MECHATRONICS

Modeling, Component Interconnection, Sensors, Actuators, and Integrated Design

Clarence W. de Silva, Ph.D. (MIT), Ph.D. (Cambridge), D.Eng. (Hon, Waterloo),

Fellow ASME, Fellow IEEE, Fellow CAE, Fellow RSC, P.Eng.

Tier 1 Canada Research Chair Professor of Mechatronics & Industrial Automation The University of British Columbia Vancouver, Canada e-mail: desilva@mech.ubc.ca www.mech.ubc.ca/~ial

Plan
MECHATRONIC ENGINEERING DYNAMIC MODELS AND ANALOGIES MODEL LINEARIZATION LINEAR GRAPHS EQUIVALENT CIRCUITS LINEAR GRAPH REDUCTION COMPONENT INTERCONNECTION AND SIGNAL CONDITIONING PERFORMANCE SPECIFICATION AND ANALYSIS ANALOG SENSORS AND TRANSDUCERS DIGITAL TRANSDUCERS STEPPER MOTORS CONTINUOUS-DRIVE ACTUATORS INTEGRATED DESIGN Material is from the Books:
De Silva, C.W., MECHATRONICS An Integrated Approach, CRC Press, Taylor & Francis, Boca Raton, FL, 2005. De Silva, C.W., MODELING AND CONTROL of Engineering Systems, CRC Press, Taylor & Francis, Boca Raton, FL, 2009.

www.crcpress.com

Mechatronics
Origin of Term: MECHAnics and elecTRONICS, Yasakawa Electric Co., 1969 Synergistic application of mechanics, electronics, control engineering, and computer science in the development of electromechanical products and systems, through integrated design.
M o de lin g, A n a lysis In te grated D e sign T e stin g an d R e fine m e nt S yste m D ev e lo p m en t T ask s

S en sors a nd T ra n sd u c ers A c tu ato rs C o n trolle rs S tru c tu ra l C om po n en ts E n erg y S ou rc es H yd rau lic an d P ne u m a tic D e v ic es T he rm a l D e v ic es M e ch an ic al E ng in e erin g E lec tro n ics (A n a lo g/D ig ital S o ftw are M e ch atro n ic S ystem

S ign a l P ro ce ssing In p ut/O utp u t H a rd w a re E lec tric al an d C o m p u te r E n gine e rin g

Information Technology
(Control & Computer Engineering)

Mechatronics

M En

ec

gin

ha

ee

nic g

rin

al

ctr le

al ic ee in

ing r

ng E

Example of Mechatronic System

The space shuttle is a multi-domain system consisting of multiple subsystems.

NASA-Houston.

Another Mechatronic System

Computer hard disk drive (HDD) showing disk and read/write head
Courtesy of Quantum Corp.

Motivation for Mechatronic Design

Sequentially designed components of existing mixed systems are not optimally matched There is potential for improvement through concurrent and optimal design Benefits of a mechatronic design:

Optimality and better component matching Increased efficiency Cost effectiveness Ease of system integration Compatibility & ease of cooperation with other systems Improved controllability Increased reliability Increased product life

Mechatronic Design Procedure

Modeling and Design
Identification of the need Conceptual design/Functional specification First principle modular mathematical modeling Topology/structure and component design Detailed mathematical modeling and simulation Component selection Design refinement and optimization

Prototyping
Hardware-in-the-loop simulation Design development Prototypeoptimization Testing and system refinement

Deployment
Production and deployment Life cycle evaluation and optimization

Information for future upgrades

Modeling

Analytical Model Example: Computer Hard Disk Drive (HDD)

Dynamic System
Environment Outputs/ Responses

Dynamic System
(State Variables, System Parameters)

Inputs/ Excitations System Boundary

Note: Rates of Changes of Outputs/State Variables are Not Negligible

TERMINOLOGY
System: Collection of interacting components of interest, demarcated by a system boundary Dynamic System: Rates of changes of response/state variables cannot be neglected Plant or Process: System to be controlled Inputs: Excitations (known, unknown) to the system Outputs: Responses of the system State Variables: Completely identify the dynamic state of system Note: If state variables at one state and inputs up to a future state are known, the future state can be completely determined Control System: Plant + controller, at least (May include sensors, signal conditioning, etc.)

Examples of Dynamic Systems

System
Human body

Typical Input
Neuroelectric pulses

Typical Outputs
Muscle contraction, body movements Decisions, finished products

Company

Information

Power plant

Fuel rate

Electric power, pollution rate

Automobile

Steering wheel movement

Front wheel turn, direction of heading joint Joint motions, effector motion

Robot

Voltage motor

to

A DYNAMIC SYSTEM EXAMPLE: A VISION-BASED WAFER INSPECTION CONTROL SYSTEM

Dynamic Model
It is a representation of a system Useful in analysis, design, modification, and control An engineering (e.g., Mechatronic) physical system consists of a mixture of different types of components An engineering (e.g., Mechatronic) system is typically a multidomain (mixed) system Integrated and unified development of model is desirable (i.e., all domains are modeled together using similar approaches) Use analogous procedures to model all components (in analytical modeling) Analogies exist in mechanical, electrical, fluid, and thermal systems

MODEL TYPES
Model: A representation of a system. Types of Models: 1. Physical Models (Prototypes) 2. Analytical Models 3. Computer (Numerical) Models (Data Tables, Curves, Programs, Files, etc.) 4. Experimental Models (use input/output experimental data for model identification ) Note: Dynamic System: Response variables are functions of time, with non-negligible rates of changes.

MODEL COMPLEXITY
Universal Model (which considers all aspects of the system) is unrealistic E.g.: An automobile model that represents ride quality, energy consumption, traction characteristics, handling, structural strength, capacity, load characteristics, cost, safety, etc. is not very practical Model may address a few specific aspects of interest/application Model should be as simple as possible (Approximate modeling, model reduction, etc. are applicable here)

ADVANTAGES OF ANALYTICAL MODELS OVER PHYSICAL MODELS Modern, high-capacity, high-speed computers can accommodate complex analytical models Models can be modified quickly, conveniently, and high speed at low cost High flexibility of making structural and parametric changes Natural use in computer simulations Can be integrated with computer/numerical/experimental models Can be done well before a prototype is built (and can be instrumental in deciding whether to prototype)

ANALYTICAL MODEL TYPES

Nonlinear: Nonlinear differential equations (principle of superposition does not hold) Linear: Linear differential equations (principle of superposition holds) Distributed (Continuous)-parameter: Partial differential equations (Dependent variables are functions of time and space) Lumped-parameter: Ordinary differential equations (Dependent variables are functions of time, not space) Time-varying (Non-stationary, Non-autonomous): Differential equations with time-varying coefficients (Model parameters vary with time) Time-invariant (Stationary, Autonomous): Differential equations with constant coefficients (Model parameters are constant) Random (Stochatic): Stochastic differential equations (Variables and/or parameters are governed by probability distributions) Deterministic: Non-stchastics differential equations Continuous-time: Differential equations (Time variable is continuously defined) Discrete-time: Difference equations (Time variable is defined as discrete values at a sequence of time points)

Through Variables and Across Variables

Across Variable: Varies Across Element (E.g., Velocity, Voltage, Temperature, Pressure) Through Variable: Remains Unchanged Through Element (E.g., Force, Current, Heat Transfer Rate, Fluid Flow Rate) Mechanical Elements:

Electrical Elements:

Thermal and Fluid Elements:

Mechanical Elements
Variables; Velocity (across variable) and force (through variable) Mass (Inertia) Element (A-Type Element) Constitutive Equation (Newton s 2nd Law): m dt = f m = mass (inertia) Power = fv = rate of change of energy 1 2 dv E = fv dt = m v dt = mv dv Energy E = mv (Kinetic Energy) 2 dt
dv

Note: Energy storage element

Integrate constitutive equation
+

1 v (t ) = v(0 ) + m
-

f
0
-

dt

Set t = 0 v (0 ) = v (0 ) unless force is infinite. Note: 0- denotes instant just before t = 0 and 0+ denotes instant just after t = 0. Observations: 1. Velocity (across variable) represents the state of an inertia element A=Type Element Notes: 1. Velocity at any t is completely determined from initial velocity and the applied force; 2. Energy of inertia element is represented by v alone. 2. Velocity across an inertia element cannot change instantaneously unless infinite force/torque applied. 3. A finite force cannot cause an infinite acceleration in inertia element. A finite instantaneous change (step) in velocity needs an infinite force v is a natural output (or state) variable and f is a natural input variable for inertia element.

Mechanical Elements
Spring (Stiffness) Element (T-Type Element) Constitutive Equation (Hook s Law): dt = kv k = stiffness Note: Differentiated version of familiar force-deflection Hooke s law in order to use velocity (as for inertia element) Energy
E = fv dt =
t

df

1 df f dt = k dt

1 k f df

Energy

1 f2 E= 2 k

(Elastic potential energy)

f (t ) = f (0 ) + k v dt
0-

f (0 + ) = f (0 - )

unless an infinite velocity is applied

Note: Energy storage element Observations: 1. Force (across variable) represents state of spring element T-Type Element Justified because: 1. Spring force of a spring at time t is completely determined from initial force and applied velocity; 2. Spring energy is represented by f alone. 2. Force through stiffness element cannot change instantaneously unless an infinite velocity is applied to it. 3. Force f is a natural output (state) variable and v is a natural input variable for spring.

Mechanical Elements
Damping (Dissipation) Element (D-Type Element)

Constitutive Equation: f = bv b = damping constant (damping coefficient); for viscous damping Observations: 1. This is an energy dissipating element (D-Type Element) 2. Either f or v may represent the state 3. No new state variable is defined by this element.

Electrical Elements Variables: Voltage (across variable) and the current (through variable) Capacitor Element (A-Type Element) dv Constitutive Equation: C dt = i C = capacitance Power = iv Energy Energy
E= 1 C v2 2

E = iv dt = C

dv v dt = Cv dv dt

(electrostatic energy)
t 0-

Note: Energy storage element

1 Integrate constitutive equation: v(t ) = v(0 ) + C i dt

v(0 + ) = v (0 - )

unless an infinite current is applied

Observations: 1. Voltage (across variable) is state variable for a capacitor A-Type Element 2. Voltage across a capacitor cannot change instantaneously unless an infinite current is applied. 3. Voltage is a natural output variable and current is a natural input variable for a capacitor.

Electrical Elements
Inductor Element (T-Type Element) Constitutive Equation: L = inductance
E= 1 2 Li 2
L di =v dt

Energy (Electromagnetic energy) Note: Energy storage element Integrate constitutive equation:
i (t ) = i (0 - ) + 1 v dt L 0 t

+ i (0 ) = i (0 ) unless an infinite voltage is applied.

Observations: Current (through variable) is state variable for an inductor T-Type Element Current through an inductor cannot change instantaneously unless an infinite voltage is applied. Current is a natural output variable and voltage is a natural input variable for an inductor.

Electrical Elements
Resistor Element (D-Type Element) Constitutive Equation: R = resistance
v = Ri

(Ohm s law)

Observations: 1. This is an energy dissipating element (D-Type Element) 2. Either i or v may represent the state 3. No new state variable is defined by this element.

Thermal Elements
Variables: Across variable temperature (T) and through variable heat transfer rate (Q).

Thermal Capacitor (A-Type Element) Consider control volume V of fluid with, density r, and specific heat c. Constitutive Equation: Net heat transfer rate into the control volume
dT =Q dt Ct = rvc = thermal Ct

Q = rVc

dT dt

capacitance of control volume

Observations: Temperature T is state variable for thermal capacitor (from usual argument) A-Type Element Heat transfer rate Q is natural input and temperature T is natural output for this element This is a storage element (stores thermal energy) Note: There is no thermal inductor like storage element with state variable Q .

Thermal Elements
Thermal Resistance (D-Type Element)
Three basic processes of heat transfer three different types of thermal resistance Constitutive Relations Conduction:
Q= kA T Dx

k = conductivity; A = area of cross section of the heat conduction element; Dx = length of heat conduction that has a temperature drop of T. Conductive resistance
Rk = Dx kA

Convection: Q = hc AT hc = convection heat transfer coefficient; A = area of heat convection surface with temperature drop T Conductive resistance
Rc = 1 hc A

Radiation: Q = sFE FA A(T14 - T24 ) a nonlinear thermal resistor s = Stefan-Boltzman constant FE = effective emmisivity of the radiation source (of temperature T1) FA = shape factor of the radiation receiver (of temperature T2) A = effective surface area of the receiver.

