ECEg4322: Digital Control

Systems
Nigusu Teshome,
Lecturer/Electrical and Computer Engineering Department,
Institute of Technology,
University of Gondar, Ethiopia.
eMail: nigusteshome@gmail.com
 Chapter 2
Design of Digital Control Systems
Using Transform Techniques
 Contents
2.1 Control System Specifications
2.2 Design by Emulation
2.2.1 Discrete Equivalent Controllers
2.2.2 Evaluation of the Design
2.3 Stability Analysis of closed-loop systems in z-plane
2.4 Direct design by using root locus in the z-plane
2.5 Frequency response methods with z-transform and w-transform
2.6 Direct design method of Ragazzini, PID controller
 2.1 Control System Specifications
 A controller is designed to achieve a given performance.
 The performance is given in terms of specifications.
 The most common specifications are:
• Steady state error:
 Depends on forward loop gain.
 Smaller is better.
• Transient performance:
 Location of dominant poles
 Small rise time, small overshoot, small settling time
• Relative stability:
 Location of dominant poles and forward loop gain
 Phase margin, gain margin
• Sensitivity to parameter variation:
 Forward loop gain
 The lesser the sensitivity, the better is the system performance.
• Control effort needed:
 Smaller is better.
Cont’d…
 The central concerns of controller design are:
 for good transient and steady state response and
 for sufficient robustness.
 Dynamic performance of systems in the time-domain is defined in terms
of parameters of system response to a unit step in command input.
 The most commonly used time domain specification parameters are:
 Rise time (tr):- The time at which the response of the system rises from 0 to
100%, 5 to 95%, 10 to 90% of its steady state value.
 Delay time (td):- The time at which the response of the system reaches to 50% of
its final value.
𝑦 𝑡𝑑 = 50% y 𝑡𝑓 = 0.5 *1 = 0.5
Cont’d ...
 Peak time (tp):- The time at which the response of the system reaches
in to its peak value.
𝑑
𝑦 𝑡 |𝑡 = 𝑡 = 0
𝑑𝑡 𝑝

 Settling time (ts):- The time at which the response of the system
reaches in to 98% or 95% of its final value.
• For 2% settling (98% reaching):
𝑦 𝑡𝑠 = 98% 𝑦 𝑡𝑓
= 0.98
• For 5% settling (95% reaching):
𝑦 𝑡𝑠 = 95% y 𝑡𝑓
= 0.95
Cont’d ...
 Maximum Percent Overshoot (Mp):- It measures the deviation between the peak
value of the response and the final value.
𝑦 𝑡𝑝 − 𝑦(∞)
𝑀𝑝 = * 100%
𝑦(∞)
Cont’d ...
 The s-plane expressions of the time domain specification requirements
are:
- Natural frequency (𝜔𝑛 ):
1.8
𝜔𝑛 ≥
𝑡𝑟
- Magnitude of the real part of the pole (𝜎):
4.6
𝜎 = 𝑅𝑒 𝑆𝑖 = 𝜁𝜔𝑛 ≥
𝑡𝑠

- Damping ratio (𝜁):

𝜁 ≈ 0.6 ( 1 − 𝑀𝑝 )
Cont’d ...
 Steady state error to the polynomial inputs (ess):
• Type 1 systems are systems that have zero steady state error to a step input and
finite error to a ramp input of slope ro.
𝑟𝑜
𝑒𝑠𝑠 =
𝐾𝑣
Where 𝑒𝑠𝑠 is steady state error, 𝑟𝑜 is slope of ramp input and 𝐾𝑣 is velocity constant.
• For a single loop system with unity feedback gain and forward transfer function of
𝐷 𝑠 𝐺(𝑠), the system is Type 1 if and only if 𝐷 𝑠 𝐺(𝑠) has a simple pole at 𝑠 = 0.

Figure: A unity feedback system.

Cont’d ...
• The velocity constant for Type 1 system is:
𝑟𝑜
𝐾𝑣 = = lim 𝑠𝐷 𝑠 𝐺(𝑠)
𝑒𝑠𝑠 𝑠→0
 The relation between sampling frequency and system natural frequency for 𝑘
samples in a rise time will be defined as the following:
-Sampling frequency (𝜔𝑠 ):
1
𝜔𝑠 = 2𝜋𝑓𝑠 but 𝑓𝑠 =
𝑇
2𝜋
⇒ 𝜔𝑠 =
𝑇
- Natural frequency (𝜔𝑛 ):
1.8
𝜔𝑛 = but 𝑡𝑟 = 𝑘𝑇
𝑡𝑟

1.8 1.8
𝜔𝑛 = ⇒𝑇=
𝑘𝑇 𝑘𝜔𝑛
Cont’d ...
• Then
2𝜋 2𝜋 2𝜋𝑘𝜔𝑛
𝜔𝑠 = = =
𝑇 1.8 𝑘𝜔𝑛 1.8
⟹ 2𝜋𝑘𝜔𝑛 = 1.8 𝜔𝑠
1.8
⟹ 𝜔𝑛 = 𝜔
2𝜋𝑘 𝑠

0.29
⟹ 𝜔𝑛 = 𝜔𝑠
𝑘
𝜔𝑠
∴ 𝜔𝑛 =
3.5𝑘
 In other words, the sample rate 𝜔𝑠 , should be 3.5𝑘 times faster than the natural
frequency 𝜔𝑛 . That means,
𝝎𝒔 = 𝟑. 𝟓𝒌𝝎𝒏
2.2 Design by Emulation
 A controller design is done as if the system is to be continuous and after a
sample period is selected, a discrete equivalent is computed and used in place of
the continuous design. This discrete controller may then be simulated and tested
in the discrete control loop and modifications made if necessary.
2.2.1 Discrete Equivalent Controllers
 Design steps of continuous time lead compensator;
 General transfer function of lead compensator:
𝑠+1 𝜏
𝐷 𝑠 = 𝐾𝑐 1 where 0 < 𝛼 < 1
𝑠+ 𝛼𝜏
Step 1:- From the given performance specifications of the system, determine the location
of the desired closed-loop poles.
𝑠𝑑 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2
Step 2:- From the root locus of uncompensated system, check whether gain adjustment
alone can yield the requirement, if not calculate the angle deficiency (𝜙).
Cont’d…
Step 3:- Calculate the poles and zeros of compensator by assuming the zeros angle
contribution is equal to 90𝑜 .
Step 4:- Find the open-loop gain of the compensator from magnitude criteria.
Example 1: Consider the gain of the system:
4
𝐺𝐻(𝑠) = where H(s) is unity feedback
𝑠(𝑠+2)

