CHAPTER 2 TIME RESPONSE ANALYSIS 21 TIME RESPONSE The time response of the system is the output of the closed loop system as a function of time. It i denoted by c(t). The time response can be obtained by solving the differential equation governing the system. Alternatively, the response c(t) can be obtained from the transfer function of the system and the input to the system. won C8) Gs) The closed loop transfer function, —S) - 9) _ 1e closed loop transfer function, Ris) Ge) HG The Output or Response in s-domain, C(s) is given by the product of the transfer function and the input, R(s). On taking inverse Laplace transform of this product the time domain response, ¢(t) can be obtained. -Me) v2} Response in s-domain, C(s) = R(s) M(s) (2.2) Response in time domain, o(t) = C'{C(s)}= £4 Ris)xMis)} sen(23) where, M(s)=—-S@)_ 1+G(S)H(s) The time response of a control system consists of two parts : the transient and the steady state response. The transient response is the response of the system when the input changes from one state to another. The steady state response is the response as time, t approacties infinity. Ris) Cs) RS) ey > ( ips ST espome 8 fsa \(or Output). | =>. M(s)=-_2)__ [H(s}j«—— G(s)A(s) Fig 2.1 : Closed loop system 2.2 TESTSIGNALS ‘The knowledge of input signal is required to predict the response of a system. In most of the systems the input signals are not known ahead of time and also it is difficult to express the input signals mathematically by simple equations. The characteristics of actual input signals are a sudden shock, a sudden change, a constant velocity and a constant acceleration. Hence test signals which resembles these characteristics are used as input signals to predict the performance of the system. The commonly used test input signals are impulse, step, ramp, acceleration and sinusoidal signals. The standard test signals are, 1. a) Step signal 2. a) Ramp signal 3. a) Parabolic signal b) Unit step signal b) Unit ramp signal —_b) Unit parabolic signal 4, Impulse signal 5. Sinusoidal signal(Gieptar F Time Response Analysis) 2.2 Since the test signals are simple functions for time, they can be easily generated in laboratories. The mathematical and experimental analysis of control systems using these signals can be carried out easily. The use of the test signals can be justified because of a correlation existing between the response characteristics of a system to a test input signal and capability of the system fo cope with actual input signals. STEP SIGNAL The step signal is a signal whose value changes from zero to A at t= 0 and remains constant at A for t> 0. The step signal resembles an actual steady input to a system. A special case of step signal is unit step in which A’is unity. 10 The mathematical representation of the step signal is, W=1 5 t20 =0;t<0 RAMP SIGNAL The ramp signal is a signal whose value increases linearly with time 1) from an initial value of zero at t= 0. The ramp signal resembles a constant velocity input to the system. A special case of ramp signal is unit ramp signal in which the value of A is unity. The mathematical representation of the ramp signal is, r()=At ; t20 0 PARABOLIC SIGNAL In parabolic signal, the instantaneous value varies as square of the (9 time from an initial value of zero at t = 0. The sketch of the signal with respect to time resembles a parabola, The parabolic signal resembles a. 4,54}-------- ++ constant acceleration input to the system. A special case of parabolic signal is unit parabolic signal in which A is unit Qn) > Fig 2.2 : Step signal. 012? Fig 2.3 : Ramp signal. ‘The mathematical representation of the parabolic signal is, 05a : ae ois st WaT 120 Fig 2.4 : Parabolic signal. =0 ;t The unit impulse signal is a special case, in which A is unity. The impulse signal is denoted by &(t) and > te > mathematically it is expressed as, Fig 2.5 + Impulse signal. a(t)=00; t=0 and fomar-a =t40 ed (2.7)2.3 Gaara stoma Trgtneerng > Since a perfect impulse cannot be achieved in practice itis usually approximated by a pulse of small width but with area, A. Mathematically an impulse signal is the derivative of a step signal. Laplace transform of the impulse function is unity. TABLE tandard Test Signals Name of the signal Time domain equation Laplace transform