations . or two complex roots. Y = R1. If the values for Y are complex. a complex solution when solving the homogeneous equation results in a sinusoidal oscillation. Therefore there are three fundamentally different results. The most notable part of the solution is that there is both a frequency of oscillation and a phase shift.

( ( X1 + X 2 ) cos ( Ct ) + j ( X1 – X 2 ) sin ( Ct ) ) 2 2 2 ( X1 + X2 ) + j ( X 1 – X 2 ) 2 2 2 2 X1 + 2X 1 X 2 + X2 –  X1 + – 2 X1 X2 + X 2   Rt x = e ------------------------------------------------------------------------------------------------------------.    ( X1 + X2 )  x = e x = e Rt X3 cos ( Ct + X4 ) phase shift where. as shown below: ( R + C j )t ( R – Cj )t x = X1 e + X2 e Rt Cjt Rt – Cjt x = X1 e e + X2e e Rt Cjt – Cjt x = e  X1e + X2e   Rt x = e ( X1 ( cos ( Ct ) + j sin ( Ct ) ) + X2 ( cos ( – C t ) + j sin ( – C t ) ) ) Rt x = e ( X1 ( cos ( Ct ) + j sin ( Ct ) ) + X 2 ( cos ( ( Ct ) – j sin ( Ct ) ) ) ) Rt x = e ( ( X 1 + X 2 ) cos ( Ct ) + j ( X1 – X 2 ) sin ( Ct ) ) Rt x = e ( ( X 1 + X 2 ) cos ( Ct ) + j ( X1 – X 2 ) sin ( Ct ) ) 2 2 2 Rt ( X1 + X2 ) + j ( X 1 – X 2 ) x = e -----------------------------------------------------------------------.7 Consider the situation where the results of a homogeneous solution are the complex conjugate pair.cos ( Ct ) + j ----------------------. Y = R ± Cj This gives the general result.differential equations .( ( X 1 + X2 ) cos ( Ct ) + j ( X 1 – X2 ) sin ( C 2 2 2 2 X1 + 2X 1 X 2 + X2 –  X1 + – 2 X1 X2 + X 2   Rt 4X1 X 2 -------------------. X = 3 4X1 X2  ( X 1 – X2 )  X4 = atan  -----------------------.  ( X1 + X2 ) frequency Figure 3.6 Phase shift solution for a second-order homogeneous differential equation .sin ( Ct )  4X X  4X1 X2 1 2   ( X 1 – X2 )   Rt x = e 4X 1 X 2 cos  Ct + atan  -----------------------.3.( ( X 1 + X 2 ) cos ( Ct ) + j ( X 1 – X2 ) sin ( Ct ) ) 4X1 X 2 ( X1 – X2 )  ( X1 + X2)  Rt x = e 4X 1 X2  -----------------------.

x ( t ) = e + 3te –t –t Figure 3.7 Drill Problem: Second order homogeneous differential equation The methods for solving non-homogeneous differential equations builds upon the methods used for the solution of homogeneous equations. To find the homogeneous solution the nonhomogeneous part of the equation is set to zero.3. An example of the solution of a first-order nonhomogeneous equation is shown in Figure 3. This process adds a step to find the particular solution of the equation.8 Solve the following diffferential equation given the initial condition.differential equations .8. · · x ( 0) = 1 x(0 ) = 2 ·· x + 2x + x = 0 ans. To find the particular solution the final .

and if an incorrect guess is made. The overall response of the system can be obtained by adding the homogeneous and particular parts.9 form must be guessed. Finally the homogeneous and particular solutions are added to get the final equation. An example is given below. 4t 4t · xp = C1 e \x p = 4C 1 e Substitute these into the differential equation and solve for A. guess the particular solution by looking at the form of ’f(t)’. The final result is converted into a phase shift form. and the values of the coefficients are found. 4t 4t 4t 6  4C 1 e  + 2  C 1 e  = 5e     524C 1 + 2C 1 = 5 ∴C 1 = ----26 Combine the particular and homogeneous solutions. it will be unsolvable. if we are given 4t · 6x + 2x = 5e A reasonable guess for the particular solution is. The homogeneous equation deals with the response to initial conditions. in Figure 3. In this example the forcing function is sinusoidal. –B t -A xh = x0 e Next.e 4t + x e 2 x = x p + x h = ----0 26 Figure 3. In the case below the guess should be similar to the exponential forcing function. and the principle of superposition applies. This step is highly subjective. .9. This is acceptable because the equations are linear.8 Solution of a first-order non-homogeneous equation The method for finding a particular solution for a second-order non-homogeneous differential equation is shown in Figure 3. Generally. –6 t -5. This is then substituted into the equation. and the particular solution deals with the response to forced inputs.3. just make another guess and repeat the process.differential equations . For example.3. find the homogeneous solution as before. When this happens. so the particular result should also be sinusoidal. · Ax + Bx = Cf ( t ) x ( 0 ) = x0 First.

Use the initial conditions to determine the coefficients in the homogeneous solution. When this happens. This step is highly subjective.3333 ( – 0.differential equations . just make another guess and repeat the process. x p = A sin ( 3t ) + B cos ( 3t ) · x p = A cos ( 3t ) – B sin ( 3t ) ·· x p = – A sin ( 3t ) – B cos ( 3t ) Substitute these into the differential equation ans solve for A and B.2523 Next.10 · x ( 0 ) = x0 and x ( 0 ) = v 0 1. For the purpose of illustration an example is given below.3333 ( sin ( – 0. x p = – 0.7568 sin ( 3t ) + 0. σt x h = X3 e cos ( ω t + X4 ) or σt σt x h = X1 e + X 2 te or σ1 t σ2 t xh = X 1e + X2e 2.8584 ) cos ( 3t ) ) x p = 0.2179 6A = 2 sin 4 A = – 0. rearrange the equation to phase shift form.3. 2 ( A sin ( 3t ) + B cos ( 3t ) ) + 6 ( A cos ( 3t ) – B sin ( 3t ) ) + 2 ( – A sin ( 3t ) – B cos ( 3t ) ) = 2 sin ( 3t + 4 ) ( 2A – 6B – 2A ) sin ( 3t ) + ( 2B + 6A – 2B ) cos ( 3t ) = 2 sin ( 3t + 4 ) ( – 6B ) sin ( 3t ) + ( 6A ) cos ( 3t ) = 2 ( sin 3t cos 4 + cos 3t sin 4 ) ( – 6B ) sin ( 3t ) + ( 6A ) cos ( 3t ) = ( 2 cos 4 ) sin ( 3t ) + ( 2 sin 4 ) cos ( 3t ) –6B = 2 cos 4 B = 0. and if an incorrect guess is made it will be unsolvable. if we are given ·· · 2x + 6x + 2x = 2 sin ( 3t + 4 ) A reasonable guess is. ·· · Ax + Bx + Cx = Df ( t ) Figure 3. Guess the particular solution by looking at the form of ’f(t)’.2523 sin ( 3t ) + 0.9 Solution of a second-order non-homogeneous equation . In the case below it should be similar to the sine function.2179 cos ( 3t ) x p = 0.6536 cos ( 3t ) ) x p = 0. Generally.8584 ) sin ( 3t ) + cos ( – 0. For example.8584 ) 3. Find the homogeneous solution as before.3333 cos ( 3t – 0.

the forms in Figure 3.10 General forms for particular solutions .10 can be helpful.differential equations .11 When guessing particular solutions.3. Forcing Function A Ax + B Ax e B sin ( Ax ) or B cos ( Ax ) Guess C Cx + D Ax or Ax Ce Cxe C sin ( Ax ) + D cos ( Ax ) or Cx sin ( Ax ) + xD cos ( Ax ) Figure 3.

As expected.11 Drill Problem: Second order non-homogeneous differential equation An example of a second-order system is shown in Figure 3. This is followed with the general solution of . x ( t ) = – e – te + 1 –t –t Figure 3.12 Solve the following diffferential equation given the initial condition.12.3. · x ( 0) = 0 x(0 ) = 0 ·· · x + 2x + x = 1 ans.differential equations . it begins with a FBD and summation of forces.

R2 R1t R2t yh = C1 e + C2 e Figure 3. with no oscillation.13 the homogeneous equation. Kd and Ks lead to the case of two different positive roots.3. At time ’t=0’ the system is released and allowed to move. Real roots are assumed thus allowing the problem solution to continue in Figure 3.13.differential equations . . The final result is a secondorder system that is overdamped.13. M Assume the system illustrated to the right starts from rest at a height ’h’. This would occur if the damper value was much larger than the spring and mass values. Ks Mg ·· My + Ks y · K dy Fg Kd y ∑ Fy · ·· = – Mg + K s y + K d y = – M y ·· · My + K d y + K s y = Mg Find the homogeneous solution. Thus.12 Second-order system example The solution continues by assuming a particular solution and calculating values for the coefficients using the initial conditions in Figure 3. At At 2 At · ·· yh = e y h = Ae yh = A e ·· · My + K d y + K s y = 0 2 At At At M  A e  + K d ( Ae ) + K s ( e ) = 0   2 MA + K d A + K s = 0 2 – K d ± K d – 4MK s A = ------------------------------------------------2M Let us assume that the values of M. A = R 1.

14 Next. find the particular solution. combine the solutions and solve for the unknowns using the initial conditions.+  ---------------------.+ C 1 e 1 + C 2 e 2 Ks y(0) = h y' ( 0 ) = 0 Mg 0 0 h = ------. ---------------------.3. ---------------------. · yh = 0 yp = C ·· yh = 0 M ( 0 ) + K d ( 0 ) + K s ( C ) = Mg Mg C = ------Ks Now. K s h – Mg – R1 – R1 R t – R 2 K s h – Mg R t Mg y ( t ) = ------.13 Second-order system example (continued) . add the homogeneous and particular solutions and solve for the unknowns using the initial conditions.  ----------------- e 2  Ks  K s   R1 – R2 K s   R1 – R2 Figure 3.+  ---------------------.  ----------------- e 2  R – R   R – R  Ks  K R  K s 1 2 1 s 1 2 – R1 R2 K s h – Mg K s h – Mg R t R t Mg y ( t ) = ------.C 2 R1 R2 Mg – ----.  -----------------   R – R  R1 Ks 1 2 Now.+ C 1 e + C 2 e Ks Mg C 1 + C 2 = h – ------Ks y' ( t ) = R1 C 1 e 0 R1 t + R2C2e 0 R 2t 0 = R 1 C1 e + R 2 C2 e 0 = R 1 C1 + R 2 C 2 – R2 C 1 = -------.differential equations .C 2 + C 2 = h – ------Ks R1 K s h – Mg – R1 C 2 =  ---------------------.  ----------------- e 1 + -------.  -----------------   R – R  Ks 1 2 – R 2 K s h – Mg – R1 C 1 = -------. R t R t Mg y ( t ) = y p + y h = ------.  ----------------- e 1 +  ---------------------.

9273 ) Figure 3.3.differential equations . A sin ( ω t + θ ) A ( cos ω t sin θ + sin ω t cos θ ) ( A sin θ ) cos ω t + ( A cos θ ) sin ω t B cos ω t + C sin ω t where.= ----------sin θ cos θ To find A. The second-order forms may . and only oscillate if forced to do so. B A sin θ --. 3 cos 5t + 4 sin 5t 2 2 3 + 4 sin  5t + atan 4 --  3 5 sin ( 5t + 0.= tan θ C A cos θ B θ = atan  ---  C To find A. (method #1) BCA = ---------. The e-to-the-negative-t forms are the first-order responses and slowly decay over time.3 RESPONSES Solving differential equations tends to yield one of two basic equation forms. B = A sin θ C = A cos θ (the desired final form) (the form we start with) To find theta.= --------------. (method #2) 2 2 A = B +C For example.15 Given.14 Proof for conversion to phase form 3. They never naturally oscillate.

2% of the way from the initial to final values for the system.15. 3. When an input to a system has changed. The response function for these systems is natural decay or growth as shown in Figure 3. we can calculate the time constant. It is a measure of how quickly the system responds to a change. the system output will be approximately 63% of the way to its final value when the elapsed time equals the time constant. Assuming the change started at t=0. The other method is to look for the time when the output value has shifted 63. That time at the intersection is the time constant. one by extending the slope of the first part of the curve until it intersects the final value line.3.16 include natural oscillation.3. the time at this point corresponds to the time constant. The initial and final values of the function can be determined algebraically to find the first-order response with little effort. initial and final values. The time constant can be found two ways.1 First-order A first-order system is described with a first-order differential equation.differential equations . The time constant for the system can be found directly from the differential equation. If we have experimental results for a system. .

632 Figure 3. τ –τ y ( τ ) = y 1 + ( y 0 – y 1 )e y ( τ ) = y 1 + ( y 0 – y 1 )e –1 y0 τ t y ( τ ) = y 1 + ( y 0 – y 1 )0.8 s.16 calculates the coefficients for a first-order differential equation given a graphical output response to an input. and finding the point where it intercepts the steady-state value. and will be examined in a later chapter.368 y ( τ ) = 0.differential equations . Some additional numerical calculation leads to the final differential equation as shown. This can then be used to calculate the remaining coefficient. as shown below. If we consider the steady state when the speed is steady at 1400RPM.17 y y0 y ( t ) = y 1 + ( y 0 – y 1 )e t –τ · 1 y + -. This gives an approximate time constant of 0.y = f ( t ) τ y1 time constant τ OR y y1 t Note: The time will be equal to the time constant when the value is 63.368 y ( τ ) = 1 + ( 0 – 1 )0. This simplifies the equation and allows us to calculate a value for the parameter K in the differential equation.3. the first derivative will be zero.15 Typical first-order responses The example in Figure 3. The differential equation is for a permanent magnet DC motor. .2% of the way to the final value. The time constant can be found by drawing a line asymptotic to the start of the motor curve.

6rads  dt  2 K0 +  K - 146.328 JR 1· ω + -----. 10 ----- JR  JR K = 0.differential equations .0682  ----.8s 1s  K 2 1 ------ = --------0.8s K----.328Vs 0. V ------   dt  JR s  JR 1s 2s 3s For steady-state –1 d ---- ω = 0 ω = 1400RPM = 146.16 Finding an equation using experimental data .= 18.0682 1400 RPM τ ª 0. = -------- JR 0. use the differential equation and the speed curve when Vs=10V is applied: 1400 RPM 2 dK ---- ω +  K - ω =  ----.8 Figure 3.8s  JR  K10.6 =  ----.18 For the motor.ω = 18.3.

19 Find the differential equation when a step input of Vs=12V is applied: 1800 RPM dKK ---- ω +  ----.15 12 ( 0. After this.2s 0.6s 2 ans.3. The modeling starts with a FBD and a sum of forces.17 Drill problem: Find the constants for the equation A simple mechanical example is given in Figure 3.4s 0. Next. and the two solutions are combined. d11800 -V  ---- ω + --------. The initial conditions are used to find the remaining unknown coefficients. . ω =  ----.18. the homogenous solution is found by setting the nonhomogeneous part to zero and solving.ω = ------------------- dt 0. V  dt  JR  JR s 0.differential equations .15 ) s Figure 3. the particular solution is found.

20 Find the response to the applied force if the force is applied at t=0s. Bt Bt · y h = Ae y h = ABe Bt Bt K d ( ABe ) + K s ( Ae ) = 0 K d B + Ks = 0 Ks y –K s B = -------Kd Next. and find the remaining unknown.3.differential equations . –Ks ---------t Kd Fy ( t ) = y p + y h = Ae + ----Ks y ( 0) = h 0 F Fh = Ae + ----∴A = h – ----Ks Ks The final solution is. Assume the system is initially deflected a height of h. F y dK d  ---- y  dt Find the homogeneous solution. F y ( t ) =  h – ----- e  K s –K s -------. yp = C K d ( 0 ) + Ks ( C ) = F · yp = 0 + F ∑ Fy d= – F + K s y + K d  ---- y = 0  dt · Kd y + Ks y = F Ks Kd F∴C = ----Ks Combine the solutions. find the particular solution.18 First-order system analysis example .t Kd F + ----Ks Figure 3.

· y + 10y = 5 y ( t ) = y 1 + ( y 0 – y 1 )e t –τ ans.5 – 19. Assume the system starts at y=-20. The output response lags the input and the delay is determined by the system’s time constant.19 Drill problem: Developing the final equation using the first-order model form A first-order system tends to be passive. .21 Use the general form given below to solve the problem in Figure 3. A first-order system will not oscillate unless the input forcing function is also oscillating.differential equations .5e – 10t Figure 3.3. y ( t ) = 0.18 without solving the differential equation. meaning it doesn’t deliver energy or power.

and the system will not oscillate because it is overdamped. and never any oscillation (more like a first-order system). If the damping coefficient is greater than one the damped frequency becomes negative. as seen in Figure 3.3. A second-order system.Approximate frequency of control system oscillations. The actual frequency of oscillation . A typical sinusoidal second-order response is shown in Figure 3. The damped frequency will be lower than the natural frequency when the damping coefficient is between 0 and 1. If > 1 overdamped.2 Second-order A second-order system response typically contains two first-order responses. ωd Figure 3. and overall response time will slow. Notice that the damped frequency of oscillation is the actual frequency of oscillation. the first peak becomes infinite in height. the frequency of oscillation.20 The general form for a second-order system When only the damping coefficient is increased.21. Notice that the coefficients of the differential equation include a damping coefficient and a natural frequency. and system will oscillate. As damping factor approaches 0.3. 2 ·· · y + 2 ζω n y + ω n y = f ( t ) σ = ζω n y = A + ( B – A )e ωd = ωn 1 – ζ –σ t cos ( ω d t ) 2 B A ωn ξ Natural frequency of system .22 3.If < 1 underdamped. and a typical response to a stepped input. or a first-order response and a sinusoidal component.20. Critical damping occurs when the damping coeffi- .differential equations . Damping factor of system . If =1 critically damped. These can be used to develop the final response. given the initial conditions and forcing function. When the damping coefficient is 0 the system will oscillate indefinitely.It is below the natural frequency because of the damping.

differential equations .21 The effect of the damping coefficient When observing second-order systems it is more common to use more direct measurements of the response. At this point both roots of the differential equation are equal. The system will not oscillate if the damping coefficient is greater than or equal to 1.= 0. Some of these measures are shown in Figure 3.23 cient is 1.3.707 2 ξ = 1 (critical) (overdamped) ξ»1 Figure 3. The rise .5 (underdamped) 1ξ = -----.22. ξ = 0 ξ = 0.

1 ∆ x –σ t t p = time tortek fs a ip e ss = steady state error x = ∆ xe cos ( ω d t ) b ts ∆x 0. the inverse is the damping frequency. tp 1 fd = -T T e ss 0.9 ∆ x tr where. The period of oscillation can be measured directly as the time between peaks of the oscillation. f d = period and frequency . (Note: don’t forget to convert to radians. The overshoot is the height of the first peak.damped b = overshoot 0.) The damped frequency can also be found using the time to the first peak. but if it were larger the system would have a shorter settling time.) ∆ x = total displacement T. Using the time to the first peak. and the overshoot the damping coefficient can be found.5 ∆ x π ω d ≈ ---tp σ = ξω n –σ tp b = ----e ∆x 1 ξ = ----------------------------π ------- 2 + 1  t σ p (1) (2) (3) (4) .3.02 ∆ x 0. as half the period. The settling time indicates how long it takes for the system to pass within a tolerance band around the final value. The permissible zone shown is 2%.02 ∆ x 0. Note: This figure is not to scale to make details near the steady-state value easier to see.differential equations . t r = rise time (from 10% to 90%) t s = settling time (to within 2-5% typ.24 time is the time it takes to go from 10% to 90% of the total displacement. and is comparable to a first order time constant.

= ω d ----2 2 ξ ξ 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 σω n = ----–ξ (1) ωn 1 – ξ = ω d 2 (2) ωd 1---. A + 2ξω n A + ω n = 0 – 2ξωn ± 4ξ ω n – 4ω n A = ----------------------------------------------------------.3. and the generic second order equation form.differential equations .= ----.  ∆x σ = – ----------------tp b ≈ ∆xe – σtp – σt cos ( ω d t + C 2 ) (4) (5) Figure 3.22 Characterizing a second-order response (not to scale) Note: We can calculate these relationships using the complex homogenous form.= jω d 2 4ξ ω n – 4ω n = 4 ( –1 )ω d ω n – ξ ωn = ωd 2 2 σ----.+ 1 2 σ (3) The time to the first peak can be used to find the approximate decay constant x ( t ) = C1 e π ω d = --tp (1) bln  ----.– ξ 2 σ .25 Figure 3.= σ = – ξω n 2 4ξ ω n – 4ω n ----------------------------------.23 Second order relationships between damped and natural frequency .= σ ± jω d 2 – 2ξω n --------------.+ 1 2 2 ξ σ 2 1 ξ = ------------------2 ωd ----.

0 Write a function of time for the graph. using a ruler.0 4. Using the dashed lines determine the settling time. to get values.3. (Note: measure. ans.2 y(t) 2.) Find the natural frequency and damping coefficient to develop the differential equation.differential equations .26 2. t<4 t≥4 y(t ) = 0 y(t ) = Figure 3.0 t 0 5.24 Drill problem: Find the equation given the response curve .

The example in Figure 3. In this case the resulting homogeneous solution yields four roots. one with about one period showing.differential equations . The underlying function is a firstorder response that drops at the beginning.3. A possible higher order system response is seen in Figure 3. There are two sinusoidal functions superimposed.3 Other Responses First-order systems have e-to-the-t type responses. the other with a much higher frequency. Second-order systems add another e-to-the-t response or a sinusoidal excitation. As we move to higher order linear systems we typically add more e-to-the-t terms. Figure 3. The result in this case are two real roots. The two real roots result in e-to-the-t terms.25 An example of a higher order system response The basic techniques used for solving first and second-order differential equations can be applied to higher order differential equations.27 3.3. but levels out.25.26 shows a fourth order differential equation. . and a complex pair. while the complex pair results in a damped sinusoid. The particular solution is relatively simple to find in this example because the non-homogeneous term is a constant. and/or more sinusoidal terms. although the solutions will start to become complicated for systems with much higher orders.

At xh = e dAt  d  2 2 At  ---- 3 x = A3 e At d---.x p = 0  dt p  dt p dt 0 + 13 ( 0 ) + 34 ( 0 ) + 42 ( 0 ) + 20A = 5 A = 0. and the solve for the coefficient.28 Given the homogeneous differential equation d.4  ---- x = 0  dt p Guess a particular solution.25 Figure 3.2 d. .x h = A e ---.4  ---- x = A4 e  dt h Substitute the values into the differential equation and find a value for the unknown. dd.26 Solution of a higher order differential equation The example is continued in Figure 3. – 1 – j.differential equations .4 d. – 10.28 where the initial conditions are used to find values for the coefficients in the homogeneous solution. 4 At 3 At 2 At At At A e + 13A e + 34A e + 42Ae + 20e = 0 4 3 2 A + 13A + 34A + 42A + 20 = 0 A = – 1. – 1 + j xh = C1 e –t + C 2e –10t –t + C 3 e cos ( t + C 4 ) d.3.3  ---- x = 0  ---- x = 0 xp = A ---.27 and Figure 3.xh = Ae  dt h  dt dt At d.2 d ---- x + 13  ---- x + 34  ---- x + 42  ---- x + 20x = 5  dt  dt  dt  dt Guess a solution for the homogeneous equation.3 d.

–t – 10t –t x ( t ) = C 1 e + C2 e + C 3 e cos ( t + C 4 ) + 0.3.25 0 = C 1 + C 2 + C 3 cos ( C 4 ) + 0.x = – C 1 e – 10C 2 e – C 3 e cos ( t + C 4 ) – C 3 e sin ( t + C 4 ) dt h 0 = – C 1 – 10C 2 – C 3 cos ( C 4 ) – C 3 sin ( C 4 ) Equations (1) and (3) can be added to get the simplified equation below.2  ---- x = C e –t + 100C e – 10t + C e – t cos ( t + C ) + C e – t sin ( t + C ) + 1 2 3 4 3 4  dt h –t –t C 3 e sin ( t + C 4 ) – C 3 e cos ( t + C4 ) 0 = C 1 + 100C 2 + C 3 cos ( C 4 ) + C 3 sin ( C 4 ) + C 3 sin ( C 4 ) – C 3 cos ( C 4 ) (5) 0 = C 1 + 100C 2 + 2C 3 sin ( C 4 ) Equations (4) and (5) can be combined.5 (6) d.25 –t – 10t –t –t d---. 0 = – 9C 2 – C 3 sin ( C 4 ) + 0.25 C 3 sin ( C 4 ) = – 9C 2 + 0.25 (3) (4) d. assuming the system starts at rest and undeflected.25 (1) (2) C 3 cos ( C 4 ) = –C 1 – C 2 – 0.29 Solve for the unknowns.27 Solution of a higher order differential equation . 0 = C 1 + 100C 2 + 2 ( – 9C 2 + 0.differential equations .25 ) 0 = – 17C 1 + 100C 2 + 0.3  ---- x = – C e – t + ( – 1000 )C e – 10t – 2C e – t sin ( t + C ) + 2C e – t cos ( t + C )  dt h 1 2 3 4 3 4 (7) 0 = – C 1 + ( – 1000 )C 2 – 2C 3 sin ( C 4 ) + 2C 3 cos ( C 4 ) Figure 3.

00109 984 1 C 1 =  – -------- ( – 0. which will be covered in later chapters. In other . obtaining the response is important.25 ) + 2 ( – C 1 – C 2 – 0.0242e + ( – 0.25 --------------------------. but evaluating the results is more important.3.00109 ) – 0.760 ) + 0.25 Figure 3.377 The final response function is. The most critical design consideration is system stability.760 tan ( C 4 ) = --------------------------------------------------------------------------– ( 0.= --------------------------------------C 3 cos ( C 4 ) – C 1 – C 2 – 0.0242 ) – ( – 0.4 RESPONSE ANALYSIS Up to this point we have mostly discussed the process of calculating the system response.760 ) = – 9 ( – 0.00109 )e + ( – 0. –t – 10t –t x ( t ) = 0. C 3 sin ( C 4 ) – 9C 2 + 0.28 Solution of a higher order differential equation (cont’d) In some cases we will have systems with multiple differential equations. such as a car "cruise control". or nonlinear terms. 984 1 0 = – 17   – -------- C 2 – -- + 100C 2 + 0.5  3  3 0 = 5676C 2 + 6. 3.25 C 3 = –0.00109 ) + 0.30 Equations (2 and (4) are substituted into equation (7).25 Equation (4) can be used the find the remaining unknown. In most cases a system should be inherently stable in all situations.25 C 4 = –0.25 – 9 ( – 0. In these cases we may use other techniques.25 ) 0 = – 3C 1 + ( – 984 )C 2 – 1 984 1 C 1 =  – -------- C – - 3  2 3 Equations (6) and (8) can be combined.differential equations . C 3 sin ( – 0.00109 ) + 0. 0 = – C 1 + ( – 1000 )C 2 – 2 ( – 9C 2 + 0. As an engineer.0242  3  3 Equations (2) and (4) can be combined. such as numerical integration.377 )e cos ( t – 0.00109 ) – -C 1 = 0. In these cases explicit analysis of the equations may not be feasible.1666667 (8) C 2 = –0.

without overshoot. Beyond establishing the stability of a system.31 cases an unstable system may be the objective. possibly leading to motion sickness. The steady-state effects occur after some period of time when the system is acting in a repeatable or non-changing form. The transient effects at the beginning include a quick rise time and an overshoot.differential equations . such as an explosive device. If a step input causes the system to go to infinity. This includes the time constant for a first-order system. we must also consider general performance. If it is critically damped it will reach the floors quickly. the system is probably resonating. Simple methods for determining the stability of a system are listed below: 1. The steady-state response settles down to a constant amplitude sine wave. and will become unstable. or worse. it will be inherently unstable. If it is under damped the elevator will oscillate. A ramp input might cause the system to go to infinity. The transient effects are closely related to the homogeneous solution to the differential equations and the initial conditions. Engineers distinguish between initial setting effects (transient) and long term effects (steady-state).29 shows a system response. If the response to a sinusoidal input grows with each cycle. If the elevator is over damped it will take longer to get to floors. or damping coefficient and natural frequency for a second-order system. . assume we have designed an elevator that is a second-order system. 3. Figure 3. 2. if this is the case. the system might not respond well to constant change.3. For example.

A basic linear differential equation has coefficients that are constant. .3. and the derivatives are all first order. Figure 3. Examples of non-linear differential equations are shown in Figure 3.5 NON-LINEAR SYSTEMS Non-linear systems cannot be described with a linear differential equation.30.29 A system response with transient and steady-state effects 3. The steady state response in mainly predicted with the particular solution.32 Steady-state Transient Note: the transient response is predicted with the homogeneous solution. such as a non-decaying oscillation. although in some cases the homogeneous solution might have steady state effects. 2 · x+x = 5 ·2 x +x = 5 · x + log ( x ) = 5 · x + ( 5t )x = 5 Note: the sources of nonlinearity are circled.differential equations .

• aerodynamic drag • forces that are a squared function of distance • devices with non-linear responses Explicitly solving non-linear differential equations can be difficult. In this case the terminal velocity is calculated by setting the acceleration to zero.33 Figure 3. 3. and will typically involve complex solutions for simple problems. and the resulting differential equation.differential equations . .1 Non-Linear Differential Equations A non-linear differential equation is presented in Figure 3.30 Examples of non-linear differential equations Examples of system conditions that lead to non-linear solutions are. and so it cannot be analyzed with the previous methods.5.31.3. (Note: This coefficient is calculated using the drag coefficient and other properties such as the speed of sound and cross sectional area.) The FBD shows the sum of forces. The velocity squared term makes the equation non-linear. It involves a person ejected from an aircraft with a drag force coefficient of 0. This results in a maximum speed of 126 kph.8.

81 ------------. The term is then integrated in Figure 3. . –2 –3 – 1 · 2 ( 0 ) + 8 × 10 m ( y ) = 9.8 ( y· ) – Mg = –My 2 m 2 2 N·· 100kgy + 0.· 2 kg m 2 2 ·· 100kgy + 0. and time on the right.( y ) = 981kg ---2 2 2 s m –1 s 2 –2 ·· · 100y + 0.8m ( y ) = 981ms –2 –3 – 1 · 2 ·· y + 8 × 10 m ( y ) = 9.81 9. The aerodynamic drag will introduce a squared variable. 2 2 2 ·· 0. as we would expect.differential equations . therefore making the equation non-linear.8 Ns ( y ) = 981N --------.33 and Figure 3. but contains e-to-t terms.8 Ns ( y' ) --------∑ Fy = 0.81ms · = -----------------------------. as shown in Figure 3.81ms The terminal velocity can be found be setting the acceleration to zero.34.34 Consider the differential equation for a 100kg human ejected from an airplane.8 Ns ( y ) = 100kg9.8kg ---.31 Development of a non-linear differential equation The equation can also be solved using explicit integration.· 2 m ·· 100kgy + 0.· y 2 Mg m 2 ms.32. The final form of the equation is non-trivial. In this case the equation is separated and rearranged to isolate the ’v’ terms on the left.3.-----.= ------------------.m 2 s – 2 = 35.0 m = 126 km -------y –3 s h –3 – 1 8 × 10 8 × 10 m Figure 3.81ms –2 9.

8m ( y ) = 981ms –1 2 –2 · 100v + 0.8m v = 981ms dt –2 –1 2 dv 100 ----.25m 2 s –2 dv = t + C1 ∫ -----------------------------------------------------------------.8m – 125m -------------------------------------------∫ v 2 – 1226.8m v –1 2 100 ------------------–1 – 0. –1 · 2 –2 ·· 100y + 0.32 – 125m = t + C1 Development of an integral .3.02 s  Figure 3.8m v = 981ms –1 2 –2 dv 100 ----.35 An explicit solution can begin by replacing the position variable with a velocity variable and rewriting the equation as a separable differential equation.02 s   v – 35.dv = ∫ dt –2 2 981 ------------------.dv m m   ---- v + 35.8m ∫ -------------------------------------------.ms + v –1 – 0.= 981ms – 0.differential equations .+ 0.8m v dt 100 -------------------------------------------------dv = dt 981ms –2 – 0.

02 m + Bv + B  35.dv = t + C 1  v + 35.3.= t + C 1 ∫ ( a + bx ) dx = -----------------------.  --.02 --s Figure 3.s 35.785s s --------------------------.02 m --1.  --. --.36 This can be reduced with a partial fraction expansion.02 m – 1. In this case the integration constants can be left off because they are redundant with the one on the right hand side.differential equations .785s – 1.02 m ( – A + B ) = – 125m --s A+B = 0 A = –B ∫ 35.02 m --C 1 1.02 m --.785s ln v + 35.= e v – 35.785s -------------------------------.02 m  v – 35.dv = t + C 1  v + 35.785s A = 1.02 m --. A B -------------------------------.785s ln v – 35.  s s The integral can then be solved using an identity from the integral table.02 m ( – A + B ) = – 125m --s 125 ( – ( – B ) + B ) = – -----------.+ C b v – 35.02 m --s t --------------v + 35.02 m = – 125m --.   s s v ( A + B ) + 35.785s ∫ 1.02 m = t + C 1 ----s s v + 35.02 B = – 1.785s s--------------------------.02 m --s t --------------.  s s Av – A  35.+ ------------------------------.785s ln --------------------------.+ C 1 v + 35.33 Solution of the integral .= e e m v – 35.02 m  v – 35.02 m --s–1 ln a + bx 1.+ ------------------------------. 1.

=  ------------ = -0 = 35.02 --s 1.= C e 1.02 --. m t v + 35.785s m 1.02 --s An initial velocity of zero can be assumed to find the value of the integration constant m 0 --------------0 + 35.  1 + e 1.02 m s 1 = C2 This can then be simplified.02 --.785s --------------------------2 --0 – 35.34 Solution of the integral and application of the initial conditions As evident from the example.785s m − − v 1 + e – 35. Typically the numerical methods discussed in the next chapter are preferred.= ± ve + 35.e s s t t --------------- -------------- 1. + --.785s m − v + 35.02 --.02 --.02 --s. 3.02 --s.02 -- = + 35.5.---------------------------- v = 35.02 --. t s --------------. non-linear equations are involved and don’t utilize routine methods.785s − e 0 m − 1 – 1 1–1 – 1 .785s e m – 1 v = 35.2 Non-Linear Equation Terms If our models include a device that is non-linear and we want to use a linear tech- .02 ----------------1.  1 − e 1.differential equations .785s t --------------- t --------------- Figure 3.785s  + t  -------------- 1.--------------.= C e 1.= ± e m v – 35. and the absolute value sign eliminated. ----------------------.02 m  +  1+1   1 + 1 − 2 s t s --------------.785s s--------------------------.3.e s s   t  -------------- 1.37 m t --------------v + 35.785s m 1.785s --------------------------2 m v – 35.

we will need to linearize the model before we can proceed. A linearized differential equation can be approximately solved using known techniques as long as the system doesn’t travel too far from the linearized point. This equation will be approximately correct as long as the first derivative doesn’t move too far from 100. 3.35 shows the linearization of a non-linear equation about a given operating point. 2. Develop a linear equation. When this value does. The example in Figure 3. . 1.38 nique to solve the equation. the new velocity can be calculated. Use the linear equation for the analysis. Pick an operating point or range for the component. Find a constant value that relates a change in the input to a change in the output.3. A non-linear system can be approximated with a linear equation using the following method. 4.differential equations .

65 ( 0 ) + 4A = 200 Initial conditions: – 0. Homogeneous: · 12.39 Assume we have the non-linear differential equation below.316t y ( t ) = – 40e + 50 A = –0. ·2 y + 4y = 200 y ( 0 ) = 10 We can make the equation linear by replacing the velocity squared term with the velocity times the actual velocity. If the velocity (first derivative of y) changes significantly.316t y ( t ) = Ce + 50 0 10 = Ce + 50 – 0. It can be solved by linearizing the value about the operating point Given.65y + 4y = 0 12.316 A = 50 C = – 40 Figure 3.316t y h = Ce Particular: yp = A 12.35 Linearizing a differential equation .65 · 12.3. · y = ± 200 – 4y · y ( 0 ) = ± 200 – 4 ( 10 ) = ± 12. As long as the system doesn’t vary too much from the given velocity the model should be reasonably accurate.differential equations .65y + 4y = 20 This system may now be solved as a linear differential equation.65A + 4 = 0 – 0. then the differential equation should be changed to reflect this.

1 ) + 50 = 11.327t y h = Ce Particular: yp = A 12. Homogeneous: · 12.327 A = 50 C = –35.24 Note: a small change – 0.5. Another example is tension in a cable.1s.24 = Ce + 50 – 0. the forces at work are continually changing. but once motion starts it is replaced with a smaller kinetic friction.327t y ( t ) = Ce + 50 0. But.36 Linearizing a differential equation 3.25 ( 0 ) + 4A = 200 Initial conditions: – 0. For example a system often experiences a static friction force when motion is starting.1 ) = – 40 ( – 0. then the differential equation should be changed to reflect this. y ( 0.25A + 4 = 0 – 0.25 dt 12.316t y ( t ) = – 35.1 ) d---.y ( 0. The .37.25y' + 4y = 20 Now recalculate the solution to the differential equation.1 11.316 ( 0. A mass is pulled by a springy cable.3 Changing Systems In practical systems. When in tension a cable acts as a spring.1 ) = – 40e – 0.3. when in compression the force goes to zero.316 )e = 12. Higher accuracy can be obtained using smaller steps in time.25y + 4y = 0 12. Figure 3.316 ( 0. Consider the example in Figure 3. and as the analysis progresses the equations will adjust slowly.070575 Notice that the values have shifted slightly. For example we could decide to recalculate the equation value after 0.40 If the velocity (first derivative of y) changes significantly.07e + 50 A = – 0.differential equations .

– -------------kgm kgs 100kg FF –2 ·· m x 2 + 10s x 2 = 1 ---.t – -------------3 100kg s For the cable/spring in compression x1 – x2 < 0 FF ·· M = 100kg ∑ Fx = –FF = Mx2 ·· ·· Mx 2 – FF = 100kgx 2 ·· 100kgx 2 = –F F FF ·· x 2 = – -------------100kg Figure 3.1 µ s = 0.3 N K s = 1000 --m M = 100kg x 1 = 0. only restricted by friction forces and inertia.1 m t – x 2 = 100kgx 2 . while the block is free to move.3.x 2 = 1000 --.0. -- m s ·· N N --100kgx 2 + 1000 --.37 A differential equation for a mass pulled by a springy cable . The friction force will be left as a variable at this point. 0.1 m t – FF m m s FF ·· N -x N -t x 2 + 10 --------.2 = 1 ------.1 m t --s An FBD and equation can be developed for the system.41 right hand side of the cable is being pulled at a constant rate.differential equations . At the beginning all components are at rest and undeflected. For the cable/spring in tension x 1 – x 2 Š 0 FF ·· K (x – x ) M = 100kg s 1 2 ∑ F x = – FF + K s ( x 1 – x 2 ) = Mx 2 ·· Mx 2 ·· N – F F + 1000 --. x2 µk = 0.

3kgm ----------------------3 2 s 100kgs t = 2.39 Analysis of the object before motion begins .differential equations .t = 294. In this case the cable/spring will be undeflected with no force.3 N K s = 1000 --m M = 100kg x 1 = 0.= 981N kg static friction d---. NN = 100kg9.1 m t --s An FBD and equation can be developed for the system.81 ----.1N Figure 3.943s Figure 3.3N kinetic friction d---. ·· x2 = 0 x2 = 0 F F = 294.38 Friction forces for the mass The analysis of the system begins by assuming the system starts at rest and undeflected.42 x2 µk = 0.x 2 £ 0 m --dt s 0N £ F F < µ s N < 294.3N FF –2 ·· m x 2 + 10s x 2 = 1 ---.t – -------------3 100kg s –2 m t – 294.943s Therefore the system is static from 0 to 2.3. The friction force will be left as a variable at this point.x 2 > 0 m --dt s F F = µ k N = 98.1 µ s = 0. Therefore the block will stay in place until the cable stretches enough to overcome the static friction.3N 0 + 10s 0 = 1 ---. and the mass will be experiencing static friction.----------------100kg 3 s m 1 ---.

This stage of motion requires the solution of a differential equation.1N m ( At + B ) = 1 ---.16js –1 98.3.1N 10s B = – -------------100kg Analysis of the object after motion begins A = 0.43 After motion begins the object will only experience kinetic friction.t – -------------3 100kg s –2 m 10s A = 1 ---3 s –2 98. –2 ·· 98. x p = At + B 0 + 10s –2 · xp = A ·· xp = 0 A = ± 3.1N m x 2 + 10s x 2 = 1 ---. and continue to accelerate until the cable/spring becomes loose in compression.0981m Figure 3.16t + C 2 ) For the particular.differential equations .40 . –2 ·· x 2 + 10s x 2 = 0 –2 A + 10s = 0 x h = C 1 sin ( 3.t – -------------100kg 3 s For the homogeneous.1 m --s B = – 0.

943s ) + C 2 ) + 0.16t – 7.1962 -----------------C 1 cos ( 9.199 ( 3.16 ( 2.0316 C 1 sin ( 9.differential equations .1 m --dt s 0 = 3.29988 + C 2 ) = – 0.1 m --s C 1 cos ( 9.943s ) = 0 m --dt s x ( t ) = C 1 sin ( 3.16t + C 2 ) + 0.29988 + C 2 ) = 6.1962 C 1 = --------------------------------------------------.16t – 7.0316 tan ( 9.29988 – 7.889 + π n )rad – 0.1 m ( 2.943s ) – 0.16C 1 cos ( 3.199 m sin ( 3.44 For the initial conditions.0981m --s C 1 sin ( 9.0981m --s d---.x ( t ) = – 0.1 m t – 0.= – 0.1 m t – 0.= – 0.16 )m cos ( 3.41 Analysis of the object after motion begins nŒI .0981m --s 0 = C 1 sin ( 3.209 C 2 = ( – 7.29988 + C 2 ) = – 0.943 ) + C 2 ) + 0.1 m --dt s Figure 3.889rad ) + 0.889 ) x ( t ) = – 0.199m sin ( 9.x ( t ) = 3.29988 + C 2 ) – 0.1962 d---.889rad ) + 0.16t + C 2 ) + 0.3.x ( 2. x ( 2.943s ) = 0m d---.16 ( 2.29988 + C 2 ) -----------------------------------------------------.16C 1 cos ( 3.

648 m --dt s After this the differential equation without the cable/spring is used.981 ---2 100kg s · m x 2 =  – 0.913 --.199 m sin ( 3. considering when to switch the analysis conditions. t + 3.913 --. This point in time can be determined when the displacement of the block equals the displacement of the cable/spring end.16t – 7.981 --.328s ) + 3.16t – 7.=  – 0. m m 0 =  – 0.648 --.913 --.x 2 ( t ) =  – -----------.1N m x 2'' = – -------------.981 m 2 .328 ) = 0.889rad ) = 0.42 Determining when the cable become slack .0981m t = 3.981 ---- ( 3.45 The equation of motion changes after the cable becomes slack.913 m s m 0.328s ) + C 1  2 s s --C 1 = 3.t + C2  s 2 s 2 2 0.981 m m 0.differential equations .51549413 x ( 3.--.= – 0.889 + π n = – 0.0981m s s – 0.453m 0. t + 3.t – 7.137m d---. ----0.x ( 2.981 ---- t + C 1  2 s m m 0.--.1 m t = – 0.( 3. t + 3.989s The solution can continue.333 ) = 0.328s 3.453m  2 2 s s This motion continues until the block stops moving.889rad ) + 0.x 2 =  – -----------.3.1 m t – 0.199 m sin ( 3.913 -- 2 s s t = 3.137m =  – -----------.16t – 7.328s ) + C 2  2 2 s s C 2 = – 7.---- ( 3. 98.981 m 2 m . Figure 3.

In addition. In total there will be four mounts mounted around the machine.43.0004Kg K d y + 0. but as usual the best place to begin is by developing a free body diagram.44. Each isolator will be composed of a spring and a damper.+ -----------. F M y The model to the left describes a piece of reciprocating industrial equipment.43 A vibration control system There are a number of elements to the design and analysis of this system. The equipment operates such that a force of 1000N with a frequency of 2Hz is exerted on the mass.81ms . The mass of the equipment is 10000kg.81ms 10000Kg 10000Kg 10000Kg Mg –1 · –1 –2 –2 ·· y + 0.0004Kg K s y = 0.46 3.= --------------------. This is done in Figure 3. and a differential equation. We have been asked to design a vibration isolation mounting system.differential equations . the oscillations while the machine is running cannot be more than 2cm total.+ g y M M M · 4K d y 4K s y –2 1000N ·· y + --------------------.1ms sin ( 4 π t ) + 9. and 25cm when loaded with the mass.3. The criteria we are given is that the mounts should be 30cm high when unloaded. Figure 3.+ --------------------. F + ∑ Fy · ·· = F – 4K s y – 4K d y + Mg = My M y · · My' + 4K d y + 4K s y = F + Mg · 4K d y 4K s y · F 4K s y 4K d y ·· + -----------.= ---.sin ( 2 ( 2 π )t ) + 9.6 CASE STUDY A typical vibration control system design is described in Figure 3.

47 Figure 3. When the system is at rest the equation is simplified.differential equations . 0.05m. in the positive y direction.3. . This will result in a downward compression of 0.45.05 ) –1 –2 ∴K s = 491KNm Figure 3. This simplifies the differential equation by eliminating several terms.0004 ( 0. we will assume that the cyclic force is not applied for the unloaded/loaded case.05m ) = 9.46.44 FBD and derivation of equation Using the differential equation.25m long.0004Kg K s y = 9.0004Kg K s ( 0. At this point we have determined the range of motion of the mass.30m long. the acceleration and velocity terms both become zero. This is done in Figure 3. and after loading the spring is 0.45 –1 Calculation of the spring coefficient The remaining unknown is the damping coefficient. 0. In addition.Kgms m 0.81ms –2 –1 9. the spring values can be found by assuming the machine is at rest. This can be done by developing the particular solution of the differential equation. as it will contain the steady-state oscillations caused by the forces as shown in Figure 3.81ms –1 –2 Now we can consider that when unloaded the spring is 0.81 K s = ------------------------------.

24 ) ( – 5.81ms The particular solution can now be found by guessing a value. (Note: The units in the expression are uniform (i.46 Particular solution of the differential equation The particular solution can be used to find a damping coefficient that will give an overall oscillation of 0.24 ) ( –5.81 – 16 π B + 0.1ms sin ( 4 π t ) + 9. and finding the roots of the resulting polynomial.02m.124 B = -------------------------------------------------------------------------------------------------------2 –9 –3 K d ( – 159 × 10 ) + K d ( – 6.1 0.0004Kg ( 491KNm )y = 0.18 × 10 K d – 0.47.0 × 10 K d ) = 0.1 A = -------------------------------------------------------------------------------------------------------------------------------------2 –6 –3 – 16 π + 196 + ( 31.81ms –1 –2 –2 –2 · ·· y + 0.1 A = -------------------------------------------------------------------------------------------------------2 –9 –3 K d ( – 159 × 10 ) + K d ( – 6.1 A ( –16 π + 196 ) + A ( 31. .81ms –2 –6 2 –6 –3 2 –3 2 2 2 Figure 3. In this case Mathcad was used to find the solution.0 × 10 K d ) = 0.1ms sin ( 4 π t ) + 9.48 –1 –1 –1 –2 –2 ·· · y + 0.. as shown in Figure 3.1 A ( – 16 π + 196 ) + B ( – 5.0004K d ( 4 π A cos ( 4 π t ) – 4 π B sin ( 4 π t ) ) + 196 ( A sin ( 4 π t ) + B cos ( 4 π t ) + C ) = 0.0 × 10 K d ) 0.1 sin ( 4 π t ) + 9.0004Kg K d y + 196s y = 0.0004Kg K d y + 0.2 × 10 ) + 38. and solving for the coefficients.0004K d 4 π A + 196A = 0 –6 B = A ( 31.24 ) 2 – 16 π A + 0.e.) y = A sin ( 4 π t ) + B cos ( 4 π t ) + C y' = 4 π A cos ( 4 π t ) – 4 π B sin ( 4 π t ) y'' = – 16 π A sin ( 4 π t ) – 16 π B cos ( 4 π t ) 2 2 ∴( – 16 π A sin ( 4 π t ) – 16 π B cos ( 4 π t ) ) + 0.differential equations .8 × 10 K d + 1. although it could have also been found by factoring out the algebra.3.1 3.8 × 10 K d + 1.1 C = 9.0004K d ( – 4 π B ) + 196A = 0.2 × 10 ) + 38. the same in each term) and will be omitted for brevity.8 × 10 K d + 1.

2 2 A +B 2 2 –6 ( 0.2 ⋅ 10 ) ) + 38.differential equations .18 ⋅ 10 )K d – 0. or requesting custom designs from other companies. Components can also be acquired by searching catalogs. The magnitude of oscillation can be calculated with the Pythagorean formula.3.2 ⋅ 10 ) ) + 38.47 Determining the damping coefficient The values of the spring and damping coefficients can be used to select actual components.01 = -------------------------------------------------------------------------------------------------------------2 –9 –3 K d ( – 159 ⋅ 10 ) + K d ( – ( 6.1 The design requirements call for a maximum oscillation of 0. or a magnitude of 0.1 ) +  ( 3. 2 2 –6 ( 0.1 ) +  ( 3.124   0.124   magnitude = -------------------------------------------------------------------------------------------------------------2 –9 –3 K d ( – 159 ⋅ 10 ) + K d ( – ( 6. .18 ⋅ 10 )K d – 0.49 In the previous particular solution the values were split into cosine and sine components. magnitude = sK d = 3411N --m Aside: the Mathcad solution Figure 3. Some companies will design and build their own components.02m.1 A given-find block was used in Mathcad to obtain a damper value of.01m.

+ N K s = 100 --m M = 1Kg y ∑ Fy ·· = K s y – Mg = – My N N··  100 --. • A case study looked at a second-order system. Solve the following differential equation with the three given cases. is applied.7 SUMMARY • First and second-order differential equations were analyzed explicitly. Write the differential equation for the mass. The following differential equation was derived for a mass suspended with a spring.81N ------. The mass. All of the systems have a . when a constant force. It then oscillates. • Non-linear systems can be analyzed by making them linear. y – ( 1Kg )  9. M. Leave the results in variable form.3. but the motion is opposed by viscous friction (damping) with the coefficient B. Solve the differential equation to find the motion as a function of time.81ms –1 · y 0 = 0ms y = 0m 0 FBD: M Ksy Mg 3.··  2  m s Kgm Kgm ·· ( 1Kg )y +  100 ----------- y = 9. x M B F 2. • The topic of analysis was discussed.50 3. • First and second-order responses were examined.differential equations . It can slide across a surface. = ( – 1Kg )y   m Kg N  1 Nm y +  100 --.81 -----. At time 0s the system is released and allowed to drop.81 ---------- 2 2 ms s –2 –2 ·· y + ( 100s )y = 9. illustrated below starts at rest. 3.8 PRACTICE PROBLEMS 1. and solve the differential equation. y = 9. F. Initially the system starts at rest.

Solve the equations for x assuming that . The input is a step function that turns on at t=0.5 ξ = 1 ξ = 2 ω n = 10 ω n = 10 ω n = 10 · x = 0 x = 0 y = 1 4. ·· · initial conditions V i = 5 0. The input is a step function that turns on at t=0. 2 ·· · x + 2ξω n x + ω n x = y initial conditions case 1: case 2: case 3: ξ = 0.6Vo + 2. Solve the following differential equation with the given initial conditions and draw a sketch of the first 5 seconds.3. a) Write the differential equations for the system below.6Vo + 2.1V o = 3Vi + 2 Vo = 0 · Vo = 0 5.5Vo + 0. Solve the following differential equation with the given initial conditions and draw a sketch of the first 5 seconds. ·· · 0.1V o = 3Vi + 2 initial conditions V i = 5 Vo = 0 · Vo = 1 6.51 step input ’y’ and start undeflected and at rest.5Vo + 0.differential equations .

If it has a spring constant of 20.5 ω n = 10 case 1: case 2: case 3: ξ = 1 ξ = 2 ω n = 10 ω n = 10 8. A spring damper system supports a mass of 34N.6N/cm. If the system if first-order find the time constant. Using a standard lumped parameter model the weight is 36N.differential equations . What are the natural frequency (Hz) and damping ratio? .52 the system is at rest and undeflected before t=0. All of the systems have a sinusoidal input ’y’ and start undeflected and at rest.06*103 N/m and damping is 100Ns/m. · 2 ·· · initial conditions x = 0 x + 2ξω n x + ω n x = y x = 0 y = sin ( t ) ξ = 0. Kd1 = 1 Ns/m Ks1 = 1 N/m Kd = 1 Ns/m x1 Kd2 = 1Ns/m Ks2 = 1 N/m F=1N M = 1 kg x x2 F=1N Ks = 1 N/m b) State whether each system is first or second-order. If it is second-order find the natural frequency and damping ratio. Also assume that gravity is present. Solve the following differential equation with the three given cases. 7. what is the systems natural frequency? 9. stiffness is 2.3.

Solve the differential equation below using homogeneous and particular solutions.) 14. Determine the first order differential equation given the graphical response shown below.differential equations . (Hint: Find the response as a function of time. Calculate the undamped natural frequency (Hz) of the system.53 10. 16.4 seconds to the first peak? 11. Assume an initial displacement of 50mm. ··· ·· · θ + 40θ + 20θ + 2θ = 4 13. A system is to be approximated with a mass-spring-damper model using the following parameters: weight 28N. the damping ratio and describe the type of response you would expect if the mass were displaced and released.3. What additional damping would be required to make the system critically damped? ·· · Mx + K d x + K s x = F 12. Assume the system starts undeflected and at rest. What would the displacement amplitude after 100ms for a system having a natural frequency of 13 rads/sec and a damping ratio of 0. What is the differential equation for a second-order system that responds to a step input with an overshoot of 20%. Assume the input is a step function. viscous damping 6Ns/m. and a steady state displacement of 0mm. Explain with graphs how to develop first and second-order equations using experimental data. x 4 t(s) 0 1 2 3 4 15. and stiffness 36N/m.20. Determine the parameters of . The second order response below was obtained experimentally. with a delay of 0.

x (m) 5m 0m 0. Develop equations (function of time) for the first and second order responses shown below.6s 0.8s t (s) -10m .2s 0.54 the differential equation that resulted in the response assuming the input was a step function.3.7s 0.3s 0.5s 0.4s 0. 0.1s x (m) 5m 0m 1s 3s t (s) 0.differential equations .5s 1s 2 10 0 t(s) 17.

differential equations - 3.55

3.9 PRACTICE PROBLEM SOLUTIONS
1. x ( t ) = –FM e ----------2 B 2. homogeneous: guess yh = e
2 At At –2 B – ---- t M

FM – F + ---------t 2 B B

· At y h = Ae
At

·· 2 At yh = A e

A e + ( 100s )e
2 –2

= 0
–1

A = ± 10 js A = – 100s y h = C 1 cos ( 10t + C 2 ) particular: guess y p = At + B
–2

· yp = A

·· yp = 0
–2

( 0 ) + ( 100s ) ( At + B ) = 9.81ms –2 –2 ( 100s ) ( At + B ) = 9.81ms –2 9.81ms A = 0 B = ---------------------- = 0.0981m –2 100s y p = 0.0981m initial conditions: y = yh + y p = C 1 cos ( 10t + C 2 ) + 0.0981m y' = – 10C 1 sin ( 10t + C 2 ) for d/dt y0 = 0m: 0 = – 10C 1 sin ( 10 ( 0 ) + C 2 ) for y0 = 0m: – 0.0981m = C 1 cos ( 0 ) y ( t ) = ( – 0.0981m ) cos ( 10t ) + 0.0981m C2 = 0

0 = C 1 cos ( 10 ( 0 ) + ( 0 ) ) + 0.0981m C 1 = – 0.0981m

differential equations - 3.56

3. case 1: case 2: case 3: x ( t ) = – 0.0115e x ( t ) = – 0.010e
– 5t

cos ( 8.66t – 0.524 ) + 0.010 – 0.10te
– 10t

– 10t

+ 0.010
– 2.679t

x ( t ) = 775 ⋅ 10 e

– 6 – 37.32t

– 0.0108e

+ 0.010

4. Vo ( t ) = –8.465e 5. V0 ( t ) = –8.331e 6. a) x1 ( t ) = e – 1 x 2 ( t ) = 2e – 2 τ = 1 7. case 1: case 2: case 3: 8. 24.37 rad/sec 9. fn=3.77Hz, damp.=.575 10. ·· · x + 8.048x + 77.88x = F ( t ) x ( t ) = – 0.00117e
– 5t –t –t – 0.6t – 0.6t

sin ( 1.960t + 1.274 ) + 8.095

cos ( 1.96t – 0.238 ) + 8.095

b) x ( t ) = – 12.485e ζ = 0.5
– 0.5t

cos ( 0.866t – 0.524 ) + 10.81

ωn = 1

sin ( 8.66t – 1.061 ) + 0.0101 sin ( t – 0.101 )
– 10t

x ( t ) = ( 1.96 ⋅ 10 )e x ( t ) = ( 3.5 ⋅ 10 )e
–3

–3

+ ( 9.9 ⋅ 10 )te – ( 18 ⋅ 10 )e
–6

–3

– 10t

+ ( 9.9 ⋅ 10 ) sin ( t + 0.20 ) + ( 9.4 ⋅ 10 ) sin ( t + 0.382 )
–3

–3

– 2.679t

– 37.32t

differential equations - 3.57

11. Given 28N M = ---------------- = 2.85kg N9.81 ----kg The typical transfer function for a mass-spring-damper systems is, sK d = 6N --m N K s = 36 --m Ks F ·· · K d x + x  -----  + x  ----- = --- M  M M The second order parameters can be calculated from this. 2 ·· · x + x ( 2ζω ) + x ( ω ) = y ( t )
n n

ωn =

Ks ----- = M

N 36 --m ---------------- = 2.85kg

kgm 36 --------2 ms - = ---------------2.85kg

12.63s

–2

rad = 3.55 -------- = 0.6Hz s

s K d ----6N -- M m ξ = ----------- = -------------------------------------------- = 0.296 2ω n rad 2 ( 3.55 ) -------- 2.85kg s 2 ω d = ω n 1 – ξ = 3.39 rad -------s If pulled and released the system would have a decaying oscillation about 0.52Hz A critically damped system would require a damping coefficient of....  K d ---- M Kd ξ = ----------- = -------------------------------------------- = 1 2ω n 2 ( 3.55 ) rad 2.85kg -------s Ns K d = 20.2 ----m

12. θ ( t ) = – 66 ⋅ 10 e
– 6 – 39.50t

– 3.216e

0.1383t

+ 1.216e

– 0.3368t

+ 2.00

13. y ( t ) = 0.05e
– 2.6t

cos ( 12.74t ) = 0.05e

– 2.6t

π sin  12.74t + --  2

differential equations - 3.58

14. x 4

t(s) 1 τ = 1 Given the equation form, 0 2 3 4

· -x + 1x = A τ The values at steady state will be · x = 0 x = 4 So the unknown ‘A’ can be calculated. 1 0 + -- 4 = A 1 · -x + 1x = 4 1 · x+x = 4 15. Key points: First-order: find initial final values find time constant with 63% or by slope use these in standard equation Second-order: find damped frequency from graph find time to first peak use these in cosine equation A = 4

differential equations - 3.59

16. For the first peak: – σt b----- = e p ∆x
– σ0.5 2----- = e 10

π ω d ≈ --tp

2 ln  ----- = – σ0.5  10 2σ = – 2 ln  ----- = 3.219  10 For the damped frequency: 2π ω d = ----- = 2π 1s 3.219 ω n = -----------ξ
2

1 ξ = ----------------------------π ------- 2 + 1  t σ p

These values can be used to find the damping coefficient and natural frequency σ = ξω n

ωd = ωn 1 – ξ

2 3.219 2π = ------------ 1 – ξ ξ 2 2 2π  ------------  = 1 – ξ ------------- 2  3.219 ξ

12π - 2  ------------  + 1 = ---2  3.219 ξ

ξ =

1 ------------------------------ = 0.4560 2π - 2  ------------  + 1  3.219

3.219 3.219ω n = ------------ = --------------- = 7.059 ξ 0.4560 This leads to the final equation using the steady state value of 10
2 ·· · x + 2ξω n x + ω n x = F 2 ·· · x + 2 ( 0.4560 ) ( 7.059 )x + ( 7.059 ) x = F ·· · x + 6.438x + 49.83x = F ( 0 ) + 6.438 ( 0 ) + 49.83 ( 10 ) = F ·· · x + 6.438x + 49.83x = 498.3

F = 498.3

differential equations - 3.60

17.
---------   0.18 x ( t ) = 5 1 – e    –t

x ( t ) = – 10e

– 0.693 t

cos ( πt )

3.10 ASSIGNMENT PROBLEMS

numerical methods - 4.1

4. NUMERICAL ANALYSIS

Topics: • State variable form for differential equations • Numerical integration with software and calculators • Numerical integration theory: first-order, Taylor series and Runge-Kutta • Using tabular data • A design case Objectives: • To be able to solve systems of differential equations using numerical methods.

4.1 INTRODUCTION
For engineering analysis it is always preferable to develop explicit equations that include symbols, but this is not always practical. In cases where the equations are too costly to develop, numerical methods can be used. As their name suggests, numerical methods use numerical calculations (i.e., numbers not symbols) to develop a unique solution to a differential equation. The solution is often in the form of a set of numbers, or a graph. This can then be used to analyze a particular design case. The solution is often returned quickly so that trial and error design techniques may be used. But, without a symbolic equation the system can be harder to understand and manipulate. This chapter focuses on techniques that can be used for numerically integrating systems of differential equations.

4.2 THE GENERAL METHOD
The general process of analyzing systems of differential equations involves first putting the equations into standard form, and then integrating these with one of a number of techniques. State variable equations are the most common standard equation form. In this form all of the equations are reduced to first-order differential equations. These firstorder equations are then easily integrated to provide a solution for the system of equations.

numerical methods - 4.2

4.2.1 State Variable Form
At any time a system can be said to have a state. Consider a car for example, the state of the car is described by its position and velocity. Factors that are useful when identifying state variables are: • The variables should describe energy storing elements (potential or kinetic). • The variables must be independent. • They should fully describe the system elements. After the state variables of a system have been identified, they can be used to write first-order state variable equations. The general form of state variable equations is shown in Figure 4.1. Notice that the state variable equation is linear, and the value of x is used to calculate the derivative. The output equation is not always required, but it can be used to calculate new output values.

d ---- x = Ax + Bu  dt y = Cx + Du

state variable equation output equation

where, x = state/output vector (variables such as position) u = input vector (variables such as input forces) A = transition matrix relating outputs/states B = matrix relating inputs to outputs/states y = non-state value that can be found directly (i.e. no integration) C = transition matrix relating outputs/states D = matrix relating inputs to outputs/states

Figure 4.1

The general state variable form

An example of a state variable equation is shown in Figure 4.2. As always, the FBD is used to develop the differential equation. The resulting differential equation is second-order, but this must be reduced to first-order. Using the velocity variable, ’v’ the second-order differential equation can be reduced to a first-order equation. An equation is also required to define the velocity as the first derivative of the position, ’x’. In the example the two state equations are manipulated into a matrix form. This form can be useful, and may be required for determining a solution. For example, HP calculators require the matrix form, while TI calculators use the equation forms. Software such as Mathcad can use either form. The main disadvantage of the matrix form is that it will only work for lin-

numerical methods - 4.3

ear differential equations.

Given the FBD shown below, the differential equation for the system is, F x ·· Mx M · Kd x Ks x +

∑ Fx

· ·· = – F – K d x – K s x = Mx

· ·· – F – K d x – K s x = Mx

The equation is second-order, so two state variables will be needed. One obvious choice for a state variable in this equation is ’x’. The other choice can be the velocity, ’v’. Equation (1) defines the velocity variable. The velocity variable can then be substituted into the differential equation for the system to reduce it to first-order. · x = v (1) ·· · Mx = – F – K d x – K s x · Mv = – F – K d v – K s x –Ks –Kd –F · v = x  --------  + v  ---------  +  ------  M  M   M

(2)

Equations (1) and (2) can also be put into a matrix form similar to that given in Figure 4.1. 0 1 0 d- x = x + ---–Ks –K d –F dt v -------------- --------- v M M M Note: To have a set of differential equations that is solvable, there must be the same number of state equations as variables. If there are too few equations, then an additional equation must be developed using an unexploited relationship. If there are too many equations, a redundancy or over constraint must be eliminated.

Figure 4.2

A state variable equation example

numerical methods - 4.4

Put the equation into state variable and matrix form.

∑F

d= F = M  ---- x  dt

2

· ans. x = v

F · v = ---M

0 d---- x = 0 1 x + F dt v ---0 0 v M

Figure 4.3

Drill problem: Put the equation in state variable form

Consider the two cart problem in Figure 4.4. The carts are separated from each other and the wall by springs, and a force is applied to the left hand side. Free body diagrams are developed for each of the carts, and differential equations developed. For each cart a velocity state variable is created. The equations are then manipulated to convert the second-order differential equations to first-order state equations. The four resulting equations are then put into the state variable matrix form.

numerical methods - 4.5

x1 F M1 K s1

x2 K s2 M2

F M1

K s1 ( x 1 – x 2 ) ·· M1 x 1

+

∑ Fx

·· = F – K s1 ( x 1 – x 2 ) = M1 x 1 (1)

·· M1 x 1 + K s1 x 1 – K s1 x 2 = F · x1 = v1 · M 1 v 1 + K s1 x 1 – K s1 x 2 = F K s1 · F - K s1 v 1 = ------- – ---------x1 + ---------x 2 M 1 M1 M1 +

(2)

K s1 ( x 1 – x 2 )

M2

K s2 ( x 2 ) ·· M2 x 2

∑ Fx

·· = K s1 ( x 1 – x 2 ) – K s2 ( x 2 ) = M 2 x 2

·· M2 x 2 + ( K s1 + K s2 )x 2 – K s1 x 1 = 0 · x2 = v2 (3) · M2 v 2 + ( K s1 + K s2 )x 2 – K s1 x 1 = 0 K s1  K s1 + K s2 · v 2 = ---------x 1 –  -------------------------- x 2 (4) M2 M2  

The state equations can now be combined in a matrix form. 0 1 0 0 x1 x1 0 – K s1 K s1 ------------ 0 --------0 FM1 v1 ------d- v 1 = M 1 ---+ M1 dt x 0 0 0 1 x2 2 0 K s1 – K s1 – K s2 v2 v2 0 --------- 0 ---------------------------- 0 M2 M2

Figure 4.4

Two cart state equation example

numerical methods - 4.6

x Kd M Ks

ans. · x = · v =

v Kd Ks v  -----  + x  -----  M  M

Figure 4.5

Drill problem: Develop the state equations in matrix form

numerical methods - 4.7

x1 K d1 M1 K s1

x2

F

K s2 M2

ans. x· = v 1 1 · x2 = v2 –K d1 – K s1 K d1 K s1 · Fv 1 = v 1  ----------- + x 1  ----------- + v 2  -------- + x 2  -------  + ----- M   M  M   M  M1 1 1 1 1 –K d1 – K s1 – K s2 K d1 K s1 · v 2 = v 2  ----------- + x 2  --------------------------  + v 1  -------- + x 1  -------   M    M  M  M2 2 2 2

Figure 4.6

Drill problem: Convert the system to state equations

numerical methods - 4.8

In some cases we will develop differential equations that cannot be directly reduced because they have more than one term at the highest order. For example, if a second-order differential equation has two second derivatives it cannot be converted to a state equation in the normal manner. In this case the two high order derivatives can be replaced with a dummy variable. In mechanical systems this often happens when masses are neglected. Consider the example problem in Figure 4.7, both ’y’ and ’u’ are first derivatives. To solve this problem, the highest order terms (’y’ and ’u’) are moved to the left of the equation. A dummy variable, ’q’, is then created to replace these two variables with a single variable. This also creates an output equation as shown in Figure 4.1.

Given the equation, · · 3y + 2y = 5u Step 1: put both the first-order derivatives on the left hand side, · · 3y – 5u = – 2y Step 2: replace the left hand side with a dummy variable, q = 3y – 5u · q = – 2y

Step 3: solve the equation using the dummy variable, then solve for y as an output eqn. q + 5u · q = – 2y y = -------------3

Figure 4.7

Using dummy variables for multiple high order terms

At other times it is possible to eliminate redundant terms through algebraic manipulation, as shown in Figure 4.8. In this case the force on both sides of the damper is the same, so it is substituted into the equation for the cart. But, the effects on the damper must also be integrated, so a dummy variable is created for the integration. An output equation was created to calculate the value for x1.

numerical methods - 4.9

x1 F Kd

x2 M

The FBDs and equations are; F · · Kd ( x1 – x2 ) +

∑ Fx

· · = F – Kd ( x1 – x2 ) = 0 (1)

· · Kd ( x1 – x2 ) = F q = x 1 – x2 Kd ( q ) = F F· q = -----Kd x1 = x2 + q M + · · ·· = K d ( x 1 – x 2 ) = Mx 2 ·· F = Mx 2 · x 2 = v2

(2) (3)

· · Kd ( x1 – x2 ) ·· Mx 2

∑ Fx

(4) (5)

· F v 2 = ---M

The state equations (2, 4, 5) can be put in matrix form. The output equation (2) can also be put in matrix form. F-----q q Kd 0 0 0 d- x = x2 + ---- 2 0 0 1 0 dt v2 0 0 0 v2 F ---M

q x1 = 1 1 0 x2 + 0 v2

Figure 4.8

A dummy variable example

numerical methods - 4.10

4.3 NUMERICAL INTEGRATION
Repetitive calculations can be used to develop an approximate solution to a set of differential equations. Starting from given initial conditions, the equation is solved with small time steps. Smaller time steps result in a higher level of accuracy, while larger time steps give a faster solution.

4.3.1 Numerical Integration With Tools
Numerical solutions can be developed with hand calculations, but this is a very time consuming task. In this section we will explore some common tools for solving state variable equations. The analysis process follows the basic steps listed below. 1. Generate the differential equations to model the process. 2. Select the state variables. 3. Rearrange the equations to state variable form. 4. Add additional equations as required. 5. Enter the equations into a computer or calculator and solve. An example in Figure 4.9 shows the first four steps for a mass-spring-damper combination. The FBD is used to develop the differential equations for the system. The state variables are then selected, in this case the position, y, and velocity, v, of the block. The equations are then rearranged into state equations. The state equations are also put into matrix form, although this is not always necessary. At this point the equations are ready for solution.

numerical methods - 4.11

Step 1: Develop equations KS Kd y M F Step 2: We need to identify state variables. In this case the height is clearly a defining variable. We will also need to use the vertical velocity, because the acceleration is a second derivative (we can only have first derivatives). Using the height, y, and velocity, v, as state variables we may now proceed to rewriting the equations. (Note: this is just an algebraic trick, but essential when setting up these matrices.) Step 3: d ---- y = v  dt – F – Kd v – Ks y d ---- v = -------------------------------------- dt M ·· My F Ksy M · Kd y

dd- 2 F y = – F – K d  ---- y – K s y = M  ---- y ∑  dt  dt dd- 2 F + K d  ---- y + K s y + M  ---- y = 0  dt  dt

Step 4: We put the equations into a state variable matrix form. d ---- y  dt 0 1 F = –K – K y + 0 ---S d –1 M --------- --------- v d- v  ---M M  dt

Figure 4.9

Dynamic system example

Figure 4.10 shows the method for solving state equations on a TI-86 graphing calculator. (Note: this also works on other TI-8x calculators with minor modifications.) In the example a sinusoidal input force, F, is used to make the solution more interesting. The next step is to put the equation in the form expected by the calculator. When solving with the TI calculator the state variables must be replaced with the predefined names Q1, Q2, etc. The steps that follow describe the button sequences required to enter and analyze the equations. The result is a graph that shows the solution of the equation. Points can then be taken from the graph using the cursors. (Note: large solutions can sometimes take a few minutes to solve.)

numerical methods - 4.12

First, we select some parameter values for the equations of Figure 4.9. The input force will be a decaying sine wave. d ---- y = v y  dt d ---- v = – F – K d v y – K s y = – 4e –0.5t sin ( t ) – 2v – 5y ---------------------------------------y  dt y M Next, the calculator requires that the state variables be Q1, Q2, ..., Q9, so we replace y with Q1 and v with Q2. Q1' = Q2 Q2' = – 4e
– 0.5t

sin ( t ) – 2Q2 – 5Q1

Now, we enter the equations into the calculator and solve. To do this roughly follow the steps below. Look at the calculator manual for additional details. 1. Put the calculator in differential equation mode [2nd][MODE][DifEq][ENTER] 2. Go to graph mode and enter the equations above [GRAPH][F1] 3. Set up the axis for the graph [GRAPH][F2] so that time and the xaxis is from 0 to 10 with a time step of 0.5, and the y height is from +3 to -3. 4. Enter the initial conditions for the system [GRAPH][F3] as Q1=0, Q2=0 5. Set the axis [GRAPH][F4] as x=t and y=Q 6. (TI-86 only) Set up the format [GRAPH][MORE][F1][FldOff][ENTER] 7. Draw the graph [GRAPH][F5] 8. Find points on the graph [GRAPH][MORE][F4]. Move the left/right cursor to move along the trace, use the up/down cursor to move between traces. Figure 4.10 Solving state equations with a TI-85 calculator

This time is chosen because of the general . where ’t’ is the time and ’Q’ is the state variable vector.. Q9. the calculator requires that the state variables be Q1. in sequence are. but it is not the matrix form.. The first step is to enter the state equations as a function. The input force will be a decaying sine wave. 1.= – 4e – 0.13 First. The arguments for the function. we enter the equations into the calculator and solve.4. the number of steps. The ’rkfixed’ function is then used to obtain a solution.5. and the state equation function. so we replace y with Q1 and v with Q2. Q1' = Q2 Q2' = – 4e – 0. Move the left/right cursor to move along the trace. (TI-86 only) Set up the format [GRAPH][MORE][F1][FldOff][ENTER] 7. In this case the 10 second time interval is divided into 100 parts each 0.12. as shown in Figure 4. Draw the graph [GRAPH][F5] 8. Look at the calculator manual for additional details. Q2=0 5. and the y height is from +3 to -3. 4.5t sin ( t ) – 2v – 5y  dt y y M Next. Go to graph mode and enter the equations above [GRAPH][F1] 3.11 Solving state equations with a TI-89 calculator State equations can also be solved in Mathcad using built-in functions. Enter the initial conditions for the system [GRAPH][F3] as Q1=0. Set the axis [GRAPH][F4] as x=t and y=Q 6.1s in duration.9. use the up/down cursor to move between traces. UPDATE FOR TI-89 Figure 4. the end time.5t sin ( t ) – 2Q2 – 5Q1 Now. Q)’. Q2. d ---- y = v  dt y – F – K d vy – Ks y d ---- v = -----------------------------------------.) The state variables in the vector ’Q’ replace the original state variables in the equations. Set up the axis for the graph [GRAPH][F2] so that time and the xaxis is from 0 to 10 with a time step of 0. .numerical methods . the state vector.. (Note: the equations are in a vector. Put the calculator in differential equation mode [2nd][MODE][DifEq][ENTER] 2. the start time. Find points on the graph [GRAPH][MORE][F4]. we select some parameter values for the equations of Figure 4. ’D(t. To do this roughly follow the steps below.

run the calculator again using a smaller time step.14 response time for the system. If the time step is too large the solution may become unstable and go to infinity. A time step that is too small will increase the computation time marginally.4.numerical methods . Figure 4. . When in doubt. while in Mathcad the arrays start at Q0.12 Solving state variable equations with Mathcad Note: Notice that for the TI calculators the variables start at Q1. Many students encounter problems because they forget this.

y(t + h) y(t) d---. were smaller. Basically. Notice that the function being integrated curves downward.13. This is then used to calculate values iteratively in the table. h. . If the time step. creating an error between the actual and estimated values at time ’t+h’.14 shows the solution of Newton’s equation using Euler’s method.y ( t ) dt Note: here the h value is the time step between integrations points. The acceleration is put directly into Euler’s equation. Given the current value of a function and the first derivative.3. Notice that the values start before zero so that initial conditions can be used.) The equation shown is known as Euler’s equation. later. the error would decrease. In this example we are determining velocity by integrating the acceleration caused by a force.13 t+h First-order numerical integration The example in Figure 4. (Note: Recall that the state equations allow us to calculate first-order derivatives. t Figure 4. we can estimate the function value a short time later.2 Numerical Integration The simplest form of numerical integration is Euler’s first-order method. h. as shown in Figure 4. A smaller time step will increase the accuracy. using a known position and first derivative we can calculate an approximate value a short time.y ( t ) dt dy ( t + h ) ≈ y ( t ) + h ---.4.15 4. If the system was second-order we would need two previous values for the calculations.numerical methods .

14 First order numerical integration example .16 Given the differential equation.4. We will assume that the system is initially at rest and that a force of 1N will be applied to the 1kg mass for 4 seconds.  M  first rearrange equation put this in the Euler equation finally substitute in known terms We can now use the equation to estimate the system response. A time step of 2 seconds will be used. i -1 0 1 2 3 4 5 6 7 8 t (sec) -2 0 2 4 6 8 10 12 14 16 F (N) d/dt vi 0 1 1 2 2 2 2 2 2 2 0 1 1 2 2 2 2 2 etc vi 0 0 2 4 8 12 16 20 etc Figure 4. After this time the force will rise to 2N. dF = M  ---- v  dt we can create difference equations using simple methods.numerical methods . dF  ---- v = --- dt M dv ( t + h ) = v ( t ) + h  ---- v ( t )  dt F(t) v ( t + h ) = v ( t ) + h  --------.

The integration is done with a separate function.15 Drill problem: Numerically integrate the differential equation An example of solving the previous example with a traditional programming language is shown in Figure 4.17 Use first-order integration to solve the differential equation from 0 to 10 seconds with time steps of 1 second. x ( t + h ) = x ( t ) + h ( – 0. In this example the results will be written to a text file ’out.16. .1s.1x ( t ) + 5 ) x ( 10 ) = Figure 4.txt’. The force value is varied over the time period with ’if’ statements.1x = 5 ans.numerical methods . · x + 0.4. The solution iteratively integrates from 0 to 10 seconds with time steps of 0.

double slope){ double v_new.0. v. M). F/M). v_new = v + h * slope. int main(){ double h = 0. double v.0. } Figure 4. if(t > 4. F. double h.txt". "%f. FILE *fp. } } fclose(fp). double).0) F = 2. if( ( fp = fopen("out. %f\n".0.0. M = 1. } double step(double v. t += h ){ if((t >= 0.4.16 Solving state variable equations with a C program .0. double. "w")) != NULL){ v = 0.numerical methods . %f. t. t. v = step(v.1. F.0.18 double step(double. fprintf(fp.0) && (t < 4.0)) F = 1. h. t < 10. return v_new. for( t = 0.

M). %f\n".1.0. t < 10. } fp.0) && (t < 4. . The state variable function is defined first. v = step(v. double). "%f. h. t += h ){ if((t >= 0. } } Figure 4.0. Finally the function is integrated with rectangular. F/M).writeStatus != fp. if(fp. fp. FileOut fp = new FileOut("out. public class Integrate extends Object public void main() { double h = 0. } fclose(fp). This is followed by a definition of the parameters to be used for the numerical integration.0)) F = 1.0) F = 2. double h.17 Solving state variable equations with a Java program The program below is for Scilab (a Matlab clone).4.19 double step(double.IO_EXCEPTION){ double v = 0.0.numerical methods .0. M = 1. F.txt"). return v_new. v_new = v + h * slope.printf(fp.0.close(). trapezoidal and Simpson’s rule forms. double slope){ double v_new. t. v. %f. double.0. for( double t = 0. } public double step(double v. if(t > 4. F.

by: H.numerical methods . force/mass]. Jack Sept.18 First order integration with Scilab . x0 = 8. 2002 // System component values mass = 10. 2). h = (t_end . Figure 4. // d/dt x = v.t_start) / steps. 16. v] function foo=f(state.20 // // // // // // // // // first_order.4.sce A first order integration of an accelerating mass To run this in Scilab use 'File' then 'Exec'.. v0 = 12. // initial conditions // define the state matrix function // the values returned are [x. d/dt v = F/M endfunction // Set the time length and step size for the integration steps = 100. v0]. t_end = 100. X=[x0. force = 100.t) foo = [ state(\$. t_start = 1.

.19 First order integration with Scilab (continued) 4.:) + h*f(X. v) = (%f.20 for approximating a function. ... velocity(v0. force/mass..2)). // // // // // // // // // // The results should be first order integration = (49710. X = [X .numerical methods . X(\$. %f)\n". The Taylor series equation shown in Figure 4. force/mass. This is relatively small. Note: increasing the number of steps increases the accuracy Figure 4. but shows a clear case of the innacuracy of the numerical solutions.4. when a highly curved function is encountered we can use a higher order integration equation. . t_end))..5 * a0 * t^2) + (v0 * t) + x0. endfunction printf("The value with integration is (x. end printf("The value at the end of first order integration is (x.21 // // Loop for integration // for i=1:steps. Notice that the first part of the equation is identical to Euler’s equation. X(\$. v0. endfunction function v=velocity(v0.. v0. position(x0. 1002) explicit = (51208.. v) = (%f.3 Taylor Series First-order integration works well with smooth functions. a0. But. t) x = (0. but the higher order terms add accuracy. . %f)\n". .1). 1012) The difference is 1498 for position and 10 for velocity.3. t) v = (a0 * t) + v0. . i*h)]. a0. X(\$. // // Explicit equation // function x=position(x0. t_end)..

x(t + h) = x(t ) + h(1 + e Thus x0 = 0 h = 0.  ---- x = 1 + e  dt d.h  ---- x ( t ) + ---.4 d.5 0.( 0. The resulting equation is then used to iteratively calculate values.h ( – e + 3t ) + ---.3 0.( 0.1 ( 2 ) + ---.2 d.g.1 2 3 11x ( 0 + 0.4 0.1 ) ( 1 ) = 2! 3! Figure 4.20 The Taylor series An example of the application of the Taylor series is shown in Figure 4.4. for t=0.21.3 d.3 1.h ( e + 6t 2! 3! x(t) 0 0 e.h  ---- x ( t ) + …  dt 2!  dt 3!  dt 4!  dt Figure 4.21 Integration using the Taylor series method .numerical methods .2  ---- x = – e –20t + 3t 2  dt d.1 0..h  ---- x ( t ) + ---.7 0.2 1.2 1.22 d1.1 ) = 0 + 0.1 t (s) 0 0.2 0. Given the differential equation.6 0.3 – t + t + x ) + ---.9 – 20t 3 –t 2 1. we must first determine the derivatives and substitute these into Taylor’s equation.4 x ( t + h ) = x ( t ) + h  ---- x ( t ) + ---. Given – 20t 3 · x–x = 1+e +t – 20t 3 d+t +x We can write.8 0.1 ) ( –1 ) + ---.3  ---- x = e – 20t + 6t  dt In the Taylor series this becomes.

thus providing higher accuracy without requiring other than first-order differential equations. However. Notice that the two central values in time are more heavily weighted. x = the state variables f = the differential function or (d/dt) x t = current point in time h = the time step to the next integration point Note: in this case the state equation function f(t.22 Fourth order Runge-Kutta integration .4 Runge-Kutta Integration First-order integration provides reasonable solutions to differential equations. Figure 4. x) includes the state variables. in simpler systems the state equations may not include time and it could be replaced with f(x). and time. The equations in Figure 4.22 are for fourth order Runge-Kutta integration. to obtain accuracy the Taylor method also requires higher order derivatives. t. That accuracy can be improved by using higher order derivatives to compensate for function curvature. For each time step the values ’F1’ to ’F4’ are calculated in sequence and then used in the final equation to find the next value. 4. x + F3 ) x ( t + h ) = x ( t ) + 1 ( F 1 + 2F2 + 2F3 + F4 ) -6 where. The function ’f()’ is the state equation or state equation vector. x + ------ - 2 2 F2 F3 = hf  t + h.3. F 1 = hf ( t. The Runge-Kutta technique uses first-order differential equations (such as state equations) to estimate the higher order derivatives. thus making is unsuitable for use with state variable equations. x + ------ - 2 2 F 4 = hf ( t + h. The ’F1’ to ’F4’ values are calculated at different time steps.4.numerical methods . The final summation equation has a remote similarity to the first order integration equation. and values from previous time steps are used to ’tweak’ the estimates of the later states.23 Recall that the state variable equations are first-order equations. x. x ) F1 F2 = hf  t + h. But.

x = v dt d---.202 + 0 0 2  = 0.7 -----  2     3 + 0.04 2.7     3 + 0. the four integrating factors are calculated.1  0 1 3 + 0 0 2  = 0.202 + 0.numerical methods . these are combined to get the final value after one time step.5432  3.108 5 3 1 2. y = 2 (assumed input) v0 = 1 x0 = 3 h = 0.108   1 + 2.1 -----2 + 0 0 2  = 0.185 + 2 0.3108  0 4 1 + 2.1 + 2 0. Finally.185 -----------2 F3 = 0.v = 3 + 4v + 5y dt d---.4.1  0 1 + 0 0 2  = 0.1  0 1 3 + 0.04 --------  2   F 4 = 0.3108  = 3. d---.1  0 1  0 4 5 3 1 2.24 An example of a Runge-Kutta integration calculation is shown in Figure 4.108 2. The solution begins by putting the state equations in matrix form and defining initial conditions.1974667 -6  1.7 vi + 1 1 2.x = 0 1 x + 0 0 y dt v 0 4 v 5 3 1 For the first time step.23.23 Runge-Kutta integration example .04   1 + 1.185 F 2 = 0.202  0 4 5 3 1 2. After this.5432 xi + 1 vi + 1 xi + 1 = xi + 1 ( F 1 + 2F2 + 2F3 + F4 ) -6 vi   = 3 + 1  0. The number of calculations for a single time step should make obvious the necessity of computers and calculators.1  1 + 0  = 0.0898667 Figure 4.1     F1 = 0.1  0 4 1  4 5 3 1 13  1.

24 Drill problem: Integrate the acceleration function .25 d. x(0) = 1 v(0) = 2 h = 0.numerical methods .2 F = M  ---- x  dt use.5 s F = 10 M=1 Figure 4.4.

26 Kd Leave the solution in variable form. –1 K d = 10Nsm –1 K s = 10Nm M = 1kg y M Ks y 0 = 1m · y 0 = 0m Figure 4. The ’main’ program loops through the time steps and writes the value to a file. The ’step’ function performs one timestep integration for a .25 Drill problem: Integrate to find the system response The program in Figure 4.numerical methods .28 is used to perform a Runge-Kutta integration of a mass-spring-damper system. and then change values to see how the system response changes.4.26 and Figure 4.

} } fclose(fp). double[]. derivative(double. } Figure 4.// set initial condition for x X[1] = 0. X[0].4.5f\n".// create state variable list X[0] = 0. add(double[]. double t.27 second order Runge-Kutta integration. t. double X[SIZE]. if(j == 0) fprintf(fp.txt". X). double. /* A program to do Runge Kutta integration of a mass spring damper system */ #include <stdio.0. double[]. step(double.// set initial condition for v if( ( fp = fopen("out. double h = 0. It uses the functions ’add’ and ’multiply’ to manipulate the state matrix.numerical methods . double[]). The ’derivative’ function updates the state matrix with the new derivative values. double[]. t < 50. j++.26 Runge-Kutta integration C program . X[1]).001.5f %9. SIZE Ks Kd Mass Force 2 /* the length of the state vector */ 1000 /* the spring coefficient */ 10000 /* the damping coefficient */ 10 /* the mass coefficient */ 100 /* the applied force */ int main(){ FILE *fp.0. int j = 0. " t(s) x v \n\n"). for( t = 0.h> void void void void #define #define #define #define #define multiply(double. double[]). double[]).5f %9. t += h ){ step(t. double[]). if(j >= 10) j = 0. "%9. h. "w")) != NULL){ fprintf(fp.

derivative(t+h/2. tmp). double V[]. multiply(1. add(tmp. dX. } /* A subroutine to add vectors to simplify other equations */ void add(double X1[]. tmp). multiply(2. tmp). derivative(t+h/2. multiply(h. tmp.0. /* Calculate F1 */ derivative(t. F1. dX). tmp. i < SIZE.0.27 Runge-Kutta integration C program (cont’d) A Scilab program is given in Figure 4. tmp). tmp. tmp. F3. F2[SIZE]. double dX[]){ dX[0] = X[1]. dX[1] = (-Ks/Mass)*X[0] + (-Kd/Mass)*X[1] + (Force/Mass). F3[SIZE].4. dX. i++) R[i] = X*V[i]. tmp). double X[]. tmp). /* Calculate F3 */ multiply(0. F2).28 /* First order integration done here (could be replaced with runge kutta)*/ void step(double t. /* Calculate F2 */ multiply(0. add(X. dX). dX. i < SIZE. multiply(h. add(X. F2.28 to perform a Runge Kutta integration. dX[SIZE].numerical methods . tmp). } Figure 4. tmp. tmp. double h.5. } /* State Equations Calculated Here */ void derivative(double t. tmp). double R[]){ for(int i = 0. F4). tmp. F1[SIZE]. tmp. double R[]){ for(int i = 0. double X2[]. F1). tmp. } /* A subroutine to multiply a vector by a scalar to simplify other equations*/ void multiply(double X. double X[]){ double tmp[SIZE]. X. tmp). /* Calculate F4 */ add(X. . dX). dX. F3.5. X. multiply(h.0.0/6. F4[SIZE]. X). multiply(h.0. i++) R[i] = X1[i] + X2[i]. dX). F3). add(F1. add(F4. derivative(t+h. tmp). /* Calculate the weighted sum */ add(F2.

:)). -5]. . // Graph the values plot2d(t.0*F2 + 2.:) + h/2. t = [t_start].sce A first order integration of an accelerating mass To run this in Scilab use 'File' then 'Exec'.:) + h]. 2).4 SYSTEM RESPONSE In most cases the result of numerical analysis is graphical or tabular.:) + h). X. t(\$. X(\$. t(\$. by: H. . 0. force/mass].grid lines for the graph xtitle('Time (s)'). t(\$.0). 15. Figure 4.0 + v0*(t_end .0].2)).:) + h/2.. printf("The position (using an equation) should be %f\n". v0]. v] function foo=f(state. // Loop for integration for i=1:steps. v) = (%f.numerical methods .0.1).:). // leg .29 // // // // runge_kutta.. h = (t_end . t_end = 100. 2003 // System component values mass = 10. end // print some results to compare printf("The value at the end of first order integration is (x. X = [X .0*F3 + F4)/6.. [-2.4.draw lines with marks // nax . t(\$. // initial conditions // define the state matrix function // the values returned are [x. t_start = 0. F3 = h * f(X(\$. t(\$. F2 = h * f(X(\$.0)..:) + (F1 + 2. X(\$. %f)\n". v0 = 12.:) + F2/2.t_start) / steps. leg="position@velocity")..t) foo = [ state(\$.t_start) + x0). Jack Sept. force = 100. d/dt v = F/M endfunction // Set the time length and step size for the integration steps = 1000.:) + F1/2. X=[x0.0. F1 = h * f(X(\$. x0 = 8.5*force/mass*(t_endt_start)^2. // d/dt x = v. X(\$. t = [t .28 Runge-Kutta integration Scilab program 4. F4 = h * f(X(\$.:) + F3.the legend titles // style . In both cases .

= ----------.– ---M M 0 v = -----------------------------.1 Steady-State Response The state equations can be used to determine the steady-state response of a system by setting the derivatives to zero.numerical methods . In addition to these methods there is a technique that can determine the steady-state response of the system. Given the state variable form: 0 1 0 dx + ---. and then solving the equations. and the equations are rearranged and solved with Cramer’s rule.29 0 0 0 Ks F .v M M M Set the derivatives to zero Solve for x and v 1 F Kd F  ---- – ---.4. Consider the example in Figure 4.  M Ks K d – ----.5 DIFFERENTIATION AND INTEGRATION OF EXPERIMENTAL . (Note: this can also be done without the matrix also.– -----.30 details such as time constants and damped frequencies can be obtained by the same methods used for experimental analysis.v M M M 0 1 0 0 = x + Ks Kd F ---0 – ----.= 0 Ks 0 1  ----.4.  M K s Kd – ----. The solution begins with a state variable matrix.– -----M M Finding the steady-state response 4.) The derivatives on the left hand side are set to zero.– ----- M M M Fx = -----------------------------.– -----.x = Ks Kd F dt v ---– ----.v M M M 0 1 0 x = K s Kd F – ---– ----. 4.= -----Ks KS 0 1  ----.– -----.– -----M M Figure 4.– ----.= ----------.29.

y(t) yi + 1 yi yi – 1 ti – 1 ti T ti T ti + 1 yi + yi – 1 y ( t i ) dt ≈  ------------------------ ( t i – t i – 1 ) = T ( y i + y i – 1 ) -∫t i – 1   2 2  y i – y i – 1  y i + 1 – y i d1 1 ---.y ( ti ) ≈  ----------------------- =  ------------------------ = -.( y i + 1 – y i ) – -. Ideally the time steps would be as small as possible to increase the accuracy of the estimates.= ---------------------------------------------------i  dt T 2 T Figure 4.31 and Fig- .y ( t ) ≈ -----------------------------------------------------------------------.( y i – y i – 1 ) –2 yi + yi – 1 + yi + 1 d- 2 T T  ---. or integral.30 Integration and differentiation using data points An example of numerical integration using Scilab is given in Figure 4.30. T. Given data points. The integral is basically the average height of the two points multiplied by the width to give an area. It is often necessary to integrate or differentiate these values. The first derivative is basically the slope between two points. we can integrate and differentiate using the given equations. collected at given times. In a computer based system the time points are often equally spaced in time.31 DATA When doing experiments. y. data is often collected in the form of individual data points (not as complete functions).4.numerical methods . so the difference in time can be replaced with a sample period.( y i – y i – 1 ) = -. t. The second derivative is the change in slope values for three points.( y i + 1 – y i ) dt T T  ti – ti – 1   ti + 1 – ti  1 1 -. The basic equations for integrating and differentiating are shown in Figure 4.

numerical methods - 4.32

ure 4.32.
// // // // // // // // // integrate.sce A simple program to integrate a function To run this in Scilab use 'File' then 'Exec'. by: H. Jack Sept., 9, 2002

// define the function function foo=f(x) foo = 5 * x + 2 * log(sin(x) / x + 2); endfunction // Set the time length and step size steps = 10; x_start = 1; x_end = 10; x_delta = (x_end - x_start) / steps;

// // Loop for rectangular integration // total = 0; // set the initial sum to zero for i=0:steps, x = x_start + i * x_delta; total = total + f(x); end total = total * x_delta; printf("Rectangular integration value %f\n", total);

Figure 4.31

Integration with a Scilab program

numerical methods - 4.33

// Loop for trapezoidal integration // total = 0; // set the initial sum to zero for i=0:steps, x = x_start + i * x_delta; if i == 0 then total = total + f(x); elseif i == steps then total = total + f(x); else total = total + 2 * f(x); end end total = total * x_delta / 2; printf("Trapezoidal integration value %f\n", total);

// // Loop for Simpson's rule integration // total = 0; // set the initial sum to zero even = 0; for i=0:steps, x = x_start + i * x_delta; if i == 0 then total = total + f(x); elseif i == steps then total = total + f(x); else even = even + 1; if even > 1 then total = total + 4 * f(x); even = 0; else total = total + 2 * f(x); end end end total = total * x_delta / 3; printf("Simpsons rule integration value %f\n", total);

Figure 4.32

Integration with a Scilab Program (cont’d)

4.6.1 Switching Functions
When analyzing a system, we may need to choose an input that is more complex than inputs such as steps, ramps, sinusoids and parabolae. The easiest way to do this is to use switching functions. Switching functions turn on (have a value of 1) when their argu-

numerical methods - 4.34

ments are greater than or equal to zero, or off (a value of 0) when the argument is negative. Examples of the use of switching functions are shown in Figure 4.33. By changing the values of the arguments we can change when a function turns on or off.

u(t) 1 t u ( –t ) 1 t u(t – 1) 1 t 1 u(t + 1) 1 -1 u(t + 1) – u(t – 1) 1 -1 Figure 4.33 1 t t

Switching function examples

These switching functions can be multiplied with other functions to create a complex function by turning parts of the function on or off. An example of a curve created with switching functions is shown in Figure 4.34.

numerical methods - 4.35

f(t) 5

t seconds 0 1 3 4 f ( t ) = 5t ( u ( t ) – ( t – 1 ) ) + 5 ( u ( t – 1 ) – u ( t – 3 ) ) + ( 20 – 5t ) ( u ( t – 3 ) – u ( t – 4 ) ) or f ( t ) = 5tu ( t ) – 5 ( t – 1 )u ( t – 1 ) – 5 ( t – 3 )u ( t – 3 ) + 5 ( t – 4 )u ( t – 4 ) Figure 4.34 Switching functions to create a non-smooth function

The unit step switching function is available in Mathcad and makes creation of complex functions relatively trivial. Step functions are also easy to implement when writing computer programs, as shown in Figure 4.35.

For the function f ( t ) = 5tu ( t ) – 5 ( t – 1 )u ( t – 1 ) – 5 ( t – 3 )u ( t – 3 ) + 5 ( t – 4 )u ( t – 4 )
double u(double t){ if(t >= 0) return 1.0; return 0.0; } double function(double t){ double f; f = 5.0 * - 5.0 - 5.0 + 5.0 t * * * * u(t) (t - 1.0) * u(t - 1.0) (t - 3.0) * u(t - 3.0) (t - 4.0) * u(t - 4.0);

return f; }

Figure 4.35

A subroutine implementing the example in Figure 4.34

numerical methods - 4.36

4.6.2 Interpolating Tabular Data
In some cases we are given tables of numbers instead of equations for a system component. These can still be used to do numerical integration by calculating coefficient values as required, in place of an equation. Tabular data consists of separate data points as seen in Figure 4.36. But, we may need values between the datapoints. A simple method for finding intermediate values is to interpolate with the "lever law". (Note: it is called this because of its’ similarity to the equation for a lever.) The table in the example only gives flow rates for a valve at 10 degree intervals, but we want flow rates at 46 and 23 degrees. A simple application of the lever law gives approximate values for the flow rates.

valve angle (deg.) 0 10 20 30 40 50 60 70 80 90

flow rate (gpm) 0 0.1 0.4 1.2 2.0 2.3 2.4 2.4 2.4 2.4 Given a valve angle of 46 degrees the flow rate is, 46 – 40 Q = 2.0 + ( 2.3 – 2.0 )  ----------------- = 2.18  50 – 40 Given a valve angle of 23 degrees the flow rate is, Q = 0.4 + ( 1.2 – 0.4 )  23 – 20 = 0.64 ---------------- 30 – 20

Figure 4.36

Using tables of values to interpolate numerical values using the lever law

The subroutine in Figure 4.37 was written to return the numerical value for the data table in Figure 4.36. In the subroutine the tabular data is examined to find the interval that the flow rate value falls inside. Once this is found the valve angle is calculated as the ratio between the two known values.

numerical methods - 4.37

#define double

SIZE 10; data[SIZE][2] = {{0.0, {10.0, {20.0, {30.0, {40.0, {50.0, {60.0, {70.0, {80.0, {90.0,

0.0}, 0.1}, 0.4}, 1.2}, 2.0}, 2.3}, 2.4}, 2.4}, 2.4}, 2.4}};

double angle(double rate){ int i; for(i = 0; i < SIZE-1; i++){ if((rate >= data[i][0]) && (rate <= data[i+1][0]){ return (data[i+1][1] - data[i][1]) * (rate - data[i][0]) / (data[i+1][0] - data[i][0]) + data[i][1]; } } printf("ERROR: rate out of range\n"); exit(1); }

Figure 4.37

A tabular interpolation subroutine example

4.6.3 Modeling Functions with Splines
When greater accuracy is required, smooth curves can be fitted to interpolate points. These curves are known as splines. There are multiple methods for creating splines, but the simplest is to use a polynomial fitted to a set of points. The example in Figure 4.38 shows a spline curve being fitted for three data points. In this case a second order polynomial is used. The three data points are written out as equations, and then put into matrix form, using the coefficients as the unknown values. The matrix is then solved to obtain the coefficient values for the final equation. This equation can then be used to build a mathematical model of the system.

numerical methods - 4.38

The datapoints below might have been measured for the horsepower of an internal combustion engine on a dynamometer. S (RPM) 1000 4000 6000 P (HP) 105 205 110

In this case there are three datapoints, so we can fit the curve with a second (3-1) order polynomial. The major task is to calculate the coefficients so that the curve passes through all of the given points. 2 P ( S ) = AS + BS + C Data values can be substituted into the equation, 2 P ( 1000 ) = A1000 + B1000 + C = 105 2 P ( 4000 ) = A4000 + B4000 + C = 205 2 P ( 6000 ) = A6000 + B6000 + C = 110 This can then be put in matrix form to find the coefficients, 2 1000 1000 1 A 105 2 = 205 4000 4000 1 B 110 C 2 6000 6000 1 –1 A 1000000 1000 1 105 = 16000000 4000 1 B 205 C 36000000 6000 1 110 Note: in this example the inverse matrix is used, but other methods for solving systems of equations are equally valid. If the equations were simpler, substitution might have been a better approach.

–8 –7 –7 A 6.667x10 – 1.667x10 1.000x10 105 = B –4 –3 – 4 205 – 6.667x10 1.167x10 – 5.000x10 C 110 1.600 –1.000 0.400 –5 A – 1.617x10 B = 0.114 C 7.000 –5 2 P ( S ) = ( – 1.617x10 )S + 0.114S + 7.000

Figure 4.38

A spline fitting example

numerical methods - 4.39

The order of the polynomial should match the number of points. Although, as the number of points increases, the shape of the curve will become less smooth. A common way for dealing with this problem is to fit the spline to a smaller number of points and then verify that it matches the remaining points, or use a least squares method to find the best approximation.

