Chapter 4: Stability of Digital

Control Systems

Dr Majd Saied 1
Chapter Objectives
After completing this chapter, you will be able to do the following:
1. Determine the input-output stability of a z-transfer function.
2. Determine the asymptotic stability of a z-transfer function.
3. Determine the internal stability of a digital feedback control system.
4. Determine the stability of a z-polynomial using the Routh-Hurwitz criterion.
5. Determine the stable range of a parameter for a z-polynomial.
6. Determine the closed-loop stability of a digital system using the Nyquist
criterion.
7. Determine the gain margin and phase margin of a digital system.

Dr Majd Saied 2
 4.1. Stable z-domain pole locations
𝑧
Consider the sampled exponential and its 𝑧-transform: 𝑝𝑘 , 𝑘 = 0,1,2, … ⟷ 𝑧−𝑝
where 𝑝 is real or complex. Then the time sequence for large 𝑘 is given by:
0, 𝑝 <1
|𝑝|𝑘 ⟶ ൞ 1, 𝑝 =1
∞, 𝑝 >1
Any time sequence can be described by:
𝑛 𝑛
𝑧
𝑓 𝑘 = ෍ 𝐴𝑖 𝑝𝑖𝑘 , 𝑘 = 0,1,2, … ⟷ 𝐹 𝑧 = ෍ 𝐴𝑖
𝑧 − 𝑝𝑖
𝑖=1 𝑖=1
𝐴𝑖 : partial fraction coefficients, 𝑝𝑖 : z-domain poles. Hence, we conclude that the sequence is
bounded if its poles lie in the closed unit disc (i.e., on or inside the unit circle) and decays
exponentially if its poles lie in the open unit disc (i.e., inside the unit circle).

This conclusion allows us to derive stability conditions based on the locations of the system
poles.

Dr Majd Saied 3
 4.2. Stability Conditions
Asymptotic stability
Theorem 4.1 gives conditions for asymptotic stability.

THEOREM 4.1: ASYMPTOTIC STABILITY

In the absence of pole-zero cancellation, an LTI digital system is asymptotically
stable if its transfer function poles are in the open unit disc and marginally stable if
the poles are in the closed unit disc with no repeated poles on the unit circle.

Theorem 4.1 applies even if pole-zero cancellation occurs, provided that the poles
that cancel are stable. This follows from the fact that stability testing is essentially a
search for unstable poles, with the system being declared stable if none are found.

Dr Majd Saied 4
 4.2. Stability Conditions
Example:
Determine the asymptotic stability of the following systems:
4(𝑧−2)
a) 𝐻 𝑧 = (𝑧−2)(𝑧−0.1)
4(𝑧−0.2)
b) 𝐻 𝑧 = (𝑧−0.2)(𝑧−0.1)
5(𝑧−0.3)
c) 𝐻 𝑧 = (𝑧−0.2)(𝑧−0.1)
8(𝑧−0.2)
d) 𝐻 𝑧 = (𝑧−0.1)(𝑧−1)

Solution
Theorem 4.1 can only be used for transfer functions (a) and (b) if their poles and zeros are
not canceled. Ignoring the zeros, which do not affect the response to the initial conditions,
(a) has a pole outside the unit circle and the poles of (b) are inside the unit circle. Hence, (a)
is unstable, whereas (b) is asymptotically stable.
Theorem 4.1 can be applied to the transfer functions (c) and (d). The poles of (c) are all
inside the unit circle, and the system is therefore asymptotically stable. However, (d) has one
Dr Majd
pole on the unit circle and is only marginally Saied
stable. 5
 4.2. Stability Conditions
BIBO stability
BIBO stability concerns the response of a system to a bounded input.
Theorem 4.2 establishes necessary and sufficient conditions for BIBO stability of a
discrete-time linear system.

THEOREM 4.2:
A discrete-time linear system is BIBO stable if and only if its impulse response
sequence is absolutely summable—that is,
∞

෍ |ℎ(𝑖)| < ∞
𝑖=0
Because the z-transform of the impulse response is the transfer function, BIBO
stability can be related to pole locations as shown in Theorem 4.3.