of of signal, r(t) the signal, R(s) A Step A Unit step l | i s A Ramp At = 1 Unitramp t zg 2 A Parabolic = 3 e 1 Unitparabotic = ra Impulse a(t) 1 | 2.3 IMPULSE RESPONSE The response of the system, with input as impulse signal is called weighing function or impulse response of the system: It is also given by the inverse Laplace transform of the system transfer function, and denoted by m(1). Impulse response, m(t) = £" {R(s) M(s)} = £7 [MSP (2.8) GG) = where, MO)= SaSHD R(s) = I, for impulse] Since impulse response (or weighing function) is obtained from the transfer function of the system, itshows the characteristics of the system. Also the response for any input can be obtained by convolution of input with impulse response. 24 ORDEROF ASYSTEM The input and output relationship of a control system can be expressed by n® order differential equation shown in equation (2.9). e ge ee 4 a Boga PO + aa Gar PUL) +e ee PO) +. Farge PO + 2a PCD = bo a(t) at gt d aT Ut) +b, 5 g(t) ct By a(t) + wn f2.9) +b Tara) +b, Gee Ot Ba Ge AE) + Bn (ED where, p(t) = Output /Response ; q(t)=Input / Excitation. The order of the system is given by the order of the differential equation governing the system. If the system is governed by n® order differential equation, then the system is called 2 order system.‘Response 2.4 Alternatively, the order can be determined from the transfer function of the system. The transfer function of the system can be obtained by taking Laplace transform of the differential equation governing the system and rearranging them as a ratio of two polynomials in s, as shown in equation (2.10). Pls) _ bys" + bys™ "+b, baa St bm se2e10) S) ays*+ajs a3" +0, 1843, Transfer function, T(s) where, P(s) = Numerator polynomial Q(s) = Denominator polynomial The order of the system is given by the maximum power of s in the denominator potynomial, Q(s). Here, Q(s) =a, s"-+a, “+ as? + sta sta, Now, n is the order of the system When n = 0, the system is zero order system. When n = 1, the system is first order system. When n = 2, the system is second order system and so on. Note : The order can be specified for both open loop system and closed loop system. | ‘The numerator and denominator polynomial of equation (2.10) can be expressed in the factorized form as shown in equation (2.11). PES) _ (S42 SF 2g )onnelSt Zp) vne(Q11) QS) (S+Py)(S+ Pa).--(S+ Pa) where, 2, 2, are zeros of the system. Pr» Pp» ~ Now, the value of m gives the number of poles in the transfer function. Hence the order is also given by the number of poles of the transfer function. Ti)= P, are poles of the system. Nate : The zeros and poles are critical value, of 8, at which the fimction Ti) attains extreme values 0 or co When s takes the value of a zero, the function T(s) will be zero. When s takes the value of la pole, the function T(s). will be infinite. 2.5 REVIEW OF PARTIAL FRACTION EXPANSION The time response of the system is obtained by taking the inverse Laplace transform of the product of input signal and transfer function of the system. Taking inverse Laplace transform requires the knowledge of partial fraction expansion. In control systems three different types of transfer function are encountered. They are, Case 1: Functions with separate poles. Case 2. : Functions with multiple poles. Case 3: Functions with complex conjugate poles. The partial fraction of all the three cases are explained with an example. Case 1: When the transfer function has distinct poles K Let, M9) = Tp) Gp)2.5 By partial fraction expansion, T(s) can be expressed as, no-—K— B 8 s(S+p,)(S+P2) S S+P) S+Pz ‘The residues A, B.and C are given by, A=T6) x4... B=T6) x G+P) C= Ts) x (+ Pake py == Example Let, T(s) = a NS) ED By partial fraction expansion, T(s) can be expressed as, T(s)= 2 + S(S+1)(8+2) 8 stl st+2 A is obtained by multiplying T(s) by s and letting s — 0. 2 -—2 x} -— a =o gt FD Ho GENEID,» 1x2 B is obtained by multiplying T(s) by (s +1) and letting s = -1. 