4.6.4 Non-Linear Elements
Despite our deepest wishes for simplicity, most systems contain non-linear components. In the last chapter we looked at dealing with non-linearities by linearizing them locally. Numerical techniques will handle non-linearities easily, but smaller time steps are required for accuracy. Consider the mass and an applied force shown in Figure 4.39. As the mass moves an aerodynamic resistance force is generated that is proportional to the square of the velocity. This results in a non-linear differential equation. This equation can be numerically integrated using a technique such as Runge-Kutta. Note that the state equation matrix form cannot be used because it requires linear equations.

x F M

·2 20x

∑ Fx

·2 ·· = F – 20x = Mx (1) (2)

· x = v 2 F · v = – 20 v + ----------M M

Figure 4.39

Developing state equations for a non-linear system

4.7 CASE STUDY
Consider the simplified model of a car suspension shown in Figure 4.40. The model distributes the vehicle weight over four tires with identical suspensions, so the mass of the vehicle is divided by four. In this model the height of the road will change and drive the tire up, or allow it to drop down. The tire acts as a stiff spring, with little deflection. The upper spring and damper are the vibration isolation units. The damper has been designed to stiffen as the damper is compressed. The given table shows how the damping

numerical methods - 4.40

coefficient varies with the amount of compression.

Mc

yc

K s1

Kd

N K s1 = 50000 --m N K s2 = 200000 --m M c = 400Kg M t = 20Kg

L = yc – yt L (cm) 30 0 -20 -30 Kd (Ns/m) 1000 1600 2000 2250

Mt

yt

K s2 yr

Figure 4.40

A model of a car suspension system

For our purposes we will focus only on the translation of the tire, and ignore its rotational motion. The differential equations describing the system are developed in Figure 4.41.

numerical methods - 4.41

·· Mc y c

Mc

yc

· · K s1 ( y t – y c ) K d ( y t – y c ) · · · = K s1 ( y t – y c ) + K d ( y t – y c ) – M c g = M c y c ·· · · Mc y c = ( K s1 )y t + ( – K s1 )y c + ( K d )y t + ( –K d )y c – M c g d---- y c = v c dt d---- yt = v t dt  K s1  – K s1  Kd  – K d · v c =  --------- y t +  ------------ y c +  ------- v t +  ---------  v c + ( – g )  Mc   Mc   Mc  Mc 

∑F

(1) (2) (3)

K s1 ( y t – y c )

· · Kd ( yt – yc )

Mt

·· M t yt

K s2 ( y r – y t ) · · ·· = K s2 ( y r – y t ) – K s1 ( y t – y c ) – K d ( y t – y c ) – Mt g = M t y t ·· · · M t y t = ( K s2 )y r + ( – K s2 – K s1 )y t + ( K s1 )y c + ( – K d )y t + ( K d )y c – Mt g

∑F

 K s2  –K s2 – K s1  K s1  – K d  K d · v t =  --------- y r +  ---------------------------- y t +  --------- y c +  ---------  v t +  ------  v c + ( – g ) Mt  Mt     Mt   Mt   M t

(4)

Figure 4.41

Differential and state equations for the car suspension system

The damping force must be converted from a tabular form to equation form. This is done in Figure 4.42.

numerical methods - 4.42

There are four data points, so a third order polynomial is required. K d ( L ) = AL + BL + CL + D
3 2

L (cm) 30 0 -20 -30

Kd (Ns/m) 1000 1600 2000 2250

The four data points can now be written in equation form, and then put into matrix form. 1000 = A ( 0.3 ) + B ( 0.3 ) + C ( 0.3 ) + D 3 2 1600 = A ( 0 ) + B ( 0 ) + C ( 0 ) + D 2000 = A ( –0.2 ) + B ( – 0.2 ) + C ( –0.2 ) + D 3 2 2250 = A ( –0.3 ) + B ( – 0.3 ) + C ( –0.3 ) + D 0.027 0.09 0.3 0 0 0 – 0.008 0.04 –0.2 – 0.027 0.09 –0.3 1 1 1 1 A 1000 B = 1600 C 2000 D 2250
3 2 3 2

The matrix can be solved to find the coefficients, and the final equation written. A – 2778 B = 277.8 C – 1833 D 1600 K d ( L ) = ( – 2778 )L + ( 277.8 )L + ( –1833 )L + 1600 Figure 4.42 Fitting a spline to the damping values
3 2

The system is to be tested for overall deflection when exposed to obstacles on the road. For the initial conditions we need to find the resting heights for the tire and car body. This can be done by setting the accelerations and velocities to zero, and finding the resulting heights.

numerical methods - 4.43

The initial accelerations and velocities are set to zero, assuming the car has settled to a steady state height. This then yields equations that can be used to calculate the initial deflections. Assume the road height is also zero to begin with. – K s1 Kd –Kd K s1 0 =  -------  y t +  ----------  y c +  ------ 0 +  --------  0 + ( – g ) M   M  M  M  c c c c Mc y c = y t – g ------K s1 K s2 – K s2 – K s1 K s1 –Kd Kd 0 =  -------  0 +  ------------------------- y t +  -------  y c +  --------  0 +  -----  0 + ( –g ) M    M  M  M  Mt t t t t gM t = ( – K s2 – K s1 )y t + ( K s1 )y c Mc gM t = ( – K s2 – K s1 )y t + ( K s1 )  y t – g -------   K 
s1

Mc + Mt Mc y c = – g ------------------- – g ------K s2 K s1

Mc + M y t = – g -------------------t K s2 M c + Mt M c y c = – g  ------------------- + -------   K s2 K s1

Figure 4.43

Calculation of initial deflections

The resulting calculations can then be written in a computer program for analysis, as shown in Figure 4.44.

numerical methods - 4.44

#include <stdio.h> #include <math.h> #define #define #define #define #define SIZE y_c y_t v_c v_t 4 0 1 2 3 /* define state variables */

#define N_step 10000 #define h_step 0.001 #define #define #define #define #define Ks1 Ks2 Mc Mt grav

// number of steps // define step size

50000.0 /* define component values */ 200000.0 400.0 20.0 9.81

void integration_step(double h, double state[], double derivative[]){ int i; for(i = 0; i < SIZE; i++) state[i] += h * derivative[i]; } double damper(double L){ return (-2778*L*L*L + 277.8*L*L - 1833*L + 1600); } double y_r(double t){ /* return 0.0; /* a zero input to test the initial conditions */ /* return 0.2 * sin(t);/* a sinusoidal oscillation */ return 0.2; */ /* a step function */ /* return 0.2 * t; */ /* a ramp function */ }

Figure 4.44

Program for numerical analysis of suspension system

numerical methods - 4.45

void d_dt(double t, double state[], double derivative[]){ double Kd; Kd = damper(state[y_c] - state[y_t]); derivative[y_c] = state[v_c]; derivative[y_t] = state[v_t]; derivative[v_c] = (Ks1/Mc)*state[y_t] - (Ks1/Mc)*state[y_c] + (Kd/Mc)*state[v_t] - (Kd/Mc)*state[v_c] - grav; derivative[v_t] = (Ks2/Mt)*y_r(t) - ((Ks2+Ks1)/Mt)*state[y_t] + (Ks1/Mt)*state[y_c] - (Kd/Mt)*state[v_t] + (Kd/Mt)*state[v_c] - grav; } main(){ double state[SIZE]; double derivative[SIZE]; FILE *fp_out; double t; int i; state[y_c] state[y_t] state[v_c] state[v_t] = = = = - grav * ( (Mc/Ks1) + (Mt + Mc)/Ks2 ); /* initial values */ - grav * (Mt + Mc) / Ks2; 0.0; 0.0;

if((fp_out = fopen("out.txt", "w")) != NULL){ /* open the file */ fprintf(fp_out, " t Yc Yt Vc Vt \n"); for(t = 0.0, i = 0; i < N_step; i++, t += h_step){ if((i % 100) == 0) fprintf(fp_out, "%f %f %f %f t, state[y_c], state[y_t], state[v_c], state[v_t]); d_dt(t, state, derivative); integration_step(h_step, state, derivative); } } else { printf("ERROR: Could not open file \n"); } fclose(fp_out); }

%f \n",

Figure 4.45

Program for numerical analysis of suspension system (continued)

This program was then used to test various design cases by selecting input types for changes in the road height, and then calculating how the tire and vehicle heights would change as a result. Some of these results are seen in Figure 4.46. These results were obtained by running the program, and then graphing the results in a spreadsheet program. The input of zero for the road height was used to test the program. As shown the height of the vehicle changes, indicating that the initial height calculations are correct, and the model is stable. The step function shows some oscillations that settle out to a stable final value. The oscillation is relatively slow, and is fully transmitted to the automobile. The ramp function shows that the car follows the rise of the slope with small transient effects at the start.

numerical methods - 4.46

y r ( t ) = 0.0
0 1 8 15 22 29 36 43 50 57 64 71 78 85 92 99

y r ( t ) = 0.2m
0.3 0.25

-0.02 0.2 -0.04 0.15 0.1 0.05 0 1 -0.05 -0.1 -0.1 -0.12 -0.15 8 15 22 29 36 43 50 57 64 71 78 85 92 99

-0.06

Series1 Series2

Series Series

-0.08

y r ( t ) = 0.2m sin ( t )
0.3 0.2

y r ( t ) = ( 0.2m )t
2.5

2

0.1
1.5

0 1 -0.1
0.5

8

15

22

29

36

43

50

57

64

71

78

85

92

99

Series1 Series2
1

Series Series

-0.2

-0.3

0 1 8 15 22 29 36 43 50 57 64 71 78 85 92 99

-0.4

-0.5

Figure 4.46

Graphs of simulation results

4.8 SUMMARY
• State variable equations are used to reduced to first order differential equations. • First order equations can be integrated numerically. • Higher order integration, such as Runge-Kutta increase the accuracy. • Switching functions allow functions terms to be turned on and off to provide more complex function. • Tabular data can be used to get numerical values.

numerical methods - 4.47

4.9 PRACTICE PROBLEMS
1. Convert the following differential equations to state variable form. ·· · x + 2x + 3x + 5y = 3 ·· · y + y + 6y + 9x = sin ( t ) 2. a) Put the differential equations given below in state variable form. b) Put the state equations in matrix form ·· · · · y 1 + 2y 1 + 3y 1 + 4y 2 + 5y 2 + 6y 3 + t + F = 0 · · ·· y 1 + 7y 1 + 8y 2 + 9y 2 + 10y 3 + 11y 3 + 5 cos ( 5t ) = 0 · ·· · y 1 + 12y 2 + 13y 3 = 0 y 1, y 2, y 3 = outputs F = input

3. Develop state equations for the mass-spring-damper system below. x Kd F M Ks

4. The system below is comprised of two masses. There is viscous damping between the masses and between the bottom mass and the floor. The masses are also connected with a cable that is run over a massless and frictionless pulley. Write the differential equations for the system, and put them in state variable form. x1 M1 B1 M2 B2 x2 F Ks

numerical methods - 4.48

5. Do a first order numerical integration of the derivative below from 0 to 10 seconds in one second steps. Assume the system starts undeflected.
2 d---- x ( t ) = 5 ( t – 4 ) dt

6. Given the differential equation below integrate the values numerically (show all work) for the first ten seconds in 1 second intervals. Assume the initial value of x is 1. You may use first order or Runge Kutta integration. · x + 0.25x = 3 7. Write a Scilab program to calculate the area under the function below using a numerical method, such as Simpson’s rule. Test it by using the range from x=1 to x=1.2. Assume the sine function is in radians sin x f ( x ) = 5x + 2 ln  ---------   x  8. Write a computer program in a language of your choice (C/C++, Java, Scilab script, etc.) that will numerically integrate the differential equation below. ·· · θ + 3θ + 9θ = 10 9. Given the following differential equation and initial conditions, draw a sketch of the first 5 seconds of the output response. The input is a step function that turns on at t=0. Use at least two different methods, and compare the results. ·· · initial conditions Vi ( t ≥ 0 ) = 5V 0.5Vo + 0.6Vo + 2.1V o = 3Vi + 2 Vo ( 0 ) = 0V · Vo ( 0 ) = 1V 10. a) For the mass-spring-damper system below solve the differential equation as a function of time. Assume the system starts at rest and undeflected. b) Also solve the problem using your calculator (and state equations) to verify your solution. Sketch the results. x Kd K s = 10 K d = 10 F M = 10 M Ks F = 10

11. The mechanical system below is a mass-spring-damper system. A force ’F’ of 100N is applied to the 10Kg cart at time t=0s. The motion is resisted by the spring and damper. The spring

numerical methods - 4.49

coefficient is 1000N/m, and the damping coefficient is to be determined. Follow the steps below to develop a solution to the problem. Assume the system always starts undeflected and at rest. a) Develop the differential equation for the system. b) Solve the differential equation using damping coefficients of 100Ns/m and 10000Ns/m. Draw a graph of the results. c) Develop the state equations for the system. d) Solve the system with a first order numerical analysis using Scilab for damping coefficients of 100Ns/m and 10000Ns/m. Draw a graph of the results. e) Solve the system with a Runge Kutta numerical analysis using Scilab for damping coefficients of 100Ns/m and 10000Ns/m. Draw a graph of the results. f) Write a computer program (in C, Java or Fortran) to do the Runge Kutta numerical integration in step e). Draw a graph of the results. g) Compare all of the solutions found in the previous steps. h) Select a damper value to give an overall system damping coefficient of 1. Verify the results by numerically integrating. x Kd F M Ks

12. For the mechanism illustrated in the figure below the values are K s1=Ks2=100N/m, M1=M2=1kg, F=1N. Assume that the system starts at rest, and the springs are undeformed initially. x1 x2

F M1

K s1 M2

K s2

a. Derive the differential equations for the system. b. Put the equations in state variable form. c. Put the equations in state variable matrices. d. Use a calculator to find values for x 1 and x2 over the first 10 seconds. Provide the results in a table in 1 second intervals. e. Use Scilab to plot the values for the first 10 seconds. f. Use a Scilab program and the Runge-Kutta method to produce a graph of the first 10 seconds.

11 – ----. Use a C program to produce a graph of points for the first 10 seconds. 4.– ----.10 PRACTICE PROBLEM SOLUTIONS 1.9.v 3 12 12 –v1 7 8911 5= ------.y 2 – ----.50 g.– ----.y = 0 dt v –3 u –9 2. using the first-order approximation method.– ----10 10 10 . · x · y · v · u = v = u = – 2v – 3x – 5y + 3 = – u – 6y – 9x + sin t 0 0 –5 –6 1 0 –2 0 0 1 0 –1 x 0 y + 0 v 3 u sin t x 0 d---. h.v 2 – ----.0 10 y1 y2 y3 v1 v2 v3 0 0 0 + –t–F 0 5– ----. Repeat step g. · y1 · y2 · y3 · v1 · v2 · v3 = v1 = v2 = v3 = – 2v 1 – 3y 1 – 4v 2 – 5y 2 – 6v 3 – t – F – v 1 13 = ------.3 = dt v 1 v2 v3 0 0 0 0 0 0 – 3 –5 0 0 0 0 0 0 0 1 0 0 –2 1– ----12 1– ----10 0 0 1 0 0 1 –4 –6 13 0 – ----12 8– ----.y 3 – ----.cos ( 5t ) 10 10 10 10 10 10 y1 y2 d.– ----.4.y 1 – ----.y ---.cos t 10 7.numerical methods .

 +  -----------------  M -  M -  M  2 2 2 5. M  M  M  M  1 1 1 1 · · B 1 ( x1 – x2 ) T M2 F · B2 x2 ∑F · · · ·· = – T + B1 ( x 1 – x 2 ) + F – B2 x 2 = M2 x 2 B1 – B1 – B2 · –T+F v 2 = v 1  -----.51 3. +  -----.4. + x  -------. +y +x T M1 · · B 1 ( x1 – x2 ) Ks x1 ∑F · · ·· = – T – K s x1 – B1 ( x1 – x2 ) = M 1 x1 · v1 = x1 · v2 = x2 –K s –B 1 B1 · –T v 1 = x 1  -------- + v 1  -------- + v 2  -----. x · x = v –Kd –K s F · v = v  -------. + v 2  -------------------.numerical methods . +  ---- -  M  M-  M F M · Kdx Ks x 4. h = 1 x ( t + h ) = x ( t ) + h5 ( t – 4 ) 2 t 0 1 2 3 4 5 6 7 8 9 10 x 0 80 125 145 150 150 155 175 220 300 425 .

870 0. x' = 3 – 0.25x dx ( t + h ) = x ( t ) + h  ---.652 0.16 0.36 8.4 x’ 2.9 11.0 10.275 0.25x ( t ) ) x ( t + h ) = 0.55 1.06 1.206 .numerical methods .52 6.81 7.75 5.52 9.2 11.5 10.75x ( t ) + 3 t 0 1 2 3 4 5 6 7 8 9 10 x 1 3.75 2.367 0.x ( t )  dt  x ( t + h ) = x ( t ) + 1 ( 3 – 0.39 10.489 0.4.

even = 0.numerical methods . x = x_start + i * x_delta. // set the initial sum to zero for i=0:steps. // set the initial sum to zero for i=0:steps. else total = total + 2 * f(x). end end total = total * x_delta / 2. // Loop for Simpson's rule integration total = 0. total = total + f(x). total). x_delta = (x_end . x_start = 1. elseif i == steps then total = total + f(x). else even = even + 1. printf("Rectangular integration value %f\n". . x = x_start + i * x_delta.. if i == 0 then total = total + f(x). end end end total = total * x_delta / 3. total). printf("Trapezoidal integration value %f\n". // Loop for rectangular integration total = 0. Jack Sept. else total = total + 2 * f(x). endfunction // Set the time length and step size steps = 10. // by: H. // integrate. printf("Simpsons rule integration value %f\n". // set the initial sum to zero even = 0. if i == 0 then total = total + f(x). elseif i == steps then total = total + f(x). 9. x_end = 10.53 7. total).sce . // Loop for trapezoidal integration total = 0. 2002 // define the function function foo=f(x) foo = 5 * x + 2 * log(sin(x) / x + 2). for i=0:steps. if even > 1 then total = total + 4 * f(x).x_start) / steps. x = x_start + i * x_delta.4. end total = total * x_delta.A simple program to integrate a function // To run this in Scilab use 'File' then 'Exec'.

step_t = 1. double theta.h> int main(){ int steps = 100.54 8. } } . omega_last = omega. step_t.0. theta_last.0. omega. omega_last. i. theta = theta_last + step_t * omega_last. #include <stdio. theta = 0. for(i = 0. i < steps. omega = omega_last + step_t*(-3 * omega_last . printf("%f %f %f \n".0. theta. omega).numerical methods .4. omega = 0. i+step_t.9 * theta_last + 10). i++){ theta_last = theta.

endfunction // Set the time length and step size for the integration steps = 1000. h = (t_end . // initial conditions // define the state matrix function // the values returned are [x.2) . [-4].095 · Y o = – 1. for i=1:steps. t = [t_start].095]. [-2.238) + 8. F4 = h * f(X(\$. v] function foo=f(state. t(\$. t(\$.:) + (F1 + 2. t(\$. tt = i * h.:) + h]. t_start = 0.:).t) foo = [ state(\$.6t cos ( 1. 2003 v0 = 0. a) b) Vo ( t ) = – 8.0*F3 + F4)/6. -8. -5].0. // // Loop for integration // for i=1:steps.0]. t(\$.:)). xtitle('Time (s)').:) + F3. t_end = 10.. // // Generate points from the given function // XX = [v0].331 * exp(-0. . t(\$.331e · Vo = Y o – 0. F1 = h * f(X(\$. end // // Graph the values // plot2d(t.sce // by: H.2Y o – 4.960t – 0.0.0).238 ) + 8. XX = [XX .:) + h/2.55 9. c) The two curves produced by the scilab program overlap.4.:) + h/2.2*state(\$. 2).0).4. X = [X . -1. X=[v0. y0 = 1. F3 = h * f(X(\$. leg="explicit"). y0].0. so the results agree.1) + 34].numerical methods . XX.:) + F1/2.:) + F2/2. 23. X(\$.2*state(\$.0*F2 + 2.6 * tt) * cos( 1. leg="position@velocity").2V o + 34 // assign4_1.:) + h). end plot2d(t. X.960 * tt . t = [t . Jack Sept.t_start) / steps. F2 = h * f(X(\$.

= – 1.63s on graph .56 10.4.t + C 2 + 1  2  – 0.sin ( C 2 ) = 0.5C 1 cos ( C 2 ) = 0 2 1 3 – -----.t + C 2  2  particular: xp = B 0+0+B = 1 xp = 1 initial conditions: x = xh + xp = C1 e – 0.t + 1.5t 3 sin  -----. F M · Kd x Ks x x homogeneous: ·· · x+x+x = 0 A +A+1 = 0 – 1 ± 1 – 4( 1 ) ( 1 ) 3 A = --------------------------------------------.5236 ) –1 tan ( C 2 ) = -----3 x = –1.C sin ( C 2 ) – 0.5236 ) + 1 = 0 –1 C 1 = -------------------------------.= – 0.5t 3 3 3 x' = – -----.C e sin  -----.t + C 2 – 0.5 ± j -----2(1 ) 2 – 0.5C 1 e cos  -----.219.155 cos ( – 0.numerical methods .047 + 1  2  First peak is at x=1. t=3.5236  3 x ( 0 ) = C 1 cos ( – 0.5 cos ( C 2 ) 2 –1 C 2 = atan  ------ = – 0.t + C 2  2   2  2 1 3 x' ( 0 ) = – -----.5t 2 · ·· = F – K d x – K s x = Mx ·· · Mx + K d x + K s x = F ·· · 10x + 10x + 10x = 10 ∑F ·· · x+x+x = 1 3 cos  -----.5t 3 xh = C1 e cos  -----.155 e – 0.5t – 0.

000s e) 0. b) 0.10 x 1 ( t ) = – 0.001s d) 0.361s e) 0.X 1 1 M M M For 100N/m: all solutions are underdamped and overshoot at.1166 at t = 0.10 K d = 100 --m – 0.524 ) + 0.1e K d = 10000 --m 0 1 X 0 d X0 0 + ---= –K –K F s d dt X ----------.57 11.-------.0s h) K d = 200 Ns ----m (verify with numerical integration also) c) g) .06321 at t = 10.363s d) 0.06321 at t = 10.115e cos ( 5 3t – 0.000s f) 0.360s For 10000N/m: all solutions are overdamped.06321) is.1t – 5 – 999. x +  ----- x = --- M  M M – 5t N x 1 ( t ) = – 0.1163 at t = 0. a) b) Kd · Ks F ·· x +  ----.1163 at t = 0. The time to reach the time constant (at 0.06321 at t = 10.363s f) 0.9t N + 10 e + 0.4.06321 at t = 10.1163 at t = 0. b) 0.numerical methods .

1 = dt x 2 v2 0 1 – K s1 ---------.v ---. M    M 2 2 c) x1 d. + x 2  ------.  M   M  M  1 1 1 · x2 = v2 K s1 – K s1 – K s2 · x 1 = x 1  ------.58 12.0 M1 0 K s1 ------M1 0 0 x1 v1 0 F-----+ M1 0 0 0 0 0 1 x2 K s1 – K s1 – K s2 ------. + x 2  -------------------------. +  -----.0 v 2 M2 M2 .4.numerical methods .0 -------------------------. b) · x1 = v1 – K s1 K s1 · Fx 1 = x 1  ---------.

4). t_end = 10. 2).:). end // Graph the values for part e) plot2d(t. M2 = 1. end . end // First order integration for part h) X=[x0. -Ks1/M1*state(\$.0). v1=%f. xtitle('Time (s)'). t_start = 0. [-2. // define the state matrix function the values returned are [x.0]. t(\$. X=[x0. v2=%f \n". X(i. g) // System component values Ks1 = 100. for time_count=0:20. x2=%f.:) + F3.1)-(Ks1+Ks2)/M2*state(\$.:) + h/2. 3). printf("Point at t=%f x1=%f. v] function foo=f(state.0. // Loop for integration for i=1:steps. 4)). state(\$.:) + h].3)+F/M1. v0. // initial conditions v0 = 0. // printf the values for part f) printf("\n\nPart e output\n\n"). for time_count=0:20. time_count/2.0*F3 + F4)/6. X(i. -7. 3).2). i = (time_count/2) / h + 1. X(i.3)]. t(\$. time_count/2. printf("Point at t=%f x1=%f.:) + h/2. t = [t_start]. -5.59 e). t(\$.:) + F1 ]. t(\$. X(i. -9]. X = [X .:). h = (t_end . Ks1/M2*state(\$. X(i.:)).t_start) / steps. F4 = h * f(X(\$. x1 = 0.:) + F1/2. v1]. X(i.0). endfunction // Set the time length and step size for the integration steps = 10000.numerical methods .t) foo = [ state(\$.:) + F2/2.:)). 1). v1]. x1. v2=%f \n". v1=%f. x0 = 0. leg="position1@velocity1@position2@velocity2"). M1 = 1. F1 = h * f(X(\$. for i=1:steps.0*F2 + 2. X. X(i. t = [t . X(i. x2=%f. t(\$.1)+Ks1/M1*state(\$.:) + h). F1 = h * f(X(\$. X(\$. F3 = h * f(X(\$. t = [t . Ks2 = 100. X(\$. F2 = h * f(X(\$. X = [X . F = 1.0. 4)). t = [t_start].:) + h].:) + (F1 + 2. 1). end printf("\n\nPart g output \n\n").4. t(\$. 2). f). i = (time_count/2) / h + 1. v0. x1. t(\$. v1 = 0.

··· ·· · x + 4x + 2x + 5x = 10 2.27.4. Find the state at two milliseconds using a) explicit integration. find the state of the system 1 second later. double X[]. With a numerical method of your choice. b) first order numerical integration and c) Runge-Kutta integration. // // State Equations Calculated Here // void derivative(double t.001s. sin x fx ) = 5x + 2 ln  --------. dX[3] = Ks1 / M2 * X[0] . Write a Scilab program to integrate the area under the function below using a numerical method. Numerically integrate one time step of the differential equation below using a) first order integration and b) Runge Kutta integration. dX[2] = X[3].  x  3.numerical methods . dX[1] = -Ks1 / M1 * X[0] + Ks1 / M1 * X[2] + Force / M1. For the numerical methods use a timestep of h=0. · k + 10k = 5 k ( 0 ) = 20 4. ·· · θ + 3θ + 9θ = 10 . The differential equation below describes a first order system that starts with an initial value of k=20.60 h) The following subroutine is used in place of the subroutine in the program shown in Figure 4. double dX[]){ dX[0] = X[1]. Convert the third order differential equation below to state equation form. Show all calculations.(Ks1 + Ks2) / M2 * X[2].11 ASSIGNMENT PROBLEMS 1. such as Simpson’s rule. Find the area from 1 to 10.26 and Figure 4. ----> The final results must be put in a table for easy comparison. } 4.

M1=M2=1kg. Kd1=10Nm/s. Assume that the system starts at rest. For the mechanism shown in the figure below the values are Ks1=Ks2=100N/m. Repeat step g. g. Derive the differential equations for the system.numerical methods . f. b. F=1N. h. and the springs are undeformed initially.61 5. Use a calculator to find values for x 1 and x2 over the first 10 seconds. . x1 K d1 M1 K s1 M2 x2 F K s2 a. Put the equations in state variable form. Use Scilab to plot the values for the first 10 seconds. Use a C program to produce a graph of points for the first 10 seconds.4. c. d. using the first-order approximation method. Use a Scilab program and the Runge-Kutta method to produce a graph of the first 10 seconds. Put the equations in state variable matrices. Provide the results in a table in 1 second intervals. e.

the angular acceleration can be found by differentiating again. dampers.1 Basic properties of rotation . levers. θ. ω. The angular acceleration can be integrated to find the angular velocity.2 α =  ---- ω =  ---- θ  dt  dt (1) (2) (3) (4) θ ( t ) = ∫ ω ( t ) dt = ω ( t ) = ∫ α ( t ) dt ∫ ∫ α ( t ) dt dt (5) α( t ) = T ( t) --------JM where. springs. ROTATION Topics: • Basic laws of motion • Inertia.1. 5.rotation . gears and belts • Design cases Objectives: • To be able to develop and analyze differential equations for rotational systems. equations of motion θ T OR dω =  ---- θ  dt dd. The angular acceleration is proportional to an applied torque. Given the angular position. the angular velocity can be found by differentiating once. the angular velocity can be integrated to find the angular position.5. α = position. but inversely proportional to the mass moment of inertia.1 5.1 INTRODUCTION The equations of motion for a rotating mass are shown in Figure 5. velocity and acceleratio n J M = second mass moment of inertia of the bd oy T = torque applied to bd oy Figure 5.

find the state 5 seconds later.opposes acceleration and deceleration • springs . as they were for translating systems.change rotational speeds and torques .2 MODELING Free Body Diagrams (FBDs) are required when analyzing rotational systems.oppose velocity • levers . • inertia . θ ( 5 ) = 86rad ω(5 ) = Figure 5. Given the initial state of a rotating mass.rotation .2 Drill problem: Find the position with the given conditions 5. The main difference is that ’torque’ normally refers to a rotating moment.rotate small angles • gears and belts . θ 0 = 1rad ω 0 = 2 rad -------s α = 3 rad -------2 s ans. The force components normally considered in a rotational system include.5.2 Note: A ’torque’ and ’moment’ are equivalent in terms of calculations.resist deflection • dampers .

1 Inertia When unbalanced torques are applied to a mass it will begin to accelerate. Figure 5.3. ω . Moment of inertia values are typically calculated about the centroid. When dealing with rotational acceleration it is important to use the mass moment of inertia.3 5. or found in tables. The center of rotation for free body rotation will be the centroid. If the object is constrained to rotate about some point. θ. The sum of applied torques is equal to the inertia forces shown in Figure 5. The parallel axis theorem provides the method to shift a moment of inertia from a centroid to an arbitrary center of rotation. other than the centroid.3 Summing moments and angular inertia The mass moment of inertia determines the resistance to acceleration. α ∑T = JM α (6) (7) (8) (9) J T J M = I xx + Iyy 2 Ixx = ∫ y dM 2 Iyy = ∫ x dM Note: The ’mass’ moment of inertia will be used when dealing with acceleration of a mass.4. . as shown in Figure 5. in rotation. This can be calculated using integration.rotation . the moment of inertia value must be recalculated.2. Later we will use the ’area’ moment of inertia for torsional springs. not the area moment of inertia.5.

are then added to give the polar moment of inertia. In this problem the density of the material is calculated for use in the integrals. . J A = area moment about the new point J˜ = area moment about the centroi d A A = mass of the object r = distance from the centroid to the new point Figure 5.5. J M = mass moment about the new point ˜ J M = mass moment about the centroid M = mass of the object r = distance from the centroid to the new point Figure 5. If the gravity varies over the length of the (very long) object then the center of gravity should be used.5 Parallel axis theorem for shifting a area moment of inertia Aside: If forces do not pass through the center of an object. If the object is made of a homogeneous material. the area and volume centroids can be used as the center.rotation . An example of calculating a mass moment of inertia is shown in Figure 5. it will rotate.4 2 J M = J˜ + Mr M where. The integrals are then developed using slices for the integration element dM.6.4 Parallel axis theorem for shifting a mass moment of inertia 2 J A = J˜A + Ar where. If the object is made of different materials then the center of mass should be used for the center. This is then shifted from the centroid to the new axis using the parallel axis theorem. The integrals for the moments about the x and y axes.

5m ) + ( –1m ) ) = 209. Find the mass moment of inertia.67Kgm 2 2 2 The centroid can now be shifted to the center of rotation using the parallel axis theorem.2Kgm M Figure 5.6 Mass moment of inertia example .33Kgm  3 3  I yy = ∫–5 x 5 2 dM = ∫– 5 x 3 5 2 ρ2 ( 4m ) dx = 1Kgm ---3 3 –1 x 3 5 –5 – 1 ( 5m ) 2 .5 The rectangular shape to the right is constrained to rotate about point A. The given dimensions are in meters. 2 2 2 2 2 J M = J˜ + Mr = 136.33Kgm + 83.25Kgm  -------------.– ----------------- = 53.67Kgm + ( 10Kg ) ( ( – 2.125Kgm 2 ( 5m )2 ( 4m ) I xx = ∴ 4 -5 -2. The total mass of the object is 10kg. 10Kg –2 ρ = ------------------------------.= 0.– ----------------- = 83.( – 4m ) = 1.5.33Kgm  3 3  J M = Ixx + I yy = 53.( – 5m ) ∴ = 1Kgm  -------------.5 -1 -4 5 ∫–4 y 4 2 dM = ∫– 4 y 4 2 ρ2 ( 5m ) dy = 1. First find the density and calculate the moments of inertia about the centroid.33Kgm = 136.25Kgm ---3 3 3 –1 y 3 4 –4 – 1 ( 4m ) 2 .rotation .

Find the mass moment of inertia WITHOUT using the parallel axis theorem. -5 4 -2. The given dimensions are in meters.8Kgm 2 J M = 209. The total mass of the object is 10kg.5 -1 -4 5 ans.6 The rectangular shape to the right is constrained to rotate about point A.2Kgm 2 Figure 5.7 Drill problem: Mass moment of inertia calculation .33Kgm 2 I My = 145. I Mx = 66.rotation .5.

7 The 20cm diameter 10 kg cylinder to the left is sitting in a depression that is effectively frictionless. of the rod.2.9. G. A simple example of a solid rod torsional spring is shown in Figure 5. The constant parameters can be lumped into a single spring coefficient similar to that used for translational springs. the shear modulus. If a torque of 10 Nm is applied for 5 seconds. what will the angular velocity be? ans. This torque increases as the deformation increases. JA. and the length.2 Springs Twisting a rotational spring will produce an opposing torque.5rad rad ω ( 5s ) = 125 -------s Figure 5. L.8 Drill problem: Find the velocity of the rotating shaft 5. the area moment of inertia.5. T. . θ ( 5s ) = 312. The angle of rotation is determined by the applied torque.rotation .

9 A solid torsional spring The spring constant for a torsional spring will be relatively constant. (8) (9) Figure 5.8 L θ JA G T =  --------. except for the formulation of the integration elements. When dealing with strength of material properties the area moment of inertia is required. The calculation for the area moment of inertia is similar to that for the mass moment of inertia. θ  L - T = K S ( ∆θ ) T Note: Remember to use radians for these calculations. or the geometry of the spring changes significantly.6. Note: This calculation uses the area moment of inertia. An example of calculating the area moment of inertia is shown in Figure 5. The calculations are similar to those for the mass moments of inertia.rotation . unless the material is deformed outside the linear elastic range.5. In fact you are advised to use radians for all calculations. Note the difference between the mass moment of inertia and area moment of inertia for the part. . Don’t forget to set your calculator to radians also. and based on the previous example in Figure 5.10. Also note the units. The area moment of inertia can be converted to a mass moment of inertia simply by multiplying by the density.

Keep track of units when doing this.4m A Next. Therefore if you have a table of area moments of inertia.5 -1 -4 5 3 3  ( –4m ) - = 10m  ( 4m ) – ------------------ = 426.7 + 666.5. I xx = I yy = ∫y ∫x 2 2 dA dA (8) (9) (10) (11) J A = Ixx + I yy 2 J A = J˜ + Ar A Note: You may notice that when the area moment of inertia is multiplied by the density of the material.7m --------------3   3 – 5m 4 4 J˜ = I xx + Iyy = ( 426. the mass moment of inertia is the result.7 --------------3   3 – 4m 5m 3 3  4 ( – 5m ) - = 8m  ( 5m ) – ------------------ = 666. shift the area moment of inertia from the centroid to the other point of rotation. 2 J A = J˜ + Ar A 4 2 2 ∴ = 1093. Figure 5. multiplying by density will yield the mass moment of inertia. 3 2 2 Ixx = ∫4m y dA = ∫4m y 2 ( 5m ) dy = 10m y ----– 4m – 4m 3 3 2 2 Iyy = ∫5m x dA = ∫5m x 2 ( 4m ) dx = 8m x ----– 5m – 5m 3 4m -2.4m + ( ( 4m – ( – 4m ) ) ( 5m – ( –5m ) ) ) ( ( – 1m ) + ( – 2. the area moment of inertia is calculated about the centroid by integration.rotation .9 4 -5 First.10 Area moment of inertia . All dimensions are in m.7 )m = 1093.5m ) ) ∴ = 1673m 4 Note: The basic definitions for the area moment of inertia are shown to the right.

Because the torsional spring is considered massless the torque will be the same at the other end of the spring. A torque is applied to the left hand spring.10 For a 1/2" 1020 steel rod that is 1 yard long. There are three torsional springs between two rotating masses.rotation . The right hand spring is anchored solidly in a wall. and will not move.12. ans. and forces are summed. at mass J1. find the torsional spring coefficient. Nm K s = 215 -------rad Figure 5. and for translational springs.5. FBDs are drawn for both of the masses. .11 Drill problem: Find the torsional spring coefficient An example problem with torsional springs is shown in Figure 5. and finally put in matrix form.) These equations are then rearranged into state variable equations. (Note: the similarity in the methods used for torsional.

 θ 1   J  J M2 M2 (3) (4) θ1 d.0 ω 2 JM2 J M2 0 + τ 0 0 Figure 5. Leave the results in state variable form.0 -------------------------. θ 2 +  ------. θ 1   J  JM2 M2 · θ2 = ω2 – K s3 – K s2 K s2 · ω 2 =  -------------------------.11 τ K s1 JM1 K s2 JM2 K s3 Model the system above assuming that the center shaft is a torsional spring. θ 2 +  ------.1 = dt θ 2 ω2 0 0 0 1 θ2 K s2 – K s3 – K s2 ------. θ1 K s2 ( θ 1 – θ 2 ) + ∑M ·· = τ – K s2 ( θ 1 – θ 2 ) = J M 1 θ 1 JM1 τ θ2 ·· J M 1 θ 1 = – K s2 θ 1 + K s2 θ 2 + τ · θ1 = ω1 – K s2 K s2 · ω 1 =  ----------- θ 1 +  ------.rotation .ω ---.5. θ 2 + τ J  J  M1 M1 K s2 ( θ 2 – θ 1 ) + (1) (2) ∑M ·· = – K s2 ( θ 2 – θ1 ) – K s3 θ 2 = J M 2 θ 2 JM2 K s3 θ 2 0 1 – K s2 ---------.12 A rotational spring example . and that a torque is applied to the leftmost disk.0 JM1 0 K s2 ------J M1 0 0 θ1 ω1 – K s3 – K s2 K s2 ·· θ 2 =  -------------------------.

141 -------2 s Figure 5.2. The first equation is used for a system with one rotating and one stationary part. It opposes angular velocity with the relationships shown in Figure 5.15. such as oils.13 The rotational damping equation If a wheel (JM=5kg m2) is turning at 150 rpm and the damping coefficient is 1Nms/rad. Visual comparison of the final matrices in this and the previous example reveal that the damping terms are the . The second equation is used for damping between two rotating parts.12 5. used for lubrication.12 is extended to include damping in Figure 5. ·· rad θ = – 3.14 Drill problem: Find the deceleration The example in Figure 5.rotation . In this case the damping coefficients are indicated with ’B’. The primary addition from the previous example is the addition of the damping forces to the FBDs.13. but ’Kd’ could have also been used. T = Kd ω T = Kd ( ω1 – ω2 ) Figure 5. what is the deceleration? ans.5.3 Damping Rotational damping is normally caused by viscous fluids. The state equations were developed in matrix form.

 θ1 - - J   J - JM2 M2 M2 0 0 θ1 ω1 0 + τ 0 0 (4) θ1 d.15 A System Example . θ 1 J    J  JM2 M2 M2 · (3) θ2 = ω2 JM2 K s3 θ 2 0 –K s2 ---------J M1 0 K s2 ------J M2 1 – B1 -------J M1 0 · B2 θ2 0 K s2 ------JM1 –B2 – K s3 – K s2 K s2 · ω 2 =  -------. θ2 + τ (2) J  J  J  M1 M1 M1 θ2 · ·· + ∑ M = – K s2 ( θ 2 – θ 1 ) – B2 θ 2 – K s3 θ 2 = J M 2 θ 2 –B2 · – K s3 – K s2 K s2 ·· θ 2 =  -------. τ K s1 B1 JM1 K s2 JM2 B2 K s3 Model the system above assuming that the center shaft is a torsional spring. θ1 +  ------. ω 2 +  -------------------------.rotation . and that a torque is applied to the leftmost disk. θ 2 +  -------------------------. θ1 K s2 ( θ 1 – θ 2 ) + ∑M · ·· = τ – K s2 ( θ 1 – θ 2 ) – B1 θ 1 = J M1 θ 1 JM1 τ · B1 θ1 K s2 ( θ 2 – θ 1 ) ·· · J M 1 θ 1 = – B1 θ 1 – K s2 θ 1 + K s2 θ 2 + τ · (1) θ1 = ω1 – B1 – K s2 K s2 · ω 1 =  -------- ω 1 +  ---------.ω ---. Leave the results in state variable form.13 only addition. θ 2 +  ------.1 = dt θ 2 ω2 0 – K s3 – K s2 0 -------------------------JM2 1 θ2 – B2 ω 2 -------J M2 Figure 5.5. θ2 +  ------.

if the lever is 2m long and the distance from the fulcrum to the rock is 10cm. F2 d1 d2 δ2 F1 δ1 d1 F2 δ1 ---. Figure 5.= ----. The lever above would become ineffective if it was vertical. how much force is required to lift it? ans.2.17 Drill problem: Find the required force .4 Levers The lever shown in Figure 5. Although theoretically a lever arm could rotate fully. This allows the lever to work over a range of angles. The amplification is determined by the ratio of arm lengths to the left and right of the center.16 Force transmission with a level Given a lever set to lift a 1000 kg rock .rotation . it typically has a limited range of motion.14 5.3N Figure 5.= ---d2 F1 δ2 Note: As the lever rotates the length ratio will be maintained because of similar triangles.16 can be used to amplify forces or motion.5. F = 516.

T1 = F c r 1 T2 = – Fc r2 n1 n2 ---. Spur . Bevel . The number of teeth on a gear is proportional to the diameter. To do this. as shown in Figure 5. The gear ratio is used to measure the relative rotations of the shafts. In a compound gear set two or more gears are connected on a single shaft.The gear has a conical shape.The teeth follow a helix around the rotational axis.= ---. In this example the gear ratio on the left is 4:1. and the ratio for the set on the right is 4:1. compound gear sets are required. Together they give a gear ratio of 16:1.= -----------. Gear teeth are carefully designed so that they will mesh smoothly as the gears rotate. as shown in Figure 5. Rack . . Some of the basic gear forms are listed below. The forces on gears acts at a tangential distance from the center of rotation called the pitch diameter. Helical .= ---T2 r2 n2 V c = ω1 r1 = –ω2 r2 where. For higher gear ratios more gears can be added.18.2.15 5. For example a gear ratio of 20:1 would mean the input shaft of the gear box would have to rotate 20 times for the output shaft to rotate once. allowing forces to be transmitted at angles.= -------.rotation .5 Gears and Belts While levers amplify forces and motions over limited ranges of motion. gears can rotate indefinitely.Round gears with teeth parallel to the rotational axis.= --------.A straight gear (used with a small round gear called a pinion).18 Basic Gear Relationships For lower gear ratios a simple gear box with two gears can be constructed.= ---r1 r2 –T 1 r1 n1 -------. The ratio of motions and forces through a pair of gears is proportional to their radii.5. r2 –ω1 –α1 –∆ θ 1 n2 --.19.= ---r1 ω2 α2 ∆θ 2 n1 n = number of teeth on respective gears r = radii of respective gears Fc = force of contact between gear teeth Vc = tangential velocity of gear teeth T = torques on gears Figure 5.

20.= 16 N3 N5 In this case the output shaft turns 16 times faster than the input shaft.19 A compound gear set A manual transmission is shown in Figure 5.rotation . In the transmission the gear ratio is changed by sliding (left-right) some of the gears to change the sequence of gears transmitting the force.16 N 2 = 60 N 5 = 15 N2 N4 e = -----------. . Notice that when in reverse an additional compound gear set is added to reverse the direction of rotation. Input shaft Compound gear set Output shaft N 3 = 15 N 4 = 60 Figure 5. If we reversed directions the output (former input) would now turn 1/16 of the input (former output) shaft speed.5.

At this point all of the needed gears will be meshed and turning.20 A manual transmission Rack and pinion gear sets are used for converting rotation to translation.21. Figure 5.rotation .17 9 clutch stem gear 2 Motor shaft reverse idler 7 8 3 4 10 11 6 5 Speed (gear) 1 2 3 4 reverse Gear Train 2-3-6-9 2-3-5-8 2-3-4-7 bypass gear train 2-3-6-10-11-9 In this manual transmission the gear shifter will move the gears in and out of contact. The basic relationships are shown in Figure 5. . The final step is to engage the last gear in the gear train with the clutch (plate) and this couples the gears to the wheels.5. A rack is a long straight gear that is driven by a small mating gear called a pinion.

A gear train has an input gear with 20 teeth.5.5 -------s Figure 5. A belt around two or more pulleys will act like gears.18 T = Fr where. and an output gear that has 40 teeth.rotation . a center gear that has 100 teeth. rad case 1: ω 3 = 2.22 Drill problem: Find the gear speed . Vc = ωr ∆l = r∆θ r = radius of pinion F = force of contact between gear teeth Vc = tangential velocity of gear teeth and velocity of rack T = torque on pinion Figure 5.21 Relationships for a rack and pinion gear set Belt based systems can be analyzed with methods similar to gears (with the exception of teeth). A belt wound around a drum will act like a rack and pinion gear pair. If the input shaft is rotating at 5 rad/sec what is the rotation speed of the output shaft? What if the center gear is removed? ans.5 -------s rad case 2: ω 3 = – 2.

The friction value must be calculated using the appropriate state equation. The friction force is left as a variable for the derivation of the state equations.6 Friction Friction between rotating components is a major source of inefficiency in machines. .2. There is friction between the shaft and the hole in the wall. It is the result of contact surface materials and geometries.rotation . Calculating friction values in rotating systems is more difficult than translating systems.5.19 5.23. If the friction force exceeds the maximum static friction the mechanism is then analyzed using the dynamic friction torque. An example problem with rotational friction is shown in Figure 5. The result of this calculation and the previous static or dynamic condition is then used to determine the new friction value. Basically these problems require that the model be analyzed as if the friction surface is fixed. Normally rotational friction will be given as static and kinetic friction torques.

· J M ω = τ – K s θ – T test · T = τ–K θ–J ω test s M (4) cases: Not slipping previously T test ≤ 10Nm T test > 10Nm Slipping previously T test < 6Nm T test ≥ 6Nm TF = Ttest T F = 6Nm TF = Ttest T F = 6Nm Figure 5. the torque force must be calculated. Each loop the friction conditions are checked and then used for a first-order solution to the state equations.24. θ  J  J  M M Next.23 can be analyzed using the C program in Figure 5. τ JM Ks Ts ≤ 10Nm T k = 6Nm θ Ks θ JM τ FF + ∑M ··· = τ – Ks θ – T F = JM θ (1) (2) (3) ·· J M θ = τ – K s θ – TF · θ = ω τ – TF Ks · ω =  -------------- +  ----.rotation . and then used to determine the new torque force.20 Model the system and consider the static and kinetic friction forces on the shaft on the right hand side. The program loops to integrate the state equations. .23 A friction system example The friction example in Figure 5. For the purposes of the example some component values are selected and the system is assumed to be at rest initially.5.

w = w + h * acceleration. theta = theta + h * w. "w")) != NULL){/* open a file to write the results */ for( t = 0.starting at rest here */ TF = 0. slip = 1.TF + Ks*theta) / J. t < 10.J*acceleration. w). /* time step */ theta. w. } Figure 5. t. /* the friction test force */ J = 10.21 int main(){ double int FILE h = 0. "%f. } } fclose(fp). } else { TF = Ttest.24 A C program for the friction example in Figure 5.0. t += h ){/* loop */ Ttest = tau . /* the applied torque (I picked the value) */ Ks = 10. fprintf(fp. /* the moment of inertia (I picked the value) */ tau = 5. /* the system starts with no slip */ *fp.txt". /* friction force */ Ttest. /* the acceleration */ TF. %f\n". /* the state variables */ acceleration.0. } } else { /* slipping */ if(Ttest < 6){ TF = Ttest. theta.Ks*theta .0.5. */ acceleration = 0.rotation .} } acceleration = (tau .23 ./* set the initial acceleration to zero also */ if( ( fp = fopen("out.1. theta = 0. /* set the initial friction to 0. /* the spring constant (I picked the value) */ slip = 0.0. slip = 0. } else {TF = 6.0. w = 0. if(slip == 0){ /* not slipping */ if(Ttest >= 10){ TF = 6./* the initial conditions . %f.

2. T voltage/current increases ω ss Figure 5.22 5. = V  ----. 2 d K K  ---- ω + ω  ----.5. The steadystate velocity will be a straight line function of the torque applied to the motor. In a permanent magnet motor there are magnets mounted on the stator. The differential equation for a motor is shown in Figure 5. as shown in Figure 5. The moment of inertia ’J’ should include the motor shaft. The value of the coil resistance ’R’ can be directly measured from the motor. while the rotor consists of wound coils. The value of the constant ’K’ is a function of the motor design and will remain fixed.25 Model of a permanent magnet DC motor The speed response of a permanent magnet DC motor is first-order. ω = the angular velocity of the motor K = the motor speed constant J M = the moment of inertia of the motor and attached loads R = the resistance of the motor coils T load = a torque applied to a motor shaft Figure 5. – T load ----------∴ s  -  dt  JR JM JR where. and will be derived in a later chapter. In addition the line shifts outwards as the voltage applied to the motor increases. When a voltage is applied to the coils the motor will accelerate. but when a load is added this should be added to the value of ’J’.rotation .25.7 Permanent Magnet Electric Motors DC motors create a torque between the rotor (inside) and stator (outside) that is related to the applied voltage or current.26 Torque speed curve for a permanent magnet DC motor .26.

rotation .3 OTHER TOPICS The energy and power relationships for rotational components are given in Figure 5.27 Energy and power relations for rotation Note: The units for various rotational quantities are listed to the right. E = E K + EP E K = JM ω E P = Tθ P = Tω 2 (5) (6) (7) (8) Figure 5. The pinion has a pitch diameter of 6 inches. The unit ’rad’ should be ignored as it appears/disappears sporadically. A 20:1 gear box is used to reduce the speed and increase the torque of the motor. These can be useful when designing a system that will store and release energy. The mass moves on rails with static and dynamic coefficients of friction of 0. .27.23 5. B JM units Nm -------rad Nms ---------rad Kgm 2 5. A 10 foot long shaft is required between the gear box and the rack and pinion set. coefficient Ks K d. We want to select a shaft diameter that will keep the system critically damped when a voltage step input of 20V is applied to the motor. The motor drives a 10000kg mass in translation through a rack and pinion gear set.1 respectively. They may be used to check equations by doing a unit balance.2 and 0.5.4 DESIGN CASE A large machine is to be driven by a permanent magnet electric motor.

rotation . .5.28 Motor speed curve and the derived differential equation The remaining equations describing the system are developed in Figure 5.29. These calculations are done with the assumption that the inertial effects of the gears and other components are insignificant.25 Figure 5.

141 × 10 m Kg K s θ pinion Figure 5.768 × 10 in Kg K s 2 ·· –6 –2 –1 1θ pinion =  ----.1 × 10 m Kg K s θ m – 1.θ m 20 1ω gear = ----.26 The long shaft must now be analyzed.5. This ratio can also be used to find the force applied to the mass. and the shaft be considered as a spring. then we can say that the torque load on the motor is equal to the torque in the shaft.0m  ·· –9 –2 –1 1θ pinion =  ----. This will require that angles at both ends be defined.rotation . θ gear. ω gear = angular position and velocity of the shaft at the gearo bx θ pinion. ω pinion = angular position and velocity of the shaft at the pinio n 1θ gear = ----.29 Additional equations to model the machine (3) (4) If the gear box is assumed to have relatively small moment of inertia.= M mass θ pinion π6in  6in ------ 2  ·· –6 –2 –1 θ pinion = ( θ gear – θ pinion )1.  2  ·· ∑ F mass = Fmass = Mmass xmass K s ( θ gear – θ pinion ) ·· -------------------------------------------.0254in -------------------- 20   1.141 × 10 )m Kg K s  20  · θ pinion = ω pinion · – 12 – 2 –1 –9 –2 –1 ω pinion = 57.768 × 10 in Kg K s  0. x mass = θ pinion π6in T shaft = F mass  6in ------ 2  6in K s ( θ gear – θ pinion ) = F mass  ------.θ m – θ pinion 1. This then allows .ω m 20 T shaft = K s ( θ gear – θ pinion ) The rotation of the pinion is related to the displacement of the rack through the circumferential travel.θ m – θ pinion ( 1.

30.27 the equation for the motor shaft to be put into a useful form.θ m – θ pinion  20  · –1 – 2 –1 –2 ω m = ( Vs 50.3V s rad – ω m 2s – 50505Kg m K s  ----. Having this differential equation now allows the numerical analysis to proceed.3 0 0 1 θ pinion K s 0 –1.1 × 10 – 12 –1 –1 –2 1 –2 0 0 50505K s 0 –9 0 0 θm ωm + 0 Vs 50.3V s rad ) + θ pinion ( 50505Kg m K s ) + ω m ( – 2 s ) + θm ( – 2525 Kg m K s ) The state equations can then be put in matrix form for clarity. as shown in Figure 5.30 Numerical analysis of system response These results indicate that a spring value of XXX is required to have the system . but acknowledging that they are consistent. The analysis involves iteratively solving the equations and determining the point at which the system begins to overshoot.3V s rad – ω m 2s – 50505Kg m K s ( θ gear – θ pinion ) · –1 –2 –1 –1 –2 1ω m = V s 50.141 × 10 K s 0 ω pinion The state equations for the system are then analyzed using a computer for the parameters below to find the Ks value that gives a response that approximates critical damping for a step input from 0 to10V. indicating critical damping.rotation . θm d. The units will be eliminated for brevity. Ks (rad/Nm) 100 Overshoot (rad) Figure 5. The Tload term is eliminated from equation (2) · –1 –2 –1 –1 –2 ω m = V s 50.5.ω m ---= dt θ pinion ω pinion 0 – 2525 K s 0 57.

The right most shaft is fixed in a wall. 5.rotation . Τ θ1 θ2 J M1 Ks1 B J M2 Ks2 .) 5.5 SUMMARY • The basic equations of motion were discussed.5.6 PRACTICE PROBLEMS 1. Draw the FBDs and write the differential equations for the mechanism below. • Rotational dampers and springs. • Mass and area moment of inertia are used for inertia and springs. (Note: Clearly this system is not second order. but in the absence of another characteristics we approximate it as second order. • A design case was presented.28 behave as if it is critically damped.

. and a spring damper combination on the other side with a suspended mass. A lever arm has a force on one side.5.29 2. Draw the FBDs and write the differential equations for the mechanism below. x2 F M1 Ks1 R2 R1 J1 Kd1 x1 3. θ2 R1 J M1 θ1 R2 J M2 Ks Kd x1 x2 M1 4.rotation . a) Derive the differential equations for the following system. There is a spring in the cable near M2. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form. The cable doesn’t slip on ‘J’. The system below consists of two masses hanging by a cable over mass ‘J’.

30 b) Convert the differential equations to state variable equations θ JM R K s1 K s2 x1 F M2 x2 M1 5.rotation . K d1 F x JM r θ M K s1 K s2 K s3 K d2 6. Write the state equations for the system to relate the applied force ’F’ to the displacement ’x’. For the system pictured below a) write the differential equations (assume small angular deflec- .5.

rotation . x1 8. N 2 Kd1 M3 R2 x3 Ks1 Two gears with fixed centers of rotation and lever arms. The slot drives a lever arm with a suspended mass. A round drum with a slot. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form. A force is applied to a belt over the drum. R1 J1 R2 R3 R4 F1 θ1 Kd1 Ks1 M1 x1 7. N 1 J M 2.5.31 tions) and b) put the equations in state variable form. R1 x2 F1 J M 1. For the system pictured below a) write the differential equations (assume small angular deflec- .

For the system pictured below a) write the differential equations (assume small angular .5. Kd1 Ks1 M1 Ks2 J M 1. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form. Assume that the cable is always tight. It is connected to a round mass that rolls and does not slip. A mass is hung in the middle of the arrangement with springs and dampers on either side.rotation .32 tions) and b) put the equations in state variable form. A pulley system has the bottom pulley anchored.3).J2. x2 θ2 F1 θ3 9. Kd3 Ks3 M3 Ks2 R2.M2 x2 x3 F1 θ1 R1. R 1 M1 x1 A mass slides on a plane with dry kinetic friction (0.J1 10.

and the wall where it is rigidly held.rotation . Forces are applied to the lower side of the moment arm through a spring damper pair.5. The rotational spring is connected between a mass ‘J’. a) Derive the differential equation for the rotational system shown. b) Put the equation in state variable form (using variables) and then plot the position (not velocity) as a function of time for the first 5 seconds with your calcula- . Kd2 Ks2 F1 x3 11. and a mass Kd1 N2 on the other. and also experiences damping ‘B’. There is a rotaJ1 J2 tional damping on one F1 Ks1 θ1 of the gears. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form. both susN1 pended from moment arms. R2 R1 Two gears have a force on one side. How are they related? 13. The mass has an applied torque ‘T’. θ2 M1 x1 12.33 deflections) and b) put the equations in state variable form. M1 x1 Kd1 R1 Ks1 R2 x2 A mass is suspended over a lever arm. Find the polar moments of inertia of area and mass for a round cross section with known radius and mass per unit area.

Assume the system starts at rest.e. Assume the system starts at rest. Find the response as a function of time (i.34 tor using the parameters below. Assume the system starts undeflected and at rest. θ τ = 10Nm τ JM Ks J M = 1Kgm Ns K d = 3 ----m N K s = 9 --m 2 Kd . solve the differential equation to get a function of time. Solve the differential equation to get a function of time.5.rotation . –1 –2 –2 θ'' + ( 1s )θ' + ( 10s )θ = 10s 14.). Nm K s = 10 -------rad J M = 1Kgm 2 Nms B = 1 ---------rad T T = 10Nm θ J Ks B c) A differential equation for the rotating mass with a spring and damper is given below.

 + θ 2  ---------.35 5. θ1 T JM1 K s1 ( θ 1 – θ 2 ) K s2 θ 2 K s1 ( θ 1 – θ2 ) JM2 · Bθ 2 θ2 + ∑ M1 ·· = T – K s1 ( θ 1 – θ 2 ) = J M 1 θ1 T ------J M1 ·· = J M2 θ 2 + K s1 – K s1 ·· - θ 1 + θ 1  ------. = J  J - M1 M1 · ∑ M 2 = Ks1 ( θ1 – θ2 ) – Bθ2 – Ks2 θ2 K s1 + K s2 –K s1 ·· · B θ 2 + θ 2  ------. + θ 1  ---------.5. + g  M  M  1 1 · x2 = v2  – R 2 K d1  – R2 K s1  R2 K s1 – FR1 · v 2 = v 1  ----------------. · x1 = v1 – K s1 K s1 · x 1 = x 1  ---------. + x 1  ----------------- + x 2  -------------- + -----------J1  J1   J1   J1  2 2 2 .7 PRACTICE PROBLEM SOLUTIONS 1. + x 2  ------.rotation . = 0 J   J  J  M2 M2 M2 2. + θ 2  ---------------------.

6 unknowns .rotation .T2.Kdx1 > 0 JM2 T2 · Kd x1 T1 –x1 θ1 = ------R2 x1 θ 2 = ----R1 T 1 = Ks ( x2 – x1 ) + ∑ Fy ∑ M1 ∑ M2 ·· = T 1 – gM1 = – M1 x 2 T1 ·· x 2 = g – -----M1 ·· = – T 1 R 1 + T2 R 1 = – J M 1 θ2 T1 R 1 – T 2 R1 ·· θ 2 = ----------------------------J M1 + + · ·· = T 2 R 2 – R 2 K d x 1 = –J M2 θ 1 T 2 R2 ·· · –R2K d θ 1 + x 1  --------------. = -----------  J JM M2 2 6 equations.5. T1 T2 M1 JM1 gM 1 if T1.36 3.

 = g M   M  2 2 (2) b) · θ = ω  – K s1 – R K s2  RK s2   – RM 1 g – RF · ω = θ  --------------------------------- + x 2  -------------------------- +  --------------------------------  JM + R2 M 1   J M + R 2 M 1  J M + R2 M 1  · x2 = v2 RK s2 – K s2 · v 2 = θ  ----------. + x 2  ---------. θK s1 a) RT JM M1 RK s2 ( x 2 – θR ) F M1 g M2 M2g if(T < 0) T=0 if(T >= 0) Rθ = – x 1 T K s2 ( x 2 – θR ) ∑ F M1 ∑ J ·· = T – M 1 g – F = –M 1 x1 ·· ·· T = – M1 x 1 + M 1 g + F = M 1 Rθ + M 1 g + F ·· M = – RT – θK + RK ( x – θR ) = J θ s1 s2 2 2 M 2 ·· – R ( M 1 g + F ) – θ ( K s1 + R K s2 ) + ( RK s2 )x 2 = ( J M + R M 1 )θ 2 ) K s1 + R K s2 – R K s2 ·· + θ  ----------------------------- + x  -------------------------- = – R ( M1 g + F -------------------------------θ -   2 2  J M + R2 M1   J M + R2 M 1 J M + R M1 ·· = K s2 ( x 2 – θR ) – M2 g = – M 2 x 2 (1) ∑ FM2 K s2 – R K s2 ·· x 2 + x 2  ------. + θ  -------------. + g  M -  M - 2 2 2 .5.rotation .37 4.

38 5. · x1 = v1 – K d1 – K s1 K d1 R 2 R 4 K s1 R 2 R 4 · v 1 = v 1  ----------- + x 1  ---------. + g  M -  M -  M R -  M R - 1 1 1 3 1 3 · θ1 = ω1  – K d1 R2 R4  – K s1 R2 R4 – F1 R1 K d1 R2 R4 K s1 R 2 R 4 · ω 1 = v 1  --------------------- + x 1  -------------------. · θ = ω  – K s1 – r K s2 – K d1 K s2 r Fr · ω = θ  ------------------------------.rotation .  M   M    M 6. + ω 1  ------------------------ + θ 1  ----------------------. + ω 1  --------------------- + θ 1  -------------------.5. + ω  ----------- + x  ---------- + ---- JM   J M  JM JM   · x = v K s2 r – K d2 – K s2 – K s3 · v = θ  ---------- + v  ----------- + x  -------------------------. + -------------- J R   J R  J1  J R2   J R2  1 3 1 3 1 3 1 3 2 2 2 2 2 .

+ J M2 2 N1 τ N2 τ ----N1 x3 θ 2 = ----R2 JM2 θ2 · · R2 K d1 ( x 1 – x 3 ) ·· J M2 θ 2 M3 R 2 K s1 ( x 1 – x 3 ) ·· M3 x 1 M3g · · K d1 ( x 1 – x 3 ) K s1 ( x 1 – x 3 ) .1 N  · 1 ω 2 = ----------------------------------------------------------------------------------------------------------------------------------------------------------------2 N2 J M1 ----. + θ 1  -------------- – g  M   M   M   M  3 3 3 3 · θ2 = ω 2 –N2 2 2 -R v 1 ( R 2 K d1 ) + x 1 ( R2 K s1 ) + ω 2 ( – R2 K d1 ) + θ1 ( – R2 K s1 ) + F  -------.39 7. + x 1  ---------. + ω 2  -------------. θ1 F 1 R1 JM1 ·· J M1 θ 1 θ1 N1 = θ2 N 2 · x1 = v1 R2 K s1 – K d1 – K s1 R2 K d1 · v 1 = v 1  ----------.rotation .5.

· K d1 x 1 K s1 x 1 x1 x1 µ k N ------x1 N = M 1 g cos ( θ 2 ) M1 g sin ( θ 2 ) · x1 = v1 · v1 = · x2 = v2     F1    – K s2  K s2 · v 2 = x 1  ------------------------. + x 3  -------------------------.  M2 + ----------.rotation . M 2 + ----------R1 R2 R1 R2 R1 R 2 9.40 8. + x 2  ------------------------. · x2 = v2 R1 R 2 ( 2F 1 + M2 g )  R 1 R 2 ( – K s2 – K s3 )    R1 R2 K s2 · v 2 = x 3  ----------------------------------------------------------- + x 2  ----------------------------------------------------------- + ---------------------------------------------------------- 4J 1 R 2 + J 2 R2 + R2 R2 M 2  4J 1 R 2 + J 2 R2 + R2 R2 M 2 4J 1 R2 + J 2 R 2 + R 2 R2 M2 · 2 1 1 2 2 1 1 2 2 1 1 2 x3 = v3 – K d3 – K s2 – K s3 K s2 · v 3 = v 3  ----------. + ------------------------J M2 J M2  J M2     M 2 + ----------.5. R 2 M2 x2 θ2 F1 M1 K s2 ( x 1 – x 2 ) . + g  M    M  M 3 3 3 2 2 2 2 2 2 K s2 ( x 1 – x 2 ) ·· M2 x 2 Fc x2 θ2 = ----R1 J M 2. + x 2  ------.

state equations · x1 = v1 R1 + R 2 · v 1 = F1  ----------------. K   K ( R + R ) R K  d1 d1 1 2 1 d1 output equations R1 + R 2 x 2 = ( x 1 – p )  ----------------. + ---------------------------- J 1 N 2 + J 2 N 2  J 1 N 2 + J 2 N 2  J 1 N 2 + J 2 N 2 J 1 N 2 + J 2 N 2 2 1 2 1 2 1 2 1 12.5. R R R 4 R 0 2 2 2 2 For area: J area r = ∫ r dA = ∫ r ( 2πr dr ) = 2π ∫ r dr = 2π --4 2 2 3 0 0 0 πR = --------2 4 M For mass: ρ = M = -----------2 A πR J mass = ∫ r dM = ∫ r ( ρ2πr dr ) = 2πρ ∫ r dr = 2πρ r ---4 2 2 3 0 0 0 R R R 4 R 0 MR = ρ  πR  = --------- --------- 2 2 4 2 The mass moment can be found by multiplying the area moment by the area density. · x1 = v1 – K s1 R2 K s1 N 1 · v 1 = x 1  ---------.  R M  1 1 F1 – K s2 K s2 · q = x 2  ---------. + x 3  -------- + -------K   K  K d1 d2 d2 – K s1 R 1 K s1 R1 + R2 · p = x 1  ---------.  R  1 x3 = – q + x2 11.rotation . + θ 1  --------------------- + g  M   M N  1 1 2 · θ1 = ω1 F1 R1 N2  R2 K s1 N 1 N 2   – K d1 N 1   – R2 K s1 N 1  · ω 1 = x 1  ----------------------------. + ω 1  ----------------------------.41 10. + x 2  -------------------------------- + F 1  ----------------. . + θ 1  ----------------------------.

ω JM J M JM Nm Nms 10 -------1 ---------10Nm rad rad · = ----------------. 10 – -------.5. 10 – -------.K s B· ω = ----.+ θ ----.– -------.– ----.ω  rad rad  .rotation .ω - 2 rad rad  Kgm 10θ –2 s· ω = s  10 – -------.– -------.– -----------------θ – -----------------ω ω 2 2 2 1Kgm 1Kgm 1Kgm 10θ sNm · ω = ------------.= ----JM JM JM ∑M · ·· = T – K s θ – Bθ = J M θ b) θ' = ω 2 1 0 5s T. a) JM · Bθ Ks θ T + ·· · J M θ + Bθ + K s θ = T Ks T·· · Bθ + θ ----.ω 2 rad rad  Kgm Kgm  ----------- m  2 - 10θ s s· ω = --------------------.θ – ----.42 13.– -------.

5 cos ( C 2 ) – 3.5t 27 θ ( t ) = ----.43 (c homogeneous: –1 · –2 ·· θ + ( 1s )θ + ( 10s )θ = 0 · At guess: θ = e At θ h = Ae h A e + ( 1s )Ae + ( 10s )e A + ( 1s )A + 10s –1 2 –1 –2 2 At –1 At –2 ·· 2 At θh = A e At = 0 – 1s ± ( 1s ) – 4 ( 1 ) ( 10s ) A = -----------------------------------------------------------------------------2(1 ) –1 –1 2 –2 = 0 –2 – 1s ± 1s – 40s –1 A = ------------------------------------------------------.123 C 1 cos ( – 0.5s t –1 10s A = -----------.5 ± j3.= 1 –2 10s –2 cos ( 3.rotation .013 e 14.123s 0 + C 2 ) + 1 = 0 cos ( 3.5s C 1 ( 1 ) cos ( C 2 ) – 3.5.5s 0 –1 –1 C 1 cos ( C 2 ) + 1 = 0 θ' ( t ) = – 0.123s t + C 2 ) –1 θ' ( 0 ) = – 0. 10 – 1.013 cos ( – 0.159 ) + 1 = 0 θ ( t ) = – 1.123s t + C 2 ) –1 ·· θp = 0 –2 ( 0 ) + ( 1s ) ( 0 ) + ( 10s ) ( A ) = 10s θp = 1 Initial conditions: θ ( t ) = C1 e – 0.= -----------.123s t + C 2 ) + 1 cos ( 3.= ( – 0.123 sin ( C 2 ) = 0 sin ( C 2 ) –0.5s t –1 –1 sin ( 3.123s t + C 2 ) – 3.524  2  9 – 0.159 ) + 1 –1 .123s C 1 e –1 –1 –1 – 0.t – 0.159 ) cos ( 3.159 –1 C 1 = ----------------------------.5s t –1 –2 cos ( 3.123s C 1 ( 1 ) sin ( C 2 ) = 0 – 0.5------------------.5s t –1 –1 θ( 0) = C1e –1 – 0.123s t – 0.283 )e cos  --------.= – 1.= tan ( C 2 ) cos ( C 2 ) 3.123 )s 2( 1) θh = C 1e particular: θ'' + ( 1s )θ' + ( 10s )θ = 10s · θp = A θp = 0 guess: –1 –2 –1 –2 –2 – 0.5s t –1 C 2 = – 0.5s C 1 e – 0.+ ( – 1.

A torque is applied to one gear. J. On one side is a mass.44 5. 2.rotation .R Ks1 J. the other side a spring damper pair. R2 R1 N1 x2 Kd1 θ1 Ks1 J1 N2 J2 T1 θ2 M1 x1 Two gears have levers attached. Assume the mass remains in contact with the lever.5.R y M Ks2 Kd2 . for the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form. Develop adifferential equation of motion for the system below assuming that the cable always remains tight.8 ASSIGNMENT PROBLEMS 1.

4.45 3. d) Write the equations in state variable matrix form.2 θ=45 Ks2 M2 Kd2 M1 = 1kg y M2 = 1kg R = 0. d) Write the equations in state variable matrix form.R F = 10N M1 Kd1 µk=0. b) Combine the equations in input-output form with y as the output and F as the input.R N K s1 = 100 --m N K s2 = 100 --m Ns K d1 = 50 ----m Ns K d2 = 50 ----m J. Analyze the system pictured below assuming the rope remains tight. J. Analyze the system pictured below assuming the rope remains tight.5. c) Use Runge-Kutta to find the system state after 1 second.R F y M1 Ks1 Kd1 M2 Ks2 Kd2 a) Draw FBDs and write the differential equations for the individual masses. . b) Combine the equations in input-output form with y as the output and F as the input.1m J = 10Kgm 2 Ks1 N K s1 = 100 --m N K s2 = 100 --m Ns K d1 = 50 ----m Ns K d2 = 50 ----m a) Draw FBDs and write the differential equations for the individual masses.rotation .R F = 10N M1 = 1kg M2 = 1kg R = 0.1m J = 10Kgm 2 J. F J.

rotation . .5.46 c) Use Runge-Kutta to find the system state after 1 second.

For a symbolic solution the system of differential equations must be manipulated into a single differential equation. In basic terms the operator can be manipulated as if it is a normal variable. The basic definition of this operator.vibration isolation Objectives: • To be able to develop input-output equations for mechanical systems. The first-order axiom can be used to help solve a first-order differential equation. INPUT-OUTPUT EQUATIONS Topics: • The differential operator.6. For convenience we will add a third. In this book there have been two forms used thus far. .1. dividing by ’D’ results in an integral.2 THE DIFFERENTIAL OPERATOR The differential operator ’d/dt’ can be written in a number of forms. d/dt x and x-dot. solve them numerically or symbolically. In this chapter we will look at methods for manipulating differential equations into useful forms.1 6. input-output equations • Design case . ’D’. 6. 6.1 INTRODUCTION To solve a set of differential equations we have two choices. Multiplying by ’D’ results in a derivative. and related operations are shown in Figure 6.input output equations .

= e –A t ( e y ( t )dt + C ) ∫ D+A Figure 6.( e x ) = e ---.6.at ---.x = D ∫ x dt algebraic manipulation Dx + Dy = D ( x + y ) Dx + Dy = Dy + Dx Dx + ( Dy + Dz ) = ( Dx + Dy ) + Dz simplification n D x = Dn – m x --------m D x(D + a) = x --------------------(D + a) – at at x(t ) -----------------.= Dn ------n dt 1 --.= e  ∫ x ( t )e dt + C   (D + a) first-order axiom Figure 6.x + ae x dt dt –A t ( ∫ e y ( t )dt + C ) De x = e Dx + ae x at at at At y(t) ------------.x = Dx dt n d .2 Proof of the first-order axiom .1 General properties of the differential operator Note: · x + Ax = y ( t ) xD + Ax = y ( t ) e xD + e Ax = e y ( t ) e xD = e y ( t ) e xD = e y ( t ) e x = x = e At At At At At At At At xD + Ax = y ( t ) x(D + A ) = y( t) y( t )x = ------------D+A ∫e At y ( t )dt + C At at d at d.input output equations .2 basic definition d---.

d. This begins by replacing the differential operator and rearranging the equation. Notice that the manipulation follows the normal rules of algebra.2 d ---- x + ---.4.x + 5x = 5t  dt dt D x + Dx + 5x = 5t x ( D + D + 5 ) = 5t 5t x = -------------------------2 D +D+5 5 x = t  --------------------------  2  D +D+5 2 2 Figure 6.6.3 Figure 6.3 contains an example of the manipulation of a differential equation using the ’D’ operator. The solution begins by replacing the ’d/dt’ terms with the ’D’ operator. The first-order axiom is then used to obtain the solution. .input output equations .3 An example of simplification with the differential operator An example of the solution of a first-order differential equation is given in Figure 6. The initial conditions are then used to calculate the coefficient values. After this the equation is rearranged to simplify the expression.

6 + 10. (0)(1 ) – (1 ) x ( 0 ) = ( 1 )  3  ----------------------------- + C = 10     5 C = 10.input output equations .6. These equations can be solved to find a single differential equation that can then be integrated.6 ( t – 1 ) + 10. + 10.5.5t 5t ---.6 x = e – 5t  te – e -------------------.6 3   5 –5 t 5t 5t x = 0.6e x = 0.( te – e ) dt = 5te + e – e 5t 5t 5t 5t guess.5t 5t ---. and these are always shown on the left hand side. and act as the non-homogeneous forcing function. . d---.6e –5 t Figure 6.4 Given. such as that in Figure 6.( te – e ) = 5e t dt Initial conditions. + C     5 5t 5t 5t = 5te – 5t  5t d. The input variables (there can be more than one) are all on the right hand side of the equation.= e 5t t dt ∫ 5 5t 5t te – e  -------------------.4 An example of a solution for a first-order system 6.3 INPUT-OUTPUT EQUATIONS A typical system will be described by more than one differential equation.6t – 0. These equations will have only a single output variable. ( ∫ e 3tdt + C ) te – e 3  -------------------. The basic technique is to arrange the equations into an input-output form.x + 5x = 3t dt Dx + 5x = 3t 3t x = -----------D+5 x = e x = e –5 t x ( 0 ) = 10 d.

input output equations . This begins by replacing the differential operator and combining the equations to eliminate one of the output variables. Given the differential equations.– 2 – 0.5u 2 + u 3 where. · · y 1 = – 3y 1 + 2y 2 + u 1 + 2u 2 · · y 2 = 2y 1 + y 2 + u 1 Find the input-output equations. ··· ·· · · ·· 2y 1 + y 1 + y 1 + 4y 1 = u 1 + u 1 + 3u 2 + u 3 + u 3 ·· · ·· · · y 2 + 6y 2 + y 2 = u 1 + 3u 2 + u 2 + 0.5D y 1 + Dy 1 – 3.5y 1 = Du 1 + 0.5u 1 – Du2 ----------- 2  (2) Dy 2 = 2y 1 + y 2 + Du 1 ∴y 2 ( D – 1 ) = 2y 1 + Du 1 (1) (2) D+3 ∴ y 1  -----------. The solution ends by rearranging the equation to input-output form.5u 1 + u 2' – u 2' 2 2 2 .5y 1 = u 1' + 0.5 e.g.5 Developing input-output equations A sample derivation of an input-output equation from a system of differential equations is given in Figure 6.5u 1 – Du 2 ( D – 1 ) = 2y 1 + Du 1   2   D + 2D – 3 2 ∴y 1  ----------------------------.6.5Du 1 – 0.5u 1 – D u 2 + Du 2 = Du 1   2 ∴0..6. (1) Dy 1 = – 3y 1 + 2y 2 + u 1 + 2Du 2 ∴y 1 ( D + 3 ) = 2y 2 + u 1 + 2Du 2 ∴y 2 = y 1  D + 3 – 0.5u 1' – 0.5u 1 + D u 2 – Du 2 ∴0.5y 1'' + y 1' – 3. y = outputs u = inputs Figure 6. – 0.5Du1 + 0.

7 Drill problem: Find the second equation in the previous example 6.6 for the output y2.6 An input output equation example Find the second equation for the example in Figure 6.6 Figure 6.6. In particular if . Figure 6.8 shows the development of an input-output equation for two freely rolling masses joined by a spring. The example starting in Figure 6. The final equation has a derivative on the right hand side that would prevent it from being analyzed in many cases.input output equations .1 Converting Input-Output Equations to State Equations In some instances we will want to numerically integrate an input-output equation.3.

6. as shown in Figure 6.7 the input force ’F’ was a step function the first derivative would yield an undefined (infinite) value that could not be integrated.9.8 Writing an input-output equation as a differential equation The equation is then converted to state variable form.  K s x1  2 x 1 ( M 1 D + K s ) = F + K s  --------------------------  M2 D 2 + K s x 1 ( ( M1 D + Ks ) ( M2 D + Ks ) – Ks ) = F ( M2 D + Ks ) x 1 ( D M 1 M 2 + D K s ( M 1 + M2 ) + K s – Ks ) = F ( M2 D + K s ) x 1 ( D M 1 M 2 + D K s ( M 1 + M2 ) ) = F ( M2 D + Ks ) d.2  ---- x M M +  ---- x K ( M + M ) =  ---- FM + FK 2 s  dt  dt 1 1 2  dt 1 s 1 2 4 2 2 4 2 2 2 2 2 2 2 2 Figure 6. including a step to calculate a second derivative of the input. .input output equations . x1 F M1 Ks M2 x2 Equations of motion can be derived for these masses.4 d. F M1 Ks ( x2 – x1 ) ∑ Fx ∑ Fx = F + Ks ( x2 – x1 ) = M 1 D x1 2 2 x 1 ( M1 D + Ks ) = F + Ks x 2 K s ( x2 – x1 ) M2 = –Ks ( x 2 – x1 ) = M 2 D x2 2 2 Ks x 1 = x 2 ( M2 D + Ks ) The equations can be combined to eliminate x2.2 d.

6.8 This can then be written in state variable form by creating dummy variables for integrating the function ’F’.input output equations . step.-------------0 d 1 M1 M1 M 2 Figure 6.1 dt a 1 d1 0 1 0 0 = 0 0 0 1 0 –Ks ( M 1 + M 2 ) 0 0 ---------------------------------M 1 M2 0 x 0 0 1 0 0 0 v1 aF + 0 1 0 a1 F 1 .= 1 -------.Ks -----.9 Writing state equations for equations with derivatives . 1  2 2 2 2 2 T2 2 T2 These equations can then be written in matrix form. + F  -------------- ---------------------------------1 F  M M   dt 1 M 1 M2 M 1 1 2 The approximate value of the second derivative of a unit time step can be calulated using the time step. it is turned on) (after the first timestep) (to verify) 11 v F = TaF 0 + Ta F1 = T  ----- + T  – ---. 1 a F 0 = ----2 T 1a F1 = – ---2 T (when the timestep. 1  ----. d ---- x = v 1  dt 1 d  ---- v = a 1  dt 1 d ---- a = d 1  dt 1 d ---- d M M + a K ( M + M ) = a M + FK 1 s 1 2 F 2 s  dt 1 1 2 Ks d1 ---- d = a  – K s ( M 1 + M 2 ) + a  -----.+ -. = 0  2  2 T T 2 2 2 2 1 1 T .x F = -----a F + T -a F = T  ----- + T  ----- = -. x1 d.v ---.

Guess.10 and Figure 6. · ·· · x + 3x + 2x = 4F + 5F The solution begins by evaluating the homogeneous equation as normal.1  dt 0  dt  d---.1 + 5 ( 0 )  dt 0  dt  dd ---- x = 4  ---.3. and the right hand side is the particular part.x 0 = 4 dt Note: This will be used as an initial condition 2 2 0 + 3 ( 0 ) + 2A = 2 ( 0 ) + 5 ( 1 ) for a step function 1· u ( t ) = ---dt Figure 6. with the left hand side being the homogeneous part.input output equations . The solution begins by evaluating the homogeneous equation.= – 1.2  ---- x + 3  ---- x + 2x = 4  ---. x = A · x = 0 ·· x = 0 2 2 5 A = -2 The derivative in the non-homogeneous solution must now be used to find the initial conditions. ·· · x + 3x + 2x = 0 A + 3A + 2 = 0 –3± 3 – 4 ( 2 ) A = -----------------------------------.11.9 6.2 Integrating Input-Output Equations An input-output equation is already in a form suitable for normal integration techniques. However the initial position and velocity are known to be zero. ddd. assuming F is is unit step function. these determine the values of initial conditions.6. If the non homogeneous part includes derivatives. –2 2 –t – 2t xh = C1 e + C2 e The particular solution can also be found as normal. An example of explicitly solving such an equation is shown in Figure 6.10 Integrating an input-output equation .F  + 5F 0 0  dt 0  dt 1  dt 0 dd ---- x + 3 ( 0 ) + 2 ( 0 ) = 4  ---.

In this case the initial acceleration is found to be non-zero.6.= 0 2 5 C 1 = – C 2 – -2 + C2 e –0 –t – 2t · x ( t ) = –C 1 e – 2C 2 e · x ( 0 ) = – C 1 – 2C 2 = 4 5 –  – C 2 – -- – 2C 2 = 4  2 3 C 2 = – -2 3 5 C 1 = –  – -- – -.10 The initial conditions can then be used to find the values of the coefficients. In practical terms.12.= – 1  2 2 The final equation can then be written. .11 Integrating an input-output equation (cont’d) The example in Figure 6. It will be assumed that the system starts undeflected and at rest. x(t ) = – e – 3 e -2 –t – 2t 5 + -2 Figure 6.input output equations . x ( t ) = C1 e + C2 e x ( 0 ) = C1 e –0 –t – 2t 5 + -2 5 + -. this can be ignored because only the initial position and velocity will be used to find the coefficients.10 is reconsidered with a sinusoidal input in Figure 6.

Figure 6.11 Assume a sinusoidal input.6.input output equations . The transfer function is given in two different forms because the system is reversible and the output could be either ’F’ or ’x’. and the equation is manipulated into transfer function form.12 Initial conditions for a sinusoidal input 6. but it will not affect the initial position or velocity. In this example the forces are summed to provide an equation. with the system initially at rest.2  ---- x + 3  ---- x + 2x = 4 ( F ( 0 ) ) + 5F ( 0 ) 0  dt 0  dt 0 d ---- x + 3 ( 0 ) + 2 ( 0 ) = 4 ( 1 ) + 5 ( 0 )  dt 0 d ---- x = 4  dt 0 2 2 This indicates that the acceleration will have an initial value. .13. F ( t ) = sin ( t ) F( 0) = 0 · F ( t ) = cos ( t ) · F(0) = 1 · dd. The differential operator is replaced.4 DESIGN CASE The classic mass-spring-damper system is shown in Figure 6.

We know this could easily be reversed mathematically and practically. others are not (consider a internal combustion engine. . Also keep in mind that if we were given a force applied to the system it would become the input (action force) and the output would be the displacement (resulting motion).+ Ks x = –M -------2 dt dt d x dx F = M ------.+ K s x 2 dt dt x F = MD x + K d Dx + K s x F -. Numerical values are substituted and the homogeneous solution to the equation is found. For the example above ‘F’ over ‘x’ implies that we are changing the input ‘x’. and there is some change in ‘F’. An input is typically something we can change.= ----------------------------------------2 F MD + K d D + K s 2 2 2 F Aside: An important concept that is ubiquitous yet largely unrecognized is the use of functional design.+ K d ----. A system experiencing a sinusoidal oscillating force is given in Figure 6. Some systems (such a mechanisms) are reversible. When shown with differentials it is obvious that the ratio is not simple.input output equations . an output is the resulting change in a system.14.6. The first stage of such analysis involves finding the actual displacement for a given displacement or force.= MD 2 x + K d Dx + K s x x OR x 1 -. and is a function of time. Mass-spring-damper systems are often used when doing vibration analysis and design work.12 Kd M Ks d x dx ∑ F y = – F + Kd ----.13 A transfer function for a mechanical system Aside: Keep in mind that the mathematical expression ‘F/x’ is a ratio between input (displacement action) and output (reaction force). We look at parts of systems as self contained modules that use inputs to produce outputs. turning the crank does not produce gasoline). To do this all we need to do is flip the numerators and denominators in the transfer function. Figure 6.

6. ----.+  0.25 – 8 )s A = ( – 0.5 ( – 0.----.5 ----m F = 5 sin ( 6t ) N The differential equation for the mass-spring damper system can be written.  dt  m 2 m dt The homogeneous solution can be determined.15 where the particular solution is found and put in phase shift form.25t –1 –1 cos ( 1.5 Ns dx +  2 --- x = 5 sin ( 6t ) N ----. . 2 d x Ns dx N .392j )s xh = C1 e Figure 6.5 ----.±  0.5 ± 0.392t + C 2 ) Explicit analysis of a mechanical system The solution continues in Figure 6. M = 1Kg N K s = 2 --m Ns K d = 0.+  2 --.input output equations .1Kg ------.13 Given the component values input force. 2 d x N 1Kg ------.+  0. – 4 ( 1Kg )  2 ---    m m m A = -----------------------------------------------------------------------------------------2 ( 1Kg ) A = 0.5 ----. x = 0   dt  m 2 m dt Ns Ns 2 N – 0.14 – 0.5 ----.25 ± 1.

15 Explicit analysis of a mechanical system (continued) The system is assumed to be at rest initially.1460 ) sin 6t + ( – 0.1466 sin  6t + atan  --------------------.6. and this is used to find the constants in the homogeneous solution in Figure 6. The steady-state force at 6rad/s will have a magnitude of . ----.1466 ( –0.08788 cos 6t ) – 0.01288 ) cos 6t ) 2 2 ( – 0. In this case there are two force frequency components at 1.16.1460 ) sin 6t + ( – 0.( – 0.01288 3 – 34 ( 34 ) ------------------.1460 ) + ( – 0.+  0.01288 ) x p = -----------------------------------------------------------------. The transient effects have a time constant of 4 seconds (1/0.5 ( 6A cos 6t – 6B sin 6t ) + 2 ( A sin 6t + B cos 6t ) = 5 sin ( 6t ) – 36B + 3A + 2B = 0 A = 34 B ----3 – 36A – 3B + 2A = 5 34 – 34  ----. 2 d x Ns dx N .input output equations .( ( –0.08788  x p = 0.    – 0.1460 ) + ( – 0.B – 3B = 5 3  5 34 A = ----.392rad/s and 6rad/s.483 ) 2 2 Figure 6.= – 0. .5 ----.01288 ) x p = 0.9961x p = 0. Finally the displacement of the mass is used to find the force exerted through the spring on the ground.2932N.+  2 --- x = 5 sin ( 6t ) N  2 m  dt  m dt x p = A sin 6t + B cos 6t x p' = 6A cos 6t – 6B sin 6t x p'' = – 36 A sin 6t – 36B cos 6t – 36 A sin 6t – 36B cos 6t + 0.01288 ) = – 0.– 3 3 x p = ( –0.1460 B = ---------------------------.14 The particular solution can now be found with a guess.1466 sin ( 6t + 1.01288 ) cos 6t ( – 0. and should be negligible within a few seconds of starting the machine.9961 sin 6t –0.25).1Kg ------.

483 ) )m The displacement can then be used to calclulate the force transmitted to the ground.392 ( C 1 e sin ( 1.07713  cos ( C )   cos ( C )  2 2 0 = 0.483 ) ) The initial conditions can be used to find the unknown constants.3346 e – 0.392t – 0.392 ( C 1 sin ( C 2 ) ) + 0.483 ) – 0.25t cos ( 1. This can be done by adding a mass-spring isolator.392t + C 2 ) – 1.1460 – 0.07713 0.25t N F =  2 --.2932 sin ( 6t + 1.5099 ) x = ( – 0.1673 e cos ( 1. as shown in Figure 6.483 ) )N Figure 6.input output equations .1466 sin ( 6t + 1.25C 1 e cos ( 0 + C 2 ) – 1.1466 cos ( 6t + 1. In the figure the bottom mass-spring-damper combination is the original system.392t – 0. 0 = C 1 e cos ( 0 + C 2 ) + 0.= – 0.483 ) C 1 cos ( C 2 ) = – 0.6.1673 e – 0.16 Explicit analysis of a mechanical system (continued) A decision has been made to reduce the vibration magnitude transmitted to the ground to 0.1466 sin ( 0 + 1.392 ( C 1 e sin ( 0 + C 2 ) ) + 6 ( 0.483 ) ) 0 = –0.392t + C 2 ) + 0. F = Ksx – 0.25t 0 0 0 cos ( 1. assuming the spring is massless.392t – 0.25C 1 e – 0.5099 ) + 0.cos ( C 2 ) – 1. The . x = xh + xp = C1 e x' = – 0. ( – 0.1466 sin ( 6t + 1.17.0365 + 0.25t cos ( 1.07713 C 2 = atan  ----------------------------------------.1466 sin ( 6t + 1.392t + C 2 ) ) + 6 ( 0.1N.25t – 0.0365 + ( 0. = – 0.2032 ) tan ( C 2 ) + 0.2032 – 0.5099 ) + 0.1460 C 1 = -------------------------------.sin ( C 2 ) + 0.1460 0 = – 0.5099   – 0.1460 C 1 = ------------------cos ( C 2 ) 0 = – 0.25C 1 cos ( C 2 ) – 1.15 The particular and homogeneous solutions can now be combined.25t cos ( 1.5099 ) + 0.1466 cos ( 0 + 1.1460 – 0.392  ------------------.07713 – 0.25  ------------------.483 ) )m  m F = ( – 0.1673 cos ( – 0.

The design begins by developing the differential equations for both masses.18. K s2 M2 x2 Kd K s1 M1 F – K s2 x 2 M2 · · Kd ( x2 – x1 ) K s1 ( x 2 – x 1 ) x1 ∑F · · ·· = – K s2 x 2 – K d ( x 2 – x 1 ) – K s1 ( x 2 – x 1 ) = M2 x 2 2 – K s2 x 2 – K d ( x 2 D – x 1 D ) – K s1 ( x 2 – x 1 ) = M 2 x 2 D 2 x 2 ( – K s2 – K d D – K s1 – M2 D ) = x 1 ( – K d D – K s1 )  K s2 + K d D + K s1 + M2 D  x 1 = x 2  ------------------------------------------------------------- K d D + K s1   2 (1) · · Kd ( x2 – x1 ) M1 F K s1 ( x 2 – x 1 ) ∑F · · ·· = K d ( x 2 – x 1 ) + K s1 ( x 2 – x 1 ) – F = M1 x 1 2 K d ( x 2 D – x 1 D ) + K s1 ( x 2 – x 1 ) – F = M1 x 1 D 2 x 1 ( – K d D – K s1 – M 1 D ) + x 2 ( K d D + K s1 ) = F (2) Figure 6.6. Values must be selected for the mass and spring.16 mass and spring above have been added to reduce the vibration that will reach the ground. as it determines the force on the ground. The .input output equations .17 Vibration isolation system For the design we are only interested in the upper spring. An input-output equation for that spring is developed in Figure 6.

input output equations .5 ----m M 2 = 1Kg  K s2 + 0.5D + 2 + K s2 ) ( D – 0.17 given values for the mass-spring-damper system are used.6. In addition a value for the upper mass is selected. We can also limit the problem by selecting a mass value for the upper mass. This is arbitrarily chosen to be the same as the lower mass.5D + 2 ) = F 0.2 ddd.5D + 2 ) x 2 ( D ( – 1 ) + D ( K s2 ) + D ( – 0.5K s2  ---- x 2 – 2K s2 x 2 = 0.2 dd–  ---- x + K s2  ---- x 2 – 0.4 d.5K s2  ---- x 2 – 2K s2 x 2 = 0. The solution begins by combining equations (1) and (2) and inserting the.5  ---- 5 sin ( 6t ) + 2 ( 5 ) sin ( 6t )  dt 2  dt  dt  dt d. This can then be used to find the needed spring coefficient.18 Developing an input output equation 4 2 1 2 2 2 This particular solution of the differential equation will yield the steady-state displacement of the upper mass.5D – 2 ) + x 2 ( 0. ( – 0.4 d.5D + 2 ) = F ( 0.4 –  ---- x + K s2  ---- x 2 – 0.numerical values for the lower mass.2 d–  ---- x + K s2  ---- x 2 – 0.5K s2  ---- x 2 – 2K s2 x 2 = 15 cos ( 6t ) + 10 sin ( 6t )  dt 2  dt  dt Figure 6. . spring and damper.5  ---- F + 2F  dt 2  dt  dt  dt d.5K s2 ) + ( – 2K s2 ) ) = F ( 0.5D – 2 – D ) + x 2 ( 0.5D + 2 + D  2 x 2  -----------------------------------------------. d. This choice may need to be changed later if the resulting spring constant is not practical.5D + 2 ) This can now be converted back to a differential equation and combined with the force.  K s2 + K d D + K s1 + M2 D  2 x 2  ------------------------------------------------------------- ( – K d D – K s1 – M1 D ) + x 2 ( K d D + K s1 ) = F K d D + K s1   M 1 = 1Kg N K s1 = 2 --m 2 2 Ns K d = 0.5D + 2   x 2 ( D + 0.

5 )K s2 6A – 2K s2 B ) = 15 cos ( 6t ) A ( – 3K s2 ) + B ( – 1296 – 38K s2 ) = 15 10 + A ( 1296 + 38K s2 ) A ( – 3K s2 ) + ----------------------------------------------------.input output equations .– 10 ( – 1296 – 38K s2 ) A = -------------------------------------------------------------------------------------2 –9K s2 ----------------------------------------.2 d–  ---- x + K s2  ---- x 2 – 0.+ ( 1296 + 38K s2 ) ( – 1296 – 38K s2 ) 45K s2 + 10 ( 1296 + 38K s2 ) A = -----------------------------------------------------------------------------------------------------2 – 9K s2 + ( 1296 + 38K s2 ) ( – 1296 – 38K s2 ) 425K 2s + 12960 A = --------------------------------------------------------------------------------2 – 1453K 2s – 98496K 2s – 1679616 2 2 Figure 6.( – 1296 – 38K s2 ) = 15 3K s2 A ( – 9K s2 ) + ( 10 + A ( 1296 + 38K s2 ) ) ( – 1296 – 38K s2 ) = 45K s2 45K s2   – 9K s2 A  ----------------------------------------- + A ( 1296 + 38K s2 ) = ----------------------------------------.4 d.3  ---- x = – 216 A cos 6t + 216B sin 6t  dt p d 4  ---- x = 1296A sin 6t + 1296B cos 6t  dt p sin ( 6t ) ( – 1296A – 36AK s2 + 0.5K s2  ---- x 2 – 2K s2 x 2 = 15 cos ( 6t ) + 10 sin ( 6t )  dt 2  dt  dt x p = A sin 6t + B cos 6t d ---- x = 6A cos 6t – 6B sin 6t  dt p d.6.19 Finding the particular solution .2  ---- x = – 36 A sin 6t – 36B cos 6t  dt p d.18 The particular solution begins with a guess.– 10 ( – 1296 – 38K s2 )  ( – 1296 – 38K s2 ) 45K s2 ----------------------------------------.5K s2 6B – 2K s2 A ) = 10 sin ( 6t ) A ( – 1296 – 38K s2 ) + B ( 3K s2 ) = 10 10 + A ( 1296 + 38K s2 ) B = ----------------------------------------------------3K s2 cos ( 6t ) ( – 1296B – 36BK s2 + ( –0. d.

20 Finding the particular solution (cont’d) Finally the magnitude of the particular solution is calculated and set to the desired amplitude of 0. This is then used to calculate the spring coefficient.097563 ×10 K 2s B = -------------------------------------------------------------------------------------------------2 3K s2 ( – 1453K 2s – 98496K 2s – 1679616 ) 540K 2s + 69918768 B = --------------------------------------------------------------------------------2 – 1453K 2s – 98496K 2s – 1679616 2 8 2 Figure 6. . amplitude = 0. +  ---------------------------------------------------------------------------------  – 1453K 2 – 98496K 2s – 1679616  – 1453K 2 – 98496K 2s – 1679616 2s 2s A value for the spring coefficient was then found using Mathcad to get a value of 662N/m. ( 1296 + 38K s2 )  – 1453K 2 – 98496K 2s – 1679616 2s B = -----------------------------------------------------------------------------------------------------------------------------------------3K s2 10 ( – 1453K 2s – 98496K 2s – 1679616 ) + ( 425K 2s + 12960 ) ( 1296 + 38K s2 ) B = ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------2 3K s2 ( – 1453K 2s – 98496K 2s – 1679616 ) 1620K 2s + 2.1 = A +B 2 2  2  2 425K 2s + 12960 540K 2s + 69918768  --------------------------------------------------------------------------------.6.input output equations .1N.19 The value for B can then be found.   425K 2s + 12960 10 +  --------------------------------------------------------------------------------.

20 Figure 6. A force is applied to cause the mass the accelerate.input output equations .6. Find the input-output form for the following equations.21 Calculation of the spring coefficient 6. There is viscous damping between the block and the ground. Use Cramer’s rule to find an input-output equation for ‘y’. Develop the input-output equation for the mechanical system below. ·· · · x 1 + x 1 + 2x 1 – x 2 – x 2 = 0 · ·· · – x 1 – x1 + x2 + x2 + x2 = F 4. F·· y + 2x = ----10 · ·· 7y + 4y + 9x + 3x = 0 Fy = ----10 2 ( 7D + 4 ) ( 9D + 3 ) x 0 (D ) 2 (2) . · y+y+x = 3 · x+x+y = 0 3.5 SUMMARY • The differential operator can be manipulated algebraically • Equations can be manipulated into input-output forms and solved as normal differential equations 6. The following differential equations were converted to the matrix form shown. Find the input-output form for the following equations. x F M B 2.6 PRACTICE PROBLEMS 1.

Ignore the effects of gravity. Find the input-output equations for the systems below. Find the input output equation for y2. Write the input-output equations for the mechanical system below. and the output is ‘y’ or the angle theta (give both equations). and . The input is force ‘F’. Here the input is the torque on the left hand side. K s1 M1 y1 K s2 M2 F y2 6. Include the inertia of both masses.6. θ1 θ2 τ K s1 B1 J1 K s2 J2 B2 7.input output equations .21 5.

The applied force ‘F’ is the input to the system.6. 6. b) Using numerical methods.input output equations . and the output is the displacement ‘x’. given F(t) = 10N for t >= 0 seconds. find the steady-state response for an applied force of F(t) = 10cos(t + 1) N ? a) Solve the differential equation to find the explicit response for an applied force of F(t) = 10cos(t + 1) N ? d) Set the acceleration to zero and find an approximate solution for an applied force of F(t) = 10cos(t + 1) N. K1 = 500 N/m K2 = 1000 N/m M = 10kg x w F a) Find x(t).22 gravity for mass ‘M’. Compare the solution to the previous solutions. K R JM θ K y M F 8. .7 PRACTICE PROBLEM SOLUTIONS 1.

