You are on page 1of 72
Cuarien 1 = Intmowucrion ‘The edited book by Gupta covers (1) the history of the subject of electrical noise, (2) an in- troduction to physical mechanisms, mathematical methods, and applications, (3) the prin- Cipal noise-generating processes, (4) the types of electronic devices in which noise phenomena have been studied, and, finally, (5) noise generators. 9. For an introductory treatment of MATLAB, see Appendix F [4o>monat Prosems 1.42 Determine whether the following signals are period- ic. If they are periodic, find the fundamental period. fa) x(t) = (cos(2me))? (b) x(t) = Sf_sw(t — 2k) for w(t) depicted in Fig. P1.42b. (ce) x(t) = Zpe-ew(t — 3k) for w(t) depicted in Fig. P1.42b. rr) of C) (@) {x)= (-1" (e) x{n] = (-1)" (f) x[n] depicted in Fig. P1.42f. (g) x(t) depicted in Fig, P1.42g. (a) x[n] = cos(2n) (i) x["] = cos(2mn) Ficune P1.42 1.43 The sinusoidal signal x(t) = 3.0s(200t + 1/6) is passed through a square-law device defined by the input-output relation y(t) = 2°(2). Using the trigonometric identity cos? @ = 3 (cos 26 + 1), show that the output y(t) consists of a de component and a sinusoidal component. (a) Specify the de component. (b) Specify the amplitude and fundamental frequen- ‘gy of the sinusoidal component in the ourput y(t). Consider the sinusoidal signal x(t) = Acos(ut + 4). Determine the average power of x(). 144 1.45 The angular frequency 02 of the sinusoidal signal x{n] = Acos(n + ) satisfies the condition for x[7t] to be periodic. De- termine the average power of x[]. 41.46 The raised-cosine pulse x(t) shown in Fig. P1.46 is defined as cos(ut) +1), -a/w sts m/w xt) {it otherwise Determine the total energy of x(t). #0 19 The oO who ‘ Additional Probleme 1.47 The trapezoidal pulse x(t) shown in Fig. P1.47 is defined by 4 4stss _] 4 -4e8e4 OV ees sans —4 0, otherwise Determine the total energy of x(t). x) © (d) (S) cc) (es) x(t + 1y(t - 2) x(t)y(-1~ 8) x(e)y(2 ~ 1) x(2t)y(}e + 1) x(4- y(t) © + Ficune PL.S2 4 Ficure P147 321012345 1.48 The trapezoidal pulse x(t) of Fig. P1.47 is applied to a differentiator, defined by we) = Zs(0. (a) Determine the resulting output y(¢) of the dif- ferentiator (b) Determine the total energy of y(t). 1.49 A rectangular pulse x(t) is defined by pu{& OstsT == Lo, otherwise * ‘The pulse x(t) is applied to an integrator defined by w(t) = [ x(t) de. Find the total energy of the output y(t). 1.50 The trapezoidal pulse x(t) of Fig. P1.47 is time scaled, producing the equation y(t) = x(at). Sheth y(6) for (a) a — 5 aud (b) a — 0.2. pulse y(t) related to that of 1.51. Sketch the trapezoi Fig. P1.47 as follows: y(t) = x(10¢ - 5) 1.52 Let x(t) and y() be given in Figs. P1.52{a) and (b), 1.55 respectively. Carefully sketch the following signals: (a) x(t)y(t - 1) (b) x(# — 1)y(-#) 1.53 Figure P1.53(a) shows a staircaselike signal x() that may be viewed as the superposition of four rectan- gular pulses. Starting with a compressed version of the rectangular pulse g(r) shown in Fig. P1.53(b), construct the waveform of Fig, P1.53(a), and express x(t) in terms of g(t). Ficune P1.53 1.54 Sketch the waveforms of the following signals: (a) x(t) = u(t) — u(t — 2) (b) x(t) = w(t + 1) — 2u(t) + u(t — 1) (c)_ x(t) = —m(t + 3) + 2u(e + 1) = 2u(t ~ 1) + u(t - 3) (d) y(t) = r(t + 1) — F(t) + r(t- 2) fe) 94) = (t+ 2) - re + 1) H(t 1) + (t= 2) Figure P1.5S{a) shows a pulse x(t) that may be ‘viewed as the superposition of three rectangular puls- ¢s. Starting with the rectangular pulse g(t) of Fig. P1.55(b), conseruct the waveform of Fig. P1.SS, and express x(t) in terms of g(t). 1.57 Determine whether the following signals are peri- dic, and for those which are, find the fundamen- tal period: (a) x{n] = cos(ean) cos(rn) cos(2t) + sin(3t) Zhe (—1)*8(¢ — 2k) Bio (8[n — 3k] + 3m — &]} cos(t)u(e) v(t) + v(t), where = cos(t)u(t) v(t) + v(t), where in(t)se(t) cos(}n) sin(}#n) Ficure PL55 1.56 Letx[7] and y{7] be given in Figs. P1.56(a) and (b), on respectively. Carefully sketch the following signals: 1.58 The sinusoidal signal x{] has fundamental period 'N = 10 samples. Determine the smallest angular fre- @ aie 1 quency 2 for which xn] is periodic. O sln 7 1.59 A complex sinusoidal signal x(t) has the following © sits) components: ia a - aie on +2) Re{x(t)} = xa(t) = A cos(ut + 4); (8) x[2n] + y[n — 4] Im{x(#)} = x(t) = Asin(wt + 4). (g) x[n + 2yfn - 2) The amplitude of x(t) is defined by the square root (hy x[3 — mbln] of <}(¢) + x}(0). Show tha this amplitude equals A and is therefore independent of the phase angle @. 1.60 Consider the complex-valued exponential signal x(t) = Ae, a > 0, Evaluate the real and imaginary components of x(t). 