THEOREM 4.3:
A discrete-time linear system is BIBO stable if and only if the poles of its transfer
Dr Majd Saied 6
function lie inside the unit circle.
 4.2. Stability Conditions
Example:
Investigate the BIBO stability of the class of systems with the impulse response
𝐴, 0≤𝑘≤𝑚<∞
ℎ 𝑘 =ቊ
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
where 𝐴 is a finite constant.

Solution
The impulse response satisfies
∞ 𝑚

෍ |ℎ(𝑖)| = ෍ |ℎ(𝑖)| = 𝑚 + 1 𝐴 < ∞

𝑖=0 𝑖=0
These systems are all BIBO stable.

Dr Majd Saied 7
 4.2. Stability Conditions
Example:
Investigate the BIBO stability of the systems.
4(𝑧−2)
a) 𝐻 𝑧 = (𝑧−2)(𝑧−0.1)
4(𝑧−0.2)
b) 𝐻 𝑧 = (𝑧−0.2)(𝑧−0.1)
5(𝑧−0.3)
c) 𝐻 𝑧 = (𝑧−0.2)(𝑧−0.1)
8(𝑧−0.2)
d) 𝐻 𝑧 = (𝑧−0.1)(𝑧−1)

Solution
After pole-zero cancellation, the transfer functions (a) and (b) have all poles inside
the unit circle and are therefore BIBO stable. The transfer function (c) has all poles
inside the unit circle and is stable; (d) has a pole on the unit circle and is not BIBO
stable.

Dr Majd Saied 8
 4.2. Stability Conditions
The preceding analysis and examples show that for LTI systems, with no pole-zero
cancellation, BIBO and asymptotic stability are equivalent and can be investigated
using the same tests.

Hence, the term stability is used in the sequel to denote either BIBO or asymptotic
stability with the assumption of no unstable pole-zero cancellation. Pole locations
for a stable system (inside the unit circle) are shown below.

Dr Majd Saied 9
 4.2. Stability Conditions
Internal Stability
• So far, we have only considered stability as applied to an open-loop system.
• For closed-loop systems, these results are applicable to the closed-loop transfer
function. However, the stability of the closed-loop transfer function is not always
sufficient for proper system operation because some of the internal variables
may be unbounded.
• In a feedback control system, it is essential that all the signals in the loop be
bounded when bounded exogenous inputs are applied to the system.
• Consider the unity feedback digital control scheme where, for simplicity, a
disturbances input is added to the controller output before the ADC. We
consider that system as having two outputs, 𝑌 and 𝑈, and two inputs, 𝑅 and 𝐷.

Dr Majd Saied 10
 4.2. Stability Conditions
Internal Stability
Thus, the transfer functions associated with the system are given by:
𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 𝐺𝑍𝐴𝑆 (𝑧)
𝑌(𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 𝑅(𝑧)
=
𝑈(𝑧) 𝐶(𝑧) 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 𝐷(𝑧)
−
1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧)
Clearly, it is not sufficient to prove that the output of the controlled system Y is
bounded for bounded reference input R because the controller output U can be
unbounded.

In addition, the system output must be bounded when a different input is applied
to the system—namely, in the presence of a disturbance.

Dr Majd Saied 11
 4.2. Stability Conditions
Internal Stability

DEFINITION 4.3: INTERNAL STABILITY

If all the transfer functions that relate system inputs (R and D) to the possible
system outputs (Y and U) are BIBO stable, then the system is said to be internally
stable.

Because internal stability guarantees the stability of the transfer function from 𝑅 to
𝑌, among others, it is obvious that an internally stable system is also externally
stable (i.e., the system output 𝑌 is bounded when the reference input 𝑅 is
bounded). However, external stability does not, in general, imply internal stability.

THEOREM 4.4
The system with feedback system is internally stable if and only if all its closed-loop
poles are in the open unit disc.
Dr Majd Saied 12
 4.2. Stability Conditions
Internal Stability

THEOREM 4.5
The system is internally stable if and only if the following two conditions hold:
1. The characteristic polynomial 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) has no zeros on or outside the
unit circle.
2. The loop gain 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) has no pole-zero cancellation on or outside the unit
circle.