Fluid Elements
Variables: Pressure (across variable) P and volume flow rate (through variable) Q Fluid Capacitor (A-Type Element) dP Constitutive Equation: C f dt = Q Note 1: Stores potential energy (a fluid spring ) Note 2: Pressure (across variable) is state variable for fluid capacitor A-Type Element Three Types: Fluid compression; Flexible container; Gravity head V 1a. For liquid control volume V of bulk modulus b : Cbulk = b 1b. For isothermal (constant temperature, slow-process) gas of volume V and pressure:
Ccomp = V P

1c. For adiabatic (zero heat transfer, fast-process) gas:

k= cp cv

Ccomp =

V kP

= ratio of specific heats at constant pressure and constant volume

Celastic A2 = k

2. For incompressible fluid in a flexible vessel of area A and stiffness k:

Note: For a fluid with bulk modulus, the equivalent capacitance = Cbulk + Celastic . A 3. For incompressible fluid column of area of cross-section A and density r: C grav = rg

Fluid Elements
Fluid Inertor (T-Type Element)
dQ =P dt

Constitutive Equation:

If

Note 1: Volume flow rate Q (through variable) is state variable for fluid inertor T-type Element Note 2: It stores kinetic energy, unlike the mechanical T-type element (spring), which stores potential energy. With uniform velocity distribution across A over length segment Fluid inertance
If = r Dx A
Dx :

For a non-uniform velocity distribution: Fluid inertance I f = ar A (correction factor a) For a pipe of circular cross-section with a parabolic velocity distribution, a = 2.0
Dx

Fluid Elements
Fluid Resistor (D-Type Element)

Constitutive Equation (Linear):

P = Rf Q

n Constitutive Equation (Nonlinear): P = K R Q (KR and n are parameters of nonlinearity)

For Viscous Flow Through a Uniform Pipe: (a) With circular cross-section of diameter d:
R f = 128 m

Dx pd 4
R f = 12 m

(b) With rectangular cross-section of height b << width w:

Dx wb 3

Note: m = absolute viscosity (or, dynamic viscosity); u = kinematic viscosity with m = ur

Analogies and Constitutive Relations

System Type Constitutive Relation for Energy Storage Elements A-Type (Across) Element Mass
m dv = f dt

TranslatoryMechanical v = velocity f = force Electrical v = voltage i = current Thermal T = temperature difference Q = heat transfer rate Fluid P = pressure difference Q = volume flow rate

T-Type (Through) Element Spring

df = kv dt

Energy Dissipating Elements D-Type (Dissipative) Element Viscous Damper

f = bv

(Newton s 2 Law) m = mass Capacitor

C dv =i dt

nd

(Hooke s Law) k = stiffness Inductor

L di =v dt

b = damping constant Resistor R = resistance Thermal Resistor

Rt Q = T
Ri = v

C = capacitance Thermal Capacitor

Ct dT =Q dt

L = inductance None

Ct = thermal capacitance Fluid Capacitor

Cf dP =Q dt

Rt = thermal resistance

Fluid Inertor
If dQ =P dt

Fluid Resistor
Rf Q = P

Cf = fluid capacitance

If = inertance

Rf = fluid resistance

Through and Across Variables

System Type
Hydraulic/Pneumatic Electrical Mechanical Thermal

Through Variable Across Variable

Flow Rate Current Force Heat Transfer Pressure Voltage Velocity Temperature

Force-Current Analogy
System Type SystemVariables: ThroughVariables AcrossVariables System Parameters Mechanical Force f Velocity v m k b Electrical Current i Voltage v C 1/L 1/R

Natural Oscillations
Require two-types of energy One type of stored energy is converted into the other type, repeatedly back and forth oscillatory natural response. Some energy will dissipate (through the dissipative mechanism of Dtype element) Natural oscillations will decay Mechanical Systems: Kinetic Energy and Potential Energy (Elastic or Gravitational) Electrical Systems: Electrostatic Energy and Electromagnetic Energy Fluid Systems: Kinetic Energy of Flow and Potential Energy of Fluid Storage or Head Thermal Systems: Only one type of energy storage element (A-type) with only one type of energy (thermal energy) purely thermal systems cannot naturally oscillate (unless with external force, or integrated with other systems (e.g., fluid systems).

STEPS OF MODEL (ANALYTICAL) DEVELOPMENT

1. Identify system of interest (purpose, boundary) 2. Identify/specify variables of interest (excitations/inputs, responses/outputs, etc.) 3. Approximate various segments (processes, phenomena) by ideal elements, suitably interconnected. 4. Draw a free-body or circuit diagram, with suitably isolated components. (a)Write constitutive equations (physical laws) for elements. (b)Write continuity (or conservation) equations for through variables (equilibrium of forces at joints; current balance at nodes, etc.) (c)Write compatibility equations for across (potential or path) variables (loop equations for velocities geometric connectivity; voltages potential balance) (d)Eliminate auxiliary (unwanted) variables 5. Express boundary conditions and initial conditions using system variables.

State Space Model

Note: Inputs (u) and outputs (y) are given Step 1: State Variable (x) Selection Across variables for A-type energy storage elements Through variables for T-type energy storage elements Step 2: Write constitutive equations for all the elements (both energy storage and dissipative) Step 3: Write compatibility equations and continuity equations Step 4: Eliminate redundant variables Step 5: Express the outputs in terms of state variables Linear Form:
& x = Ax + Bu ;
y = Cx

State Space Model Example

Diesel Engine Driving a Pump Linear Model

B1 W w1

B1
J w B2

J B2

w1

Free Body Diagrams

T K J T B1(W - w1)

w B2 w
& Jw

B2w

State Space Model

For K:
dT Constitutive relation: dt = K (w1 - w ) Torque balance (Continuity): B1 (W - w1 ) - T = 0

(i) (ii) (iii)

Substitute (ii) into (i): dt = - B T - Kw + KW 1 This is one state equation. For J: & Constitutive relation: Jw + B2w - T = 0 dw B 1 =- 2w+ T => dt J J This is the second state equation. Vector-matrix form of the state-space model:
dT - K dt B 1 = dw 1 dt J -K T K + W B - 2 w 0 J

dT

(iv)

Analogous Translatory System

Model Linearization

Model linearization
Real systems are nonlinear They are represented by nonlinear analytical models Common techniques (e.g., response analysis, frequency domain analysis, eigenvalue problem analysis, simulation, control) use linear models Method: Linearize each nonlinear term (1st order Taylor series approximation) Note: A nonlinear term may be a function of more than one independent variable.

Example

Induction Motor-Pump Combination in Spray Painting System of Automobile Assembly Plant Light gear with efficiency h and speed ratio 1:r; flexible shaft torsional stiffness kp Moments inertia of motor rotor and pump impeller: Jm and Jp Dissipation in motor and bearings: viscous damping of damping constant bm Pump load (paint load plus dissipation): viscous damping constant bp T qw (w - w ) Tm = 0 0 2 0 2 m Magnetic torque Tm of induction motor (qw0 - w m )

w m = motor speed; T0 depends quadratically on phase voltage supplied to motor; w 0 line frequency of the AC supply; q > 1.0

Related Questions
(a) Comment about the accuracy of model (b) State variables: motor speed w m , pump-shaft torque Tp, pump speed w p . Derive three state equations: nonlinear state model. What are the inputs to the system? (c) What do parameters w0 and T0 represent with regard to motor behavior? Determine Tm Tm , and Tm and verify that first is negative and other two are positive. Note:
w m T0 w0

under normal operating conditions 0 < wm < w0 . (d) Consider steady-state operating point: steady motor speed wm . Obtain expressions for w p , Tp, and T0 at operating point, in terms of wm and w0 (e) Let Tm = -b , Tm = b1 , and Tm = b 2 at operating point. Voltage control: vary T0;
wm T0 w0

Frequency control: vary w0 . Linearize state model about operating point and express it in terms of incremental variables wm , Tp , w p , T0 , w0 . Let (incremental) output variables be incremental pump speed w p and incremental angle of twist of the pump shaft. Obtain linear state space model A, B, C and D. (f)For frequency control only ( T0 = 0) obtain input-output differential equation relating
w p and w0 .

Show that if w0 is suddenly changed by step Dw0 then

b 2 rk p Jm J p Dw0 ,

d 3w p dt 3

will

instantaneously change by step

but not the lower derivatives of w p

(a) Backlash and inertia of gear transmission neglected (not accurate in general). Gear efficiency h is assumed constant but usually varies with speed There is some flexibility in the shaft (coupling) between gear and drive motor Energy dissipation (in pump load, bearings) is lumped into a single linear viscous-damping element (In practice nonlinear and distributed) dq p dq m wp = (b) Motor speed w m = ; Load (pump) speed ; qm = motor rotation; qp = pump rotation dt dt Gear efficiency h =
Tp r wm Tg w m = Output Power with Tg = reaction torque on motor shaft from gear Input Power

Solution

r = gear ratio rw m = output speed of gear Power = torque speed Tg = r Tp

h

(i) (ii) (iv)

& Newton s 2nd law (Torque = Inertia Angular Acceleration) for motor: Tm - Tg - bmw m = J mw m & Newton s 2nd law for pump: T p - b pw p = J pw p (iii)
Hooke s law (Torque = torsional stiffness angle of twist) for flexible shaft: Tp = k p (rq m - q p ) State Equations: (ii), (iii), and derivative of (iv); Specifically Substitute (i) into (ii): Differentiate (iv): Equation (iii):
r & J mw m = Tm - bmw m - T p h

(v) (vi) (vii)

T0 qw0 (w0 - w m ) (viii) 2 2 (qw0 - w m )

& Tp = k p (rwm - w p )

& J pw p = Tp - bpw p

Note: Nonlinear model; State vector = [wm Tp w p ]T ; Input Tm = Strictly, two inputs: w 0 (speed of rotating magnetic field

line frequency); T0

(phase voltage)2

Induction Motor Characteristic Curve

Solution (Cont d)
(c) From (viii): When w m = 0 Tm = T0 T0 = starting torque of motor From (viii): When Tm = 0 w m = w 0 w0 = no-load speed = synchronous speed Note: Under no-load conditions, no slip Actual speed of induction motor = speed w 0 of rotating magnetic field Differentiate (viii) with respect to T0 , w 0 , and w m :
Tm qw 0 (w 0 - w m ) = = b1 2 2 T0 (qw 0 - w m )

(say) (ix) (x)

2 2 2 Tm T0 q[( qw0 - wm )(2w0 - wm ) - w0 (w0 - wm )2qw0 ] T0 qwm [(w0 - wm ) 2 + (q - 1)w0 ] = = = b 2 (say) 2 2 2 2 w0 (qw0 - wm ) 2 ( qw0 - wm )2 2 2 2 Tm T0 qw0 [( qw0 - wm )( -1) - (w0 - wm )( -2wm )] T0 qw0 [( q - 1)w0 + (w0 - wm )2 ] = == -b (say) 2 2 2 2 wm (qw0 - wm )2 ( qw0 - wm )2

(xi)

Note: b1 > 0 ; b2 > 0 ; b > 0 (d) Steady-state operating point rates of changes of state variables = 0
& & & Set wm = 0 = Tp = w p in (v)
r Tp ; 0 = k p ( rwm - w p ) ; h

(vii) (xiii)

0 = Tm - bmw m -

0 = Tp - bpw p

w p = rwm (xii);
Tm = bmwm + r b pwm
2

Tp = bp rwm
T qw (w - w ) h = 0 02 0 2 m ( qw0 - wm )

(from (2.62) or (viii))

T0 =

2 2 w m (bm + r 2 b p h )(qw0 - wm )

qw0 (w0 - w m )

(xiv)

Solution (Cont d)
(e) Take increments of state equations (v) (vii):
r & J mwm = -bmwm - T p - bwm + b1T0 + b 2w0 h

& & (xv); Tp = k p (rwm - w p ) (xvi); J pw p = Tp - bpw p (xvii)

Where

Tm =

T T Tm wm + m T0 + m w0 = -bwm + b1T0 + b 2w0 wm T0 w0

(xviii)

Linearized State Equations: (xv)-(xvii) subject to (xiii) State vector A=

x = [wm Tp w p
-(b + b ) J m m rk p 0

; Input vector u = T0 w0 ; Output vector y = w p Tp k p b1 J m b 2 J m - r (h J m ) 0 0 1 0 0 0 ; C = 0 -k p ; B = 0 1 k p 0 ; D = 0 0 0 1 Jp -b p J p

(f) For frequency control, T0 = 0 Substitute (xvi) into (xv) to eliminate w m ; Substitute (xvii) into the result to eliminate Tp Input-Output Differential Equation:
Jm J p kp ( d 3w p dt 3 + [ J mbp + J p (bm + b)] d 2w p dt 2 + [k p ( J m + r2J p h ) + bp (bm + b)] dw p dt +

r 2bp h

(xix)
+ bm + b)w p = b 2 rk pw0

3rd order differential equation; system is 3rd order; state-space model is also 3rd order.

Solution (Cont d)
Observation From (xix): When w0 is changed by finite step Dw0 RHS of (xix) will be finite LHS, and particularly highest derivative ( Further Verification: If as a result,
d 3w p
d 2w p dt 2

d 3w p dt 3

) also must change by a finite value.

or lower derivatives also change by a finite step

dt 3 should change by an infinite value (Note: derivative of a step = impulse) Contradicts the fact that RHS of (xix) is finite
d 2w p dt 2
d 3w p dt
3

, dt , and w p will not change instantaneously will change instantaneously by a finite value due to finite step change of w 0
d 3w p dt
3

dw p

Only

From (xix): Resulting change of

is J J Dw0 m p

b 2 rk p

DC Motor Operating Curves

Torque-Speed Curves: (a) Shunt-Wound; (b)Series-Wound; (c) Compound-Wound; (d) General Case Note: In each curve, excitation voltage vc is maintained constant

DC Motor Control Circuit and Mechanical Loading

(a) Equivalent circuit of a dc motor (separately excited) (b) Armature mechanical loading diagram

Linear Model for Motor Control

Figure (d): One curve is at control voltage vc and the at vc+Dvc. Tangent drawn at a selected point (operating point O). Magnitude b of slope (negative)
T b=- m Damping Constant (including electrical and mechanical damping effects): w m
vc = constant

Note: Included mechanical damping effects depends on mechanical damping present during test (primarily bearing friction) Draw a vertical line through operating point O Torque intercept DTm between two curves Note: Vertical line is a constant speed line Voltage Gain:
T d Tm = m wm
kv = Tm vc =
wm = constant

DTm Dvc

dwm +
vc

Tm vc

wm

d vc = -bdwm + kvd vc Linearized model of motor.

Note: Torque needed to drive rotor inertia is not included in this equation (because steady-state curves are used in determining parameters) Inertia term is explicitly present in Mechanical Equation of motor rotor (See Figure (b)): d dw m Jm = d Tm - d TL dt Note: Overall damping constant of motor (electrical and mechanical damping) is included

Linear Graphs

Linear Graphs
Graphical representation of a model Allow visualization of system structure (before formulation of an analytical model) Help identify similarities (structure, performance) in different types of systems Applicable for Lumped-parameter systems They use interconnected line segments (branches) Linear graph use of line segments (Nothing to with system linearity) A convenient tool for developing state-space model

Single-Port Elements
Port: Place where energy/power is exchanged environment (input or output of energy/power) with

A Single-Port (Single Power or Energy Port) Element: Represented by a single branch (line segment) Has a through variable f and a corresponding across variable v, given as the ordered pair f,v on the branch The relationship between f and v can be nonlinear Possesses only one power variable, which is the product of f and v; with just one power port, hence the name Has two terminals
f, v

Sign Convention (Single-Port Element)

One end of branch: Point of action; Other end of branch: Point of reference (Somewhat arbitrary, can reflect the physics of system) Oriented branch Direction is assigned using an arrowhead Positive direction of power flow Into element at point of action, and out of element at point of reference (except for a source ) Arrowhead of branch is pointed toward point of reference Note: Does not represent the positive direction of f or v (See Figure (a) and Figure (b) Across variable: Given relative to point of reference Pre-establish positive direction of any one of f and v Positive direction of the other variable Convention: Assign direction of v and power flow the same at point of action (Figure (a) is preferred over Figure (c)).

