Frequency Response Dost thou love Life? Then do not squander Time; for that is the stuff Life is made. Benjamin Franklin AWA Career in Control Systems Control systems are another area of electrical engineering where circuit analysis is used. A control system is designed to regulate the behavior of one or more variables in some desired manner. Control systems play ‘major roles in our everyday life. Household appliances such as heat- ing and air-conditioning systems, switeh-controlled thermostats, wash- ers and dryers, cruise controllers in automobiles, elevators, traffic lights, manufacturing plants, navigation systems—all utilize control systems. In the aerospace field, precision guidance of space probes, the wide range of operational modes of the space shuttle, and the ability to maneuver space vehicles remotely from earth all require knowledge of control systems. In the manufuecturing sector, repetitive production line operations are increasingly performed by robots, which are pro- grammable control systems designed to operate for many hours with- out fatigue. Control engineering int it theory and communication theory. It is not limited to any specific engineering discipline but may involve environmental, chemical, aeronautical, mechanical, civil, and al engineering, For example, a typical task for a eontrol system neer might be to design a speed regulator for a disk drive head. ‘A thorough understanding of control systems techniques is essen- tial to the electrical engineer and is of great value for designing con. trol systems to perform the desired task. tes ci le {A welding robot. © Vol. 1 PhotoDise Getty Images 63 oOon The frequency response of a circuit may also be considered as the ‘variation of the gain and phase with frequency. Xo) Linearnetwork |__ Yo) Input Mie) Output Figure 14.1 [A block diagram representation ofa linear network In this context, X(w) and ¥(w) denote the input and output phasors of a net- ‘work they should not be confused ‘with the seme symbolism used for re actance and admittance. The multiple usage of symbols is convertionaliy| permissible due to lack of encugh let- tes in the English language to express all crcuttvaisbles dstincty. Chapter 4 Frequency Response. 441) Introduction In our sinusoidal circuit analysis, we have learned how to find volt- ages and currents in a circuit with a constant frequency source. If ‘we let the amplitude of the sinusoidal source remain constant and vary the frequency, we obtain the circuit's frequency response. The frequency response may be regarded as a complete description of the sinusoidal steady-state behavior of a circuit as a function of frequency. ‘The frequency response of a circuit is the variation in its behavior wth change in signal frequency ‘The sinusoidal steady-state frequency responses of circuits are of significance in many applications, especially in communications and control systems, A specific application is in electric filters that block. ‘out or eliminate signals with unwanted frequencies and pass signals of the desired frequencies, Filters are used in radio, TV, and telephone systems to separate one broadcast frequency from another ‘We begin this chapter by considering the frequency response of simple circuits using their transfer functions, We then consider Bode plots, which are the industry-standard way of presenting frequency response, We also consider series and parallel resonant circuits and encounter important concepts such as resonance, quality factor, cut? frequency, and bandwidth, We discuss different kinds of filters and net- work scaling. In the last section, we consider one practical application of resonant circuits and two applications of files. 42) Transfer Function The transfer function H(w) (also called the network function) is a useful analytical tool for finding the frequency response of a circuit In fact, the frequency response of a circuit is the plot of the circuit's transfer function H(w) versus @, with @ varying from w = 0 to A transfer function is the frequency-dependent ratio of a forced function to a forcing function (or of an output to an input). The idea of a transfer function was implicit when we used the concepts of impedance and admittance to relate voltage and current. In general, 2 linear network can be represented by the block diagram shown in Fig. 14.1 ‘The transfer function H(w) of a circuit is the frequency-dependent Fatio of a phasor cutput ¥(w) (an element voltage or curent) to a phasor input X(w) (source voltage or curent). Thus, _ Yo) X(o) (asap142 Tarsferfurction assuming zero initial conditions, Since the input and output can be either voltage or current at any place in the eireuit, there are four possible trans- fer functions H(o) = Voltage gain = Y®? (14.22) seen Vio) - H(o) = Current gain = (14.20) H(o) = Transfer Impedance = (1420) Le) “ H(o) = Transfer Admitance = ¥O (1424) Vio) ” ‘where subseripts and o denote input and output values. Being a complex quantity, H(q) has a magnitude H() and a phase «5 that is, H(o) Hw)fo. ‘To obtain the transfer function using Eq. (14.2), we first obtain the frequency-domain equivalent of the circuit by replacing resistors, inductors, and capacitors with their impedances R, jes, and 1/jaC. We then use any circuit technique(s) to obtain the appropriate quantity in Eq, (14.2). We can obtain the frequency response of the circuit by plot- ting the magnitude and phase of the transfer funtion as the frequency ‘varies. A computer isa real time-saver for plotting the transfer function, ‘The transfer function H(w) can be expressed in terms of its numer ator polynomial N(wo) and denominator polynomial D(w) as Nw) Ho) (Deo (143) where N(w) and D(w) are not necessarily the same expressions for the input and output functions, respectively. The representation of Hw) in Eq, (14.3) assumes that common numerator and denominator factors in H(oo) have canceled, reducing the ratio to lowest terms. The roots of N(w) = 0 are called the zeras of HMw) and are usually represented as jo Similarly, the roots of Dw) = 0 are the poles of Hw) and are represented as joo = ps, p>, ‘A zeto, a8 2 root of the numerator polynomial, is 8 value that results in 2 ze10 value ofthe function. A pole, 28a root of the denominator poly- nomial is a value for which the function is infinite. ‘To avoid complex algebra, itis expedient to replace je temporarily with s when working with H(i) and replace s with jo at the end. For the RC cireuit in Fig, 14.2(a), obtain the transfer function V,/V, and its frequency response. Let v, = Vy cost Sotuti The frequency-domain equivalent ofthe circuit is in Fig, 14:2(b). By votage division, the transfer function is given by Vs. ioc 1 No) = "Rt Ifo” 1+ joRC os ‘Some authors use H(i) for trenster instead of Hw), since w end jare an inseparable pair ‘A zero may also be regarded as the: value of s = jw thot makes H(s) zer0, ‘ard a pole as the value of s = jon that makes H(s) infinite.616 0707 co) as: igure 14.3, Frequency response ofthe RC circuit: (a) amplitude response, (6) phase response. Figure 14.4 RL circuit for Practice Prob. 14.1 rapier 4 Frequency Response @ ® Figure 14.2 For Example 14.1: (a) time-domain RC circuit, ( frequency-domain RC circuit ‘Comparing this with Eq, (9.18), we obtain the magnitude and phase of H(o) as L oy H : tan V1 + (@/o0)? where wp = 1/RC. To plot H and ¢ for 0.< <*, we obtain their values at some eritical points and then sketch. Ato =0,H=1 and 6 = 0. Atw = », H = 0 and d = —90", Also, at @ = wp, H = 1/-V2 and —45°. With these and a few ‘more points as shown in Table 14.1, we find thatthe frequency response is as shown in Fig. 14.3, Additional features of the frequency response in Fig. 14,3 will be explained in Section 14.6.1 on lowpass filters. ears For Example 14.1 w/a H 6 w/e 4 ° 1 0 10 on 1 on 45 2 0.05 “7 2 04s -8 100 ool “7 8 032 - 2 ° Obtain the transfer function V,/V, of the RE circuit in Fig. 1 assuming v, = Vp, cosiot. Sketch its frequency response. Answer: jul /(R + joL); see Fig, 14.5 for the response, Be as © Figure 14.5 Frequency response ofthe RZ circuit in Fig. 144,443. TheDecibelScate on For the circuit in Fig. 14.6, calculate the g and zeros. ()/I,(w) and its poles : a du By current division, o® ase Le) = Pw) = O)~ F4 pw + 177050" Figure 14.6 For Example 142. or Lo) _ i504 +20) _ fs +2) _ (@) 1+ fo+ Guy eae * The zeros are at se+2=0 + The poles are at S++ 1=6+IP=0 Thus, there is «repeated pole (or double pole) at p = —1. nd the transfer function V,(w)/I,(w) for the circuit in Fig. 14.7. Pre Obtain its zeros and poles “vo 10s + 18 +3 Answer E+ WET) oj, 3; poles: ~0.683, we See Pek tS 2 r Figure 14.7 For Practice Prob. 142. 