Control Systems Engineering

Handout Part 2:

Time Domain Analysis

Contents

1 Introduction .............................................................................................................. 2
2 Time Response for First Order Systems ...................................................................... 2
2.1 Step Response of First Order System - ILT .................................................................. 2
2.2 Time Response for Second Order Systems.................................................................. 7
3 Poles, Zeros and Stability ......................................................................................... 11
3.1 Poles and Zeros.......................................................................................................... 11
3.2 Stability ...................................................................................................................... 12
3.3 Summary of pole positions for a second order example (2 poles) ........................... 13
3.4 Examples of poles and stability ................................................................................. 14
4 The Final Value Theorem ......................................................................................... 15
5 Time Response Examples & Introduction of P+I Control ........................................... 16
5.1 Ex1: FVT for steady-state-error (Motor with Proportional Control) ......................... 16
5.2 Proportional Control + Integral Control .................................................................... 17
6 Example: Time domain Calc’s for Water Tank System .............................................. 19

Summary

This part of the course material (part 2) focusses on the time domain responses. These can
be calculated accurately using the Inverse Laplace Transform. Alternatively, for 1st and 2nd
order systems, we can get a good approximation (sketch) by recognising some key elements
of the transfer functions. Being able to “imagine” the response like this is useful for practicing
engineers. We also introduce the final value theorem – which is used to find the steady-state
response after an input. Along the way we will briefly discuss poles and zeros and the impact
of the pole locations on stability.

R Dixon 1 of 23
1 Introduction
In Maths, you previously have covered Laplace and Inverse Laplace Transforms. We have
also mentioned these in Part 1 of the Control Systems course.

Here sections 2.1 and 2.2 will show how the time responses can be calculated for first and
second order systems, respectively. We will also discover how their responses can be
estimated/sketched based on knowledge of the key parameters in the transfer function. Both
will be analysed in open and closed-loop (the latter with proportional control).

In section 3 Then we will move on to consider the poles and zeros of a system and the effect
of pole location on stability.

In section 4 we will consider the Final Value Theorem which provides a short-cut to analyse
the steady state response of a system model in transfer function form. (I.e. without having to
calculate the Inverse Laplace Transform and evaluate the resulting time function).

Finally, in section 5, water-level control will be used as an example: including closed-loop

with proportional + integral action.

2 Time Response for First Order Systems

We have already seen the common form of a generic 1st order TF:

𝐾𝐾
𝐺𝐺(𝑠𝑠) = where K = 1 / K e and τ = RJ / K t K e
𝜏𝜏𝜏𝜏+1

Previously it was stated that, K is the steady-state gain of the system and τ is the time taken
for the system to go from 0 to 63% of the final value. Here we will see why…..

2.1 Step Response of First Order System - ILT

Y (s)
G(s) is the transfer function between the input and output. I.e., G ( s ) =
U (s)

1
Now for a unit step input we know (from LT Tables) that U(s) = .
s

So in the Laplace domain, the expression for Y(s) given a unit step is given by:

K 1
Y ( s ) = G ( s )U ( s ) = .
τs + 1 s
If we want the time function, then we need to do Inverse Laplace Transform using the tables
to obtain,

 K 1  −
t

y (t ) = L−1 {G ( s ).u ( s )} = L−1  .  = K 1 − e τ 

τs + 1 s   

{Note: We need to do partial fraction expansion then use tables (item 4) for Inverse Laplace}

R Dixon 2 of 23
 t
−
If we expand out the expression, we see that y (t ) = K − Ke τ

From inspection we can see that, as t becomes very large and the exponent term tends to
zero, the system will settle to the value K in the steady state.

Hence, our steady-state gain is known as K.

( )
If we set t = τ , we find that at that time the response will be y (τ ) = K 1 − e −1 = 0.63K .

Clearly, these two terms are as stated above (and as presented without explanation in the
motor modelling, in the part 1).

2.1.1 Typical First Order Response

So it is now possible to picture the general first order response as from the above equation.

It will be as shown below (with time normalised with respect to τ and y(t) normalised to K).

The rise time and settling time are shown. But most notably we can see that when we set t
equal to the time constant, τ, y (τ ) = 0.63K (as shown) and that the settling time occurs at
time t = 4τ.

R Dixon 3 of 23
2.1.2 Example for DC Motor (First Order Model)

The first order model

for a DC motor has
been introduced
previously in Part 1.
Here the system is as
shown on right.