Sign Convention (Single-Port Element)

(a)
Energy Stored v
2

or Dissipated

f 2 Point of Action (Work in)

Element 1

v = v2 - v1

Point of Reference (Work out)

(b)
f, v

(c)

v2

v1

f 2 1 2

Element 1

v = v2 - v1

Note: Reversing the directions of both f and v won t affect power flow direction (See (a) and (c); Convention (a) is preferred (A) Work done on element at point of action (by an external device) is positive (B) Work done by element at point of reference (on an external load) is positive. Note: (A) (B) Stored in element (e.g., kinetic or potential) Has capacity to do work or dissipated (damping) through mechanisms manifested as heat transfer, noise, etc.

Linear Graphs of Mechanical Elements

(Single-Port Elements) Need Three Single-port Elements: Mass, Spring, Dashpot/Damper Corresponding rotary elements are easy to visualize f = applied torque; v = relative angular velocity in same direction. Note: Linear graph of an inertia element has a broken line segment. Force does not physically travel from one end of linear graph to other but felt Note: Analogous electrical, fluid, and thermal elements may be represented similarly

Linear Graphs of Mechanical Elements

(Single-Port Elements)

Linear Graphs of Electrical Elements

(Single-Port Elements)

Source Elements
T-Type Source (e.g., Force Source, Current Source) A-Type Source (e.g., Velocity Source, Voltage Source)

(a) T-Type Source (through-variable input); (b) Linear graph representation (Arrowhead: positive direction of through variable) (c) A-Type Source; (d) Linear graph representation (+ sign: point of action; - sign: point of reference; Arrow head: direction of across variable drop)

Effects of Source Elements

Source (independent) variable is unaffected by dynamics of system to which the source is connected. But co-variable (dependent) will change Source elements can serve to inhibiting interactions between systems. Dynamic behavior of a system not affected by connecting a new system: (a) in series with an existing T-source (b) in parallel with an existing A-source. The modified system may be treated (analyzed) as two separate uncoupled systems driven by the same input source Order of the overall system = sum of the order of individual subsystems.

Effects of Source Element

(a) Two systems connected in series to a T-source (b) Two systems connected in parallel to an A-source.

Mechanical Transformer (Two-Port Element)

(a) vi fo

fi (b)

vo

fi vi

vo

fo

(c) fi , v i r fo , v o

vo = rvi fo = 1 fi r

Reference g

Electrical and Fluid Transformers

Electrical Transformer

Fluid Transformer

Mechanical/Electrical Transformer (Two-Port Element)

Arrow on each branch (line segment) Direction when power = force velocity > 0 Note: vi and fi are velocity/voltage and force/current at input port vo and fo are velocity/voltage and force/current at output port (Linear) transformation ratio r is given by

vo = r vi

Conservation of power: fi vi + f o vo = 0
fo = 1 fi r

Note: r is a dimensional parameter Note: Electrical transducers and transmission devices are also treated as transformers (idealized by neglecting dissipation)

Gyrator (Two-Port Element)

Mechanical Gyrator: Ideal gyroscope Spinning top rotating about own axis at high angular speed w (positive in x direction) Assumed to remain unaffected by other motions
(a) y fo , v o J z x

w
L fi , v i

(c) fi , vi fo , v o
vo = Mfi fo = 1 vi M

R e f e r e nc e g

Gyrator (Two-Port Element)

Moment of inertia about x axis of rotation = J Corresponding angular momentum = Jw Velocity vi is given to free end in the y direction (result in force fi at that point, whose positive direction is y) Corresponding rate of change of angular momentum = J w vi L about positive y-axis Note: vi L is the angular velocity due to vi Newton s second law: Required torque = J w vi L in the same direction. fo = corresponding force at free end in z-direction Corresponding torque = foL acting in negative y-direction - f o L = J w vi L L = length of gyroscope With
L2 M = Jw
fo = 1 vi M

Conservation of power f i vi + f o vo vo = Mfi

=0

Note: M is a mobility parameter (velocity/force).

Compatibility (Loop) Equations

Loop: Closed path formed by two or more branches Loop Equation (Compatibility Equation): Sum of across variables in a loop is zero. Sign Convention: 1.Go in counter-clockwise direction of loop 2.In direction of branch arrow (except in a T-source) Across variable drops Positive Opposite to arrow Across variable increases negative Note: Loop equation (compatibility) across variables are the same (i.e., unique) at any given point in the loop at a given time (E.g., a mass and spring connected to the same point must have the same velocity point is intact; does not break or snap; system is compatible). Number of Primary Loops Minimal set of loops from which any other loop can be determined. Primary loop set is an "independent" set. It will generate all the independent loop equations. Note: Loops closed by broken-line (inertia) branches should be included in counting primary loops.

Mechanical Example

Three primary loops; Force source E.g., Loop 1: m-k; Loop 2: k-b; Loop 3: b-f Other Choices: (b-k, m-b and m-f) or (b-k, m-b and f-k), etc. Note: m-k loop = (m-b loop) - (b-k loop); or Loop 1 = (m-b loop) Loop 2 f-m loop = (f-b loop) + (b-m loop); or f-m loop = Loop 3 + (b-m loop) Verify these relations using loop equations Loop 1 Equation: v1 v2 = 0; Loop 2 Equation: v2 v3 = 0 Loop 3 Equation: v3 v = 0; m-b Loop Equation: v1 v3 = 0 f-m Loop Equation: v1 v = 0

Continuity (Node) Equations

Node: Point where two or more branches meet Node Equation: Sum of all through variables at node = 0 "What goes in must come out" continuity of through variables at a node Mechanical Systems: Force balance or equilibrium equation; Newton s third law; etc. Electrical Systems: Current balance; Kirchoff s current law; conservation of charge; etc. Hydraulic Systems: Conservation of matter Thermal Systems: Conservation of energy Sign Convention: Into node is positive Previous Example: Two nodes. Corresponding node equations are identical: Node 2 Equation: -f1-f2-f3+f = 0 Node 1 Equation: f1+f2+f3-f = 0

Required number of node equations = Number of nodes - 1

Series and Parallel Connections

Series Connected System Through variables are the same Across variables are additive Parallel Connected System Across variables are the same Through variables are additive

Spring (k), Damper (b): (a) Connected in parallel; (b) Connected in series (a) has two primary loops (two elements in parallel) (b) has only one loop (all elements in series with force)

Amplifiers
Accomplish signal amplification Are active devices (need external power to operate) Electrical Signals: Voltage, Current, Power Voltage amplifiers, Current amplifiers, Power amplifiers Note: Analogous across variables, through variables, and power variables exist for other types of systems Mechanical Signals: Velocity, Force, Power Signal levels at component interface locations may have to be adjusted for satisfactory performance E.g.: Input to an actuator should possess adequate power to drive the load Use a power amplifier

Linear Graph Representation of Amplifier

Amplifier uses an external power source Amplifier has a decoupling effect on systems (interaction reduction) Its is a Dependent Source or Modulated Source Amplifier output depends on (modulated by) amplifier input

Across Variable Amplifier

Through Variable Amplifier

Multi-Domain Example Using Linear Graphs

Robotic Door Opener

Linear Graph
State Space Model
Node Equations:
i - iR = 0 iR - ia = 0 ia - ib = 0 -Tm - Td - Tb - Tk = 0

Independent Loop Equations:

va (t ) - vR - vL - vb = 0 wm - wd = 0 wd - wb = 0 wb - wk = 0

Constitutive Equations:

dia = vL dt dw d Jd = Td State-space shell dt dTk = kd w k dt La

vR = RaiR Auxiliary constitutive equations Tb = bd wb

wm =

1 vb km Electro-mechanical transformer Tm = - kmib

State Space Model

State vector x = [ia w d Input vector u = [ va (t ) ] Output vector y = [wd
Tk ]
Tk ]
T

Eliminate auxiliary variables from sate-space shell, by substitution:

vL = va (t ) - vR - vb = va (t ) - Ra ia - kmwm = - Ra ia - kmwd + va (t )
Td = -Tk - Tm - Tb = -Tk + kmib - bd wb = kmia - bd wd - Tk

w k = wb = w d

State Space Model

State Model: & x = Ax + Bu

y = Cx + Du
where
- Ra La k J m d A= 0 -km La -bd J d kd 0 -1 J d ; 0
1 La 0 ; C= B= 0

0 1 0 0 0 1 ;

D=0

Note 1: This is a multi-domain (Electro-mechanical model) Note 2: Multi-functional devices (e.g., piezoelectric device that serves as both actuator and sensor) may be modeled similarly, using an electro-mechanical transformer (or, reciprocity principle ).

Useful Mechanical Transfer Functions

Transfer Function Dynamic Stiffness Receptance, Dynamic Flexibility, Compliance Impedance (Z) Mobility (M) Dynamic Inertia Accelerance Force Transmissibility (Tf) Motion Transmissibility (Tm)

Definition (Laplace or Frequency Domains) Force/Displacement Displacement/Force Force/Velocity Velocity/Force Force/Acceleration Acceleration/Force Transmitted Force/Applied Force Transmitted Velocity/Applied Velocity

Impedance and Mobility Functions of Mechanical Elements

Element Time-Domain Model Impedance Mobility (Generalized Impedance)

Mass m Spring k Damper b

dv m = f dt df = kv dt f = bv

Z m = ms
k Zk = s Zb = b

Mm =
Mk =
Mb =

1 ms
s k
1 b

Note: Frequency domain is a special case of Laplace domain

Commonly, frequency domain is used when dealing with impedance approaches

Impedance and Admittance Functions of Electrical Elements

Element Capacitor C Inductor L Resistor R Time-Domain Model Impedance (Z) Admittance (W)

dv C =i dt
L di =v dt

1 ZC = Cs

WC = Cs
1 WL = Ls 1 WR = R

Z L = Ls

Ri = v

ZR = R

Note: Frequency domain is a special case of Laplace domain Commonly frequency domain is used in the impedance methods

Interconnection Laws for Mechanical Impedance and Mobility

Series Connections
f Z1 M1 v1 v Z2 M2 v2 f f2

Parallel Connections
f1 Z1 M1 Z2 M2 v

v = v1 + v 2
v v1 v 2 = + f f f

f = f1 + f 2
f f f = 1 = 2 v v v

M = M1 + M 2
1 1 1 = + Z Z1 Z 2

Z = Z1 + Z 2
1 1 1 = + M M1 M 2

Interconnection Laws for Electrical Impedance and Admittance

Series Connections Parallel Connections

v = v1 + v 2
v v1 v2 = + i i i
Z = Z1 + Z 2
1 1 1 = + W W1 W2

i = i1 + i2
i i1 i2 = = v v v

W = W1 + W2
1 1 1 = + Z Z1 Z 2

Note: Electrical Impedance and Mechanical Mobility are A-Type Transfer Functions [Across Variable/Through Variable] Same Interconnection Law Electrical Admittance and Mechanical Impedance are T-Type Transfer Functions [Through Variable/Across Variable] Same Interconnection Law

Equivalent Circuits Linear Graph Reduction

Thevenin s Theorem for Electrical Circuits

(a) Circuit segment with impedances and sources Ze = equivalent impedance with source killed (i.e., voltage sources shorted and current source opened) = Thevenin resistance

(b) Thevenin Equivalent Circuit

Voc(s) = open-circuit voltage at cut terminals Isc(s) = short-circuit current at cut terminals (c) Norton Equivalent Circuit

Note: Variables are expressed in Laplace or frequency domains Note 2: For multiple sources, use superposition (linear system)

General Procedures for Mechanical Circuit Analysis Using Linear Graphs

1.For each branch mark the Mobility function (not mechanical impedance) 2.Carry out graph analysis and reduction as if you are dealing with an electrical circuit, by using the following analogies Mechanical Circuit Mobility Function Force Velocity Electrical Circuit Analogy Electrical Impedance Current Voltage

For Example: 1.For parallel branches: mobilities are combined by inverse relation (M = M
M 1M 2 1 + M2

);Velocity is common; Force is divided inversely to branch mobilities

2.For series branches: Mobilities add ( M = M 1 + M 2 ); Force is common; velocity is divided in proportion to mobility 3.Killing a force source means open-circuiting it (so, transmitted force = 0) 4.Killing a velocity source means short-circuiting it (so, velocity across = 0)

Example 1: Ground-based Mechanical Oscillator

Mechanical System

Linear Graph

Equivalent Circuit
Mm =

Thevenin Equivalent Circuit

Mm 1 ( s / k )(1/ b) s F (s) F ( s ) ; Voc ( s ) = M s F ( s ) = ; Me = Ms = = (ii) ; Fs = (i) (M m + M s ) ms ( s / k ) + (1/ b) (bs + k ) (b + k / s )

Note: Norton equivalent circuit is same as (b) or (c) No additional useful information.

Example 1 (Continued)
Check Velocity Across and Force Through Mm for All Circuits For Circuits (b) and (c):
Velocity Across Inertia Element = Mobility of Circuit Source Force

Velocity across mass element = M

M sM m F (s) s + Mm

(iii)

(Note: Suspension and mass are connected in parallel) Force through mass element =
Ms F ( s) Ms + Mm

(iv)

Note: Simply divide (iii) by mass mobility; or since force is divided inversely to mobilities in parallel paths For Circuit (d): Velocities are divided in proportion to mobilities in series paths M Velocity across mass = M + M V (s)
m oc e m

Substitute (i) and use (ii): Velocity across mass element = M s + M m F (s) This is identical to (iii) Force through mass = This is identical to (iv)
Voc ( s) Me + Mm

MsMm

Substitute (i) and (ii): Force through mass element =

Ms F (s) Ms + Mm

Example 2: Oscillator with Support Motion

Mechanical System

Linear Graph

Equivalent Circuit
Mm =

Norton Equivalent Circuit

V (s) 1 s Mm = (b + k / s )V ( s ) (i) ; Me = Ms = (ii) ; Vm ( s) = V ( s) ; Fsc ( s) = Ms ms (bs + k ) Mm + Ms

Note: Norton equivalent circuit is same as (b) or (c) No additional useful information.

