4.

8 The transfer function G(s)

Lecture Notes
Study Unit 4: Lecture 8

Presented by

Charles Nsukukazifani Msipha

Outcome(s):
By the end of this lecture you should be able to:
 explain what you understand by a transfer function.
 explain in your own words the purpose of the transfer function.
 explain what are zeros and poles of a given system in relation to the transfer function.
 determine the transfer function given a differential equation modelling a time-
dependent system.
 solve problems involving the INPUT-OUTPUT relationship involving the transfer
function.
 explain what the initial value of a time signal is.
 explain what the final value of a given time signal is
 compute both the initial and final values of given signal.

Compiled by: Msipha CN Page 1 of 7

 4.8.0 Introduction
In our introduction to differential equations we introduce differential equations as
mathematical entities/objects to model time-dependent input-output systems. It is
common practice to represent input-output systems schematically as follows:

Figure 1

We saw that given a differential equation, as a model of a time dependent system, it serves
as a model depicting and expressing mathematically the interconnection(s) between the
mathematical model of the input to system and that of the output from the system.

We recall that given any differential equation of the following form:

dmy d m 1 y d3y d2y dy

am m
 a m 1 m 1
 . . .  a3 3
 a 2 2
 a1  a0 y  x (t ) (1)
dt dt dt dt dt

we are never supposed to lose sight of the fact that the left hand-side usually comes in
terms of the derivatives of unknown function of time ( t ), in this particular case y , which
is the mathematical model of the output of the time-dependent system, while the right-
hand is the mathematical model of the input to the system. It is important also to take
note of the fact that the right-hand side can be given in terms of a function of time ( t ) say
x and its derivatives.
In this presentation we look closely at the above input-output flow diagram in Figure 1 and
to make sense how this time ( t - domain) situation is modelled in the frequency or say s -
domain. This will take us to the concept of the transfer function G ( s) , which is the
subject of this discussion.

4.8.1 The transfer function G ( s)

We consider the following differential equation of order m:
dmy d m 1 y d3y d2y dy d nx d n1 x dx
am  a m 1  . . .  a 3  a 2  a1  a0 y  bn  bn 1  . . .  b1  b0 x (2)
dt m dt m 1 dt 3 dt 2 dt dt n dt n 1 dt

which is to say am  0 and we assume that bn  0 , which tells us that the input is modelled
by the sum of the function of time ( t ) namely x and its derivatives of up to order n . We
seek to model the INPUT-OUTPUT relationship modelled by the given differential equation
above in the s -domain. We are also going to consider that the system is at rest at time t  0
which means that ALL the initial conditions are equal to zero, that is:
d m 1 y d3y d2y dy d n1 x dx
(0)  . . .  (0)  (0)  (0)  y (0)  (0)  . . .  (0)  x(0)  0 .
dt m 1 dt 3 dt 2 dt dt n 1 dt

Compiled by: Msipha CN Page 2 of 7

 Applying the Laplace transformation on both sides of the equation (2) taking into account
that ALL the initial conditions above are equal to zero we have:
dmy d m 1 y d3y d2y dy d nx d n1 x dx
am m
 am 1 m 1  . . .  a3 3  a2 2  a1  a0 y  bn n  bn 1 n 1  . . .  b1  b0 x
dt dt dt dt dt dt dt dt

  am s m  am 1 s m 1  . . .  a3 s 3  a2 s 2  a1s  a0  Y ( s )   bn s n  bn 1s n 1  . . .  b1 s  b0  X ( s )   1
X (s)

Y ( s)
am s m  am1 s m1  . . .  a3 s 3  a2 s 2  a1s  a0 
X (s)
  bn s n  bn1s n 1  . . .  b1s  b0  (3)

Now dividing by am s m  am 1 s m 1  . . .  a3s 3  a2 s 2  a1s  a0 on both sides of the equation

(3) above we have the rational function of s :

Y (s) bn s n  bn 1s n 1  . . .  b1s  b0

 .
X ( s ) am s m  am 1s m1  . . .  a3 s 3  a2 s 2  a1s  a0

The rational function of s above is called the transfer function of the system modelled by
the given differential equation (2) above and in short is symbolised by G ( s) that is to say:

Y ( s)
G( s)   Y ( s)  G ( s) X (s ) (4)
X ( s)

