Stability of Digital

Control Systems
M. Sami Fadali
Professor of Electrical Engineering
UNR

1
 Outline
• Asymptotic stability.
• Input-output stability.
• Internal stability.
• Stability conditions.
• Routh-Hurwitz criterion.
• Jury test.

2
 Asymptotic Stability
Response due to any initial conditions
decays to zero asymptotically in the steady
state
Lim 𝑦 𝑘 = 0
𝑘→∞
Marginal Stability: response due to any
initial conditions remains bounded but does
not decay to zero.

3
Bounded-Input-Bounded-Output
(BIBO) Stability
The response due to any bounded input
𝑢 𝑘 < 𝑏𝑢 , 𝑘 = 0,1,2, …
0 < 𝑏𝑢 < ∞
remains bounded.
𝑦 𝑘 < 𝑏𝑦 , 𝑘 = 0,1,2, …
0 < 𝑏𝑦 < ∞
4
 Stable Z-Domain Pole
Locations
Sampled exponential and its z-transform
with 𝑝 real or complex
𝑘
Z 𝑧
𝑝 , 𝑘 = 0, 1, 2, …
𝑧−𝑝
0, 𝑝 < 1
Lim 𝑝 𝑘 = ቐ 1, 𝑝 = 1
𝑘→∞
∞, 𝑝 > 1

5
 𝑛
Time Sequence 𝑛
Z 𝐴𝑖 𝑧
𝑓 𝑘 = ෍ 𝐴𝑖 𝑝𝑖𝑘 , 𝑘 = 0, 1, 2, … 𝐹 𝑧 =෍
𝑧 − 𝑝𝑖
𝑖=1 𝑖=1
 Bounded sequence for poles in the closed unit
disc (i.e. on or inside the unit circle).
 Sequence decays exponentially for poles in the
open unit disc (i.e. inside the unit circle).
 Unbounded sequence for repeated poles on the
unit circle.
 For real time sequences poles and partial fraction
coefficients are either real or complex conjugate
pairs.
6
 Theorem: Asymptotic Stability
Asymptotic Stability: response to any ICs
decays to zero asymptotically.
Thm. In the absence of pole-zero cancellation, a
LTI digital system is (i) asymptotically stable if its
transfer function poles are in the open unit disc
and (ii) marginally stable if the poles are in the
closed unit disc with no repeated poles on the
unit circle.
LTI Model:
𝑦 𝑘 + 𝑛 + 𝑎𝑛−1 𝑦 𝑘 + 𝑛 − 1 + ⋯ + 𝑎1 𝑦 𝑘 + 1 + 𝑎0 𝑦 𝑘
= 𝑏𝑚 𝑢 𝑘 + 𝑚 + 𝑏𝑚−1 𝑢 𝑘 + 𝑚 − 1 + ⋯ + 𝑏1 𝑢 𝑘 + 1
+ 𝑏0 𝑢 𝑘
𝑘 = 0,1,2,7 …
 Proof
• Response of the system due to ICs 𝑦 0 , 𝑦 1 , … , 𝑦 𝑛
𝑁 𝑧
𝑌 𝑧 = 𝑛
𝑧 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0
1. Assume no pole-zero cancellation and consider all
the poles.
2. 𝑌 𝑧 denominator= z-transfer function denominator.
3. Poles of 𝑌 𝑧 = poles of transfer function.
4. 𝑌 𝑧 due to ICs: bounded for system poles in the
closed unit disc unit (no repeated poles on the unit
circle) & decays exponentially for system poles in
the open unit disc (inside the unit circle).
8
 Example
Determine the asymptotic stability of the
following systems:
4 𝑧−2 4 𝑧 − 0.2
𝑎) 𝐻 𝑧 = , 𝑏) 𝐻 𝑧 =
𝑧 − 2 𝑧 − 0.1 𝑧 − 0.2 𝑧 − 0.1
5 𝑧 − 0.3 8 𝑧 − 0.2
𝑐) 𝐻 𝑧 = , 𝑑) 𝐻 𝑧 =
𝑧 − 0.2 𝑧 − 0.1 𝑧 − 0.1 𝑧 − 1
• Use Theorem 1 (a) and (b) without pole-zero
cancellation.
• Ignore zeros, (do not affect response due to ICs)
9
 Solution (𝑎, 𝑏)
4 𝑧−2
𝑎) 𝐻 𝑧 =
𝑧 − 2 𝑧 − 0.1
• Pole outside the unit circle  unstable
4 𝑧 − 0.2
𝑏) 𝐻 𝑧 =
𝑧 − 0.2 𝑧 − 0.1
• All poles inside the unit circle 
asymptotically stable.