Example 2 (Continued)
Check Velocity Across and Force Through Mm for All Circuits For Circuits (b) and (c): Velocity across mass element
Vm = Mm V (s) Ms + Mm

(iii)

(Note: Suspension and mass are connected in series velocities are divided in proportion to mobilities) Force through mass element =
V (s) Ms + Mm

(iv)

For Circuit (d): Equivalent mobility of the two parallel elements = Velocity across mass = This is identical to (iii) Force through mass element =
Me Fsc (s ) Me + Mm

MeM m Me + Mm

Mm V ( s) s + Mm

MeM m Fsc ( s) Me + Mm

Substitute (i) and use (ii): Velocity across mass element = M

(Note: Force is divided inversely to mobilities in a parallel connection) V ( s) Substitute (i) and (ii): Force through mass element = M + M
s m

This is identical to (iv)

Example 3: Ground-based 2DOF Mechanical System

Mechanical System

Linear Graph

Linear graph showing suspension components

Cutting Out the Suspension Unit of Interest

Thevenin Equivalent Circuit

Example 3: (Continued)

Cutting Out the Suspension Unit of Interest

Thevenin Equivalent Circuit

F1 =
Fs =

M m 2 ( M s1 + M m1 ) M m1 M m1M m 2 F ( s) ; M e = F ( s) ; Voc ( s) = M m 2 F1 = M m 2 + ( M s1 + M m1 ) ( M s1 + M m 2 + M m1 ) ( M s1 + M m 2 + M m1 )
Voc ( s) M m1M m 2 1 = F ( s) M ( M + M m1 ) ( M e + M s ) ( M s1 + M m 2 + M m1 ) [ m 2 s1 + Ms] M m 2 + ( M s1 + M m1 )

Fs =

M m1M m 2 F ( s ) M m 2 ( M s1 + M m1 ) + M s ( M s1 + M m 2 + M m1 )

Tf =

M m1M m 2 M m 2 ( M s1 + M m1 ) + M s (M s1 + M m 2 + M m1 )

Show that this is identical the previous result.

Example 4: 2DOF Mechanical System with Support Motion

Mechanical System

Linear Graph

Linear Graph Showing Suspension Components

Norton Equivalent Circuit

Example 4: (Continued)

Linear Graph Showing Suspension Components Force from source velocity =

M e = M s1 +

Norton Equivalent Circuit

V (s ) M m2 M m2 V ( s) Fsc ( s) = = V ( s) M s 1M m 2 M s1M m 2 ( M s1 + M m 2 ) M s1M m 2 + M s ( M s1 + M m 2 ) [ + Ms] [ + Ms] ( M s1 + M m 2 ) ( M s1 + M m 2 )

M m2M s M ( M + M s ) + M m 2 M s M s1M m 2 + M s ( M s1 + M m 2 ) V = M m1M e F ( s) = s1 m 2 = sc ; m (M m1 + M e ) M m2 + M s M m2 + M s M m2 + M s

Note:
Vm =

Fsc ( s ) M e =
M m1

M m2 V ( s) (M m2 + M s )

M m2 M m1M m 2 V ( s) = V ( s) M m1 ( M m 2 + M s ) + M s1M m 2 + M s (M s1 + M m 2 ) M s1M m 2 + M s ( M s1 + M m 2 ) ( M m 2 + M s ) M m1 + M m2 + M s

Vm =

M m1M m 2 V ( s) M m 2 ( M m1 + M s1 ) + M s ( M m1 + M s1 + M m 2 )

Tm =

M m1M m 2 M m 2 ( M m1 + M s1 ) + M s (M m1 + M s1 + M m 2 )

Show that this is identical to the earlier result

Summary of Thevenin Approach for Mechanical Circuits General Steps

1. Draw the linear graph for the system and mark the mobility functions for all the branches (except the source elements) 2. Simplify the linear graph by combining branches as appropriate (series branches: add mobilities; parallel branches; inverse rule applies for mobilities) and mark the mobilities of the combined branches 3. Based on the problem objective (e.g., determine a particular force, velocity, transfer function) determine which part of the circuit (linear graph) should be cut (i.e., The variable or function of interest should be associated with this part of the circuit) 4. Based on the problem objective establish whether Thevenin equivalence or Norton equivalence is needed (specifically: Use Thevenin equivalence if a through variable needs to be determined, because this gives two series elements with a common through variable; Use Norton equivalence if an across variable needs to be determined, because this gives two parallel elements with a common across variable) 5. Determine the equivalent source and mobility of the equivalent circuit 6. Using the equivalent circuit determine the variable or function of interest

Components Interconnection, Matching

Mechanical Components
Load bearing/structural components (strength and surface properties) (Bearings, springs, shafts, beams, columns, flanges) Fasteners (strength) (Bolts and nuts, locks and keys, screws, rivets, and spring retainers, welding, bracing, soldering) Dynamic isolation components (transmissibility)
(Springs, dampers, shock and vibration mounts, inertia blocks, suspension systems)

Transmission components (motion conversion) (Gears, friction or traction drives, lead screws and nuts, power screws, racks and pinions, cams and followers, chains and sprockets, belts and pulleys or drums, differentials, kinematic linkages, flexible couplings, fluid transmissions) Mechanical actuators (generated force/torque) (Hydraulic pistons and cylinders or rams, hydraulic motors, their pneumatic counterparts) Mechanical controllers (controlled energy dissipation) (Clutches, brakes, hydraulic and pneumatic servo valves)

Harmonic Drive
Fixed Rigid Spline with Internal Teeth (nr)

Wave Generator (Cam) Driven by Input (Actuator) Shaft Annular Spline (Flexispline) with External Teeth (nf) (Output)

Motor Shaft (Input)

2p Tooth pitch of: rigid spline = n radians r

2p 2p 2p nr = n f - nr nr n f n f

2p flexispline = n radians . f

Rotation of flexispline during one revolution of wave generator (around the rigid spline):

=>

Gear reduction ratio (r : 1):

r=

nf n f - nr

Component Matching & Interconnection

Mechatronic system consists of interconnected components Behavior of components in integrated system component operating independently Considerations: Impedance; Signal type; Signal conditioning and modification Significant implications on system performance Example: Electric Motor + Electronic Drive + Mechanical Load 1.Motor: Electrical power Mechanical power 2.Load: Coupled through gear transmission; Mechanical power flows at component interfaces (angular velocity, torque)
3. Electronic

behavior of each

Drive System: Electrical power flows at component

interfaces (voltage, current)

Impedance
Input Impedance Zi: [Rated input voltage]/[Current through input terminals] (while output terminals are maintained in open circuit) Output Impedance Zo: [Open-circuit (no-load) voltage at output port]/ [Short-circuit current at output port] (while rated voltage is applied at input)

Note: Generalized Impedance = [Across Variable]/ [Through Variable] Mechanical Impedance = Force/ Velocity Mechanical Mobility = Velocity/ Force => Generalized Impedance: Mechanical Mobility or Electrical Impedance
vo = G vi + Input vi Zi G + vo Output Zo

Cascade Connection of Devices

+ vi Zi1 G1 + Zo1 vo1 + Zi2 vi2 G2 + Zo2 vo

v o1 = G1 vi

Zi 2 vi 2 = v o1 Z o1 + Z i 2
Zi 2 G2 G1 vi Z o1 + Z i 2

v o = G2 v i 2

Overall input/output relation:

vo =

=> Differs from ideal (G2G1) by factor

Zi2 1 = Z o1 + Z i 2 Z o1 / Z i 2 + 1

If Z o1 / Z i 2 <<1, deviation is insignificant.

Impedance Matching (Electrical Example)

DC power supply (voltage vs, output impedance (resistance) Rs) Load of resistance Rl
+ Rs Rl + vs Vl

vs il = Current through circuit Rl + Rs

Voltage across load

vl = il Rl =

v s Rl Rl + R s
2

v s Rl p l = il v l = Power absorbed by the load [ Rl + Rs ] 2

l For maximum power dR = 0 => l

dp

Rl = Rs

Impedance Matching (Mechanical Example)

Torque source (motor) of torque T and moment of inertia Jm Drives a purely inertial load of moment of inertia JL Torque source is connected to load through a gear of speed reduction ratio r : 1
Torque Source T q T

Jm

JL

Jm

qg

JL Motor Load T rqg

Without gear transmission:

T & & ( J m + J L )q& = T => q& = Jm + JL

Gear r:1 Tg

(i)

Jm Tg/r

qg JL

With gear transmission:

rT T & & & J m rq&g = T - g and J Lq&g = Tg => q&g = 2 (r J m + J L ) r
& q&g r ( J + J ) r (1 + J / J ) Divide (ii) by (i): && = a = 2 m L = 2 L m q (r J m + J L ) (r + J L / J m )

(ii)

r (1 + p ) a= 2 => (r + p)

with p = JL/Jm.

For r = 0, a = 0; For r , a 0.
a (1 + p )[( r 2 + p ) - r 2r ] = = 0 => Peak value of a => r (r 2 + p ) 2

rp =

p and a p = 2 p

1+ p

Mechanical Example (Cont d)

Direct Drive Case (r = 1): a = r = 1. => All curves pass through (1,1). 1. When JL = Jm, pick a direct-drive system (no gear transmission; i.e., r = 1). 2. When JL < Jm, pick a step-up gear with peak value r = 3. When JL > Jm, pick a step-down gear with peak value r.
J L / Jm

p 0.1

rp

ap

0.31 1.74 6

1.0

1.0

1.0

10.0 3.16 1.74

Nondimensional Load Acceleration a 1.74 p = 0.1

p = 10.0

1.0

p = 1.0

0 0.316

1.0

3.16 Transmission Speed Ratio r

Signal Modification
Amplification Filtering Modulation & Demodulation Bridge Circuitry ADC and DAC etc.

Signal Modification Devices

Signal Conditioning (amplification, analog and digital filtering, etc.) Signal Conversion (analog-to-digital conversion, digital-to-analog conversion, voltage-to-frequency conversion, frequency-to-voltage conversion, etc.) Modulation (amplitude modulation, frequency modulation, phase modulation, pulse-width modulation, pulse-frequency modulation, pulse-code modulation, etc.) Demodulation (reverse process of modulation). Other Useful Signal Modification Operations: Sample and hold circuits (in digital data acquisition systems). Analog and digital multiplexers Comparators Curve shaping Offsetting Linearization

Operational Amplifier
(a) +vs (Positive Power Supply Collector vCC) + vi -

vip
(Noninverting)

Inputs vin
(Inverting)

Zi

Zo + Kv i -

Output vo = K v i

(Negative Power Supply Emitter vEE) -vs (b)

1 2 3 4
(c)

Pin Designations:

8 7 6 5

1 2 3 4 5 6 7

Offset Null Inverting Input Noninverting Input Negative Power Supply vEE Offset Null Output Positive Power Supply v

(c) vin vip vo +

Operational Amplifier
v o = Kvi
with

vi = vip - vin

Very high open loop voltage gain K (105 to 10 9) Very high input impedance Zi (2 MW to 10 MW) Low output impedance is low (10 W to 75 W) Typical output vo is 1 to 15 V => vi @ 0 => vip @ vin Note: If we apply a large voltage differential vi (say, 10 V) => op amp saturates => output voltage ~ 15 V Properties of Op-amp: 1. Voltages of the two input leads are (almost) equal (due to high open-loop gain). 2. Currents through each input lead is (almost) zero (due to high input impedance).

Operational Amplifier Performance

Definitions Open-loop gain = Input impedance =
Output voltage with no feedback. Voltage difference at input leads

Voltage between an input lead and ground Current through that lead

(with other input lead grounded and the output in open circuit) Output impedance =
Voltage between output lead and ground in open circuit Current through that lead

(with normal input conditions) Bandwidth = frequency range in which the frequency response is flat (gain is constant). Gain bandwidth product (GBP) = Open-loop gain x Bandwidth at that gain Input bias current = average (DC) current through one input lead Input offset current = difference in the two input bias currents Differential input voltage = voltage at one input lead with the other grounded when the output voltage is zero. Common-mode gain =
Output voltage when input leads are at the same voltage Common input voltage
Open loop differential gain Common-mode gain

Common-mode rejection ratio (CMRR) =

Slew rate = rate of change of output of a unity-gain op-amp, for a step input

Amplifier Performance Ratings

1. Stability 2. Speed of response (bandwidth, slew rate) 3. Unmodeled signals. Unmodeled Signals (a major source of amplifier error): 1. Bias currents 2. Offset signals 3. Common mode output voltage 4. Internal noise.

Voltage and Current Amplifiers (a)

v o = Kv vi
Input vi

Kv = voltage gain
Kv = 1 + Rf R
R

+ A -

Output vo

Rf (b)
R

i o = Ki ii
Ki = current gain
Ki = 1 + Rf R

B ii A Input ii + io (Output) Rf RL Load

Note: Feedback

Instrumentation Amplifiers
(a) Rf

vi1 Inputs vi2

A B + Output vo

vo =

Rf R

( vi 2 - v i 1 )
(b) vi1 + A R1 1

Rf

R4

2R v B - v A = 1 + 1 ( vi 2 - vi1 ) R2

R3

+ R4+d R4 Output vo

R vo = 4 ( v B - v A ) R3

Inputs 2 -

R2
R3 R1 B

vi2

Filter Summary
Active Filters (Need External Power) Advantages: Smaller loading errors and interaction (have high input impedance and low output impedance, and hence don t affect the input circuit conditions, output signals and other components). Lower cost Better accuracy Passive Filters (No External Power, Use Passive Elements) Advantages: Useable at very high frequencies (e.g., radio frequency) No need of a power supply

Filter Types
Low Pass: Allows frequency components up to cutoff and rejects the higher frequency components. High Pass: Rejects frequency components up to cutoff and allows the higher frequency components. Band Pass: Allows frequency components within an interval and rejects the rest. Notch (or, Band Reject): Rejects frequency components within an interval (usually, a narrow band) and allows the rest.

Definitions
Filter Order: Number of poles in the filter circuit or transfer function Anti-aliasing Filter: Low-pass filter with cutoff at less than half the sampling rate (i.e., at less than Nyquist frequency), for digital processing. Butterworth Filter: A high-order filter with a quite flat pass band. Chebyshev Filter: An optimal filter with uniform ripples in the pass band Sallen-Key Fitler: An active filter whose output is in phase with input.

Filters (Ideal)
(a) w
Magnitude

G( f )
1 0

fc = Cutoff frequency

fc

Frequency f

(b)

G( f )
1 0

fc

(c)

G( f )
1 0

fc1

fc2

(d)

G( f )
1 0

fc1

fc2

(a) Low-pass; (b) High-pass; (c) Band-pass; (d) Band-reject (Notch)

Tracking Filter

Input Channel 1

Output Channel 1

Tracking Filter
Input Channel 2 Output Channel 2

Carrier Input (Tracking Frequency)

Modulation
(a)
Time t

(b)
t

(c)
t

(d)
t

(e)
t

(a) Modulating signal (data signal); (b) Amplitude-modulated (AM) signal; (c) Frequency-modulated (FM) signal; (d) Pulse-width-modulated (PWM) signal (pulse period is fixed); (e) Pulse-frequency-modulated (PFM) signal (pulse width is fixed).

