UNIT 4 FREQUENCY RESPONSE 4. 1 SINUSOIDAL TRANSFER FUNCTION AND FREQUENCY RESPONSE, ‘The response of a system for the sinusoidal input is called sinusoidal response. ‘The ratio of sinusoidal response and sinusoidal input is called sinusoidal transfer function of the system and in general, it is denoted by T(j). The sinusoidal transfer fiction is the frequency domain representation of the system, and so itis also called frequency domain transfer function. ‘The sinusoidal transfer function can be obtained as shown below. 1. Construct a physical model of a system using basic elements/parameters. 2. Determine the differential equations governing the system from’ the physical model of the system, 3. Take Laplace transform of differential equations in order to convert them to s-domain equation, 4. Determine s-domain transfer function, T(s), which is ratio of s-domain output and input. 5, Determine the frequency domain transfer- function, T(jw) by replacing § by jo in the s-domain transfer function, T(s). it FREQUENCY RESPONSE: ‘The frequency domain transfer funetion TG jis a complex function of w. Hence it ean be separated into magnitude function and phase function, Now, the magnitude and phase funetions will be real functions of to, and they are called frequency response. The frequency response can be evaluated for open loop system and closed loop system. The frequency domain transfer function of open loop and closed loop systems can be obtained from the s-domain transfer function by replacing s by jw shown below. Open loop transfer function G(S) _s=jo Ga) = |G Go)| 26jo) Go) Ro) Closed loop transfer function = M(jw) = |MGw)| 2a) ‘The advantages of frequency response analysis are the following, 1. The absolute and relative stability of the closed loop system can be estimated from the knowledge of their open loop frequency response.2. The practical testing of systems can be easily carried with available sinusoidal signal generators and precise measurement equipments. 3. The transfer function of complicated systems can be determined experimentally by frequency response tests. 4, The design and parameter adjustment of the open loop transfer function of a system for specified closed loop performance is carried out more easily in frequency domain. 5. When the system is designed’ by use of the frequency response analysis, the effects of noise disturbance and parameters variations are relatively easy to visualize and incorporate corrective measures, 6. The frequency response analysis and designs can be extended to certain nonlinear control systems. 4.2 FREQUENCY DOMAIN SPECIFICATIONS ‘The performance and characteristics of a system in frequency domain are measured in terms of frequency domain specifications. The requirements of a system to be designed are, usually specified in terms of these specifications. The frequency domain specifications are, 1. Resonant peak, Mr 2. Resonant Frequency, oo . Bandwidth 4. Cut-off rate v .. Gain margin, Kg 6 Phase margin, y Resonant Peak: The maximum value of the magnitude of closed loop transfer function is called the resonant peak. A large resonant peak corresponds to a large overshoot in transient response. Resonant Frequency: ‘The frequency at which the resonant peak occurs is called resonant frequency. Th related (0 the frequency of oscillation in the step response and thus it is indicative of the speed of transient responseBandwidth: The Bandwidth is the range of frequencies for which normalized gain of the system is more than -3db. The frequency at which the gain is -3db is called cut-cff frequency. Band width is usually defined for closed loop system and it transmits the signals whose frequencies ate less ‘than the cut-off frequency. The Bandwidth is a measure of the ability of a feedback system to reproduce the input signal, noise rejection characteristics and rise time. A large-bandwidth corresponds to a small rise time or fast response. ‘Cut-off Rate: ‘The slope of the log-magnitude curve near the cut off frequency is called cut-off rate. The ‘cut -off rate indicates the ability of the system to distinguish the signal from noise. Gain Margin, Kg: ‘The gain margin, Kg is defined as the reciprocal of the magnitude of open loop transfer function at phase cross over frequency Phase cross-over frequency, eye ‘The frequency at which the phase of open loop transfer function is 180° is called phase cross-over frequency, dpe. Gain Margin, Kg Kewl 7) ‘The gain margin in db can be expressed as, 1 Kg in db =20 log Kg = 200g 755 Phase Margin, y ‘The phase margin , is defined as the additional phase lag to beaded at the gain cross ‘over frequency in order to bring the system to the verge of instability. Gain cross over frequency dye: ‘The frequency at which the magnitude of the open loop transfer function is unity (or itis the frequency at which the db magnitude is zero) is called gain cross over frequency @yc. ‘The phase margin, is obtained by adding 180° to the phase angle of the open loop transfer function at the gain cross over frequency. Phase margin y =180° + Pye4.3 BODE PLOT ‘The Bode plot is a frequency response plot of the sinusoidal transfer function of a system. ee 5 Bode plot consists of two graphs. One isa plot of the magnitude of a sinusoidal transfer function versus log@, The other is a plot of the phase angle of sinusoidal transfer function versus logw. ‘The Bode plot can be drawn for both open loop and closed loop system. Usually the bode plot is drawn for open loop system. The standard representation of the logarithmic magnitude of open loop transfer function of Gjw) is 20 Log |G(jw)|, Where the base of the logarithm is 10. The unit used in this representation of the magnitude is the decibel, usually abbreviated db, The curves are drawn on semilog paper, using the log scale (abcissa) for frequency and the linear scale (ordinate) for either magnitude (in decibels) ‘or phase angle (in degrees). ‘The main advantage of the bode plot is that multiplication of magnitudes can be converted into addition, Also a simple method for sketching an approximate log-magnitude curve is available. K(1+sT1) Consider the open loop transfer function, GO)= sry pTay , K(1+j0T1) S6e) = TararnaHars) 4.3.1 PROCEDURE FOR MAGNITUDE PLOT OF BODE PLOT From the analysis of previous sections the following conclusions can be obtained. 