 The control specifications of the system are:

i. Overshoot to a step input is less than 25.46%.
ii. Settling time to 2% to be less than 4.6 sec.
iii. Tracking error to a ramp input of slope 0.01 rad/sec to be less than 0.01 rad.
iv. Sampling time to give at least 2 samples in a rise time.
 Design a controller D(z) for this system using the method of emulation.
Solution:
 Acceptable regions in the s-plane for the closed-loop poles:
i) 𝑀𝑝 ≤ 25.46%
⟹ 𝜁 ≈ 0.6 ( 1 − 𝑀𝑝 ) ≈ 0.6(1 − 0.2546) ≈ 0.44724
⟹ 𝜁 ≥ 0.44724
ii) 𝑡𝑠 ≤ 4.6
4.6 4.6
⟹ 𝜎 = 𝜁𝜔𝑛 = ⇒𝜎= =1
𝑡𝑠 4.6
as 𝑡𝑠 ↓, 𝜎 ↑
⟹𝜎≥1
iii) 𝑒𝑠𝑠 ≤ 0.01
𝑟𝑜 0.01
⟹ 𝐾𝑣 = = =1
𝑒𝑠𝑠 0.01
as 𝑒𝑠𝑠 ↓, 𝐾𝑣 ↑
⟹ 𝐾𝑣 ≥ 1
iv) 𝑡𝑟 = 2𝑇
Cont’d…
Continuous Time Lead Compensator Design
Step 1: Determining the location of the desired closed-loop poles:
• From the requirements, let 𝜁 = 0.44724, 𝜎 = 1.5
Then 𝜎 = 𝜁𝜔𝑛 = 1.5
𝜎 1.5
⟹ 𝜔𝑛 = = = 3.354
𝜁 0.44724
• The desired closed-loop poles will be
𝑠𝑑 = −𝜁𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜁 2
= −𝜎 ± 𝑗𝜔𝑛 1 − 𝜁 2
= −1.5 ± 𝑗3.354 1 − 0.44724 2

= −1.5 ± 𝑗3
Step 2: Calculating the angle deficiency (𝜙):
• From angle criteria
⟹ 𝜙 − 𝜃1 + 𝜃2 = −180°
⟹ 𝜃1 = 90° + tan−1 1.5 3 = 116.57°
⟹ 𝜃2 = tan−1 3 0.5 = 80.54°
Cont’d…
⟹ 𝜙 − 𝜃1 + 𝜃2 = −180°
⟹ 𝜙 − 116.57° + 80.54° = −180°
⟹ 𝜙 − 197.11° = −180°
⟹ 𝜙 = −180° + 197.11° = 17.11°
 This angle deficiency should be corrected by the compensator. This means the angle
contributed by the compensator should be 17.11°.
⟹ 𝜙 = ∠𝐷 𝑠 |s=𝑠𝑑
Step 3: Calculating the poles and zeros of compensator:
𝑠+1 𝜏
𝐷 𝑠 = 𝐾𝑐 1 zero = −1 𝜏 , pole = −1 𝛼𝜏
𝑠+ 𝛼𝜏
• From our general design steps (refer to step 3), the angle contribution of the
compensator zero is equal to 90𝑜 .
⟹ 𝑅𝑒 𝑠𝑑 = 𝑧𝑒𝑟𝑜 ′ 𝑠 𝑜𝑓 𝐷 𝑠
⟹ −1.5 = −1 𝜏
⟹ 1 = 1.5
𝜏
Cont’d…
• The pole of the compensator can be obtained from the above s-plane map by using a
trigonometric approach.
1 1
−
𝛼𝜏 𝜏
⟹ tan 𝜙 =
3
1
𝛼𝜏
−1.5
⟹ tan 17.11° =
3
1
⟹ 3 tan 17.11° = − 1.5
𝛼𝜏
1
⟹ 1.5 + 3 tan 17.11° =
𝛼𝜏
1
⟹ = 2.4235
𝛼𝜏
• Then
𝑠 + 1.5
𝐷 𝑠 = 𝐾𝑐
𝑠 + 2.4235
Step 4: Open-loop gain of the compensator
• It can be obtained from the magnitude criteria as follows:
⟹ |𝐷 𝑠 𝐺𝐻 𝑠 |s=𝑠𝑑 = 1
Cont’d…
𝑠+1.5
⟹ |4𝐾𝑐 | =1
𝑠+2.4235 𝑠(𝑠+2) s=𝑠𝑑

(−1.5+𝑗3+1.5)
⟹ |4𝐾𝑐 | =1
−1.5+𝑗3+2.4235 (−1.5+𝑗3)(−1.5+𝑗3+2)

𝑗3
⟹ 4𝐾𝑐 =1
0.9235+𝑗3 −1.5+𝑗3 0.5+𝑗3

3
⟹ 4𝐾𝑐 =1
3.139∗3.354∗3.041

12𝐾𝑐
⟹ =1
32.016

⟹ 𝐾𝑐 = 2.668
 In general, the transfer function of the continuous time lead compensator is
𝑠 + 1.5
𝐷 𝑠 = 2.668
𝑠 + 2.4235
Discrete equivalent of the controller
• It will be done by using pole-zero mapping technique.
• Each poles and zeros are mapped according to the following relation,
𝑧 = 𝑒 𝑠𝑇
• The value of the sampling period T can be obtained from the desired specifications:
𝜔𝑛 = 3.354
1.8
⟹ 𝜔𝑛 =
𝑡𝑟
𝑏𝑢𝑡 𝑡𝑟 = 2𝑇
1.8
⟹ 𝜔𝑛 =
2𝑇
1.8 1.8
⟹ 𝑇= =
2𝜔𝑛 2∗3.354

⟹ 𝑇 = 0.268 𝑠𝑒𝑐
Cont’d…
• The discrete controller will have the following form:
𝑧 − 𝑧1
𝐷 𝑧 =𝐾
𝑧 − 𝑝1
where
𝑧1 = 𝑒 𝑠𝑇 |𝑠=−1.5 𝑝1 = 𝑒 𝑠𝑇 |𝑠=−2.4235

= 𝑒 −1.5∗0.268 = 𝑒 −2.4235∗0.268
= 0.669 = 0.522

• Then 𝐷 𝑧 will be:

𝑧 − 0.669
𝐷 𝑧 =𝐾
𝑧 − 0.522
Cont’d…
• To make the dc gain of 𝐷 𝑧 and 𝐷 𝑠 be identical, we require that
⟹ 𝑑𝑐 𝑔𝑎𝑖𝑛 = lim 𝐷 𝑧 = lim 𝐷 𝑠
𝑧→1 𝑠→0

