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

# 1 s etting s jw in this relationship yields eiwt h jwe

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: o _ _ o o r o-- o ro _ _ - 7t/2 ( a) (b) (c) F ig. 7 .2 F requency response of an ideal (a) delay (b) differentiator (c) integrator. Consequently IH(jw)1 ;rhiS = w and L H(jw) =~ (7.7) amp~itude an~ p hase response is depicted in Fig. 7.2b. T he amplitUde response ;~:eases h nearly WIth frequency, a nd p hase response is constant (7r / 2) for all frequencies h IS r esult c an be .explai~ed physically by recognizing t hat if a sinusoid cos w t is p assed t rough an Ideal dlfferentlator, t he o utput is - w sin w t = w cos (wt + !!:) T h C h o utp t ' 'd I' d . 2' erelOre, t e u .smusm aImp ItU e IS w t imes t he i nput a mplitude'' t hat is , t he a mpl't Ud e response . I (gam) mcreases &quot;mearly WIth frequency w Moreover t he 0 t t ' 'd d h 'f'&quot; ., U p u smusm un ergoes a h p( /ase) s .Iht 2' WIth r espect t o t he i nput cos wt. Therefore, t he p hase response is c onstant 7r 2 WIt frequency. e H a~ideal d ifferentiator,.the a mplitude response (gain) is p roportional t o frequency [I (J ) 1.- wJ.' so t hat t he hIgher-frequency components are enhanced (see Fig. 7.2b). All ~ractJcal .slgnals are c ontaminated w ith noise, which, by its nature, is a broad-band (rap~dly varymg) signal.containing components of very high frequencies. A differentiator ca~ I~crease t he nOIse ~Isprop?rtionately t o t he p oint of drowning o ut t he desired signal. ThIS IS t he reason Why Ideal dlfferentiators are avoided in practice. ( c) A n i deal i ntegrator: T he t ransfer function of a n ideal integrator is [see Eq. (6.56)J 1 H (s) = s Therefore H (jw) = ~ = - j = ~e-j&quot;/2 w JW Consequently IH(jw)1 =; a nd w L H(jw) = -% (7.8) ! his a mplitude ~nd p hase response is i llustrated in Fig. 7.2c. T he a mplitude response is mversely proportIOnal t o frequency, a nd t he p hase shift is c onstant ( -7r / 2) w ith frequency. 476 7 Frequency Response a nd Analog Filters T his r esult c an b e e xplained physically by recognizing t hat if a sinusoid cos wt is passed t hrough a n i deal i ntegrator, t he o utput is s in wt = cos (wt - ~). T herefore, t he a mplitude r esponse is inversely p roportional t o w, a nd t he p hase response is c onstant ( - &quot;/r / 2) w ith f requency. Because i ts g ain is l /w, t he i deal i ntegrator s uppresses higher-frequency c omponents b ut e nhances lower-frequency c omponents w ith w &lt; 1. C onsequently, noise signals ( if t hey d o n ot c ontain a n a ppreciable a mount o f very low frequency components) a re s uppressed ( smoothed o ut) b y a n i ntegrator. • t l:!. t 7.2 Bode P lots 7.2 4 77 Bode Plots Sketching frequency response plots is considerably facilitated by t he use of logarithmic scales. T he a mplitude a nd p hase response plots as a function of w on a logarithmic scale are known as t he B ode p lots. B y using t he a symptotic behavior o f t he amplitude a nd t he p hase response, we can sketch these plots with remarkable ease, even for higher-order transfer functions. Let us consider a system with t he t ransfer function E xercise E...
View Full Document

Ask a homework question - tutors are online