10
 Solution 𝑐, 𝑑
5 𝑧 − 0.3
𝑐) 𝐻 𝑧 =
𝑧 − 0.2 𝑧 − 0.1
• All poles inside the unit circle 
asymptotically stable.
8 𝑧 − 0.2
𝑑) 𝐻 𝑧 =
𝑧 − 0.1 𝑧 − 1
• One pole on the unit circle  marginally
stable: not asymptotically stable.

11
 BIBO Stability
𝑘

𝑦 𝑘 = ෍ℎ 𝑘 − 𝑖 𝑢 𝑖 , 𝑘 = 0, 1, 2, …
𝑖=0
Is system BIBO stable if its impulse response ℎ 𝑘 is
bounded? NO
Counterexample: bounded & strictly positive impulse
response 0 < 𝑏ℎ1 < ℎ 𝑘 < 𝑏ℎ2 < ∞, 𝑘 = 0, 1, 2, …
Bounded input: 𝑢 𝑖 = 1, 𝑖 = 0, 1, 2, …
Unbounded output:
𝑘 𝑘

𝑦 𝑘 = ෍ ℎ 𝑘 − 𝑖 𝑢 𝑖 > 𝑏ℎ1 ෍ 1 , 𝑘 = 0, 1, 2, …
𝑖=0 𝑖=0 12
 Theorem 4.2:BIBO Stability

A discrete-time linear system is BIBO

stable if and only if its impulse response
sequence is absolutely summable i.e.
∞

෍ℎ 𝑖 <∞
𝑖=0

13
 Proof of Necessity (Only if)
𝑘

𝑦 𝑘 = ෍ℎ 𝑖 𝑢 𝑘 − 𝑖 , 𝑘 = 0, 1, 2, …
𝑖=0
• Assume the system is BIBO stable but the
impulse response is not absolutely summable.
1, ℎ 𝑖 ≥ 0
• Input 𝑢 𝑘 − 𝑖 = ቊ
−1, ℎ 𝑖 < 0
𝑘

𝑦 𝑘 =෍ℎ 𝑖 , 𝑦 𝑘 → ∞ 𝑎𝑠 𝑘 → ∞
𝑖=0
• Contradiction: unbounded output with a bounded
input.
14
 Proof of Sufficiency (If)
• Assume an absolutely summable impulse
response and show that the system is BIBO
stable.
• Use the input bound 𝑏𝑢 in the convolution
summation 𝑢 𝑘 < 𝑏𝑢
𝑘

𝑦 𝑘 ≤෍ℎ 𝑖 𝑢 𝑘−𝑖
𝑖=0
𝑘

< 𝑏𝑢 ෍ ℎ 𝑖 < ∞, ∀𝑘 ≥ 0
𝑖=0 15
 Theorem 4.3: BIBO Stability
A discrete-time linear system is BIBO
stable if and only if the poles of its transfer
function lie inside the unit circle.

