Documente Academic
Documente Profesional
Documente Cultură
1
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
analog signal by using a digital–to–analog converter and hold circuit, which actuates the
actuator. Then the actuator controls the input to the plant so as to give the desired output.
The operation that transforms continuous time signals into discrete time data is called
sampling or discretization. The reverse-operation, which transforms discrete-time data into a
continuous–time signal, is called data hold. The A/D conversion process is called coding or
encoding. The D/A conversion process is called decoding.
4. Principle of signal conversion
2
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
The signal shown in fig.2b can therefore be stored, if the sampled values are quantified to
four quantization levels. Fig.2c shows a quantified discrete–time signal. It is assumed that
any value between 0.5 and 1.5 is round to 1. The signals, for which both time and amplitude
are discrete, are called digital signals.
After sampling and quantization, it is converted into a form acceptable by the
computers, which is called as coding or encoding. Fig.2d gives coded digital signal
corresponding to the analog signal in fig.2a for a 2-bit machine .
The resolution of the converter can be improved by the number of digits. In an 8-bit
A/D converter an input signal can be resolved to one part in 28, or 1 in 256. If the input signal
has a range of 10V, the resolution is 10/256 or approximately 0.04V. Thus it is enough to
change at least by 0.04V, to make a change in the output.
5. Types of sampling
A sampling operation is basic in transforming a continuous–time signal into a discrete-time
signal. The different types of sampling are:
1. Periodic sampling: Here the sampling instants are equally spaced, i.e, tk = KT
(k = 0,1,2,...). Periodic sampling is the most conventional type of sampling operation.
2. Multiple order sampling: The pattern as the tk is repeated periodically, that is, tk+r - tk is
constant for all K.
3. Multiple rate sampling: In multiple loop control systems, the loop which has largest time
constant needs slow sampling rate, where as the loop which has smaller time constant needs
faster sampling. A digital control system many have different sampling periods for different
feedback paths or may have multiple sampling rates.
4. Random sampling: Here the sampling instants are random in nature or tk is a random
variable.
6. Impulse sampling
3
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
The sampling process is equivalent to multiplying the analog signal X(t), with an unit
impulse train to produce the sampled signal as shown in fig 3a.
χ * (t) = x(t ) δ (t − kT ) k=0, 1,2,3,4,………
*
Where χ (t) = impulse sampled signal
χ(t) = analog signal
δ(t-KT) = impulse train
∞
At t = kT, χ * (t) = ∑ χ (kT).δ (t − kT)
k =0
*
where X (t) is the sampled signal, which can be written as
χ*(t) = χ (0) δ(t) + χ (T) δ (t – T)... + χ (kT) (t – kT) + .......... (1)
Let the train of unit impulses be defined as
∞
δ T (t) = ∑ δ (t − kT)
k =0
Since the output is equal to the product of the continuous–time input χ(t) and the train of unit
impulses δT(t), the sampler may be considered as a modulator with the input χ(t) of the
modulator signal and the train of unit impulses δT(t) as the carrier as shown in fig.4
1
Let eTS = z (or) s=
ln z
T
Substituting in equation (2), then the equation becomes,
∞ 1 ∞
− kT. .ln z
χ * (s) = ∑ χ (kT) e
k =0
T
= ∑ χ (kT). Z
k =0
−k
(3)
It is the z-transform of the sequence χ(0), χ(T), χ(2T),... generated from χ(t) at t = kT where k
= 0,1,2,.... Hence we may write
1
At s = ln z
T
χ *(s) = χ (z)
∞
⎛1 ⎞
Equation (3) become, χ * (s) = χ * ⎜ ln z ⎟ = χ (z) = ∑ χ (kT) z − k (4)
⎝T ⎠ k =0
4
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
7. Data–hold circuits
Data-hold is a process of generating a continuous-time signal h(t) from a discrete-time
sequence χ(kT). A hold circuits converts the sampled signal into a continuous-time signal,
which approximately reproduces the signal applied to the samples. The data-hold circuit
output signal h(t) during the time interval kT ≤ t < (k+1)T may be approximated by a
polynomial in τ as follows:
h(kT + τ) = an τn + an-1 τn-1 + ... + a1 τ + a0 (5)
5
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
8. Zero-order hold
Since the integral of an impulse function is a constant, we may assume that the zero-order
hold is an integrator and the input to the zero-order hold circuit is a train of impulses.
Consider the sampler and zero-order hold shown in fig 5a. Then the output h1(t) is related to
χ(t)as follows:
h1(t) = χ (o) [1(t) – 1 (t – T)] + χ(T) [1(t – T) – 1 (t – 2T)]
+ χ(2T) [1(t – 2T) – 1 (t – 3T) + ......
∞
= ∑ χ (kT )[1(t − kT ) − 1(t − (k + 1)T ]
k =0
(8)
Since
e − kTS
Laplace transform £ [1(t − kT ) =
s
The Laplace transform of equation (8) becomes
∞
e − kTS − e − ( k +1)TS
£ [h1 (t )] = H 1 ( s ) = ∑ χ (kT ).
k =0 s
∞
e − kTS − e − kTS .e −TS
= ∑ χ (kT ).
k =0 s
1 − e −TS ∞
= .∑ χ (kT ).e − kTS (9)
s k =0
∞
But ∑ χ (kT ).e
k =0
− kTS
= χ ( s ) [Laplace transform of input]
H 1 ( s ) 1 − e −TS
∴ = .
χ ( s) s
1 − e −Ts
Therefore transfer function of zero-order hold = (10)
s
8.1 Frequency response characteristic of zero-order hold.
1 − e − sT
The transfer function of ZOH GOh ( s ) =
s
Substituting s = jω
1 − e − jω T
The sinusoidal transfer function of ZOH Goh (jω) = (10a)
jω
6
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
− jω T jωT
Since e 2
.e 2
= 1 , equation (10a) becomes,
− jω T jωT − jω T jωT − jωT − jωT
e 2
.e 2
− e − jωT e 2
.e 2
−e 2
.e 2
Goh ( jω ) = =
jω jω
⎡ jω2T − jωT
⎤ − jωT
e −e 2 ⎥ 2
= ⎢ e (10b)
⎢ jω ⎥
⎣⎢ ⎦⎥
Since
e jθ − e − jθ
Sinθ = Equation (10b) becomes,
2j
− jωT
2 ωT
Goh ( jω ) = . sin .e 2
… (10c)
ω 2
ωT
Multiply the numerator and denominator by . Equation (10c) become
2
ωT
T . sin − jωT
Goh ( jω ) = 2 .e 2
(10d)
ωT
2
2π 2π
Since the sampling frequency ωs = , on substituting T = , equation (10d) becomes
T ωs
⎛ πω ⎞
sin ⎜⎜ ⎟ − jπω
2π ⎝ ω s ⎟⎠ ωs
Goh ( jω ) = . .e (10e)
ω s ⎛ πω ⎞
⎜ ⎟
⎝ ωs ⎠
The magnitude part and phase angle part of equation (10e) is
⎛ πω ⎞
sin ⎜⎜ ⎟
2π ⎝ ω s ⎟⎠
Goh ( jω ) = . ,
ω s ⎛ πω ⎞
⎜⎜ ⎟⎟
ω
⎝ s⎠
− πω
angle ∠Goh ( jω ) = .
ωs
The frequency response curves for the zero-order hold and the equivalent Bode diagram.
When T = 1 sec is shown in fig. 6.
Note that the magnitude curve approaches -∞ decibels of frequency points that are integral
multiples of sampling frequency. ωs = 2π/r = 6.28 rad/sec. Discontinuous of the phase curve
of ± 180° occurs at these frequency points.
7
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
The frequency contents of a signal can be obtained from the fourier transform of the signal.