1. The constant gain K, integral and derivative factors contribute gain (magnitude) at all frequencies. 2. In approximate plot the first, quadratic and higher order factors contribute gain (magnitude) only when the frequency is greater than the comer frequency. Hence the low frequency response upto the lowest comer frequency is decided by K or K / (je)" or K(j)" term. Then at every comer frequency the slope of the magnitude plot is altered by the first, quadratic and higher order terms. Therefore the magnitude plot can be started-with K or Kj)" or KG)” term and, then the db magnitude of every first and higher order terms are added one by one in the increasing order of the comer frequency. ‘This is illustrated in the following example. Kass? OS) SG esray +ska+Gort)? “SU+Gor2)G+Gar3) Go) Let, T1 SUTIN 4285) KQ+jort) GUa)=——— Jott jor2)a~ 24262) Step 2: List the comer frequencies in the increasing order and prepare a table as shown below. Term | Corner frequency Slope | Change in slope rad/sec db/dec db/dec In the above table enter K or K / (jo)" or K(j)" as the first term and the other terms in the increasing order of comer frequencies. Then enter the comer frequency, slope contributed by cach term and change in slope at every corner frequency. Step 3: Choose an arbitrary frequency (1)1 which is lesser than the lowest corer frequency. Calculate the db magnitude of K or K / (j@)" or K(jw)" at o and at the lowest comer frequency. Step 4: Then calculate the gain (db magnitude) at every corner frequency one by one by using the formula, Gain ateoy = change in gain from ws to @y + Gain at en = [slope from o, to oy Hog(ooy/ @.)] + gain at oMagnitude in db ———» a 8. e ‘otlog sale) Figure:4.1 Step 5: Choose an arbitrary frequency, os which is greater than the highest frequency. Calculate the gain at «, by using the formula in step 4. Step 6 : In a semi log graph sheet mark the required range of frequency on x-axis and the range of db on y-axis (ordinary scale) after choosing proper unit. Step 7: Mark all the points obtained in steps3.4, and 5 on the graph and join the plot by straight lines, Mark the slope at every part of the graph.Figure4.2 4.3.2 PROCEDURE FOR CONSTRUCTING PHASE PLOT OF BODE PLOT: ‘The phase plot is an exact plot and no approximations are made, while drawing the phase plot. Hence the exact phase angles of G(jo) are computed for various values of and tabulated. ‘The choice of frequencies are preferably chosen for the magnitude plot. Usually the magnitude plot and phase plot are drawn in a single semi log Sheet on a common frequency scale. Take another y axis in the graph where the magnitude plot is drawn and in this y-axis mark the desired range of phase angles after choosing proper units from the tabulated values of @ and phase angles, mark all the points on the graph, Join the points bya smooth curve. 4.3.3 DETERMINATION OF GAIN MARGIN AND PHASE MARGIN OF BODE PLOT: =e log scale) on ‘olog scale) Figures.3 ‘The gain margin in db is given by the negative of db magnitude of G(jm) at the phase cross over frequeney, Mpc. The wp. is the frequency at which the phase of GQ) is 180°. If the db magnitude of G(jq) is negative then the gain mar n is positive and vice versa.Let @ye be the phase angle of Gijw) at gain cross over frequency @g.. The ay is the frequency at which the magnitude of G(jw) is zero. Now the phase margin is given by ‘Y=180° Gy If Gis less than -180° then the phase margin is positive and vice versa. The positive and negative gain and phase margins are illustrated in above figure. 4.3.4 GAIN ADJUSTMENT IN BODE PLOT: In the open loop transfer function Gj) the constant K contributes only magnitude. Hence by changing the value of K the system-gain can be adjusted to meet the desired specifications. ‘The desired specifications are gain margin, phase margin, In a system transfer function if the value of K required to be estimated to a desired specification then draw the-bode plot of the system with K = 1. The constant K can add, 20 log K to every point of the magnitude plot and due to this addition the magnitude plot will shift vertically up or down, Hence shift the magnitude plot vertically up or down to meet the desired specification. Equate the vertical distance by which the magnitude plot is shifted to 20logK and solve for K. Let, x = change in db Now, 20logK = log = x/20: K=10*/20 EXAMPLE 4.1 Sketch Bode plot for the following transfer function and determine gain cross over frequency and phase cross over frequency. . 10 SO) = SGyoas\av045) SOLUTION The sinusoidal transfer function G(jo) is obtained by replacing s by jo in the given twansfer function. MAGNITUDE PLO’ ‘The comer frequencies are, wel = 0.4 =2.5 us 1/0. we? Ors The various terms of G(jo) are listed in Tatile-1 in the increasing order of their comer frequency. Also the table shows the slope contributed by each term and the change in slope at the comer frequency.‘TABLE-1 Chose a low frequency « such that @< @21 and chose a high frequency «y such that we 2 Let 20.1 r/s and on-50 1s. Let A=[G(ja)] in db. Let us calculate A at ©), Oe1, Og and Op, At o= a A= 200g] AL @= we2, A= [slope from wet coaxlog==] + A(at w= wat) = -40xlog( 10/2.5)+12 = -12db. AL @= wn, A= [slope from wero: conxlog ==} + A(at @= we2) = -60xlog(50/10)+(-12) = -S4db. Let the points a, b, ¢ and d be the points corresponding to frequencies @, 1, 2 and © respectively on the magnitude plot in a semilog graph sheet choose a scale of Lunit-10 db on y- axis. The frequencies are marked in decades from 0.1 to 100 #/s on logarithmic scales in x-axis, Fix the points a, b, ¢ and d on the graph. Join the points by a straight line and mark the slope in the respective region PHASE PLOT: The phase angle of G(/w) as a function of w is given by, @ =-90°-tan™! 0.-4w-tan™? 