3 )  dt  dt  dt  dt d.3 ) d.4 d. F----10 (2) 2 F2 ----.6.2 ------------------------------------------------------------ ---- y +  ---- y  K s1 M 2 + K s2 M 2 + K s2 M1 + y  K s1 K s2 = 2  ---------------.2 d ---- x + 2  ---- x + 3  ---- x +  ---- x + x =  ---- F +  ---- F + 2F 2  dt 2  dt 2  dt 2  dt 2  dt  dt 4.4 d ---- x + 2  ---- x + 3  ---- x +  ---- x + x =  ---- F + F 1  dt 1  dt 1  dt 1  dt 1  dt d. ·· · x + 2x = – 3 ·· · y + 2y = 3 3.input output equations .= --------------------------------------------------2 2 4 2 D ( 9D + 3 ) – 2 ( 7D + 4 ) 9D + 3D – 14D – 8 2 (D ) (2) ( 7D + 4 ) ( 9D + 3 ) y ( 9D + 3D – 14D – 8 ) = F ( 0.  dt 2  dt 2   M M M M 1 2 1 2 4 2 2 2 M1 d.2 dd. d.4 d.1 -------d.4 d.= -------------------------------------------------------------.2 d.2 dd.2  ---- F  -------------.1 d.( 9D + 3 ) 2 0 ( 9D + 3 ) F ( 0.3 ) 10 y = ------------------------------------------------------.3 d. + F  K s1 + K s2 --------------------- dt  M M   M M  1 2 1 2 .3 d.2 1 ---- y +  ---- y  1 +  ---- y  –14 + y  – 8 =  ---- F  ----- + F  1 ----- dt  dt  3  dt  9   9  dt  10  3 5.23 F ·· · B x + x  ---- = --- M M 2.2 d. d.9D + 0.9 ) + F ( 0.4  ---- y ( 9 ) +  ---- y ( 3 ) +  ---- y ( – 14 ) + y ( – 8 ) =  ---- F ( 0.9D + 0.2 -d.

input output equations .3 d.  dt J 1 J2 J1 J2 d. + τ  --------. 1  ----------------------------------.  dt  dt  J   dt  J J  J J  J J 1 2 1 1 2 1 2 .2  ---- θ +  ---- θ  J 1 B2 + J 2 B 1 +  ---- θ  J 1 K s2 + J 2 K s2 + B1 B2 + ---------------------------------------------------1  - 1  ---------------------------.24 6. d.6.4 d 3 d.2  ---- θ +  ---- θ  J 1 B2 + J 2 B 1 +  ---- θ  J 1 K s2 + J 2 K s2 + B1 B2 + ------------------------------------------------------------------------------ dt 2  dt 2    dt 2   J J J J 1 2 1 2 K s2 d ---- θ  B 1 K s2 + B 2 K s2  = τ --------2  ----------------------------------.2 1d ---- θ  B 1 K s2 + B 2 K s2 =  ---- τ  ---- +  ---- τ  --------.  dt  dt 1   dt J J J J 1 2 1 2 B2 K s2 dd.4 d.

4 ---------- ---- θ  J M M +  ---- θ ( 2R 2 M + J ) + θ ( K 2 R2 ) = FKR + MgKR M  dt  KR   dt 3 2 2 3 3 2 2 K R d.4 ------------.2 ------------.+ K2R + ------------.4 d.– R K = K K   2 2  2 JM D   2 JM D  – F  2R + ------------- – Mg  2R + ------------- K  K    4 J M 2 J M ddM  ---- y  ----------- +  ---- y  2R2 M + ----.------ ---- θ +  ---- θ  2R K + KR + θ  ------------ = F K R + MgK R - ------------------------------   dt  dt  J M M JM M JM M   JM M  for the y output equation.25 7.+ y ( R 2 K ) =  dt  K -  dt  K d. 2 2 θ ( KR ) + y ( MD + K ) = – F – Mg 2  2R K + J M D  2 θ ( KR ) + θ  --------------------------------- ( MD + K ) = – F – Mg – RK   2 2 2 2 2 θ ( – K R ) + θ ( 2R K + JM D ) ( MD + K ) = FKR + MgKR θ ( – K R + 2R MKD + 2R K + J M MD + J M KD ) = FKR + MgKR d. θ JM R ( K ( – y – θR ) ) y ∑M ·· = – R ( K ( θR ) ) + R ( K ( – y – θR ) ) = J M θ ·· R Kθ + RKy + R Kθ = – J M θ 2 2 R ( K ( θR ) )  2R K + J M D  θ  --------------------------------- = y – RK   2 2 K ( – y – θR ) M ∑F ·· = K ( – y – θR ) – F – Mg = My K ( y + θR ) + F + Mg = – M yD 2 2 θ ( KR ) + F + Mg = y ( – MD – K ) F+Mg for the theta output equation.   2 – RK y  --------------------------------- ( KR ) + y ( MD + K ) = – F – Mg  2R2 K + J M D 2  2 2 JM D M 2 JM D 2  y  2R MD + -----------------.   dt  M   JM M   JM  4 2 2 2 2 2 2 2 4 2 .2 – J M  ---- F  --------- + F ( – 2 R2 ) + ( – 2 MgR2 )  dt  K  2 2 2 d.2 ---- ---- F  –1 + F  –2 KR - +  – 2 gKR - ----------------  ------------------.input output equations .6.4 d.4 d.1 ---- y +  ---- y  2KR + ---- + y  R K  =   ----------- dt  dt  J M M   JM M  2 2 d.

1962 x ( t ) = 0.26 8. · y + y + 5x = 3 · x+x+y = 7 2.1962 6.8 ASSGINMENT PROBLEMS 1.input output equations . K s1 M1 y1 N K s1 = 100 --m M1 = 1Kg N K s1 = 1000 --m K s2 M2 F y2 M 2 = 0.6.  M ( K + K )  M ( K + K ) K + K  1 2 1 2 1 2 a) b) c) x ( t ) = –0. b) numerical analysis.2168 cos ( 5.774t ) + 0. = F  ---------------------------. For the system pictured below find the output response as a function of time for y1 and y2 using a) integration to find an explicit function.02 cos ( t + 1 ) + 0.2162 x ( t ) = – 0. Find the input-output equation for the differential equation below if both ’x’ and ’y’ are outputs.774t ) + 0.02 cos ( t + 1 ) + 0.1Kg . + g  -----------------.2168 cos ( 5. K 1 K2 K2 K2 ·· x + x  ---------------------------.

. and Frederick.. second edition. Close.K. J. . 1979.R. and Graf. D. “Modeling and Analysis of Dynamic Systems. C.D. John Wiley and Sons. 1995.27 6. Industrial Noise and Vibration Control.input output equations .9 REFERENCES Irwin. E. Inc.M.. Prentice Hall Publishers..6.

capacitors. ∑ I node = 0 ∑ V loop = 0 Figure 7.1 7.7. The more freely the electrons can flow. The node current law holds true because the current flow in and out of a node must total zero.1 INTRODUCTION A voltage is a pull or push acting on electrons. If the sum of currents was not zero then electrons would be appearing and disappearing at that node. inductors and op-amps • Device impedance • Example circuits Objectives: • To apply analysis techniques to circuits 7. The loop voltage law states that the sum of all rises and drops around a loop must total zero. power sources.2 MODELING Kirchoff’s voltage and current laws are shown in Figure 7. ELECTRICAL SYSTEMS Topics: • Basic components. The voltage will produce a current when the electrons can flow through a conductor. In AC circuit analysis we consider the steady-state response to a . resistors.circuits .1. Most electrical components are used to control this flow. the lower the resistance of a material. 7. typically only requiring algebraic manipulation.1 node current loop voltage Kirchoff’s laws The simplest form of circuit analysis is for DC circuits. thus violating the law of conservation of matter.

In this circuit an input voltage is supplied on the left hand side.deliver power to a circuit • capacitors .7. such as resistors. often requiring integration.find the response for a constant input.2 sinusoidal input. To prove this the cur- . capacitors and inductors. If the output resistance is very large.2 Ohm’s law The voltage divider example in Figure 7. There is a wide range of components used in circuits. these block high frequencies • op-amps .2.pass current based on current flow. A resistor will oppose current flow as described by ohm’s law in Figure 7. I + V R V I = -R V = IR Figure 7. Finally the most complex is transient analysis. and the ratio of output to input voltages is determined by the ratio of the resistance between R1 and R2. The output voltage on the right hand side will be some fraction of the input voltage.find the initial response to changes.resist changes in current flow.circuits .3 illustrates the methods for analysis of circuits using resistors.reduce current flow as described with ohm’s law • voltage/current sources . • resistors .very high gain amplifiers useful in many forms 7. • Transient .2. these block DC currents • inductors . The simplest components are passive. • DC (Direct Current) . no current will flow. • AC (Alternating Current) . or similar techniques. A list of components that will be discussed in this chapter are listed below. Active components are capable of changing their behaviors. such as op-amps and transistors. but in actuality it varies with conductor temperature. The resistance value is assumed to be linear.find the steady-state response to an AC input.1 Resistors Resistance is a natural phenomenon found in all materials except superconductors.

 = 0  R   R  1 2 Assume the output resistance is large.3 A voltage divider circuit If two resistors are in parallel or series they can be replaced with a single equivalent resistance. so I2 is negligible.4. = 0  R  R  1 2 1. I 1 + I 2 + I3 = 0 + R1 Vi I2 R2 I3 + Vo I1 0 – Vo Vi – Vo ∴ ---------------. Vi – Vo – Vo ∴ ---------------. +  -------.3 rents into the center node are summed and set equal to zero.= ----------------R 1 + R2 Vi Figure 7.+ ----- = V i  -----  R1 R 2 R  1 R1 + R 2 1 ∴V o  ----------------. = V i  -----  R R  R - 1 2 1 R2 Vo ∴----.circuits . . + I 2 +  -------------. as shown in Figure 7. The equations are then manipulated to produce the final relationship.11∴V o  ----.7.

The symbol with two horizontal lines is a battery.4 R1 series resistors R eq = R1 + R2 R2 parallel resistors R1 R2 1 1 1 ------. . with the symbols indicating polarity.----1----.5.+ R1 R2 R1 R2 Req = ----------------R1 + R2 Figure 7. A current source will supply a current to a circuit.circuits . The last symbol with an arrow inside the circle is a current supply. The circle with a sine wave is an AC voltage supply. by varying the current as required. The supplies with ’+’ and ’-’ symbols provide DC voltages. The schematic symbols for voltage and current sources are shown in Figure 7. The arrow indicates the direction of positive current flow.7.2 Voltage and Current Sources A voltage source will maintain a voltage in a circuit.4 Equivalent resistances for resistors in parallel and series 7.2. by varying the voltage as required.= ----.+ ----Req R1 R2 1 Req = -----------------1.

5 + V - + V - V+ - V I Figure 7.+ ----.11Vo  ----.circuits .7. Find the output voltage Vo. and in as negative.+ ----. R1 + Vi + - R2 R3 Vo - Examining the circuit there are two loops. so the node current methods is the most suitable for calculations.+ ----.5 Voltage and current sources A circuit containing a voltage source and resistors is shown in Figure 7.1.= 0 R1 R2 R3 1. = V i  ------  R R R   R1 1 2 3 R2R3 Vo = V i  ------------------------------  R + R + R - 1 2 3 Figure 7. will be solved. This will reduce the chances of careless mistakes. The circuit is analyzed using the node voltage method.6.6 A circuit calculation Aside: when doing node-current methods. The currents into the upper right node. select currents out of a node as positive.Vo. = V i  ------  R 1 R2 R3  R1 R 1 + R2 + R3 1Vo  -----------------------------. ∑I Vo – V Vo Vo = ----------------i + ----. . but only one node.

7 Drill problem: Mesh solution of voltage divider .6 using the loop voltage method. R2 R3 Vo = Vi  ------------------------------------------------. R R + R R + R R  1 2 1 3 2 3 Figure 7.7.6 Evaluate the circuit in Figure 7.circuits .

7 Evaluate the voltage divider in Figure 7.8 Drill problem: Mesh solution of voltage divider Dependant (variable) current and voltage sources are shown in Figure 7.9 Dependant voltage sources .circuits . while the current source has an arrow inside.3 using the loop current method. The voltage and current values of these supplies are determined by their relationship to some other circuit voltage or current.9.7. Hint: Put a voltage supply on the left. and an output resistor on the right. The dependant voltage source will be accompanied by a ’+’ and ’-’ symbol. Remember that the output resistance should be infinite. V = f( ) + - I = f( ) Figure 7. Figure 7.

if Ii = 1A.8 Ii 2ohm + V1 - + .3 Capacitors Capacitors are composed of two isolated metal plates very close together.10 Drill problem: Find the currents in the circuit above 7.circuits . But. electrons will be forced into one plate. Temporarily this creates a small current flow until the plates reach equilibrium. high frequency AC currents will pass through the device. When a voltage is applied across the capacitor. So. . except for transient effects. The equation for a capacitor and schematic symbols are given in Figure 7. f(V1) = 3V1 What if the input current is Ii = 1sin(2t)A? Figure 7. any voltage change will result in some current flow.11. In practical terms this means that the capacitor will block any DC voltages. and forced out of the other plate.2.V =f(V ) 2 1 3ohm Io Find the output current.7. Io.

circuits .9 + + C C V I dI = C  ---- V = CDV  dt Figure 7. Find the current as a function of time.12 Drill problem: Current through a capacitor .) The other capacitor symbol is for a regular capacitor. fail and possibly explode. These contain a special fluid that increases the effective capacitance of the device but requires that the positive and negative sides must be observed in the circuit.11 Capacitors The symbol on the left is for an electrolytic capacitor. normally with values under a microfarad. (Warning: reversing the polarity on an electrolytic capacitor can make them leak.7. + V= 5cos(10t)V C=1uF I Figure 7.

The core can be hollow. designers normally try to avoid using inductors in circuits.13 An inductor An inductor is normally constructed by wrapping wire in loops about a core. When possible.. Inductors usually cost more than capacitors. The schematic symbol and relationship for an inductor are shown in Figure 7.7. When a current flows through the coils. .4 Inductors While a capacitor will block a DC current. otherwise the field is maintained without effort (i. no voltage).13. an inductor will pass it. + L V I dV = L  ---- I = LDI  dt Figure 7.2.circuits . Inductors are basically coils of wire. or be made of ferrite to increase the inductance. Therefore the inductor resists changes in the current. If the current through the inductor changes then the magnetic field must change. In addition. inductors are susceptible to interference when metals or other objects disturb their magnetic fields. a magnetic field is generated.e.10 7.

On the left hand side are the inverting and non-inverting inputs.increase the amplitude of a signal • impedance isolators . and dividers . In circuits op-amps are used with feedback to perform standard operations such as those listed below. multipliers. Both of these inputs are assumed to have infinite impedance. • adders. + V= 5cos(10t)V L=1mH I Figure 7.hide the resistance of a circuit while passing a voltage . subtractors.5 Op-Amps The ideal model of an op-amp is shown in Figure 7.15.simple analog math operations • amplifiers .2.14 Drill problem: Current through an inductor 7.11 Find the current as a function of time.circuits . Op-amp application circuits are designed so that the inverting and non-inverting inputs are driven to the same voltage level.7. The output of the op-amp is shown on the right. and so no current will flow.

so it will force both of the inputs to0V. I.15 An ideal op-amp A simple op-amp example is given in Figure 7.12 II+ V+ V+ Vo Note: for analysis use. so it is normally reasonable to assume they are the same.7.) The non-inverting input is connected directly to ground.circuits .= V+ Figure 7. or amplifier. As expected both of the op-amp input voltages are the same. This is a function of the circuit design. When the currents are summed at the inverting input.16.= I+ = 0 V. (Note: most op-amp circuits are designed to force both inputs to have the same voltage. The gain of the amplifier is determined by the ratio of the input and feedback resistors. an equation including the input and output voltages is obtained. . The final equation shows the system is a simple multiplier.

7.– V V.circuits .17.– V o = ---------------i + ---------------.13 R2 R1 + Vi + + Vo - The voltage at the non-inverting input will be0V. .16 A simple inverting operational amplifier configuration An op-amp circuit that can subtract signals is shown in Figure 7.= 0 R1 R2 0 – V 0 – Vo -------------i + -------------. V+ = 0V V. Vi R  1 Figure 7.= V+ = 0V The currents at the inverting input can be summed.= 0 R1 R2 –R2V Vo = --------------i R1 – R2 Vo =  -------. by design the voltage at the inverting input will be the same. ∑ I V- V.

Both the inverting and non-inverting input voltages are then set to be equal. The general approach to this solution is to sum the currents into the inverting and non-inverting input nodes.7. The equations for the circuit in Figure 7.14 R2 Find the input/output ratio. After that. Notice that the current into the op-amp is assumed to be zero. . R1 + + Vi R5 + Vref R3 R4 + Vo - Figure 7.17 and derived in Figure 7.17 Op-amp example For ideal op-amp problems the node voltage method is normally the best choice. Notice that if all of the resistor values are the same then the circuit becomes a simple subtractor. algebraic manipulation results in a final expression for the op-amp.18.circuits .

 ----.circuits . First sum the currents at the inverting and non-inverting op-amp terminals.= 0 R5 R4 1 1 1V .= Vref  ----------------.= 0 R1 R2 1.000 times. A typical op-amp will work for signals from DC up to about .+ ----- = V i  ----- + Vo  -----  R 1 R 2 R  R  1 2 R1 + R2 11V +  ----------------. – V ref  ----------------------------- - R  R ( R + R ) 1 1 4 5 (3) Figure 7.7. typically 100. = V i  ----- + Vo  -----  R R  R  R  1 2 1 2 R2 R1 V + = V i  ----------------.18 Op-amp example (continued) An op-amp (operational amplifier) has an extremely high gain. – Vref  ----------------. + Vo  ----------------. The gain is multiplied by the difference between the inverting and non-inverting terminals to form an output.+ ----- = V ref  -----  R4 R 5 R  5 R4 V . + V o  ----------------. although others can be used if the non-ideal op-amp model is used. - - R + R R + R  R + R - 4 5 1 2 1 2 R1 R2 R4 V o  ----------------. = Vi  ----------------. ∑ IV+ V+ – V V + – Vo = ----------------i + ----------------.+ ----. = V i  ----------------. R + R  R + R  R + R  1 2 1 2 4 5 R2 R 4 ( R1 + R2 ) V o = Vi  ----.111V +  ----.= V+ R4 R2 R1 V ref  ----------------. R + R  4 5 (2) Now the equations can be combined. R + R  R + R  1 2 1 2 (1) ∑ IV- V . V .– Vref V= -------------------.15 Note: normally node voltage methods work best with op-amp circuits.

16 100KHz.19 A non-ideal op-amp model 7.3 IMPEDANCE Circuit components can be represented in impedance form as shown in Figure 7.19 should be used. the model of the op-amp in Figure 7. ’Z’.20. instead of resistances. ’R’. In this form we can use impedances as if they are resistances.I ( t ) dt Impedance Z = R 1Z = ------DC Z = LD Note: Impedance is like resistance. V = ZI Capacitor Inductor Figure 7. When the op-amp is being used for high frequencies or large gains.∫ I ( t ) dt C dV ( t ) = L ---. The voltage difference drives a dependent voltage source with a large gain. This model includes a large resistance between the inverting and non-inverting inputs.circuits . The output resistance will limit the maximum current that the device can produce.20 Impedances for electrical components . except that it includes time variant features also. normally less than 100mA. V– rn + A ( V+ – V .) Vo ro typically. Device Resistor Time domain V ( t ) = RI ( t ) 1 V ( t ) = --. When represented this way the circuit solutions can focus on impedances. Notice that the primary difference is that the differential operator has been replaced. r n > 1M Ω 5 A > 10 r o < 100 Ω V + Figure 7.7.

21 A impedance example for a circuit .circuits . The impedances are written beside the circuit elements.= -----------------. Notice that the two impedances at the right (resistor and capacitor) are equivalent to two resistors in parallel.21.+ DC + -R 1 ------.7. DL 50VDC + - 1 1 R -----------------------.= 50V  -----------------------------------------.17 When representing component values with impedances the circuit solution is done as if all circuit components are resistors. and the overall circuit is a voltage divider.= --------------------1 . R  DC Treat the circuit as a voltage divider.  2  R  --------------------- D RLC + DL + R DL +  1 + DCR Figure 7. R  ---------------------  1 + DCR R Vo = 50V ----------------------------------------.-1 1 RCD + 1 ------------. An example of this is shown in Figure 7. DL t=0sec 50VDC + R 1/DC + Vo - Find the equivalent for the capacitor and resistor in parallel.

If the circuit contains fewer loops. Write the equations for the loops or nodes. Before continuing. 5. If so. The voltages around each loop are summed to provide equations for each loop. select the node voltage method. Identify the desired value and eliminate unwanted values using algebra techniques. For the current loop method define current loops and indicate voltage rises or drops by adding ’+’ or ’-’ signs.7. If the circuit contains fewer nodes. 6. current divider or an op-amp inverting amplifier. Look at the circuit to determine if it is a standard circuit type such as a voltage divider. coefficient C L R units As ----V Vs ----A V -A The circuit in Figure 7. consider the nodes and loops in the circuit. 2. They may be used to check equations by doing a unit balance. Otherwise. . or two nodes. An arbitrary decision is made to use the current loop method. 1.22 could be solved with two loops. The most important concept to remember is that a minute of thinking about the solution approach will save ten minutes of backtracking and fixing mistakes. Use numerical values to find a final answer. 3. select the current loop method.4 EXAMPLE SYSTEMS The list of instructions below can be useful when approaching a circuits problem. Note: The units for various electrical quantities are listed to the right. verify that the select method can be used for the circuit. 4. use the standard solution to solve the problem. For the node voltage method define node voltages and current directions.circuits .18 7.

First. This current can be used to find the current through the output resistor R2. ∑ VL1 = – V + R 1 I1 + L ( DI 1 – DI 2 ) = 0 (1) (2) ( R1 + LD )I1 = V + ( LD )I 2 V LD I 1 = ------------------. .19 R1 C + V - L I1 I2 + Vo - R2 Note: when summing voltages in a loop remember to deal with sources that increase the voltage by flipping the sign. sum the voltages around the loops and then eliminate I1.+ ------- I 2 2 LD   CLD (4) Figure 7.23 to develop an input-output equation for the second current loop. The output voltage can then be found by multiplying the R2 and I2.22 are manipulated further in Figure 7.+ R 2 I 2   CD ∑ VL2 (3) R2   1 I 1 =  1 + -------------.7.circuits .+ R2 I2 = 0 CD I2 L ( DI 2 – DI1 ) + ------. I2 R1 + LD  R1 + LD I2 = L ( DI 2 – DI 1 ) + ------.22 Example problem The equations in Figure 7.+ R2 I 2 = 0 CD 1( DL )I1 =  LD + ------.+  ------------------.

.=  1 + -------------.+  ------------------. and the output variable must be calculated separately.+ ------- I2 2 LD R1 + LD  R1 + LD   CLD R2  V 1 LD - ------------------. In this case a dummy variable is required to replace the two first derivatives in the first equation.24.=  --------------------------------------------------------------------------------------------------- I2 2 R 1 + LD   CLD 2   CLD I 2 =  --------------------------------------------------------------------------------------------------------------------------------. as shown in Figure 7. which now becomes an output variable.+ ------.20 First.– ------------------. V  ( R1 + LD ) ( ( CLD 2 + 1 + CDR2 ) ( R 1 + LD ) – CL2 D 3 ) 2   CLD I2 =  --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- V 2  CL ( R1 + R 2 )D 3 + L ( CR 2 + 2CR 1 R 2 + L )D 2 + R 1 ( CR1 R 2 + 2L )D + ( R1 ) 1 2 2 3 Convert it to a differential equation. 2 CL ( R 1 + R2 )I2''' + L ( CR 1 + 2CR 1 R 2 + L )I 2'' + R1 ( CR 1 R2 + 2L )I2' + ( R 1 )I 2 = CLV'' 2 Figure 7. In the remaining state equations I1 is replaced by q1.7. The dummy variable is used in place of I1. I2 =  1 + -------------.circuits . I 2 2 R 1 + LD LD R 1 + LD  CLD  ( CLD + 1 + CDR 2 )( R1 + LD ) – CL D  V ------------------. In the final matrix form the state equations are in one matrix.23 Example problem (continued) The equations can also be manipulated into state equations. sum the voltages around the loops and then eliminate I1. R2   V 1 LD I 1 = ------------------.

I 1 L L q 1 = I1 – I 2 · V R1 q 1 = -.+ R2 I2 = 0 C I2 · V – R1 ( q 1 + I 2 ) + --.I1 L L I1 = q1 + I2 · V R1 q1 = -.+ R2 I 2 = 0 C I2 · V – R1 I1 + --.( q 1 + I2 ) L L R1 R1 V · q 1 = q 1  – ----- + I 2  – ----- + - L  L L ·· ·· I2 · LI 2 – LI 1 + --.24 Example problem (continued) .– ----. R1 R1 V – ----.– ----.2 C (13) I1 = 1 1 q1 I2 Figure 7. · · (1) becomes R1 I 1 + LI1 = V + LI2 · · LI1 – LI 2 = V – R 1 I1 · · V R1 I 1 – I 2 = -.= – R2 I 2 C (12) (11) (10) (3) becomes · 1 I 2 = I2  – R 1 + --- + q 1 ( – R1 ) + V  C These can be put in matrix form.21 State equations can also be developed using equations (1) and (3).7.– ----.– ----q1 q1 d-L L ---= + L dt I 1 I 2 V – R1 – R 1 + --.circuits .

circuits .22 Evaluate the circuit in Figure 7.25 Drill problem: Use the node voltage method .22 using the node voltage method.7. Figure 7.

23 Find the equation relating the output and input voltages.circuits .7. Rf C Vin Ri + Vout Figure 7.26 Drill problem: Find the state equation .

and convert it to a differential equation.27 is continued in Figure 7.27 Circuit solution using impedances The first-order differential equation in Figure 7. .28 where the equation is integrated. t=0 V s = 3 cos t + - C R + Vo - As normal we relate the source voltage to the output voltage.   R  ------------------ V o = Vs  1 -  R + ------.7. In this case the result is a first-order differential equation. DC CRD V o = Vs  ---------------------  CRD + 1 V o ( CRD + 1 ) = V s ( CRD ) · · V o ( CR ) + Vo = V s ( CR ) · · 1V o + Vo  ------- = V s  CR Figure 7. The we find the values for the various terms in the frequency domain.circuits .24 The circuit in Figure 7. we may combine the equations. The solution is left in variable form. ZR V o = Vs  ------------------ Z + Z  R C where. except for the supply voltage.27 can be evaluated as a voltage divider when the capacitor is represented as an impedance. ZR = R 1ZC = ------DC Next.

t · 1dV o + Vo  ------- =  ---- ( 3 cos t ) = –3 sin t  CR  dt Vp = A sin t + B cos t Vp' = A cos t – B sin t 1( A cos t – B sin t ) + ( A sin t + B cos t )  ------- = –3 sin t  CR 1A + B  ------- = 0  CR – 1A = B  -------  CR 1– B + A  ------- = – 3  CR –1 . starting with a guess.  ------- = --------------------.  A 0 2 2 0 = C 1 e + A + B sin  0 + atan  B  -  A  t – ------CR 2 2 C 1 = – A + B sin  0 + atan  B    -. - 2 2 2 2 1+C R  1 + C 2 R 2 CR 1+C R 2 2 B A + B sin  t + atan  -.circuits . write the homogeneous solution using the known relationship. The system will be assumed to be at rest initially.+ 1 = 3  2 2  C R 2 2  3C 2 R 2  – 1 3C R – 3 CR B = --------------------A =  --------------------. – ------· 1CR yields V o + Vo  ------- = 0 Vh = C1 e  CR Next. the particular solution can be determined.25 First.    A  The homogeneous and particular solutions can now be combined. Vp = V0 = V h + Vp = C1 e 2 2 + A + B sin  t + atan  B    -.1– B + B  -------  ------- = – 3  CR  CR 1 B  -----------.28 Circuit solution using impedances (continued) .7.  A Figure 7.

29. And. The rotational inertia J1 is the motor rotor. On the left hand side is the resistance of the motor and the ’back emf’ dependent voltage source. When a voltage is applied the torque will cause the rotor to accelerate. for a given voltage the ratio between steady-state torque and speed will be a straight line. and the second inertia is an attached disk.7. On the right hand side the inertia components are shown. T voltage/current increases ω Figure 7. .30. For any voltage and load on the motor there will tend to be a final angular velocity due to friction and drag in the motor.29 Torque speed curve for a permanent magnet DC motor The basic equivalent circuit model is shown in Figure 7.circuits . includes the rotational inertia of the rotor and any attached loads. as shown in Figure 7.26 7.5 PERMANENT MAGNET DC MOTORS DC motors apply a torque between the rotor and stator that is related to the applied voltage/current.

= ------------------R K T load dKK∴ ---- ω + ω  ----. ω Vm + - J1 J2 Because a motor is basically wires in a magnetic field.31. Tm T m = KI ∴I = -----K Next. And.7. P = V m I = Tω = KIω ∴Vm = Kω Consider the dynamics of the rotating masses by summing moments. Vs – Vm I = -----------------R V s – Kω Tm ∴ ----.27 I Voltage Supply + Vs - R T.31 The first-order model of a motor . The current-voltage relationship for the left hand side of the equation can be written and manipulated to relate voltage and angular velocity.= ------------------R K dJ  ---- ω + T load  dt V s – Kω ∴ -----------------------------------. = Vs  ----.30 d= T m – T load = J  ---- ω  dt d∴T m = J  ---- ω + T load  dt The torque and inertia in a basic motor model These basic equations can be manipulated into the first-order differential equation in Figure 7.circuits . – ---------- dt  JR  JR J 2 Figure 7. the electron flow (current) in the wire will push against the magnetic field. the torque (force) generated will be proportional to the current. consider the power in the motor. ∑M Figure 7.

and hence opposing magnetic fields in the stator. The rotating fields generate currents. These can be combined with the equivalent circuit model to determine the response of the motor to a load.32. The maximum motor speed is a function of the frequency of the AC power. and compatibility with the power distribution system. IL. but always less than the rotating AC fields. . Ls Rs Lr Rr V Lm IL RL RL = 1 – .33. and the number of pole of the machine. It is a function of the fraction. The motors are constructed with windings in the stator (outside of the motor). For example.circuits . of full speed.32 Basic model of an induction motor The torque relationship for AC motors is given in Figure 7. The equivalent circuit for an AC motor is given in Figure 7. f.6 INDUCTION MOTORS AC induction motors are extremely common because of the low cost of construction. The motor does not have bushes to the rotor.28 7. The rotor normally has windings. an induction motor with three poles being used with a 60Hz AC supply would have a maximum speed of 2*(60Hz/3) = 40Hz = 2400RPM.7. or a squirrel cage. The slip of the motor determines the load current.R r ---------f f Figure 7. The motor speed is close to.

33. 7.ω dt CONTINUE TO STATE VARIABLE MODEL.29 First the torques on the motor are summed. with windings on the stator.7 BRUSHLESS SERVO MOTORS Brushless servo motors are becoming very popular because of their low maintenance requirements.. as shown in Figure 7. The windings on the stator are switched at a given frequency to produce a desired rotational speed. ∑M d= Trotor – T load = J ---. or held static to provide a holding torque.33 The construction of a brushless servo motor .circuits .. Va windings on stator Vb N S permanent magnet rotor Vc Figure 7.7.. The motors eliminate the need for brushes by using permanent magnets on the rotor.

V t = terminal voltage across motor windings Rm = resistance of a motor winding L = phase to phase inductance Im = current in windin g E = back e.f.L Im + E  dt  E = Keω T = Kt I m where.circuits .30 The basic relationships for brushless DC motors are given in Figure 7. of motor K e = motor speed constant ω = motor speed K t = motor torque constant T = motor torque Figure 7.m.34. dVt =  Rm + ---.34 Basic relationships for a brushless motor .7.

increasing the gain of the op-amp circuit. At high frequencies the capacitor.circuits . Low pass often used to eliminate noise.37 Low-Pass Filter A high pass filter is shown in Figure 7.37.8 FILTERS Filters are useful when processing data signals.38. C. A low pass filter is shown in Figure 7. Band pass filters reject all frequencies outside a desired frequency band. At low frequencies the capacitor impedance is high. This is easier to conceptualize if the R 1-C pair are viewed as a voltage divider. In this circuit the gain will increase for signals with higher frequencies. has a very low impedance. In this case the voltage divider in the previous circuit is reversed. Rf Ri + R1 + Vi C + Vo - Figure 7.36 Typical supply voltages 7. .7.32 Figure 7. and grounds the input signal. high pass filters eliminate static signals and leave dynamic signals.

such as inductors and capacitors.38 High-Pass Filter 7. • node and loop methods can be used to analyze circuits.33 Rf Ri + C + Vi R + Vo - Figure 7. as resistances will dissipate power in the form of heat.39 can be used to calculate the power and energy in a system. inductors op-amps.7. .39 Electrical power and energy 7. store energy. Other devices. but don’t dissipate it.circuits . capacitors.9 OTHER TOPICS The relationships in Figure 7. V2 P = IV = I R = ----R 2 E = Pt Figure 7. Notice that the power calculations focus on resistance.10 SUMMARY • Basic circuit components are resistors.

11 PRACTICE PROBLEMS 1. b) Put the equation in state variable matrix form. L Vi + C R a) Develop a differential equation for the circuit. R Vi + R + R R Vo - 4. Derive the equations for combined values for resistors. L R + + C Vo - Vi 3.7. 7. 2. Find the output voltage as a function of input voltage. Consider the following circuit.circuits . Write the differential equation for the following circuit. + Vo - . capacitors and inductors in series and parallel.34 • Capacitor and inductor impedances can be used as resistors in calculations.

a) Find the differential equation for the circuit below where the input is Vi. .7.circuits . and then find the natural frequency and damping coefficient. Consider the following circuit. R1 Vi + R2 L C + Vo - 8. L Vi + + C R Vo b) Convert the equation to an input-output equation. Develop differential equations and the input-output equation for the electrical system below. Develop a differential equation for the circuit.35 5. L Vi + R1 + R3 Vo C R2 - 7. Find the input-output equation for the circuit below. and the output is Vo. R1 Vi + L C R2 - + Vo 6.

circuits . a) Write the differential equations for the system pictured below. Simplify the results.36 c) Solve the differential equation found in part b) using the numerical values given below. find the ratio of the output over the input (this is also known as a transfer function). L = 10mH C = 1µF R = 1KΩ Vi = 10V Vo = 2V Vo ′ = 3 V -at t=0s s 9. the circuit has the voltage Vo and the first derivative shown below. R2 R1 Vi C + Vo 10. Assume at time t=0. Given the circuit below. b) Put the equations in input-output form.7. R2 Vi R1 + C Vo .

001H 1000Ω + Vi + 10Ω + Vo - .circuits . a) Find the differential equation for the circuit below.7. Examine the following circuit and then derive the differential equation. 1000Ω 1µF 0.37 11. L R2 R1 + Vi + + Vo - 13. Examine the following circuit and then derive the differential equation. L R1 + Vi + + Vo R2 12.

and the input is Vi=5V. d) Plot a graph of gain against the frequency of the input. a) Write the differential equations for the system pictured below. and see what the magnitude of the output is. b) Put the equations in state variable form. R2 R1 Vi C C=1uF R1=1K R2=1K + Vo 7.12 PRACTICE PROBLEM SOLUTIONS 1. The general method is put in a voltage such as Vi=1sin(___t). Note: This should act as a high pass or low pass filter. Divide the magnitude of the output sine wave by the input magnitude. Assume the system starts at rest.circuits . R1 R2 Rparallel = ----------------R1 + R2 C1 C 2 C series = -----------------C 1 + C2 L1L2 L parallel = ----------------L 1 + L2 Rseries = R1 + R 2 C parallel = C 1 + C 2 L series = L 1 + L 2 2. .7.38 b) Put the differential equation in state variable form and a numerical method to produce a detailed sketch of the output voltageVo. c) Use numerical methods to find the ratio between input and output voltages for a range of frequencies. 14.

+ --------. + Vo  -----------------------------.  RC  LC  LC · V o = Xo · –1–1 1Xo = X o  ------- + Vo  -----. a) b) ·· · · 1 111Vo + V o  --------. = V i  --------.= 1 -Vi 5 4. = Vi  -----.circuits .  CR1 CR 2  LC  CR  1 6. ·· · 11· 1 1Vo + Vo  --------. + V o  -----. = Vi  -------------------------- - -  LC ( R + R )  LC ( R + R )  L ( R + R ) 2 3 2 3 2 3 7. + Vi  -----. = Vi  -----.7. = Vi  --------. R2 R3 R2 · R2R3 C + L ·· · Vo + V o  -----------------------------.  RC  LC  LC 5.39 ·· · 11Vo + Vo  R + V o  -----. a) b) ·· · 1 11Vo + V o  ------- + Vo  -----. Vo ----. - L  LC  LC 3.  CR 1 CR2  LC  CR  1 ωn = 1-----LC L ( R1 + R2 ) ζ = -----------------------------2 CR1 R 2 . + Vo  -----.+ --------.

Sum currents at node Vo ( Vo – Vi) ( Vo ) I Vo = --------------------. (a.circuits .= 0 ∑ DL R Vo DL 2 V o – Vi + Vo D LC + ------------.40 8.7. ·· · 1 11Vo + V o  ------- + Vo  -----. = Vi  -----.+ ( Vo )DC + ---------.= 0 R 111V o'' + Vo'  ------- + V o  -----.  LC  LC  CR .  CR  LC  LC (b. = V i  -----.

Vp = A guess ( 0 ) + ( 0 ) ( 10 ) + ( A )( 10 ) = ( 10 ) 3 8 9 A = 10 Vp = 10 for initial conditions.7.circuits .01 cos ( C 2 ) Vo = – 8.01 e – 500t cos ( 9987t – 0.41 (c.= ------------------. Vo=2V · – 500t – 500t V o = – 500C 1 e cos ( 9987t + C 2 ) – 9987C 1 e sin ( 9987t + C 2 ) for t=0.050 ) + 10 .= tan ( C 2 ) C 2 = – 0.050 8 ( 9987 ) cos ( C 2 ) –8 C 1 = ------------------. d/dt Vo=3V 3 = – 500C 1 cos ( C 2 ) – 9987C 1 sin ( C 2 ) 3 = – 500C 1 cos ( C 2 ) – 9987C 1 sin ( C 2 ) –8 3 = – 4000 – 9987  ------------------- sin ( C 2 )  cos ( C ) 2 sin ( C 2 ) – 3997 ------------------. ·· · L V o ( LC ) + Vo  -.= – 8.= – 500 ± 9987j 2 – 500t Vh = C 1 e cos ( 9987t + C 2 ) 2 3 8 3 3 2 8 2 At At 3 At 8 particular. A e + Ae ( 10 ) + e ( 10 ) = 0 – 10 ± ( 10 ) – 4 ( 10 ) A + A ( 10 ) + ( 10 ) = 0 A = ------------------------------------------------------------. + V o = V i  R ·· ·  10 – 2 –2 – 6 Vo ( 10 10 ) + Vo  ---------- + V o = 10  103  ·· · –8 –5 Vo ( 10 ) + V o ( 10 ) + Vo = 10 ·· · 3 8 9 V o + Vo ( 10 ) + V o ( 10 ) = ( 10 ) homogeneous. Vo = C1 e – 500t cos ( 9987t + C 2 ) + 10 cos ( 9987 ( 0 ) + C 2 ) + 10 2 = C1 e – 8 = C 1 cos ( C 2 ) (1) – 500 ( 0 ) for t=0.

circuits .+ ( – V o ) ( 0. + V i  -------.  1000000 + D V o ( – 1000000 – D – 1000D – 0.000001D ) = 0  1000000 + D 1000 1000 1000000 Vo ( – 1 – 0.000001D ) = 0 1000 1000 ( – Vo ) ( –1 ) 1000000 ----------. Create a node between the inductor and resistor Va.= --------------------------------Vi R1 + DR 1 R 2 C 11.42 9.001D ) = Vi  -----------------------------. R2 · · – R2 Vo + V o  ----- = V i  -------. R  R  1 1 12.  1000000 + D V– = V + = 0V XXXXADD UNITSXXXXX ∑ IV – ( V– – V A ) ( V – – V o ) = ------------------------. Vo –R2 ----.001D 1000 1000000 ( V A – V i ) + VA D = 0 1000000 V A = Vi  -----------------------------. a) b) R1 · V i + V i ( R 1 C ) + V o  ----- = 0  R 2 – R2 · V o = Vi ( –CR2 ) + V i  -------.7.= 0 ∑ 0. (a.  L R  1 13.+ ( V – – V o ) ( 0. R  1 10.001 ( 10 ) ) + V o ( – 1 ) = Vi –9 –3 2 . – R2 · –L Vo = V i  ----.Vi  ------------------------------- + ------------.001D ) = 1000000V i V o'' ( – 10 ) + V o' ( – 1. and use the node voltage method ( VA – V i ) ( VA – V– ) I VA = --------------------.+ ------------------------.+ -----------------------.

Vo = – 1000000000Vi – 1001000V o – 1000000000V o dt 14. and use them to find the response to .    1000ω  7. · d---.circuits .43 (b.· ---. Develop the differential equation(s) for the system below. a) b) c) –R2 · Vo = V i ( – CR 2 ) + Vi  -------.7.V o = Vo dt · d. R  1 Not a state equation assume Vi=1sin(wt) Vo = d) 2 2 1 1 + ( 1000ω ) sin  ωt + atan  --------------.13 ASSIGNMENT PROBLEMS 1.

R3 + Vi + R4 C Vo R1 R2 R1=R2=R3=R4=1Kohm C=1uF a) V i = 5 sin ( 100t ) b) c) V i = 5 sin ( 1000000t ) Vi = 5 .7.circuits .44 the following inputs. Assume that the circuit is off initially.

8. A manual transmission has an input shaft from the motor and from the shifter. consider a mechanical gear box with an input and output shaft. When analyzing systems we will often use transfer functions that describe a system as a ratio of output to input. A function can be described as a transformation of inputs to outputs.= -------------------------------2 F D + 4D + 16 Figure 8.1 A transfer function example . block diagrams and simplification • Feedback controllers • Control system design Objectives: • To be able to represent a control system with block diagrams. The general form output ---------------.1 8. FEEDBACK CONTROL SYSTEMS Topics: • Transfer functions. Or.= f ( D ) input An example x 4+D -.1 INTRODUCTION Every engineered component has some function.1. An example of a transfer function is shown below in Figure 8.8. The general form calls for output over input on the left hand side.feedback control . while ’F’ is the input. In the example ’x’ is the output.2 TRANSFER FUNCTIONS Transfer functions are used for equations with one input and one output variable. For example it could be an amplifier that accepts a signal from a sensor and amplifies it. 8. The right hand side is comprised of constants and the ’D’ operator. • To be able to select controller parameters to meet design objectives.

Approximate frequency of control system oscillations. In reality many systems are not reversible. If < 1 overdamped. This ability to invert a transfer function is called reversibility. The homogeneous equation (the left hand side) ends up as the denominator of the transfer function. If =1 critically damped.= ------------------------------------------2 2 f D + 2ζωn D + ω n particular homogeneous ωn ξ Natural frequency of system .If < 1 then underdamped. the first peak becomes infinite in height. As damping factor approaches 0.2 If both sides of the example were inverted then the output would become ’F’.8.3. There is a direct relationship between transfer functions and differential equations. As before the homogeneous and non-homogeneous parts of the equation becomes the denominator and the numerator of the transfer function. Damping factor of system . . 2 f· ·· x + 2ζω n x + ω n x = ---M f2 2 xD + 2ζωn xD + ω n x = ---M 2 f2 x ( D + 2ζω n D + ω n ) = ---M 1 ----  M x .2 The relationship between transfer functions and differential equations for a mass-spring-damper example The transfer function for a first-order differential equation is shown in Figure 8. and the system will oscillate. Figure 8. and never any oscillation (more like a first-order system). and the input ’x’. This is shown for the second-order differential equation in Figure 8.feedback control . The nonhomogeneous solution ends up as the numerator of the expression.2.

Negative feedback is the term used for this type of controller.3 A first-order system response 8. In standard operation the gas pedal angle is controlled by the driver. is the velocity of the car.3 CONTROL SYSTEMS Figure 8. . or result.= ------------f 1 D + -τ Figure 8. in this figure it is shown inside the dashed line. The system output. The subtraction occurs in the summation block (the circle on the left hand side).4 shows a transfer function block for a car. The input. This error is used by the controller function to adjust the control variable in the system. In this control system the output velocity is subtracted from the setpoint to get a system error. When a cruise control system is engaged the gas pedal must automatically be adjusted to maintain a desired velocity setpoint.3 · -x + 1x = f τ xD + 1 x = f -τ 1 x  D + -- = f  τ x 1 . or control variable is the gas pedal angle.feedback control .8. To do this a control system is added.

Larger differences between the setpoint and output will result in larger errors. . The rules in the figure give a general idea of how a control function might work for a cruise control system. Negative feedback also makes systems more immune to random variations in component values and inputs.4 can be defined many ways. the error is -10mph. a car) OUTPUT (e.g. and the car should be slowed down. The control function in Figure 8.g.feedback control . For example if the desired velocity is 50mph and the actual velocity 60mph.4 Control variable INPUT (e. θgas) SYSTEM (e. In a negative feedback control the setpoint and output values are subtracted.4 An automotive cruise control system There are two main types of feedback control systems: negative feedback and positive feedback. Figure 8. Recall that the system error is the difference between the setpoint and actual output. A possible set of rules for controlling the system is given in Figure 8. When the system output matches the setpoint the error is zero.8.g. As a rule negative feedback systems are more stable than positive feedback systems. Each block in the diagram represents a transfer function. In a positive feedback control system the setpoint and output values are added.5. velocity) vdesired + _ verror control function θgas car vactual Note: The arrows in the diagram indicate directions so that outputs and inputs are unambiguous.

keep θgas the same 4. to provide a longterm output for the motor (a ballpark guess). .5 Example control rules In following sections we will examine mathematical control functions that are easy to implement in actual control systems. 8. then increase/decrease θgas. The resulting error is used in the PID function. When there is a sudden change occurs in the error value the differential branch will give a quick response. In the proportional branch the error is multiplied by a constant. At the output the motor shaft drives a potentiometer.1 PID Control Systems The Proportional Integral Derivative (PID) control function shown in Figure 8. Figure 8.7 shows a basic PID controller in block diagram form. and has been positive/negative for a while. The voltages from the setpoint and output are subtracted at the summation block to calculate the feedback error.6 is the most popular choice in industry. The overall performance of the system can be changed by adjusting the gains in the three branches of the PID function. u = K p e + K i ∫ edt + K d  de  ----- dt Figure 8. 5. This result is then amplified to drive the motor.8. The results of all three branches are added together in the second summation block. also used as a voltage divider. The result is a control variable value. In the equation given the ’e’ is the system error. If verror suddenly becomes bigger/smaller. If verror is very big/small increase/decrease θgas 3. In this case the potentiometer on the left is used as a voltage divider. If verror is not zero. If an error is largely positive or negative for a while the integral branch value will become large and push the system towards zero.feedback control .5 Human rules to control car (also like expert system/fuzzy logic): 1. If verror is near zero.3. etc. and there are three separate gain constants for the three terms. increase/decrease θgas 2.6 A PID controller equation Figure 8. providing a setpoint voltage.

In a P controller only the proportional gain in non-zero. but it is prone to overshoot. This controller will generally work.8.) This controller is generally good for eliminating long term errors.e  dt  + + + u PID function V +V amp -V motor Figure 8. For a PI Controller θ gas = K p v error + K i ∫ v error dt For a P Controller θ gas = K p v error For a PD Controller dv error θ gas = K p v error + K d  ----------------  dt  Figure 8.feedback control . (Recall the second order response.8 Some other control equations . but is very responsive to system changes.7 A PID control system There are other variations on the basic PID controller shown in Figure 8. A PI controller results when the derivative gain is set to zero. The PD controller does not deal with longterm errors.6 proportional V + e Kp ( e ) integral Ki ( ∫ e ) derivative dK d  ---.8. but often cannot eliminate errors.

3. equivalent blocks for a negative feedback loop are shown in Figure 8. c  1 + G ( D )H ( D ) Figure 8.8.10 to Figure 8.9. 8. The proportional gain is then adjusted until the system is responding to input changes without excessive overshoot.16. After that the integral gain is increased until the longterm errors disappear.9 A negative feedback block reduction Other block diagram equivalencies are shown in Figure 8. For example.2 Manipulating Block Diagrams A block diagram for a system is not unique. along with an algebraic proof. except for the cases where the equivalence is obvious. The diagram will then be simplified through a process that is both graphical and algebraic.feedback control . Typically a block diagram will be developed for a system. The differential gain will be increased last to make the system respond faster. meaning that it may be manipulated into new forms.7 Aside: The manual process for tuning a PID controlled is to set all gains to zero. XXXXXXXXXXXXXXX SHOW HOW UNITS MATCH c + - G(D) r is equal to c G(D ) -----------------------------------1 + G ( D )H ( D ) r H(D) r = G ( D )e r = G ( D ) ( c – H ( D )r ) r ( 1 + G ( D )H ( D ) ) = G ( D )c G ( D) r =  -----------------------------------. . Proofs are provided. In all cases these operations are reversible.

8.feedback control .10 A positive feedback block reduction c G(D ) r is equal to r = G ( D )c 1 c = r  -------------  G ( D ) r 1 ------------G(D ) c Figure 8.11 Reversal of function blocks c G(D ) r is equal to c G(D ) G(D) r r r Figure 8.8 c + + G(D) r is equal to c G(D ) -----------------------------------1 – G ( D )H ( D ) r H(D) r = G ( D )e r = G ( D ) ( c + H ( D )r ) r ( 1 – G ( D )H ( D ) ) = G ( D )c G(D ) r =  ------------------------------------ c  1 – G ( D )H ( D ) Figure 8.12 Moving branches before blocks .

15 Moving summation functions before blocks .feedback control .14 Moving branches after blocks c G(D) ± ± e c is equal to ± ± c G(D) e b e = ± G ( D )c ± b b e = G ( D )  ± c ± ----------------  ± G ( D ) 1 ------------G(D ) b Figure 8.13 Combining sequential function blocks c G(D ) r c G(D ) is equal to 1 ------------G(D ) c r c Figure 8.9 c G(D) x H(D ) b is equal to c G ( D )H ( D ) b x = G ( D )c b = H ( D )x b = H ( D )G ( D ) )c Figure 8.8.

feedback control .17 to include mathematical models for each of the function blocks. This block diagram is first simplified by multiplying the blocks in sequence.16 Moving summation function past blocks Recall the example of a cruise control system for an automobile presented in Figure 8. This example is extended in Figure 8. so by default the function is ’1’ .10 c G(D ) r is equal to c G(D) r d r = G( D ) ( ±c ± d ) r = ± cG ( D ) ± dG ( D ) G(D ) d Figure 8.everything that goes in. . The feedback loop is then reduced to a single block. comes out. Notice that the feedback line doesn’t have a function block on it.4.8.

18 to a final transfer function for the whole system.17 An example of simplifying a block diagram The function block is further simplified in Figure 8.8. ----i d   p D MD H(D) = 1 vactual vdesired 10 K + K-i + K D  -------. .11 e. The block diagram of the car speed control system vdesired + verror K K p + ----i + K d D D θgas 10 F 1-------MD vactual vdesired + - verror K K p + ----i + K d D D θgas 10-------MD vactual = G(D) vdesired + verror 10 K + K. ( 1 )    MD D vactual Figure 8. ---d   p D MD -------------------------------------------------------------------------K 101 +  K p + ----i + K d D  -------.+ K D  -------.g.feedback control .

19 shows a reasonable control system arrangement.feedback control .18 An example of simplifying a block diagram (continued) 8.3 A Motor Control System Example Consider the example of a DC servo motor controlled by a computer. Some elements such as power supplies and commons for voltages are omitted for clarity.12 vdesired   K + K-i + K D  ---d    p D   ------------------------------------------------------- K  Ms   -----.3.8.+ K p + ----i + K d D   10   D vactual vdesired   2  D ( Kd ) + D ( K p ) + ( Ki )   ---------------------------------------------------------------------- M  2  ----   D  10 + K d + D ( K p ) + ( K i ) vactual   2 v actual  D ( Kd ) + D ( K p ) + ( Ki )  ---------------. The system in Figure 8.=  ---------------------------------------------------------------------- v desired M  2  ----   D  10 + K d + D ( K p ) + ( K i ) Figure 8. The purpose of the controller is to position the motor. .

13 2.feedback control .19 A motor feedback control system The feedback controller can be represented with the block diagram in Figure 8.20 A block diagram for the feedback controller .20. desired position voltage + Vd - gain Kp op-amp motor shaft potentiometer Figure 8.2K Computer Running Labview PCI-1200 data acquisition card from National Instruments gain Kp 1K + LM675 op-amp shafts are coupled 12Vdc motor X + +5V 5K potentiometer desired position voltage Vd -5V Figure 8.8.

Two of the values must be provided by the system user. so there is no need to calculate ’J’. θ Vo = 5V  --------------  5 ( 2π ) Vo ----.feedback control .159Vrad –1 θ For the shaft.8. The potentiometer is connected as a voltage divider and the equation relates angle to voltage. it integrates the angular velocity into position: dω = ---.22 using experimental data.= 0.2 times.21 Transfer functions for the power amplifier.+ ----------.θ dt θ 1 --. Finally the velocity of the shaft is integrated to give position.= 0V – Vi – Vo ------. So there are 5 turns in the negative and positive direction.desired potentiometer voltage Vd .= --ω D Figure 8.14 The transfer functions for each of the blocks are developed in Figure 8.= 0 ∑ 1K 2.21.= – 2.2 Vi For the potentiometer assume that the potentiometer has a range of 10 turns and 0 degrees is in the center of motion.2K V + = V. potentiometer and motor shaft The basic equation for the motor is derived in Figure 8.2K Vo ----. The op-amp is basically an inverting amplifier with a fixed gain of -2.gain K For the op-amp: V + – V V+ – V o I V+ = ----------------i + ----------------. . In this case the motor was tested with the full inertia on the shaft. Given or selected values: .= 0 1K 2.

0682  ----. The block diagram is then simplified for the entire system to a single transfer function relating the desired voltage (setpoint) to the angular position (output).22 Transfer function for the motor The individual transfer functions for the system are put into the system block diagram in Figure 8. = 1 - JR τ 1K0.ω = 18.33V s 0.8s 1s 2 K ----. 10  JR  JR K = 0.6rads  dt 2 KK0 +  ----.8.= -------------------Vs D + 1.= 18.328 JR 1Dω + -----.25 Figure 8.23. .6 =  ----.8s K----. ω =  ----. V  dt  JR  JR s 1s 2s 3s 2 For steady-state –1 d  ---- ω = 0 ω = 1400RPM = 146.33 ---.0682 τ ≈ 0. The transfer function contains the unknown gain value ’Kp’.feedback control . 146.15 For the motor use the differential equation and the speed curve when Vs=10V is applied: 1400 RPM dK K ---- ω +  ----.8 ω 18. = -------- JR 0.

or critically damped. This is done by recognizing that the bottom (homogeneous) part of the transfer function is second-order and then extracting the damping factor and natural frequency. .33Kp 1 + ( 0.159 ) -----------------------------( D + 1. and simplification The value of ’Kp’ can be selected to ’tune’ the system performance.23 The system block diagram.33 -------------------D + 1.25 )D – 40.159 Vd – 40.2 18.33Kp --------------------------------------------------------( D + 1.33Kp -----------------------------------------------------2 D + 1.33Kp ----------------------------( D + 1.33Kp -----------------------------( D + 1.25 D – 6. but this makes sense when the negative gain on the op-amp is considered.25 )D – 6.412Kp θ θ Vd Vd θ Figure 8.159 θ Vd + - – 40.16 Vd + - Kp -2.25 1 --D θ 0.0.25 )D 0.24 the gain value is calculated to give the system an overall damping factor of 1. In Figure 8. The final result of ’Kp’ is negative.feedback control .8.412Kp – 40.25 )D ---------------------------------------------------------– 40.

0 )ω n ω n = 0.25.412K p = x'' + 2ξω n x' + ω n x 1.412K p = ω n – 6. The equations for calculating these errors are shown in Figure 8.17 We have specified.24 Calculating a gain Kp 8.feedback control .0 The denominator of the system transfer function can be compared to the standard second-order response. The two common types of error are system error and feedback error.3. If the feedback function ’H’ has a value of ’1’ then these errors will be the same.625 K p = – 0.8.0609 2 2 2 2 Figure 8.412K p = 0.4 System Error System error is often used when designing control systems. ξ = 1.625 – 6.25 = 2ξω n 1. c + b e G r System error.25 D – 6.25 = 2 ( 1. or been given the damping coefficient as a design objective. D + 1.25 Controller errors . H e = c–b Figure 8. S = c–r Feedback error.

The response is then compared to the input to find the system error. The system is a simple integrator. Kp G ( D ) = ----D H(D ) = 1 Kp G r ∴. Given.feedback control . S = c–r S = A – ( –A e Figure 8.8. In this case the error will go to zero as time approaches infinity. r = rh + rp = C 1 e 0 = C1 e + A C 1 = –A r = –A e –Kp t 0 –K p t +A +A The error can now be calculated.18 An example of calculating these errors is shown in Figure 8.= ----------------.26. with a unity feedback loop. The overall transfer function for the system is calculated and then used to find the system response.26 –Kp t + A ) = Ae –K pt System error calculation example for a step input . rp = C2 · rp = 0 0 + Kp C2 = Kp A C2 = A The solutions can be combined and the remaining unknown found for the system at rest initially.= ---------------D + Kp 1 + GH c · ∴r + K c = K c p p c = A The homogeneous solution is. rh = C1 e –K pt The particular solution is found with a guess.

8.19 Solve the previous problem using a ramp input. c = At Figure 8.27 Drill problem: Calculate the system error for a ramp input .feedback control .

8. R.feedback control .28 Drill problem: Calculate the errors . 1 G ( D ) = ----------------------------2 D + 4D + 5 H(D ) = 5 c = 4t Figure 8.20 Find the system error ’e’ for the given ramp input.

3. as pictured in Figure 8. When this is used together with a more traditional feedback function the overall system can outperform more traditional controllers function.3.29.+ τD   τD 1 + ατD G c = K  --------------------  1 + τD  1 + τD G c = K  --------------------  1 + ατD α>1 α>1 α>1 τ1 > τ2 Lead Lag Lead-Lag 1 + τ 1 D 1 + ατ2 D G c = K  ---------------------.  ----------------------  1 + ατ D  1 + τ D  1 2 Figure 8. .5 Controller Transfer Functions The PID controller.6 Feedforward Controllers When a model of a system is well known it can be used to improve the performance of a control system by adding a feedforward function.feedback control .30.21 8. such as the PID controller. A more complete table is given in Figure 8. Type Transfer Function Proportional (P) Proportional-Integral (PI) Proportional-Derivative (PD) Proportional-Integral-Derivative (PID) Gc = K 1G c = K  1 + ------  τD G c = K ( 1 + τD ) 1G c = K  1 + -----.8. The feedforward function is basically an inverse model of the process. and simpler variations were discussed in earlier sections.29 Standard controller types 8.

d---. then the angular velocity.32 that implements a second order state equation.22 feedforward function feedback function process Figure 8. .3.feedback control . These can also be used to form a control system. The system uses two integrators to to integrate the angular acceleration. to get the position. as shown in Figure 8.X = AX + BU dt Y = CX + D + B + d---.31 A state variable control system An example is shown in Figure 8.31.7 State Equation Based Systems State variable matrices were introduced before.X dt X C + + Y U 1 --D A D Figure 8.8.30 A feed forward controller 8.

ω dt + + K State Based Controller JR ----K 2 + - Ki K P + ---.32 A second order state variable control system The previous block diagrams are useful for simulating systems.+ DK d D Feedback Plant .θ dt θ 1 --D d---.ω = Aω + Bθ + Cτ dt B Figure 8. The model.ω dt + 1 --D A d---.  JR  JR · JR V s = ω  ----. = V s  ----.feedback control .23 τ d---. KK· ω + ω  ----. These can then be used in feedforward control systems to estimate system performance and then predict a useful output value.θ = ω dt + C + d---.8. + ω ( K )  K ω A d---.

while the outer loop controls a conveyor feeding parts into the oven. • Block diagrams can be used to describe and simplify systems. • Other control types are possible for more advanced systems.feedback control . There is viscous damping between the block and the ground.33 A state based feedforward controller 8. • System errors can be used to determine the long term stability and accuracy of a controlled system.3.4 SUMMARY • Transfer functions can be used to model the ratio of input to output.24 Figure 8.8 Cascade Controllers When controlling a multistep process a cascade controller can allow refined control of sub-loops within the larger control system.5 PRACTICE PROBLEMS 1.34 A cascade controller 8. Develop differential equations and then transfer functions for the mechanical system below. Most large processes will have some form of cascade control. A force is applied to cause the .8. 8. the inner loop may be for a heating oven. Gc1 Gc2 Gp1 Gp2 Figure 8. For example. such as damping ratio and natural frequency. • Controllers can be designed to meet criteria.

8. The height is measured with an ultrasonic proximity sensor. N K s = 10 --m sK d = 5N --m M = 0. and the output is the angle of the second mass. Here the input is a torque. When the spring is undeflected y=0. Neglect gravity. θ1 θ2 τ K s1 B1 J1 K s2 J2 B2 . the output Vo=0V. If y=20cm then Vo=2V and if y=-20cm then Vo=-2V.feedback control . The mass is suspended by a spring and a damper. Find the transfer functions for the systems below.25 mass the accelerate. Develop a transfer function for the system below. The input is the force ‘F’ and the output is the voltage ‘Vo’. x F M B 2.5Kg Ks Kd M Ultrasonic Proximity Sensor F y 3. When y = 0.

Given the transfer function below. a) Draw a sketch of what the actual system might look like. Find the transfer functions for the system below where Vi is the input and Vo is the output. Given the transfer function. C + Vi R1 R2 L + Vo - 5.8. develop a mechanical system that it could represent.rad D + 1 sV 1 --D θa b) Convert the system to a transfer function. Given a mass supported by a spring and damper.+ rad Va V2 -------rad Ve K Vm 10 . x(D) 1 -----------. G(s). and gravity pulls it downward. Kd = 5Ns/m. Identify components.= ---------------------F( D) 10 + 20D where: x = displacement F = force 7. M=10kg. c) solve the input output equation to find an explicit equation of the position as a function of time for Ks = 10N/m. 8. find the displacement of the supported mass over time if it is released from neutral at t=0sec. (Hint: Differential Equations).d 2 -------. determine the time response output Y(t) to a step input X(t). The system uses a proportional controller. d) solve part c) numerically.ω -----------.26 4. a) develop a transfer function for y/F. .= -X ( t ) = 20 When t >= 0 D+2 X 6. and what is it used for? b) What does feedback do in control systems? 9. θd V V. 4 Y G = -----------. b) find the input function F. a) What is a Setpoint. The block diagram below is for a servo motor position control system.-------.feedback control .

) R Vf R Vd R + - R Ve b) Develop an equation for the system below relating the input to the output.8.) + Vi Vn A + + C Vo 12.feedback control .27 10. Simplify the block diagram below to a transfer function. Simplify the block diagram below. (Hint: think of a summation block. Put it in block diagram format. a) Develop an equation for the system below relating the two inputs to the output. (Note: Vn is an input and cannot be combined in a transfer function. A x + B + C + + y 11. Put .

 – T load -----------  s  dt JR J R R M  J R R M Vs Km -----RM + e RM --------------------------------2 J R R M D + Km T load d) Write the transfer function for the system below relating the input torque to the output angle theta2. θ1 θ2 ω 2 τ K s1 J1 K s2 B J2 e) The system below is a combination of previous components. and a tachometer .28 the result in block diagram format. 1K 1K Ve RP + 1K Vs c) The equation below can be used to model a permanent magnet DC motor with an applied torque. Then write the transfer function for the angular velocity of mass 2.8. = V  -----------. An equivalent block diagram is given. Prove that the block diagram is equivalent to the equation. Km Km d  ---- ω + ω  -----------.feedback control .

The PID controller will examine the value ‘e’ and then control the speed of the lift motor with a control voltage ‘c’. all of these equations can be combined into a single system transfer .7 (for a step input). It uses a desired height ‘d’ provided by a user.feedback control .8. R(D) + 5 ---------------------D(D + 1) C(D) 1 + 0. The following system is a feedback controller for an elevator.5t. The difference between these two is called the error ‘e’. select a system gain K that will give the overall system a damping ratio of 0. What is the resulting damped frequency of the system? + K 2 ----------------------------2 D + 3D + 2 15. feedback erro and system error when the input is a ramp with the function c(t) = 0.29 for velocity feedback. Given the block diagram below. The elevator and controller are described with transfer functions.2D 14. Sketch the system errors as a function of time. Find the system output. Vd + 2R P -------------------3 R P + 10 Km -----RM + RM ---------------------------------2 JM RM D + K m ω K s2 ----------------------------------------------------------------------------------------------------------------------3 2 D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + ( BK s2 ) KT 13. Simplify the block diagram. as shown below. and the actual height of the elevator ‘h’.

8. e = d–h error 2 K r . b) Calculate the damping coefficient and natural frequency of the results in part a). . and the input (desired height) changes to 20 as a step input.= K p + ----i + K d D = 2D + 1 + D .10 2  e  r  2 e D D(D + 1) D D D +D h 10 ( D + 1 ) eliminate ‘e’ ----------.e.---------------------.30 equation as shown. b) Find the output of the circuit using explicit integration (i. = -----------------------------------.000t) is applied.-----------------------------------------------.feedback control .. homogeneous and particular solutions).= ( D + 1 ) .   2 10D + 10 h D  --------------------------------.= ---------------2 r D +D elevator r -10  . Study the circuit below.= -----------------------2 d–h D h =  10 ( D + 1 ) ( d – h )  ------------------------ 2 D 10 ( D + 1 ) 10 ( D + 1 ) h  1 + ------------------------ =  ------------------------ ( d ) 2     2 D D 10 ( D + 1 )  -----------------------. 16. Assume that for t<0s the circuit is discharged and off.---------------.PID controller ----------------------------e D D combine the transfer functions 2 2 h 10 -.= 2   d 10 ( D + 1 ) D + 10D + 10  1 + ------------------------ 2   D system transfer function a) Find the response of the final equation to a step input.= 10 ( D + 1 ) ------------------------. Starting at t=0s an input of Vi=5sin(100.  h = h = 2D + 1 + D . The system starts at rest on the ground floor. 1K Vi + 1uF + Vo a) Write a differential equation and then a transfer function describing the circuit.

= ----. θ2 K s2 ---.= -------------------------------------------------------------------------------------------------2 Vi D ( R 1 LC ) + D ( L + CR1 R 2 ) + ( R1 + R 2 ) 5. Vo D( R1R2C ) ----.8. V V 10  --.6 PRACTICE PROBLEM SOLUTIONS 1.-. – 2t + 40 N K s = 10 --m F Ns K d = 20 ----m . y ( t ) = – 40e 6.= ---------------------------F D ( B – DM ) 2.= -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------4 3 2 τ D ( J 1 J 2 ) + D ( B1 J 2 + B2 J 1 ) + D ( K s2 ( J 1 + J 2 ) + B1 B2 ) + D ( K s2 ( B 1 + B2 ) ) 4.= --------------------------------------------------------------2 2 F Ns – 10 --N yF – MD – K d D – K s – 0.  m  m Vo Vo y ----.5kgD – 5 -----D m m 3. x 1 -.31 8.feedback control .= --------------------------------------------. 10  --.

= ----------------------------------------2 F D M + DK d + K s F = Mg y ( t ) = – 10. a) A setpoint is a system input that correlates to a desired output for a system with negative feedback control.253 ) + 9. the error.32 7.968t – 0. a) b) c) d) y 1 -. b) Feedback in a control system allows an out to be compared to an input. A+B -----------------------------------1 + CB + A + B .81 y(t + h ) = y(t) + h 0 1 y(t ) + 0 v(t + h ) v(t) – 1 – 0.25t cos ( 0.8.81 8. is used to adjust the control variable and increase the accuracy. a) 20K --------------------------------2 D + D + 20K b) 10.5 v ( t ) 9.13e – 0. 9. The difference.feedback control .

– V f V . Vn 1 -A + + Vo CA -----------1+A Vi 12.– Vf Ve = 2 ( 0.33 11.1V+  ------.= --------------.= 0.+ -------------. = ------3  3 RP 10 10 substitute in (1) Ve  RP + 1K 0.5Vs  ------------------.– Ve I V.= ------------.– 0 V.feedback control .– Vs IV .+ ----. = ------3  RP ( 10 3 ) 10 Vs 2R P ----.= 0 ∑ 3 3 10 10 V .5Vs V+ – Ve V+ – 0 IV + = ----------------.5Vd ) – Vf Ve = Vd – Vf Vd + Vf Ve b) V.= -------------------3 Ve RP + 10 (1) Ve 2R P -------------------3 R P + 10 Vs note: a constant set by the variable resistor Rp .+ ---------------.8.= 0 ∑ 3 RP 10 Ve 1 .= 0 ∑ R R Ve = 2V.+ ---------------. R a) V + = V d  -----------. = 0.5Vd  R + R V .

 – T load -----------  s  dt JR J R R M  J R R M 2 2 .8. = V  -----------. = V s  -----------. – ----------J R  J R RM JR  R M Tload   Km JRM ω =  --------------------------------.  Vs  -----------. – ----------- JR   J R RM D + K m 2   J R R M   RM Km ω =  --------------------------------.feedback control .  Vs  -----.34 c) Km  T load  Km ω  D + -----------. – Tload  R    J R RM D + K m 2 M Km Km d ---- ω + ω  -----------.

feedback control .= ----------------------------------------------------------------------------------------------------------------------3 2 τ D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + ( BK s2 ) 4 3 2 e) ωD ( ) + D ( ) + D( ) + ( ) ----.35 d) τ J1 K s2 ( θ 1 – θ 2 ) + ∑M = τ – K s2 ( θ 1 – θ 2 ) = J 1 θ 1 D 2 2 τ – K s2 ( θ 1 – θ2 ) = J 1 θ 1 D 2 θ 1 ( J 1 D + K s2 ) + θ 2 ( – K s2 ) = τ K s2 ( θ 1 – θ 2 ) + J2 Bθ 2 D (1) 2 ∑M = K s2 ( θ 1 – θ 2 ) – Bθ 2 D = J 2 θ 2 D 2 K s2 ( θ 1 – θ2 ) – Bθ 2 D = J 2 θ 2 D 2 θ 2 ( J 2 D + K s2 + BD ) = θ1 ( K s2 )  J 2 D + K s2 + BD θ 1 = θ 2  ----------------------------------------- K s2   2 (2) substitute (2) into (1) 2  J 2 D + BD + K s2 2 θ 2  ----------------------------------------- ( J 1 D + K s2 ) + θ 2 ( – K s2 ) = τ K s2    D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + D ( BK s2 ) θ 2  -----------------------------------------------------------------------------------------------------------------------------. = τ K s2   θ2 K s2 ---.8.= -----------------------------------------------------------------------------------------------------------------------------4 3 2 τ D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + D ( BK s2 ) ω 2 = Dθ 2 ω2 K s2 ----.= -----------------------------------------------------------------------------------------4 3 2 Vd D ( ) + D ( ) + D ( ) + D( ) + ( ) 3 2 .

464 – -.224e cos  --------.224e cos  --------.91e ζ = 1.162 -------s 16. ω d = 1.t – 0.050e – 1000t b) + 0. a) b) h ( t ) = 2.+ --- 2  5 2t –t 1 15 S ( t ) = 0.112e cos  --------.30 – 22.464 + ---- 2  10 14.530 rad -------s 15.36 13.464 – - 2  5 –t 115 e ( t ) = –0.873t K = 1.561 ) 5 .t – 0.= --------------------D + 1000 Vi V o = 0.8.58 – 8.91e – 1.feedback control .127t + 20 rad ω n = 3. –t 1 115 R ( t ) = – 0. a) Vo 1000 ----.050 sin ( 10 t + 1.t – 0.

+ 1 -----------D+2 3.8. L R Vi R + C Vo 2. Develop a transfer function for the circuit below. G1 R + G4 G2 G3 + G6 + G5 C 1 -----------D+1 .37 8.feedback control . Simplify the following block diagram.7 ASSIGNMENT PROBLEMS 1. Simplify the following block diagram.

as in Figure 9.1 9. input and output.9. Initially there is a substantial transient response. the complex part corresponds to the oscillations of the system.1 INTRODUCTION When a system is stimulated by an input it will respond. Therefore to do a steadystate sinusoidal analysis we can replace the ’D’ operator with jω . 9.1 Transient and steady-state parts of the differential operator . These include the integration of differential equations. The real component of the number corresponds to the natural decay (e-tothe-t) of the system.frequency response . But. Techniques for finding the combined steady state and transient responses were covered in earlier chapters. while the complex part will represent the sinusoidal steady-state. In other words the real part of the number will represent the transient effects of the system. D = differential operator σ = decay constant ω = oscillation frequency Phasor transform D = jω Figure 9. this is the phasor transform. PHASOR ANALYSIS Topics: • Phasor forms for steady state analysis • Complex and polar calculation of steady state system responses • Vibration analysis Objectives: • To be able to analyze steady state responses using the phasor transform 9. that is eventually replaced by a steady state response.2 PHASORS FOR STEADY-STATE ANALYSIS When considering the differential operator we can think of it as a complex number. and numerical solutions. D = σ + jω where. These techniques involve using the phasor transform on the system transfer function. Phasor analysis can be used to find the steady state response only.1.

the expression is quickly reduced to a simple complex number. The differential operator is replaced with jω.2. x( ω) 1 x ( ω ) = ------------F ( ω ) = ---------------------------------------------------------------------.2 values for the sinusoidal input force are assumed. Some component values are assumed.( A cos θ input + jA sin θ input ) 2 F(ω) ( 2000 – 1000ω ) + j ( 3000ω ) Figure 9.frequency response . We start with a transfer function for a mass-spring-damper system. This equation then described the overall response of the system to an input based upon the frequency of the input. A generic form of sinusoidal input for the system is defined. and the equation is simplified to a complex number in the denominator.) The steady state response of the system is then obtained by multiplying the transfer function by the input.= ----------------------------------------------------------------.2 An example of the phasor transform is given in Figure 9. After this the method only requires the simplification of the complex expression.= ---------------------------------------------------------------------2 2 2 F(ω) 1000j ω + 3000jω + 2000 ( 2000 – 1000ω ) + j ( 3000ω ) A given input function can also be converted to phasor form.= ----------------------------------------------------------2 2 F( D) 1000D + 3000D + 2000 MD + K d D + K s x( ω) 1 1 -----------. but it will be used later.9.= ----------------------------------------. and then finally back into a function of time. In this example numerical values are assumed to put the equation in a numerical form. (Note: the frequency of the input does not show up in the complex form of the input. To simplify this expression it is multiplied by the complex conjugate. and also converted to phasor (complex) form. In particular having a complex denominator makes analysis difficult and is undesirable. The complex number is then converted to polar form. A phasor transform can be applied to a transfer function for a mass-spring-damper system. But it is needed for the transfer function. to obtain the output.2 A phasor transform example To continue the example in Figure 9. The response of the steady state output ’x’ can now be found for the given input. N Ns K s = 2000 --K d = 3000 ----M = 1000kg m m 1 x( D) 1 -----------. F ( t ) = A sin ( ω input t + θ input ) F ( ω ) = A ( cos θ input + j sin θ input ) F ( ω ) = A cos θinput + jA sin θ input Note: the frequency is not used when converting an oscillating signal to complex form. . After this.

671 This can then be converted to a function of time.( – 9998000 – j300000 ) – 9998000 + j300000 ( – 9998000 – j300000 ) Note: This is known as the complex conjugate.505 ×10-6 -6 2 -6 2 ( – 0.794 ) ( – 9998000 ) + j ( 300000 ) 8. x ( ω ) = 0.9.776 + j4.3 Assume the input to the system is. The value is equivalent to 1 so it does not change the value of the expression 2. so the angle should be between 180 and 270 degrees.5rad s These can be applied to find the steady state output response.0005 ×10 x ( ω ) = ( – 0.776 + j4. A = 10N F ( t ) = 10 sin ( 100t + 0. Only the denominator is used top and bottom ( – 86304248 ) + j ( – 50563212 ) x ( ω ) = ----------------------------------------------------------------------14 1.5 )N rad ω = 100 -------θ input = 0.863 ×10-6 -6 -6 Note: the signs of the components indicate that the angle is in the bottom left quadrant of the complex plane.frequency response .( 10 cos 0.863 ×10 ) + ( – 0. 1.9999 ×10 sin ( 100t + 3. The complex component is now negative 3.5 ) 2 ( 2000 – 1000 ( 100 ) ) + j ( 3000 ( 100 ) ) 1 x ( ω ) = ----------------------------------------------------------. To correct for this pi radians are added to the result of the calculation.505 ×10 ) x(ω ) =  – 0.794 ------------------------------------------------------x ( ω ) = --------------------------------------------------.505 ×10 ) ∠ atan  ----------------------------- + π  – 0. 1 x ( ω ) = -----------------------------------------------------------------------------------------.671 )m -6 -6 Figure 9.( 8.863 ×10 ) + j ( –0.5 + j10 sin 0.3 A phasor transform example (cont’d) . x ( t ) = 0.9999 ×10 ∠3.

= A ∠( θ 1 – θ 2 ) -B ∠θ 2 B ( A ∠θ 1 ) ( B ∠θ 2 ) = AB ∠( θ 1 + θ 2 ) For example. After this the output is found by multiplying the transfer function by the input.7854 + 0.6325 ∠0. A + B ∠ atan  B 2 2  -.∠0.4636 – 0.7854 -----------------------------25 ∠0. The polar form for complex numbers simplifies many calculations. But.9.4 Calculations in polar notation The cartesian form of complex numbers seen in the last section are well suited to operations where complex numbers are added and subtracted. The calculations for magnitudes involve simple multiplications. Unfortunately when the numbers are only added or subtracted they need to be converted back to cartesian form to perform the operations. The previous example started in Figure 9. 2 + j( 1 + j )  ------------.∠ atan  B – atan  D  --- A  C  C + jD 2 2  2 2  D C +D C + D ∠ atan  --- C 2 2 A ∠θ 1 ------------.9273 25 = 0. The input frequency is substituted into the transfer function and it is then converted to polar form.= ----------------------. In this example the input is directly converted to polar form.3217 = 0. After this the polar form of .= ------------------------------------------------.46362 ∠0.5. This method eliminates the need to multiply by the complex conjugate.  3 + 4j = 5 ∠0. The angles are simply added. when complex numbers are to be multiplied and divided these become tedious and bulky.4 Note: when dividing and multiplying complex numbers in polar form the magnitudes can be multiplied or divided. and the angles added or subtracted.2 is redone using polar notation in Figure 9.2 Figure 9.9273 2 5 = ------------. without the need for calculation.frequency response . A A + jB A +B ---------------.6 + j0.

6. -7 x(ω ) x ( ω ) = ------------F ( ω ) = ( 0.= ----------------------------------------.5 the result is converted directly back to a function of time.= --------------------------------------------------2 F( ω) – 9998000 + j300000 ( 2000 – 1000 ( 100 ) ) + j ( 3000 ( 100 ) ) x(ω ) 1∠0 -----------.5 Correcting quadrants for calculated angles Consider the circuit analysis example in Figure 9.9.= ------------------------------------------------------------------------------------------------------------------------------F( ω) 2 2  300000 - ( – 9998000 ) + ( 300000 ) ∠ atan  ----------------------- + π  – 9998000  x(ω ) 1 ∠0 1 -----------. x ( ω ) = 0.612 ) -6 -7 -6 Re Note: recall that tan θthe ----but = . imaginary +/+/+ real -/- -/+ Figure 9. (Note: this is a useful point to convert all magnitudes to powers of 10. If it lies beyond the -90 to 90 degree range the correct angle can be obtained by adding or subtracting 180 degrees.5 The transfer function can also be put in polar form.9998 ×10 ∠– 2.∠( 0 – 3.5 )N F ( ω ) = 10 ∠ 0.112 10002500 The output can now be calculated. 1 x(ω ) 1 -----------. In this example the component values are converted to their impedances.612 The output function can be written from this result.112 F(ω) 10002500 ∠3. and the input voltage is converted to phasor form.) After this the .112 ) = 0.frequency response .9998 ×10 ( 10 ) ∠( – 3.112 + 0.function in atan θ Im calculators and software only returns values between 90 to 90 degrees.= -----------------------------------------------------------------------------------------. To compensate for this the sign of the real and imaginary components must be considered to determine where the angle lies. F ( t ) = 10 sin ( 100t + 0.5 ) F(ω) x ( ω ) = 0.5 ) = 0.9998 ×10 sin ( 100t – 2.9998 ×10 ∠– 3. Consider the input function from the previous example in polar form it becomes.112 ) ( 10 ∠0.9998 ×10-7 ∠– 3.= ----------------------.

6 Phasor analysis of a circuit The analysis continues in Figure 9.3 + 4 Z eq + Vo - 11 1 1------.7 as the output is found using a voltage divider.6 three output impedances are combined to a single impedance.= -------------------. In this case the calculations were simpler in the cartesian form. 10 5 –0.frequency response . 10  –7 6  10 10 j 1------. .+ ------–3 6 4 Zeq 1 10 10 j  -------------------. 10K Ω + 1mH 100nF 10K Ω + Vo - 6 V i = 5 sin ( 10 t + 0.099 ) Z eq 1 Z eq = ----------------------------------------–4 ( 10 ) + j ( 0.3 ) The impedances and input voltage can be written in phasor form.= – 10 –3 j + 10 –1 j + 10 –4 = ( 10 –4 ) + j ( 0.3 + 4 10 –3 6 10 j 1 ----------------------–7 6 10 10 j 10 4 + Vo - The three output impedances in parallel can then be combined.099 ) Figure 9. 10 5 ∠ 0.9. Given the circuit.+ -------------------------.

frequency response .7 Phasor analysis of a circuit (cont’d) Phasor analysis is applicable to systems that are linear.3 )  --------------------------------------------------------------------.  Z eq  V o ( ω ) = ( 5 ∠0.099 ) ) + 1 5 ∠0.05 sin ( 10 t – 1.9. The results are then converted back to functions of time. . and then the results simply added.5 and 20 rad/s.05 × 10 ∠– 1.3 1 V o ( ω ) = ( 5 ∠0.∠( 0.5687761 ) = 5. The example considered in Figure 9.   4 –4 ( 10 ) ( ( 10 ) + j ( 0.  –4  ( 10 ) + j ( 0. The output components are found by multiplying the inputs by the response at the corresponding frequency.269 990 Finally. as well as frequencies at 0. Therefore.099 )  1 V o ( ω ) = ( 5 ∠0.269 ) mV 6 Figure 9.7 In this case a combination of cartesian and polar forms are used to simplify the calculations. This means that the principle of superposition applies.2 is extended in Figure 9.3 5 ∠0.5687761 990 2 2 2 + 990 ∠ atan -------2 –3 5V o ( ω ) = -------.8. if an input signal has more than one frequency component then the system can be analyzed for each component. the output voltage can be written. The transfer function is analyzed for each of these frequencies components. The final result is then converted back from phasor form to a function of time.099 )  -------------------------------------------------------------- V o ( ω ) = ( 5 ∠0. The output can be found using the voltage divider form.3 )  ----------------------  10 4 + Zeq 1   ----------------------------------------. V o ( t ) = 5. = --------------------------------------------------.3 )  4  1   10 +  ----------------------------------------. and added together. In this example the input has a static component.= ------------------------------------- 2 + j990 990 ∠1.      –4 ( 10 ) + j ( 0.3 – 1.3 )  ------------------.

= ----------.5 ∠0 -4 1 ∠0 x ( 20 ) =  ----------------------------------.709  2305 ∠0.frequency response .992 ( 2000 – 1000 ( 20 ) ) + j ( 3000 ( 20 ) ) x ( 0.3 VIBRATIONS Oscillating displacements and forces in mechanical systems will cause vibrations. In some cases these become a nuisance.0005 ∠0 2 F( 0) 2000 ( 2000 – 1000 ( 0 ) ) + j ( 3000 ( 0 ) ) 1 1 ∠0 x ( 20 ) 1 -------------.5 ) ) These can then be multiplied by the input components to find output components.8 Given the transfer function for the system. x ( 0 ) = ( 0.34 × 10 sin ( 0. The equations for transmissibility and isolation is shown in Figure 9. or possibly lead to premature wear and failure in mechanisms.992  402497 ∠2.34 × 10 ∠– 0. x(ω ) 1 -----------.709 Therefore the output is.= -----------------------------------------------------------------------------.= --------------------------------------------. 20 ∠0 = 0.= ----------------------------------2 F ( 20 ) – 398000 + j60000 402497 ∠2.992 –3 1∠0 x ( 0.2) 9.5 ) 1750 + j1500 2305 ∠0.= 0.= -----------------------------------------------------------------------------------.709 ( 2000 – 1000 ( 0. x ( t ) = 0. The calculation is easy to perform with a transfer function or Bode plot.5 + 49.497 ×10 ∠– 2. These equations can compare the ratio of forces or displacements through an isolator. x(0 ) 1 1 ----------. 10 ∠0 = 4. F ( t ) = 1000 + 20 sin ( 20t ) + 10 sin ( ( 0.= ---------------------------------------------------------------------2 F(ω) ( 2000 – 1000ω ) + j ( 3000ω ) and an input with multiple frequency components. the transfer function for each frequency can be calculated.5t ) ( N ) ) Note: these are gains and phase shifts that will be used heavily in Bode plots later.9.709 ) –6 –3 Figure 9. .5t – 0.= ----------------------------2 F ( 0.7 × 10 sin ( 20t – 2.5 ) =  ----------------------------.5 ) ) + j ( 3000 ( 0.8 A example for a signal with multiple frequency components (based on the example in Figure 9.0005 ∠0 )1000 ∠0 = 0.9.= ------------------------------.992 ) + 4.5 ) 1 1 1 ∠0 --------------. A common approach to dealing with these problems is to design vibration isolators.= --------------------------------------------------------------------------------------.

= ----------------F in ( ω ) x in ( ω ) %I = ( 1 – T )100% The gain of a transfer function gives transmissibility Figure 9. find a value of K that will give 50% isolation for a 10Hz vibration.9 Given a vibration force in. x out 5D + 10 -------.9.10 Drill problem: Select a K value .9 Transmissibility Given the transfer function for a vibration isolator below.= ---------------------------------------2 x in 4D + 20KD + 4 Figure 9. x out ( ω ) Fout ( ω ) T = -----------------.frequency response . to a force out.

4 SUMMARY • Phasor transforms and phasor representations can be used to find the steady state response of a system to a given input. model with a weight of 1.2 kN and a stiffness of 340 N/m has a steady-state harmonic excitation force applied at 95 rpm (revolutions per minute).10 9.5 PRACTICE PROBLEMS 1.frequency response . one at at each corner of a piece of equipment. Calculate the mass of the equipment and determine the amount of isolation the springs would afford if the equipment operating frequency is twice the natural frequency of the system. 9. • Vibration analysis determines frequency components in mechanical systems.9. 1mH 1KΩ + Vi + 1KΩ 1KΩ 1KΩ Vo - 2.o.000 t) using phasor transforms. Four helical compression springs are used. a) For the circuit below find the transfer function and the steady state response for an input of . 4.f. Develop a transfer function for the system pictured below and then find the response to an input voltage of Vi = 10sin(1. A single d. The spring rate is 240 N/m for each spring and the vertical static deflection of the equipment is 10mm.000. What damper value will give a vibration isolation of 92%? 3.

e. . with homogeneous and particular solutions).frequency response . 1K 1uF 1mH Vi + + + Vo - b) Verify the results in question 1 by solving the differential equation below (i.9.11 Vi=5sin(1000t)V.

001D + 1KΩ V o  ------------------------------------ = V a   1KΩ (1) (2) substitute (2) into (1) 3 1 0.12 9.+ ----------------- + V i  ----------.+ ----------.= 0 1KΩ 1KΩ 1KΩ 0.001D  1KΩ  0.001D  0.+ ----------.= 0 0.001D 0.001D  1KΩ  0.001D + 1KΩ  ----------.9.  0.001D 1KΩ 1 1 1 V o  ----------------. = V o  -----------------  1KΩ 0. = V o  -----------------    1KΩ 0.+ ----------------- –  -----------------  ------------------------------------  1KΩ 0.+ ----------------- + V i  ----------.001D + 1KΩ –1 1 Vo  ------------------------------------  ----------.+ ----------------.= -----------------------------------------------------------------------------------------------------------------Vi 3 1 1  0.frequency response .001D 1KΩ .001D 1KΩ 1 ----------Vo 1KΩ ----.001D 1KΩ  0.001D Vo – Va Vo = ----------------. = V a  ----------------.001D 3 1 –1 1 V a  ----------.+ ----------. 1KΩ Va 1mH Vi + 1KΩ 1KΩ 1KΩ + Vo - ∑ I Va ∑ I Vo Va – V Va Va Va – Vo = ----------------i + ----------.6 PRACTICE PROBLEM SOLUTIONS 1.

---------------------------.2  y   F F D M + DK d + K s Ffloor – ( K s + K d ωj ) -----------.= ----------------------------------------------------------------------------------–3 –3 –3 10 + 0j ( j + 1 ) ( 3 × 10 – j10 ) + j ( 10 ) Vo 1 ---------------.=  -----------.= ---------------------------------------------2 F – ω M + ωjK d + K s Ks + ( Kd ω ) F floor -----------.001 ( j10 ) –3 Vo 10 ---------------.001j10 + 1KΩ 1  ------------------------------------------  ----------.644 )V 2.= ----------------------------------------------------------------------------------------------------------------------------------------------6 10 + 0j 3 1 0.=  ------------.= --------------------------------------10 + 0j ( j + 1 )( 3 – j ) + j Vo 1 .frequency response .  -.= K s + K d D y Ffloor –( K s + K d D ) Ffloor y -----------.= ----------------------------------------------------------2 F 2 2 ( K s – ω M ) + ( ωK d ) 2 2 . = ----------------------------------------.13 for the given input of Vi(t) = 10sin(1.00 sin ( 10 t – 0.000 t).  4 – 3j  4 + 3j  4 – 3j 10 + 0j 4 – 3j Vo = 10  ------------- = 2 ∠– 0. F y M 6 ∑ Fy = – K s y – K d yD – F = MyD 2 y –1 -. 6      1KΩ 6  1KΩ 0.001 ( j10 ) 0.= ----------------------------------------2 F D M + DK d + K s Ks Kd F floor = K s y + Kd yD F floor -----------.000. 1 ----------Vo 1KΩ ---------------.644  25 - V o ( t ) = 2.9.+ ---------------------------- –  ---------------------------.

BODE PLOTS Topics: • Bode plots Objectives: • To be able to describe the response of a system using Bode plots 10.1 10. At different frequencies the transfer function value will change. . In the previous chapter these values were calculated for a single frequency and then multiplied by the input values to get an output value. The magnitude is the gain.10.1 INTRODUCTION When a phasor transform is applied to a transfer function the result can be expressed as a magnitude and angle that are functions of frequency. and the angle is the phase shift. The transfer function gain and phase angle can be plotted as a function of frequency to give an overall picture of system response.bode plots .

2K 3. The magnitude of the resulting transfer function is the gain. Note that to correct for the quadrant of the phase shift angles pi radians is subtracted for certain frequency values.7K-20K gain f(Hz) Figure 10. and might sound somewhat muffled.2. In this example the transfer function is multiplied by the complex conjugate to eliminate the complex number in the denominator. and the phase shift is the angle. The position of the sliders adjusts the envelope that the audio signal is filtered through. The spectrum can be adjusted so that high or low tones are emphasized or muted. The sliders trace out a Bode gain plot.1 Commonly seen Bode plot The mass-spring-damper transfer function from the previous chapter is expanded in Figure 10. and the bode plot shifts.bode plots . 40-120 120-360 360-1K 1K-3. The high frequencies would not be passed clearly. In theoretical terms the equalizer can be described with a transfer function.7K 9.2 Aside: Consider a ’graphic equalizer’ commonly found on home stereo equipment. In the example below the slides are positioned to pass more of the lower frequencies.10.2K-9. . As the slides are moved the transfer function is changed.

= ----------------------------------------------------------------------------. Also note that the frequencies are changed in multiples of tens. The magnitude of the output wave can be calculated by multiplying the input wave magnitude by the gain. (Note: recall this example was used in the previous chapter) The phase angle can be added to the input wave to get the phase of the output wave.= ---------------------------------------------------------------------2 F(ω) j ( 3000ω ) + ( 2000 – 1000ω ) x(ω ) 1 -----------. or magnitudes. Gain is normally converted to ’dB’ so that it may cover a larger range of values while still remaining similar numerically. These need to be converted from Hz to rad/s before use. An example of this type of analysis is done in Figure 10.2 are normally left in variable form so that they may be analyzed for a range of frequencies.3 x( ω) 1 -----------. A set of frequencies is used for calculations.bode plots .10.2 A phasor transform example for ( ω ≤ 2 ) for ( ω > 2 ) 2 2 2 2 ( – j ) ( 3000ω ) + ( 2000 – 1000ω ) -----------------------------------------------------------------------------2 ( – j ) ( 3000ω ) + ( 2000 – 1000ω ) 2 The results in Figure 10.3.= ---------------------------------------------------------------------2 F( ω ) j ( 3000ω ) + ( 2000 – 1000ω ) ( 2000 – 1000ω ) – j ( 3000ω ) x(ω) -----------. .= ------------------------------------------------------------------------F(ω) 2 2 2 ( 2000 – 1000ω ) + ( 3000ω ) ( 2000 – 1000ω ) + ( 3000ω ) 1 x(ω ) -----------.= -----------------------------------------------------------------------------2 2 2 F( ω ) 2 2 2 ( 2000 – 1000ω ) + ( 3000ω ) ( 2000 – 1000ω ) + ( 3000ω ) – 3000ω – 3ωθ = atan  ----------------------------------------- = atan  ------------    2 2 – 1000ω + 2000 2–ω – 3ωθ = atan   ------------ – π   2 2 –ω Figure 10. The gain gives a ratio between the input sine wave and output sine wave of the system. For each one of these the gain and phase angle is calculated.

83 628.4.) θ (deg. where an input waveform is supplied with three sinusoidal components. Figure 10. If the input and output are sinusoidal.001 0. and x is the resulting output.06283 0. relative to the input wave. Note: The frequencies chosen should be chosen to cover the points with the greatest amount of change.006283 0.01 0. Therefore F is the input to the system.3 6283 Gain Gain (dB) θ (rad.bode plots .10. there is a difference in phase between the input and output wave of θ (the phase angle). These are then used to calculate the resulting output wave.1 1 10 100 1000 (rad/sec) 0 0. The resulting output represents the steady-state response to the sinusoidal output. The gain means that for each unit of F input to the system. .3 A phasor transform example (continued) In this example gain is defined as x/F. To help keep the graphs rational we will "compress" the values by converting them to dB (decibels) using the following formula.6283 6.283 62.) Note: the gain values will cover many magnitudes of values. For each of the frequencies a gain and phase shift are calculated. gain db = 20 log ( gain ) Note: negative phase angles mean that the mass motion lags the force. This is shown in Figure 10. there will be gain*F=x output.4 f(Hz) 0 0.

as defined by the equation. the next major division would be 1. F ( t ) = 1000 + 20 sin ( 20t ) + 10 sin ( 0. These plots are normally done on semi-log graph paper. . Along the linear axis (the short one) the gains and phase angles are plotted. such as that seen in Figure 10. This means that if the paper started at 0. Along the longer axis of this paper the scale is logarithmic (base 10).5.5t – ________  --------    360  gain from calculations or Bode plot phase angle from calculations or Bode plot x ( t ) = 0.5 + ______ sin ( 20t + _____ ) + _______ sin ( 0. and finally 1000 on the other side of the paper.5t + ______ ) Figure 10. then 10.2 BODE PLOTS In the previous section we calculated a table of gains and phase angles over a range of frequencies. Graphs of these values are called Bode plots.10. The basic nature of logarithmic scales prevents the frequency from being zero. then 100. normally with two graphs side-by-side on a single sheet of paper.5 Assuming there is excitation from two sinusoidal sources in addition to the static load.bode plots .5t ) ( N ) The result for each component can be evaluated separately · 2πx ( t ) = 1000 ( _________ ) sin  0t + ______  --------    360  2π+ 20 ( _________ ) sin  20t + _______  --------    360  2π+ 10 ( __________ ) sin  0.4 A phasor transform example (continued) 10.1 on one side.

6 .10.1 2 3 4 5 6 789 1 2 3 4 5 6 789 1 2 3 4 5 6 789 1 2 3 4 5 6 789 bode plots .

7 Figure 10. gain (dB) -40dB/dec phase (deg) 0.1 1 10 f(Hz) 100 1000 Figure 10.6 Drill problem: Plot the points from Figure 10.5 4 cycle semi-log graph paper Plot the points from Figure 10. The general layout is pictured below.5.10.3 in the semi-log graph paper in Figure 10.3 on graph paper .bode plots .

such as Mathcad or a spreadsheet.2 with a computer .8 Drill problem: Plot the points from Figure 10.10. and then draw Bode plots. to calculate the points in Figure 10. for step = 0:(steps_per_dec * decades). phasedeg).1.// find the gain and convert it to mag and angle gaindb = 20 * log10(gain). f = start_freq * 10 ^ (step / steps_per_dec). gaindb. ’Sample Transfer Function’). %f \n". "w").// convert magnitude to dB phasedeg = 180 * phase / %pi. bode(h. start_freq = 0.1. phase] = polar(G(w)).2. Figure 10. 1000. // the transfer function function foo=G(w) D = %i * w.txt".bode plots . "%f. // convert the frequency to radians [gain. f. endfunction // this section writes the values to a datafile that may be graphed in a spreadsheet fd = mopen("data. Figure 10. foo = (D + 5) / (D^2 + 100*D + 10000). %f. ’D’). h = syslin(’c’.8 steps_per_dec = 6. // to graph it directly the following is used D = poly(0.// convert to degrees mfprint(fd. // calculate the next frequency w = f * 2 * %pi). (D + 5) / (D^2 + 100*D + 10000) ) . decades = 6. 0.7 Aside: Bode Plot Example with Scilab Use computer software. end mclose(fd). Most software will offer options for making one axis use a log10 scale.

bode plots .9 Drill problem: Draw the Bode plot for gain and phase . D+3 ----------------------------------------------------2 D + 10000D + 10000 Figure 10.10.9 Draw the Bode plot for the transfer function by hand or with computer.

1.20dB/decade for each term. Also note that squared (second-order) terms would have a slope of +/-40dB/decade.bode plots .01=-40dB. When the smooth curve is drawn. An initial gain is also calculated to shift the results up or down. Once in that form.10 An approximate technique for constructing a gain Bode plot is shown in Figure 10. When done. At each corner frequency a numerator term causes the graph to turn up. Plot the straight line pieces. A smooth curve is then drawn over top of this curve. a) The gain at 0rad/sec is calculated and used to find an initial offset. b) Put the transfer function in root form to identify corner frequencies. Figure 10. 2. This method involves looking at the transfer function and reducing it to roots in the numerator and denominator.10. In second-order systems the damping coefficient make may the corner flatter or peaked. a straight line approximation for each term can be drawn on the graph. For example (D+1)/(D+1000) will have corner frequencies at 1 and 1000 rad/sec. Bode plots for transfer functions can be approximated with the following steps. c) Curves that turn up or down are drawn for each corner frequency. The slope up or down is generally +/. each term in the dominator causes the graph to turn down. there should commonly be a 3dB difference at the corner frequencies.10 The method for Bode graph straight line gain approximation . the straight line segments are added to produce a more complex straight line curve. The effect of each term is added up to give the resulting straight line approximation. For example this transfer function starts at 10(D+1)/ (D+1000)=10(0+1)/(0+1000)=0. 3.10.

= 20 log ( ω ) = – 20 log ( ω ) jω 1----ωc ωc = ω 1 π 1 = --------------------.= ------------.= – 3. and the phase angle is 45 degrees of the way to +/. Figure 10. ∠– -. thus resulting in a gain change of -20 dB.bode plots .01dB ∠– - 2 4 4 G(0) 2 4 Why does a first order pole go down at 20dB/dec? 1 G ( jω ) = -----------------ω c + jω before the corner frequency.∠– -.11 Note: Some of the straight line approximation issues are discussed below. Why is there 3dB between a first order corner and the smooth plot.11 Why the straight line method works .= 20 log  -----.10. ω c > jω –1 1G ( jω ) = ----.= 20 log ( ω c ) = – 20 log ( ω c ) ωc after the corner frequency. the log value becomes 1 larger. ω c < jω –1 1G ( jω ) = ----.= ω c + j0 at the corner frequency 1 G ( jω ) = -------------------ω c + jω c Therefore the difference is π π G ( jω ) 1. 1 G ( jω ) = -----------------ω c + jω the initial gain is 1 G ( 0 ) = ---------------.90 degrees.∠– -π ωc 2 4 -ωc 2 ∠ 4 each time the frequency increases by a multiple of 10.π 1diff = -------------.= -----.

10. ω c < jω π ωangle ( G ( jω ) ) = – atan  ----- = – atan ( ∞ ) = – -ω  2 c Figure 10.12 Why the straight line method works (cont’d) .bode plots .12 Why does a pole make the phase angle move by -90deg after the corner frequency? 1 1 ∠0 1 ωG ( jω ) = -----------------. = 0 ω  ω  c c after the corner frequency. ω  ω c + jω 2 2 2 ω 2 c ω c + ω ∠ atan  ----. ωc + ω ω c before the corner frequency.∠– atan  ----. ω c > jω ω 0angle ( G ( jω ) ) = – atan  ----- = – atan  ----.= --------------------------------------------------.= ----------------------.

. 1. In total there are three roots.13 Show that the second order transfer function below would result in a slope of +/.40 dB/decade. The single root in the numerator will cause the curve to start upward with a slope of 20dB/dec after 1rad/sec. In this example the transfer function is first put into a root form. but if Hz are used then it is necessary to convert the values.14. An example of the straight line plotting technique is shown in Figure 10. The initial gain of the transfer function is also calculated. The two roots will cause two curves downwards at -20dB/dec starting at 10 and 100 rad/sec.13 Drill problem: Slope of second order transfer functions. The frequency axis is rad/sec by default. 10 and 100 rad/sec.10.bode plots . and converted to decibels. 1 G ( jω ) = ------------------------2 ( ω c + jω ) Figure 10.

0dB 1 10 100 freq.bode plots .14 100D + 100 10 ( D + 1 ) G ( D ) = -------------------------------------------------.01D + 0.10.11D + 10 4 (the equation is put in root form) Step 1: Draw lines for each of the terms in the transfer function.14 An approximate gain plot example The example is continued in Figure 10. .(rad/sec) (could be Hz also) -20dB/dec.= 10 = 20dB ( 10 ) ( 100 ) 4 20dB/dec. -20dB/dec.= --------------------------------------------2 ( D + 10 ) ( D + 100 ) 0. 1 --------------D + 10 1 -----------------D + 100 Figure 10. Gain (dB) D+1 10 ( 1 ) GAIN 0 = -----------------------.15 where the straight line segments are added to produce a combined straight line curve.