Hence the frequency response of the impulse sampled signal can be obtained by taking
fourier transform of equation (11).
α
− jωt
f [ x (t )] = F * (ω ) = ∫ x ∗ (t ).e
∗
dt
−α
α α
∫ x (t ).∑ δ (t − kT ) e
− jω t
= . dt (12)
−α −α
Mathematically the equation (12) is the convolution of χ(t) and δ(t – kT). By convolution
theorem of fourier transform, the convolution of two time domain signals is equivalent to the
produce of their individual Fourier transform. Therefore
1 ⎡ α ⎤
∴ F * (ω ) = .F ( x(t )).F ⎢ ∑ δ (t − kT )⎥ (13)
2π ⎣ k = −α ⎦
Let F(χ(t)) = F(ω). (14)
⎡ α ⎤ α
and F ⎢ ∑ δ (t − kT )⎥ = ωs ∑ δ (ω − kωs ) (15)
⎣k = −α ⎦ k = −α
2π
Where ω S = = Sampling frequency in rad/sec.
T
Substituting equation (14) and (15) in equation (13), the equation becomes
8
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
α
1
F * (ω ) = .F (ω ) × ω s ∑ δ (ω − kω s ).
2π k =α
1 2π α
= . . ∑ F (ω ).δ (ω − kω s ) .
2π T k = −α
Since F(ω). δ(ω - kωs) = F(ω - kωs).
1 α
F * (ω ) = ∑ F (ω − kωs ).
T k = −α
(16)
The equation (16) gives the frequency spectrum of the impulse sampled signal.
Figure 7 Fourier spectra of input signal and its impulse sampled version.
If F(ω) is a band-limited signal with a maximum frequency ωm, the frequency spectrum of
F(ω) is shows in fig 7a, which is plot of | F(ω) | VS ω. The frequency spectrum of impulse
sampled signal is shown in fig 7b, which is a plot of | F*(ω) | VS ω, when ωs > 2 ωm. Fig 7c is
a frequency spectrum plot when ωs < 2 ωm.
Shanon’s sampling theorem: Shanon’s sampling theorem states that a band limited
continuous time signal with highest frequency ωm, can be uniquely recovered from its
samples provided the sampling rate ωs is greater than or equal to 2ωm samples per second. (In
practice ωs should be 10–20 times of ωm).
9. The z-transform
The z transform of a time function χ(t), where ‘t’ is non-negative, or of sequence of values
χ(kT), where k takes zero or positive integers and T is the sampling period is defined by the
following equation:
α
X ( z ) = Z [( x(t )] = Z [ x(kT ) = ∑ x(kT ) z − k (17)
k =0
For sequence of numbers x(k), the z transform is defined by
∞
X ( z ) = Z [ x(k )] = ∑ x(k ) z − k (18)
k =0
10. Transforms of elementary functions
9
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
10
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
11
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z. sin ωT
X(z) = 2 (31)
z − 2 z cos ωT + 1
10.6 Cosine function
⎧ cos ωt t ≥ 0
x(t ) = ⎨ (32)
⎩ 0 t<0
1
X(z) = Z[cos ωt] = Z ( e jωt + e − jωt )
2
1⎡ 1 1 ⎤
= ⎢ +
2 ⎣1 − e j ω T
z −1
1− e − j ω T
z ⎥⎦
−1
1⎡ z z ⎤
= ⎢ ω
+ − ω ⎥
2 ⎣z − e j t
z−e j T
⎦
z ⎡ z − e − jωT + z − e jωT ⎤
= ⎢ ⎥
2 ⎣ ( z − e jωT )( z − e − jωT ) ⎦
z 2 − z (e jωT + e − jωT ) / 2
=
z 2 − ze jωT − ze − jωT + e − jωT e jωT
z 2 − z cos ωT z ( z − cos ωT )
= 2 jω T − jω T
= 2
(33)
z − z (e +e ) +1 z − 2 z cos ωT + 1
Table of z transforms
X(s) x(t) x(kT) or x(k) X(z)
1. - - Kronecker delta 1
δ0(k)
1 k=0
0, k≠0
2. - - δ0(n-k) z-k
1, n=k
0, n≠k
3. 1 1(t) 1(k) 1
s 1 − z −1
4. 1 e-at e-akT 1
s+a − aT −1
1− e z
5. 1 t kT Tz −1
2
s (1 − z −1 ) 2
6. 2 t2 (kT)2 T 2 z −1 (1 + z −1 )
s3 (1 − z −1 ) 3
12
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
7. 6 t3 (kT)3 T 3 z −1 (1 + 4 z −1 + z −2 )
s4 (1 − z −1 ) 4
8. a 1-e-at 1-e-akT (1 − e − aT ) z −1
s( s + a)
(1 − z −1 )(1 − e −aT z −1 )
9. b−a e-at – e-bt e-akT – e-bkT (e − aT − e −bT ) z −1
( s + a )( s + b)
(1 − e −aT z −1 )(1 − e −bT z −1 )
13
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
27. k (k − 1) k −2 z −2
a
2!
(1 − az −1 ) 3
Example 1
Obtain the z-transfer of
1
X ( s) =
s ( s + 1)
Solution
By partial fraction
1 A B
= +
s ( s + 1) s s + 1
1 1
A= × s s =0 = s =0 = 1
s ( s + 1) s +1
1 1
B= × ( s + 1) s = −1 = s = −1 = −1
s ( s + 1) s
Substituting
⎡1 1 ⎤
X ( s) = ⎢ −
⎣ s s + 1⎥⎦
The inverse Laplace transform of X(s) is
x(t) = 1 – e-t
1 1
X ( z ) = Z [1 − e −t ] = −1
−
1− 2 1 − e −T z −1
z z
= −
z −1 z − e −T
z[ z − e −T − z + 1] z (1 − e −T )
= =
( z − 1)( z − e −T ) ( z − 1)( z − e −T )
14
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Since x(mT) = 0 for m < 0 we may change the lower limit of the summation from m = -n to
m = 0, then
∞
Z [ x(t − nT )] = z − n ∑ x(mT ) z − m = z − n X ( z )
m=0
Multiplication of a z-transform by z -n has the effect of delaying the time function x(t) by time
nT. Further, by comparing equation (40)
∞
Z [ x(t + nT )] = ∑ x(kT + nT ) z −k
k =0
∞
= zn ∑ x(kT + nT ) z −( k + n)
k =0
15
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∞ n −1 n −1
= zn[ ∑ x(kT + nT ) z −( k +n) + ∑ x(kT ) z −k − ∑ x(kT ) z −k ]
k =0 k =0 k =0
⎡∞ n −1 ⎤
= z n ⎢ ∑ x( kT ) z − k − ∑ x( kT ) z − k ⎥
⎣k =0 k =0 ⎦
⎡ n −1 ⎤
= z n ⎢ x( z ) − ∑ x(kT ) z − k ⎥ .
⎣ k =0 ⎦
Example 2
Find the z transform of unit step functions that are delayed by 1 sampling period and 4
sampling periods, as shown in fig. 8a and 8b.
16
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∞
y (k ) = ∑ x(h), k = 0,1,2,...
h =0
Solution
y(k) = χ(o) + χ(1) + .... χ(k-1) + (k)
y(k – 1) = x(0) + x(1) + ... x(k-1)
Therefore
y(k) – y(k – 1) = x(k) .... k = 0, 1, 2.....
Z[y(k) – y(k – 1)] = Z[x(k)]
Y(z) – z-1 Y(z) = χ (z).
Y(z) [1 – z-1] = χ (z)
1
y(z) = . χ (z).
1 − z −1
11.6 Complex translation theorem
If χ(z) is the z-transfer of χ(t), then χ(zeaT) is the z transfer of e-at χ(t).
