This preview shows page 1. Sign up to view the full content.
Unformatted text preview: t hat t his e quation
is v alid for any value o f k. T herefore, replacing k b y k + 2, we o btain lOT
(e) (d) ( c) y[k] T t I~
~ y[k] = y{kT) a nd (B.27) T he signals f [k] a nd y[k] a re t he i nput a nd t he o utput for t he d iscretetime s ystem G.
Now, we r equire t hat yet) =
T herefore, a t t = kT t (see Fig. 8.20a) dJI y(kT) = dt = l im . =kT ~ [J(kT)  J[(k  I)T]] T oT (8.28) 566 8 D iscretetime S ignals a nd S ystems By using the notation in Eq. (8.27), the above equation can be expressed as T { I[k]  f[k  I]} (8.29) T he approximation improves as T approaches O. A discretetime processor G t o realize
Eq. (8.29) is shown inside t he shaded box in Fig. 8.20b. T he system in Fig. 8.20b acts as
a differentiator. This example shows how a continuoustime signal can be processed by a
d iscretetime system.
To determine t he sampling interval T , we note t hat t he highest frequency t hat will
a ppear a t t he input is 20 kHz; t hat is, Fh = 20,000. Hence, according t o Eq. (8.17c)
1 T:'O 40,000 = 25 J1.s To gain some insight into this method of signal processing, let us consider t he differentiator in Fig. 8.20b with a ramp input f (t) = t, depicted in Fig. 8.20c. I f t he system
were to act as a differentiator, t hen t he o utput yet) o f t he s ystem should b e t he unit step
function u(t). L et us investigate how the system performs this particular operation and
how well it achieves t he objective.
T he samples of t he i nput f (t) = t a t t he interval of T seconds a ct as the input to
t he discretetime system G. These samples, denoted by a compact notation J[kJ, are,
therefore, f[k] = f(t)I'=kT = tl'=kT
= kT t 5 67 1. T~O T his is t he inputoutput relationship for G required t o achieve our objective. In practice,
t he s ampling interval T c annot be zero. Assuming T t o be sufficiently small, the above
e quation can be expressed as
y [k] ' " E xamples o f D iscreteTime S ystems We s how i n C hapter 11 t hat t his is g enerally t rue o f d iscretetime s ystems. T he
d iscretetime s ystems c an b e r ealized i n t wo ways: l
y [k] = lim T { I[k]  f[k  I]} 1 8.5 2.: 0 k 2.:0 F igure 8.20d shows the sampled signal f [k]. This signal acts as a n i nput t o t he discretetime system G. Figure 8.20b shows t hat t he operation of G consists of subtracting a
sample from the previous (delayed) sample and t hen multiplying t he difference with l iT.
F rom Fig. 8.20d, i t is clear t hat t he difference between t he successive samples is a constant
k T  (k  I)T = T for all samples, except for t he sample a t k = 0 (because there is no
previous sample a t k = 0). T he o utput o f G is l iT times t he difference T , which is unity
for a ll values of k, except k = 0, where it is zero. Therefore, t he o utput y[k] of G consists
of samples of unit values for k 2.: 1, as illustrated in Fig. 8.20e. T he Die (discretetime
t o continuoustime) converter converts these samples into a continuoustime signal yet),
as shown in Fig. 8.20f. Ideally, t he o utput should have been yet) = u (t). T his d...
View
Full
Document
This note was uploaded on 04/14/2013 for the course ENG 350 taught by Professor Bayliss during the Spring '13 term at Northwestern.
 Spring '13
 Bayliss
 Signal Processing, The Land

Click to edit the document details