Modulation Theorem

(a) A transient data signal and its Fourier spectrum magnitude; (b) Amplitude-modulated signal and its Fourier spectrum magnitude; (c) A sinusoidal data signal; (d) Amplitude modulation by a sinusoidal signal.

Amplitude Modulation and Demodulation

Modulating Input (Data)

Multiplier Modulated Signal Out

Carrier SIGNAL

(a)

~ X( f )
M

M 2
-2 fc -fb 0 fb 2 fc
Frequency f

(b)
Modulated Signal

Multiplier
~ (t ) x
OUT

x a (t )
Carrier

Original Signal

Low-Pass Filter
Cutoff = fb

x (t )

2 cos 2pf c t ac

Wheatstone Bridge
(a) R1 A R2 R4 B
vref (Constant Voltage)

Small i

+ RL

vo Load (High)

R3

(b) R1

A R2 R4 B
iref (Constant Current)

Small i

+ RL

vo Load (High)

R3

(c) A R C R+ R B + vref
+

Rf

R + D R

vo

(a) Constant-voltage resistance bridge; (b) Constant-current resistance bridge; (c) A linearized bridge.

Bridge Equations
(a) Constant-Voltage Bridge (Wheatstone Bridge)
vo = v A - vB = R1vref (R1 + R2 ) R3vref (R3 + R4 ) = (R1 R4 - R2 R3 ) vref (R1 + R2 )(R3 + R4 )

Set R1 = R2 = R3 = R4 = R (Balanced bridge) and increase R1 by dR:

dv o dR / R = v ref (4 + 2dR / R)

Linearize:

dvo dR = v ref 4R

(b) Constant-Current Bridge

vo = (R1 R4 - R2 R3 ) iref (R1 + R2 + R3 + R4 )

dv o dR / R = Riref (4 + dR / R) => Less nonlinear than (a)

(Compare denominators). (c) Linearized Bridge

dvo R f dR = vref R R

Half Bridge Circuit

+vref Rf
R1

(Active Element) A R2 (Dummy) -vref

+ Output vo

R dR dvo R f dR = ; f vref R ( R + dR ) R R

(when linearized)

Note: Actually more nonlinear than Wheatstone bridge

Impedance (AC) Bridge Circuits

(a) General Impedance Bridge
(a) Z1 Z2

Output vo

Z3

Z4

~ vref (AC Supply)

(b) Owen Bridge

(b)

C1 L4 R3 C3 ~ vref R4

R2

Output vo

(c) Wien Bridge

(c) R1
C4

R2

vo

R3 C3 ~ vref

R4

Performance Specification

Performance Specification Parameters

Categories:
Speed of Response Stability Accuracy (e.g., Steady-state Error or Offset)

Time-domain Performance Specification

Response y Mp Percentage Overshoot Boundary

Steady-State Error Boundary

P.O.

2%
100

1.0 0.9 Settling Time Boundary 0.5 Speed of Response Boundary 0.1 0

0 T T d r Trd

Tp

Ts

Time t

Rise Time Peak Time Peak Value

Tr =

p -f with cos f = z wd Tp = p wd
p z 1- z 2

Mp = 1 - e
Percentage Overshoot (P.O.)
-

p z 1- z

P. O.= 100 e
Time Constant
t = 1 zw n

Settling Time (2%)

Ts = -

ln 0.02 1 - z 2 zw n

] 4t =

4 zw n

Automobile Suspension System Example

Automobile weight = 1000 kg Equiv. stiffness at each wheel (including suspension system) = 60.0 103 N/m. Performance spec for suspension: Percentage overshoot = 1% Estimate the required damping constant at each wheel.

&& & Simple Oscillator Model: my + by + ky = ku (t )

2 2 & y => && + 2zw n y + w n y = w n u(t )

m = equivalent mass = 250 kg; b = equivalent damping constant k = equivalent stiffness = 60.0 103 N/m; u = displacement excitation at the wheel.
b z = Damping ratio 2 km

Note: Equivalent mass at each wheel = 1/4th the total mass. For a P.O of 1%: Substitute: 0.83 =
1 = 100exp(pz 1-z
2

=> z = 0.83

b 2 60 10 250.0
3

=> b = 6.43 10 3 N/m/s.

Frequency-domain Performance Specification

Useful Frequency Range (operating interval)
Gain (Magnitude)

Bandwidth (speed of response) Static Gain (steady-state performance) Resonant Frequency (speed and critical frequency region) Magnitude at Resonance (stability) Input Impedance (loading, efficiency, interconnectability)

Mo = static gain (DC Gain) fr = resonant frequency Slop within 1 dB/decade Mo Useful Frequency Range Df fmin fmax fr Frequency (Hz)

Phase Angle

Output Impedance (loading, efficiency, interconnectability) Gain Margin (stability) Phase Margin (stability)

Stability (in Frequency Domain)

Gain Margin (GM) and Phase Margin (PM) should be high
(a)
Phase lag = f1

Input u
Phase lag = p

G(jw)
-

Output y

H (jw)
(b)
Gain 0 dB

GM

Phase

-p

PM

wc

wp

Frequency (log)

(a) A feedback system; (b) Bode diagram (Illustration of GM and PM.

Manifestations of Nonlinearity
(a) Saturation Level (b) (c)

Device Input Saturation Level

Linear Range

-d

Device Input

Device Input

(d) Hardening Nonlinearity D A (Jump) C (Jump) B Frequency Softening Nonlinearity C (Jump) B

(e)

Speed Limit Cycle (Stable)

A (Jump)

Displacement

Frequency

(a) Saturation; (b) Dead zone; (c) Hysteresis; (d) The jump phenomenon; (e) Limit cycle response.

Other Considerations
Frequency Creation At steady state, nonlinear devices can create frequencies that are not present in the excitation signals. Methods to Reduce/Eliminate Nonlinear Behavior 1. Calibration (static case) 2. Use of linearizing elements (resistors, amplifiers) 3. Use of nonlinear feedback 4. Avoid operating the device over a wide range of signal levels. 5. Avoid operation over a wide frequency band. 6. Use devices that do not generate large mechanical motions. 7. Minimize Coulomb friction and stiction (e.g., by using proper lubrication). 8. Avoid loose joints and gear coupling (i.e., use direct-drive mechanisms).

Instrument Ratings Parameters

1. Sensitivity (Measured by magnitude of output for a unit input. May be expressed as: (incremental output) / (incremental input) or a partial derivative. sensitivity should be specified; e.g., direct or cross) 2. Dynamic Range (Determined by lower and upper limits of operating variable. Usually expressed as a ratio, in decibels Typically, lower limit of dynamic range = resolution of the device => Dynamic range = (range of operation) / (resolution). 3. Resolution (smallest change in operating variable that can be detected/measured. Usually expressed as a % of maximum range or inverse of the dynamic range ratio). 4. Linearity (closeness of calibration curve to a straight line; e.g., least squares fit) 5. Zero Drift (drift from the null reading) and Full-scale Drift (drift while input is maintained at full-scale value) 6. Useful Frequency Range (flat gain curve and a zero phase curve in the frequency response characteristics. Typically less than half; say, one-fifth the dominant resonant frequency. Measure of bandwidth) 7. Bandwidth (measure of maximum speed or frequency at which the instrument is capable of operating. Directly related to the useful frequency range) 8. Input and Output Impedances. Direction of

Output Full scale reading Static calibration curve

Linear approximation

Maximum deviation from the Calibration curve

Dynamic range

Input

dr Linearity = Full scale reading 100%

Bandwidth
Has different interpretations depending on context and application. Example Interpretations: 1. Useful Frequency Range: Flat region (static region) of spectrum (e.g., fBW = 0.25 fr ; slope smaller than 1 dB/decade; etc.). In this range, instrument dynamics will not corrupt signal. 2. For Measuring Instruments: Frequency range within which measurements are accurate 3. Relates to Fundamental Resonant Frequency of system System speed (response speed for a given excitation) 4. For Band-pass Filter: Frequency band within which the signal components are allowed through it 5. In Digital Communication Network: Capacity (information rate) 6. Frequency Uncertainty in Data. Note: Signal through a band-pass filter Signal frequency content not known exactly. The larger the filter bandwidth, less certain the actual frequency content of a passed signal

Control Bandwidth
Maximum possible speed of control. Frequency range within which a system can be controlled (assuming that all the devices in the system can operate within this bandwidth). Important specification in both analog control and digital control. In digital control, the data sampling rate (in samples/second) has to be several times higher than the control bandwidth (in hertz) so that sufficient data would be available to compute the control action. From Shannon's sampling theorem, control bandwidth is given by half the rate at which the control action is computed (to reduce aliasing).

Aliasing Distortion Due to Signal Sampling

Sampling Theorem
Signal x(t) is sampled at equal time steps DT no information regarding its frequency spectrum X(f) is obtained for frequencies higher than fc = 1/(2DT). Note: Limiting (cut-off) frequency f is called Nyquist Frequency.
(a)
Spectral Magnitude

0 (b)
Spectral Magnitude

fc

Frequency f

fc = Nyquist frequency

Aliasing Folded High-Frequency Spectrum 0 Original Spectrum

fc

Frequency f

Illustration of Aliasing
(a)

f1 = 0.2 Hz f2 = 0.8 Hz
Signal

1 2 3 4

Sampling rate fs = 1 sample /s Nyquist frequency fc = 0.5 Hz (b) Amplitude Spectrum

f1
0 0.2

fc
0.5

f2
0.8 Frequency (Hz)

Methods of Reducing Aliasing

1. Increase sampling rate (Disadvantages: Increases the signal processing time and computer storage requirements; Can result in stability problems in numerical computations). 2. Add a low-pass filter with cut-off frequency Example: Consider 1024 data points from a signal, sampled at 1 millisecond (ms) intervals. Sampling rate fs = 1/0.001 samples/s = 1000 Hz = 1 kHz Nyquist frequency = 1000/2 Hz = 500 Hz Due to aliasing, approximately 20% of the spectrum (i.e., spectrum beyond 400 Hz) will be distorted. Digital Fourier transform computation provides 1024 frequency points of data up to 1000 Hz. Half of this number is beyond the Nyquist frequency, and will not give any new information about the signal. Spectral line separation = 1000/1024 Hz = 1 Hz (approx.) Keep only the first 400 spectral lines as the useful spectrum. Note: Almost 500 spectral lines may be retained if an anti-aliasing filter is used. Nyquist frequency fc (Anitf c / 128( @ 0.8 f c ) .

aliasing filter). Typically, filter cutoff frequency =

Bandwidth Design Example

Mechanical position application with digital control
ke Drive amplifier and motor electromagnetic circuit (torque generator): S 2 + 2z w s + w 2 e e e

km Mechanical system including motor rotor inertia): s 2 + 2z w s + w 2 m m m

k = equivalent gain; z = damping ratio; w = natural frequency

Power Supply Damper DAC Drive Amplifer Control Computer Encoder (Position Sensor) Feedback Motor Inertia Spring

DTc = tim taken to com each control action; DTp = pulse period of the position sensing encoder. e pute
we = 1000p rad/s; z e = 0.5; wm = 100p and 500 p rad/s ; z m = 0.3 ; DTc = 0.02 s

Solution
Drive system resonant frequency ~ 500 Hz => flat region of drive system spectrum ~ 1/10th = 50 Hz = limit of max spectral component of drive signal = hardware limit on control bandwidth
1 1 = Digital control signal rate DT 0.02 Hz = 50 Hz c

1 Effective (useful) spectrum of digital control signal ~ 2 x50 Hz = 25 Hz (sampling theorem)

Control bandwidth ~ 25 Hz (Note: Governed by digital control even though drive system hardware can accommodate a bandwidth of about 50 Hz)

DTp = sampling period of measurement signal (for feedback) =>

Useful spectrum limit of measured signal = 2DT (Shannon s sampling theorem) p (Feedback signal will not carry useful process info beyond frequency 2DT ) p To generate a control signal at up to DT Hz c => guarantees
1
1 1
1 DTc

samples/s, process information has to be provided at least

=> 2DT DT or DTp 0.5 DTc . p c 2 points of sampled sensor data for computing a control action.

Solution (Cont d)
100p % Case 1: Plant (positioning system) resonant frequency ~ 2p Hz - 50 Hz :

At frequencies near 50 Hz, resonance will interfere with control (avoid if possible) At frequencies >> 50 Hz, process will not significantly respond to control action (will not be of much use because plant will be felt like a rigid wall). => operating BW has to be sufficiently < 50 Hz (say 25 Hz), to avoid plant dynamics. Note: Is a matter of design judgment, based on application (e.g., excavator, disk drive). Typically, plant dynamics must be controlled => use entire control BW (i.e., max possible control speed) as operating BW (regardless the plant resonance). In the present case, even if the entire control BW (i.e., 25 Hz) is used as the operating BW, it still avoids the plant resonance (50 Hz).
Case 2: Plant resonance ~ 2p Hz ~ 250 Hz 500p

=> Operating bandwidth ~

250p Hz - 125 Hz , % 2

to avoid plant dynamics.

DTc = 0.02 sec ).

=> But, control bandwidth ~ 25 Hz (because

Operating bandwidth cannot be greater than this value => Operating bandwidth ~ 25 Hz.

Bandwidth Design of a Mechatronic System

Step 1: Decide on max frequency of operation (BWo) of system based on requirements of particular application. Step 2: Select plant components (electro-mechanical) that have the capacity to operate at BWo and perform the required tasks. Step 3: Select feedback sensors with a flat frequency spectrum (operating frequency range) > 4 BWo. Step 4: Develop a digital controller with Sensor sampling rate > 4 BWo (keeping within flat spectrum of sensors) Direct-digital control cycle time (period) of 1/(2 BWo). Note: digital control actions are generated at a rate of 2 BWo. Step 5: Select control drive system (interface analog hardware, filters, amplifiers, actuators, etc.) with flat frequency spectrum > BWo. Step 6: Integrate the system and test the performance. If performance specs are not satisfied, make necessary adjustments and test again.