0.10rs los- <8 Figure:4.4‘The phase angles of various values of w are calculated and listed in table 2. TABLED ° tan 04 @ tan? 0.1 @ Points in radisec deg deg phase plot an 229 os? ° 1 21.80 an 1 | 28 450 40 8 | oo4 5799 28 h 0 75.98 450 1 2 e2s7 3.43 On the same semilog graph sheet choose a scale of 1 unit=20° on the y-axis on the right side of semilog graph sheet. Mark the phase angles on the graph sheet and join the graph by smooth curve. RESULT: 1. Gain cross over frequeney = 5r/s 2. Phase cross over frequency = 5r/s. EXAMPLE 42 Sketch Bode plot for the following transfer function and determine system gain K for the gain cross over frequency to be 51/s. G(S) = EST SOLUTION The sinusoidal transfer function G(jo) is obtained by replacing s by jo in the given transfer function. Let, K=I get, G jo) = KGuy G+0.2)w)(1+0.02}0) MAGNITUDE PLOT: ‘The comer frequencies are, wel = 110.2 coc = 10.02 = 5 ‘The various terms of G(jw) are listed in Tatile-1 in the increasing order of their corner frequency. Also the tahle shows the slope contributed by each term and the change in slope at the comer frequency. Chose a low frequency «, such that ac ox Let o=0.1 r/s and on =104/s. TABLE-1 Term ‘Comer frequency Slope Change in slope rad/sec db/dec db/deg 5 S 1 Pe, Tio -20 0-20=-20 1+)20 ete weg Te 0.250 3 = O55 Let A= |G(a)] in db. Let us calculate A at 1, @e1, @2 and wn At o= a, A= 20log|(jo)| = 2010g(5) = 14db. At o= wa, A= 20log| (jw)| = 20log(5) = 14db. Ato <2, A= [slope from 0» Oxxlog#=2] + A(at o= 1) = -20xlog((.5/0.25)+14 = 8db. Ato= os A= [slope from wna nenxlog 22] + ACat = wa) = Oxlog(4/0.5)4(8) = 8db.Do5/pes Uy <— ool os. on ¥ Figure:4.6 H [estos £ z We8l9s% € Wesiose € 7% Toso € 1Let the points a, b, c and d be the points corresponding to frequencies 1, «1, a2 and oy respectively on the magnitude plot in a semilog graph sheet choose a scale of Lunit-Sdb on y- axis. The frequencies are marked in decades from 0.1 to 100 #/s on logarithmic scales in x-axis, Fix the points a, b, c and d on the graph, Join the points by a straight line and mark the slope in the respective region. PHASE PLOT: The phase angle of G(ja) as a function of « is given by, @ =-tan! 2otan™! 4u-tan™ 0.250 ‘The phase angles of various values of « are calculated and listed in table 2, TABLE-2 ° tant 20 tan 40 tan 0.250 ¢ = 2G(j0) Points in seg deo deg phase plot 01 na 218 143, -11.93~-12 1 0.25 26.56 450 35 -21.94*-22 9 05 450 63.43 7A -25.53~-26 h 2 7596 || 8287 26.56 -33.47=-33 i 4 8287 86.42 450 48 55~-49 i 10 87.13 88.56 68.19 -69.62=-70 k 50 89.42 sort | 8542 85.71 =-86 i On the same semilog graph sheet choose a scale of 1 unit=10° on the y-axis on the right side of semilog graph sheet. Mark the phase angles on the graph sheet and join the graph by smooth curve. 4.4 POLAR PLOT. ‘The polar plot of a sinusoidal transfer fu versus the phase angle of G(je) on polar coordi polar plot is the locus of vectors [G(j)] 26(j@) as «is varied from zero to infinity. The polar plot is also called Nyquist plot. The polar plot is usually plotted on a polar graph sheet. The polar graph sheet has concentric circles and radial lines. The circles represent the magnitude and the radial lines represent the phase angles. Each point on the polar graph has a magnitude and phase angle the magnitude of a point is given by the value of the circle passing through that point and the phase angle is given by the radial line passing through that point. In polar graph sheet a positive phase angle is measured in anti clockwise from the reference axis (0°) and a negative angle is measured jon G{jw) a-plot of the magnitude of Gu) as is varied from zero to infinity. Thus the clockwise from the reference axis(0”).Figure:4.7 In order to plot the polar plot, magnitude and phase of G(jw) are-computed for various values of o and tabulated, Alternatively, if GGw) can be expressed in rectangular coordinates as, G(j@) = GrjoGiljo) where, Gr(jo}=Real part of Gio); Gi(j@) = Imaginary part of GGw). ‘Then the polar plot can be plotted in ordinary graph sheet between Ge(ja)and Gi(jo) by varying w from o to <, In order to plot the polar plot on ordinary graph sheet, the magnitude and phase of Gj) are computed for various values of o. Then convert the polar coordinates to rectangular coordinates using P+R conversion (polar to rectangular conversion) in the calculator. Sketch the polar plot using rectangular coordinates. Start of ype-s | End of 3% End of 4” system order system order system Start of type-2—> = Start of type-0 ‘system system End of 2% End of 1* order system order system ‘Start of type-1 system Start of polar plot ofall pole Start of polar plot of all pole minimum phase system. minimum phase system. Figure:4.8 For minimum phase transfer function with only poles, type number of the system determines the quadrant at which the polar plot starts and the order of the system determines. the quadrant at which the polar plot ends. The minimum phase systems are systems with all polesand zeros on left half of s- plane. The start and end of polar plot of all pole minimum phase system are shown in figures respectively. Some typical sketches of polar plot are shown in table. ‘The change in shape of polar plot can be predicted due to addition of a pole or zero. 1. When a pole is added toa system, the polar plotend point will shift by -90°. 2. When azem is added toa system the polar plot end point will shift by +90°. 44.1 TYPICAL SKETCHES OF POLAR PLOT: 20, Order : 1 SO" et Iino Zatot iva Asa 90, G(jo) > 120" Asam, G{jo) > 02-90" i ‘Dei l, Order :2 Gi)= @, |-270 8) oa 0 ty 1 1 1 G(jo) = 2. ee £(-90" ~ i“ let) O°)" aI 9 290 va Zan et vise : 8 As@ 40, Gljo) > 22-90" Aso a, ,G(jo) > 02-180" 1 tH aaa oye Wo) That) Cr jah) aa eal leo Zan 'ah, tan" 01, ~tan!0) Tape Aso $0, G(ja) -» 120" As@ +2, G{jo) + 02-180" ‘Type: 0, Order: 3 G= 1 Tp jet) GssT)OrsT +s) Gije)= *Yieo*h Zia let, fi+0"? Zim 'or, Yio" Zot 1 - £(-tan oT, - tan” or - tan“orT) [iro] (--0) (+0) ee Asa 0, G(jo) + 1x0" Aso >, G(ja) + 04-270"‘Type: t, Order :3 2 C= Tea) 180% Cf es J Ss jo0+50%) OF) g 207 five Zan oy, yin" Zan oh, eaten ila) \ often) = As@>0, G(jo) + 22-90" 20 Aso, G(jo)—> 02-270" 1 00 Sia 1 1 G50) = en GoF (507) (507) @?Z—180 fire Zan tay, ftw Zeno = é A- 180° — lat, - telat) oe [ron fimo) As@ 90, Gljo) > 2-180" Aso, Olja) + 02-360" 1 = Fieaeapae) 1 0)" FoF Got (I= oR HET) ae 1 “ORL-180 [iso Leh fio Zan oh, five Zio 1 iad . A¢- 180 ~ tao ~ tot — tft) of [ron ire n)ven) A800, Glo) > 22-100" Aso >, Gf) + 02-450" 02-90" ‘ume:2, Onder: § ‘Type: 1, Order: 1 Ga)=5 ye te tt G(je)2 + -—_1 Ne) aor a AS@>0, Gia) > mZ-90" Aso, G(jo) > 02-90" Figure:4.9442 DETERMINATION OF GAIN MARGIN AND PHASE MARGIN FROM POLAR PLOT: The gain margin is defined as the inverse of the magnitude of G(jw) at phase crossover frequency. The phase crossover frequency is the frequency at which the phase of Gj) is 180°. Let the polar plot cut the 180° axis at point B and the magnitude circle passing through the point B be Gn. Now the Gain margin, Kg = 1/Gu. Ifthe point B lies within unity cirele, then the Gain margin is positive otherwise negative. (If the polar plot is drawn in ordinary graph sheet using rectangular coordinates then the point B is the cutting point of G(ja) locus with negative real axis and K,=I/|Gpl where GB is the magnitude corresponding (o point B). The following figure showing positive gain and phase margin. Phase margin, y =180° +4, Figure:4.10 ‘The following figure showing negative gain and phase margin. Figure:4.11The phase margin is defined as, phase margin, y=180°+d, where dx is the phase angle ‘of Gia) at gain crossover frequency. The gain crossover frequency is the frequency at which the magnitude of Gj) is unity. 