15 Step 2: Sum the individual lines.10. Higher order functions will be discussed later. Gain (dB) 40dB 20dB 0dB 1 10 100 freq. Step 3: Draw the smooth curve (leaving 3dB at the corners).(rad/sec) . For a simple first-order term there is a 3dB gap between the sharp corner and the function. Gain (dB) 40dB 3dB 20dB 0dB 1 10 100 freq. and get the straight line approximation.(rad/sec) Figure 10.bode plots .15 An approximate gain plot example (continued) Finally a smooth curve is fitted to the straight line approximation. When drawing the curve imagine that there are rubber bands at the corners that pull slightly and smooth out.

(rad/sec) 1 -----------------D + 100 4 1 --------------D + 10 Figure 10.17 10 ( D + 1 ) G ( D ) = --------------------------------------------( D + 10 ) ( D + 100 ) Step 1: Draw lines for each of the terms in the transfer function Phase angle (deg. the concept of an rubber band will smooth the curve. Again. .) D+1 +90 +45 0 -45 -90 0.1 1 10 100 1000 freq.10. A smooth line approximation is drawn using the straight line as a guide.19 to produce a straight line approximation of the final plot.18 An approximate phase plot example The straight line segments for the phase plot are added in Figure 10.bode plots .

This can be handled by determining the damping coefficient and natural frequency as shown in Figure 10.19 An approximate phase plot example (continued) The previous example used a transfer function with real roots.1 1 10 100 1000 freq.) +90 +45 0 -45 -90 0.707 to 0 where the peak will be infinite. The peak will occur at the damped frequency. If the roots for the transfer function are complex (underdamped the corner frequencies will become peaked. In a second-order system with double real roots (overdamped) the curve can be drawn with two overlapping straight line approximations.20.(rad/sec) 0.bode plots .1 1 10 100 1000 freq.10.18 Phase angle (deg. . The peaking effect will become more pronounced as the damping coefficient goes from 0.(rad/sec) Figure 10.) +90 +45 0 -45 -90 Phase angle (deg.

1 =0.bode plots . and the Bode plots get a peak.19 Gain Mn Resonant Peak =0. This can be seen mathematically because the roots of the transfer function become complex. In addition these curves provide an understanding of the system that makes design easier. a designer will often describe a system with a Bode plot. Figure 10. .707 -40dB/decade ωd freq A ------------------------------------------------------2 2 D + ( 2ξω n )D + ( ω n ) ωd = 1 – ξ ωn 2 dB = A Note: If ζ is less than 1 the roots become complex.20 Resonant peaks The approximate techniques do decrease the accuracy of the final solution. and then convert this to a desired transfer function. For example.10. but they can be calculated quickly.5 ξ = 0.

10.20 Draw the straight line approximation for the transfer function.21 Drill problem: Draw the straight line approximation .bode plots . D+3 ----------------------------------------------------2 D + 10000D + 10000 Figure 10.

The three frequency components are clearly identifiable spikes.0005  5 t (ms) frequency 1 freq.t + 0. 3 freq.10.22. The overall sinusoidal shape is visible. To determine the function other tools are needed to determine the frequencies.005   0.t + 0.01A sin  --------------.01   0.3A sin  -----------. along with a significant amount of ’noise’.21 10. x(t) displacement A 2π2π 2π x ( t ) = A sin  --------.22 is shown in the spectrum in Figure 10.23.22 A vibration signal as a function of time A signal spectrum displays signal magnitude as a function of frequency. When this is considered in greater detail it can be described with the given function. The time based signal in Figure 10. The height of the peaks indicates the relative signal magnitude. instead of time. and magnitudes of the frequency components. .t  0.bode plots . 2 Figure 10.3 SIGNAL SPECTRUMS If a vibration signal is measured and displayed it might look like Figure 10.

= ( D + 10 ) ( D + 5 ) 2 x(D) (D + 5) a) draw the straight line approximation of the Bode (gain and phase shift) plots. ( D + 1 ) ( D + 1000 ) --------------------------------------------2 ( D + 100 ) AND 5----2 D 2.4 SUMMARY • Bode plots show gain and phase angle as a function of frequency. Given the transfer function below.23 The spectrum for the signal in Figure 10. y (D) ------------------------------------------------. Draw a Bode Plot for both of the transfer functions below using the phasor transform.3) using the .0005 =2000Hz Figure 10.01 =100Hz 1 -----------0. b) determine the steady-state output if the input is x(t) = 20 sin(9t+0.5 PRACTICE PROBLEMS 1.22 Amplitude A log(freq. • Bode plots can be constructed by calculating point or with straight line approximations.bode plots .22 10.) (Hz or rads.10. • A signal spectrum shows the relative strengths of components at different frequencies.) 1--------0.005 =200Hz 1 --------------0. 10.

23 Bode plot.10. Given the transfer function below. phasor transforms) if the input is Vi=5sin(100. b) Draw an approximate Bode plot for the circuit. The applied force ‘F’ is the input to the system. Neglect the effects of gravity. c) Use the Bode plot to find the response to an input of 5sin(624t) + 1sin(6. K1 = 500 N/m K2 = 1000 N/m M = 10kg x F b) What is the steady-state response for an applied force F(t) = 10cos(t + 1) N ? c) Give the transfer function if ‘x’ is the input.. Vo 1000 ----.= ----------------------------------2 x D + 5D + 100 4. d) Draw the bode plots for the transfer function found in a). 3. b) Verify the Bode plot by calculating values at a few points. and the output is the displacement ‘x’. F D + 1000 G = -. D ( D + 2π ) --------------------------------------------------------2 2 D + 300πD + 62500π a) Use the straight line method and the attached log paper to draw an approximate Bode plot.= --------------------D + 1000 Vi a) Find the steady state response of the circuit using phasors (i. given F(t) = 10N for t >= 0 seconds. 6. For the transfer function. 5. a) find the transfer function.000t).bode plots .2t). e) Find x(t). Use the straightline approximation techniques to draw the Bode plot for the transfer function below.e. .

d) Find the frequency response (gain and phase) for the transfer function using the phasor transform.bode plots . 7.010 f) Give the expected ‘x’ response of this first-order system to a step function input for force F = 1N for t > 0 if the system starts at rest. The following differential equation is supplied.10.008 0. F ( t ) = 10 sin ( 100t ) e) Put the differential equation in state variable form and use a calculator to find values in time for the given input.0 0.002 0. b) Solve the differential equation numerically. with initial conditions. 8.004 0. given F(t) = 10N for t >= 0 seconds considering the effects of gravity. d) Use the Bode plot to find the response to. c) Draw a Bode plot for the system using either approximate or exact techniques. dF ( t ) = 10 sin ( 100t ) 5x +  ---- x = F  dt a) Write the transfer function for the differential equation if the input is F.006 0.24 f) Find x(t). You are given the following differential equation for a spring damper pair. ·· · · y + y + 7y = F y(0 ) = 1 y(0) = 0 F ( t ) = 10 t>0 a) Write the equation in state variable form. Hint: Use the canonical form. Sketch the bode plots. . F = 10 sin ( 100t ) x t 0. b) Apply the phasor transform to the transfer function to find magnitude and phase as functions of frequency. c) Solve the differential equation using calculus techniques.

20dB 0dB -20dB 1----2π 90deg 100 -------2π 1000 ----------2π 0deg 0dB -40dB/dec 5 -----2π -45deg -180deg .25 10.6 PRACTICE PROBLEM SOLUTIONS 1.10.bode plots .

11 sin ( 9t – 0.236 – 0.3312 ) ) y ( t ) = 26.=  9j + 10  5 – 9j = 50 + 81 – 45j = 1.307 ) sin ( 9t + 0.------------ 9j + 5   5 – 9j x (D + 5) 25 + 81 y .425j ------------------------------------------------- ---------------. = 1.1 – 0.425j ----------------------------------------------------------------.33dB ∠– 9.3312rad  1.91j y ---------------------------.3 + ( – 0.= ( D + 10 .91j ) = 26.813j = 26.425j ) ( 19.3 ) Aside: the numbers should be obtained from the graphs.91j y ( t ) = 26.425 ∠ atan  --------------.= ( D + 10 ) ---------------------------------------------------------2 x(D) (D + 5) (D + 5) 6dB 0dB 0.236  y .425 1.3rad ) + j sin ( 0.1 + 5.26 2. y (D) ) ----------.1 + 5.=  9j + 10  5 – 9j = 50 + 81 – 45j = 1.8 1.1 sin ( 9t – 0.= 2.3rad ) ) = 19.6 1.236 + 0.6 -90 x ( t ) = 20 cos ( 9t + 0.1 ∠– 0.49° x y ( t ) = 20 ( 1.bode plots . but I have calculated them y ) .236 – 0.= ( 1.236 – 0.1 + 5.  ------------- x ( D + 5) 25 + 81 9j + 5 5 – 9j x = 20 ( cos ( 0.= ( D + 10 ) ( D + 5 .8 90 0.= ( D + 10 .031 ) .031 ) Aside: This can also be done entirely with phasors in cartesian notation y ) .10.307 ∠– 0.031 19.= x 2 2 – 0.

= ( D + 10 .45 ∠0.30 ∠1.bode plots .3 y = 1.3 ) = 26.064 10.31 ∠–0.3 y ----------------.733 – 1.27 (cont’d Aside: This can also be done entirely with phasors in polar notation y ) .=  9j + 10 = 13.331 20 ∠0.733 = 13.45 ∠0.2 sin ( 9t – 0.031 y ( t ) = 26.31 ∠– 0.= 1.331 --------------------------------------------------------------------------- 9j + 5  x (D + 5) 10.331 + 0.31 ( 20 ) ∠( – 0.10.30 x = 20 ∠0.2 ∠– 0.031 ) .064 = 1.

54 10 15.10.59Hz s 2ω n ζ = 5 5 ∴ζ = ------------.154 1 1.= 0.54Hz for the initial gain Gain (dB) 20 0 0.1 -90 0. for the numerator. (poles) 1000 rad = 159Hz -------s D + 2ω n ζD + ω n = D + 5D + 100 ω n = 100 ∴ω n = 10 rad -------f n = 1.54 10 100 -40dB/dec -60 -80 phase (deg) 0 0.1 1 1. (zero) for the denominator.28 3.bode plots .= 10 = 20dB 2 0 + 5 ( 0 ) + 100 2 2 2 2 2 -20dB/dec -180 .15.9 100 159 1000 1590 f(Hz) 159 1000 f(Hz) 0 + 1000 G ( 0 ) = ------------------------------------.25 (underdamped) 2 ( 10 ) f d = f n 1 – ζ = 1.4.

a) b) c) 8.561 ) 16Hz 159Hz 1.6KHz -20dB/dec 5 0 deg -90 deg 5.141 ×10 sin ( 6.89 sin ( 624t + 1.bode plots .050 sin ( 10 t – 1.2t + 2.571 ) + 0.29 4. a) b) Vo ( t ) = 0.10.356 ) –3 .

0667 -.0667 m .10.---. a) x 0.0207 cos ( t + 1 ) m c) 2 F -.07 × 10 –3 ∠0rad m --( 10 ∠1rad )N N –3 m ∴x ( ω ) = 2.07 × 10 –3 ∠0rad m --2 F N ( 1j ) + 33.= ---------------------.3 -----------------------.07 × 10 ∠0rad --.( 10 ∠1rad )N N –3 ∴x ( ω ) = ( 10 )2.m 2 F N D + 33.0667 .= ---------------------------.N .x 0.= 2.30 6.3 b) ω = 1 ∴D = 1j x 0.= D + 33.= 2.bode plots .3 F ( t ) = 10 cos ( t + 1 ) N ∴F ( ω ) = ( 10 ∠1rad )N x ------------------------------.07 × 10 ∠( 0rad + 1rad )m ∴x ( ω ) = 0.0207 ∠1radm ∴x ( t ) = 0.

1 -90 1 10 -20dB/dec f(Hz) 100 -180 e) f) x ( t ) = ( – 0.31 d) for the denominator.77t ) + 0.3 -134 phase (deg) 0 0.918Hz 2 2 2 (undamped) 2 f d = f n 1 – ζ = 0.= 2 = – 54 dB 33.020 cos ( 5.2162 )m .bode plots . (poles) D + 2ω n ζD + ω n = D + 33.0667 G ( 0 ) = --------------.2162 cos ( 5.918Hz for the initial gain Gain (dB) -54 0 0.77 -------s ζ = 0 fn = 0.10.77t ) + 0.3 rad ω n = 5.020 )m x ( t ) = ( – 0.1 -94 1 10 -40dB/dec f(Hz) 100 ∞ 0.

598t – 0.000 m --s cos ( 2.10. until = 1 + 0..437e – 0.bode plots .429 )m .428m 0.001s y1 v1 etc.32 7.5t b) = +h 0 1 yi + 0 –7 –1 vi F F = 10 = 1 0 = 1. a) · y = v · v = – v – 7y + F yi + 1 vi + 1 given yi vi y0 v0 using h=0.001 0 1 1 + 0 0 –7 – 1 0 10 y 100 v 100 c) y ( t ) = ( – 0.19 ) + 1.

= ∠( 0 – angle ( ω.01Hz 1Hz 100Hz -180deg .= ---------------------------------.= -------------------------.= ------------------------------------2 2 2 F D +D+7 ( jω ) + jω + 7 (7 – ω ) + j(ω) 1 y -.33 d) y 1 1 1 -. 7 – ω ) 2 ∠angle ( ω. 7 – ω ) -20dB -40dB 0. 7 – ω ) ) = ∠– angle ( ω.= --------------------------------------F 2 2 2 (7 – ω ) + ω ∠0 2 2 ∠θ = ---------------------------------------------.10.01Hz 1Hz 100Hz 0deg 0.bode plots .

521 ) = 0.bode plots .34 8.01 phase = – 90° = – π rad -2 π x ( t ) = 10 ( 0.521 ------- 5  10 ∠0 2 2 5 + 100 x = ( 10 ∠0 ) ( 0.∠– atan  ω -- 5 F 5+D 5 + jω 2 2 2 2  ω 5 +ω 5 + ω ∠ atan  --- 5 gain(dB) -14dB 1 initial gain = 20 log  ----------.= 0.08Hz 0.∠– atan  100 = 0.8Hz 100 f = -------.= 16Hz 2π From the Bode plot.= ------------F 5+D x 1 1 1 ∠0 1 -.00999 ∠– 1.01 ) sin  100t – --  2 Aside: verified by calculations.0999 ∠–1.521 x ( t ) = 0.= ------------.521 ) .10.8Hz . 2π -20dB/dec c) phase(deg) 0° – 45° – 90° 0.= -------------------------.= --------------------. = – 14dB  5 + 0 5 corner freq = ----.= ----------------------------------------------. a) b) x 1 -. 8Hz d) gain = – 40dB = 0.00999 ∠– 1.= -------------.0999 sin ( 100t – 1. x 1 ------------.

0 x 0 9.010 . y (D) ----------.026 0.186 0.– -------5 5 – 5t 10.004 0.006 0.35 e) 0.002 0.015 0.099 0.7 ASSIGNMENT PROBLEMS 1.02π ) ( D + 2π ) -----------------------------------------------------2 2 D + 50πD + 10000π 2.92e-3 0. Draw Bode plots for the following functions using straight line approximations.2 t 0. b 0 0 0. 1 -----------D+1 1 --------------2 D +1 1 -------------------2 (D + 1) 1 ----------------------------2 D + 2D + 2 3.5 i h 1 0.008 0. For the transfer functions below.2 f) 1 e x = -.10. ( D + 2π )D --------------------------------------------------------( D + 200π ) ( D + 0.3) using the striaght line plots.999 0.041 Xi 0 0. Given the transfer function below. draw the bode plots using computer software.98e-4 5.bode plots .= ( D + 10 ) ( D + 5 ) --------------------------------------2 x(D) (D + 5) a) draw the straight line approximation of the bode and phase shift plots. c) use an exact method to verify part b). b) determine the steady state output if the input is x(s) = 20 cos(9t+.

and if more cycles are required.36 10. as well at most large office supply stores. make additional copies if required. 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 .10.8 LOG SCALE GRAPH PAPER Please notice that there are a few sheets of 2 and 4 cycle log paper attached. sheets can be cut and pasted together.bode plots . Also note that better semi-log paper can be purchased at technical bookstores.

1 2 3 4 5 6 789 1 2 3 4 5 6 789 1 2 3 4 5 6 789 1 2 3 4 5 6 789 bode plots .10.37 .

.2 ROOT-LOCUS ANALYSIS In a engineered system we may typically have one or more design parameters. 11. and a time response is shown to the right. This is generally undesirable and possibly unsafe. If the roots are real the result will be e-to-the-t terms. 11. Recall the general solution to a homogeneous differential equation.root locus . These roots are shown on the complex planes on the left.7. Root-locus plots are used to plot the system roots over the range of a variable to determine if the system will become unstable.1 11. The rule is that any roots on the right hand side of the plane make a system unstable. or user settings. ROOT LOCUS ANALYSIS Topics: • Root-locus plots Objectives: • To be able to predict and control system stability.11. Complex roots will result in a sinusoidal oscillation. adjustments. It is important to determine if any of these will make the system unstable. or a high speed aircraft that fails due to resonant vibrations. think of a washing machine that vibrates so much that it ‘walks’ across a floor. while positive roots will result in terms that grow exponentially and become unstable.1 INTRODUCTION The system can also be checked for general stability when controller parameters are varied using root-locus plots. Notice that in these figures (negative real) roots on the left hand side of the complex plane cause the response to decrease while roots on the right hand side cause it to increase. as shown in Figure 11. For example.1 to Figure 11. or oscillate. Also note that the complex roots cause some amount of oscillation. Consider the roots of a second-order homogeneous differential equation. If the real roots are negative then the terms will tend to decay to zero and be stable.

2 Complex roots make a system oscillate R = – A ± Bj jw B -A sigma -B t x(t) Figure 11.root locus .1 Negative real roots make a system stable R = ± Aj jw A x(t) sigma -A t Figure 11.3 Negative real and complex roots cause decaying oscillation .2 R = – A.11. – B jw x(t) x0 -A -B sigma t Figure 11.

A jw x(t) A.3 R = – A ± Bj jw B -A sigma -B t x(t) Figure 11.A sigma t Figure 11.-A sigma t Figure 11.11.root locus .5 Overlapped roots are possible R = A.6 Positive real roots cause exponential growth and are unstable .4 More negative real and complex roots cause a faster decaying oscillation R = – A. – A jw x(t) -A.

By analyzing the function in the denominator of a transfer function the general system response can be found. In this example the transfer function is found and the roots of the equation are written with the quadratic equation. recall that the denominator of a transfer function is the homogeneous equation. The range of values for the spring coefficient should be determined by practical and design limitations. At this point there are three unspecified values that can be manipulated to change the roots.8. For example.root locus . .4 R = A ± Bj jw B A sigma x(t) -B t Figure 11.7 Complex roots with positive real parts have growing oscillations and are unstable Next. and the spring value will be varied. The mass and damper values are fixed. the spring coefficient should not be zero or negative. An example of root-locus analysis for a mass-spring-damper system is given in Figure 11.11.

.± -----. In addition complex roots will indicate oscillation.5 x ( D) 1 -----------. These will show how the values of the roots change as the design parameter is varied.D + ----M M –K d Kd Ks -------. If any of these roots pass into the right hand plane we will know that the system is unstable. B = -------------------------------------2a C 0 2 2 Figure 11.root locus .8 A mass-spring-damper system equation The roots of the equation can then be plotted to provide a root locus diagram.– 4 ----2 M M M B. C = -----------------------------------------2 M 0 100 100 100 Kd 0 100 100 100 Ks 0 100 1000 10000 B 0 2 Ks Kd M Aside: ax + bx + c = ( x + A ) ( x + B ) – b ± b – 4ac A.= ----------------------------------------2 F( D) MD + K d D + K s Note: We want the form below A -------------------------------------( D + B )( D + C ) 1---x(D ) M -----------.= -----------------------------------F (D ) Ks 2 Kd D + ----.11.

and the system can go where it pleases. . so the system is stable and doesn’t oscillate. In real terms this will mean that the controller becomes unresponsive. Keep in mind that gain values near zero put the control system close to the right hand plane. The root of the denominator is then calculated and plotted for a range of ’K’ values.root locus . It would be advisable to keep the system gain greater than zero to avoid this region. The variable gain ’K’ necessitates the evaluation of controller stability over the range of operating values.10.9 Drill problem: Plot the calculated roots on the axes above A feedback controller with a variable control function gain is shown in Figure 11. In this case all of the roots are on the left side of the plane.6 Imaginary Real Figure 11.11. This analysis begins by developing a transfer function for the overall system.

11.. we use the characteristic equation of the denominator to find the roots as the value of K varies.= -------------------------. K  ---  D G(D) K G S ( D ) = -----------------------------------. make the system more stable.. root -0 -1 -2 -3 jω K→∞ K = 0 σ Note: This system will always be stable because all of the roots for all values of K are negative real.= ------------1 + G ( D )H ( D ) D+K K 1 +  --- ( 1 )  D Next. D+K = 0 K 0 1 2 3 etc.root locus . Also. and it will always have a damped response. we must develop a transfer function for the entire control system.7 Note: This controller has adjustable gain. Figure 11. These can then be plotted on a complex plane. Note: the value of gain ’K’ is normally found from 0 to +infinity. After this design is built we must anticipate that all values of K will be used. It is our responsibility to make sure that none of the possible K values will lead to instability. + K 1 K G ( D ) = --D H(D ) = 1 1 --D First. larger values of K.10 Root-locus analysis in controller design .

= – 1..8 Given the system elements (assume a negative feedback controller). K 0 1 2 3 roots jω σ Figure 11.5 ± ------------------2 2 Next. find the characteristic equation. K - 1 +  ----------------------------- ( 1 ) = 0  2 D + 3D + 2 D + 3D + 2 + K = 0 2 Note: For a negative feedback controller the denominator is.root locus . and an equation for the roots.11.11 Drill problem: Complete the root-locus analysis . find values for the roots and plot the values. K H(D ) = 1 G ( D ) = ----------------------------2 D + 3D + 2 First. 1 + G ( D )H ( D ) – 3 ± 9 – 4(2 + K) 1 – 4K roots = ----------------------------------------------.

12 Drill problem: Draw a root locus plot 11.root locus .9 K(D + 5) G ( D )H ( D ) = --------------------------------------2 D ( D + 4D + 8 ) Figure 11.11.3 SUMMARY • Root-locus plots show the roots of a transfer function denominator to determine stability .

Plot the Gain and magnitude as a function of frequency. Kd1 = 1 Ns/m Ks1 = 1 N/m Kd = 1 Ns/m M = 1 kg M = 1kg y F F Ks = 1 N/m y b) If the input force is a step function of magnitude 1N.root locus .10 11.0 + 1 -------------------2 ( D + 1) Kd D 2. and use it to find the time response for ‘y’ by solving a differential equation. c) Draw the poles on a real-complex plane. For each of the systems below. specify all points and values. d) Apply the phasor transform function and make it a function of frequency.11. + 3. Draw the root locus diagram for the system below. . convert the input to a transfer function.4 PRACTICE PROBLEMS 1. a) write the differential equation and convert it to a transfer function.

= --------------------------------2 θd D + D + 20K a) Draw the root locus diagram and state what values of K are acceptable. Given the system transfer function below. c) Given a gain of K=10 find the steady-state response to an input step of 1 rad.= ---------------------------------------------------------------------------------------------2 X D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i ) 2 a) Verify the close loop controller function given. A feedback control system is shown below. c) Draw a Bode plot for the feedback system if Kd=Kp=Ki=1. The closed loop transfer function is given. Identify any values of Kd that would leave the system unstable.11 3. 4. The block diagram below is for a motor position control system. b) Select a gain value for K that has either a damping factor of 0. d) Given a gain of K=0. The system incorporates a PID controller. 5. Use either the approximate or exact techniques.root locus .11. b) Draw a root locus plot for the controller if Kp=1 and Ki=1. b) Draw the Root-Locus diagram for the system (as K varies).707 or a natural frequency of 3 rad/sec.01 find the response of the system to an input step of 0. d) Select controller values that will result in a natural frequency of 2 rad/sec and .+ K d D D+9 D 4 D ( 3K d ) + D ( 3K p ) + ( 3K i ) Y -. The system has a proportional controller with a variable gain K. State if the Root-Locus diagram shows that the system is stable for the chosen K.1rad. c) Select a K value that will result in an overall damping coefficient of 1. θo 20K ---. Y X + 3 Ki -----------K p + ---. θd Vd + Ve Vs ω θa 2 K 100 -----------D+2 1 --D Va 2 a) Simplify the block diagram to a single transfer function.

e) For the parameters found in the last step can the initial values be found? f) If the values of Kd=1 and Ki=Kp=0.root locus .5.12 damping coefficient of 0. b) Draw a root locus diagram for the variable parameter ‘P’.1 -------------------------------------2 D + 10D + 100 Y 7. find the response to a unit ramp input as a function of time. including critical points. X + 1 K  5 + --. Vd + - e D + PD + 1 -----------------------------D Va 2 Vs 100 -----------------D + 100 ω 1 --D θ 2 a) Find the transfer function for the system. Draw a root locus plot for the control system below and determine acceptable values of K.11. Verify that the controller will be stable. c) Find the response of the system in Question #5 to a step input using explicit integration. The feedback loop below is for controlling a DC motor with a PID controller.+ D   D + 10 ----D 0. 6. .01 0.

732j -1.936j -1 +/. + 3.13 11.372 -0.000.0 = 0 Kd -10 -6 -2 -1 0 1 2 5 10 100 1000 roots 7. -11.5 PRACTICE PROBLEM SOLUTIONS 1.323j -2. 2.11. Any gains between -4 and -2 will result in oscillations.1. 0.0 ------------------------------------2 ( D + 1 ) + Kd D 3.039.5 +/.1. 2 2 2 2 2 2 Gains larger than -2 will result in a stable system.0 D + D ( K d + 2 ) + 4.657 -0.0 ) D = -------------------------------------------------------------------------2 – K d – 2 ± K d + 4K d – 12 D = ---------------------------------------------------------------2 Critical points: (this is simple for a quadratic) The roots becomes positive when 0 > – K d – 2 ± K d + 4K d – 12 2 + K d > ± K d + 4K d – 12 16 > 0 0 > – Kd – 2 Kd > –2 The roots becomes complex when 0 > K d + 4K d – 12 – 4 ± 16 – 4 ( – 12 ) K d = ---------------------------------------------2 K d = – 6.0 + Kd D 1 -------------------2 ( D + 1) + - 3. .536 2.0 -0.1.0 --------------------------------------------------2 D + D ( K d + 2 ) + 4.000 0 +/.5 +/. -102. -1000 2 – K d – 2 ± ( K d + 2 ) – 4 ( 4.628. -2. -6.root locus .343.2.000 -0.004.464.000j -0.000.

75t ) +  ----------------------.001 0.75  Note: a similar answer can be found with a single sin or cos 0.14 2.75t )  0.-1. a) 200K --------------------------------------2 D + 2D + 200K b) – 2 ± 4 – 4 ( 200K ) roots = ----------------------------------------------. sin ( 0.5t 1 – 0.1. – ω ) 2 3.= -------------------------------2 4 F 1–ω +ω θ = angle ( 1 – ω .5 -0.005 0.= -------------------------2 F D +D+1 y = 1–e – 0.-2 -0.root locus .11. K=0.= – 1 ± 1 – 200K 2 Im K roots 0 0.5 ( 1 ) cos ( 0.1 1 5 10 0.-1 etc.005 Re -2 -1 . (the same for both) a) b) y 1 -.866 Re Im c) -0.866 d) 1 y -.9 -1.

444j -0.< 0 2 ± 1 – 80K < 1 K>0 1K > ----80 2 b) Matching the second order forms.= 2. -1.000.11.5 +/.65 4. 2ω n ξ = 1 ω n = 20K 2 The gain can only be used for the natural frequency 20 20 K = ----.5 +/.005 From the root locus graph this value is critically stable. -5.15 c) D + 2D + 200K = D + 2ζω n D + ω n 2 2 2 ∴ω n = 1 ∴K = 0.13j -0.141.4.500.500 -0. a) D + D + 20K = 0 – 1 ± 1 – 4 ( 20K ) D = -------------------------------------------2 K -10 -1 0 1/80 1 10 1000 roots 13. 4.000 0.65.root locus . -14. -0.14.4j For complex roots 1 – 80K < 0 For negative real roots (stable) – 1 ± 1 – 80K -----------------------------------.= ----.000.22 2 2 ωn 3 .000 -0.5 +/.

1t + C 2 ) – 14.5 – 1rad– 1rad C 1 = -----------------.16 c) θo 20 ( 10 ) ---.1t + C 2 ) θ o ( 0 ) = C 1 ( 1 ) sin ( 14.1t + C 2 ) + 1rad C 1 sin ( C 2 ) = –1rad (1) cos ( 14.5t – 0.54 sin ( 14.5t θ'o ( t ) = –0.5t 2 A = – 0.54 ) θo ( t ) = ( e – 0.root locus .5t C 2 = –1.5C 1 e sin ( 14.= ---------------------------------------2 θd D + D + 20 ( 10 ) ·· · θ o + θ d + θ d 200 = 200θ d Homogeneous: A + A + 200 = 0 – 1 ± 1 – 4 ( 200 ) A = ------------------------------------------2 θo ( t ) = C1 e Particular: θ = A 0 + 0 + A200 = 200 ( 1rad ) θ o ( t ) = 1rad Initial Conditions (assume at rest): θo ( t ) = C1 e – 0.1 ( 0 ) + C 2 ) + 1rad = 0 – 0.5 ± 14.11.5C 1 sin ( C 2 ) – 14.= tan ( C 2 ) – 0.1t + C 2 ) A = 1rad sin ( 14.1C 1 cos ( C 2 ) 14.5t – 0.= 1.5 sin ( C 2 ) 14.1C 1 e 0 = –0.54 ) + 1 ) ( rad ) .1 cos ( C 2 ) = – 0.1t – 1.= ------------------------.1j sin ( 14.1 --------.000rad sin ( C 2 ) sin ( – 1.

X + KpD + Ki + KdD -----------------------------------------D 4 X + 3K p D + 3K i + 3K d D ---------------------------------------------------D(D + 9 ) 4 X 3K p D + 3K i + 3K d D ------------------------------------------------------------------------------------------2 D ( D + 9 ) + 12K p D + 12K i + 12K d D 3K p D + 3K i + 3K d D ---------------------------------------------------------------------------------------------2 D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i ) 2 2 2 2 3 -----------D+9 Y Y Y X Y .root locus .11.18 5. (ans.

roots -0.52j -0.682 .808 +/.418 -0.1 0 1 10 100 Stable for.19 b) D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i ) = 0 – 9 – 12K p ± ( 9 + 12K p ) – 4 ( 12K d + 1 )12K D = ------------------------------------------------------------------------------------------------------------------i 2 ( 12K d + 1 ) Kd -100 -10 -1 -0.57. 0.root locus .303j -0.0. -20.588.1j – 9 – 12K p ± ( 9 + 12K p ) – 4 ( 12K d + 1 )12K i < 0 ± ( 9 + 12K p ) – 4 ( 12K d + 1 )12K i < 9 + 12K p ( 9 + 12K p ) – 4 ( 12K d + 1 )12K i < ( 9 + 12K p ) – 4 ( 12K d + 1 )12K i < 0 –1 K d > ----12 Becomes complex at. 2.0.087 +/.11.241.109 -0. 0.41 -0.0.6 -0.092.0087 +/.46.369 -0. 0 > ( 9 + 12K p ) – 4 ( 12K d + 1 )12K i 576K d K i > ( 9 + 12K p ) – 48K i ( 9 + 12K p ) – 48K K d > ----------------------------------------------i 576K d K i 2 2 2 2 2 2 2 2 2 K d > 0. 105.

521 -12dB .923 = 0. 2  13  2 X D 13 + D21 + 12 D + D1.840 2ω n 2 2 ωd = ωn 1 – ξ = for the denominator. ωn = 0.=  -----  -------------------------------------------------.866 1.= ----------------------------------------.= 0.923 3 final gain = 20 log  ----- = – 12.961 2 ω d = ω n 1 – ξ = 0.7  13 3initial gain = 20 log  ----- = – 12.20 c) Kp = 1 Ki = 1 Kd = 1 2 3K p D + 3K i + 3K d D Y -. ωn = 1 = 1 2 1 ξ = --------.5 2ω n 1 – 0.root locus .11.840 = 0.615 + 0.615 ξ = -----------.0  12 for the numerator.961 1 – 0.= 0. D +D+1 -.5 = 0.= ---------------------------------------------------------------------------------------------2 X D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i ) 2 2  Y 3D + 3D + 3 3.

5 .21 3K p D + 3K i + 3K d D Y -.= 2 12K d + 1 12K i = 48K d + 4 24K d = 7 + 12K p 2 9 + 12K p 2ξω n = --------------------.root locus .292 ) + 4 = 258.0 Now to check for stability D ( 12 ( 5.5 ) 2 2 2 K d = 5.= 20.= – 1 ± 1.73j 2 ( 64.292 K i = 21.504D + 129D + 258 = 0 – 129 ± 129 – 4 ( 64.= ---------------------------------------------------------------------------------------------2 X D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i ) ωn = 12K i --------------------.11.5 )258 D = -------------------------------------------------------------------. one value must be selected to continue.5 ( 2 ) 12K d + 1 At this point there are two equations and two unknowns.5 ) ) = 0 64. K p = 10 24K d = 7 + 12K p = 7 + 12 ( 10 ) = 127 12K i = 48K d + 4 = 48 ( 5. therefore.292 ) + 1 ) + D ( 9 + 12 ( 10 ) ) + ( 12 ( 21.

= 0 13 It is a first order system. Y ( t ) = C1 e Y(0) = 0 0 = C1 1 + C2 9Y' ( t ) = – ----.C e 13 1 90 = – ----.C 1 13 1 9 – ----.= -------------------------2 X 13D + 9D Y ( 13D + 9D ) = X ( 3D ) ·· · ·· Y 13 + Y 9 = X 3 ·· · 9 Y + Y ----.= --------------------------------------------------------------------------------------------------2 X D ( 12 ( 1 ) + 1 ) + D ( 9 + 12 ( 0 ) ) + ( 12 ( 0 ) ) Y 3D -.t 13 2 2 2 2 X = t · X = 1 ·· X = 0 + C2 Y' ( 0 ) = 0 C1 = –C2 starts at rest/undeflected C1 = 0 C2 = 0 no response .11.t 13 9 – ----.22 e) Cannot be found without an assumed input and initial conditions f) Y 3 ( 0 )D + 3 ( 0 ) + 3 ( 1 )D -.root locus .

1D ) ) K ( 0.1D ) ------------------------------------------------------------------------------------------------------------------------------------3 2 2 D + 10D + 100D + 1 + 0.1 + 0.1D ) ------------------------------------------------------3 2 D + 10D + 100D + 1 2 Y Y X Y .11.5D + 0.1 + 0.+ D   D + 10 ----D 0. X + 1 K  5 + --.1  ----- D Y K ( 0.01 ( K ( 0.23 6.root locus .01 X + 0.1 -------------------------------------2 D + 10D + 100 Y 5D + 1 + D K  -----------------------------   D 2 0.01 X K ( 0.01 X + 0.1 + 0.1 ---------------------------------------------------------------2  10 D + 10D + 100 + 0.5D + 0.005K ) + ( 1 + 0.1D ) ---------------------------------------------------------------------------------------------------------------------------------------------3 2 D + D ( 10 + 0.1 + 0.5D + 0.001K ) + D ( 100 + 0.001K ) 2 2 0.5D + 0.

52.019. -5+/-8.263 0.root locus . -5.11. Draw a root locus diagram for the feedback system below given the variable parameter ‘P’.01. -4.005K ) + ( 1 + 0.3. -4.01.64j -0. . or equivalent.66j -0.3. -5. Vd + - e P Vs 100 -----------------D + 100 ω 1 --D θ Va 2 2. and draw an approximate time response for each.001K ) + D ( 100 + 0.49+/-8.546 -0. -0.000 94.66j -0.6 ASSIGNMENT PROBLEMS 1.0174. -3. -104.001K ) = 0 The roots can be found with a calculator. -4.995+/-8. Mathcad.65j -0.5+/-8.0099.657j -0.12 100.99+/-8.24 Given the homogeneous equation for the system. D + D ( 10 + 0. a) b) c) θ D ( 100 ) + D ( 100P ) + ( 100 ) ----. draw the root locus plots.= ---------------------------------------------------------------------------------3 2 Vd D + D ( 300 ) + D ( 200P ) + ( 200 ) 2 11. For the transfer functions below. -13. -13.992.099.000 -1000 -10 0 10 1000 17165. K roots notes -100.572 3 2 roots become negative roots become real 7.

25 1 -----------D+1 1 --------------2 D +1 1 -------------------2 ( D + 1) 1 ----------------------------2 D + 2D + 2 .11.root locus .

no transfer functions .can be overcome by control strategies 12.2.input/output relationships are not linear .1 12.new elements needed in block diagrams 12.1 INTRODUCTION • how they are different from linear . NONLINEAR SYSTEMS Topics: • Objectives: • 12.1 Non-Linear Relationships .nonlinear system .2 SOURCES OF NONLINEARITY .important for analysis .12.

5  e – 1   θ 12. 4.2  – - π Q ( θ ) = 8. Pick an operating point or range for the component.3 NON-LINEAR ELEMENTS • If our models include a device that is non linear we will need to linearize the model before we can proceed. 3. Develop a linear equation. 1. • A non-linear system can be approximated with a linear equation using the following method. In this case the relationship between pressure drop and flow are non-linear. 2. Find a constant value that relates a change in the input to a change in the output. We need to develop an equation that approximates the local operating point. Use the linear equation in the analysis (Laplace or other) • Consider the example below. q ∆p ≈ R ( q – q ) R ≈ ∆ ( ∆p ) --------------∆q qR = 2 ----2 K q ∆p ∆ ( ∆p ) ∆q .12.nonlinear system .

nonlinear system .1 Linearizing non-linear elements 12. Show for a system with first .3.cables in tension/compression . Kd Ks F ·· · x + x  ---------------------------.0001t 12.system components turned on/off . PWM is used to generate analog voltage equivalents.2 Switching .3 Figure 12. + x  ---------------------------.3.0001t  M – 0.show an example where input conditions change · y + 4y = f ( t ) f(t < 0) = 0 f ( 0 ≤ t < 5s ) = 2 f ( t ≥ 5s ) = 0 .1 Time Variant .give PWM (Pulse Width Modulation) example with ripple showing equivalent voltage.0001t M – 0.12. = ----------------------------  M – 0.system parameters vary as a function of time.

so it is better to compensate in software .The static friction. and when a motor is driving a mechanical system.relationship in figure below. If the systems breaks free and starts turning. .3 Deadband .nonlinear system . will prevent initial motion.4 order response with tau = 0.1s for a frequency of 1KHz.important to consider when doing system analysis 12.the region where the applied voltage has no effect is called the deadband. .3. .small actuation signals not large enough to overcome friction .In systems there are two type of friction that must be considered. Point out the ripple and effective voltage.12. 10Hz and 1Hz.Friction in all components . ’stiction’.This effect is normally known as ’stiction’. the kinetic friction will provide a roughly constant friction resistance. .costs money to reduce friction. . . a combination of the words static and friction. .Friction is common in less expensive motors.

12.3 Deadband approximation for a bidirectional motor .nonlinear system .equations for these are shown in figure .2 Motor deadband for a bidirectional motor .deadband compensation as shown in figure below.5 The kinetic friction results in a different curve while slowing down ω Vapplied deadband motor starts to turn as friction is overcome Figure 12. Vadjusted Vstick Vwanted -Vstick Figure 12.

4 Hysteresis and Slip .correct by tracking the previous motion direction and taking extra steps when reversing direction .12.windup resulting from springiness and friction .3.backlash • .nonlinear system .7 12.

7 ASIGNMENT PROBLEMS .8 12.4 SUMMARY 12.12.5 PRACTICE PROBLEMS 1.6 PRACTICE PROBLEM SOLUTIONS 12. 12.nonlinear system .

analog IO .1.1 13. quantization error. not discrete.1 INTRODUCTION An analog value is continuous. 13. Voltage logical continuous t Figure 13. Actuators and sensors that can be used with analog inputs and outputs will be discussed in later chapters. Inputs: • oven temperature • fluid pressure • fluid flow rate Outputs: . aliasing.1 Logical and Continuous Values Typical analog inputs and outputs for computers are listed below. techniques were discussed for designing continous control systems. In this chapter we will examine analog inputs and outputs so that we may design continuous control systems that use computers. In the previous chapters. resolution Objectives: • To understand the basics of conversion to and from analog values.13. ANALOG INPUTS AND OUTPUTS Topics: • Analog inputs and outputs • Sampling issues. as shown in Figure 13.

it must be quantized into an available integer value. current or other value by a transducer.2 Analog inputs Analog to digital and ditial to analog conversion uses integers within the computer. a desired analog output value is limited to available quantized values.2. Consider when a continuous analog voltage is being read. A signal conditioner converts the signal from the transducer to a voltage or current that is read by the analog input.analog IO . but the possible integer values are quantied with a ’staircase’ set of values.. sensor) signal conditioning analog input integer Figure 13.2 • fluid valve position • motor position • motor velocity A basic analog input is shown in Figure 13. Integers limit the resolution of the numbers to a discrete. Likewise. In this type of system a physical value is converted to a voltage.13. or quantized range. In general the difference between the analog and quantized integer value is an error based upon the resolution of the analog I/O. analog continuous quantized integer . physical phenomenon transducer (ie.3 where the desired or actual analog value is continuous. The effect of using integers is shown in Figure 13.

2 ANALOG INPUTS To input an analog voltage (into a computer) the continuous voltage value must be sampled and then converted to a numerical value by an A/D (Analog to Digital) converter (also known as ADC). but reasonable ranges are. The maximum (Vmax) and minimum (Vmin) voltages are a function of the control hardware. but a common sampling rate is 100KHz. The sampling frequency is specified when buying hardware. A/D converters can only acquire a limited number of samples per second. The time between samples is called the sampling period T.13. The sampling time is often much smaller than the sampling period. Figure 13.3 Figure 13. There are three samples shown on the figure. These are often specified when purchasing hardware. The process of sampling the data is not instantaneous. and the inverse of the sampling period is the sampling frequency (also called sampling rate). . The data value sampled will be somewhere between the voltage at the start and end of the sample. Voltage is sampled during these time periods voltage time T = (Sampling Frequency)-1 Sampling time Figure 13. and even between the start and end of the data sample there is a significant change in the voltage value. This data is noisier.4 Sampling an Analog Voltage A more realistic drawing of sampled data is shown in Figure 13.3 Quantization error 13.analog IO .5. The time required to acquire the sample is called the sampling time. so each sample has a start and stop time.4 shows a continuous voltage changing over time.

6. If the A/D converter is 8 bit then the result can read up to 256 different voltage levels.end of sampl e V ( t1 ).13. V( t ) Vmax V( t2 ) V( t1 ) V min t τ t1 t2 where. end of sampl e Vmin.analog IO . V ( t 2 ) = voltage at start.5 Parameters for an A/D Conversion The parameters defined in Figure 13. These equations are summarized in Figure 13. V ( t ) = the actual voltage over time τ = sample interval for A/D converter t = time t1. Most A/D converters have 12 bits. t 2 = time at start. Equation 1 relates the number of .4 0V to 5V 0V to 10V -5V to 5V -10V to 10V The number of bits of the A/D converter is the number of bits in the result word.5 can be used to calculate values for A/D converters. Vmax = input voltage range of A/D converter N = number of bits in the A/D converter Figure 13. 16 bit converters are used for precision measurements.

 ( Vmax – Vmin ) + V min  R – 1 where.564V is input into the PLC.570V-4. where 0 is -10V and 1023 is +10V. R = 2 N (1) (2) Vmax – V min VERROR = ±  ---------------------------- -  2R Vin – Vmin VI = INT  ---------------------------- ( R – 1 ) V –V  max min (3) VI VC =  ----------. and a voltage in the computer.006V. The resulting quantization error is 4. This gives a resolution of 1024.570V.analog IO .6 A/D Converter Equations Consider a simple example. Equation 2 gives the error that can be expected with an A/D converter given the range between the minimum and maximum voltages. Finally.8mV. equation 4 allows a conversion between the integer value from the A/ D converter. the A/D converter converts the voltage to an integer value of 746. Because there are only 1024 steps there is a maximum error of ±9. If a voltage of 4. Each bit halves the quantization error. This error can be reduced by selecting an A/D converter with more bits.5 bits of an A/D converter to the resolution. When we convert this back to a voltage the result is 4.13. . a 10 bit A/D converter can read voltages between 10V and 10V.564V=+0. Equation 3 relates the voltage range and resolution to the voltage input to estimate the integer that the A/D converter will record. R = resolution of A/D converter (4) V I = the ingra eersnnte ptoae t evl rpeet g i u l g e u i hn v t VC = the voltage calculated from the integer value VERROR = the maximum quantization error Figure 13. and the resolution (this is commonly called the quantization error).

and 9 samples are taken. In the upper graph the waveform completes seven cycles. The sampling frequency was too low. .564V Calculate. R = 2 N = 1024 V max – V min V ERROR =  ---------------------------- = 0.570V  R Figure 13.13.7 Sample Calculation of A/DValues If the voltage being sampled is changing too fast we may get false readings.analog IO . as shown in Figure 13. so the signal read appears to be different that it actually is. N = 10 Vmax = 10V V min = – 10V V in = 4. this is called aliasing.8.0098V   2R V in – V min V I = INT  ---------------------------- R = 746 V –V  max min V V C =  ----I ( Vmax – Vmin ) + V min = 4. The bottom graph plots out the values read.6 Given.

analog IO - 13.7

Figure 13.8

Low Sampling Frequencies Cause Aliasing

The Nyquist criterion specifies that sampling frequencies should be at least twice the frequency of the signal being measured, otherwise aliasing will occur. The example in Figure 13.8 violated this principle, so the signal was aliased. If this happens in real applications the process will appear to operate erratically. In practice the sample frequency should be 4 or more times faster than the system frequency.

where,

f AD = sampling frequency fsignal = maximum frequency of the input

There are other practical details that should be considered when designing applications with analog inputs; • Noise - Since the sampling window for a signal is short, noise will have added effect on the signal read. For example, a momentary voltage spike might result in a higher than normal reading. Shielded data cables are commonly used to reduce the noise levels. • Delay - When the sample is requested, a short period of time passes before the final sample value is obtained. • Multiplexing - Most analog input cards allow multiple inputs. These may share

analog IO - 13.8

the A/D converter using a technique called multiplexing. If there are 4 channels using an A/D converter with a maximum sampling rate of 100Hz, the maximum sampling rate per channel is 25Hz. • Signal Conditioners - Signal conditioners are used to amplify, or filter signals coming from transducers, before they are read by the A/D converter. • Resistance - A/D converters normally have high input impedance (resistance), so they affect circuits they are measuring. • Single Ended Inputs - Voltage inputs to a PLC can use a single common for multiple inputs, these types of inputs are called single ended inputs. These tend to be more prone to noise. • Double Ended Inputs - Each double ended input has its own common. This reduces problems with electrical noise, but also tends to reduce the number of inputs by half. • Sampling Rates - The maximum number of samples that can be read each second. If reading multiple channels with a multiplexer, this may be reduced. • Quantization Error - Analog IO is limited by the binary resolution of the converter. This means that the output is at discrete levels, instead of continuous values. • Triggers - often external digital signals are used to signal the start of data collection. • Range - the typical voltages that the card can read. Typical voltage ranges are 10V to 10V, 0V to10V, 0V to 5V, 1Vto 5V, -5Vto 5V, 4mA to 20mA. • DMA - a method to write large blocks of memory directly to computer memory. This is normally used for high speed data captures. • Filters - some A/D input cards will provide built in functions to filter the incoming data to remove high frequency noise components. • Input impedance - most analog inputs have very high input resistances, in the range of Mohms.

analog IO - 13.9

ASIDE: This device is an 8 bit A/D converter. The main concept behind this is the successive approximation logic. Once the reset is toggled the converter will start by setting the most significant bit of the 8 bit number. This will be converted to a voltage Ve that is a function of the +/-Vref values. The value of Ve is compared to Vin and a simple logic check determines which is larger. If the value of Ve is larger the bit is turned off. The logic then repeats similar steps from the most to least significant bits. Once the last bit has been set on/off and checked the conversion will be complete, and a done bit can be set to indicate a valid conversion value. Vin above (+ve) or below (-ve) Ve Vin +Vref successive approximation logic 8 D to A converter

+ -

clock reset

Ve done

-Vref

8

data out

Quite often an A/D converter will multiplex between various inputs. As it switches the voltage will be sampled by a sample and hold circuit. This will then be converted to a digital value. The sample and hold circuits can be used before the multiplexer to collect data values at the same instant in time.

Figure 13.9

A Successive Approximation A/D Converter

13.3 ANALOG OUTPUTS
Analog outputs are much simpler than analog inputs. To set an analog output an integer is converted to a voltage. This process is very fast, and does not experience the timing problems with analog inputs. But, analog outputs are subject to quantization errors. Figure 13.10 gives a summary of the important relationships. These relationships are

analog IO - 13.10

almost identical to those of the A/D converter.

R = 2

N

(5) (6)

Vmax – V min VERROR = ±  ----------------------------   2R V desired – V min V I = INT  ----------------------------------- R  V  max – V min V Voutput =  ----I ( Vmax – Vmin ) + V min  R

(7)

(8)

where, R = resolution of A/D converter VERROR = the maximum quantization error VI = the integer value representing the desired voltage V output = the voltage output using the integer value Vdesired = the desired output voltae g

Figure 13.10 Analog Output Relationships Assume we are using an 8 bit D/A converter that outputs values between 0V and 10V. We have a resolution of 256, where 0 results in an output of 0V and 255 results in 10V. The quantization error will be 20mV. If we want to output a voltage of 6.234V, we would specify an output integer of 160, this would result in an output voltage of 6.250V. The quantization error would be 6.250V-6.234V=0.016V.

analog IO - 13.11

Given, N = 8 V max = 10V V min = 0V V desired = 6.234V Calculate, R = 2
N

= 256

Vmax – Vmin VERROR =  ---------------------------- = 0.020V   2R V in – V min V I = INT  ---------------------------- R = 160 V –V 
max min

V V C =  ----I ( Vmax – Vmin ) + V min = 6.250V  R The current output from a D/A converter is normally limited to a small value, typically less than 20mA. This is enough for instrumentation, but for high current loads, such as motors, a current amplifier is needed. This type of interface will be discussed later. If the current limit is exceeded for 5V output, the voltage will decrease (so don’t exceed the rated voltage). If the current limit is exceeded for long periods of time the D/A output may be damaged.

analog IO - 13.12

ASIDE: 5K Ω MSB bit 3 10K Ω 20K Ω 40K Ω 80K Ω V– V + + 0 Vss + Vo LSB bit 0

bit 2 Computer bit 1

First we write the obvious, V
+

= 0 = V–

Next, sum the currents into the inverting input as a function of the output voltage and the input voltages from the computer, V b3 Vb 2 Vb1 V b0 Vo -------------- + -------------- + -------------- + -------------- = ----------10KΩ 20KΩ 40KΩ 80KΩ 5KΩ ∴Vo = 0.5V b 3 + 0.25Vb 2 + 0.125Vb1 + 0.0625V b 0 Consider an example where the binary output is 1110, with 5V for on, ∴Vo = 0.5 ( 5V ) + 0.25 ( 5V ) + 0.125 ( 5V ) + 0.625 ( 0V ) = 4.375V

Figure 13.11 A Digital-To-Analog Converter

13.4 NOISE REDUCTION

13.4.1 Shielding
When a changing magnetic field cuts across a conductor, it will induce a current flow. The resistance in the circuits will convert this to a voltage. These unwanted voltages

analog IO - 13.13

result in erroneous readings from sensors, and signal to outputs. Shielding will reduce the effects of the interference. When shielding and grounding are done properly, the effects of electrical noise will be negligible. Shielding is normally used for; all logical signals in noisy environments, high speed counters or high speed circuitry, and all analog signals. There are two major approaches to reducing noise; shielding and twisted pairs. Shielding involves encasing conductors and electrical equipment with metal. As a result electrical equipment is normally housed in metal cases. Wires are normally put in cables with a metal sheath surrounding both wires. The metal sheath may be a thin film, or a woven metal mesh. Shielded wires are connected at one end to "drain" the unwanted signals into the cases of the instruments. Figure 13.12 shows a thermocouple connected with a thermocouple. The cross section of the wire contains two insulated conductors. Both of the wires are covered with a metal foil, and final covering of insulation finishes the cable. The wires are connected to the thermocouple as expected, but the shield is only connected on the amplifier end to the case. The case is then connected to the shielding ground, shown here as three diagonal lines.

Two conductor shielded cable cross section

Insulated wires Metal sheath Insulating cover

Figure 13.12 Shielding for a Thermocouple A twisted pair is shown in Figure 13.13. The two wires are twisted at regular intervals, effectively forming small loops. In this case the small loops reverse every twist, so any induced currents are cancel out for every two twists.

analog IO - 13.14

1" or less typical

Figure 13.13 A Twisted Pair When designing shielding, the following design points will reduce the effects of electromagnetic interference. • Avoid “noisy” equipment when possible. • Choose a metal cabinet that will shield the control electronics. • Use shielded cables and twisted pair wires. • Separate high current, and AC/DC wires from each other when possible. • Use current oriented methods such as sourcing and sinking for logical I/O. • Use high frequency filters to eliminate high frequency noise. • Use power line filters to eliminate noise from the power supply.

13.4.2 Grounding
- ground voltages are based upon the natural voltage level in the physical ground (the earth under your feet). This will vary over a distance. Most buildings and electrical systems use a ground reference for the building. Between different points on the same building ground voltage levels may vary as much as a few hundred millivolts. This can lead to significant problems with voltage readings and system safety. - A signal can be floating, or connected to a ground - if floating a system normally has a self contained power source, or self reference such as a battery, strain guage or thermocouple. These are usually read with double ended outputs. The potential for floating voltage levels can be minimized by connecting larger resistors (up to 100K) from the input to ground. - a grounded system uses a single common (ground) for all signals. These are noramlly connected to a single ended inputs. - the analog common can also be connected to the ground with a large resistor to to drain off induced voltages.

analog IO - 13.15

- calbe shields or grounds are noramlly only connected at one side to prevent ground loops.

13.5 CASE STUDY
- data smoothing by averaging inputs - conversion back to an input voltage - use it to calculate an output voltage for control.

13.6 SUMMARY
• A/D conversion will convert a continuous value to an integer value. • D/A conversion is easier and faster and will convert a digital value to an analog value. • Resolution limits the accuracy of A/D and D/A converters. • Sampling too slowly will alias the real signal. • Analog inputs are sensitive to noise. • Analog shielding should be used to improve the quality of electrical signals.

13.7 PRACTICE PROBLEMS
1. What is the difference between an A/D input and D/A output? 2. An analog voltage that has a range of -10V to 10V and is to be read to a precision of +/-0.05V. How is the minimum number of bits required for the A/D converter? 3. If given a 12 bit analog input with a range of -10V to 10V. If we put in 2.735V, what will the integer value be after the A/D conversion? What is the error? What voltage can we calculate?

13.8 PRACTICE PROBLEM SOLUTIONS
1. an A/D converter will convert an analog input voltage (or current) to an integer value. A D/A output will convert an integer value to an output voltage.

analog IO - 13.16

2. 10V – ( – 10V ) R = --------------------------------- = 200 0.1V 7 bits, R = 128 8 bits, R = 256 The minimum number of bits is 8.

3. N = 12 R = 4096 V min = – 10V Vmax = 10V Vin = 2.735V

V in – V min V I = INT  ---------------------------- ( R – 1 ) = 2607 V – Vmin max VI V C =  -----------  ( V max – V min ) + Vmin = 2.733V  R – 1 error = Vc – Vin = ( – 0.002 )V

13.9 ASSIGNMENT PROBLEMS
1. An analog output needs to be between -4V and 8V, in 0.005V intervals. How many bits are required for the D/A converter? 2. Discuss methods for reducing electrical noise in analog inputs.

continuous sensors - 14.1

14. CONTINUOUS SENSORS
Topics: • Continuous sensor issues; accuracy, resolution, etc. • Angular measurement; potentiometers, encoders and tachometers • Linear measurement; potentiometers, LVDTs, Moire fringes and accelerometers • Force measurement; strain gages and piezoelectric • Liquid and fluid measurement; pressure and flow • Temperature measurement; RTDs, thermocouples and thermistors • Other sensors • Continuous signal inputs and wiring • Glossary Objectives: • To understand the common continuous sensor types. • To understand interfacing issues.

14.1 INTRODUCTION
Continuous sensors convert physical phenomena to measurable signals, typically voltages or currents. Consider a simple temperature measuring device, there will be an increase in output voltage proportional to a temperature rise. A computer could measure the voltage, and convert it to a temperature. The basic physical phenomena typically measured with sensors include; - angular or linear position - acceleration - temperature - pressure or flow rates - stress, strain or force - light intensity - sound Most of these sensors are based on subtle electrical properties of materials and devices. As a result the signals often require signal conditioners. These are often amplifiers that boost currents and voltages to larger voltages. Sensors are also called transducers. This is because they convert an input phenomena to an output in a different form. This transformation relies upon a manufactured device with limitations and imperfection. As a result sensor limitations are often charac-

continuous sensors - 14.2

terized with; Accuracy - This is the maximum difference between the indicated and actual reading. For example, if a sensor reads a force of 100N with a ±1% accuracy, then the force could be anywhere from 99N to 101N. Resolution - Used for systems that step through readings. This is the smallest increment that the sensor can detect, this may also be incorporated into the accuracy value. For example if a sensor measures up to 10 inches of linear displacements, and it outputs a number between 0 and 100, then the resolution of the device is 0.1 inches. Repeatability - When a single sensor condition is made and repeated, there will be a small variation for that particular reading. If we take a statistical range for repeated readings (e.g., ±3 standard deviations) this will be the repeatability. For example, if a flow rate sensor has a repeatability of 0.5cfm, readings for an actual flow of 100cfm should rarely be outside 99.5cfm to 100.5cfm. Linearity - In a linear sensor the input phenomenon has a linear relationship with the output signal. In most sensors this is a desirable feature. When the relationship is not linear, the conversion from the sensor output (e.g., voltage) to a calculated quantity (e.g., force) becomes more complex. Precision - This considers accuracy, resolution and repeatability or one device relative to another. Range - Natural limits for the sensor. For example, a sensor for reading angular rotation may only rotate 200 degrees. Dynamic Response - The frequency range for regular operation of the sensor. Typically sensors will have an upper operation frequency, occasionally there will be lower frequency limits. For example, our ears hear best between 10Hz and 16KHz. Environmental - Sensors all have some limitations over factors such as temperature, humidity, dirt/oil, corrosives and pressures. For example many sensors will work in relative humidities (RH) from 10% to 80%. Calibration - When manufactured or installed, many sensors will need some calibration to determine or set the relationship between the input phenomena, and output. For example, a temperature reading sensor may need to be zeroed or adjusted so that the measured temperature matches the actual temperature. This may require special equipment, and need to be performed frequently. Cost - Generally more precision costs more. Some sensors are very inexpensive, but the signal conditioning equipment costs are significant.

14.2 INDUSTRIAL SENSORS
This section describes sensors that will be of use for industrial measurements. The sections have been divided by the phenomena to be measured. Where possible details are provided.

continuous sensors - 14.3

14.2.1 Angular Displacement
14.2.1.1 - Potentiometers Potentiometers measure the angular position of a shaft using a variable resistor. A potentiometer is shown in Figure 14.1. The potentiometer is resistor, normally made with a thin film of resistive material. A wiper can be moved along the surface of the resistive film. As the wiper moves toward one end there will be a change in resistance proportional to the distance moved. If a voltage is applied across the resistor, the voltage at the wiper interpolate the voltages at the ends of the resistor.

V1 resistive film Vw V1 Vw V2 wiper V2 schematic physical

Figure 14.1

A Potentiometer

The potentiometer in Figure 14.2 is being used as a voltage divider. As the wiper rotates the output voltage will be proportional to the angle of rotation.

continuous sensors - 14.4

V1 θ max

θw

V out

θw Vout = ( V2 – V1 ) ----------  + V1 θ 
max

V2

Figure 14.2

A Potentiometer as a Voltage Divider

Potentiometers are popular because they are inexpensive, and don’t require special signal conditioners. But, they have limited accuracy, normally in the range of 1% and they are subject to mechanical wear. Potentiometers measure absolute position, and they are calibrated by rotating them in their mounting brackets, and then tightening them in place. The range of rotation is normally limited to less than 360 degrees or multiples of 360 degrees. Some potentiometers can rotate without limits, and the wiper will jump from one end of the resistor to the other. Faults in potentiometers can be detected by designing the potentiometer to never reach the ends of the range of motion. If an output voltage from the potentiometer ever reaches either end of the range, then a problem has occurred, and the machine can be shut down. Two examples of problems that might cause this are wires that fall off, or the potentiometer rotates in its mounting.

14.2.2 Encoders
Encoders use rotating disks with optical windows, as shown in Figure 14.3. The encoder contains an optical disk with fine windows etched into it. Light from emitters passes through the openings in the disk to detectors. As the encoder shaft is rotated, the light beams are broken. The encoder shown here is a quadrature encode, and it will be discussed later.

continuous sensors - 14.5

light emitters

light detectors

Shaft rotates

Note: this type of encoder is commonly used in computer mice with a roller ball.

Figure 14.3

An Encoder Disk

There are two fundamental types of encoders; absolute and incremental. An absolute encoder will measure the position of the shaft for a single rotation. The same shaft angle will always produce the same reading. The output is normally a binary or grey code number. An incremental (or relative) encoder will output two pulses that can be used to determine displacement. Logic circuits or software is used to determine the direction of rotation, and count pulses to determine the displacement. The velocity can be determined by measuring the time between pulses. Encoder disks are shown in Figure 14.4. The absolute encoder has two rings, the outer ring is the most significant digit of the encoder, the inner ring is the least significant digit. The relative encoder has two rings, with one ring rotated a few degrees ahead of the other, but otherwise the same. Both rings detect position to a quarter of the disk. To add accuracy to the absolute encoder more rings must be added to the disk, and more emitters and detectors. To add accuracy to the relative encoder we only need to add more windows to the existing two rings. Typical encoders will have from 2 to thousands of windows per ring.

continuous sensors - 14.6

absolute encoder

Figure 14.4

Encoder Disks

When using absolute encoders, the position during a single rotation is measured directly. If the encoder rotates multiple times then the total number of rotations must be counted separately. When using a relative encoder, the distance of rotation is determined by counting the pulses from one of the rings. If the encoder only rotates in one direction then a simple count of pulses from one ring will determine the total distance. If the encoder can rotate both directions a second ring must be used to determine when to subtract pulses. The quadrature scheme, using two rings, is shown in Figure 14.5. The signals are set up so that one is out of phase with the other. Notice that for different directions of rotation, input B either leads or lags A.

continuous sensors - 14.7

Note the change as direction is reversed

total displacement can be determined by adding/subtracting pulse counts (direction determines add/subtract)

Note: To determine direction we can do a simple check. If both are off or on, the first to change state determines direction. Consider a point in the graphs above where both A and B are off. If A is the first input to turn on the encoder is rotating clockwise. If B is the first to turn on the rotation is counterclockwise. Aside: A circuit (or program) can be built for this circuit using an up/down counter. If the positive edge of input A is used to trigger the clock, and input B is used to drive the up/down count, the counter will keep track of the encoder position.

Figure 14.5

Interfaces for encoders are commonly available for PLCs and as purchased units. Newer PLCs will also allow two normal inputs to be used to decode encoder inputs.

continuous sensors - 14.8

Normally absolute and relative encoders require a calibration phase when a controller is turned on. This normally involves moving an axis until it reaches a logical sensor that marks the end of the range. The end of range is then used as the zero position. Machines using encoders, and other relative sensors, are noticeable in that they normally move to some extreme position before use.

14.2.2.1 - Tachometers Tachometers measure the velocity of a rotating shaft. A common technique is to mount a magnet to a rotating shaft. When the magnetic moves past a stationary pick-up coil, current is induced. For each rotation of the shaft there is a pulse in the coil, as shown in Figure 14.6. When the time between the pulses is measured the period for one rotation can be found, and the frequency calculated. This technique often requires some signal conditioning circuitry.

rotating shaft

pickup coil Vout

Vout

t magnet 1/f

Figure 14.6

A Magnetic Tachometer

Another common technique uses a simple permanent magnet DC generator (note: you can also use a small DC motor). The generator is hooked to the rotating shaft. The rotation of a shaft will induce a voltage proportional to the angular velocity. This technique will introduce some drag into the system, and is used where efficiency is not an issue. Both of these techniques are common, and inexpensive.

14.2.3 Linear Position
14.2.3.1 - Potentiometers

.14. 14.2. so in the figure more current will be induced in the left hand coil. and the signal out is zero (0Vac). These are capable of measuring linear displacement over long distances. But. Figure 14. V2 a V out = V 1 + ( V2 – V1 )  -.7 shows the output voltage when using the potentiometer as a voltage divider. but potentiometers are also available in linear/sliding form. Near the center of motion the change in voltage is proportional to the displacement. The magnitude of the signal out voltage on either line indicates the position of the core. The basic device is shown in Figure 14.8.7 Linear Potentiometer Linear/sliding potentiometers have the same general advantages and disadvantages of rotating potentiometers. The outside coils are wound in opposite directions so that when the core is in the center the induced currents cancel.  L Vout L a V1 Figure 14.3.Linear Variable Differential Transformers (LVDT) Linear Variable Differential Transformers (LVDTs) measure linear displacements over a limited range.2 . The core will pull the magnetic field towards it. It consists of outer coils with an inner moving magnetic core. High frequency alternating current (AC) is applied to the center coil.9 Rotational potentiometers were discussed before. further from the center the relationship becomes nonlinear. This generates a magnetic field that induces a current in the two outside coils.continuous sensors .

The two diodes convert the AC wave to a half wave DC wave. Typical applications for these devices include measuring dimensions on parts for quality .9 A Simple Signal Conditioner for an LVDT These devices are more accurate than linear potentiometers. Vac out Vac in LVDT Vdc out Figure 14. The final capacitor should be large enough to smooth out the voltage ripple on the output.continuous sensors . and have less friction. ∆ V = output voltage K = constant for devic e ∆ x = core displacement signal out AC input Figure 14.10 A rod drives the sliding core ∆x ∆V = K∆x where.14. The capacitor and resistor values can be selected to act as a low pass filter.8 An LVDT Aside: The circuit below can be used to produce a voltage that is proportional to position.

The stationary strip is offset at an angle so that the strips interfere to give irregular patterns.10. Both of the strips are transparent (or reflective).3 . . You will notice that shifting one sheet will cause the stripes to move up or down. depending upon the speed and direction of motion.14. or down. Moving Stationary Note: you can recreate this effect with the strips below. overlay the sheets and hold them up to the light. A major disadvantage of these sensors is the high cost.11 control. with black lines at measured intervals. Photocopy the pattern twice.11. Two sensors. This is then passed through the fringes to be detected by light sensors.2. can act as a quadrature pair.3. and the same method used for quadrature encoders can be used to determine direction and distance of motion.continuous sensors . as shown in Figure 14.Moire Fringes High precision linear displacement measurements can be made with Moire Fringes. 14. often in the thousands. The spacing of the lines determines the accuracy of the position measurements. They are often used for pressure measurements with Bourdon tubes and bellows/ diaphragms. A light source is collimated by passing it through a narrow slit to make it one slit width. close enough. Figure 14.10 The Moire Fringe Effect A device to measure the motion of the moire fringes is shown in Figure 14. At least two light sensors are needed to detect the bright and dark locations. As the moving strip travels by a stationary strip the patterns will move up.

12 on off on off Figure 14. typically used for measuring vibrations . but the cost will be high. 14. often meters.11 Measuring Motion with Moire Fringes These are used in high precision applications over long distances. Typical applications include Coordinate Measuring Machines (CMMs).12. A base and housing enclose the sensor.12 A Cross Section of an Accelerometer Accelerometers are dynamic sensors. Base Mounting Stud Force Sensor Mass Housing Figure 14. as shown in Figure 14. the inertial resistance of the mass will cause the force sensor to deflect.Accelerometers Accelerometers measure acceleration using a mass suspended on a force sensor.14. A small mounting stud (a threaded shaft) is used to mount the accelerometer. In this case the mass is cantilevered on the force sensor. By measuring the deflection the acceleration can be determined. When the sensor accelerates.2.3.continuous sensors .4 . They can be purchased from a number of suppliers.

for example a 1g shaker will hit a peak velocity of 9. • The control vibrations should be less than 1/3 of the signal for the error to be less than 12%). • These devices can be calibrated with shakers. and are fixed to the surface as often as possible to prevent noise from the cable striking the surface.13 Piezoelectric Accelerometer Sensitivities The force sensor is often a small piece of piezoelectric material (discussed later in this chapter). (Hz) 22 KHz 180KHz sensitivity 4. The smaller frequency results in a smaller sensitivity. Standard accelerometers can be linear up to 100. The accelerometer is electrically isolated from the vibration source so that the sensor may be grounded at the amplifier (to reduce electrical noise). Piezoelectric based accelerometers typically have parameters such as. A smaller resonant frequency limits the maximum frequency for the reading. Each accelerometer is different.5 pC/(m/s**2) . • Mass of the accelerometers should be less than a tenth of the measurement mass. Cables are fixed to the surface of the vibration source. resonant freq. Temperature variations will affect the accuracy of the sensors.continuous sensors .004 Figure 14.14.14. close to the accelerometer. Background vibrations can be detected by attaching control electrodes to non-vibrating surfaces. but some general application guidelines are.13 shows the sensitivity of two accelerometers with different resonant frequencies.000 m/s**2.81 m/s**2.000 m/s**2: high shock designs can be used up to 1. The units for sensitivity is charge per m/s**2.13 between 10Hz to 10KHz.000. There is often a trade-off between a wide frequency range and device sensitivity (note: higher sensitivity requires a larger mass). Figure 14. . The piezoelectic material can be used to measure the force in shear or compression. -100to250°C operating range 1mV/g to 30V/g sensitivity operate well below one forth of the natural frequency The accelerometer is mounted on the vibration source as shown in Figure 14.

15 Typical Connection for Accelerometers Accelerometers are commonly used for control systems that adjust speeds to reduce vibration and noise.14 Mounting an Accelerometer Equipment normally used when doing vibration testing is shown in Figure 14.14. Sensor preamp Source of vibrations.continuous sensors .15. The sensor needs to be mounted on the equipment to be tested. The signal from accelerometers can be . The voltage can then be analyzed to determine the vibration frequencies. and detect tool failure. or site for vibration measurement signal processor/ recorder control system Figure 14.14 Sealant to prevent moisture hookup wire isolated stud accelerometer isolated wafer surface Figure 14. A pre-amplifier normally converts the charge generated by the accelerometer to a voltage. Computer Controlled Milling machines now use these sensors to actively eliminate chatter.

14.4 Forces and Moments 14.4.16 The Electrical Properties of a Wire . the resistance of the wire is a function of the resistivity. advances in micromachining are already beginning to provide integrated circuit accelerometers at a low cost.15 integrated to find velocity and acceleration. w L V + I t V L LR = -. length.= ρ -. 14.2. Basically.Strain Gages Strain gages measure strain in materials using the change in resistance of a wire. and cross sectional area. The wire is glued to the surface of a part. R = resistance of wire V.= ρ ----I A wt where. Their current use is for airbag deployment systems in automobiles.1 . t = width and thickness A = cross sectional area of conductor ρ = resistivity of material Figure 14. Figure 14.continuous sensors . so that it undergoes the same strain as the part (at the mount point).2. But.16 shows the basic properties of the undeformed wire. Currently accelerometers cost hundreds or thousands per channel. I = voltage and current L = length of wir e w.

as defined below.16 has been deformed it will take on the new dimensions and resistance shown in Figure 14. If a force is applied as shown.= ----. as predicted by Young’s modulus.= Eε A wt F∴ε = --------Ewt F L'L(1 + ε) R' = ρ ------.  w ( 1 – νε )t ( 1 – νε ) w't' (1 + ε) ∴∆R = R' – R = R --------------------------------------.14. ν = poissons ratio for the material F = applied force E = Youngs modulus for the material σ.17 The Electrical and Mechanical Properties of the Deformed Wire . as predicted by Poison’s ratio. the wire will become longer. the cross sectional area will decrease.16 After the wire in Figure 14.17.– 1 ( 1 – νε ) ( 1 – νε ) where.continuous sensors .= ρ  --------------------------------------------. ε = stress and strain of material Aside: Gauge factor.  ∆R ----- R GF = -----------ε Figure 14. The new length and cross sectional area can then be used to find a new resistance. But. is a commonly used measure of stain guage sensitivity. w’ L’ t’ F Fσ = -.

continuous sensors . . so the wire is actually wound in a uniaxial or rosette pattern. and soldering leads to the strain gages. it must be placed in the direction of the normal stress. Then the resistance of the strain gage can be calculated using the given equation. In this circuit the variable resistor R2 would be tuned until Vo =0V. (Note: the gages cannot read shear stress.when Vo = 0V R3 R1 R4 R2 + R3 R5 Vo Rstrain Figure 14. The process of mounting strain gages involves surface cleaning.14. and if a shear force is present the gage will measure the resulting normal force at 45 degrees.) Rosette gages are less sensitive to direction. As a result.18 Measuring Strain with a Wheatstone Bridge A strain gage must be small for accurate readings. V+ R2 R 1 R strain = ------------. When using uniaxial gages the direction is important.19. Wheatstone bridges are used to amplify the small change. application of adhesives. as shown in Figure 14.17 Aside: Changes in strain gauge resistance are typically small (large values would require strains that would cause the gauges to plastically deform). These gauges are sold on thin films that are glued to the surface of a part.

2. so it may be used to measure force.18 stress direction uniaxial Figure 14.4. In the narrow neck the strain is proportional to the load on the member.continuous sensors .14. This often involves making a reading with no load. This also changes the capacitance of the crystal.19 Wire Arrangements in Strain Gages A design techniques using strain gages is to design a part with a narrowed neck to mount the strain gage on.20.20 Using a Narrow to Increase Strain Strain gauges are inexpensive.Piezoelectric When a crystal undergoes strain it displaces a small amount of charge. and can be used to measure a wide range of stresses with accuracies under 1%. An example application includes using strain gages to measure die forces during stamping to estimate when maintenance is needed.2 . In other words. These parts are often called load cells. when the distance between atoms in the crystal lattice changes some electrons are forced out or drawn in. This is known as . 14. rosette mounted in narrow section to increase strain effect F F Figure 14. Gages require calibration before each use. or a known load applied. as shown in Figure 14.

F b + q a c εab dC = -------i = εg ---. such as audio speakers. The change in capacitance is proportional to the change in the thickness. . 8. the strain in the material. Applying an electrical charge can induce strain.6*10**10 N/m**2) Figure 14. but they are also used for applications such as microphones and pressure sensors. allowing them to be used as actuators.06*10**-11 F/m) i = current generated F = force applied g = constant for material (quartz typ. C = capacitance chne ag a.19 the Piezoelectric effect. when used for static measurements they tend to drift or slowly lose charge.21 shows the relationships for a crystal undergoing a linear deformation. These sensors are best suited to dynamic measurements. 50*10**-3 Vm/N) E = Youngs modulus (quartz typ. 4. c = geometry of material F ε = dielectric constant (quartz typ. Figure 14.continuous sensors .21 The Piezoelectric Effect These crystals are used for force sensors.F c dt where. and the signal value will change.14. The charge generated is a function of the force applied. b. When using piezoelectric sensors charge amplifiers are needed to convert the small amount of charge to a larger voltage. and a constant specific to the material.

P = the gas pressure V = the volume of the gas n = the number of moles of the gas R = the ideal gas constant = T = the gas temperature When flowing. the flow may be smooth. Normally a fluid is considered incompressible. . The equation below is for calculation the Reynold’s number for fluid flow in a pipe. given sufficiently high enough temparatures. At a value between 7000 and 8000 the flow will become turbulent. In case of high flow rates or unrestricted flow.continuous sensors . or laminar.20 14.2.5 Liquids and Gases There are a number of factors to be considered when examining liquids and gasses. while a gas normally follows the ideal gas law. The Reynold’s number is used to determine the transition to turbulence.14. A value below 2000 will result in laminar flow. PV = nRT where. Also. turbulence may result. • Flow velocity • Density • Viscosity • Pressure There are a number of differences factors to be considered when dealing with fluids and gases. At a value of about 3000 the fluid flow will become uneven. or low enough pressures a fluid can be come a liquid.

connected to the end of the tube. A position sensor.5.14.) the tube will straighten.7psi approx.continuous sensors .Pressure Figure 14. pressure increase position sensor pressure increase position sensor pressure pressure a) Bourdon Tube Figure 14.1 . The Bourdon tube uses a circular pressure tube. R = Reynolds number V D ρ u = = = = velocity pipe diameter fluid densit y viscosity 14.22 Pressure Transducers b) Baffle .2.21 VDρ R = ----------u where. will be elongated when the pressure increases.22 shows different two mechanisms for pressure measurement. When the pressure inside is higher than the surrounding air pressure (14.

differential pressure transducer fluid flow Figure 14.2 . If there is no flow the pressure before and after the neck will be the same.22 These sensors are very common and have typical accuracies of 0.14. This is known as a Venturi valve.5.continuous sensors . the greater the pressure difference before and after the neck. 14.23 A Venturi Valve .Venturi Valves When a flowing fluid or gas passes through a narrow pipe section (neck) the pressure drops. The fluid flow rate will be proportional to the pressure difference before and at the neck (or after the neck) of the valve. Figure 14. The faster the fluid flow.5%.2.23 shows a Venturi valve being used to measure a fluid flow rate.

14.+ gz = C = ----------.14. And. and the difference will be proportional to the velocity squared. which makes them very reliable and durable.+ gz = C ρ 2 where. 2 2 p before v before p after v after --------------.continuous sensors . p v-.= -----------.+ gz ρ 2 ρ 2 p before v before p after v after --------------. It is also common to use Venturi valves to generate vacuums for actuators.24 The Pressure Relationship for a Venturi Valve Venturi valves allow pressures to be read without moving parts.+ -------------ρ 2 ρ 2 2 2 2 pbefore – p after  v after v before  = ρ  -------------. p = pressure 2 ρ = density v = velocity g = gravitational constant z = height above a reference C = constant Consider the centerline of the fluid flow through the valve.+ -------------.3 .+ ----------------. assume that the center line of the valve does not change. When leaving the point it decelerates.– ----------------- 2   2 2 The flow velocity v in the valve will be larger than the velocity in the larger pipe section before.+ ----------------.23 Aside: Bernoulli’s equation can be used to relate the pressure drop in a venturi valve. causing them to vibrate.5.2. As the fluid approaches the point of maximum vibration it accelerates. Figure 14. They work well for both fluids and gases. the right hand side of the expression will be positive. that relates the velocit and pressure before and after it is compressed. So. Assume the fluid is incompressible. This will mean that the pressure before will always be higher than the pressure after. such as suction cups. The . so the density does not change. as shown below.Coriolis Flow Meter Fluid passes through thin tubes.+ ---. This gives us a simpler equation.

2. but this also increases the flow resistance. If the sound is against the flow it will take longer to arrive.Ultrasonic Flow Meter A transmitter emits a high frequency sound at point on a tube. and the other in the opposite direction.14.5.6 . The signal must then pass through the fluid to a detector where it is picked up.5 . As the fluid moves. 14. The typical accuracy for these sensors is 0.4 . one traveling forward. Measuring the frequency allows an estimate of the flow rate.2. and therefore don’t result in pressure drops.1%. 14. The amount of bending is proportional to the velocity of the fluid flow. These sensors tend be low cost and are popular for low accuracy applications. The typical accuracy for a corriolis flowmeter is 0. These flowmeters don’t oppose fluid flow. then the detected frequency will be lowered. the electrons in the fluid experience an electromotive force. The frequency of the vortices will be proportional to the flow rate.continuous sensors . If the particle is moving away from the emitter and detector pair.24 result is a distributed force that causes a bending moment.5. The higher the flow rate.Vortex Flow Meter Fluid flowing past a large (typically flat) obstacle will shed vortices. If the fluid is flowing in the same direction as the sound it will arrive sooner. A doppler flowmeter bounces a soundwave off particle in a flow. The difference in travel time for the sounds is used to determine the flow velocity.5%.Magnetic Flow Meter A magnetic sensor applies a magnetic field perpendicular to the flow of a conductive fluid. In a transit time flow meter two sounds are used. The result is that a potential (voltage) can be measured perpendicular to the direction of the flow and the magnetic field. Some of the devices also use bent tubes to increase the sensitivity.5. . and so they don’t result in pressure drops. and result is significant loses. The transmitter and receiver have a minimal impact on the fluid flow. the greater the voltage. if it is moving towards them the frequency will be higher. 14.2. These devices typically have a large constriction on the flow. and hence twisting of the pipe.

8 .14.e.6 Temperature Temperature measurements are very common with control systems.5.Pitot Tubes Gas flow rates can be measured using Pitot tubes.e.g. very low temperatures <-60 deg C .Positive Displacement Meters In some cases more precise readings of flow rates and volumes may be required.g. as shown in Figure 14.e.e.g. 14.2. environmental controls high temperature measurements <3000 deg F . superconductors in MRI units low temperature measurement -60 to 0 deg C . gas flow pitot tube connecting hose pressure sensor Figure 14.25 14.25. In effect these meters are like pumps run in reverse. metal refining/processing .g. The diameter of the tube is small (typically less than 1/8") so that it doesn’t affect the flow. The temperature ranges are normally described with the following classifications.7 . These are small tubes that project into a flow.5. These can be obtained by using a positive displacement meter. normally on a rotating shaft.25 Pitot Tubes for Measuring Gas Flow Rates 14.2.2. As the fluid is pushed through the meter it produces a measurable output. freezer controls fine temperature measurements 0 to 100 deg C .continuous sensors .

But.2 .) The current through the RTD should be kept to a minimum to prevent self heating. plasma systems 14. (Note: when designing assemblies.g. a temperature can be measured using the resistance of a wire. and can have accuracies of 0. Material Temperature Range C (F) Typical Resistance (ohms) 100 120 10 Platinum -200 . and covered for protection. The basic calculations for thermocouples are shown in Figure 14.Thermocouples Each metal has a natural potential level.27.26 RTD Properties These devices have positive temperature coefficients that cause resistance to increase linearly with temperature.26 very high temperatures > 2000 deg C . This calculation provides the measured voltage using a reference temperature and a constant specific .572) Copper -200 . dissimilar metals should not touch.1 . they can be expensive 14.e. These devices are more linear than thermocouples. copper or nickel-iron alloys.) Thermocouples use a junction of dissimilar metals to generate a voltage proportional to temperature. Seebeck. and when two different metals touch there is a small potential difference. The total resistance of an RTD might double over the temperature range. The resistances of these alloys are shown in Figure 14.2. that will increase by 0.6.2.4 ohms/°C. Resistive Temperature Detectors (RTDs) normally use a wire or film of platinum. So.Resistive Temperature Detectors (RTDs) When a metal wire is heated the resistance increases.260 (-328 .J. a voltage.850 (-328 . This principle was discovered by T.500) Figure 14. A platinum RTD might have a resistance of 100 ohms at 0C. this will lead to corrosion.continuous sensors .300 (-112 . nickel.05%. The metals are wound or wrapped over an insulator. A current must be passed through the RTD to measure the resistance.1562) Nickel -80 .26.14.6. (Note: a voltage divider can be used to convert the resistance to a voltage.

V out V out = α ( T – Tref ) Vout ∴T = --------. measuring device + . and relatively inexpensive.(typical) °C T. The equation can also be rearranged to provide a temperature given a voltage.97 to 50. most PLC vendors sell thermocouple input cards that will allow multiple inputs into the PLC.63 0 to 16. For example. µV 50 -----.82 0 to 42. and the normal temperature ranges.07 .28 -8. Table 1: Thermocouple Types ANSI Type T J E K R S C Materials copper/constantan iron/constantan chromel/constantan chromel/aluminum platinum-13%rhodium/platinum platinum-10%rhodium/platinum tungsten-5%rhenium/tungsten-26%rhenium Temperature Range (°F) -200 to 400 0 to 870 -200 to 900 -200 to 1250 0 to 1450 0 to 1450 0 to 2760 Voltage Range (mV) -5.14.74 0 to 14. Tref = current and reference temperatures α = constant (V/C) Figure 14.27 Thermocouple Calculations The list in Table 1 shows different junction types.60 to 17. and signal conditioners are commonly available.78 -5.+ T ref α where.82 to 68.27 to the device.97 0 to 37.continuous sensors . Both thermocouples.

6. A series of thermocouples connected together in series produces a higher voltage and is called a thermopile.14. Readings can approach an accuracy of 0. a second thermocouple can be kept at a known temperature for reference. . to predict the resistance at another temperature.2.28 mV 80 E 60 J 40 K C 20 T R S 0 0 500 1000 1500 2000 2500 (°F ) Figure 14.28 Thermocouple Temperature Voltage Relationships (Approximate) The junction where the thermocouple is connected to the measurement instrument is normally cooled to reduce the thermocouple effects at those junctions. The basic calculation is shown in Figure 14.Thermistors Thermistors are non-linear devices. often metal oxide semiconductors The calculation uses a reference temperature and resistance. (Note: this is because the extra heat reduces electron mobility in the semiconductor.5%.29.3 . their resistance will decrease with an increase in temperature. 14. When using a thermocouple for precision measurement.continuous sensors . with a constant for the device. The expression can be rearranged to calculate the temperature given the resistance.) The resistance can change by more than 1000 times.

30 Thermistor Signal Conditioning Circuit .14.– ---. T o.29 Rt = Ro e 1 1 β  -.29 Thermistor Calculations Aside: The circuit below can be used to convert the resistance of the thermistor to a voltage using a Wheatstone bridge and an inverting amplifier. Ro.continuous sensors . R t = resistances at reference and measured temps.  T T o βTo ∴T = -------------------------------Rt  ----- + β T o ln  - Ro where. T = reference and actual temperatures β = constant for devic e Figure 14. +V R1 R3 R5 + R2 R4 Vout Figure 14.

The change in resistance is non-linear.6. But. and they have a higher resistance. inexpensive devices that are often made as beads. 14.7 Light 14.continuous sensors .1 .Light Dependant Resistors (LDR) Light dependant resistors (LDRs) change from high resistance (>Mohms) in bright light to low resistance (<Kohms) in the dark. have a limited temperature/resistance range and can be self heating.30 Thermistors are small. but the devices are not linear.7.2. they have limited temperature ranges. They output a digital reading and can have accuracies better than 0.2. and is also relatively slow (ms).01%.4 . or for production lines where it is not possible to mount other sensors to the material. The devices respond quickly to temperature changes. 14. .14. Pyrometers are non-contact temperature measuring devices that use radiated heat.2. These are normally used for high temperature applications. so junction effects are not an issue. Typical accuracies are 1%. or metallized surfaces. and require some knowledge of interfacing methods for serial or parallel data.Other Sensors IC sensors are becoming more popular.

Vout Vlow Figure 14. These measurements are normally made with electrodes that are in direct contact with the fluids.continuous sensors .8.2.31 A Light Level Detector Circuit 14.1 . often a liquid is often used to detect impurities.2. . V high These are common in low cost night lights.31 Aside: an LDR can be used in a voltage divider to convert the change in resistance to a measurable voltage.14. to a strong acid.pH The pH of an ionic fluid can be measured over the range from a strong base (alkaline) with pH=14. High frequency inductive fields is another alternative. pH=0.8. This can be measured directly be applying a voltage across two plates submerged in the liquid and measuring the current. 14.Conductivity Conductivity of a material.2 . pH=7.2. to a neutral value.8 Chemical 14.

This is often caused by electromagnetic fields on the factory floor inducing currents in exposed conductors.32 14. This sheath needs to be connected to the measuring device to allow induced currents to be passed to ground. then the amount of noise induced is reduced.2. This prevents electromagnetic waves to induce voltages in the signal wires.9 Others A number of other detectors/sensors are listed below. as shown in Figure 14.normally in gases Dew Point . Analog signal are prone to electrical noise problems.shared or isolated reference voltages (commons). Combustion . When a signal is transmitted through a wire. twisted pairs . . Some of the techniques for dealing with electrical noise include.3 INPUT ISSUES Signals from sensors are often not in a form that can be directly input to a controller. If the wires are twisted. A shielded cable has a metal sheath.the wires are twisted to reduce the noise induced by magnetic fields. This technique is common in signal wires and network cables.14.shielding is used to reduce the effects of electromagnetic interference.gases such as CO2 can be an indicator of combustion Humidity . If the wires have an area between them the magnetic flux enclosed in the loop can induce current flow and voltages. but it may also include noise filters. shielding . This section will discuss the electrical and electronic interfaces between sensors and controllers. single/double ended inputs .to determine when condensation will form 14. In these cases it may be necessary to buy or build signal conditioners. a signal conditioner is an amplifier. a few times per inch. it must return along another path. Normally.32.continuous sensors . and circuitry to convert from current to voltage.

33. this is called single ended. The least expensive method uses one shared common for all analog signals. this is called double ended. or reference voltage can be connected different ways.14. when double ended inputs are used the number of available inputs is halved. But. Most analog output cards are double ended. as shown in Figure 14.continuous sensors . The more accurate method is to use separate commons for each signal. . Most analog input cards allow a choice between one or the other.33 A Shield is a metal sheath that surrounds the wires Analog voltage source + - Analog Input IN1 REF1 SHLD Figure 14.32 Cable Shielding When connecting analog voltage sources to a controller the common.

14. Amplifiers are used to increase the magnitude of these signals. .continuous sensors . and well suited to simple design and construction projects. so the output will have an opposite sign from the input.34.with this arrangement the signal quality can be better. When purchasing a commercial signal conditioner.33 Single and Double Ended Inputs Signals from transducers are typically too small to be read by a normal analog input card. Adjustments are provided for gain and offset adjustments. common.34 device + #1 device + #1 - Ain 0 Ain 1 Ain 2 Ain 3 Ain 4 Ain 5 Ain 6 Ain 7 COM device + #1 device + #1 - Ain 0 Ain 0 Ain 1 Ain 1 Ain 2 Ain 2 Ain 3 Ain 3 Single ended . Figure 14. An example of a single ended signal amplifier is shown in Figure 14. but fewer inputs are available. Note: op-amps are used in this section to implement the amplifiers because they are inexpensive. The amplifier is in an inverting configuration. but more inputs are available. the circuitry will be more complex. Double ended . and include other circuitry for other factors such as temperature compensation.with this arrangement the signal quality can be poorer.

14.continuous sensors . Note that Rc converts a current to a voltage. The voltage is then amplified to a larger voltage. R1 Iin Rc R2 + Rf Vout R3 R4 Figure 14.35.35 +V Ro -V + Vout offset Rf Rg gain Vin Ri Rf + Rg V out =  ---------------. Vin + offset  R  i Figure 14.34 A Single Ended Signal Amplifier A differential amplifier with a current input is shown in Figure 14.35 A Current Amplifier .

If the resistor values are all made the same (and close to the value of R3) then the equation can be simplified. as shown in Figure 14.continuous sensors .36 The circuit in Figure 14. .37.14. The following amplifier amplifies the voltage difference. The two input op-amps are used as unity gain followers.36 will convert a differential (double ended) signal to a single ended signal. + Vin + Vout + CMRR adjust Figure 14. to create a high input impedance.36 A Differential Input to Single Ended Output Amplifier The Wheatstone bridge can be used to convert a resistance to a voltage output.

A meter to indicate electrical current.continuous sensors .37 +V R1 R3 R5 + R2 R4 Vout R2 1.  ----. By measuring the displacement of the tip the pressure can be measured. It can be set up so that when pressure changes.transducer used to measure rotational speed.14. 2") but it is very thin (type 1/4").37 A Resistance to Voltage Amplifier 14. It is normally part of a DMM Bellows . Bourdon tube . Chromatographic instruments . Out- .+ ----. This often looks like a cylinder with a large radius (typ.+ ----- – -----   R + R   R3 R 4 R5 R3 1 2 or if R = R1 = R2 = R4 = R 5 RVout = V  -------.This is a flexible volumed that will expand or contract with a pressure change. When the pressure inside changes the moon shape will tend to straighten out.Widely used industrial gage to measure pressure and vacuum. Inductance-coil pulse generator .4 SENSOR GLOSSARY Ammeter .laboratory-type instruments used to analyze chemical compounds and gases. the displacement of one side can be measured to determine pressure.11Vout = V ( R5 )   ----------------.1. It resembles a crescent moon.  2R  3 Figure 14.

Widely used to indicate torque.transducer to measure flow rate.device to measure temperature by sensing the thermal radiation emitted from the object. and other variables. Ohmmeter .These use the interference of light waves 180 degrees out of phase to determine distances. which can be converted into voltage. pressure. Operation is based on change in resistance as a function of temperature. • Angular positions can be measured with potentiometers and encoders (more accurate). in which a junction of two dissimilar metals emits emf related to temperature. This is like a rotary transformer. Typical sources of the monochromatic light required are lasers. Output is pulse train. an instrument used to measure temperature.meter to indicate electrical resistance Optical Pyrometer .instrument used to measure voltage Pressure Transducers . Orifice Plate . Interferometers . Operation of the transducer can be based on strain gages or other devices. Typical output is voltage.this device is similar to an incremental encoder. . • Tachometers are useful for measuring angular velocity.device used to measure flow rates.14.38 put is pulse train.5 SUMMARY • Selection of continuous sensors must include issues such as accuracy and resolution.A class of transducers used to measure pressure.Also called a resistance thermometer. Typical output is pulse train. except that it uses coils to generate magnetic fields.device to measure temperature of an object at high temperatures by sensing the brightness of an objects surface. Thermistor . Pitot Tube . Output is emf.continuous sensors . Turbine Flowmeter . Output is Voltage Manometer . Venturi Tube . Strain Gage . Radiation pyrometer .Laboratory device used to measure flow. Piezoelectric Accelerometer . and photoconductors. Resolver . Linear-Variable-Differential transformer (LVDT) electromechanical transducer used to measure angular or linear displacement.widely used temperature transducer based on the Seebeck effect.Transducer used to measure vibration.liquid column gage used widely in industry to measure pressure. force. photodiodes. Output is change in resistance due to strain. including phototubes. Potentiometer . 14. Positive displacement Flowmeter . Thermocouple .Variety of transducers used to measure flow.a class of transducers used to sense light.widely used flowmeter to indicate fluid flow rates Photometric Transducers . phototransistors.

• Accelerometers measure acceleration of masses.765V. • Flow rates can be measured with Venturi valves and pitot tubes. If possible identify prices for the units. what voltage will be generated at 1200F? 6. What is the resolution of an absolute optical encoder that has six binary tracks? nine tracks? twelve tracks? 4. The wiper arm is directly connected to the rotational joint so that a given rotation of the joint corresponds to an equal rotation of the wiper arm.continuous sensors . 14. a) If the joint is at 42 degrees. 2. thermocouples. • Strain gauges and piezoelectric elements measure force. and Bryan. Industrial Text Co. A Systems Approach to Programmable Controllers.. allen bradley. The power supply connected across the potentiometer is 5. LVDTs (limited range).39 • Linear positions can be measured with potentiometers (limited accuracy). • Input signals can be single ended for more inputs or double ended for more accuracy. moire fringes (high accuracy). Sensor manufacturers include (hyde park.7 PRACTICE PROBLEMS 1. omron. and the potentiometer output is 2. Name two types of inputs that would be analog input values (versus a digital value). If a thermocouple generates a voltage of 30mV at 800F and 40mV at 1000F. 14.A. E.14.. F. • Temperatures can be measured with RTDs.A. Theory and Implementation... what voltage will be output from the potentiometer? b) If the joint has been moved. what is the position of the potentiometer? 7. A potentiometer is to be used to measure the position of a rotating robot link (as a voltage divider).6 REFERENCES Bryan. Search the web for common sensor manufacturers for 5 different types of continuous sensors. Delmar Publishers Inc. Programmable Controllers. The motor is driving a reducing gear box with a 50:1 . etc. 1988. and the total wiper travel is 300 degrees. banner. A motor has an encoder mounted on it.0 V. Suggest a couple of methods for collecting data on the factory floor 5. 1992. • Pressure can be measured indirectly with bellows and Bourdon tubes. Swainston.) 3. L. and thermistors.

030 = α ( 800 – T ref ) 0.continuous sensors .040 50µV α = -------------------------. What is the difference between a strain gauge and an accelerometer? How do they work? 9. If the position of the geared down shaft needs to be positioned to 0. Temperature and displacement 2.40 ratio. 360°/512steps.= ------------1000 – 200 F . The potentiometer can rotate 300 degrees. etc 3. what is the minimum resolution of the incremental encoder? 8. 14. Write a ladder logic program to read the voltage from the potentiometer and convert it to an angle in radians stored in F8:0.ab.= -------------------------α 0. Use the equations for a permanent magnet DC motor to explain how it can be used as a tachometer.050V 0.040 = α ( 1000 – T ref ) 800 – T ref 1000 – T ref 1 -.com. 360°/4096steps 4. A potentiometer is connected to a PLC analog input card.com. 360°/64steps.040 800 – T ref = 750 – 0.25T ref Tref = 200F V out = 0. 10. Sensors can be found at www. V out = α ( T – Tref ) 0.8 PRACTICE PROBLEM SOLUTIONS 1. smart machines. What are the trade-offs between encoders and potentiometers? 11. and the voltage supply for the potentiometer is +/-10V.75T ref 50 = 0.= ----------------------.030 0. PLCs with analog inputs and network connections 5.omron. data bucket. www.1 degrees.00005 ( 1200 – 200 ) = 0.14.

 + V 1 = ( 5V – 0V )  ------------------ + 0V = 0.1 ------------. thus minizing the loses in resistor R.7V θ   300deg max θw 2.14. deg θ output = 0. 9. = 5 -------------  count count deg 360 -------rot count R = -----------------.= 50 ----θ output 1 deg deg θ input = 50  0. DMM KK· ω + ω  ----.  JR  JR · JR Vs = ω ( K ) + ω  ----. a) b) θw 42deg V out = ( V 2 – V 1 )  ---------.765V = ( 5V – 0V )  ------------------ + 0V  300deg θ w = 165.= 72 -------------deg rot 5 ------------count 8.1 ------------count θ input --------------. This is the reverse emf. strain gauge measures strain in a material using a stretching wire that increases resistance .41 6.9deg 7.765V = ( 5V – 0V )  ------------------ + 0V  300deg θw 2.accelerometers measure acceleration with a cantilevered mass on a piezoelectric element.continuous sensors . A dmm. = Vs  ----. or other high impedance instrument can be used to measure this.  K 2 . R V = Ks ω + When the motor shaft is turned by another torque source a voltage is generated that is proportional to the angular velocity.

FS BTW Rack: 0 Group: 0 Module: 0 BT Array: BT9:0 Data File: N7:0 Length: 37 Continuous: no BT9:1/EN BTR Rack: 0 Group: 0 Module: 0 BT Array: BT9:1 Data File: N7:37 Length: 20 Continuous: no CPT Dest F8:0 Expression "20.9 ASSIGNMENT PROBLEMS 1.0 . Potentiometers have limited ranges of motion 11.0 * N7:41 / 4095.0 * (F8:0 + 10) / 20" RAD Source F8:0 Dest F8:1 BT9:0/EN BT9:1/DN 14. . Note: use the quadrature encoding to determine the position of the motor. Write a simple C program to read incremental encoder inputs (A and B) to determine the current position of the encoder.10" CPT Dest F8:0 Expression "300. encoders cost more but can have higher resolutions.continuous sensors .42 10.14.

it is better to use discrete actuators to reduce costs and complexity.15. Typical electric motors use at least one electromagnetic coil. Even in their simplest form. and a stationary outside. The cost for such actuators can easily be higher than for the control system itself. These actuators also require sophisticated control techniques that will be discussed in later chapters.1 INTRODUCTION Continuous actuators allow a system to position or adjust outputs over a wide range of values. as the voltages supplied to the motors will vary the speeds and torques that they will provide. called the stator.2 ELECTRIC MOTORS An electric motor is composed of a rotating center. There are a variety of motor configuration the yields motors suitable for different applications. These motors use the attraction and repulsion of magnetic fields to induce forces.continuous actuators . Most notably. 15. In general. For example. A control system is required when a motor is used for an application that requires . continuous actuators tend to be mechanically complex devices. called the rotor.1 15. When a voltage is applied to these coils the result is a torque and rotation of an output shaft. CONTINUOUS ACTUATORS Topics: • Servo Motors. and sometimes permanent magnets to set up opposing fields. a linear slide system might be composed of a motor with an electronic controller driving a mechanical slide with a ball screw. AC and DC • Stepper motors • Single axis motion control • Hydraulic actuators Objectives: • To understand the main differences between continuous actuators • Be able to select a continuous actuator • To be able to plan a motion for a single servo actuator 15. when there is a choice. and hence motion.

2.continuous actuators . . The controller with then generate an output.1. The encoder is connected directly to the motor shaft to provide feedback of position.2 shows a simplified model of a motor.15. which in turn drives the motor. A typical controller is shown in Figure 15.2 continuous position or velocity. The output is then passed through a power amplifier. based on the system error. The magnetics provide a permanent magnetic field for the rotor to push against. When current is run through the wire loop it creates a magnetic field. command generator (e.g.1 A Typical Feedback Motor Controller 15. Figure 15.1 Basic Brushed DC Motors In a DC motor there is normally a set of coils on the rotor that turn inside a stator populated with permanent magnets.. In any controlled system a command generator is required to specify a desired position. PLC) desired position or velocity controller voltage/ current power amp amplified voltage/ current encoder motor Figure 15. The controller will compare the feedback from the encoder to the desired position or velocity to determine the system error.

