Benha University Electrical Engineering Department

Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Chapter # 2 Pulse Transfer Function

After completing this chapter, the students will be able to:

• Obtain the open-loop transfer function for digital control systems,

• Obtain the closed-loop transfer function for digital control systems,

1. Pulse Transfer Function for Open-Loop Systems

In continuous control systems, which represented in S-domain as shown in Fig. 1, the
transfer function can be obtained as:
𝑌(𝑠)
𝑌(𝑠) = 𝑅(𝑠). 𝐺(𝑠) → = 𝐺(𝑠)
𝑅(𝑠)

Input G(s) Output

R(s) Y(s)
Fig. 1 Analog control system

If the input signal R(s) is sampled, it becomes R(s)* or R(z). Also, If the output signal
Y(s) is sampled, it becomes Y(s)* or Y(z). Therefore, the plant G(s) is only defined at
sampling instants only, this means it becomes sampled G(s)* or G(z). The block
diagram of such system is shown in Fig. (2).
1 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein
 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Fig. 2 Digital control system

In that case the value of Y(z)/R(z) is called the transfer function of digital control
system or Pulse Transfer Function as given below.
𝑌(𝑧)
𝑌(𝑧) = 𝑅(𝑧). 𝐺(𝑧) → = 𝐺(𝑧)
𝑅(𝑧)
The pulse transfer function is defined as the ratio of the z-transform of the output of a
linear Time Invariant (LTI) system to the z-transform of its input, when the initial
conditions are zero.
Example:
A unit step signal is applied to the following system. Calculate and draw the output
response of the system, assuming a sampling period of T = 1 s.

U(s) U(s)* Y(s) Y(s)*

1
𝑠+1

2 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

It is important to know, if there are samplers between blocks as shown in Fig. 3, then
the pulse transfer function of the two blocks can be combined as:
𝑌(𝑧)
= 𝐺1 (𝑧)𝐺2 (𝑧)
𝑅(𝑧)

Fig. 3 Cascaded blocks in digital system

If the sampler between blocks is removed as shown in Fig. 4, then the pulse transfer
function of the two blocks can be combined as:
𝑌(𝑧)
= 𝐺1 𝐺2 (𝑧)
𝑅(𝑧)

Fig. 4 Two blocks without sampler in between

3 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein
 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Both transfer functions are different.

𝐺1 (𝑧)𝐺2 (𝑧) ≠ 𝐺1 𝐺2 (𝑧)
Note that the presence of samplers complicates the algebra of block diagrams, since
the existence and expression of any input-output function depend on the number and
location of the samplers.
Example:
The system 1/(S+1) is preceded by a ZOH. What will the system output be if the applied
input is a unit step, assuming a sampling period of T = 1 s.

U(s) U(s)* Y(s) Y(s)*

1
ZOH
𝑠+1

From Z-transform tables:

Example:
Consider the cascaded blocks with

4 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

1 2
𝐻1 (𝑠) = 𝑎𝑛𝑑 𝐻2 (𝑠) =
𝑆+2 𝑆+4
Calculate the equivalent pulse T.F in the following two cases:
a) Both blocks are connected directly,
b) Both blocks are separated by a sampler
In case (a), the equivalent pulse T.F is given as H1H2(Z)
2 1 1
𝐻(𝑆) = 𝐻1 (𝑠)𝐻2 (𝑠) = = −
(𝑆 + 2)(𝑆 + 4) 𝑆 + 2 𝑆 + 4
ℎ(𝑡) = 𝑒 −2𝑡 − 𝑒 −4𝑡
ℎ(𝑘𝑇) = 𝑒 −2𝑘𝑇 − 𝑒 −4𝑘𝑇 𝑘 = 0, 1, 2, 3, …
Taking Z Transform:
𝑍 𝑍 𝑍(𝑒 −2𝑇 − 𝑒 −4𝑇 )
𝐻(𝑍) = − =
𝑍 − 𝑒 −2𝑇 𝑍 − 𝑒 −4𝑇 (𝑍 − 𝑒 −2𝑇 )(𝑍 − 𝑒 −4𝑇 )
In case (b), the equivalent pulse T.F is given as H1(Z) H2(Z)
1 𝑍
𝐻1 (𝑠) = → ℎ1 (𝑡) = 𝑒 −2𝑡 → 𝐻1 (𝑍) =
𝑆+2 𝑍 − 𝑒 −2𝑇
2 2𝑍
𝐻2 (𝑠) = → ℎ2 (𝑡) = 2𝑒 −4𝑡 → 𝐻2 (𝑍) =
𝑆+4 𝑍 − 𝑒 −4𝑇
𝑍 2𝑍 2𝑍 2
𝐻(𝑍) = =
𝑍 − 𝑒 −2𝑇 𝑍 − 𝑒 −4𝑇 (𝑍 − 𝑒 −2𝑇 )(𝑍 − 𝑒 −4𝑇 )
Using partial fractions:
𝐻(𝑍) 2𝑍 𝐴 𝐵
= = +
𝑍 (𝑍 − 𝑒 )(𝑍 − 𝑒 ) (𝑍 − 𝑒 ) (𝑍 − 𝑒 −4𝑇 )
−2𝑇 −4𝑇 −2𝑇