4.4.3 GAIN ADJUSTMENT USING POLAR PLO! Draw G(jo) locus with K = 1. Let it cut the -180° axis at point B corresponding to a gain of G. let the specified gain margin be x db. For this gain margin, the G(jo) locus will cut -180° at point A whose magnitude is Ga Figure:4.12 Now, 200g x 1 lores x20 2-10 1 Gaw 1020 Now the value of K is given by, K = G/Gu EXAMPLE 4.4 Sketch the polar plots of I/s SOLUTION jo Given that, G(s) = 1/s Puts= jo, we get ° GGjw) = bho . If o=0, then [G(ja)| = ‘ I a=, then [G(ja)] =0 ; Figure:4.13EXAMPLE 4.5 Sketch the polar plots of I/s(L+sT). SOLUTION Given that, Gis) Puts= jo, we get G(i@) = Ljio(1+j0T) ‘ 1 166) = Sar = Us(14sT) LG(jw) =-90°-tan“! wT If «=0, then |G(ja)| =» and 2G(joo) =-90° If w=:0, then [G(jw)| = 0 and ZGGw) =-180° Figure:4.14 EXAMPLE 4.6 The open loop transfer function of a unity feedback system is given by G(s)=1/s(145)(142s). Sketch the polar plot and determine the gain margin and phase margin, SOLUTION Given that, G(s) = /s(1+s)(142s) Puts= jo G(w) = 1 jo +joy(142j0) ‘The comer frequencies are wel = 1/2= 0.5 1/s we2 = 1 fs. The magnitude and phase angle ‘of G{jw) are calculated for the comer frequencies and for frequencies around comer frequencies and tabulated in table, The polar plot using polar coordinates is sketched on a polar graph sheet as shown in figure. Gia) Jas fay +2 fa) a e290" (Gta?) tan wfGtarz tan" 20 1 a £00 — tan w — tan! 20 ota) (+40) 4 669)! =" aananae)30° 20" 10° 3s0e 40" 330° 20 “40° 40° 320" shoe “50° soe 310 300° 60" a 300" 290° 0" 70° -290¢ Pe Eve oo 20" : ay. iat Je ae 60" 250" ios ite Soh 240" 120: be Bee ae a0 130" SoM ae no. 140 ar 20° 20 Pe a Tr ae =130° ws Vor ora wor = Boek Figure:4.151 oN Hattanrdar 1 1G) | = aratraat 26(jw) =-90%tan“ w-tan“! 20 Table: ‘Magnitude and phase of G(jw) at various frequencies @ raaisec os | o4 | 04s | os 06 o7 | 10 Bi 22 18 18 12 [08 o7 | 03 2660) am | 0 | tse [te | 1m | -1705 | -198 eg =-180. RESULT: 1. Gain margin, Kz= 1.42 2. Phase margin, y = +12° EXAMPLE 4.7 ‘The open loop transfer function of a unity feedback system is given by G(s)=1/s(1+ s)*. Sketch the polar plot and determine the gain margin and phase margin SOLUTION Given that, G(s)=1/s(1+ 5° Puts jo Gj) = I jo 1+ jo)” ‘The comer frequency we1= Ir/s. The magnitude and phase angle of G(joo) are calculated for the corner frequencies and for frequencies around comer frequencies and tabulated in table, ‘The polar plot using polar coordinates is sketched on a polar graph sheet as shown in figure. 1 08) oar iaF 1 pis (re) ctan a) Gra?) ctan tw30" 20" 10" 350" 340 330 =330° 40" 350" o = 20" =30" 40" 30° =o" or 300° 70 290° 80" 280° 90° a0 100° 260" nor 280" ( 120° toe 130" 30" 40° 0 200" 190" 180" =170" =130" 150" 160" v0" 180" 190° 210" Figurest.16= = = tan-t @ — tan-? = eae 290° = tan™? @ = tan“ w 1 |GG)|_ =———> Ol =) ae 1 * oGrut) 1 I6Ga) | == 2G(jm) =-90%-2tan“! w Table: “Magnitude and phase of G(jw) at various frequencies ° ] | | radtsec | o4 | 05 | 06 | o7 | os | op | 1 [as | Re 22 | 16 | 12 | 1 [os | 06 | 05 | on | How can you determine gain margin and phase margin in polar plot? i Draw the polar plot of G18) = sara Draw the polar plot of G(S) = /S? yeep TEN MARKS 1. Explain the procedure for constructing Bode plot. 2. Sketch Bode plot for the following transfer function and determine gain cross over frequency and phase cross over frequency. G(S) = aS 3. Sketch Bode plot for the following transfer function and determine gain cross over frequency and phase cross over frequency. G(S) = ET 4. Sketch Bode plot for the following transfer function and determine gain cross over 105410) Ssens+5) ‘The open loop transfer function of a unity feedback system is given by G(s)=1/?(14s)(1428). Sketch the polar plot and determine the gain margin and phase margin. 6. The open loop transfer function of a unity feedback system is given by G()E1SG+4)is+8). Sketch the polar plot and determine the gain margin and phase margin. frequency and phase cross over frequency. G(S) =UNIT 5 STABILITY 5.1 STABILITY DEFINITIONS OF STABITY The term stability refers to the stable working condition of @ control system, Every working system is designed to be stable. In a stable system, the response or output is predictable, finite and stable for a given input (or for any changes in input or for any changes in system parameters). ‘The different definitions of the stability are the following. 1.Stable system: “A system is stable, if its output is bounded (finite) for any bounded (finite) input.” 2. Asymptotically stable system: “A system is asymptotically stable, if in the absence of the input the output tends towards zero (or to the equilibrium state) irrespective of initial conditi 3. Unstable system: “A system is unstable if for a bounded disturbing input signal the output is of infinite amplitude or oscillatory” 4. Limitedly stable system: “For 2 bounded input signal, if the output has constant amplitude oscillations then the system may be stable or unstable under some limited constraints. Such a system is called limitedly stable system” 5, Absolutely stable system: “If a system output absolutely stable system” table for all variations of its parameters, then the system is called 6. Conditionally stable system: “If a system output is stable for a limited range of variations of its parameters, then the system is called conditionally stable system” 7. BIBO stabilit“A linear relaxed system is sai results in a bounded (finite) output’ to have BIBO stability if every bounded (finite) input - 8. Relative stability: The Relative stability indicates the closeness of the system to stable region, It is an indication of the strength or degree of stability” RESPONSE OF A SYS ‘EM. Let the Closed loop transfer function mc M(S) ‘The response or output in s domain C(S)= M(S)R(S) C(s) = Output Response in s-domain, R(s) = Input Response in s-domain. 