3 I I magnetic field axis of rotation ω Figure 15.3. . In the figure the power is supplied to the rotor through graphite brushes rubbing against the commutator.continuous actuators . The commutator is split so that every half revolution the polarity of the voltage on the rotor.15. and the induced magnetic field reverses to push against the permanent magnets.2 A Simplified Rotor The power is delivered to the rotor using a commutator and brushes. as shown in Figure 15.

In addition. A feedback controller is used with these motors to provide motor positioning and velocity control. and the speed is proportional to the voltage. the use of brushes increases resistance.continuous actuators . The brushes are subject to wear.3 A Split Ring Commutator The direction of rotation will be determined by the polarity of the applied voltage. These motors are losing popularity to brushless motors.15. and lowers the motors efficiency. which increases maintenance costs. .4 motor shaft brushes Top Front split commutator split commutator motor shaft brushes V+ power supply V- Figure 15.

15. if the voltage is only on half the time.Vmax 100 t V max 0 Figure 15. above the range of human hearing.4 Pulse Width Modulation (PWM) For Control . As shown below the signal produces an effective voltage that is relative to the time that the signal is on. V max 0 50% duty cycle 50V eff = -------. This method is popular because it can produce a variable effective voltage efficiently. the effective voltage is half the maximum voltage. So.5 ASIDE: The controller to drive a servo motor normally uses a Pulse Width Modulated (PWM) signal. When the voltage is on all the time the effective voltage delivered is the maximum voltage.continuous actuators . The percentage of time that the signal is on is called the duty cycle. The frequency of these waves is normally above 20KHz.Vmax 100 t V max 0 V max 0 100% duty cycle V eff = 100 Vmax -------100 t 0% duty cycle 0V eff = -------.Vmax 100 t 20% duty cycle 20V eff = -------.

V+ signal source com DC motor V+ V- power supply Figure 15.continuous actuators . thus switching power to the motor.5 PWM Unidirectional Motor Control Circuit .6 ASIDE: A PWM signal can be used to drive a motor with the circuit shown below. so the motor may only turn in one direction.15. In this case the voltage polarity on the motor will always be the same direction. The PWM signal switches the NPN transistor.

. As the AC current reverses. To reverse the direction the PWM signal is applied to the Vb inputs. while the Vb inputs are held low. • In AC motors the AC current is used to create changing fields in the motor. In this arrangement the positive voltage is at the left side of the motor.2 AC Motors • Power is normally generated as 3-phase AC. the magnetic field in the rotor appears to rotate. To turn the motor in one direction the PWM signal is applied to the Va inputs. This applies the positive voltage to the right side of the motor.7 ASIDE: When a motor is to be controlled with PWM in two directions the H-bridge circuit (shown below) is a popular choice.continuous actuators . so using this increases the efficiency of electrical drives. while the Va inputs are held low. Each pole is a pair of windings.6 PWM Bidirectional Motor Control Circuit 15.2. or purchased as integrated circuits for smaller motors. +Vs Va Vb Vb Va -Vs Figure 15. • Typically AC motors have windings on the stator with multiple poles.15. These can be built with individual components.

continuous actuators .7 A 2 Pole Single Phase AC Motor .15.8 L1 stator windings rotor Neut. Figure 15.

Figure 15. The difference is called the slip. Their function is to intersect changing magnetic fields from the stator. The changing field induces currents in the rotor.the rotor has coils wound. The result would be a collapse of the rotor magnetic field. This is typically a few percent. More poles results in a lower rotation of the motor. it will not see a changing magnetic field. . As the motor is loaded the slip will increase until the motor stalls. Neut. As a result an induction motor always turns slightly slower than the stator field.8 A 6 Pole 3-Phase AC Motor • The number of windings (poles) can be an integer multiple of the number of phases of power.has the shape of a wheel with end caps and bars Wound Rotor .15. • Rotor types for induction motors are listed below. Squirrel cage .9 L2 L1 L3 Neut.continuous actuators . If the motor turns at the precise speed of the stator field. These may be connected to external contacts via commutator • Induction motors require slip. generating a torque. These currents in turn set up magnetic fields that oppose fields from the stator. Neut.

Synchronous motor drives control the speed of the motors by synthesizing a variable frequency AC waveform.9 AC Motor Speed Control These drives should be used for applications that only require a single rotational direction. When an AC voltage is applied to the stator coils an AC magnetic field is created.9. It is possible to control the speed of the motor by controlling the frequency of the AC voltage.10 An induction motor has the windings on the stator. Controller Figure 15. . The motor actually has the maximum torque below the synchronous speed. The squirrel cage is a cast aluminum core that when exposed to a changing magnetic field will set up an opposing field. The rotor is normally a squirrel cage design.continuous actuators . The motor will turn at a frequency close to that of the applied voltage. When the motor is used with a fixed frequency AC source the synchronous speed of the motor will be the frequency of AC voltage divided by the number of poles in the motor. When a feedback controller is used the issue of slip becomes insignificant. For example a motor 2 pole motor might have a synchronous speed of (2*60*60/2) 3600 RPM.10. but there is always some slip. as shown in Figure 15.15. the squirrel cage sets up an opposing magnetic field and the resulting torque causes the motor to turn. The torque speed curve for a typical induction motor is shown in Figure 15. but be rated for 3520 RPM.

etc. 1 – ------------. and number of poles.) RPM = motor speed in rotations per minut e S = motor slip • Single phase AC motors can run in either direction. 6..continuous actuators . 4. f = power frequency (60Hz typ. • Universal motors were presented earlier for DC applications. To compensate for this a shading pole is used on the stator windings. An ideal motor with no load would have a slip of 0%.  100% p where. S f120 RPM = ---------.. slip.) p = number of poles (2.15.11 torque operating range synchronous speed speed Figure 15.10 Torque Speed Curve for an Induction Motor • The figure below shows the relationship between the motor speed and applied power. It basically acts as an inductor to one side of the field which slows the filed buildup and collapse. but they can also be . • Thermal protection is normally used in motors to prevent overheating. The result is that the field strength seems to naturally rotate.

• Starting AC motors can be hard because of the low torque at low speeds. At low speeds other coils or capacitors are connected into the circuits. To deal with this a switching arrangement is often used. The coils are switched on to attract or repel the permanent magnet rotor. as shown in Figure 15.15. At higher speeds centrifugal switches disconnect these and the motor behavior switches. The lack of brushes means that these motors require less maintenance than the brushed DC motors.continuous actuators . each corresponding to one power input. • Synchronous motors are different from induction motors in that they are designed to rotate at the frequency of the fields.12 used for AC power sources.3 Brushless DC Motors Brushless motors use a permanent magnet on the rotor. 15. in other words there is no slip. This is because the field polarity in the rotor and stator both reverse as the AC current reverses. Each of coils is separately controlled.11. • Synchronous motors use generated fields in the rotor to oppose the stators field.2. . Therefore there is no need to use brushes and a commutator to switch the polarity of the voltage on the coil. and use windings on the stator. A typical Brushless DC motor could have three poles.

continuous actuators . The speed of the motor is proportional to the frequency of the signal. The applied voltage can also be trapezoidal. A typical torque speed curve for a brushless motor is shown in Figure 15.13 V1 V2 N S V3 Figure 15.11 A Brushless DC Motor To continuously rotate these motors the current in the stator coils must alternate continuously.12. the motor will rotate continuously.15. The changing waveforms are controller using position feedback from the motor to select switching times. If the power supplied to the coils was a 3-phase AC sinusoidal waveform. . which will give a similar effect.

4 Stepper Motors Stepper motors are designed for positioning.13.5.continuous actuators . as shown in Figure 15. 15 and 30 degrees. 1 a b 1a 1b a b 2a 2b unipolar 2 bipolar .14 torque speed Figure 15. There are two basic types of stepper motors. unipolar and bipolar. Other motors are designed for step sizes of 2.8 degrees giving 200 steps per revolution. The unipolar uses center tapped windings and can use a single power supply. The bipolar motor is simpler but requires a positive and negative supply and more complex switching circuitry. 5.15.12 Torque Speed Curve for a Brushless DC Motor 15.2. They move one step at a time with a typical step size of 1. 2.

4.. 4.14. When the motor slips it will move a number of degrees from the current position. when the motor is turned on we might apply the voltages as shown in line 1.. 4.15 Figure 15. then 3.14 Stepper Motor Control Sequence for a Unipolar Motor Stepper motors do not require feedback except when used in high reliability applications and when the dynamic conditions could lead to slip. If a set of outputs is kept on constantly the motor will be held in position. Stepper motors are relatively weak compared to other motor types. 3. this is normally a few thousand steps per second. When not turning the output voltages are held to keep the motor in position. then 4. then 1. The slip cannot be detected without position feedback.15. 3.. 1. 2.. etc. . . The torque speed curve for the motors is shown in Figure 15. Figure 15. A stepper motor slips when the holding torque is overcome. The voltage change patterns for a unipolar motor are shown in Figure 15. and then back to 1. To reverse the direction of the motor the sequence of steps can be reversed. These motors are also prone to resonant conditions because of the stepped motion control. 1a controller 2a 1b 2b stepper motor Step 1 2 3 4 1a 1 0 0 1 2a 0 1 1 0 1b 1 1 0 0 2b 0 0 1 1 To turn the motor the phases are stepped through 1. or it is accelerated too fast. For example. Reversing the sequence causes the motor to turn in the opposite direction. In addition they have different static and dynamic holding torques. 2. The dynamics of the motor and load limit the maximum speed of switching. To rotate the motor we would then output the voltages on line 2. eg.15.continuous actuators .13 Unipolar and Bipolar Stepper Motor Windings The motors are turned by applying different voltages at the motor terminals.

and are considered to have a relatively constant speed.15 Stepper Motor Torque Speed Curve The motors are used with controllers that perform many of the basic control functions. .16 torque speed Figure 15. or by putting a rheostat/resistor in series with the stator windings.the user indicates maximum velocity and acceleration and a distance to move indexer . .have the rotor and stator coils connected in parallel.the user indicates direction and number of steps to take microstepping . and slowing the motor. A more sophisticated indexing controller will accept motion parameters. no permanent magnets) • Shunt motors .each step is subdivided into smaller steps to give more resolution 15. and convert them to individual steps..e. .these motors have a relatively small variation in speed as they are varied.when the load on these motors is reduced the current flow increases slightly. increasing the field. such as distance. .the speed of the motor can be controlled by changing the supply voltage.2. At the minimum a translator controller will take care of switching the coil voltages.15.14 and converts them to sinusoidal steps.continuous actuators . Other types of controllers also provide finer step resolutions with a process known as microstepping. This effectively divides the logical steps described in Figure 15. translators .5 Wound Field Motors • Uses DC power on the rotor and stator to generate the magnetic field (i.

continuous actuators . Ia. Ra = Armature current. .17 Va I a = ----Ra T = K t Ia φ where.have the rotor and stator coils connected in series. voltage and resistanc e T = Torque on motor shaft K t = Motor speed constant φ = motor field flux ω operating range T • Series motors\ . V a.as the motor speed increases the current increases.15. the motor can theoreti- .

18 cally accelerate to infinite speeds if unloaded. Rf = Armature and field coil resistance T = Torque on motor shaft K t = Motor speed constant 2 φ = motor field flux ω stall torque T .these motors typically have greater starting torques that shunt motors Va I a = ---------------Ra + R f T = KtIaφ = KtIa where.continuous actuators . I a. Va = Armature current. .15. This makes the dangerous when used in applications where they are potentially unloaded. voltae g R a.

.16 Equations for an armature controlled DC motor • Compound motors\ .continuous actuators .have the rotor and stator coils connected in series.19 The XXXXXXX e f = r a i a + Dl a i a + e m e m = K e θD T = KTia e a = ( r a + l a D )ia + K e Dθ T e a = ( r a + l a D )  ----. + K e Dθ  K - T Figure 15.15.cumulative compound motors have the shunt and series winding fields aligned so that they add ω cumulative differential T .differential compound motors have the shunt and series winding field aligned so that they oppose each other. .

15. The remainder of the provides the hydraulic power to drive the system.= θ --.20 e f = rf i f + l f i f D T = KTif T -.f  r + l D ef if ef f f Figure 15. This controls the flow of fluid to the actuator.=  ----------------------. When used for continuous actuation they are mainly used with position feedback. .= ----------------------2 T JD + BD KT θ θ --. and drivers a servo valve..= JD 2 + BD θ θ 1 -.18..= K T  ----------------- -.3 HYDRAULICS Hydraulic systems are used in applications requiring a large amount of force and slow speeds.continuous actuators .= -.T = -----------------------2 if T if JD + BD  i KT  θ 1 --.f  r + l D 2 ef if ef  JD + BD f f i T 1 --.17 Equations for a controlled field motor 15.  ----------------- -. The controller examines the position of the hydraulic system. An example system is shown in Figure 15.= T --.

15. The primary difference is that they have a limited travel and their cost is typically much higher than other linear actuators. valve designs use a scheme like pulse with modulation (PWM) which open/close the valve quickly to adjust the flow rate.continuous actuators . Ball Screws . .rotation is converted to linear motion using balls screws.21 controller valve hydraulic power supply position sensor position hydraulic actuator sump Figure 15. Heaters . Newer. or an SCR can be used to supply part of an AC waveform. more expensive. Linear Motors .a linear motor works on the same principles as a normal rotary motor.18 Hydraulic Servo System The valve used in a hydraulic system is typically a solenoid controlled valve that is simply opened or closed.air controlled systems can be used for positioning with suitable feedback.15.4 OTHER SYSTEMS The continuous actuators discussed earlier in the chapter are the more common types. Velocities can also be controlled using fast acting valves.to control a heater with a continuous temperature a PWM scheme can be used to limit a DC voltage. These are low friction screws that drive nuts filled with ball bearings. For the purposes of completeness additional actuators are listed and described briefly below. Pneumatics . These are normally used with slides to bear mechanical loads.

025” a) to achieve this control resolution how many step angles are required on the stepper motor? b) What is the corresponding step angle? c) Determine the pulse rate that will be required to drive a given joint at a velocity of 3. a pulse train is to be generated by the robot controller. For the stepper motor in the previous question. 2. It is desired that the control resolution of each of the axes be 0.6 PRACTICE PROBLEMS 1. . what pulse rate must be generated by the robot controller? 3.5 SUMMARY • AC motors work at higher speeds • DC motors work over a range of speeds • Motion control introduces velocity and acceleration limits to servo control • Hydraulics make positioning easy 15. a) How many pulses are required to rotate the motor through three complete revolutions? b) If it is desired to rotate the motor at a speed of 25 rev/min. The motor output shaft will be connected to a screw thread with a screw pitch of 0. Explain the differences between stepper motors.125”. A stepping motor is to be used to drive each of the three linear axes of a cartesian coordinate robot.continuous actuators .15.22 15.0”/sec. variable frequency induction motors and DC motors using tables.

  5 step = 125 steps = 125  1min steps = 2.= 120 steps -----------in s 0.025 --------stepR rotθ = -.2 --------. rot pulses rot--------------.8 ASSIGNMENT PROBLEMS 1.025 --------step 2.  rot b) c) rotdeg θ = 0.15.7 PRACTICE PROBLEM SOLUTIONS 1.23 15. a) pulses = ( 3rot )  5 step = 15steps  --------.125  -----.2 --------step in 0. For each pulse received from the pulse train source the motor rotates through a distance of one step angle. .2 --------P step in0. a pulse train is to be generated by a motion controller. a) What is the resolution of the stepper motor? b) Relate this value to the definitions of control resolution.025 --------step Thus 1 -----------------.continuous actuators . A stepper motor is to be used to actuate one joint of a robot arm in a light duty pick and place application.125  -----.08 step ------------------- min  --------. The step angle of the motor is 10 degrees.  rot in R = 0.  ------------ -----------s rot min 60s min s b) 3. spatial resolution.= 72 --------step step 3 in ---s PPS = -----------------------.= 0.=  25 -------. as discussed in class. and accuracy. speed very low speeds limited speed range wide range torque low torque good at rated speed decreases at higher speeds stepper motor vfd dc motor 15.= -------------------------.= 5 step --------rotrot 0. c) For the stepper motor. a) inP = 0.

24 How many pulses are required to rotate the motor through three complete revolutions? If it is desired to rotate the motor at a speed of 25 rev/min.continuous actuators . . Describe the voltage ripple that would occur when using a permanent magnet DC motor as a tachometer.15. Hint: consider the use of the commutator to switch the polarity of the coil. what pulse rate must be generated by the robot controller? 2.

In practical applications this setpoint needs to be generated automatically.16. . trapezoidal and smooth • Gain schedulers Objectives: • To understand single and multi axis motion control systems.multiaxis motion . A simple motion control system is used to generate setpoints over time.1. In earlier chapters we simply chose step inputs. The motion profile is then used to generate a set of setpoints. maximum acceleration and maximum velocity. such as a velocity.1 16. The setpoint scheduler will then use a realtime clock to output these setpoints to the motor drive. MOTION CONTROL Topics: • Motion controllers • Motion profiles. ramp inputs and other simple inputs to determine the system response. and times they should be output. The motion controller will accept commands or other inputs to generate a motion profile using parameters such as distance to move. An example of a motion control system is shown in Figure 16.1 INTRODUCTION A system with a feedback controller will attempt to drive the system to a state described by the desired input. 16.

.3 0. etc.) For position control we want a motion that has a velocity of zero at the start and end of the motion.4 .8 t (s) 0 0. drive and actuator is called an axis. When there is more than one drive and actuator the system is said to have multiple axes.4 0. . setpoint 0.16.4 0.2 0.2 MOTION PROFILES 16...multiaxis motion .1 A motion controller The combination of a motion controller.. and accelerates and decelerates smoothly..2.4 .2 0.6 0.. In this example the motion starts at 20 deg and ends at 100 deg.2 motion controller motion commands Motor Setpoint Generator Setpoint Scheduler Servo Drive setpoint schedule motion profile ω 0. Complex motion control systems such as computer controlled milling machines (CNC) and robots have 3 to 6 axes which must be moved in coordination.1 Velocity Profiles A simple example of a velocity profile for a point-to-point motion is shown in Figure 16. instead of degrees. 16. Figure 16.4 t 0 0.2 0.1 0.2. for positions velocities.0 0. (Note: in motion controllers it is more common to used encoder pulses..

16.3 θ ( deg ) 100 20 t (s) Figure 16. The area under the curve is the total distance moved. The slope of the initial and final ramp are the maximum acceleration and deceleration. This profile gives a continuous acceleration.3 An example of a velocity profile ω max = the maximum velocity .2 An example of a desired motion (position) A trapezoidal velocity profile is shown in Figure 16. Some controllers allow the user to use the acceleration and deceleration times instead of the maximum acceleration and deceleration. t dec = the acceleration and deceleration times t max = the times at the maximum veloci t y ttotal = the total motion time Figure 16.3. The top level of the trapezoid is the maximum velocity. ω ( deg ) ω max αmax 0 0 where.multiaxis motion . tacc t max t dec ttotal – αmax t (s) α max = the maximum acceleration t acc. but there will be a jerk (third order derivative) at the four sharp corners.

The system moves for 7.16. and the acceleration and deceleration times must be decreased. If this time is negative it indicates that the axis will not reach the maximum velocity. Figure 16. .t dec ω max = ω max  ------. The equations can be used to find the acceleration and deceleration times. ω max t acc = t dec = ----------αmax t total = t acc + t max + tdec t acc t dec 1 1 θ = -. The resulting velocity profile will be a triangle.+ t max + ------.5 seconds at the maximum velocity.– ---------.t acc.t dect max = ----------.  2 2  2 2 θ .4 The basic relationships for these variables are shown in Figure 16.– ---------ω max 2 2 (1) (2) (3) (4) Note: if the time calculated in equation 4 is negative then the axis never reaches maximum velocity.4 Velocity profile basic relationships For the example in Figure 16. and the velocity profile becomes a triangle.5 the move starts at 100deg and ends at 20 deg.t acc ω max + t max ω max + -. The acceleration and decelerations are completed in half a second.multiaxis motion .4. These equations can also be used to find the time at the maximum velocity.

= 7.5s = deg αmax 20 -------2 s θ = θ end – θ start = 20deg – 100deg = – 80deg θ .t acc.5s + 0. .5s 0.– --------.5 Given.– --------.multiaxis motion .5s t max = ----------.5s = 8. θ start = 100deg deg ω max = 10 -------s θ end = 20deg deg α max = 20 -------2 s The times can be calculated as.= 0.= --------------.5s Figure 16. The result in this example is a motion that accelerates for 0. t acc = t dec deg 10 -------ω max s ----------.5s + 7.16. This is made obvious by the negative time at maximum velocity.5 Velocity profile example The motion example in Figure 16.6 is so short the axis never reaches the maximum velocity.t dec80deg 0.= --------------.– ---------. In place of this the acceleration and deceleration times can be calculated by using the basic acceleration position relationship.5s ω max 2 2 deg 2 2 10 -------s t total = t acc + tmax + t dec = 0.316s and then decelerates for the same time.– ---------.

= --------------.– ---------.6 Velocity profile example without reaching maximum velocity Given the parameters calculated for the motion. .= 0.3 s ω max 2 2 deg 2 2 10 -------s The time was negative so the acceleration and deceleration times become. θ start = 20deg deg ω max = 10 -------s θ end = 22deg deg α max = 20 -------2 s The times can be calculated as. 2 θ -. the setpoints for motion can be calculated with the equations in Figure 16.– --------. deg 10 -------ω max s ----------.5s t max = ----------.= αmax 2deg --------------.t acc 2deg 0.6 Given.– --------.5s 0.5s t acc = t dec = deg αmax 20 -------2 s θ = θ end – θ start = 22deg – 20deg = 2deg t dec θ .7.= –0.= deg 20 -------2 s 0.316s 2 t max = 0s Figure 16.16.– ---------.multiaxis motion .1s = 0.= 1 αmax t acc -2 2 t acc = θ ----------.= --------------.

αmax t + θ start 2 t acc ≤ t < t acc + t max 2 θ ( t ) = 1 αmax t acc + ω max ( t – t acc ) + θ start -2 t acc + t max ≤ t < t acc + tmax + t dec 2 2 θ ( t ) = 1 α max t acc + ω max t max + 1 αmax ( t – t max – t acc ) + θ start --2 2 t acc + tmax + t dec ≤ t θ ( t ) = θ end Figure 16. In this subroutine the time is looped with fixed time steps. which will then be used elsewhere to guide the mechanism. .8.7 Generating points given motion parameters A subroutine that implements these is shown in Figure 16. 0s ≤ t < t acc 2 1 θ ( t ) = -. The position setpoint values are put into the setpoint array.7 Assuming the motion starts at 0s.16.multiaxis motion .

t_1) + theta_start.16. t += t_step){ if( t < t_1){ setpoint[*count] = 0. t <= t_total. This can be achieved by replacing the acceleration ramps with a smooth polynomial. *count = 0. double t_step.5*acc_max*t_acc*t_acc + vel_max*(t . } else if ( (t >= t_1) && (t < t_2)){ setpoint[*count] = 0. t_1 = t_acc. t_2 = t_acc + t_max. as shown in Figure 16.8 Subroutine for calculating motion setpoints In some cases the jerk should be minimized. t_total. double acc_max.multiaxis motion . double setpoint[]. and another two for the deceleration. } else { setpoint[*count] = theta_end. double t_max. t_1. } else if ( (t >= t_2) && (t < t_total)){ setpoint[*count] = 0. } setpoint[*count] = theta_end. .0.5*acc_max*t*t + theta_start. double vel_max. } *count++. for(t = 0.5*acc_max*t_acc*t_acc + vel_max*(t_max) + 0. } Figure 16.9. int *count){ double t. double theta_start.5*acc_max*(t-t_2)*(t-t_2) + theta_start. t_total = t_acc + t_max + t_acc.8 void generate_setpoint_table( double t_acc. t_2. In this case two quadratic polynomials will be used for the acceleration. *count++. double theta_end.

The curve found is for the first half of the acceleration. tdec = the acceleration and deceleration times t max = the ti eahmx u vlcy mst e ai me i m ot t total = the to l oot e t mt nm a ii Figure 16.16. t acc t max ω ( t ) = At + Bt + C 2 – α max t (s) t dec t total αmax = the maximum acceleration t acc.10.9 A smooth velocity profile ω max = the maximum velocity An example of calculating the polynomial coefficients is given in Figure 16.multiaxis motion . .9 ω ( deg ) ω max αmax 0 0 where. It can then be used for the three other required curves.

( t acc – t ) 4ω max α max 2 2 ω ( t ) = ωmax – -------------. α max The equation for the first segment is. 0 = A0 + B0 + C 0 = 2A0 + B ω max = Atacc 2 2 ∴C = 0 ∴B = 0 ω max A = ----------2 t acc α max A = ----------2t acc 2ω max tacc = -------------α max 2 α max = 2Atacc α max ω max A = ----------.10 Given.10 A smooth velocity profile example .t acc ---.( t – 2tacc t + t acc ) 4ω max 2 2 2 t acc 0 ≤ t < ------2 The equation for the second segment can be found using the first segment.= ----------2 2t acc t acc α max αmax α max A = ----------. θ start θ end ω max αmax The constraints for the polynomial are.≤ t < t acc 2 Figure 16.ω  ------.= -------------2t acc 4ω max  2ω max  2  -------------. t acc ------.= -----------------------. α max 2 -t ω ( t ) = -------------4ω max α max 2 ω ( t ) = ωmax – -------------. = α max dt dt  2  These can be used to calculate the polynomial coefficients. = -----------  2 2 dd.multiaxis motion .16. ω max t acc ω(0) = 0 ω  ------.ω ( 0 ) = 0 ---.

This equation can then be used in equation (1) to generate a smooth motion path between any arbitrary start . ------.multiaxis motion .dt + ∫  ω max – -------------. --. is.t acc + -------------------2 96ω max 2 2 2 2 2 3 t a cc ------2 3 3 2 α max  t acc tacc 3 so the time required at the maximum velocity is.– t acc + tacc – ω max ------. (3). 4 2  12ω max 8 4ω max  3 2 4ω max  24  θ acc θ acc ω max t acc αmax 3 α max 3 7α max 3 -t -t -t = ----------------.11 The distance covered during acceleration.+ ω max t acc – --------------. t acc ------2 2 2 2 θ acc αmax 3 -t = ----------------12ω max 2 3 2 0   α max t3 2 2 -------------.= ----------------. the area under the curves.acc .+ ------.2.+ -------------.– t acc t + t acc t  . is fairly simple.– ----------------. These methods are normally used when a controller does not have any velocity or acceleration limits. ( θ – 2θ acc ) t max = -------------------------ω max Figure 16. However. In this case the final polynomial equation.11 A smooth velocity profile example (cont’d) 16.acc + -------------------.+  ω max t –  4ω max  3  2 t a cc acc t  αmax t acc αmax  t acc 3 t acc 3  .2 Position Profiles A motion can be described using position points along a path. the parameters of the function are selected so that the motion starts and stops with a velocity of zero. The function value varies from 0 to 1 as the parameter ’u’ varies from 0 to 1. The method shown in Figure 16.( t – 2t acc t + t acc ) dt -t t acc 4ω max 4ω max  ------.acc – ----------------.------.16.– ------.acc 2 96ω max 12ω max 96ω max ω max t acc – 14αmax 3 = -------------------.12 controls motion using a parametric function ’p(u)’. ------. θ acc = ∫0 tacc ------2 t acc   αmax 2 αmax 2 2 -------------.

13 shows the use of a trigonometric function.12 and end point. This function was used to generate the points in the following sample program in Figure 16.multiaxis motion . t – t start θ ( t ) = θ start + ( θ end – θ start )p  ------------------------- t – t  end start where. p(0) = 0 d---.p ( 0 ) = 0 dt p(1) = 1 d---. t end = start and end times for the motion p ( u ) = Au + Bu + Cu + D The constraints for the polynomial are.p ( 1 ) = 0 dt 3 2 (1) (2) These can be used to calculate the polynomial coefficients. with arbitrary start and end times.12 Generating smooth motion paths The example in Figure 16.16. . instead of a polynomial. θ start. θ end = start and end positions of motio n t start. 0 = A0 + B0 + C0 + D 0 = 3A0 + 2B0 + C 0 = 3A1 + 2B1 1 = A1 + B1 + ( 0 )0 + 0 3 2 2 2 3 2 ∴D = 0 ∴C = 0 3 ∴B =  – -- A  2 ∴A = – 2 ∴B = 3 p ( u ) = – 2 u + 3u 3 2 (3) Figure 16.14.

14 generates a motion table that can then be used to update setpoints. The function ’table_init()’ must be called once when the program starts to set up global time and table values.sin  πt – -. π 1 1 p ( u ) = -. ’point_current’ at scheduled times. p ( u ) = A sin ( Bt + C ) + D The coefficients can be calculated using the conditions used previously.13 Generating smooth motion paths The program in Figure 16. + D = 1  2 ∴A ( 1 ) + A = 1 The final relationship is.p ( 1 ) = AB cos ( B ( 1 ) + C ) = 0 dt ∴ cos ( B + C ) = 0 π ∴B + C = -2 π p ( 0 ) = A sin  B ( 0 ) – -. When a new target position has been specified the ’table_generate()’ function is called to generate the setpoint table. . + -2 2 2  ∴B = π ∴A = D 1 ∴A = -2 Figure 16. d---.p ( 0 ) = AB cos ( B ( 0 ) + C ) = 0 dt π ∴C = – -∴ cos ( C ) = 0 2 d---. The ’table_update()’ function is called once every interrupt scan to check the setpoint table.16. This function also includes a simple clock to keep track of the system time. and update the global setpoint variable. + D = 0  2 π p ( 1 ) = A sin  π ( 1 ) – -.13 where.multiaxis motion .

for(i = 1. 794. 206. int end. int duration_sec){ unsigned i.3 MULTI AXIS MOTION In a machine with multiple axes the motions of individual axes must often be coordinated.start) * point_master[i] / 1000. point_position[TABLE_SIZE]. point_time[0] = ticks + 10. i < TABLE_SIZE. ticks. /* delay the start slightly */ point_position[0] = start. point_position[i] = start + (long int)(end . } int table_update(){/* interrupt driven encoder update */ ticks++.1).16. 976.multiaxis motion . outint16(point_current). 500.14 Subroutines for motion profile generation and use 16. } } return point_current. putch("\n"). } point_index = 0. /* a global variable to track position */ int table_init(){ /* ticks = 0. point_time[TABLE_SIZE].14 #define int int int int int int int TABLE_SIZE 11 point_master[TABLE_SIZE] = {0. /* mark the table as empty */ void table_generate(int start. } Figure 16. 345. A simple example would be a robot that needs to move two joints to reach a new position. i++){ point_time[i] = point_time[0] + (unsigned long)i * duration_sec * 250 / (TABLE_SIZE . /* update the clock */ if(point_index < TABLE_SIZE){ if(point_time[point_index] == ticks){ point_current = point_position[point_index++]. point_index. /* start the system at zero */ TABLE_SIZE. 24. 95. 905. 1000}. /* point_current point_index = } initialize the setpoint table */ set the clock to zero */ = 0. /* variables to keep a system clock count */ point_current. . We could extend the motion of the slower joints so that the motion of each joint would begin and end together. 655. point_start_time.

150) degrees/sec/sec. All of the joints reach the maximum acceleration. 40. 100.13s. 0)deg. and the maximum velocities are (20. 0. 50) degrees/sec. 80. These are done in vector format for simplicity.1 Slew Motion When the individual axis of a machine are not coordinated this is known as slew motion. Each of the axes will start moving at the same time. Consider the example in Figure 16. The maximum absolute accelerations and decelerations are (50. A three axis motion is required from the starting angles of (40. while the longest motion takes 4. and must end at (120. but finish at separate times.3.15.15 Multi-axis slew motion The calculations for the motion parameters are shown in Figure 16.multiaxis motion . Joint angle (degrees) 180 90 time(sec) θ3 -90 Joint velocity (degrees/sec) ω max – αmax θ2 θ1 t acc t max t dec Figure 16. The fastest motion is complete in 1. .15 16.16.16. -40)deg.4s.

4. ω max 20 40.4. ---------------------.333 ( 50 ) θ acc. -------------------------------. 0 – ( – 40 ) ) = ( 80.. but setting a new maximum velocity. = --------------------------.13 )s Figure 16. t total = t acc + t max + t dec = ( 4.=  ----------------.. 8. If this did not occur a straight line cut in the x-y plane would actually be two straight lines. . First we can determine the distance covered during acceleration.4 ( 40 ) 0. ----------------.=  -----.6.333 ) t max = ---------------------------------. and find the travel time at maximum velocity.4 ( 20 ) 0.50t acc = t dec = ----------.46668 )sec.1. The slew motion example can be extended to be slew motion where all joints finish their motion at 4. and deceleration.16. = ( 4. and deceleration and the time during acceleration.17 using the results from the example in Figure 16.4s.1 . This is shown in the example in Figure 16.16 Calculated times for the slew motion 16.  2  2 2 2 The next step is to examine the moves specified..16.6. 40 )deg. 0.4.vel.. θ move = θ end – θ start = ( 120 – 40. Remove the angles covered during accel./deccel.Interpolated Motion In interpolated motion the faster joints are slowed so that they finish in coordination with the slowest.4. -------- = ( 0. – 80. = θ dec.=  ---------------------. 0 – 80. --------. 2.. θ move – 2θ acc 80 – 2 ( 4 ) 80 – 2 ( 8 ) 40 – 2 ( 8. 8.3.multiaxis motion . This is essential in devices such as CNC milling machines.333 )sec.  50 100 150 α max t acc ω max.   ω max 20 40 50 t max = ( 3. This can be done by accelerating at the maximum acceleration. ----------------------.33 )deg. 1. 0. 0. 1.16 The area under the velocity curve is the distance (angle in this case) travelled. 0.

= 0.= 0.19s 100 A new maximum velocity can be calculated for joint 3 using this equation.17 The longest motion time is 4.22 t acc' = t acc' = --------.4s for joint 1. 19.-----------.3.ω max'  2 αmax αmax  2 αmax ω max' ∴∆ θ = ω max'  t total – -----------.0 2 19.ω max' +  t total – 2 -----------. 9.= 421.multiaxis motion .4 ) ( 100 ) – 4 ( 80 ) ( 100 ) ∴ω max' = ------------------------------------------------------------------------------------------------2 440 ± 401.22 9.-----------. The calculation begins by rewriting the velocity/position relationship using a new maximum velocity. it is executed by the setpoint scheduler.4 ( 150 ) ± ( 4. 4.2 Motion Scheduling After the setpoint schedule has been developed.16 16. ω max' + -.4 ( 100 ) ± ( 4.4 ) ( 150 ) – 4 ( 40 ) ( 150 ) ∴ω max' = ------------------------------------------------------------------------------------------------2 ∴ω max' = 651. The setpoint scheduler will use a clock to determine when an output setpoint .99502 ∴ω max' = ---------------------------------------.17 Interpolated motion based upon Figure 16. ω max' 1 ω max-' 1 ω max-' ∆θ = -.0 t acc' = t acc' = --------. and this can be used to prolong the other motions.16.  α max  ∴( ω max' ) + ω max' ( – t total αmax ) + ∆θαmax = 0 t total α max ± t total α max – 4∆θαmax ∴ω max' = ------------------------------------------------------------------------------------2 ω max' t acc' = t acc' = -----------αmax A new maximum velocity can be calculated for joint 2 using this equation. 4.092s 100 2 2 2 2 2 2 2 Figure 16.

16. A diagram of a scheduler is shown in Figure 16.18 A setpoint scheduler The output from the scheduler updates every time step.18. The servo drive is implemented with an algorithm such a PID control. Time based interrupt routine Setpoint table Choose new point from trajectory table Servo motor routine runs for each axis Read θ desired θ desired Interrupt Clock Compute error Output actuator signal Return Figure 16. .19. When enough time has elapsed the routine will move to the next value in the setpoint table.multiaxis motion . In this system the setpoint scheduler is an interrupt driven subroutine that compares the system clock to the total motion time.18 should be updated. This then leads to a situation where the axis is always chasing the target value. This leads to small errors. as shown in Figure 16. The frequency of the interrupt clock should be smaller than or equal to the time steps used to calculate the setpoints.

. paths are easy to generate. • Typically paths for these will repeat as shown below.19 speed actual position trajectory table time trajectory table time step position required actual time Figure 16.etc.pockets .19 Errors in path following 16. .4 PATH PLANNING • When we have simple features.multiaxis motion . These features include.holes .steps .16.

16.multiaxis motion . These surfaces will almost always be represented with spline patches. • Recall that a spline patch can be represented parametrically .20 • For complex surfaces we want to contour appropriately.

v ) = y p zp (u=0. a direction flag n=10 .2) .xp.start) .zp+0.the controller described in the block diagram below uses a model for a DC permanent magnet DC motor to estimate a voltage based upon a predicted velocity and position. . number of passes to cut the surface step=1.zp+0. calculate the start position print(“G00 X”. step sizes for u and v directions start=step/2 .zp] = p(start.” Z”.start+(n-j)*step) print(“G01 X”.v=0) (u=1.yp. dirn_flag = 1. will increment in the u direction for j=0 to (n-1) . move the tool to above the start position for i=0 to (n-1) . move to next cut line in the u direction print(“G00 Z”. instruction to cut to next point next j .2) .start+j*step) if dirn_flag=1 then [xp. move the tool to above the end position 16. reverse direction to cut in opposite direction next i .zp) .v=0) • A simple algorithm to cut the surface is shown below. make next step in v direction until done dirn_flag = -dirn_flag .zp]=p(start+i*step.zp]=p(start+i*step.the desired position and velocity are given in figure xxx based upon the motion .xp.5 CASE STUDIES .0/n .” Y”.” Z”. will increment in the v direction . calculate next point if dirn_flag=-1 then [xp.v=1) xp p ( u.multiaxis motion .21 (u=0.yp.yp.yp. the start offset in the u and v directions [xp.” Y”.16.v=1) (u=1.yp.

22 position control derived in the earlier section t – tstart θ ( t ) = θ start + ( θ end – θ start )p  ------------------------- t – t  end start p ( u ) = – 2 u + 3u 3 2 · t – t start ω ( t ) = ( θend – θ start )p  ------------------------- t – t  end start 2 · p ( u ) = – 6 u + 6u Figure 16.multiaxis motion .21 Feedforward trajectory controller .the controller that uses the desired position and velocity is shown in figure xxxx realtime control loop ωd JRD + K ----------------------K 2 physical system θd + - + P + K ----------------------2 JRD + K θa 1---------K enc K enc Figure 16.20 Motion position control equation .16.

The motion is to have a maximum acceleration of 300 counts/s/s. Find a motion profile that satisfies these constraints. • Trapezoidal and smooth motion profiles were presented. • Motion profiles can be used to generate setpoint tables. 2. The motor will start at a position count of 100.5 seconds for the entire motion. 16. Generate a table of setpoints for the desired position every 2 seconds. a) Develop a motion profile for a joint that moves from -100 degrees to 100 degrees with a maximum velocity of 20 deg/s and a maximum acceleration of 100deg/s/s. and end the motion at a count of 3000.16. b) Develop a setpoint table that has values for positions every 0.multiaxis motion . . Consider a basic servo controller with encoder feedback. • Values from the setpoints can then be output by a scheduler to drive an axis.23 16. and a maximum velocity of 100 counts/s.7 PRACTICE PROBLEMS 1.6 SUMMARY • Axis limits can be used to calculate motion profiles.

0 2.5 2.2s = 10.2s ) + 20 -------.5 5.5 3.= 9.5s + 20 deg ( 0.16. Given.5s – 0.0 deg 200deg – 20 -------.( 0.5 1.2s ∆θ – ω max t acc s --------------------------------. deg 20 -------ω max s t acc = t dec = ----------.5s = -.0 9.multiaxis motion .2s angle (deg) -100 -92 -82 -72 -62 -52 -42 -32 -22 -12 -2 8 18 28 38 48 58 68 78 88 98 100 100 2 1 deg deg θ 0.5 11.0 0.5 8.2s ) – 100deg 2 2 s s θ 0.8 PRACTICE PROBLEM SOLUTIONS 1.5s = – 92deg θ 1.5 9.24 16.0 10.5 4.= 0.2s αmax deg 100 -------s 2 ∆θ = 100 – ( – 100 ) = 200° t max t total t (s) 0.0.0 4.0 8.5s ) = – 92deg + 10deg -------s .0s = θ 0.0 7.100 -------.( 0.8s + 0.0 5.5 6.= -------------------------------------------------.= -----------------.0 3. deg ω max = 20 deg -------α max = 100 -------2 s s The motion times can be calculated.8s = ω max deg 20 -------s = t acc + t max + t dec = 0.2s + 9.0 1.5 7.5 10.0 6.

An axis has a maximum velocity of 2. a) slew motion b) interpolated motion 3.9 ASSIGNMENT PROBLEMS 1.0 seconds along the path for a total motion time of 10 seconds.25 2. and a maximum acceleration of 30 deg/s/s.multiaxis motion . Develop a smooth velocity profile for moving a cutting tool that starts at 1000 inches and moves to -1000 inches. Do this by developing an equation then calculating points every 1.5 rad/s^2. 100deg and -50deg respectively. The maximum velocity is 100 in/s and the maximum acceleration is 50in/s/s. Find a smooth path for a robot joint that will turn from θ= 75° to θ = -35° in 10 seconds. Assuming all of the joints start at an angle of 0 degrees. Each of the joints has a maximum velocity of 20 deg/s. Sketch a trapezoidal motion profile and find the motion time for a -4.16. Develop the motion profiles assuming.0 rad/s and a maximum acceleration/deceleration of 0. Do not assume a maximum velocity. 4. 2 and 3 move to 40 deg. Joints 1. t (s) 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 counts 100 283 483 683 883 1083 1283 1483 1683 1883 2083 2283 2483 2683 2883 3000 16. Paths are to be planned for a three axis motion controller. 2.0 rad motion. .

As with the other representations the Laplace s is related to the rate of change in the system.17. The normal convention is to show the function of time with a lower case letter.1 INTRODUCTION Laplace transforms provide a method for representing and analyzing linear systems using algebraic methods. . D = s s = σ + jω (if the initial conditions/derivatives are all zero at t=0s) Figure 17.2. In systems that begin undeflected and at rest the Laplace ’s’ can directly replace the d/dt operator in differential equations. for the steady state response. but also a real part.1 17. 17.4 for a first order derivative.numerical methods . It is a superset of the phasor representation in that it has both a complex part. while the same function in the s-domain is shown in upper case. Another useful observation is that the transform starts at t=0s. LAPLACE TRANSFORMS Topics: • Laplace transforms • Using tables to do Laplace transforms • Using the s-domain to find outputs • Solving Partial Fractions Objectives: • To be able to find time responses of linear systems using Laplace transforms.1 The Laplace s The basic definition of the Laplace transform is shown in Figure 17. Examples of the application of the transform are shown in Figure 17. representing the transient part.3 for a step function and in Figure 17.

numerical methods .2 F( s ) = where.3 Proof of the step function transform .2 The Laplace transform Aside: Proof of the step function transform. F( s) = ∞ ∞ ∞ 0 – s0 ∫0 f ( t )e – st dt = ∫0 5e – st dt = – 5 e -s – st = –5 e -s – s∞ 5e 5 – – -----------. ∫0 f ( t )e ∞ – st dt f ( t ) = the function in terms of time t F ( s ) = the function in terms of the Laplace s Figure 17.= -s s Figure 17. For f(t) = 5.17.

2 APPLYING LAPLACE TRANSFORMS The process of applying Laplace transforms to analyze a linear system involves the basic steps listed below.f ( t ) = ∫ (d/dt)f ( t )e dt dt 0 we can use integration by parts to go backwards.17. b = uv a – ∫ v du a b b ∫0 (d/dt)f ( t )e du = df ( t ) u = f( t ) ∴∫ f ( t ) ( – s )e 0 ∞ 0 ∞ – st ∞ – st dt v = e – st dv = – se dt = f ( t )e – st ∞ 0 – st dt ∞ 0 – st – ∫ (d/dt)f ( t )e – f ( t )e – 0s dt – st ∴∫ (d/dt)f ( t )e – st dt = [ f ( t )e – ∞s ] + s ∫ f ( t )e 0 ∞ dt d∴L ---. 17. ∞ – st dG ( s ) = L [ g ( t ) ] = L ---. ∫a udv therefore.numerical methods . however tables of values will be presented in a later section for the most popular transforms.f ( t ) = – f ( 0 ) + sL [ f ( t ) ] dt Figure 17.3 Aside: Proof of the first order derivative transform Given the derivative of a function g(t)=df(t)/dt. .4 Proof of the first order derivative transform The previous proofs were presented to establish the theoretical basis for this method.

2. Use partial fractions to reduce the output function to simpler components. Convert the system transfer function. Algebraically combine the input and transfer function to find an output function.5.5 are general properties normally used for manipulating equations. . The transforms shown in Figure 17. 4.8. 3.17. to the s-domain by replacing ’D’ with ’s’. 17.4 1.1 A Few Transform Tables Laplace transform tables are shown in Figure 17. 5. Figure 17.) 2.7 and Figure 17. (Note: If any of the initial conditions are non-zero these must be also be added. or differential equation. Convert the input function(s) to the s-domain using the transform tables. Convert the output equation back to the time-domain using the tables. These are commonly used when analyzing systems with Laplace transforms. and for converting them to/from the s-domain.numerical methods .

a > 0 e – at t L[f(t )] f(t ) f( s – a ) 1  s -.17.– … – ------------------n dt dt – n – ∫0 f( t ) dt f ( t – a )u ( t – a ).a a –df ( s ) -------------ds f(s) ( – 1 ) -------------n ds nd n f ( at ).f  -- .5 TIME DOMAIN f(t) Kf ( t ) f 1 ( t ) + f 2 ( t ) –f 3 ( t ) + … df ( t ) ----------dt d f(t ) ------------2 dt d f(t ) ------------n dt n 2 FREQUENCY DOMAIN f( s ) KL [ f ( t ) ] f1 ( s ) + f 2 ( s ) – f 3 ( s ) + … sL [ f ( t ) ] – f ( 0 ) df ( 0 ) 2 – s L [ f ( t ) ] – sf ( 0 ) – ----------------dt s L[ f( t ) ] – s L[ f( t ) ] ---------------s e – as n n–1 – – f(0 ) – s – n – 2 df ( 0 d f( 0 ) ) ----------------. a > 0 tf ( t ) t f(t ) f( t ) ------t n ∞ ∫ f ( u ) du s Figure 17.5 Laplace transform tables .numerical methods .

numerical methods .8 are normally used for converting to/from the time/s-domain. .6 ·· · L [ x + 7x + 8x = 9 ] = where.17. ·· x(0) = 1 · x(0) = 2 x(0) = 3 Figure 17.7 and Figure 17.6 Drill Problem: Converting a differential equation to s-domain The Laplace transform tables shown in Figure 17.

7 Laplace transform tables (continued) .n>0 e – at n exponential decay 1 ---------s+a ω ----------------2 2 s +ω s ----------------s +ω 2 2 sin ( ωt ) cos ( ωt ) te – at 1 -----------------2 (s + a) 2! -----------------3 (s + a) t e 2 – at Figure 17.17.7 TIME DOMAIN FREQUENCY DOMAIN δ(t) A t t 2 unit impulse step ramp 1 A -s 1 ---s 2 ---3 s n! ----------s n+1 2 t .numerical methods .

numerical methods .+ ------------------------------------s + α + βj s + α – βj A -----------------------------.8 Laplace transform tables (continued) .17.8 TIME DOMAIN e e – at FREQUENCY DOMAIN ω ------------------------------2 2 (s + a) + ω s+a ------------------------------(s + a) + ω 2 2 sin ( ωt ) cos ( ωt ) sin ( ωt ) – at e – at ω ------------------------------2 2 (s + a) + ω Bs + C ------------------------------2 2 (s + a ) + ω A A ----------------------.+ A .------------------------------------2 2 ( s + α – βj ) ( s + α + βj ) – bt complex conjugate complex conjugate e – at C – aB B cos ωt +  ---------------- sin ωt  ω  – αt 2Ae cos ( βt + θ ) cos ( βt + θ ) – at 2t A e – αt ( c – a )e – ( c – b )e --------------------------------------------------------b–a e –e ----------------------b–a – at – bt s+c -------------------------------( s + a )( s + b ) 1 -------------------------------( s + a )( s + b ) Figure 17.

These can be converted to the s-domain.numerical methods .10 Drill Problem: Converting from the s-d to time domain 17. and then transfer functions. were derived for mechanical and electrical systems. as .3 MODELING TRANSFER FUNCTIONS IN THE s-DOMAIN In previous chapters differential equations.17.9 f ( t ) = 5 sin ( 5t + 8 ) f ( s ) = L [f ( t ) ] = Figure 17.+ ---------s s+7 f(t) = L [f(s)] = –1 Figure 17.9 Drill Problem: Converting from the time to s-domain 5 6f ( s ) = -.

I ( t ) dt s-domain V ( s ) = RI ( s ) 1 I( s) V ( s ) =  --- ------- C s V ( s ) = LsI ( s ) Impedance Z = R 1Z = ----sC Z = Ls Figure 17. In this case we assume the system starts undeflected and at rest. Although this is very convenient. the ’D’ operator would be replaced with a more complex expression that includes the initial conditions. it is only valid if the initial conditions are zero.– … – ------------------L ------------.∫ I ( t ) dt C dV ( t ) = L ---. – n d f(0 ) n n–1 – n – 2 df ( 0 ) d f(t ) ----------------.11 A mass-spring-damper example Impedances in the s-domain are shown in Figure 17.12.numerical methods . so the ’D’ operator may be directly replaced with the Laplace ’s’.12 Impedances of electrical components . otherwise the more complex form.10 shown in the mass-spring-damper example in Figure 17. F = MD x + K d Dx + K s x Kd M Ks F( t ) --------. If the system did not start at rest and undeflected. shown below.= ---------.17. must be used.= Ms + K d s + K s x( s) x(t) 2 x F ASIDE: Here ’D’ is simply replaced with ’s’. Device Resistor Capacitor Inductor Time domain V ( t ) = RI ( t ) 1 V ( t ) = --.= MD 2 + K d D + K s x(t ) F(s) 2 F(t) L --------.= s L [ f ( t ) ] – s f( 0 ) – s n n dt dt dt n – Figure 17.11. As before these assume that the system starts undeflected and at rest.

numerical methods .11 Figure 17. To state the obvious. An example of this conversion for a step function is shown in Figure 17. these relate an output and an input. . L t=0sec Vi + R + Vo - C Treat the circuit as a voltage divider.4 FINDING OUTPUT EQUATIONS An input to a system is normally expressed as a function of time that can be converted to the s-domain. The calculation result is a transfer function.13 A circuit example At this point two transfer functions have been derived. using the impedances of the devices. 17.  1 + DCR R R Vo = -------------------------------------. R Vo R ----.=  ---------------------------------------------  2 2  Vi s R LC + sLR + R Figure 17.14. To find an output response.= V i  -------------------------------------------------  2 2    R D R LC + DLR + R DLR +  --------------------- 1  ------------------  1 + DCR DL +  1  DC + -.   1  ------------------ Vi  R 1 Vi  ---------------------  DC + -.= --------------------------------------------. The switch that closes at t=0s ensures that the circuit starts at rest.13 shows an example of circuit analysis using Laplace transforms. an input is needed.17. The circuit is analyzed as a voltage divider.

12 Apply a constant force of A.- F( s )  Ms 2 + K d s + K s s K d = 3000 Ns ----m N K s = 2000 --m M = 1000kg A = 1000N Therefore. Assume. starting at time t=0 sec. to transfer functions in the s-domain. (*Note: a force applied instantly is impossible but assumed) F(t) = 0 for t < 0 = A for t >= 0 Perform Laplace transform using tables A F ( s ) = L [ F ( t ) ] = -s Figure 17.numerical methods . Partial fraction methods allow the functions to be broken into .= ------------------------------------2 F( s ) Ms + K d s + K s A F ( s ) = -s   x(s) 1 x ( s ) =  ---------- F ( s ) =  ------------------------------------- A - . 1 ∴x ( s ) = --------------------------------2 ( s + 3s + 2 )s Figure 17. Given. If the transfer function is multiplied by the input function. both in the s-domain. the result is the system output in the s-domain. x(s) 1 ---------.14 An input function In the previous section we converted differential equations.17. for systems. These transfer functions are a ratio of output divided by input. It is often necessary to simplify the output function before it can be converted back to the time domain.15 A transfer function multiplied by the input function Output functions normally have complex forms that are not found directly in transform tables.

+ ( s + 1 ) ---------.= ( s + 1 ) A + B + ( s + 1 ) ----------( s + 2 )s s s+2 1 lim -----------------.= -----. the output function is easily .+ ---------.= A + ----------.5 –1 0. In this example the roots of the third order denominator polynomial.+ ---------2 s s+1 s+2 ( s + 3s + 2 )s Figure 17.= ( s + 1 ) -. are calculated. and then finding the limit as s approaches the root.+ ---------.= lim ( s + 1 ) A -( s + 2 )s s s → –1 C + lim B + lim ( s + 1 ) ----------s+2 s → –1 s → –1 s → –1 1 lim -----------------.+ ----------( s + 1 ) ( s + 2 )s s s+1 s+2 A BC 1 ( s + 1 ) ----------------------------------. The residues (numerators) of the partial fraction terms must still be calculated. 1 B C----------------------------------.= lim B = B s → – 1 ( s + 2 )s s → –1 1 0. The example shows a method for finding resides by multiplying the output function by a root term. 1 A BC1 x ( s ) = --------------------------------.numerical methods .+ ( s + 1 ) ----------( s + 1 ) ( s + 2 )s s s+1 s+2 1 C-----------------. The previous example in Figure 17.17.16 Partial fractions to reduce an output function After simplification with partial fraction expansion.5 x ( s ) = --------------------------------. A simple proof for finding ‘B’ above is given in this box.16 using a partial fraction expansion.= ----------------------------------.15 is continued in Figure 17. simpler components.13 smaller.+ ---------2 ( s + 1 ) ( s + 2 )s s s+1 s+2 ( s + 3s + 2 )s 1 A = lim s  -----------------------------------  ( s + 1 ) ( s + 2 )s s→0 1 = -2 = –1 1 = -2 1 B = lim ( s + 1 )  -----------------------------------  ( s + 1 ) ( s + 2 )s s → –1 1 C = lim ( s + 2 )  -----------------------------------  ( s + 1 ) ( s + 2 )s s → –2 Aside: the short cut above can reduce time for simple partial fraction expansions.= -. These provide three partial fraction terms.

+ L ---------.5 -----.17.5 ---------.+ L ---------s s+1 s+2 –t – 2t x ( t ) = [ 0.+ 0. In some cases the function is simple enough to immediately use a transfer function table.5 ] + [ ( – 1 )e ] + [ ( 0.5e –t – 2t ] Figure 17. .5 0.5 INVERSE TRANSFORMS AND PARTIAL FRACTIONS The flowchart in Figure 17.+ – 1 .14 converted back to a function of time as shown in Figure 17.17. partial fraction expansion is normally used to reduce the complexity of the function.numerical methods .--------------.5 – e + 0.5 )e x ( t ) = 0.17 Partial fractions to reduce an output function (continued) 17. Otherwise.18 shows the general procedure for converting a function from the s-domain to a function of time.5s s+1 s+2 –1 – 1 – 1 0. x(t ) = L [x(s) ] = L x(t ) = L –1 –1 –1 0.

NOTE: This does not apply for transfer functions.numerical methods . Calculators or computers are normally used when the order of the polynomial is greater than second order.17. yes is the function in the transform tables? no can the function be simplified? no simplify the function yes Use partial fractions to break the function into smaller parts Match the function(s) to the form in the table and convert to a time function Done Figure 17.18 The methodology for doing an inverse transform of an output function Figure 17. the overall order of the expression can be reduced by long division. In cases where the numerator is greater than the denominator.15 Start with a function of ’s’.19 shows the basic procedure for partial fraction expansion. . This results in a number of terms with unknown resides that can be found using a limit or algebra based technique. After this the denominator can be reduced from a polynomial to multiplied roots.

19 The methodology for solving partial fractions Figure 17. If there are higher order roots (repeated terms) then derivatives will be required to find solutions use algebra technique Done Figure 17.20 shows an example where the order of the numerator is greater than the denominator. This method is used infrequently because this type of output function normally occurs in systems with extremely fast response rates that are infeasible in practice.numerical methods . Long division of the numerator is used to reduce the order of the term until it is low enough to apply partial fraction techniques. .17.16 start with a function that has a polynomial numerator and denominator is the order of the numerator >= denominator? yes use long division to reduce the order of the numerator no Find roots of the denominator and break the equation into partial fraction form with unknown values OR use limits technique.

– 12s – 6 x ( s ) = 5s + 3 + --------------------2 s +4 solve – 12s – 6 A B --------------------. Here it produces partial two partial fraction terms divided by s and s-squared.21.+ -----------2 s + 2j s – 2j s +4 Figure 17. 5s + 3 s +4 2 3 2 5s + 3s + 8s + 6 3 5s + 20s 3s – 12s + 6 2 3s + 12 2 3 2 – 12s – 6 This can now be used to write a new function that has a reduced portion that can be solved with partial fractions. This pattern is used whenever there is a root to an exponent. .17.17 5s + 3s + 8s + 6 x ( s ) = ------------------------------------------2 s +4 This cannot be solved using partial fractions because the numerator is 3rd order and the denominator is only 2nd order.= ------------.numerical methods . Therefore long division can be used to reduce the order of the equation. The s-squared term requires special treatment.20 Partial fractions when the numerator is larger than the denominator Partial fraction expansion of a third order polynomial is shown in Figure 17.

numerical methods .+ --------------3 2 3 2 (s + 1) 2 s (s + 1 ) (s + 1) s s (s + 1) Figure 17.= --. The example in Figure 17. In this case each of the repeated roots is given with the highest order exponent. down to the lowest order exponent.x ( s ) = -------------------.+ ---------2 2 s s+1 s s ( s + 1) 1 C = lim ( s + 1 )  -------------------.+ -----------------. The equivalent forms are simplified algebraically.  2  s→0 s (s + 1) = 1 1 = lim ---------.2 B = lim ---.22 shows another example with a root to an exponent.= 1 s→0 s +1 d.  2  s → –1 s ( s + 1) 2 1 A = lim s  -------------------.23 extends the example begun in Figure 17.18 A B C1 . ---------.22 Partial fractions with repeated roots Algebra techniques are a reasonable alternative for finding partial fraction residues. until the point where an inverse matrix solution is used to find the residues.21 A partial fraction example Figure 17. so the resulting partial fraction expansion has five first order terms. ds  s 2 ( s + 1 ) s→0 Figure 17. ds  s + 1 s→0 = lim [ – ( s + 1 ) ] = – 1 s→0 –2 1 d.17.+ -.s  -------------------. The reader will note that the order of the denominator is fifth order.= ---. . 5 F ( s ) = ----------------------3 2 s (s + 1 ) A -C D E 5 ----------------------.1 = lim ---.+ B + -----------------.22.

+ B + -----------------. the example in Figure 17. .= ---.23 Solving partial fractions algebraically For contrast.24 using the limit techniques.= ---.+ -----------------. despite the need to solve simultaneous equations.+ -----------------.19 5 A -C D E ----------------------. In this case the use of repeated roots required the differentiation of the output function.+ – 15 + -----------------.+ --------------3 2 2 s ( s + 1 ) 3 ( s + 1 )2 ( s + 1 ) s s (s + 1) A ( s + 1 ) + Bs ( s + 1 ) + Cs + Ds ( s + 1 ) + Es ( s + 1 ) = ----------------------------------------------------------------------------------------------------------------------------------------3 2 s (s + 1) s ( B + E ) + s ( A + 3B + D + 2E ) + s ( 3A + 3B + C + D + E ) + s ( 3A + B ) + ( A ) = -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------3 2 s (s + 1) 0 1 3 3 1 1 3 3 1 0 0 0 1 0 0 0 1 1 0 0 1 2 1 0 0 A 0 B 0 C = 0 D 0 E 5 A 0 1 0 B 1 3 0 C = 3 3 1 D 3 1 0 E 1 0 0 0 1 1 0 0 1 2 1 0 0 –1 4 3 2 3 3 2 2 2 2 0 5 0 – 15 0 = 5 0 10 5 15 5 5 -------5 10 15 ----------------------.numerical methods .+ --------------3 2 3 2 (s + 1 ) 2 s s (s + 1) (s + 1) s (s + 1) Figure 17.23 is redone in Figure 17. In these cases the algebra techniques become more attractive.17.

d.= ---.2 5 1. ----------------------- ( s + 1 ) = lim ---. ------------------ 3 ds  2 ds  ( s + 1 )3 s→0 s→0 s (s + 1 ) 3 5 5 C = lim  ----------------------- ( s + 1 ) = lim ---.------------D = lim ---.= lim ---.2 5 1.24 Solving partial fractions with limits An inductive proof for the limit method of solving partial fractions is shown in Figure 17.= ---. 3 2 3 2 ( s + 1 )  2 s (s + 1 ) (s + 1) s → –1 s s (s + 1 ) 5 A( s + 1) .+ C + D ( s + 1 ) + E ( s + 1 ) 2 s s → –1 s 2 s 3 3 .---.+ -----------------.2 d5 B = lim ---.d.+ -----------------.– 2 ( 5 ) = 10  2 3 2 s → – 1 1! ds s → – 1 1! ds s s → – 1 1! s 3 s (s + 1 ) 3 1.+ --------------.----.= 15 3 2! ds  2 2! ds s 2 2! s 4 s → –1 s → –1 s → –1 s (s + 1) 5 5 – 15 5 10 15 ----------------------.17.3 2 2 s ( s + 1)3 ( s + 1)2 ( s + 1) s s (s + 1) 5 .+ -----------------.---.= lim ---.= 5  2 3 s → –1 s2 s → –1 s (s + 1) 3 1.30 E = lim ---.A B C D E lim ----------------------.= --.+ -----------------.+ -----------------.+ -------.numerical methods .---.+ -----------------.= 5  2 3 s → 0 ( s + 1 )3 s→0 s (s + 1 ) d5 .+ --------------3 2 2 s ( s + 1)3 ( s + 1)2 ( s + 1) s s (s + 1 ) 5 .20 5 A -C D E ----------------------.+ -----------------.= ---.+ -----------------.25.d.= --.= ---------------------. ----------------------- ( s + 1 ) = lim ---. 5 A -C D E ----------------------.---.+ --------------3 2 3 2 (s + 1) s ( s + 1) s → –1 2 s (s + 1) s ( s + 1) 5 .---.B ( s + 1 )2 lim ---.+ -.+ --------------.+ -.2 5 A = lim  ----------------------- s = lim -----------------.A B C D E 3 lim ( s + 1 )  ----------------------.+ B + -----------------.--.+ ---------------------.51.d5 1. ----------------------- s = lim ---.= – 15 s → 0 ( s + 1 )4 Figure 17.+ B + -----------------.+ --------------3 2 3 2 2 s (s + 1) s (s + 1 ) (s + 1) s (s + 1) 5 ( –3 )= lim -----------------.

differentiate once.+ --------------------. 3 3 d.+ B ( s + 1 ) .= 3 s → –1 s –2 ( 5 ) ------------. evaluate now.+ --------------------.+ --------------------------.B ( – 1 + 1 ) 2 2 –1 ( –1 ) ( –1 ) 5 -----------. then evaluate (the terms for A and B will be ignored to save space.17. then evaluate 3 3 2 d.+ C + D ( – 1 + 1 ) + E ( – 1 + 1 ) 2 --------------------------.26 with a sinusoidal input.2 5 ---------------------.numerical methods .+ B ( 0 ) .+ ---------------------. 5 A ( s + 1 ) .B ( s + 1 ) lim ---.--------------------- ----  ---.= A ( … ) + B ( … ) + D + 2E ( s + 1 )  dt  3  s → –1 s –3 ( –2 ( 5 ) ) lim ------------------------. + D + 2E ( s + 1 ) 3 2 2 s     s s s 10 For E.+ C + D ( s + 1 ) + E ( s + 1 )  2 s s → – 1 dt  s 2  s 3 3 3 3 C = 5 –2 ( 5 ) lim ------------. but these will drop out anyway).+ C + D ( s + 1 ) + E ( s + 1 ) 2 lim   2 s s → – 1  dt  s 2  s ) d.= D = 3 ( –1 )  2 ( s + 1 )3 3 ( s + 1 ) 2  ( s + 1 )3 3 ( s + 1 ) 2 A  – --------------------.6 EXAMPLES 17. differentiate twice.= A ( … ) + B ( … ) + 2E 4 s → –1 s –3 ( –2 ( 5 ) ) ------------------------.= A ( – 1 + 1 ) . ---. + B  – -----------------.= A ( s + 1 ) .= A ( 0 ) + B ( 0 ) + 2E 4 ( –1 ) E = 15 Figure 17.25 A proof of the need for differentiation for repeated roots 17.+ C + D ( 0 ) + E ( 0 ) 2 ------------. .------------2 2 –1 ( –1 ) ( –1 ) For D.= ---------------------.= A ( 0 ) . 5 -----------.6.21 For C.– 2 ( 5 lim  ----  ------------.1 Mass-Spring-Damper Vibration A mass-spring-damper system is shown in Figure 17.

F ( t ) = 5 cos ( 6t )N 5s ∴F ( s ) = --------------2 2 s +6 This can be combined with the transfer function to obtain the output function.5 – 1.39j s – 0.26 are calculated and converted to a function of time in Figure 17.+ --------------------------------s + 6j s – 6j s – 0.5s + 2 5s ∴x ( s ) = --------------------------------------------------------2 2 ( s + 36 ) ( s + 0. 1---M x (s ) ∴---------. x( s) = F (s )  2  2  2 F ( s ) s + 6  s + 0. .  ----------------------------.5s + 2 ) A B C D ∴x ( s ) = -----------.17.numerical methods .22 Given.+ ---------------------------------. 1 -  5s x (s ) 1  ---------- =  --------------.5 ----m The sinusoidal input is converted to the s-domain.5 + 1. In this case the roots of the denominator are complex.= -------------------------------Ks F (s ) 2 Kd s + ----.s + ----M M Component values are.26 A mass-spring-damper example The residues for the partial fraction in Figure 17.39j Figure 17.+ -----------. so the result has a sinusoidal component. M = 1kg N K s = 2 --m Ns K d = 0.27.

00516j ∴x ( s ) = ------------------------------------------.00516 cos  0.= ---------------------------------------------2 2 ( – 12j ) ( 36 – 3j + 2 ) ( s – 6j ) ( s + 36 ) ( s + 0.  + …   0.= ---------------------------------.× ---------------------.36 – 456j ∴A = ---------------------------------------.numerical methods .0654   Figure 17.6.00516 e 2 2 – 0t 0.2 Circuits It is not necessary to develop a transfer functions for a system.0654 + 0.28.23 A = lim s → – 6j – 30j ( s + 6j ) ( 5s ) --------------------------------------------------------------------------. An inverse Laplace transform is used to convert the equation into a function of time using the tables.17.0654 – 0.= 0.00516j 0.27 A mass-spring-damper example (continued) 17. Impedance values and the input voltage are converted to the s-domain and written in the equation.+ ------------------------------------------. 232 36 + 456j 36 – 456j --Continue on to find B. The equation for the voltage divider is shown in Figure 17. .0654 + 0. The resulting output function is manipulated into partial fraction form and the residues calculated.00516t + atan  – -----------------.5s + 2 ) – 30j 13680 + 1080j 30j .0654 + 0.+ … s + 6j s – 6j Do inverse Laplace transform x ( t ) = 2 0. D same way 0.= ---------------------.00516j – 432j – 36 – 24j 209. C.

 1.5s – 1.= ------------------------------------------------------V o = ------------. The numerator of resulting output function must be reduced by long division.5 1. equations are combined. 3s 3s 3s .  = 2 2 3 –3 2 1 R + ----.5 1. relate the source voltage to the output voltage using component values in the s-domain.---------------sC ------------------------------------------.5V o = ---------------------.5 ∴A = 1.5 ∴C = 1. 2 13s Z C = ----ZR = R Vs ( s ) = ------------sC 2 R C s +1 Next. 2 2 2  ----. 2 2 2 3s C----------------------------------.numerical methods .28 A circuit example .24 R = 10 Ω t=0 V s = 3 cos t + + Vo - 3 C = 10 F –3 As normal. ( s + 1 ) ( 1 + sRC ) ( s + 1 ) ( s10 10 + 1 ) s +1  sC The output function can be converted to a partial fraction form and the residues calculated.17.= ---------------------------------------------------------------------------2 2 (s + 1)(s + 1 ) ( s + 1)(s + 1) B+C = 0 A+B = 0 A+C = 3 ∴C = – B ∴A = – B ∴– B – B = 3 2 2 ∴B = – 1.= As + B + ---------.= As + As + Bs + B + Cs + C ----------------------------------------------------------------------------------Vo = 2 2 2 s+1 ( s + 1 )( s + 1) s +1 (s + 1)(s + 1 ) ZC V o = Vs  ------------------ Z + Z  3s s (A + C ) + s (A + B ) + ( B + C) Vo = ----------------------------------.+ ---------2 s+1 s +1 Figure 17.

– 1.5L ------------. This is easily converted to the s-domain using the e-to-the-s functions.17.5 + L –1 ------------------------------2 2 s+1 s+1 s +1 s +1 –1 –t s 1 ------------. + 1.5e   1.5---------------------. as shown in Figure 17.= L– 1 1.+ 1.5 1.5s – 1.numerical methods .5 cos  t + atan  --------. .5L –1 –1 1. Vo ( t ) = L [ V o ( s ) ] = L ∴V o ( t ) = 1.29 A circuit example (continued) 17.121 cos  t – -.5   π –t ∴V o ( t ) = 2.1 Input Functions In some cases a system input function is comprised of many different functions.30.5s – 1.5 sin t + 1.5 ---------1.5e ∴V o ( t ) = 2 2 –t – 1.7.5e  4 Figure 17.5 cos t – 1.5.+ 1. as shown below.  + 1.5 + 1.5e 2 2 s +1 s +1 –1 –t ∴V o ( t ) = 1.7 ADVANCED TOPICS 17. The step function can be used to switch function on and off to create a piecewise function.25 The output function can be converted to a function of time using the transform tables.

17.= --.= -2 2 s → 0 ( s 2 + 3s + 2 )s s → 0 s 2 + 3s + 2 (0 ) + 3(0) + 2 x ( t → 0 ) = lim [ sx ( s ) ] s→∞ Initial value theorem 1 1 1(s) ∴x ( t → 0 ) = lim --------------------------------.numerical methods .= 0 2 2 ∞ s → ∞ ( s + 3s + 2 )s ((∞ ) + 3(∞ ) + 2 ) Figure 17. x ( t → ∞ ) = lim [ sx ( s ) ] s→0 Final value theorem 1 1 1s 1 ∴x ( t → ∞ ) = lim -------------------------------.7.= lim ------------------------.5e f ( s ) = ---.30 Switching on and off function parts –s – 3s – 4s 17.= ------------------------------------------.31.= -----------------------------------.– --------.– -----------.26 f(t) 5 t seconds 0 1 3 4 f ( t ) = 5tu ( t ) – 5 ( t – 1 )u ( t – 1 ) – 5 ( t – 3 )u ( t – 3 ) + 5 ( t – 4 )u ( t – 4 ) 5e 5 5e .+ -----------2 2 2 2 s s s s Figure 17.31 Final and initial values theorems .2 Initial and Final Value Theorems The initial and final values an output function can be calculated using the theorems shown in Figure 17.

numerical methods . System Model (differential equations) Laplace Transform Transfer Function Some input disturbance in terms of time Input described with Laplace Equation We can figure out from plots Experiments Root-Locus for stability Substitute Output Function (Laplace form) Bode and Phase Plots .Straight line .32 A map of Laplace analysis techniques .27 17.Exact plot Phasor Transform Output Function (Laplace terms) Partial Fraction Inverse Laplace Equations for gain and phase at different frequencies Approximate equations for steady state vibrations Steady state frequency response Time based response equation Figure 17.17.8 A MAP OF TECHNIQUES FOR LAPLACE ANALYSIS • The following map is to be used to organize the various topics covered in the course.

2 –3t L ---. a) b) c) d) e) f) g) h) i) j) k) l) m) n) L[5] L[e – 3t o) ] ] ] p) q) r) s) t) u) –( t + 1 ) ·· · · L [ x + 5x + 3x ]. x ( 0 ) = 7 dL ---.numerical methods .7t ] L [ sin ( 5t ) ] L [ sinh ( 3t ) ] L [ t sin ( 2t ) ] d.28 17.10 PRACTICE PROBLEMS 1. Convert the following functions from time to laplace functions.2t + 3 ) ] . 17.t e dt 2 cos ( 9. x ( 0 ) = 8.sin ( 6t ) dt d 3  ---- t 2 L  - dt L L [ 5e – 3t L [ 5te L [ 5t ] – 3t ∫0 y dt 3 – 5t t L [ 3t ( t – 1 ) + e ] L [ 4t ] L [ cos ( 5t ) ] L[3(t – 1 ) + e L [ 5e L [ 5e – 3t – 3t 2 L[u(t – 1 ) – u(t – 2) ] L[e L[e – 2t u(t – 2 ) ] u(t – 1) ] ] v) w) x) y) z) aa) –( t – 3 ) – 3t cos ( 5t ) ] cos ( 5t + 1 ) ] L [ 5e + u(t – 1 ) – u( t – 2 )] t–3 L [ cos ( 7t + 2 ) + e L [ cos ( 5t + 1 ) ] L [ 6e – 2.17.9 SUMMARY • Transfer and input functions can be converted to the s-domain • Output functions can be calculated using input and transfer functions • Output functions can be converted back to the time domain using partial fractions.

5s ) -s 4 + 3j 4 – 3j ---------------------.+ ---------------------.29 2.+ ---------------------s + 1 – 2j s + 1 + 2j 66 --.+ ---------------------s s 2 + 4 s + 2 – 3j s + 2 + 3j . Convert the output function below Y(s) to the time domain Y(t). 3 3 5 12 Y ( s ) = -.numerical methods . a) b) c) d) e) f) L L L L L L –1 1---------s+1 5---------s+1 6 ---2 s 6 ---s 3 g) h) i) j) k) L L L L L –1 5 ( 1 – e – 4.+ ------------.+ ------------4 2 s s +9 6 ------------------------s + 5s + 6 6 ---------------------------------4s + 20s + 24 2 2 –1 –1 –1 –1 –1 –1 –1 –1 s+2 -------------------------------(s + 3)(s + 4 ) 6 ------------2 s +6 –1 3.17. Convert the following functions below from the laplace to time domains. a) b) c) d) e) f) L –1 s+2 -------------------------------(s + 3)(s + 4 ) g) h) i) j) k) l) L L L L L L –1 L [ ] L [ ] L [ ] L L –1 –1 –1 –1 –1 s + 9s + 6s + 3 ---------------------------------------3 2 s + 5s + 4s + 6 9s + 4-----------------(s + 3) 9s + 4 ----------------------3 3 3 3 2 –1 –1 s (s + 3) 2 s + 2s + 1 ------------------------s + 3s + 2 2 s + 3s + 5 ------------------------6s + 6 2 s + 2s + 3 ------------------------s + 2s + 1 2 2 2 –1 6 ---------------2 s + 5s 2 9s + 6s + 3 ---------------------------------------s + 5s + 4s + 6 3 2 –1 –1 4. Convert the following functions below from the laplace to time domains using partial fractions.

 m F(s) ( s + 3 ) ( s + 4 )  N F ( t ) = 5N x ( t ) = 5m 8. –4 t L [ 5e cos ( 3t + 2 ) ] = a) b) c) L[e – 2t + 5t ( u ( t – 2 ) – u ( t ) ) ] = where at t=0 y0 = 1 y 0'' = 3 y 0' = 2 y 0''' = 4 d. find the response in time. ·· · 5x + 6x + 2x = 5F a) · y + 8y = 3x b) · y – y + 5x = 0 c) 6.-----------. G(s). Do the following conversions as indicated. Convert the output function to functions of time. m F(s) ( s + 3 ) ( s + 4 )  N x( s) s + 2 .= -------------------------------.30 5.17. Solve the differential equation using Laplace transforms. determine the time response output Y(t) to a step input X(t).+ ---------------------------.+ --------------------. Transfer Function Input a) b) x( s) s + 2 .= 2 s + 2 s + 4s + 40 –1 9.= s + 3 + 4j s + 3 – 4j 13 s + ---------. 4Y(s) G ( s ) = ---------.numerical methods .-----------. 3 2 a) s + 4s + 4s + 4 ---------------------------------------3 s + 4s 2 b) s +4 -----------------------------------------------------------------s + 10s + 35s + 50s + 24 10. Given the transfer function.= ---------X ( t ) = 20 When t >= 0 s+2 X(s) 7. Assume the system starts undeflected and at rest. Given the following input functions and transfer functions. Convert the following differential equations to transfer functions.3 dL  ---- y + 2  ---- y + y =  dt  dt L L –1 d) e) 1+j 1–j ---------------------. ··· ·· · θ + 40θ + 20θ + 2θ = 4 4 3 2 .= -------------------------------.

37 + --------------------------------------------.– ---------s–3 s+3 –4 16s -------------------.– ----.– -------s s e ------------s+2 e ---------s+1 –s – 2s 5 ----------.+ ---------.5.416s – 6.+ e .31 17.4s + 91.e -----.– 3 + ---------2 s s+1 s 5(s + 3) -----------------------------2 2 (s + 3) + 5 2.= 12s – 16 ---------------------2 3 3 2 2 2 (s + 4) ( s + 4) (s + 4) 2s -----------------3 (s + 3) ( s x – 7s – 8 ) + 5 ( sx – 7 ) + 3x 6s ---------------2 s + 36 2 y s 72 18 -------------.5 ∠– 1 ---------------------.– e -----.5 ∠1 2.2j s + 5.93 2 z) .+ ------------------------------.+ -------------------.e 2 2 s–1 s–1 s + 49 s + 49 s cos 1 – 5 sin 1 ---------------------------------s + 25 3∠3 3 ∠– 3 s( ) + ------------------------------.7 – 9.+ 1 5 4 s+5 s s e .+ ---------------------s + 3 – 5j s + 3 + 5j s( ) + = ---------------------------2 s + 6s + 34 5 ---------------s + 25 x) y) 2 s) t) u) v) w) k) cos ( 2 ) s – sin ( 2 ) 7 e ------------------------------------------.7 + 9.11 PRACTICE PROBLEM SOLUTIONS 1.17.= ---------------------------------------2 s + 2.5---------.e ---.numerical methods .2j s + 2.0.= – 0. a) 5 -s 1---------s+3 5---------s+3 5 -----------------(s + 3) 5 ---2 2 l) m) n) o) p) q) r) –1 0.-------s s s+3 –3 –3 2–s 4 – 2s –s – 2s 2 2 2 b) c) d) e) f) g) h) i) j) s 8 ---3 s s ---------------2 s + 25 3 -.

y ( t ) = 5 + 6 sin ( 2t ) + 2 ( 3 )e 5.927 ) 6 cos ( 1.32 2.834 cos ( t + 0.5 ) 2 ( 5 )e 3 – ( 1 )t b) 5e c) 6t d) 3t 3 cos  2t + atan  – --    4  2 i) j) – 4t t + 2 sin ( 3t ) 6e – 2t e) – e f) 3.3t δ(t) k) --------. a) b) c) d) e) f) –e – 3t – 3t – 6e – 3t – 3t + 2e k) 0.2e 8.4t – 5t – 0.17.13t + 1.= ---------------------------2 F 5s + 6s + 2 y 3 .5e – 2t – 0. a) b) c) x 5 -.= ---------x s–1 – 2t cos ( 3t – 0 ) .99 )e 4.= ----------x s+8 y –5 .34e – 4.2 – 1. a) e –t –t g) h) 5 – 5u ( t – 4.+ 0.5e 6 sin ( 6t ) + 2e – 4t g) h) i) j) δ( t ) – e – 2t 1.numerical methods .23 ) l) δ ( t ) + 2te –t + 2 ( 0.

040 – 2.6 3 2 ( 5δ ( t ) + 30 – 5e – 2t – 2t )N L [ 5e – 4t cos ( 3t + 2 ) ] = L [ 2 A e – αt cos ( βt + θ ) ] α = 4 A = 2.e .17.5j sin 2 = – 1.+ 5e L [ t ] + 10e L [ 1 ] – ---2 s+2 s – 2s – 2s 1 .numerical methods .5 β = 3 θ = 2 A = 2.+ 5L [ tu ( t – 2 ) ] – ---.+ -.040 + 2.273j --------------------------------------.e – -. a) 5 5 –3t 5 –4t -.-----------------------------------------------------------.273j A ------------------------------------.= ---------.– ---2 2 s+2 s s s c) d ---- y = s 3 y + 1s 2 + 2s 1 + 3s 0  dt d ---- y = s 1 y + s 0 1  dt 3 2 d.+ -----------.5e 10e 5 = ---------.3 dL  ---- y + 2  ---- y + y = ( s y + 1s + 2s + 3 ) + ( sy + 1 ) + ( y )  dt  dt 3 = y ( s + s + 1 ) + ( s + 2s + 4 ) 3 2 .+ 5L [ ( t – 2 )u ( t – 2 ) ] + 10L [ u ( t – 2 ) ] – ---2 s+2 s 15 – 2s – 2s = ---------.33 6.040 + 2.+ 5L [ ( t – 2 )u ( t – 2 ) + 2u ( t – 2 ) ] – --2 2 s+2 s+2 s s 5 1= ---------. y ( t ) = 40 – 40e 7.= – 1.+ --------------.+ A s + α + βj s + 4 – 3j s + 4 + 3j s + α – βj complex conjugate b) L[e – 2t + 5t ( u ( t – 2 ) – u ( t ) ) ] = L [ e – 2t ] + L [ 5tu ( t – 2 ) ] – L [ 5tu ( t ) ] 15 1 5= ---------.5 cos 2 + 2.273j + – 1. a) b) 8.

= L [ s ] + L ---------.141 – αt 2 2 complex conjugate A = π θ = atan  – 1 = – - ----.5e sin ( 6t ) dt 9. 4 1 – 3t α = 3 β = 4 = 2Ae cos ( βt + θ ) = 2.+ A s + α + βj s + 3 + 4j s + 3 – 4j s + α – βj 1 + 1 = 1.5L -----------------------------2 2 dt dt ( s + 2 ) + 36 ( s + 2 ) + 36 –1 – 2t – 2t d= ---.= 1 + ----------------------------------------------------------3 2 3 s s +4 s + 4s s + 4s 1 3s = 1 + -.δ ( t ) + e + 0.numerical methods .= L –1 ----------------------.+ --------------------.+ L ---------------------------2 s + 2 s 2 + 4s + 40 s+2 s + 4s + 40 – 2t – 2t d3 d6 = ---.δ ( t ) + e + L -----------------------------.+ ---------------------------.= ---.δ ( t ) + e + 0.+ --------------.  4 e) L 13 1 3 s + ---------.34 d) L –1 1+j 1–j A ---------------------------------------------------------.= 1 + -.+ ------------2 s s +4 2 A = 1 C = 0 B = 3 = δ ( t ) + 1 + 3 cos ( 2t ) .17. a) s + 4s + 4s + 4 ---------------------------------------3 s + 4s 3 2 3 s + 4s 1 3 2 s + 4s + 4s + 4 3 –( s + 4s ) 4s + 4 2 2 4s + 4 A Bs + C s ( A + B ) + s ( C ) + ( 4A ) = 1 + ---------------.282e π cos  4t – -.

50t – 3.35 b) s +4 ABCD -----------------------------------------------------------------.F  -- a  a L[e – at t→∞ lim f ( t ) = lim sF ( s ) s→0 f(t)] = F(s + a) dL [ tf ( t ) ] = – ---.+ ---------. = ----D = lim  - –6 s → –4 ( s + 1 ) ( s + 2 ) ( s + 3 ) 5 e –t – 4e –2t + 13 e –3t – 10 e –4t ---------6 2 3 10.+ ----------4 3 2 s+1 s+2 s+3 s+4 s + 10s + 35s + 50s + 24 2 5 s +4 A = lim  ------------------------------------------------. = ---- ( s + 1 ) ( s + 3 ) ( s + 4 ) –2 s → –2 2 2 13 s +4 C = lim  ------------------------------------------------. = - ( s + 2 ) ( s + 3 ) ( s + 4 ) 6 s → –1 2 8s +4 B = lim  ------------------------------------------------. Prove the following relationships.= ---------. θ ( t ) = – 66 ⋅ 10 e – 6 – 39.+ ---------.00 17.L [ f ( at ) ] = -. a) b) c) tL f  --  a = aF ( as ) d) e) f) t→∞ lim f ( t ) = lim sF ( s ) s→0 1 s .F ( s ) dt .216e 0.3368t + 2.1383t + 1.numerical methods .12 ASSIGNMENT PROBLEMS 1. = ---- ( s + 1 ) ( s + 2 ) ( s + 4 ) 2 s → –3 2 20 s +4  ------------------------------------------------.17.216e – 0.

17. 1979. The applied force ‘F’ is the input to the system. and the output is the displacement ‘x’. including initial conditions. second edition. C. The following differential equation is supplied.D.M. Inc. Close. and Frederick. 3.13 REFERENCES Irwin..36 2. John Wiley and Sons. a) find the transfer function.R. 17. and Graf. c) Find the frequency response (gain and phase) for the transfer function using the phasor transform.. Solve to find the time response. Prentice Hall Publishers. K1 = 500 N/m K2 = 1000 N/m M = 10kg x F b) What is the steady state response for an applied force F(t) = 10cos(t + 1) N ? c) Give the transfer function if ‘x’ is the input. 1995. d) Convert the differential equation to the Laplace domain.. E.K. b) Write the equation in state variable form and solve it numerically. Roughly sketch the bode plots.. d) Find x(t). given F(t) = 10N for t >= 0 seconds using Laplace methods. D. ·· · · y + y + 7y = F y(0 ) = 1 y(0) = 0 F ( t ) = 10 t>0 a) Solve the differential equation using calculus techniques. . “Modeling and Analysis of Dynamic Systems.numerical methods . Industrial Noise and Vibration Control. J. with initial conditions.

.they become less sensitive to variation in component values .they typically become more stable . • In general we use negative feedback systems because. and how it is enhanced by the addition of a control system. CONTROL SYSTEM ANALYSIS Topics: Objectives: 18.1 18.linear feedback controller analysis .2 CONTROL SYSTEMS • Control systems use some output state of a system and a desired state to make control decisions.it makes systems more immune to noise • Consider the system below. .1 INTRODUCTION 18.18.

If verror suddenly becomes bigger/smaller.2 Rules for a feedback controller . then increase/decrease θgas. increase/decrease θgas 2. If verror is near zero. etc.linear feedback controller analysis .2 Control variable INPUT (e. a car) vdesired + _ verror Driver or cruise control θgas car vactual The control system is in the box and could be a driver or a cruise control (this type is known as a feedback control system) Figure 18. 5. If verror is very big/small increase/decrease θgas 3.g.1 An example of a feedback controller Human rules to control car (also like expert system/fuzzy logic): 1. Figure 18.18. and has been positive/negative for a while. θgas) OUTPUT (e. keep θgas the same 4.g. velocity) SYSTEM (e. If verror is not zero.g.

3 The PID control equation • The figure below shows a basic PID controller in block diagram form.3 • Some of the things we do naturally (like the rules above) can be done with mathematics 18.2. This function will try to compensate for error in a controlled system (the difference between desired and actual output values). de u = K c e + K i ∫ edt + K d  -----  dt  Figure 18. proportional V + e Kp ( e ) integral Ki ( ∫ e ) derivative dK d  ---.linear feedback controller analysis .4 A block diagram of a feedback controller .1 PID Control Systems • The basic equation for a PID controller is shown below.e  dt  + + + u PID Controller V +V amp -V motor Figure 18.18.

18.4 e. .linear feedback controller analysis .g. dv error θ gas = K c v error + K i ∫ v error dt + K d  ----------------  dt  Rules 2 & 3 (general difference) Rule 4 (Immediate error) Rule 1 (Long term error) Kc Ki Kd Relative weights of components This is a PID Controller Proportional Integral Derivative For a PI Controller θ gas = K c v error + K i ∫ v error dt For a P Controller θ gas = K c v error For a PD Controller dv error θ gas = K c v error + K d  ----------------  dt  • The PID controller is the most common controller on the market.

dv error θ gas = K c v error + K i ∫ v error dt + K d  --------------.18.= K c + K i dt + K d  ---- ∫  dt v error Then do a Laplace transform d---.→ sx dt ∫ dt 1 = -s ∫ xdt x = s K θ gas L -----------. We can rewrite the control equation as a ratio of output to input.= K c + ----i + K d s s v error The transfer function .2.  dt  θgas d-----------.linear feedback controller analysis .→ s dt dx ----.5 18.2 Analysis of PID Controlled Systems With Laplace Transforms 1.

θgas 10-----Ms vactual . θgas 10 F 1-----Ms vactual • The ‘system model’ is shown above. F = Aθ gas = 10θ gas F--------. (Laplace transform would be the same as initial value.= M ---v dt 1v L -. • The two blocks above can be replaced with a single one.linear feedback controller analysis . • If θgas is specified directly. but much simpler.= 10 θ gas Transfer function for engine and transmission. this is called ‘open loop control’.) 2 dv d x F = Ma = M ------. This is not desirable.18. We can also develop a transfer function for the car.= -----Ms F Transfer function for acceleration of car mass 3. We want to draw the system model for the car.= M ----2 dt dt F d-.6 2.

linear feedback controller analysis . If we have an objective speed. verror Controller θgas K θ gas L -----------.18.= K c + ----i + K d s s v error *The coefficients can be calculated using classical techniques. the difference is the ‘system error’ v error = v desired – v actual ‘set-point’ . 6. and we can control θgas (the control variable). knowing the error is verror.7 4. but they are more commonly approximated by trial and error. Finally.desired system operating point 5. and an actual speed. For all the components we can now draw a ‘block diagram’ vdesired + - verror K K c + ----i + K d s s θgas 10 F 1-----Ms vactual • A ‘negative feedback loop that is the fundamental part of this ‘closed loop control system’ . we can select a control system.

..8 18. ..3 Finding The System Response To An Input • Even though the transfer function uses the Laplace ‘s’.linear feedback controller analysis .. it is still a ratio of input to output.. • Find an input in terms of the Laplace ‘s’ vdesired 100 step vdesired(t) = 100 for t >= 0 sec ramp vdesired(t) = 50t for t >= 0 sec t(sec) 0 Input type STEP Time function f ( t ) = Au ( t ) f ( t ) = Atu ( t ) f ( t ) = A sin ( ωt )u ( t ) f ( t ) = A ( u ( t ) – u ( t – t1 ) ) Laplace function A f ( s ) = -s A f ( s ) = ---2 s Aω f ( s ) = ---------------2 2 s +ω f(s ) = 2 RAMP SINUSOID PULSE etc.18.2.

and transfer function to find the output of the system.9 Therefore to continue the car example.------- s 2 ( 10100 ) + s ( 10000 ) + 1000  s  . some numbers will be selected for the values.18. Kd = 10000 Kc = 10000 Ki = 1000 M = 1000  2  v actual =  s ( 10000 ) + s ( 10000 ) + 1000  100 ----------------------------------------------------------------------. lets assume the input below.linear feedback controller analysis . v desired v  desired v actual   2  s ( K d ) + s ( Kc ) + K i   100 =  ------------------------------------------------------------- -------M  2  ---- s    s  10 + K d + s ( K c ) + K i To go further. lets use the input. v desired ( t ) = 100 t ≥ 0 sec 100 v desired ( s ) = L [ v desired ( t ) ] = -------s Next. v actual v actual =  ---------------.

ax + bx + c = 0 2 – b ± b – 4ac x = -------------------------------------2a 2 – 1 ± 1 – 4 ( 1.= – 0.+ --------------------.1 = 10  ------------------------------------------------.  s + s + 0.795 2 2 .10 At this point we have the output function.1 ) 2 2 v actual Aside: We must find the roots of the equation.877 2 ( 1.114 s + 0. – 0.01 ) + s + 0.1 ) ∴x = -----------------------------------------------------------.18.  s ( s 2 ( 1.01  s ( s + 0. and then find inverse Laplace transforms for each.01 ) ( 0.1 v actual = --------.+ --------------------s s + 0.linear feedback controller analysis .113 )s + 0. To do this we break up the function into partial fractions. before we can continue with the partial fraction expansion.01 ) 2 10 s + s + 0.113. --------------------------------------------------- 1. recall the quadratic formula. but not in terms of time yet.877 A B C v actual = -.

 --------------------------------------------------------  ( s + 0.113 )   99.877 )  s→0 ∴A = 99. --------------------------------------------------------   1.877 ) ( – 0.– --------------------s s + 0.877 -  ) 2 2 2 C = s → – 0.01 2 B = s → – 0.877 + 0.877 ) + ( – 0.01  s ( s + 0.113 ) + 0.11 s + s + 0.113 ) ( s + 0.113 + 0. -----------------------------------------------------------------  = 0.113 ) ( – 0.18. -----------------------------------------------------------------  = – 1.01  ( – 0.113 ) ( 0.1 10 = --------.264 1.linear feedback controller analysis .1  --------.877 .877 )  1.9 2 2 0.01  s ( s + 0.1 ∴C =  --------.877 ) + 0.( – 0.113 ) + ( – 0.877 )  2 2 2 10 . --------------------------------------------------------  ( s + 0.1  --------.877 ) 1.113 ) ( s + 0.9 0.1 ∴B =  --------.264  1.113 ) ( s + 0.01  ( – 0.877 )   10 s + s + 0.( –0.+ --------------------.113 )  1.1 10 A = lim s  --------.113 2 lim 10 s + s + 0.01  s ( s + 0.113 s + 0. ------------------------------------  ( 0.16  1.16 v actual = --------.877 2 lim 10 .

+ --------------------.9 + 0.9 0.16e – 0.– --------------------s s + 0.113 s + 0.16 v actual = --------.113t – 1. f( t) A f(s) A -s A ---2 s A----------s+α Aω ----------------2 2 s +ω 2 At – αt Ae A sin ( ωt ) e – ξω n t sin ( ω n t 1 – ξ ) ωn 1 – ξ --------------------------------------2 2 s + 2ξω n s + ω n etc.12 Next we use a list of forward/inverse transforms to replace the terms in the partial fraction expansion. 2 for ( ξ < 1 ) To finish the problem.18.877t . 99.877 v actual = 99.264 1.264e – 0.linear feedback controller analysis . we simply convert each term of the partial fraction back to the time domain.

we need to determine if any of the adjustable gains will lead to an unstable system.  τs G c = K ( 1 + τs ) 1G c = K  1 + ---. consider that the roots of the denominator directly impact the partial fraction expansion and the following inverse Laplace transfer. • Root locus plots allow us to determine instabilities (poles on the right hand side of the plane).13 18.linear feedback controller analysis . overdamped systems (negative real roots) and oscillations (complex roots).+ τs   τs 1 + ατs G c = K  ------------------  1 + τs  1 + τs G c = K  ------------------  1 + ατs α>1 α>1 α>1 τ1 > τ2 Lead Lag Lead-Lag 1 + τ 1 s 1 + ατ2 s G c = K  --------------------  --------------------  1 + ατ s  1 + τ s  1 2 18. Type Transfer Function Proportional (P) Proportional-Integral (PI) Proportional-Derivative (PD) Proportional-Integral-Derivative (PID) Gc = K 1G c = K  1 + ---.3 ROOT-LOCUS PLOTS • Consider the basic transform tables. A superficial examination will show that the denominator (bottom terms) are the main factor in determining the final form of the solution. • When designing a controller with variable parameters (typically variable gain).2.4 Controller Transfer Functions • The table below is for typical control system types. To explore this further.18. .

but the results are very important when designing a control system.14 • Note: this procedure can take some time to do. .linear feedback controller analysis . • Consider the example below.18.

the transfer function for the whole system was found. • Consider the previous example. After this design is built we must anticipate that all values of K will be used. Note: the value of gain ’K’ is normally found from 0 to +infinity.= -------------------------. root jω K→∞ K = 0 σ Note: because all of the roots for all values of K are real negative this system will always be stable. the more stable the system becomes.  K -- s G(s) KG S ( s ) = -------------------------------. s+K = 0 K 0 1 2 3 etc. It is our responsibility to make sure that none of the possible K values will lead to instability.linear feedback controller analysis .18. we use the characteristic equation of the denominator to find the roots as the value of K varies.15 Note: This controller has adjustable gain. and it will always tend to have a damped response. These can then be plotted on a complex plane. we must develop a transfer function for the entire control system.. So in general we do not need to find the transfer function for the whole system.= ----------1 + G ( s )H ( s ) s+K 1 +  K ( 1 )  --- s Next. The large the value of K. . but then only the denominator was used to determine stability. + K 1 K G ( s ) = --s H(s) = 1 1 -s First.

1 + G ( s )H ( s ) = 0 It is typical. The system response is a function of the denominator. (especially in textbook problems) to be given only G(s) or G(s)H(s). K ( s + z 0 ) ( s + z 1 )… ( s + z m ) G ( s )H ( s ) = -----------------------------------------------------------------( s + p 0 ) ( s + p 1 )… ( s + p n ) • Consider the example. The transfer function values will often be supplied in a pole zero form.linear feedback controller analysis . and that when not given the feedback gain is 1.16 Consider the general form for a negative feedback system. .18. and it’s roots. G( s) G S ( s ) = -------------------------------1 + G ( s )H ( s ) Note: two assumptions that are not often clearly stated are that we are assuming that the control system is a negative feedback controller.

= – 1. This includes writing the poles and zeros of the .5 ± ------------------2 2 Next.3. ( 1 ) = 0  2  s + 3s + 2 s + 3s + 2 + K = 0 – 3 ± 9 – 4(2 + K) 1 – 4K roots = ----------------------------------------------. write the characteristic equation. find values for the roots and plot the values.linear feedback controller analysis . K 1 +  ------------------------. find the characteristic equation. 1..18. and an equation for the roots.17 Given the system elements (you should assume negative feedback).1 Approximate Plotting Techniques • The basic procedure for creating root locus plots is. K G ( s ) = ------------------------2 s + 3s + 2 H(s) = 1 First. K 0 1 2 3 root 0 -1 -2 -3 jω σ 2 ******CALCULATE AND PUT IN NUMBERS 18.

= 0 ( jω + p 1 ) ( jω + p 2 )… ( jω + p n ) • Consider the example in the previous section. Plot the asymptotic curves to pass through the imaginary axis at this point. ( s + z 1 ) ( s + z 2 )… ( s + z m ) 1 + G ( s )H ( s ) = 1 + K --------------------------------------------------------------.linear feedback controller analysis . ( jω + z 1 ) ( jω + z 2 )… ( jω + z m ) 1 + K -------------------------------------------------------------------------. Breakin points exist between zeros. and solve for the frequencies. = 0  ds   ds  6. the breakaway and breakin points are found next. draw the asymptotes on the graph. B = ( s + z 1 ) ( s + z 2 )… ( s + z m ) . plot the root loci that lie on the real axis.= 0 ( s + p 1 ) ( s + p 2 )… ( s + p n ) 2. n – m – 1 ] ( p1 + p2 + … + pn ) ( z1 + z2 + … + z m ) σ = ------------------------------------------------------------------------------------------n–m 5. Find the points where the loci lines intersect the imaginary axis. Breakaway points exist between two poles on the real axis. To do this substitute the phasor for the laplace variable. to calculate these the following polynomial must be solved. B – A  ---. Points will be on a root locus line if they have an odd number of poles and zeros to the right. Finally. Next. The resulting roots are the breakin/breakout points. A = ( s + p 1 ) ( s + p2 )… ( s + p n ) d-A d-B  ---. Draw these lines in. 3. ± 180° ( 2k + 1 ) β ( k ) = ----------------------------------n–m k ∈ [ 0. The difference (n-m) will indicate how many root loci lines end at infinity (used later).18 equation. determine the asymptotes for the loci that go to infinity using the formula below. count the number of poles and zeros. determine where the asymptotes intersect the real axis using the second formula.18. 4.

19 Given the system elements (you should assume negative feedback).= 270° 2 (0)(– 1 – 2) σ = ---------------------------.linear feedback controller analysis .= 0 2 asymptotes σ -2 -1 . 1 ] jω 180° ( 2 ( 0 ) + 1 ) β ( 0 ) = -----------------------------------.18. K G ( s ) = ------------------------2 s + 3s + 2 H(s) = 1 Step 1: (put equation in standard form) K 1 1 + G ( s )H ( s ) = 1 +  ------------------------- ( 1 ) = 1 + K ------------------------------- 2  (s + 1)(s + 2) s + 3s + 2 Step 2: (find loci ending at infinity) m = 0 n–m = 2 Step 3: (plot roots) jω n = 2 (from the poles and zeros of the previous step) (loci end at infinity) σ -2 -1 Step 4: (find asymptotes angles and real axis intersection) 180° ( 2k + 1 ) β ( k ) = ------------------------------2 k ∈ I [ 0.= 90° 2 180° ( 2 ( 1 ) + 1 ) β ( 1 ) = -----------------------------------.