1.61. Consider the continuous-time signal YO+05, -A/2sts 3/2 x(t) = ti) x[-n}y[-n] 4) x[m)pf-2 - (ky x(n + 2916 ao) 3 4 t> ad > 0 <-a/2 1 ‘which is applied to a differentiator. Show that the ‘output of the differentiator approaches the unit im- 7 pulse 6(t) as A approaches zero. saayras 1.62 In this problem, we explore what happens when a « unit impulse is applied to a differentiator. Consider we a triangular pules =(2) of duration A and amplinade 2/A, as depicted in Fig. P1.62. The area under the ' pulse is unity. Hence, as the duration A approaches aaa | zero, the triangular pulse approaches a unit impulse. isd ” (a) Suppose the triangular pulse x(¢) is applied to | a differentiator. Determine the output y(#) of IN the differentiator. a {b) What happens to the differentiator output y(t) as A approaches zero? Use the definition of a Ficune P1.56 unit impulse 6() to express your answer. Additional Problems (c) What is the total area under the differentiator ourput y(t) for all A? Justify your answer. Based on your findings in parts (a) through (c), de- scribe in succinct terms the result of differentiating a unit impulse. x) ay “a0 ae Ficune P1.62 1.63 A system consists of several subsystems connected x) 4s shown in Fig. P1.63. Find the operator H relating (0) 10 y(¢) for the following subsystem operators: Hyyi(t) = xa(t)x(t ~ 1)5 Healt) = ba(Ok Hyys(e) = 11 2x4(2)s Heya(t) = cos(x(t)). x00, nO xo Ficure P1.63 1.64 ‘The systems that follow have input x(t) or x{7] and output y(t) or y{n]. For each system, whether it is (i) memoryless, (i) stable, (ii) causal, (iv) linear, and (v) time invariant. (a) (0) = costx(e) tb) afm] = 2efmdel] (©) (0) = logit td) y(1) Lex) de (e) y[n] = x[k + 2) © 9 = Za) (e) y["] = cos(2mx[m + 1]) + x[n] (hy 0) a x0 = Ztex(o} re) = x(2— 1) yo] = x[n1] Zio 8lm — 2k] 1.65 1.66 1.67 1.68 1.69 1.70 417i mu ‘The output of a discrete-time system is related to its input x[] as follows: yn] = aox{n] + ayx{n — 1] + azx[n — 2] + ayx[n — 3]. Let the operator S* denote a system that shifts the input x[7t] by k time units to produce x[n ~ k]. Formulate the operator H for the system relating yl] to x{7]. Then develop a block diagram repre- ‘sentation for HY, using (a) cascade implementation and (b) parallel implementation. Show that the system described in Problem 1.65 is BIBO stable for all ao, a4, dz, and a. How far does the memory of the discrete-time system described in Problem 1.65 extend into the past? [sit possible for a noncausal system to possess mem- cory? Justify your answer ‘The output signal y{] of a discrete-time system is re- lated to its input signal x{n] as follows: yl) = xn) + x(n — 1] + x[n - 2). Letthe operator S denote a system that shifts its input by one time unit. (a) Formulate the operator H for the system relat- ing y[n} to x{n). (b) The operator H*™ denotes a discrete-time sys- tem thats the inverse of the given system. How is H™ defined? ‘Show that the discrete-time system described in Prob- Jem 1.65 is time invariant, independent of the coef- ficients ay, a, 42, and ay, (a) Isit possible for a time-variant system to be lin- car? Justify your anwer. (b) Consider the RC circuit of Fig. P1.71, in which the resistive component R(t) is time varying. For all time t, the time constant of the circuit is large enough to justify approximating the cir- ‘cuit as an integrator. Show that the circuit is in- deed linear. Ro mo ea Ficure P1.71 4.72 Show that a pth powerlaw device defined by the input-output relation y(t) = 3°(2), is nonlinear pintegerand p ¥ 0,1, 2 (Cuarren 1 = INTRODUCTION 1.73 A linear time-invariant system may be causal ornon- 1.75 A system H has its input-output pairs given. Deter- causal. Give an example of each of these possibilities. mine whether the system could be memoryless, 1.74 Figure 1.56 shows two equivalent system configu- causal, linear, and time invariant for (a) signals de- rations on condition that the system operator H is picted in Fig. P1.75(a), and (b) signals depicted in linear. Which of these configurations is simpler to Fig. P1.75(b). For all cases, justify your answers. implement? Justify your answer. x nee) ia 1 mi /N ' ' oft of 1 2 ol) ft) 23, —-ii B oft ot NGA 4 4 0 we t 1 —i— TIF, ' o| iia aes 1 NOGA a @ x0 no 7 1 — Hs 2 ' FLU. aa ol ot 0 2 weft) 1 i He of t23 4 o i234 ae ao 1 ‘NO, me ate, ora 3s 4 oH 123 xo) wl) 1 — = [1 ' of 1234 Ficune P1.75 a) ‘Addssional Problems 1.76 A linear system H has the input-output pairs de- picted in Fig. P1.76(a). Answer the following ques- tions, and explain your answers: (a) Could this system be causal? x0 93 (b) Could this system be time invariant? (©) Could this system be memoryless? (4) What is the output for the input depicted in Fig. P1.76(b)? Ficure P1.76 1.77 A discrete-time system is both linear and time in- Variant. Suppose the output due to an input x(n] = 8] is given in Fig. P1.77(a). (a) Find the output due to aninputx{) = 6[m ~ 1]. (b) Find the ourpur due to an input x{n) = 26{n] ~ afm - 2]. (c) Find the output due to the input depicted in Fig. P1.77(b). Ficure P1.77 oF [ Avvaxce Prost Cnarien = Iymovection 1S 1.78 (a) An arbitrary real-valued continuous-time signal ay be expressed as x(t) = x(t) + x(t), where x,(t) and x,() are, respectively, the even and odd components of x(t). The signal x(t) ‘occupies the