5.2 LOCATION OF POLES ON s-PLANE FOR STABILITY ‘The closed loop transfer function, M(s) can be expressed as a ratio of two polynomials in 8. The Denominator polynomial of closed loop transfer function is called characteristic equation. ‘The roots of characteristic equation are poles of closed loop transfer function, For BIBO stability the integral of impulse response should be finite, which implies that the impulse response should be finite as t tends to infinity. [The impulse response is the inverse Laplace transform of the transfer function]. This requirement for stability can be linked to the location of roots of characteristic equation in the s-plane. ‘The closed loop transfer function M(s) can be expressed a ratio of two polynomials, (s#21) (5422) (s422)...(st2m) (GDI) (+2) (+3) .S¥pn) ‘The roots of numerator polynomial z1z2,23. ..... zn are zeros. ‘The roots of denominator polynomial p!,p2,p3....pm are poles ‘The denominator polynomial is the characteristic equation and so the poles are ‘roots of characteristic equation. From table, the following conclusions are drawn based on the location of roots of characteristic equation 1. IF all the roots of characteristic equation have negative real parts (i., lying-on left half s- plane) then the impulse, response is bounded (Le... it decreases to zero, as t tends to =). Hence, the system has bounded-input and bounded-output so the system is stable.2. If any root of the characteristic equation has a positive real part (i.e., lying on right half s- plane) then impulse response is unbounded,(i.e.,it increases to as t tends to 2). Hence the system is unstable. 3. If the characteristic equation has repeated roots on the imaginary axis then impulse response is unbounded (i... it increases to % as t tends to »).Hence the system is unstable, Transfer function, M(s) and location of roots on s-plane Impulse response, m(t) my At m4 m= cf Ahan 2 aa Complex conjugate roots on left half of s-plane f Root on negative real axis ing Salensten Moya A joy mo, Mee cif Ah aac ts ‘ ‘ Root on positive real axis Impulse response is exponentially increasing, Unrate sistem [a _, mnee aS = Aes Aree “impulse response is damped simusoidal (ie., Damped oscillatory). Stable systemM(s)= ae on Complex confugate roots on right half of s-plane La, = mon of A =. | mA OR ate oie = 2Ac™ cosbt = 2Ac" sin (bt +90") Impulse response is exponentially increasing sinusoidal (Le,, Amplitude of oscillations exponentially increases swith ime). Unstable system. Single pair of roots on imaginary axis cf A mi) ole As =Ae™ + Ate (0 = 2A cos bt =2A sin (bt +90") Impulse response is oscillatory Marginally stable Double pair of roots on imaginary axis m(t) =" aoe + A s { ow te Meare = 2At cos bt=2At sin (br +90") me) Impulse response is linearly increasing sinusoidal (ce., amplitude of oscillations linearly increases with time). Unstable system.io mF miy= e{4} = Single root at origin Paelnsened fs constant farginalty stable system. t Impulse reponse linearty increases Double root at origin | with time. Unstable system Figure:5.1 4. Ifone or more non ~ repeated roots of the characteristic equation are lying on the imaginary axis, then impulse response is bounded (ie., it has constant amplitude oscillations) hence the system-is unstable 5. If the characteristic equation has single root at origin then the impulse response is bounded (i.e, it has constant amplitude) hence the system is unstable. 6. If the characteristic equation has repeated roots at origin then the impulse response is unbounded (i.e, it linearly increases to infinity as t tends to 0) and so the system is unstable 7.In system with one or more non-repeated rots in imaginary axis or with single root at origin, the output is bounded for bounded inputs except for the inputs having poles matching the system poles. These cases may be treated as acceptable or non-acceptable. Hence when the system has non repeated poles on imaginary axis or single pole at origin, it is referred as limitedly or marginally stable system.In summary, the following three points may be stated regarding the stability of the system depending on the location of roots of characteristic equation. 1. If all the roots of characteristic equation have negative real parts, then the system is stable. 2. If any root of the characteristic equation has a positive real part or if there is a repeated root on the imaginary axis then the system is unstable. 3. If the condition (i) is satisfied except for the presence of one or more non repeated roots on the imaginary axis, then the system is limitedly or marginally stable. In summary, the following conclusions can be made about coefficients of characteristic polynomial 1. If all the coefficients are positive and if no coefficient is zero, then all the roots are in the lett half of s- plane. 2. If any coefficient a is equal to zero then, some of the roots may be on the Imaginary axis or the right half of s- plane. 3. If any coefficient a: is negative, then at least one root is in the right half of $- plane. 5.3 ROUTH HURWITZ CRITERION ‘The Routh-Hurwitz stability criterion is an analytical procedure for determining whether all the roots of a polynomial have negative real part or not The first step in analyzing the stability of a system is to examine its characteristic ‘equation. The necessary condition for stability is that all the coefficients of the polynomial be positive. If some of the coefficients are zero or negative it can be concluded that the system is not stable. When all the coefficients are positive, the system is not necessarily stable. Even though the coefficients are positive, some of the roots may lie on the right half of s-plane or on the imaginary axis. In order for all the roots to have negative real parts, it is necessary but not sufficient that all coefficients of the characteristic equation be positive. If all the coefficients of the characteristic equation are positive, then the system may be stable and one should proceed further to examine the sufficient conditions of stability. A. Hurwitz and EJ. Routh independently published the method of investigating the sufficient conditions of stability of a system. ‘The Hurwitz criterion is in terms of determinants and Routh criterion is in terms -of array formulation. The Routh stability criterion is presented here.