ELE8311: The z-Transform
Dr. Hassan A. Bashir
ELE8311: Digital Control Engineering
2.5 Module 2B - Inversion of the z-transform
While z-transformation mainly simplifies analysis of time domain problems, it is essential to
invert z-transforms back to time domain in order to make sense of their analytic behavior.
There are a number of approaches for determining the inverse of z-transforms. Two
approaches that will be covered in this course are the power series (long division) and partial
fraction approaches.
a. Power series – Long division method
Since z-transforms are generally obtained as a ratio of exponentials in z, it is possible to
apply long division to obtain as many terms as desired of the z-transform expansion; then
the coefficients of this expansion can be used to write the corresponding time sequence.
Procedure:
i. Expand 𝐹𝐹(𝑧𝑧) using long division to obtain the following series
𝑡𝑡
−1 −2 −𝑡𝑡
𝐹𝐹𝑡𝑡 (𝑧𝑧) = 𝑓𝑓0 + 𝑓𝑓1 𝑧𝑧 + 𝑓𝑓2 𝑧𝑧 + ⋯ + 𝑓𝑓𝑡𝑡 𝑧𝑧 = � 𝑓𝑓𝑘𝑘 𝑧𝑧 −𝑘𝑘
𝑘𝑘=0
ii. Write the inverse transform as the sequence
{𝑓𝑓0 , 𝑓𝑓1 , … , 𝑓𝑓𝑡𝑡 , … }
The number of terms obtained via long division is chosen to obtain sufficient number of
points in the time sequence.
Thus,
𝑥𝑥(𝑘𝑘𝑘𝑘) = 𝑓𝑓0 + 𝑓𝑓1 𝛿𝛿(𝑡𝑡 − 𝑇𝑇) + 𝑓𝑓2 𝛿𝛿(𝑡𝑡 − 2𝑇𝑇) + ⋯.
Example 2.13
Determine the inverse z-transform of the function
1
ELE8311: The z-Transform
𝑧𝑧 + 1
𝐹𝐹(𝑧𝑧) =
𝑧𝑧 2 + 0.2𝑧𝑧 + 0.1
Solution
𝑧𝑧 + 1
𝑥𝑥(𝑘𝑘𝑘𝑘) = 𝑍𝑍 −1 � � = 𝑍𝑍 −1 [0 + 𝑧𝑧 −1 + 0.8𝑧𝑧 −2 − 0.26𝑧𝑧 −3 + ⋯ ]
𝑧𝑧 2 + 0.2𝑧𝑧 + 0.1
= 1𝛿𝛿(𝑡𝑡 − 𝑇𝑇) + 0.8𝛿𝛿(𝑡𝑡 − 2𝑇𝑇) − 0.26𝛿𝛿(𝑡𝑡 − 3𝑡𝑡) + ⋯
Hence, the time sequence is:
{𝑓𝑓𝑘𝑘 } = {0,1,0.8, −0.25, … }
Exercise 2.1
Determine the inverse z-transform of the function
𝑧𝑧 3 + 2𝑧𝑧 2 + 1
𝐹𝐹(𝑧𝑧) = 3
𝑧𝑧 − 1.5𝑧𝑧 2 + 0.5𝑧𝑧
Ans: 𝑥𝑥(𝑘𝑘𝑘𝑘) = 1 + 3.5𝛿𝛿(𝑡𝑡 − 𝑇𝑇) + 4.75𝛿𝛿(𝑡𝑡 − 2𝑇𝑇) + 6.375𝛿𝛿(𝑡𝑡 − 3𝑇𝑇) + ⋯
b. Partial Fraction method
Partial fraction permits expanding a complex z-function 𝐹𝐹(𝑧𝑧) into a sum of simpler functions
that are the z-transforms of known discrete-time functions. However, it is often convenient to
expand 𝐹𝐹(𝑧𝑧)/𝑧𝑧 rather than 𝐹𝐹(𝑧𝑧).