entire interval ~20 Figure P1.81 1.82 It is proposed that the unit impulse 6() be approx- imated by the symmetric double exponential pulse chawn in Fig, P12, which is defined hy s(t) = A (erat +e u(0)) (a) Determine the amplitude A attained by x3(1) at t= -A/2. (b) Find the necessary condition that the time con stant r must satisfy for x3(1) to approach 8(1) asthe parameter A approaches zero, as given by 8(¢) = lim x3(¢). (c) Mlustrate graphically the nature of chis approx- imation for A = 1, 0.5, 0.25, and 0.1 Ficure P1.82 Adultivreal Problems 1.83 The operations of differentiation and integration xo are closely related. In light of this close relation- ship, it is tempting to state that they are the inverse of each other. (a) Explain why it would be wrong, in a rigorous sense, to make this statement. ‘The simple LR circuits of Figs. P1.83(a) and P1,83(b) may be used as approximators to dif- ferentiating and integrating circuits. Derive the conditions the elements of these two circuits would have to satisfy for them to fulfill heir ap- proximate functions. Use the examples of Figs. P1.83(a) and P1.83(b) in support of the explanation in part (a) of the problem. ib «) x L 13 x9 x0) RE) ea » Ficune PL.83 1.84 Figure P1.84 shows the block diagram of a linear time-varying system that consists simply of a multi- plier that multiplies the input signal x(t) by the out- put of an oscillator, Ay cos(wut + 6), thereby producing the system output ¥(t) = Aycos(aot + b)x(2). Demonstrate the following: (a) The system is linear; that is, it satisfies both the principle of superposition and the property of homogeneity. ‘The system is time variant; that is, it violates the time-shift property. To show this, you may use the impulse input x(t) = 6(t). x) —+ B—+> 0 I Ap costant +6) Fioure P1.84 1.85 In Problem 1.$4 we considered an example ofa linear time-varying system. In this problem, we consider a more complex nonlinear time-varying system. The out Put of the system, ¥(¢), 15 related to the input x(2) as v= sa(arte ef tere) 1.86 1.87 1.88 95 where k is a constant parameter. (a) Show that the system is nonlinear. (b) Evaluate y(t) for x(t) = 8(t) and its time- shifted version x(#) = 6(f— f), where ty > 0. Hence, demonstrate that the system is time variant. In this problem, we explore a useful application of nonlinearity: A nonlinear device provides a means for mixing two sinusoidal components. Consider a square-law devic y(t) = x7(t). Let the input (0) = Aycos(wyt + 6s) + Arcos(ust + 3). Determine the corresponding output y(t). Show that +y(2) contains new components with the following fre- ‘quencies: 0, 20,5 2a, oo) ¥ om. What are their re- spective amplitudes and phase shifts? In this problem, we explore another application of nonlinearity: A nonlinear device provides a basis for harmonic generation. Consider a euhie-law device: y(t) = x(t). Let the input x(t) = Acos(wt + 4). Determine the corresponding ourput y(1). Show that {y(4) contains components wid the frequen ies w aud 3w, What are their respective amplitudes and phase shifts? ‘What form of nonlinearity would you use to gener- ate the pth harmonic of a sinusoidal component w? Justify your answer (a) The step response ofa second-order system pro- duced by the input x() is given by Wt) = [1 = e%cos(at)}u(t), x(t) = a(t), where the exponential parameter a and the fre~ quency parameter w, are both real. Show that response of the system is given by # cos(w,t) + one sin(oo,t)]u(t) for x(6) = 8(6). Suppose next that the parameter wis imagi- nary—say, o = jat,, where ay = a. Show that the impulse response of the corresponding sec~ ond-order system consists of the weighted sum of two decaying exponentials, y(t) = [Serr Fs Seely, x(t) = 8(t), where @, = a ~ a, anda; = @ + a. 96 Chapter 1 = Eyrropuction 1.89 Figure P1.89 shows the block diagram of a first-order recursive discrete-time filter. This filter differs from thar of Fig. 1.69 in thar the output y(t] also requires knowledge of the past input x{7 ~ 1) for its evalu- ation. Building on the solution given in Eq, (1.116), dorive an expression for the output y/rt] in eorme of the input [7], x11 > ~~ B—yind ‘t at os va os (net eae Ficune PL.89 1.90 Itis proposed that the MEMS acceleometer described in the block diagram of Fig. 1.64 be simulated by a second-order discrete-time system that would lend itself for use on a digital computer. Derive the dif- ference equation that defines the input-output be- havior of this simulator. Hint: Use the approximation of a derivative given by e-D-A-D} [ Comreren Experiments where T, denotes the sampling interval. For the sec- ond derivative d°z/dt’, apply the approximation ewice. Typically, the received signal of a radar or commu- nication receiver is corrupted by additive noise. To combat the degrading effect of the nuise, the sigsl- processing operation performed at the front end of the receiver usually involves some form of integra- sion. Explain why, in such an application, integra- tion is preferced over differentiation. Consider the parallel RC circuit shown in Fig. P1.92 The source of current is denoted by if), and the re sulting currents through the capacitor C and resistor R are respectively denoted by iy(t) and 1:1). By for- rmulating i () in terms of if) and the voltage across the resistor, (1), the circuit may be viewed as a feed- back system. Develop the block diagram of this par- ticular method of representation. 