Either by Laplace Transform and then algebra or by drawing the diagram and using block
diagram algebra. The model can be shown to be:

𝜃𝜃̇ (𝑠𝑠) 𝐾𝐾𝑇𝑇

or =
𝑉𝑉(𝑠𝑠) 𝑅𝑅𝑅𝑅𝑅𝑅+𝐾𝐾𝑇𝑇 𝐾𝐾𝑣𝑣

The closed-loop with a proportional gain can then be drawn (complete the diagram):

Again we can use block diagram reduction (or normal algebra) to show that the closed loop
transfer function can be written in the form:

𝐺𝐺𝐾𝐾𝑇𝑇
𝜃𝜃̇(𝑠𝑠) 𝐺𝐺𝐾𝐾𝑇𝑇 + 𝐾𝐾𝑇𝑇 𝐾𝐾𝑣𝑣
𝐺𝐺𝐶𝐶𝐶𝐶 = =
𝜃𝜃̇𝑖𝑖𝑖𝑖 (𝑠𝑠) 𝑅𝑅𝑅𝑅
�𝐺𝐺𝐾𝐾 + 𝐾𝐾 𝐾𝐾 � 𝑠𝑠 + 1
𝑇𝑇 𝑇𝑇 𝑣𝑣

R Dixon 4 of 23
We can then analyse the closed-loop step response using Inverse Laplace Transform or
using our knowledge of the general 1st order response as discussed in the previous section.
Here we will take the latter approach to analyse the closed-loop response. If we put in the
numbers for the system parameters we obtain:

G
θ( s) G + 0.1
GCL = =
θin ( s)  0.0015 
 s + 1
 G 0.08 + 0.008 

And we can, by comparing this closed-loop TF with our general model, begin to see the
effect of the proportional gain G.

K G 0.0015
G (s) = where now, K = and τ =
τs + 1 G + 0.1 G 0.08 + 0.008

Assuming, for example, a proportional controller gain G with a gain of 2 V/rads-1. Then we
have

K = 0.952 and τ = 0.009

From here we can sketch the closed-loop system response to a step of 100rad/s-1.

R Dixon 5 of 23
2.1.3 Exercise 2.1
In your own time, perform the same analysis using the Inverse Laplace Transform. Calculate
the values of the output speed for suitable time values and check that you get the same
100
response curve as above. (Remember, a step of 100 rad/s is modelled by ).
𝑠𝑠

t (seconds)

Theta (rad/s)

120

100
Motor Velocity /rads -1

80

60

40

20

0
0 0.01 0.02 0.03 0.04 0.05 0.06
Time /s

R Dixon 6 of 23
2.2 Time Response for Second Order Systems
2
𝜔𝜔𝑛𝑛
There is a general form of 2nd order TF: 𝐺𝐺(𝑠𝑠) = 2
𝑠𝑠2 +2𝜁𝜁𝜔𝜔𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

Where,

ζ is the damping ratio

and

ω n is the undamped natural frequency.

The damping ratio reflects the level of damping in the system as a fraction of critical damping
(see below) and the undamped natural frequency is the frequency at which the system would
oscillate if ζ were zero.

2.2.1 Inverse Laplace for Step Response of Second Order System

Complex poles can be defined in terms of their real and imaginary parts s = −σ ± jω d . In
terms of analysing 2nd order performance, it can be useful to consider rewriting the transfer
ω n2
function as G ( s ) = .
( s + ζω n ) 2 + ω n2 (1 − ζ 2 )

This form arises from looking at the characteristic polynomial of the system with two complex
poles. i.e. ( s + σ + jω d )( s + σ − jω d ) = ( s + σ ) 2 + ω d2

It allows the step and impulse responses of the second order system to be described in
terms of 𝜔𝜔𝑛𝑛 , 𝜁𝜁, and ωd (where ω d = ω n 1 − ζ 2
).

So now, in the Laplace domain, the expression for Y(s) given a unit step is given by:

ω n2 1
Y ( s ) = G ( s )U ( s ) = .
( s + ζω n ) + ω n (1 − ζ ) s
2 2 2

If we want the time function, then we need to do Inverse Laplace Transform using the tables
to obtain,

 ω n2 1
y (t ) = L−1 {G ( s ).u ( s )} = L−1  . 
 ( s + ζω n ) + ω n (1 − ζ ) s 
2 2 2

After some maths we can eventually show that

 ζω 
y (t ) = 1 − e −ζω nt  cos ω d t + n sin ω d t 
 ωd 

{Note: First, do partial fraction expansion, then use tables for Inverse Laplace – this is a very
longwinded procedure and so is not explained in detail here. You will not be expected to do
step response calculations via the ILT for 2nd order systems in the exam.}

R Dixon 7 of 23
Note that it should be clear that the oscillation (the cosine and sine terms in the bracket) will
be reduced to zero (for positive values of 𝜁𝜁 and 𝜔𝜔𝑛𝑛 ) depending on the magnitude of 𝜁𝜁. Very
small values will allow the oscillations to continue for longer. We can see the effect of
different values of the damping factor 𝜁𝜁 below.