𝑧−0.669 𝑠+1.5
⟹ lim 𝐾 = lim 2.668
𝑧→1 𝑧−0.522 𝑠→0 𝑠+2.4235

1−0.669 2.668∗1.5
⟹𝐾 =
1−0.522 2.4235

0.331 4.002
⟹𝐾 =
0.478 2.4235

4.002∗0.478
⟹𝐾=
2.4235∗0.331

⟹ 𝐾 = 2.385
In general, the discrete equivalent of the designed continuous time lead compensator is:
𝑧 − 0.669
𝐷 𝑧 = 2.385
𝑧 − 0.522
Implementing the Controller
Example 2: Give the difference equation that corresponds to the discrete transfer
function of the controller 𝐷(𝑧) which can be used for implementation in a computer.
Solution: The difference equation of the discrete time compensator is
𝑈 𝑧 𝑧−0.669
⟹𝐷 𝑧 = 𝐸 𝑧
= 2.385 𝑧−0.522

𝑈 𝑧 2.385𝑧−1.596
⟹ =
𝐸 𝑧 𝑧−0.522

⟹ (𝑧 − 0.522)𝑈(𝑧) = (2.385𝑧 − 1.596)𝐸(𝑧)

⟹ 𝑧𝑈(𝑧) − 0.522𝑈(𝑧) = 2.385𝑧𝐸(𝑧) − 1.596𝐸(𝑧)

⟹ 𝑢(𝑘 + 1) − 0.522𝑢(𝑘) = 2.385𝑒(𝑘 + 1) − 1.596𝑒(𝑘)

⟹ 𝑢 𝑘 + 1 = 0.522𝑢 𝑘 + 2.385𝑒(𝑘 + 1) − 1.596𝑒(𝑘)

 This equation can be directly evaluated by a computer.
2.2.2 Evaluation of the Design
 A block diagram of the sampled-data system with discrete controller is given below:

 The 𝒵 -transform of the continuous plant preceded by a ZOH is:

𝑧−1 −1
𝐺 𝑠
𝐺 𝑧 = 𝔷 ℒ |𝑡 = 𝑘𝑇
𝑧 𝑠
𝑧−1 4
= 𝔷 ℒ −1
𝑧 𝑠 2 (𝑠+2)
𝑧−1 2 1 1
= 𝔷 ℒ −1 − +
𝑧 𝑠2 𝑠 𝑠+2
Cont’d…
𝑧−1
= 𝔷 2𝑡 − 1 𝑡 + 𝑒 −2𝑡 |𝑡 = 𝑘𝑇
𝑧
𝑧−1
= 𝔷 2𝑘𝑇 − 1 𝑘𝑇 + 𝑒 −2𝑘𝑇
𝑧
𝑧−1 2𝑇𝑧 𝑧 𝑧
= − +
𝑧 (𝑧−1)2 𝑧−1 𝑧−𝑒 −2𝑇
2𝑇 𝑧−1
= − 1 + but T = 0.268 sec
𝑧−1 𝑧−𝑒 −2𝑇
2∗0.268 𝑧−1
= −1+
𝑧−1 𝑧−𝑒 −2∗0.268
0.536 𝑧−1
= −1+
𝑧−1 𝑧−0.585
0.536 𝑧−0.585 − 𝑧−1 𝑧−0.585 +(𝑧−1)2
=
(𝑧−1)(𝑧−0.585)
0.536𝑧−0.314−𝑧 2 +1.585𝑧−0.585+𝑧 2 −2𝑧+1
=
𝑧 2 −1.585𝑧+0.585
0.121𝑧+0.101
=
𝑧 2 −1.585𝑧+0.585
Cont’d…
 With the transfer function of the plant-plus-zero order hold and the discrete
controller, we can obtain the system difference equation and compute the step
response.
 Alternatively, it can be solved by using a computer (Computer Aided Design
Package) as the following:
% Continuous plant {G(s)} formulation
𝑁𝑝 = [0 0 4]; % Numerator coefficients of the continuous plant
𝐷𝑝 = [1 2 0]; % Denominator coefficients of the continuous plant
𝐺𝑠 = tf(𝑁𝑝, 𝐷𝑝); % Transfer function of the continuous plant
% Discrete equivalent of the continuous plant with T = 0.268
𝐺𝑧 = c2d(𝐺𝑠, 0.268); % Discrete equivalent transfer function of the plant
Cont’d…
% Continuous controller {D(s)} formulation
𝑁𝑐 = [2.668 4.002]; % Numerator coefficients of the continuous controller
𝐷𝑐 = [1 2.4235]; % Denominator coefficients of the continuous controller
𝐷𝑠 = tf(𝑁𝑐, 𝐷𝑐); % Transfer function of the continuous controller
% Discrete equivalent of the controller by using pole-zero mapping principle
𝐷𝑧 = c2d(𝐷𝑠, 0.268, ‘matched’); % Discrete equivalent transfer function of the controller
% The product of the discrete controller with the discrete equivalent of the plant is
pro = 𝐷𝑧*𝐺𝑧; or
pro = series(𝐷𝑧, 𝐺𝑧);
% Constructing the transfer function of the closed-loop system
sys = connect(pro, [1 -1]); or
sys = feedback(pro, 1);
% Step response of the closed-loop system
step(sys)
 Assignment #2
1. Design a digital controller for
1
𝐺 𝑠 =
𝑠(𝑠+0.4)
preceded by a 𝑍𝑂𝐻 so that the response has a rise time of approximately 0.5 𝑠𝑒𝑐,
overshoot < 25%, and zero steady-state error to a step command. [Hint: Cancel
the plant pole at 𝑠 = 0.4 with a compensator zero; a second-order closed-loop
system will result, making the transient response comparison between experiment
and theory much easier.]
a) Determine a 𝐷(𝑧) using emulation with the matched pole-zero mapping technique.
(Use a sample period of 𝑻 = 𝟏𝟎𝟎 𝒎𝒔𝒆𝒄.)
b) Give the difference equation that corresponds to the discrete transfer function of the
controller D(z) which can be used for implementation in a computer.
c) What is the discrete transfer function of the continuous plant preceded by a ZOH.
2.3 Stability Analysis of closed-loop systems in z-plane
 for a SISO LTI discrete time control systems, let us consider the following closed-loop
pulse transfer function system:
𝐶(𝑧) 𝐺(𝑧)
=
𝑅(𝑧) 1+𝐺𝐻(𝑧)