∞ ∞
Z [e − at x(t )] = ∑ x(kT )e − akT .z − k =∑ x(kT )( ze aT ) − k
k =0 k =0
= χ(zeaT)
Example 3
Obtain the z transfer of e-at sin ωt and e-at cos ωt.
we know that
z -1 . sin ωT
Z [sin ωt] =
1 − 2 z −1 cos ωT + z − 2
e − aT .z −1 . sin ωT
Z [e − at . sin ωt ] = .
1 − 2e − aT z −1 . cos ωT + e − 2 aT .z − 2
Similarly
1 − z −1 . cos ωT
Z [cos ωt ] =
1 − 2 z −1 . cos ωT + z − 2
1 − e − aT z −1 . cos ωT
Z [e − at . cos ωt ] =
1 − 2e − aT .z −1 cos ωT + e − 2 aT z − 2
Example 4
Obtain the z-transfer of t e-at
Tz −1
Z [t ] =
(1 − z −1 ) 2
Then
− at Te − aT .z −1
Z [te ] =
(1 − e − aT .z −1 ) 2
11.7 Initial value theorem
If χ(z) is the z-transfer of χ(t), then the initial value x(0) is given by
x(0) = lim X ( z ) (42)
z →∞
Proof
∞
x( z ) = ∑ x(k ).z − k = x(0) + x(1) z −1 + x(2) z − 2 + .... (43)
k =0
17
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∴ x(0) = lim X ( z )
z →∞
Example 4
Find the initial value χ(0), if the z-transfer of x(t) is given by
(1 − e −T ) z −1
x( z ) =
(1 − z −1 )(1 − e −T z −1 )
By initial value theorem
(1 − e −T ) z −1
x(0) = lim =0
z →∞ (1 − z −1 )(1 − e −T z −1 )
11.8 Final value theorem
lim x(k ) = lim[(1 − z −1 ) X ( z )] (44)
k →∞ z →1
Proof
∞
z[ x(k )] = X ( z ) = ∑ x(k ) z − k (45)
k =0
∞
z[ x(k − 1)] = z −1 X ( z ) = ∑ x(k − 1) z − k (46)
k =0
Subtracting equation (46) from equation (45).
∞ ∞
∑ x(k ) z
k =0
−k
− ∑ x(k − 1) z − k = x( z ) − z −1 X ( z )
k =0
Taking the limit as z approaches unity on both sides
⎡∞ ⎤
[ ]
∞
lim ⎢∑ x(k ) z − k − ∑ x(k − 1) z − k ⎥ = lim (1 − z −1 ) X ( z )
z →1
⎣ k =0 k =0 ⎦ z →1
Applying the limit and expanding the left side of the above equation
∞
∑ [ x(k ) − x(k − 1)] = [ x(0) − x(−1)] + [ x(1) − x(0)] + [ x(2) − x(1)] + ....... = x(∞) = lim x(k )
k =0
k →∞
18
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
9. x(n – k) z-kX(z)
10. tx(t) d
− Tz X (z )
dz
11. kx(k) d
−z X (z )
dz
12. e-at x(t) X(zeaT)
13. e-ak x(k) X(zea)
14. ak x(k) ⎛z⎞
X⎜ ⎟
⎝a⎠
20. n 1
∑ x(k ) 1 − z −1
X ( z)
k =0
21. ∂ ∂
x(t , a ) X ( z, a)
∂a ∂a
22. kmx(k) m
⎛ d ⎞
⎜ − z ⎟ X (z )
⎝ dz ⎠
24. ∞ X(1)
∑ x(k )
k =0
Example 5
Determine the final value x(∞) for the function
1 1
x( z ) = −1
− − aT −1
for a > 0
1− z 1− e z
19
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Solution
[
x(∞) = lim (1 − z −1 ) X ( z )
z →1
]
⎡ ⎛ 1 1 ⎞⎤
= lim ⎢(1 − z −1 )⎜ − ⎟
z →1
⎣ ⎝1− z
−1
1 − e z ⎠⎥⎦
− aT −1
⎡ 1 − z −1 ⎤
= lim ⎢1 − − aT −1 ⎥
=1
z →1
⎣ 1− e z ⎦
Example 6
Obtain the z-transfer of Gk, where G is an n × n constant matrix.
Solution
∞
Z [G k ] = ∑ G k .z − k
k =0
= I + Gz −1 + G 2 z −2 + G 3 z −3 + ...
= ( I − Gz −1 ) −1
= (ZI-G)-1 Z.
Example 7
Find the z-transform of (i) k2 (ii) kak-1
Solution
(i) x(k) = k2
Multiply the given discrete sequence by a discrete unit-step sequence
x(k) = k2 x(k)
where x(k) is unit-step sequence.
By the property of z – transfer.
m
⎡ d⎤
Z [k x(k )] = ⎢− z ⎥ X ( z )
m
⎣ dz ⎦
z
Z [ x(k )] = (unit − step)
z −1
d ⎡ d ⎡ z ⎤⎤ ⎡ z −1− z ⎤ z
− z X ( z) = − z ⎢ ⎢ ⎥ ⎥ = −z⎢ 2 ⎥
=
⎣ dz ⎣ z − 1⎦ ⎦ ⎣ ( z − 1) ⎦ ( z − 1)
2
dz
2
⎡ d⎤ d ⎡ d ⎤
⎢⎣− z dz ⎥⎦ X ( z ) = − z dz ⎢⎣− z dz X ( z )⎥⎦
d ⎡ z ⎤ ⎡ ( z − 1) 2 − z.2( z − 1) ⎤
= −z ⎢ ⎥ = − z ⎢ ⎥
dz ⎣ ( z − 1) 2 ⎦ ⎣ ( z − 1) 4 ⎦
⎡ (z − 1)(z − 1 − 2z) ⎤ ⎡ − (z + 1) ⎤
= − z⎢ ⎥ = −z ⎢ 3 ⎥
⎣ (2 − 1) 4
⎦ ⎣ (z − 1) ⎦
z(z + 1)
=
(z − 1)3
(ii) x(k) = kak-1
z
We know that Z[a k ] =
z−a
Or
20
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∞
z
∑a z
k =0
k −k
=
z−a
Expending the left side of the equation
z
1 + az −1 + a 2 z − 2 + a 3z − 3 + ... =
z−a
Differentiating both sides of the above equation with respect to z
( z − a) × 1 − z × 1 −a
− az − 2 − 2a 2 z −3 − 3a 3 z − 4 .... = =
( z − a) 2
( z − a) 2
On multiplying both side by (-z/a), we get
z
z −1 + 2az − 2 + 3a 2 z −3 + ... =
( z − a) 2
The left side of the infinite series can be expressed as summation and can be written as
follows.