Procedure:
i. Find the partial fraction expansion of 𝐹𝐹(𝑧𝑧)/𝑧𝑧 or 𝐹𝐹(𝑧𝑧).
ii. Obtain the inverse transform 𝑓𝑓(𝑘𝑘) from the z-transform tables.
Case I: Real and simple roots
Example 2.14
Obtain the inverse z-transform of the function
𝑧𝑧 + 1
𝐹𝐹(𝑧𝑧) =
𝑧𝑧 2 + 0.3𝑧𝑧 + 0.02
Solution
2
ELE8311: The z-Transform
Let’s first obtain the expansion of 𝐹𝐹(𝑧𝑧)/𝑧𝑧
𝐹𝐹(𝑧𝑧) 𝑧𝑧 + 1
= 2
𝑧𝑧 𝑧𝑧(𝑧𝑧 + 0.3𝑧𝑧 + 0.02)
𝐴𝐴 𝐵𝐵 𝐶𝐶
= + +
𝑧𝑧 𝑧𝑧 + 0.1 𝑧𝑧 + 0.2
Thus,
50𝑧𝑧 90𝑧𝑧 40𝑧𝑧
𝐹𝐹(𝑧𝑧) = − +
𝑧𝑧 𝑧𝑧 + 0.1 𝑧𝑧 + 0.2
From the z-transform tables, we have
𝑘𝑘 𝑘𝑘
𝑍𝑍 −1 = 𝑓𝑓(𝑘𝑘) = �50𝛿𝛿(𝑘𝑘) − 90(−0.1) + 40(−0.2) , 𝑘𝑘 ≥ 0
0 𝑘𝑘 < 0
Since 𝑓𝑓(0) = 0, the time sequence can be rewritten as
𝑘𝑘 𝑘𝑘
𝑓𝑓(𝑘𝑘) = �−90(−0.1) + 40(−0.2) , 𝑘𝑘 ≥ 1
0 𝑘𝑘 < 1
Exercise 2.2
Verify that the above sequence can also be obtained by directly expanding 𝐹𝐹(𝑧𝑧) rather than
𝐹𝐹(𝑧𝑧)/𝑧𝑧.
Exercise 2.3
Determine the time sequence corresponding to the following function
𝑧𝑧
𝐹𝐹(𝑧𝑧) =
(𝑧𝑧 + 0.1)(𝑧𝑧 + 0.2)(𝑧𝑧 + 0.3)
Ans:
50(−0.1)𝑘𝑘 − 100(−0.2)𝑘𝑘 + 50(−0.3)𝑘𝑘 , 𝑘𝑘 ≥ 0
𝑓𝑓(𝑘𝑘) = �
0, 𝑘𝑘 < 0
Case II: Complex Conjugate and Simple Real Roots
3
ELE8311: The z-Transform
If 𝐹𝐹(𝑧𝑧) has real coefficients with complex roots which normally occur in complex conjugate
pairs. Such roots can be combined to yield a function with real coefficients and a quadratic
denominator. In general the z-transform pair of such function is given by:
𝑒𝑒 −𝑎𝑎 sin(𝑤𝑤𝑑𝑑 ) 𝑧𝑧
𝑍𝑍{𝑒𝑒 −𝑎𝑎𝑎𝑎 sin(𝑘𝑘𝑤𝑤𝑑𝑑 )} =
𝑧𝑧 2 − 2𝑒𝑒 −𝑎𝑎 cos(𝑤𝑤𝑑𝑑 ) 𝑧𝑧 + 𝑒𝑒 −2𝑎𝑎
𝑧𝑧[𝑧𝑧 − 𝑒𝑒 −𝑎𝑎 cos(𝑤𝑤𝑑𝑑 )]
𝑍𝑍{𝑒𝑒 −𝑎𝑎𝑎𝑎 cos(𝑘𝑘𝑤𝑤𝑑𝑑 )} =
𝑧𝑧 2 − 2𝑒𝑒 −𝑎𝑎 cos(𝑤𝑤𝑑𝑑 ) 𝑧𝑧 + 𝑒𝑒 −2𝑎𝑎
Notice that the denominators are identical with complex conjugate roots.