16
 Proof of Necessity
𝑛 𝑛
Z 𝐴𝑖 𝑧
ℎ 𝑘 = ෍ 𝐴𝑖 𝑝𝑖𝑘 , 𝑘 = 0, 1, 2, … 𝐻 𝑧 =෍
𝑧 − 𝑝𝑖
𝑖=1 𝑖=1
• Impulse response:
– Bounded if the poles of the transfer function are in
the closed unit disc, no repeated poles on the disc
– Decays exponentially if the poles are in the open unit
disc.
• Systems with a bounded impulse response that
does not decay exponentially are not BIBO
stable.
17
 Proof of Sufficiency
• Assume exponentially decaying impulse
response (i.e. poles inside the unit circle).
𝐴𝑟 is the coefficient of largest magnitude 𝑝𝑠 < 1
is the largest pole magnitude.
• The impulse response is bounded by
𝑛 𝑛

ℎ 𝑘 = ෍ 𝐴𝑖 𝑝𝑖𝑘 ≤ ෍ 𝐴𝑖 𝑝𝑖 𝑘 ≤ 𝑛 𝐴𝑟 𝑝𝑠 𝑘 ,
𝑖=1 𝑖=1
𝑘 = 0,1,2, …
∞ ∞
𝑘
𝑛 𝐴𝑟
෍ℎ 𝑘 ≤ 𝑛 𝐴𝑟 ෍ 𝑝𝑠 = <∞
1 − 𝑝𝑠
𝑘=0 𝑘=0
18
 Example 4.2
Investigate the BIBO stability of systems
with the impulse response
𝐾, 0≤𝑘≤𝑚<∞
ℎ 𝑘 =ቊ
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
where 𝐾 is a finite constant.

19
 Solution
BIBO stable since the impulse response
satisfies
∞ 𝑚

෍ℎ 𝑘 =෍ℎ 𝑘 = 𝑚+1 𝐾 <∞

𝑘=0 𝑘=0

• Let 𝐾 = upper bound for any impulse

response of finite duration.
• Any FIR system is BIBO stable.
20
 Example 4.3
Investigate BIBO stability for Example 4.1
4 𝑧−2 4 𝑧 − 0.2
𝑎) 𝐻 𝑧 = , 𝑏) 𝐻 𝑧 =
𝑧 − 2 𝑧 − 0.1 𝑧 − 0.2 𝑧 − 0.1
5 𝑧 − 0.3 8 𝑧 − 0.2
𝑐) 𝐻 𝑧 = , 𝑑) 𝐻 𝑧 =
𝑧 − 0.2 𝑧 − 0.1 𝑧 − 0.1 𝑧 − 1
• Use Theorem 4.3 with pole-zero cancellation.

21
 Solution 𝑎, 𝑏, 𝑐
4 𝑧−2 4 𝑧 − 0.2
𝑎) 𝐻 𝑧 = , 𝑏) 𝐻 𝑧 =
𝑧 − 2 𝑧 − 0.1 𝑧 − 0.2 𝑧 − 0.1
5 𝑧 − 0.3
𝑐) 𝐻 𝑧 =
𝑧 − 0.2 𝑧 − 0.1

After pole-zero cancellation

a) BIBO stable, all poles inside unit circle.
b) BIBO stable, all poles inside unit circle.
c) BIBO stable, all poles inside unit circle.
22
 Solution 𝑑
8 𝑧 − 0.2
𝑑) 𝐻 𝑧 =
𝑧 − 0.1 𝑧 − 1
• Not BIBO stable, a pole on unit circle.

23
Z-plane Stable Pole Locations
Im[z]

Unit Circle

STABLE
Re[z]

24
MATLAB Stability Determination
Obtain roots of polynomial:
» roots(den) % denominator coeffts. den
» zpk(g) %g = transfer function

Stable for roots inside the unit circle.

25
 MATLAB: ddamp
Pole Magnitude Damping Frequency Time Constant
(rad/TimeUnit) (TimeUnit)
-5.00e-02 + 1.41e-01 7.11e-01 2.75e+00 5.11e-01
1.32e-01i
-5.00e-02 - 1.41e-01 7.11e-01 2.75e+00 5.11e-01
1.32e-01i

>> ddamp([1,.1,.02])
Gives the pole locations, 𝜁 and 𝜔𝑛
• Closed-loop transfer function
>> H = feedback(gforward, gfeedback, 1)
• Default: negative feedback
26
 Internal Stability
If all the transfer functions that relate the system
inputs (𝑅 and 𝐷) to the possible system outputs
(𝑌 and 𝑈) are BIBO stable, then the system is
internally stable.
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
D(z
R(z) + E(z) U(z) ) Y(z)
C(z) GZAS(z)


27
 Theorem 4-4
The system of Figure 4-3 is internally stable if
and only if all the closed-loop poles are in the
open unit disc.