14.3 1 ‘The Decibel Scale It is not always easy to get a quick plot of the magnitude and phase of the transfer function as we did above, A more systematic way of obtain- ing the frequency response is to use Bode plots. Before we begin to construct Bode plots, we should take care of two important issues: the use of logarithms and decibels in expressing gain, Since Bode plots are based on logarithms, it is important that we keep the following properties of logarithms in mind: 1. log P,P, = log Py + log Ps 2, log Pi /P2 = log P, — log P; nog P In communications systems, gain is measured in bels. Historically, J Historical note: The bel is nemed ater the bel is used to measure the ratio of two levels of power or power | Alexender Graham Bell the inventor of gain G; that is, the telephone. 4a)Grapter 4 Frequency! Alexander Graham Bell (1847-1922) inventor of the telephone, ‘was a Scottish-American scientist. Bell was bom in Edinburgh, Scotland, a son of Alexander Melville Bell, a well-known speech teacher. Alexander the younger also became fa speech teacher after graduating from the University of Edinburgh and the University of London, In 1866 he became interested in trans ting speceh electrically. After his older brother died of tuberculosis, hi father decided to move to Canada. Alexander was asked to come to Boston to work at the School for the Deaf. There he met Thomas A. Watson, who became his assistant in his electromagnetic transmitter experiment. On March 10, 1876, Alexander sent the famous first tele- phone message: “Watson, come here I want you.” The bel, the loga- rithmic unit introduced in Chapter 14, is named in his honor. i , \ oa P Figure 14. Voltage-current relationships fora four- terminal network The decibel (4B) provides us with a unit of less magnitude. It is 1/10th of| abel and is given by ?, Gag = logo 5 ass) When P, = Ps, there is no change in power and the gain is 0 dB. IF Pa = 2P\, the gain is Gan and when P: = 0.5P, the goin is Gay = 10 logy 0.5 = -3 dB (14.7) Equations (14.6) and (14.7) show another reason why logarithms are areatly used: The logarithm of the reciprocal of a quantity is simply negative the logarithm of that quantity Alternatively, the gain G ean be expressed in terms of voltage ‘or current ratio, To do $0, consider the network shown in Fig. 148, If P, isthe input power, P. is the output (load) power, Ry isthe input stance, and Ry is the load resistance, then P, = 05/2) and OSV3/Ra, and Eq, (148) becomes 0 logo 2 ~ 3 4B (14.6) Gap = Wlogyo'p: = 10 loei0 v3 9 i UR, 48) 10 logio a4) R, For the case when R> = Rj, a condition that is often assumed when comparing voltage levels, Eq. (14.9) becomes Vs Gag = Wlogs0 7 (14.10)14a Bode Mots Instead, if P; = 77R; and Py = FRe, for R, = Re, we obtain 0 Josie 2 a4. Can ‘Three things are important to note from Eqs. (14.5), (14.10), and (14.11): 1. That 10 logio is used for power, while 20 logyo is used for voltage or current, because of the square relationship between them (P= V7/R= PR) 2. That the dB value is a logarithmic measurement of the ratio of one variable to another of the same type. Therefore, it applies in expressing the transfer funetion 7 in Eqs, (14.2a) and (14.2b), which are dimensionless quantities, but not in expressing in Eqs. (14.2c) and (14.20. 3. Itis important to note that we only use voltage and current mag- nitudes in Eqs. (14.10) and (14.11). Negative signs and angles will be handled independently as we will see in Section 14.4. this in mind, we now apply the concepts of logarithms and deci- bels to construct Bode plots. 44) Bode Plots Obtaining the frequency response from the transfer function as we did in Section 14.2 is an uphill task. The frequency range required in frequency response is often so wide that it is inconvenient to use a linear scale for the frequency axis. Also, there is a more systemati way of locating the important features of the magnitude and phase plots of the transfer function. For these reasons, it has become stan- dard practice to plot the transfer function on a pair of semilogarith- mic plots: the magnitude in decibels is plotted against the logarithm of the frequency; on a separate plot, the phase in degrees is plotted against the logarithm of the frequency, Such semilogarithmic plots of the transfer function—known as Bode plots—have become the indus- tty standard. Bode plots are semilog plots ofthe magnitude (in decibels) and phase (in degrees) of a varsfer function versus frequency Bode plots contain the same information as the nonlogarithmic plots discussed in the previous section, but they are much easier to construct, as we shall see shortly. ‘The transfer function ean be written as H= H/b = He (14.12) Taking the natural logarithm of both sides, InH= In + Ine = In +j6 (04.13) Thus, the real part of In H is a function of the magnitude while the imaginary part is the phase. In a Bode magnitude plot, the gain Ho = 20 logo H (414) a9 Historical note: Named after Hench \W. Bode (1905-1989), an engineer ‘withthe Bell Telephone Laboratories, forhis pioneering workin the 1930s and 19406,rere) Specific gain and their decibel values.* Magnitude 7 20 login H1 (dB) 0.001 = 60 01 ~40 Od -20 08 -6 ava -3 L 0 vi 3 2 6 10 20 20 26 100 40 1000 0 * Some ofthese values ate approxima J] Teorginiswherew = 1 orloge and the gain is zero, Acdecade isan intenal between two, frequencies with aratio of 1033, bbetween on and Thos, or between 1Qand 100 Hz, Thus, 90 dB/decadk means thatthe magnitude changes £20 dB whenever the frequency ‘changes tenfold or one decade, Chapter 4 Frequency Response. is ploted in decibels (AB) versus frequeney. Table 14.2 provides a few values of H with the corresponding values in decibels. In a Bode phase plot, & is plotted in degrees versus frequency. Both magnitude and phase plots are made on semilog graph paper. ‘A transfer function in the form of Eq, (14.3) may be written in terms fof factors that have real and imaginary parts. One such representation might be LI + j2Cro/org + Gio/n)"I jole Ml + P2erw/ay + (iw/a,)} which is obiained by dividing out the poles and zeros in H(w). The representation of H(w) as in Eq, (14.15) is called the standard form. H(o) may include up to seven types of different factors that can appear in various combinations in a transfer funetion. These are: 1A gain K 2. A pole (jo)~' or zero (joo) at the origin 3. A simple pole 1/(1. + jw/py) or zero (L + je A.A quadratic pole 1/1 + (220/04 + (ie/w,)'| or zero U1 + 2éyo/an + Go/oy)"] (14.15) In constructing a Bode plot, we plot each factor separately and then add them graphically. The factors can be considered one at a time and then combined additively because of the logarithms involved. It is this mathematical convenience of the logarithm that makes Bode plots. a powerful engineering tool. We will now make straight-line plots of the factors listed above. We shall find that these straight-line plots known as Bode plots approx- imate the actual plots to a reasonable degree of accuracy. Constant term: For the gain K, the magnitude is 20 logy K and the phase is 0°; both are constant with frequency. Thus, the magnitude and phase plots of the gain are shown in Fig. 14.9. If K is negative, the ‘magnitude remains 20 logyo |&| but the phase is +180". Pole/zero at the origin: For the zero (jw) at the origin, the magni- tude is 20 logio « and the phase is 90°. These are plotted in Fig. 14.10, ‘where we notice that the slope of the magnitude plot is 20 dB/decade, while the phase is constant with frequency. ‘The Bode plots for the pole (jw) | are similar except that the slope of the magnitude plot is ~20 dB/decade while the phase is 90°, In ‘general, for (jw)®, where Nis an integer, the magnitude plot will have a slope of 20N dBidecade, while the phase is 90 degrees. " 20 089k -_——— 6 ———— a a « © Figure 14.9 Bode plots for gain K: (a) magnitude plot, (b) phase plo14a Bode Mots Simple poleizero: For the simple zero (1 + j/z), the magnitude is 20 logyo [1 + jeo/2| and the phase is tan”! w/2,. We notice that Hyg = 20 og | > lest = 416) as @>0 Hay = 20 ogo |1 + 4 201 lan = 2010810 = lozi0 qa as o> @ showing that we ean approximate the magnitude as zero (a straight line with zero slope) for small values of w and by a straight line with slope 20 dBidecade for lage values of w. The frequency @ = 2, where the 0 asymplotic lines meet is called the corner frequency or break frequency. Thus the approximate magnitude plot is shown in Fig. 14.11(a), where the actual plot is also shown. Notice thatthe approximate plot is close to the actual plot except at the break frequency, where « = =) and the deviation is 20 logio|(1 + j1)| = 20 logya V2 ~ 3B, ‘The phase tan” '(w/z,) can be expressed as $= aw'(2) s {es ae _ (14.18) 90°, w > ‘Asa straight-line approximation, we let ~ 0 for @ < 2,/10,@ = 45° for © = 21, and = 90° for @ = 10:). As shown in