∞
z
∑
k =0
ka k −1.z − k =
(z − a ) 2
This is
z
Z[ka k −1 ] =
(z − a ) 2
Example 8
⎡k ⎤ 1
Show that (i) Z ⎢∑ x ( h) ⎥ = −1
X ( z)
⎣ h =0 ⎦ 1− z
⎡ h −1 ⎤ z −1
(ii) Z ⎢∑ x ( h) ⎥ = −1
X ( z)
⎣ h =0 ⎦ 1− z
Solution
k
(i) Let y(k ) = ∑ x (h ) k = 0, 1, 2, 3…. (48)
h =0
So that
y(0) = x(0)
y(1) = x(0) + x(1)
y(2) = x(0) + x(1) + x(2)
y(k) = x(0) + x(1) + x(2) + …. + x(k)
Then
Y(k) – y(k – 1) = x(k)
Taking z-transform respectively
Y(z) – z-1Y(z) = X(z)
Y(z) [1 – z-1] = X(z)
1
Y(z) = X (2)
1 − z −1
⎡k ⎤ 1
∴ Z ⎢∑ x(h)⎥ = Z [ y (k )] = Y ( z ) = X ( z ). (49)
⎣ h =0 ⎦ 1 − z −1
(ii) By shifting theorem
Z [x(t – nT)] = Z-n X(z)
Therefore
21
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
⎡ k −1 ⎤ 1
Z ⎢∑ x(h)⎥ = Z [ y (k − 1)] = z −1Y ( z ) = X ( z)
⎣ h =0 ⎦ 1 − z −1
Example 9
∞
⎡k ⎤ 1 ⎡ i −1
⎤
Show that (i) ∑ x(k ) = lim X ( z ) and (ii) Z ⎢∑ x(h)⎥ = −1 ⎢ X ( z ) − ∑ x ( h) z − h ⎥
k =0
z →1
⎣ h =i ⎦ 1− z ⎣ h =0 ⎦
Solution
(i) Consider the equation (48)
k
y( k ) = ∑ x ( h ) , k = 0, 1, 2, 3…
h =0
⎡k ⎤
lim y(k ) = lim ⎢∑ x (h )⎥
k →∞ k →∞
⎣h =0 ⎦
By equation (49)
⎡k ⎤ 1
Z ⎢∑ x(h)⎥ = Z [ y (k )] = Y ( z ) = X ( z)
⎣ h =0 ⎦ 1 − z −1
Therefore using final value theorem
⎡k ⎤ ⎡ 1 ⎤
lim y (k ) = lim ⎢∑ x(h)⎥ = lim ⎢(1 − z −1 ). −1
X ( z )⎥
k →∞ k →∞
⎣ h=0 ⎦ z →1
⎣ 1− z ⎦
∞ ∞
(ii)
k
Define y (k ) = ∑ x(h) = x(i) + x(i + 1) + ... + x(k ) where 1 ≤ i ≤ k – 1
h =i
Also define
X ( z ) = x(i) z −1 + x(i + 1) z −1(i +1) + ... + x(k ) z − k + ...
And
∞
X ( z ) = Z [ x(t )] = ∑ x(k ) z − k = x(0) + x(1) z −1Tx(2) z − 2 + ...
k =0
we get
i −1
X ( z ) = X ( z ) − ∑ x ( h) z − h
h =0
Since
y (k ) − y (k − 1) = x(k ) k = i, i+1, i+2,i+3....
The z-transfer of the above equation is
Y ( z ) − z −1 Y ( z ) = X ( z )
Since the z-transform of χ(k) which begins with k = i, is X ( z ) and not = X ( z )
⎡k ⎤ 1 1 ⎡ i −1
⎤
Z ⎢∑ x ( h) ⎥ = Y ( z ) = −1
X = ⎢ X ( z ) − ∑ x ( h) z − h ⎥
⎣ h =i ⎦ 1− z 1 − z −1 ⎣ h =0 ⎦
Example 10
Find the z-transfer of an output with T = 1 sec for the sampled values given below:
x(0) = 0; x(1) = 0.25; x(2) = 0.50; x(3) = 0.75 and x(k) = 1 for k = 4, 5, 6....
22
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∞
We know X ( z ) = ∑ x(k ) z − k
k =0
= 0.25 z-1 + 0.50 z-2 + 0.75 z-3 + z-4 + z-5 + z-6 +.....
⎛ 1 ⎞
= 0.25 (z-1 + 2z-2 + 3z-3) + z-4 ⎜ ⎟
⎝ 1 − z −1 ⎠
z −1 + 2 z −2 + 3z −3 z −4
= +
4 1 − z −1
( z −1 + 2 z −2 + 3z −3 )(1 − z −1 ) + 4 z −4
=
4(1 − z −1 )
z −1 + z −2 + z −3 + z −4
=
4(1 − z −1 )
1 z −1 (1 + z −1 + z −2 + z −3 )(1 − z −1 )
=
4 (1 − z −1 ) 2
1 z −1 (1 − z −4 )
=
4 (1 − z −1 ) 2
12. Inverse z - transform
When X(z), the z- transform of x(kT) or x(k), is given, the operation that determines the
corresponding x(kT) or x(k) is called the inverse z-transformation. Four methods are
available for obtaining the inverse z-transform.
1. Direct division method
2. Computational method
3. Partial-fraction-expansion method
4. Inversion integral method
Out of the above mentioned four methods, direct decision method and partial-fraction-
expansion method are commonly used.
23
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Example 11
Find x(k) for k = 0, 1, 2, 3, 4 when X(z) is
10 z + 5
X ( z) =
( z − 1)( z − 0.2)
Solution
10 z + 5 10 z + 5
Given that X ( z ) = = 2
( z − 1)( z − 0.2) z − 1.2 z + 0.2 z
Multiply both numerator and denominator by z -2, then the equation become
10 z −1 + 5 z −2
X ( z) =
1 − 1.2 z −1 + 0.2 z − 2
Dividing the numerator by the denominator, we have
10 z −1 + 17 z −2 + 18.4 z −3 + 18.68 z −4
10 z −1 + 5 z − 2
1 − 1.2 z −1 + 0.2 z − 2
10 z −1 − 12 z − 2 + 2 z −3
17z -2 − 2 z −3
17z -2 − 20.4 z −3 + 3.4 z − 4
18.4z -3 − 3.4 z − 4
18.4z -3 − 22.08 z − 4 + 3.68 z −5
18.68z -4 − 3.68 z −5
18.68z -3 − 22.416 z −5 + 3.736 z −6
Example 12
1
Determine the inverse z-transform of X ( z ) =
3 −1 1 − 2
1− z + z
2 2
when (a) Roc: |z| > 1.0 and (b) Roc: |z| < 0.5.
Solution
(a) When Roc: |z| > 1.0
1
X ( z) =
3 −1 1 − 2
1− z + z
2 2
24
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
3 −1 7 − 2 15 −3 31 − 4
1+ z + z + z + z + ...
2 4 8 16
1
3 −1 1 − 2
1− z + z 3 1
2 2 1 − z −1 + z − 2
2 2
3 −1 1 − 2
z − z
2 2
3 −1 9 − 2 3 −3
z − z + z
2 4 4
7 − 2 3 −3
z − z
4 4
7 − 2 21 − 3 7 − 4
z − z + z
4 8 8
15 − 3 7 − 4
z − z
8 8
15 − 3 45 − 4 15 −5
z − z + z
8 16 10
31 − 4 15 −5
z − z
16 16
#
Thus
3 −1 7 − 2 15 −3 31 − 4
x( z ) = 1 + z + z + z + z ...
2 4 8 16
∞
By comparing with X ( z ) = ∑ x(k ) z − k
k =0
3 7 15 31
x(0) = 1; x(1) = ; x(2) = ; x(3) = . x(4) = ....
2 4 8 16
When region of convergence is less than one, rewrite the denominator polynomial of X(z) in
the reverse order and then the numerator is dvided by the denominator as shown below:
25
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
3z - 2z 2
3z - 9z 2 + 6z 3
7z 2 − 6z 3
7z 2 − 21z 3 + 14z 4
15z 3 − 14z 4
15z 3 − 45z 4 + 30z 5
31z 4 − 30z 5
#
X(z) = 2z2+6z3+14z4+30z5+62z6+....
0
Comparing with X ( z ) = ∑ x(k ) z
k = −∞
−k
x(0) = 0, x(-1) =0; x(-2) =2; x(-3) = 6; x(-4) = 14; x(-5) = 30; x(-6) = 62 .....