= 0 2 s + 3s + 2 s + 3s + 2 + K = 0 ( jω ) + 3 ( jω ) + 2 + K = 0 – ω + 3jω + 2 + K = 0 ω + ω ( – 3j ) + ( – 2 – K ) = 0 3j ± ( – 3j ) – 4 ( – 2 – K ) 3j ± – 9 + 8 + 4K 3j ± 4K – 1 ω = -------------------------------------------------------------.= ------------------------------2 2 2 In this case the frequency has an imaginary value.A = 0 ds B = s + 3s + 2 d-B ---. = 0  ds   ds  1 ( 2s + 3 ) – ( s + 3s + 2 ) ( 0 ) = 0 2s + 3 = 0 s = –1.= --------------------------------------------.5 -1 2 jω dd-A A  ---.linear feedback controller analysis . This means that there will be no frequency that will intercept the imaginary axis.= 2s + 3 ds σ -2 -1. 2 2 2 2 2 • Plot the root locus diagram for the function below.18. but we it will be done for illustrative purposes.20 Step 5: (find the breakout points for the roots) A = 1 d---. so step 6 is not necessary.B – B  ---. we know the system will be stable. .5 2 Note: because the loci do not intersect the imaginary axis. Step 6: (find the imaginary intercepts) 1 + G ( s )H ( s ) = 0 1 1 + K ------------------------.

4 DESIGN OF CONTINUOUS CONTROLLERS 18.18.5 SUMMARY • .linear feedback controller analysis .21 K(s + 5) G ( s )H ( s ) = --------------------------------2 s ( s + 4s + 8 ) 18.

linear feedback controller analysis . Given the transfer function below. Be careful to specify angles of departure. Use either the approximate or exact techniques. find the output ‘y(t)’ as a function of time. . Draw the root locus diagram for the transfer function below.5 ) ( s + 2s + 2 ) G ( s ) = ----------------------------------------------------------3 s (s + 1)(s + 2) 10. Draw the root locus diagram for the transfer function below. and the input ‘x(s)’. 200K ----------------------------------2 s + 2s + 200K b) Draw the Root-Locus diagram for the system (as K varies).18.6 PRACTICE PROBLEMS 3.22 18. and gains and frequency at stability limits. The system has a proportional controller with a variable gain K. θd Vd + Ve Vs ω θa 2 2 2 K 100---------s+2 1 -s Va 2 a) Simplify the block diagram to a single transfer function. K( s + 1)(s + 2) G ( s ) = ------------------------------------3 s 12. 2K ( s + 0. y( s) 5 --------. The block diagram below is for a motor position control system. K(s + 4) G ( s ) = ---------------------2 s (s + 1) 11.= ----------x( s) s+2 x ( t) = 5 t ≥ 0 sec 8. Draw a detailed root locus diagram for the transfer function below. ranges for breakout/breakin points. ans.

The following questions ask you to find the system response to a . 13. The equation below describes a dynamic system. It has the initial values specified. State if the Root-Locus diagram shows that the system is stable for the chosen K.= ----------------------------2 θd s + s + 20K a) Draw the root locus diagram and state what values of K are acceptable.23 ans. (ans.= – 1 ± 1 – 200K 2 Im K r o o K=0. Given the system transfer function below. Draw a Bode Plot for either one of the two transfer functions below. c) Given a gain of K=10 find the steady state response to an input step of 1 rad. s + 2s + 200K = s + 2ζω n s + ω n 2 2 2 ∴ω n = 1 ∴K = 0.005 From the root locus graph this value is critically stable. ( s + 1 ) ( s + 1000 ) ----------------------------------------2 ( s + 100 ) OR 5 ---2 s 15. θo 20K ---.linear feedback controller analysis . – 2 ± 4 – 4 ( 200K ) roots = ----------------------------------------------.005 ts 0 0 -2 -1 Re 0 c) Select a K value that will result in an overall damping coefficient of 1. The input is ‘F’ and the output is ‘V’.707 or a natural frequency of 3 rad/sec.18. d) Given a gain of K=10 find the response of the system as 17. b) Select a gain value for K that has either a damping factor of 0.

find the response to a ramp input as a function of time.= -----------------------------------------------------------------------------------2 X s ( 4 + 3K d ) + s ( 36 + 3K p ) + ( 3K i ) 2 a) Verify the close loop controller function given. all of these equations can be combined into a single system transfer . The difference between these two is called the error ‘e’.24 unit step input using various techniques. The following system is a feedback controller for an elevator. 19. The PID controller will examine the value ‘e’ and then control the speed of the lift motor with a control voltage ‘c’. Verify that the controller will be stable. The elevator and controller are described with transfer functions. It uses a desired heigh ‘d’ provided by a user. The system incorporates a PID controller. V(0) = 1 V' ( 0 ) = 2 V'' + 10V' + 20V = – 20F a) Find the response using Laplace transforms. 18. c) Draw a Bode plot for the feedback system if Kd=Kp=Ki=1. b) Find the response using the homogenous and particular solutions.+ K d s s+9 s 4 s ( 3K d ) + s ( 3K p ) + ( 3K i ) Y -. Sketch the result accurately below. b) Draw a root locus plot for the controller if Kp=1 and Ki=1.5. The closed loop transfer function is given. as shown below. c) Put the equation is state variable form. and the actual height of the elevator ‘h’. e) For the parameters found in the last step find the initial and final values. A feedback control system is shown below. f) If the values of Kd=1 and Ki=Kd=0. d) Select controller values that will result in a natural frequency of 2 rad/sec and damping coefficient of 0. Identify any values of Kd that would leave the system unstable.linear feedback controller analysis . and solve it using your calculator.18. Y X + 3 Ki ---------K p + ---.

   2 10s + 10 h s  ------------------------------. c) verify the solution using the initial and final value theorems.= K p + ----i + K d s = 2s + 1 + s ------------------------e s s combine the transfer functions 2 h =  10 ( s + 1 ) ( d – h )  ---------------------- 2 s 10 ( s + 1 ) 10 ( s + 1 ) h  1 + ---------------------- =  ---------------------- ( d ) 2     2 s s 10 ( s + 1 )  ---------------------.linear feedback controller analysis . b) Write find the damping coefficient and natural frequency of the results in part a). .-----------------.25 equation as shown.-.= 10 ( s + 1 ) 2 d–h s 2 system transfer function a) Find the response of the final equation to a step input.= -----------2 c s +s elevator 2 K c .= ( s + 1 ) . and the input (desired height) changes to 20 as a step input.18. The system starts at rest on the ground floor.= 2   d 10 ( s + 1 ) s + 10s + 10  1 + ---------------------- 2   s 2s + 1 + s 10  c  h = h = ------------------------.-----------. e = d–h error PID controller h 10 -. = -------------------------------.= 10 ( s + 1 ) ----------------------.10 2  e  c  2 e s s(s + 1) s s s +s h eliminate ‘e’ -------------------------------.-----------------.

18.6j h ( t ) = 20 + L –1 – 2.873j s + 10s + 10 200s + 200 A = lim  ------------------------------.= 0.= ---------------. 20 = 0  2  s  2 s→∞ s→∞ s + 10s + 10 s ( 10 )20 10s + 10 .873t + 4.26 (ans.5 – 22.6j  s ( s + 5 + 3.– 2.5 + 22.+ ------------------------------- 2  s s s + 5 – 3.873 h ( t ) = 20 + 2 ( 22.= ------------------------------2 d s + 10s + 10 d ( t ) = 20u ( t ) 20 d ( s ) = ----s A B C 10s + 10 .6j-------------------------------.873j s + 5 + 3.602 ) 5 ζ = ----ωn 1 – ζ ωn = 2 25 1 – ----.6 = 22.873 = – 5t cos ( 3. a) h 10s + 10 -.+ -------------------------------s + 5 – 3.873j ) s → – 5 + 3.5 + 22.602 α = 5 β = 3.873j s + 5 + 3.20 h =  ------------------------------.6j.20 10s h ( 0 ) = lim s  ------------------------------- ----. = 20  2 s → 0 s + 10s + 10 B = 200s + 200  ---------------------------------------- = – 2.845 35 c) 10s + 10 .20 h ( ∞ ) = lim s  ------------------------------. ----. ----.= -.= 20  2  s 10 s→0 s + 10s + 10 .873j θ = 4.ω n 2 ωn ωn = 35 = 5.916  2 25  2 15 =  1 – ----.= lim  ------.73 )e b) –5 = –ωn ζ 3. ω n = ω n – 25 2  ω n 5ζ = --------.5 – 22.+ -------------------------------.873j lim A = 2.73 2 2 C = – 2.linear feedback controller analysis .5 + 22.

25 Vd + Potentiometer Va 1--------0.1 + 400K d ) + s ( 400K p ) + ( 400K i ) 2 c) The function below has a step input of magnitude 1. This is compared to a feedback voltage from a potentiometer.18.= -----------------------------------2 θd s + 2. θa s ( 0. A PID controller is used to generate an output voltage to drive a DC motor.= -----------------------------------2 θd s + 2.9 ) + ( 4 ) ---.5s + 100 d) The function below has a step input of magnitude 1.9 ) + ( 4 ) ---.1 ω 1 -s θa b) Given the transfer function below. a) The block diagram below is for an angular positioning system. Find the output as a function of time using numerical methods.125 and a natural frequency of 10rad/s.linear feedback controller analysis .7 PRACTICE PROBLEM SOLUTIONS 18. select values for Kp. PID Controller θd 1--------0. (Hint: eliminate Ki).5s + 100 .e. θa s ( 400K d ) + s ( 400K p ) + ( 400K i ) ---. Simplify the block diagram. which is converted to a desired voltage.25 e K K p + ----i + sKd s Vs Motor 100 --------------s + 0. The setpoint is a desired angle. Give the results in a table OR graph. using the homogeneous and particular solutions).0. Ki and Kd that will result in a second order response that has a damping coefficient of 0.27 18.= ---------------------------------------------------------------------------------------------------------3 2 θd s + s ( 0.. θa s ( 0.8 ASSIGNMENT PROBLEMS 1. Find the output as a function of time by integrating the differential equation (i.

b) calculate a set of values and present them in a table. c) use the values calculated in step b) to develop a frequency response plot on semi-log paper.= ---------θd s+2 e Gc Vs Motor 100 --------------s + 0.5s. θa s ( 0.25 Vd + Potentiometer Va 1--------0. Gc. that will reduce the system to a first order system with a time constant of 0.= -----------------------------------2 θd s + 2. Find the output as a function of time using Laplace transforms. θa ---.9 ) + ( 4 ) ---------------------------2 θd s +s+4 2. Controller θd 1--------0. a) apply a phasor/Fourier transform and express the gain and phase angle as a function of frequency. as shown below. d) draw a straight line approximation of the Bode plot on semi-log paper.= s ( 0.1 ω 1 -s θa desired transfer function .9 ) + ( 4 ) ---. Select a controller transfer function.28 e) The function below has a step input of magnitude 1.25 θa 1---.18.5s + 100 f) Given the transfer function below.linear feedback controller analysis .

CONVOLUTION Topics: Objectives: 19.1 INTRODUCTION • Can be used to find normal input reponses for linear systems • It is most useful for finding an output response for a system given an arbitrary input function • It is also the basis for other methods that come later for system analysis.2 UNIT IMPULSE FUNCTIONS • A unit impulse function .convolution .19. 19.1 19.

.convolution .u ( t ) = δ ( t ) dt • If we look at an input signal (force here) we can break it into very small segments in time.19. it goes to infinity for an instant. but it does have an area of 1.2 Unit impulse 1 Area = ε  -- = 1  ε P(t) 1 -ε ε t Dirac delta function δ ( t ) = lim P ( t ) ε→0 ∞ Area = ∫–∞ δ ( t ) dt = 1 This function is theoretically undefined. As the time becomes small we can approximate it with a set of impulses. u(t < 0 ) = 0 u(t ≥ 0 ) = 1 d---. t 0 • For a unit step function.

2 Response of the system to a single pulse • If we add all of the impulse responses together we will get a total system response. F(t) x(t) ∫ g ( t )F ( t) dt t t Figure 19. .3 F(t) ∆t → 0 An impulse t Figure 19.convolution . This operation is called convolution.19.1 An impulse as a brief duration (instant) pulse 19.3 IMPULSE RESPONSE • If we put an impulse into a system the output will be an impulse response.

4 F( t ) t x(t) sum of responses ∫ g( t – τ )F ( τ ) dτ 0 t impulse responses The convolution integral t t c(t) = ∫ g ( t – τ )r ( τ ) dτ 0 Figure 19.19.convolution . The preferred method is shown later. . Note: This method is only valid for trivial differential equations with only one homogeneous term.3 A set of pulses for a system gives summed responses to give the output • Consider the unit impulse of a system with the given differential equation.

11 ---- X + 0. The input can be replaced with a unit impulse function. · X + 0. assuming a zero initial condition.5 The following first order differential equation has an input ’F’. Xp = A · Xp = 0 X ( 0 ) = 2 = Ce X ( t ) = 2e – 0.5x = 2F · X + 0.5 ( 0 ) = 2  ----  dt 0  dt X0 = 2 The initial condition caused by the impulse function found.5X = 0 A + 0.5X = 2δ ( t ) 0 + 0.5X = 2δ ( t ) The homogeneous solution can be found.5 t guess. X ( t ) = Ce – 0.5 t • Note that the derivation of the unit impulse function assumed zero initial conditions.5 = 0 X h = Ce – 0.5 t 0 guess.5A = 2 ( 0 ) Xp = A = 0 The initial condition caused by the impulse function found. so the process of convolution must also assume systems start at rest and undeflected. · X + 0. · x + 0.convolution . 19.19. assuming a zero initial condition. Xh = e At · At X h = Ae The particular solution is found.4 CONVOLUTION • The following example shows the use of the convolution integral to find a num- .

5t  e 0.convolution . t t x ( t) = ∫ X ( t – τ )F ( τ ) dτ 0 – 0.5 ( 0 ) ∫ ( 2e – 0.5τ t --------0. .5t ∫e 0 0.– e .5τ dτ = 2e – 0.5τ sin ( τ ) dτ FINISH THE INTEGRAL 19.5τ τ dτ FINISH THE INTEGRAL The unit impulse response to a sinusoidal input can be calculated using the convolution integral.5 ( t – τ ) t sin ( τ ) dτ = 2e – 0.5t e 0.5t – 1 ) = 4 ( 1 – e – 0. This is particularly useful for systems with arbitrary inputs. t t x ( t) = x(t ) = ∫ X ( t – τ )F ( τ ) dτ 0 = ∫ 2e 0 – 0. t t x ( t) = x(t ) = ∫ X ( t – τ )F ( τ ) dτ 0 = ∫ 2e 0 – 0.5t = x ( t ) = 2e --------.5  0 0.5t )  0. The unit impulse response to a step input can be calculated using the convolution integral.5t ( e 0.19. = 4e – 0.5t ∫e 0 0.5 0 The unit impulse response to a unit ramp input can be calculated using the convolution integral.5t ∫e 0 0.6 ber of responses.------------.5 ( t – τ ) t ) ( 1 ) dτ = 2e – 0.5 ( t – τ ) t τ dτ = 2e – 0.5 NUMERICAL CONVOLUTION • The convolution integral can also be solved numerically.5 0.

7 x ( t ) = h ∑ X ( t – ( i + 0.5 )h )F ( ( i + 0.19.5 )h ) i=0 n–1 where.convolution . . with a 2 second time step. t n = -.= number of steps h h = step size (s) X ( t ) = the unit impulse response functio n F ( t ) = the inunin pf c uo tt • This can be applied to the previous example for a unit step input to find the system position at 10 seconds.

006738 ) = 3.22313 + 0.5 + 0 +e – 4.5 +e – 2. .5 +e – 3.6065 ) x ( 10 ) = 3.5 + i x ( 10 ) = 4 ( e x ( 10 ) = 4 ( e – 4.5 + 2 +e – 4.08208 + 0.973048 • A Scilab program to perform the previous calculation numerically is shown below.= 5 2 x ( 10 ) = 2 ∑ 2e i=0 5–1 – 0.8 x ( t ) = h ∑ 2e i=0 n–1 – 0.5 ( t – ( i + 0.5 ( 10 – 2 ( i + 0.5 +e – 1.5 ) x ( 10 ) = 4 ( 0.01111 + 0.81208 This value can be compared to the exact value calculated below.5 )h ) F ( ( i + 0.5 ( 10 ) ) = 4 ( 1 – 0. The accuracy of the numerical value would increase substantially if the step size were decreased.5 + 1 +e – 4.convolution .5 )h ) 10 n = ----.5 + 4 ) – 4.19.5 ) ) (1) x ( 10 ) = 4 ∑ e i=0 4 – 4. x ( 10 ) = 4 ( 1 – e – 0.03020 + 0.5 + 3 +e – 4.5 +e – 0.

for i=0:n-1 foo = foo + h * X(t .9 // A numerical convolution example function foo=X(t) // The impulse response function foo = 2 * exp(-0. t. convolution(t. endfunction function foo=F(t) // The input function foo = 1.convolution . But.5*t). foo = 0.19. 19. // the time point to calculate printf("The estimated value x(%fs) = %f \n". x_calc = 4*(1-exp(-0. printf("The actual value is x(%fs) = %f \n". the Laplace transform for the convolution integral turns it into a simple multiplication. . h)). h) // The integration function n = t / h.5)*h). t.5)*h) * F((i+0.6 LAPLACE IMPULSE FUNCTIONS • The convolution integral can be difficult to deal with because of the time shift. // the time step for the integration t = 10. // a step function endfunction // define the variables function foo = convolution(t.5*t)). x_calc).(i+0. end endfunction h = 1.

convolution .19.10 ASSIGNMENT PROBLEMS 1.7 SUMMARY • 19.10 t c(t ) = ∫ g ( t – τ )r ( τ ) dτ 0 C ( s ) = G ( s )R ( s ) Figure 19.9 PRACTICE PROBLEM SOLUTIONS 19.4 The convolution integral in the Laplace s-domain 19.8 PRACTICE PROBLEMS 19. .

The derivation follows.1 20.state space analysis . . STATE SPACE ANALYSIS Topics: Objectives: 20.20.1 INTRODUCTION .state equations can be converted to transfer functions.

state equation coefficient matrices can be transformed to another equivalent for.state space analysis . if the state vector is rearranged.= ( C ( sI – A ) –1 B + D ) U (the transfer function) –1 –1 –1 –1 –1 –1 –1 .20. .2 State equations as functions of time · x = Ax + Bu y = Cx + Du In the s-domain sX – X 0 = AX + BU X ( sI – A ) = BU + X0 X = ( sI – A ) BU + ( sI – A ) X0 Y = CX + DU Y = C ( ( sI – A ) BU + ( sI – A ) X 0 ) + DU Y = ( C ( sI – A ) B + D )U + C ( sI – A ) X 0 Assuming the system starts at rest. Y = ( C ( sI – A ) B + D )U Y --.

· z = Az + Bu y = Cz + Du A = T AT C = CT –1 C = CT D = D B = T B D = D –1 • .state space analysis .3 State equations as functions of time · x = Ax + Bu y = Cx + Du Map one state vector to another one x = Tz This can be used to calculate new coefficient matrices –1 · z = T ( Ax + Bu ) –1 · z = T ( ATz + Bu ) –1 –1 · z = T ATz + T Bu –1 –1 · z = T ATz + T Bu A = T AT –1 B = T B –1 · z = Az + Bu The output equation becomes y = Cx + Du y = CTz + Du y = CTz + Du y = Cz + Du The equivalent set of state equations is.20.

( sI – A ) – B C D ----------------------------------.4 The two following transfer functions are equivalents G = C ( sI – A ) B + D G = C ( sI – A ) B + D This can be shown by applying the transformation matrix G = CTT ( sI – A ) TT B + TT D G = ( CT ) ( T ( sI – A ) T ) ( T B ) + D G = ( CT ) ( ( sI – A ) ) ( T B ) + D G = C ( ( sI – A ) )B + D = G –1 –1 –1 –1 –1 –1 –1 –1 –1 –1 –1 –1 • The transfer function form can be put into a matrix form.= ------------sI – A zeros sI – A = characteristic equationhomogeneous = • The free (homogeneous) response of a system can be used to find the state transition matrix.state space analysis .20.= ( sI – A )D – ( –B )C = ( sI – A )D + BC ( sI – A ) --------------------------------------------sI – A sI – A ( sI – A ) – B C D poles –1 G = C ( sI – A ) B + D = ----------------------------------. The transfer function can be said to be equivalant to the determinants of the matrix form. In this case the denominator is the characteristic equation. .

----.+ A .3 3 = I + At +  ---- A t +  ---- A t + …  2!  3! • The forced response (particular) response of the system can be found using convolution.20. · x = Ax + Bu x(t ) = e x(0 ) + ∫ e 0 At t A( t – τ ) Bu ( τ ) dτ y = Cx + Du y ( t ) = Ce x ( 0 ) + ∫ Ce 0 At t A(t – τ) Bu ( τ ) dτ + Du ( t ) initial response impulse response • As an example the homogeneous/free response of the system is shown below. and then converted to time. .+ … X0 s s s2 s3 e At 2 3 x h ( t ) = e x0 At = transition matrix aside: This expansion is a McLaurin (Taylor) series.5 The homogeneous equation can be written in the s-domain.----.state space analysis .2 2 1. The homogeneous equation can be written in the s-domain. X h = sI – A X0 x h ( t ) = L [ sI – A X 0 ] xh ( t ) = L –1 –1 I -. and then converted to time.+ A + A . e At 1.

--0 1 -2 2 s s s The function of time can be found assuming an initial position of 10 and velocity of 5.= 2  0 1 0 0 0 s 0. the state matrix can be found. . –1 –1 e At –1 –1 = L [ ( sI – A ) ] = 1 t 0 1 At x h ( t ) = x = e 10 = 1 t 10 = 10 + 5t v 5 0 1 5 5 • The forced/particular solution is shown below.6 Given a second order differential equation. ·· F = Mx · x = v 0 d---.If a matrix is diagonizable. ·· F = Mx . the diagonal matrix can be found with the following technique.state space analysis . the state matrix can be found.---. Given a second order differential equation.20.1 2 2  1 0  0 s –1 0 1  = s –1 s s = s s2 ( sI – A ) =  s – = ------------.x = 0 1 x + 1 F A = 0 1 F · = ---dt v ---0 0 v 0 0 v M M This can be used to find the inverse matrix. s 1 s.ss –0 --.11 -----.

2 2 e = I + At + ---. e At –1 –1 A = TΛT Λ = … … λn λ1 … = Te T Λt – 1 e This can also be applied to the At 1. The Eigenvalues are then used to find the Eigenvectors. T AT = Λ This may also be written as.state space analysis . ( λI – A )v = 0 Av = λv The Eigenvectors are then combined into a single matrix. a purely diagonal matrix may be found. • example. this is the transition matrix.20. .A t + … 2! At –1 –1 2 2 1e = I + TΛT t + ---.7 Given a matrix A.( TΛT ) t + … 2! Λt e 1 … = … … e λn t λ t Eigenvalues can be found using the following relationship. If any of the Eigenvalues are repeated the Jordan normal form is required. λI – A = 0 Note: the Eigenvalues are the values found in the characteristic/homogeneous equation solution.

1 = 4. 1 – 3 4 3 –3 4 1 6 1 4.state space analysis . – 1 2 u 11 = λ u 11 = 2 u 11 1 – 3 4 u 21 u 21 u 21 – u 11 + 2u 21 = 2u 11 ∴2 u 21 = 3u11 v1 = u 11 u 21 = 2 3 λ1 = 2 λ2 = 1 Another eigenvector can be calculated using the Eigenvalue 1.8 Given the state matrix A = –1 2 –3 4 The eigenvalues can be calculated. – 1 2 u 12 = λ u 12 = 1 u 12 2 u 22 u 22 – 3 4 u 22 – u 12 + 2u 22 = u 12 ∴u 22 = u12 v2 = u 12 u 22 = 1 1 The Eigenvectors can be combined to a single matrix.20.1 1 3 1 – 1 2 2 . T = v1 v2 = 2 1 3 1 To check. –1 2 1 = 4 . λI – A = λ 1 0 – – 1 2 0 1 –3 4 = λ + 1 –2 3 λ–4 2 ∴ = ( λ + 1 ) ( λ – 4 ) – ( – 6 ) = λ – 3λ + 2 = ( λ – 2 ) ( λ – 1 ) Therefore the Eigenvalues are. .A u12 u22 = 2 2 . An eigenvector can be calculated using the Eigenvalue 2. A u 11 u 21 Av i = λ i v i . 1 6 1 6 1 verified • The Eigenvectors can be used to calculate the system response.

.20. λ1 = 2 n λ2 = 1 v1 = 2 3 2t v2 = 1 1 t xh( t ) = ∑ vi e i=1 λi t 2t 1t = 2 e + 1 e = 2e + e 2t t 3 1 3e + e • zeros of state space functions can be found using the state matrices.state space analysis .9 Given the following Eigen values and vectors the free response is.

x = 0 1 x + – 5 F dt v –3 –4 v 5 A = 0 1 – 3 –4 B = –5 5 x = 1 0 x + 0F v C = 1 0 D = 0 This can be put into the matrix form.20.10 The following relationship locates the zero frequencies in state equations.state space analysis . s0 I – A –B C D = 0 Consider the example beginning with state equations. · x = v – 5F · v = – 3x – 4v + 5F d---. then there are no zeros. The order of the polynomial determines the number of zeros.  1 0  – 0 1  – –5  s0  0 1 –3 –4  5 1 0 s0 –1 5 0 = 0 3 s0 + 4 –5 = 0 1 0 0 s 0 ( 0 ) – ( –1 ) ( – 4.9 rad/sec. .9 There is one zero at -4. If the polynomial has no ’s’ terms.5 ) + 5 ( –( s 0 + 4 ) ) = 0 – 24.5 = 5s 0 s 0 = – 4.

11 20. the second state equation would be said to be not observable.If an input is to be observable it must be detectable in the output.x = 1 0 x + 3 u dt v 0 –a v b y = 5 0 x + 6 u v   G = C ( sI – A ) B + D = 5 0  s 0 – 1 0   0 s 0 –a  –1 –1 3 + 6 b   G = 5 0  s–1 0   0 s+a 1---------s–1 0 3 1---------. .e. This will happen when an output parameter is zero in the C matrix and an equation is decoupled in the A matrix. For example.state space analysis . d---. the same variable column is zero. i.+ 6 s–1 b---------s+a Note: the variables ’a’ and ’b’ both dissapeared because this system is not observable. d---. .2 OBSERVABILITY . d or f were not in the final expression for y.a system is observable iff the system state x(t) can be found by observing the input u and output y over a period of time from x(t) to x(t+h). For example consider the following state equations.20.b s+a 0 –1 G = 5 0  s+a 0   0 s–1  3 + = 5 0  -------------------------------- 3 + 6 6  ( s – 1 ) ( s + a ) b b     3---------15 + 6 = 5 0 s – 1 + 6 = ---------.x = a b x + e u dt v c d v f y = g h x + j u v k if the parameters c.

12 • This often happens when a system has elements that are decoupled. . • Another theorem for testing observability is given below. 0  = { 3e . • Observability can be verified formally for an LTI system with the following relationship. 1 . so both states are observable. Ce x = 0 At * t≥0 1 0 x = 0 . A = 1 0 0 2 C = 3 0 At * t 3t 2t 4t t     At * t Ce x = 3 0 e 0  1 .   0 1   0 1    At * Ce x =  ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t ) 1 . 0 } 2t  0 1  0 e  0 1  The second state is not observable.… … … 0 0 * 0 0 … 1 One example. or when a pole and zero cancel each other. ( 5e + 6e ) } Both values are non-zero. 0  = 3e t 0  1 .state space analysis . A = 1 2 3 4 At * C = 5 6 e e e 3t t 2t 4t Ce x = 5 6 e  1 0   1 0   . If any of the states satisfies the equation it is unobservable. ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t ) 0  0 1   Ce x = { ( 5e + 6e ).  = ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t )  . One example. The system is unobservable if the equation below is true for any state.20.

. it is possible to make it observable by changing the model. • If all unstable modes are observable. 20. . the system is detectable.stabilizable if it is controllable or if the uncontrollable nodes are stable. . rank ( M o ) = rank C CA … CA n–1 = n • If a system in unobservable. • A pole-zero cancellation is often the cause of the loss of observability. For example consider the following state equations.a system is controllable iff there is an input u(t) that will cause the system to go from any initial state to any final state in a finite time.3 CONTROLLABILITY .13 C CA … CA n–1 x = 0 * • Yet another test for observability is.state space analysis .20.If an input is to be observable it must be detectable in the output.

i. d---.0 s+b Note: the variables ’a’ and ’b’ both dissapeared because this system is not controllable.e.4 .---------5 = ---------. For example.x = a b x + e u dt v c d v f y = g h x + j u v k if the parameters c. d or f were not in the final expression for y.14 d---. the second state equation would be said to be not uncontrollable.20.x = 1 0 x + 2 u dt v a –b v 0 y = 3 4 x + 5 u v   G = C ( sI – A ) B + D = 3 4  s 0 – 1 0   0 s a –b  –1 –1 –1 2 + 5 0 G = 3 4 s–1 0 0 s+b 1---------s–1 0 0 s+B 0 0 s–1 2 2 + + 5 5 = 3 4 -------------------------------(s – 1)(s + b ) 0 0 G = 3 4 62 + 3.state space analysis .2 + 5 = ---------.+ 5 s–1 s–1 s+b 0 1---------. • This can be verified with the . This will happen when an output parameter is zero in the C matrix and an equation is decoupled in the A matrix. the same variable column is zero.

• Another test for controllability is. so both states are controllable. (x ) e B = 0 An example is. A = 1 0 0 2  1 0   . so it in not controllable.15 The basic relationship is. 5e + 6e }  1 0. . 0 1  3   0 e 2t 3   0 2t = { 0.state space analysis . 0 1    3e 2t The first state has a zero. 3e }  1 0. A = 1 2 3 4  1 0   .20.  e  0 1  T 1 0 t 0 2 B = 0 3  e 1t 0 0 0 =   1 0. 0 1  6   e 3t e 4t 6   5e 1t + 6e 2t 1t 2t 3t 4t = { 5e + 6e . 0 1    5e 3t + 6e 4t The results are both non-zero. ( x ) B AB … A n – 1 B * T • Yet another test for controllability is. An example is.  e  0 1  T 1 2 t 3 4 * T At t≥0 B = 5 6  e 1t e 2t 5 5 =   1 0.

.state space analysis .4 SUMMARY • 20.20. 20. it is possible to make it observable by changing the model. the systems is said to be stabilizable. • A pole-zero cancellation is often the cause of the loss of observability. • If a system in uncontrollable.7 ASSIGNMENT PROBLEMS 1.5 PRACTICE PROBLEMS 20. all of the states must be controllable. • if all unstable modes are controllable.16 rank ( M c ) = rank B AB … An – 1 B = n • For a system to be controllable.6 PRACTICE PROBLEM SOLUTIONS 20.

J.17 20.20. "16. MIT Opencourseware website.state space analysis .31 Feedback Control Course Notes".8 BIBLIOGRAPHY How. .

1 INTRODUCTION • 21.state space control .21. STATE SPACE CONTROLLERS Topics: Objectives: 21.2 FULL STATE FEEDBACK • The basic definition of a feedback controller is below.1 21. .

21.state space control . with D=0.2 The system state equations. . · x = Ax + Bu · y = Cx + 0u The Eigenvalues of A give the system poles u = r – Kx r = setpoint u = control output/plant input K = proportional gain values · x = Ax + B ( r – Kx ) · x = ( A – BK )x + Br · x = A c x + Br · y = Cx • An example of a controller design follows.

Check for the stability of the system without a controller. 1 – 2j s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = ( s + 1 + 2j ) ( s + 1 – 2j ) s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = s + s ( 1 + 2j + 1 – 2j ) + ( ( 1 + 2j ) ( 1 – 2j ) ) s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = s + s ( 2 ) + ( 5 ) 4 + 4k 2 = 2 2 + 4k 1 = 5 k 2 = – 0. M c = B AB = 0 4 0 1 0 – 2 –3 4 = 0 4 4 – 12 The rank is 2.....3 The controller design involves calculating K values to meet control objectives. u = –Kx = – k 1 k 2 x Ac = 0 1 0 1 – 0 0 1 – 0 0 = k1 k2 = 4k 1 4k 2 ( – 2 – 4k 1 ) ( – 3 – 4k 2 ) – 2 –3 4 –2 –3 0 1 s 0 – ( – 2 – 4k 1 ) ( – 3 – 4k 2 ) 0 s = 2 sI – A c = s –1 s + 2 + 4k 1 s + 3 + 4k 2 sI – Ac = s ( s + 3 + 4k 2 ) + ( s + 2 + 4k 1 ) = s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) The system is clearly second order. and the order of the matrix A is 2. p = 1 + 2j. why don’t we select a system that has poles at.state space control . so the second order methods may be used to select parameters.5 k 1 = 0. · x = 0 1 x+ 0 u –2 –3 4 · y = 1 x 0 Check for controlability.. the uncompensated system is. Select an input vector with variables. so the system is controllable.75 2 2 2 2 2 . sI – A = s 0 – 0 1 0 s –2 –3 2 = s –1 –2 s – 3 sI – A = s ( s – 3 ) – 2 = s – 3s – 2 = ( s – ) ( s – ) Therefore..21. However for variety.

Φ d ( s ) = the desired response –1 • Consider the previous example.4 r + - u B + + · x 1 -s A K x C y • The value of K in the previous example could also be picked using Ackermann’s formula. K = 0 0 … 1 Mc Φ d ( A ) where.21. .state space control .

25 3 – 1 = – 0.4 PRACTICE PROBLEMS . TRACK DOWN ALGEBRA PROB 21.25 – 0.75 0.5 The state equations and the desired response of the system are.3 SUMMARY • 21. Mc = 0 4 0 1 0 –2 –3 4 –1 = 0 4 4 –12 K = 0 1 0 4 4 – 12  0 1 2  + 0 1 ( 2 ) + ( 5 )I   –2 – 3  –2 –3 – 12 – 4  –4 0  K = 0 1 --------------------. – 2 – 3 + 0 2 + 5 0  0 – 16  6 7 –4 –6 0 5 K = 0 1 – 0.state space control .75 – 0. · x = 0 1 x+ 0 u –2 –3 4 2 · y = 1 x 0 Φd ( s ) = s + s ( 2 ) + ( 5 ) The Markov (?) parameter can be calculated first.21.25 0 3 – 1 = – 0.25 0 2 6 2 6 NOTE: DOESN’T MATCH.

.5 PRACTICE PROBLEM SOLUTIONS 21.state space control .6 ASSIGNMENT PROBLEMS 1.6 21.21.

22.2.2 MATHEMATICAL PROPERTIES • Magnetic fields have direction.1 Induction • Magnetic fields pass through space. .1 22.electromagnetics . ELECTROMECHANICAL SYSTEMS Topics: Objectives: 22. 22. The fields can allow energy storage. or transmit forces. • 22. and vector calculations. As a result we must pay special attention to directions.1 INTRODUCTION • Magnetic fields and forces are extremely useful.

B = µH = µr µ 0 H where. I = ∫0 2π Hr dθ ∴I = 2πHr I∴H = -------2πr • Flux density can be calculated for low H values.2 • resistivity of materials decreases with temperature • Amperes Circuit Law I = ∫ °Hdl where. . I = current flowing along a line (A) H = magnetic field intensity (A/m) l = A perpendicular path around the current flow (m) For example. at a fixed radius (r) around a wire.22. As the value climbs the relationship becomes non-linear.electromagnetics . B = Flux density (Wb/m*2 orT) µ = permeability of material – 7 Henry µ 0 = permeability of free spac = e4π ×10 --------------m µ r = relative permeability • Permeability.

3 B µ = --H –7 H µ 0 = 4π10 --m µ = µ r µ0 where.1 Saturation for a mild steel (approximately) . but with larger magnetic fields the materials saturate and the value of B reaches a maximum value. B 1.5 linear saturation H 4000 Figure 22.electromagnetics .22. A H = Magnetic field intensity  --.  m µ 0 = permeability of free spac e µ r = relative permeability of a material µ = permeability of a material • Permeability is approximately linear for smaller electric fields.

22.electromagnetics .5 silicon sheet steel cast steel 1.4 1.5 cast iron 0 0 500 H (At/m) Figure 22. 1989) 1000 • Flux density about a wire .2 Magnetization curves (Sen.0 B (T) 0.

LR = -----µA where.electromagnetics .or Tesla y  2  m I = Current in the conductor(A) r = radial distance from the conductor • Flux and flux density. Φ = ∫ B dA Φ = Flux density ( Wb ) A = Cross section area perpendicular to flux where.5 IB = -------2πr where. For an infinitely long straight conductor Wb B = Flux densi t ------.22. • When a material is used out of the saturation region the permeabilities may be written as reluctances. R = reluctance of a mania ge h ct t p L = length of a magnetic path A = cross section area of a magnetic path .

= -R l -----. .electromagnetics . I lg lc Rc = ----------µc Ac lg R g = ----------µg Ag Ni Φ = ----------------Rc + Rg lc • Faraday’s law.  µA R • Example.6 • Electric circuit analogy Φ F Ni F Φ = -----------.22.

B F = ------A 2µ 2 • The basic property of induction is that it will (in the presence of a magnetic field) convert a changing current flowing in a conductor to a force or convert a force to a current flow from a change in the current or the path. B B -Al W = ------V = ----2µ 2µ 2 2 • Force can be derived from the energy. For a coil e = the potential voltage across the coil N = the number of turns in the coil • Field energy.electromagnetics .Φ dt where. .22.7 de = N ---.

L = conductor length F = force (N) The FBD/schematic equivalent is. I.22. This is also called emf (Electro-Motive Force) .8 F = ( I × B )L F = ( L × B )I B NOTE: As with all cross products we can use the right hand rule here. L F where.3 The current and force relationship • We will also experience an induced current caused by a conductor moving in a magnetic field. I M F Figure 22.electromagnetics .

By applying voltage the wires push back against the magnetic field.4 Electromagnetically induced voltage • Hysteresis 22. .webers) v = velocity of conductor The FBD/schematic equivalent v e + - B v e M Figure 22.3 EXAMPLE SYSTEMS • These systems are very common.electromagnetics . The simplest motor has a square conductor loop rotating in a magnetic field.22. e m = electromotive force (V) φ = magnetic flux (Wb . take for example a DC motor.9 e m = ( v × B )L where.

22.10 y x z 2 4 b 3 a axis of rotation ω Figure 22.electromagnetics .5 B 1 5 I I A motor winding in a magnetic field .

electromagnetics . the two wires together will act like wire 3.6 Calculation of the motor torque .11 For wire 3. b – -2 b 2 = Bω sin ( ωt )  b –  – -- = 0 - 2  2 By symmetry.to -2 2 de m3 = ( V × B )dL b -2 V3 = – rω sin ( ωt ) rω cos ( ωt ) 0 0 B = –B 0 e m3 = ∫– b-2 – rω sin ( ωt ) 0 × –B dr rω cos ( ωt ) 0 0 0 dr 0 Brω sin ( ωt ) b -2 e m3 = ∫– b-2 2 b -2 e m3 = B r sin ( ωt ) ----ω 2 For wires 1 and 5. r cos ( ωt ) P3 = r sin ( ωt ) b b .– -. Therefore they both have an emf (voltage) of0V. e m1 = e m5 = 0V Figure 22.22.

sin ( ωt ) -.ω cos ( ωt ) 2 0 0 0 0 b B -. • In this arrangement we have to change the polarity on the coil every 180 deg of rotation.ω cos ( ωt ) 2 0 B = –B 0 e m2 = ∫0 a e m2 = ∫0 a b dl = aB -. but not rotate fully. b cos ( ωt ) -– b ω sin ( ωt ) -2 2 P2 = b V2 = b -.ω cos ( ωt ) 2 2 0toa 0 de m2 = ( V × B )dL – b ω sin ( ωt ) -2 0 × – B dl b -. The result would be that the motor would swing back and forth.22.ω cos ( ωt ) + 0 + aB -. If we didn’t do this the loop the torque on the loop would reverse for half the motion. To make the torque push consistently in the same direction we need to reverse the .12 For wire 2 (and 4 by symmetry). or a given voltage will cause the loop to rotate (motor). as the loop is rotated a voltage will be generated (a generator).7 Calculation of the motor torque (continued) • As can be seen in the previous equation.electromagnetics .ω cos ( ωt ) + 0 2 2 e m = aBbω cos ( ωt ) Figure 22. e m = e m1 + e m2 + e m3 + e m4 + e m5 b b e m = 0 + aB -.cos ( ωt ) 2 e m4 = e m2 For the total loop.

e m = aBbω cos ( ωt ) • Real motors also have more than a single winding (loop of wire). The equivalent circuit for a motor shows the related components.22. and hence the force. To add this into the equation we only need to multiply by the number of loops in the winding. etc. The device that does this is called a commutator. . e m = NaBbω cos ( ωt ) • As with most devices the motor is coupled. • Consider a motor that is braked with a constant friction load of Tf.13 applied voltage for half the cycle. But a change in voltage will also change the current in the windings.electromagnetics . This means that one change. Fw = ( I × B )L = IBa b T w = 2  F × -- = Fb = IBab  2 ∑M = T w – T f = Jα IBab – T f = Jα Figure 22.8 Calculation of the motor torque (continued) • We still need to relate the voltage and current on the motor. say in torque/force will change the velocity and hence the voltage. It is basically a split ring with brushes.

.electromagnetics . I A.9 Calculation of the motor torque (continued) • Practice problem. e A = voltage and current applied to the armature (motor supply) RA. L A = equivalent resistance and inductance of windings ∑V d = e A – IA RA – L A ---. de A – I A RA – L A ---.I A – NaBbω cos ( ωt ) = 0 dt and recall the previous equation.IA – e m = 0 dt we can now add in the other equations.22.14 IA RA + LA em eA + - where. IBab – T f = Jα Figure 22.

electromagnetics .22.10 Drill problem: Electromotive force • Consider a motor with a separately excited magnetic field (instead of a permanent magnet there is a coil that needs a voltage to create a magnetic field). The model is similar to the previous motor models. but the second coil makes the model highly nonlinear.15 Write the transfer function relating the displacement ’x’ to the current ’I’ a N M Ks S Kd R x I Figure 22. .

7 ASSIGNMENT PROBLEMS .22.electromagnetics .5 PRACTICE PROBLEMS 1.6 PRACTICE PROBLEM SOLUTIONS 22.4 SUMMARY • 22.16 22. 22.

we can induce flow through a pipe/value/orifice.fluids .23.1 23. 23. w = flow rate p = pressure . The pressure may be expressed as an absolute value. More correctly we should consider atmospheric pressure (gauge pressure). • When we deal with fluids we approximate them as incompressible. by applying a pressure at one point.1 SUMMARY • Fluids are a popular method for transmitting power (hydraulics). FLUID SYSTEMS Topics: Objectives: 23. • Fluids observe some basic laws. Basically.2 MATHEMATICAL PROPERTIES • Fluids do work when we have a differential pressure on a surface.

we say it is without resistance. w ∆p ≈ R( w – w ) ∆ ( ∆p ) R ≈ --------------∆w wR = 2 ----2 K w ∆p ∆ ( ∆p ) ∆w Figure 23. We have two choices if we want to analyze this system. 1. We can approximate the equation with a linear equation.fluids .2. Valves usually restrict flow by reducing an area for fluid to flow through.g. Of similar interest is the resistance of a value. The valve below is called a two way . w = K ∆p where. • A simple form of valve is a sliding plunger. As the flow rate changes significantly the accuracy of the equation will decrease.2 23. integration) 2. This is only good for operation near the chosen valve position.1 Fluid flow resistance • Resistance may also result from valves. Even a simple pipe has resistance. In reality. We can do a non-linear analysis (e.23.1 Resistance • If fluid flows freely. w = flow rate through the pipe ∆p = pressure difference across valve/pipe K = a constant specific to the pipe/valve/orific e NOTE: In this case the relationship between pressure drop and flow are non-linear. every fluid flow experiences some resistance.

The fluid out is typically not under pressure. and just returns fluid to a reservoir tank. or some other force to return it and force the fluid out.3 valve because it will allow fluid to flow in or out. Fluid in to cylinder Fluid out NOTE: For this type of valve the fluid in is typically under pressure. When the valve slide is up the pressure is applied to the cylinder. Figure 23.2 Fluid servo valve • Four way valve allow fluid force to be applied in both directions of a cylinder motion.fluids . and causes it to actuate.23. When used this way a cylinder needs a spring. .

2.3 Fluid flow control valves 23. such as the shape of the tank. or whether or not the top of the tank is open.4 to reservoir fluid in fluid in to reservoir advance no motion retract Figure 23.fluids . In a tank we have little pressure near the top.2 Capacitance • Fluids are often stored in reservoirs or tanks.23. • To calculate the pressure we need to integrate over the height of the fluid. Other factors also affect the pressure. but at the bottom the mass of the fluid above creates a hydrostatic pressure. .

a cylinder).. and letting fluid out recovers the energy.g.p dt . When designing these tanks we should be careful to keep the cross section constant (e.4 Pressures on fluid elements • Consider a tank as a capacitor. Water is pumped into these tanks.5 A Fp h F p + Fg Fg Fp + F g F p + Fg Fg P = ----A F g = ρV = ρgAh ρgAh P = ------------. and the hydrostatic pressure increases. Laplace). Hence we can pump fluid into a tank to store energy. As consumers draw water through the system these tanks provide pressure to the system.= ρgh A Figure 23. As fluid is added the height of the fluid rises.fluids . AC = ----ρg dw = C ---. If the cross section varies then the fluid pressure will not drop at a linear rate and you won’t be able to use linear analysis techniques (eg. • The mathematical equations for a constant cross section tank are.23. A very common application of this principle is a municipal water tower.

and forcing the inlet vale closed. In hydraulics these are pumps that will provide pressure and/or flow to the system. and piston strokes Figure 23.3 Power Sources • As with most systems we need power sources.5 A piston driven hydraulic pump • A geared hydraulic pump is pictured below. and pulls fluid into the chamber. The fluid volume can be controlled by using the cylinder size. Fluid In Normally closed inlet valve Fluid Out piston Normally Closed Valve In this common form of piston pump.fluids . Other types use vanes and pistons. creating a pressure.2. and the outlet valve open. . When the cylinder is full of fluid the piston motion is reversed. the piston rod is drawn back creating suction that holds the valve closed.6 23. and the fluid is pumped out. • One type of pump uses a piston.23.

Fluid out . As the pump rotates the vanes move to keep a good seal with the outer pump wall.6 A gear driven hydraulic pump • Vane based pumps can be used to create fluid flow.23. the vanes will slide in and out. The displacement on the advance and return sides are unequal (aided by the sliding vanes). Fluid in As the center core of the pump rotates.7 a geared hydraulic pump Figure 23. The relative displacement across the pump determines the fluid flow. On the inlet side the volume enclosed by the vanes expands. forcing fluid out. drawing fluid in. On the outlet side the volume between vanes decreases.fluids . Fluid is trapped between vanes.

8 Linearizing a hydraulic valve ????? 23. these are not linear devices.7 A vaned hydraulic pump • As with the resistance of valves.8 Figure 23.fluids . It is essential that we linearize the devices. To do this we look at the pressure flow curves. (Note: most motors and engines have this problem) ∆p ∆p ∆ ( ∆p ) K = --------------∆w 1 w ≈ --. Consider the example below.23.3 EXAMPLE SYSTEMS • We can model a simple hydraulic system using the elements from before.( ∆p – ∆p ) K ∆ ( ∆p ) w ∆w Figure 23. pump valve tank Figure 23.9 A hydraulic system example .

23. and it will operate at the point marked. w ∆p Figure 23.fluids . ∆p w For the valve the relationship is given below w ∆p The pipes are all equal length and have the relationships shown below.9 For the pump the relationship is shown in the graph.10 A hydraulic system example (continued) .

5 PRACTICE PROBLEMS 1. 23.23.6 PRACTICE PROBLEMS SOLUTIONS 23.10 23.4 SUMMARY • 23.7 ASSIGNMENT PROBLEMS .fluids .

THERMAL SYSTEMS Topics: Objectives: 24.1 Resistance • A universal property of energy is that it constantly strives for equilibrium.) .24.thermal systems . Materials have a measurable conductivity or resistance (note: this is different than electrical conductivity and resistance. 24.1 24.2 MATHEMATICAL PROPERTIES 24. This means that a concentration of energy will tend to dissipate. When material separates regions of different temperatures it will conduct heat energy at a rate proportional to the difference. Thermal energy (heat) is stored and transmitted through all forms of matter.1 INTRODUCTION • Energy can be stored and transferred in materials in a number of forms.2.

thermal systems .+ ----RT R 1 R2 RT = R 1 + R2 Figure 24.2 Parallel and series thermal resistance . If we have thermal systems in parallel or series they add as normal resistors do.( θ j – θ i ) R where.2 1 q = -. dq 1 --------. R1 R2 R1 R2 11.24.1----.= -. The flow of heat (current) is proportional to the thermal difference (voltage).= ----. dθ .dl R  dl  q = heat flow rate from j to i (J/s or watts) R = thermal resistance θ = temperature d R = ------Aα d = length of thermal conductor A = cross section area of thermal conductor α = thermal conductivity (W/mK) Figure 24.1 Thermal resistance • When dealing with thermal resistance there are many parallels to electrical resistance.

( q in – q out ) C where. 1 ∆θ = --. dθ 1 --------.this rate is referred to as thermal capacitance. When we want to increase the rate of heat absorption we can use a mixer (with a gas or fluid). C = thermal capacitance C = Mσ where.dt C dt M = mass of thermal body σ = specific heat of material in mass Figure 24.4 A mixer to prevent uneven temperature distribution in thermal storage . Figure 24.3 24.e. A mixer is shown in the figure below.24.2. This mixer is just a rotating propeller that will cause the liquid to circulate.= --. And as long as the material has no major changes in structure (i.dq . gas-solid or phase transition) this number is relatively constant..thermal systems .2 Capacitance • Heat energy is absorbed at different rates in different materials .3 Thermal capacitance • One consideration when dealing with heating capacitance is that the heat will not instantly disperse throughout the mass.

3 Sources • If we plan to design a thermal system we will need some sort of heat source. I resistance + V temp P = IV where.3 EXAMPLE SYSTEMS • We know that as we heat materials at one point.4 24. As these metals get hotter the resistance rises.2.thermal systems . hence the temperature is self regulating. there is some delay (consider a spoon in a hot bowl of soup). This delay is a function of heat capacitance (yes there is a parallel to the electrical description).24.5 Electrical heating elements 24. • heating coils are normally made out of some high resistance metal/alloy such as nichrome (nickel and chrome) or tungsten. P = power generated as heat I = current into coil V = voltage across coil Figure 24. • If we run a current through the wire the resistance will generate heat. One popular heating source is an electrical heating element. . they do not instantly heat at all points. If we control the voltage and current we can control the amount of heat delivered by the coil.

thermal systems . C 1 R θ 2.( θ 1 – θ 2 ) RC 2 θ 1 = θ 2 + sRC 2 θ 2 Figure 24.θ 2 = ----.24.( q 0 – q 1 ) dt C1 11 sθ 1 = ----.5 • Consider an insulated sealed chamber with a resistive barrier between sides. C 2 Given q 0 = heat flow into the left chamber from the heating element θ 1.( q 1 ) dt C2 1sθ 2 = --------.( θ 1 – θ 2 )   C1 R Rq o – sRC 1 θ 1 = θ 1 – θ 2 θ 1 ( 1 + sRC 1 ) = θ 2 + Rq o θ 2 + Rq o θ 1 = --------------------1 + sRC 1 d1---. q 0 – -. C 1 = initial temperature and heat capacitance of left sid e d θ 2.θ 1 = ----. q 1 = heat flow through the center barrier 1 q 1 = -.( θ 1 – θ2 ) R 1d---.6 A two chamber thermal system . and a heating element in one side. qi θ 1. C 2 = initial temperature and heat capacitance of right sie Next.

= s ( RC 1 + RC 2 ) + s 2 R 2 C 1 C 2 θ2 qo ---.   RC C   RC C  –qo 1 2 1 2  -------------------------------------.24.   ---------------.+ --------------------------s C1 + C 2 s + -----------------RC1 C 2 Figure 24. RC 1 C 2  RC 1 C 2  RC 1 C 2   qo –qo ----------------------------------C1 + C2 C1 + C2 θ 2 = -----------------.= ----------------- C +C  C1 + C2 s→0  1 2 2  C 1 + C 2 + ( 0 )  s -----------------.7 A two chamber thermal system (continued) .+ s   -----------------.--------------------2 --------------------.+ --------------------------s C1 + C2 C 1 + C 2  -----------------.= 1 + sRC 1 1 + sRC 2 1 + sRC 2 Rq o 2 2 1 +  -------.+ s 2 s s + -----------------RC1 C 2  RC 1 C 2 qo   qo    ---------------.+ s  –  -----------------.+ . s -----------------.  RC C  A B 1 2 θ 2 = -------------------------------------. = 1 + s ( RC 1 + RC 2 ) + s R C 1 C 2  θ  2 Rq o -------.   RC C  qo RC 1 C 2 1 2 A = lim  --------------------------------------s = -------------------------------------.= -. s + C 1 + C 2  = --------------------------.= θ 2 + sRC 2 θ2 1 + sRC 1 θ 2 + Rq o -----------------------------.thermal systems .6 θ 2 + Rq o θ 1 = --------------------.= ----------------------------------lim B =  RC 1 C 2   C1 + C2 C 1 + C 2   C 1 + C 2 2  C 1 + C 2 s → – -----------------. ---------------.= 1 + sRC 1 θ 2 + sRC 2 θ 2  Rqo ------- θ  1 .= s ( C 1 + C 2 ) + s 2 RC 1 C 2 θ2 qo  ---------------. RC1 C 2   RC1 C 2   qo   qo    ---------------.

7 qo –1 –1 1 –1 1 L ( θ 2 ) = -----------------.24.5 PRACTICE PROBLEMS 1.6 PRACTICE PROBLEM SOLUTIONS 24.thermal systems .8 A two chamber thermal system (continued) 24.4 SUMMARY • 24.L -.t  RC 1 C 2  qo θ 2 ( t ) = -----------------.– L --------------------------C1 + C 2 s C1 + C2 s + -----------------RC 1 C 2 – -----------------. 24.1 – e C1 + C2  C 1 + C 2 Figure 24.7 ASSIGNMENT PROBLEMS .

multiaxis motion . OPTIMIZATION Topics: • Objectives: • 25.1 INTRODUCTION • For simple examples use derivatives and find the maxima/minima • Consider the example below to find the L = • local and global optimums .25.1 25.

money .25.2 OBJECTIVES AND CONSTRAINTS • Objectives are those things we want to minimize. simplex 25.volume . or maximize. gradient descent. there is a reduced value.2 • search functions. .power consumption . In this case when the area becomes too large.time . random walk. We want to maximize the value of V. .some combination of factors • Expressed as a function of variables that provides a value • Consider the example of building a fenced pasture.multiaxis motion .mass .

. ( wd – 300m ) + 60\$  2  m yr wd < 300m wd ≥ 300m 2 2 where.  2 m C land where.multiaxis motion .  ( 2w + 2d )  m  where. R = revenue generated by pasture lan d V = R – C fence – C land where.05  ----------. wd  2  m yr 2 \$ R = 0..35  ----.25.1 Example cost function for building a fence around a pasture • The cost function can be written as. V = total value Figure 25. C land = cost for the land \$ R = 0. C fence = cost to construct the fence (\$) w = width of the pasture (m) d = depth of the pasture (m) \$= 0.3 \$ C fence = 20  --.01  ----------.

double d){ double value. the pasture cannot be larger than one 1600m be 1600m beacuse of the constraints of an existing road system. . R. cland. } else { R = 0. if (w*d < 300){ R = 0.01 * (w * d .05*w*d.25. } Figure 25.Cfence . double cfence.multiaxis motion . • Example of constraints.Cland return value. } value = R . Cfence = 40*(w + d). Cland = 0.3 Example constraint functions for a pasture • The cost function can be written as..4 double cost(double w.05 * w * d. w ≤ 1600m d ≤ 1600m Figure 25.2 A subroutine for cost function calculation • Constraints are boundaries that cannot be crossed.300) + 60.

 +  --------------.4 A subroutine for cost function calculation • Slack variables allow constraints to be considered as part of the cost function. double d){ double value.5 Example of slack variables for including constraints • The cost function can be written as. } else { R = 0. Cfence = 40*(w + d).multiaxis motion .Cfence .05*w*d. . Cland = 0.4 d .Cland if(w > 1600) value = 1000000.4 C penaly =  --------------.5 double cost(double w.300) + 60. cland.. } value = R . } Figure 25. w . if(d > 1600) value = 1000000. return value. R.01 * (w * d . if (w*d < 300){ R = 0. Helps with a system with many local minimum.  1600m  1600m V = R – C fence – C land – C penaly Figure 25.05 * w * d.25. double cfence.

} // slack variable slack = pow((w/1600). } else { R = 0.01 * (w * d .Cland . 4) + pow((d/1600). cland. } Figure 25. .05*w*d. if (w*d < 300){ R = 0.6 double cost(double w.6 A subroutine for cost function calculation 25.multiaxis motion . value = R . double cfence.Cfence .slack. Cland = 0. double d){ double value.25. R.3 SEARCHING FOR THE OPTIMUM • Local Search Space • A topographical map shows the relationship between search parameters and cost values. Cfence = 40*(w + d). double slack. return value.05 * w * d. 4).300) + 60.

multiaxis motion .7 Local searches • Global Search space. In this case the system becomes ’stuck’ in a local mimina. .7 Maximum (or minimum) Objective Space Constraints Search path Figure 25.25.

25.8 Local Optima Global Optima Figure 25.8 Global searches • Global Search space.multiaxis motion . . In this case the system searches all maxim.

4.9 Local Optima Global Optima Figure 25. 25.25.1 Random Walk .4 OPTIMIZATION ALGORITHMS • The search algorithms change system parameters and try to lower system parameters.multiaxis motion . • The main question is how to change the system paramters to minimize the system value.9 A Global Search 25.

2 Gradient Decent 25.25.3 Simplex 25.5 SUMMARY • 25.6 PRACTICE PROBLEMS 1.multiaxis motion .4.10 25. . 25. 25.4.8 ASSIGNMENT PROBLEMS 1.7 PRACTICE PROBLEM SOLUTIONS 1.

1 INTRODUCTION .state space analysis .26.1 26. FINITE ELEMENT ANALYSIS (FEA) Topics: Objectives: 26.

2 FINITE ELEMENT MODELS • Consider a central node ’i’ connected to neighboring nodes with springs. nb Ks na Ks ni Ks nc Fi Ks y x nd • Equations can be written to relate the position of node i to the surrounding nodes .26.2 26.state space analysis .

state space analysis . K s ( yb – yi ) Ks ( xi – xa ) F i = ( Fi x.26. Fi y ) Ks ( xc – x i ) + Ks ( yi – y d ) + ∑ Fx = Fix – Ks ( xi – xa ) + Ks ( xc – xi ) = 0 x i ( 2K s ) + x a ( – K s ) + x c ( –K s ) = Fi x ∑ Fx = Fiy – Ks ( yi – yd ) + Ks ( yb – yi ) = 0 y i ( 2K s ) + y b ( – K s ) + y d ( – K s ) = Fi y xi yi xa ya F ix F iy = 2K s 0 – K s 0 0 0 – K s 0 0 0 xb xc yc xd yd 0 2K s 0 0 0 – K s 0 0 0 – K s y b .3 and the applied force.

Fb y ) K s ( yb – yc ) + ∑ Fx ∑ Fx = Fbx – K s ( xb – xa ) = 0 ∴F b x = x a ( – K s ) + x b ( K s ) + = F by – Ks ( yb – yc ) = 0 ∴F b y = y b ( K s ) + y c ( – K s ) . Ks ( xb – xa ) + ( F ax. The ’x’ and ’y’ values are deflections from the unloaded state. Fa y ) ∑ Fx ∑ Fx = F ax + Ks ( xb – xa ) = 0 Ks ( y a – yd ) ∴F a x = x a ( K s ) + x b ( – K s ) + = F ay – Ks ( ya – yd ) = 0 ∴F a y = y a ( K s ) + y d ( – K s ) Ks ( xb – xa ) ( F bx. na Fa Ks Ks nd Fd Fc nc Ks Fb Ks nb y x • Equations can be written to relate the position of node i to the surrounding nodes and the applied force.4 26. The ’Ks’ value is based on the material stiffness and the geometry of the elements.state space analysis .26.3 FINITE ELEMENT MODELS • Consider a central node ’i’ connected to neighboring nodes with springs.

state space analysis . Fd y ) + ∑ Fx ∑ Fx = F dx + Ks ( xc – xd ) = 0 ∴F d x = x c ( – K s ) + x d ( K s ) + = Fdy + K s ( ya – yd ) = 0 ∴F d y = y a ( – K s ) + y d ( K s ) F ax F ay F bx F by F cx F cy F dx F dy = Ks 0 0 0 0 0 0 –K s 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks xa xb yb xc yc xd yd 0 –Ks ya –Ks 0 0 –K s 0 Ks 0 0 Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 –Ks 0 0 –Ks 0 local stiffness matrix • This can be combined into a more .26.5 Ks ( yb – yc ) K s ( xc – xd ) ( F c x. F c y ) + ∑ Fx ∑ Fx = Fcx – K s ( xc – xd ) = 0 ∴F c x = x c ( K s ) + x d ( – K s ) + = Fcy + K s ( yb – yc ) = 0 ∴F c y = y b ( – K s ) + y c ( K s ) Ks ( ya – yd ) K s ( xc – xd ) ( F dx.

6 • A four element mesh n1 Ks Ks #1 Ks n4 Ks n5 n2 Ks #2 Ks Ks n6 n3 Ks n7 #3 Ks Ks n8 #4 Ks Ks n9 .state space analysis .26.

state space analysis .7 For element #1 Local stiffness matrix F 1x F 1y F 2x F 2y F 4x F 4y F 5x F 5y Global stiffness matrix F 1x F 1y F 2x F 2y F 3x F 3y F 4x F 4y F 5x F 5y F 6x F 6y F 7x F 7y F 8x F 8y F 9x F 9y = Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ks 0 0 0 0 0 0 00 0 0 0 0 0 00 0 0 0 0 0 00 0 0 0 0 x1 y1 x2 y2 x3 = Ks 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks x1 x2 y2 x4 y4 x5 y5 0 –Ks y1 –Ks 0 0 –Ks 0 Ks 0 0 Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 –Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 –Ks 0 0 –Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 y3 0 0 0 00 0 0 0 0 x 0 0 0 00 0 0 0 0 4 0 –Ks 0 0 0 0 0 0 0 y4 K 0 0 0 0 0 0 0 0 x5 s 0 0 0 –Ks 0 0 0 0 0 0 0 0 Ks 0 0 0 0 0 0 0 0 –Ks 0 0 0 0 0 0 0 0 y5 K 0 0 0 0 0 0 x6 s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 y6 x7 y7 x8 y8 x9 y9 .26.

state space analysis .26.8 For element #2 local stiffness matrix F 2x F 2y F 3x F 3y F 5x F 5y F 6x F 6y = Ks 0 0 0 0 0 0 –K s 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks x2 x3 y3 x5 y5 x6 y6 0 –Ks y2 –Ks 0 0 –K s 0 Ks 0 0 Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 –Ks 0 0 –Ks 0 added to global matrix F1x F1y F2x F2y F3x F3y F4x F4y F5x F5y F6x F6y F7x F7y F8x F8y F9x F9y = Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 x1 0 0 0 0 0 0 y1 0 –Ks 0 0 0 0 0 0 x 2 0 0 0 0 0 0 0 0 y2 0 0 0 0 0 0 0 0 x3 0 0 0 0 0 0 0 0 0 0 0 0 y3 0 0 0 0 0 0 x4 0 0 0 0 0 0 y4 0 0 0 0 0 0 x5 – K s 0 2K s 0 – K s 0 0 2K s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 0 0 0 0 0 0 –Ks 0 Ks 0 0 Ks 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –Ks 0 0 –K s 0 0 – K s 0 2K s 0 0 –Ks 0 0 0 0 0 0 0 y5 0 2K s 0 0 0 0 0 0 x 6 0 0 Ks 0 0 0 0 0 y 6 0 0 0 K 0 0 0 0 x7 s 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 y7 x8 y8 x9 y9 .

state space analysis .9 For element #3 F4x F4y F5x F5y F7x F7y F8x F8y = Ks 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks x4 x5 y5 x7 y7 x8 y8 0 –Ks y4 –K s 0 0 –Ks 0 Ks 0 0 Ks 0 0 –K s 0 0 –K s 0 0 0 0 –Ks 0 0 –Ks 0 .26.

26. . then the resulting displacements of the nodes can be calculated by inverting the matrix.10 For element #4 F5x F5y F6x F6y F8x F8y F9x F9y = Ks 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks x5 x6 y6 x8 y8 x9 y9 0 –Ks y5 –Ks 0 0 –Ks 0 Ks 0 0 Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 –Ks 0 0 –Ks 0 • If the input forces are known.state space analysis . Consider the matrix for a single node.

the matrix can reflect this by setting the appropriate matrix rows to zero.11 xa ya xb yb xc yc xd yd = Ks 0 0 0 0 0 0 –Ks 0 Ks 0 0 0 0 0 Ks 0 0 0 0 0 Ks 0 0 0 0 Ks 0 0 0 0 0 0 0 0 0 0 Ks –1 Fax Fay Fbx Fby Fcx Fcy Fdx Fdy 0 –K s –Ks 0 0 –Ks 0 Ks 0 0 Ks 0 0 –Ks 0 0 –Ks 0 0 0 0 –Ks 0 0 –Ks 0 • If we assume that node ’c’ is fixed in the ’x’ and ’y’ directions. We can select a value of 1000 for the stiffness.26.state space analysis . and a force of 10 will be . xa ya xb yb xd yd = Ks 0 0 0 0 –Ks 0 0 Ks 0 0 0 Ks 0 0 0 Ks 0 0 0 0 0 –1 Fax Fay Fbx Fby Fdx Fdy 0 0 –Ks –Ks 0 0 Ks 0 0 0 Ks 0 –Ks 0 • The displacements can then be found by selecting values for the coefficents and solving the matrix.

• More complex elements are commonly used depending upon the stress conditions. • Strains can be found by calculating the relative displacements of neighboring points. These can then be used to calculate the stresses. part geometry and other factors.12 applied at point ’a’ in the positive ’x’ direction.state space analysis . easily into the millions.4 SUMMARY • .26. • There are a variety of finite element element methods and applications Computation Fluid Dynamics Nonlinear deformation 26. xa ya xb yb xd yd 10 0 = – 10 0 0 0 3 3 0 10 0 0 0 –10 3 3 – 10 0 10 0 0 0 3 –1 0 0 0 0 3 0 –10 0 3 3 0 10 0 0 3 0 0 10 0 0 0 10 3 10 0 0 0 0 0 • This method generates very large matrices. • To reduce the matrix size techniques such as symmetry are commonly used.

MIT Opencourseware website.8 BIBLIOGRAPHY How.26.31 Feedback Control Course Notes". 26.7 ASSIGNMENT PROBLEMS 1.6 PRACTICE PROBLEM SOLUTIONS 26.5 PRACTICE PROBLEMS 26. "16.13 26. J. .state space analysis .

The other bucket states have similar functions. what is meant by filling the bucket slowly or quickly. The sets are shifted to the right. FUZZY LOGIC <TODO . half full. or full?" And. Rules and membership sets are used to make a decision. the set membership is 0. or a bit of both (0-1). or systems that cannot be modelled with linear differential equations.Find an implementation platform and add section> Topics: • Fuzzy logic theory. If (bucket is empty) then (fill quickly) Figure 27. There is no definite line stating that the bucket is full.1 A Fuzzy Logic Rule Set The outstanding question is "What does it mean when the bucket is empty.1 INTRODUCTION Fuzzy logic is well suited to implementing control rules that can only be expressed verbally.1. but after that it increases quickly. In reality this would probably mean that the valve would have to be turned a large angle before flow begins. When the height is 0. Consider the bucket is full set.Fuzzy logic . 1. If (bucket is full) then (stop filling) 2.1 27. false (0). . These rules concern how fast to fill a bucket.27.2. Notice that the angle function relates the valve angle to the fill rate. • Be able to implement a fuzzy logic controller. As the height increases more people think the bucket is full until they all think it is full. 27. as shown in Figure 27. A simple verbal rule set is shown in Figure 27. If (bucket is half full) then (fill slowly) 3. sets. so nobody would think the bucket is full. based upon how full it is. We can define sets that indicate when something is true (1). rules and solving • Objectives: • To understand fuzzy logic control.

if we are given a height we can examine the rules.75 ( a1 ) + 1.75 membership value is found on the stop filling and a value of a1 is found for the valve angle..2 1 0 1 bucket is half full 0 1 bucket is empty 0 height Figure 27.0.0 . The result for rule 1 is 0. so the center point is chosen to get angle a2. and find output values. as shown in Figure 27.2 Fuzzy Sets bucket is full 1 stop filling 0 1 fill slowly 0 1 0 fill quickly angle Now.3.0 ( a2 ) angle = --------------------------------------------.0 and bucket is empty at 0.27. This begins be comparing the bucket height to find the membership for bucket is full at 0. so the fill slowly set is examined to find a value. Rule 3 is ignored because the membership was 0.75. For rule 2 the result was 1.Fuzzy logic .0. so the 0. bucket is half full at 1. 0. In this case there is a range where fill slowly is 1.75 + 1.75. These two results can then be combined with a weighted average to get 0.

but this will have a minimal effect on the controller performance. If (bucket is half full) then (fill slowly) 1 bucket is half full 0 height 3.Fuzzy logic . If (bucket is empty) then (fill quickly) 1 bucket is empty 0 height 1 stop filling 0 a1 angle 1 fill slowly 0 a2 angle 1 0 fill quickly angle Figure 27. . In this example the set memberships are defined with straight lines. This controller rules examines the system error.3 Fuzzy Rule Solving An example of a fuzzy logic controller for controlling a servomotor is shown in Figure 27.4 [Lee and Lau. 1988].27.3 1. If (bucket is full) then (stop filling) 1 bucket is full 0 height 2. and the rate of error change to select a motor voltage.

If verror is LN and d/dtverror is any then Vmotor is LN.5. verror 1 1 -100 -50 0 50 100 1 rps 0 -6 1 -100 -50 0 50 100 1 rps 0 -6 1 -100 -50 0 50 100 1 rps 0 -6 1 rps -100 -50 0 50 100 1 1 -100 -50 0 50 100 rps 0 -6 -3 0 3 6 rps/s 0 -6 -3 0 3 6 1 0 0 6 12 18 24 V rps/s -3 0 3 6 rps/s -3 0 3 6 rps/s -3 0 3 6 rps/s d/ dtverror 1 0 0 1 0 0 1 0 0 1 0 0 6 12 18 24 V 6 12 18 24 V 6 12 18 24 V 6 12 18 24 V Vmotor LN 0 SN 0 ZE 0 SP 0 LP 0 Figure 27. Rule 1to 6 are calculated in Figure 27. The sets for verror. 5. 6.Fuzzy logic . If verror is SN and d/dtverror is SN then V motor is SN. If verror is SP and d/dtverror is SP or ZE then Vmotor is SP. If verror is LP and d/dtverror is any then Vmotor is LP. 2.4 A Fuzzy Logic Servo Motor Controller Consider the case where verror = 30 rps and d/dt verror = 1 rps/s. d/dtverror. 3. If verror is ZE and d/dtverror is SP then Vmotor is ZE. and Vmotor are.4 vdesired + - verror Fuzzy Vmotor Motor Imotor Logic Power Controller Amplifier vactual Servo Motor The rules for the fuzzy logic controller are.27. 1. . If verror is ZE and d/dtverror is SN then V motor is SN. 4.

If verror is LP and d/dtverror is any then Vmotor is LP. 1 0 -100 -50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 0 6 12 18 24 V 30rps 1rps/s the lowest results in 0 set membership This has about 0. the OR means take the highest of the two memberships 1 0 -100-50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 -6 -3 0 3 6 the AND means take the lowest of the two memberships 1 rps/s 0 V 0 6 12 18 24 1rps/s 30rps 1rps/s 14V This has about 0.0 (out of 1) membership .27. If verror is SP and d/dtverror is SP or ZE then Vmotor is SP.4 (out of 1) membership 3.5 1.Fuzzy logic . If verror is ZE and d/dtverror is SP then Vmotor is ZE. 1 0 -100 -50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 0 6 12 18 24 V 30rps ANY VALUE (so ignore) This has about 0.6 (out of 1) membership 17V (could also have chosen some value above 17V) 2.

If verror is ZE and d/dtverror is SN then Vmotor is SN.5 Rule Calculation The results from the individual rules can be combined using the calculation in Figure 27. so only two terms are used. to give a .0 (out of 1) membership 6.6 4. 1 0 -100 -50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 0 6 12 18 24 V 30rps 1rps This has about 0. If verror is SN and d/dtverror is SN then V motor is SN. If verror is LN and d/dtverror is any then Vmotor is LN.Fuzzy logic .27.0 (out of 1) membership the lowest results in 0 set membership 5. 1 0 -100 -50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 0 6 12 18 24 V 30rps 1rps/s This has about 0.6. In this case only two of the rules matched. 1 0 -100 -50 0 50 100 rps 1 0 -6 -3 0 3 6 rps/s 1 0 0 6 12 18 24 V 30rps ANY VALUE This has about 0 (out of 1) membership Figure 27.

2 COMMERCIAL CONTROLLERS At the time of writing Allen Bradley did not offer any Fuzzy Logic systems for their PLCs.4 SUMMARY • Fuzzy rules can be developed verbally to describe a controller. But.3 REFERENCES Li.27.4 Figure 27. ∑ ( V motor ) ( membershipi ) i n Vmotor = i=1 --------------------------------------------------------------------n i=1 ∑ ( membershipi ) 0. pp. possible in structured text.7 final motor control voltage of 15. 27.6 + 0. Y. • Solving fuzzy logic involves finding fuzzy set values and then calculating a value for each rule. These values for each rule are combined with a weighted average. and Lau.4 ( 14V ) V motor = --------------------------------------------------. “Application of Fuzzy Control for Servo Systems”. • Fuzzy sets can be developed statistically or by opinion.C.= 15.F. 1511-1519.6 ( 17V ) + 0..Fuzzy logic .. . C. It will accept up to 128 rules that operate on sets defined with polygons with up to 7 points.8V 0. IEEE International Conference on Robotics and Automation. It is also possible to implement a fuzzy logic controller manually.8V.6 Rule Results Calculation 27. 27. Philadelphia. other vendors such as Omron offer commercial controllers. 1988. Their controller has 8 inputs and 2 outputs.

8 27. and the sets they use are given below. Develop a fuzzy logic control algorithm and implement it in structured text. If (verror is SP) or (d/dtverror is SP) then (Vmotor is SP). verror 1 1 -100 -50 0 50 100 1 rps 0 -6 1 -100 -50 0 50 100 1 rps 0 -6 1 rps -100 -50 0 50 100 0 -6 -3 0 3 6 rps/s -3 0 3 6 rps/s -3 0 3 6 rps/s d/ dtverror 1 0 0 1 0 0 1 0 0 6 12 18 24 V 6 12 18 24 V 6 12 18 24 V Vmotor SN 0 ZE 0 SP 0 4. If verror = 30rps.6 PRACTICE PROBLEM SOLUTIONS 27. find Vmotor. The fuzzy rule set below is to be used to control the speed of a motor. Develop a set of fuzzy control rules adjusting the water temperature in a sink.5 PRACTICE PROBLEMS 27. 1. Two fuzzy rules. When the error (difference between desired and actual speeds) is large the system will respond faster. When the difference is smaller the .27. Find products that include fuzzy logic controllers in their designs. Suggest 5 control problems that might be suitable for fuzzy logic control. 5. 3. 2. and d/dtverror = 3rps/s. 2.Fuzzy logic . If (verror is ZE) and (d/dtverror is ZE) then (Vmotor is ZE).7 ASSIGNMENT PROBLEMS 1.

9 response will be smaller. 100% Big Error 0% 10 100% Small Error 0% 10 20 Big Output error 50 30 error 30 error Small Output 5 error 50 if (big error) then (big output) if (small error) then (small output) .Fuzzy logic . 20 and 40.27. Calculate the outputs for the system given errors of 5.

inherent in neural networks. inertia. Artificial neural networks behave in much the same manner as biological neural networks. giving many of the same benefits. springs. physiology. Pattern recognition is performed by classifying an unknown pattern through comparisons with previously learned pat- . achieves the increase in speed by distributing the calculations among many neurons. which allows the network to withstand component failures without having the entire network fail. the connection weights are adjusted to minimize the difference between the desired and actual output. friction. Artificial neural nets are fault tolerant. and pattern matching. This activation level becomes the neuron’s output and can be either an input for other neurons. drag. NEURAL NETWORKS Topics: • Basic laws of motion • Gravity. and multiplied by a set of connection weights. the network should produce a good output match for the inputs.pattern recognition. cables and pulleys. and have the ability to generalize based on a limited set of data. After a network is trained with a set of training vectors. Introduction Over the past decade. exhibit the ability to learn and adapt to new situations. the artificial intelligence community has undergone a resurgence of interest in the research and development of artificial neural networks. I. This paper focuses on a popular feedforward model of neural networks. Learning in this network is done by adjusting the connection weights based upon training vectors (input and corresponding desired output). FBDs • System analysis techniques • Design case Objectives: • To be able to develop differential equations that describe translating systems. All of the weighted inputs to the neuron are then summed and an activation function is applied to the summed value.28. dampers. This parallel nature. When a training vector is presented to a neural net. This arises because of the structure which allows neural nets to process information simultaneously. as opposed to the serial nature of traditional digital computers. or an output for the network. Artificial neural networks are mainly used in two areas --. In this model a set of inputs are applied to the network. and psychology [14]. The network structure provides a level of fault tolerance. An artificial neural network is an attempt to simulate the manner in which the brain interprets information as determined by the current knowledge of biology.

A brief survey of these various algorithms is given in [2] and [15]. hence. Many of these solutions require that the kinematics be performed off-line because of the computational demands of the algorithms. however. (ii) Iterative. there is a need for conversion between the two coordinate systems. This ability is termed associative recall. within a quarter of the robot workspace. Human users are more suited to working in cartesian coordinates than the joint coordinates of robots. and a set of control outputs would drive the current manufacturing process. In pattern recognition. Both of the forms are generally rigid and do not account for uncontrollable variables such as manufacture tolerances. Pattern matching uses continuous input patterns to evoke continuous output patterns (i. The inverse kinematics problem is described. the problem of solving the inverse kinematics has been done through the application of various methods. The paper begins with a general introduction to the artificial neuron. A neural network approach could adapt to changes in the robot due to wear from extended use over time. In a popular model which will be . Such algebraic and algorithmic procedures cannot incorporate these situations. such as. calibration error. In the past. one-to-one mappings). Iterative techniques. the network can generalize and choose the closest match [3][4][10][13]. The Basic Artificial Neuron The basic building block of an artificial neural network is the neuron. These solution types are more desirable since explicit equations are available. An example of this is the use of neural networks to recognize handwritten digits [4].2 terns. II. and a five degree of freedom robot working in three dimensional space. and wear. This paper uses the pattern matching strategy for the inverse kinematics of a typical three link manipulator. An example is the use of a neural network as a basic controller for a plant. the training region for the two link robot was a small square in the centre of the workspace. and a description of the backpropagation learning algorithm. This paper focuses on the kinematic control of a three link manipulator working in three dimensional space. when a particular pattern is noisy or distorted. The manipulators were controlled with a high accuracy. and the application of a two-layer neural network is discussed.translation .28. Previous research in the application of neural networks to kinematic control was performed for a to a twolink planar manipulator (working in a plane). This paper explains research extended to investigate the performance throughout the entire work space of the robot. When manipulators interact with the environment.e. A closed form solution is found by decoupling the joints responsible for position and those responsible for orientation. Investigations of a more complete representation of inverse kinematics throughout the entire workspace is provided. The conversion from world coordinates to robot joint angles is called inverse kinematics. Newton-Gauss and Linearization are available when no closed form solutions can be obtained. The implications of the neural network approach to the kinematic control problem are presented at the conclusion of this paper. Newton-Raphson. The controller would accept the plant conditions as the inputs. These methods take one of two basic forms (i) Closed form equations or. The use of compensation networks is explored and the results of the experimental work are presented. a brief description of feedforward neural networks. This approach is discussed in more detail in [11]. the position of the end effector must be described. and the five link robot was limited to a small wedge volume within the robot workspace [5][6][7].

wij opi Neuron uj activation function opj Σ where. These linear functions.28. In order to establish a bias value θj. The bias values are then self-adjusted as the other neurons learn. however. Figure 1 shows the structure of the basic neuron. f(netpj).translation . netpj.3 used in this paper. which gives the output opj . do not take advantage of multi-layer networks [14]. without the need for extra considerations. the hyperbolic tangent is unbounded and hard to implement in hardware. Included in the summation is a bias value θj to offset the basic level of the input to the activation function. After the inputs have been multiplied by the connection weights for input pattern p. where wij is the connection between neurons ui and uj. the connection weights between neurons are adjusted. Each input is multiplied by a connection weight wij. . The connection weights correspond to the strength of the influence of each of the preceding neurons. or connection weight between neuron ui and neuron uj output of neuron ui (input of neuron uj) bias value of neuron uj output of neuron uj (function of the summation f(netpj)) Figure 1 Basic structure of an artificial neuron. Figure 2 shows some commonly used activation functions. their values are summed. in which the output of the neuron is +1 if a threshold level is reached and 0 otherwise. the bias term can appear as an input from a separate neuron with a fixed value (a value of +1 is common). Hyperbolic tangents and the sigmoidal functions are similar to real neural responses. In this paper. the activation function may be in the form of a threshold function. In calculating the output of the neuron. which can be implemented in hardware in future research areas. Squashing functions limit the linear output between a maximum and minimum value. however. the Sigmoidal function is used because of its ability to produce continuous non-linear functions. Each neuron requiring a bias value will be connected to the same bias neuron. wij opi θj opj ≡ ≡ ≡ ≡ 1 opj = f(netpj) netpj netpj netpj = Σ wijopi + θj synaptic weight. The neuron receives inputsopi from neuron ui while the network is exposed to input pattern p.

4 linear sigmoidal logistic 1 hyperbolic tangent 1 Squashing Function 1 linear threshold 1 f(netpj) = 1 . and thus. The input layer does not process information. the architecture becomes an important consideration. The circular nodes in figure 3 represent basic neurons. and an output layer. . A one-layer neural net consists of a layer of input neurons and a layer of output neurons. A simple one-layer network can effectively map many sets of inputs to outputs. thus. [14] and [16]. The input neurons are shown as squares because they only act as terminal points (i. one or more hidden layers.net 1-e -1 pj Figure 2 Common activation functions. Neural Network Architectures in Feedforward Systems A single neuron can only simulate 14 of the 16 basic boolean logic functions. It cannot emulate the X-OR. Each neuron has inputs that only come from neurons in preceding layers. networks propagate the inputs forward through the neural net. please refer to introductions to the subject in [3]. In Feedforward systems. For more in depth information on neural networks. If such a set of connection weights cannot be found using one-layered networks.28. opi = input). [8]. In pattern matching problems. and topography of the function. this depends upon the linear separability of the problem domain. or the X-NOR gates [9].e. this depends upon continuity. A multi-layer network consists of an input layer of neurons. as described in the previous section. then multi-layered networks must be considered. In pattern recognition problems. These limitations require that more than a single neuron be used. it is not considered to be a part of the structure and is numbered layer 0. III.translation .

Supervised learning is required for pattern matching. When a system starts with all connection weights equal.opj) f ´(netpj ) (1) where. In the initial state.28. and opj is the actual output. the network has a random set of connection weights.translation . the estimated error contribution. The delta value for any output neuron is computed as. In order to estimate the delta of a hidden neuron. and f ´(netpj) is the first derivative of the activation function with respect to the total input (netpj) for the given input pattern p evaluated at neuron uj. the connection weights are adjusted to reduce the output error. In order for the network to learn. and will not converge to the global solution. and backpropagation is the most popular of the supervised learning techniques. The calculations begin at the output layer of the network. tpj is the desired output of output neuron uj. a set of inputs are presented to the system and a set of outputs are calculated. In the backpropagation Training Paradigm. δ. the desired outputs of these hidden neurons becomes more difficult to estimate. Backpropagation of Errors Learning methods may be either supervised. For the output layer. the error signal from the output . With the introduction of hidden layers. or unsupervised. IV. by each neuron must be calculated.5 input layer output layer input layer hidden layer (b) two-layer network output layer (a) one-layer network Figure 3 Simple feedforward artificial neural networks. the change in connection weights can easily be calculated since the δ of the output neurons is easily determined. δpj = (tpj . After inputs have been applied and the network output solution has been calculated. A difference between the actual outputs and desired outputs is calculated and the connection weights are modified to reduce this difference [14]. the network begins at a sort of local optimum.

then the weights will overshoot their optimum values. k) means that neuron ui precedes neuron uj. By introducing the momentum term. For an in-depth derivation of the back propagation method. the gradient descent approach requires that steps be taken in weight space. ∆pwij(n+1) = η δpj opi + α ∆wij(n) new delta momentum term (2) (3) where η is the learning rate. The delta.1 to 0. The robotic controller must then determine the correct joint coordinates to locate the manipulator. has been well researched and many good solutions exist. Inverse Kinematics. and ∆wij(n) is the weight change in the previous training interval. opi is the output of preceding neuron ui.28. In order to adjust the connection weights. etc. it may take longer to converge and may become caught in local minima. . The estimated weight change ∆pwij(n+1) is calculated for each input connection to neuron uj from neuron ui by.9. This method helps avoid local optimum points in weight space. the weights may be adjusted. a user must specify a location in three dimensional space. the learning rate can be increased while preventing oscillation. j. The momentum term is introduced by adding to the weight change some percentage of the last weight change. f ´(netpj) is the first derivative of the activation function with respect to netpj at hidden neuron uj.8 The training set is comprised of points which have inputs and desired outputs. δpk is the delta value for the subsequent neuron uk. and gives a much faster global convergence. This process of calculating the δs is performed for all layers until the input layer is reached After all the error terms have been calculated for all the neurons. This area of robotics. should be large enough to come up with a solution in a reasonable amount of time without having the solution oscillate within weight space.2 to 0. The learning rate. and wjk is the connection weight for the link between hidden neuron uj and subsequent neuron uk. In all cases the * note: The subscript notation (i. δpj . A Typical Three Link Manipulator In order for a robotic manipulator to perform tasks within space. and common values of α range from 0. α is the smoothing term. δpj = f ´(netpj ) δ Σpkwjk k where. This allows a weight correction after each point is trained. δpj is the delta of neuron uj. If the steps are too large. If the steps are too small.translation . V. which varies between 0 and 1. Common values for η range from 0. the training points were chosen at random from an ordered table.6 layer must be propagated backwards to all preceding layers. which varies between 0 and 1. for the hidden neurons is calculated as. by creating a random walk type of weight optimization. This process is then repeatedly performed for all training examples until a satisfactory output error is achieved. which may or may not be in a random order. which precedes uk. In this paper. The step size is then a function of the learning rate. The estimated errors are used to estimate the weight changes with the least mean squares method. the reader is suggested to consult Rumelhart and McClelland [14]. The backpropagation method is essentially a gradient descent approach which minimizes the errors.

An example of a robot which may be used with an artificial neural network control system is the three link manipulator as shown in figure 4. These changes may be caused by poor tolerances in manufacture. link 1 2 3 variable θ1 θ2 θ3 α 90o 0o 0o a 0 a2 a3 d 0 0 0 Z θ2 1 Z θ3 2 a3 (px. Z 0 θ1 a2 Table 1 Link parameter table. are given in the equations (4) and (5). The forward and inverse kinematic equations. pz) X Denavit-Hartenburg Transformations c1c23 s1c23 s23 0 -c1s23 -s1c23 c23 0 s1 -c1 0 0 c1(a3c23 + a 2c2) s1(a3c23 + a 2c2) a3s23 + a 2s2 1 T 3 = A 1A2A3 = Figure 4 A typical three link manipulator. and the negative sign (-ve) corresponds to an elbow up solution.7 resultant solutions are highly specific to a particular robot configuration. with exact dimensions. and poor calibration. Such explicit solutions do not tolerate changes over time. . wear over extended periods of operation. damage to the robot. The positive sign (+ve) corresponds to an elbow down solution. and require a computer capable of performing complex mathematical functions.translation . py.28. used by the authors in testing. The ± sign in θ2 indicates the general configuration of the robot arm. can be quite slow. Computations are often complex.

Neural systems are fault tolerant and robust. and • when the end of the arm is directly over the origin. VI. The sources of indeterminacy for this manipulator occur: • for Elbow up and Elbow Down Configurations. When implementing a neural network approach. The network will not fail if a few neurons are damaged.translation . Neural systems can generalize to approximate solutions from small training sets. Other examples of indeterminacy occur in PUMA style robots where left and right arm configurations would give indeterminacy.28. Training a Neural Network for Inverse Kinematics Neural Network solutions have the benefit of having faster processing times since information is processed in parallel.β 2 2 - tan-1 inverse kinematic equations (5) θ3 = tan -1 p z . These sources of indeterminacy may all be overcome by constraining the inverse kinematics to a particular case. complex computers are not essential and robot controllers need not be spe- .a 32 2a2 It should be noted that the inverse kinematics is not determinate.a2c2 - θ2 where s1 = sin θ1 c1 = cos θ1 c12 = cos (θ1 + θ2) s12 = sin (θ1 + θ2) α = + and β = p √ x2 + p y2 α2 + p z2 + a 22 .a 2 s2 α . θ1 is undefined. and the solutions may still retain accuracy. The solutions may be adaptive and still be implemented in hardware by using specialized electronics.8 px = c 1(a 3c23 + a 2c2) py = s1(a3c23 + a2c2) pz = a3s23 + a2s2 θ1 = tan -1 py px β ± α pz forward kinematic equations (4) θ2 = tan -1 √pz 2 + α .

points should be evenly distributed. In this case. 20. and 40.9 cific to any one manipulator. thus. the network is shown with a small amount of arrows. For simplicity. If the points are not presented in a random order.28. The neural network architecture chosen for this problem is a simple network with one hidden layer. θ and θ ). and three output neurons for the estimated joint angles. y. 1 2 3 then the network will be very accurate when dealing with points near the cluster. however. 28. the network is actually fully connected. This network architecture is pictured in figure 5. and to all of the output neurons.1 SUMMARY • . z) and the corresponding outputs (θ . The order in which points are presented to the network also affect the speed and quality of convergence. input layer hidden layer output layer x y z θ1 θ2 θ3 +1 bias neuron Figure 5 Neural architecture for solving the inverse kinematic problem. The network has three input neurons for the desired position. The nature of neural networks require that a set of training points be chosen which represent the nature of the inputs (x. If the training points chosen tend to be clustered. The number of neurons in the hidden layer varied from 10. the robot should be familiar with points throughout the entire workspace. each training update will tend to train the network for the current section of space which the points are from. A bias neuron (with a value of +1) is attached to all of the neurons in the hidden layer.translation .

3 PRACTICE PROBLEM SOLUTIONS 28.10 28.4 ASSIGNMENT PROBLEMS 28.translation .5 REFERENCES .28.2 PRACTICE PROBLEMS 28.

11 .translation .28.

28.translation .12 .

Project . These then support the conclusions drawn by the author. Executive . “Would I skip or skim this sentence/paragraph/section/report because it is so obvious?”. discuss details and then draw conclusions from them. could somebody pick up my report and continue?” if the answer to this question is no the report requires more detail.Theses ‘Lab Reports’ describe one or more experiments. 29. and the conclusions drawn from them. The credibility of a report is based upon the evidence it contains.A summary of current advances in a topic.A technical document describing a project for the purposes of obtaining permission to proceed. Engineers often write reports for other engineers. Laboratory . This should end with some comparison of alternatives. or technically aware readers. The depth of detail is another common issue. the results. The author should provide enough detail to allow others to follow and reproduce the results. reports give necessary background.3 THE TECHNICAL DEPTH OF THE REPORT Knowing the level of reader is very important when writing. appendices can be added to accommodate readers with variable backgrounds.29. observations. “What happens if I am hit by a car. Typically they will also contain a recommendation. and a set of conclusions. Interim .A brief summary of the report.A report to apprise supervisors and others as to the progress of a project or other major undertaking. Proposals . and any implications for decision making at the management levels. In general. They .4 TYPES OF REPORTS The list below shows many of the reports written by engineers.5 LABORATORY REPORTS • Purpose: These reports should outline your procedure and results in detail. Consulting . As a result an engineer will often write at their own knowledge level.A description of work done in a project to inform other engineers who may be asked to take up further work on the project. When necessary. 29. or moved to and appendix. The two questions below should form a good basis for evaluation.3 29. Research .writing guide . test results. If the answer is yes section can be condensed.A summary of details.

will the counter increment/decrement between 0 to 9? • NOTE: These reports are much easier to write if you prepare all of the calculations. If you sit down and decide to do things as you .what is the bandwidth of an amplifier? . We want to use this as a scientific protocol. Title. It will also be very helpful to note other events that occur (e. Any design work is done at this stage 4. etc.writing guide .the results are recorded in tables. As a result the work should be past tense • Laboratory reports should have one or more hypotheses that are to be tested. Examples might be. power loss. The level of detail should be enough that somebody else could replicate the procedure. the lab instructor and the lab section.these are sequential operations that describe what was done during the experiment. Discussion . before you start to write. Conclusions . Results (Note: sometimes procedure and results are mixed) . etc. etc. 8. Author. • Standard Format: 1.a brief one line statement that allows a quick overview of what the experiment is about. Procedure . graphs. . and make general statements either upholding or rejecting our purpose. These have been historically used for hundreds of years and are accepted as a form of scientific and legal evidence. Purpose .these make it clear what the labs contain. and when it was done. Specific identifying numbers should be listed when possible. 2. If testing designs these are the specifications. graphs. as appropriate. high humidity. Equipment .29.what is the thermal capacity of a material? . 3. who did the work. or an instrument is found to be out of calibration. 6. Theory . 5.4 should also contain the analysis and conclusions. At this point we want to consider the scientific method. • Style: These are meant to be written AS the work is done. If there are problems in the data.g.a list of the required equipment will help anybody trying to replicate the procedure. Date .To conclude we will summarize the significant results. we can track the problems to specific sets of data and equipment.At this stage the results are reviewed for trends and other observations.) 7. If the report is done for a course it should also include the course name/number. It is completely unacceptable to make incorrect entries or leave out important steps or data.a review of applicable theory and calculations necessary. This is best written in the form of a scientific goal using the scientific methods. The completeness of detail allows you (and others) to review these and verify the correctness of what has been done.

In a typical use. when coffee is cooled kinetic energy is removed. We can also assume that the atmosphere is so large that the heat transfer will not change the temperature.5.. 29. Wyred Date: Dec. coffee cools as heat is lost through convection and conduction to the air and solids in contact. 23.. . M.5 write it will take twice as long and get you half the marks. The greater the temperature difference. I have written many in the past and I mark them now.1 . the higher the rate of heat flow out of the coffee. believe me... 1998 Purpose: To derive a theoretical model of the rate at which coffee cools and experimentally verify the model and find coefficients.An Example First Draft of a Report Grand Valley State University Padnos School of Engineering EGR 345 Dynamics Systems Modeling and Control Laboratory Exercise 7 Title: The Cooling of Coffee Author: I. Theory: When coffee is heated kinetic energy is added..writing guide . The factors involved in this convection/conduction can be difficult to measure directly. Consider the temperature difference between the coffee and the ambient temperature.29..0. but we can approximate them with a simple thermal resistance. This relationship can be seen formally in the equation below.

and into the atmosphere is governed by the resistance. C coffee = thermal capacitance C coffee = M coffee σ coffee where. The basic relationships are given below. and that there is no energy flowing into the coffee. This is known as the thermal capacitance. Mcoffee = mass of thermal body σ coffee = specific heat of material in mass dθ coffee –1 -----------------. And. q = heat flow rate from coffee to air (J/s) R = thermal resistance between air and coffee θ = temperatures in the coffee and air We can also consider that coffee has a certain thermal capacity for the heat energy. I will assume that the energy flow rate into the coffee is negligible. As the amount of energy rises.( q in – q out ) = --------------. 1 –1 ∆θcoffee = --------------. We will make two assumptions. there will be a corresponding temperature increase.29.6 1 q = -.q dt C coffee The temperatures can be found by consider that the energy flowing out of the cup. .= --------------.writing guide . and this value is unique for every material.q C coffee C coffee where. that the thermal capacitance of the atmosphere is infinite.( θ coffee – θ air ) R where. the temperature in the coffee and air are governed by the two capacitances.

( θ coffee – θ air ) - 1  dt Mcoffee σ coffee R dθ coffee 1 1 ∴-----------------. θ coffee =  -----------------------------------.+  -----------------------------------.+  -----------------------------------. ( A + Be ) =  -----------------------------------.θ = BCe dt ∴BCe Ct 1 1 +  -----------------------------------. θ = θ air + ( θ 0 – θ air )e The time constant of this problem can be taken from the differential equation .= -------------------------------.= 0 Mcoffee σ coffee R A –1 -----------------------------------.+  -----------------------------------.= --------------. -. θ air M  M  coffee σ coffee R coffee σ coffee R B A Ct –1 ∴e  BC + -----------------------------------. θ air = 0    Mcoffee σ coffee R  M   M coffee σ coffee R coffee σ coffee R B BC + -----------------------------------.7 dθ coffee –1 -----------------. Guess θ = A + Be Ct Ct Ct d---. the initial temperature of the coffee should be used. +  -----------------------------------.29.q dt C coffee dθ coffee –1 ∴-----------------. θ 0 = A + Be C(0) = θ air + B –t ----------------------------------M co ffee σ coffee R B = θ 0 – θ air The final equation is.writing guide . θ air = 0 Mcoffee σ coffee R  Mcoffee σ coffee R –1 C = -----------------------------------Mcoffee σ coffee R A = θ air To find B. θ air M  M  dt coffee σ coffee R coffee σ coffee R This differential equation can then be solved to find the temperature as a function of time.

During this period the mass of the full coffee cup was measured and found to be 478g. The temperature was 65C. 3. τ = Mcoffee σ coffee R Equipment: 1 ceramic coffee cup (14 oz. Coffee was poured into the cup and. The coffee pot was filled with water and this was put into the coffee maker.8 above. 4. ground coffee 1 coffee maker . 2. after allowing 1 minute for the temperature to equalize. . Readings of the coffee temperature were taken every 10 minutes for the next 60 minutes. During this period the cup was left on a table top and allowed to cool in the ambient air temperature.writing guide . The mass of the empty coffee cup was measured on the scale and found to be 214g. These values were recorded in Table 1 below. and the pot was full. the temperature was measured again.) 2 oz.Proctor Silex Model 1234A 1 thermocouple (gvsu #632357) 1 temperature meter (gvsu #234364) 1 thermometer 2 quarts of tap water 1 standard #2 coffee filter 1 clock with second hand 1 small scale (gvsu# 63424) Procedure and Results: 1. After five minutes approximately the coffee was done. The temperature of the coffee in the pot was measured using the thermocouple and temperature meter and found to be 70C.29. and the machine was turned on. The air temperature in the room was measured with the thermometer and found to be 24C. The coffee filter and grounds were put into the machine.

A graph of coffee temperature measured at 10 minute intervals . temp (deg C) 60 40 24 20 0 20 τ ≈ 32min 40 60 t (min) Figure 1 . The readings for temperature over time are graphed in Figure 1 below. and from these we can graphically estimate the time constant at approximately 32 minutes.writing guide .29.9 Table 1: Coffee temperatures at 10 minute intervals time (min) 0 10 20 30 40 50 60 temperature (deg C) 65 53 43 35 30 28 26 Discussion: The difference between the temperature of the coffee in the pot and in the cup was 5C. and this will have no effect on the data collected afterwards. This indicates that some of the heat energy in the coffee was lost to heating the cup. but I will assume that the heating of the cup was complete within the first minute. This change is significant. These show the first-order response as expected.

temp (deg C) 60 max.29. and have the time constant. C experimental data mathematical model 40 20 0 20 40 60 t (min) Figure 2 .100 = 8. τ = M coffee σ coffee R τ sC R = -------------------------------. compared to the total range of the data.2 --------KgJ τ = 32min = 1920s .writing guide . If we know the mass of the coffee and assume that the coffee has the same specific heat as water. The graph clearly shows that there is good agreement between the two curves.5% between these two curves.264Kg Cσ coffee = 4. the results can be used to calculate a thermal resistance. difference of 3.Comparison of experimental and modelled curves This gives an overall error of 8.10 We can compare the theoretical and experimental models by using plotting both on the same graph. where there is a difference of 3.5 error = ----------------. the thermal resistance is found to be 1731sC/J.5 deg.5% 65 – 24 Finally. except for the point at 30 minutes.= 1731 ----M coffee σ coffee J M coffee = 478g – 214g = 0.5 degrees C. 3.

writing guide .7 DRAFT REPORTS . Clearly define the objectives for the report 2. The error value is greater than the theoretical value. Do research to find the missing information 4. 29. After reading a research report another reader should be able to discuss advanced topics in general terms.draft reports will be edited for contect.5% of the entire data range. Rearrange the points into a logical structure 6. except for a single data point. This error was most likely caused by a single measurement error.11 Conclusion: In general the models agreed well.2 .5. 29. which probably resulted in a variation in measurement location. • Strategy A: 1. and it will be distributed in the lab. Outline what you know on a word processor in point form and find the ‘holes’ 3. which suggests that the temperature might have been read at a "hot spot". This error was relatively small. Convert point form into full text 7.An Example Final Draft of a Report • A final draft of the report is available on the course website in Mathcad format. As a result they must be double spaced to make room for editing comments. . only being 8.6 RESEARCH • Purpose: After looking at a technical field we use these reports to condense the important details and differences.29. Incorporate the new and old information (still in point form) 5. In the procedure the temperature measuring location was not fixed.0. Proof read and edit 29.