1.91 1.92 iat io) te : to Ficure P1.92 1.93 (a) The solution of a linear differential equation is given by x(t) = Wet — 56°, Using MATLAB, plot x(t) versus t for t= 0:0.01:5. (b) Repeat Part (a) for x(t) = le + Se, 1.94. Anexponentially damped sinusoidal signal is defined by x(1) = 20sin(2a x 1000t ~ 1/3) e™, where the exponential parameter a is variable, tak- ing on the set of values a = S00, 750, 1000. Using MATLAB, investigate the effect of varying a on the signal x(¢) for -2 = ¢ = 2 milliseconds. Write a set of MATLAB commands for approximat- ing the following continuous-time periodic waveforms: 195 (a) Square wave of amplitude 5 volt, fundamental frequency 20 Hz, and duty cycle 0.6. (b) Sawtooth wave of amplitude 5 volts and fun- damental frequency 20 Hz, Plot five cycles of each of these two waveforms. 1.96 A raised-cosine sequence is defined by cos(2aFn), -(1/2F) Sm = (1/2F) 9, otherwise ‘Use MATLAB to plot zo(] versus n for F = 0.1. A rectangular pulse x(t) is defined by 10, ost=5 at) -{ 0, otherwise 1.97 Generate x(#), using (a) A pair of time-shifted step functions (b) An Mefile ‘Addirtonal Problems 183 3. Applications of difference equations to signal-processing problems and block diagram. descriptions of discrete-time systems are described in the following texts: » Proakis, J.G., and D. G. Manolakis, Digital Signal Processing: Principles, Algorithms and Applications, 3ed ed. (Prentice Hall, 1995) » Oppenheim, A. V., R. W. Schafer, and J. R. Buck, Discrete Time Signal Processing, 2nd ed. (Prentice Hall, 1999) Both of the foregoing texts address numerical issues related to implementing discrete-time TI systems in digital computers. Signal flow graph representations are often used to describe ‘implementations of continuous- and discrete-time systems. They are estentially the same as a block diagram representation, except for a few differences in notation. 4, In this chapter, we determined the input-output characteristics of block diagrams by ma- nipulating the equations representing the block diagram. Mason’s gain formula provides 2 direct method for evaluating the input-output characteristic of any block diagram repre- sentation of an LTT system. The formula is described in detail in the following two texts: » Dorf, R.C., and R. H. Bishop, Modern Control Systems, 7th ed. (Addison-Wesley, 1995) > Phillips, C. L, and R. D. Harbor, Feedback Control Systems, 3rd ed. (Prentice Hall, 1996) 5. The role of differential equations and block diagram and state-variable descriptions in the analysis and design of feedback control systems is described in Dorf and Bishop and in Phillipe and Harbor, hoth just mentioned 6. More advanced treatments of state-variable-description-based methods for the analysis and design of control systems are discussed in » Chen, C. T., Linear System Theory and Design (Holt, Rinehart, and Winston, 1984) » Friedland, B., Control System Design: An Introduction to State-Space Methods (McGraw-Hill, 1986) AA thorough, yet advanced application of state-variable descriptions for implementing discret time LTI systems and analyzing the effects of numerical round-off is given in » Roberts, R. A., and C. T. Mullis, Digital Signal Processing (Addison-Wesley, 1987) [L4oprmonat Prostems 2.32 2.33 ‘A discrete-time LTI system has the impulse response ‘b(n depicted in Fig. P2.32(a). Use linearity and time invariance to determine the system output y{7) ifthe input is (a) x(x] = 36[n] - 28{n - 1] (b) x[n] = afm + 1) - wf - 3] (c)_x[m] as given in Fig. P2.32(b). Evaluate the following discrete-time convolution Ficure P2.32 CHAPTER 2 & TiME-DOMAIN REPRESENTATIONS OF LINEAR TIME-INVARIANT SYSTEMS. (8) yf) = cos(En) « (2)"mln — 2) (@) yn] = Aru[n] «mtn 3}, [al 05 0 In Problem 2.81, we determined that the solution of the equations of motion for a vibrating string have the form y4(l,1) = du(Ifi(t),0 SI's a, where and oo, = km/a. Since »4(I,9) ie a solution for any a4 and bg, the most general solution has the form whe) = Binet ar cos(onct) + bysin(ayet)). ‘We can use the initial conditions y(/,0) = x(I) to. find the ay and -£-y(,t)|.0 to find the by. (a) Express the solution for a, in terms of the FS coefficients of x(I). (Hint: Consider Eqs. (3.25) and (3.26), with / replacing t.) Express the solution for b, in terms of the FS coefficients of g(!). Find y(l,t), assuming that g(I) = 0 and x(I) is ae chown in Fig, P3.96 ) ( x) Ficure P3.96 3.97 In this problem, we explore a matrix representation for the DTES. The DTFS expresses the N time-do- main values of an