‘The Routh stability criterion is based on ordering the coefficients of the characteristic quation, aos™+ars™ 4ars" 2+... +an-18+an=0, where ao>0 into a schedule, called the Routh array as shown below Some Oe sora Re RL < eee str bm mm fr yg 4 & YF eG sory ' 5 2 hy ‘The Routh stability criterion can be stated as follows. "The necessary and sufficient condition for stability is that all of the elements in the first column of the Routh array be positive. If this condition is not met, the system is unstable and the number of sign changes in the elements of the first column of the Routh array corresponds to the number of roots of the characteristic equation in the right half of the s-plane". Note: If the order of sign of first column element is +, +, -, + and +. Then + to -is considered as ‘one sign change and -to + as another sign change 5.3.1 CONSTRUCTION OF ROUTH ARRAY Let the characteristic polynomial be, “rays! as Fas The coefficients of the polynomial are arranged in two rows as shown below. e Ao az ay as st ay a3 as a When n is even, the s* row is formed by coefficients-of even order terms (ic., coefficients of even. powers of s) and s™ ‘row is formed by coefficients of odd order terms (ic., coefficients of odd powers of s). When n is odd, the s* row is formed by coefficients of odd order terms (i.c., coefficients of odd powers of s) and s“! row is formed by coefficients of even order terms(ie., coefficients of even powers of s)The other rows of routh array upto S° row can be formed by the following procedure. Each row of Routh array is constructed by using the elements of previous two rows. ‘Consider two consecutive rows of Routh array as shown below, FMM Xe Ky Xe Kyron Yo Yi: Y2 Ys Let the next row be, Pa he Be ‘The -lements of s**? row are given by, ko o_¥il _ YoXs~ Yio Yo Yo Ko x, z= —Pe_Yal Yo = Yao Yo Yo * ke ‘o_Ys} _ YoXs~YsXo Yo Yo x4| fo_Yal _ YoXa—Ys Yo Yo x 1y-——No_ Yah Yo%s=¥5% and so on. Yo Yo ‘The elements 70,21,22,22.... are computed for all possible computations as shown above. ) %= In the process of constructing Routh array the missing terms are considered as zeros. Also, all the elements of any row can be multiplied or divided by a Positive constant to simplify the computational work In the construction of Routh array one may come across the following three cases. Case-I : Normal Routh array (Non-zero elements in the first column of routh array). Case-II : A row of all zeros. Case-III_: First element of a row is zero but some or other elements are not zero.Case-I: Normal Routh Array In this case, there is no difficulty in forming routh array. The routh array can be constructed as explained above. The sign changes are noted to find the number of roots lying on the right half of s-plane and the stability of the system can be estimated. In this case, 1. If there is no sign change in the first column of Routh array then all the roots are lying on left half of s-plane and the system is stable. 2.1 there is sign change in the first column of routh array, theri the system is unstable and the~ number of roots lying on the right half of s-plane is equal to number of sign changes, The remaining roots are lying on the left half of s-plane. Case-II: A row of all zeros ‘An all zero row indicates the existence of an even polynomial as a factor of the given characteristic equation. In an even polynomial the exponents of s are even integers or zero only. ‘This even polynomial factor is also called auxiliary polynomial. ‘The coefficients of the auxiliary polynomial will always. be the - elements of the row directly above the row of zeros in the array. 's that are equal in magnitude and opposite in sign. Hence, these roots can be purely imaginary, purely real or complex. The purely ‘The roots of an even polynomial occur in pa imaginary and purely real roots occur in pairs. The complex roots occur in groups of four and the complex roots have quadrantal symmetry, that is the roots are symmetrical with respect to both the real and imaginary axes. The following figure shows the roots of an even polynomial jo: = a Figure:5.2 ‘The case-II polynomial can be analyzed by anyone of the following two methods. METHOD, 1. Determine the auxiliary polynomial, A(s) 2. Differentiate the auxiliary polynomial with respect to s, to get d A(sy/ds3. The row or zeros is replaced with coefficient of dA(s)/ds. 4. Continue the construction of the array in the usual manner (as that of ease-L) and the array is interpreted as follows. a. If there are sign changes in the first column of routh array then the system is unstable, ‘The number of roots-lying on right half of s-plane is equal to number of sign changes. ‘The number of roots on imaginary axis can be estimated from the roots of auxiliary polynomial. ‘The remaining roots are lying on the left half of s-plane. b.Ifthere are no sign changes inthe first column of routh array then the all zeros row indicate the existence of purely imaginary rootsand so the system is I-tedly Or marginally stable, The’ roots of auxiliary equation lies on imaginary axis and the remaining roots lies on left half of s-plane, METHOD-2 1, Determine the auxiliary polynomial, A(s). 2. Divide the characteristic equation by auxiliary polynomial. 3. Construct Routh array using the coefficients of quotient polynomial. 