Lecture 12
Lecture 12
3. DISCRETE SYSTEMS
3.6. Pulse Transfer Function
The TF of a sampled-data system is
called the Pulse Transfer Function
Consider the continuous system of
Fig.3.9 (a).
By sampling the input as shown in Fig.
3.9b, an output that is still a continuous
signal is obtained.
Now, we assume that the output is to be
obtained only at the sampling instants.
This is shown in Fig. 3.9 (c) in which the
output is sampled by a sampler
synchronized with the input, called a
phantom sampler.
Fig. 3.9 Sampled-data systems: a. continuous; b.
sampled input; c. sampled input and output
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 1
3. DISCRETE SYSTEMS
3.6.1. Derivation of the Pulse Transfer Function
From Eq. (3.1), the sampled input to the system of Fig. 3.9 (c) is given as a sum of impulses
by
𝑟∗ 𝑡 = 𝑟 𝑛𝑇 𝛿 𝑡 − 𝑛𝑇 (3.25)
Recalling that the impulse response of a system,𝐺(𝑠), is 𝑔(𝑡), the time output of 𝐺(𝑠)
denoted by 𝑐 𝑡 can be written as the sum of impulse responses generated by the input of
Eq. (3.25), as
𝑐 𝑡 = 𝑟 𝑛𝑇 𝑔 𝑡 − 𝑛𝑇 (3.26)
𝐶 𝑧 = 𝑐 𝑘𝑇 𝑧 (3.27)
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 2
1
5/19/2024
3. DISCRETE SYSTEMS
3.6.1. Derivation of the Pulse Transfer Function
Letting 𝑡 = 𝑘𝑇 in Eq. (3.26), we get
𝑐 𝑘𝑇 = 𝑟 𝑛𝑇 𝑔 𝑘𝑇 − 𝑛𝑇 (3.28)
𝐶 𝑧 = 𝑟 𝑛𝑇 𝑔[ 𝑘 − 𝑛 𝑇] 𝑧 (3.29)
𝐶 𝑧 = 𝑟 𝑛𝑇 𝑔 𝑚𝑇 𝑧 ( ) = 𝑔 𝑚𝑇 𝑧 𝑟 𝑛𝑇 𝑧 (3.30)
where the lower limit,𝑚 + 𝑛, was changed to 𝑚 since 𝑔(𝑡) = 0 for 𝑡 < 0, then 𝑛 = 0 in the
first sum's lower limit.
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 3
3. DISCRETE SYSTEMS
3.6.1. Derivation of the Pulse Transfer Function
By definition of the z-transform, Eq. (3.30) becomes
𝐶 𝑧 = 𝑔 𝑚𝑇 𝑧 𝑟 𝑛𝑇 𝑧 = 𝐺 𝑧 𝑅(𝑧) (3.31)
Comments
Eq. (3.31) shows that the transform of the sampled output is the product of the
transforms of the sampled input and the pulse transfer function of the system.
Remember that despite the output of the system being a continuous function,
we had to assume a sampled output (phantom sampler) so as to realize the
compact result of Eq. (3.31).
A common way to find the pulse transfer function, 𝐺 𝑧 , is to start with G(s),
find 𝑔(𝑡), and then use Table 3.1 to find 𝐺(𝑧).
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 4
2
5/19/2024
3. DISCRETE SYSTEMS
3.6.Pulse Transfer Function
Example 3.3: Converting 𝐆 𝐬 in Cascade with z.o.h. to 𝐆(𝐳)
Given a z.o.h in cascade with 𝐺 𝑠 = (𝑠 + 2)/(𝑠 + 1) or
1−𝑒 (𝑠 + 2)
𝐺 𝑠 = (3.32)
𝑠 (𝑠 + 1)
find the sampled-data transfer function,𝐺(𝑧), if the sampling time, 𝑇, is 0.5
second.
Solution
Equation (3.32) is a common occurrence in digital control systems where a
transfer function is in cascade with a zero-order hold. In this case, 𝐺 𝑠 =
(𝑠 + 2)/(𝑠 + 1)is in cascade with a zero-order hold, (1 − 𝑒 )/𝑠.
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 5
3. DISCRETE SYSTEMS
3.6.Pulse Transfer Function
Example 3.3. Soln’ Cont’
For this type of problem, we can formulate a general solution by moving the 𝑠
in the denominator of the zero-order hold to the transfer function 𝐺 𝑠 ),
yielding
𝐺 𝑠
𝐺 𝑠 = (1 − 𝑒 ) (3.33)
𝑠
from which
𝐺 𝑠 𝑧−1 𝐺 𝑠
𝐺 𝑧 = 1−𝑧 𝑧 = 𝑧 (3.34)
𝑠 𝑧 𝑠
Thus, begin the solution by finding the impulse response (inverse Laplace
transform) of 𝐺 𝑠 /𝑠. Hence,
𝐺 𝑠 𝑠+2 𝐴 𝐵 2 1
𝐺 𝑠 = = = + = − (3.35)
𝑠 𝑠 𝑠+1 𝑠 𝑠+1 𝑠 𝑠+1
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 6
3
5/19/2024
3. DISCRETE SYSTEMS
3.6.Pulse Transfer Function: Example 3.3. Soln’ Cont’
Taking the inverse Laplace transform, we get
𝑔 𝑡 =2−𝑒 (3.36)
from which
𝑔 𝑘𝑇 = 2 − 𝑒 (3.37)
Using Table 3.1, we get
2𝑧 𝑧
𝐺 𝑧 == − (3.38)
𝑧−1 𝑧−𝑒
Substituting 𝑇 = 0.5 yields
𝐺 𝑠 2𝑧 𝑧 𝑧 − 0.213𝑧
𝐺 𝑧 =𝑧 = − = (3.39)
𝑠 𝑧−1 𝑧 − 0.607 (𝑧 − 1) 𝑧 − 0.607
From Eq. (3.34),
𝑧−1 𝑧 − 0.213
𝐺 𝑧 = 𝐺 𝑧 = (3.40)
𝑧 𝑧 − 0.607
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 7
3. DISCRETE SYSTEMS
3.6.2. Block Diagram Reduction
The objective here is to determine the closed-loop pulse transfer function of
an arrangement of subsystems incorporating a computer in the loop.
