Signal Processing and Linear Systems-B.P.Lathi copy

907981008 182 1414213568 182 1782013058 182 t ag b

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 2983 + 43382 + 37328 + 16081 T his is t he a lternate s olution where t he p assband specifications are exceeded, b ut t he s topband specifications are satisfied exactly. O n t he o ther h and, t he s olution in Example C7.5 exceeds t he s topband specifications, b ut satisfies t he p assband specifications exactly because of use of Eq. (7.40) [ rather t han Eq. (7.41)]. To plot a mplitude response, we can use t he l ast t hree functions from Example C7.5. 0 {;, E xercise E 7.3 Determine n , the order of the lowpass Butterworth filter to meet the following specifications: Gp = - 0.5dB, Gs = - 20 dB, Wp = 100, andws = 200. Answer: 5. 'V 514 7 .6 7 Frequency Response and Analog Filters 7.6 Chebyshev F ilters Chebyshev Filters 515 t t 1:H(j(j)) I 1:H(j(j)) I T he a mplitude response of a normalized Chebyshev lowpass filter is given by 1 (7.42) 11t(jw)1 = - , = = = = VI + E 2C n 2 b 1 b 1 n =7 (w) where C n(w), t he nth-order Chebyshev polynomial, is given by Cn(w) = cos ( ncos- I w) (7.43a) o o F ig.7.24 A mplitude r esponse o f n ormalized s ixth- and s eventh-order l owpass C hebyshev filters. An alternative expression for Cn(w) is (7.43b) The form (7.43a) is most convenient t o c ompute Cn(w) for Iwl < 1 a nd form (7.43b) is convenient for computing Cn(w) for Iwl > 1. We c an show l t hat Cn(w) is also expressible in polynomial form, as shownt in Table 7.3 for n = 1 t o 10. T he normalized Chebyshev lowpass amplitude response [Eq. (7.42)] is depicted in Fig. 7.24 for n = 6 a nd n = 7. We make the following general observations: 1. T he Chebyshev a mplitude response has ripples in the passband and is smooth (monotonic) in the stopband. The passband is 0 ::; w ::; 1, a nd there is a t otal of n m axima and minima over the passband 0 ::; w ::; 1. 2. From Table 1.3, we observe t hat C~(O) == {~ n o dd n even (7.44) Therefore, the dc gain is T able 1 .3: C hebyshev P olynomials 11t(O)1 == {I n o dd I 7 rP' 2 3 4 5 6 7 8 9 10 1 w 2w 2 - 1 4 w 3 - 3w 8w 4 - 8w 2 + 1 16w 5 - 20w 3 + 5w 32w 6 - 48w 4 + 18w 2 - 1 64w 7 - 112w 5 + 56w 3 - 7w 128w 8 - 256w 6 + 160w 4 - 32w 2 + 1 256w 9 - 576w 7 + 432w 5 - 120w 3 + 9w 512w IO - 1280w 8 + 1120w 6 - 400w 4 + 50w 2 t The Chebyshev polynomial C n(w) has the propertyl C n(w) = 2WCn _l(W) - C n _2(W) Thus, knowing t hat a nd Co(W) = 1 we can construct C n(w) for any value of n . For example, and so on. (7.45) 3. T he p arameter E controls the height of ripples. In the passband, r , t he ratio of the maximum gain t o t he minimum gain is n o n even r==~ (7.46a) This ratio r , specified in decibels, is (7.46b) so t hat (7.41) Because all the ripples in the passband are of equal height, the Chebyshev polynomials Cn(w) are known as e qual-ripple f unctions. - 1 n> 2 4. T he ripple is p resent only over the passband 0 ::; w ::; 1. A t w == 1, t he amplitude response is 1 /Vl+? == l /r. For w > 1, t he gain decreases monotonically. 5. For Chebyshev filters, the ripple f dB takes the place of G ( the minimum gain p in t...
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.

Ask a homework question - tutors are online