To obtain the partial fraction expansion of such functions we use the residue approach. With
complex conjugate poles we expand:
𝐴𝐴𝐴𝐴 𝐴𝐴∗ 𝑧𝑧
𝐹𝐹(𝑧𝑧) = + .
𝑧𝑧 − 𝑝𝑝 𝑧𝑧 − 𝑝𝑝∗
Then, the inverse z-transform yields
𝑓𝑓(𝑘𝑘) = 𝐴𝐴𝑝𝑝𝑘𝑘 + 𝐴𝐴∗ 𝑝𝑝∗𝑘𝑘
= |𝐴𝐴||𝑝𝑝|𝑘𝑘 �𝑒𝑒 𝑗𝑗 �𝜃𝜃𝑝𝑝 𝑘𝑘+𝜃𝜃 𝐴𝐴 � + 𝑒𝑒 −𝑗𝑗 �𝜃𝜃𝑝𝑝 𝑘𝑘+𝜃𝜃 𝐴𝐴 � �
∴ 𝑓𝑓(𝑘𝑘) = 2|𝐴𝐴||𝑝𝑝|𝑘𝑘 cos�𝜃𝜃𝑝𝑝 𝑘𝑘 + 𝜃𝜃𝐴𝐴 �,
where 𝜃𝜃𝑝𝑝 and 𝜃𝜃𝐴𝐴 are the angle of the pole 𝑝𝑝 and the angle of the partial fraction coefficient 𝐴𝐴,
respectively.
Example 2.15
Determine the discrete-time sequence corresponding to the following function.
𝑧𝑧 3 + 2𝑧𝑧 + 1
𝐹𝐹(𝑧𝑧) =
(𝑧𝑧 − 0.1)(𝑧𝑧 2 + 𝑧𝑧 + 0.5)
Solution
Completing the squares of the denominator yields
𝐹𝐹(𝑧𝑧) 𝑧𝑧 3 + 2𝑧𝑧 + 1
=
𝑧𝑧 𝑧𝑧(𝑧𝑧 − 0.1)[(𝑧𝑧 + 0.5)2 + 0.52 ]
4
ELE8311: The z-Transform
𝐴𝐴 𝐵𝐵 𝐶𝐶 𝐶𝐶 ∗
= + + +
𝑧𝑧 𝑧𝑧 − 0.1 𝑧𝑧 + 0.5 − 𝑗𝑗0.5 𝑧𝑧 + 0.5 − 𝑗𝑗0.5
Therefore, substituting the coefficients yields
19.689𝑧𝑧 (0.656 + 𝑗𝑗2.213)𝑧𝑧 (0.656 − 𝑗𝑗2.213)𝑧𝑧
𝐹𝐹(𝑧𝑧) = −20 + + +
𝑧𝑧 − 0.1 𝑧𝑧 + 0.5 − 𝑗𝑗0.5 𝑧𝑧 + 0.5 + 𝑗𝑗0.5
Converting the coefficient 𝐶𝐶 from Cartesian to Polar form yields
19.689𝑧𝑧 2.308𝑒𝑒 𝑗𝑗 1.283 𝑧𝑧 2.308𝑒𝑒 −𝑗𝑗 1.283 𝑧𝑧
𝐹𝐹(𝑧𝑧) = −20 + + +
𝑧𝑧 − 0.1 𝑧𝑧 + 0.5 − 𝑗𝑗0.5 𝑧𝑧 + 0.5 + 𝑗𝑗0.5
The inverse z-transform is then
3𝜋𝜋𝜋𝜋
𝑓𝑓(𝑘𝑘) = −20𝛿𝛿(𝑘𝑘) + 19.689(0.1)𝑘𝑘 + 4.616(0.707)𝑘𝑘 cos � + 1.283� .
4
Case III: Repeated Roots
For a function 𝐹𝐹(𝑧𝑧) with a repeated root of multiplicity 𝑟𝑟, 𝑟𝑟 partial fraction coefficients are
associated with the repeated root.