This requires that we know how to manipulate block diagrams for sampled-
data systems, with its transfer function clearly remembered to avoid making
mistakes.
For instance,𝑧 𝐺 𝑠 𝐺 𝑠 ≠ 𝐺 𝑧 𝐺 𝑧 . In the subsequent discussion, the
notation 𝐺 𝐺 𝑠 is used to denote a single function 𝐺 𝑠 𝐺 𝑠 after
evaluating the product.
Thus,𝑧 𝐺 𝑠 𝐺 𝑠 = 𝐺 𝐺 𝑧 ≠ 𝐺 𝑧 𝐺 𝑧 .
4
5/19/2024
3. DISCRETE SYSTEMS
3.6.2. Block Diagram Reduction
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 9
3. DISCRETE SYSTEMS
3.6.2. Block Diagram Reduction
Fig. 3.10 (a) is that of the standard system we derived earlier, whose pulse
transfer function is
𝐺 𝑧 = (3.41)
and which forms the basis for the other entries.
5
5/19/2024
3. DISCRETE SYSTEMS
3.6.2. Block Diagram Reduction
Fig. 3.10 (c) shows a cascade of two subsystems of the type depicted in Fig.
3.10 (a) and we see that the z-transform is the product of the two z-
transforms, or 𝐺 (𝑧)𝐺 𝑧 .
Lastly, Fig. 3.10 (d) presents a case in which the continuous signal
entering the sampler is 𝑅 𝑠 𝐺 𝑠 .
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 11
3. DISCRETE SYSTEMS
3.6.2. Block Diagram Reduction
The z-transform of the input to 𝐺 𝑠 is
z { 𝑅 𝑠 𝐺 𝑠 } = z { 𝑅𝐺 𝑠 } =𝑅𝐺 𝑧 .
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 12
6
5/19/2024
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function
Consider the error sampled system shown in Fig. 3.11.
Since there is no sampler between 𝐺 𝑠 and 𝐻 𝑠 in the closed-loop system
shown in Fig. 3.11, the arrangement is similar to that shown in Fig. 3.10 (b).
The closed-loop pulse transfer function can be written as
𝐶 𝐺 𝑧
𝑧 = (3.43)
𝑅 1 + 𝐺𝐻 𝑧
where 𝐺𝐻 𝑧 = 𝑧 𝐺𝐻 𝑠 .
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function
Consider the error and output sampled system shown in Fig. 3.12, with a sampler
between 𝐺 𝑠 and 𝐻 𝑠 , which is similar to Fig. 3.10 (c). In this case, the closed-
loop pulse transfer function is now written as
𝑧 = (3.44)
( )
7
5/19/2024
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function
Example 3.4: Pulse Transfer Function of a Feedback System
Find the z-transform of the system shown in Fig. 3.13.
Fig. 3.13
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 15
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function
Example 3.4: Pulse Transfer Function of a Feedback System
Solution
The objective of this problem is to illustrate the steps in block diagram
reduction of a sampled-data system.
One of the operations to be always performed is that of placing a phantom
sampler at the output of any subsystem with a sampled input, provided the
nature of the signal transmitted to any other subsystem remains the same.
Thus, for instance, in Fig. 3.14 (a), the phantom sampler 𝑆 can be added.
Another operation that can be carried out is the addition of phantom
samplers 𝑆 and 𝑆 at the input to the summing junction whose output is
sampled.
The reason for this is that provided all samplers are synchronized, the
sampled sum is equivalent to the sum of the sampled inputs.
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 16
8
5/19/2024
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function
Example 3.4:Solution
Next, move sampler 𝑆 and
𝐺(𝑠) to the right past the
pickoff point, as shown in Fig.
3.14(b) in order to yield a
sampler at the input of
𝐺(𝑠)𝐻(𝑠) that matches Fig.
3.10(b).
(a)
Also, 𝐺(𝑠) with sampler 𝑆 at
the input and sampler 𝑆 at
the output matches Fig.
3.10(a).
The closed-loop system now
has a sampled input and a
sampled output. (b)
EEE 2502 Control Engineering III Lecture Notes by Dr. Linus. A. Aloo
5/19/2024 17
3. DISCRETE SYSTEMS
3.6.3 Closed-loop Pulse Transfer Function: Example 3.4:Solution Cont’
𝐺(𝑠)𝐻(𝑠) with samplers 𝑆 and 𝑆 becomes 𝐺𝐻(𝑧), and 𝐺(𝑠) with samplers 𝑆 and 𝑆
becomes𝐺(𝑧), as shown in Fig. 3.14(c).
Converting 𝑅∗ (𝑠) to 𝑅(𝑧) and 𝐶 ∗ (𝑠) to 𝐶(𝑧), gives the system represented fully in the z-domain.
The equations developed for continuous time systems (with Laplace transform transfer
functions) can be used for sampled-data transfer functions by only changing the variables
from 𝑠 to 𝑧.
Hence, using the feedback formula, the first block of Fig. 3.14(d) is realized.
Finally, multiplying the cascaded sampled-data systems, we obtain the final result shown in
Fig. 3.14(e).