2.2.2 Typical Second Order Response

The shape of the step response for the second order system, as we change damping factor
and natural frequency, comes directly from the maths above.

The general second order response takes the form shown in the figure below. Here four step
response examples with different damping factors are shown.

Reminder of general TF:

2
𝜔𝜔𝑛𝑛
𝐺𝐺(𝑠𝑠) = 2
𝑠𝑠 2 +2𝜁𝜁𝜔𝜔𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

These second order responses

can be described by the
following terms, commonly
used among control systems
engineers:

𝜁𝜁 < 0.5 → under-damped

𝜁𝜁 > 0.5 → well-damped

𝜁𝜁 = 1 → critically damped

𝜁𝜁 > 1 → over-damped

In terms of the step

response, ωn is related to
the rise time by,

Tr ~ 1.8 / ωn

and ζ governs the amount of

overshoot and oscillation

Figure showing general form for second order step response

R Dixon 8 of 23
2.2.3 Example: Analysis DC Motor – Position Control
The first order model for a DC motor has been introduced previously. Here the motor system
is identical, but we are interested in controlling the position rather than the speed of the
motor.

Make the required change in the block diagram model below:

The (correctly modified) model reduces to a single transfer function

𝜃𝜃(𝑠𝑠)) 𝐾𝐾𝑇𝑇
=
𝑉𝑉𝐴𝐴 (𝑠𝑠) 𝑠𝑠(𝑅𝑅𝑅𝑅𝑅𝑅+𝐾𝐾𝑇𝑇 𝐾𝐾𝑣𝑣 )

The closed-loop with a proportional gain can then be drawn (complete the missing elements
of the diagram):

Again we can use block diagram reduction (or normal algebra) to show that the closed loop
transfer function can be written in the form:

GK T
θ (s) RJ
GCL = =
θ in ( s ) K K GK T
s2 + T v s +
RJ RJ

R Dixon 9 of 23
We can then insert the motor parameters and analyse the closed-loop step response using
Inverse Laplace Transform or using our knowledge of the general second order response as
discussed in the previous section. Here we will again take the second approach to analyse
the closed-loop response.

θ (s) 53.3G
G CL = = 2
θ in ( s ) s + 5.33s + 53.3G

And we can, by comparing this closed-loop TF with our general model, begin to see the
effect of the proportional gain G.

2
𝜔𝜔𝑛𝑛
𝐺𝐺(𝑠𝑠) = 2 where we can see (by inspection) that
𝑠𝑠2 +2𝜁𝜁𝜔𝜔𝑛𝑛 𝑠𝑠+𝜔𝜔𝑛𝑛

𝜔𝜔𝑛𝑛 = √53.3𝐺𝐺 and 𝜁𝜁 = 5.33/2𝜔𝜔𝑛𝑛 = 5.33⁄2√53.3𝐺𝐺

Assuming, for example, a proportional controller gain G with a gain of 1 V/rad. Then we have

𝜔𝜔𝑛𝑛 = 7.30 and 𝜁𝜁 = 5.33/14.6 = 0.36

From this we can work out that,

θ ss 53.3G
1) The steady state gain will be unity: GCL ss = = . (set 𝑠𝑠 = 𝑗𝑗ω = 0)
θ in ss 0 + 0 + 53.3G
2) The rise time will be Tr = 1.8/7.3 = 0.25 seconds

3) The overshoot of the order of ~30% (interpolating between

0.3 and 0.5 from table)

So we could sketch the unit step response out with this in

mind….

Of course the above is produced using Simulink and actually we see that whilst the rise time
is 0.25 seconds the overshoot is about 30% (close enough!).

R Dixon 10 of 23
3 Poles, Zeros and Stability
The “natural response” of the system describes the way that the system inherently behaves;
the nature of the natural response depends only upon the system, not the type of input. The
“forced response” of the system describes the way that the system responds to the input; the
nature of the response depends upon the input (although the system has an influence also).