 The stability of this system can be determined from the locations of the closed-loop poles
in the z-plane or the roots of the characteristic equation 𝑃 𝑧 = 1 + 𝐺𝐻 𝑧 = 0 as follows:
1. For the system to be stable, the closed-loop poles or roots of the characteristic equation must
lie within the unit circle in the 𝑧-plane. Any closed-loop pole outside the unit circle makes
the system unstable.
2. If a simple pole lies at 𝑧 = 1, then the system becomes critically stable. Also, the system
becomes critically stable if a single pair of conjugate complex poles lies on the unit circle in
the 𝑧-plane. Any multiple closed-loop pole on the unit circle makes the system unstable.
3. Closed-loop zeros do not affect the absolute stability and therefore may be located anywhere
in the 𝑧-plane.
Jury Stability Test
 In applying a Jury stability test to a given characteristic equation 𝑃 𝑧 = 0, we
construct a table whose elements are based on the coefficients of 𝑃 𝑧 . Assume
that the characteristic equation 𝑃 𝑧 is a polynomial in 𝑧 as follows:
n 1
P ( z )  an z  an 1 z
n
   a1 z  a0  0, an  0
Then, the Jury table becomes as given in table below:

z0 z 1 z2 z n k z n1 zn
a0 a1 a2 an  k an 1 an
an an 1 an  2 ak a1 a0
b0 b1 b2 bn  k bn 1
bn 1 bn  2 bn 3 bk 1 b0
c0 c1 c2 cn  k
cn  k c2 c1 c0
Cont’d…
• The entries in the third row and above are calculated by:

a0 a n  k b0 bn 1k c0 cn  2  k
bk  ck  dk 
an ak bn 1 bk cn  2 ck
• The necessary and sufficient condition for stability is then:
b0  bn 1
a0  a n c0  cn  2

P(1)  0
(1) d 0  d n 3 (2)


(1) P(1)  0
n
m0  m2
Cont’d…
• Steps
i. From the given system description, obtain the characteristic equation.
ii. Apply group 1 constraints, if any one fail, then the system is unstable, stop.
iii. If group 1 constraints are valid, start filling Jury’s table and check the
condition at each row. If condition fails, the system is unstable, stop.
iv. If all constraints are fulfilled, then the system is stable.
Example: for a system having the following characteristic equation, check
stability using Jury’s test.
P ( z )  z  1.75 z  0.625 z  0.25
3 2

Solution:
 Number of rows of Jury table: 2𝑛 − 3 = 2 ∗ 3 − 3 = 3
 From the given characteristic equation: 𝑎3 = 1 , 𝑎2 = −1.75 , 𝑎1 = −0.625 , 𝑎0 = 0.25
Cont’d…
 Jury table:
Row 𝑧0 𝑧1 𝑧2 𝑧3

1 𝑎0 𝑎1 𝑎2 𝑎3

2 𝑎3 𝑎2 𝑎1 𝑎0

3 𝑏0 𝑏1 𝑏2
 Then
a0 a3 0.25 1
b0    (0.25 * 0.25)  (1*1)  0.0625  1  0.9375
a3 a0 1 0.25

a0 a2 0.25  1.75
b1    {0.25 * (0.625)}  (1.75 *1)  0.15625  1.75  1.59375
a3 a1 1  0.625

a0 a1 0.25  0.625
b2    {0.25 * (1.75)}  (0.625 *1)  0.4375  0.625  0.1875
a3 a2 1  1.75
Cont’d…
 Let us examine the condition for stability:
1) 𝑎0 < 𝑎3 ⟹ 0.25 < 1

⇝ 𝑇ℎ𝑖𝑠 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑖𝑠 𝑠𝑎𝑡𝑖𝑠𝑓𝑖𝑒𝑑.

2) 𝑃 1 > 0 ⟹ 13 − 1.75(1)2 −0.625 1 + 0.25 > 0

⟹ 1 − 1.75 − 0.625 + 0.25 > 0

⟹ −1.125 > 0
⇝ 𝑇ℎ𝑖𝑠 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛 𝑖𝑠 𝑣𝑖𝑜𝑙𝑎𝑡𝑒𝑑.

 Therefore, the system is unstable. We may stop the test here.

Cont’d…
Exercise 1: For a system having the following characteristic
equations, check stability using Jury’s test.

i) 𝑃 𝑧 = 𝑧 4 − 𝑧 3 + 2𝑧 2 + 0.5

ii) 𝑃 𝑧 = 𝑧 2 − 1.5𝑧 + 0.5

Exercise 2: For a system that is described by the following

characteristic equation, find the range of k for stability by using Jury
stability test.
𝑃 𝑧 = 𝑧 3 + 0.5𝑧 2 − 2𝑘 + 1 𝑧 − 0.5
2.4 Direct Design by using Root Locus in the z-plane
 The root locus introduced by W. Evans is based on graphical rules for
plotting the roots of a polynomial as a parameter is varied.
 The most common root locus is a plot of the roots of a closed-loop
characteristic polynomial in the s-plane as the loop gain is varied from
0 to ∞.
 In linear discrete systems also the dynamic performance is largely
determined by the roots of the closed-loop characteristic polynomial, in
this case a polynomial in 𝑧 with stability represent by having all roots
inside the unit circle.
 The consequences for direct digital design are that one can use Evans’
root locus rules unchanged, but that the performance specifications must
first be translated into the 𝑧-plane.
2.4.1 z-plane specifications
 Mapping of the s-plane to the z-plane, 𝑧 = 𝑒 𝑠𝑇
• The left half of the s-plane maps to a unit circle in the z-plane.
• The right half maps to a point out of the unit circle in the z-plane.
Cont’d…
 Generally, for a system given by transfer function:
 n2
H (s)  2
s  2n s   n2
• s-plane poles are given by:
𝑠 = −𝜉𝜔𝑛 ± 𝑗𝜔𝑛 1 − 𝜉 2
= −𝜎 ± 𝑗𝜔𝑑
• z-plane poles are:
𝑠𝑇 −𝜉𝜔𝑛 𝑇±𝑗𝜔𝑛 1−𝜉 2 𝑇
𝑧=𝑒 =𝑒
−𝜉𝜔𝑛 𝑇 ±𝑗𝜔𝑛 1−𝜉 2 𝑇
=𝑒 𝑒
= 𝑟 𝑒 ±𝑗𝜃 = 𝑟 ∠ ± 𝜃
where
𝑟 = 𝑒 −𝜉𝜔𝑛 𝑇 , 𝜃 = 𝜔𝑛 1 − 𝜉 2 𝑇 = 𝜔𝑑 𝑇
Cont’d…
 If the z-plane complex pole is given as:
z  r  
• The damping ratio, natural frequency and time constant can be obtained as:

 ln r

ln 2 r   2

1
n  ln 2 r   2
T

1 T
 
n ln r
Cont’d…
 In summary, to get the specifications on acceptable pole locations in the
z-plane:
• Estimate the desired 𝜔𝑛 , 𝜉, and 𝑀𝑝 from the continuous-time response
specifications. Compute σ = 𝜉𝜔𝑛 .
• Compute the radius r = 𝑒 −𝜉𝜔𝑛 𝑇 = 𝑒 −σ𝑇 .
• Obtain a plot of the z-plane showing lines of fixed damping and 𝜔𝑛 . The
MATLAB command zgrid will do this, plotting 𝜉 in steps of 0.1 from 0.1
to 0.9 and
𝑁𝜋
𝜔𝑛 =
10𝑇
for integer 𝑁 from 1 to 10.
• Mark the region of acceptable closed-loop pole locations on the plane.
 The specification of steady-state error:
 consider the following discrete closed-loop system.