Example 13
1
Determine the inverse z-transform of the following function, (a) X(z) =
−1
1 − 1.5z + 0.5z − 2
z2 1 + z −1 1
(b). X(z) = (c). X(z) = (d). X(z) =
z 2 − z + 0.5 1 − z −1 + 0.5z − 2 (1 + z −1 )(1 − z −1 ) 2
Solution
1
(a) Given that, X(z) =
1 − 1.5z −1 + 0.5z − 2
26
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
1 1
X(z) = =
1 − 1.5z −1 + 0.5z − 2 1−
1.5 0.5
+ 2
z z
z2 z2
= =
z 2 − 1.5z + 0.5 (z − 1)(z − 0.5)
X(z) z
∴ =
z (z − 1)(z − 0.5)
X(z)
By partial fraction expansion, can be expressed as:
z
X(z) A1 A2
= +
z z − 1 z − 0.5
X(z) z
A1= (z − 1) z =1 = (z − 1)
z (z − 1)(z − 0.5) z =1
z 1
= = =2
( z − 0.5) Z =1
1 − 0.5
X(z) z 0.5
A2= (z − 0.5) z =0.5 = (z − 0.5) = = -1
z (z − 1)(z − 0.5) z = 0.5
0.5 − 1
X(z) 2 1
∴ = −
z z − 1 z − 0.5
2z z
X(z) = −
z − 1 z − 0.5
z z
We know that Z{ak}= and Z{u(k)} =
z−a z −1
On taking inverse z-transform of X(z) we get,
x(k) = 2 u(k) -(0.5)k; k ≥ 0
z2 The roots of the
(b) Given that, X(z) =
2
z − z + 0.5 quadratic
z2 z 2
z2 – z + 0.5 = 0 are
X(z) = =
z 2 − z + 0.5 (z − 0.5 − j0.5)(z − 0.5 + j0.5)
1 ± 1 − 4 × 0.5
X(z) z z=
∴ = 2
z (z − 0.5 − j0.5)(z − 0.5 + j0.5)
By partial fraction expansion, we can write, = 0.5 ± j0.5
*
X(z) A A
= +
z z − 0.5 − j0.5 z − 0.5 + j0.5
X(z)
A= (z − 0.5 − j0.5)
z z = 0.5 + j0.5
z
= (z − 0.5 − j0.5)
(z − 0.5 − j0.5)(z − 0.5 + j0.5) z = 0.5 + j0.5
∴ A = (0.5 – j0.5)
27
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
(z + 1)
= (z − 0.5 − j0.5)
(z − 0.5 − j0.5)(z − 0.5 + j0.5) z = 0.5 + j0.5
0.5 + j0.5 + 1 1.5 + j0.5
= = = -j1.5 + 0.5 = 0.5– j1.5
0.5 + j0.5 − 0.5 + j0.5 j1
A* = (0.5 – j1.5)* = 0.5 + j1.5
X(z) 0.5 − j1.5 0.5 + j1.5
∴ = +
z z − 0.5 − j0.5 z − 0.5 + j0.5
28
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z z
X(z) = (0.5 − j1.5) + (0.5 + j1.5)
z − (0.5 + j0.5) z − (0.5 − j0.5)
z
We know that Z{ak} =
z−a
On taking inverse z-transform of F(z) we get,
x(k) = (0.5 – j1.5) (0.5 + j0.5)k + (0.5 + j1.5) (0.5 – j0.5)k; for k ≥ 0
1
(d) Given that, X(z) =
(1 + z )(1 − z −1 ) 2
−1
1 1
X(z) =
−1 −1 2
= 2
(1 + z )(1 − z ) ⎛ 1 ⎞⎛ 1 ⎞
⎜1 + ⎟⎜1 − ⎟
⎝ z ⎠⎝ z ⎠
1 z3
= =
(z + 1) ⎛ z − 1 ⎞
2
(z + 1)(z − 1) 2
⎜ ⎟
z ⎝ z ⎠
2
X(z) z
∴ =
z (z + 1)(z − 1) 2
By partial fraction expansion, we can write,
X(z) A1 A2 A
= + + 3
z z + 1 (z − 1) 2
z −1
X(z) z2
A1 = (z + 1) = (z + 1)
z z = −1 (z + 1)(z − 1) 2 z = −1
z2 (−1) 2
= = = 0.25
(z − 1) 2 z = −1
(−1 − 1) 2
X(z) z2
A2 = (z − 1) 2 = (z − 1) 2
z z =1 (z + 1)(z − 1) 2 z = +1
2
z 1
= = = 0.5
z +1 1+1
z =1
d ⎡ F(z) 2⎤ d ⎡ z2 2⎤
A3 = (z − 1) = ⎢ (z − 1) ⎥
dz ⎢⎣ z ⎥⎦
z =1 dz ⎢⎣ (z + 1)(z − 1) 2 ⎥⎦ z =1
d ⎡ z2 ⎤ (z + 1)2z − z 2
= ⎢ ⎥ =
dz ⎣⎢ z + 1⎦⎥ (z + 1) 2
z =1 z =1
(1 + 1) × 2 − 1 3
= = = 0.75
(1 + 1) 2 4
X(z) 0.25 0.5 0.75
∴ = + +
z z + 1 (z − 1) 2 z −1
29
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z z z
X(z) = 0.25 + 0.5 + 0.75
z +1 (z − 1) 2 z −1
z z z
= 0.25 + 0.5 + 0.75
z − ( −1) (z − 1) 2 z −1
z ⎪⎧ az ⎫⎪ z
We know that Z{ak} = ; Z⎨ ⎬ = kak and Z{u(k)} =
z−a ⎪⎩ (z − a) 2 ⎪⎭ z −1
On taking inverse z-transform of F(z) we get,
x(k) = 0.25(-1)k + 0.5k(1)k + 0.75 u(k)
x(k) = 0.25(-1)k + 0.5k + 0.75 u(k); for k ≥ 0
Example 14
Determine the inverse z-transform of the following z-domain functions.
3z 2 + 2z + 1 3z 2 + 2z + 1
(a) X(z) = (b) X(z) =
z 2 − 3z + 2 z 2 + 3z + 2
z − 0.4 z−4
(c) X(z) = (d) X(z) =
z2 + z + 2 (z − 1)(z − 2) 2
Solution
3z 2 + 2z + 1
(a) Given that, X(z) =
z 2 − 3z + 2
11z − 5 3
X(z)= 3 +
z − 3z + 2
2
3z 2 + 2z + 1
11z − 5 z 2 − 3z + 2
= 3+ 3z 2 − 9z + 6
(z − 1)(z − 2)
11z − 5
By partial fraction expansion we get,
A1 A
X(z) = 3 + + 2
z −1 z − 2
11z − 5 11z − 5 11 − 5
A 1= (z − 1) = = = −6
(z − 1)(z − 2) z =1
(z − 2) z =1 1 − 2
11z − 5 11z − 5 11 × 2 − 5
A 2= (z − 2) = = = 17
(z − 1)(z − 2) z=2
(z − 1) z=2
2 − 1
6 17
∴ X(z) = 3 − +
z −1 z − 2
1 z 1 z
= 3−6 + 17
z z −1 zz−2
z z
= 3 − 6z −1 + 17z −1
z −1 z−2
z z
We know that, Z{δ(k)} = 1; Z{u(k)} = and Z{ak} = .
z −1 z−a
By time shifting property we get,
30
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z z
Z{u(k – 1)} = z −1 and Z{a(k – 1)} = z −1
z −1 z−a
On taking inverse z-transform of X(z) we get,
x(k) = 3 δ(k) – 6u(k – 1) + 17(2)(k – 1) u(k – 1); for k ≥ 0
Note:The term 2(k – 1) is multiplied by u(k – 1), because this term have samples only after k ≥ 1.