The characteristics of the natural and forced responses are determined by the values of the
transfer function’s poles and zeros.

3.1 Poles and Zeros

The poles of a transfer function are the values of the Laplace operator ‘s’ that cause the
transfer function to become infinite – these determine the system’s natural response
characteristics.

The zeros of a transfer function are the values of the Laplace operator ‘s’ that cause the
transfer function to become zero - determine the system’s forced response characteristics

B( s)
For a TF, G ( s ) = , where A(s) and B(s) are polynomials in s.
A( s )

The poles are found by setting the denominator polynomial, A(s)=0 and solving for s.
Whereas the zeros are found by setting the numerator polynomial, B(s)=0 and solving for s.

These poles and zeros can be plotted on the complex s-place as shown below. By
convention the zeros are marked with “o” and the poles with “x”. Note that the real axis is
denoted as σ and the imaginary as jω.

Example for pole-zero plot on the s-plane

s+2
G(s)=
s+5
jω

X
-5 -4 -3
O
-2 -1 0 1 2 3
σ

R Dixon 11 of 23
Why are they called poles?

We can see, if we plot a 3D diagram with the s-plane

and the magnitude of our transfer function, H(s).

1
e.g. 𝐻𝐻(𝑠𝑠) =
𝜏𝜏𝜏𝜏+1

set the denominator to zero: 𝜏𝜏𝜏𝜏 + 1 = 0

1
Therefore, the pole is at 𝑠𝑠 = −
𝜏𝜏 and we see that as s
1
approaches −
𝜏𝜏 the magnitude of H(s) tends to infinity:
1
i.e. 𝐻𝐻(𝑠𝑠) = = ∞.
1−1

You can follow a similar process with the numerator for zeros – try it if you like.

3.2 Stability
A linear transfer function system is called bounded-input bounded-output (BIBO) stable if its
output will stay bounded for any bounded input.

For a linear time invariant system to be stable, all of the poles must be located in the left-half
of the s-plane (i.e. all the poles have negative real parts).

When any one (or more) of the poles of the transfer function are located in the right-half of
the s-plane (i.e. poles with positive real parts), the system will be unstable. If a pole lies on
the boundary (when the real part is zero) mean the system will be marginally stable.

3.2.1 Summary of pole positions (single pole)

Imaginary Axis

s-plane
s = σ+ jω

x x x Real Axis

system with all poles on system with any pole on

the LHP = stable RHP is unstable
A pole on the imaginary Step Response
Step Response
1.5 axis is marginally stable 0
Step Response
10
Amplitude
Amplitude

1
8 -5
Amplitude

0.5 6
-10
4
0 0 1 2 3 4 5
0 1 2 3 4 5 2
Time (sec) Time (sec)
0
0 2 4 6 8 10
Time (sec)

R Dixon 12 of 23
3.3 Summary of pole positions for a second order example (2 poles)
If we revisit the motor closed-loop (position control with proportional controller in section
2.2.3). We get the following pole positions and responses as the gain G is altered.

G = 0.1 V/rad G = 1 V/rad G = 10 V/rad

jω jω Χ jω

Χ
Χ Χ σ σ σ
Χ
Χ

As the gain is increased, the poles take a larger imaginary part and the oscilation magnitude
increases (more oscillatory).

Note: Typically control engineers are interested in stability of the closed-loop system (after
they have added the controller). There are a number of tests which deal with closed-loop
stability: The Routh-Hurwitz Criteria, the Root Locus method, and the Nyquist Stability
Criteria all test whether there are poles of the closed-loop transfer function lie in the RHP.
These can be read about in the recommended text book.

The Nyquist Stability Criteria will be dealt with in this course (in the section on Frequency
Domain (part 3). It is favoured because it leads directly to useful concepts such as Gain and
Phase Margins. These are of interest as they give engineers information about the relative
stability which is far more useful than absolute stability.

Relative stability is about how stable a system is rather than just stating that it is stable (or
not).

R Dixon 13 of 23
3.4 Examples of poles and stability
1
1) Show that the poles of G ( s ) = are at 𝑠𝑠 = −1, −5, sketch them and
s + 6s + 5
2

comment on stability.

1
2) Show that the poles of G ( s ) = are at 𝑠𝑠 = −2 ∓ 1𝑖𝑖 sketch them and
s + 4s + 5
2

comment on stability.