 Proceeding as we for the continuous system, suppose the input r is a step, r 𝑘 = 1(𝑘), and
the disturbance w is zero. The transform of the error is computed using the same block-diagram
reduction tools that apply for continuous systems represented by their Laplace transforms,
except that now we use 𝐷 𝑧 and 𝐺(𝑧).
 Doing this yields the transform of the error,
𝑅(𝑧)
𝐸 𝑧 =
1+𝐷 𝑧 𝐺 𝑧
(i) Type 0 system:
𝑧 1
𝐸 𝑧 =
𝑧−1 1+𝐷 𝑧 𝐺 𝑧
Cont’d…
• The steady-state error can be obtained by using final value theorem:
𝑒 ∞ = lim 𝑧 − 1 𝐸(𝑧)
𝑧→1
𝑧 1
= lim 𝑧 − 1
𝑧→1 𝑧−1 1+𝐷 𝑧 𝐺 𝑧
𝑧
= lim
𝑧→1 1+𝐷 𝑧 𝐺 𝑧
1
=
1+𝐷 1 𝐺 1
1
≜
1+𝐾𝑝

Where 𝐾𝑝 is position-error constant of Type 0 system if the limit is finite.

𝐾𝑝 = lim 𝐷 𝑧 𝐺(𝑧) = 𝐷 1 𝐺(1)

𝑧→1

 If 𝐷 𝑧 𝐺(𝑧) has a pole at 𝑧 = 1, then the steady state error of Type 0 system for unit step
input is zero. Suppose there is a single pole at 𝑧 = 1. Then we have a Type 1 system and we
can compute the steady state error to a unit ramp input, r 𝑘𝑇 = 𝑘𝑇 1(𝑘𝑇) as follows.
Cont’d…
(ii) Type 1 system:
𝑇𝑧 1
𝐸 𝑧 = 2
𝑧−1 1+𝐷 𝑧 𝐺 𝑧
• Then, the steady-state error can be obtained by using final value theorem:
𝑒 ∞ = lim 𝑧 − 1 𝐸(𝑧)
𝑧→1
𝑇𝑧 1
= lim 𝑧 − 1
𝑧→1 𝑧−1 2 1+𝐷 𝑧 𝐺 𝑧
𝑇𝑧
= lim
𝑧→1 𝑧−1 (1+𝐷 𝑧 𝐺 𝑧 )
1
≜
𝐾𝑣

 Where 𝐾𝑣 is velocity-error constant of Type 1 discrete system.

(𝑧−1)(1+𝐷 𝑧 𝐺 𝑧 )
𝐾𝑣 = lim
𝑧→1 𝑇𝑧
Truxal’s Rule
 Because systems of Type 1 occur frequently, it is useful to observe that the
value of 𝐾𝑣 is fixed by the closed-loop poles and zeros.
• Suppose that the overall transfer function is:
𝑌(𝑧)
𝐻 𝑧 =
𝑅(𝑧)
and that 𝐻 𝑧 has poles 𝑝𝑖 and zeros 𝑧𝑖 . Then we can write:
𝑧 − 𝑧1 𝑧 − 𝑧2 𝑧 − 𝑧3 … 𝑧 − 𝑧𝑛
𝐻 𝑧 =𝐾
𝑧 − 𝑝1 𝑧 − 𝑝2 𝑧 − 𝑝3 … 𝑧 − 𝑝𝑛
• Now suppose that 𝐻 𝑧 is the closed-loop transfer function that results from a Type 1 system,
which implies that the steady-state error of this system to a step input is zero.
• Then, the error on the system be:
𝐸 𝑧 =𝑅 𝑧 −𝑌 𝑧 but 𝑌 𝑧 = 𝐻 𝑧 𝑅(𝑧)
Cont’d…
Then
𝐸 𝑧 =𝑅 𝑧 −𝐻 𝑧 𝑅 𝑧
= [1 − 𝐻 𝑧 ]𝑅 𝑧
• The steady-state error of the system for a unit step input can be obtained by using final value theorem as:
𝑒 ∞ = lim 𝑧 − 1 𝐸(𝑧)
𝑧→1
= lim 𝑧 − 1 [1 − 𝐻 𝑧 ]𝑅 𝑧
𝑧→1
𝑧
= lim 𝑧 − 1 1 − 𝐻 𝑧
𝑧→1 𝑧−1
= lim 1 − 𝐻 𝑧 𝑧
𝑧→1
= lim [1 − 𝐻 𝑧 ]
𝑧→1
• Since the system is a Type 1 system, its steady-state error for a unit step input is zero. Then
𝑒 ∞ = lim [1 − 𝐻 𝑧 ]
𝑧→1
0 = lim [1 − 𝐻 𝑧 ]
𝑧→1
0=1−𝐻 1
∴ 𝐻 1 =1
Cont’d…
• In addition to this, the steady-state error of the system for a ramp input can be expressed as the
following:
𝑒 ∞ = lim 𝑧 − 1 𝐸(𝑧)
𝑧→1
= lim 𝑧 − 1 1 − 𝐻 𝑧 𝑅 𝑧
𝑧→1
𝑇𝑧
= lim 𝑧 − 1 1 − 𝐻 𝑧
𝑧→1 𝑧−1 2
𝑇𝑧
= lim (1 − 𝐻 𝑧 )
𝑧→1 𝑧−1
1
≜
𝐾𝑣
• Therefore (omitting a factor of z in the numerator, which makes no difference in the result)
1 1−𝐻 𝑧
= lim
𝑇𝐾𝑣 𝑧→1 𝑧 − 1
1−𝐻 1 1−1 0
= = =
1−1 1−1 0
Cont’d…
Since the limit is indeterminate, we can use L’Hopital’s rule
𝑑
1 1−𝐻 𝑧
= lim 𝑑𝑧
𝑇𝐾𝑣 𝑧→1 𝑑 (𝑧 − 1)
𝑑𝑧
𝑑
= lim − 𝐻(𝑧)
𝑧→1 𝑑𝑧