Dr Majd Saied 13
 4.2. Stability Conditions
Example:
An isothermal chemical reactor where the product concentration is controlled by
manipulating the feed flow rate is modeled by the following transfer function:
0.5848(−0.3549𝑠 + 1)
𝐺 𝑠 =
0.1828𝑠 2 + 0.8627 + 1
Determine 𝐺𝑍𝐴𝑆 (𝑧) with a sampling rate 𝑇 = 0.1, and then verify that the closed-
loop system with the feedback controller
−10(𝑧 − 0.8149)(𝑧 − 0.7655)
𝐶 𝑧 =
(𝑧 − 1)(𝑧 − 1.334)
is not internally stable.

Solution:
The discretized process transfer function is:
−1
𝐺 𝑠 −0.075997(𝑧 − 1.334)
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 Ζ =
𝑠 (𝑧 − 0.8149)(𝑧 − 0.7655)

Dr Majd Saied 14
 4.2. Stability Conditions
Solution:
The transfer function from the reference input to the system output is given by:
𝑌(𝑧) 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 0.75997
= =
𝑅(𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 𝑧 − 0.24
The system appears to be asymptotically stable with all its poles inside the unit
circle. However, the system is not internally stable as seen by examining the
transfer function
𝑈(𝑧) 𝐶(𝑧) −10(𝑧 − 0.8149)(𝑧 − 0.7655)
= =
𝑅(𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) (𝑧 − 0.24)(𝑧 − 1.334)
which has a pole at 1.334 outside the unit circle. The control variable is unbounded
even when the reference input is bounded.

Dr Majd Saied 15
 4.3. Routh-Hurwitz Criterion
• The Routh-Hurwitz criterion determines conditions for left half plane (LHP)
polynomial roots and cannot be directly used to investigate the stability of
discrete-time systems.
• The bilinear transformation:
1+𝑤 𝑧−1
𝑧= ⇔𝑤=
1−𝑤 𝑧+1
transforms the inside of the unit circle to the LHP.
• To verify this property, consider the three cases shown in the figure below. They
represent the mapping of a point in the LHP, a point in the RHP, and a point on
the 𝑗𝑤 axis.

Dr Majd Saied 16
 4.3. Routh-Hurwitz Criterion
• The angle of 𝑤 after bilinear transformation is:
∠ 𝑤 = ∠ 𝑧 − 1 − ∠ (𝑧 + 1)
For a point inside the unit circle, the angle of 𝑤 is of a magnitude greater than 90°,
which corresponds to points in the LHP.

For a point on the unit circle, the angle is ±90°, which corresponds to points on the
imaginary axis, and for points outside the unit circle, the magnitude of the angle is
less than 90°, which corresponds to points in the RHP.

The bilinear transformation allows the use of the Routh-Hurwitz criterion for the
investigation of discrete-time system stability. For the general 𝑧-polynomial, we
have the transform pairs:
1+𝑤 𝑛 𝑛−1
𝑧= 1+𝑤 1+𝑤
1−𝑤
𝐹 𝑧 = 𝑎𝑛 𝑧 𝑛 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎0 𝑎𝑛 + 𝑎𝑛−1 + ⋯ + 𝑎0
1−𝑤 1−𝑤

Dr Majd Saied 17
 4.3. Routh-Hurwitz Criterion
Example:
Find stability conditions for:
1. The first-order polynomial: 𝑎1 𝑧 + 𝑎0 , 𝑎1 > 0
2. The second-order polynomial: 𝑎2 𝑧 2 + 𝑎1 𝑧 + 𝑎0 , 𝑎2 > 0

Solution
1. The stability of the first-order polynomial can be easily determined by solving for its root.
Hence, the stability condition is:
𝑎0
<1
𝑎1
2. Substituting the bilinear transformation in the second-order polynomial gives:
2
1+𝑤 1+𝑤
𝑎2 + 𝑎1 + 𝑎0
1−𝑤 1−𝑤
which reduces to:
𝑎2 − 𝑎1 + 𝑎0 𝑤 2 + 2 𝑎2 − 𝑎0 𝑤 + (𝑎2 + 𝑎1 + 𝑎0 )
By the Routh-Hurwitz criterion, it can be shown that the poles of the second-order w-
polynomial remain in the LHP if and only if its coefficients are all positive.
Dr Majd Saied 18
 4.3. Routh-Hurwitz Criterion
Solution
Hence, the stability conditions are given by:

𝑎2 − 𝑎1 + 𝑎0 > 0
2 𝑎2 − 𝑎0 > 0
𝑎2 + 𝑎1 + 𝑎0 > 0

Without loss of generality, the coefficient 𝑎2 can be assumed to be unity, and the stable
parameter range can be depicted in the 𝑎0 versus 𝑎1 parameter plane.

