INSTRUMENTATION AND CONTROL

TUTORIAL 4 – SYSTEM RESPONSE

This tutorial is of interest to any student studying control systems and in particular the EC
module D227 – Control System Engineering.

On completion of this tutorial, you should be able to do the following.

 Explain the output response of an ON/OFF control system.

 Explain and define the standard models for 1st and 2nd order systems.

 Explain and define the standard time dependant inputs to a system.

 Define a linear system and a linear time invariant system.

 Explain and calculate the response of a standard 1st order system to a step change in
the input.

 Explain and calculate the response of a standard 1st order system to a ramp
(velocity) change in the input.

 Explain and calculate the response of a standard 1st order system to a sinusoidal
change in the input.

 Explain how to find the overall gain of a system.

 Use partial fractions to solve responses.

If you are not familiar with instrumentation used in control engineering, you should
complete the tutorials on Instrumentation Systems.

In order to complete the theoretical part of this tutorial, you must be familiar with basic
mechanical and electrical science.

You must also be familiar with the use of transfer functions and the Laplace Transform
(see maths tutorials).

You must be able to convert polynomials into partial fractions.

 D.J.DUNN 1
1. INTRODUCTION

A system may be defined as a set of connected things. We are concerned with engineering systems and
we may define this more precisely as a set of connected elements designed to produce specified outputs
when a given input is applied. System may include analogue elements or digital elements. First we will
examine a special kind of simple system that uses ON/OFF control.

2. ON/OFF CONTROL

Consider a system made up of a tank of liquid, a heater

and a switch.

When the switch is operated, power is supplied to the

heater and the water gets hotter. The operation of the
switch is the input action and the temperature of the water
is the output action. The output temperature is related to
the power by the following formula.

Mass x Specific Heat x Temperature Change mcθ

POWER   Figure 1
Time t

We may represent our system as a simple block.

Figure 2

Often it is possible to represent the ratio of output/Input as a simple constant and give this constant the
symbol G (it might be thought of as gain but this would not be strictly correct). In this example we have
the following.
Output θ t
G  
Input P mc

The system described is known as an OPEN LOOP SYSTEM because the signal path does not form a
closed loop and there is no control over the temperature. When the switch is closed, the temperature will
keep rising with time.

Now consider an improvement to the system by using a thermostatic controller and a thermocouple to
measure the output. The desired temperature i (input) is set on the controller and the actual temperature
o is the output of the system. The controller
compares the measured temperature with the
set temperature. If the water is too cold, the
heater is turned on. If the temperature is too
hot it is switched off.

Suppose the temperature is low and correct

and the set temperature is suddenly increased.
This is called a step change. The thermostat
will turn the heater on and the temperature
will rise. Sometime later, the temperature
reaches the correct value. Figure 3

Ideally, the thermostat would switch off the heater at the precise moment that the correct temperature is
reached.

 D.J.DUNN 2
In reality this cannot happen as heat will go on being put out after it is switched off and the thermostat has
to detect that the temperature is too hot before it switches off. Consequently, the temperature rises above
the correct value before being switched off and then the liquid starts to cool. It will cool to just below the
correct value before the thermostat detects the error and switches the heater back on. The liquid is heated
again and the process continues indefinitely producing a temperature – time graph as shown. This graph is
called the SYSTEM TIME RESPONSE GRAPH.

Figure 4

The block diagram of this system shows that the thermocouple turns the signal path into a closed loop so
it is a CLOSED LOOP SYSTEM.

Figure 5

The problem with all ON/OFF control is that the output must cycle between an OFF value and an ON
value which are just above and below the SET value. If you bring the ON and OFF levels closer together
the system must switch on and off more quickly. It is impossible to get precise control with this method.
This is adequate for many systems such as central heating where the exact temperature is not important.
Systems that need precision control (such as the movement of a machine tool) require a more
sophisticated method of regulation and these are what we must study in detail.

3. RESPONSE OF CONTINUOUS CONTROL SYSTEMS

We are going to discuss how to set about solving the way a given system behaves with respect to time. It
should be noted, however, that time is not always the main variable but we will assume it is. We could,
for example, be discussing how a robot manipulator moves in response to an input or how a valve on a
pipe line moves in response to an input.

3.1 STANDARD MODELS

The first step is to derive a mathematical model along the lines shown in the previous tutorials. In those
tutorials it was shown that the following models applied to several different systems that were analogues
of each other. These are the standard first order and second order equations.