4. The array is interpreted as follows. If there are sign changes in the first column of routh array of quotient polynomial then the system is unstable, The number of roots of quotient polynomial lying on right half of s- plane is given by number of sign changes in fist column of routh array. The roots of auxiliary polynomial are directly calculated to find whether they are purely imaginary or purely real or complex. ‘The total number of roots on right half of s-plane is given by the sum of number of, sign changes and the number of roots of auxiliary polynomial with positive real part. The number of roots on imaginary axis can be estimated from the roots of auxiliary polynomial. The remaining roots are lying on the left half of s-plane, b. I there is no sign change in the first column of routh array of quotient polynomial then the system is limitedly or marginally stable, Since there is no sign change all the roots of ‘quotient polynomial are lying on the left half of s-plane.The roots of auxiliary polynomial are directly calculated to find whether they are purely imaginary or purely real or complex. The number of roots lying on imaginary axis and on the right half of s-plane can be estimated from the roots of auxiliary polynomial. The remaining roots are lying on the left half of s-plane. Case-I First element of a row is zero While constructing routh array, if a zero is encountered as first element of a row then all the elements of the next row will be infinite. To overcome this problem let 0 € and complete the construction of array in the usual way (as that of case-I) Finally let € 0 and determine the values of the elements of the array which are functions of € ‘The resultant array is interpreted as follows, Note: If all the elements of a row are zeros then the solution is attempted by considering the polynomial as case-Il polynomial. Even if there is a single element zero on S! row, it is considered as a row of all zeros. a. Ifthere is no sign change in first column of routh array and if there is no row with an zeros, then all the roots are lying on left half of s-plane and the system is stable. b. If there are sign changes in first column of routh array and there is no row with all zeros, then some of the roots. are lying on the right half of s-plane and the system is unstable. The number of toots lying on the right half of s-plane is equal to number of sign changes and the remaining roots are lying on the left half of s-plane. c. If there is a row of all zeros after letting Oe then there is a possibility of roots on imaginary axis. Determine the auxiliary polynomial and divide the characteristic equation by auxiliary polynomial to eliminate the imaginary roots. The routh array is, constructed using the coefficients of quotient polynomial and the characteristic equation is interpreted as explained in method-2 of case-II polynomial. EXAMPLE 5.1 Using Routh criterion, determine the stability of the system represented by the characteristic equation, s4+8s!418s°+16s+5=0Comment on the location of the roots of characteristic equationSOLUTION ‘The characteristic equation of the system is, s'+8s"+ 1887+ 16s+5=0 The given characteristic equation is 4" order equation and so it has 4 roots. Since the highest power of s is even number, form the first row of routh array using the coefficients of even powers of s and form the second row using the coefficients of odd powers of s. * 15 Rom 2, RAB-2at 15-01 s 8 6 on Row? . Sa ‘The elements of s* row can be divided by 8 to simplify the 6 5 cautions. 1, 1632-504 s ® 5 Coe ° 2 $5 18875017 Oe 5 7x5-0016 s ANE ¢ cotumn-t On examining the elements of first column of routh array it is observed that all the elements are positive and there is no sign change. Hence all the roots are lying on the left half of plane and the system is stable. RESULT 1. Stable system 2. All the four roots are lying on the left half of s-plane. EXAMPLE 5.2 Construct Routh array and determine the stability of the system whose characteristic equation is s°+255+8s"+ 12s3+20s?+1 6s+ 1620. Also determine the number of roots lying on right half of s-plane, left half of s-plane and on imaginary axis. SOLUTION ‘The characteristic equation of the system is, s°+2sS+8s4+ 12s1420s"+16s+ 16=0. The given characteristic polynomial is 6" order equation and so it has 6 roots. Since the highest power of s is even number, form the first row of routh array using the coefficients of even powers of S and form the second row using the coefficients of odd powers of s.eo: 1 8 2 6 Row-1 e > 2 2 6 on Row2 ‘The elements of s* row can be divided by 2 to simplify the calculations. s 16... Row o-ROW2 % OD, % Te6-6x1 1x8-8x1 1 7 x: 0 o ‘The auxiliary equation is, A = s‘+6s%+8. On differentiating A with respect to s we get, #4 43412 = +12 = 8 Thecsetcirect S ermumd eee 8:4 12 divide by 4On examining the elements of 1" column of routh array it is observed that there is no sigh change. The row with all zeros indicates the possibility of roots on imaginary axis. Hence the system is limitedly or marginally stable, ‘The auxiliary polynomial is, S* 465548 = 0 Let, $2=X X°H6N+8=0 The roots of quadratic are, x = ==AEPE =341 2or-4 ‘The roots of auxiliary polynomial is, sey =4 V=Zand + V=t stivZ, -jvZ.4)2and 2 The roots of a are lying on imaginary axis and the remaining two roots are lying on the left half of s-plane, xiliary polynomial are also roots of characteristic equation, Hence 4 roots RESULT 1. The system is limitedly or marginally stable. 2. Four roots are Lying on imaginary axis and remaining two roots are lying on left half of, s-plane, EXAMPLE 83 Construct Routh array and determine the stability of the, system represented by the jon, s°4s'+2s'+2s'+3s+5=0. Comment on the location of the roots of characteristic equa characteristic equation SOLUTION ‘The characteristic equation of the system is, s°+s"#2s'42s7435+The given characteristic polynomial is 5" order equation and so it has 5 roots. Since the highest power of s is odd number, form the first row of routh atray using the coefficients of odd powers of s and form the second row using the coefficients of even powers of s. es: 1 23 ...Rowl eo: 1 2 5 ...Row2 e: « 2 ROWS é 2en 5 Le 12-201 1x3-5xt € 7 T -(52 +4e+4) ~ eo 2 die Be ow Rowe Zed Remntuse e: 8 oo ROWS = Onletting <->0, we get PT 2 8 Rowd spd y 2 5 Rowe 5. erp oy2 ROW tae gectets Row goes 4 Heital er : ee = eiisi gi 4 ed) * ded Be Her) 9 202 eg OS = “(52 144) Te 5 On observing the elements of first column of routh array, it is found that there are two sign changes. Hence two roots are lying on the right half of s- plane and the system is unstable. The remaining three roots are Lying on the left half of s-plane. RESULT 1. The system is unstable. 