Scribd
Upload
1K views

Solution Part1

A) Determine the function f(t) whose Fourier transform is shown in figure P-3.l.a. B). C) Sketch f (t) and g (t) near t = 0. What effects does the phase has on the symmetry of this waveform?

Uploaded by

Valeri Suarez
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
1K views

Solution Part1

A) Determine the function f(t) whose Fourier transform is shown in figure P-3.l.a. B). C) Sketch f (t) and g (t) near t = 0. What effects does the phase has on the symmetry of this waveform?

Uploaded by

Valeri Suarez
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 30
Problem 3.1.1 a) Determine the function f(t) whose Fourier transform is shown in figure P-8.La. b) Determine the function f(t) whose Fourier transform is shown in figure P-8.1.b. ©) Sketch f(t) and g(t) near t = 0. What effects does the phase has on the symmetry of this waveform? F(a [eles | rie Olea wea Ae eee e(w)) (0) 2 w/t @ ® Figure 9: P-3.1.1 "The function f(t) can be obtained from F(w) by doing an inverse Fourier transform, ie, S(t) = F{Fw)} = zl. Flw)e*at - Lh ice sinWe f(t) = let de = Wr b) f= Lf rene a f'n Sr eittdy = wi (75) (76) c) The sketches for the time functions are given in the figures 10, 11. In the first case, the Fourier transform is real and even resulting in a real, even time function. In the second case, the Fourier transform is pure imaginary and odd resulting in a real, odd time function. Show that if F(w) = F{f(#)}, then Fo) = [Hoe 24 (78) Figure 10: The function f(t) = sin(Wt)/t for W = 1, even symmetry. Figure 11: The function f(t) = [1 — cos(W't)]/t for W = 1, odd symmetry. Pel sf \eeplas (7) °) ‘ Bg [ree (80) _ i ie S(t-+ re" drduodt = 2x F(0) (81) F(0) = Lf Noe**at| al. FjePar= [Fea (82) b) Using the triangle inequality, |A + B| < |A| + |B), and considering integration as a limit of a summation, it can be shown that: [fF ocmar| < J railem|ar= J iri. (83) Note that Je~#*| = 1, 5) Fla | 2 [% otrwrertas] s 2 [ot rtoyiae (es) Note that the same inequality used in part b) was used. d) Note that, [User netar = or) (85) then, we have, LEL Slt} re" drducdt f. [oPeyetavat =2n [ S()dt =2nF(0) (86) Problem 3.2.1] a) Find the Fourier transform for of the raised cosine pulse signal defined by: ltcosrt if -11. (93) c) Using Buler’s identity, we get directly an expansion of our function in exponential form as 1 S(t) = 1+ cos nt) =14 5( eM) (94) from which it is evident that Fo , Fy = Fy =1/2 and F, =0 for |n| > 1. Problem 3.2.2) a) Find the spectral density of the real-valued function: aexp(-at) t>0 (t) = (95) ti={ Sa tee ) b) Examine your answer to part a) for the special cases 6 = a and 6 = ~a, particularly with respect to the following chart: f(t) F@) Fe) even even real odd odd imaginary 27 Figure 12: Raised cosine pulse. Solution: a) The Fourier Transform can be written as: ~ a(6-+0) + ju(b— Fo) =f elorevat yf acW(HieNgy = sb toy folio} (98) ») From the above expression we obtain the following cases: dat eeu F( (97) pd if a=-5 ‘These conclusions are consistent with the chart given above, with 6 = a corresponding to the first row and 6 = —a to the second row. Problem 3.2.3] A pulse signal described by f(t) = exp(—alt|)rect(t/T) is repeated periodically with period T- a) Find the exponential