2𝑍 = 𝐴(𝑍 − 𝑒 −4𝑇 ) + 𝐵(𝑍 − 𝑒 −2𝑇 )

At Z = e-2T → 2e-2T = A(e-2T – e-4T) → A = 2e-2T / (e-2T – e-4T)
At Z = e-4T → 2e-4T = B(e-4T – e-2T) → B = 2e-4T / (e-4T – e-2T)
𝐻(𝑍) 2𝑒 −2𝑇 2𝑒 −4𝑇
= −2𝑇 +
𝑍 (𝑒 − 𝑒 −4𝑇 )(𝑍 − 𝑒 −2𝑇 ) (𝑒 −4𝑇 − 𝑒 −2𝑇 )(𝑍 − 𝑒 −4𝑇 )

5 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Taking inverse Z transform

The above example clearly shows the effect of placing a sampler between analog blocks
on the pulse transfer function.
Consider the control system shown in Fig. 5, the relation between input and output is
given as:
𝑌(𝑠) = 𝑅(𝑠). 𝐺(𝑠)
∗
𝑌(𝑠)∗ = (𝑅𝐺(𝑠)) → 𝑌(𝑧) = 𝑅𝐺(𝑧)

Fig. 5 open-loop system with one sampler

Therefore, the pulse T.F. is not defined, but we can calculate the system response only.

2. Pulse Transfer Function for Closed-Loop Systems

Consider the closed-loop system with samplers indicated as shown in Fig. 6

Fig. 6, closed-loop system with sampler at feedback

To obtain the pulse T.F, we assume the sampler output is Y(s)*
E(s) = R(s) − H(s)Y(s)*
Y(s) = G(s)E(s)=G(s)[R(s) − H(s)Y(s)*]
Y(s) = G(s) R(s) − G(s)H(s)Y(s)*
6 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein
 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Y(s) = G(s) R(s) − G(s)H(s)Y(s)*

Y(s)* = RG(s)* − GH(s)*Y(s)*
𝑌(𝑧) = 𝑅𝐺(𝑧) − 𝐺𝐻(𝑧)𝑌(𝑧)
𝑌(𝑧)[1 + 𝐺𝐻(𝑧)] = 𝑅𝐺(𝑧)
𝑅𝐺(𝑧)
𝑌(𝑧) =
1 + 𝐺𝐻(𝑧)
It is clear that we can obtain the system response Y(z) but we can NOT obtain the pulse
transfer function.
To solve this problem, a sampler must be added before the summing point as shown in
Fig. 7.

Fig. 7 closed-loop system with sampler at feedback and sampler at input

In that case the pulse T.F. is
𝑌(𝑧) 𝐺(𝑧)
=
𝑅(𝑧) 1 + 𝐺𝐻(𝑧)
Consider the discrete control system with two samplers shown in Fig. 8

Fig. 8, closed-loop system with sampler at error signal

To obtain the pulse T.F, we assume the sampler output is E(s)*
E(s) = R(s) − H(s)Y(s)
7 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein
 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