D(z
R(z) + E(z) U(z) ) Y(z)
C(z) GZAS(z)


28
 Transfer Functions
𝐺𝑍𝐴𝑆 𝑧 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧
𝑌 𝑧 𝐶 𝑧 −𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 𝑅 𝑧
=
𝑈 𝑧 1 + 𝐺𝑍𝐴𝑆 𝑧 𝐶 𝑧 𝐷 𝑧
• Ratios of coprime polynomials (no cancellation)
𝑁𝐶 𝑧 𝑁𝐺 𝑧
𝐶 𝑧 = , 𝐺𝑍𝐴𝑆 𝑧 =
𝐷𝐶 𝑧 𝐷𝐺 𝑧
D(z
R(z) + E(z) U(z) ) Y(z)
C(z) GZAS(z)


29
 Substitute
𝐺𝑍𝐴𝑆 𝑧 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧
𝑌 𝑧 𝐶 𝑧 −𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 𝑅 𝑧
=
𝑈 𝑧 1 + 𝐺𝑍𝐴𝑆 𝑧 𝐶 𝑧 𝐷 𝑧
• No cancellation of poles on or outside the
unit circle
𝑁𝐶 𝑧 𝑁𝐺 𝑧
𝐶 𝑧 = , 𝐺𝑍𝐴𝑆 𝑧 =
𝐷𝐶 𝑧 𝐷𝐺 𝑧
𝑁𝑐 𝑁𝐺 𝐷𝑐 𝑁𝐺
𝑌 𝑧 𝑁𝑐 𝐷𝐺 −𝑁𝑐 𝑁𝐺 𝑅 𝑧
=
𝑈 𝑧 𝐷𝑐 𝐷𝐺 + 𝑁𝑐 𝑁𝐺 𝐷 𝑧
30
 Proof (Necessity: only if)
𝑁𝑐 𝑁𝐺 𝐷𝑐 𝑁𝐺
𝑌 𝑧 𝑁𝑐 𝐷𝐺 −𝑁𝑐 𝑁𝐺 𝑅 𝑧
=
𝑈 𝑧 𝐷𝑐 𝐷𝐺 + 𝑁𝑐 𝑁𝐺 𝐷 𝑧
• Internally stable (all transfer functions are
asymptotically stable) + No cancellation of
poles on or outside the unit circle
⇒ characteristic polynomial 𝐷𝑐 𝐷𝐺 + 𝑁𝑐 𝑁𝐺 has
no zeros on or outside the unit circle.
31
 Proof (Sufficiency: if)
𝑁𝑐 𝑁𝐺 𝐷𝑐 𝑁𝐺
𝑌 𝑧 𝑁𝑐 𝐷𝐺 −𝑁𝑐 𝑁𝐺 𝑅 𝑧
=
𝑈 𝑧 𝐷𝑐 𝐷𝐺 + 𝑁𝑐 𝑁𝐺 𝐷 𝑧
• Characteristic polynomial 𝐷𝑐 𝐷𝐺 + 𝑁𝑐 𝑁𝐺
• No zeros on or outside the unit circle
⇒ the systems with the 4 TFs are asymptotically
stable
⇒ closed-loop system is internally stable.
32
 Theorem 4-5
The system is internally stable if and only if:
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 .
D(z)
R(z) + E(z) U(z) Y(z)
C(z) + GZAS(z)