Fourier Series beginning with the Fourier transform of exp (~alt|) and the converting to the series. b) Compare your answer with problem 2.7.1. ©) Under what conditions can equation 3.15 (equation 3.15 is as follows: F, = +F(nw)) be used to obtain the Fourier series coefficients ? a) For one pulse , 2a oral (from table 3.1) (98) 28 Using equation 3.15: 2a/T leenee = (tat T Tat (99) b) For T ae (100) But, problem 2.7.1 gives: py = Slo ecteosn) acy a nia? Relationship (100) is indeed the general term of the Fourier series representation of the periodic function obtained by repeating the function exp(—ale|) while (101) is the general term of the Fourier series representation of the periodic function obtained by repeating the function f(t) = exp (~alt|)rect(t/T) ¢) For a give function f(t), the formula F, = 4F(nwo) gives the Fourier coefficients of the periodic signal S7f2_o. f(t — KT). These are equal to the Fourier coefficients of f(t) as long as J(t) is zero outside of the time interval T. In this case, the replica of the original signal, namely F(t- kT), do not have overlap and the resulting periodic signal, namely D7... f(t— KT), is equal to the original signal f(t) over the time interval T yblem 3.2.4] The time function f(t) = (1/ov2m)e~*/2* (o= constant) is known as the Gaussian function. This function has finite energy and thus is Fourier transformable. Fin Fourier transform. In carrying out your solution it will be helpful to combine exponents, complete "du = /®. Note that f(t) and ‘F(w) have the same mathematical form; i.e., the Gaussian function is its own Fourier transform. the square in the exponent, and then use the finite integral [, Solution: al ~ ovam © prl(Pitetetnotatsote?)/2e%ldy (192) (108) [Problem 3.4.2] Show that a more general statement of Parseval’s theorem for for energy signals than Bq. (3.21) is f F()g" (at = z £ F(w)G"(w)dw. (104) 29 Solution: [-sortwoa= [~ [= [- Feer'a] [Ef ewe *a] de (105) “Allowing an interchange in the order of integration, this becomes: = re S Pee : ~ite-w)t £ M09" Ote= Ges Fw) Ce dusdedt, (106) Using Eq, (8.25) of the main text, this becomes: f. F(t)g" (at = z £. £. F(w)G"(2)5(w — 2)dedw (107) Problem 3.4.3] Evaluate the following definite integrals using Parseval’s theorem: a) [isn 2/2}dz, b) _ de/(a? +2), 3 f dz/(a? +22)? Solutions] We know that, a) rect(t/2) <> 2Sa(w) (108) then, f. ldt = sl. [sin w/w)Pdo (109) resulting in, = f cin w/uo]Pdao = (110) b) ettult) = 2 fa (aut) dy = a [ (edt = n/a, a>0. (112) °) ott (13) a z [~e lye = Sf Mat= a> 0 (ud) Use the result of problem 3.4.2 to evaluate the following integrals, for a > 0, > 0: 30 ’ i @F ayer Faye tery Sa(bi) Loo 2+ a? da. Solution: a) , ele (115) Feu % a(attiegy = 7% a = ey 9) rier le l. Cray =o a ea a Lrect(t/x) = Salwr/2) (17) [Bihar ate ft Bea(B) Brena Problem 3.5.1] Use Eq. (3.25) and an interchange in the order of integration to show that: a) FNFLF(O} = F(t), b) FAF{F(O} = F(-2) ») dt = a0 —e%), (118) Solution: 3) 7 Furon =z f[f sneer] de as) FeO = ge [ se) [dade = J peryse-nidr = 409). (120) ») FALFUFON = £ i sey ar] eds = £ roy [fe etter] dr, (221) FEF = [ floyste + rar = 1-9 (122) [Problem 3.6.1} a) Find F(w) for the f(t) shown in Fig. P-8.6.1 (a), (b). b) Sketch |F(w)| for r < 7 for both cases, and compare = Are*TSa(wr/2), (128) = Arei*™Sa(wr/2) + Are" Sa(wr/2) = 2Acos(wT)Salwr/2). (124) 31 A) fle) a T T ° Tr Figure 13: P-3.6.1 b) Graphs for | (w)| and |F,(w)| are shown if Figures (14) and (15), respectively. $5 it Figure 14: The function ASa(wr/2) for A= 1,7 Use the modulation property to find the function f(t) whose Fourier transform is shown in Fig. P-3.6.2 for the conditions: a) B=A; b) B=-A. Solution: a) For A= B, we have, 32 { i V 1 {fife y Sa Figure 15: The function 2A cos(wP)Sa(wr/2) for A= 1,7 =1, T= 2. Fw) 2A » |b aw Figure 16: P-3.6.2 Pw) = 2A rectw/2W] + Arect{(w ~ p)/2W] + Areet[(w + 0)/2W) (128) Using Table 3.1 and the modulation property of the Fourier transform, F0) = 2A saquve + AM emtsacive + Ame tsa(wt) (126) = oan + cos wot}Sa( WW). (27) b) Similarly, for A = —B, we have, Fw) = 2A rect |w/2W] - Arect{(w — wo)/2W] — Arect{(w + wo)/2W] (128) and f(t) = 2a saws) - aL oa'sawe = awe Juntga(Wt) (129) = 24% [1 ~ cos wotlsa(we) (130) 33. If {(®) © F(w), determine the Fourier transform of a) f(2-1), b) Fl(t/2) - 1), ©) F(t) cos a(t -1), &) S120) Solution: a) Assuming € = 2- t, we obtain, £ (2— thet = [7 Hey HO-Bagw oe) ‘b) Assuming € = (t/2) — 1, we obtain, fllt/2) — tje~**at = af” S(gje +2) ag = 20-599 F(Qw), c) Substituting, cos(mt ~ m) = [eil**— + eal") /2, we obtain, fem [seta +5e" f. He etde = TF w—m) + Feta) Note that e/* = e9" = -1. d) We have, Fe) [Feces fet) =f F@e*ag Stren) =f urereerae Assuming w = 2¢, we obtain, Suen = [* jartenetae/2 ‘This results in, 4 qe] = (u/2)F(u/2) Problem 3. (131) (132) (133) (134) (135) (136) Find the Fourier transform of the pulse waveform f(t) shown in Fig. P-3.6.5 by differentiating to obtain impulse functions, then writing the transform using the delay and integration properties. [Hint: Consider use of superposition.) 34 Figure 17: P-3.6.5 Solution: ‘Taking two derivatives of f(t), we obtain f"(t) = (—A/r)4(t + r) + (A/r)d(t — 7) + 246"(t) ‘Therefore, we can write, o= ff ieammoc rn + (armme—mdcag + [* aan ae (as7) which has a Fourier transform (A/r)ei*" + (A/ne" 2A de Gey? je = Ho = Sa(wr)] (138) ‘wo functions of time, f(t) and g(t), are known to satisfy the following integral equation: ate) = [7 alr) fle dr + 509) (139) a) If f(t) = exp(—at)u(t), find g(t) b) If f(t) = exp(~altl), find g(t). [Solution: | Computing the Fourier transform of the two sides of our main relationship, we obtain, G(w) = Cw) Fw) +1 (140) (a1) a) From Table 3.1, for f(t) = exp(—at)u(t), we have F(w) = 1/(a+ jw) and, jute jute gt) = S(t) te (Mut), a> GW) = b) From Table 3.1, for f(t) = exp(—alt]), we have, w+ 2a orat—aa + eat ae g(t) = 8+ Ve, a > 2. Problem 3.7.4] Consider the following (volume) integral 2 L S(a)h(w)5|t — (u+ v)]dudv. 1) Show that this integral describes the convolution integral, f(t) + h(t). b) Using this integral and the result of (a), show that f(t) + A(t) = A(t) + f(t). (142) (143) (144) (145) (148) (147) c) Use this integral to show that the area under the convolution result of two given functions is equal to the product of the areas under the two functions. 4) Repeat part (c) using the inverse Fourier transform of one function in f(t) h(t) and then interchanging the order of integration. a) Noting that the impulse function is at v = t— u, the integration over v yields: [some du = f(t) * h(t). b) We could just as well have chosen the integration in (a) over u f f(t — v)h(v)du= h(t) * fe). 