. Title page .writing guide .e.list page numbers .if you must include this section. unviersity) name .8 PROJECT REPORT XXXXX UPDATE USING PROJECT DESCRIPTION ..title of project ..list greek and other variables separately Table of contents .29.date of submission Executive Summary/Overview ..what were the results.. Body. it should be brief .briefly describe the purpose of the work . Title page ...... with upper case first .the introduction is expected to outline details to be discussed in the report.given numbers .the purpose is to let people know if they want ro read the report .. be detailed .. then write Basic elements are.list all variables used in the report .right justify the numbers Introduction .list all major sections.author list .keep it short Nomenclature . and minor sections to a reasonable level .12 29.do exhibits first.Department/Company (i.write it last after all the details are done .put in alphaetical order.course/name .

Quite often the projects that we start will be handed off to others after a period of time. Basically these save a manager from having to read a complete report to find the details that interest him/her. The report typically ends with conclusions.writing guide .9. As work continues on the project add notes and figures 6. suggestions or recommendations.29. central document that gathers all relevant information. Make engineering decisions.13 • Purpose: These reports allow the developer or team to document all of the design decisions made during the course of the project.9. Examine available options and also add these in point form 3. and summarize the results. These reports serve as a well known.1 Executive Purpose: These reports condense long topics into a very brief document. Start to examine engineering aspects of the options 4. do tests as required. 29. and add point form to the document 5. The consultant will review the details of the problem. or we may not have the time. convert the point form to full text. Define the goals for the project clearly in point form 2. 7.9 OTHER REPORT TYPES 29.9. 29.3 Memo(randum) Purpose: An internal business letter designed to convey a business policy or infor- . When the project is complete. • Strategy A: 1. Proof read and edit 29.2 Consulting Purpose: These reports are typically commissioned by an independent third party to review a difficult problem. This report should also mention avenues not taken. In many cases they will not have the opportunity to talk to us. typically less than one page in length.

begin with Date.indicate others who played a significant role in the results.9.14 mation. it will be given a timeline to follow. When a project is initially planned.6 Progress Report Purpose: To provide details of project progress to a supervisor or customer. or worst. . The interim report will indicate progress relative to the initial timeline. A summary of the major work detials and outcomes.put on foan core backing board. approach . as well as major achievements and problems.The format uses bullets (still full sentences) to abreviate the presentation.use a minmimum of 16pt. Subject:. and be able to review the key concepts in under one minute. . or higher font. a purchased cardboard backing board. .use color in a tasteful way.be neat . . . .abundant usage of figures .a brief summary of the purpose of the work. To. 29. 29. cc: 29.9.writing guide .5 Poster Purpose: A presentation format that conveys the key concept visually so that a spectator would have a concept of the project in a glance.the work that was done conclusions . purpose/motivation . Typical sections include.9.4 Interim Purpose: This report is normally a formal report to track the progress of a project.29.the results of the work that was assigned acknowledgements . From.the poster should be self-explanatory .

A techibcally detailed description of the patentable work.writing guide . presenting then reviewing. but is assigned to the employer.9. It is common for the description to include drawings and equations.a patent only gives the holder the right to sue. Abstract . . Normally the are at least a dozen claims.An itemized list of features that make the design unique from previous ideas. .it gives them the right to sue those who use the idea and ’infringe’ .7 Oral Purpose: Work may be presented to an audience in a formal presentation.Normally a patent lists the inventor (normally an engineer).use color. .the presentation should be self-explanatory . If the reviewers are satisfied. or higher font.the patent is submitted to the patent office and reviewed. . then the patent is granted.a brief description of the patent References . visuals and sounds in a tasteful way.A patent is only available for ideas that can be patented. they assign the rights to the employer in exchange for \$1.9. .A list of other patents. . Claims .15 29.The format uses bullets (still full sentences) to abreviate the presentation. or other publications that are related. Description of the Drawings. . Description . .Patents have a life of 20 years from the date of application. by introducing. . These are the items that are the legal basis for the patent. .Tihis normally includes sections such as Summary of the Invention. but it can also be very cruel because an audience may be forced to sit through an uncomfortable experience.use a minmimum of 16pt.abundant usage of figures . . Detailed Description of the Invention. and a patent number is assigned.8 Patent Purpose: A document to describe a design for legal purposes.This can be a very effective method of comminucation when done well. .the patent is structured into the follwing sections. 29.present concepts. .29. These are expected as most patents are built upon previous concepts.Patents are publically available after they are granted to encourage innovation. . If the inventor is an employee.a patent documents that an individual/company owns the idea.

16 29. . drawings. equations. results.10 LAB BOOKS (This is adapted from a document developed by Dr. and preliminary interpretation of the results. Well-kept laboratory notebooks are very important. They facilitate the statistical analysis and interpretation of the data and the eventual writing of a final technical report. every member of the team is expected to keep an independent record (in his/her own laboratory notebook) of the experimental procedure. As a result we use a number of common items. This notebook should be used to record all of the details of the various experiments and tests that are performed no matter how trivial they may seem at the time.writing guide . the notebook should be witnessed and signed by either the instructor or a laboratory assistant. P. The entries should include the following: (i) Date the test was performed (ii) Objective of the test or experiment (iii) A brief description of the experimental or test apparatus or machine and the procedure used (iv) Results obtained . This will facilitate team discussions. Anyalebechi) Each student is required to maintain a well-kept and up-to-date laboratory notebook.11 REPORT ELEMENTS • In reports we must back up our opinions with data. etc. For team studies. The notebook should be submitted to your instructor for grading within 24 hour of the end of the lab. They provide a legal record of events and often make the difference between winning or losing legal cases on patents and on a company's right to use a certain technology in their manufacturing facilities.29.including visual observations and computer data file names (if the results were obtained with a data acquisition system) (v) Preliminary thoughts on the significance of the results At the end of each lab period. Losing one of these cases could bankrupt a company or cost it millions of dollars in royalty! 29.

17 . .use legends that can be seen in black and white . linear regression) .clearly label units and scales .screen captures are clipped to include only relevant data 29.g. charts. • These days it is common to cut and paste figures in software. etc.scanned photographs are clear and cropped to size . graphs. there MUST be a mention of them in the written text.the smallest features are visible . For example “Hardness (RHC)” instead of “RHC” . 1’ • The figures do not need to immediately follow the reference. schematics.figures . We will often move figures to make the type setting work out better. but they should be kept in sequence.label axes with descriptive term.writing guide .tables .the resolution is appropriate .equations • When these elements are included.digital photographs should be properly lit.scanned drawings are clean and cropped to size .try not to use more than 5 curves on the same graph . • If drawing graphs by computer.if fitting a line/curve to the points indicate the method used (e.1 Figures • Figures include drawings. For example “Figure 1 -Voltage and currents for 50 ohm resistor” • In the body of the report the reference may be shortened to ‘Fig. Make sure . • They should be labelled underneath sequentially and given a brief title to distinguish it from other graphs. and cropped to size .29.11.the colors print properly in the final form or print well as black and white .

indicate why the content of the figure is significant. • Smooth lines are drawn for functions or fitted curves. Figure 2 .writing guide . If a curve has been fitted the fitting method should be described.titles should clearly. .18 .11.titles should clearly. indicate why the content of the figure is signif- . and disctinct.29.Various Techniques for Making a Sphere with AMP 29.2 Graphs • Units on axes • Should have points drawn and connected with straight (or no) lines if experimental.avoid overly busy graphs . and disctinct. • if using graphing software don’t put a title on the graph.scale the curve to make good use of the graph .

11.writing guide .titles should clearly. use a good equation editor.3 Tables • Tables are often treated as figures.4 Equations • When presenting equations. 29. 29. • Number equations that are referred to in the text. and watch to make sure subscripts. Table 1: A Comparison of Toy Vehicle Properties Description car truck motorcycle 3 6 2 Number red blue green Color Shape rectangular long small Material die cast polyprop.29. and disctinct. indicate why the content of the figure is significant. • They allow dense information presentation. • table columns should have units as appropriate .19 icant. typically numerical in nature.11. • Box in equations of great significance. • left justify equations . etc are visible. wood • Legends can be added to tables to help condense size.

.= tan θ R cos θ R 1 + cos 60° 0.8 below the theoretical value. Absolute difference ." Percent difference . For example if the theoretical value is 100 and the measured value is 50 then you would say that ".0 then you would say that "..866 ∴ tan θ R = ---------------1 + 0.5 Experimental Data • When analyzing the results from an experiment there are a few basic methods that may be used.writing guide .Useful for multiple data readings.29. .look at each point.11.= ----------------------------------sin θ R cos θ R sin θR sin 60° ∴-------------------------.= -------------. the measured value was 1. the theoretical value was 200% of the experimental value. and compare it to another.8 and the measured value is 104.For example if the theoretical value is 105." OR ".20 • express results in engineering notation • use care with subscripts + ∑ Fx ∑ Fy = – T 1 sin 60° + FR sin θ R = 0 (1) (2) sub (1) into (2) + = – T 1 – T 1 cos ( 60° ) + F R cos θ R = 0 T1 + T 1 cos 60° T 1 sin 60° ∴F R = ---------------------." Mean and standard deviation ..Avoid using this. the measured value was 50% below the theoretical value. but if must remember to assign a sign to the percentage value.. Point by point .5 ∴θ R = 30° 98 sin 60° = FR sin 30° ∴F R = 170N 29...

29. the best rule is be consistent. The numbers are used in the body of the paper (eg.11. 29. 29. It is best to use a graphical format such as a table or list.8 Acknowledgments • When others have contributed to the work but are not listed as authors we may .11.Make a recommendation about changes that would improve that data quality.11.Discuss the fit between theory and measured.Try to develop a conclusion about the quality of the experimental data.6 Result Summary A results summary allows both the author and the reader a chance to review the resuls before a final conclusion. a date viewed and the page author if available. • Reference formats vary between publication sources. • Footnotes are not commonly used in engineering works. [14]). 1997]) and then list the references at the end of the report. • Another method is to list the author name and year (eg.Don’t make vague references to human or experimental error. or the reader may want to get additional detail. Explain differences in terms of hard facts and with numbers.29. . .21 • After data is presented.7 References • References help provide direction to the sources of information when the information may be questioned. In general a few good strategies are. . it should be discussed. . • One popular method for references is to number them. • References on the internet should include a working web address. and the references are listed numerically at the end. But.writing guide . a description of the page (normally the page title). These should include the important results presented elsewhere in the report. [Yackish.

It is customary to provide a summary of the results in the appendix.long listings of computer programs are often put in appendices.11.tables of numerical data are often put in appendices. Typically a sample of the table is included in the body for discussion purposes. 29. These are often provided for the reader who wants to use the data beyond the uses in the report. . It should repeat a number of key details. program listings .multiple sets of data graphs are often put in appendices and summarized in a report body. These are uncommon in student reports. The graphs are often provided so that the reader may use the graphs for verification or further analysis. as if it were an advertisement for a movie.writing guide .11. An abstract should not omit key findings. long tables of data . unless it is tutorial in nature. • Acknowledgments are brief statements that indicate who has contributed to a work.22 choose to recognize them.sample calculations and results. It us used by potential readers to determine if they are interested in reading a report. These listings are provided for readers who want to use the program. 29. When putting material in an appendix it should stand alone and be referred to in the body of the report.these are often referenced in the body of the report for readers who may not have seen a topic previously.9 Abstracts An abstract is a brief summary of the results of the project.10 Appendices Reports often include appendices to reduce bulk from the body. reviews of basic theory . multiple data graphs . Examples of common appendices are given below.29. sample calculations . or a prolonged derivation of equations. They are referenced in the body on the report near the algorithm/calculation/ method they implement. The body of the report has a summary of key assumptions.these are redundant numerical calculations. The calculations are often provided so that the reader may verify the calculations.

blind holes made by drilling must have a drill point shown . number pages by section.put a ’0’ before a point.5 - 29.take care to distinguish between frequency in hertz and radians/sec .11 Page Numbering . Orthographic views .345432in for a dimension measured with a ruler is ridiculous.11. roman numerals starting from ‘i)’ on the first page arabic numerals starting from ‘1’ on the . units used. Avoid number formats such as ’0. .number sections sequentially with roman or Arabic numerals 29.11. kilo.units are required always . milli.a title block with a part name.all parts must be manufacturable Dimensioning .aligned correctly .11. General Requirements . employer/client/etc.or.13 Engineering Drawings The Padnos School of Engineering basic drawing requirements are summarized below.any hole/arc over 180 degrees must have center lines and marks .23 29. For example a length of 0.12 Numbers and Units . tolerances (unless noted on each dimension). 0. giga.number all pages sequentially.use significant figures to round the numbers.greek symbols or the unit name spelled out .29.critical assembly dimensions must be directly readable and not require . designers name. . This is very useful for multi part manuals for example ‘4-7’ would be the 7th page in the 4th section .three views unless axial symmetry allows fewer. date of creation.5 not .writing guide .The location and size of each feature must be clearly defined .00000456’ . etc.a standard border .if pages are blank label them ‘this page left blank’ .the front view should be the most descriptive . mega.use engineering notation (move exponents 3 places) so that units are always micro.

16 Recomendations . .discussions are expected to state all of the details.You must not make statements such as ’this lab was a wonderful learning experience’ or ’all students should enjoy this lab’.writing guide .clear.hole sizes and dimensions should be on the profile view .11. .This is the section to restate the key numbers that support the purpose of the lab.holes that form patterns must be dimensioned relative to each other and relative to a major feature .14 Discussions .clear. 29.This is the section to restate the key numbers that support the purpose of the lab.Smaller tolerances should be used for mating parts .Tolerances must be reasonable for manufacturing . and smaller tolerances indicated for critical dimensions to reduce clutter 29. - .24 addition.mating parts should not have identical dimensions.a general part tolerance should be defined for the part.redundant dimensions should be eliminated Tolerancing . . Sometines these might be obvious when the data is observed. 29. concise recommendations are needed.29. .11.chained dimensions must be aligned . . but they must be stated.smaller dimensions should be closer to the part .arcs/circles more than 180 degrees are sized by diameter.15 Conclusions . concise recommendations are needed.11. they should be free running or press fits.Tolerances must ensure proper assembly and operation at maximum/minimum material conditions . otherwise radius is used. .

. but the software can often be reconfigured.Many technical terms are not in the dictionaries used for checking spelling.11.Put a apce between numbers an units . it must be summarized in the body of the report.’Grammar checkers’ can be used to look for obvious problems.Some material is too bulky for a report body. not words used inappropriately.29.17 Appendices . Basic Spelling . these are normally moved to an appendix. or visually verify.A number of technical terms and acronyms have been .Normally grammar checking will reject text written in passive voice.when material is placed in an appendix. Jargon and Acronyms .A document should always be checked for spelling. .18 Units .it is expected that there will be some duplication between the appendix and the summary in the appendix.radians are one of the units that may not observe norma.11. Basic Grammar .25 29. Considering that utilities for checking spelling are available in most software this is expected. Grammar checking software should not be used as a replacement for proofreading. Technical Spelling . 29.writing guide . so you should also proofread. You may add these terms to the dictionary.12 GENERAL WRITING ISSUES This list below indicates some of the general problems students encounter when writing technical reports and some strategies for fixing these problems.The report must briefly summarize (ussually a figure or equation or more) and then refer to the appendix.verify that units match the numerical results] . This software will also be confused by the interchangeable use of nouns and verbs common in technical english. 29. Be aware that ’smell checkers’ will only point out mis-spelled words. such as input. Using simple sentence structures will reduce problems and speed the writing process. Technical Grammar . . conventions. Be very careful when using the ’autoreplace’ options in software.

read Bad .Current knowledge is based upon knowledge discovered and used before.26 developed for efficiency and clarity. Skip that great opening paragraph . introduction and conclusions last. etc. or leave it out.Simply write what you mean to say.14 TECHNICAL ENGLISH The table below indicates some good and poor choices for words and terms to use in technical papers. Table 2: Technical Words and Phrases Good calculated measured found.writing guide . It doesn’t need to sound impressive .It is quite acceptable to start by writing central sections of a report. If you are having trouble saying it skip it and come back later.29. This can be caused by a lack of understanding about the topic. parted. confuse some readers and make material easily dated. Reorganizing the material often helps to create clarity. All acronyms should be defined at their first use. or possibly how to express the point. In the cases where you don’t know everything you should define what you do/don’t know. This means that no matter how simple something apparently is. it has more layersw of knowledge than could be known by any one person. Use of informal language such as ’show me the beef’ will look unprofessional.Avoid informal language in technical reports.In the case you are not sure what you are writing about you should spend time clarifying your knowledge before returning to writing. Examples include DMM. Colloquialisms . HTTP. 29.13 WRITERS BLOCK Everybody has suffered from writers block before. 29. Knowledge . The obvious cause is not being sure what to write. Many authors will write the abstract. Lack of knowledge . kitted.

29.27 Table 2: Technical Words and Phrases Good chose parallax error optimized human error Bad Table 3: Plain English Good was Bad it became obvious that came in at representing .writing guide .

28 Table 3: Plain English Good Bad Generally bad phrases. in order to be needed to be ----> is needed to be used so as to decided to be can be located found to have found through it was found that .writing guide .29.

29 implementation of important precise exact perfect noted to be invoved allowed for it was found to be was looked thorough ---> reviewed along with .29.selected found .writing guide .used to grade weekly lab reports Design Report Form .measured.with also then decided on .this is used by students for team mates . The Lab Report Form .used to evaluate the final reports Peer Evaluation Form . calculated 29.15 EVALUATION FORMS The forms on the following pages are examples of forms that may be used to evaluate written reports during the semester.

30 EGR 345 . graphs and tables use of english professional style passive voice/past tense suitable use of formatting B = _____ / 10 Overall: A * B = ______ / 100 .writing guide .29.Laboratory Report Evaluation Student: Lab Number/Title: Technical: technically accurate theory section is complete complete procedure and collected data data analysis theory is verified with data conclusions based on data adequate use of numbers A = _____ / 10 Communication: clear and concise good use of figures.

227 Olson April 9.writing guide .31 EGR 345 .29.Design Project Evaluation Sheet Team: no Follows defined formats and requirements yes Writing style is clear and concise and has a logical flow The paper is technically accurate and complete The writing is professional Overall writing quality and style Comments: 29.16 PATENTS United States Patent 6.368. 2002 Method of swinging on a swing Abstract .

The method of claim 1.S.32 A method of swing on a swing is disclosed.119. c) having the user pull alternately on one chain to induce movement of the user and the swing toward one side.writing guide . Agent or Firm: Olson.122. Attorney. 2000 Current U. MN 55104) Appl. No. wherein the method is practiced independently by the user to create the side-to-side motion from an initial dead stop.: 715198 Filed: November 17.123. and then on the other chain to induce movement of the user and the swing toward the other side.125 References Cited [Referenced By] U. Steven (337 Otis Ave. A method of swinging on a swing. Patent Documents 242601Jun.. . 1995Perreault248/228. resulting in a swinging path that is generally shaped as an oval. Primary Examiner: Nguyen. 1881Clement472/118. 3. b) positioning a user on the seat so that the user is facing a direction perpendicular to the tree branch. Inventors: Olson. Paul. relative to the user.29. Kien T. Peter Lowell Claims I claim: 1.S.. 5413298May.. wherein the method further comprises the step of: e) inducing a component of forward and back motion into the swinging motion. The method of claim 1. that is parallel to the tree branch. and d) repeating step c) to create side-to-side swinging motion.120. the method comprising the steps of: a) suspending a seat for supporting a user between only two chains that are hung from a tree branch. 2. Class: 472/118 Intern'l Class: A63G 009/00 Field of Search: 472/118. St.121. in which a user positioned on a standard swing suspended by two chains from a substantially horizontal tree branch induces side to side motion by pulling alternately on one chain and then the other.

Another method of swinging on a swing involves twisting the seat around repeatedly so that the chains or ropes are wound in a double helix. the swing spins quickly. Perhaps the most common is one that includes a seat suspended between two ropes or chains that are hung from a tree branch or other substantially horizontal support. Young children often need help to climb onto a swing. Description TECHNICAL FIELD The present invention relates to a method of swinging on a swing. the swinging method may be practiced independently by the user to create the side-to-side motion from an initial dead stop. which can be entertaining for the user. SUMMARY OF THE INVENTION In accordance with one embodiment of the present invention.writing guide . A new method of swinging on a swing would therefore represent an advance of great significance and value. for example. The method of claim 3. In another embodiment of the invention. BRIEF DESCRIPTION OF THE DRAWINGS FIG. a method is provided for swinging on a swing. Others may be able to begin the swinging movement on their own by pushing with their feet against the ground. 1. 1 is a schematic top view of the swinging path of a swing used in accordance with conventional swinging methods. . These and other features of the invention are described in greater detail below. and may need a push (sometimes even an "underdog" push) to begin swinging.29.33 4. When swinging in this manner. The method comprises the steps of: a) positioning a user on the seat. and then on the other chain to induce movement of the user and the swing toward the other side. a school playground. These methods of swinging on a swing. The swing comprises a seat for supporting a user that is suspended between two chains that are hung from a substantially horizontal tree branch. When allowed to unwind. can lose their appeal with age and experience. and b) having the user pull alternately on one chain to induce movement of the user and the swing toward one side. These swings are often found in side-by-side sets of two or three or more on. wherein the magnitude of the component of forward and back motion is less than the component of side-to-side motion. the user travels along a path as generally shown in the cross-section of FIG. although of considerable interest to some people. BACKGROUND OF THE INVENTION A few basic types of swings have been around for generations. to create side-to-side motion. and once moving may coordinate the motion of their legs and body in what may be called "pumping" to sustain the movement of the swing.

and from any other swing. and begin from an initial dead stop to pull first on one chain. Another improvement on the swinging method described above is the induction into the side-toside swinging movement of a component of forward-and-back motion. or other obstruction or threat to safety that may be present. That is. by skillful manipulation of the body.34 FIG. In contrast to the conventional method of swinging. and is illustrated in FIG. This "forward and back" movement has been known for generations. 1. This side to side swinging method has the added benefit that it can be continued for long periods of time simply by alternately pulling on one chain and then the other. alternately until the user and the swing have begun to swing side-to-side in accordance with the inventive swinging method described herein.writing guide . That is. either or both of which can be used in conjunction with the swinging method described immediately above. the present inventor has discovered that much greater satisfaction can be obtained by alternately pulling on one chain to move the swing and the user toward that side. the swing should be suspended a sufficient distance away from the trunk of the tree from which it suspended.29. support. As is apparent to those of ordinary skill in the area of swinging. The first is that the inventive swinging method can be initiated from a dead stop without pushing. 2. or the like. which is of great benefit to all. The swing is of the type described above. FIG. cables. DETAILED DESCRIPTION OF THE INVENTION The present inventor has created. 2 is a front view of a swinging path of a swing used in accordance with one embodiment of the swinging method of the present invention. The importance of sufficient clearance between the swing and any obstructions or threats to the user's safety is apparent. This enables even young users to swing independently and joyously. 3 is a schematic top view of a swinging path of a swing used in accordance with a second embodiment of the swinging method of the present invention. and is illustrated in FIG. The standard method of swinging on a swing is defined by oscillatory motion of the swing and the user along an axis that is substantially perpendicular to the axis of the tree branch from which the swing is suspended. the present inventor has found it possible to add a relatively minor component . overhead wire. the chains could be replaced with ropes. building. or the tree branch could be replaced with another substantially horizontal support such as a metal bar or pole. the user can climb onto the swing. a new and improved method of swinging. and without the user having to contact the ground. The present inventor has discovered certain other improvements in the art of swinging on a swing. That is. in which a seat is suspended between two chains that are hung from a substantially horizontal tree branch. This side-to-side oscillatory motion of the swing and the user is thus along an axis that is substantially parallel to the axis of the tree branch from which the swing is suspended. and then on the other chain. through experimentation on a standard swing. and then pulling on the other chain to move the swing and the user toward that side. The standard swing should be a single swing that is suspended sufficiently far away from obstructions to make the practice of the inventive swinging method completely safe.

which more accurately replicates swinging on vines in a dense jungle forest. so that the latter predominates. the swinging method of the present invention may be referred to by the present inventor and his sister as "Tarzan" swinging. Lastly. it should be noted that because pulling alternately on one chain and then the other resembles in some measure the movements one would use to swing from vines in a dense jungle forest. 3. It is preferred that the magnitude of the forward-and back motion (shown in FIG. the motion can be more easily continued simply by alternately pulling on one chain and then the other in the manner described. Actual jungle forestry is not required. 3 as being along the Y axis) be less than the magnitude of the side-to side motion (shown in FIG. . 3 as being along the X axis).35 of forward-and-back motion to the side-to-side swinging motion. Licenses are available from the inventor upon request.writing guide . resulting in a swinging path that is generally shaped like an oval. In this manner.29. as is shown in FIG. The user may even choose to produce a Tarzan-type yell while swinging in the manner described.

1. An electric .1 30. This design project will endeavor to build a proof-of-concept prototype for an antisway system for a gantry crane as shown in Figure 30.30. 30.writing guide . At the end of the motion the load is put down. This delay is undesirable because it increases the total cycle time for the crane. PROJECTS 30. The crane will ride on a wooden beam constructed with 2x4 lumber (actually smaller). If the load is swaying.1 The project is intended to emphasize proper project management and team skills to produce a complex engineering system from concept to completion.2 OVERVIEW Cranes are used to move large loads in industrial environments. the operator may have to pause before putting down the load. To do this they normally lift a load and then move it to a new position.

Additional factors used in assessment of performance will be the mass.1 The Objectives and Constraints The design will be mounted on 2x4 lumber and controlled through a wire harness to external electronics and power supplies.30. as it is meant to be equivalent to a cable suspending the mass. It will be expected to carry a 1Kg mass.2 motor will drive the cart left or right. with a 1/4" diameter. The overall system performance will be evaluated on how quickly the crane can move the load. motor driven cart pivot point fixed rail (wood 2x4) 1Kg mass Figure 30. The width of the mass will be less than or equal to the . The 2x4 will be oriented so that the longer dimension is vertical.writing guide . The mass will be provided with a rod that will be between 1/4" and 1/2". cost and overall design quality. and then eliminate sway. The mass will hang below the cart.2.1 The equipment The system will use active compensation to move the cart between positions. The apparatus must be easily mounted and removed without damaging the experimental setup. and to move the cart to compensate for swaying of the suspended mass. The arm must be able to swing freely. 30. suspended 40cm below the top of the 2x4. The mass will be a cylinder that is free to rotate to reduce dynamic effects rotation. The mass will be mounted with a carriage bolt that is at least 3 inches long.

3 width of the 2x4.The equipment cost should be minimized.0.2 B T score =  --s ( 4 ) ( 10 ) ( 10 ) ( 2 )  d where.Teams matched with EGR 101 students Sept 14 . This will be judged before the competition. These may evolve as the semester progresses. 1 = poor) M = mass of the apparatus (Kg) 2 C -------M ------ 30. .3 MANAGEMENT The management details for the project are outlined below.EC616 (6pm) to formalize contracts Sept 16 . The gantry system must move the mass to another position once started. the target value is 0. t s = the time to settle (s) C = total cost of part (\$) d = distance moved in test (m) B = build quality score assigned by judges (0=best.Contract submitted .30. .The test will start with the mass stationary at one position. The total movement range will be 2 to 20 inches in 2 inch increments.3.The settling time. ts.Teams assigned Sept 12 . should be less than 2cm below the bottom of the 2x4. The objectives for the design are itemized below. None of the cart. .EGR 345 Teams visit EGR 101 lab . t 200 0.2Kg.writing guide .Mass should be minimized.5 inch margin around the commanded position.1 Timeline .A clearly justified and supported design process. 1=worst) T = theory quality score assigned by judges (0=best. The following equation will be used to assess the overall score for the design. 30. .Tentative Sept 12 . This time will be determined using the best ts/d for two arbitrary target motions. except for the arms. to less than a +/. . The goal of \$150 or less has been specified as very important.

In the event that team members cannot resolve differences with a team member.when problems arise.Design concept approved Oct 24 . Personal conflicts must be resolved by team members in a professional manner. 345/101 building begins Nov 12 .Preliminary design concept submitted with specifications materials list estimate budget estimate Gantt charts Oct 13 .Final report draft posted to the web 30.First draft of report posted to the web used to determine the score for competition Nov 25 .30. except . even if they are not your fault. . All team members are expected to work in a professional manner.2 Teams The teams are composed of up to 4 students from EGR 345 and one student from EGR 101.Second test Nov 20 . The EGR 345 teams are grouped using the skills survey attached in the appendices.communicate expectations and problems clearly .Proposal submitted detailed drawings (CAD) materials list budget calculations/simulations Oct 29 . Teams are expected to divided tasks for members to work in parallel. This is particularly true of calculations.First test Nov 19 .writing guide .Proposals approved. help to solve them. the team may ’fire’ the team member by a vote (it must be unanimous. In the case of non-participation the penalty may be up to 100% of the project grade. It is also expected that team members will review the work of others to ensure accuracy and completeness.Competition and judging Dec 4 . Don’t lay the blame for problems on others. materials lists drawings and budgets.be polite and accommodating . The general rules of conduct in a team are. The performance of the team will be assessed using peer evaluations on a regular basis and may impact individual grades.treat others as you want to be treated .4 Oct 4 .3.

Although the students are at different points in their academic careers they are all considered peers.Lists of components and budgets to indicate the major parts of the system. A good set of sketches will include 2-D. As with most contracts. Components that would be expected for this type of design are. The firing mechanism is intended to deal with individuals who ’harm’ the progress of the team. 30.1 Conceptual Design Concepts are normally communicated with sketches that make the overall design clear. isometric and pictorial views. it outlines a process that is mutually beneficial. To help clarify issues of differences between the EGR 101 and 345 students.4 DELIVERABLES The following items are to be submitted at various stages of the project. . a ’contract’ will be written that formalizes the responsibilities of the students in EGR 101.5 for the member in question).2 EGR 345/101 Contract This project involves participants from EGR 101 and 345. marks will be deducted to hold the responsible party(s) accountable.30. 30. .Calculations should be provided that support the design concept. In the event they cannot find another team to join. they will be expected to perform all of the work themself.4. . . In the event of a failure. not for non-participants.Sketches should be done using normal drafting practices.writing guide . and should have equal voices in the team. 30.Electrical schematics to describe the control system. .4. such as flowcharts. In this case both parties are expected to benefit in an educational sense.Other items.Block diagram to describe the control system. The contract will be written in such a manner that if all requirements are satisfied work will be marked ’as-is’. . In this case the ’fired’ student is responsible for finding another team that will accept (or ’hire’) them. are often required to clarify the design concept. .

but complete sentences must be used. other items will be requested. Cover Page . Software . calculations and/or simulations to verify the new design. Testing . including the EGR 101 team members. Budget . project and tema numbers. Mass Table . these should normally be accompanied by a new set of models. These reports will include the following elements divided into sections with a heading for each.Design changes should be indicated. and a skills inventory. including major accomplishments and issues. The names of all team members should be listed on the cover.writing guide . . but a thorough presentation of the design details.30.a cover page indicating all of the team members and all other pertinent informa- .The status of ordered items should be indicated. schematics. If there is nothing to be said about a category use ’no changes’. or ’complete’ as appropriate.a cover sheet indicating the course. include an updated mass table.4. including any numerical results when available. Gantt chart .4.when changes are made. such as a combined timetable for all team members. Typical elements are listed below in a typical sequence.The current status of software development should be indicated. Point form is preferred. and current action items. Purchasing . Other Issues .4 Design Proposal The design proposal is used to present all of the design details in a single document. ’nothing done’.Items that may impact the success of the team should be indicated. Fabrication . When appropriate. The focus of this document is a MINIMAL AMOUNT OF TEXT.6 30.’) Each section should include items completed since the last report.3 Progress Reports Teams are expected to submit progress reports on a weekly basis. include an updated budget. (Note: ’Parts purchased’ should be ’Parts were purchased for the cart assembly. or equivalent should be included. 30. Early in the semester.The testing progress should be indicated.The status of items being built/assembled should be indicated.when changes are made. including the overall performance equation.updated on a weekly basis and included each time. Appropriate drawings. These should only be included in the reports the weeks they are requested.A prediction of performance. Performance . Design .

In the final report. . In general the format of the report is as outlined below.a Scilab program that verifies the operation of the system using the equations of motion.the equations of motion for the system . electrical and software design aspects. copies of receipts. with the addition of the following elements. 30.additional calculations for mechanical design issues. Drawing summary .writing guide . Normally these result in a factor of safety. or catalog pages will be required.circuit schematic . such as stress that may result in failure.the drawings.a C program that implements the controller as designed. etc. . as shown in the diagrams.block diagrams showing the system architecture . including a BOM .7 tion. it should be documented .4.block diagram of the control system .5 The Final Report The final design will follow the same structure as the Design Proposal. Catalog pages and quotes can be used to validate the budget.30. drawings.one page or less that summarizes the design and results.a budget listing each of the parts that must be purchased/acquired. Subsections should focus on the following elements. calculations. itemized by each part of the design .selected isometric and assembly drawings System block diagrams Description of control scheme. etc.an assembly drawing of the mechanism. The report should concisely and clearly describe the design.this section should describe the mechanical.three view drawings of each significant part . Cover page Executive summary . calculations. such as the motion profile Schematics . should be based upon the final design. Sections and subsections should be numbered. . but it is a minimal/optional part of the report. . .a weight inventory. Table of contents Design description .if a motion profile will be used.test results . It is reasonable to write a page or less about the modifications that were required.a table of contents .

Appendix . . including sequencing.htm. The results of formal tests should also be described. if one task is preparing mechanical drawings.Controller C program Appendix . theory to backup the design (does the theory match the actual design?). did the theory and actual match? 30. There should be a comparison of the device with and without sway compensation.Most project parts should have a development. In general a Gantt chart should include.A brief description of the overall results indicating what the strengths and weaknesses of the design. otherwise a larger sheet can be used.5 REPORT ELEMENTS 30.e. A comparison of overall score estimates. this should be followed by a review task .Stress and other similar calculations Appendix .done by another team member.8 Calculations .. how clear is what has been done?).Simulation program Appendix .Drawings Appendix .30. . . If it is small enough this can be one page. Conclusions . a brief tutorial on project management with Microsoft Project is available at http:// www.writing guide . For example.Suggestions for improvement. and then a review stage by another team member.5. .stylusinc.On a weekly basis a Gantt chart should be updated to include the completion of tasks.1 Gantt Charts Gantt charts are used to track major project tasks.shtml. You may also download free project management software from http://www.smartworks. .Receipts and cost evidence Final reports will be evaluated on numerous factors inculding the clarity for the design documentation (i.FBDs and differential equations Project budget and BOM Weight inventory Test results .us/htm/downloads.this section should describe Simulation results The tests that were done to describe the overall performance. Reccomendations .net/ms_project_tutorial/project_management.A Gantt chart should be presented on one single sheet.Most tasks that stretch more than a week should be broken into smaller sub tasks.

and prices provided. others are tougher. similar devices can be identified from catalogs and those prices may be used.5. Consumables.30. such as Ryder’s Hobbies on 28th St. late 1 week). not ordered. A BOM is normally found on assembly drawings.5. due 2 weeks. some are more brittle. This includes dimensions and tolerances that can be produced using the available equipment and materials. while drawing means it is done formally in a CAD package. Shaded views have very little value and should be avoided. approximately. wireframe drawings are much more useful. and then generate multiview drawings. you may be required to purchase these yourself. . Normally you should create solid models. The budget should also list the quantity of parts/material. bolts. change the background to change the quantity of toner used. However. Please note that ’sketch’ means that it is done by hand. 68HC11 Axiom board \$89 Grayhill encoder est.2 Drawings All drawings will observe the standards used in EGR 101 (see the EGR 101 or 345 course pages). Assume the power supply is provided as part of the crane system and therefore has no cost is associated. such as bolts are normally listed under a ’miscellaneous’ heading. . under a miscellaneous category with a general cost estimate. If the components have been drawn from the engineering stores.. List simple commodity items such as wires.there are different type of plastic. Some of the commonly available components are listed below. Notes..9 30. \$8 Power Supply. list it but do not give a price Gearhead motors assume \$10 A Bill of Materials (BOM) lists all of the parts required to produce or assemble some other device. An excellent local source of small parts is hobby stores. 30. If your design calls for parts not commonly issued to engineering students. Don’t forget to include cost of the controller and other components used. received. price.3 Budgets and Bills of Material Budgets should list all substantial components. Note: All drawings must have a title block. If there is a definite need to include a shaded drawing. all other components should be listed. source/supplier and status (eg. In other words. This is different from parts listed in a budget in that some of the parts will be work in process. the original material has been worked on to produce new parts.writing guide . etc.

30.5. These should follow the conventions used in EGR 345.Sample System The example system in Figure 30.2 is to control a cart with a hanging pendulum. As the cart moves the mass underneath may sway.4 Calculations Calculations are required to justify the design work.2 System Architecture A block diagram for the system is shown in Figure 30. The two control signals are added to produce a command signal the drives the load to the final position and then reduces vibrations. 30. The position of the hanging mass is measured with a potentiometer.writing guide .3. There are two feedback control loops.6. 68HC11 PWM H-bridge V s Motor ω cart x hanging θ mass mass Digital in motor position Analog in mass angle quadrature input encoder Vθ 1 --D θ motor potentiometer Figure 30.6 APPENDICES 30.1 Appendix A . The first loop adjusts the motor for general positioning of the system. The inner loop adjusts the output to reduce sway of the cart.10 30. they should be commented and included. A motion generator is used to generate a smooth path for the mass. The motor has an attached encoder to measure the motor position. . When computer programs are written. The cart is driven by a motor.

3 System Block Diagram .30. C p = commanded positio n C d = desired position setpoint Cp encoder K enc C pc K pp + + C Lc K sp CL A/D VL Cc PWM Vs motor θp θL Cd C pe cart and mass potentiometer K pot θ L – Vzero where.writing guide .11 Cp generate setpoint table update setpoints Cd where. C pe = the position error K pp = the proportional gain for position control C p = the position count from the encoder θ p = the rotational position of the motor C pc = output command for position control C c = the combined motor control output V s = effective voltage to the motor θ p = position of the motor (cart) θ L = angle of the load from vertical d V L = voltage proportional to angle of loa C L = integer value for load angle C Lc = output command for sway control Figure 30.for a stationary cart .

12 ·· x θL Fp ·· M p lθ L x l Fw Mc Fp θL where.writing guide .4 FBDs for the suspended mass and the compensator .30. M c = mass of cart M p = mass of payload F p = force in suspension arm ·· Mc x Mp · 2 θL Mpl Mp g cos θ L ·· x sin θ L Mp p M p g sin θ L ·· x cos θ M L F w = force from wheels θ L = angle of payload from vertical l = length of suspension ar m r w = radius of cart wheel ·· ·· ∑ F = – Mp g sin θL – Mp lθL – x cos θL Mp = 0 ·· ·· lθ L = – g sin θ L – x cos θ L (1) · 2 ·· = Fp – Mp g cos θ L – θ L M p l + x sin θ L Mp = 0 · 2 ·· F p = M p g cos θ L + θ L M p l – x sin θ L Mp ∑F ∑ Fx ·· = – M c x + Fp sin θ L + Fw = 0 ·· – M c x + Fp sin θ L + F w = 0 · 2 ·· ·· – Mc x + ( Mp g cos θ L + θ L Mp l – x sin θ L Mp ) sin θ L + F w = 0 (2) Figure 30.

 – x  -------.writing guide . + θ L  -------------------------------------------------------------. = Vs   w  w r R r  r R w w w x = θw rw 2 K·· J- ·  K - ∴F w = Vs  -------. – x  -------.  r R  r2   r 2 R w w w (1) ---> 2 ·2 ·· ·· K ·· J- ·  K - – M c x + ( M p g cos θ L + θ L Mp l – x sin θ L M p ) sin θ L + Vs  -------.13 2 Fw rw · KK ----. + x  ----------------------------------------------------------------------. – ω  ---. – x  -------.30. – ----------ωw + ωw   s J JR JR ∴F w given 2 · J KK -------.  r R 2  r 2 R   w rw w    M p gr w M p l sin θ L r w · 2 ·· x = ( cos θ L sin θ L )  -------------------------------------------------------------. = ( cos θ L sin θ L )Mp g + θ L ( Mp l sin θ L ) + V s  -------. = 0  r R  r2   r 2 R w w w 2  J·2 2 ·· K·  K - – x  M c + ---. – x  -------. = 0 2  r R    r 2 R w rw w 2 2 2 2  Mc r w + J + ( sin θ L ) Mp r w 2 · K·· ·  K - x  -------------------------------------------------------------. – x  ---.5 Adding the differential equation for the motor .+ ( sin θ L ) Mp + cos θL sin θ L M p g + θL ( M p l sin θ L ) + Vs  -------. – ω  -------.  R ( M c r 2 + J + ( sin θ L )2 M p r 2 )  R ( M c r 2 + J + ( sin θ L )2 Mp r 2 ) w w w w 2 2 Figure 30. +  Mc r 2 + J + ( sin θ L ) 2 Mp r 2   M c r 2 + J + ( sin θ L ) 2 M p r 2  w w w w 2   ·  Kr w –K V s  ----------------------------------------------------------------------. – x  ---. = V  ----.

writing guide .14 State Equations: · x = v 2 M p gr w · = v  ----------------------------------------------------------------------.6 State equations for the system A Scilab program that simulates the given system is given in Figure 30.30. + ( cos θ sin θ )  -------------------------------------------------------------.  R ( Mc r 2 + J + ( sin θ L ) 2 Mp r 2 ) w w · v cos θ L · –g ----.  M c r 2 + J + ( sin θ L )2 M p r 2  w w   Kr w + Vs  ----------------------------------------------------------------------. .7.sin θ L – ----------------ωL = l l Figure 30. –K - - v   L L  R ( Mc r 2 + J + ( sin θ L ) 2 Mp r 2 )  Mc r 2 + J + ( sin θ L )2 Mp r 2  w w w w 2 + · θL = ω L 2  M p l sin θ L r w 2 ω L  -------------------------------------------------------------.

.// The number of steps to use t_start = 0.3)). Cd)). Xd = 10.// the end time h = (t_end .// the setpoint 10 turns == 160 pulses Cd = ppr * Xd / (2 * PI * rw) . 6.:) + h]. endfunction // Integration Set the time length and step size for the integration steps = 5000. h) Vs=control(motion(0. X = [X ..// precalculate some repeated terms to save time term4 = sin(state(\$. foo = [ state(\$. .t_start)... v_dot = -state(\$. 4).// Precalculate these terms to save time term2 = R*term1. state(\$. h)]. .state(\$.:)..// the step size t = [t_start]..writing guide .. .// d/dt v + term3*term4*Mp*g*rw^2 / term5 ..t_start) / steps.t_start ) / (t_end . C_end) if t_now < t_start then foo = C_start. end endfunction // define the state matrix function term1 = Mc*rw^2 + J.30.t. elseif t_now > t_end then foo = C_end.16 // The motion profile generator function foo=motion(t_start. C_start.3)).// the start time . 2) * term3 / l . 0. t(\$...2)*(K^2) / (term5 * R) . function foo=derivative(state.// an array to store time values for i=1:steps. else foo = C_start + (C_end . t(\$. + Vs*K*rw / term5.// d/dt theta -g * term4 / l . //printf("%f %f \n". Cd.8 A Scilab program to simulate the swaying mass on the crane . + state(\$.// d/dt x = v v_dot. 1).:) + h * derivative(X(\$. X(\$. term5 = term1 + Mp * (term4 * rw)^2. // Vs=control(Cd)....// first order end Figure 30.// d/dt omega ].C_start) * (t_now ..2).normally use zero t_end = 10.4)^2 * Mp*l*term4*rw^2 / term5 . term3 = cos(state(\$. t = [t . Vs). t_end.:). t_now. t(\$.

1) + l*sin(X(i. leg="position@theta (X 10)"). xmass = X(i. if xmass > (Cd + tolerance) then ts = i*h + t_start. X(i.3))].30.writing guide .1) + l*sin(X(i. i * h. end end printf("\nTheoretical settling time %f \n". for i = 1:steps. [X(:.9 A Scilab program to simulate the swaying mass on the crane . [X(:. 3)].1)/intervals * steps + 1).17 // Graph the values for part e) plot2d(t.tolerance) then ts = i*h + t_start. end // find the settling time in the results array ts = 0. -5]. //plot2d(t. X(i.3)). end if xmass < (Cd . printf("Mass location at t=%f x=%f \n". i * h. xmass). v=%f. for time_count=1:intervals.1).3)).1) + l*sin(X(:. omega=%f \n". 1). [-2]. 3). 10*X(:. X(i. X(i. [-2. xmass = X(i. ts). i = int((time_count . xtitle('Time (s)'). // printf the values over time intervals = 20. 4)). 2). Figure 30. theta=%f. // printf("Point at t=%f x=%f. leg="mass position").

2.writing guide . 12. The 345 students will prepare the Design Report Draft using the ProE drawing submitted by the 101 student. Jack and/or Farris will comment on the Formal Proposal. 5. 9. 11. as shown on the Schedule of Actions. The result must be a fully functional systems that meets the published design objectives. They are responsible for submitting the report by the date in the Schedule of Actions. 101 and 345 students are expected to resolve any conflicts informally and mitigate any losses.30. This will be signed by all parties and submitted to Dr. Pete Zaa. 4. In the event that one or both parties fundamentally breach the . The 345 students are to submit the final report with all necessary changes by the date listed in the Schedule of Actions. 10. At the end of this period the design must be fully documented in ProE and be ready for inclusion in the Design Report Draft. Both the 101 and 345 students will participate in the Competition listed in the Schedule of Actions. This list will be finalized and signed. Articles: 1. Any changes made to the design must be updated and submitted to the 345 students for inclusion in the Design Report. mass. 7. The cart will be build according to the Cart Build Approval before the date specified on the Schedule of Actions. according to the date on the Schedule of Actions. The design work will result in the submission of a Formal Proposal. Both the 101 and 345 students will participate in the first tests to verify the operation of the system and develop a First List of Deficiencies. according to the date on the Schedule of Actions. Farris by the Scheduled date. The 345 students are expected to prepare a functional design for an anti-sway system for a crane. The 101 student is expected to produce a cart that is built to professional standards. This design will be documented fully by the 101 student using accepted CAD practices and ProE.2 Appendix B . This will be labelled Cart Build Approval. The 101 student is to participate in the design and construction of a cart as outlined below. The 101 student will be responsible for all other design details related to the cart including the geometry. The 345 students are to prepare a design and construct a multicomponent system that uses the cart as described below. Anne Nyther and Robert Sochs. The First List of Deficiencies will include a list of remedies to be performed by the 101 and 345 students. They must ensure that this will lead to a system that functions within the rules of the competition. 8. budget bill of materials and construction. 3. at all times. generally accessible files formats should be used. The 345 students are to.6. 6.18 30. This contract has been entered into this date by the parties of the first part Joe Junior. with Virve Meurte. maintain a functional design concept. Both the 101 and 345 students will participate in the final tests to verify the operation of the system and develop a Final List of Deficiencies. This design will include a cart that is designed in coordination with the 101 student. This list will be finalized and signed. The 345 students will be required to do all calculation including system dynamics and strength of materials. to be referred to as ’101 student(s)’. The Final List of Deficiencies will include a list of remedies to be performed by the 101 and 345 students. When communicating drawings. In the event of a dispute. Dr. to be referred to as ’345 students’. Based upon these comments the 101 and 345 students will revise the design and agree upon a design for the cart.EGR 345/101 Contract The text below outlines the general form of a contract between students in EGR 101 and 345. All drawings are expected to observe professional standards.

6.Final test completed and Final List of Deficiencies submitted Nov 25 .writing guide . This may involve actions as severe as receiving a failing grade in the project.Forms Skills Self Evaluation Peer Evaluations .Initial build completed Nov 12 .Design Report Draft submitted for review Nov 19 .Competition Dec 4 .Cart Build Approval submitted Nov 11 .First test completed and First List of Deficiencies submitted Nov 15 . Schedule of Actions Oct 15-24 . one or both of the parties will be penalized.30.19 contract Dr. Jack will acts as arbiters. Exhibits: 1.Submit Final Report 30. If this occurs. Farris and Dr.3 Appendix C . 2003 .Cart designs are developed by 101 and 345 students resulting in submission of the Formal Proposal Oct 30.

creating web pages. work. Leadership Skills: The ability to act as a role model that teammates will follow. metal or other materials. etc. . etc. Hands-on Computer Application: Programming and computer interfacing Mathematical Problem Solving: Ability to formulate and solve complex problems Writing: Layout and write complex documents Teamwork Skills: The ability to work with others in a team environment. etc.30. proficient 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 1 1 1 1 1 1 1 1 1 Hands-on Electrical: Basic wiring skills.give hours for each) Other Items of Interest: . Hands-on Computer Usage: CAD.writing guide . plastic. soldering. Personal/Technical Strengths: Personal/Technical Weaknesses: People you would like to work with: People you would NOT like to work with: Other Commitments (courses.20 Skills Self Evaluation Your Name: none Hands-on Mechanical: The ability to build components with wood. Design Skills: Work in unstructured/semistructured problem solving. Spreadsheets.

explain and evaluate the technical aspects of the project in a clear concise manner? poor 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 1 1 1 1 1 Did your teammate come to meetings on time? Did the teammate participate in all aspects of the project? How much did the teammate’s efforts contribute to the overall success of the project? Did you teammate complete individual tasks on time? Did the teammate keep the project progressing forward in a timely manner with a consistent effort throughout the project or was the teammate only available when the team was in trouble? Was you r teammate willing to accept and carry out individual tasks on time? How well were these individual tasks carried out? Did your teammate do his or her fair share of the work? Would you be happy working with the person again? Would you give this person a job reference? Would you hire this person: yes / no Other Comments: .writing guide .21 EGR 101 / 345 Project Peer Evaluation Your Name: Your Class: EGR 101 / 345 Person Being Evaluated: good Communicates well: Works in team environment: Meets deadlines: Quality of work: Overall: Did the teammate return e-mails and other forms of communication promptly? Could the teammate understand.30.

1 ] [ 0. ts = s Overall Score: based upon other values ------------2 s  t- ( 4 ) 200 ( 10 ) B ( 10 )T ( 2 ) 0. 0 is the best C = [ 0. d C M SCORE = Comments: .5in.22 Competition Check Sheet Team Number: Arms Swing Freely: arms should swing freely.0.30.writing guide .2 score =  -. 1 ] Technical Analysis: 0 is the best T = Test Distance: the distance specified in the test d = m Settling Times: The time required to move within +/. from the target. with the exception of measurement apparatus YES / NO Height of Payload: 40cm below the beam = 40cm Mass of Cart: not including the harness M = Kg Cost: proof of costs C = \$ Build Quality: assigned by judges.

30.writing guide .23 EGR 101 / 345 Project Judging Your Name: Work Quality: excellent 1 2 3 4 5 6 7 8 9 10 11 12 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 2 3 2 3 2 3 2 3 2 3 2 3 2 3 2 3 2 3 2 3 2 3 2 3 4 4 4 4 4 4 4 4 4 4 4 4 average 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 5 6 7 8 8 8 8 8 8 8 8 8 8 8 8 poor 9 10 9 10 9 10 9 10 9 10 9 10 9 10 9 10 9 10 9 10 9 10 9 10 Team Number: .

Don’t just look at the final grade for an assignment. 31. .Don’t start a long problem at the bottom of a page. . don’t be careless with these.Use engineering notation.solving problems is important..Problems should be solved in a logical order from the top to the bottom of the page (not left to right).Box final answers .Always starts mechanics/circuits problems with FBDs or schematics .When comparing results. . but eliminate them .Eliminate redundant variables . problem number and date on the top of each sheet. 31. not scientific notation .Use significant figures as appropriate .verify that numerical magnitudes are reasonable. many problems will take less than one page.Note: a calculator is not a method . .Variable subscripts are very important.Show derivations for all problems .1 BASIC RULES OF STYLE The basic rules are listed below. .Create temporary variables when necessary. use different methods . This is also expected when different calculations should lead to the same result. . Ussually comments will include elements that should be fixed.writing guide . .2 EXPECTED ELEMENTS Students in the junior year are expected to..Use green engineering paper and write your name.Always simply the results of calculations .Don’t cram things into one page. read the comments.1 31.. .verify that units make sense .31. . ENGINEERING PROBLEM SOLVING .Compare the results of calculations when asked to do so. In later assignments these items normally result in reduced marks..

This means if answers don’t match.writing guide .when used to solve problems.If answers are provided. . 31. . students should check them.31.2 .when solving problems using computational tools. 31. such as C programs.Equations are plotted as smooth functions.3 SEPCIAL ELEMENTS 31. .3.Frequency responses should be in phase angle forms.Data is plotted as discrete points.Label all axis. solutions should match normally. include program listings . provide copies of the programs. including units .4 SCILAB Basic ruules are. or numerical simulations.3.1 Graphs .2 EGR 345 Specific .Provide a legend for surves with multiple traces . 31.

an expression. graph: a plot.31.writing guide .3 31.5 TERMINOLOGY The terminology is significant. numbers: a value equation: a function. a relationship. proof: a derivation .

quadratic equation . cosine. homogeneous.second-order differential equation . ellipses.1 INTRODUCTION • This section has been greatly enhanced.1 32. and tailored to meet our engineering requirements. but for reference.conics. angles and triple product .integration of volumes .integration of areas . small angle approximation . .maxima.triangles. non-homogeneous. conjugate. the following topics are commonly on the exam. etc. • This is not intended for learning. • We use math in almost every problem we solve. substitution . Andrew Sterian.integration .solving equations using inverse matrices. sine. polar form. determinants. adjoint. eigenvectors . inverse.math guide . minima and inflection points . second-order . • For those planning to write the first ABET Fundamentals of Engineering exam.solenoidal and conservative fields . . multiplication . MATHEMATICAL TOOLS ***** This contains additions and sections by Dr.complex numbers.32. . 32. As a result the more relevant topics of mathematics are summarized here.eigenvalues.straight line equations .first-order differential equations . addition of polar forms . etc.matrices. L’Hospital’s rule.divergence and curl .limits.dot and cross products.centroids . but it is designed to be a quick reference guide to support the engineer required to use techniques that may not have been used recently. cofactors.linear. Cramer’s rule.slop and perpendicular .by parts and separation . areas of parallelograms. • The section outlined here is not intended to teach the elements of mathematics. circles.guessing and separation .

partial fractions 32.basics.mean.2 .taylor series .trig identities . mode.permutations and combinations . chain rule. spherical . cylindrical. .1 Constants and Other Stuff • A good place to start a short list of mathematical relationships is with greek letters name lower case α β γ δ ε ζ η θ ι κ λ µ ν ξ ο π ρ σ τ υ φ χ ψ ω upper case Α Β Γ ∆ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω alpha beta gamma delta epsilon zeta eta theta iota kappa lambda mu nu xi omicron pi rho sigma tau upsilon phi chi psi omega .1.probability ..partial fractions . small angles approx.log properties . natural log. standard deviation.derivative . etc.cartesian.basic coordinate transformations .32.math guide .integration using Laplace transforms .

• Basic properties include. other values can be derived through calculation using modern calculators or computers. and are described in sufficient detail for our use.57721566 = Eulers constant 1radian = 57.7182818 = lim  1 + -- = natural logarithm bas e n n → ∞ π = 3.1.2 Some universal constants 32.29578° Figure 32.3 Basic algebra properties .1 The greek alphabet • The constants listed are amount some of the main ones.3 Figure 32.math guide . The values are typically given with more than 15 places of accuracy so that they can be used for double precision calculations. 1 n e = 2.1415927 = pi γ = 0. commutative distributive associative a+b = b+a a ( b + c ) = ab + ac a ( bc ) = ( ab )c a + ( b + c ) = ( a + b) + c Figure 32.2 Basic Operations • These operations are generally universal.32.

.1.3 Exponents and Logarithms • The basic properties of exponents are so important they demand some sort of mention ( x )( x ) = x (x ) ---------.1 .4 The basic factorial operator 32.5 Properties of exponents • Logarithms also have a few basic properties of use.1.2. n! = 1 ⋅ 2 ⋅ 3 ⋅ 4 ⋅ … ⋅ n 0! = 1 Figure 32.4 32.math guide .Factorial • A compact representation of a series of increasing multiples. if x is not 0 x –p 0 x = x m --n 1 -n n x x m 1= ---p x n n n = n = x n⋅m ( xy ) = ( x ) ( y ) n x = n x -----y n y Figure 32.= x n – m m (x ) (x ) n m n n m n+m x = 1 .32.

7 Properties of logarithms 32.5 The basic base 10 logarithm: log x = y The basic base n logarithm: logn x = y x = n y x = 10 y The basic natural logarithm (e is a constant with a value found near the start of this section: ln x = log e x = y x = e y Figure 32.math guide .32.6 Definitions of logarithms • All logarithms observe a basic set of rules for their application.4 Polynomial Expansions • Binomial expansion for polynomials. log n ( xy ) = log n ( x ) + log n ( y ) x log n  .1. . = logn ( x ) –log n ( y )  y logn ( x ) = ylog n ( x ) log m ( x ) log n ( x ) = -----------------log m ( n ) ln ( A ∠θ ) = ln ( A ) + ( θ + 2πk )j k∈I y log n ( n ) = 1 log n ( 1 ) = 0 Figure 32.

8 A general expansion of a polynomial 32.3  9  5 - 4 ------5 (y + 4 ) (y – 2) x y x+1 ----------.32.C+D C D log ( ab ) = log ( a ) + log ( b ) 5(5 ) = 5 4 5 3 log ( 4 ) = log ( 16 ) ( x + 6 ) ( x – 6 ) = x + 36 10 log ( 5 ) 10 = ----5 6 ( x + 1 ) .a x +…+x 2! Figure 32.math guide . Simplify the following expressions. a) b) c) d) e) f) g) x ( x + 2 ) – 3x 2 ( x + 3 ) ( x + 1 )x -------------------------------------( x + 1) x 3 log ( x ) 64 ----16 15 3 ----. Are the following expressions equivalent? a) b) c) d) e) f) g) h) A ( 5 + B ) – C = 5A + B – C A+B A B ------------.1.+ --.= x 2 + 2x + 1 -----------------(x + 1) 2 2 2.5 Practice Problems 1.-.+ ----21 28 (x y ) 2 2 3 4 4 2 2 h) i) 5  8 -.= --.6 ( a + x ) = a + na n n n–1 n n(n – 1) n – 2 2 x + ------------------.= 4 x+2 2 3 j) k) l) 4x – 8y .

Simplify the following expressions. Simplify the following expressions.  x2  d) log ( x 5 ) + log ( x 3 ) (ans.+ -----.= -----. A.+ -.32.AB AB AB B A b) -----------AB A + B cannot be simplified 4 5 c)  ( x y )  -------------.  x2  3 = (x y ) = x y 2 5 3 6 15 d) log ( x 5 ) + log ( x 3 ) = 5 log ( x ) + 3 log ( x ) = 8 log ( x ) 4.B 1 1 a) A + B -----------.math guide .7 3.= -. a) n log ( x ) + m log ( x ) – log ( x ) 2n log ( x ) + 3m log ( x ) – 4 log ( x ) ( 2n + 3m – 4 ) log ( x ) ( 2n + 3m – 4 ) log ( x ) 2 3 4 . a) n log ( x ) + m log ( x ) – log ( x ) 2 3 4 (ans. a) A + B -----------AB b) AB -----------A+B 4 5 3 c)  ( x y )  -------------.

= 5 3 t → 0 5t 3 + 1 5( 0 ) + 1 3 3  t3 + 5  ∞ +5 ∞ ----------------------.= x + 2 y+z (ans.8 5. y+x ---------. a)  t3 + 5  lim  ---------------- t → 0 5t 3 + 1  t3 + 5  lim  ---------------- t → ∞ 5t 3 + 1 b) (ans.32.= -------------. Find the limits below.2 lim  ---------------- = 3 3 t → ∞ 5t 3 + 1 5( ∞ ) + 1 5( ∞) . Rearrange the following equation so that only ‘y’ is on the left hand side.= 0. y+x ---------.math guide . a) b) 3  t3 + 5  0 +5 lim  ---------------- = ---------------------.= x + 2 y+z y + x = ( x + 2 )( y + z ) y + x = xy + xz + 2y + 2z y – xy – 2y = xz + 2z – x y ( – x – 1 ) = xz + 2z – x xz + 2z – x y = ------------------------–x–1 6.

2.9 32.32.1 Discrete and Continuous Probability Distributions Binomial P (m ) = Poisson P (m ) = t≤m ∑  t p q   n t –λ t n–t q = 1–p q.∫ e dt 2π – ∞ Normal Figure 32.10 Quadratic equation . r 2 = -------------------------------------2a 2 Figure 32. 1 ] λe ∑ -----------t! t≤m λ>0 Hypergeometric  r  s   t  n – t P ( m ) = ∑ ----------------------r + s t≤m   n  1 x –t2 P ( x ) = ---------.2.9 Distribution functions 32.2 FUNCTIONS 32.2 Basic Polynomials • The quadratic equation appears in almost every engineering discipline. p ∈ [ 0. ax + bx + c = 0 = a ( x – r 1 ) ( x – r 2 ) 2 – b ± b – 4ac r 1.math guide . therefore is of great importance.

find the roots of the 2 equations below.10 • Cubic equations also appear on a regular basic. z +  ----------------------------.( S – T ) 2 3 2 2 3 2 9ab – 27c – 2a R = -------------------------------------54 3 S = 3 R+ Q +R 3 2 T = 3 R– Q +R 3 2 Figure 32. calculate.math guide .11 Cubic equations • On a few occasions a quartic equation will also have to be solved. 3b – a Q = ----------------9 Then the roots. = 0 2 2     3 2 2 2 4 3 2 Figure 32. y – by + ( ac – 4d )y + ( 4bd – c – a d ) = 0 Next. a r 1 = S + T – -3 S+T a j 3 .( S – T ) 2 3 2 S+T a j 3 . x + ax + bx + cx + d = 0 = ( x – r1 ) ( x – r 2 ) ( x – r 3 ) ( x – r 4 ) First.32. solve the equation below to get a real root (call it ‘y’). x + ax + bx + c = 0 = ( x – r 1 ) ( x – r 2 ) ( x – r 3 ) First. and as a result should also be considered.+ -------.r 2 = ----------.12 Quartic equations .r 3 = ----------. This can be done by first reducing the equation to a quadratic.– -. r 2 = z +  ------------------------------------------- z +  ------------------------------ = 0 2 2     2 2  y – y 2 – 4d 2  a – a – 4b + 4y r 3.– -------.– -. r 4 = z +  -----------------------------------------.  a + a 2 – 4b + 4y  y + y 2 – 4d r 1.

32. If there are higher order roots (repeated terms) then derivatives will be required to find solutions use algebra technique Done Figure 32.13 The methodolgy for solving partial fractions • The partial fraction expansion for.3 Partial Fractions • The next is a flowchart for partial fraction expansions.math guide .2. . start with a function that has a polynomial numerator and denominator is the order of the numerator >= denominator? yes use long division to reduce the order of the numerator no Find roots of the denominator and break the equation into partial fraction form with unknown values OR use limits technique.11 32.

  2  s→0 s (s + 1) = 1 1 = lim ---------.math guide .  2  s → –1 s (s + 1) 2 1 A = lim s  -------------------.= ---.12 A B 1 Cx ( s ) = -------------------. s → 0 ds  s + 1 = lim [ – ( s + 1 ) ] = –1 s→0 –2 d.+ -.32.1 = lim ---.  2  s → 0 ds s ( s + 1) Figure 32.2 1 B = lim ---. .= 1 s→0 s +1 d.s  -------------------.14 A partial fraction example • Consider the example below where the order of the numerator is larger than the denominator. ---------.+ ---------2 2 s s+1 s (s + 1) s 1 C = lim ( s + 1 )  -------------------.

– 12s – 6 x ( s ) = 5s + 3 + --------------------2 s +4 solve – 12s – 6 A B --------------------.= ---.math guide .= ------------.15 Solving partial fractions when the numerator order is greater than the denominator • When the order of the denominator terms is greater than 1 it requires an expanded partial fraction form.+ B + -----------------.+ --------------3 2 3 2 (s + 1) 2 s (s + 1 ) (s + 1) s s (s + 1) Figure 32.13 5s + 3s + 8s + 6 x ( s ) = ------------------------------------------2 s +4 This cannot be solved using partial fractions because the numerator is 3rd order and the denominator is only 2nd order.16 Partial fractions with repeated roots . Therefore long division can be used to reduce the order of the equation. 5s + 3 s +4 2 3 2 5s + 3s + 8s + 6 3 5s + 20s 3s – 12s + 6 2 3s + 12 2 3 2 – 12s – 6 This can now be used to write a new function that has a reduced portion that can be solved with partial fractions.+ -----------------. 5 F ( s ) = ----------------------3 2 s (s + 1 ) 5 A -C D E ----------------------.+ -----------2 s + 2j s – 2j s +4 Figure 32. as shown below.32.

b ≥ .14 • We can solve the previous problem using the algebra technique. 5 A -C D E ----------------------.17 An algebra solution to partial fractions 32.+ -----------------.+ --------------3 2 3 2 (s + 1 ) 2 s s (s + 1) (s + 1) s (s + 1) Figure 32.32.+ --------------3 2 3 2 (s + 1) 2 s (s + 1) s (s + 1 ) s (s + 1) A ( s + 1 ) + Bs ( s + 1 ) + Cs + Ds ( s + 1 ) + Es ( s + 1 ) = ----------------------------------------------------------------------------------------------------------------------------------------3 2 s (s + 1) s ( B + E ) + s ( A + 3B + D + 2E ) + s ( 3A + 3B + C + D + E ) + s ( 3A + B ) + ( A ) = -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------3 2 s (s + 1) 0 1 3 3 1 1 3 3 1 0 0 0 1 0 0 0 1 1 0 0 1 2 1 0 0 A 0 B 0 C = 0 D 0 E 5 A 0 1 B 1 3 C = 3 3 D 3 1 E 1 0 0 0 1 0 0 0 1 1 0 0 1 2 1 0 0 –1 4 3 2 3 3 2 2 2 2 0 5 0 – 15 0 = 5 0 10 5 15 5 5 -------5 10 15 ----------------------.= ---.2.math guide .+ -----------------.= ---.aThe index variable is a placeholder whose name does not i • Operations on summations: b a i=a ∑ xi = i=b ∑ xi .+ B + -----------------.+ – 15 + -----------------.4 Summation and Series b • The notation is equivalent to ∑ xi i=a x a + x a + 1 + xa + 2 + … + xb assuming a and b integers and are matter.

15 ∑ i=a b b αx i = α ∑ x i i=a b b i=a b ∑ xi + ∑ yj j=a = i=a ∑ ( xi + yi ) ∑ xi d c b i=a ∑ xi + ∑ c xi = i = b+1 i=a b  b  d   ∑ x i  ∑ y j = i = a  j = c  ∑ ∑ x i yj i=a j=c • Some common summations: N ∑i i=1 N–1 1 = -. α < 1 for both real and complex α . the summation 1–α does not converge. α ≠ 1 α =  1–α for both real and complex α . For α ≥ 1 . ∑  i=0 N.N ( N + 1 ) 2  1 – αN  ---------------.32.math guide . .. α = 1  i i=0 ∑α ∞ i 1 = ----------.

2x + 8x = –8 x + 4x + 4 = 0 2 ( x + 2) = 0 x = – 2.2. e. x + 4---------------2 x + 8x (ans.5 x+4 ---------------. Solve the following equation to find ‘x’.= Ax + 8A + Bx --------------------------------2 2 x(x + 8 ) x x+8 x + 8x x + 8x ( 1 )x + 4 = ( A + B )x + 8A 8A = 4 A = 0.32. x+4 Bx + 4---------------..math guide .= 0.= ------------------.0. Convert the following polynomials to multiplied terms as shown in the example. Reduce the following expression to partial fraction form. a) b) c) x + 2x + 1 = ( x + 1 ) ( x + 1 ) x – 2x + 1 x –1 x +1 2 2 2 2 d) e) f) x + x + 10 2 2.= A + ----------. –2 2 2 3.g.5 1 = A+B B = 0.5 2 x x+8 x + 8x .5 + ---------------.16 32. 2x + 8x = – 8 2 (ans.5 Practice Problems 1.

= ---------r csc θ x 1 cos θ = .32.= ---------.17 32.sin 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° .3 SPATIAL RELATIONSHIPS 32.1 Trigonometry • The basic trigonometry functions are. Sine .= ----------r sec θ y 1sin θ tan θ = . y 1sin θ = .= ----------x cot θ cos θ Pythagorean Formula: r = x +y 2 2 2 r y θ x • Graphs of these functions are given below.math guide .3.

32.tan 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° Cosecant .csc 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° .cos 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° Tangent .math guide .18 Cosine .

32.math guide .sec 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° Cotangent . that any value that does not lie in the right hand quadrants of cartesian space. may need additions of ±90° or ±180°.cot 1 -270° -180° -90° 0° -1 90° 180° 270° 360° 450° • NOTE: Keep in mind when finding these trig values.19 Secant . .

These are often best used when attempting to manipulate equations.= ------------.math guide .20 Cosine Law: c = a + b – 2ab cos θ c c Sine Law: a b c ------------.32. .= ------------sin θ A sin θ B sin θ C θB θC θA b 2 2 2 a • Now a group of trigonometric relationships will be given.

0 0.= -------------------2 1 + cos θ sin θ ( cos θ ) + ( sin θ ) = 1 2 2 OR OR sin ( 2θ ) = 2 sin θ cos θ cos ( 2θ ) = ( cos θ ) + ( sin θ ) 2 2 -ve if in left hand quadrants • Numerical values for these functions are given below. θ (deg) -90 -60 -45 -30 0 30 45 60 90 sin θ -1.866 0.5 0 0.32.5 0. sin ( θ 1 ± θ 2 ) = sin θ 1 cos θ 2 ± cos θ 1 sin θ 2 cos ( θ 1 ± θ 2 ) = cos θ 1 cos θ 2 − sin θ 1 sin θ 2 + tan θ 1 ± tan θ 2 tan ( θ 1 ± θ 2 ) = -----------------------------------− 1 + tan θ 1 tan θ 2 cot θ 1 cot θ 2 − 1 + cot ( θ 1 ± θ 2 ) = -----------------------------------tan θ 2 ± tan θ 1 θ sin -.866 -0.5 0.707 -0.0 tan θ -infinity -1 0 1 infinity .707 0.0 cos θ 0.= ± 1 + cos θ -------------------2 2 θ sin θ 1 – cos θ tan -.0 -0.math guide .= ± 1 – cos θ -------------------2 2 θ cos -.866 1.21 sin ( – θ ) = – sin θ cos ( – θ ) = cos θ tan ( – θ ) = – tan θ sin θ = cos ( θ – 90° ) = cos ( 90° – θ ) = etc.5 0.= -------------------.707 0.866 1 0.707 0.

e +e cos θ = --------------------2 jθ –jθ e –e sin θ = --------------------2j jθ –j θ 32.= ----------------.= ----------------. .math guide .= ----------------.3.= hyperbolic secant of x x –x cosh ( x ) e +e cosh ( x ) e +e coth ( x ) = -----------------.32.= hyperbolic tangent of x x –x cosh ( x ) e +e 1 2 csch ( x ) = ----------------.= ----------------.= hyperbolic sine of x 2 e +e cosh ( x ) = ----------------.22 • These can also be related to complex exponents. e –e sinh ( x ) = ----------------.= hyperbolic cosine of x 2 sinh ( x ) e –e tanh ( x ) = -----------------.2 Hyperbolic Functions • The basic definitions are given below.= hyperbolic cotangent of x x –x sinh ( x ) e –e x –x x –x x –x x –x • some of the basic relationships are.= hyperbolic cosecant of x x –x sinh ( x ) e –e 1 2 sech ( x ) = -----------------.

2.23 sinh ( – x ) = – sinh ( x ) cosh ( – x ) = cosh ( x ) tanh ( – x ) = – tanh ( x ) csch ( –x ) = – csch ( x ) sech ( –x ) = sech ( x ) coth ( – x ) = – coth ( x ) • Some of the more advanced relationships are.1 .math guide .32. ( cosh x ) – ( sinh x ) = ( sech x ) + ( tanh x ) = ( coth x ) – ( csch x ) = 1 sinh ( x ± y ) = sinh ( x ) cosh ( y ) ± cosh ( x ) sinh ( y ) cosh ( x ± y ) = cosh ( x ) cosh ( y ) ± sinh ( x ) sinh ( y ) tanh ( x ) ± tanh ( y )tanh ( x ± y ) = ---------------------------------------------1 ± tanh ( x ) tanh ( y ) • Some of the relationships between the hyperbolic. and normal trigonometry functions are.Practice Problems . 2 2 2 2 2 2 sin ( jx ) = j sinh ( x ) cos ( jx ) = cosh ( x ) tan ( jx ) = j tanh ( x ) j sin ( x ) = sinh ( jx ) cos ( x ) = cosh ( jx ) j tan ( x ) = tanh ( jx ) 32.3.

= Ax + 2A + Bx + B = ( 2A + B ) + ( A + B )x -----------------------------------------------------------------------------------------2 2 2 x+1 x+2 x + 3x + 2 x + 3x + 2 x + 3x + 2 A+B = 0 A = –B B = –4 A = 4 4 –4 ----------.+ ----------.5 (ans.= 2 x + 3x + 2 1 Note: there was a typo here.3 Geometry • A set of the basic 2D and 3D geometric primitives are given.+ ----------x+1 x+2 2A + B = 4 = – 2B + B = – B 32. cos θ cos θ – sin θ sin θ = cos ( θ + θ ) = cos ( 2θ ) ( s + 3j ) ( s – 3j ) ( s + 2j ) = ( s – 9j ) ( s + 4js + 4j ) s + 4js + 4j s – 9j s – 9j 4js – 9j 4j s + ( 4j )s + ( 5 )s + ( 36j )s + ( –36 ) 4 3 2 4 3 2 2 2 2 2 2 2 2 2 2 2 2 3. cos θ cos θ – sin θ sin θ = ( s + 3j ) ( s – 3j ) ( s + 2j ) = 2 (ans.3. 5 3 10° 5 3 2. Simplify the following expressions.= ----------. Solve the following partial fraction 4 -------------------------.math guide . 4 AB-------------------------. so --------------an acceptable answer is also.32. x + 0. and the notation used is described below.24 1. Find all of the missing side lengths and corner angles on the two triangles below. .

z = centre of mass x. y. y.32.= centroid location along the y-axis dA ∫ A ∫ y dA .25 A = contained area P = perimeter distance V = contained volum e S = surface are a x. I y. z = centroid Ix.math guide . I z = moment of inertia of area (or second moment of inertia) AREA PROPERTIES: Ix = A ∫y ∫x A 2 dA = the second moment of inertia about the y-axis Iy = 2 dA = the second moment of inertia about the x-axis A I xy = ∫ xy dA A = the product of inertia 2 2 JO = ∫r 2 dA = A ∫ (x + y ) dA = Ix + Iy = The polar moment of inertia x = A -----------A ∫ x dA ∫ = centroid location along the x-axis dA A y = -----------.

32.26 Rectangle/Square: A = ab P = 2a + 2b y a x b Moment of Inertia (about origin axes): ba Ix = -------3 b a Iy = -------3 Ixy b a= ---------4 2 2 3 3 Centroid: b x = -2 a y = -2 Moment of Inertia (about centroid axes): ba I x = -------12 b a I y = -------12 I xy = 0 3 3 .math guide .

32.( 2a – b ) 24 2 Circle: A = πr 2 y r P = 2πr x Moment of Inertia Moment of Inertia (about centroid axes): (about origin axes): 4 πr I x = ------Ix = 4 πr I y = ------4 I xy = 0 4 Centroid: x = r y = r Mass Moment of Inertia (about centroid): Mr J z = --------2 2 Iy = I xy = .( a + b – ab ) 36 3 Moment of Inertia (about origin axes): bh Ix = -------12 2 2 I y = bh ( a + b – ab ) ----12 3 bh I xy = -------.( 2a – b ) 72 2 bh I xy = -------.math guide .27 Triangle: bh A = ----2 P = y a θ h x b Centroid: a+b x = ----------3 h y = -3 Moment of Inertia (about centroid axes): bh Ix = -------36 2 bh 2 I y = ----.

01647 r 4 Ixy r = ---8 4 . r  8 9π πr I y = ------8 I xy = 0 4 Moment of Inertia (about origin axes): πr I x = ------8 πr I y = ------8 I xy = 0 4 4 Quarter Circle: πr A = ------4 πr P = ----.28 Half Circle: πr A = ------2 P = πr + 2r 2 y r x Centroid: x = r 4r y = ----3π Moment of Inertia (about centroid axes): π 8.05488r 4 Moment of Inertia (about origin axes): 4 πr Ix = ------16 πr Iy = ------16 4 4 Ixy = – 0.math guide .05488r I y = 0.4 I x =  -.+ 2r 2 2 y r x Centroid: 4r x = ----3π 4r y = ----3π Moment of Inertia (about centroid axes): I x = 0.32.– ----.

( θ + sin θ ) 8 I xy = 0 4 4 .( θ – sin θ ) 8 r Iy = ---.32.29 Circular Arc: θr A = ------2 P = θr + 2r 2 y r x θ Centroid: θ 2r sin -2 x = ---------------3θ y = 0 Moment of Inertia (about centroid axes): Ix = Iy = I xy = Moment of Inertia (about origin axes): r I x = ---.math guide .

( sin θ ) dθ a 2 2 2 x r1 + r 2 P ≈ 2π --------------2 Centroid: x = r2 y = r1 Moment of Inertia Moment of Inertia (about centroid axes): (about origin axes): πr 1 r 2 Ix = -----------4 3 πr 1 r 2 Iy = -----------4 Ixy = 3 2 Ix = Iy = Ixy = .math guide .32.30 Ellipse: A = πr 1 r 2 π -2 y r1 r2 P = 4r 1 ∫ 0 r1 + r2 2 1 – ------------------.

+ 2r 2 2 Centroid: x = r2 4r 1 y = ------3π Moment of Inertia (about centroid axes): Ix = 3 0.05488r 2 r 1 3 Moment of Inertia (about origin axes): πr 2 r 1 Ix = -----------16 3 πr 2 r 1 Iy = -----------16 2 2 r1 r2 I xy = --------8 3 Iy = 0.( sin θ ) dθ + 2r 2 a 2 2 r1 2 r2 r1 r2 x r1 + r2 P ≈ π --------------.31 Half Ellipse: πr 1 r 2 A = -----------2 P = 2r 1 ∫ 2 π -2 0 y + 2 1 – ------------------.01647r 1 r 2 2 2 .32.05488r 2 r 1 Ixy = – 0.math guide .

+ ----.ln  4a + b + 16a - P = 2 8a  b  2 2 2 y a x b Centroid: b x = -2 2a y = ----5 Moment of Inertia (about centroid axes): Ix = Iy = I xy = Moment of Inertia (about origin axes): Ix = Iy = Ixy = . --------------------------------------.32 Quarter Ellipse: πr 1 r 2 A = -----------4 r1 ∫ 2 π -2 0 y P = + 2 1 – ------------------. --------------------------.32.( sin θ ) dθ + 2r 2 a 2 2 r1 2 r2 r1 r2 x π r1 + r2 P ≈ -.--------------.math guide .+ 2r 2 2 2 Centroid: 4r 2 x = ------3π 4r 1 y = ------3π Moment of Inertia (about centroid axes): Ix = Iy = Ixy = Moment of Inertia (about origin axes): I x = πr 2 r 1 I y = πr 2 r 1 I xy r2r1 = --------8 2 2 3 3 Parabola: A = 2 ab -3 2 2 b + 16a b .

This for is shown below.+ -------. 4 16a  b  2 2 2 2 2 y a x b Centroid: 3b x = ----8 2a y = ----5 Moment of Inertia Moment of Inertia (about centroid axes): (about origin axes): 3 3 8ba 2ba I x = ----------Ix = ----------175 7 19b a Iy = -------------480 Ixy b a= ---------60 2 2 3 2b a Iy = ----------15 Ixy b a= ---------6 2 2 3 • A general class of geometries are conics.math guide .33 Half Parabola: ab A = ----3 b + 16a b .32. 4a + b + 16a P = --------------------------.ln  --------------------------------------. and can be used to represent many of the simple shapes represented by a polynomial. .

32.34 Ax + 2Bxy + Cy + 2Dx + 2Ey + F = 0 Conditions A = B = C = 0 B = 0. A = C 2 B – AC < 0 2 B – AC = 0 2 B – AC > 0 straight line circle ellipse parabola hyperbola 2 2 .math guide .

= centroid location along the z-axis dV ∫ V ∫ z dV .32.35 VOLUME PROPERTIES: Ix = V ∫ rx ∫ ry V 2 dV = the moment of inertia about the x-axis Iy = Iz = 2 dV = the moment of inertia about the y-axis V ∫ r z dV = the moment of inertia about the z-axis 2 V x = -----------.= centroid location along the x-axis dV ∫ V ∫ x dV V y = -----------.= centroid location along the y-axis dV ∫ V ∫ y dV V z = -----------.math guide .

36 Parallelepiped (box): V = abc S = 2 ( ab + ac + bc ) b c y z x a Centroid: a x = -2 b y = -2 c z = -2 Moment of Inertia (about centroid axes): M( a + b ) I x = -------------------------12 2 2 M (a + c ) Iy = -------------------------12 M( b + a ) I z = -------------------------12 2 2 2 2 Moment of Inertia (about origin axes): Ix = Iy = Iz = Mass Moment of Inertia (about centroid): Jx = Jy = Jz = Sphere: 4 3 V = -.πr 3 S = 4πr 2 y r z x Moment of Inertia (about centroid axes): 2Mr Ix = -----------5 2Mr Iy = -----------5 2Mr Iz = -----------5 2 2 Centroid: Moment of Inertia (about origin axes): Ix = Iy = Iz = Mass Moment of Inertia (about centroid): 2Mr J x = -----------5 2Mr J y = -----------5 2Mr J z = -----------5 2 2 2 x = r y = r z = r 2 .32.math guide .

2 Iz = -------.math guide .2 Ix = -------.πh ( 3r – h ) 3 y h z r S = 2πrh x Centroid: Moment of Inertia (about centroid axes): Ix = Iy = Iz = Moment of Inertia (about origin axes): Ix = Iy = Iz = x = r y = z = r .32.Mr 320 2Mr I y = -----------5 83.Mr 320 2 Moment of Inertia (about origin axes): Ix = Iy = Iz = x = r 3r y = ---8 z = r Cap of a Sphere: 1 2 V = -.πr 3 y z S = r x Centroid: Moment of Inertia (about centroid axes): 83.37 Hemisphere: 2 3 V = -.

+ ----  12 4  2 2 2 Mass Moment of Inertia (about centroid): M ( 3r + h ) J x = ----------------------------12 Mr J y = --------2 2 2 2 x = r h y = -2 z = r h.math guide .+ -------  80 20  3Mr I y = -----------10 3h 3r Iz = M  -------.r I z = M  ----.38 Cylinder: V = hπr 2 y r h S = 2πrh + 2πr 2 z x Moment of Inertia (about origin axis): h.r Ix = M  ---.+ ----  12 4  Mr Iy = --------2 h.r Iz = M  ---.πr h 3 S = πr r + h 2 2 y z h r x Centroid: x = r h y = -4 z = r Moment of Inertia (about centroid axes): 3h 3r Ix = M  -------.+ ----  3 4 Iy = 2 2 2 2 Centroid: Moment of Inertia (about centroid axis): h.+ ------.+ ----  3 4 2 2 M ( 3r + h ) J z = ----------------------------12 2 2 Cone: 1 2 V = -.32.  80 20  3 2 2 3 2 Moment of Inertia (about origin axes): Ix = Iy = Iz = .r Ix = M  ----.

39 Tetrahedron: V = 1 Ah -3 y h A Centroid: Moment of Inertia (about centroid axes): Ix = Iy = Iz = z x Moment of Inertia (about origin axes): Ix = Iy = Iz = x = h y = -4 z = Torus: 2 1 2 V = -.math guide .π ( r 1 + r 2 ) ( r 2 – r 1 ) 4 y r2 z r1 x Moment of Inertia (about centroid axes): Ix = Iy = Iz = Moment of Inertia (about origin axes): Ix = Iy = Iz = S = π ( r2 – r1 ) 2 2 2 Centroid: x = r2 r2 – r1 y =  -------------.32.  2 - z = r2 .

math guide .32.πr 1 r 2 r 3 3 y r2 z r3 r1 x Moment of Inertia (about centroid axes): Ix = Iy = Iz = Moment of Inertia (about origin axes): Ix = Iy = Iz = S = Centroid: x = r1 y = r2 z = r3 Paraboloid: 1 2 V = -.40 Ellipsoid: 4 V = -.πr h 2 y h z r x S = Centroid: Moment of Inertia (about centroid axes): Ix = Iy = Iz = Moment of Inertia (about origin axes): Ix = Iy = Iz = x = r y = z = r .

a b defined with a slope and intercept a slope perpendicular to a line the slope using two points as defined by two intercepts • If we assume a line is between two points in space.4 Planes. and that at one end we have a local reference frame.math guide . • The most fundamental mathematical geometry is a line.3. there are some basic relationships that can be derived. The basic relationships are given below. y = mx + b 1 m perpendicular = --m y2 – y 1 m = --------------x2 – x 1 x y -. .41 32.32. etc. Lines.+ -.= 1 .

z 1 ) d = ( x2 – x1 ) + ( y2 – y1 ) + ( z2 – z1 ) 2 2 2 d ( x 2.= --------------. y 2. x – x1 y – y1 z – z1 --------------.= -------------x2 – x1 y2 – y1 z 2 – z1 ( x. y. y 1.  d - 2 2 ( x 0.42 y θβ ( x 1. x2 – x1 θ α = acos  --------------. . z 1 ) ) Explicit Parametric t=[0. z ) = ( x 1.  d - ( cos θ α ) + ( cos θ β ) + ( cos θ γ ) = 1 The equation of the line is. y 0.math guide . z 1 ) + t ( ( x 2. z 2 ) – ( x 1. y 2. z 0 ) y2 – y1 θ β = acos  --------------.1] • The relationships for a plane are. z 2 ) θγ θα x z The direction cosines of the angles are.  d - 2 z2 – z1 θ γ = acos  -------------.32. y 1. y 1.

1 A = -a 1 B = -b 1 C = -c D = –1 N P2 = ( x 2. y 3. z ) = t ( A. z 1 ) where the coefficients defined by the intercepts are. x – x1 y – y1 z – z 1 det x – x 2 y – y 2 z – z 2 = 0 x – x3 y – y3 z – z 3 ∴det y2 – y1 z2 – z 1 y3 – y1 z3 – z 1 ( x – x 1 ) + det z2 – z1 x2 – x1 z3 – z1 x3 – x1 ( y – y1 ) c a x b + det x2 – x1 y2 – y1 x3 – x1 y3 – y1 ( z – z1 ) = 0 The normal to the plane (through the origin) is. z 2 ) P 3 = ( x 3.32. What is the circumferenece of a circle? What is the area? What is the ratio of the area to the circumference? . ( x. B. C ) 32. y 2.5 Practice Problems 1.43 y The explicit equation for a plane is.3.math guide . z 3 ) z The determinant can also be used. Ax + By + Cz + D = 0 P 1 = ( x 1. y. y 1.

0) to (2.32. 2) b) (1. a) x 2 + 3x = – 2 x + 3x + 2 = 0 – 3 ± 3 – 4( 1 )( 2 ) –3± 9–8 –3±1 x = -----------------------------------------------. a) ( cos 2θ ) sin 2θ  ---------------------.= –1. Find a line that is perpendicular to the line through the points (2.math guide .= 1 cos x tan x = 1 x = atan 1 x = …. 2). 225°. 45°. … 5. 4) to (2. 1) and (1. a) x 2 + 3x = – 2 b) sin x = cos x (ans.= ---------------. Manipulate the following equations to solve for ‘x’.= ----------------------------. What is the equation of a line that passes through the points below? a) (0. – 135°. 1) c) (3.+ sin 2θ = ( cos 2θ ) + ( sin 2θ ) = 1  sin 2θ  2 . The perpendicular line passes through (3.+ sin 2θ  sin 2θ  2 2 2 (ans. 5). 9) 3. Simplify the following expressions. 4.44 2. 0) to (0. a) ( cos 2θ ) 2 2 sin 2θ  --------------------. – 2 2(1 ) 2 2 b) sin x = cos x sin x ---------.

given.5 m y = – 0.32. ‘j’ will be the preferred notation for the complex number.math guide .= – 0.4 COORDINATE SYSTEMS 32. A line that passes through the point (1. y = mx + b m = 2 x = 1 y = 2 substituting 2 = 2(1) + b y = 2x perpendicular 1 m perp = – --. Find the equation for the line. 2) and has a slope of 2.1 Complex Numbers • In this section. (ans.4. . this is to help minimize confusion with the ‘i’ used for current in electrical engineering.45 6.5 x b = 0 32. • The basic algebraic properties of these numbers are. and for a line perpendicular to it. as in all others.

46 The Complex Number: j = Complex Numbers: a + bj where. a and b are both real numbers –1 j = –1 2 Complex Conjugates (denoted by adding an asterisk ‘*’ the variable): N = a + bj Basic Properties: ( a + bj ) + ( c + dj ) = ( a + c ) + ( b + d )j ( a + bj ) – ( c + dj ) = ( a – c ) + ( b – d )j ( a + bj ) ⋅ ( c + dj ) = ( ac – bd ) + ( ad + bc )j N a + bj N --------.math guide .------------. please consider the notation above uses a cartesian notation. but a polar notation can also be very useful when doing large calculations.= ---. N* = a – bj .---------------------------2 2  c + dj  c – dj  2 2 M c + dj M  N* c +d c +d • We can also show complex numbers graphically.= ------------.32. These representations lead to alternative representations. N* =  a + bj  c – dj = ac + bd +  bc – ad j -----------------. If it in not obvious above.

whereas cartesian notation is easier for addition and subtraction).  R POLAR FORM Ij imaginary (j) N = A ∠θ θ R real A = amplitude θ = phase angl e • We can also do calculations using polar notation (this is well suited to multiplication and division.32.47 CARTESIAN FORM Ij imaginary (j) N = R + Ij R real A = R +I 2 2 R = A cos θ I = A sin θ Iθ = atan  -. .math guide .

32.=  ----- ∠( θ 1 – θ 2 ) A  ( A2 ∠θ 2 ) 2 ( A ∠θ ) = ( A ) ∠( nθ ) n n (DeMoivre’s theorem) • Note that DeMoivre’s theorem can be used to find exponents (including roots) of complex numbers jθ • Euler’s formula: e = cos θ + j sin θ • From the above.math guide . . the following useful identities arise: jθ – jθ e +e cos θ = --------------------2 e –e sin θ = --------------------2j jθ – jθ 32.48 A ∠θ = A ( cos θ + j sin θ ) = Ae jθ ( A 1 ∠θ 1 ) ( A2 ∠θ 2 ) = ( A1 A2 ) ∠( θ 1 + θ 2 ) ( A1 ∠θ 1 ) A1 -------------------.2 Cylindrical Coordinates • Basically. these coordinates appear as if the cartesian box has been replaced with a cylinder.4.

3 Spherical Coordinates • This system replaces the cartesian box with a sphere.math guide .49 z ( x. y. θ.4. . z ) ↔ ( r.32. z ) z z z y x x y θ r = x +y 2 2 r x = r cos θ y = r sin θ θ = atan  y  x 32.

a) b) c) 16 -------------------2 ( 4j + 4 ) 3j + 5 -------------------2 ( 4j + 3 ) ( 3 + 5j )4j where. z ) ↔ ( r.  x φ = acos  z  r 32.4. y. j = –1 .4 Practice Problems 1. θ.50 z ( x.32. Simplify the following expressions. φ ) z φ r z y x x y 2 θ r = x +y +z 2 2 x = r sin θ cos φ y = r sin θ sin φ z = r cos θ y θ = atan  .math guide .

= – 0. head terminus A vector is said to have magnitude (length or strength) and direction.1 Vectors • Vectors are often drawn with arrows. a) 16 1616 -------------------.math guide .= ------.= --------------------------------. y a) find the area of the shape. b) find the centroid of the shape.5 MATRICES AND VECTORS 32. as shown below.5. origin tail . y = (x + 2 ) 2 x 4 32.= --------------------.32.5j 2 – 16 + 32j + 16 32j ( 4j + 4 ) 3j + 5 3j + 5 3j + 5 -------------------b) -------------------. c) find the moment of inertia of the shape about the centroid. – 7 – 24j = – 35 – 141j + 72 = 37 – 141j -----------------------------------------------------------2 – 16 + 24j + 9 49 + 576 625 – 7 + 24j  – 7 – 24j ( 4j + 3 ) c) ( 3 + 5j )4j = 12j + 20j 2 = 12j – 20 2.= -----------------------------------. For the shape defined below.51 (ans.

4 ) B = ( 7. 4 + 9 ) A – B = ( 2 – 7.5. numerically and graphically. y x z becomes y k j i j z x k i • Vectors can be added and subtracted. 3 – 8.52 • Cartesian notation is also a common form of usage. 9 ) A + B = ( 2 + 7. 4 – 9 ) B A+B A B Parallelogram Law A 32. A = ( 2. 3. 3 + 8. 8.32.math guide .2 Dot (Scalar) Product • We can use a dot product to find the angle between two vectors .

= -------------------.= 0.47 ) ( 6 ) θ F 2 = 5i + 3j F1 = 2i + 4j x • We can use a dot product to project one vector onto another vector. z We want to find the component of force F1 that projects onto the vector V.math guide .53 y F1 • F2 cos θ = ----------------F1 F2  (2)(5 ) + (4)(3 )  ∴θ = acos  -----------------------------------------. the component we find will be multiplied by the magnitude of V. To do this we first convert V to a unit vector. x 1j + 1k Vλ V = ----.707j + 0.  2 2 + 4 2 5 2 + 3 2 22 ∴θ = acos  ---------------------.707k V 2 2 1 +1 F1 V = λ V • F1 = ( 0.5°  ( 4.707 ) ( 4 ) + ( 0. = 32.707k ) • ( – 3i + 4j + 5k )N ∴F1V = ( 0 ) ( – 3 ) + ( 0.32.707j + 0. .707 ) ( 5 ) = 6N F 1 = ( – 3i + 4j + 5k )N V = 1j + 1k y F1 V F 1V • We can consider the basic properties of the dot product and units vectors. if we do not.

and the displaced x-y axes shown below as x’-y’. . But.54 Unit vectors are useful when breaking up vector magnitudes and direction. λ x = 1i + 0j (unit vector for the x-axis) F x = λ x • F = ( 1i + 0j ) • [ ( 10 cos 60° )i + ( 10 sin 60° )j ] ∴ = ( 1 ) ( 10 cos 60° ) + ( 0 ) ( 10 sin 60° ) = 10N cos 60° This result is obvious. but consider the other obvious case where we want to project a vector onto itself.math guide . and hence they are not unit vectors. x axis = 2i y axis = 3j x‘ axis = 1i + 1j y‘ axis = – 1i + 1j F = 10N ∠60° = ( 10 cos 60° )i + ( 10 sin 60° )j None of these vectors has a magnitude of 1. y y’ F = 10N x’ 45° 60° x We could write out 5 vectors here. if we find the equivalent vectors with a magnitude of one we can simplify many tasks. relative to the x-y axis. As an example consider the vector.32. In particular if we want to find the x and y components of F relative to the x-y axis we can use the dot product.

Not using a unit vector FF = F • F = ( ( 10 cos 60° )i + ( 10 sin 60° )j ) • ( ( 10 cos 60° )i + ( 10 sin 60° )j ) = ( 10 cos 60° ) ( 10 cos 60° ) + ( 10 sin 60° ) ( 10 sin 60° ) = 100 ( ( cos 60° ) + ( sin 60° ) ) = 100 Using a unit vector FF = F • λF = ( ( 10 cos 60° )i + ( 10 sin 60° )j ) • ( ( cos 60° )i + ( sin 60° )j ) = ( 10 cos 60° ) ( cos 60° ) + ( 10 sin 60° ) ( sin 60° ) Correct Now consider the case where we find the component of F in the x’ direction. u x' = cos 45°i + sin 45°j F x' = F • λ x' = ( ( 10 cos 60° )i + ( 10 sin 60° )j ) • ( ( cos 45° )i + ( sin 45° )j ) = ( 10 cos 60° ) ( cos 45° ) + ( 10 sin 60° ) ( sin 45° ) = 10 ( cos 60° cos 45° + sin 60° sin 45° ) = 10 ( cos ( 60° – 45° ) ) Here we see a few cases where the dot product has been applied to find the vector projected onto a unit vector.32.= -------------------------------------------------------.math guide . this can be done using the dot product to project F onto a unit vector.55 10 cos 60°i + 10 sin 60°j Fλ F = ----. = 10 ( ( cos 60° ) + ( sin 60° ) ) = 10 2 2 2 2 . Again. Now finally consider the more general case.= cos 60°i + sin 60°j F 10 Incorrect .

• -------.32. we can manipulate this expression into the dot product form.56 y V2 V1 θ2 θ1 x V2 V1 First.math guide . . V 2V1 = V2 cos ( θ 2 – θ 1 ) Next. = V2 ( cos θ 1 cos θ 2 + sin θ 1 sin θ 2 ) = V 2 [ ( cos θ 1 i + sin θ 1 j ) • ( cos θ2 i + sin θ 2 j ) ] V1 • V2 V 1 V2 V1 • V2 = V 2 -------.= V2 ----------------. V1 • V2 V2V 1 = V2 cos ( θ 2 – θ 1 ) = V2 ----------------V1 V2 V1 • V2 ∴ V2 cos ( θ 2 – θ 1 ) = V2 ----------------V1 V 2 V 1 • V2 ∴ cos ( θ 2 – θ 1 ) = ----------------V1 V 2 *Note that the dot product also works in 3D. and similar proofs are used. we can see that the component of V2 (projected) in the direction of V1 will be.= V 2 • λ V1 V1 V1 V2 V 1 V2 Or more generally.= ----------------. by inspection.

66j + 0k )N d = ( 2i + 0j + 0k )m i M = d×F = j k NOTE: note that the cross product here is for the right hand rule coordinates. . F = ( – 6.43N ) ) j + ( 2m ( 7.43N 7. consider an example. Here there is only a z component because this vector points out of the page.66N ) – 0m ( – 6. it is a vector.math guide . this result is positive.3k ( mN ) NOTE: there are two things to note about the solution. that is the positive direction.43i + 7.57 32.66N ) )i – ( 2m0N – 0m ( – 6.5.32. and a rotation about this vector would rotate on the plane of the page.43N ) )k = 15.66N 0N ∴M = ( 0m0N – 0m ( 7. and curl your fingers. • The basic properties of the cross product are. If the left handed coordinate system is used F and d should be reversed.3 Cross Product • First. Second. In this right handed system find the positive rotation by pointing your right hand thumb towards the positive axis (the ‘k’ means that the vector is about the z-axis here). because the positive sense is defined by the vector system. First. 2m 0m 0m – 6.

58 The cross (or vector) product of two vectors will yield a new vector perpendicular to both vectors.math guide . The fingers curl in the positive direction. V1 × V 2 V1 V 1 × V2 = ( x1 i + y1 j + z1 k ) × ( x2 i + y2 j + z 2 k ) i j k V 1 × V2 = x1 y1 z 1 x2 y2 z 2 V 1 × V2 = ( y 1 z 2 – z 1 y 2 )i + ( z 1 x 2 – x 1 z 2 )j + ( x1 y 2 – y 1 x 2 )k V2 We can also find a unit vector normal ‘n’ to the vectors ‘V1’ and ‘V2’ using a cross product. The positive orientation of angles and moments about an axis can be determined by pointing the thumb of the right hand along the axis of rotation.32. y x z x + z y + z + x y • The properties of the cross products are. . with a magnitude that is a product of the two magnitudes. V 1 × V2 λ n = -------------------V 1 × V2 • When using a left/right handed coordinate system. divided by the magnitude.

but not associative. • A matrix has the form seen below. u x u y uz (d × F) • u = d x d y dz Fx Fy Fz 32.as a result they are found in many computer applications.59 The cross product is distributive. we can do both steps in one operation by finding the determinant of the following.32. but we cannot change the order of the cross product. followed by a dot product (called the mixed triple product). and then the moment component twisting the pipe is found using the dot product.5. An example of a problem that would use this shortcut is when a moment is found about one point on a pipe.math guide . This allows us to collect terms in a cross product operation.5. .4 Triple Product When we want to do a cross product.5 Matrices • Matrices allow simple equations that drive a large number of repetitive calculations . r1 × F + r2 × F = ( r1 + r 2 ) × F r×F≠F×r but r × F = –( F × r ) DISTRIBUTIVE NOT ASSOCIATIVE 32.

math guide . A = 2 3 4 5 B = 6 7 8 9 10 11 12 13 14 C = 15 16 17 18 19 20 21 D = 22 23 E = 24 25 26 Addition/Subtraction 3+2 4+2 5+2 A+B = 6+2 7+2 8+2 9 + 2 10 + 2 11 + 2 B + D = not valid 3–2 4–2 5–2 B–A = 6–2 7–2 8–2 9 – 2 10 – 2 11 – 2 B – D = not valid 3 + 12 4 + 13 5 + 14 B + C = 6 + 15 7 + 16 8 + 17 9 + 18 10 + 19 11 + 20 3 – 12 4 – 13 5 – 14 B + C = 6 – 15 7 – 16 8 – 17 9 – 18 10 – 19 11 – 20 .32.many of these are indicated. There are also many restrictions . • Matrix operations are available for many of the basic algebraic expressions. We may also represent a matrix as a 1-by-3 for a vector.60 n columns a 11 a 21 … a n1 a 12 a 22 … a n2 … … … … a 1m a 2m … a nm m rows If n=m then the matrix is said to be square. examples are given below. Many applications require square matrices.