2. Two roots are lying on right half of s-plane and three roots are lying on left half of s- plane.EXAMPLE 5.4 By Routh stability criterion determine the stability of the system represented by the characteristic equation, 9s°-20s4+10s-s*-9s-10=0, Comment on the location of roots of characteristic equation. SOLUTION ‘The characteristic polynomial of the system is, 9s*-20s'+10s*-s-9s-10=0. On examining the coefficients of the characteristic polynomial, it is found that some of the coefficients are negative and so some roots will lie on the right half of-s-plane, Hence the system in is unstable. The routh array can be constructed to find the number of roots lying on right half of s-plane. The given characteristic polynomial is 5th order equation and so it has 5 roots. Since the highest power of s is odd number, form the first row of routh array using the coefficients of odd powers of s and form the second row using the coefficients of even powers of S. ies oo Row-1 +) 9 i" : Row2 Niue: soe ROWS Z| Ne oe ROWS s Row-5 s nee ROW =20x10-(-x9 -20x(-9) s:955 -135 97, 955%(-9-(-13.5)«(-20) 958(-10) 955 -293 -10-28:3x(-135)~(~10) x 955 By examining the elements of first column of routh array it is observed that there are three sign changes and so three roots are lying on the right half of s-plane and the remaining two roots are Lying on the left half of s-plane. RESULT 1. The system is unstable. 2. Three roots are lying on right half of s-plane and two roots are lying on left half of s- plane. EXAMPLE 5.5 Determine the range of K for stability of unity feedback system whose open loop transfer K function is G(s) = = SOLUTION i) _ ai) ‘The closed loop transfer function 57> = 5575 — serDeeD = © Marna a =-——*§__. “sts + D+ 24K ‘The characteristic equation is, s(s + 1+ 2)+K=0 s(s443s+2}4K=0_ Si+3s"42s4K=0 ‘The routh array is constructed as shown below.‘The highest power of s in the character! stic polynomial is odd number, hence form the first row using the Coefficients of odd powers of s and form the second row using the coefficients-of even powers of s. - . e ' ' 1 r ' ' ' For the system to be stable there should not be any sign Change in the elements of first column, Hence choose the value of K so that the first column elements are positive. From s° row, for the system to be stable, K> 0 6K From s! row, forthe system to be stable, "50 For 0, the value of K should be less than 6. ‘The range of K for the system to be stable is O Y poot jocus on real 4 i Zadss Figure:5.8 Angles of asymptote: SHEGID hence n=3 and m=0. If q=0, Angles = +82 = +60" Ig=l, Angles = + 180%«s If q=2, Angles = +5 — 460° a0? 2 If get, Angles = + = £60" sum of poles~sum of zeros Centro ‘The centroid is marked on real axis and from the centroid the angles of asymptotes are marked using a protractor. The asymptotes are drawn as dotted lines as shown in fig. Step 4: To find the break away and breakin points ‘The closed loop transfer function of the system with unity feedback is given by,oS) _ 66) __aaaaH RS) tes i = SaGHaIeK ‘The characteristic equation is $(S + 2)(S +4) + K=0 S?+6s°+88+K = 0 K =-S'-65?-8s dK /ds= ~(3s7+12s+8) put dK/ds = 0; -Gs?+12s+8)= 0 12 VTS 2x3 = -0.845 of -3.154 Check for K, When s= -0.845, K = [(-0.845)'+6(-0.845)°+8(-0.845)] = positive and real for K = -0.845, this point is actual breakaway point. When s=-3.154, K = [(-3.154)°+6(3.154)°48(3.154)] = 3.08 + positive, hence K=-3.154 is not a breakaway point. Step 5: to find angle of departure Since there are no complex pole or zero, we need not find angle of departure or arrival Step 6: To find the crossing point on imaginary axis ‘The characteristic equation is given by S346s448s+K = 0 Put s= jo (jo)'+6(j0)+8G0)+K = 0 -jo*-6jo7+8jo+K = 0Figuress.9On equating imaginary part to zero, -w +80 =0 o=8 o= 428 On equating real parts to zero, 607+K =0 K 0° K=6(8) =48 The crossing point of root locus is + j2.8. The value of K at this crossing point is K = 48, (This is the limiting value of K for the y of the system), The complete root locus sketch is shown in fig, The root locus has three branches. One branch starts at the pole at s = - 4 and travel through negative real axis to meet the zero at Infinity. The other two root locus branches starts at s = 0 and s = - 2 and travel through negative real axis, breakaway from real axis at s= -0.845, then crosses imaginary axis at s= j2.8 and travel parallel to asymptotes to meet the zeros at infinity. To find the value of K corresponding to & = 0.5 Given that 5= 05 Let a= cos"! 5 = cos“ 0.5 = 60° Draw a line OP, such that the angle between line OP and negative real axis is 60° ( (60°) as shown in fig 4.23.2. The meeting point of the line. OP and root locus gives the dominant pole, Sd.REVIEW QUESTIONS TWO MARKS Define stability What is stable system? ‘What is absolutely stable system? ‘What is limitedly stable system? Define asymptotically stable system. Define BIBO stability. What is root locus? ‘What is root locus on real axis? pera Abe |. Write the formula for find centroid. 10, Write the formula for angle of asymptote. ‘THREE MARKS 1. Comment the location of roots on s-plane for stability. State Routh Stability criterion. 3. Determine the stahility of the system S*+S* location of the roots of characteristic equation, 4, Draw the typical sketches of Root locus, 5. What is break away and break in points and how can you find it? 38°78 1S-4=0.Comment on the ‘TEN MARKS 1, Construct Routh array and determine the stability of the, system represented by the characteristic equation, $749S°+24S°424S++24S°+24S?423S+15=0.Comment on the location of the roots of characteristic equation. . Construct Routh array and determine the stability of the, system represented by the characteristic equation, S$7+5S°+9S*+9S'+4S*+20S*+36S+36=0. Comment on the location of the roots of characteristic equation. 3. Construct Routh array and determine the stability of the, system represented by the characteristic equation, S°+S5+38'+3843S7+2S+1=0.Comment on the location of the roots of characteristic equation, 4, Determine the range of K for stability of unity feedback system whose open loop transfer function is G(s) = GD 5. Explain the step by step procedure for constructing Root locus. 6. A unity feedback control system has an open loop transfer function G(S) Som Sketch the root locus. 7.A unity feedba Kissa) S(5+05)(S+2) ck control system has an open loop transfer function G(S) = ketch the root locus.