2 -—2 s(F2),__, --U-1+2) | 2. ALB A=T6) «4, B=T(s)x etn, xGrDl, 2 __ (s+ 1) (+2) C is obtained by multiplying T(s) by (s +2) and letting s = -2. _ 2 2 C=T(s) x (s+2)| “ye x (+2) weal. 21g-—2— 2 Case 2: When the transfer function has multiple poles K Let, T(s)=——————— > Oren eps By partial fraction expansion, T(s) can be-expressed as, kK A_ B c¢_, > T(s)=$ 5 = ++ + sS+p:)(stp2) § StP, +P.) (S+P2) The residues A, B, C and D are given by, A=T@) x44 B=Ts) x (S+P)],_ 5, d com) x Gem, Do G{Me xterm] Example Let, T(s)= 2 . s(st 1) (s+2)7 By partial fraction expansion, T(s) can be expressed as, a()2 —*-- 4+ 8+ So D sS+1)(st2y 8 (+1) (s+2)pera Tine Re sts 2.6 A is obtained by multiplying T(s) by s and letting s = 0. | - _ 2 . 2 | A=TO) x4, Wie GHeDL., 1. B is obtained by multiplying T(s) by (s +1) and letting s 2 2 2 Ba T(s)x(s+D|_,=————— xs]. = | =< > x6 Da Sangre 8 | ser Ce . |. C is obtained by multiplying T(s) by (s +2)’ and letting s = -2. 2 +, x(s+2 s(s+1) (s+2)?* C=Tis)x(s+ »L, D is obtained by differentiating the product T(s) (s +27 with respect to s and then letting s = -2. y 6 4f. 2] 208-1) 2262) +H) as ds | s(6+1) Sty | CC2+n? 5 15. =? __, 4 2,1, sGt(st2e os stl (+2 842 D= Shr x(st a Tis) Case 3 : When the transfer function has complex conjugate poles K Let, T(s) = —————— +p) @ + bs+e) By partial fraction expansion, T(s) can be expressed as, 1%) K A, BstC_ ee(212) “Grp @ tbstc) stp) S+bste The residue A is given by, A=T(s) x (s+p,)) Ls The residues B and C are solved by cross multiplying the equation (2.12) and then equating the coefficient of like power of s. Finally express ‘T(s) as shown below, 11) = A Be (xtyhex+2ay ty? S+p) So +bs+e Let us express, s? + bs, in the form of (x+y), This will require addition and subtraction of an extra term (b/2). Ts) = eg BS A b (sy (ey stp, of?) ae [b 22 22.7 Leontral systems ingineering Example 1 Let, Foe TO" Gap ess By partial fraction expansion, 1%) 1 A Bs+C =} __-4.44 Gt2@ stl) st2 Sesel A is obtained by multiplying T(s) by (s +2) and letting s = -2. 1 2 A=T(s) x(s+2) +3 t t Lea 4 Cyp=241 3 To solve B and C, cross multiply the following equation and substitute the value of A. Then equate the’ like power of s. L A _BstC | (8+2)(# +841) ssl =? S | 1=A(s? +8+1)+(Bs+O) (s+2) 1 2 1 ; ={s+=] +[1- =(P +8+1)+Bs? +2Bs+Cs+2C (> 3) “U ‘) 3 a =(s+0.5) +075 1a 545+ 54s! +2Bs4C5+2C 1 ‘On equating the coefficient of s’ terms, 0= 37 B; -. On equating the coefficient of stems, g=14284C ; 3 _ 3 s+1) 3(8+s+1) a 3 (s+ 0.5) +0.75 2.6 RESPONSE OF FIRST ORDER SYSTEM FOR UNIT STEP INPUT The closed loop order system with unity feedback is shown in fig 2.6. R46 1 cs) RE) Lt] og * {Ts} = 114-Ts I Fig 2.6 : Closed loop for first order system. The closed loop transfer function of first order system, © RG®) If the input is unit step then, 1(4) =1 and R(s)= * 1 1d . itus-domain, C(s)=R@)-——=1_1_ The response in s- domain, Cs) =RS)T 55 = Say ayGiapter E:T Responso Analysis > 2.8 By partial fraction expansion, L | -T_,J . “ +4) sti T WOT AHCG) xg, 1 1 OH 8)=— ——T | fers} Soe sta ‘The response in time domain is given by. =o Yew@ =H te 7 (2.13) 5 T) ‘The equation (2.13) is the response of the closed loop first order system for unit step input, For step input of step value, A, the equation (2.13) is multiplied by A. st For closed loop first order system, Unit step response = | Step response = ae] t When, 0, c(t) When, 1T, e(t) = When, 2T, oft) When, 3T, ot) =1-e When, AT, c(t) When, t = ST, c(t) When, t = », o(t) =~1-e*=1 Here T is called Time constant of the system, In a time of ST, the system is assumed to have attained steady state. The input and output signal of the first order system is shown in fig 2.7.2.9 x(t) + oft) 4 ¥ 1=0 t oO T 2T 3T AT » t Fig 2.7a : Unit step input. Fig 2.76 : Response for Unit step input. Fig 2.7 : Response of first order system to Unit step input. 2.7 SECOND ORDER SYSTEM ‘The closed loop second order system is shown in fig 2.8 Ms) R(s) 2 cs) , \s?+ Xo,s+0% ° Fig 2.8 : Closed loop Jor second order system. The standard form of closed loop transfer function of second order system is given by, RS) £8), oy (2.14) R@) 9 +2fo,s+0, where, @, = Undamped natural frequency, rad/sec. € = Damping ratio. The damping ratio is defined as the ratio of the actual-damping to the critical damping, The response c(t) of second order system depends on the value of damping ratié. Depending on the value of ©, the system can be classified into the following four cases, Case I Case 2 Undamped system, e=0 Under damped system, = 01, 5 =to9toe6! “ine system isoverdamped = (2.19) When0<<1, 8,5) =-Co, 20,0 —1 =o, to,yCD 0-0) =o, 20-1 fI-2 =o, +0, y1-F =e, jms OA complex conjugate £345 | the system is underdamped siee(2.20) where, 4 =0,y1-0 Here ©, is called damped frequency of oscillation of the system and its unit is rad/see. RESPONSE OF UNDAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, Cs) _ o, Ris) = +200,5+0, For undamped system, z= 0. - OS) _ On Ro) +0 222) When the input is unit step, 1(()= I and R(s)= + “ + Fors” Fret cftt=2 | £fc0s ot}== 25 . 