36 t—v to give: (148) (149) L. Lf seancste- (e+ eylauaat (150) Lf sean [ate (e+ oyataude = (151) £ £ f(u)h(v)dudv = (152) [sede [™ soo (153) d) We have, [ stosnna= [~~ sonar) arae= f~ [™ 5 [Pleyel darter) ar at (158) Assuming € = t ~ 7, we obtain, £ ref F(w)e*"dwh(t —r)drdt = (155) iff Pea |fZ nee a] aed = (15) [rome [oa - cs [Fey )5(w) 4a = FO)H(0) (158) Problem 3.7.3, Use frequency convolution to prove the following trigonometric identities (also see Fig. 3.3) a) 2cos? ws b) 2sin? wot ©) 2eoswit coswat = cos(iw1 +w2)t + cos(wr ~ we)f. + cos 2unt, — cos 2wyt, a) We know that cos(wot) => m[5(w ~ wo) + 5(w +w0)), and, f(t)g(t) <=> [F(w) + G(w)]/2n, then, Dees? wot > 2 {r[Slw— wy) + (w+ we}] 4 IBle — un) + Slut u0))} = (159) 2 [7 16 (ews) +6(u+erjrli(w-a~e9) + He u-ten)idu = (160) 175 (w — 2uto) + 25(w) + 15(w) + ¥5(w + 2ur0) <=> 1 + 008 Quot (asi) b) We know that sin(wot) > —jx[6(w — wo) — 6(w + wo)], then, Bsin? wot > 2 {—jn[ Bw — ay) ~ Sw + w9)] + —Jn6(w ~ wy) — Slut wn)]} = (162) 37 = £ —1[6(u— wp) ~ 6(u-+ wo)]n[6(w — u — wp) — 6(w — w+ wo)]du —b(.o — 2uvo) + w5(w) + #5 (w) — 75(w + 2p) <=> 1 — cos Qunt Booswrt cos wat —p 2 F {a[S(w —ws) + (w+ 03)] 4 x[Ble — ws) + 6(w + 3))} af. n{6(u — wr) + (w+ wi))ef6(w Blu ~ wy ~ we) + 28(w + wr — w2)+]5(w ~ 04 +02) + 75lw + oy + 2) ~ us) + 5(w — w+ w2)]du => cos(wy + w2)t + cos(ier — wn)t ‘Problem 3.8.1 (163) (164) (165) (165) (167) (168) Sketch the results of the following convolution operations (where t > 0 in all cases). Check your result by writing a Fourier transform, of each function, multiplying and writing the corresponding time function. a) A6(t) + Bé(t — to); b) Ad(t + to) * BA(t — to); ©) Ad(t ~ ty) + BO(t ~ to); d) A[d(t + to) + 5(t — to] « B[S(t + to) + 8(t ~ to)]; e) A[d(t + tr) + 4(t — th] ¥ B[S(t + to) + 6(-to)], to > ty Solution: Recall that convolution of two functions f(t) and g(t) is given by: He +a() = [~ Foralt- rar; Using this, the convolution of our given functions becomes: a) ‘ AS(t) # Bé(t~ te) = [ A5(r)BO(t— 7 — to)dr = ABB(t — to); FLAS} = A; F{BE(t—t) = Bem; The product is ABe~**, and the inverse Fourier transform of the product is, F-{ABe-#*"} = ABS(t ~ to). which agrees with our equation obtained earlier. b) wo Ab(t-+ to) # B&(t — to) yl AS(r + to) BO(t — 7 — to)dr = ABS(t); 38 F{AS(t + to)} = Ae"; F{BA(t — to)} = Be; The inverse Fourier transform of their product is given by: F{ Aci" Be“ to} — ABS(t) Which agrees with our previous result, °) 7 Ab(t ~ t) + BO(t — to) ii Ab(r — 1) BS(t — 7 ~ to)dr = ABS(t — to ~ 4); F{AS(t—ty)} = Ae"; F{B(t — to)} = Be"; ‘The inverse Fourier transform of the product is given by: Fo Ae7iet Beto} = ABS(t ~ to ~ th) Which agrees with our previous result, 4) The convolution expression given is equal to: [AB (r+ t0) + 5(7 to) BIB(t 1 + to) +3(E- 7 ~ t)]dr = ABLE(E + 2a) + 25(2) + 5(E~ 2) By evaluating the Fourier transform of the individual terms, multiplying and finding the inverse Fourier transform, as done in parts a)—c), it can be found that: Ff Acie + Fe] Bleieto 4 eIle]} = ABLS(t + 2to) + 26(t) + 4(t — 2to)] which agrees with our previous result, ¢) The convolution of the given expression is