3) Are the systems in 1) and 2) stable?

1
4) Is G ( s ) = stable?
s + 4s − 5
2

R Dixon 14 of 23
4 The Final Value Theorem
The Final Value Theorem (FVT) is an important theorem that is used to relate frequency
domain expressions to the time domain behaviour as time approaches infinity. It allows the
time domain behaviour to be directly calculated by taking a limit of a (frequency) s-domain
expression, rather than having to convert (via ILT) to a time domain expression and taking its
limit.

The FVT says that if a time function f(t) has a finite limit

i.e lim 𝑓𝑓(𝑡𝑡) is finite

𝑡𝑡→∞

Then we can find that limit by evaluating a limit in the s-domain as follows.

lim 𝑓𝑓(𝑡𝑡) = lim 𝑠𝑠𝑠𝑠(𝑠𝑠)

𝑡𝑡→∞ 𝑠𝑠→0

It is item 11 in the table of Laplace Transform Theorems (in Part 1 of the notes).

Let’s look at an example. In Part 1 we looked at a DC motor model and should have found
that the steady state response to a step of magnitude 100 would be 95.2rad/s. Also, in
section 2.1.3, you worked this out for yourself using the Inverse Laplace Transform. Let’s
confirm it here using the FVT.

The TF model for the motor, proportional gain of 2 V/(rad/s) and closed-loop was found to be

𝜃𝜃̇(𝑠𝑠) 0.952 100

𝐺𝐺𝑐𝑐𝑐𝑐 (𝑠𝑠) = = . The 100rad/s step input is 𝜃𝜃̇𝑖𝑖𝑖𝑖 (𝑠𝑠) =
𝜃𝜃̇𝑖𝑖𝑖𝑖 (𝑠𝑠) 0.009𝑠𝑠+1 𝑠𝑠

So applying FVT {need to realise that here 𝐹𝐹(𝑠𝑠) = 𝐺𝐺𝑐𝑐𝑐𝑐 (𝑠𝑠). 𝜃𝜃̇𝑖𝑖𝑖𝑖 (𝑠𝑠)}:

𝜃𝜃̇𝑠𝑠𝑠𝑠 (𝑠𝑠) = lim 𝑓𝑓(𝑡𝑡) = lim 𝑠𝑠𝑠𝑠(𝑠𝑠) = lim s

0.952 100 95.2
= = 𝟗𝟗𝟗𝟗. 𝟐𝟐 rad/s
𝑡𝑡→∞ 𝑠𝑠→0 𝑠𝑠→0 0.009 s + 1 s 1

Note: Strictly speaking, we should check that the time function f(t) does have a finite limit.
For that to be the case:

1. All roots of the denominator of system G(s) must have negative real parts (be
stable).
2. F(s) must not have more than one pole at the origin.

To me, 1 means 2 will be satisfied unless we use the ramp ( 1� 2 ) or higher order inputs. So
𝑠𝑠
FVT works for impulse or step tests (provided 1 is satisfied).

R Dixon 15 of 23
5 Time Response Examples & Introduction of P+I Control
In this section we will look again at the example from section 4 above to find the steady
state error on a dc motor with proportional controller.

Then we will introduce the concept of integral action – forming a so called Proportional +
Integral (PI) controller. We will repeat the steady state error analysis to see that the addition
of integral action in the controller has the effect of causing the steady-state error to be zero.

5.1 Ex1: FVT for steady-state-error (Motor with Proportional Control)

The the first order motor plus proportional controller G can be drawn as below (with numbers
inserted as per the previous example motor model):
.
θin .
Step = 100rad/s ε VA
53.3 θ
G

- s +System
Motor 5.33
2 V/(rad/s)

It is straightforward to show that the transfer function from the error, ε, to the command input
θ is given by the following general expression:
in

ε 1
= - where G is the controller and P is the plant model.
θin 1 + GP

This expression is known as the sensitivity function.

ε s + 5.33
For the above example: =

θ in s + (5.33 + G53.3)

Now if the input is a step of magnitude 100 we can arrive at the following analysis:

ε s + 5.33 100
= and θin =
θin s + (5.33 + G53.3) s

The Final Value Theorem States that, lim 𝑓𝑓(𝑡𝑡) = lim 𝑠𝑠𝑠𝑠(𝑠𝑠)
𝑡𝑡→∞ 𝑠𝑠→0

s + 5.33 100
Hence, ε (∞) = lim s
s →0 s + (5.33 + G53.3) s

533
∴ 𝜀𝜀(∞) =
(5.33 + 𝐺𝐺53.3)
533
= (5.33+(2𝑥𝑥53.3)) = 4.8rad/s as before

What general statement can we make in relation to this system with proportional control and
the effect on the steady state error ε (∞) ? {See the equation with G in two lines up}

R Dixon 16 of 23
5.2 Proportional Control + Integral Control
The P+I or PI controller is the most common controller used in practice. Adding integral
action results in a system (in general) as shown…..