5 | ¢ +0 s ‘Time domain response, o(t)= C'{C(s)} = £" 1 I-cos@t ot (2.24)cs ny Dh on eden apc ew een nenn eee eens, 1 Ip--f---------\e- eee kee 9 * 9 > Fig 2.9.0; Input. Fig 2.9.b : Response. Fig 2.9 : Response of undamped second order system for unit step input. Using equation (2.24), the response of undamped second order system for unit step input is sketched in fig 2.9, and observed that the response is completely oscillatory. f | Note : Every practical system has some amount of damping. Hence undamped system does not exist in practice. The equation (2.24) is the response of undamped closed loop second order system for unit step input. For step input of step value A, the equation (2.24) should be multiplied by A. | ~ For closed loop undamped second order system, Unit step response = 1— cos @,t Step response = A(1 — cos «,1) 2.7.2___ RESPONSE OF UNDERDAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, cs) o, Ris) © 22 sao Ris) s*+2ta,s+o%, For underdamped system, 0 < ¢ <1 and roots of the denominator (characteristic equation) are complex conjugate. The roots of the denominator are, s=-Co, +0,J@=1 Since¢<1, ¢ isalso less then 1, andso 1- Gre B=(s+o,)?x co), Ze = fo+0,7 xa] =et) =1-e*"(1+0,t) The equation (2.31) is the response of critically damped closed loop second order system for unit step input. For step input of step value, A, the equation (2.31) should be multiplied by A. :. For closed loop critically damped second order system, Unit step response = 1-6"*"(I+0,1) Step response = ali- eels. o,t)] Using equation (2.31), the response of critically damped second order system is sketched as shown ’n fig 2.11 and observed that the response has no oscillations. rt) Oa 1 It ay 0 Fig 2.11.a + Input. Fig 2.11.b : Response. Fig 2.11 : Response of critically damped second order system for unit step input. 2.7.4 _ RESPONSE OF OVER DAMPED SECOND ORDER SYSTEM FOR UNIT STEP INPUT The standard form of closed loop transfer function of second order system is, Cs) on Ris) s+ 2fo,s+05 For overdamped system ¢ > 1. The roots of the denominator of transfer function are real and distinct. Let the roots of the denominator be s,, s,- SS =—Go, £0, 6 [ge to/e-1| swn(232) Let ,=-sands,=-s .s)=G@,-@,y6°-1 82=00, +0, '=1 ‘The closed loop transfer function can be written in terms of s, and s, as shown below. CO) ee (2.35) RG) S+%oseor (Gty)G*S) ZO, 8+ 0, ,215 For unit step input r(t) = 1 and R(s) = 1/s. a 2 (3) = R(s)- 22 = —_2 ___ = ROE say ee (s+5) (+8) By partial fraction expansion we can write, 2 c= oO, uA, Be ae S(+5) GH) $ Sts, SHS @? A=sxC(5)jco= 8x" SESE), 8S = on [00 -eayeP=1 [Beaten 2 On = Ba (s+8) C8), = s+), ~+5) = oR - ony 5, |Co, +0, 407-1400, +onye?=1 20, =1) 5 2yer=1 C= Cs) x(s+ 5) 0 BTS ait o(t) = 1-8 fe wWe-1\ Ss A236 where, s,=60,-@,y@-1 $= SO, t+ OnyS ‘The equation (2.36) is the response of overdamped closed loop system for unit step input. For step input of value, A, the equation (2.36) is multiplied by A. -. For closed loop over damped second order system, = where, s,=C0,-@,yC?-1 = 60, +0, 16-1 Unit step response = 1- Step response0 t * t Fig 2.12. : Input. Fig 2.12.5 : Response. Fig 2.12 : Response of over damped second order system for unit step input. Using equation (2.36), the response of overdamped second order system is sketched as shown in fig 2.12 and observed that the response has no oscillations but it takes longer time for the response to reach the final steady value. 