equal to: [AUC 15) + 8 = BIB (E140) +8 7 = olde = AB[S(t + th + to) + S(t — tr + to) + A(t + tr — to) + 5(t — tr — to)]s Likewise , Ff Ale -e- Ft] Bletto + eS} = AB[S(t+t1 +10) +4(t+t) ~ to) +46(t—th +to)+4(t—t —to) |; Which also agrees with the above result. Problem 3.8.2] Evaluate the following convolution integrals; check your result by taking the Fourier transform of each function , multiplying, and finding the inverse Fourier transform. a) u(t) ¥e-*u(t); b) em ttu(t) + eM u(t); c) etltl x coswot 39 u(t) +e-tu(t) £. u(rjeu(t = r)dr = f ede ‘The integral is zero if t < 0 and is equal to (1 - e~*) if > 0. Then, we have, u(t) #e“u(t) = (1 eu). ‘The product of their Fourier transforms is given by: 4 + rae] (——]- 1, BW) 4 dg atu) s+6 [eal = joGwrt) * wrt > jo” jor tO) ‘The Inverse Fourier transform of the above expression is given by: (1-e*)u(t) ’) . ettu(t) +e Mu(t) et u(r)e Mult — rar entre MN _ [ testutty a= | [let -e*) Ha- due) a #6 ) Following the same procedure, it is seen that: ell ecosunt = £ enol" cos(wor)dr set i at costor)de + ett [ e-** cos(uyrdr tee ‘The Product of their Fourier transforms is given as: [a = etl [r6(w ~ wo) + 26(w +u0)] [rd(w ~ wo) + 5 (w + wo) ‘The Inverse Fourier transform of this expression is given as, 2a wpag coment ‘This is also consistent with the earlier result. 40 Problem 3.8.3] Evaluate f;(t) + fa(t), fa(t) + fa(t), and fo(t) * fo(t) for functions defined by waveforms shown in figure P- 3.8.3. A | fate) Alt) Solution: a) We shift fr with respect to fo, this results in, A®sh) = o 0? Solution: a) (RC) HO)= ARO) where Ry = (R1R2)/(Ri + Ra) Using the relationship between the input and output power spectral densi get the result: ofa linear system we A(Ric)? 2uF+1/(R,CP b.) Using table of Fourier transforms, we get, Sun(wo) = a_i fim VEO = ore ee Integrating S,,(w) results in the same value. d.) As C+ 0, H(w) becomes a resistive divider and because the input source is assumed to be white, the mean square output + co. 50 Problem 4.4.1]a.) Determine the power spectral density of F; exp (jwot) by first finding the autocorrelation function, then taking the Fourier Transform of the autocorrelation function. b.) Repeat part a.) for [Fy exp (jwot) + Frexp (j2uot)]. ¢.) Extend your result in part (b) to S739. Fue" ny 8 [Re minetenar = ne” Thon * . Splw) = FA Ry(7)} = 2m|Fi/?5(w ~ we) 1 ptr - >) wo (7) a) = BL remanent 4 nt = [Fier [reer uFrom this we get: S4(w) = 2n|Fy|?6(w — wo) + 2m] Fe[*5(w — 2wo) c.) The cross-terms integrate to zero and we get: Ry(r) = Caw [Falter ‘The power spectral density is then found to be given by: = S IFalt6(w — nao). Ss(w) ‘The complex pulse signal et OStch ) Ae Na ast 21. a.) Find the autocorrelation function Rg(r), and sketch the real part. b,) Find and sketch the power spectral density Sj(w). Solution: a.) We want to compute the integral, RG) =F [Poser nae We select the range of the integration as (0, 7]. We have, ewe OStct 9G < ron er 51 Note that for + < 0, the function f(t +1) is obtained by shifting f(t) towards positive ¢ axis and vice versa. For ~t; <7 <0, we have, 0 O ty. Similarly, for 0 <7 < ty, we have, Ofort <—r, ciel) O t1Sa(wty /2) => 1780 (wts/2) — tA(r/t). °) Sa(Wt) > Zrect{w/(2W)] => (=) rect[w/(2W)] => Wu) 53 (77) (178) (a79)

You might also like