However, at 𝑧 = 1, we have
𝑑 1 𝑑 1 𝑑 𝑑
ln 𝐻(𝑧) = 𝐻 𝑧 = 𝐻 𝑧 = 𝐻(𝑧)
𝑑𝑧 𝐻(𝑧) 𝑑𝑧 𝐻(1) 𝑑𝑧 𝑑𝑧

so that
1 𝑑
= lim − ln 𝐻 𝑧
𝑇𝐾𝑣 𝑧→1 𝑑𝑧
Cont’d…
𝑑 (𝑧−𝑧𝑖 )
= lim − ln 𝐾
𝑧→1 𝑑𝑧 (𝑧−𝑝𝑖 )

𝑑
= lim − ln 𝑧 − 𝑧𝑖 − ln 𝑧 − 𝑝𝑖 + ln 𝐾
𝑧→1 𝑑𝑧
𝑑
= lim ln 𝑧 − 𝑝𝑖 − ln 𝑧 − 𝑧𝑖 − ln 𝐾
𝑧→1 𝑑𝑧
1 1
= lim −
𝑧→1 𝑧−𝑝𝑖 𝑧−𝑧𝑖
𝑛 𝑛
1 1
= −
1 − 𝑝𝑖 1 − 𝑧𝑖
𝑖=1 𝑖=1
 Note:-
 The farther the poles of the closed-loop system are from 𝑧 = 1, the larger the velocity
error constant and the smaller the errors.
 Similarly, 𝐾𝑣 can be increased and the errors decreased by zeros close to 𝑧 = 1.
2.4.2 The Discrete Root Locus
 The characteristic equation of the single-loop system is:
1+𝐺 𝑧 𝐻 𝑧 =0
 The mechanics of drawing the root loci are exactly the same in the z-plane as in the s-plane.
The difference lies in the interpretation of the results because the pole locations in the z-plane
mean different things than pole locations in the s-plane when we come to interpret system
stability and dynamic response.
 Angle and magnitude conditions
 The characteristic equation can be written as:
𝐺 𝑧 𝐻 𝑧 = −1 ⟹ 𝐺 𝑧 𝐻 𝑧 ∠𝐺 𝑧 𝐻 𝑧 = 1∠180°
• Angle condition:
∠𝐺 𝑧 𝐻 𝑧 = ±180° 2𝑘 + 1 , 𝑘 = 0,1,2, …
• Magnitude condition:
𝐺 𝑧 𝐻 𝑧 =1
Cont’d…
• The value of z that fulfill both the angle and magnitude conditions are the roots of the
characteristic equation or the closed-loop poles.
• A plot of the points in the complex plane satisfying the angle condition alone is the
root locus. The roots of the characteristic equation corresponding to a given value of the
gain can be located on the root loci by the use of the magnitude condition.
 General procedures for constructing root loci:
Step 1: Obtain the characteristic equation:
1+𝐺 𝑧 𝐻 𝑧 =0
and then rearrange this equation so that the parameter of interest such as given 𝐾 appears
as the multiplying factor in the form:
𝑧+𝑧1 𝑧+𝑧2 …(𝑧+𝑧𝑚 )
1+K =0 where 𝐾 > 0
𝑧+𝑝1 𝑧+2 …(𝑧+𝑝𝑛 )
Cont’d…
• From the factored form of the closed-loop transfer function, find the transfer function of
the open-loop pulse transfer function:
𝑚
𝑖=1 𝑧 + 𝑧𝑖
𝐺 𝑧 𝐻 𝑧 =𝐾 𝑛
𝑗=1 𝑧 + 𝑝𝑗
Then locate poles and zeros of the open-loop pulse transfer function in the z-plane.
Theorem 1: A root locus starts from an open-loop poles of the pulse transfer function at
K = 0 and terminates at a finite or infinite open-loop zero at 𝐾 = ∞.
Theorem 2: The number of branches of the root locus is max(m, n), where m and n are
the number of finite zeros and poles of 𝐺 𝑧 𝐻 𝑧 respectively.
Step 2: Determine the root loci on the real axis.
o Root loci on the real axis are determined by open-loop poles and zeros lying on it.
 The conjugate complex poles and zeros of the open-loop pulse transfer function have no
effect on the location of the root loci on the real axis because the angle contribution of a
pair of conjugate complex poles or zeros is 360o on the real axis.
Cont’d…
o In constructing a root loci on the real axis, choose a test point on it.
 If the total number of real poles and zeros to the right of this test point is odd, then
this point lies on a root locus.
Step 3: Determine the asymptotes of the root loci.
• If a test point z is located far from the origin, then the angles of all the complex poles
and zeros may be considered the same. One open-loop zero and one open-loop pole
then cancel each other. Therefore, the root loci for very large values of z must be
asymptotic to straight lines whose angles are given as follows:
±180° 2𝑘 + 1
A𝑛𝑔𝑙𝑒 𝑜𝑓 𝑎𝑠𝑦𝑚𝑝𝑡𝑜𝑡𝑒𝑠 = , 𝑘 = 0,1,2, …
𝑛−𝑚

Where: 𝑛 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑖𝑛𝑖𝑡𝑒 𝑝𝑜𝑙𝑒𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑜𝑝𝑒𝑛 − 𝑙𝑜𝑜𝑝 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛

𝑚 = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑖𝑛𝑖𝑡𝑒 𝑧𝑒𝑟𝑜𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑜𝑝𝑒𝑛 − 𝑙𝑜𝑜𝑝 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
Cont’d…
• Intersection point of the asymptotes with the real axis is obtained from:
𝑝𝑜𝑙𝑒𝑠 − 𝑧𝑒𝑟𝑜𝑠
𝑧=
𝑛−𝑚
Step 4: Find the break away and break in points.
 It indicates the existence of multiple poles of the closed-loop system transfer
function.
 Break away point exists between two open-loop poles, where as the break in point
exists in between of two open-loop zeros.
 This points may lie on the real or imaginary axis or on the complex plane.
• The characteristic equation of the system 1 + 𝐺 𝑧 𝐻 𝑧 = 0 can be written as the following:
𝐵(𝑧) 𝐴(𝑧)
1+K =0 ⇒𝐾=−
𝐴(𝑧) 𝐵(𝑧)
Cont’d…
 The break away or break in point can be obtained from:
𝑑𝐾
=0
𝑑𝑧
𝑑𝐾
 If the value of 𝐾 corresponding to a roots of = 0 is positive, then the obtained roots
𝑑𝑧
are an actual break away or break in points.
𝑑𝐾
 If the value of 𝐾 corresponding to a roots of = 0 is negative, then the obtained roots
𝑑𝑧
are neither a break away nor a break in points.
Step 5: Determine the angle of departure (or angle of arrival) of the root loci.
 To sketch the root loci, we must find the direction of the root loci near to each open-loop poles
and zeros. This can be obtained by using angle condition.
Step 6: Find the points where the root loci cross the imaginary axis.
 This can be obtained by setting z = 𝑗𝑉 in the characteristic equation, equating both the real
part and the imaginary part to zero and solving for V and 𝐾 .
2.5. Frequency Response Methods
 The frequency response methods for continuous control system design were
developed from the original work of Bode (1945) on feedback-amplifier techniques.
Their attractiveness for design of continuous linear feedback systems depends on
several ideas.
1. The gain and phase curves for a rational transfer function can be easily plotted by hand.
2. If a physical realization of the system is available, the frequency response can be measured
experimentally without the necessity of having a mathematical model at all.
3. Nyquist’s stability criterion can be applied, and dynamic response specifications can be readily
interpreted in terms of gain and phase margins, which are easily seen on the plot of log gain and
phase-versus-log frequency.
4. The system error constants, mainly 𝐾𝑝 or 𝐾𝑣 , can be read directly from the low-frequency
asymptote of the gain plot.
5. The corrections to the gain and phase curves (and thus the corrections in the gain and phase
margins) introduced by a trial pole or zero of a compensator can be quickly and easily computed
using the gain curve alone.
6. The effect of pole, zero, or gain changes of a compensator on the speed of response ( which is
proportional to the crossover frequency) can be quickly and easily determined using the gain
curve alone.
2.5.1 Bilinear Transformation
 Is a transformation which transforms the z-plane to a new plane, w-plane.
• The transformation is given by:
1  (T / 2) w 2 z 1
z w
1  (T / 2) w T z 1
• By this transformation, the unit circle in z-plane transforms to the imaginary axis of w-plane.
• The stable region of the w-plane is the left half-plane.
• This enables the RH, Bode plot and other frequency domain analysis to be performed on
discrete time systems.
• On the unit circle in z-plane,

j T 2 T
ze w  j tan
T 2
Cont’d…
• Hence the frequency in s-plane and the frequency in w-plane are related by:
2 T
w  tan
T 2

Figure: Mapping from s-plane to z-plane to w-plane.

2.5.2 RH Criteria
 Is a systematic way of determining the stability of a continuous
time system without actually solving the characteristic equation.
 In digital system, RH cannot be applied directly but the following
method is used:
• Convert the 𝑧-domain transfer function to w-plane by using
Bilinear transformation.
• Determine the characteristic equation.
• Apply RH as a continuous time system.
2.5.3 Nyquist Stability Criterion
 For continuous systems, the Nyquist stability criterion seeks to determine whether there
are any zeros of the closed-loop characteristic equation
1 + 𝐾𝐷 𝑠 𝐺 𝑠 = 0
in the right half-plane. The method establishes stability by determining the excess of
zeros over poles of the characteristic equation in the right-half plane by plotting
𝐾 𝐷 𝑠 𝐺 𝑠 for s along the 𝒟 contour that encloses the entire right-hand side (unstable
region) of the s-plane as sketched in Figure.
• The specific statement of the Nyquist stability criterion for
continuous systems is
𝑍 = 𝑃 + 𝑁.
where
• 𝑍 = the number of unstable zeros (that are unstable closed-loop poles).
For stability, 𝑍 = 0.
• 𝑃 = the number of the unstable (open-loop) poles of 𝐾𝐷 𝑠 𝐺 𝑠 .
• 𝑁 = the net number of encirclements of the −1 point for the contour
evaluation of 𝐾𝐷 𝑠 𝐺 𝑠 in the same direction as that taken by s along 𝒟.
• Usually s is taken clockwise around 𝒟 and therefore clockwise encirclements
are taken as positive.
Cont’d…
• For the discrete case, the ideas are identical; the only difference is that the unstable region of
the z-plane is the outside of the unit circle and it is awkward to visualize a contour that encloses
this region. The problem can be avoided by the simple device of considering the encirclement
of the stable region and calculating the stability result from that.
• The characteristic equation of the discrete system is written as
1 + 𝐾𝐷 𝑧 𝐺 𝑧 = 0.
2.5.4 Design Specifications in the Frequency Domain
 Gain and Phase Margins:
The Nyquist plot shows the number of encirclements and thus the stability of the
closed-loop system.
 Gain and phase margins are defined so as to provide a two-point measure of how close
the Nyquist plot is to encircling the -1 point, and they are identical to the definitions
developed for continuous systems.
 Gain Margin (GM ):
• is the factor by which the gain can be increased before causing the system to go
unstable, and
• is usually the inverse of the magnitude of 𝐷(𝑧)𝐺(𝑧) when its phase is 180.
 Phase Margin (PM ):
• is a measure of how much additional phase lag or time delay can be tolerated in the
loop before instability results.
• is the difference between -180 and the phase of 𝐷(𝑧)𝐺(𝑧) when its amplitude is 1.
2.5.5 Representation Mechanisms of Frequency Pulse
Transfer Function
• Root locus method:- consider a closed loop system shown below:

• The characteristic equation of the system is:

1+KGH(z) = 0
• The root locus of this system will be drawn using the standard steps of
continuous system.
Cont’d…
• Example: Draw the root locus of the following system if 𝐷 𝑧 = 𝐾.

• The characteristic equation of the system is:

1  KG ( z )  0
where
z 1  1 
G( z)   2 
z  s  s  1 
Cont’d…
• Using z-transform table and T = 1, the final characteristic equation becomes:
K (0.3679 z  0.2642)
1  KG ( z ) 
( z  1)( z  0.3679)

• The root locus becomes:

Root Locus
1.5

0.5
Imaginary Axis

-0.5

-1

-1.5
-5 -4 -3 -2 -1 0 1 2
Real Axis
Cont’d…
• Bode plot:
 Is another frequency response technique.
 Graphical method which draws the magnitude and phase of the
closed-loop response as a function of frequency.
 For discrete time systems two methods can be followed:
 Form the characteristic equation of the system in z.