Fig. 14.11(b) along with the actual plot, the straight-line plot has a slope of 45° per decade. The Bode plots for the pole 1/(1 + jo/pi) are similar to those in Fig. 14.11 except that the comer frequency is at = p,, the magnitude has a slope of ~20 dB/decade, and the phase has a slope of ~45° per decade. Quadratic polezero:, The magnitude of the quadkatie pole 1 (1 + j2E 20/0 + (iw/arg)"| is —20 logio|I + j2E2@/0, + Gw/a)| and the phase is —tan~'(2¢,@/w,)/(1 — w/w). But oa ‘The special case of de (w = 0) does ‘not appear on Bode plots because log (0 = ~, implying that zero frequency is infiritely fr tothe left ofthe origin of Bode plots, HW 20 beveeesseesnon ° 01 t 0 -20 Slope = 20 dBidecade @ 6 ra oui. it, 01 10 ow o Figure 14.10 Bode plot for a zero (jw) at the origin: (a) magnitude plot, (6) phase plot, t= -Dhepelt «2B «(MF 0 = bil fe ‘ é di o oon o Figure 14.11 Bode plots of zero (1 + jey/z): (a) magnitude plot, (b) phase plot o622 Chapter 4 Frequency Response. and 1+ be pie (2) = rms (14.20) Hea, = ~20 logo ‘Thus, the amplitude plot consists of two straight asymptotic lines: one ‘with zero slope for « <«, and the other with slope ~40 dB/decade for « > wy, with a, as the comer frequency. Figure 14.12(a) shows the approximate and actual amplitude plots. Note thatthe actual plot ‘depends on the damping factor Z> as well as the comer frequency «©, ‘The significant peaking in the neighborhood of the comer frequency should be added to the straight-line approximation if a high level of accuracy is desired, However, we will use the straight-line approxima- tion for the sake of simplicity. 4 20 20 4 Dbte, Olea 10», 1000, 0 co) Figure 14.12 Bode plots of quadratic pole (1 + j2¢w/w, ~ w/e] |: (a) magnitude plot, (b) phase plot. The phase can be expressed as Thee is nether erocedtre for obiain- 0 w= ing Bode plots that is faster and perhaps: 1 _2hr0/ oy more effelen thnthe one we hoe = man AEE) ~90", =a, (1420 discussed, It consists in realizing that 7 180°, a> % zeros cause an increase in slope, while Palescaunee dociease Bystcingwan THE phase plot is a straight line tha slope of 90° per decade starting at @4/10 and ending at 10u,, as shown in Fig. 14.12(0). We see again the low-reauen. ofthe Pete tne centcicmeney thatthe diffrence betwssn the actual plot andthe straight-line plot is due ‘is, enditereasng cr decreasrg the tothe damping factor. Notice tha the straight-line approximations for both slope ateach comer frequency, cne can magnitude and phase plots for the quadatic pole are the same as those sketch the Bode plot immedistely trom for a double pole, ie. (1 + jeo/a,)~*. We should expect this because the venserfuncton witout theeffot’ the double pole (1 + jr/a,) ? equals the quadratic pole 1/{ + ‘ofrmakrg indica plotsandadcrg—_2f,u/w, + (jv/w,)"] when £2 = 1. Thus, the quadratic pole can be them. This procedure can be wed once ‘tyeaind asa double pole as far as straight-line approximation is concemed. youbecome proicentin Beone For the quadratic zero [1 + /2¢0/4, + (ja/a)"). the plots in siete lee Fig. 14.12 are inverted because the magnitude plot has a slope of eee eutershaveendees Fi 14.12 ae inverted because the magnitude plot has a slope th weed 40 dBidecade while the phase plot has a slope of 90° per decade 1 procedure discussed here aimost arte. Sereel abot etaaes, Table 14.3 presents a summary of Bode plots forthe seven factors SichosPSpice MATLAB Nahend, OF Course, Not every transfer Function has all seven factors. To sketch and Micro-Cop can be used te gener- the Bode plots fora function H(w) inthe form of Eq, (14.15) for exam- ae frequency resporse pts. We wil ple, we frst record the comer frequencies on the semilog graph paper, dscuss PSpicelaterin te chapter. _ sketch the factors one at a time as discussed above, and then combineHa Bode Ms on ECUETE Summary of Bode straight-line magnitude and phase plots. Factor Magnitude Phase 20 eek K ee 20NdBAlecade (jw) <— 1 o ° 1 T ° 3° Gay” 20N aBAdecade ae ON" 7 20N dB Atecede +2 3 o us an , 1 oe : . (1 + jeo/py’ Sette = a sires / itt st 1 1004 — " 17 Bolla Gola eb e eons Chapter 4 Frequency Response. additively the graphs of the factors. The combined graph is often drawn from left to right, changing slopes appropriately each time a corner fre- quency is encountered. The following examples illustrate this procedure. Semple 143 ‘Construct the Bode plots for the transfer function 200 Hw) = ie Go + 2a + 10) Solution: We fist put H(w) in the standard form by dividing out the poles and zeros. This, LO je Ho) = - (1+ jeo/2X1 + jo/10) Hence, the magnitude and phase are Hep, = 20 logo 10 + 20 logyo| fo] ~ 20 logo] 4 je ~ 20 og,0|1 + 20 log. | al = 90" — tan"! = tan We notice that there are two comer frequencies at « = 2, 10. For both the magnitude and phase plots, we sketch each term as shown by the dotted lines in Fig. 14.13, We add them up graphically to obtain the ‘overall plots shown by the solid curves. (a8) 20 104,010 20 m2 Figure 14.13 For Example 14.3: (a) magnitude plot, (b) phase plotM4 Bode lots 625 Daw the Bode plots forthe transfer Function ractice Problem 44m _ So +2) HO) = Ge ¥ 10) Answer: Sce Fig. 14.14. Xo ogiot 0 Figure 14.14 For Practice Prob, 14.3: (a) magnitude plot, (b) phase plo. Obtain the Bode plots for ~ Example Ja +10 Ho) = Os? Joljo + 3) Solution: Putting H(w) in the standard form, we get 0.401 + j/10) Hey = 21+ Jolt) Jo + jo/3y From this, we obtain the magnitude and phase as Ha = 20 logyo0.4 + 20 logio, jw 1 +22 — 20 opi £2) — 20 tox \il = 40 logit +4 © — 90° = 2 tan“! b= 0 + an wale ‘There are two comer frequencies at o = 5, 10 rads. For the pole with cor ner frequency at w = 5, the slope ofthe magnitude plot is ~40 dB decade and that ofthe phase plot is ~90° per decade due to the power of 2. The626 rapier 4 Frequency Response ‘magnitude and the phase plots forthe individual terms (in dotted lines) and the entire H(j) (in solid lines) are in Fig. 14.15. cavny|t+ 8 , sar da —4sriecae © 40 dBidesade cy “4svideende Figure 14.15 Bode plots for Example 144: (a) magnitude plot, (b) phase plo. A Sketch the Bode plots for S0jo HO) ~ Toe hie + 10F Answer: See Fig. 14.16 =f = zw ogo : 1 40" 100 wr a = ° o o ae ect 3 aS ar a TT ja01 1 180° 208 Sout @ Sera Figure 14.16 For Practice Prob. 144: (a) magnitude pl () phase plot Draw the Bode plots for stl Hy + 12s + 100 Solution: 1. Define, The problem is clearly stated and we follow the technique outlined in the chapter. 2. Present. We are to develop the approximate bode plot for the given function, H(s). 3. Alternative, The two most effective choices would be the approximation technique outlined in the chapter, which we will14a Bode Mots oa use here, and MATLAB, which can actually give us the exact Bode plots. 4, Attempt, We express His) as 1/1001 + joo) T+ joo 2/10 + (iw 10 For the quadratic pole, w,, = 10 rad/s, which serves as the comer frequency. The magnitude and phase are Hay = ~20 logyo 100 + 20 logio | + Jw) Hw) = _ jol2— ot robea)i rte a bro senor 2200] 1 = @°/100. Figure 14.17 shows the Bode plots. Notice that the quadratie pole is treated as a repeated pole at wy, that is, (+ jeo/a), ‘hich is an approximation, (8b) - 4 20 ogg lt Hol 20 “ > ne eros Nant ® o a Y 10 in one -180" » Figure 14.17 ‘Bode plots for Example 14.5: (a) magnitude pot, (b) pase plot. 5, Evaluate. Although we could use MATLAB to validate the solution, we will use a more straightforward approach. First, we ‘must realize that the denominator assumes that ¢ = 0 for the approximation, so we will use the following equation to check st n= 22 i) + 10° We also note that we need to actually solve for Hay and the corresponding phase angle . First, let « = 0. Hay = 201ogyo(1/100) = ~40 and Now ty @ = Hay, = 20 logy(1.4142/99) = ~36.9 4B which is the expected 3 dB up from the comer frequency. gel e458 fiom HY = =pa8 rapier 4 Frequency Response Now try w = 100. Hay, = 20 loz1o(100) ~ 20 logio(9900) = 39.91 dB. 4 is 90° from the numerator minus 180°, which gives —90°, We now have checked three different points and got close agreement, and, since this is an approximation, we ean feel confident that ‘we have worked the problem successflly ‘You can reasonably ask why did we not check at « = 10? If-we just use the approximate value we used above, we end up with an infinite value, which is to be expected from £ = 0 (sce Fig, 14.12a). If we used the actual value of H(j10) we will still end up being far from the approximate values, since ¢ = 0.6 and Fig. 14.12a shows a significant deviation ftom the approximation. We could have reworked the problem with £ = 0.707, which ‘would have gotten us closer to the approximation, However, we really have enough points without doing this. 6. Satisfactory? We are satisfied the problem has been worked successfully and we can present the results as @ solution to the problem. 