Instrument Error Analysis

Error = (instrument reading) - (true value) Note: Mechatronic systems contain multiple components Difficulties in Error Analysis: 1. Correct (true) value is usually unknown 2. Instrument reading may contain random error (cannot be exactly determined) 3. Error may be a complex (i.e., not simple) function of many variables (input variables, state variables, response variables) 4. The system/instrument may be made up of many components having complex interrelations (dynamic coupling, multiple dof responses

Accuracy and Precision

Accuracy => Inverse of Error Measurement Error; Instrument Error Deterministic Errors (Systematic Errors):
Caused by well-defined factors (nonlinearities, offsets in readings, etc.). Can be handled by proper calibration and analysis. Error ratings and calibration charts are used

Random (Stochastic) Errors:

Caused by uncertain factors entering into instrument response. (Device noise, line noise, and effects of random variations in operating environment, etc.). Statistical analysis using large amounts of data is necessary to estimate random errors. => Mean Value m e and a standard deviation s e Note: If s e = 0 => Assume error to be deterministic Precision = (Measurement range) / s e Note: An instrument with high deterministic error can have a high precision (e.g., precise clock with wrong time setting)

Error Combination
Important to Know How: Component errors are propagated within a multicomponent system Individual errors in system variables/parameters contribute toward overall error

Overall error in a mechatronic system depends on error levels of individual components (sensors, actuators, controller hardware, filters, amplifiers, etc.) and on the manner in which these components are physically interconnected/interrelated.

y = f(x1, x2, . . . ,xr)

into a dependent variable y.

(i)

xi = independent system variables/parameters values whose error is propagated

Note: Determination of f is not always simple, and it may be in error. Denote error in a variable by its differential (for small errors) of that variable. Differential of (i):
f f f dy = dx1 + dx 2 +K+ dx r x1 x 2 x r
r x f dx i dy = i => y i =1 y x i x i

dy y = overall fractional error

dx i x i = component error (fractional)

Error Combination Methods

Absolute Error Error dx i could be either positive or negative => upper-bound combined error
eABS =
i =1 r

x i f ei y x i

Note: In the expression, terms e are positive. When specifying error, both +ve and -ve limits should be indicated or implied (e.g.,
e ABS , ei ).

SRSS Error A nonconservative error estimate is given by

eSRSS r x f 2 = i ei i =1 y x i
1/ 2

Note: This is not an upper bound estimate for error. Note: eSRSS < eABS when there are two or more nonzero error contributions Note: SRSS method is particularly suitable when component error is represented by the standard deviation of the associated variable/parameter value and when the corresponding error sources are independent.

Example
An optical device for measuring displacement (optical potentiometer) Potentiometer element is uniform and has resistance Rc. A photoresistive layer is sandwiched between Rc and a perfect conductor. Moving light source directs a light beam of intensity is I, on to a narrow rectangular region of photoresistive layer. As a result, this region becomes resistive with resistance R, bridging the potentiometer element and the conductor element, An empirical relation between R ( k ) and I (watts
R I 2 ln = 0 per square metre or W/m ): R0 I
Rx Photoresistive Layer Resistance R Conductor Light Intensity I
14

Parameters R0 and I0 are empirical constants having some error.

Rc - Rx Signal Acquisition Circuit

DC Supply

Light Source (moves with the measured object)

Example (Cont d)
Resistance R Exponential Curve

2.7Ro Typical Operating Range 1.75Ro Ro

Io

10Io

Current I

Ro = minimum resistance provided by photoresistive bridge (at very high light intensity levels). When I = I o bridge resistance R ~ 2.7 Ro => Io ~ lower bound for intensity for proper operation of sensor. A suitable upper bound: I ~ 10 Io, for satisfactory operation => R ; 1.75Ro .

Example (Cont d)
Differentiate given relation:
d R d Ro 1 I o = R Ro 4 I
-3/ 4

d Io Io = 1 Io - 2 dI 4 I I I

1/ 4

d Io d I - => I Io
1/ 4

1 I eR = eRo + o With the absolute method of error combination: 4 I

eIo + eI

Empirical value of I0 = 2.0 W/m2. Empirical errors: eR0 = 0.01 and eI 0 = 0.01 Due to variations in supply to light source and in ambient lighting conditions, fractional error in I = 0.01 . Error ER is to be maintained within 0.02
1 Io Substitute values: 0.02 = 0.01 + 4 I
1/ 4

[0.01 + 0.01] Io 1/ 4 = 2 =>

I
1 2.0

2 2 Light source should operate at intensity level I = 16 I o = 16 W/m = 0.125 W/m

Note: For larger values of I the absolute error in Ro would be smaller. For
1 1 eR = 0.01 + example, for I = 10 Io: 4 10
1/ 4

[0.01+ 0.01] ; 0.013.

Example (Cont d)
Advantages Noncontacting Small moving mass (low inertial loading) All advantages of a potentiometer. Disadvantages Nonlinear and exponential variation of R Effect of ambient lighting Possible nonlinear behavior of the device (input-output relation) Effect of variations in the supply to the light source Effect of aging of the light source.

Sensors

Analog Sensors and Transducers

Uses of Sensors in a Mechatronic System: 1. 2. 3. 4. To measure the system outputs for feedback control. To measure system inputs (desirable inputs, unknown inputs, and disturbances) for feedforward control. To measure output signals for system monitoring, diagnosis, evaluation, parameter adjustment, and supervisory control. To measure input and output signals for system testing and experimental modeling (i.e., for system identification). Measurand: Variable that is being measured Measurement: Output/reading of the measuring device Two Stages in a Measuring Device: 1. Measurand is felt or sensed 2. Measured signal is transduced (or converted) into the form of device output. Note: A sensor which senses the response automatically converts (transduces) this into the sensor output Note: Terms sensor and transducer are used interchangeably to denote a sensortransducer unit.

Analog Sensors
Measurand (Typically analog) Signal sensor Transmittable variable Transducer (Typically Electrical)

Acceleration

Stress Voltage (Strain) (Charge) Generation Generation

Electric voltage

Piezoelectric Accelerometer

Biological Transducers

Sensors/Transducer Types
Pure Transducers: Depend on nondissipative coupling in the transduction stage (no wastage of signal power). Passive Transducers (Self-generating Transducers): Depend on their power transfer characteristics for operation (Do not need an external power source). Note: Pure transducers are passive devices. They derive their energy from the measurand => tend to distort (or load) the measured signal. Active Sensors/Transducers: Require external power for operation (Do not depend on power conversion characteristics for their operation). Note: Power refers to that in the immediate transducer stage (not the power used in subsequent signal conditioning).

Perfect Measurement Device

It Possesses Following Characteristics: 1. 2. 3. Output of the measuring device instantly reaches the measured value (Fast Response). Transducer output is sufficiently large (High Gain, Low Output Impedance, High Sensitivity). Device output remains at the measured value (without drifting or being affected by environmental effects and other undesirable disturbances and noise) unless the measurand (what is measured) itself changes (Stability and Robustness). 4. 5. 6. The output signal level of the transducer varies in proportion to the signal level of the measurand (Static Linearity). Connection of a measuring device does not distort the measurand itself (Loading effects are absent and Impedances are matched). Power consumption is small (High Input Impedance).

Motion Transducers
Motion:
Displacement (position, distance, proximity, size, gage, etc.) Velocity Acceleration Jerk. Note: Direct relationship may not exist between a measuring device and a measured variable. Example: Strain gages can be adapted to measure displacements by using a front-end auxiliary sensor element (e.g., cantilever or spring). Note: Same measuring device may be used to measure different variables through appropriate data interpretation techniques. In principle, a force sensor can be used as an acceleration sensor, velocity sensor, or displacement sensor, depending on whether the front-end auxiliary sensor is: 1. Inertia element (converts acceleration into force), 2. Damping element (converts velocity into force), or 3. Spring element (converts displacement into force)

Limitations of Sensor Signal Conversion

Displacement, Velocity, Acceleration, Jerk: Related through integration or differentiation. In theory, we can measure only one variables and through analog processing (analog circuit hardware) or digital processing (dedicated processor) convert to another motion variables. Limitations of Feasibility of the Approach: 1. 2. 3. 4. 5. 6. Nature of measured signal (steady, highly transient, periodic, narrow/broad-band), etc. Required frequency content of processed signal (frequency range of interest) Signal-to-noise ratio (SNR) of the measurement Available processing capabilities (e.g., analog or digital processing, limitations of digital processor and interface: speed of processing, sampling rate, buffer size, etc.) Controller requirements and nature of plant (e.g., control bandwidth, operating bandwidth, time constants, delays, complexity, hardware limitations) Required task accuracy (processing requirements, hardware costs depend on this). Example: Differentiation of signal (in time domain) is unacceptable for noisy and highfrequency narrow-band signals. Costly signal-conditioning hardware may be needed for pre-preprocessing. Rule of Thumb: Low-frequency applications (~ 1 Hz): Use displacement measurement. Intermediate-frequency applications (< 1 kHz): Use velocity measurement. High-frequency motions with high noise levels: Use acceleration measurement.

Potentiometer
Resistive Element Wiper Arm vref (Supply) i + vo (Measurement) No Current x (Measurand) vref i + Load % Z vo Impedance Nonzero Current

Resistive Element RC

q max
output vo Slider Arm R

q
Load RL

Supply vref

Loading Nonlinearity
vo (q q max )( RL RC ) = vref RL RC + (q q max ) - (q q max )2

1.0 0.9

Potentiometer Reading vo/vref

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Displacement /

max

RL/RC = 0.1 RL/RC = 1.0 RL/RC = 10.0

1. Increase RL/RC (increase load impedance, reduce coil impedance). 2. Use pots to measure small values of of the element for linear reading).
q q max

(or calibrate only a small segment

Performance of Potentiometer
Limitations:
1. 2. The force needed to move the slider (against friction and arm inertia) is provided by the displacement source => mechanical loading =>distorts measured signal. High-frequency (or highly transient) measurements are not feasible because of such factors as slider bounce, friction and inertia resistance, and induced voltages in the wiper arm and primary coil. 3. 4. 5. 6. Variations in the supply voltage cause error. Electrical loading error can be significant if the load resistance is low. Resolution is limited by the number of turns in the coil and by the coil uniformity. This will limit small-displacement measurements. Wearout and heating up (with associated oxidation) in the coil or film, and slider contact cause accelerated degradation.

Advantages:
1. 2. They are relatively inexpensive. Provide high-voltage (low-impedance) output signals, requiring no amplification in most applications. Impedance can be varied simply by changing the coil resistance and supply voltage.

Variable-Inductance Transducers
They employ the principle of electromagnetic induction. Electromechanical devices coupled by a magnetic field. Three Primary Types: 1. 2. 3. Mutual-induction transducers Self-induction transducers Permanent-magnet transducers.

Variable-reluctance Transducers: Variable-inductance transducers that use a non-magnetized ferromagnetic medium to alter the reluctance (magnetic resistance) of the flux path Note: Some of the mutual-induction transducers and most of self-induction transducers are of this type. Note: Permanent-magnet transducers are not considered variable-reluctance transducers.

Linear-Variable Differential Transformer (LVDT)

Variable-inductance transducer; Variable-reluctance transducer; Mutual-induction transducer; Passive transducer

vo (Measurement) Insulating Form

Primary Coil

Housing

Ferromagnetic Core

Core Displacement x (Measurand)

Secondary Coil Segment

vref (b)

Secondary Coil Segment

See the animation at: http://www.rdpe.com/displacement/lvdt/lvdt-principles.htm

Linear-Variable Differential Transformer (LVDT)

vo

Secondary Coil

Core Primary Coil vref

vo

Voltage level

Displacement x Linear Range

Signal Conditioning for LVDT

x(t)
Carrier Signal R C

v p sin w c t

v1

+ _

v2 R2

v3 R1

Output

vo
+

R1
LVDT Amplifier Multiplier

R1

Low-pass Filter

x(t )

v1

v2

v3

vo

v3 =

v p 2 rk 2

x(t ) [1 - cos 2 wc t ]
v p 2 rk 2 x (t ) [1 - cos 2 w c t ] ; t = RC = filter time constant; ko = R/R1

Frequency Domain: v3 =

Advantages of LVDT
1.

Noncontacting

device

with

no

frictional

resistance.

Near-

ideal electromechanical energy conversion and light-weight core => very small resistive forces. Hysteresis (both magnetic hysteresis and mechanical backlash) is negligible. 2. Low output impedance ~ 100 W . (Signal amplification is not usually needed beyond what is provided by the conditioning circuit.) 3. Directional measurements (positive/negative) are obtained. 4. Available in small sizes (e.g., 1 cm long with maximum travel of 2 mm). 5. Has a simple and robust construction (inexpensive and durable). 6. Fine resolutions are possible (theoretically, infinitesimal resolution; practically, much better than a coil potentiometer).

DC Tachometer
Permanent-magnet velocity sensor; Uses electromagnetic induction Rectilinear speeds or angular speeds can be measured.
Permanent Magnet

Moving Coil

Output vo (Measurement)

Velocity v (Measurand) (a)

Commutator

Speed wc

Permanent Magnet

wc

S Rotating Coil vo Rotating Coil 2r

(b)

AC Tachometers
(a)

AC Carrier Source ~ vref PermanentMagnet Rotor

Output vo

Primary Stator

Secondary Stator

(b)

AC Carrier Source ~ vref Shorted Rotor Coil

Output vo

Primary Stator

Secondary Stator

(a) An ac permanent-magnet tachometer (b) An ac induction tachometer.

Eddy Current Transducers

Coaxial Cable Output vo Calibrating Unit
Impedance Bridge Demodulator Low-Pass Filter

Compensating Coil

(Measurand) x Target Object Conducting Surface

RF Signal (100 MHz) Radio Frequency Converter (Oscillator)

Active Coil

20 V DC Supply (a)

C Compensating Coil Active Coil C L

R1 RF Generator

R2

Bridge Output (to Demodulator)

~
L+ L R1 R2

(b)

Variable-Capacitance Transducers
Capacitance Bridge DC Output vo Fixed Plate Rotation A

(a) Capacitive rotation sensor

q
Rotating Plate (a) Capacitance Bridge vo

(b) Capacitive displacement sensor

Position x Moving Plate (e.g., Diaphragm) (b)

Fixed Plate

Capacitance Bridge Fixed Plate Liquid

vo

Level h

(c) Capacitive liquid level sensor

(c)

Piezoelectric Accelerometer
Spring Direction of Sensitivity (Input) Inertia Mass Output vo Piezoelectric Element

Electrodes

Resonance

Accelerometer Signal (dB)

Useful Range

5,000 20,000 Frequency (Hz)

Charge Amplifier
Piezoelectric devices need a charge amplifier.
Primary Reasons: 1. 2. High output impedance in sensor => small output signal levels and large loading errors. The charge can quickly leak out through the load.
Rf Cf A -vo/K q C Cc + K + Output vo Charge Amplifier

Note: Feedback resistance and capacitance

R f jw vo ( jw ) =q( jw ) Rf C f jw + 1

Piezoelectric Sensor

Cable

=>

G( jw ) =

jt cw [ jtcw + 1]

tc = Rf C f

Torque Sensing in Robotic Joints

Link 1

Stator
q

Tm TJ

Motor Torque Tm Joint Angle q Link 2 Rotor J

Transmitted Torque TJ

Rotor

& Damping bq
(b)

(a) Joint Torque (Transmitted) TJ Motor Torque Tm

Joint Input Voltage vi

Drive Amplifier

Rotor

1 s ( Js + b )

Joint Response

(c)

Feedforward Signal Joint Command

TJ Torque Sensor Drive Amplifier (d)

q

Joint Controller

Rotor

Torque Sensing Methods

1. 2. Measure strain in a sensing member between drive element and driven load, using a strain gage bridge Measure displacement in a sensing member (as in Method 1)--either directly, using a displacement sensor, or indirectly, by measuring a variable, such as magnetic inductance or capacitance, that varies with displacement 3. 4. 5. 6. Measure reaction in support structure or housing (by measuring a force) and the associated lever arm length In electric motors, measure the field or armature current that produces motor torque; in hydraulic or pneumatic actuators, measuring actuator pressure Measure torque directly, using piezoelectric sensors, for example Employ a servo method-- balance the unknown torque with a feedback torque generated by an active device (say, a servomotor) whose torque characteristics are precisely known 7. Measure angular acceleration caused by torque in a known inertia element.