Dr Majd Saied 19
 4.4. Jury Test
It is possible to investigate the stability of 𝑧-domain polynomials directly using the
Jury test for real coefficients or the Schur-Cohn test for complex coefficients. These
tests involve determinant evaluations as in the Routh-Hurwitz test for s-domain
polynomials but are more time consuming. The Jury test is given next.

THEOREM 4.6
For the polynomial
𝐹 𝑧 = 𝑎𝑛 𝑧 𝑛 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0 = 0
the roots of the polynomial are inside the unit circle if and only if:
(1) 𝐹 1 >0
(2) −1 𝑛 𝐹 −1 > 0
(3) 𝑎0 < 𝑎𝑛
(4) 𝑏0 > |𝑏𝑛−1 |
(5) 𝑐0 > |𝑐𝑛−2 |
⋮
𝑛+1 𝑟0 > |𝑟2 |
where the terms in the (𝑛 + 1) conditions are calculated from the Jury Table.
Dr Majd Saied 20
4.4. Jury Test

Dr Majd Saied 21
 4.4. Jury Test
The entries of the table are calculated as follows:

𝑎0 𝑎𝑛−𝑘
𝑏𝑘 = 𝑎 𝑎𝑘 , 𝑘 = 0,1, … , 𝑛 − 1
𝑛

𝑏0 𝑏𝑛−𝑘−1
𝑐𝑘 = , 𝑘 = 0,1, … , 𝑛 − 2
𝑏𝑛−1 𝑏𝑘
⋮
⋮
𝑠0 𝑠3 𝑠0 𝑠2 𝑠0 𝑠1
𝑟0 = 𝑠 𝑠0 , 𝑟1 = 𝑠 𝑠1 , 𝑟2 = 𝑠 𝑠2
3 3 3

Based on the Jury table and the Jury stability conditions, we make the following
observations:
• The first row of the Jury table is a listing of the coefficients of the polynomial
𝐹(𝑧) in order of increasing power of 𝑧.
• The number of rows of the table 2𝑛 − 3 is always odd, and the coefficients of
each even row are the same as the odd row directly above it with the order of
the coefficients reversed. Dr Majd Saied 22
 4.4. Jury Test
• There are 𝑛 + 1 conditions that correspond to the 𝑛 + 1 coefficients of 𝐹(𝑧).
• Conditions 3 through 𝑛 + 1 of are calculated using the coefficient of the first
column of the Jury table together with the last coefficient of the preceding row.
The middle coefficient of the last row is never used and need not be calculated.
• Conditions 1 and 2 are calculated from 𝐹(𝑧) directly. If one of the first two
conditions is violated, we conclude that 𝐹(𝑧) has roots on or outside the unit
circle without the need to construct the Jury table or test the remaining
conditions.

Dr Majd Saied 23
 4.4. Jury Test
Example:
Test the stability of the polynomial.

𝐹 𝑧 = 𝑧 5 + 2.6𝑧 4 − 0.56𝑧 3 − 2.05𝑧 2 + 0.0775𝑧 + 0.35 = 0

We compute the entries of the Jury table using the coefficients of the polynomial.

Dr Majd Saied 24
 4.4. Jury Test
Solution:
The first two conditions require the evaluation of 𝐹(𝑧) at 𝑧 = ±1
1) 𝐹 1 = 1 + 2.6 − 0.56 − 2.05 + 0.0775 + 0.35 = 1.4175 > 0
2) −1 5 𝐹 −1 = −1 −1 + 2.6 + 0.56 − 2.05 − 0.0775 + 0.35 = −0.3825 < 0

Conditions 3 through 6 can be checked quickly using the entries of the first column of the
Jury table.