N-periodic signal x{rt] as a func- tion of N frequency domain values X/k}. Define vectors x{0] x(0] al | ana x=] *H x{N-1] X(N - 1] (a) Show that the DTFS representation xa win} = SB xtbpet, = 0, n) = Ext, w= 0,1, can be written in matrix vector form asx = VX, where V is an N-by-N matrix. Find the elements of V. Show that the expression for the DTFS coeffi- cents, wN=4 (b) wie) = BBSatmeie, bose 3.98 3.99 MSE, = can be written in matrix vector form as X = Wx, where W is an N-by-N matrix. Hind the elements of W. The expression x = VX implies that X = V"'x provided that V is a nonsingular matrix. Com- equation with the results of (b), we hat W = V"l Show thar thie i tne by establishing that WV = I. (Hint: Use the de- finitions of V and W determined in (a) and (b) to obrain an expression for the element in the Ith row and mth column of WY, and use the re- sult of Problem 3.85. ‘We may find the FS coefficients by forming the inner product of the series expansion and the conjugate of the basis functions. Let ) x() = Si x{kle, Using the result of Problem 3.85, derive the ex- pression for X[k] by multiplying both sides of this equation by e-"*~ and integrating over one period. In this problem, we find the FS coefficients X{k] by minimizing the mean square error (MSE) between the signal x(t) and its FS approximation. Define the Jterm FS 50 = Zane and the Jerm MSE as the average squared differ- ence over one period: 7 MSE, = t [ Ix(e) — (Pat, (a). Substitute the series representation of #(t), and expand the magnitude squared, using the iden- tity la + BP = (a + 6)(a* + 6°), to obtain 4 [wore = Sarn($ [x0ear) -3 a(t [smetvar) 3.100 Adblitevenal Probleme and use the orthogonality of e and e™! (see Problem 3.85) to show that * se, = $f heoat- Seem - Saextey + See, (c) Use the technique of completing the square to show that se, = 4 f[ kx(P ae 4 = _ 2, ~ Sate) ataar - Spc. (d) Find the value of A[k] that minimizes MSE. (c) Express the minimum MSE, as a function of x(t) and X{k]. What happens to MSE, as J Generalized Fourier Series. The concept of the Fourier Series may be generalized to sums of signals ‘other than complex sinusoids. That is, we may ap- proximate a signal x(t) on an interval [f,f] a8 2 weighted sum of N_ functions” #(t), PD ys Oxi wa x(t) = Ban ‘We assume that these N functions are mutually or thogonal on [f;.t7] that is, [esostna = te ath In this approximation, the mean squared error is MSE = 7 + 5 f ko - Zaswol a (a) Show that the MSE is minimized by choosing cx = feix()O2(¢) de. (Hint: Generalize steps ‘outlined in Problem 3.99 (a}-{d) to this problem.) (b) Show that the MSE is zero if [wore "Salar. If this relationship holds for all x(¢) in a given class of functions, then the basis functions o(t), d2(F)s---9 dy-i(F) are said to be “com- plete” for that class. (c)_ ‘The Walsh functions are one set of orthogonal functions that are used for representing a sig- nal on (0, 1]. Determine the cy and MSE ob- tained by approximating the following signals with the first six Walsh functions depicted in 337 Fig. P3.100: , fu k and thatthe limit as T approaches infinity is taken in a symmetric manner, Define the finite-duration nonperiodic 338 (Cnaries 3 = Fouen Rermesenunions oF Sronats anv Lanean Time-IWwaniant S1stEms signal x(t) as one period of the T:periodic signal 28(0)s that is, = fH, -FF (o) Graph an example of +(+) and %(#) to demon- strate that, as T increases, the periodic repli cates of x(¢) in 3(e) are moved farther and farther away from the origin. Eventually, as T approaches infinity, these replicates are re- moved to infinity. Thus, we write x(0) = Jim X(e). (b) The FS representation for the periodic signal (1)is x) = 3 xikle™, where [Compcren Exreniments Show that X[k] = }X(jkw,), where X (jo) = [lxpetae (c) Substitute the preceding definition of X[k] into the expression for ¥() in (b), and show that oy 1 (1) = Imps X(jkw,)e™e,. (d) Use the limiting expression for x(t) in (a), and define w ~ ku, to express the limiting form of ‘the sum in (c) as the integral = [Mine do. 3.102 3.103 3.104 3.105 3.106 3.107 Use MATLAB to repeat Example 3.7 for N = 50 and (a) M ~ 12, (b) M = 5, and (c) M = 20. Use MATLAB's f ft command to repeat Prob- lem 3.48. Use MATLAB's 1fft command to repeat Prob- em 3.49. Use MATLAB's ft command to repeat Exam- ple 3.8. Use MATLAB to repeat Example 3.14. Evaluate the peak overshoot for J = 29, 59, and 99. Let x(t) be the triangular wave depicted in Fig, P3.107. (a) Find the FS coefficients X[k]. (b) Show that the FS representation for x(¢) ean be expressed in the form x(t) = Blk] cos(kw,t). (c) Define the J-term partial-sum approximation to x(t) as (1) = Zak costhe.0. Use MATLAB to evaluate and plot one period of the Jth term in this sum and £,(¢) for J = 1,3,7,29, and 99. Fioune P3.107 3.108 Repeat Problem 3.107 for the impulse train given by x(t) = > a(t -n). 