Standard 1st order equation. i = T do/dt + o ......................(1)

Standard 2nd order equation. i = T2 d2o/dt2 + 2 T do/dt + o ..................(2)

T is a time constant and  is a damping ratio. These will be discussed in greater detail later. T and  are
examples of system parameters or constants.

 D.J.DUNN 3
Before you can solve how the output changes with time, you have to decide how the input changes with
time. For example, you might make a sudden change to the input or you might be changing it over a
period of time. Let’s look at this next.

3.2 STANDARD INPUTS

Standard inputs are usually listed in the following order.

i. AN IMPULSE

This is an instantaneous change in i lasting for zero length of time and returning
to the initial value. This is mostly applied to digital systems where instantaneous
values are sampled by digital to analogue converters. It is also widely used as a
standard input to a system to compare the responses of different systems.

Figure 6

ii. A STEP CHANGE

This is an instantaneous change in the input which then remains at the

new value.

i = H at all values of time after t= 0.

H is the change or height of the step.

Figure 7

iii. A RAMP OR VELOCITY CHANGE

This is when the input changes at a constant rate. It is also called a

velocity input.
di/dt = c or i = ct

c is the rate of change (velocity).

Figure 8

iv. A PARABOLIC or ACCELERATION CHANGE

This is when d2i/dt2 = a or i = at2/2

This is also known an acceleration since the rate of rate of change is a

constant a.

Figure 9

 D.J.DUNN 4
v. A SINUSOIDAL CHANGE

In this case the input changes with time sinusoidally.

i = A sin(t + )

A is the amplitude and  is the phase angle.

Figure 10

vi. AN EXPONENTIAL CHANGE

This is when the input changes exponentially with time.

i = A(1 - e-at)

There are several exponential forms e.g. representing growth and

decay.

Figure 11

Once an equation has been formed to represent how the input to a system changes with time, the main
task is to deduce how the output change with time so that you can see how accurate and how fast the
response will be. It is this aspect which we will discuss in detail and in particular the use of LAPLACE
TRANSFORMS to help us do it. Before this, however, let's look at some more definitions and
terminology used to describe types of systems.

3.3 LINEAR SYSTEM

A property of a linear system is that of SUPERPOSITION. This means that if one input produces a given
output and another input produces a second output, then both inputs together will produce the sum of the
two outputs. A good example of this is the deflection of a spring. If a force F1 is applied the resulting
extension is x1. If a force F2 is applied the resulting extension is x2. If a force F1 + F2 is applied, the
resulting extension is x1+ x2.

Another property of a linear system is that of HOMOGENEITY. This means that if an input produces a
given output, then multiplying the input will multiply the output by the same figure. In the case of the
spring if a force F produces deflection x then a force of nF produces a deflection nx.

3.4 LINEAR TIME-INVARIANT SYSTEMS

Consider the mathematical model for a differential pressure flow meter. Q = C (p)½. C is a constant for
the meter and the power of ½ is a parameter, both of which define the equation. Q and p are variables. If
the constants and parameters do not change over the period of time being studied, the system is a linear
time-invariant system. T and  in equations (1) and (2) are examples of constants for such systems.

For the moment we need to study how such equations are solved rather than how they are derived.

 D.J.DUNN 5
3.5. LAPLACE TRANSFORMS

In order to solve o as a function of time we use a mathematical transformation known as LAPLACE

TRANSFORMS. This changes the equation from a differential equation into an algebraic equation which
can then be manipulated and converted back into a differential equation by inverse transforms. You
should study the tutorial on this in the maths section unless you are already familiar with it.

First we replace ‘d/dt’ with the symbol ‘s’ so for example do/dt becomes so and d2o/dt2 becomes
s2o.

You may go on to appreciate that o on its own becomes soo and that the integral of o becomes s-1o
and so on.

Conducting this change on equations (1) and (2) produces the algebraic equations as follows.

i = T so + o
i = o(T s + 1)....................(3)
and
i = T2 s2o + 2 T so + o
i = o( T2 s2+ 2 T s + 1)..................(4)

It is now possible to write the equations as a ratio and when this is done, the equation or model is called a
TRANSFER FUNCTION. Equations 3 and 4 become

θo 1 θo 1
 and  2 2
θ i Ts  1 θ i T s  2  Ts  1

The transfer functions are now functions of s f(s) instead of functions of time f(t).