3) 0.35 < |1|

4) 0.8775 > |0.8325|
5) 0.0770 < 0.5151
6) −0.2593 < −0.3472

Conditions 2, 5, and 6 are violated, and the polynomial has roots on or outside the unit
circle. In fact, the polynomial can be factored as:
𝐹 𝑧 = 𝑧 − 0.7 𝑧 − 0.5 𝑧 + 0.5 𝑧 + 0.8 𝑧 + 2.5 = 0
and has a root at −2.5 outside the unit circle. Note that the number of conditions violated is
not equal to the number of roots outsideDrthe unit circle and that condition 2 is sufficient
Majd Saied 25
to
conclude the instability of 𝐹(𝑧).
 4.4. Jury Test
Example
Find the stable range of the gain 𝐾 for the unity feedback digital cruise control system with
the analog plant transfer function:
𝐾
𝐺 𝑠 =
𝑠+3
and with digital-to-analog converter (DAC) and analog-to-digital converter (ADC) if the
sampling period is 0.02 𝑠.

Solution
The transfer function for analog subsystem ADC and DAC is:
𝐺(𝑠)
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 −1 Ζ
𝑠
Using the partial fraction expansion:
𝐺(𝑠) 𝐾 𝐾 1 1
= = −
𝑠 𝑠 𝑠+3 3 𝑠 𝑠+3
we obtain the transfer function:
1.9412 × 10−2 𝐾
𝐺𝑍𝐴𝑆 𝑧 =
𝑧 − 0.9418
Dr Majd Saied 26
 4.4. Jury Test
Solution
For unity feedback, the closed-loop characteristic equation is:

1 + 𝐺𝑍𝐴𝑆 𝑧 = 0

which can be simplified to:

𝑧 − 0.9418 + 1.9412 × 10−2 𝐾 = 0

The stability conditions are:

𝐹 1 > 0 ⇒ 1 − 0.9418 + 1.9412 × 10−2 𝐾 > 0
−1 1 𝐹 −1 > 0 ⇒ − −1 − 0.9418 + 1.9412 × 10−2 𝐾 > 0

Thus, the stable range of 𝐾 is:

−3 < 𝐾 < 100.03

Dr Majd Saied 27
 4.4. Jury Test
Example
Find the stable range of the gain 𝐾 for the vehicle position control system with the analog
plant transfer function:
10 𝐾
𝐺 𝑠 =
𝑠(𝑠 + 10)
and with digital-to-analog converter (DAC) and analog-to-digital converter (ADC) if the
sampling period is 0.05 𝑠.

Solution
The transfer function for analog subsystem ADC and DAC is:
𝐺(𝑠)
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 −1 Ζ
𝑠
Using the partial fraction expansion:
𝐺(𝑠) 10𝐾 10 1 1
= 2 = 0.1𝐾 2 − +
𝑠 𝑠 𝑠 + 10 𝑠 𝑠 𝑠 + 10
we obtain the transfer function:
1.0653 × 10−2 𝐾(𝑧 + 0.8467)
𝐺𝑍𝐴𝑆 𝑧 =
(𝑧 − 1)(𝑧 − 0.6065)
Dr Majd Saied 28
 4.4. Jury Test
Solution
For unity feedback, the closed-loop characteristic equation is:

1 + 𝐺𝑍𝐴𝑆 𝑧 = 0

which can be simplified to:

𝑧 − 1 𝑧 − 0.6065 + 1.0653 × 10−2 𝐾 𝑧 + 0.8467 = 0
⇒ 𝑧 2 + 1.0653 × 10−2 𝐾 − 1.6065 𝑧 + 0.6065 + 9.02 × 10−3 𝐾 = 0

The stability conditions are:

𝐹 1 > 0 ⇒ 1 + 1.0653 × 10−2 𝐾 − 1.6065 + 0.6065 + 9.02 × 10−3 𝐾 > 1 → 𝐾 > 0

−1 2 𝐹 −1 > 0 ⇒ 1 − 1.0653 × 10−2 𝐾 − 1.6065 + 0.6065 + 9.02 × 10−3 𝐾 > 0

⟶ 𝐾 < 1967.582

0.6065 + 0.0902𝐾 < 1 ⇒ −1 < 0.6065 + 0.0902𝐾 < 1 ⇒ −178.104 < 𝐾

< 43.6199

Thus, the stable range of 𝐾 is:

0 < Dr
𝐾 Majd Saied
< 43.6199 29
Problems
Problem 1:
Determine the asymptotic stability and the BIBO stability of the following systems:
a) 𝑦 𝑘 + 2 + 0.8𝑦 𝑘 + 1 + 0.07𝑦 𝑘 = 2𝑢 𝑘 + 1 + 0.2𝑢 𝑘 𝑘 = 0,1,2, …
b) 𝑦 𝑘 + 2 − 0.8𝑦 𝑘 + 1 + 0.07𝑦 𝑘 = 2𝑢 𝑘 + 1 + 0.2𝑢 𝑘 𝑘 = 0,1,2, …
c) 𝑦 𝑘 + 2 + 0.1𝑦 𝑘 + 1 + 0.9𝑦 𝑘 = 3.0 𝑢 𝑘 𝑘 = 0,1,2, …

Solution:
To determine the asymptotic stability or BIBO stability, we z-transform to obtain the transfer
function. The system is both BIBO stable and asymptotically stable if its poles are all inside
the unit circle (without transfer function reduction).
2𝑧 + 0.2 2(𝑧 + 0.1)
𝑎) 𝐺 𝑧 = 2 =
𝑧 + 0.8𝑧 + 0.07 (𝑧 + 0.7)(𝑧 + 0.1)
Stable: 0.7 < 1, 0.1 < 1

2𝑧 + 0.2 2(𝑧 + 0.1)

𝑏) 𝐺 𝑧 = =
𝑧 2 − 0.8𝑧 + 0.07 (𝑧 − 0.7)(𝑧 + 0.1)
Stable: −0.7 < 1, 0.1 < 1
Dr Majd Saied 30
Problems
Solution:
3 3
𝑐) 𝐺 𝑧 = 2 =
𝑧 + 0.1𝑧 + 0.9 (𝑧 + 0.05)2 +0.9 − 0.052
Stable: Complex conjugate poles with square magnitude |0.9 | < 1

Dr Majd Saied 31
Problems
Problem 2:
Biochemical reactors are used in different processes such as waste treatment and alcohol
fermentation. By considering the dilution rate as the manipulated variable and the biomass
concentration as the measured output, the biochemical reactor can be modeled by the
following transfer function in the vicinity of an unstable steady-state operating point:
5.8644
𝐺 𝑠 =
−5.888𝑠 + 1
Determine 𝐺𝑍𝐴𝑆 (𝑧) with a sampling rate 𝑇 = 0.1 and then consider the feedback controller
𝑧 − 1.017
𝐶 𝑧 =−
𝑧−1
Verify that the resulting feedback system is not internally stable.

Solution:
The discretized process transfer function is:
𝐺 𝑠 −0.1004
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 −1 Ζ =
𝑠 𝑧 − 1.017
Although the transfer function from the reference input to the system output:
𝑌(𝑧) 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) 0.1004
= =
𝑅(𝑧) 1 + 𝐶(𝑧)𝐺 (𝑧) 𝑧 − 0.8997
𝑍𝐴𝑆Saied
Dr Majd 32
Problems
appears to be asymptotically stable (note that the pole is inside the unit circle), the controller
results in pole-zero cancellation outside the unit circle and therefore the system is not
internally stable according to Theorem 4-5. This can also be verified by considering the
transfer function between the disturbance input 𝐷 and the system output 𝑌:

𝑌(𝑧) 𝐺𝑍𝐴𝑆 (𝑧) −0.10045(𝑧 − 1)

= =
𝐷(𝑧) 1 + 𝐶(𝑧)𝐺𝑍𝐴𝑆 (𝑧) (𝑧 − 1.017)(𝑧 − 0.0997)

The pole at 1.017 is outside the unit circle and therefore the system output is unbounded
when a disturbance occurs in the system.