5 Construct the Bode plots for 10 + 80s + 400) Hs) == Answer: See Fig. 14.18. (a8) ry Figure 14.18 For Practice Prob. 145: (a) magnitude plo, (b) phase plot [iemple 14.60 Given the Bode plot in Fig. 14.19, obtain the transfer function H(), Solution: To obtain H(w) from the Bode plot, we keep in mind that a zero always ccauses an upward turn at a comer frequency, while a pole causes a145 Seves Resonance downward turn, We notice from Fig, 14.19 that there is a zero joo at the origin which should have intersected the frequency axis at «o = 1. This is indicated by the straight line with slope +20 dB/decade. The fact that this straight line is shifted by 40 dB indicates that there is a 40-dB gain; that is, logig K = 2 40= 200g. K => K= 10? = 100 In addition to the zero jes at the origin, we notice that there are three factors with comer frequencies at « = 1, 5, and 20 rad/s. Thus, we have: 1A pole at p = 1 with slope ~20 dBidecade to cause a down ‘ward tum and counteract the zero atthe origin. The pole at p= Lis determined as 1/(1 + jo/1) 2. Another pole at p = 5 with slope ~20 dB/decade causing a downward turn, The pole is 1/(1 + jo/5). 3. A thitd pole at p = 20 with slope ~20 dBdecade eausing a further downward turn. The pote is 1/(1 + jo/20) Putting all these together gives the corresponding transfer function as _ 100 MO) TF jal NC + jw/S\O + je/20) _ joi" © Go + Ge + 5)\je + 20) He) 10's : is + 1)fs + 5) + 20) Oban the transfer function Hie) corresponding t© the Bode plot in Fig. 14.20. 4,000(s + 5) Answer: Hie) = — 0 e. (166 + 100) To see how to use MATLAB to produce Bode plot, refer to Section 14.11 45 0 Series Resonance ‘The most prominent feature of the frequency response of a circuit may be the sharp peak (or resonant peak) exhibited in its amplitude char- acteristic. The concept of resonance applies in several areas of science and engineering. Resonance occurs in any system that has a complex conjugate pair of poles; itis the cause of oscillations of stored energy from one form to another. It is the phenomenon that allows frequency ” aoa 20 dBidccade x | 220 dBidecade ° mn 1s 1020 Whe, Figure 14.19 For Example 146, Prac Hw +20 dBidecade oun rs 0 1 Figure 14.20 For Practice Prob. 14.6,Figure 14.24 ‘The series resonant circuit, Note No.4 becomes evilnt from the fact that Iv:| = “out = OF, Vil = Fel = OF Ve Ov where Q's the quality factoy, defined in Eq. (14.38. Chapter 4 Frequency Response. discrimination in communications networks. Resonance oceurs in any circuit that has at least one inductor and one capacitor. Resonance is a condition in en RIC circuit in which the capacitive andl inductive reactances are equal in magnitude, thereby resulting in 2 purely resistive impedance. Resonant circuits (series or parallel) are useful for constructing filters, as their transfer functions can be highly frequeney selective. They are used in many applications such as selecting the desired stations in radio and TV receivers. Consider the series RLC circuit shown in Fig, 14.21 in the fre- quency domain, The input impedanc (14.22) 1 Z=R+ ior - +) (04.23) Resonance results when the imaginary part of the transfer function is Img) = ot - +. =0 14.24 im(Z) = wl — = 04.24) ‘The value of that satisfies this condition is ealled the resonant fre- quency ew. Thus, the resonance condition is (14.25) 1 rads (14.26) Ve " Since w = 27 fo (1427) Note that at resonance: 1. The impedance is purely resistive, thus, Z = R. In other words, the LC series combination acts like a short circuit, and the entire voltage is across R. 2. The voltage V, and the current I are in phase, so that the power factor is unity. ‘The magnitude of the transfer function H(w) = Z(w) is minimum, 4. The inductor voltage and capacitor voltage can be much more than the source voltage. ‘The frequency response of the circuit v ‘current magnitude i) (14.28)145 Seves Resonance is shown in Fig. 14.22; the plot only shows the symmetry illustrated in this graph when the frequency axis is a logarithm, The average power dissipated by the REC circuit is Po} Lin 3PR (14.29) ‘The highest power dissipated occurs at resonance, when = Vy/R, 80 that sine Peon) = 375 1430) At certain frequencies ‘maximum value; that is, 1,2, the dissipated power is half the WafVB? _ Vi 2R AR Pw) = Pls) = (14.31) Hence, «oy and cy are called the half power frequencies. ‘The halfpower frequencies are obtained by setting Z equal to V2R, and writing lr (o ts 2) vik (14.32) (14.33) ‘We ean relate the half-power frequencies with the resonant frequency. From Eqs, (14.26) and (14.33), ty = Vow; (0434) showing that the resonant frequency is the geometric mean of the half= power frequencies. Notice that «; and w> are in general not symmet- rical around the resonant fiequency «, because the frequency response is not generally symmetrical. However, as will be explained shortly, symmetry of the half-power frequencies around the resonant frequency is offen a reasonable approximation. Although the height of the curve in Fig. 14.22 is determined by R, the width of the curve depends on other factors. The width of the response curve depends on the bandwidth B, which is defined as the difference between the two halF-power frequencies, Bro.- 0, (1438) ‘This definition of bandwidth is just one of several that are commonly used. Strictly speaking, B in Eq. (14.35) is a half-power bandwidth, because it is the width of the frequency band between the half:power frequencies. The “sharpness” of the resonance in a resonant circuit is measured quantitatively by the quality factor Q. At resonance, the reactive energy on 1 Fai 0707 (8 Fo--~ o “mye Bandwidh Figure 14.99 The current amplitude versus frequency forthe series resonant citcuit of Fig, 14.21Athough the same symbol Qis used for the reactive power, the two are not ‘equal and should not be confused. @ here is dimensionless, whereas reactive power Qis in VAR. This may help dis- tinguish between the two. Amplinde ess seociviy) Os(medum selectivity) (0 greatest selesivty) Figure 14.23 ‘The higher the circuit Q, the smaller the Dandi, ‘The quality factor is 9 measure of the selectivity (or sharpness of resonance) ofthe circuit Chapter 4 Frequency Response. in the circuit oscillates between the inductor and the capacitor. The quality factor relates the maximum or peak energy stored tothe energy dissipated in the circuit per cycle of oscillation: Peak energy stored in the circuit egy diasteated by te Geet im one peiod at resonance (14.36) Its also regarded as a measure of the energy storage property of a cit- cuit in relation to its energy dissipation property. In the series REC circuit, the peak energy stored is £/°, while the energy dissipated in cone period is (PR\(I/fo). Hence, be Q= (14.37) (14.38) Notice that the quality factor is dimensionless. The relationship between the bandwidth B and the quality factor Q is obtained by sub- stituting Eq, (14.33) into Eq, (14:38) and utilizing Eq, (14:38). (14.39) or B = aGCR. Thus ‘The quality factor of a resonant circuit is the ratio of its resonant frequency to its bandwidth, Keep in mind that Egs. (14.33), (14.38), and (14.39) only apply to series REC circuit. AAs illustrated in Fig. 14.23, the higher the value of Q, the more selective the cireuit is but the smaller the bandwidth. The selectivity of an RLC circuit is the ability of the circuit to respond to a certain fre- quency and discriminate against all other frequencies. IF the band of frequencies to be selected or rejected is narrow, the quality factor of the resonant circuit must be high. If the band of frequencies is wide, the quality factor must be low. ‘A resonant circuit is designed to operate at or near its resonant fre- quency. It is ssid to be a high-Q circuit when its quality factor is equal to or greater than 10. For high-Q circuits (Q = 10), the half-power frequencies are, for all practical purposes, symmetrical around the res- ‘onant frequency and can be approximated as B B yy 5 oy + (14.40) High-Q circuits are used often in communications networks.145 Seves Resonance 33 We see that a resonant circuit is characterized by five related parameters: the two halfspower frequencies «, and ¢, the resonant fre quency a, the bandwidth 8, and the quality factor O. cuit of Fig, 1424, R= 20, L=1mH, and C= 0.4 nF. (@) Find the resonant frequency and the half:power frequencies. b) Cal- culate the quality factor and bandwidth, (¢) Determine the amplitude - a of the current ato, and o. Solution: 20 sin ot rc (a) The resonant frequency is J 1 1 Figure 14.24 « = 50 krad’s For Example 14.7 8 Vie V0 x 04 x 10% ME METHOD 1 The lower half-power frequency is R (z y 1 Rae of RE) ye a * V\ai) * ze 2 2x10 <1 + VIF 2500 krad’s Similarly, the upper