𝑟𝑟
𝑁𝑁(𝑧𝑧) 𝐴𝐴1𝑖𝑖
𝐹𝐹(𝑧𝑧) = = � ,
(𝑧𝑧 − 𝑧𝑧1 )𝑟𝑟 (𝑧𝑧 − 𝑧𝑧1 )𝑟𝑟+1−𝑖𝑖
𝑖𝑖=1
where the coefficients for repeated roots are given by:
1 𝑑𝑑𝑖𝑖−1
𝐴𝐴1𝑖𝑖 = (𝑧𝑧 − 𝑧𝑧1 )𝑟𝑟 𝐹𝐹(𝑧𝑧)� , 𝑖𝑖 = 1,2, … , 𝑟𝑟.
(𝑖𝑖 − 1)! 𝑑𝑑𝑧𝑧 𝑖𝑖−1
𝑧𝑧→𝑧𝑧
1
Example 2.16
Determine the inverse z-transform of the function
1
𝐹𝐹(𝑧𝑧) =
𝑧𝑧 2 (𝑧𝑧 − 0.5)
Solution
Partial fraction expansion gives:
𝐹𝐹(𝑧𝑧) 1 𝐴𝐴11 𝐴𝐴12 𝐴𝐴13 𝐴𝐴4
= 3 = 3 + 2 + +
𝑧𝑧 𝑧𝑧 (𝑧𝑧 − 0.5) 𝑧𝑧 𝑧𝑧 𝑧𝑧 𝑧𝑧 − 0.5
5
ELE8311: The z-Transform
Then,
𝐹𝐹(𝑧𝑧) 1
𝐴𝐴11 = 𝑧𝑧 3 � = � = −2
𝑧𝑧 𝑧𝑧=0 𝑧𝑧 − 0.5 𝑧𝑧=0
1 𝑑𝑑 3 𝐹𝐹(𝑧𝑧) 𝑑𝑑 1 1
𝐴𝐴12 = 𝑧𝑧 � = � =− � = −4
1! 𝑑𝑑𝑑𝑑 𝑧𝑧 𝑧𝑧=0 𝑑𝑑𝑑𝑑 𝑧𝑧 − 0.5 𝑧𝑧=0 (𝑧𝑧 − 0.5)2 𝑧𝑧=0
1 𝑑𝑑 2 3 𝐹𝐹(𝑧𝑧) 1 𝑑𝑑 −1 1 −1(−2)
𝐴𝐴13 = 𝑧𝑧 � =� � � =� � � = −8
2! 𝑑𝑑𝑧𝑧 2 𝑧𝑧 𝑧𝑧=0 2 𝑑𝑑𝑑𝑑 (𝑧𝑧 − 0.5) 𝑧𝑧=0
2 2 (𝑧𝑧 − 0.5)3 𝑧𝑧=0
𝐹𝐹(𝑧𝑧) 1
𝐴𝐴4 = (𝑧𝑧 − 0.5) � = 3� =8
𝑧𝑧 𝑧𝑧=0.5 𝑧𝑧 𝑧𝑧=0.5
Hence,
1 8𝑧𝑧
𝐹𝐹(𝑧𝑧) = = − 2𝑧𝑧 −2 − 4𝑧𝑧 −1 − 8
𝑧𝑧 2 (𝑧𝑧 − 0.5) 𝑧𝑧 − 0.5
From the z-transform table
8(0.5)𝑘𝑘 − 2𝛿𝛿(𝑘𝑘 − 2) − 4𝛿𝛿(𝑘𝑘 − 1) − 8𝛿𝛿(𝑘𝑘), 𝑘𝑘 ≥ 0
𝑓𝑓(𝑘𝑘) = �
0, 𝑘𝑘 < 0
Evaluating 𝑓𝑓(𝑘𝑘) at 𝑘𝑘 = 0,1,2 yields
𝑓𝑓(0) = 8 − 8 = 0
𝑓𝑓(1) = 8(0.5) − 4 = 0
𝑓𝑓(2) = 8(0.5)2 − 2 = 0
Thus, the time sequence can be rewritten as the following delayed sequence:
(0.5)𝑘𝑘−3 , 𝑘𝑘 ≥ 3
𝑓𝑓(𝑘𝑘) = �
0, 𝑘𝑘 < 3
Note that the solution can be obtained directly using the delay theorem without the need for
partial fraction expansion since 𝐹𝐹(𝑧𝑧) can be written as:
𝑧𝑧
𝐹𝐹(𝑧𝑧) = 𝑧𝑧 −3 .