3z 2 + 2z + 1 3
(b) Given that, X(z) =
z 2 + 3z + 2 3z 2 + 2z + 1
z 2 + 3z + 2
3z 2 + 2z + 1 7z + 5 3z 2 − 9z + 6
X(z) = = 3− 2
z 2 + 3z + 2 z + 3z + 2 − 7z − 5
7z + 5
= 3−
(z + 1)(z + 2)
By partial fraction expansion we get,
A1 A
X(z) = 3 − − 2
z +1 z + 2
7z + 5 7z + 5 7 × (−1) + 5
A1 = (z + 1) = = = −2
(z + 1)(z + 2) z = −1
(z + 2) z = −1
− 1 + 2
7z + 5 7z + 5 7 × (−2) + 5
A2 = (z + 2) = = =9
(z + 1)(z + 2) z = −2
z + 1 z = −2 − 2 + 1
2 9
∴ X(z) = 3 + −
z +1 z + 2
1 z 1 z
= 3+ 2 −9
z z − (−1) z z − (−2)
z z
= 3 + 2z −1 − 9z −1
z − ( −1) z − (−2)
z
We know that, Z{δ(k)} = 1 and Z{ak} =
z−a
By time shifting property,
z
Z{a(k – 1)} = z −1
z−a
On taking inverse z-transform of F(z) we get,
f(k) = 3δ(k) + 2(-1)(k – 1) u(k – 1) + 9(-2)(k – 1) u(k – 1); for k ≥ 0
(k – 1)
Note: The term a is multiplied by u(k – 1), because these terms have samples only after
k ≥ 1.
z − 0.4 The roots of the quadratic
(c) Given that, X(z) =
z2 + z + 2 z2 + z + 2 = 0 are
z − 0.4 -1 ± 1 − 4 × 2
X(z) = z=
2
z +z+2 2
7
= -0.5 ± j
2
31
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z − 0.4
=
7 7
(z + 0.5 − j )(z + 0.5 + j )
2 2
By partial fraction expansion we get,
A A*
X(z) = +
7 7
z + 0.5 − j z + 0.5 + j
2 2
z − 0.4 7
A= (z + 0.5 − j )
7 7 2
(z + 0.5 − j )(z + 0.5 + j ) 7
2 2 z = −0.5 + j
2
7
− 0.5 + j − 0.4
z − 0.4 2
= =
7 7 7
(z + 0.5 + j ) 7
− 0.5 + j + 0.5 + j
2 z = −0.5 + j 2 2
2
7 7
− 0.9 + j j
= 2 = − 0.9 + 2 = 0.5 + j 0.9 = 0.5 + j0.34
j 7 j 7 j 7 7
∴A* = (0.5 + j0.34)* = 0.5 – j0.34
0.5 + j0.34 0.5 − j0.34
∴ X(z) = +
7 7
z + 0.5 − j z + 0.5 + j
2 2
1 z 1 z
= (0.5 + j0.34) + (0.5 − j0.34)
z 7 z 7
z + 0.5 − j z + 0.5 + j
2 2
z z
= (0.5 + j0.34)z −1 + (0.5 − j0.34)z −1
7 7
z − (−0.5 + j ) z − (−0.5 − j )
2 2
z z
We know that, Z{ak} = and by time shifting property we get, Z{a(k – 1)} = z −1
z−a z−a
On taking inverse z-transform of X(z) we get,
7 (k – 1) 7 (k – 1)
x(k) = (0.5 + j0.34) (-0.5 + j ) u(k – 1) + (0.5 – j0.34) (-0.5 - j ) u(k – 1);
2 2
for k ≥ 0
Note: Since the term a(k – 1) is valid only for k ≥ 1, it is multiplied by u(k – 1).
z−4
(d) Given that, X(z) =
(z − 1)(z − 2) 2
By partial fraction expansion we get,
32
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z−4 A1 A2 A3
X(z) = = + +
(z − 1)(z − 2) 2 z − 1 (z − 2) 2 (z − 2)
z−4 z−4 1− 4
A1 = (z − 1) = = = −3
(z − 1)(z − 2) 2 z =1
(z − 2) 2 z =1
(1 − 2) 2
z−4 z−4 2−4
A2 = 2
(z − 2) 2 = = = −2
(z − 1)(z − 2) z − 1 z=2 2 − 1
z=2
d ⎡ z−4 2⎤ d ⎡z − 4⎤
A3 = ⎢ (z − 2) ⎥ =
dz ⎢⎣ (z − 1)(z − 2) 2 ⎥⎦ dz ⎢⎣ z − 1 ⎥⎦ z = 2
z=2
(z − 1) − (z − 4) 3 3
=
2
=
2
= =3
(z − 1) z=2
(z − 1) z=2
(2 − 1) 2
−3 2 3 1 z 1 2z 1 z
∴ X(z) = − + = −3 − + 3
z − 1 (z − 2) 2 z − 2 z z − 1 z (z − 2) 2 z z−2
z 2z −1 z
= − 3z −1 − z −1 + 3z
z −1 (z − 2) 2 z−2
z z az
We know that, Z{u(k)} = ; Z{ak} = and Z{kak} =
z −1 z−a (z − a) 2
By time shifting property we get,
z z az
Z{u(k – 1)} = z −1 ; Z{a(k – 1)} = z −1 and Z{(k – 1)a(k – 1)} = z −1
z −1 z−a (z − a) 2
On taking inverse z-transform of F(z) we get,
x(k) = -3u(k – 1) – (k – 1) 2(k – 1) u(k – 1) + 3(2)(k – 1) u(k – 1)
Note: Since the term a(k – 1) is valid only for k ≥ 1, it is multiplied by u(k – 1).
k =0
∞ ∞
∑x(kT)e = ∑x(kT)elogZ
−K
−K logZ
=
k =0 k =0
33
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
∞
X ( s ) = ∑ x(kT ) z − k = X ( z ) .
k =0
It is obvious that the z-transfer of a discrete sequence can be obtained form the Laplace
1
transforms of its impulse sampled version by choosing s = log z . Consider the mapping of
T
jω-axis in the s-plane i.e.,
if z = eST
Substituting s = jω, then
z = ejωT
2πω
j 2π
or z=e ωs (Since ωs = )
T
⎛ω ⎞
= 1 ∠2π ⎜⎜ ⎟⎟
⎝ ωs ⎠
⎛ ω ⎞ ⎛ ω ⎞ ⎛ jω ⎞
Thus the section (− jω s / 2) − ⎜ − j s ⎟ − 0 − ⎜ j n ⎟ − ⎜ s ⎟ maps into the unit circle in the
⎝ 4 ⎠ ⎝ 4 ⎠ ⎝ 2 ⎠
−π π
anticlockwise direction. (-π, , 0, , π) as shown is fig. 6. In fact every function of the
2 2
jω-axis which is integral submultiples of ωs. maps into the unit circle. This suggest that the s-
plane stability criterion of the poles of s-transfer function lying the left half of the s-plane will
become the z-plane stability criterion that all poles of the z-transfer function could lie within
the unit circle
34
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
k
y(kT ) = ∑ g (kT − hT ) x (hT ) (54)
h =0
Where g(kT) is the system weighting sequence.
Since x(t) = 0 for t < 0 and g(kT – hT) = 0 for h > k, we may assume that the value of h in
equation (54), can be taken from 0 to ∞ rather than from 0 to k. Therefore equation (54)
becomes
∞
y(kT ) = ∑ g (kT − hT ) x (hT )
h =0
∞
The z-transfer of y(kT) is Y(z) = ∑ y(kT ).z
K =0
−k
∞ ∞
Substituting for y(kT) Y ( z ) = ∑∑ g (kT − hT ) x(hT ).z − k
k =0 h =0
∞ ∞
Let k – h = m, then, Y ( z ) = ∑∑ g (mT ) x(hT ).z −( m + h )
k =0 h =0
∞ ∞
= ∑ g (mT ) z −m .∑ x(hT ) z −h
m=− h h =0
(56)
Since we consider only one sided z-transform the lower limit of the first summation
becomes m = 0, then the equation (56) becomes
∞ ∞
Y ( z ) = ∑ g (mT ) z − m .∑ x(hT ).z − h = G ( z ). X ( z )
m =0 h =0
Y ( z)
= G ( z ). (57)
X ( z)
The ratio of the output Y(z) and the input X(z), is called the pulse transfer function of
the discrete-time system.