hal-power frequency is + VI 2500 krad/s = 51 krad/s 3 + Va0)? + 60x 10% 19 krad/s, (b) The bandwidth is B= oy ~ = 2krad/s The quality factor is HE METHOD 2 Altematively, we could find gl _ $0 10° x 10> | “R 2 “8 From Q, we find om _ 50 10° = = A nas Ba A nk Since Q > 10, this is a high-Q circuit and we can obtain the half- power frequencies asos rapier 4 Frequency Response as obtained earlier, () At w= oo, [4.7 A series-connected circuit has R= 40) and Z = 25 mH. (a) Caleulate the value of Cthat wil produce a quality factor of 50, (b) Find cy, and B. (c) Determine the average power dissipated at w = wp, a), Take Va, = 100 V. “Answer: (a) 0.625 F,(b) 7920 rad/s, 8080 rads, 160 rad/s, (€) 1.25 kW, (0.625 kW, 0.625 KW. 14.6 I Parallel Resonance 1-,.20Q) VER Aju - JC The parallel RLC circuit in Fig. 14.25 is the dual of the series RLC cir- cuit, So we will avoid needless repetition. The admittance is Figure 14.95 i ‘The parallel resonant ciruit Y= Ho) + oC + or asa Y= + 4.42) ~ ob, a Resonance occurs when the imaginary part of ¥ is zero, 1 @C = — 4 CeO (14.43) 0 wm op Bandwidin : Figure 14.26 = ins ‘The curent amplitude versus fequeney for oye asso the series resonant circuit of Fig. 428. which is the same as Eq, (14.26) for the series resonant circuit. The ‘We can see this from the fact thet voltage |V| is sketched in Fig. 14.26 as a function of frequency. i Notice that at resonance, the parallel LC combination acts like an tn = 2 = 91, ‘open circuit, so that the entire current lows through R. Also, the we inductor and capacitor current can be much more than the source Hcl = epCla = Ow current at resonance. where Q's the qual factor, defined ‘We exploit the duality between Figs. 14.21 and 14.25 by compar- ing (i447, ing Bq, (1442) with Eq, (1423). By replacing R, Z, and C in the146 Palle Resonance expressions for the series circuit with 1/R, C, and L respectively, we obtain for the parallel circuit (14.45) (14.46) (14.47) It should be noted that Egs. (14.45) to (14.47) apply only to a parallel RLC circuit. Using Eqs. (14.45) and (14.47), we can express the half power frequencies in terms of the quality factor. The result is yt Ga) +2 (14.48) (14.49) ‘Table 14.4 presents a summary of the characteristics ofthe series and par allel resonant circuits. Besides the series and parallel REC considered here, other resonant circuits exist. Example 14.9 teats a typical example. ‘Summary of the characteristics of resonant RLC circuits. Characteristic Series circuit Parallel circuit Resonant equenc du H eens oe vie Vie ook yt R Quality fiewr, 0 Mor ne Korey oy % Bandi, 8 3 “ 5 payee oles tntpotepecsnes onfis(—hp et afin hy et 8 a For Q = 10,0, 0» a +3 mss as626 rapier 4 Frequency Response In the parallel RLC circuit of Fi and C= 8 pF. (a) Calculate w, O, and B. (b) (©) Determine the power dissipated at wy, a, and ws, ind @ and @>. mp hm @ Figure 14.97 sym ee Oe sans For Exanple 148 VIC Voaxixexi® 4 R x10 Rg BN __ agg Oe gk 25X10 X02 X10 & p= % = 15625 dls Q (©) Due to the high value of Q, we can regard this as a high-Q circuit, Hence, B 00) = tu) ~ 2 = 28,000 ~ 7812 = 24992 nals 2 os = oy +2 = 25,000 + 7.812 = 25,008 dls (©) At = a, ¥ = 1/R or Z = R= 8kO. Then y_ 10/-90" Z~ 8,000 = 1.25/-90° ma Since the entire current flows through R at resonance, the average power dissipated at @ = a is 1 P dnee yO.25 x 10P(8 x 10°) = 6.25 mW v2 _ 100 PR axsxip Om Atw = aon, A Fe = 3:125mw parallel resonant circuit has R ‘alculate «, a, ws, Q, and B. 20 mH, and C Answer: 100 krad/s, 99 krad/s, 101 krad/s, 50, 2 krad’s,147 Pasive riers or Determine the resonant frequency of the eieuit in Fig, 14.28. ‘Solution: ‘The input admittance is 2H Ipcos wt or S108 y= joo +4 201 + joo +2 ‘| 20 / 10) 2+ jo 4+ 40" At resonance, Im(¥) = 0 and Figure 14.28 5 For Example 149, ago. -— 4+ 4 Cotcalate the resonant frequency of the Grcult in Fig, 1429. Pi Answer: 100 rad/s. 100 mit ryenor@) osaF 200 Figure 14.29 147 Passive Filters ‘The concept of filters has been an integral part of the evolution of elec trical engineering from the beginning. Several technological act ‘ments would not have been possible without electrical filters. Because of this prominent role of filters, much effort has been expended on the theory, design, and construction of filters and many articles and books have been written on them, Our discussion in this chapter should be considered introductory. A filter is a circuit that is designed to pass signals with desired frequencies and reject or attenuate others. AS a frequency-selective device, a filter can be used to limit the fre quency spectrum of a signal to some specified band of frequencies, Filters are the circuits used in radio and TV receivers to allow us to select one desired signal out of a multitude of broadcast signals in the environment. ‘A filter is a passive filter if it consists of only passive elements R, L, and C. 11 is said to be an active fitr if it consists of active elements (such as transistors and op amps) in addition to passive elements R, L, and C. We consider passive filters in this section and active filters in the next section, ZC filters have been used in practical applications for more than eight decades, LC filter technology feeds related areas such as equalizers, impedance-matching networks, transformers, shaping networks, power dividers, attenuators, and direetional couplers, and is continuously providing practicing engineers with oppurtunities to inno- vate and experiment. Besides the LC filters we study in these sections, there are other kinds of filters—such as digital filters, electromechanical filters, and microwave filters—which are beyond the level of this text. For Practice Prob. 14.9|i Pasha ' . 4% % f@ Lie Stop -_ (10 (ae Figure 14.30 cal frequency response of four types of filter (a) lowpass filter, (b) highpass filter, () bandpass filter, (4) bandstop iter a #@ c + 0 Figure 14.31 lowpass fier Or th | teat amr Actua o™ Figure 14.32 Ideal and actual frequency response of a lowpass filter Chapter 4 Frequency Response. As shown in Fig, 14.30, there are four types of filters whether pas- sive or active: 1. A lonpass filter passes low frequencies and stops high frequencies, as shown ideally in Fig. 14:30(2, 2. A highpass fiter passes high frequencies and rejects low frequencies, as shown ideally in Fig. 14.300). 3. A bandpass filter passes frequencies within a frequeney band and blocks or attenuates frequencies outside the band, as shown ide- ally in Fig. 14.30(c). 4. A handstop filter passes frequencies outside a frequency band and blocks or attenuates frequencies within the band, as shown ideally in Fig. 14.30(¢), ‘Table 14.5 presents a summary of the characteristics of these filters. Be aware that the characteristics in Table 14.5 are only valid for first- or second-order filter-—but one should not have the impression that only thse kinds of filter exist. We now consider typical circuits for realiz~ ing the filters shown in Table 14.5, cere) Summary of the characteristics of ideal filters. ‘Type of Filter HO) He) Ho) or Hlo) Lowpass| 1 ° yi Higtpass 0 1 Wa Bandpass 0 0 1 Bandsiop 1 1 o ‘isthe cut fequency for lowpass and highpass ters ste center frequency for bondpass and bandstop fies. 14.7.1 Lowpass Filter ‘A typical lowpass filter is formed when the output of an RC circuit is taken off the capacitor as shown in Fig. 14.31. The transfer function (see also Example 14.1) is Mij@r Ho) Rviged 1 HW (14.50) Note that H(0) = 1, H(=) = 0. Figure 14.32 shows the plot of /1(w)|. along with the ideal characteristic. The half-power frequency, which is ‘equivalent to the comer frequency on the Bode plots but in the con- text of filters is usually known as the cutoff frequency wis obtained by setting the magnitude of H(«) equal to 1/V2, thus, Hw.) 4st)147 Passive Fiters The cutoff frequency is also called the rollof frequency: ‘A lowpass fiter is designed to pass only frequencies from dc up to the cutoff frequency «. ‘A lowpass filter can also be formed when the output of an RL cir- cuit is taken off the resistor. OF course, there are many other circuits for lowpass filters. 