33
 Proof: Necessity
(i) 1 + 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 : no zeros on or outside the
unit circle for internal stability (Theorem 4.4)
(ii) 𝑧0 > 1 cancels ⇒ unstable
𝑁𝐶 𝑧 𝑁𝐺 𝑧 𝑧 − 𝑧0 𝑁1 𝑧
𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 = =
𝐷𝐶 𝑧 𝐷𝐺 𝑧 𝑧 − 𝑧0 𝐷1 𝑧
𝐷𝐶 𝑧 𝐷𝐺 𝑧 + 𝑁𝐶 𝑧 𝑁𝐺 𝑧
1 + 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 =
𝐷𝐶 𝑧 𝐷𝐺 𝑧
𝑧 − 𝑧0 𝑁1 𝑧 + 𝐷1 𝑧 c. l. charact. poly.
= =
𝑧 − 𝑧0 𝐷1 𝑧 o. l. charact. poly.
𝑧0 is both an open-loop and a closed-loop pole
34
 Proof: Sufficiency
𝐷𝐶 𝑧 𝐷𝐺 𝑧 + 𝑁𝐶 𝑧 𝑁𝐺 𝑧
1 + 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 =
𝐷𝐶 𝑧 𝐷𝐺 𝑧
i. No zeros (c.l. poles) on or outside the unit circle
ii. No unstable pole-zero cancellation for the
closed-loop system if the loop gain has no
unstable pole-zero cancellation
𝑁𝐶 𝑧 𝑁𝐺 𝑧
𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 =
𝐷𝐶 𝑧 𝐷𝐺 𝑧
Sufficient for internal stability (Theorem 4.4)

35
 Example 4-4
Transfer function of isothermal chemical
reactor
0.5848 −0.3549𝑠 + 1
𝐺 𝑠 =
0.1828𝑠 2 + 0.8627𝑠 + 1
o Determine 𝐺𝑍𝐴𝑆 𝑧 with 𝑇 = 0.1
o Verify that the resulting feedback system is
not internally stable with the feedback
controller
10 𝑧 − 0.8149 𝑧 − 0.7655
𝐶 𝑧 =−
𝑧 − 1 𝑧 − 1.334
36
 Solution
• Discretized process transfer function
𝐺 𝑠
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 Z −1
𝑠
−0.075997 𝑧 − 1.334
=
𝑧 − 0.8149 𝑧 − 0.7655
• Transfer function from reference input to output
𝑌 𝑧 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 0.75997
= =
𝑅 𝑧 1 + 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 𝑧 − 0.24
• External stability: All its poles are inside the
unit circle (some cancel)
37
 Solution: Control Variable
𝑈 𝑧 𝐶 𝑧
=
𝑅 𝑧 1 + 𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧
−10 𝑧 − 0.8149 𝑧 − 0.7655
=
𝑧 − 0.24 𝑧 − 1.334
• Pole at 1.334 outside the unit circle.
• The control variable is unbounded even
when the reference input is bounded.
38
 Solution: Theorem 4-5
−10 𝑧 − 0.8149 𝑧 − 0.7655
𝐶 𝑧 𝐺𝑍𝐴𝑆 𝑧 =
𝑧 − 0.24 𝑧 − 1.334
−0.075997 𝑧 − 1.334
×
𝑧 − 0.8149 𝑧 − 0.7655
• Violates condition (ii) of Theorem 4-5:
Unstable pole at 1.334 cancels in the loop
gain.

39
 Routh-Hurwitz Criterion
1. Transform the inside of the unit circle to
the LHP (bilinear transformation).
2. Use the Routh-Hurwitz criterion for the
investigation of discrete-time system
stability.
1+𝑤 𝑧−1
𝑧= ⇔𝑤=
1−𝑤 𝑧+1