Consider the system shown in fig 7a. From the diagram we obtain,
U(s) = G(s) X*(s) (58)
Y(s) = H(s) U*(s) (59)
35
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Where X*(s) is the Laplace transfer of the sampled signal. By taking the starred Laplace
transform of the above two equation, we get
U*(s) = G*(s) X*(s) (60)
Y*(s) = H*(s) U*(s) (61)
Substituting equation (60) in equation (61)
Y*(s) = H*(s) G*(s). X*(s)
Or
Y*(s) = G*(s). H*(s). X*(s).
In term of z transform
Y(z) = G(z) H(z). X(z).
The pulse transfer function between the output y*(t) and input x*(t) in given by
Y ( z)
= G( z) H ( z)
X ( z)
Note: Take z-transfer of G(s) and H(s) individually and then multiply.
Now consider the system shown in fig 7b. From the diagram since there is no samples
between G(s) and H(s).
Y(s) = G(s) H(s) X*(s) = GH(s) X* (s).
Taking starred Laplace transform of Y(s).
Y*(s) = [GH(s)]* x*(s).
In term of z transform
Y(z) = GH(z) X (z).
The pulse transfer function between the output y*(t) and input x*(t) is.
Y ( z)
= GH ( z ) = z[GH ( s )]
X ( z)
Note: First multiply G(s) and H(s) and then take z- transform
G(z) H(z) ≠ GH(z)
Example 15
Consider the system shown in fig (8a) and (8b). Obtain the pulse transfer function Y(z)/X(z)
for each of these two systems.
Figure 8a. Sampler between G(s) and H(s) and 8b. No sampler between G(s) and H(s).
For the system shown in fig 8a, where there is a sampler between G(s) and H(s)
36
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Y ( z) Y ( z) U ( z)
= = H ( z ).G ( z )
X ( z) U ( z) X ( z)
Hence
Y ( z) ⎡ 1 ⎤ ⎡ 1 ⎤
= G ( z ).H ( z ) =
z⎢ .z
X ( z) ⎣ s + a ⎥⎦ ⎢⎣ s + b ⎥⎦
1 1
= .
1 − e z 1 − e −bT z −1
−aT −1
z2
=
(z − e −aT )(z − e −bT )
Note: Take z-transform first and then multiply.
For the system shown in fig 8b, where there is no sampler between G(s) and H(s).
1 1 ⎡ A B ⎤
GH ( s ) = . and Z [GH ( s )] = Z ⎢ +
( s + a ) ( s + b) ⎣ s + a s + b ⎥⎦
1 1
A= × (s + a) S =− a =
( s + a )( s + b) b−a
1 1 −1
B= × ( S + b) s = − b = =
( s + a )( s + b) a−b b−a
⎡ 1 ⎡ 1 1 ⎤⎤
Substituting the values of A and B Z [GH ( s )] = z ⎢ ⎢ − ⎥⎥
⎣ b − a ⎣ s + a s + b ⎦⎦
1 ⎡ 1 1 ⎤ 1 ⎡ z z ⎤
= ⎢ − −1
− − − 1 ⎥ = ⎢ −
− − ⎥
b − a ⎣1 − e zaT
1− e z ⎦ b − a ⎣z − e
bT aT
z−e bT
⎦
− aT −bT
Y ( z) z ⎡ z−e − z+e
− bT − aT
⎤ z (e −e )
= GH ( z ) = ⎢ ⎥ = .
X ( z) b − a ⎣ ( z − e − aT )( z − e −bT ) ⎦ (b − a)( z − e − aT )( z − e −bT )
(Note: First multiply in s-domain and then take z-transfer)
Clearly, we see that the pulse transfer function of the two systems are different; that is
G(z)H(z) ≠ GH(z).
37
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Taking starred Laplace transfer i.e., the laplace transfer of discrete signal.
E*(s) = R*(s) – GH*(s) E*(s)
(Note: Since there is no star between H(s) and G(s). H(s).G(s) = GH(s) i.e., we can multiply
in s-domain itself)
R * ( s)
Re-arranging E * ( s) =
1 + GH * ( s )
Since C * ( s) = G * (s) E * (s) .
G * ( s ).R * ( s )
C * ( s) =
1 + GH * ( s )
G ( z ).R( z )
Taking z-transfer C ( z) =
1 + GH ( z )
C ( z) G( z)
Closed-loop pulse transfer function is given by =
R ( s ) 1 + GH ( z )
Table 3 show five typical configurations for closed-loop discrete-time control systems.
38
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Example 16
Obtain the discrete-time output C(z) of the closed loop transfer function of the system shown
in fig 10.
39
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
Example 18
For the system shown in fig 12, find the response to unit step input where G(s) = 1/(s+1)
40
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
C ( z) GoG ( z )
=
R ( z ) 1 + GoG ( z )
⎡1 − e − sT 1 ⎤
GoG ( z ) = z ⎢ . ⎥
⎣ s ( s + 1) ⎦
⎡ 1 ⎤ ⎡A B ⎤
= (1 − z ) −1 Z ⎢ ⎥ = (1 − z −1 ) z ⎢ +
⎣ s ( s + 1) ⎦ ⎣ s s + 1⎥⎦
By partial fraction
1 A B
= +
s ( s + 1) s s + 1
1 1
A= × s s =0 = s =0 = 1
s ( s + 1) s +1
1 1
B= × ( s + 1) s = −1 = s = −1 = −1
s ( s + 1) s
Substituting
⎡1 1 ⎤
GoG ( z ) = (1 − z −1 ) z ⎢ −
⎣ s s + 1 ⎥⎦
z −1 ⎡ z z ⎤
= −
z ⎣ z − 1 z − e −T ⎥⎦
⎢
z − 1 ⎡ z ( z − e −1 ) − z ( z − 1) ⎤ 1 − e −1 0.632
since T = 1 sec. = ⎢ ⎥ = =
z ⎢⎣ ( z − 1)( z − e −1 ) ⎥⎦ z − 0.368 z − 0.368
Input to the system is unit-step input
z
R( z ) =
z −1
GoG ( z )
C ( z) = × R( z )
1 + GoG ( z )
⎛ 0.632 ⎞ z
⎜ ⎟
⎝ z − 0.368 ⎠ z − 1 0.632 z 0.632 z
= = =
0.632 ( z − 1)( z − 0.368 + 0.632) ( z − 1)( z + 0.264)
1+
z − 0.368
c( z ) A B
= +
z ( z − 1) ( z + 0.264)
0.632 0.632 0.632
A= × ( z − 1) z =1 = = = 0.5
( z − 1)( z + 0.264) z + 0.264 1.264
0.632 0.632
B= ( z + 0.264) z = −0.264 = z = 0.364 = −0.5
( z − 1)( z + 0.264) z −1
C ( z ) 0 .5 0 .5
= −
z z − 1 z + 0.264
z z
C ( z ) = 0 .5 − 0 .5
z −1 z − ( −0.264)
C(k) = 0.5 – 0.5 (-0.264)k = 0.5 [1 – (-0.264)k].
41
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
42
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
43
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
The reconstructed signal will be a straight line in a sampling period, whose slope is
determined by the current sample and previous sample.