14.7.2. Highpass Filter A highpass filter is formed when the output of an RC circuit is taken off the resistor as shown in Fig. 14.33. The transfer function is a Me) = VR + IoC te) = RC 145 a ake (14.52) Note that H(0) = 0, H(2) = 1. Figure 14.34 shows the plot of [H(o)]. ‘Again, the comer or cutoff frequency is 0. = (14.53) ‘A highpass filter is designed to pass all frequencies above its cutoff frequency a. ‘A highpass filter can also be formed when the output of an RZ cir- cuit is taken off the inductor, 14.7.3, Bandpass Filter ‘The RLC series resonant circuit provides a bandpass filter when the output is taken off the resistor as shown in Fig. 14.35. The transfer function is Hw = Ye-___& __ (OV, Rel ~ Woe) We observe that H(0) = 0, H(c) = 0. Figure 1436 shows the plot of | H(«)|. The bandpass filter passes a band of frequencies (0 <@ < os) centered on ao, the center frequency, which is given by (14.54) (14.55) Vv ‘A bandpass fiter is designed to pass all frequencies within a bend of frequencies, a) <@ < wo Since the bandpass filter in Fig. 14.35 is a series resonant circuit, the half-power frequencies, the bandwidth, and the quality factor are deter- mined as in Section 14.5. A bandpass filter can also be formed by cascading the lowpass filter (where «) = o,) in Fig, 14.31 with the on The cutoff frequency isthe frequency ‘at which the transfer function H drops in magnitude to 70:71% of its maximum value. tis aso regarced as the fre quency at which the power dissipated in accu is haf ofits maximum value. — 0) RE 0 Figure 14.33 A highpass filter. i400) seal 1 0707 Acta! ° Figure 14.34 Ideal and actual frequency response of highpass filter a) 2340 Figure 14.35 Atbandpass filter lana 1 0707 0 ye 6 Figure 14.36 Ideal and actual frequency response of a bandpass filter.640 Chapter 4 Frequency Response. highpass ‘iter (where «; = o,) in Fig. 14.33. However, the result ‘would not be the same as just adding the output of the lowpass filter to the input of the highpass filter, because one circuit loads the other and alters the desired transfer function, 14.7.4 Bandstop Filter 0® AA filter that prevents a band of frequencies between two designated values (1 and «) fiom passing is variably known as a bandstop, band reject, or notch filter. A bandstop filter is formed when the output REC Figure 14.37 series resonant circuit is taken off the LC series combination as shown Abandstp filter. in Fig. 14.37. The transfer function is Vo __ Hol = Woe) HO) VR Kel— od) (14.56) Lien Notice that H(0) = 1, H() = 1, Figure 14.38 shows the plot of |F1(0)}, Again, the center frequency is given by 1 0707 | Acta (147) Heal hile the hal-power frequencies, the bandwidth, and the quality fac- tor are calculated using the formulas in Section 14.5 for a series reso- 0% 9% © nant circuit. Here, a is called the frequency of rejection, while the Figure 14.38 corresponding bandwidth (B =e = «) is known as the bandwidth Ideal and actual frequency response ofa of rejection. Thus, bhandstop Bite. A bandstop filter is designed to stop or eliminate all frequencies ‘within a band of frequencies, «, < w < a, Notice that adding the transfer fumetions of the bandpass and the bbandstop gives unity at any frequency for the same values of R, L, and C. Of course, this is not true in general but true for the circuits treated here. This is due to the fact that the characteristic of one is the inverse of the other, In concluding this section, we should note that 1. From Eqs. (14.50), (14,52), (14.54), and (14.56), the maximum gain of a passive filler is unity. To generate a gain greater than unity, fone should use an active filter as the next section shows. 2. There ate other ways to get the types of filters treated in this section 3. The filters treated here are the simple types. Many other filters have sharper and complex frequency responses, ner or cutoff frequency. Take R Solution: ‘The transfer funetion is Yo Ri l/s V, ab + RAC HG) sje (14.1041)147 Passive Fiters on But abn Rise R + sC RF 1/sC 1+ 5RC o0® Ro C= eft) yo) = BAL ste R i= Figure 14.39 SL+ R/V + SRC) PRLC + SL +R For Example 14.10, R WRLC + jo +R Since H(0) = 1 and H(ee) = 0, we conclude from Table 14,5 that the circuit in Fig. 14.39 is a second-order lowpass filter. The magnitude of His Ho) = (14.10.2) (14.103) ‘The comer frequency is the same as the half-power frequency, i.e., where His reduced by a factor of 1/\V2. Since the de value of HMw) is 1, at the comer frequency, Eq. (14.10.3) becomes after squaring Substituting the values of &, L, and C, we obtain 2= (1 oF 4x 10°F + (@,107P Assuming that isin krad/s, 2= (= 4027 + @2 or Nai — Tw? — 0 Solving the quadratic equation in 02, we get? = 0.5509 and —0.1134. Since «is rel, Identify the type of filter the circuit represents and determine the cor- net frequency. Take R, = 1000 = Ra, L = 2 mil, i answer 5 als . =} highpass fier © a A = 25 kas Figure 14.40 + ROL i igure 14. Ki RD For Practice Prob, 14.10a2 Chapter 4 Frequency Response. Ifthe bandstop filter in Fi passing other frequencies, calculate the values of L and R = 180.0 and the bandwidth as 100 Hz. Solution: We use the formulas for a series resonant circuit in Section 14.5. B = 2(100) = 200 rad/s But R R_ 150 i B 2007 = 0.2387 H Rejection of the 200-Hz sinusoid means that / is 200 Hz, so that ay in Fig. 14.38 is = 2mfy = 2 (200) = 4007 ; 2.653 uF aL (400%) (0.2387) a TERREHES PODS 14.17 Design a bandpass filter of the form in Fig. 14.35 with « lower eutoft frequency of 20.1 kHz and an upper cutoff frequency of 20.3 kH¥z. Take R= 20KO. Calculate L, C, and O. Answer: 15.92 H, 3.9 pF, 101. 480 Active Filters ‘There are three major limitations to the passive filters considered in the previous section. First, they cannot generate gain greater than 1; pas- sive elements cannot add energy to the network. Second, they may require bulky and expensive inductors. Third, they perform poorly at Jrequencies below the audio frequency range (300 Hz <_< 3,000 Hz). Nevertheless, passive filters are useful at high frequencies ‘Active filters consist of combinations of resistors, capacitors, and ‘op amps. They offer some advantages over passive RLC filters. First, they are often smaller and less expensive, because they do not require inductors. This makes feasible the integrated circuit realizations of fil- ters. Second, they can provide amplifier gain in addition to providing the same frequency response as RLC filters. Third, active filters ean be combined with buffer amplifiers (voltage followers) to isolate each stage of the filter from source and load impedance effects. This isola tion allows designing the stages independently and then cascading them to realize the desired transfer function. (Bode plots, being logarithmic, may be added when transfer functions are cascaded.) However, active filters are less reliable and less stable. The practical limit of most active48 ActiveFiters filters is about 100 kHz—most active filters operate well below that frequency. Filters are often classified according to their order (or number of poles) or their specific design type. 14.8.1 First-Order Lowpass Filter One type of first-order filter is shown in Fig. 14.41, The components selected for Z, and Z; determine whether the filter is lowpass or high- pass, but one of the components must be reactive. Figure 14.42 shows a typical active lowpass filter. For this filter, the transfer function is Vo “Ly Hw = Ye =~ (48) where = Rand y= 8| oo 1 ¥s| Fac,” B+ thot Tr puGk, 4S) Therefore, He) = -Z (14.60) " Ry L + fol, Ry We notice that Eq. (14.60) is similar to Eg, (14.50), except that there is a low frequency (« > 0) gain or de gain of —Ry/R, Also, the cor- ner frequency is a BG, Which does not depend on R,. This means that several inputs with dif ferent R, could be summed if required, and the comer frequency would remain the same for each input (4.61) 14.8.2 First-Order Highpass Filter Figure 14.43 shows a typical highpass filter. As before, Yes % We = = -7 (14.62) where Z; = Ry + L/jC; and Zp= Ry so that Hw) =, = -_ oe (14.63) Ri + MjoC, 1+ joc This is similar to Eq. (14.52), except that at very high frequencies (+ ®), the gain tends to —R//R,. The comer frequency is 1 RG (14.64) 14.8.3 Bandpass Filter The circuit in Fig. 14.42 may be combined with that in Fig. 14.43 t0 form a bandpass filter that will have a gain K over the required range of frequencies. By cascading a unity-gain lowpass filter, a unity-gain Figure 14.41 ‘A general ftstorder active filter Figure 14.42 Active first-onder lowpass filter, Figure 14.43, Active first-order highs ile. ‘This way of creating @ bandpass fier, rot necessarily the best, is perhaps the easiest tounderstand.rapier 4 Frequency Response highpass filter, and an inverter with gain —R,/,, as shown inthe block diagram of Fig. 14.44(a), we can construct a bandpass filter whose fre- {quency response is that in Fig, 14.44(b). The aetual construction of the bandpass filter is shown in Fig. 14.45. ” Kk 07k [Paz] Beeline Le Fier fier : ow» «& « o © Figure 14.44 ‘Active bandpass iter: (a) block diagram, (b) frequency response. stage 1 Stage? i Low-pass filter igh pas filter An inverter sets oy value sets value provides gain Figure 14.45 Active bandpass filter ‘The analysis of the bandpass filter is relatively simple. Its transfer function is obtained by multiplying Eqs. (14.60) and (14.63) with the gain of the inverter; that is, y, (- 1 \(- JoCsR \( 4) 7 1+ joCR, 