Dr Majd Saied 33
Problems
Problem 3:
Use the Routh-Hurwitz criterion to investigate the stability of the systems:
5(𝑧 − 2)
𝑎) 𝐺 𝑧 =
(𝑧 − 0.1)(𝑧 − 0.8)
10 𝑧 + 0.1
𝑏) 𝐺 𝑧 =
(𝑧 − 0.7)(𝑧 − 0.9)
Solution:
The Routh-Hurwitz criterion requires the substitution
1+𝑤
𝑧=
1−𝑤
The systems investigated are both second order and the following conditions can be used.
For the second order polynomial:

𝐹 𝑧 = 𝑎2 𝑧 2 + 𝑎1 𝑧 + 𝑎0 = 0

The stability conditions are:

𝑖) 𝑎2 − 𝑎1 + 𝑎0 > 0
𝑖𝑖) 𝑎2 − 𝑎0 > 0
𝑖𝑖𝑖) 𝑎2 + 𝑎1 + 𝑎0 > 0 Dr Majd Saied 34
Problems
Solution:
5(𝑧 − 2) 5(𝑧 − 2)
𝑎) 𝐺 𝑧 = = 2
(𝑧 − 0.1)(𝑧 − 0.8) 𝑧 − 0.9𝑧 + 0.08
𝑖) 1 − (−0.9) + 0.08 > 0
𝑖𝑖) 1 − 0.08 > 0
𝑖𝑖𝑖) 1 + (−0.9) + 0.08 > 0
Hence, the system is stable. This is obvious since its poles are inside the unit circle.

10(𝑧+1) 10(𝑧+1)
b) 𝐺 𝑧 = (𝑧−0.7)(𝑧−0.9) = 𝑧 2 −1.6𝑧+0.63
𝑖) 1 − (−1.6) + 0.63 > 0
𝑖𝑖) 1 − 0.63 > 0
𝑖𝑖𝑖) 1 + (−1.6) + 0.63 > 0
Hence, the system is stable. This is obvious since its poles are inside the unit circle.

Dr Majd Saied 35
Problems
Problem 4:
Repeat Problem 3 using the Jury Test.

Solution
The conditions for the Jury criterion are:
𝐹 1 >0
𝐹 −1 > 0
𝑎2 > |𝑎0 |

5(𝑧 − 2) 5(𝑧 − 2)
𝑎) 𝐺 𝑧 = =
(𝑧 − 0.1)(𝑧 − 0.8) 𝑧 2 − 0.9𝑧 + 0.08
𝑖) 1 + −0.9 + 0.08 > 0
𝑖𝑖) 1 − −0.9 + 0.08 > 0
𝑖𝑖𝑖) 1 > 0.08
Hence, the system is stable.

Dr Majd Saied 36
Problems
Solution

10(𝑧 + 1) 10(𝑧 + 1)
b) 𝐺 𝑧 = =
(𝑧 − 0.7)(𝑧 − 0.9) 𝑧 2 − 1.6𝑧 + 0.63
𝑖) 1 + −1.6 + 0.63 > 0
𝑖𝑖) 1 − −1.6 + 0.63 > 0
𝑖𝑖𝑖) 1 > 0.63
Hence, the system is stable.

Dr Majd Saied 37
Problems
Problem 5:
Use the Routh-Hurwitz criterion to find the stable range of 𝐾 for the closed-loop unity
feedback systems with loop gain.
𝐾(𝑧 − 1)
𝑎) 𝐺 𝑧 =
(𝑧 − 0.1)(𝑧 − 0.8)
𝐾(𝑧 + 0.1)
𝑏) 𝐺 𝑧 =
(𝑧 − 0.7)(𝑧 − 0.9)

Solution
The characteristic equation for a system with forward gain 𝐺(𝑧) and unity feedback is:

1+𝐺 𝑧 =0⇒𝑁 𝑧 +𝐷 𝑧 =0

where 𝑁(𝑧) and 𝐷(𝑧) are the numerator and denominator of the transfer function
respectively.