Before solution of the transfer function can proceed, we must change the input i from a function of time
f(t) into a function of s f(s). This is where the main transformation occurs. You do not have to transform
the function mathematically but instead you will use standard tables. Here are two examples of how it is
done (again you will find details in the maths section)

WORKED EXAMPLE No.1

Find the Laplace transform L H when H is a constant (a step change)

SOLUTION
   st  
  e st 
L H   0 e f  t  dt  0 H dt  H 
- st
e 
 s 0
 0  1
f(s)  H 
H
 
 s s  s
H
L H   f(s) 
s
For a unit step H = 1 and the Laplace transform is 1/s

 D.J.DUNN 6
 WORKED EXAMPLE No.2.
Find the Laplace transform of e-at

SOLUTION
  0f  t  dt
L e -at   0

e st e -at dt 

0 e (s  a)t dt

Le - at
 

 e  (s  a)t  
 s  a 0

 
L e - at   
0

1 
s  a 
s  a
 
L e - at  f(s) 
s 
1
a

 D.J.DUNN 7
TABLE OF COMMON LAPLACE TRANSFORMS

Note that the use of the letters for constants is arbitrary and that often the solution is found by
interchanging ‘a’ with ‘1/T’ and ‘’. H and k are arbitrary constants.

Time domain f(t) Frequency domain f(s) Description

1 e-at f(t) dt f(s + a)

2 t 1 Unit Impulse
3 H H Step H
s
4 ct c Ramp
s2
5 H(t - T) e sT Delayed Step
H
s
6 1 - e sT Rectangular pulse
s
7 k e-at k Exponential
s  a
8 kt e-at k
(s  a) 2
9 K(e-at - e-bt) k(b  a)
(s  a)(s  b)
9 k sin (t) k Sinusoidal
s  2
2

10 k  t sin(t) 2kω 2s
(s 2  ω 2 ) 2
11 k cos (t) ks Co sinusoidal
s  ω2
2

12 k e-at sin (t) kω Damped sinusoidal

(s  a) 2  ω 2
13 k e-at cos (t) s  a Damped co sinusoidal
(s  a) 2  ω 2
14   
t ka
a = 1/T
Exponential growth
k 1  e  T
s(s  a)
 
15 
   
t
 ka
k t  1  e T  s (s  a)
2

  

16 k(1-cost) kω 2
s(s 2  ω 2 )
17 k sin (t + ) kω cos φ  s sinφ
s 2  ω2
18 a
1  e  at (cos bt  sin bt) a 2  b2 Standard second order
b 
s s  a 2  b 2  system step response

 D.J.DUNN 8
4. THE STANDARD 1ST ORDER SYSTEM

4.1 RESPONSE TO AN IMPULSE AND STEP

Any system with an impulse applied to the input with have an input of i(s) = 1. It follows that the output
in the time domain will simply be the inverse Laplace transform of the transfer function.

WORKED EXAMPLE No.3

θo 1
Compare the time response of a system with the transfer function  to a unit impulse and a
θ i Ts  1
unit step input.

SOLUTION

UNIT IMPULSE
1θ i
θ o (s)  substitute i(s) = 1 for an impulse input.
Ts  1
1(1)
θ o (t)  inverse Laplace transform of
Ts  1
1/T
rearrange into a recognisab le transform and θ o (t)  (1/T)e  t/T
s  1/T
UNIT STEP
1θ i
θ o (s)  substitute i(s) = 1/s for an step input.
Ts  1
1
θ o (t)  inverse Laplace transform of
s(Ts  1)
1/T
rearrange into a recognisab le transform and θ o (t)  1  e  t/T
s(s  1/T)
The diagram shows the two responses.

Figure 12

 D.J.DUNN 9
 WORKED EXAMPLE No.4

A R- C circuit is shown in which R = 200  and C = 15 F. The voltage Vi is suddenly changed from
0 to 10 Volts. Determine the time constant and how long it takes Vo to reach 9.99V.

Figure 13
SOLUTION

The model for the R – C circuit shown was derived in tutorial 1 and shown to be Vi = TdVo/dt + Vo
First replace dVo/dt by s Vo Vi = T s Vo+ Vo
Next rearrange into a transfer function G(s) = Vo/Vi= 1/(Ts + 1)
This may be represented diagrammatically as shown.