1+ joCgh, Ry 3 we (14.65) 1 jwGR Rl +joCR1 + joGR ‘The lowpass section sets the upper comer frequency as 1 ae (14.6) while the highpass section sets the lower comer frequency as ane (14.67) RC,48 ActiveFiters With these values of @ and «>, the center frequency, bandwidth, and quality factor are found as follows: 0) = Voor (14.68) B=o,~ 0, (14.69) o=F (14.70) To find the passband gain K, we write Eq, (14.68) in the standard form of Eq. (14.15), & jefe _ Bj, Ril + jeofers 1 + joo/ o>) Ry (1 + jos + Jo) aa7y AL the center frequency «@» = Veo\@2, the magnitude of the transfer function is Hw) = med! barista Rater ‘Thus, the passband gain is Ry wy naa (14.73) 14.8.4 Bandreject (or Notch) Filter A bandreject filter may be constructed by parallel combination of @ lowpass filter and a highpass filter and a summing amplifier, as shown in the block diagram of Fig. 14.46(a). The circuit is designed such that the lower cutoff frequency «, is set by the lowpass filter while the upper cutoff fiequency «> is set by the highpass filter, The gap between 4, and « is the bandwidth of the filter. As shown in Fig. 14.46(b), the filter passes frequencies below «; and above ws. The block diagram in Fig. 14.46(a) is actually constructed as shown in Fig. 14.47. The trans- fer function is o Ho) e-te ) (14.74), 1 i om |. a ae | Figure 14. Active bandejeet filter: (a) block diagram, (b) frequency response, a oChapter 4 Frequency Response. Figure 14.47 ‘Active bandreject filter ‘The formulas for calculating the values of «1, w>, the center frequency, ‘bandwidth, and quality factor are the same as in Eqs, (14.66) o (14.70). To determine the passband gain K of the filter, we can write Eq, (14.74) in terms of the upper and lower comer frequencies as Re jwfe, ) He a + jufag * 1+ jofen Re + Pole, + Yoy/ore) Re (+ jo/eos\ + jo/ens) ‘Comparing this with the standard form in Eq, (14.15) indicates that in the two passbands (« — 0 and w —> 2) the gain is (14.75) (14.76) We can also find the gain at the center frequency by finding the mag- nitude of the transfer funetion at w%) = Vooyon, writing Ry (+ w/e + Uies)*/ove) Ry (1 + jeoo/ws\{1 + joro/@) Ry 2a Ro +02 Ha) (4.77) ‘Again, the filters treated in this section are only typical. There are many other active filters that are more complex. Design a lowpass active of 500 Hz, and a comer frequency Solution: From Eg, (14.61), we find mf, = 2m(500) = (4.12.1) RG48 ActiveFiters on ‘The de gain is HO) (14.122) We have two equations and three unknowns. If we select C;= 0.2 F, then 159k and We use a 1.6-K(0 resistor for Ry and a 400-02 resistor for R,. Figure 14.42 shows the filter Design a highpass filter with a high-frequency gain of 5 and a comer Practice Problem 14200 frequency of 2 kHz. Use a 0.1-F capacitor in your design. Answer: &, = 8000 and Ry = 4 k0. Design a bandpass filter in the form of Fig. 14.45 to pass frequencies ss EX@MIIE AMIS between 250 Hz and 3,000 Hz and with K = 10. Select R = 200. Solution: 1. Define. The problem is clearly stated and the circuit to be used in the design is specified 2. Present. We are asked to use the op amp circuit specified in Fig, 14.45 to design a bandpass filter. We are given the value of R to use (20kQ). In addition, the frequency range of the signals, to be passed is 250 Hz to 3 kHz. 3. Alternative. We will use the equations developed in Section 14.83 to obtain a solution. We will then use the resulting transfer funetion to validate the answer. 4, Attempt, Since @, = 1/RC, we obtain L 1 Ray 2afR~ 2m X 250 20 X 10° G = 31.83 nF Similarly, since w= 1/RC,, 1 1 1 Ron afk Be X3,000 X 20 X 10 From Eq. (14.73), GQ 10(3,250) 000 = 1083Chapter 4 Frequency Response. If we select R, = 10 KO, then Ry = 10.832, ~ 108.3 KO. 5. Evaluate. The output of the first op amp is given by $2.65 x 10-%(V, ~ 0) T ~~ 1¥33x 10% ‘The output of the second op amp is given by Wa0 ,m-0 1 20ka SSTA3 nF 6366 % 10" 1 + 6.366 x 10-4 6366 X10, (1 + 6,366 x 10-451 + 5.3 x 10-*s) 20K + Va=— ‘The output of the third op amp is given by W=0, M-0 | 10kQ — 108.3 kD woh 10.83¥, > jm x 25° _ 6.894 x 10-850 (1 6.366 LOS) + 5.3 x 1075) Let j2a x 25° and solve for the magnitude of V,/¥,. v= ¥, _ ~710:829 yo a+ s)a) V./¥ = (0.7071)10.829, which isthe lower comer frequency point. Let s = j2m X 3000 = j18.849 KO. We then get ¥,__~jl2994 % Ost +) 129.94/—90° ~ (aompsznaaiazasy ~ OOM LIBS Clearly this is the upper comer frequency and the answer checks. 6, Satisfactory? We have satisfactorily designed the circuit and can present the results as a solution to the problem. ‘a notch filter based on Fig. 14.47 for wy = 20 krads, K and Q Answers C; = 4.762 nF, C, = 5.263 nP, and Ry = 50 KO. 490 Scaling In designing and analyzing filters and resonant circuits or in circuit analysis in general, it is sometimes convenient to work with element values of 1.0, 1H, or I F, and then transform the values to realistic49 Sealing values by scaling. We have taken advantage of this idea by not using realistic element values in most of our examples and problems tering circuit analysis is made easy by using convenient component val- ues. We have thus eased calculations, knowing that we could use scaling to then make the values realistic. There are two ways of sealing a circuit: magnitude or impedance sealing, and frequency scaling. Both are useful in scaling responses and circuit elements to values within the practical ranges. While mag- nitude scaling leaves the frequency response of a circuit unaltered, frequency scaling shifts the frequency response up or down the fre- queney spectrum, 14.9.1. Magnitude Scaling ‘Magnitude scaling is the process of increasing all impedances in a net- ‘work by a factor, the frequency response remaining unchanged. Recall that impedances of individual elements R, L, and C are given by Ze=R L=hl, Ee=-z aa jot In magnitude scaling, we multiply the impedance of each circuit ele rent by a factor K,, and let the frequency remain constant. This gives the new impedances as Zh, = Kis, = jOKgL 1 (14.79) joC]Ky Comparing Eq, (14.79) with Eq, (14.78), we notice the following changes in the element values: R + Ky, L—> Kyl., and C—> C/K. Thus, in ‘magnitude scaling, the new values of the elements and frequency are Ze (14.80) ‘The primed variables are the new values and the unprimed variables are the old values. Consider the series or parallel RLC circuit. We now have ey (14.81) VEulC/Kn showing that the resonant frequency, as expected, has not changed, Similarly, the quality factor and the bandwidth are not affected by mag- nitude scaling. Also, magnitude scaling does not affect transfer func- tions in the forms of Eqs. (14.2a) and (14.2b), which are dimensionless quantities. VicFrequency scaling is equivalent to rela boeing the tequency ais of afre~ ‘quency response plot Itis needed ‘when trarsatng frequencies such as @ resonant frequency, a comer frequency, a bandwidth et, toa reals level. t ‘can be used to bring capacitance and inductance values into range tht is. Ccorwerient to work with. epee 4 Frequency esporse 14.9.2 Frequency Scaling Frequency scaling is the process of siting the frequency response of ‘network up or down the frequency axis wile leaving the impedance the same, We achieve frequency scaling by multiplying the frequency by a fae- tor Ky while keeping the impedance the same. From Eq. (14.78), we see that the impedances of L and C are frequeney-dependent. If we apply frequency sealing to Z,(io) and Z(w) Eq. (14.78), we obtain Z= KOK =jol > = 4.82a) 1 L ra ee C= (14.82b) joc Kk Le since the impedance of the inductor and capacitor must remain the same after frequency scaling. We notice the following changes in the clement values: > L/K, and C—> C/K, The value of R is not affected, since its impedance does not depend on frequency. Thus, in frequency sealing, the new values of the elements and frequency are (14.83) Again, if we consider the series or parallel RLC circuit, for the reso- nant frequency 1 \ K, “VEC Vi/KNCIK) Vic and for the bandwidth Kyoy (14.84) B= KB (14.85) but the quality factor remains the same (Q" = Q). 