E(s) = R(s) − H(s)G(s)E(s)*

E(s)* = R(s)* − GH(s)* E(s)*
E(s)*[1+GH(s)*] = R(s)*

∗
𝑅(𝑠)∗
𝐸(𝑠) =
1 + 𝐺𝐻(𝑠)∗
As the sampled error signal E(s)* multiplied by G(s)* this gives Y(s)*
[Y(s)*=G(s)*E(s)*] (dashed line)

∗
𝐺(𝑠)∗ 𝑅(𝑠)∗
𝑌(𝑠) =
1 + 𝐺𝐻(𝑠)∗
The pulse T.F. is given by:
𝑌(𝑠)∗ 𝐺(𝑠)∗
=
𝑅(𝑠)∗ 1 + 𝐺𝐻(𝑠)∗
𝑌(𝑧) 𝐺(𝑧)
=
𝑅(𝑧) 1 + 𝐺𝐻(𝑧)
We can obtain the same pulse T.F with the configuration shown in Fig. 9

Fig. 9 closed-loop system with two samplers

To obtain the pulse T.F, we follow the steps:

8 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Introduce a variable, with name E(s), at the input of the first sampler and a variable,
with name E(s)*, at the output of that sampler. Moreover, introduce a variable, with
name Y(s), at the input of the second sampler and a variable, with name Y(s)*, at the
output of that sampler.
E(s) = R(s) − H(s)Y(s)
But we know that Y(s)=E(s)*G(s), by substituting by this value in above equation;
E(s) = R(s) − H(s) G(s) E(s)*
E(s)* = R(s)* − HG(s)* E(s)*
E(s)*[1+HG(s)*] = R(s)*

∗
𝑅(𝑠)∗
𝐸(𝑠) =
1 + 𝐺𝐻(𝑠)∗
Replace E(s)* by Y(s)*/G(s)*
𝑌(𝑠)∗ 𝑅(𝑠)∗
=
𝐺(𝑠)∗ 1 + 𝐺𝐻(𝑠)∗
The pulse T.F. is given by:

𝑌(𝑠)∗ 𝐺(𝑠)∗
=
𝑅(𝑠)∗ 1 + 𝐺𝐻(𝑠)∗
𝑌(𝑧) 𝐺(𝑧)
=
𝑅(𝑧) 1 + 𝐺𝐻(𝑧)
Consider the closed-loop system with samplers indicated as shown in Fig. 10

Fig. 10 closed-loop system

9 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

We can obtain the same pulse T.F with the configuration shown in Fig. 11

Fig. 11 Closed-loop discrete system

𝑌(𝑧) 𝐺(𝑧)
=
𝑅(𝑧) 1 + 𝐺(𝑧)𝐻(𝑧)

10 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Example:

Example:

11 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

Example:
For MISO Linear Time Invariant Digital (LTID) control system shown below,
calculate the pulse transfer function.

12 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

First, we assume that D(s) = 0;

forward paths:
P1 = R(z) G1(z) G2(z)

Individual loops:
L1 = −G2 H1 (z)
L2 = − G1(z) G2(z) H2 (z)

The system response is:

P1
𝑌(𝑧) =
1 − (𝐿1 + 𝐿2 )
R(z)G1 (z)G2 (z)
𝑌(𝑧) =
1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
The pulse T.F. is:
𝑌(𝑧) G1 (z)G2 (z)
=
𝑅(𝑧) 1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
Second, we assume that R(s) = 0;

forward paths:
P1 = D(z) G2(z)

Individual loops:

13 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

L1 = −G2 H1 (z)
L2 = − G1(z) G2(z) H2 (z)

The system response is:

P1
𝑌(𝑧) =
1 − (𝐿1 + 𝐿2 )
D(z)G2 (z)
𝑌(𝑧) =
1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
The pulse T.F. is:
𝑌(𝑧) G2 (z)
=
𝐷(𝑧) 1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
G1 (z)G2 (z)𝑅(𝑧) G2 (z) 𝐷(𝑧)
𝑌(𝑧) = +
1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z) 1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
G2 (z){G1 (z)𝑅(𝑧) + 𝐷(𝑧)}
𝑌(𝑧) =
1 + G2 H1 (z) + G1 (z)G2 (z)H2 (z)
Example:
Obtain the pulse T.F from the following difference equation:
𝑦(𝑘) − 0.5𝑦(𝑘 − 1) = 2𝑥(𝑘)
Taking Z-transform for both sides:
𝑌(𝑧) − 0.5𝑍 −1 𝑌(𝑧) = 2𝑋(𝑧)
𝑌(𝑧)(1 − 0.5𝑍 −1 ) = 2𝑋(𝑧)
𝑌(𝑧) 2 2𝑍
= =
𝑋(𝑧) 1 − 0.5𝑍 −1 𝑍 − 0.5
Example:
Obtain the pulse T.F from the following difference equation:
𝑦(𝑘) + 2𝑦(𝑘 − 1) − 𝑦(𝑘 − 2) = 2𝑥(𝑘) − 𝑥(𝑘 − 1) + 2𝑥(𝑘 − 2)
Taking Z-transform for both sides:
𝑌(𝑧) + 2𝑍 −1 𝑌(𝑧) − 𝑍 −2 𝑌(𝑧) = 2𝑋(𝑧) − 𝑍 −1 𝑋(𝑧) + 2𝑍 −2 𝑋(𝑧)
𝑌(𝑧)(1 + 2𝑍 −1 − 𝑍 −2 ) = 𝑋(𝑧)(2 − 𝑍 −1 + 2𝑍 −2 )

14 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

𝑌(𝑧) 2 − 𝑍 −1 + 2𝑍 −2 2𝑍 2 − 𝑍 + 2
= =
𝑋(𝑧) 1 + 2𝑍 −1 − 𝑍 −2 𝑍 2 + 2𝑍 − 1
3. Matlab Command
The Matlab command to obtain the pulse T.F. from analog T.F. is:
>> g = tf(num; den); % continuous TF
>> gd = c2d(g, T, ‘zoh’) % digital TF using zero order hold and sampling time T

3. Characteristic Equation
Characteristics equation plays an important role in the study of linear systems. As said
earlier, an nth order LTI discrete data system can be represented by an n th order
difference equation,
c(k + n) + an−1c(k + n − 1) + an−2c(k + n − 2) + ... + a1c(k + 1) + a0c(k) = bmr(k + m) +
bm−1r(k + m − 1) + ... + b0r(k)
where r(k) and c(k) denote input and output sequences respectively. The input output
relation can be obtained by taking Z-transformation on both sides, with zero initial
conditions, as
𝐶(𝑧) 𝑏𝑚 𝑧 𝑚 + 𝑏𝑚−1 𝑧 𝑚−1 + ⋯ + 𝑏0
=
𝑅(𝑧) 𝑧 𝑛 + 𝑏𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0
The characteristics equation is obtained by equating the denominator of the pulse
transfer function to 0, as
𝑧 𝑛 + 𝑏𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0 = 0
• In a causal system, the output does not precede the input. In other words, in a causal
system, the output depends only on the past and present inputs, not on the future ones.
• The transfer function of a causal system is physically realizable, i.e., the system can
be realized by using physical elements.
• For a causal discrete data system, the power series expansion of its transfer function
must not contain any positive power in z. Positive power in z indicates prediction.
Therefore, in the transfer function (given above), n must be greater than or equal to m.
m = n ⇒ proper transfer function m < n ⇒ strictly proper Transfer Function

15 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein

 Benha University Electrical Engineering Department
Faculty of Engineering at Shubra Dr. Ahmed Mustafa Hussein

References:
[1] Anastasia Veloni, Nikolaos Miridakis. Digital Control Systems: Theoretical Problems and
Simulation Tools, CRC Press, Taylor & Francis Group, 2018
[2] Bosch, R. GmbH. Automotive Electrics and Automotive Electronics, 5th ed. John Wiley & Sons
Ltd., UK, 2007.
[3] Franklin, G. F., Powell, J. D., and Emami-Naeini, A. Feedback Control of Dynamic Systems.
Addison-Wesley, Reading, MA, 1986.
[4] Dorf, R. C. Modern Control Systems, 5th ed. Addison-Wesley, Reading, MA, 1989.
[5] Nise, N. S. Control System Engineering, 6th ed. John Wiley & Sons Ltd., UK, 2011.
[6] Ogata, K. Modern Control Engineering, 5th ed ed. Prentice Hall, Upper Saddle River, NJ, 2010.

16 Chapter Two: Pulse Transfer Function Dr. Ahmed Mustafa Hussein