Figure 14

Next change the input from a function of time into a function of s by making a Laplace
transformation. For a step input Vi(t) = H the transform is Vi(s) =H/s. In this case H = 10 Volts.
Vi H
Vo  
Ts  1 sTs  1
Next manipulate the equation into a recognisable Laplace transform as follows.
H/T
Vo 
ss  1/T 

Looking in the table of transforms we see Ka/s(s+a) is the transform of K(1 - e-at). This means that if
we put a = 1/T and K = H the solution for Vo is Vo = H(1 - e-t/T)

The time constant T = RC = 200 x 15 x 10-6 = 0.003 seconds

Put Vo = 9.99 V and H = 10 V

 
t
  
t

9.99  101  e 0.003  0.999  1  e 0.003 
   
t
t

e 0.003
 0.001 ln 0.001   - 6.9077 t  0.0207 s
0.003

 D.J.DUNN 10
4.1.1 FULL ANALYSIS OF THE RESPONSE

For any standard first order system with a step input, the output response is o = H(1 – e-t/T). If we plot
both input and output against time we get the result shown.

Figure 15

4.1.2 FIRST ORDER TIME CONSTANT

The time constant ‘T’ has a real meaning. Consider the response graph. The rate of change of o is
dθ o  1 t
 H e T
dt T
 H e  0 
dθ o 1 H
At the start of the change t = 0, so the initial rate of change is
dt T T
The initial gradient (rate of change) is H/T. If the initial gradient is projected as shown, it will intercept i
at time t = T. This leads to a definition of T as the time it would take the output to match the input if it
kept changing at the initial rate.

Another way of defining the time constant T is as follows.

At time t = T the value of the output is o= H{1 - e-1} = 0.632H

In other words, the output will reach 63.2% of the desired value after T seconds.

When t = 4T the output is o = H(1 - e-t/T) = H(1 - e-4) = 0.999H. This is as close to the correct value as
we are likely to want so we normally assume that the system has reached the correct value after t = 4T
seconds.

 D.J.DUNN 11
4.2 RESPONSE OF A STANDARD 1ST ORDER SYSTEM TO A RAMP INPUT

A ramp or velocity input occurs when i (t) = ct and i (s)= c/s2

θo 1
(s) 
θi Ts  1
θ i (s) c c(1/T)
θ o (s)    2
Ts  1 s (Ts  1)
2
s (s  1/T)
If we replace ‘c’ with ‘k’ and ‘1/T’ with ‘a’ we have
ca
θ o (s)  2 and we can find this in the table of
s (s  a)
transforms. The inverse transform gives the solution
o (t) = c {t - T(1 - e-t/T)}

Figure 16
Plotting the output o against time produces the result shown. The equation may be written as:
o (t) = c t – cT(1 - e-t/T)}
At large values of time t the term ( e-t/T) becomes negligibly small and the output becomes:
o (t) = c {t - T(1)} = c(t - T)

The error becomes e (t) = i -o = ct – ct + cT e (t) = cT

In other words, a constant error cT result after an initial transient stage and this is the steady state error.

WORKED EXAMPLE No.5

A position control system has a transfer function G(s) = 1/(0.2 s + 1). The input is changed at a
constant rate of 5 degrees/s from the zero position. Calculate the error after 0.4 seconds and the
steady state error.

SOLUTION

Comparing parameters it is apparent that T = 0.2 and c = 5 deg/s.

After 0.4 seconds i = 5 x 0.4 = 2 degree.

o = c t – cT(1 - e-t/T)} = 5 x 0.4 – 5 x 0.2 x (1 – e-0.4/0.2) = 2 – 0.835 = 1.135 degrees.

e = 2-1.135 = 0.865 degrees

The steady state error is cT where T = 0.2 and c = 5 degrees/s. e = 5 x 0.2 = 1 degree.

 D.J.DUNN 12
4.3 RESPONSE OF A STANDARD 1ST ORDER SYSTEM TO A SINUSOIDAL INPUT

θo k
The standard first order system transfer function is (s) 
θi Ts  1
θo k
In the steady state s = 0 and (s)   k so it follows that k is the steady state system gain.
θi 1
Note that it must be +1 on the bottom line to make this so. T is a time constant.

For a sinusoidal input let i(t) = A sin(t)

ω
The Laplace transform is θ i (s)  2
ω  s2
θi k ω
The Output is θ o (s)  
(Ts  1) (Ts  1) (ω  s 2 )
2

The inverse transform gives the output as a function of time.

Plotting the input and output gives a typical result like this with A=1 and K=1).

Figure 17

The plot shows that the output quickly settles down into steady state response with a sinusoidal form but
with different amplitude and lagging the input with a phase shift. The transient stage only lasts for one
cycle.

A more detailed analysis of the steady state response is covered in the next tutorial.