Gs + Gi
The transfer function for this controller is often represented as . Where G is the
s
proportional gain and Gi is the integral gain. Note that you will see this represented in a
slightly different form in the design section (part 4).

5.2.1 Calculation of steady-state-error Motor with Proportional + Integral Control

Now our control structure is as follows:

.
θin .
ε VA
53.3 θ
Step = 100rad/s
GG
+sGi
s s + 5.33
- 2 V/(rad/s)
Motor System

P+I

Again we can do block diagram analysis to show that:

ε s 2 + 5.33s
=
θin s 2 + (5.33 + G53.3) s + 53.3Gi

It would be good practice for you to try to show this!

Now if the input is a step of magnitude 100 we can arrive at the following analysis with the PI
controller:

ε s 2 + 5.33s 100
= 2 and θin =

θ in s + (5.33 + G53.3) s + 53.3Gi s

R Dixon 17 of 23
The Final Value Theorem States that, lim 𝑓𝑓(𝑡𝑡) = lim 𝑠𝑠𝑠𝑠(𝑠𝑠)
𝑡𝑡→∞ 𝑠𝑠→0

s 2 + 5.33s 100
Hence, ε (∞) = lim s
s →0 s + (5.33 + G53.3) s + 53.3Gi s
2

∴ ε (∞) = 0rad/s

What can we say in relation to this system with P+I control and the effect on the steady state
error ε (∞) ?

In sum we have shown that the PI control (in particular addition of Integral action) has the
effect of driving the steady state error to zero. It does this regardless of the value of the
integral gain – though the gain affects how quickly it goes to zero.

R Dixon 18 of 23
6 Example: Time domain Calc’s for Water Tank System

Aim:
Carry out time domain response analysis of the level control example (used in of Part 1,
Section 4.4).

Description:
The exercise is concerned with the tank system shown below. The closed-loop transfer
function with Gain G allows us to calculate some dynamic performance characteristics. A
number of questions related to this are set below.

P(s)
hin ε q h
G

-
Tank System

Variables: Constants:
q- is the control input [m3s-1] c =0.1m2s-1 - the (linearised) outlet
h- is the level of the liquid [m] which coefficient
is the system output (that we A = 0.5m2 - the cross-sectional area of
would control) the tank
qout – is the flow out of the tank (given by

Earlier analysis in Part 1 gave:

10 10G
P( s) = and a closed-loop TF Gcl ( s ) =
5s + 1 5s + (1 + 10G )

R Dixon 19 of 23
Tank Control System Tasks (Questions):
1. From inspection write down the time constant and the steady state gain of the closed-
loop response with G = 1.

2. Estimate the settling-time (Ts) and the initial slope and sketch a (labelled) step
response when the input command hin is a step of amplitude 1m.

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5 3

3. Evaluate the time domain response and check (using ILT) the values at two or three
key points in time to confirm your sketch.

R Dixon 20 of 23
 4. What will the steady-state error be (calculate it using Final Value Theorem – check
against answer to 3)?

5. Assume G is still to be decided. What value of G would be needed to give a steady

state error of zero?

R Dixon 21 of 23
Additional Questions/Tasks:

𝐺𝐺𝑖𝑖
6. Show that, for the tank system above, introducing an integral controller TF 𝐺𝐺(𝑠𝑠) = 𝑠𝑠
rather than the proportional controller would result in a steady state error of zero after a
step input.

6. Check your answers above with Matlab/Simulink

R Dixon 22 of 23
Appendix 1
No. f(t) F(s)

1 δ (t ) unit impulse function 1

1
2 u (t ) unit step function
s

n!
3 tn
s n +1

1
4 e − at
s+a

ω
5 sin(ωt )
s2 + ω2

s
6 cos(ωt )
s2 + ω2

ω
7 e − at sin(ωt )
(s + a )2 + ω 2
(s + a )
8 e − at cos(ωt )
(s + a )2 + ω 2

Laplace Transform Theorems

R Dixon 23 of 23