3.109 Use MATLAB to repeat Example 3.15, with the following values for the time constant: (©) RO- 16, 3.110 This experiment builds on Problem 3.71. (a) Graph the magnitude response of the circuit depicted in Fig. P3.71, assuming that the voltage across the inductor is the output. Use logarithmically spaced frequencies from 0.1 rad/s to 1000 rad/s. You can generate N logarithmically spaced values between 10* and 10% by using the MATLAB com- mand Logspace(d1,d2,N). 3.111 3.112 ‘Addtstonal Problems (b) Determine and plot the voltage across the in- ductor, using at least 99 harmonics in a tran- cated FS expansion for the output, ifthe input is the square wave depicted in Fig. 3.21 with T= Land T, = 1/4. (c) Graph the magnitude response of the circuit depicted in Fig. P3.71, assuming that the volt- age across the resistor is the output. Use loga- rithmically spaced frequencies from 0.1 rad/s to 1000 rad/s. (d) Determine and plot the voltage across the re- sistor, using at least 99 harmonics in a trun- cated FS expansion for the output if the input is the square wave depicted in Fig. 3.21 with T= and T, = 1/4. This experiment builds on Problem 3.72. (a). Graph the magnitude response of the circuit de- picted in Fig. P3.72. Use 501 logarithmically spaced frequencies from 1 rad/s to 10° rad/s. ‘You can generate N logarithmically epaced val- ues between 10%" and 10? by using the MAT- LABcommand Logspace(d1 ,d2/N). (i) Assume that L = 10mH. (ii) Assume that L = 4mH. (b) Determine and plot the output, using at least 99 harmonics in a truncated FS expansion, if the input is the square wave depicted in Fig. 3.21 with T-2w™% 10 and T, = (m/2) x 10°. (i) Assume that L = 10 mH. (ii) Assume that L = 4 mH. Evaluate the frequency response of the truncated filter in Example 3.46. You can do this in MAT- LAB by writing an m-file to evaluate Hye) = 3 wine for a large number ( > 1000) of values of in the interval ~ < 01 =. Plot the frequency response ‘magnitude in dB (20 logyo|#;(e%)|) for the follow ‘ing values of M: (a) M=4 (b) M=10 ic) 25 (a) M= 50 Discuss the effect of increasing M on the accuracy with which H,(é") approximates H(e"). 339 3.113 Use the MATLAB command freqs or freqz to plot the magnitude response of the following systems: 8 @) 100) = Gy age)? + Bo 8 Ge)’ () 1G) ~ G54 Gu)? + Bu 41 + 36 4 gO 6+ 2678 0.02426(1 — «®)* (0196 ™)(1 + OS66IGe™ + 07657) Determine whether the system has a low-pass, high- pass, or band-pass characteristic. a 3.114 Use MATLAB to verify that the time-bandwidth product for a discrete-time square wave is approx- imately independent of the number of nonzero val- ‘ues in each period when the duration is defined as the number of nonzero values in the square wave and the bandwidth is defined as the mainlobe ‘width. Define one period of the square wave as ain = {bh O50> h = freqresp(sys,w; >> hmag = abs(squeeze(h)); >> plot(w,hmag) >> title('System Magnitu: Response!) >> xlabel(' Frequency Crads/sec)'); ylabel(*Magnitude') Use MATLAB to plot the magnitude and phase re- sponse for the systems with state-variable descrip- tions given in Problems 3.88 and 3.89, Additional Problems 413, » Heideman, M. T,, D. H. Johnson, and C. S. Burrus, “Gauss and the history of the fast Founer transform,” IEEE ASSP Magazine, vol. 1, n0. 4, pp. 14-21, October 1984. > J. W. Cooley, “How the FFT gained acceptance,” IEEE Signal Processing Magazine, vol. 9, no. 1, pp. 10-13, January 1992. The following book is devoted to a detailed treatment of the FFT algorithm: » E.O. Brigham, The Fast Fourier Transform and Its Applications (Prentice Hall, 1988) [420monat Prostems 4.16 Find the FT representations of the following peri- afr) ‘dic signals: 1 (2) 20) = Dele) + an(2e milion | I ike (b) x(t) = ZhofY cos((2k + 1)at) on SO ee (©). x(#) as depicted in Fig, P4.16(a). (d) x(t) as depicted in Fig. P4.16(b). @ Skerch the magnitude and phase specrra te PIP UTPAD I TI © Ficune P4.17 Ficune P4-16 4.18 An LTI system has the impulse response . . , qr) = 2S nc 7 4.17 Find the DTFT representations of the following pe- at riodic signals: Use the FT to determine the system output if the (a) x[#] = cos(Fm) + sin($n) input is (b) x[n] = 1+ DE. cos(Fm)s[n — m] (a) x(t) = cos(2mt) + sin(6mt) (6) *{1] ae depicted in Fig. P4.17(a). tb) x(t) = D2 a-oo(-1)"8(¢ - m) (d)_x[n} as depicted in Fig. P4.17(b). (c)_ x(t) as depicted in Fig. P4.18(a). (e) x(n] as depicted in Fis. P4.17(c). (a) x(+) as depicted in Fig. P4.18(b). ‘Sketch the magnitude and phase spectra. ()_ x(t) as depicted in Fig. P4.18(c). a8 x(t) 1 a 1 3 [ 2 > [le-, 4 ada 71 LJ ot a) x0 2 TTT ol 1A VPVi Va Vi » x0 5 AARKAAK Se 7 bia 3 © Ficune PS.18 4.19 We may design a de power supply by cascading a full- ‘wave rectifier and an RC circuit as depicted in Fig. P4.19. The full-wave rectifier output is given by a(t) = be). Let H(jw) =tm 21 Hie) = 435 = Roa be the frequency response of the RC circuit. Suppose the input is x(t) = cos(120z¢). (a) Find the FT representation of (2). (b) Find the FT representation of y(¢). (c) Find the range for the time constant RC such that the first harmonic of the ripple