 D.J.DUNN 13
 SELF ASSESSMENT EXERCISE No.1

1. Show the derivation of the transfer function for spring and damper system shown. Given that the
damping coefficient kd is 0.03 and the spring stiffness k is 4 kN/m, determine the time constant for
the system. (Answer 7.5 s)
If a force of F = 100 N is suddenly applied, calculate the value of x after T seconds. (Answer 16 mm)

Figure 18

2. A block of metal has a mass of 0.5 kg, specific heat capacity 346 J/kg K and temperature of
1 = 20oC. It is dropped into a large tank of oil at 2 = 120oC and it is found that the temperature of
the block takes 6 minutes to reach 119 oC.
θ 1
Assume that the temperature of the block is changes by the law 1 (s) 
θ2 (Ts  1)
Show that the temperature of the block changes with time by the law  = 1 + (2 - 1)(1-e-t/T)

Determine the time constant T and hence the thermal resistance between the block and the oil. (You
should see the tutorial 1 for details)

(Answer R=0.452 K/W)

3. A hydraulic motor has a nominal displacement of k 1 m3/radian. The speed  is controlled by a

simple valve such that the pressure to the motor is k 2x where x is the input position of the valve.

The motor has a moment of inertia J kg m2 and a damping coefficient of k3 Nm s/radian.

Given that the torque developed by the motor is k 1p, show that the open loop transfer function
ω 1
relating output  to input x is given by 
kmx Ts  1
J kk
T  and km  1 2
k3 k3

The input is given a step change. Sketch the response of the output. Determine the % change in the
output at t = T and t = 4T. (63.2% and 99.9%)

Show on the sketch the affect of increasing the moment of inertia.

4. A position control system has a transfer function G(s) = 1/(3 s + 1). The input is changed at a constant
rate of 4 mm/s from the zero position. Calculate the error after 2 seconds and the steady state error.
(2.161 mm and 12 mm)

 D.J.DUNN 14
4.3 GAIN OF 1st ORDER SYSTEMS

DC GAIN :- this is an electrical term and is the gain when there is no variation of the input in the time
domain. When an input is constant all the derivative values of the transfer function are zero so the D.C. is
the magnitude of G(s) when s = 0.
θ 1
Hence the gain of the standard 1st order closed loop transfer function o (s)  is unity.
θi Ts  1
θo K
When the transfer function is of the form (s)  , the D.C. gain is K.
θi Ts  1
θo A
The D.C. gain is not always obvious at first glance e derivation. Consider the case (s) 
θi Ts  B
θ A/B
Put s = 0 and the D.C. gain is A/B. We could rearrange the transfer function to o (s) 
θi (T/B)s  1
θo K
This is now in the form (s) 
θi T2s  1
T2 is a new time constant resulting from the change. In the steady state s = 0 so K is the steady state gain
of the system.

D.C. GAIN = A/B T2 = T/B

The D.C. gain also affects the magnitude of a dynamic response

The response to a step change H will hence settle down at KH. The response to a ramp input will not
settle down at a fixed value but o (t) = Kc {t - T(1 - e-t/T)}. The affect of introducing gain is shown on
the response diagram.

Figure 19

In the case of the step input i = H, the output o reaches a value of K H.

In the case of the ramp input i = ct, the output passes and exceeds the input more and more as time
progresses.

 D.J.DUNN 15
 WORKED EXAMPLE No.6
θo 10
The transfer function for a simple control system is (s) 
θi 0.5s  0.2
Solve and tabulate the values of o against time for a step input of 2 units.

SOLUTION
θo 10/0.2 50
Rearrange the transfer function. (s)  
θi (0.5/0.2)s  1 2.5s  1
From this we see the gain is 50 and the time constant is 2.5. For a step input of 2 units the transform
50(2)
gives i (s) = 2/s Hence θ o (s) 
s(2.5s  1)
The inverse transform is o= 100(1 - e -t/2.5) = 100(1 - e -0.4t)
Plotting the result we get the response shown below.

Figure 20

WORKED EXAMPLE No.7

The same system is subjected to a ramp input of 1 mm/s. Plot the output against time and determine
the steady state error.