 Change to w-plane and plot bode plot or

 Draw the bode plot in z-domain and convert it to s-domain using
the conversion formula.
2.6 Direct Design Method of Ragazzini
• Suppose we are given a discrete transfer function 𝐺 𝑧 of the plant (plus hold) and a desired
transfer function H(𝑧) between reference 𝑅(𝑧) and output 𝑌(𝑧). The structure is assumed to be
a unity feedback system and the design is to select the computer transfer function D(𝑧) to
realize 𝐻(𝑧).
• The overall transfer function is given by the formula:

𝐷(𝑧)𝐺(𝑧)
H z =
1 + 𝐷(𝑧)𝐺(𝑧)
• From which we get the direct design formula:
𝐻(𝑧)
D z =
𝐺 𝑧 {1 − 𝐻 𝑧 }
 This design calls for a 𝐷(𝑧) that will cancel the plant effects and add whatever is necessary to give the
desired result.
 The problem is to discover and implement constraints on 𝐻(𝑧).
Consideration on constraints
i. The design must be causal.
 If 𝐷 𝑧 is to be causal, then as 𝑧 → ∞ its transfer function is bounded and does not have a pole at
infinity. If 𝐺 𝑧 were to have a zero at infinity, then 𝐷 𝑧 would have a pole there. This pole should be
cancelled by 𝐻 𝑧 .
Thus the constraint for 𝐷 𝑧 to be causal is:
𝐻 𝑧 𝑚𝑢𝑠𝑡 ℎ𝑎𝑣𝑒 𝑎 𝑧𝑒𝑟𝑜 𝑎𝑡 𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑦 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑜𝑟𝑑𝑒𝑟 𝑎𝑠 𝑡ℎ𝑒 𝑧𝑒𝑟𝑜 𝑜𝑓 𝐺 𝑧 𝑎𝑡 𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑦.

ii. The roots of the characteristic equation of the closed-loop system are the roots
of equation:
1+𝐷 𝑧 𝐺 𝑧 =0
𝑐 𝑧 𝑏(𝑧)
• Let 𝐷 𝑧 = , 𝐺(𝑧) = 𝑤ℎ𝑒𝑟𝑒 𝑎, 𝑏, 𝑐, 𝑎𝑛𝑑 𝑑 𝑎𝑟𝑒 𝑝𝑜𝑙𝑦𝑛𝑜𝑚𝑖𝑎𝑙𝑠.
𝑑 𝑧 𝑎(𝑧)

• Then the characteristic polynomial is

𝑐𝑏
1+ =0
𝑑𝑎
𝑎𝑑 + 𝑏𝑐 = 0
Cont’d…
• Now suppose there is a common factor in 𝐷 𝑧 𝐺(𝑧), as would result if 𝐷 𝑧 were
called upon to cancel a pole or zero of 𝐺 𝑧 .
• Let this factor be 𝑧 − 𝛼 and suppose it is a pole of 𝐺(𝑧), so we can write
𝑎 𝑧 = (𝑧 − 𝛼)𝑎(z),
and to cancel it we have
c 𝑧 = (𝑧 − 𝛼)𝑐(z)
Then the characteristic polynomial becomes
𝑧 − 𝛼 𝑎(z) 𝑑 𝑧 + 𝑏 𝑧 (𝑧 − 𝛼)𝑐(z) = 0
𝑧 − 𝛼 [𝑎𝑑 + 𝑏𝑐] = 0
 This common factor remains a factor of the characteristic polynomial.
 If this factor is outside a unit circle, the system is unstable. So this cancellation
should be avoided.
Cont’d…
• To avoid this cancellation:
 If 𝐷 𝑧 is not to cancel a pole of 𝐺 𝑧 , then that factor of 𝑎 𝑧 must also be a factor of 1 − 𝐻 𝑧 .

 If 𝐷 𝑧 is not to cancel a zero of 𝐺 𝑧 , then that factor of 𝑏 𝑧 must also be a factor of 𝐻 𝑧 .

 Thus the constraints for stability:

• 1 − 𝐻 𝑧 𝑚𝑢𝑠𝑡 𝑐𝑜𝑛𝑡𝑎𝑖𝑛 𝑎𝑠 𝑧𝑒𝑟𝑜𝑠 𝑎𝑙𝑙 𝑡ℎ𝑒 𝑝𝑜𝑙𝑒𝑠 𝑜𝑓 𝐺 𝑧 𝑡ℎ𝑎𝑡 𝑎𝑟𝑒 𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑐𝑖𝑟𝑐𝑙𝑒.

• 𝐻 𝑧 𝑚𝑢𝑠𝑡 𝑐𝑜𝑛𝑡𝑎𝑖𝑛 𝑎𝑠 𝑧𝑒𝑟𝑜𝑠 𝑎𝑙𝑙 𝑡ℎ𝑒 𝑧𝑒𝑟𝑜𝑠 𝑜𝑓 𝐺 𝑧 𝑡ℎ𝑎𝑡 𝑎𝑟𝑒 𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑐𝑖𝑟𝑐𝑙𝑒.

iii. The constraint of steady-state accuracy:

• Because 𝐻 𝑧 is the overall transfer function, the error transform is given by:
𝐸 𝑧 = {(1 − 𝐻 𝑧 }𝑅 𝑧
Cont’d…
• If the system is to be Type 1 with velocity error constant 𝐾𝑣 , then
a) The steady-state error to a unit step input = 0.
b) The steady-state error to a ramp input = 1/𝐾𝑣 .
for (a):
𝑧
𝑒 ∞ = lim(𝑧 − 1) 1 − 𝐻 𝑧 =0
𝑧→1 𝑧−1

which implies
𝐻 1 =1
for (b): The velocity error constant requirement is that:
𝑇𝑧 1
𝑒 ∞ = lim (𝑧 − 1) 1 − 𝐻 𝑧 =
𝑧→1 (𝑧 − 1)2 𝐾𝑣

1 𝑇𝑧 1 − 𝐻 𝑧
= lim
𝐾𝑣 𝑧→1 𝑧−1
Cont’d…
 We know that 1 − 𝐻 𝑧 is zero at 𝑧 = 1, so that to evaluate the limit, it is necessary to
use L’Hopital’s rule with the result:
𝑑
1 {𝑇𝑧 1 − 𝐻 𝑧 }
= 𝑑𝑧 |𝑧 = 1
𝐾𝑣 𝑑
(𝑧 − 1)
𝑑𝑧
𝑑
𝑇 1−𝐻 𝑧 −𝑇𝑧𝑑𝑧𝐻 𝑧
= |𝑧 = 1
1
𝑑
= 𝑇 1 − 𝐻(1) − 𝑇 𝐻 𝑧 |𝑧 = 1
𝑑𝑧
𝑑
= 𝑇 1−1 − 𝑇 𝐻 𝑧 |𝑧 = 1
𝑑𝑧
𝑑
= −𝑇 𝐻 𝑧 |𝑧 = 1
𝑑𝑧
 Assignment #3
1. For a system that is described by the following characteristic equation, find the
range of k for stability by using Jury stability test.

𝑃 𝑧 = 𝑧 3 + 0.5𝑧 2 − 2𝑘 + 1 𝑧 − 0.5

2. Draw the root locus of a negative feedback system with an open-loop transfer
function of
𝐾
𝐺𝐻 𝑧 =
𝑧(𝑧 + 3)2
3. Find the Bode plot of
𝑤−2
𝐺 𝑤 =
(𝑤 + 4)(𝑤 + 8)