2.8 TIME DOMAIN SPECIFICATIONS The desired performance characteristics of control systems are specified in terms of time domain specifications. Systems with energy storage elements cannot respond instantaneously and will exhibit iransient responses, whenever they are subjected to inputs or disturbances. The desired performance characteristics of a system of any order may be specified in terms of the transient response to a unit step input signal. The response of a second order system for unit-step input with various values of damping ratio is shown in fig 2.13. e(t), ro Fig 2.13.0: Input. Fig 2.13.b : Response. Fig 2.13 : Unit step response of second order system. The transient response of a system toa unit step input depends on the initial conditions. Therefore to compare the time response of various systems it is necessary to start with standard initial conditions. The most practical standard is to start with the system at rest and so output and all time derivatives before +=0 will be zero. The transient response of a practical control system often exhibits damped oscillation ‘before reaching steady state. A typical damped oscillatory response of a system is shown in fig 2.14. ‘The transient response characteristics of a control system fo a unit step input is specified in terms of the following time domain specifications. 1. Delay time, t, 2. Rise time, t, 3, Peak time, t, 4. Maximum overshoot, M, 5, Settling time, t,Allowable error 2% or 5% out & w t Fig 2.14 : Damped oscillatory response of second order system for unit step input. ‘The time domain specifications are defined as follows. 1, DELAY TIME (t,) : It is the time taken for response to reach 50% of the final value, fo the very first time. It is the time taken for response to raise from 0 to 100% for the ver first time. For underdamped system, the rise time is calculated fron 0 to 100%. But for overdamped system it is the time taken by the response to raise from 10% to 90%. For critically damped system, i is the time taken for response to raise from 5% to 95%. It is the time taken for the response to reach the peak value the ver first time. (or) It is the time taken for the response to reach the peal overshoot, M,, 4.PEAKOVERSHOOT(M,)_: Its defined as the ratio of the maximum peak value fo the final value where the maximum peak value is measured from final value. Let, o(s0) = Final value of c(t). o(tp) = Maximum value of e(t). (ty 2. RISE TIME (f) 3. PEAK TIME t) Now, Peak overshoot, M, = 2" (23 oa) % Peak overshoot, %M, = a 100 238 ees 5, SETTLING TIME (t) : It is defined as the time taken by the response to reach and stay with a specified error. It is usually expressed as % of final value. The usua tolerable error is 2 % or 5% of the final value. EXPRESSIONS FOR TIME DOMAIN SPECIFICATIONS Rise time (t) ‘The unit step response of second order system for underdamped case is given by, Att=t, c(t) = c(t) =1 (Refer fig 2.14).nim oo ce sin(o st, +0)=1 | Note : On constructing right | y | angle triangle with ¢and Ji te + sin(o gt, +0) =0 | we get Since ~ eS" #0, the term, sin (@,t, +6) =0 tan @= When, 6=0,2,2n,37.., sin 9-0 — [ae] | Rise Time,t,-"=2 | vonf239) | | 2 — ; Damped frequency of oscillation, 0-0, J1-C (refer note) «Rise time, t, = 2 inser | Onl (2:40) ne er Note: 6 or ta BE should be measured in radians. Peak time (t,) To find the expression for peak time, |, differentiate o(t) with respect to t and equate to 0. 0 .. a be, FeO, The unit step response of under damped second order system is given by, Seat Differentiating c(t) with respect to t. c(t) sin(ogt +0) (—Co,) sin(@,t +0) { cos(o,t + O)o, gy = 2 tosh a ae 2h ae oo sinoas) Toe e*"" cos(ot +8) o,§ = ir {6 sinoyt+8)~4 i cos(o,t+€)} ®, *at[ cos sin(oyt+0)—sind cos(o,t+@)] (refer note) e="[sin(a 4t +9) cos - cos(o,t + 6)sinO]satiny oMntfsin((o,t +6)-0)] x! sin(o,t) Vi-g d py =O Note : On constructing right angle | triangle with {and f\-C , we get | oa ot? sinogt,) =0 mare owe get . ce roa 1-g sind= 1-0? 1 (le Since, oS"? 20, the term, sin(a,t,)=0 | cosa <6 A |s | When 6 =0, n, 2x, 3x, sing =0 1 @4ty=m ae | «Peak time, t, = see2.41) Ll a ‘The damped frequency of oscillation, og = =] one ~. Peak time, ty Peak overshoot (M,) ty) — A) 9g wn(2.43) %Peak overshoot, %M, = e() where, c(t.) = Peak response at t= t, . (2) = Final steady state value. ‘The unit step response of second order system is given by, eet e(t)=1- sin(@,t +8) Vv Att-0, off) = e(#) =1 sin@@t+@) = 1-0 = 1 == Att=tp, o(t)=clty)= sin(ojt, +8) 1-¢ Og =O VIS) sin (nx +6)~—sind eee [Note : On constructing right angle triangle with Cand 1-2", we get is | | sin® =2.20 Percentage Peak Overshoot, %M, (2.45) Settling time (t) The response of second order