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

# T his t erm c an b e a dded t o t he p lot by s imply

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

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: remain unchanged whether the zero(s) are in t he L HP or the RHP. However, the phase of t he R HP zero a t s = a is L(jw - a) =L - (a -jw) = + tan-l (-aw ) = IT IT - tan-l(~) 490 7 Frequency Response a nd Analog Filters whereas t he p hase of t he L HP pole a t s = - a is - tan-1(w/a). T he c omplex conjugate zeros in R HP give rise t o a t erm 8 2 - 2(W n 8 +w~, which is identical t o t he t erm 8 2 + 2 (W n 8 + w~ w ith a sign change in ( . Hence, from Eqs. (7.23) a nd (7.24) i t follows t hat t he amplitudes are identical, b ut t he phases are of opposite signs for the two terms. 7.2-1 The Transfer Function From t he Frequency Response In t he p revious section we were given t he t ransfer function o f a system. From a knowledge o f t he t ransfer function, we developed techniques for determining t he s ystem response to sinusoidal inputs. We can also reverse t he p rocedure t o determine t he t ransfer function of a system if t he s ystem response to a sinusoidal i nput is known. This problem has significant practical utility. I f we a re given a system in a black box w ith only t he i nput a nd o utput t erminals available, t he t ransfer function has t o b e d etermined by experimental measurements a t t he i nput a nd o utput t erminals. T he f requency response to sinusoidal inputs is one of t he possibilities t hat is very a ttractive because of t he simple n ature of t he m easurements involved. O ne o nly needs t o a pply a sinusoidal signal a t t he i nput a nd observe t he o utput. We find t he a mplitude gain IH(jw)1 a nd t he o utput p hase shift L.H(jw) ( with respect t o t he i nput s inusoid) for various values of w over t he e ntire range from 0 t o 0 0. T his information yields t he frequency response plots (Bode plots) when plotted against log w. From t hese p lots we d etermine t he a ppropriate asymptotes by taking advantage o f t he fact t hat t he slopes of all asymptotes must be multiples of ± 20 d B/decade if t he t ransfer function is a r ational function (function which is n ot necessarily a ratio of two polynomials in 8). From t he a symptotes, t he corner frequencies are obtained. C orner frequencies determine t he poles a nd zeros of t he t ransfer function. 7 .3 Control System Design Using Frequency Response F igure 7 .10a shows a basic closed-loop system, whose open-loop transfer function (transfer function when t he loop is opened) is K G(8)H(8). T he closed-loop transfer function is [see Eq. (6.69) 7.3 C ontrol System Design Using Frequency Response Sec. 6.7 a nd t he frequency response method should be considered as c omplementary r ather t han as alternatives or as rivals. Frequency response information can be presented in a various forms of which Bode plot is one. T he same information can be presented by t he N yquist p lot&quot; also known as t he p olar p lot or by t he N ichols p lot also known as log-magnitude versus angle plot. Here, we shall discuss t he techniques using Bode a nd N yquist plots only. Figure 7.lOb shows Bode plots for t he open-loop transfer function K /s(s+2)(s+...
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