40
 Geometric Interpretation
Im(s) z3 Im(z)
z3  1
z2  1
X s3
z3  1 z2
z2  1
s2 X
Re(s) Re(z)
1 1
s1 X z1  1
z1  1 z1

∠𝑠 = ∠ 𝑧 − 1 − ∠ 𝑧 + 1
> 90° → 𝑖𝑛𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑐𝑖𝑟𝑐𝑙𝑒
∠𝑠 ቐ = ±90° → 𝑜𝑛 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑐𝑖𝑟𝑐𝑙𝑒
< 90° → 𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑐𝑖𝑟𝑐𝑙𝑒 41
 Advantages/Disadvantages
• Easy stability test for low-order polynomials.
• Difficult for high order z-polynomials.
• For high order polynomials, use symbolic
manipulation.
𝐹 𝑧 = 𝑎𝑛 𝑧 𝑛 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0
1+𝑤
𝑧=
1−𝑤

𝑛 𝑛−1
1+𝑤 1+𝑤 1+𝑤
𝑎𝑛 + 𝑎𝑛−1 + ⋯ + 𝑎1 + 𝑎0
1−𝑤 1−𝑤 1−𝑤
42
 Example 4.4
Find stability conditions for
a) The first order polynomial
𝑎1 𝑧 + 𝑎0 , 𝑎1 > 0
b) The second order polynomial
𝑎2 𝑧 2 + 𝑎1 𝑧 + 𝑎0 , 𝑎2 > 0

43
Solution: 1st order
• Solve for the root.
• Stability conditions
𝑎0
<1
𝑎1

44
 Solution: 2nd order
−𝑎1 ± 𝑎12 − 4𝑎0 𝑎2
𝑧1,2 =
2𝑎2
 Stability determination by solving for roots is
difficult.
 Monic polynomial
constant term = product of poles
 For pole magnitudes < 1
Necessary stability condition
Sufficient for complex conjugate poles
𝑎0
< 1 ⇔ −𝑎0 < 𝑎2 & 𝑎0 < 𝑎2
𝑎2
45
 Bilinear Transformation
2
1+𝑤 1+𝑤
𝑎2 + 𝑎1 + 𝑎0
1−𝑤 1−𝑤
𝑎2 − 𝑎1 + 𝑎0 𝑤 2 + 2 𝑎2 − 𝑎0 𝑤 + 𝑎2 + 𝑎1 + 𝑎0
• Routh-Hurwitz criterion: poles of 2nd order w-
polynomial remain in the LHP iff its coefficients are
all positive.
𝑎2 − 𝑎1 + 𝑎0 > 0, 𝑎2 − 𝑎0 > 0, 𝑎2 + 𝑎1 + 𝑎0 > 0
• Adding the first and third conditions gives the
condition obtained earlier 𝑎2 + 𝑎0 > 0
46
 Comments
• Recall: −𝑎0 < 𝑎2 & 𝑎0 < 𝑎2
Sufficient for complex conjugate roots and
only necessary for real roots.
• Real roots: substituting the three conditions
in the z-domain characteristic polynomial
gives roots between 1 and +1.

47
 Stable Parameter Range for 2nd
Order z-polynomial
Conditions with a0
𝑎2 = 1:
a0 = 1
1 − 𝑎1 + 𝑎0 > 0 1

stable
a0= a11
1 − 𝑎0 > 0 1 1
a1
1 + 𝑎1 + 𝑎0 > 0 a0= a11
1

48
 Jury Test
The roots of the polynomial
𝐹 𝑧 = 𝑎𝑛 𝑧 𝑛 + 𝑎𝑛−1 𝑧 𝑛−1 + ⋯ + 𝑎1 𝑧 + 𝑎0 , 𝑎𝑛 > 0
are inside the unit circle if and only if
(1) 𝐹 1 >0
(2) (1)𝑛 𝐹 1 > 0
(3) 𝑎 0 < 𝑎𝑛
(4) 𝑏0 > 𝑏𝑛−1
(5) 𝑐0 > 𝑐𝑛−2
⋮
(n+1) 𝑟0 > 𝑟2
49
 Jury Table
Row z0 z1 z2  zn  k  zn  1 zn
1 a0 a1 a2  an  k  an  1 an
2 an an  1 an  2  ak  a1 a0
3 b0 b1 b2  bn  k  bn  1
4 bn  1 bn  2 bn  3  bk  b0