system has two components. They are, ett 1. Decaying exponential component, fe : yi- 2. Sinusoidal component, sin(og +9) In this the decaying exponential term dampens (or) reduces the oscillations produced by sinusoidal component. Hence the settling time is decided by the exponential component. The settling time can be found out by equating exponential component to percentage tolerance errors. For 2 % tolerance error band, at t = t, For least values of € , eS*e's = 0,02 On taking natural logarithm we get, ~Co,1,=1n(0.02) =>. —Ca,t=-4 3 For the second order system, the time constant, T= Settling time, t,=—--=4P (for 2% error) fo For 5% error, os" = 0.05 On taking natural logarithm we get, —Cogt,= (0.05) => —Lo,t, sseeel2AT) (for 5% exror) — 3 <, Settling time, t, =~ =3T is So x In general for a specified percentage error, Settling time can be evaluted using equation (2.48). in@aerror) meron) | 48) fo, T <. Settling time, t2.21 EXAMPLE 2.1 tian the response ofunty feedback system hose open oop tensor funcion’s G(s) = = 4 ay andwhen the input se isuni'step SOLUTION Rs) cs) ‘The closed loop system is shown in fig 1. y The closed loop transferfunction, 2S) - Sts) 4 Ris) 1+ Gis) Fig 1: Closed loop system. GS) __s(s+5)___s(s+9) ___ 4 A Ris) 4 S(St5)+4 g(gt5)+4 © gs? 45s+4 (824) (8+1) i 4 hi = —_—— Thetesponseins-demain, Os) RS) =p Since the ifiput is unit step, reat 0) = By partial fraction expansion, we can write, 4 ALB C(s)= stt(st4) ssl ss 4 _4 1 +0644), 1x4 A=C(s) x g} SGD... 4-4) 3 The time domain response c(t) is obtained by taking inverse Laplace transform of C(s).. 144 1 14 , in, t= cic c(t 42 1 fesponse in time domain, e()= £"{C(s)} = £ 5 3 se1°3 ova} RESULT Response of unity feedback system, c(t) = EXAMPLE 2.2 . A positional control system with velocity feedbacks shown in fig 1. Whatis the response ofthe system for unit step input. SOLUTION Ris) (“00 cs) (s(s+2) ‘The closed loop transfer function, Gis) __Gls)_ lo.tse1 Ris) 1+ G(s) H(s) , —— 100 Fig 1 : Positional control system. Given that, G(s)=——— and H(s)=0.1s +1 ser2) 100 ._ 100 Cs) _ s(s+2) _ 3(s#2) 100 ____100 Rs) SET 2) +100 242s+10s+100 s? Ris) uf 100 (9 1541) s(S +2) +100 S425+108+100 s?+12s+100 (s(s+2)chapter: ‘Response Analysis _) 2.22 Here (s?+ 12s + 100) is characteristic polynomial. The roots of the characteristic polynomial are, W122 Ji44— 400 _ -122)16 _ 2 207 The roots are complex conjugate. The system is underdamped and so the response of the system will have damped oscillations. Sy 8: -6+j8 The response in s-domain, C{s)=R(s) 0 _ 37 +125 +100 Since input is unit step, R(s) = 2 100 100 4 =€(s)=+ 100 _ (= 5 Fy iass (00 s(8°+ 128+ 100) By partial fraction expansion we can write, = 100 A, Bs+C “3(6?#128+100) 5 s?+128+100 The residue Ais obtained by multiplying C(s) by s and letting s = 0. 100 100 +12s+100|,., 100 ‘The residue B and C are evaluated by cross multiplying the following equation and equating the coefficients of like power of s. A=C(s)x 8. 100 A Bs+C Giets2s+100) 8 s+ 1254100 100 = A(s? +128 +100)+(Bs+C) s 100 = As* +12As+100A+Bs*+Cs On equating the coefficients of s*weget, 0=A+B ‘On equating coefficients of s we get, O=12A+C 2Os) = 14,78 s 1s? = 1__s+6+6 vom 8 §?4125+100 8 8°4125+36+64 5 (5+6)°+8 -1__s6____# os 1 _st6_ 8 8 “s (s+6 +8 (5 +6) +8" Ss (8+6)+8? 8 (s+6) The time domain response is obtained by taking invorse Laplace transform of C(s).. - f1__s+6 6 8 =o =o¥--_* > _- = Time response, eft) = £ {ows} = ene ene =1-0%* coset eo sin8t = -(f sin8t + cosét “The result can be converted to another standard form by constructing right angle triangle with ¢ end 1—¢? . The | damping ratio Sis evaluated by comparing the closed loop transfer function of the system with standard form of second order | wansfer function. Os) 100 “Rs) s?+2f@,+0, 87 +128+100 On comparing we get, @2 oi, =12 12 2G, = 102.23 Constructing right angled triangle with Cand J1—C? we get, 08 = 08 ; cos®= 06; tand=— sind =08 ; cos no= GE = 59°= 53° x rad = 0.925 cad. 810 10 - Derfsindt «06 +cosst « 08] =1-1.25 e*[sin8t coso + cos8t sind] [ Note -6 is expressed in radians. / | «. Time response, c(t) = 1-e"°| Ssinat~ cost) =1-¢ #29] Ssinat +S coset ls =1-1.25 e*[sin (8t+6)]=1- 125 © sin(8t +0.925) RESULT ‘The response in time domain, at =7=01 sin ont] or cft)=1-1.25 6% sin(Bt+0,925) EXAMPLE 2.3 ‘The response of a servomechanism is, c(t) = 1 + 0.2 e*- 1.2 e** when subject to a unit step input. Obtain an expression for closed loop transfer function. Determine the undamped netural frequency and damping ratio. SOLUTION Given that, c(t) =1+0.2e-1.2e"" On taking Laptace transform of c(t) we get, wi. 1 go 1 _ (8 +60) (810) 40.28 (8+ 10)- 1.28 (+60) Ga)=5+02 (+60) 29 's (S60) (8+ 10) 287 +708 + 600+0.2 s7+2s-12s?