Strain Gages
dR = Ss e R

Direction of Sensitivity Foil Grid Backing Film

Ss = gage factor or sensitivity Use a bridge circuit =>

dv o = Ce v ref
Solder Tabs (For Leads) (a)

(a) Strain gage nomenclature

Single Element

Two-Element Rosette

(b) Typical foil-type strain gages

Three-Element Rosettes (b) Doped Silicon Crystal (P or N Type)

Welded Gold Leads

Nickle-Plated Copper Ribbons

(c) A semiconductor strain gage

Phenolic Glass Backing Plate (c)

Torque Sensor Location

Drive Link Motor Stator Tachometer

Gearing

D Ball Bearings Resolver

Motor Rotor

Driven Link

Bearing Tf1 D

Driven Link TL C

Bearing Tf2 B A TI

Motor

Tm Torque
Tm TI Tm TI TL Tf1 = Tm TI Tf2 TL Tf2

Tm

Axial Location

Example: Force sensor

Rating Parameters of Sensors and Transducers

Measurand Frequency Transducer Measurand Max/Min Output Impedance Potentiometer LVDT Resolver Displacement Displacement Angular displacement 5 Hz/DC 2,500 Hz/ DC 500 Hz/ DC (limited by excitation freq) Tachometer Velocity 700 Hz/ DC Moderate (50 W ) Eddy current proximity sensor Piezoelectric accelerometer Semiconducto r strain gage Acceleration (and velocity, etc.) Strain (displacement, acceleration, etc.) Loadcell Force (10 - 1000 N) Laser Displacement/ Shape Optical encoder Motion 100 kHz/ DC 500
W

Typical Resolution 0.1 mm 0.001 mm or less 2 min.

Accuracy

Sensitivity

Low Moderate Low

0.1% 0.3% 0.2%

200 mV/mm 50 mV/mm 10 mV/deg

0.2 mm/s

0.5%

5 mV/mm/s 75 mV/rad/s

Displacement

100 kHz/ DC

Moderate

0.001 mm 0.05% full scale 1 mm/s2

me

0.5%

5 V/mm

25 kHz/ 1Hz

High

1%

0.5 mV/m/s2

1 kHz/ DC (limited by fatigue) 500 Hz/ DC

200

1 - 10
6

(1 me =10-

1%

1 V/ e , 2,000
me

unity strain) 0.01 N 1.0 mm 0.05%

max

Moderate

1 mV/N

1 kHz/

DC

100

0.5% bit

1 V/mm 104/rev.

10 bit

Tactile Sensors
Force distribution is measured, using a closely spaced array of force sensors (usually exploiting skin-like properties of sensor array).
Important in: 1. Grasping 2. Object identification. Considerations: 1. Spatial resolution of about 2 mm 2. Force resolution (sensitivity) of about 2 gm 3. Force capacity (maximum touch force) of about 1 kg 4. Response time of 5 ms or less 5. Low hysteresis (low energy dissipation) 6. Durability under harsh working conditions 7. Robustness and insensitivity to change in environmental conditions (temperature, dust, humidity, vibration, etc.) 8. Capability to detect and even predict slip.

Optical Tactile Sensor

Position 1 of Reflecting Surface Position 2 of Reflecting Surface Intensity at Receiver Light (or Laser) Receiver Light (or Laser) Source (a) Position x x

Tactile Forces Elastomeric Touch Pad Transparent Elastomeric Layer

Reflecting Surface Fixed Array of Optical Fiber Beam Splitter Solid-State Camera

Light (or Laser) Source

Image Processor

Deflection or Force Profile

(b)

Digital Transducer
A measuring device that produces a digital output A transducer whose output is a pulse signal A transducer whose output is a frequency Advantages: 1. Less susceptible to noise, disturbances, or parameter variation in instruments (data generated/represented/transmitted/processed as bits: two states). 2. Complex signal processing with very high accuracy and speed by digital means (Hardware implementation is faster than software implementation). 3. High reliability by minimizing analog hardware components. 4. Large amounts of data may be stored using compact, high-density methods. 5. Data can be stored/maintained for very long periods of time without any drift or being affected by adverse environmental conditions. 6. Fast data transmission over long distances without significant dynamic delays (unlike analog systems). 7. Digital signals use low voltages (e.g., 0-12 V DC) and low power. 8. Digital devices typically have low overall cost.

Incremental Optical Encoder

(a) vo Output LED Light Source t Encoder Track Transparent Windows

Phototransistor, Or Photodiode (Light Sensor)

Opaque Disk (Code Disk) (b)

Cover and Connector

Electronics Assembly Photodetector Array Code Disc and Spindle Assembly Light Source and Mask

Bearing Housing Assembly

Direction Sensing
Pick-Off Pick-Off 1 2 v1 90 Lags by 90 Time v2 Reference Window ,w Code Disk Reference Pulse Pick-Off v1 90 Lags by 90 Time t Time t t

(a)

Counting (timing) begins when the v1 signal begins to rise (i.e., when a rising edge is detected). n1 = number of clock cycles (time) up to time when v2 begins to rise n2 = number of clock cycles up to time when v1 begins to rise again. If n1 > n2 - n1 cw rotation If n1 < n2 - n1 ccw rotation.

v2

(b)

Time

Time (c)

Hardware Features
Collimating Lens Focusing Lens Photodiode Amplifier LED Two Photodiodes (Half-Pitch Offset) Photodiode Amplifier 5 V DC Supply

+ Differential Amplifier and Thresholding

Pulse Signal Output v1

Code Disk (Moving)

Stationary Masking Disk (a)

Integrated Circuit Located Within Encoder Housing

LED

Sensors (Half-Pitch Spacing)

Moving (Code) Disk

Masking Disk (b)

Encoder Resolution
Displacement Measurement

q=

n q max M

Count = n; Maximum count possible = M pulses; Range of the encoder = q max

180o 360o Digital resolution Dq d = 2r -1 = 2 r

(Number of bits = r)

Physical Resolution

Dq p =

360o 4N

(with quadrature signals)

Velocity Measurement Pulse-counting method (For high speeds): Speed w =

2p / N 2p n = (count during a time period T is n) T /n NT

Velocity resolution Dwc =

2p NT

Pulse-timing method (For low speeds): Speed w = m / f = Nm (m clock counts with clock frequency f Hz)
2p f Nw 2 Velocity resolution Dwt ; Nm2 = 2p f
2p / N 2p f

ABSOLUTE OPTICAL ENCODERS

(a)
MSB LSB

Reading: 0 1 15

1 14

(a) Binary code

3 Transparent 4 Opaque

13

12

11

5 6 7 8 9

10

(b) 1 2

15 14 13

12

11

5 6 9 7 8

10

(b) A gray code

Hall Effect Transducers

Note: Magnetic field pulses due to teeth Output voltage pulses

Actuators

Stepper Motors
Driven in fixed angular steps (increments) Each rotation step = rotor response to an input pulse (or a digital command) Three Basic Types: 1. Variable-reluctance (VR) stepper motors (have soft-iron rotors) 2. Permanent-magnet (PM) stepper motors (have magnetized rotors) 3. Hybrid stepper motors (have two stacks of rotor teeth forming the two poles of a permanent magnet located along the rotor axis).

Stepping Sequence (Half-stepping) for a Two-phase PM Stepper Motor

Start
f1 =1
N Pole S Pole S Pole

Step 1

f1 =1
N Pole N Pole

f2 = 0

S Pole

f2 = 1

Step 2

f1 = 0

Step 3

f1 = -1
S Pole

f2 = 1

S Pole

N Pole

f2 = 1

S Pole N Pole

N Pole

Step 4

f1 = -1
S Pole

f2 = 0

N Pole

Full-stepping Sequence for a Three-phase VR Stepper Motor

Start

f1 = 1 f2 = 0

Step 1

f1 = 0 f2 = 1

f3 = 0

f3 = 0

Step 2

f1 = 0 f2 = 0

Step 3

f1 = 1 f2 = 0

f3 = 1

f3 = 0

Hybrid Stepper Motor

Rotor Teeth N Phase 1 Phase 2

1 Pitch 4 Offset

Stator Teeth S

Stator Rotor Rotor Stack 2 Stack 1 (Magnetized N) (Magnetized S)

Rotor Stack 1

Stator Stator Rotor Segment Segment Stack 2 1 2 (Phase 2) (Phase 1) Surrounding Surrounding Stack 2 Stack 1

Rotor stack misalignment (1/4 pitch) in a hybrid stepper motor (Schematically shows the state where phase 1 is off and phase 2 is on with N polarity).

Microstepping

Equilibrium (detentent) Position Before the Microstep i Equilibrium (dentent) Position After the Microstep

dq

i+di

Stepper Motor Drive System

(a)

Controller/ Indexer

Driver

Motor

Response Feedback

(b) Driver Controller Position Pulse Train Power Supply

Computer/ Indexer

Translator

Amplifier

Stepper Motor

To Load

Direction Pulse Train

Current to Windings

Stepper Motor Selection Example

Speed v Objects (mL) Jg2 r Jd Jg1 Jm wm Stepping Motor Step-down Gear (p : 1) Conveyor Belt (mc) r Js

Speed v vmax

Distance d

Time t

To Determine Equivalent Inertia at Motor Rotor: Overall KE

w 1 1 1 rw 2 = ( J m + J g1 )wm + ( J g 2 + J d + J s ) p + ( mc + mL ) m 2 2 p p 2 1 1 r2 1 2 2 = [ J m + J g1 + 2 ( J g 2 + J d + J s ) + 2 ( mc + mL )] wm = J ewm 2 p p 2
2 2

J e = J m + J g1

1 r2 + 2 ( J g 2 + J d + J s ) + 2 ( mc + m L ) p p

Stepper Motor Selection Example (Cont d)

Data: d = 10 cm, T = 0.2 sec; r = 10 cm; mc = 5 kg; mL = 5 kg,; Jd = Js = 2.010-3 kg.m2 Two gear units: p = 2 and 3; Jg1 = 50 10-6 kg.m2 and Jg2 = 200 10-6 kg.m2 Overall system efficiency = 80% (for either gear unit)
1 d = v max T From triangular speed profile: 2 1 0.1 = vmax 0.2 => vmax = 1.0 m/s Substitute numerical values: 2 v 1.0 a = max = m/s 2 = 10.0 m/s 2 Max acceleration/deceleration: T / 2 0.2 / 2

pa a= Angular acceleration/deceleration of motor: r

With efficiency h:

pa 1 r2 pa hTm = J ea = J e = [ J m + J g1 + 2 ( J g 2 + J d + J s ) + 2 (mc + mL )] r r p p

Maximum speed of motor:

w max =

pv max r

v a hTm = [ J m + J d + J s + r 2 (mc + m L )] and w max = max Without gears: r r

Stepper Motor Selection Example (Cont d)

Stepper Motor Selection Example (Cont d)

Motor Performance Curves

Stepper Motor Selection Example (Cont d)

Case 1: Without Gears
-3 -3 2 For efficiency h = 0.8 (80%) => 0.8Tm = [ J m + 2 10 + 2 10 + 0.1 (5 + 5)] 0.1 N.m

10

=> Tm = 125.0[ J m + 0.104] N.m and w max = 0.1 rad/s = 10 2p rpm = 95.5 rpm Operating speed range = 0 to 95.5 rpm. Note: Torque at 95.5 rpm is < starting torque for first two motors (see speed-torque curves). In motor selection use the weakest point (i.e., lowest torque) in the operating speed range. Form the following table: Motor Model Available Torque at wmax (N.m) 0.26 0.60 2.58 7.41 Motor Rotor Inertia (kg.m2) 11.8 10-6 35.0 10-6 187.0 10-6 805.0 10-6 Required Torque (N.m) 13.0 13.0 13.0 13.1

1. 0

60

50 SM 101 SM 310 SM 1010 SM

Note: Without a gear unit, available motors cannot meet system requirements.

Stepper Motor Selection Example (Cont d)

Case 1: With Gears Note: System efficiency drops when a gear unit is added. For simplicity, neglect this. Efficiency of 80%, h = 0.8 =>
1 0.12 10 -6 -3 -3 0.8Tm = [ J m + 50 10 + 2 (200 10 + 2 10 + 2 10 ) + 2 (5 + 5)] p N.m p p 0.1 1.0 p 60 w max = rad/s = 10 p rpm and 0.1 2p
-6

or

Tm = 125.0[ J m + 50 10 - 6 +

1 104.2 10 -3 ] p N.m 2 p

and w max = 95.5 p rpm

For p = 2 we have wmax = 191.0 rpm. Form the following table: Motor Model 50 SM 101 SM 310 SM 1010 SM Available Torque at wmax (N.m) 0.25 0.58 2.63 7.41 Motor Rotor Required Inertia (kg.m2) Torque (N.m) -6 6.53 11.8 10 6.53 35.0 10-6 6.57 187.0 10-6 6.73 805.0 10-6

Note: For p = 2, motor model 1010 SM satisfies requirement. With full stepping, step angle = 1.8o. Corresponding step in conveyor motion = positioning resolution. 1.8o p -3 With p = 2 and r = 0.1 m, position resolution = 2 180o 0.1 = 1.57 10 m.