Dr Majd Saied 38
Problems
Solution
𝐾(𝑧 − 1)
𝑎) 𝐺 𝑧 =
(𝑧 − 0.1)(𝑧 − 0.8)
The closed loop characteristic equation is:

𝑧 − 0.1 𝑧 − 0.8 + 𝐾 𝑧 − 1 = 𝑧 2 + 𝐾 − 0.9 𝑧 + 0.08 − 𝐾 = 0

This is a second order polynomial and the Routh-Hurwitz test yields the stability conditions
are:
𝑖) 𝑎2 − 𝑎1 + 𝑎0 > 0
𝑖𝑖) 𝑎2 − 𝑎0 > 0
𝑖𝑖𝑖) 𝑎2 + 𝑎1 + 𝑎0 > 0
Hence:

𝑖) 1 − 𝐾 − 0.9 + (0.08 − 𝐾) > 0 ⇒ 1.98 − 2𝐾 > 0 ⇒ 𝐾 < 0.99

𝑖𝑖) 1 − 0.08 − 𝐾 > 0 ⇒ 𝐾 ≻ −0.92
𝑖𝑖𝑖) 1 + 𝐾 − 0.9 + 0.08 − 𝐾 > 0 ⇒ 0.18 > 0
The stable range of 𝐾 is: −0.92 < 𝐾 < 0.99
Dr Majd Saied 39
Problems
Solution
𝐾(𝑧+0.1)
b) 𝐺 𝑧 = (𝑧−0.7)(𝑧−0.9)
The closed loop characteristic equation is:

𝑧 − 0.7 𝑧 − 0.9 + 𝐾 𝑧 + 0.1 = 𝑧 2 + 𝐾 − 1.6 𝑧 + 0.63 + 0.1𝐾 = 0

This is a second order polynomial and the Routh-Hurwitz test yields the stability conditions
are:

𝑖) 1 − 𝐾 − 1.6 + (0.63 + 0.1𝐾) > 0 ⇒ 3.23 − 0.9𝐾 > 0 ⇒ 𝐾 < 3.589

𝑖𝑖) 1 − 0.63 + 0.1 𝐾 > 0 ⇒ 𝐾 < 3.7
𝑖𝑖𝑖) 1 + 𝐾 − 1.6 + 0.63 + 0.1𝐾 > 0 ⇒ 1.1𝐾 − 0.03 > 0 ⇒ 𝐾 ≻ −0.0273
The stable range of 𝐾 is: −0.0273 < 𝐾 < 3.589

Dr Majd Saied 40
Problems
Problem 6
Use the Jury criterion to determine the stability of the following polynomials
𝑎) 𝑧 5 + 0.2𝑧 4 + 𝑧 2 + 0.3𝑧 − 0.1 = 0
𝑏) 𝑧 5 − 0.25𝑧 4 + 0.1𝑧 3 + 0.4𝑧 2 + 0.3𝑧 − 0.1 = 0

Solution
𝑎) 𝑧 5 + 0.2𝑧 4 + 𝑧 2 + 0.3𝑧 − 0.1 = 0
The first two conditions require the evaluation of 𝐹(𝑧) at 𝑧 = ±1
1 𝐹 1 = 1 + 0.2 + 1 + 0.3 − 0.1 = 2.4 > 0
2 (−1)5 𝐹(−1) = (−1)(−1 + 0.2 + 1 − 0.3 − 0.1 ) = 0.2 > 0

Dr Majd Saied 41
Problems
Solution
3 𝑎0 = 0.1 < 𝑎2
4 𝑏0 = 0.99 > 𝑏4 = 0.32
5 𝑐0 = 0.8777 < 𝑐3 = 0.9164 condition violated.

We conclude that the polynomial is unstable. The polynomial has roots outside the unit
circle.

𝑏) 𝑧 5 − 0.25𝑧 4 + 0.1𝑧 3 + 0.4𝑧 2 + 0.3𝑧 − 0.1 = 0

The first two conditions require the evaluation of 𝐹(𝑧) at 𝑧 = ±1
1 𝐹 1 = 1 − 0.25 + 0.1 + 0.4 − 0.3 = 1.45 > 0
2 (−1)5 𝐹(−1) = (−1)(−1 − 0.25 − 0.1 + 0.4 − 0.3 − 0.1 ) = 1.35 > 0

Dr Majd Saied 42
Problems
Solution
3 𝑎0 = 0.1 < 𝑎2
4 𝑏0 = 0.99 > 𝑏4 = 0.275
5 𝑐0 = 0.9045 > 𝑐3 = 0.4664
6 𝑑0 = 0.812 > 𝑑2 = 0.2288

We conclude that the polynomial is stable. The polynomial has no roots outside the unit
circle.

Dr Majd Saied 43