14.9.3 Magnitude and Frequency Scaling Ifa circuit is scaled in magnitude and frequency at the same time, then (14.86) ‘These are more general formulas than those in Egs. (14.80) and (14.83), We set Ky = 1 in Eq. (14.86) when there is no magnitude sealing or K,= | when there is no frequency scaling,49 Sealing ost ‘A four-order Butterworth lowpass filter is shown in Fig, 14.48(a). The filter is designed such thatthe cutofT frequency o, = 1 rad/s. Seale the circuit for a cutoff frequency of 30 KHz using 10-kQ resistors. rosa 076s Wk S882 mH T ah TH a ° ¢ Tt sFlawr S1Q yy =2435pk —=Se82pF S10KQ », @ © Figure 14.48 For Example 14.14: (a) Normalized Butterworth lowpass filter, (b) scaled version ofthe same lowpass filter. Solutic If the cutoff frequency is to shift from w, = I rad/s to w{, = 277(50) rad/s, then the frequency scale factor is L100 x10 = x 10° Ky Also, if each 1-0 resistor isto be replaced by a 10-kO resistor, then the magnitude scale factor must be = Rx 10 108 wR I Using Eq, (14.86), Kn 108 x 108 La = 100765) = 2438 mit X10 8.82 mH (1.848) = CG _ 0.765 KK 7X10 Ch 1848 © Roky X10? = 243.5 pF = 588.2 pF ‘The sealed circuit is shown in Fig, 14.48(b). This eircuit uses practical values and will provide the same transfer function as the prototype in Fig. 14.48(a), but shifted in frequency. A third-order Butterworth filter normalized 10 «, = Iradls is shown Practice Problem 4am in Fig. 14.49, Scale the circuit to a cutoff frequency of 10 kHz. Use 15-nF eapacitors. lo aH Answer: 2 = L061 KO, Ch = Cy = SMF L' = 33.77 mH. ip tir Zia Figure 14.49 For Practice Prob, 14.14Figure 14.50 For Example 14.15, Chapter 4 Frequency Response. 14.10) Frequency Response Using PSpice PSpice is a useful tool in the hands of the modem circuit designer for obtaining the frequeney response of circuits. The frequency response js oblained using the AC Sweep as discussed in Section D.5 (Appen- dix D). This requires that we specify in the AC Sweep dialog box Total Prs, Start Freq, End Freq, and the sweep type. Total Pis is the num- ber of points in the frequency sweep, and Start Freq and End Freq are, respectively, the starting and final frequencies, in hertz. In order to know what frequencies to select for Start Freq and End Freq, one must have an idea of the frequency range of interest by making a rough sketch of the frequency response. In a complex citcuit where this may not be possible, one may use a trial-and-error approach, There are three types of sweeps: Linear: The frequency is varied linearly from Start Freq to End Freq with Total Pts equally spaced points (or responses). Octave: The frequency is swept logarithmically by octaves from ‘Start Freq to End Freq with Total Pas per octave. An octave is a factor of 2 (e.g. 2 t0 4,4 10 8, 8 to 16). Decade: The frequency is varied logarithmically by decades from ‘Start Freq to End Freq with Toral Pts per decade, A decade is a factor of 10 (e.g. from 2 Hz to 20 Hz, 20 Hz to 200 Hz, 200 Hz to 2 kHz). It is best to use a linear sweep when displaying a narrow frequency range of interest, as a linear sweep displays the frequency range well in a narrow range. Conversely, it is best to use a logarithmie (octave for decade) sweep for displaying a wide frequency range of interest— if a linear sweep is used for a wide range, all the data will be crowded at the high- or low-frequency end and insufficient data at the other end. With the above specifications, PSpice performs a steady-state sinu- soidal analysis of the circuit as the frequency of all the independent sourees is varied (or swept) from Start Freg to End Freq. The PSpice A/D program produces a graphical output. The output data type may be specified in the Trace Command Bax by adding one of the following suffixes to V or I: M Amplitude of the sinusoid P_ Phase of the sinusoid. 4B Amplitude of the sinusoid in decibels, ie., 20 logio (amplitude). Determine the frequency response of the eire Solution: ‘We let the input voluge v, be « sinusoid of amplitude 1 V and phase 0° Figure 14.51 is tho schematic forthe cireuit. The capacitor i rotated 270° counterclockwise to ensure that pin 1 (the positive terminal) is on top. The voltage marker is inserted 10 the output voltage across the capacitor. To perform a linear sweep for | > bode (num, den); % determine and draw Bode plots ‘The step response (0) of a system is the output when the input x(0) is the unit step function. The command step plots the step response of a sys tem given the numerator and denominator of the transfer function of that system. The time range and number of points are automatically selected, For example, consider a second-order system with the transfer function 2 me) 38+ 12 We obtain the step response of the system shown in Fig, 14.60 by using the following commands. 2: 13121; >> step(n/d) ; ‘We can verify the plot in Fig. 14.60 by obtaining y() = x() * u(0) oF ¥69) = XVM, Sep reponse 2 Bode diagrams : gm» 30 {o ° uo 2-2 a0 02 26 4s oe ep eg Z o 0s 1 is 2 a5 333 4 wo “rine () ‘Frequency (rads) Figure 14.60 Figure 14.59 The Step response of Magnitude and phase plots. His) = 12/(2 + 3s + 12). ‘The command Isim is a more general command than step. It cal- cculates the time response of a system to any arbitrary input signal. The format of the command is y = Isim (num, den, x, 2), where x(2) is the input signal, is the time vector, and (2) is the output generated. For example, assume a system is described by the transfer function s+4 S42 + 55+ 10 HG) =14.12 Applications To find the response (0) of the system to input x(0) = 10e™'u(), we use the following MATLAB commands. Both the response y() and the input x(9 are ploted in Fig, 14.61 >>t=0:0.02: 0.02 >>x= 10*exp(-t) 5 >>num= [1 4]; >>den=[1 2 5 10]; >>y=1sim(num,den,x,t) ; >> plot (t,x,t,y¥) ; 4 time vector 0 @ > 0) Sealing is the process whereby unrealistic element values are magnitude-scaled by a factor K,, and/or frequency-scaled by a fac~ tor K; to produce realistic values. Kn % ay Kuk; R= KR, 13. PSpice can be used to obtain the frequency response of a circuit if a frequency range for the response and the desired number of points within the range are specified in the AC Sweep. 14, The radio receiver—one practical application of resonant circuits — employs a bandpass resonant circuit to tune in one frequency ‘among all the broadcast signals picked up by the antenna. 15, The touch-tone telephone and the erossover network are two typi- cal applications of filters. The touch-tone telephone system employs filters to separate tones of different frequencies to activate elec- tronie switches. The crossover network separates signals in differ- ‘ent frequency ranges so that they can be delivered to different devices such as tweeters and woofers in a loudspeaker system. ———————— 14a 142 14.3 144 14s 14.6 Azo of the rane faeton @o-% — @o-n ng = 10049 wo-8 won4 HO) B+ D 14.7 Ina parallel RLC circuit, the bandwidih B is dretly ew proportional oR @@ w-1 @-2 3 (@) Tne (6 False 14.8 When the elements of'an RC circuit are both ‘magnitude-sealed and fiequeney-sealed, which quality is unaffected? ‘On the Bode magnitude pot, the slope of 1/(5 + jo)” For lange values of wis (@)20dBidecade (6) 40 dBdecade (©) ~40aB/decade (8) ~20 dBidecade ‘On the Bode phase pot for 05 < « < 50, the slope off + flO — oF /25P 8 (@)45%idecade (6) 90°/decade (6) 138%decade—_(@) 180Hdecade How much inductance is needed to resonate at 5 kHz witha eapacitance of 12 nF? (a) resistor (b) resonant frequency (c) bandwidth (4) quality factor 14.9 What kind of filter can be used to selet a signal of ‘one particular radio station? (a) lowpass, (b) highpass (©) bandpass (€) bandstop 14.10. A voliage source supplies signal of constant amplitude, from 0 co 40 kHz, tan RC lowpass fier, (0)2.652H (ss ‘load resistor, connceted in parallel across the ()3333H () $4.43 mH capacitor, experiences the maximum vollag a: The difference between the half-power frequ (ee ()10kitz called the (©) 20 kHz (40 ki (2) quality factor (b) resonant fequeney (e)bandwidth —_(€) cutoff frequency Ina series REC cieuit, which ofthese quality factors —_____ has the steepest magnitude response curve near Answers: 14.16, 14.2, 14.3, 144d, 14.5c, 14.60, resonance? 14.7b, 14.8d, 14.96, 14.10.Problems 60s ——— Section 14.2 Transfer Function 14.1. Find the transfer function V,/V, ofthe RC cireuit in Fig. 14.68, Express it using wy = 1/RC. c —i— o0® R Figure 14.68 For Prob. 14.1. 14.2 ead Using Fig, 14.69, design a problem to help other students better understand how to determine transfer functions. Figure 14.69 * & i vO - Lr ForProb. 142. 14.3. Given tho circuitin Fig, 18.70, Ry 0,8, = 50, C; = 0.1 F,and C; = 0.2 F, determine the transfer funtion H(s) = V,6s)/¥4). Ry R ¥; a Figure 14.70 For Prob. 143. 14.4 Find the transfer function H(w) = V/V, of the circuits shown in Fig. 14.71, L v, c gr Y% o c 3 3 R vy, v, ae o Figure 14.71 For Prob. 14.4, 14.5. Foreach of the circuits shown in Fig, 14.72, find HG) = V/V). Figure 14.72 For Prob. 145 14.6 For tho cireuit shown in Fig, 14.73, find H(s) = 1,6)/16). L@ ef tain Figure 14.73 For Prob. 14.6. Section 14.3 The Decibel Scale 14.7 Calculate |H(o)| if Hon equals (005 4B (b)-624B (©) 104.748 14.8 Design a problem to help other students efaed caleulate the magnitude in dB and phase in degrees ofa variety of transfer functions at a single value of o. Section 14.4 Bode Plots 14.9 A ladder notwork has a voltage gain of 10 (+ joy(10 * jo) Sketch the Bode plots for the gain. Ho) =