𝑧𝑧 − 0.5
6
ELE8311: The z-Transform
2.6 z-Transform Solution of Difference Equations
Similar to solution of differential equations with Laplace transformation in time domain, z-
transform can also be used to solve difference equations in discrete domain. The preceding
properties of z-transform play a key role in this regard.
Example 2.17
For initial conditions 𝑥𝑥(0) = 1, 𝑥𝑥(1) = 5/2, solve the linear difference equation
3 1
𝑥𝑥(𝑘𝑘 + 2) − � � 𝑥𝑥(𝑘𝑘 + 1) + � � 𝑥𝑥(𝑘𝑘) = 1(𝑘𝑘).
2 2
Solution
Using the time-advance property, we determine the z-transform as
3 1 𝑧𝑧
[𝑧𝑧 2 𝑋𝑋(𝑧𝑧) − 𝑧𝑧 2 𝑥𝑥(0) − 𝑧𝑧𝑧𝑧(1)] − [𝑧𝑧𝑧𝑧(𝑧𝑧) − 𝑧𝑧𝑧𝑧(0)] + 𝑋𝑋(𝑧𝑧) =
2 2 𝑧𝑧 − 1
Solving for 𝑋𝑋(𝑧𝑧) yields
𝑧𝑧[1 + (𝑧𝑧 + 1)(𝑧𝑧 − 1)] 𝑧𝑧 3
𝑋𝑋(𝑧𝑧) = =
(𝑧𝑧 − 1)(𝑧𝑧 − 1)(𝑧𝑧 − 0.5) (𝑧𝑧 − 1)2 (𝑧𝑧 − 0.5)
Now, partial fraction expansion of 𝑋𝑋(𝑧𝑧)/𝑧𝑧 is
𝑋𝑋(𝑧𝑧) 𝑧𝑧 2 𝐴𝐴11 𝐴𝐴12 𝐴𝐴3
= = + +
𝑧𝑧 (𝑧𝑧 − 1)2 (𝑧𝑧 − 0.5) 𝑧𝑧 − 1 (𝑧𝑧 − 1)2 𝑧𝑧 − 0.5
Thus, 𝐴𝐴11 = 0, 𝐴𝐴12 = 2, and 𝐴𝐴3 = 1.
2𝑧𝑧 𝑧𝑧
𝑋𝑋(𝑧𝑧) = +
(𝑧𝑧 − 1)2 𝑧𝑧 − 0.5
Now, the inverse transform from the z-transform table is
𝑥𝑥(𝑘𝑘) = 2𝑘𝑘 + (0.5)𝑘𝑘 .
Notice that, by applying z-transform to the difference equation, a closed-form solution of the
LTI difference equation is obtained.
7
ELE8311: The z-Transform
Exercise 2.4
Determine a closed-form solution for the following difference equation given the initial values
𝑦𝑦(−2) = 1, 𝑦𝑦(−1) = −2 and the driving force function 𝑢𝑢(𝑘𝑘) is a unit-step sequence.
3𝑦𝑦(𝑘𝑘) + 2𝑦𝑦(𝑘𝑘 − 1) − 𝑦𝑦(𝑘𝑘 − 2) = 2𝑢𝑢(𝑘𝑘 − 1) − 3𝑢𝑢(𝑘𝑘 − 2).
19 1 𝑘𝑘 9 1
Ans: 𝑦𝑦(𝑘𝑘) = 24 �3� + 8 (−1)𝑘𝑘 − 4 (1)𝑘𝑘 ; 𝑘𝑘 = 0,1,2, …