The essential purpose of the transfer function is to characterise the system in the frequency
domain and the equivalent equations (4) above express the INPUT-OUTPUT relationship in
the s -domain which is never supposed to be out of sight in this context. The above flow
diagram in Figure 1 in the time-domain is equivalently given by the one below in the
frequency / s -domain:

Figure 2

Very important in the process OUTPUT-INPUT SYSTEMS ANALYSIS are zeros and poles of
the transfer function G ( s) . The zeros of the system occur where:
bn s n  bn 1s n 1  . . .  b1s  b0  0 and the poles of the system occur where:
m m 1 3 2
am s  am 1s  . . .  a3 s  a2 s  a1s  a0  0 . Note that zeros and poles are both set of
zeros of some sort. The zeros of the system are zeros of the numerator of the s -rational
expression G ( s) and in particular the poles are zeros of the denominator of the same s -
rational expression G ( s) . At the zeros of the system, G ( s)  0 and it is expected that the
output should be also equal to zero, however of the other hand at the poles of the system,
G ( s) is undefined or say has G ( s) no definite value, hence the system is not expected to
be operational at such points.

Compiled by: Msipha CN Page 3 of 7

 As a frequency domain function by its own right the transfer function has its own time-
domain representation. This talks to the inverse Laplace transform of G ( s) which is usually
referred to as the impulse response of the system and is supposed to characterise the time-
dependent behaviour of the system.

4.8.2 The Impulse Response

We can simply describe the impulse response of a time-dependent INPUT-OUTPUT
SYSTEM as the time-dependent OUTPUT when the system INPUT is an impulse function.
In this case if the INPUT is x(t )   (t ) the flow diagram in Figure 2 adopts the form:

and the INPUT-OUTPUT relationship given in equations (4) above becomes:

Y ( s )  G ( s ) 1  G ( s ) . (4’)
where the sub-index  is meant to communicate that we are in the presence of the
frequency / s-domain representation of the impulse function of time  (t ) . We note that in
this particular case the inverse Laplace transformation of G ( s) is the time-domain
function:

y (t )  L1 Y ( s )  L1 G ( s )  g (t ) . (5)

and that is what we precisely call the impulse response of a time-dependent system when
it is excited by means of an impulse at some given arbitrary instant of time t  t0 We would
like to take note at this point that the delayed and advanced versions of the impulse function
 (t ) that is  (t   ) and  (t   )    t  ( )  are taken care of here by just recalling that:

L  (t )    (t )e st dt  1
0


(6)
L  (t   )    (t   )e  st dt  1  e  s  L  (t ) e  s  e  s
0

In the following examples we look at particular examples and the upcoming homework
we practice the process of determining the transfer function G ( s) given a differential
equation modelling a given system and how to determine the output given the transfer
function and the input time-domain signal. We reiterate that in case of determining the
transfer function we always hold on the standing assumption that the system in
consideration is at rest, that is to say ALL initial conditions are zero.

Compiled by: Msipha CN Page 4 of 7

 Example 1. Determine the transfer function of the time-dependent system modelled by the
2 2
equations: d 2y  3 dy  2 y  d 2x  x(t )
dt dt dt

2 2 1
Solution: d 2y  3 dy  2 y  d 2x  x(t )  Y ( s )( s 2  3s  2)  X ( s )(s 2  1) 
dt dt dt X ( s)
Y (s ) 2   s 2  3s  2
 (s  3s  2)  ( s 2  1)
X (s )

Y ( s) s2 1
 G(s)   2
X ( s) s  3s  2

Example 2. Consider the flow diagram give below:

and that the transfer function of the time-dependent system is given by:
1
G( s)  and the input time signal is x(t )  3u (t )
s2
i. Determine the output time domain signal y (t )
ii. Hence determine the initial value y (t ) of the time signal.
iii. What is the value of the signal as time t is passing by approaching infinite (  )
Solution:
Determining the output time domain signal y (t )

The input-output relationship in terms of the transfer function is given by the equation:
Y (s )  G ( s ) X ( s )

3
We are given that the input time domain signal is x(t )  3u (t ) hence X (s )  L 3u (t ) 
s
and the transfer function is given by:
1
G( s) 
s2
Now substituting the data given in Y (s )  G ( s ) X ( s ) we have:

3 A B 31 3 1 
Y ( s)    =     
s( s  2) s s  2 2 s  2 s2

3 3  st
 y (t )   e  1.5 1  e 2t 
2 2

Compiled by: Msipha CN Page 5 of 7

 Determining initial value y (t ) of the time signal

This talks to evaluating the y (t ) at time t  0 hence y (0)  1.5 1  e 2(0)   0 .