SOLUTION
50
θ o (s)  For a ramp input i = ct and c = 1 mm/s. i(s) = c/s2 hence
(2.5s  1)
50c 
  -
t 

θ o (s)  2 
The inverse transform of this is θ o (s)  50ct - 2.51 - e 
s (2.5s  1) 
2.5

  

Figure 21

 D.J.DUNN 16
 SELF ASSESSMENT EXERCISE NO.2

1. Find the D.C. gain and time constant for the following transfer functions.

2
i. G(s)  (4 and 0.4)
0.2s  0.5

0.2
ii. G(s)  (2 and 0.5)
0.05s  0.1
2
iii. G ( s)  (2 and 3)
3s  1
16
iv. G( s)  (4 and 2)
8s  4

2. The output speed of a motor ( rad/s) is related to the angle of the input sensor ( radian) by the
ω km
transfer function (s) 
θ Tm s  2
-1
Where km = 15 s and Tm = 4 s

Determine the D.C. gain and time constant of the system. (7.5 s-1 and 2 seconds)

 D.J.DUNN 17
5. USING PARTIAL FRACTIONS TO SOLVE RESPONSES

The response to most types of inputs can be solved by the use of partial fractions. Here are two examples
using the standard first order transfer function.
θo 1 1/T
First Order System  G(s)  
θi Ts  1 s  1/T
θo p
Put 1/T = p and (s) 
θi s  p
5.1 UNIT STEP INPUT
p
A unit step input in Laplace form is i(s) = 1/s θ o (s) 
ss  p 
A A1
Change into partial fractions θ o (s)  0 
s s  p
Solve A0 and A1 in the normal way for partial fractions.
A0 p A1
 
s s  p ss  p 
A 0 s  p   A1s p

ss  p  ss  p 
The numerators must equate so A 0 s  p  A1s  p and this is true for any value of s
Put s = -p and A 0 0  A1p  p so A1 = -1
Now put s = 0 A 0 p  0  p so A0 = +1
1 1
Substitute back into the solution θ o (s)  
s (s  p)
Now conduct an inverse Laplace transform. 
θ o (t)  1  e  tp  1  e  tp 
Put p = 1/T θ o (t)  1  e  t/T
And this is the solution found by other methods in previous tutorials.

5.2 UNIT RAMP

θo p
(s) 
θi s  p
B B B2 p
Partial fractions θ o (s)  0  21   2
s s s  p s s  p 
ss  p B0  s  p B1  s 2 B2 p
θ o (s)  
s s  p 
2
s s  p 
2

ss  p B0  s  p B1  s 2 B 2  p

Equate the numerator and first let p = -s s0B0  0B1  p 2 B 2  p

1
put s  - p and B 2 
p
put s  0 and B1  1
Next put s = 1 and substitute for B1 and B2 to solve B0 and find B0 = -1/p
1 1
θ o (t)  -  t  e - pt Let 1/T = p and the solution is
p p
-
t
 - 
t
θ o (t)  -T  t  Te T  θ o (t)  t  T1  e T 
 
This is the solution found by other methods previously.
 D.J.DUNN 18
 θo K
 G(s) 
Let’s extend this to a 3rd order system or higher.
θi 
s  p1 s  p 2 s  p 3   
K
UNIT STEP i(s) = 1/s θ o (s) 
 
s s  p1 s  p 2 s  p 3  
A0 A1 A2 A3
Change into partial fractions θ o (s)    
s s  p1 s  p2 s  p3
Inverse Laplace Transform θ o (s)  A 0  A1e-pt  A 2e-pt  A3e-pt
A0, A1, A2 and A3 etc must then be solved.

K
UNIT RAMP i(s) = 1/s2 θ o (s) 
 
s s  p1 s  p 2 s  p 3
2
 
Change into partial fractions
B0 B1 B2 B3 B4
θ o (s)     
s s2 s  p1 s  p2 s  p3
Inverse Laplace Transform θ o (t)  B0  tB1  B2e-p1 t  B3e-p 2 t  B4e-p 3 t

B0, B1, B2 and B3 etc must then be solved.

WORKED EXAMPLE No.8

θo 1
A system has a transfer function (s)  Determine the time response to a unit step
θi (s  1)(s  2)
input.

SOLUTION
θi 1
θ o (s)   Convert to partial fractions
(s  1)(s  2) s(s  1)(s  2)
1 1 1 1 1
θ o (s)    Inverse transforms θ o (t)   e  t  e  2t
2s s  1 2(s  2) 2 2

Figure 22

 D.J.DUNN 19
 SELF ASSESSMENT EXERCISE NO.3

θo 1
1. A system has a transfer function (s)  Determine the output time response to a unit
θi (s  2)(s  3)
1 1 1
step input. Answer θ o (t)   e  2t  e  3t
6 2 3

θo 1
2. A system has a transfer function (s)  Determine the output time response to a unit
θi (s  1)(s  2)
t 3 t 1
ramp input. θ o (t)   e  e  2t
2 4 4

 D.J.DUNN 20