DC Motors
Current i Force F Field B Magnetic Field B (a) Field Supply vf

B 3 2

1 Torque Tm

N pole

S pole

N Pole Stator

Speed wm

2 S Pole = Current In = Current Out

1 i Current Through Conductor Commutation Plane

3 Rotor

(b) Split Ring Commutator 3 1 3 2 1 Brush Armature Supply va 2

2 1 1 3

2 3

Torque in DC Motors
(a)

(a)Torque generated in a planar rotor (One coil segment)

Torque T

With Commutation Without Commutation

2p

Commutation Point

Commutation Point

Commutation Point

(b)

T 1 2 3 Resultant

(b) Resultant torque from a rotor with three coil segments.

2 3

4 3

5 3

2p

Commutation Point

Electronic Commutation for a (Brushless) DC Motor

Rotor Poles

DC Power Supply

Stator Windings

S Pole

Power to Stator Segments

Microelectronic Drive Circuit (Stator Switching)

Timed Pulses (Switching Commands)

qm

Angular Position Sensor (Hall-Effect)

Pulse Generator

Speed Setting

Equivalent Circuit for a DC Motor

(a) + vf -

if

Rf

La

Ra

ia +

Lf

Vb

Tm
wm TL

va -

Stator (Field Circuit)

Rotor (Armature Circuit)

(b) Magnetic Torque Armature Load TL Load Shaft bm

Tm Jm Damping

DC Servomotor
(Optional) Torque Sensor Transmitted Torque T

Armature Current Reference Input

ia

Servo Controller

Drive Amplifier DC Motor Load

Tachometer or Encoder

wm (Speed)

qm (Position)

Encoder

DC Motor Drive/Control
Control Software Encoder

Interface (I/O) Board Control Computer

Drive Amplifier

DC Motor

PWM Drive System

Velocity Reference Velocity Amplifier Current Sensing Velocity Feedback Encoder Torque Amplifier PWM Amplifier DC Motor

To Brushless Motor

Stator Switching Electronics

Commutation Logic

DC Motor Selection
Useful Operating Region
Motor T orque T Peak T orque Curve

Motor Selection to Match a Load

Continuous Torque Curve Useful Operating Region

Torque T

AB = acceptable operating range A = optimal operating point Continuous Torque Curve Load Curve A Operating Region of Motor + Drive Unit

Tmin wmin Motor Speed wm

Tmin w min

B Speed w

DC Motor Data
1. Mechanical data Peak torque (e.g., 65 N.m) Continuous torque at zero speed or continuous stall torque (e.g., 25 N.m) Frictional torque (e.g., 0.4 N.m) Maximum acceleration at peak torque (e.g., 33 103 rad/s2 ) Maximum speed or no-load speed (e.g., 3,000 r.p.m.) Rated speed or speed at rated load (e.g., 2,400 r.p.m.) Rated output power (e.g., 5100 W) Rotor moment of inertia (e.g., 0.002 kg.m ) Dimensions and weight (e.g., 14 cm diameter, 30 cm length, 20 kg) Allowable axial load or thrust (e.g., 230 N) Allowable radial load (e.g., 700 N) Mechanical (viscous) damping constant (e.g., 0.12 N.m/krpm) Mechanical time constant (e.g., 10 ms)
2

2. Electrical data Electrical time constant (e.g., 2 ms) Torque constant (e.g., 0.9 N.m/A for peak current or 1.2 N.m/A rms current) Back emf constant (e.g., 0.95 V/rad/s for peak voltage) Armature/field resistance and inductance (e.g., 1.0 W, 2 mH) Compatible drive unit data (voltage, current, etc.) 3. General data Brush life and moto r life (e.g., 5 108 revolutions at maximum speed) Operating temperature and other environmental conditions (e.g., 0 to 40 C) Thermal resistance (e.g., 1.5 C/W) Thermal time constant (e.g., 70 minutes) Mounting configuration

Drive Amplifier and Power Supply Selection Objective: Determine Ratings (Current, Voltage Rating, Power) for PWM Amplifier; Select a Power Supply at These Ratings Method: Required motor torque Tm = J ma + TL + T f a = highest required angular acc.; TL = worst-case load torque; Tf = frictional torque on motor Note: For a pure inertia load (JL) Tm = ( J m + J L )a + T f Required current i =
Tm km

km = torque constant of the motor ' Required voltage v = k mw m + Ri k 'm = k m = back e.m.f. constant; R = winding resistance w m = highest operating speed of motor Voltage Rating = [Required Voltage]/[Max Duty Cycle of PWM Amp] Note 1: Pick amplifier and power supply with these ratings (voltage, current, power) Note 2: If several amplifiers use the same single power supply: Increase the power rating of power supply in proportion

Duty Cycle of PWM Amplifier

Duty cycle =

Average output 100% Peak output

Voltage T
To

vref

Time t On Off On

Hydraulic Control System

hm hh (i, v) (T , w ) (Q, P)

Main Components of a Hydraulic Control System 1. Servo-valve 2. Hydraulic actuator 3. Load 4. Feedback control elements Valve: q = kq u - kc p Hydraulic Actuator:

q=A

dy V dp + dt 2 b dt

d2y dy m 2 +b = Ap - f L Load: dt dt

Hydraulic Spool Valve

(a)

P = P2 - P 1
To Hydraulic Actuator

From Actuator

(Q2 , P2 )
Control Input (Displacement)

(Q1 , P ) 1
Spool

Electrical Input (i, v) Torque Motor or Proportional Solenoid (Valve Actuator)

Land Port Pressure @ 0

Ps
Supply (from Pump)

Discharge (to Reservoir)

(b) Land

(c) Land

Port

Port

(a) A four-way spool valve; (b) An overlapped land; (c) An underlapped land.

Hydraulic Actuator

( P2 , Q2 )

Fluid Flow Ports

( P , Q1 ) 1

Equivalent Inertia of the Actuator

Area A m Load FL Damping b

V2
Cylinder Piston

V1

Advantages and Disadvantages of Hydraulic Systems

Advantages Over Electrical Motor Systems: High pressures (e.g., 5,000 psi) Can provide very high forces (torques) at high power levels simultaneously to several actuating locations (flexible). Quite stiff when viewed from load side (because a hydraulic medium is mechanically stiffer than an electromagnetic medium). Heat generated at the load can be quickly transferred to another location away from the load, by the hydraulic fluid itself. Self-lubricating Low friction in valves, cylinders, pumps, hydraulic motors, etc. Safety considerations will be less (e.g., no possibility of spark generations unlike motors with brush mechanisms). Disadvantages: More nonlinear Noisier Synchronization of multi-actuator operations may be more difficult More expensive and less portable

Mechatronic Design

Mechatronic Design Quotient (MDQ)

Task Requirements ControlFriendliness Matching and Optimization

MDQ
Power and Efficiency Production and Reliability Intelligence

Mechatronic Design
Use of MDQ

Concurrent Mechatronic Design

Evaluation Index (MDQ)

Optimization (maximize MDQ)

Some Considerations
Mechatronic design: concurrent and multi criteria design, has many benefits For criteria aggregation 1. interaction between criteria, 2. human experience should be considered. (e.g., use fuzzy integrals) MDQ represents the degree of satisfaction of mechatronic design criteria MDQ can be used in optimization and automation of concurrent design MDQ is useful as a design evaluation index MDQ may be used as fitness function in GP Bond graph method is appropriate for modeling mechatronic systems (Multi-domain mixed systems) Considerations of topology/structure optimization are important

Mechatronic Design
Component Sizing MDQ Optimization

System Topology Design

Conceptual Example
Electrical System
i+ v +

Mechanical System

Electrical Dynamics

Ideal Energy Transformer

Mechanical Dynamics

Energy Dissipation

Energy Dissipation

Conceptual Example (Cont d)

MDQ =

2 e e 2 e se

I + I +

2 m m 2 m sm

I I

Ise = Best Electrical Design Index in sequential design Ism= Best Mechanical Design Index in sequential design Ie = Electrical Design Index in the coupled (concurrent) design Im = Mechanical Design Index in the coupled (concurrent) design

Note: The I terms incorporate various design attributes

Example: Motor Selection for a Robot

Data Base MDQ Attribute 1: MDQ Attribute 2: MDQ Attribute 3: Speed Weight Cost 245-6089: Solution 1 80 55 39.30 245-6095: Solution 2 40 57 39.30 Stock No. 245-6102: Solution 3 18 58 39.30

MDQ INDEX VALUES Speed Weight Cost MDQ Index Index Index Value Solution 1 80 % 70% 80% 78% Solution 2 40 % 80% 80% 60% Solution 3 20 % 85% 80% 51% Weighting Parameter 50% 20% 30% Design Index
Note: Sum of weighting values satisfies

W
i =1

= 100%
n

n = total number of MDQ attributes ( n=3 in this example)

MDQ value Mj for solution j, is computed according to: M j =

W I
i =1

i ij

; for j=1, 2

m, 0< Mj <1

m = total number of feasible solutions

Largest value of Mj optimal design = Solution 1 in this example

Evolutionary Computing
It is a broad class of optimization methods (discrete and continuous) Inspired by biological (natural) evolution Uses derivative-free and population based search techniques Starts from multiple searching points (solution candidates)

Advantages:
Can handle complex and ill-modeled problems Global minima can be determined Multiple objective functions can be handled

Possible Approach:
Use MDQ as the Fitness Function Use multiple MDQ functions for complex multi-criteria optimization

Evolutionary Computing
(for MDQ Optimization)

Evolutionary Computing
Nomenclature Individual of a population is a chromosome (a sequence of genes). The position of a gene in a chromosome is called locus. The values a gene can take are called alleles. Initial population is created randomly. Solution evolves using evolutionary operations (selection, recombination or crossover, and mutation). Every chromosome is evaluated and given a fitness value. Most successful chromosomes are selected for crossover to produce offspring. Offspring would have a better fitness values. Mutation adds diversity to potential solutions. Components of an Evolutionary Algorithm 1. Encoding: representation of potential solutions (binary, floating-point, or gray encoding). 2. Initialization: creation of initial population. 3. Fitness Function: assigning fitness values to chromosomes. 4. Evolutionary Operators: selection, crossover, mutation, etc. 5. Working Parameters: population size, chromosome length, etc.

Genetic Algorithms
Problem: Maximize (or, minimize f) m = number of optimization parameters Steps Choose a random population of n individuals within range of optimization. Using binary encoding, represent each parameter as a string of q binary digits. => population represented by nqm matrix Apply genetic operators to the matrix => new population with larger f value Nomenclature Individual: potential solution Genotype: genetic traits of an individual = encoded potential solution = chromosome Gene: single bit of genetic information = one element of a chromosome string Alleles: possible values of a gene (Note: a binary gene has two alleles) Fitness Function Represents the value or goodness or fitness of a solution. Same as objective function. Input: genotype of parameters; Output: a real value representing goodness of solution. Fitness is comparative not absolute (serves to differentiate solutions) May not be analytic (e.g., physical performance measurements; heuristics) Note: Multiple objectives may be needed. Example: A Manufacturing Process Genotype: operating conditions, control parameters Objective Function: Productivity or Efficiency or Product Quality or Cost Effectiveness

Genetic Algorithm Operations

Selection (Selection of fit individuals for reproduction) Crossover (Mating of selected individuals for reproduction) Mutation (Introduction of new alleles into chromosomes in the population, to create completely new solutions)

Operations of an Evolutionary Algorithm

New population is produced by mating the best individuals Over generations, desirable characteristics are spread throughout population Mutation is used to escape from a local minimum.

Initialization of Individuals

Evaluation

Selection of Parents Crossover Mutation

Selection

Elitist Model: Top 10 to 20 individuals of the population are selected. Ranking Model: Fitness of each member of the population is ranked. Those above a specified fitness level are selected. Roulette Wheel Procedure: Probabilities pi are assigned to individuals i according to their fitness, cumulative probabilities ci = p j are computed, and the first individual with
j =1 i

ci random number r is selected. Repeat until a sufficient number of individuals selected.

Crossover
Exchanges some genes of the two parents to create the genotypes of the offspring.

Method: Select points along parents chromosomes (randomly) and exchange genes between these points. Note: In Simple Crossover , only one point is chosen. See Figure.

Mutation
Introduces completely new alleles into a population of chromosomes. Creates completely new solutions (avoids stagnation). Method: Select one or more genes in an individual at random and change their alleles. Note: Allele change itself can be random or deterministic fashion.

Mutation of a Single Gene.

Important Steps of Genetic Algorithm

Note: Roulette-wheel, cumulative probability method is used for selection, crossover, and mutation.

Important Steps of Genetic Algorithm (Cont d)

Function Optimization Example

Use GA to find the maximum of Note: Answer is x = 0.5 The range of optimization = [0, 1.5]; Required precision = 3 decimal points Required number of bits (genes) r in a chromosome is given by:
(1.5 - 0) 103 2r - 1 r = 11

Crossover probability = 0.25; Mutation probability = 0.1 Note: Use roulette wheel (cumulative probability) method for selection, crossover, and mutation.

Note: Fitness = f value; Selection probability = individual fitness/total fitness

Example (Cont d)
Population After Selection Operation (Using Roulette-Wheel Method)

Note: Some individuals are repeatedly selected (X1 and X2; X4, X7, and X9; X5, X6, and X8) by this probabilistic operation.

Example (Cont d)
Population After Crossover and Mutation Operations

Note: Roulette wheel method used to determine the crossover chromosomes pair and mutation chromosome. Note: Gene mutation is done randomly.

Example (Cont d) Population After 16 Generations

Note: Solution is chromosome X7.

Example(Cont d)

Genetic Algorithms in Mechatronic Design

Problem: Maximize MDQ(p1, p2, pm) where, pi = ith design attribute Individual Chromosome: possible design Individual Gene: element of information in a design (e.g., system component, connection structure, set of parameters) = one element of a chromosome string Alleles: possible values of a gene (e.g., available choices for a particular component of the system) Fitness Function: Represents the value or goodness or fitness of a design = MDQ (computable for a given design once the element information of the design is known)

Further Study
Try problems at the end of chapters 2-9. Read case studies in chapter 13 Book: De Silva, C.W., MECHATRONICS An Integrated Approach, CRC Press, Taylor & Francis, Boca Raton, FL, 2005. (ISBN 0-8493-1274-4): www.crcpress.com Also:
De Silva, C.W., MODELING AND CONTROL of Engineering Systems, CRC Press, Taylor&Francis, Boca Raton, FL, 2009. Karray, F. and de Silva, C.W., Soft Computing and Intelligent Systems Design, Addison Wesley, New York, 2004.

Thank you!
www.mech.ubc.ca/~ial