in y(¢) is less than 1% of the average value. om. Ts i. oe Figure P4.19, (CHAPTER @ 1 APPLICATIONS OF FOURIER REPRESENTATIONS TO MIXED SIGNAL CLASSES 4.20 Consider the system depicted in Fig. P4.20(a). The] of the input signal is depicted in Fig. P4.20(b). L x(t) + Z(ja) and y(t) + ¥(jw). Sket Z(jw) and (jw) for the following cases: (a) w(t) = cos( Sart) and b(t) = #822 (b) w(t) = cos( Sat) and b(t) (c) w(t) depicted in Fig. P4.20(c) and b(t) = 24 cos(Smt) - Xo) x0 RS hi) + R00 1 19 coset ZN + @ «0 5 ~LELIN SLI 1-, 3 1 foi 235 4 ~s © Ficure P4.20 4.71. Consider the eystem depicted in Fig, P4.21 The i pulse response is given by and we have x(t) = Dpoowtsen and a(t) ~ 3 cos(k8n2) Use the FT to determine y(t). x0 —> i) — R— by —- x0 s—> ny —J Ficune P421 Adllttumul Problems a5 4.22 The input to a discrete-time system is given by 4.25 Consider sampling the signal x(t) = 4; sin(2mt). 7 i (a) Sketch the FT of the sampled signal for the fol- x(n] ~ cos(Zn) + sio( Zn), lowing sampling interval Use the DTFT to find the ourput of the system, y[t], if the impulse response is given by sin(3r) (@) bn) = (b) Let x{x1] = x(n). Sketch the DTFT of x(r], (b) bom) = (-1)" X(e), for each of the sampling intervals alt ) given in (a). (6) bln) = cos($n): eure 4.26 The continuous-time signal x(t) with FT as depict 77 cd in Fig. P4.26 is sampled. 4.23 Consider the discrete-time system depicted in le) Sleek EET of toe serie eal ete bok lowing sampling intervals: sin(n) oT Fig. P4.23. Let h(n] = . Use the DTFT to, a (i) T (ii) In cach case, identify whether aliasing occurs. (b) Let x{n] = x(nT,). Sketch the DTFT of x{x], X(c), for each of the sampling intervals given in (a). determine the output y{1] forthe following cases: Xu) (4) x{n] = 1+ sin(Zn) + 2cos(*n), ulm) = cos( Fn) Also, sketch G(e), the DTFT of g{). sto} xi. —> tin) — irk Fioune P4.26 win) Ficune P4.23 a ee 4.27 Consider subsampling the signal x{1#) = 4.24 Determine and sketch the FT representation, Xs(jw), alia for the following discrete-time signals with the sam- that y(n] = x[q7]. Sketch ¥(e') for the following pling interval T, as given: choices of q: (a) a T=2 ib) g=4 a= 4.28 The discrete-time signal x{11] with DTET depicted in Fig. P4.28 is subsampled to obtain y[n] = x(qr]. Sketch ¥(e) for the following choices of q: OT (4) x{1] depicted in Fig. P4.17(a) with T, = 4. (b) (e) x(n] = Z7--25[" - 4p], T= 5 eq (©) x[1] = cos(3n). 416 Carin ¢ © Arruc suns ur Fourier Rernesey nuns 10 MIXED SiuvaL CLasses xe) 4.32 A reconstruction system consists of a zero-order hold followed by a continuous-time ant-magang filter with frequency response H, (ja). The original signal x(1) is band limited to 0, (ie. X(jwo) = 0 foro] > wm) and a is sampled with a sampling interval of T,. Determine the constraints on the magnitude response ofthe anti- imaging filter so that the overall magnitude response of this reconstruction system is between 0.99 and 1.01 in Ficure 4.28 the signal passhand and lessthan 10 “onal bands con- taining the images of the signal spectrum for the fol- Towing values: Sr -7 ie 2 4 4.29 For cach of the following signals, sampled with sam- pling interval T,, determine the bounds on T,, which guarantee that there will be no aliasing: {a) @, = 107, T= 0.1 (a) x(t) = }sin3mt + cos(27t) {b) w= 10, T= 0.05 sin(wt) (6) @_= 107, T, = 0.02 (by x(2) = cos(12m1) “S70 (a) o,=2, T= 0.05 (ce) x(t) = e*u(e) « (We) 4.33 The zero-order hold produces a stairstep approxi- eens at mation to the sampled signal x(t) from samples (4) x(t) = w(t)2(#), where the ET's W(jw) and x(nT,). A device termed a first-order hold Z(jw) are depicted in Fig. P4.29. cry eal ‘output of the first-order hold may be described as —™. Ae xt) = S x[nJh(e — 21), ~ fo» 45 fo a where y(t) is the triangular pulse shown in ! Fig. P4.33(a). The relationship between x[m] and Ficunt P4.29 1x)(¢) is depicted in Fig. P4.33(b). 4.30 Consider the system depicted in Fig. P4.30. Let [X(jeo)| = 0 for || > w,. Find the largest value of bao T such that x(#) can be reconstructed from y(t). De= termine a system that will perform the reconstruc- tion for this maximum value of T. x) —>R— 0) t wt) wat) LATA T ot 3 FicurE P4.30 431 Let (X(jo)| = 0 for || > 0. Form the signal ylt) = x(0)[cos(3zt) + sin({0zt)]. Determine the maximum value of »,, for which x(t) can be re- constructed from y(1), and specify a system that will © perform the reconstruction. Ficune P4.33, Adddstional Problems (a) Identify the distortions introduced by the first-order hold, and compare them with those introduced by the zero-order hold. [Hint: by(t) = b(t) * h.(t).] (b) Consider a reconstruction system consisting of a first-order hold followed by an anti-imaging fil- ter with frequency response H jn). Find H, (jen) so that perfect reconstruction is obtained. (©) Determine the constraints on |H.