5 c0 c1 c2   cn  2
6 cn  2 cn  3 cn  4   c0

. . . .  
. . . .  
. . . .  
2 n  5 s0 s1 s2 s3
2 n  4 s3 s2 s1 s0

2 n  3 r0 r1 r2

50
 Table Entries
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
𝑟0 = 𝑠 𝑠0 , 𝑟1 = 𝑠 𝑠1
3 3
𝑠0 𝑠1
𝑟2 = 𝑠 𝑠2
3
51
 Remarks
1. First row of Jury table: list the coefficients of the
polynomial 𝐹 𝑧 in order of increasing power of 𝑧.
2. Number of rows of 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.
3. There are 𝑛 + 1 conditions for 𝑛 + 1 coefficients of 𝐹 𝑧 .
4. Conditions 3 through 𝑛 + 1 are calculated using the
coefficient of the first column of the Jury table, together
with the last coefficient of the row.
5. The middle coefficient of the last row is never used and
need not be calculated.
6. Conditions (1) and (2) are calculated from 𝐹 𝑧 directly. If
one of the first two conditions is violated, 𝐹 𝑧 has roots
on or outside the unit circle (no need to construct the Jury
table or test the remaining conditions). 52
 Remarks (cont.)
7- Condition (3) with 𝑎𝑛 = 1, requires the constant term of
the polynomial to be less than unity in magnitude. The
constant term is simply the product of the roots and
must be smaller than unity for all the roots to be inside
the unit circle.
8 For first and second order systems, the Jury stability
conditions reduce to the conditions derived earlier.
9- For higher order systems, applying the Jury test by
hand is laborious and it is preferable to test the stability
of a polynomial 𝐹 𝑧 using a CAD package.
10- If the coefficients of the polynomial are functions of
system parameters, the Jury test can be used to obtain
the stable ranges of the system parameters.

53
 Example
Test the stability of the polynomial
𝐹 𝑧 = 𝑧 5 + 2.6𝑧 4 − 0.56𝑧 3 − 2.05𝑧 2 + 0.0775𝑧 + 0.35

Row 𝑧0 𝑧1 𝑧2 𝑧3 𝑧4 𝑧5
1 0.35 0.0775 2.05 0.56 2.6 1
2 1 2.6 0.56 2.05 0.0775 0.35
3 0.8775 2.5729 0.1575 1.854 0.8325
4 0.8325 1.854 0.1575 2.5729 0.8775
5 0.0770 0.7143 0.2693 0.5151
6 0.5151 0.2693 0.7143 0.0770
7 0.2593 0.0837 0.3472

54
 Solution
(1) 𝐹 1 > 0? Yes
𝐹 1 = 1 + 2.6 − 0.56 − 2.05 + 0.0775 + 0.35 = 1.4175
(2) 1 𝑛 𝐹 1 > 0? No
−1 5 𝐹 −1
= −1 −1 + 2.6 + 0.56 − 2.05 − 0.0775 + 0.35
= −0.3825 < 0
(3) 𝑎0 < 𝑎𝑛 ? Yes 0.35 < 1
(4) 𝑏0 > 𝑏𝑛−1 ? Yes −0.8775 > 0.8325
(5) 𝑐0 > 𝑐𝑛−2 ? No 0.0770 < 0.5151
(6) 𝑟0 > 𝑟2 ? No −0.2593 < −0.3472
55
 Solution
• Conditions (2), (5) & (6) violated .
• Condition (2) is sufficient to conclude
instability.
No. of conditions violated  No. of roots outside
the unit circle.
• Polynomial has roots on or outside the unit
circle.
𝐹 𝑧
= 𝑧 − 0.7 𝑧 − 0.5 𝑧 + 0.5 𝑧 + 0.8 𝑧 + 2.5
• Root at −2.5 outside the unit circle.
56
 Example 4.6
Find the stable range of the gain 𝐾 for
the unity feedback digital control system
with analog plant
𝐾
𝐺 𝑠 =
𝑠+3
with DAC and ADC if the sampling
period is 0.02 s.