17. Define acquisition time.
In analog to digital conversion process, the acquisition time is defined as the total
time required for obtaining a signal sample and the time for quantizing and coding. It
is also called conversion time.
18. Define aperture time.
The duration of sampling the signal is called aperture time.
19. Define settling time.
In digital to analog conversion process the settling time is defined as the time required
for the output of the D/A converter to reach and remain within a given fraction of the
final value, after application of input code word.
20. What is “Hold mode droop”?
The changes in signal voltage level in the hold circuits during hold mode (or hold
period) is called hold mode droop.
21. What are the problems encountered in a practical hold circuit?
The problems encountered in practical hold circuit are:
1. Errors in the periodicity of sampling process.
2. Nonlinear variations in the duration of sampling aperture.
3. Droop (changes) in the voltage held during conversion.
22. How the high frequency noise signals in the reconstructed signal are eliminated?
The high frequency noise signals (or unwanted signals) introduced by hold circuits in
the reconstructed signal are easily filtered out by the various elements of the control
system, because the control system is basically a low – pass filter.
23. What is discrete sequence?
A discrete sequence or discrete time signal, f(k) is a function of independent variable,
k, which is an integer.
A two sided discrete-time signal f(k) is defined for every integer value of k in the
range -∞ < k < ∞.
A one sided causal discrete-time signal f(k) is defined for every integer value of k in
the range of 0 ≤ k < ∞.
A one sided anticausal discrete-time signal f(k) is defined for every integer value of k
in the range of -∞ < k ≤ 0.
24. Define one sided and two – sided z-transform.
The z-transform (two-sided z-transform) of a discrete sequence, f(k) is defined as the
power series,
∞
F(z) = Z{f(k)} = ∑ f(k)z -k
k=-∞
where z is a complex variable.
The notation Z{f(k)} is used to denote z-transform of f(k).
The one sided z-transform of f(k) is defined as the power series,
∞
F(z) = Z{f(k)} = ∑ f(k)z -k
k=0
where z is a complex variable.
25. What is region of convergence (ROC)?
The z-transform of a discrete sequence is an infinite power series; hence the z-
transform exists only for those values of z for which the series converges. If F(z) is a
44
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
z-transform of f(k) then the region of convergence (ROC) of F(z) is the set of all
values of z, for which F(z) attains a finite value.
26. Write the infinite and finite geometric series sum formula.
The infinite geometric series sum formula is
∞ 1
∑ Ck = 1 − C ; Where |C| < 1
k=0
The finite geometric series sum formula is
M-1 M
k = C -1 ; When C ≠ 1
∑ C
C -1
k=0
= M; when C = 1
27. State the final value theorem with regard to z-transform.
If f(k) is causal & stable signal and F(z) exists with z = 1 included in the ROC then
the final value theorem is given by
f(∞) = Lt (1- z -1)F(z) ; where F(z) = Z{f(k)}
z →1
The final value theorem can be applied only if F(z) is analytic for |Z| > 1.
28. State the initial value theorem with regard to z-transform.
If f(k) is a causal signal and F(z) exists then the initial value of the signal is given by
f(0) = Lt F(z) ; where F(z) = Z{f(k)}
z →∞
29. Define z-transform of unit step signal.
The unit step signal, u(k) = 1 for k ≥ 0
The z-transform at u(k) = Z{u(k)}
∞ z
= ∑ z-k =
k=0
z -1
30. Find z-transform of ak.
By the definition of z-transform.
∞ ∞ 1 1 z
Z{ak} = ∑ a k z-k = ∑ (az-1)k = = =
k=0 k=0 1- az-1 1- a / z z - a
31. Find z-transform of e-akT.
By the definition of z-transform, Z{e-akT} is
∞ ∞ 1 1 z
= ∑ e -akT z = ∑ (e-aT z-1)k =
-k = =
-aT
1- e z -1 1- e -aT / z z - e-aT
k=0 k=0
32. What are the different methods available for inverse z-transform?
The inverse z-transform of a function, F(z) can be obtained by any one of the
following methods.
1. Direct evaluation by contour integration (or) complex inversion integral.
2. Partial fraction expansion.
3. Power series expansion.
33. What is linear (time-invariant) discrete time system (LDS)?
A discrete-time system is a device or algorithm that operates on a discrete-time signal
called the input or excitation, according to some well-defined rule, to produce another
discrete-time signal called the output or the response of the system.
45
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
A discrete time system is linear if it obeys the principle of superposition and it is time
invariant if its input – output relationship do not change with time.
34. What is weighting sequence?
The impulse response of a linear discrete-time system is called weighting sequence.
The impulse response is the output of the system when the input is unit impulse.
35. Explain how a discrete-time signal can be expressed as a summation of impulses.
Multiplication of a discrete signal, r(k) by an nit impulse at some delay m, picks out
the value r(m) of the signal r(k) at k = m, where the unit impulse is non-zero. If we
repeat this multiplication overall possible delays in the range of 0 ≤ m < ∞ and sum
all the product sequences, the result will be a sequence that is equal to the sequence
r(k). Hence r(k) can be expressed as
∞
r(k) = ∑ r(m)δ(k - m)
m=0
36. How the output of a linear discrete-time system (LDS) is related to impulse response?
The output or response c(k) of a linear discrete time system (LDS) is given by
convolution of the input r(k) with the impulse response h(k) of the system. It is
expressed as,
c(k) = r(k) ∗ h(k)
where the symbol ∗ represents convolution operation.
37. What is discrete convolution?
The convolution of two discrete-time signals (or sequences) is called discrete
convolution. The discrete convolution of sequences f1(k) and f2(k) is defined as
∞
c(k) = ∑ f1(m)f 2 (k - m)
m=0
38. Write any two properties of discrete convolution.
The discrete convolution obeys the commutative property and associative property.
Commutative property : r(k) ∗ h(k) = h(k) ∗ r(k)
Associative property : [r(k) ∗ h1(k)] ∗ h2(k) = r(k) ∗ [h1(k) ∗ h2(k)]
39. What is pulse transfer function?
The transfer function of linear discrete time system is called pulse transfer function or
z-transfer function.
It is given by the z-transform of the impulse response of the system. It is also defined
as the ratio of z-transform of output to z-transform of input of the linear discrete time
system
Pulse transfer function = H(z) = C(z) / R(z)
where H(z) = Z{h(k)} = z-transform of impulse response, h(k)
C(z) = z-transform of output of LDS
R(z) = z-transform of input to LDS.
Fig. Q3.40a Pulse sampler with ZOH
40. Sketch the frequency response curve of ZOH device.
The frequency response characteristic curve of ZOH device is shown in Fig.
46
Digital Control System Dr.V.Dharmalingam, Prof/SEEE, SASTRA
41. When the z-transfer function of the system can be directly obtained from s-domain
transfer function?
When the input to the system is an impulse sampled signal, the z-transfer function can
be directly obtained by taking z-transform of the s-domain transfer function.
42. Give the steps involved in determining the pulse transfer function of G(z) from G(s).
The following are the steps involved in determining the pulse transfer function G(z)
from G(s).
1. Determine g(t) from G(s), where g(t) = L-1[G(s)].
2. Determine the discrete sequence g(kT0 by replacing t by kT in g(kT).
3. Take z-transform of g(kT), which is the required z-transfer function, G(z).
43. How the s-plane is mapped into z-plane? or What is the relation between s and z domain?
The transformation, s = (1/T) ln z maps the s-plane into the z-plane. Every section of
jω axis of length Nωs of s-plane maps into the unit circle in the anticlockwise
direction, where N is an integer and ωs is sampling frequency. Every strip in s-plane
of width ωs, maps into the interior of the unit circle. The mapping of s-plane into z-
plane is shown in Fig.
47