61 Multiplication/Division 3(2) 4( 2) 5(2) A ⋅ B = 6(2) 7( 2) 8(2) 9 ( 2 ) 10 ( 2 ) 11 ( 2 ) 3 -2 B -... D.math guide .= 6 -A 2 9 -2 4 -2 7 -2 10 ----2 5 -2 8 -2 11 ----2 ( 3 ⋅ 21 + 4 ⋅ 22 + 5 ⋅ 23 ) B ⋅ D = ( 6 ⋅ 21 + 7 ⋅ 22 + 8 ⋅ 23 ) ( 9 ⋅ 21 + 10 ⋅ 22 + 11 ⋅ 23 ) D ⋅ E = 21 ⋅ 24 + 22 ⋅ 25 + 23 ⋅ 26 B⋅C = ( 3 ⋅ 12 + 4 ⋅ 15 + 5 ⋅ 18 ) ( 3 ⋅ 13 + 4 ⋅ 16 + 5 ⋅ 19 ) ( 3 ⋅ 14 + 4 ⋅ 17 + 5 ⋅ 20 ) ( 6 ⋅ 12 + 7 ⋅ 15 + 8 ⋅ 18 ) ( 6 ⋅ 13 + 7 ⋅ 16 + 8 ⋅ 19 ) ( 6 ⋅ 14 + 7 ⋅ 17 + 8 ⋅ 20 ) ( 9 ⋅ 12 + 10 ⋅ 15 + 11 ⋅ 18 ) ( 9 ⋅ 13 + 10 ⋅ 16 + 11 ⋅ 19 ) ( 9 ⋅ 14 + 10 ⋅ 17 + 11 ⋅ 20 ) B B -----. etc = not allowed (see inverse) . . the first matrix must have the same number of columns as the second matrix has rows. --.32.C D B Note: To multiply matrices.

E = not valid (matrices not square) Transpose 3 6 9 T B = 4 7 10 5 8 11 D = 21 22 23 T 24 E = 25 26 T .math guide .62 Determinant B = 3⋅ 7 8 –4⋅ 6 8 +5⋅ 6 7 10 11 9 11 9 10 = ( 7 ⋅ 11 ) – ( 8 ⋅ 10 ) = –3 = 3 ⋅ ( –3 ) – 4 ⋅ ( – 6 ) + 5 ⋅ ( – 3 ) = 0 7 8 10 11 6 8 9 11 6 7 9 10 = ( 6 ⋅ 11 ) – ( 8 ⋅ 9 ) = –6 = ( 6 ⋅ 10 ) – ( 7 ⋅ 9 ) = –3 D .32.

math guide . Note the sign changes on alternating elements. and the matrix is indeterminate. and then finding the determinant of what is left. B D = B –1 –1 x ⋅B⋅ y z x x B D = I⋅ y = y z z –1 –1 B B= -------B In this case B is singular. D –1 = invalid (must be square) .63 Adjoint 7 8 10 11 B = – 4 5 10 11 4 5 7 8 – 6 8 10 11 3 5 9 11 – 3 5 6 8 6 7 9 10 – 3 4 9 10 3 4 6 7 T The matrix of determinant to the left is made up by getting rid of the row and column of the element. so the inverse is undetermined.32.z we need to move B to the left hand side. D = invalid (must be square) Inverse x D = B⋅ y z To solve this equation for x. To do this we use the inverse.y.

64 Identity Matrix This is a square matrix that is the matrix equivalent to ‘1’. B⋅I = I⋅B = B D⋅I = I⋅D = D B –1 ⋅B = I 1 0 0 1 0.6 Solving Linear Equations with Matrices • We can solve systems of equations using the inverse matrix.5. then rearrange.32.math guide . 0 1 0 . A – λI = 0 32. and solve. Given. 2 4 3 x 5 = 7 9 6 8 y 11 13 10 z 12 x 2 4 3 ∴ y = 9 6 8 z 11 13 10 –1 5 7 = 12 . 0 1 0 0 1 • The eigenvalue of a matrix is found using. etc=I 1. 2⋅x+4⋅y+3⋅z = 5 9⋅x+6⋅y+8⋅z = 7 11 ⋅ x + 13 ⋅ y + 10 ⋅ z = 12 Write down the matrix.

= A 2 5 3 9 7 8 11 12 10 y = ----------------------------. 2 4 3 9 6 8 11 13 10 5 7 12 A = B = Calculate the determinant for A (this will be reused).math guide . .= A 2 4 5 9 6 7 11 13 12 z = ----------------------------. 2⋅x+4⋅y+3⋅z = 5 9⋅x+6⋅y+8⋅z = 7 11 ⋅ x + 13 ⋅ y + 10 ⋅ z = 12 Write down the coefficient and parameter matrices. Given.32. and calculate the determinants for matrices below.5. 5 4 3 7 6 8 12 13 10 x = ----------------------------.65 • We can solve systems of equations using Cramer’s rule (with determinants).= A 32. Note: when trying to find the first parameter ‘x’ we replace the first column in A with B. Perform the matrix operations below.7 Practice Problems 1.

167 = 0. 1 A = 2 3 6 B = 2 1 Cross Product Dot Product A×B = A•B = ANS.167 0. A × B = ( – 4. – 10 ) A • B = 13 4.32. Solve the following set of equations using a) Cramer’s rule and b) an inverse matrix.5 0.167 – 0.167 – 0.333 0. Solve the following equations using any technique. Perform the vector operations below.167 0.66 Multiply 1 2 3 10 4 5 6 11 = 7 8 9 12 Determinant 12 3 42 6 78 9 Inverse 12 3 42 6 78 9 –1 ANS. 5x – 2y + 4z = – 1 6x + 7y + 5z = – 2 2x – 3y + 6z = – 3 ANS.833 0.072 z= -0.math guide . 1 2 3 10 68 = 167 4 5 6 11 7 8 9 12 266 = 1 2 3 4 2 6 7 8 9 = 36 = 1 2 3 4 2 6 7 8 9 –1 – 0.167 2. 17. 2x + 3y = 4 5x + 1y = 0 . x= 0.273 y= -0.627 5.

Perform the matrix calculations given below.math guide . A B C X L + M D E F Y GH I Z N T T T = AX + BY + CZ L + M DX + EY + FZ GX + YH + IZ N AX + BY + CZ + L = DX + EY + FZ + M GX + YH + IZ + N = AX + BY + CZ + L DX + EY + FZ + M GX + YH + IZ + N 7. Show all work.= -------– 13 2 3 5 1 T 2 4 5 0 – 20 20 y = --------------.= ----– 13 13 2 3 5 1 b) x = 2 3 y 5 1 –1 1 –( 5 ) 1 –3 4– ----–( 3 ) 2 –5 2 4 1- 4 4 = ------------------------------.67 (ans. Perform the following matrix calculation.32. a) A B C X D E F Y = GH I Z b) A B C D E F X Y Z = GH I .4 = -----------------13 =  – -----  13 – 20 = 20 – 13 0 0 0 2 3 ----13 5 1 6. 2 3 x = 4 5 1 y 0 a) 4 3 0 1 4x = --------------. T A B C X L + M D E F Y GH I Z N (ans.= -------.

68 8.= -----------------.–c ----------------.----------------c d –b a = -----------------. of the vectors A and B below.math guide . x A = y z p B = q r (ans.= ad – bc ad – bc ad – bc –b a a b ----------------. and the cross product. Find the dot product. A • B = xp + yq + zr A×B = i j k x y z p q r = i ( yr – zq ) + j ( zp – xr ) + k ( xq – yp ) = yr – zq zp – xr xq – yp 9. a) a b c T d e f g h k mnp b) a b c d c) a b c d –1 (ans. Perform the following matrix calculations.32.----------------ad – bc ad – bc c d adj . a) a b c T d e f d e f = a b c g h k = ( ad + bg + cm ) ( ae + bh + cn ) ( af + bk + cp ) g h k mnp mnp = ad – bc b) a b c d c) a b c d –1 a b d –c d .

= – 7 1 ( 4 – 16 ) + 2 ( 32 – 1 ) + 3 ( 2 – 16 ) – 12 + 62 – 42 8 1 2 3 1 4 8 4 2 1 11. 1 2 3 x 5 1 4 8 y = 0 4 2 1 z 1 5 2 3 0 4 8 1 2 1 5 ( 4 – 16 ) + 2 ( 8 – 0 ) + 3 ( 0 – 4 ) – 60 + 16 – 12 – 56 x = ----------------. x + 2y + 3z = 5 x + 4y + 8z = 0 4x + 2y + z = 1 (ans.32. a) A B C X D E F Y = GH I Z A B C D E F X Y Z = GH I det A B = CD A B CD A B CD A b) c) d) = –1 e) = .= --------------------------------------------------------------------------------. Perform the matrix calculations given below. Find the value of ‘x’ for the following system of equations.math guide .= ---------------------------------.69 10.= -------.

.6. Each one has to be learned separately.1. 3x + 5y = 7 4x – 6y = 2 a) Inverse matrix b) Cramer’s rule c) Gauss-Jordan row reduction d) Substitution 32.70 12.Differentiation • The basic principles of differentiation are. 32.math guide . or know where to look. the secret to calculus is to remember that there is no single "truth" . Solve the following set of equations with the specified methods. and when needed you must remember it.6 CALCULUS • NOTE: Calculus is very useful when looking at real systems.6. But. Many students are turned off by the topic because they "don’t get it".it is more a loose collection of tricks and techniques.32.1 Single Variable Functions 32.1 .

( v ) + ( vw ) ----.32.( u + v + … ) = ----. and all angles are in radians.( x ) du • Differentiation rules specific to basic trigonometry and logarithm functions . Also note that C is used as a constant.( x ) du d----.( y ) ----.( uvw ) = ( uv ) ----.( u ) dx dx dx dx ddd----.71 Both u.( uv ) = ( u ) ----.( u ) dx dx dx v d d.( v ) + … dx dx dx n–1 d d.( u ) = chain rule dx du dx 1 d----.( w ) + ( uw ) ----.( v ) + ( v ) ----.d----.( y ) du d.-u.( C ) = 0 dx dd----.( y ) = ----.( Cu ) = ( C ) ----.( u ) + ----. v and w are functions of x.( u ) dx dx ddd----.( u ) = ( nu ) ----. but this is not shown for brevity. u =  ---- ----.( y ) = ----------------ddx ----. v  2 dx  2 dx dx v v dddd----.( u ) = ------------ddx ----.( v ) .n ----.( u ) dx dx ddd----.( u ) –  ---- ----.math guide . d----.

( u ) dx dx dd----.( u ) dx dx dd----.( sin u ) = ( cos u ) ----.( u ) dx dx d1 2 d----.g ( x )   dt    ---- g ( x ) dt • Some techniques used for finding derivatives are.( u ) dx dx dd----.( cot u ) = ( – csc u ) ----.32.u ----. 2 .( tanh u ) = ( sech u ) ----.( ln x ) = -x dx • L’Hospital’s rule can be used when evaluating limits that go to infinity.( u ) dx dx 2 d d----.( u ) dx dx 1 d----.( u ) dx dx dd----. dd  ---- f ( x )    ---- f ( x )    dt    dt  f(x) lim  ---------- = lim  --------------------- = lim  ----------------------. = … x → a g ( x ) x → a  d  x → a  d  2  ---.( sec u ) = ( tan u sec u ) ----.( sinh u ) = ( cosh u ) ----.72 dd----.( csc u ) = ( – csc u cot u ) ----.( u )  cos u dx dx u d d.( e ) = ( e ) ----.( cos u ) = ( – sin u ) ----.( tan u ) =  ----------- ----.( u ) dx dx dd----.( u ) dx dx 2 d d----.( cosh u ) = ( sinh u ) ----.math guide .

n d.1.∫ f ( u ) du u = Cx C d F(u) ∫ F ( f ( x ) ) dx = ∫ F ( u ) -----. ∫ C dx = ax + C = C ∫ f ( x ) dx = ∫ u dx + ∫ v dx + ∫ w dx + … ∫ Cf( x ) dx ∫ u dv ∫ ( u + v + w + … ) dx ∫ f ( Cx ) dx n = uv – ∫ v du = integration by parts 1 = --.32. and w are functions of x.math guide .6. u.73 Leibnitz’s Rule. in addition to this.2 .2 d.du du f' ( x ) x ∫ x dx = -----------.n d.1 d. v. C is a constant.Integration • Some basic properties of integrals include.n – 1 ( v )  ----- ( uv ) =  ----- ( u )  ----- ( v ) +  n  ----- ( u )  -----  dx  dx  dx  dx  1  dx d.dx x 1 = ln x + C x ∫ e dx x = e +C • Some of the trigonometric integrals are.+ C ln a x x n+1 u = f(x) ∫ -.0 d. (notice the form is similar to the binomial equation) can be used for finding the derivatives of multiplied functions.( x ) du = ∫ ----------.+ C n+1 a ∫ a dx = -------. and all angles are radians. . In the following expressions.n – 2 ( v ) + … +  n  ----- n ( u )  ----- 0 ( v ) dd n  ----- ( u )  -----  dx  dx  dx  dx  2  n 32. d.

+ C n+1 ∫ ( cos x ) 4 ∫ sinh x dx ∫ cosh x dx ∫ tanh x dx = cosh x + C = sinh x + C = ln ( cosh x ) + C ∫ ( cos x ) 2 cos x ( ( sin x ) + 2 ) ∫ ( sin x ) dx = – ------------------------------------------.+ C 8 4 32 n+1 ( sin x ) n ∫ cos x ( sin x ) dx = -----------------------.+ -.32.+ a x – 2 sin ( ax ) + C ∫ 2 3 a a • Some other integrals of use that are basically functions of x are.+ C 2 2 3x sin 2x sin 4x dx = ----.+ ------------. .------------------x cos ( ax ) dx = ------------------------.+ C 3 3 sin x ( ( cos x ) + 2 ) ∫ ( cos x ) dx = ------------------------------------------.+ C 2 sin x cos x + x dx = -----------------------------.+ C 3 3 2 ∫ x cos ( ax ) dx cos ( ax ) x = ------------------.+ ------------.sin ( ax ) + C 2 a a 2 2 2x cos ( ax ) 2 .math guide .74 ∫ sin x dx ∫ cos x dx ∫ ( sin x ) 2 = – cos x + C = sin x + C sin x cos x + x dx = – -----------------------------.

b < 0 ∫  a – x –b  2 ( – b )a ln ( bx + a ) 2 –1 x ( a + bx ) dx = --------------------------.ln  -------------------------.ln  ----------- + C. c < 0 ∫ 2 –c b – 4ac .+ C n+1 n n+1 ∫ ( a + bx ) –1 ln a + bx dx = ----------------------.atan  x ab + C ----------- a  b b ab 2 2 –1 2 2 1a+x ( a – x ) dx = ----.– ------------. a > 0. + C.math guide .+ C b 2 –1 1 a + 2 –b ( a + bx ) dx = --------------------.ln ∫ c b 2 a + bx + cx + x c + --------.asin -----------------------.+ C.+ C ∫ 2b 2 ∫x 2 x 2 –1 a ( a + bx ) dx = -.75 x ∫ x dx = -----------.32. c > 0 2 c 2 –1 1– 2cx – b( a + bx + cx ) dx = --------.+ C b x ±a +C 2 2 CORRECT?? ± a ) dx = 2 1 – -2 2 –1 1 ( a + bx + cx ) dx = -----.+ C. a > x ∫  a – x 2a ∫ ( a + bx ) ∫ x( x 2 –1 2 a + bx dx = ---------------------.

 ---.76 2 2 ∫ ( a + bx ) dx = -----.( a – x ) + 1 x ( a – x ) + a asin  -- - a 4 8 1 – -2 1 -1 -2 2 3 -2 2 1 -2 2 ∫x 2 12 2 1 ( 1 + a x ) dx = -.– x  ∫  3a 2 2 2 x 2 x ( a – x ) dx = – -. a2 + x  – -------.+ x 2 2 ln  3 1 1  2   ---a 2   12 2 2 2 2 2 2 2 ax  ---8 ∫ x ( 1 + a x ) dx = ----.32.asin ( ax ) = – -.+ x 2 2 ln  1  2   -a   2 2 2 1 -x ( 1 + a x ) + -------------------------------------------2 a 2 2 ∫ ( 1 + a x ) dx = ---------------------------------------------------------------------------------2 1 a  ---.acos ( ax ) a a .math guide .( a + bx ) 3b 2 2 ∫ ( a + bx ) dx = -----.+ x   2 a ∫ x ( 1 + a x ) dx = ---------------------------3 1 -2 2 2 --  1  x +  ---.2 x ( 1 – a x ) dx = – -.x ( 1 + a x ) – --------------------------------------------3 2 4 8a 8a 1 -2 2 2 1 3 -2 2 ∫ (1 – a asin ( ax ) x ) dx = 1 x ( 1 – a x ) + --------------------a 2 1 -2 2 2 3 -- 1 -2 2 2 2 2 a 1.+ x  ∫  2  a a 2 2 ∫ (1 – a 1 – -2 2 2 1 1 x ) dx = -.ln x +  ---.( a + bx ) 3b 1 -2 1 -2 1 -2 3 -- 3 -- ∫ x ( a + bx ) 2 ( 2a – 3bx ) ( a + bx ) dx = – ---------------------------------------------------2 15b 1 3 -2 --  1  x +  ---.

77 • Integrals using the natural logarithm base ‘e’.math guide .+ C a e ∫ xe dx = ------. • Consider a basic function and how it may be represented with partial derivatives.( ax – 1 ) + C 2 a ax ax ax 32.32. and as a whole. .2 Vector Calculus • When dealing with large and/or time varying objects or phenomenon we must be able to describe the state at locations.6. To do this vectors are a very useful tool. e ax ∫ e dx = ------.

z ) ∂x ∂y ∂z  We then replace these vectors with the operators below. In this form we can manipulate the equation easily (whereas the previous form was very awkward). y. z ) • dX ( d )y = ∇f ( x. we get a simpler form. but the right hand side must contain partial derivatives. z ) We can write this in differential form. z ) dz   ∂ x    ∂ y    ∂ z  ∂ ∂ ∂ ( d )y =   dx +   dy +   dz f ( x. V.math guide . and divide it into two vectors. This is very useful for dealing with magnetic fields. ∂ ∂ ∂ ( d )y =    f ( x.78 y = f ( x. We can then look at them as the result of a dot product. If we separate the operators from the function. z ) dy +    f ( x. z ) dx +    f ( x. y. ( d )y = [ ∇ • dX ]f ( x. y. Both sides give the flux across a surface. y. y. y. y . z ) dX cos θ In summary. V ∫ ( ∇ • F ) dV = ∫ °FdA A where. y. or out of a volume.32. ∇ = ∂ ∂ ∂ i+ j+ k ∂x ∂y ∂z ∇ • F = the divergence of function F ∇ × F = the curl of functio n F F = Fx i + F y j + F z k • Gauss’s or Green’s or divergence theorem is given below. z )  ∂ x  ∂ y  ∂ z ( d )y =  ∂ i + ∂ j + ∂ k • ( dxi + dyj + dzk ) f ( x. A = a volume V enclosed by a surface area A F = a field or vector value over a volume . y. z ) ( d )y = ∇f ( x.

d.g. • Another complication that often occurs is that the solution of the equations may vary depending upon boundary or initial conditions. In the example below ’x’ is the variable with derivatives.if all the derivatives are of a single variable.79 • Stoke’s theorem is given below.g.2 d ---- x +  ---- x = y  dt  dt e. A.6. then they will oscillate indefinitely. . A ∫ ( ∇ × F ) dA = ∫ °FdL L where. with a bounding parimeter of lengt L h F = a field or vector value over a volume 32. first-order differential equations . An example of this is a mass spring combination. This is very useful for dealing with magnetic fields. • Note: These equations are typically shown with derivatives only. but there are a number of forms that need to be remembered. • We can judge the order of these equations by the highest order derivative in the equation..32. but if there is some disturbance. ordinary differential equations . when integrals occur they are typically eliminated by taking derivatives of the entire equation. or out of a volume. L = A surface area A. • Some of the terms used when describing differential equations are..3 Differential Equations • Solving differential equations is not very challenging. dd ---- x +  ---- y = 2  dt  dt e.have only first-order derivatives. Both sides give the flux across a surface.math guide . If they are initially at rest then they will stay at rest.

guessing then testing separation . eg. partial differential equations . • Examples of these equations are given below..1 .have at least one derivative that is higher than second-order.3.have at least on second derivative.6.First-order Differential Equations • These systems tend to have a relaxed or passive nature in real applications. • linearity of a differential equation is determined by looking at the dependant variables in the equation. higher order differential equations .math guide . In these cases backtrack and try another approach.80 second-order differential equations . y' + 2xy – 4x = 0 y' – 2y = 0 2 3 • Typical methods for solving these equations include.these equations have partial derivatives • Note: when solving these equations it is common to hit blocks. y'' + y' + 2 = 5x ( y'' ) + y' + 2 = 5x y'' + ( y' ) + 2 = 5x y'' + sin ( y' ) + 2 = 5x 3 2 linear non-linear non-linear non-linear 32. The equation is linear if they appear with an exponent other than 1.2 d ---- x +  ---- y = 2  dt  dt e.32.g. d.

it is correct y = Ce –t • The previous example showed a general solution (i.math guide .Guessing • In this technique we guess at a function that will satisfy the equation. This typically means there is only a single derivative term. .1 .81 homogeneous 32..e.Separable Equations • In a separable equation the differential can be split so that it is on both sides of the equation. y' + y = 0 y = Ce –t the given equation the guess now try to substitute into the equation y' = – Ce –t –t –t y' + y = – Ce + Ce = 0 therefore the guess worked . and test it to see if it works. We can also find a particular solution.6.2 .3. the value of ’C’ was not found).1. We then integrate to get the solution. y = Ce y = 5e –t a general solution a particular solution –t 32.6.1.3.32.

– y – 3y + C 3 3 2 e.+ x = 0 dy 1 ∴ – -- dx = dy  x ∴ ln ( – x ) = y 32.3 . dx ----. .Homogeneous Equations and Substitution • These techniques depend upon finding some combination of the variables in the equation that can be replaced with another variable to simplify the equation..g.g.6.32.+ y 2 + 2y + 3 = 0 dy ∴dx = ( – y – 2y – 3 )dy –y 2 ∴x = ------.1. dx ----.82 e.math guide . This technique requires a bit of guessing about what to substitute for.3.. and when it is to be applied.

• In engineering we will encounter a number of forms.1 ..3.g. . .2.= y – 1 dx x y u = x the equation given the substitution chosen Put the substitution in and solve the differential equation.6.2 .6.math guide .= u – 1 dx ∴u + x du = u – 1 ----dx –1 ∴du = --------x dx du 1 ∴– ----.= -dx x ∴– u = ln ( x ) + C Substitute the results back into the original substitution equation to get rid of ’u’.3.32.Second-order Differential Equations • These equations have at least one second-order derivative. dy ----.homogeneous .= ln ( x ) + C x ∴y = – x ln ( x ) – Cx 32.Linear Homogeneous • These equations will have a standard form. dy ----.83 e. y – .nonhomogeneous 32.

32.2 d ---- y + 6  ---- y + 3y = 0  dt  dt Guess. y = e y = e y = e ( – 3 + 2.2  ---- y + A  ---- y + By = 0  dt  dt • An example of a solution is.84 dd.449j )t – 3 t 2.449jt 2 2 Bt Bt Bt 2 = 0 e – 3t ( cos ( 2. y = e Bt d ---- y = Be Bt  dt d ---- y = B 2 e Bt  dt substitute and solve for B. .449j.3..2 . d. y = C 1 e + C 2 te Bt Bt 32.2.math guide . – 3 – 2.449t ) + j sin ( 2.449j substitute and solve for B.g.Nonhomogeneous Linear Equations • These equations have the general form.6. B e + 6Be + 3e B + 6B + 3 = 0 B = – 3 + 2.449t ) ) Note: if both the roots are the same. e.

Cx form A Ax + B Ax e B sin ( Ax ) or B cos ( Ax ) Guess C Cx + D Ax Ax Ce Cxe C sin ( Ax ) + D cos ( Ax ) or Cx sin ( Ax ) + xD cos ( Ax ) • Consider the example below.2 d ---- y + A  ---- y + B = 0  dt  dt to find yp guess at a value of y and then test for validity.32.85 d.2 d ---- y + A  ---- y + By = Cx  dt  dt • to solve these equations we need to find the homogeneous and particular solutions and then add the two solutions. y = yh + yp to find yh solve. A good table of guesses is. d. .math guide .

4 .32. solve for the particular part.3.2 d ---- y +  ---- y – 6y = 0  dt  dt try y = e Bx d ---- y = Be Bx  dt d ---- y = B 2 e Bx  dt 2 B e 2 2 Bx + Be Bx – 6e Bx = 0 B +B–6 = 0 B = – 3.86 d.Higher Order Differential Equations 32.math guide .3 . y = e – 3x – 2x +e 2x + 0. .25 y p = 0.25e – 2x 32. y = Ce – 2x d ---- y = – 2C e –2x  dt d ---- y = 4Ce–2x  dt 4Ce – 2x 2 + – 2C e – 2x – 6Ce – 2x = e – 2x 4C – 2C – 6C = 1 C = 0. etc.Partial Differential Equations • Partial difference equations become critical with many engineering applications involving flows.2 d ---- y +  ---- y – 6y = e –2x  dt  dt First solve for the homogeneous part.6.6.3. We will guess the function below. d.25e Finally. 2 yh = e – 3x 2x +e Next.

32.  d   3t 2t –2 –2 –3 2t dd.6.math guide .+ e 2t 2 ( 2 + t) (ans. 3t ----------------.2t ----  ----------------. Solve the following differential equation. x'' + 4x' + 4x = 5t x0 = 0 x'0 = 0 .4 Other Calculus Stuff • The Taylor series expansion can be used to find polynomial approximations of functions. ( a )( x – a) f'' ( a ) ( x – a ) f f ( x ) = f ( a ) + f' ( a ) ( x – a ) + -----------------------------.5 Practice Problems 1.+ … + -----------------------------------------------2! ( n – 1 )! 2 ( n – 1) n–1 32. Find the derivative of the function below with respect to time.6.87 32.+ e  =  ---- ( 3t ( 2 + t ) ) +  ---- ( e ) = 3 ( 2 + t ) – 6t ( 2 + t ) + 2e 2  dt   dt  dt (2 + t) 2. given the initial conditions at t=0s.

25 C 1 = 1.25 dd) ---. Find the following derivatives.25t – 1.25 x ( t ) = 1.25t – 1.25 + C 2 te –2 ( 0 ) – 2t + 1. homogeneous solution: guess: x'' + 4x' + 4x = 0 At xh = e 2 At x h' = Ae At At At x h'' = A e 2 2 At A e + 4Ae + 4e xh = C 1 e – 2t = 0 = A + 4A + 4 = ( A + 2 ) ( A + 2 ) + C 2 te – 2t particular solution: guess: x p = At + B x p' = A x p'' = 0 ( 0 ) + 4 ( A ) + 4 ( At + B ) = 5t ( 4A + 4B ) + ( 4A )t = ( 0 ) + ( 5 )t 5 A = -.25 –2 ( 0 ) 0 = C1 e + C 2 ( 0 )e + 1.25e + 1.( 5te ) dt C 2 = 1.25 ) + C 2 + 1.25t – 1.25 = C 1 – 1. da) ---.= 1.math guide .25 combine and solve for constants: x ( t ) = xh + xp = C1 e for x(0) = 0 – 2t 4 ( 1.( 5 ln t ) dt .( ( t + 2 ) ) dt 8t dc) ---.32.25 = 0 = – 2 ( 1.25 ) + 4B = 0 B = – 1.( sin t + cos t ) dt –2 db) ---.25 –2 C 1 + C 2 + 1.88 (ans.25 –2 ( 0 ) for (d/dt)x(0) = 0 2t 0 = ( – 2 )C 1 e 2t –2 ( 0 ) + ( – 2 )C 2 ( 0 )e + C2 e –2 ( 0 ) + 1.25 4 x p = 1.25 3.25 ( 0 ) – 1.25te + 1.

( sin t ) + ---.89 (ans. Find the following derivative. + C = 2e + C  7 7t 3 c) d) ∫ sin ( 0. Find the following integrals ∫ 6t dt 7t b) ∫ 14 e dt a) c) ∫ sin ( 0.( sin t + cos t ) = ---.( cos t ) = cos t – sin t dt dt dt –2 –3 db) ---.5t ) dt ∫ -.5t ) + C 0.( 5te ) = 5e + 40te dt 5 dd) ---. 3 ta) ∫ 6 t dt = 6  --- = 2t + C  3 2 7t 7t eb) ∫ 14 e dt = 14  ----.5t ) dt d) ∫ -.( 5te + ( t + 4 ) ) dt .5t ) = ------------------------.dx x 5 – cos ( 0.+ C = – 2 cos ( 0.dx x 5 2 (ans.5 = 5 ln ( x ) + C 5.( ( t + 2 ) ) = – 2 ( t + 2 ) dt 8t 8t 8t dc) ---. 4t –3 d---.32. ddda) ---.( 5 ln t ) = -t dt 4.math guide .

a) ∫e 2t dt = 0.32.–t ---.( e sin ( 2t – 4 ) ) dt b) (ans. Find the following derivatives. a) ∫e 2t dt b) ∫ ( sin θ + cos 3θ ) dθ (ans.–t ---.1 ----.math guide . ----------- = -----------------2  x + 1 dx (x + 1) –t –t d.( e sin ( 2t – 4 ) ) = – e sin ( 2t – 4 ) + 2e cos ( 2t – 4 ) dt b) 7. ----------- dx  x + 1 d.5e + C 2t b) ∫ ( sin θ + cos 3θ ) dθ = – cos θ + 1 sin 3θ + C -3 8. x'' + 5x' + 3x = 3 x(0 ) = 1 x' ( 0 ) = 1 . Solve the following differential equation.1 ----.90 6. a) –1 d. a) d. Solve the following integrals.

math guide . – 1 2 2 xh = C1e Particular: xp = A x' p = 0 x'' p = 0 A = 1 – 4t + C2 e –t 0 + 5 ( 0 ) + 3A = 3 xp = 1 Initial values: x = xh + xp = C1 e – 4t – 4t –t + C2e + 1 C1 = –C2 1 = C1 ( 1 ) + C2 ( 1 ) + 1 x' = – 4C 1 e – C2 e –t 1 = – 4C 1 ( 1 ) – C 2 ( 1 ) 4 ( –C2 ) + C 2 = – 1 1 –4t –t x = – -. Set up an integral and solve it to find the volume inside the shape below.= – 4.e + 1 e + 1 -3 3 1 C 2 = -3 1 C 1 = – -3 9. .91 (ans. The shape is basically a cone with the top cut off. x'' + 5x' + 3x = 3 Homogeneous: A + 5A + 4 = 0 2 x(0 ) = 1 x' ( 0 ) = 1 – 5 ± 25 – 16 –5±3 A = ----------------------------------.32.= ---------------.

 a + -.math guide . x + π  ---------. x  a   a  2 2 a c–b r = b +  ---------. V = ∫ dV = ∫0 A dx 2 2 c–b c–b 2 = πb + 2π  ---------. Solve the second order non-homogeneous differential equation below. -. x   a   V =  2  c – b  c – b 2 ∫0  πb + 2π  ----------- x + π  ----------- x  dx a a a a 2 2 3 2 V = πb x + π  c – b x + π  c – b x ---------. 2x'' + 4x' + 2x = 5 where. Solve the first order non-homogeneous differential equation below.92 z b y x a c (ans. x(0) = 2 x' ( 0 ) = 0 . a  a  3 a  2 2 2 V = πb a + π ( c – b )a + π ( c – b ) a -3 0 10. ---------. ---------.32. x  a  2 c–b A = πr = π  b +  ---------. Assume the system starts at rest.  a 3 a 2 c–b 2 π c–b 2 3 V = πb a + π  ---------. 2x' + 4x = 5 sin 4t 11.

( y i – y i – 1 ) –2 yi + yi – 1 + yi + 1 T T y ( ti ) ≈ ----------------------------------------------------------------. y(t) yi + 1 yi yi – 1 ti – 1 ti T ti T ti + 1 yi + yi – 1 T y ( t i ) ≈  -------------------. and will result in small errors.( y i + y i – 1 ) ∫t i – 1   2 2 y i – yi – 1 yi + 1 – y 1 1 d---.y ( ti ) ≈  -------------------- =  --------------------i = -. 32.( y i + 1 – y i ) t –t  t dt – ti  T T i i–1 i+1 d ----  dt 2 1 1 -. etc. Numerical techniques are not as elegant as solving differential equations.( y i – y i – 1 ) = -.7 NUMERICAL METHODS • These techniques approximate system responses without doing integrations.= -------------------------------------------2 T T .32.7.math guide .this means by doing repeated calculations to solve the equation.1 Approximation of Integrals and Derivatives from Sampled Data • This form of integration is done numerically . ( t i – ti – 1 ) = -.( y i + 1 – y i ) – -. • This method uses forward/backward differences to estimate derivatives or integrals from measured data.93 32. But these techniques make it possible to solve complex problems much faster.

3 d. d1.h  ---- x ( t ) + ---.h  ---- x ( t ) + …  dt 2!  dt 3!  dt 4!  dt • When h=0 this is called a MacLaurin series.7 0.8 0.h  ---- x ( t ) + ---.6 0. dy ( t + h ) ≈ y ( t ) + h ---.4 0.1 0. d ---- x = 1 + x 2 + t 3  dt d ---- x = 0  dt 0 x0 = 0 t (s) 0 0.4 d.3 Taylor Series Integration • Recall the basic Taylor series.y ( t ) dt 32.3 0.2 0.2 d.94 32.7.3 1.32.1 .2 Euler First-order Integration • We can also estimate the change resulting from a derivative using Euler’s equation for a first-order difference equation.4 x ( t + h ) = x ( t ) + h  ---- x ( t ) + ---.math guide .7.9 x(t) 0 d/dt x(t) 0 h = 0. • We can integrate a function by.5 0.2 1.

. One simple technique uses an instantaneous slope of the function.math guide .f ( x )  dx i  • The function f(x) is supplied by the user. x = the state variables f = the differential function t = current point in time h = the time step to the next integration point 32.( F1 + 2F 2 + 2F 3 + F4 ) 6 F1 = hf ( t.5 Newton-Raphson to Find Roots • When given an equation where an algebraic solution is not feasible. and takes iterative steps towards a solution.7. x + F 3 ) where.4 Runge-Kutta Integration • The equations below are for calculating a fourth order Runge-Kutta integration. x + -----  2 2 F4 = hf ( t + h.95 32. f ( xi ) x i + 1 = x i – ---------------------d ----. x ) F1 h F 2 = hf  t + --.32. x + -----  2 2 F2 h F 3 = hf  t + --.7. a numerical solution may be required. 1 x ( t + h ) = x ( t ) + -.

• The basic Laplace transform equations is shown below. And.32. F( s ) = where. ∫0 f ( t )e ∞ – st dt f ( t ) = the function in terms of time t F ( s ) = the function in terms of the Laplace s 32. as you recall from before the inverse of time is frequency. the Laplace transform is much more useful in system analysis.math guide .8 LAPLACE TRANSFORMS • The Laplace transform allows us to reverse time. Because we are normally concerned with response.96 • This method can become divergent if the function has an inflection point near the root.1 Laplace Transform Tables • Basic Laplace Transforms for operational transformations are given below.8. • This calculation should be repeated until the final solution is found. 32. • The technique is also sensitive to the initial guess. .

32.97 TIME DOMAIN Kf ( t ) f 1 ( t ) + f 2 ( t ) –f 3 ( t ) + … df ( t ) ----------dt ) d f(t ------------2 dt d f(t ) ------------n dt n 2 FREQUENCY DOMAIN Kf ( s ) f1 ( s ) + f2( s )–f3( s ) + … sf ( s ) – f ( 0 ) df ( 0 ) 2 – s f ( s ) – sf ( 0 ) – ----------------dt s f(s) – s f( s ) -------s e – as n n–1 – – f(0 ) – s – n – 2 df ( 0 d f(0 ) ) ----------------.a  a – df ( s ) --------------ds f(s) ( – 1 ) -------------n ds nd ∞ n f ( at ).math guide .f s . .– … – ------------------n dt dt – n – ∫0 f( t ) dt f ( t – a )u ( t – a ). a > 0 e – at t f(s) f(t ) f( s – a ) 1  -- -. a > 0 tf ( t ) t f(t ) f( t ) ------t n ∫ f ( u ) du s • A set of useful functional Laplace transforms are given below.

98 TIME DOMAIN FREQUENCY DOMAIN A -s 1 ---2 s 1 ---------s+a ω ----------------2 2 s +ω s ----------------2 2 s +ω 1 -----------------2 (s + a) ω ------------------------------2 2 (s + a) + ω s+a ------------------------------(s + a) + ω 2 2 A t e – at sin ( ωt ) cos ( ωt ) te e e – at – at sin ( ωt ) cos ( ωt ) – at Ae – at A---------s–a A ----------------2 (s – a) cos ( βt + θ ) cos ( βt + θ ) A ----------------------. .+ A .32.math guide .------------------------------------2 2 ( s + α – βj ) ( s + α + βj ) complex conjugate complex conjugate Ate – at 2Ae – αt 2t A e – αt • Laplace transforms can be used to solve differential equations.+ A ------------------------------------s + α – βj s + α + βj A -----------------------------.

In both cases. similar to the Laplace transform. of finite magnitude). ..∫ X ( z )z dz . the two-sided z-transform is defined by X(z ) = n = –∞ ∑ ∞ x [ n ]z . • Along with a z-transform we associate its region of convergence (or ROC).99 32. the z-transform is a polynomial in the complex variable z . This is usually avoided by partial fraction inversion techj2π° niques.9 z-TRANSFORMS • For a discrete-time signal x [ n ] . • The inverse z-transform is obtained by contour integration in the complex plane n–1 1x [ n ] = ------. These are the values of z which isX ( z ) for bounded (i.32.e.math guide . The one-sided z-transform is defined by X ( z ) = –n ∑ x [ n ]z n=0 ∞ –n .

32. Table 1: Common z-transforms Signal x [n ] δ[n] u[n ] z-Transform X(z ) 1 1 --------------–1 1–z z ---------------------–1 2 (1 – z ) z (1 + z ) ---------------------------–1 3 (1 – z ) 1 -----------------–1 1 – az az -------------------------–1 2 ( 1 – az ) 1 -----------------–1 1 – az az -------------------------–1 2 ( 1 – az ) 1 – z cos ω 0 -----------------------------------------------–1 –2 1 – 2z cos ω 0 + z z sin ω 0 -----------------------------------------------–1 –2 1 – 2z cos ω 0 + z 1 – az cos ω 0 --------------------------------------------------------–1 2 –2 1 – 2az cos ω 0 + a z –1 –1 –1 –1 –1 –1 –1 –1 ROC All z z >1 nu [ n ] z >1 n u[ n] 2 z >1 a u[ n] n z > a na u [ n ] n z > a ( – a )u [ – n – 1 ] n z < a ( – na )u [ – n – 1 ] n z < a cos ( ω 0 n ) u [ n ] z >1 sin ( ω 0 n ) u [ n ] z >1 a cos ( ω 0 n ) u [ n ] n z > a .100 • Some common z-transforms are shown below.math guide .

32.< z < ---r1 r2 r2 < z < r1 z-Domain Scaling Time Reversal a x [ n] x [ –n ] nx [ n ] n X(a z) X(z ) – z dX ( z ) ------------dz –1 –1 z-Domain Differentiation . The table below lists properties for the two-sided z-transform.math guide .101 Table 1: Common z-transforms Signal x [n ] a sin ( ω 0 n ) u [ n ] n z-Transform X(z ) az sin ω 0 --------------------------------------------------------–1 2 –2 1 – 2az cos ω 0 + a z z ----------------------------–1 k + 1 (1 – z ) –k –1 ROC z > a n! ---------------------. instead ] Table 2: Two-sided z-Transform Properties Property Notation Time Domain x[n ] x1 [ n ] x2 [ n ] Linearity αx 1 [ n ] + βx 2 [ n ] X(z ) X1( z ) X2( z ) αX1 ( z ) + βX 2 ( z ) z-Domain ROC r2 < z < r1 ROC 1 ROC 2 At least the intersection of ROC 1 and ROC 2 Time Shifting x[n – k ] z X(z ) –k That of X ( z ) .u [ n ] k! ( n – k )! z >1 • The z-transform also has various properties that are useful. The one-sided z-transform properties can be x[n ] derived from the ones below by considering the signal x [ n ]u [ nof simply . except z = 0 if k > 0 and z = ∞ if k < 0 a r 2 < z < a r1 11 --.

 dx  L  L –L 1 L nπx b n = -.  L -  L - 1 L nπx a n = -.11 TOPICS NOT COVERED (YET) • To ensure that the omissions are obvious.∫ f ( x ) sin  -------.32.10 FOURIER SERIES • These series describe functions by their frequency spectrum content. dx  L  L –L 32.∫ f ( x ) cos  -------. + b n sin  -------.Fourier. I provide a list of topics not covered below.∫ X1 ( v )X 2  . Bessel .102 Table 2: Two-sided z-Transform Properties Property Convolution Time Domain x 1 [ n ]*x 2 [ n ] z-Domain X 1 ( z )X2 ( z ) ROC At least the intersection of ROC 1 and ROC 2 Multiplication x 1 [ n ]x 2 [ n ] x [ n ] causal 1z –1 ------. • Frequency domain . For example a square wave can be approximated with a sum of a series of sine waves with varying magnitudes. Some of these may be added later if their need becomes obvious.math guide . a0 f ( x ) = ---. • The basic definition of the Fourier series is given below.+ 2 n=1 ∑ ∞ nπx nπx a n cos  -------. v dv  v j2π° x [ 0 ] = lim X ( z ) z→∞ At least r 1l r 2l < z < r 1u r 2u Initial value theorem 32.

R. McGraw-Hill Book Company. Mathematical Handbook of Formulas and Tables. Schaum’s Outline Series. 1968.math guide .103 32.12 REFERENCES/BIBLIOGRAPHY Spiegel.32. M.. .

• You may easily customize ’C’ to your own needs.1 WHY USE ‘C’? • ‘C’ is commonly used to produce operating systems and commercial software. which means that it requires only small changes to run on other computers. • Suited to Large and Complex Programs.1 33. dBase. • Very Fast. • Machine Portable. and some ‘C’ compilers. • Emphasizes structured programming.33. Lotus-123. . A BASIC INTRODUCTION TO ‘C’ 33. Some examples of these are UNIX.C programming . by focusing on functions and subroutines. almost as fast as assembler.

and commercial compilers became available in the Late 70’s. and can never be ignored. • The main program is treated like a function. and thus it has the name main().C programming .3 PROGRAM PARTS • /* is the start of a comment. Over 90% of UNIX is written in ‘C’. allows you to create your own functions. AT&T originally developed ‘C’ with the intention of making it an in-house standard. 33. • */ is the end of comment. • lower/UPPER case is crucial.33.2 • Very Flexible. . 33.Recnetly has become more popular because of its ties to UNIX.2 BACKGROUND • Developed at Bell Laboratories in the Early 70’s.

This function will do a formatted print. another space. another integer. y. } • There are no line numbers. /* define three variables and give one a value */ x = 3. in their respective orders. x.C programming . z. y = 2. or a set of statements between curly brackets {. /*print the results */ } Results (output): • lines may be of any length. and z are the three integers to be printed. and another space.3 • Statements are separated by semi-colons ‘. /* give another variable a value */ z = x + y. /* add the two variables */ printf(“%d + %d = %d\n”. then a line feed ‘\n’. . In this case the format says print an integer %d followed by a space then a ‘+’ then another space. z). Program to Add two Numbers: /* A simple program to add two numbers and print the results */ main() { int x. • A very common function in ‘C’ is printf(). x.’ • Statements consist of one operation. ‘=’. The format is the first thing which appears between the brackets. All variables that follow the format statement are those to be printed.33. another integer. y.

double (8 byte IEEE floating point standard). char can be modified by the addition of unsigned. This function is capable of accepting an argument list. An unsigned integer will not use 1 bit for number sign. and it will speed things up (this is only available for integers). float (4 byte IEEE floating point standard).C programming . •int. and register. if possible.4 • Major Data Types for variables and functions are (for IBM PC): int (2 byte integer). char (1 byte integer). A register variable will use a data register in the microprocessor. short. and returning a single value. sin() and read()). long (4 byte integer). double m. . Example of Defining Different Data Types: main() { unsigned int i. register j. which has been assigned a name. (e. long. etc • A function consists of a sub-routine or program. short (1 byte integer). The function must be defined before it is called from within the program.33. short k. char l.g.

/* define three variables and give values */ z = add(x. x. y). /*print the results */ } int add(a. b) /* define function and variable list */ int a. static variables can be used for a variable that must keep the value it had the last time the function was called. This determines which functions are able to use that variable.33.5 Program to add numbers with a function: /* A simple program to add two numbers and print the results */ int add(). y = 2. . If a variable is defined in a function. Using extern will allow the variable types from other parts of the program to be used in a function. /* add the numbers */ return(c). /* describe types of variable lists */ { int c. then it will be local to that function. /* pass the two values to ‘add’ and get the sum*/ printf(“%d + %d = %d\n”. y. /* define a work integer */ c = a + b. /* Return the number to the calling program */ • Every variable has a scope. If the variable needs to be initialized every time the subroutine is called. then it may be used by any function. If a variable is global. /* Declare a integer function called ‘add’ */ main() { int x = 3. These can be modified by the addition of static. and is not used by any other function. extern and auto.C programming . b. z). this is an auto type. z.

/* Return the sumto the calling program */ • Other variable types of variables areunion. /*print the results */ } int add() /* define function */ { return(x + y). switch(). x. struct. add(). i = i + 1){ printf(“number %d \n”. etc. /*print the number */ } } . for(). for(i = 1. i <= 5.C programming . do-while(). add()).33. • Some basic control flow statements are while(). /* Declare an integer function called ‘add’ */ main() { printf(“%d + %d = %d\n”. Program example with a for loop: /* A simple program toprint numbers from 1 to 5*/ main() { int i. and if(). i). /* Define global x and y values */ y = 2. A couple of example programs are given below which demonstrate all the ’C’ flow statements. enum.6 Program example using global variables: /* A simple program to add two numbers and print the results */ int x = 3. y.

i). i = i + 1. do{ printf(“number %d \n”. i).C programming . while(i <= 5){ printf(“number %d \n”. i = i + 1. i = i + 1.7 or example with a while loop: main() { int i = 1. }while(i <= 5) } or example with a do until loop: main() { int i = 1.33. }until(i > 5) } . i). } } or example with a do while loop main() { int i = 1. do{ printf(“number %d \n”.

y). x-2). but a start */ printf(“There is one parent\n”). break.C programming . case 2: /* You need two to start something */ printf(“There are two parents\n”). break. break. case 1: /* Only one person. y = 3. } } Example Program using switch-case: main() { int x = 3. /* Number of People in Family */ switch(x){ /* choose the numerical switch */ case 0: /* Nobody */ printf(“There is no family \n”). if(x > y){ printf(“Maximum is %d \n”. } else { printf(“Both values are %d \n”. } } . break. x). } else if(y > x){ printf(“Maximum is %d \n”.33.8 Example Program with an if else: main() { int x = 2. default: /* critical mass */ printf(“There are two parents and %d kids\n”. x).

or else the compiler would not know what type of value that sin() is supposed to return. #include. This almost always includes “stdio. x = x + (TWO_PI / STEPS)){ printf(“%f = sin(%f) \n”.h extension is used to indicate a header file which contains ‘C’ code to define functions and constants. before the compiler touches the program. #undef CONSTANT will undefine the CONSTANT.h” contains a line which says ‘int printf(). /* and */ are all handled by the Preprocessor.h into the program. #if.h” #include “math.C programming . before compilation. . •#define CONSTANT TEXT will do a direct replacement of CONSTANT in the program with TEXT. If we needed to use a math function like y = sin(x) we would have to also use #include <math. x).9 • #include <filename. sin(x). The *.h” #define TWO_PI 6.h>. • #define. “stdio. /* Current x value*/ for(x = 0.283185307 #define STEPS 5 main() { double x. #else and #else can be used to conditionally include parts of a program. } } • #ifdef. #else.h> will insert the file named filename. x <= TWO_PI. As we saw before. This is use for including and eliminating debugging lines in a program. A Sample Program to Print Some sin() values (using defined constatnts) #include “stdio. #if. #ifndef. #ifndef.0.’. #ifdef.33. a function must be defined (as with the ‘add’ function).h”. this is normally done by using #include <stdio. We did not define printf() before we used it.h> at the top of your programs.

string[i]. ASCII 71 pos 0. Arrays may be any data type. or dimensions. i < STRING_LENGTH.10 • Matrices are defined as shown in the example. ASCII 72 pos 0.C programming . char G. i++){ printf(“pos %d.h” #define STRING_LENGTH 5 main() { int i. } } INPUT: HUGH<return> OUTPUT: pos 0. /* Input string from keyboard */ for(i = 0. Strings are stored as arrays of characters. • i++ is the same as i = i + 1. char H. char %c. string[i]). ASCII 0 . char U. char string[STRING_LENGTH]. i. ASCII %d \n”. ASCII 72 pos 0. ASCII 85 pos 0. /* character array */ gets(string). char H.33. char . A Sample Program to Get a String Then Print its ASCIIValues (with matrix) #include “stdio. In ‘C’ there are no limits to the matrix size.

i. char H. ASCII %d \n”. char G. First and Second Pass . points to. } } INPUT: HUGH<return> OUTPUT: pos 0. /* Input string from keyboard */ for(i = 0.C programming . char U. The computer remembers where the location of that variable is. This pointer is always hidden from the programmer.4 HOW A ‘C’ COMPILER WORKS • A ‘C’ compiler has three basic components: Preprocessor. In ‘C’. We may use some of the operations of ‘C’ to get the variable that the pointer. ASCII 72 pos 0. char H.h” main() { int i.11 • Pointers are a very unique feature of ‘C’. string[i]). string[i]. /* character pointer */ gets(string). char *string. ASCII 85 pos 0. This allows us to deal with variables in a very powerful way. A Sample Program to Get a String Then Print its ASCIIValues (with pointers): #include “stdio. the pointer to a variable may be used. ASCII 72 33. char %c.33. this memory of location is called a pointer. ASCII 71 pos 0. First recall that each variable uses a real location in memory. and uses it only in the background. string[i] != 0. i++){ printf(“ pos %d.

replace strings which have a defined value.lib) Executable Code (*. Library files (*. Object Code (*. and Linker.12 Compiler.h”) ASCII Text Code The First and Second Pass The compiler will parse the program and check the syntax. and remove unneeded characters. Source code “filename. #include files (like “stdio.C programming . TheSecond Pass produces some simple machine language.exe) . An optimization operation may be used to reduce execution time.obj) The Linker The compiler will combine the basic machine language from the first pass and combine it with the pieces of machine language in the compiler libraries. include programs.c” The Preprocessor Will remove comments.33. which performs the basic functions of the program.

33. and give it a block style. and defined constants make the purpose of the variable obvious.13 33.C programming . • Descriptive variable names. spaces and blank lines.5 STRUCTURED ‘C’ CODE • A key to well designed and understandable programs. • Use indents. • All declarations for the program should be made at the top of the program listing. . to make the program look less cluttered. • Comments are essential to clarify various program parts.

} exit(0).for(.h> #define COUNT 10 /* Number of counts in loop */ main() { int i.14 A Sample of a Bad Program Structure: main(){int i.i<10.33. /* counter */ for(i = 0. . } 33. i).6.6 ARCHITECTURE OF ‘C’ PROGRAMS (TOP-DOWN) 33.} A Good Example of the same Program: #include <stdio. Each function consists of programs written using the previous simpler functions.1 How? • A program should be broken into fundamental parts (using functions for each part) and then assembled using functions. i++){ /* loop to print numbers */ printf(“age:%d\n”.i). i < COUNT.C programming .i++)printf(“age:%d\n”.

15 Example with a Car Frame Suspension Wheel Body Frame is like one subroutine which also calls other subroutines like Suspension Engine • A Clear division should be maintained between program levels. and are thus much easier to locate. Functions are much easier to use. 33. It is much easier to find an error in a few lines of code. . • Try to isolate machine specific commands (like graphics) into a few functions. once written.2 Why? • A top-down design allows modules to be tested as they are completed. • Never use goto’s.C programming . errors tend to be associated with modules. • When programs are complete.33. than in a complete program. they are a major source of logic errors.6.

Define Problem . which are necessary for a complete solution to the problem. when we only need to change one function. calculations and figures. .This is the main code that decides when general operations occur. 3.C programming .Write out the relevant theory. Write Flow Program . as it is to change a function. • It is just as easy to change the overall flow of a program.The layout of the screen(s) must be done on paper. Define Objectives .Make a written description of what the program is expected to do.33. This is the most abstract part of the program. Application of ‘C’ to a CAD Program 33.7 CREATING TOP DOWN PROGRAMS 1.) 4. This description should include variables. Design User Interface . User options and help are also considered here.16 • Updates to programs are much easier. as outlined in the course notes. The method of data entry must also be considered. From this we make a list of required data (inputs) and necessary results (output). 2. and is written calling dummy ‘program stubs’. (There are numerous factors to be considered at this stage.

then the program is revised. as outlined in the ES488 course notes.C programming . and checked to make sure that it meets the objectives. This continues until all of the stubs are completed. 7. Golden Rule: If you are unsure how to proceed when writing a program.33.17 5. 6. Document . and a written description of functions should also be given. Expand Program .At this stage. Note: Always consider the basic elements of Software Engineering. 33. If any bugs are encountered. After the completion of any new function. Testing and Debugging. before you commit yourself to your programmed solution. the program is compiled. For Programmers. then work out the problem on paper. These functions will call another set of dummy ‘program stubs’.8 HOW THE BEAMCAD PROGRAM WAS DESIGNED . a top-down diagram can be drawn.The program operation is tested.The dummy ‘stubs’ are now individually written as functions. the operation of the program is formally described. and then retested. tested and debugged.

Beam Length. Weight.C programming . • The possible outputs are Cross Section Area. and be under a single point load.1 . and then providing immediate feedback. Load Force. • The inputs were determined to be few in number: Beam Type.33. and Beam Deflection.3 User Interface: 33. Load Position. Beam Thickness.3. The beam will be simply supported.18 33.8. Beam Width.8. Bending Stiffness. 33.1 Objectives: • The program is expected to aid the design of beams by taking basic information about beam geometry and material. a visual display of Beam Geometry. The program should also provide a printed report on the beam.8. Axial Stiffness.8. Beam Material. In this example it will not be given.2 Problem Definition: • The basic theory for beam design is available in any good mechanical design textbook. 33.Screen Layout (also see figure): . Beam Height. a display of Beam Deflection.

2 . and Beam Deflection(5).8. F2. F4. • The cursor keys could be used to cursor the input selector up or down. and updated. Beam Cross section(3). on a single screen. • The left side of the screen was for inputs. • In a separate Prompt Box(2). this input could be made. • Keys required: UP/DOWN Cursors.19 • The small number of inputs and outputs could all be displayed. ‘-’.3.’. numbers from ‘0’ to ‘9’.33. • Single keystroke operation. • The screen is divided into regions for input(2). F1.C programming . . 33. the right side for outputs. ‘. Numerical Results(4).Input: • Current Inputs were indicated by placing a box around the item on the display(1). input display and prompts(1).

with the prompt area replaced with the date and time.3. In the spirit of robustness it was decided to screen all other keys.C programming .8.3. material types.Output: • Equations. • For a printed report. screen information would be printed to a printer. • A default design was given.4 .Help: • A special set of help information was needed. and other relevant information were obtained from a text.3 . • Proper textual descriptions were used to ensure clarity for the user.8. It was decided to ensure that the screen always displays all information necessary(2). 33. calculations.33. which the user could modify.3.8. 33. 33. .Error Checking: • Reject any input which violates the input limits.20 and <RETURN>.5 .

21 • An error checking program was created. • Colors were used to draw attention.Miscellaneous: • The screen was expressed in normalized coordinates by most sub-routines. . and highlight areas.8. 33.C programming .6 . which gives error messages.3.33.

while((error = input()) != DONE) { if(error == REVISED) { screen(NEW). if((error = setup()) != ERROR) { screen(NEW). revision * and termination.C programming .33. } } error = NO_ERROR. */ { static int error. } kill(). if(error == ERROR) { printf(“EGA Graphics Driver Not Installed”). } 33.8. screen(UPDATE).8. 1989. screen(UPDATE).4 Flow Program: main() /* * EXECUTIVE CONTROL LEVEL * * This is the main terminal point between the * various stages of setup. input. * * January 29th.5 Expand Program: .22 33.

draws the beam cross section and deflection of beam.23 • The routines were written in a top down fashion.C programming . Routines Used In Package: • main() . . • setup() .to set up graphics mode and printer.33. These routines are listed below. For the sake of speed. or refresh part of the screen. this section will use low level commands. • calculations() . and thus speed up response. con- • Condition and error flags were used to skip unnecessary operations. this function uses some low level commands.perform the calculations of outputs from the inputs • picture() . in a time of about 30 hours. • screen() .A function to draw.A function which controls the main input loop for numbers. A response of more than 0.5 seconds will result in loss of attention by the user. In the interest of program speed.to be used as the main program junction. • input() .

input() and kill() directly. In this case main() will call setup().24 High Level Executive Subroutine Top main() screen() input() box() printer() Low Level Specific Subroutines Down setup() kill() picture() In this case we see that most of the routines are at the bottom of the design tree.33. This structure shows a clear division of tasks. to their basic parts. enter() draw_line() text() calculations() printes() Machine Dependence Increases Consideration of detail Consideration of flow 33.C programming . On the above diagram.8.6 Testing and Debugging: . screen(). only to the right. none of the functions calls any of the functions to the left of it.

1 . inputs and outputs.Users Manual: • The documentation included an Executive Summary of what the Program does. This took a couple of hours.7 Documentation 33.7.C programming . • Program Specifications were also given. • The theory for beam design was given for the reference of any program user. 33. This allowed the user to follow an example which displayed all aspects of the program. who wanted to verify the theory. • A manual was given which described key layouts.8. and possible use it. • The Objectives of the program were described. screen layout. basic sequence of operations.8. .25 • The testing and debugging was very fast.33. • A walk through manual was given. with only realignment of graphics being required.

7.33. • Written for turbo ‘C’ 33.Programmers Manual: • Design Strategy was outlined and given.8.26 33.9 PRACTICE PROBLEMS 1.8. and what do they do? . 33. What are the basic components of a ‘C’ compiler. • A complete program listing was given (with complete comments).C programming .8 Listing of BeamCAD Program. • Complete production of this Documentation took about 6 hours.2 .

27 2. You are required to consider that the two pieces are experiencing a single force. and how to do it. State your assumptions about the problem. 3. Describe some of the reasons for Using Top-Down Design. Each piece of sheet metal will have a hole drilled in it that is the size of the screw. then describe how you would produce this program with a Top Down design. You have been asked to design a CAD program which will choose a bolt and a nut to hold two pieces of sheet metal together.33.C programming . . What are some reasons for using ‘C’ as a programming language? 4.

d x = 10m Find the distance ‘d’.092903 )m 102. ask the instructor to show you how.3048m --------------------  1ft 2 2 m 100m + 25ft ( 0. • Good engineering practice demands that each number should always be accompanied with a unit.3048m ∴1 = -------------------1ft 2 2 2 100m + 25ft  0.1 34. but instead a minimal (but fairly complete) set is given.34.3048m 0. From the values below any conversion value can be derived. • A simple example of unit conversion is given below. 34.32m = 10. UNITS AND CONVERSIONS • Units are essential when describing real things.1 HOW TO USE UNITS • This table does not give an exhaustive list of conversion factors. If you are not sure about this.12m 2 2 2 .units .092903 ) -----2 ft 2 cancel out units 100m + 25 ( 0. **a simple unit conversion example: Given. keep the units in the equation d = dx + dy 2 2 d y = 5ft· ∴d = ∴d = ∴d = ∴d = ∴d = ∴d = ( 10m ) + ( 5ft ) 100m + 25ft 2 2 2 2 multiply by 1 From the tables 1ft = 0.

.g. e.2 HOW TO USE SI UNITS 1. = 1852 m = 1. (feet) = 12 in.609km 1 in.units .3 THE TABLE Major Division Distance 1 ft. 34. (inches) = 0. e. Use a slash or exponents. N•m 5.34. = 0. Use a dot in compound units. e. mega 2. (kg•m/s2) or (kg•m•s-2) 4. commas are avoided because their use is equivalent to decimal points in some cultures. e.g.560 ft. Beware upper/lower case letter in many cases they can change meanings. m = milli. There are two major variations US units are marked with ‘US’ and Imperial units are marked with ‘IMP’.g. N/cm or KN/m 3.540 cm 1 yd (yard) = 3 ft. Try to move prefixes out of the denominator of the units.150782 mi 1 micron = 10-6 m 1 angstrom = 10-10 m 1 mil = 10-6 m 1 acre = 43.4047 hectares . • In some cases units are non-standard.(inch) = 2.3048 m (meter) 1 mile = 1760 yards = 5280 ft = 1. 1 nautical mile = 6080 ft. Use spaces to divide digits when there are more than 5 figures.2 34...g.

785 l = 4 quarts = 8 pints = 16 cups 1 liter (l) = 0.000 m2 1 Hectare (ha) = 10.000 dyne 1 dyne = 2.040842 IMP oz = 2 tablespoons = 6 teaspoons 1 IMP gal.559. drams = 456.248*10-6 lb. 1 US gal.units .0129 drops = 480 US minim = 1.448N .2365882 l = 8 USoz 1 US oz = 8 U.201 U.8689762 knot Angle 1 rev = 2PI radians = 360 degrees = 400 gradians 1 degree = 60 minutes 1 minute = 60 seconds Volume 1 US gallon = 231 in3 1 CC = 1 cm3 1 IMP gallon = 277. (ounce) = 4.1 pints (pt) = 1.34. 1 US pint = 16 US oz 1 IMP pint = 20 IMP oz 1tablespoon = 0.274 in3 1 barrel = 31 IMP gal.000 m2 1 Hectare (ha) = 10.06 quarts (qt) = 0.000 m2 1 Hectare (ha) = 10.S.3 1 furlong = 660 ft 1 lightyear = 9.000 m2 Velocity 1 mph = 0.5 US gal. gal.66 ft2 1 Hectare (ha) = 10. = 31. = 3.085678e16 m Area 1 acre = 43. 1 bushel = 32 quarts 1 peck = 8 quarts Force/Mass 1 N (newton) = 1 kg•m/s2 = 100.5 oz.460528e15 m 1 parsec = 3.26 gallons (gal) 1 qt (quart) = 0.2046 lb 1lb = 16 oz.9464 l 1 cup (c) = 0.81 N (on earth surface) = 2. = 1.S.001 m 3 = 2. (pound) 1 kg = 9.

1 W = 1 J/s .) =760 mmHg at 0°C=14.03108 slug 1 kip = 1000 lb.7 W (watts) = 2. = 550 ft. Hg at 32F = 2. Hg at 62F 1 microbar = 0. 1 slug = 14. (horsepower) = 745.35 g (gram) = 0. = 28.2 kg 1 metric tonne = 1000 kg 1 troy oz = 480 grain (gr) 1 g = 5 carat 1 pennyweight = 24 grain 1 stone = 14 lb 1 long ton = 2240 lb 1 short ton = 2000 lb Pressure 1 Pascal (Pa) = 1 N/m2 = 6.1 N/m2 Scale/Magnitude atto (a) = 10-18 femto (f) = 10-15 pico (p) = 10-12 nano (n) = 10-9 micro (µ) = 10-6 milli (m) = 10-3 centi (c) 10-2 deci (d) = 10-1 deka (da) = 10 hecto (H) = 102 kilo (K) = 103 mega (M) = 106 giga (G) = 109 tera (T) = 1012 peta (P) = 1015 exa (E) = 1018 Power 1 h.223 lb/in2=1.0416 in.545 BTU/hr.4 1 oz.895 kPa 1 atm (metric atmos. 1 ft•lb/s = 1.356 W 1 J (joule) = 1 N•m = 107 ergs = 0.0132*105 N/m2 1 psi = 2.2780N 1 lb = 0./sec.units .2389 cal.0355 in.p.59 kg 1 imperial ton = 2000 lb = 907.lb.34.

34.7 °F = 0 K = 0 R = absolute zero 0 °C = 32 °F = 273.62x10-34 J. pressure) 1 therm = 100.02x1023 atoms/atomic weight density of water = 1 g/cm3 electron charge = 1.5 1 ev = 1.000 BTU Mathematical π radians = 3.666 0.252 calories = 1 BTU (British Thermal Unit) -273.7 R = Water Freezes 100°C = 212°F = 373.sec Avagadro’s number = 6.718 Time 1 Hz (hertz) = 1 s-1 1 year = 365 days = 52 weeks = 12 months 1 leap year = 366 days 1 day = 24 hours 1 fortnight = 14 days 1 hour = 60 min.3 K = 671.units .2 °C = -459. 1 min = 60 seconds 1 millenium = 1000 years 1 century = 100 years 1 decade = 10 years Physical Constants R = 1. F = Fahrenheit K = Kelvin Rankine (R) = F .459. °C = Celsius (Centigrade). e = 2.3 K = 491. 1 cfm (cubic foot per minute) = 1 ft3/min.11*10**-31 kg proton rest mass = 1. electron rest mass = 9.3x10-23 J/K h = Planck’s constant = 6.987 cal/mole K = ideal gas law constant K = Boltzmann’s constant = 1.62x10 -27 erg-sec = 6.60x10-19 coul. 1 rpm (revolutions per minute) = 60 RPS (Revolutions per second) = 60Hz 1 fps (foot per second) = 1 ft/sec 1 mph (miles per hour) = 1 mi.3x10-16 erg/K = 1.5 cycles 1 Hz = 1 cycle/sec.60219*10-19 J 1 erg = 10-7 J Temperature °F = [(°C*9)/5]+32.67*10**-27 kg ./hr.1416 radians = 180 degrees = 0.7 R = Water Boils (1 atm.

The values shown only go up to 127.4 ASCII.05564 lux = 43.6 speed of light (c) = 3.67*10**-11 Nm**2/kg**2 permittivity of free space = 8. binary. ASCII values above this are not commonly used in robust applications.05564 meter-candles illumination = lux resistance = ohm 34.00x1010 cm/sec speed of sound in dry air 25 C = 331 m/s gravitational constant = 6.34. BINARY CONVERSION • The table below will allow conversions between decimal.26*10**-6 henry/m mean radius of earth = 6370 Km mass of earth = 5. hexadecimal.units . and ASCII values. HEX.85*10**-12 farad/m permeability of free space = 1.004*PI oersted electric flux density = coulomb/m**2 capacitance = farad permeability = henry/m electric field strength = V/m luminous flux = lumen luminance = candela/m**2 1 flame = 4 foot candles = 43.98*10**24 kg Electromagnetic magnetic flux = weber (We) = 10**8 maxwell inductance = henry magnetic flux density = tesla (T) = 10**4 gauss magnetic intensity = ampere/m = 0. .

34.7 hexadecimal hexadecimal decimal decimal ASCII 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 0 1 2 3 4 5 6 7 8 9 A B C D E F 10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 00000000 00000001 00000010 00000011 00000100 00000101 00000110 00000111 00001000 00001001 00001010 00001011 00001100 00001101 00001110 00001111 00010000 00010001 00010010 00010011 00010100 00010101 00010110 00010111 00011000 00011001 00011010 00011011 00011100 00011101 00011110 00011111 NUL SOH STX ETX EOT ENQ ACK BEL BS HT LF VT FF CR S0 S1 DLE DC1 DC2 DC3 DC4 NAK SYN ETB CAN EM SUB ESC FS GS RS US 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 20 21 22 23 24 25 26 27 28 29 2A 2B 2C 2D 2E 2F 30 31 32 33 34 35 36 37 38 39 3A 3B 3C 3D 3E 3F 00100000 00100001 00100010 00100011 00100100 00100101 00100110 00100111 00101000 00101001 00101010 00101011 00101100 00101101 00101110 00101111 00110000 00110001 00110010 00110011 00110100 00110101 00110110 00110111 00111000 00111001 00111010 00111011 00111100 00111101 00111110 00111111 space ! “ # \$ % & ‘ ( ) * + . / 0 1 2 3 4 5 6 7 8 9 : . < = > ? ASCII binary binary . .units .

l arr.34.8 hexadecimal hexadecimal decimal decimal ASCII 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 40 41 42 43 44 45 46 47 48 49 4A 4B 4C 4D 4E 4F 50 51 52 53 54 55 56 57 58 59 5A 5B 5C 5D 5E 5F 01000000 01000001 01000010 01000011 01000100 01000101 01000110 01000111 01001000 01001001 01001010 01001011 01001100 01001101 01001110 01001111 01010000 01010001 01010010 01010011 01010100 01010101 01010110 01010111 01011000 01011001 01011010 01011011 01011100 01011101 01011110 01011111 @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ yen ] ^ _ 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 60 61 62 63 64 65 66 67 68 69 6A 6B 6C 6D 6E 6F 70 71 72 73 74 75 76 77 78 79 7A 7B 7C 7D 7E 7F 01100000 01100001 01100010 01100011 01100100 01100101 01100110 01100111 01101000 01101001 01101010 01101011 01101100 01101101 01101110 01101111 01110000 01110001 01110010 01110011 01110100 01110101 01110110 01110111 01111000 01111001 01111010 01111011 01111100 01111101 01111110 01111111 ‘ a b c d e f g h i j k l m n o p q r s t u v w x y z { | } r arr. and may vary between machines. ASCII binary binary .5 G-CODES • Note that G and M codes not universal standards.units . the manuals for the machine should be checked. 34. In any case of doubt.

return from reference point G31 .Pause (for operator intervention) G08 .return to reference point G29 .Deceleration G17 .Stop on input (INROB1 is high) (Prolight Mill) G33-35 .thread cutting cycle or metric data specification G24 .face turning cycle G25 .Rapid move (not cutting) G01 . positive G44 .wait for input #2 to go low (Prolight Mill) G36 .1 .fixed cycle cancel .Linear move G02 .set inch based units or finishing cycle G71 .wait for input #2 to go high (Prolight Mill) G40 .disable 360 deg arcs (Prolight Mill) G75 .tool length compensation.wait for input #1 to go low (Prolight Mill) G26 .units .y-z plane for circular interpolation G20 .facing cycle contour (Emco Lathe) G74.Counterclockwise circular motion G04 .3D circular interpolation counter clockwise (Prolight Mill) G74 .Dwell G05 .cutter compensation to the left G42 . These direct motion of the tool.pattern repeating (Emco Lathe) G75.cut-in cycle in x-axis G78 .thread cutting functions (Emco Lathe) G35 .multiple threading cycle G80 .cutter compensation to the right G43 . G00 . cut cycle in z-axis G77 .cutter compensation cancel G41 .34.9 • A basic list of ‘G’ operation codes is given below.enable 360 degree arcs (Prolight Mill) G76 .deep hole drilling.tool length compensation.wait for input #1 to go high (Prolight Mill) G28 .Preset position G70 .set metric units or stock removal G72 .3D circular interpolation clockwise (Prolight Mill) G73 .turning cycle contour (EMCO Lathe) G73 . negative G50 .z-x plane for circular interpolation G19 .Clockwise circular motion G03 .Acceleration G09 .indicate finishing cycle (EMCO Lathe) G72 .turning cycle or inch data specification G21 .x-y plane for circular interpolation G18 .1 .

open chuck (EMCO Lathe) M25 .boring cycle (Prolight mill) G86 .Feed rate in ipr (EMCO Lathe) G96 .boring cycle with dwell (Prolight Mill) G90 .taping cycle (EMCO Lathe) G85 .program stop M01 .drilling cycle (EMCO Lathe) G83 .tailstock forward (EMCO Lathe) M22 .turn on accessory #2 (120VAC outlet) (Prolight Mill) M11 . M00 .optional stop using stop button M02 .units .mist with coolant M08 .turn off accessory #2 (120VAC outlet) (Prolight Mill) or tool change M17 .absolute dimension program G91 .peck drilling cycle (Prolight Mill) G84 .incremental dimensions G92 .fixed cycles specified by machine tool manufacturers G81 .withdraw the tool to a safe plane or feed per revolution G101 .boring with spindle off and dwell cycle (Prolight Mill) G89 .end of program M03 .Surface cutting speed (EMCO Lathe) G97 .Rotational speed rpm (EMCO Lathe) G98 .set output #1 off (Prolight Mill) .Chain to next program (Prolight Mill) M21 .spindle on CCW M05 .Spline interpolation (Prolight Mill) • M-Codes control machine functions and these include.withdraw the tool to the starting point or feed per minute G99 .turn off accessory #1 (120VAC outlet) (Prolight Mill) M10 .reaming cycle (EMCO Lathe) G85 .coolant off M09 .Spindle speed limit G93 .turn on accessory #1 (120VAC outlet) (Prolight Mill) M09 .subroutine end M20 .spindle on CW M04 .flood with coolant M08 .10 G81-89 .tool change M07 .Write current position to data file (Prolight Mill) M25 .spindle off M06 .tailstock back (EMCO Lathe) M20 .Coordinate system setting G94 .drilling cycle (Prolight Mill) G82 .Feed rate in ipm (EMCO Lathe) G95 .straight drilling cycle with dwell (Prolight Mill) G83 .34.

arc angle.close chuck (EMCO Lathe) M26 . not ipr).x-axis center for circular interpolation.a feed value (in ipm or m/s. or thread pitch Innn .subprogram call M99 .11 M26 .relative motion in x Vnnn .an orientation. .return from subprogram.y-axis center for circular interpolation.34.an orientation. loop counter and program cycle counter Nnnn . or a fixed number of times (Prolight Mill) M71 .a clearance plane for tool movement.peck depth for pecking cycle Snnn .restart a program continuously.an x-axis value Ynnn .an orientation. or second z-axis spline control point.z-axis center for circular interpolation.cutting speed (rpm).end of tape (rewind) M35 .compensate for sharp external curves M98 .puff blowing on (EMCO Lathe) M72 . or taper value Qnnn . or first z-axis spline control point Lnnn .units . or arc radius.a tool number Unnn . or second x-axis spline control point Bnnn .move z-axis home (Prolight Mill) • Other codes and keywords include.set output #2 on (Prolight Mill) M38 . spindle speed Tnnn .a sequence/line number Onnn . or first x-axis spline control point Jnnn . or chamfer Fnnn .relative motion in y Wnnn .puff blowing off (EMCO Lathe) M96 . Annn .relative motion in z Xnnn .set output #1 on (Prolight Mill) M30 .a z-axis value .subprogram reference number Rnnn .starts a comment (proLight Mill).compensate for rounded external curves M97 .move y-axis home (Prolight Mill) M103 .put stepper motors on low power standby (Prolight Mill) M47 . or end of block (EMCO Lathe) . or first y-axis spline control point Knnn .move x-axis home (Prolight Mill) M102 . jump instruction M101 .a y-axis value Znnn .set output #2 off (Prolight Mill) M36 . or second y-axis spline control point Cnnn .subprogram block number Pnnn .

33 35-500 100-550 1-45 2.57 55.comp. ATOMIC MATERIAL DATA density g/cm3 8.01 47.33 2.86 2.37 28.7 73 70 72 36-41 26 26 0.09 12.shear ten .6-2.97 183.1 35. stress Material Density poisson Type G E aluminum typical 1100-h14 2024-T6 5456-h116 2014-T6 6061-T6 7075-T6 2.8 2.yield stress ten .comp.96 7. stress .ult.yield strain ten .materials . ult.8 70-79 26-30 0.13 2.1 4. thermal expansion 23 ult.9 Valances Hw 1/2 2/3 3 6/8 2 4 2/4 3/4 F = 96.500 coulombs 35.67 19.85 65.3 7.8 2. stress .33 0.51 Element copper iron aluminum tungsten zinc silicon carbon titanium Atomic # 29 26 13 74 30 14 6 22 Atomic weight 63.35.33 0.33 410 270 480 480 310 550 13 17 11 10-6/deg C (Mg/m3) (MPa) (MPa) (GPa) (GPa) (%) .85 26. MECHANICAL MATERIAL PROPERTIES Table 1: yield stress .

1-0.1-0.330.4 1.32 1701100 100-620 4501200 310-760 2-50 Nickel 8.2 Table 1: yield stress .3 82-690 120 200-830 69-480 5-60 0-1 Concrete typical reinforced lightweight 2.121.2-8.4-8.27 30-1000 0 Plate Fibers 70 700020000 1.0-7.617.2 18-21 3401400 10-70 9.comp.1-1. thermal expansion 19.170.ult.8 7.36 55-760 230-830 4-50 Copper typical annealed hard drawn Glass typical 2.comp.9 17-31 17-31 17-31 110-120 na na na 40-47 0.761.2 0.8 210 80 0.30% Cu 170 66 0.9-12 7-14 16.8 8.2-0.83 8.6 96-110 36-41 0.8 48-83 19-35 0.3 2.6 5-11 26.34 70-550 200-620 4-60 Yellow Brass cold rolled annealed Red Brass cold rolled annealed Bronze Cast Iron typical typical 8.materials . stress Material Density poisson Type G E Brass typical 8. stress .35 80-280 140-340 2-20 Magnesium typ.4-2.1-0.shear ten .2 0.8 14 13 ult.8 41-45 15-17 0.yield stress ten . ult.2 0.4 96-120 83-170 36-44 32-69 0.yield strain ten . stress .128.34 0.31 2-50 10-6/deg C (Mg/m3) (MPa) (MPa) (GPa) (GPa) (%) .35. alloys Monel 67%Ni.

3 Quartz 2.9 20-70 na 0.004 0.6-2.9 40-100 na 0.4 0.50 Sand.materials .4 na na 0.35.ult.9 40-100 na 0.6-2.4 2.0-2.2-0.3 Table 1: yield stress .9 40-100 na 0.3 Limestone 2.soil.comp.96-1.2-0.4 40-80 7-28 20-100 15-300 Marble 2.6-2. stress .96-1.3 Rubber typical .001 0. stress Material Density poisson Type G E Plastics Nylon Polyethylene Rock .4 0.9 20-70 na 0.00070.2 7.30 cold rolled annealed high strength 3401000 340-700 4001600 280-700 5501200 550-860 7001900 4001000 900 340-830 5-25 machine spring 5-25 3-15 stainless 5-40 tool structural astm-a48 520 200-700 8 10-40 10-6/deg C (Mg/m3) (MPa) (MPa) (GPa) (GPa) (%) . thermal expansion 70-140 140-290 5-9 Granite 2.1 .3 50280 50280 50280 20200 20200 1-7 7-20 100-800 130-200 10-18 14 12 ult.yield strain ten .2-0.00020.1-3.2-2.comp.yield stress ten .0-2.7-1.88-1.85 190-210 75-80 0.2-0.shear ten . ult.2-0.3 Sandstone 2.3 0. stress .450.270.gra vel Steel 1.

yield stress ten .comp.comp.5 100-120 39-44 0. stress Material Density poisson Type G E astm-a47 astm-a36 astm-a5242 astm-a441 astm-a572 astm-a514 wire 340 700 2801000 500 830 5501400 20 15 5-40 250 400 30 Stainless Steel Titanium aisi 302 typ.1-11 4.9 340-380 140-160 0.33 7601000 9001200 14004000 10 Tungsten 1. stress .4 Table 1: yield stress .56-. ult.02 Wood (dry) Douglas fir Oak Southern pine .72 .3 ult.0 1.35.materials .56 .2 0-4 Water fresh sea 1.64 11-13 11-12 11-14 30-50 40-60 40-60 50-80 50-100 50-100 30-50 30-40 30-50 40-70 30-50 40-70 35.yield strain ten . alloys 4.1 FORMULA SHEET • A collection of the essential mechanics of materials formulas are given below 10-6/deg C (Mg/m3) (MPa) (MPa) (GPa) (GPa) (%) . stress .shear ten .64-.ult. thermal expansion 12 12 12 17 8.48-.

materials .35.– ν ----.( r o – r i ) 2 P = 2πfT 4 (cylinder) (hollow tube) .– ν ----z E E E E = 2G ( 1 + ν ) Torsion Tc τ max = ----J TL φ = -----JG σ ten = – σ comp = τ γ ε = -2 τ max γmax = ---------G πr J = ------2 π 4 4 J = -.5 Axial/Normal Stress/Strain P σ = εE = -A PL δ = εL = -----AE Shear Stress/Strain τ = γG τ xy γ xy = -----G Poisson’s ratio ε lateral ν = – -------------------------ε longitudinal σx σy σ ε x = ----.

0 (both ends pinned) k=2.1 0.9 1. one free) K 3.materials .0 P d/2 r P d/2 2.2 0.0 r/d .5 0.7 0.5 (both ends fixed) k=0.8 0.5 2.5 0. one pinned) k=1.0 (one end pinned.6 0.7 (one end fixed.0 1.4 0.3 0.(for rectangle) 12 πr I = ------.35.(for circle) 2 c ε max = -ρ EI ρ = ----M 4 3 Buckling EI ----------P<π 2 2 k L 2 k=0.6 Beams My σ = – ------I VQ τ = ------Ib L = ρθ My σ = – ------I bh I = -------.

9 1.3 0.25 D/d = 1.35.5 0.6 0.5 P 2.1 r/d 0.5 D/d = 2.materials .7 0.8 0.2 0.0 .0 D/d = 1.0 1.0 D r d P 1.7 K 2.1 0.4 0.5 D/d = 1.

materials .01 0.0 D/d = 1.33 D/d = 1.1 1.02 0.0 .03 0.0 1.06 0.8 K T D 2.05 0.35.09 0.5 D/d = 2.2 D/d = 1.07 0.1 r/d 0.04 0.5 r d T 2.08 0.

.. Second Edition. "Applied Nonlinear Control".1. 1990. 36. Topics/Chapters: Introduction Phase plane analysis Fundamentals of Lyapunov theory Advanced stability theory Describing function analysis Feedback linearization Sliding control Adaptive control Control of Multi-input physical systems Comments: Directed to mechanical undergraduate students. W. "Feedback Control Systems". E.1 TEXTBOOKS 36.1 Slotine and Li Citation: Slotine. 1991.1.1 36. Li. BIBLIOGRAPHY 36.2 VandeVegte Citation: VandeVegte. Topics/Chapters: Linearized dynamic models Transfer function models of physical systems Modeling of feedback systems and controllers Transient performance and the s-plane The perfoarmance and dynamic compensation of feedback systems The root-locus method . Prentice Hall.bibliography . J. John. Prentice Hall.36.

. pneumatics. Chapters 9-10 present z-transform methods for controller analysis.bibliography . some design issues are presented.2 Frequency response analysis Frequency response design Digital control systems Digital control system analysis and design State space analysis Introduction to state space design Multivariable systems in the frequency domain Nonlinear control systems App: Vectors matrices and determinants App: Computer aids for analysis and design Comments: Directed to mechanical undergraduate students. Chapters 1-8 use classical design techniques. fluids.36. Chapters cover a range of topics systems including motors. thermo. etc.

( r a + l a D ) + K E ω KT ( JDω + C d ω + T L ) v a = -----------------------------------------------.1 37. TOPICS IN DEVELOPMENT Topics: Objectives: 37.input output equations .( r a + l a D ) + K E ω KT ( JDω + C d ω + TL ) ( r a + la D ) + K E K T ω = K T v a Jr a Dω + C d r a ω + T L r a + Jl a DDω + l a DC d ω + l a DT L + K E K T ω = K T v a · 2 · ·· ω ( Jl a ) + ω ( Jr a + la ωC d ) + ω ( K E K T ) + ω ( C d r a ) = v a ( K T ) + T L ( – r a ) + T L ( – l a ) 2 2 2 2 2 2 .1 UPDATED DC MOTOR MODEL ea = KE ω Te = KTia T e – T L – C d ω = JDω T e = JDω + C d ω + TL v a = r a i a + la Di a + e a v a = i a ( ra + l a D ) + ea v a = i a ( ra + l a D ) + KE ω Te v a = ----.37.

0 ( Jl a ) + 0 ( Jr a + la ωC d ) + ω ( K E K T ) + 0 = v a ( K T ) + T L ( – r a ) + 0 ( – l a ) ω ( KE KT ) = va ( KT ) + T L ( –ra ) ra ω ( K E ) + TL  ----. 0 ( Jl a ) + 0 ( Jr a + l a ωC d ) + 0 ( K E K T ) + 0 ( C d r a ) = v a ( K T ) + TL ( – r a ) + 0 ( – la ) KT T L = v a  ----.2 Measure the motor resistance. r  a 2 2 2 . 0 ( Jl a ) + 0 ( Jra + l a ωC d ) + 0 ( K E K T ) + 0 ( C d r a ) = v a ( K T ) + T L ( – r a ) + 0 ( –l a ) T L ( ra ) = va ( KT ) T L ( ra ) K T = --------------va At a steady state speed. = v a K  T The static friction limit. 0 ( Jla ) + 0 ( Jr a + l a ωC d ) + ω ( K E K T ) + ω ( C d r a ) = v a ( K T ) + TL ( – r a ) + 0 ( – la ) ω ( K E K T ) + ω ( C d ra ) = v a ( K T ) + TL ( – ra ) ω ( Cd ra ) + ω ( KE KT ) + ( T L ( ra ) – va ( KT ) ) = 0 At lower steady state speeds (little drag).input output equations .37. r a = measured When stalled a voltage is applied and the torque is measured.

37.3 2 · ·· · ω + ω A + ωB + ω C = v a D + T L E + T L F .input output equations .

2 ANOTHER DC MOTOR MODEL TL + 1 -------------------JD + Bm ωr vs + - 1 -------------------La D + r a ia Ka Te Ka where.input output equations .37. K a = motor speed constant L a = motor armature inutne dc c a r a = motor armature resistance J = inertia of rotor Bm = damping of rotor v s = voltage applied to armatur e i a = armature current T e = effective mo r ru t t qe oo T L = load torque ω r = motor speed .4 37.

ia 1 1 1 --. allowing the simplified system block diagram. This allows the calculation of the motor speed coefficient directly.= -------------------------. the load and effective torques are equal.5 Motor parameters can be measured by applying a voltage and holding (stalling) the shaft.= -----------------. and the current are measured.input output equations .= -----------------L a s + ra vs XXXXXXXXXXXXXXX . The stalled motor torque.= ---L a s + ra vs L a ( j0 ) + r a ra v r a = ---s ia The inductance of the windings can be found using the time constant . When stalled.37.phase shift ???????????????? ia 1 --. and the feedback loop is zero. but can also be verified by using the steady state current and voltages. vs 1 -------------------La D + r a ia Ka TL TL K a = ----ia The resistance of the windings can be found by direct measurement with a DMM.

.. the derivatives of motor speed will become zero.– ----r a ω r ra 2 2 2 2 2 The rotor inertia can be found using .. vs + 1 -------------------L a D + ra ia Ka Te 1 -------------------JD + B m ωr Ka D ω r ( La J ) + Dω r ( L a Bm + Jr a ) + ω r ( r a Bm + K a ) = Vs K a ( 0 ) ( L a J ) + ( 0 ) ( La Bm + Jr a ) + ω r ( r a Bm + K a ) = Vs K a ωr ( ra Bm + Ka ) = V s K a Bm V s Ka Ka = ----------....input output equations .. .... If the system is allowed to reach steady state....6 If the motor is unloaded (TL = 0) the following differential equation can be derived....37.... D ω r ( La J ) + Dω r ( La Bm + Jr a ) + ω r ( r a Bm + K a ) = Vs K a 2 2 • Another approach to finding the motor parameters is system identfication..

--------J J + vs –1 TL 0 ----J .a = dt ω r ia K a – Bm ωr ----.input output equations .37.i ---. + T L  ----- + ω r  --------. + i a  ------- L  L  L  a a a 1 ω r = ( K a i a – T L ) -------------------JD + B m Dω r ( J ) + ω r ( Bm ) = K a ia – T L Ka – Bm –1 Dω r = i a  ----.7 TL + 1 -------------------JD + Bm ωr vs + - 1 -------------------La D + r a ia Ka Te Ka 1 i a = ( v s – K a ω r ) -------------------L a D + ra Di a ( L a ) + i a ( r a ) = v s – K a ω r –Ka –ra 1 Di a = v s  ----- + ω r  -------.-------L a La 1 ----.  J  J  J  – r a –K a ------.0 La d.

-------La L a ia 1 ----.0 La vs vs i = A B a + E 0 0 F TL C D ωr Ka –B m ωr ----.i ---. d.input output equations .8 The matrix can be parameterized.3 STATE SPACE IDENTIFICATION d---.a = dt ω r –r a –Ka ------.Yout = CXout + DXin dt .X out = AX out + BXin dt d---.--------J J + – 1 TL 0 ----J 1 E = ----La –ra A = ------La –Ka B = -------La Ka C = ----J – Bm D = --------J La = E r a = – AL a = – AE K a = – BL a = – BE Ka – BE J = ----.= ---------C C – BED Bm = -------------C 37.37.

x = v dt F d---.input output equations .9 Given a system model for a force applied to a mass.x = 0 1 x + F dt v ---00 v M .37.v = ---M dt 0 d---. d---.

xi – x i – 2 xi – x i – 2 -------------------. d---.i = a b i + e F dt v c d vi f i The state variables can be replaced with backwards difference equations.x = a b x + e F dt v c d v f x d.10 For an unknown second order linear system.37.= ax i – 1 + b -------------------.+ eF T T xi – xi – 2 xi – xi – 2 a ( x i – 1 ) + b  -------------------- + c ( 0 ) + d ( 0 ) + e ( F ) + f ( 0 ) = ------------------- T T  xi – 3 – xi – 5 xi – 3 – xi – 5 a ( x i – 4 ) + b  --------------------------- + c ( 0 ) + d ( 0 ) + e ( F ) + f ( 0 ) = --------------------------  T T xi – 6 – xi – 8 xi – 6 – xi – 8 a ( x i – 7 ) + b  --------------------------- + c ( 0 ) + d ( 0 ) + e ( F ) + f ( 0 ) = --------------------------  T T x i – 2x i – 1 + x i – 2 xi – x i – 2 ---------------------------------------.x ---. xi – xi – 2 -------------------xi – 1 T a b e F = xi – xi – 2 + x i – 2x i – 1 + x i – 2 c d -------------------f ---------------------------------------T 2 T The equations can then be written for three different set of points to yield six equations.+ fF 2 T T x i – 2x i – 1 + x i – 2 x i – xi – 2 a ( 0 ) + b ( 0 ) + c ( x i – 1 ) + d  -------------------- + e ( 0 ) + f ( F ) = ---------------------------------------- 2  T T x i – 3 – 2x i – 4 + x i – 5 x i – 3 – xi – 5 a ( 0 ) + b ( 0 ) + c ( x i – 4 ) + d  --------------------------- + e ( 0 ) + f ( F ) = ----------------------------------------------2   T T x i – 6 – 2x i – 7 + x i – 8 x i – 6 – xi – 7 a ( 0 ) + b ( 0 ) + c ( x i – 7 ) + d  --------------------------- + e ( 0 ) + f ( F ) = ----------------------------------------------2   T T .input output equations .= cx i – 1 + d -------------------.

5 x ( 6 ) = 18 x ( 7 ) = 24.t --.5 xi – 6 = 2 x i – 5 = 4.5 x(4) = 8 x ( 5 ) = 12.input output equations .0 F ----------------------------------------------T 2 T Some sample data points can be calculated for F=1 and M=1. 2 2 x = 1 at = 1 ---.0 F T e T x i – 3 – 2x i – 4 + x i – 5 f xi – 3 – x i – 5 -----------------------------------------------------------------------.37.5 x(2) = 2 x ( 3 ) = 4.5 x i = 32 M = 1 .11 The equations can be rewritten in matrix form.5 x i – 2 = 18 x i – 1 = 24. xi – 1 xi – 4 xi – 7 0 0 0 xi – xi – 2 -------------------T xi – 3 – x i – 5 -------------------------T xi – 6 – x i – 8 -------------------------T 0 0 0 0 0 0 xi – 1 xi – 4 xi – 7 xi – xi – 2 -------------------0 F 0 T xi – 3 – xi – 5 --------------------------0 F 0 T a xi – 6 – xi – 8 b --------------------------0 F 0 T c = x i – 2x i – 1 + x i – 2 xi – xi – 2 d ---------------------------------------2 -------------------.5 x ( 8 ) = 32 xi – 8 = 0 x i – 7 = 0.F .0 F 2 T T xi – 6 – x i – 7 x i – 6 – 2x i – 7 + x i – 8 -------------------------.T = 1 F = 1 2 2M x(0) = 0 x ( 1 ) = 0.5 xi – 4 = 8 x i – 3 = 12.

5 – 2 ( 8 ) + 4.0 1 2 – 0.5 – 4.37. x d.i = a b i + e F dt v c d vi f i xi x = 1 0 i + 0 0 F 0 0 vi 0 0 1 0 state equation output equation This can be incorporated as a realtime algorithm to calculate the data parameters based upon realtime updates.5 ---------------------------------------1 f 2 1 2 – 2 ( 0.5 ) + 18 -----------------------------------------2 1 d 1 e 12.5 0 f 14 8 2 1 1 1 a 0 b 1 c = 0 d 0 e 0 f 1 Substituting these values back into the state equation results in the following expression.5 14 0 0 1 8 8 0 0 1 b c = 0. which matches the exact state equation.5 ----------------1 12.5 32 – 18 0 ----------------1 12.5 8 ----------------------.5 14 0 d e 0 0 8 8 0 0 0 0.5 ----------------------0 1 a 2–0 b ----------0 1 c = 32 – 2 ( 24.5 0.5 8 ----------------------1 2–0 0.5 2 0 0 1 0 0 24.5 ----------1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 32 – 18 ----------------1 12.5 ) + 0 1 --------------------------------1 2 24.input output equations .5 – 4.0 1 0 0 0 1 1 1 –1 a 24.x ---.5 1.5 – 4.12 The values can be substituted into the matrix. . and the matrix solved 32 – 18 24.5 --------------.

input output equations .0 F T xi – 6 – xi – 7 x i – 7 -------------------------.37.0 F T xi – 3 – xi – 5 x i – 4 -------------------------.0 F T xi – x i – 2 -------------------T xi – 3 – xi – 5 -------------------------T xi – 6 – xi – 8 -------------------------T x i – 2x i – 1 + x i – 2 ---------------------------------------2 T x i – 3 – 2x i – 4 + x i – 5 ----------------------------------------------2 T x i – 6 – 2x i – 7 + x i – 8 ----------------------------------------------2 T This can then be used as a model to calculate a transfer function as shown below using the previous example.13 xi – 1 xi – 4 a b x c = i–7 d 0 e f 0 0 xi – xi – 2 -------------------T xi – 3 – xi – 5 -------------------------T xi – 6 – xi – 8 -------------------------T 0 0 0 –1 0 0 0 0 0 0 F 0 F 0 F 0 xi – xi – 2 x i – 1 -------------------. .

14 The state equation variables are.37.   G = C ( sI – A ) B + D = 1 0  s 1 0 – a b  0 0  0 1 c d –1 –1 e + 0 0 f 0 0   G = 1 0  s – a –b  0 0  –c s – d  –1 s–d b     0 0 ( s – d )e + bf G =  -------------------------------------------- e = ------------------------------------------- ( s – a ) ( s – d ) + bc f ( s – a ) ( s – d ) + bc     ( s – a ) ( s – d ) + bc –1 G = -------------------------------------------( s – d )e + bf s–d b     c s–a e = 1 0  -------------------------------------------- e f 0 0  ( s – a ) ( s – d ) + bc f     A block diagram for the system may then be written. x d. G + xd - –1 + + controller F plant xa 1 .x ---.input output equations . and its inverse.i = a b i + e F dt v c d vi f i xi A = a b c d B = e f x = 1 0 i + 0 0 F C = 1 0 D = 0 0 0 0 vi 0 0 1 0 0 0 0 0 these can be used to calculate a SISO system transfer function.

F d.15 Consider the controller with the previous example.= s ( s – d )e + bf ( s – 0 )0 + 1 ( 1 ) The identified coefficients for the controller are substituted into the inverse transfer func.= ------------------------------------------------.= F - dt M 1 1 x -. a 0 b 1 c = 0 0 d 0 e 1 f G –1 ( s – a ) ( s – d ) + bc (s – 0)(s – 0 ) + 1(0) 2 = -------------------------------------------. s + xd 1 P 2 + + F 1 ---2 s xa .= ---2 F s A state diagram for the system may then be written using a proportional controller.2  ---- x = ---.input output equations . The identified coefficients for the controller are substituted into the inverse transfer func.37.

16 The state diagram is simplified. such as sensor noise. . This also assumes that the force actuator can generate any requested force.37.+ 1 P P  s + P  -------------- ------------- P  2  s +P 1 2 2 xa xd xa xd xa Note: This suggests that the result is an ideal controller. imperfect actuators. however in practice there are errors.input output equations . s + xd 1 s ---P + xd 1 s ---P + xd 2 2 2 1 -P + + P 1 --2 s xa + P ---2 s xa + P -------------2 s +P P -------------2 s +P xa xd s ---. and changing system conditions such as mass.

input output equations . 37. 1 1–e G h ( s ) = -----------------s – Ts t 0 T .t.4 BLOCK DIAGRAMS AND UNITS .5 SIGNAL FLOW GRAPHS .these invert block diagrams.6 ZERO ORDER HOLD .17 37.each block in a block diagram has an input and output. Making sure that the units are consistent is useful for detecting mistakes and ensuring consitency.37. ·· x 1 --D · x 1 --D x 37.

state diagrams can be used with multiple input/multiple output (MIMO) systems .if a system input (setpoint) is zero.18 37.to track an input/feedback signal disturbance rejection . then the control system is called a regulator. or not used. state feedback .7 TORSIONAL DAMPERS - 37.Linear systems allow superposition . For example.8 MISC .determines the system state using limited data. state coefficient matrices are all constant values .Linear Time Invariant (LTI) systems are linear.in an ideal controller the full state variables can be measured observer-estimator .leaf springs are non-linear and have a cubic relationship to displacement .state based controller designs.input output equations .to control the system about a central design value follow .37. .controller goals stabilize .to minimize the effects of noise .to make the system stable regulate . having constant coefficients. .

10 Nichols Plot . 37.9 Nyquist Plot . although trapezoidal profiles are more common. .19 . 2001].a plot of the real vs.11 BESSEL POLYNOMIALS . the stepper motor should follow the path.put in root locus chapter? . the angle of a system transfer function after a phasor transform. 37.a plot of the gain vs.37. imagniary values of the system transform after phasor transform - 37.The following table of functions provides polynomial roots normalized to a motion that is complete in 1second [How.stepper motor motion plans should not be approximate.input output equations .

37.-7.1682±4.6824±2. ∞ J ITAE = ∫ t e ( t) dt 0 .-7.6786j.-7.6200 -4.7519j.-3.9898±6.3609±9.7845j -4.1104±6.5696±11.3400j -5.8554±6.0271.-4.1557j.-8.3361±8.-5.5714±5.3142j.-8.is used to select poles to minimize the integral value when selecting a characteristics equation. -6.4554±9.0093.-6.4278±1.The Integral of Time Absolute Error (ITAE) integral equation becomes large for errors that persist for a long time.6835±12.5692±5.2613±4.0530±2.-5.-4.0723j.-8.10.-8.-7.1205±1. or decrease the reposnse time the root values are multiplied by the desired settling time.3655j.9268±3.-9.6057j.20 polynomial order 1 2 3 4 5 6 7 8 9 10 roots -4.9342j.-6.0156±5.4540j -8. .9715j.3071j .6585.8081j -4. 37.input output equations .4018j.3616j -9.5300j.3777j.4033.-4.5281±1.7693±1.9668±3.4013±2.0813j -4.9278j.1838j.-9.To increase.1057j. .the roots are used to construct the charateristic equation of the system.6655j -4.12 ITAE .6553j -6.2169±7.9657±3.1145±8.4480.

13 LQR . It is assumed that D=0. The integral is evaluated to reduce the value.Linear Quadratic Regulator (LQR) . ∞ J LQR = ∫ [ ( y ( t )) 0 T T y ( t ) + r ( u ( t ) ) ] dt 2 ( y ( t ) ) y ( t ) = state cost 2 ( u ( t ) ) = control cost r = control penalty ∞ J LQR = ∫ [ ( x ( t )) 0 T C Cx ( t ) + r ( u ( t ) ) ] dt T 2 .the .The basic function for evaluating the controller is defined below. and the plant input will be minimized.input output equations . Ideally the state value will go to zero (it is a regulator).21 37.37.

. where a(s) is the characteristic eqn..input output equations .22 Assume a simple controller form.37. u ( t ) = – K LQR x ( t ) Solve using an Algebraic Riccati Equation Consider a simple system · x = Ax + Bu y = Cx The equivalent transfer function can be written..= ---------------a( s) sI – A This is followed by... UNKNOWN STUFF The desired charactersitic equation can be used to find the controller gain values.. b( s) b( s) –1 C ( sI – A ) B = --------. sI – A + BK LQR = i=1 n ∏ ( s – pi ) –1 ∆ ( s ) = a ( s )a ( – s ) + r b ( s )b ( – s ) The poles are then found using Ackermann’s method. ( u ( t ) ) = control cost r = control penalty ∞ 2 J LQR = 0 ∞ ∫ [ ( x ( t )) ∫ [ ( x ( t )) 0 T C Cx ( t ) + r ( u ( t ) ) ] dt T 2 J LQR = T C Cx ( t ) + r ( u ( t ) ) ] dt T 2 .

ea.ea.ea.20 XXXXX .37. 37. 37. 37.ea. .14 XXXXX . 37.ea.17 XXXXX .ea.input output equations .ea.19 XXXXX .18 XXXXX .15 XXXXX . 37.23 37.16 XXXXX . 37.

input output equations . 37.ea. MIT Opencourseware website.24 37.25 PRACTICE PROBLEM SOLUTIONS 1.21 XXXXX . 37.24 PRACTICE PROBLEMS 1.23 SUMMARY • 37.27 REFERENCES 37. 37.37. J. 37.22 XXXXX .31 Feedback Control Course Notes". .28 BIBLIOGRAPHY How.26 ASSGINMENT PROBLEMS 1. 37.ea. "16.

37.input output equations .25 .