57
 Solution
• Transfer function of analog subsystem, ADC and DAC
𝐺 𝑠
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 −1 Z
𝑠
𝐾
= 1−𝑧 Z −1
𝑠 𝑠+3
• Partial fraction expansion
𝐾 𝐾 1 1
= −
𝑠 𝑠+3 3 𝑠 𝑠+3
• Transfer function (𝑇 = 0.02 𝑠)
1.9412  10−2 𝐾
𝐺𝑍𝐴𝑆 𝑧 =
𝑧 − 0.9418
58
 Closed-loop System
• Unity feedback, closed-loop characteristic
equation 1 + 𝐺𝑍𝐴𝑆 𝑧 = 0
𝑧 − 0.9418 + 1.9412 × 10−2 𝐾 = 0
• Stability conditions 𝑎0 < 1 ⇒ −1 < 𝑎0 < 1
0.9418  1.9412 10−2 𝐾 < 1
0.9418 + 1.9412 10−2 𝐾 < 1
• Stable range of 𝐾 is
3 < 𝐾 < 100.03
59
 Example
Find the stable range of the gain 𝐾 for the digital
position control system with the analog plant
transfer function
𝐾
𝐺 𝑠 =
𝑠 𝑠 + 10
and with DAC and ADC if 𝑇 = 0.05 𝑠.

60
 Solution
𝐺 𝑠
𝐺𝑍𝐴𝑆 𝑧 = 1 − 𝑧 −1
Z
𝑠
𝐾
= 1−𝑧 −1
Z 2
𝑠 𝑠 + 10
• Partial fraction expansion
𝐾 10 1 1
2
= 0.01𝐾 2 − +
𝑠 𝑠 + 10 𝑠 𝑠 𝑠 + 10
• Transfer function (𝑇 = 0.05 𝑠)
𝐾 𝑧 + 0.8467
𝐺𝑍𝐴𝑆 𝑧 = 1.0653 × 10−2
𝑧 − 1 𝑧 − 0.6065
61
 C.L. Characteristic Equation
−2
𝐾 𝑧 + 0.8467
𝐺𝑍𝐴𝑆 𝑧 = 1.0653 × 10
𝑧 − 1 𝑧 − 0.6065
1 + 𝐺𝑍𝐴𝑆 𝑧 = 0
𝑧 − 1 𝑧 − 0.6065 + 1.0653 × 10−2 𝐾 𝑧 + 0.8467
= 𝑧 2 + 1.0653 × 10−2 𝐾 − 1.6065 𝑧 + 0.6065
+9.02 × 10−3 𝐾 = 0

62
 Stability Testing
𝐹 𝑧 = 𝑧 2 + 1.0653 × 10−2 𝐾 − 1.6065 𝑧
+0.6065 + 9.02 × 10−3 𝐾
1) 𝐹 1 > 0?
𝐹 1 = 1 + 1.0653  102𝐾 − 1.6065 + 0.6065
+ 9.02103𝐾 > 0, ∀𝐾 > 0
2) 1 𝑛 𝐹 1 > 0?
𝐹 1 = 1 1.0653102 𝐾1.6065 + 0.6065
+9.02103𝐾 > 0  𝐾 < 1967.582

63
 Stability Testing
𝐹 𝑧 = 𝑧 2 + 1.0653 × 10−2 𝐾 − 1.6065 𝑧
+0.6065 + 9.02 × 10−3 𝐾
3) 𝑎0 < 𝑎𝑛 , 0.6065 + 0.0902 𝐾 < 1
 + 0.6065 + 0.0902 𝐾 < 1
&  (0.6065 + 0.0902 𝐾 ) < 1
 178.104 < 𝐾 < 43.6199
Earlier conditions: 0 < 𝐾, 𝐾 < 1967.582
• The stable range is 0 < 𝐾 < 43.6199
64