(je)| so that the overall magnitude response of this recon- struction system is between 0.99 and 1.01 in the signal passband and is less than 10~* on all bands containing the images of the signal spec- trum for the following values: () T= 0.05 (ii) T, = 0.02 Assume that x(t) is band limited to 12; that is, X(jw) = 0 for lol > 120. Determine the maximum factor q by which a discrete- time signal x[n] with DTFT X(e) depicted in Fig, P4.34 can be decimated without aliasing, Sketch the DTFT of the sequence that results when x[7] is decimated by the factor g. eel) §P4.34 A discrete-time system for processing continuous- time signals is shown in Fig. P4.35. Sketch the mag- nitude of the frequency response of an equivalent continuous-time system for the following cases: (a) 9, (b) 2 © xo Hw) “lr 1Or “7 £P435 a7 sin() Find and sketch [1], where Z[] < for the following values of ,: (a) 0, = 5 b) 9, = % (c) O, 436 LetX(e) = and define X{k] = X(eH) DIFS.M. , 314) 5 4.37 Let X(jw) = 2) and define X[k] = X(jkeg). Find and sketch ¥(t), where ¥(t) 22% X[k] for the following values of w,: (a) @ =F (b) () w, 4.38 A signal x(t) is sampled at intervals of T, = 0.01s. One hundred samples are collected, and a 200-point DTFS is taken in an attempt to approximate X(jw). Assume that |X(jw)| = 0 for Jo] > 120m rad/s. Determine the frequency range 0, 127 rad/s. De- termine the frequency range ~w, < w x(i(o- £e, i) ~ 75,22 G2) (b) The goal is to express ¥s(io) as a function of X,(jw) so that ¥(e'") can be expressed in terms of X(e!), To that end, write k/q in Yg(je) as the proper fraction k £ai+%, a 4 where | is the integer portion of k/q and m is the remainder. Show that we may thus rewrit Ye(jw) as, . wi = ASE $ a((o- te -2e))} ww) = BSLx((»- 2a). (6) Now we convert from the FT representation back to the DTFT in order to express ¥(e) as a function of X(e"), The sampling interval as- sociated with Ye(jo) is qT. Using the relation- ship © = wq7,In ¥(e%) = Kio)I,.9 + Ma F)) (d) Last, use X(e) = X,(j%) to obtain x(e-m), ‘A band-limited signal x(t) satisfies |X(jwo)| = 0 for lol < eo and [a] > op. Assume thar io, > ay — 0. In this case, we can sample x(t) at a rate less than that indicated by the sampling interval and still per- form perfect reconstruction by using a band-pass re- construction filter H,(jw). Let xm] = x(nT,). Determine the maximum sampling interval T, such that x(t) can be perfectly reconstructed from x]. ‘Sketch the frequency response of the reconstruction filter required for this case. 4.44 Suppose a periodic signal x(t) has FS coefficients {Q es 4 0, otherwise” show that ¥(e%) 1 ey = yer) = 3 443 XA 445 oy @ ‘Additional Problems ‘The period of this signal is T = 1. (a) Determine the minimum sampling interval for the signal that will prevent aliasing. ‘The constraints of the sampling theorem can be relaxed somewhat in the case of periodic signals if we allow the reconstructed signal to be a time- scaled version of the original, Suppose we choose a sampling interval 7, = #9 and use a re~ construction filter Filia) = te anes 0, otherwise’ Show that the reconstructed signal is a time- scaled version of x(t), and identify the scaling factor. Find the constraints on the sampling interval T, so that the use of H,(jw) in (b) results in the re construction filter being a time-scaled version of x(t), and determine the relationship between the scaling factor and T,. In this problem, we reconstruct a signal x(t) from its samples x{n] = x(nT,), using pulses of width less than T, followed by an anti-imaging filter with fre~ ‘quency response H,(jw). Specifically, we apply 20) = alelbp(t ~ nT) to the anti-imaging filter, where hp(t) is a pulse of width 1,, as depicted in fig, #4.49 (a). An example ‘of x(t) is depicted in Fig. P4.45 (b). Determine the constraints on |H,(j)| so that the overall magnitude response of this reconstruction system is between 0,99 and 1.01 in the signal passband and less than 10~ on the band containing the images of the signal spectrum for the following values, with x(t) band limited to 10%—that is, X(jw) = 0 for » > 10m: (a) T= 0.08, 7, = 0.04 (b) T= 0.08, T, = 0.02 () T= 0.04, 7, = 0.02 (d) T= 004, 7, = 0.01 (by (C) xa Figure P4.45 4.46 447 419 ‘A nonideal sampling operation obtains x(n] from x(t) as [ a(t) de ln-vt, (a) Show that this equation can be written as ideal sampling of a filtered signal y(t) = x(t) * h(t) fice., x[m] = y(nT,)], and find h(+), Express the FT of x[n] in terms of X(jw), H(fo), and T,. Assume that x(t) is band limited to the fre- quency range lol < 32/(4T,). Determine the frequency response of a discrete-time system that will correct the distortion in x{1] introduced by nonideal sampling, ‘The system depicted in Fig. P4.47(a) converts a con- inuous-time signal x(¢) to a discrete-time signal y{n). We have (b) 0 my ft, lle § HD {i otherwise’ Find the sampling frequency @, = 2/T, and the constraints on the anti-aliasing filter frequency re- sponse H,(jw) so that an input signal with FT (jw) shown in Fig, P4.47(b) results in the out- put signal with DTFT ¥(e"*). x0) x0) Sample at 20> Hf) —> SO ——> Hie!) > yi Ficure P4.47

You might also like