Determining value of the signal as timet is passing by approaching infinite (  )

This talks to determining the limit of the time signal y (t ) when the time t tends to infinite:

t  t 
 
Hence, we compute:  lim y (t )  lim 1.5 1  e 2 t   1.5 lim 1  e2 t   1.5 1  0   1.5
t 

We shall call this value or similar values the final value of the output signal y (t ) in this
case or say eventually as time is passing by the output signal settles at a definite value 1.5.

Closing comments: Lets us look at briefly at the equation of straight line for example
y  2t . We may replace t by the generic variable x it does not make a difference. The
gradient m  2 has the same kind of interpretation we should have for our G ( s) above.
We observe that when plotting this straight line on a cartesian system of coordinates:

Figure 4

The gradient characterizes the straight line and for every 1-unit increase in the independent
variable the values of the dependent variable y have a gain of 2. In a similar fashion G ( s)
in the equation Y (s )  G ( s ) X ( s ) essentially plays the same role of a gain of the system. We
hope that makes closer to home the essence of G ( s) as a model characterizing the system
modelled by the given differential equation just like m  2 characterises the linear model
modelled by the equation of the lime: y  2t. That also motivates for the naming of the
transfer function as G ( s) .

The comment above concludes this lecture in our next discussion we shall look at plotting
zeros and poles of the system which are precisely the zeros and poles of the transfer
function. We urge you to do further reading to educate yourselves about the significance
of these mathematical entities / objects for as the end of it all, mathematics and its symbols
must communicate some indisputable meaning in our fields of specializations. Otherwise
what would be the UNHOLY or HOLY purpose of doing it in the first place.

Compiled by: Msipha CN Page 6 of 7

 Homework:
A. Given the following differential equations modelling time-dependent systems
determine the corresponding transfer function:
3 2 2 s2
1. d 3y  2 d 2y  3 dy  y  d 2x ; Ans: G ( s)  3
dt dt dt dt s  2 s 2  3s  1
3 2 2 s 2  2s  3
2. d 3y  2 d 2y  dy  d 2x  2 dx  3x ; Ans: G ( s)  3
dt dt dt dt dt s  2s 2  s
3 2 s 1
3. d 3y  2 d 2y  5 dy  dx  x ; Ans: G ( s )  3
dt dt dt dt s  2 s 2  5s
4 3 2 2 s 2  2s  1
4. d 4y  7 d 3y  20 d 2y  26 dy 12 y  d 2x  2 dx  x ; Ans: G ( s )  4
dt dt dt dt dt dt s  7 s 3  20 s 2  26s  12
3 2 2 s2  1
5. d 3y  2 d 2y  dy  d 2x  x ; Ans: G ( s) 
dt dt dt dt s3  2 s 2  3s  1

B. Solve the following problems involving the transfer function G ( s) .

1
1. Given that the input to a system characterised by G ( s )  is x(t )  tu (t )
s3
determine
1.1 the explicit expression the output signal y in terms of the independent variable
t.
 2 2t 2e 3t 
Ans: y (t )       u (t )
 9 3 9 
1.2 both the initial and the final values of the output signal y .
2. A time-dependent time system is modelled in the frequency domain by the transfer
2
function G ( s )  s 2  4s  5
s  4s  3
2.1 Show that the transfer function G ( s) can be expressed as follows:
2 .
G (s )  1 
( s  1)( s  3)
2.2 Determine the impulse response of the system.
Ans: g (t )   (t )   e t  e 3t  u (t )
2.3 Determine the initial value of the impulse response of the system.
2.4 Eventually what is the final value of the impulse response.
2.5 Determine the zeros.
2.6 Determine the poles of the system.
2.7 Draw the pole-zero map or diagram.
3
3. Given the time dependent input signal s (t )  e 4 t and G ( s )  2
determine
s  2s  5
initial and final values of the output signal f (t ) .
 3e 4t 6e t cos 2t 
Ans: f (t )     u (t ) ,
 13 13 

Compiled by: Msipha CN Page 7 of 7