-72s » 600. _i 600 8 (6 +60) (s+ 10) ‘s(s+60)(s+10) $s (8+60)(s+10) ‘Since inputis unit step, R(s)= /s. 60 600 © AS)=FS) 560) (5>10) |) 575 705+ 600 . The closed loop transfer function of the system, ee oe we ‘The damping ratio and natural frequency of osciliation can be estimated by comiparing the system transfer function with standard form of second order transfer function. Os) _ EH 600 “RS) s?+2f0,8+0% 87+708+600 ‘On comparing we get, 2 = 600 70 600 = ' — oe 600 =24.49 rad/sec | eas 43 RESULT : as) 600 The closed loop t1 sfer function of the system, ——- = —;—————_— e P transfer tu Sy Ris) _ s?+70s+600 Natural frequency of oscillation, «,= 24.49 radisec. Damping ratio, g2.24 EXAMPLE 2.4 ‘The unity feedback system is characterized by an open loop transfer function G(s) =K/s ( +10). Determine the gain K, so thatthe system will have a damping ratio of 0.5 for this value of K. Determine peak overshoot and time at peak avershoot for unit step input. SOLUTION RE) 6 Gey Cs) The unity feedback system is shown in fig 1. x tion AS) -_GS)_ The closed loop transfer function SE = Fig 1 : Unity oedback system. Given that, G(s)=/s (s +10) Kk _Os)_ 3+) KK “RIS) 44K “'s(st10)+K s?+10s ss +10) The value of K can be evaluated by comparing the system transfer function with standard form of second order transfer function. . fs) oR K ” Res) 2a,s+os s*+10s+K On comparing we get, oz=K 2a, =10 K=100 PutG=0.5 anda, =/K ©, = 10rad/sec 2.2x0.5x¥K =10 vK=10 The value of gain, K=100, Percentage peak overshoot, %M, = 0°" 100 = 2TH85? 100 = 0163x100 = Peaktime, = oe whe oe RESULT The value of gain, K = 10 Percentage peak overshoot, cM, = 163% Peaktime, t, = 0.263sec. EXAMPLE 2:5 ‘The open toop transfer function of a unity feedback systemis given by G(s) = K/s (sT + 1), where K and T are positive constant. By what factor should the ampitfier gain K be reduced, so that the peak overshoot of unit step response of the system is reduced from 75% to 25%. cl SOLUTION RO 6 sa—-“? The unity feedback system is shown in fig 1. } Gl ‘The closed loop transfer function, oot = cal Fig 1 : Unity feedback system.2.25 CBeartrol Systems nginvering >) Given that, G(s) = Kis (ST +1) Os)_ _KisT+1) KK KIT “RS) A K/s6THD SGT+)K STH54K , 1, Expression for ¢ and «, can be obtained by comparing the transfer function with the standard form of second order ‘transfer function, - Os) es Ris) 20,8+o5 On comparing we get, o2=KIT VKIT ‘The peak overshoot, M, is reduced by increasing the damping ratio. The damping ratio Cis increased by reducing the gain. When M,=0.75, Let¢=¢, and K=K, When M, = 0.25, Lett =, Peak overshoot, M, = @-°"" ‘Taking natural logarithm on both sides, in M, (Pn? On squaring we get, (in M,)* = 2 On crossing multiplication we get, On equating, equation (1) & (2) we get, 2 1 (im M,)? 2 2 (Fe) ayy AT Fn ME 2 etm Mj? er2—? 1 ATU MY (00 My? Can Mh)? = PHP 2 2 2 = cn 2 M2 +(In M,) (in M)° = Con® +6°(in M,) “erin (in M,)? =¢? [x2 +(in M,)*] (in M,) Pep PO onl 7+ (in MY o puget, et sn) KT 4KT 2 2 Ke =07 x? +(In0.75)? _ 9952 _ 30.06 Wen KS Ks My=0.78, * Ke= "er Gng7? ~ oaait 7 = - — 2 +(in0.25) _ 1179 _ 153 When, KK M,=0.25, «. Ke= Se a eae _ Ki _ (7) 30.08 _ “Ke (7) 1.53 196 K,=19.6K, (or) Keats ky To reduce peak overshoot from 0.75 to 0.25, K should be reduced by 19.6 times (approximately 20 times). RESULT The value of gain, K should be reduced approximately 20 times to reduce peak overshoot from 0.75 to 0.25.2.26 EXAMPLE 2.6 positional control system with velocity feedback is shown in fig 1. Whatis the response c(t) to the unit step input. Given that = 0.5. Also calculate rise time, peak time, maximum overshoot and settling time. SOLUTION cls) ats) Ris) cs) s s ‘The closed loop transfer function, = ; - T Given that G(s) = 16/s(s+ 0.8) and H(s)=Ks +1 Ks Fig 1 16 5 (5+08) 16 _ 16 s ($+ 0.8) + 16(Ks +1) sey Wet) 16 16 “2085+ 10K +16 +(08+16K) $+16 ‘The values of K and a, are obtained by comparing the system transfer function with standard form of second order transferfunction, of 16 + 260,8+ 0% (0.8+ 16K) $+16 On comparing we get 2 =16 0.8+16K = 260, 2o,-08 _ 2x0.5x4-0.8 _ 16 6 16 16 87+ (08+ 16x02)s+16 57+ 48+16 Given that the damping ratio, ¢= 0.5. Hence the system is underdamped and so the response of the system will have damped oscillations. The roots of characteristic polynomial will be complex conjugate. _—s__ 8445416 * 0, = 4 rad /sec Ke 0.2 ‘The response in s-domain, C(s) = Rls) For unit step input, R(s) = 1/s. creat 16 16 OSS Fe yasrte s 4s+16) By partial fraction expansion we can write, 16 A Bs+C Os) = 3 (e+ 45+16) 8 s?+4s+16 ‘The residue A is obtained by multiplying C(s) by sand lettings=0. 16 16 A=Cls)x 4 wren, 6 The residues B and C are evaluated by cross multiplying the following equation and equating the coefficients of like powers ofs. 16 A, Bs+C s@14s116) 6 s?+45+16 On-ross multplicationweget, 16=A(s?+4s+16)+(Bs+C)s 16= As?+4As+16A+Bs?+Cs2.27 ‘On equating the coefficients of s* we get, 0= A+B ‘On equating the coefficients of swe get, O=4A+C -.C 1, -s-4_ 1 +4 8 s'445+16 §° s?445+4+12 1 s+2+2 1 s+2_ 2 vi2 Ss (s+2)°-+12 8 (8+2)°+12 12 (+2)?+12 The time domain response is obtained by taking inverse Laplace transform of C(s). ‘The response in time domain, < Cs)= ct) = £C(5)} = of s+2 2 V2 | 8 (S+2°+12 fi2 (+2) +12 +107 eas VIB -37p @ snd pena 1) +o08(Vi2 v| ‘The result can be converted to another standard form by constructing ightangle triangle with Zand /1—? « On constructing right angle triangle with & and /1—¢ weget, FI Ke ice | sind =0866=V3/2; cos@=05 =1/2>; tand=1732 28 | 0x66 | | g=05 I +. @=tan"'1732 = 60°= 1047 rad +. The response in time domain, ag=si0el can Bed ren Btx =t-e leone teos8+cosyi2 tsino] ‘Note: 9 is expressed in radians. # e*{sin( it +0)]=1- Zo fsin( SB t+ 1047] V3 Damped feauency = 0,102 = 4108? = 3.464 rad/ sec n-8_x-1047 - Rise time, 1, = = 0.8048 ise ay ange 7 05046 sec Peak time, t, = =~ = 0907 sec a, 3464 uM ae -O5e2 % Maximum =a =o =etrost - =163% ova =8 5" x100= e%F9% x 100 = 0.163 x 100 = 16.3% Time constant, T=-*-= 5On For 5% error, Settling time, t, For 2% error, Satling time, t= RESULT 1 The time domain response, c(t) =1- [en 1) +cos(/72t) | Zerfsn liz v4 1047)] (on) eft)= 12.28 Rise time, 1,= 0.6046 sec Peak time, t,=0.907 sec % Maximum overshoot, %M,= 16.3% Setting time, t= 1.5se¢, for 5% error =2sec, for2%eror EXAMPLE 2.7 Aunity feedback control system is characterized by the following open loop trensfer function G(s) = (0.48 +1)/s(s +06). Determine its tansient response for unit step input and sketch the response. Evaluate the maximum overshoot and the corresponding peak time. SOLUTION Ris) 1+ Gis) His) The closed loop transfer function, S18). 6) (6 Given that, G(s) = (0.4.5 +1)/s(s +0.6) For unity feedback system, H(s) = 1. O4s+1 As) __ Gls) 55 +06) _ O4s+1 “RS 1+G(s) 04837 ~5(s+06)+048+1 4S*0.6) O4s+t O4s+4 +068+048+1 s?4841 Tho domain response, C(s)=Ris) x 248*1 s4sei For step input, R(s) = 1/s. 1 04841 04s+4 O82 Pygit aetesiD By partial fraction expansion C(s) can be expressedas, Seed The residue A is solved by multiplying C(s) by s and letting s = 0. O4stt wae) x sh ge SAS ay 18) « s, 87454 leg The residues B and C are solved by cross multiplying the following equation and equating the coefficients of ike powers ofs. O4s+1 A Bs+C as?+s+) Ss siyse1 On cross multiplication we get, 04841 =A(s?+8-41)+(BS+C)s 0.45+1=Ast+As+A+Bst+Cs On equating coefficients of s* we get, 0 On equating coefficients of swe get, 0.4=A+C 1 -s-06 1 $+0.6 2. C(8)= ase ot Pests se8+0254075 (8? +2x05s+057)+0.75 1 s+05+01 1 s+05 Ot 0.75 8 (84054075 8 (6+05)+0.75 VO75 (840.5) 10.752.29 The time domain response is obtained by taking inverse | aplace transform of C(s). :. The response in time domain, = cies} = ctt-—s#05____04 __v076 __ ot) = EOS} -«f (e+ 05) +075 Pee cy on| o8* coe 075 t= eS sin 0.75 t =t-e4 fats sin( {0.75 1) cos(V0.75 t)] The transient response is the part of the output which vanishes as t tends to infinity. Here as t tends to infinity the ‘exponential component e%*'tends to zero. Hence the transient response is given by the damped sinusoidal component. ‘The transient response of c(t)=e™ s|a1 155 sin(/0.75 t)+cos(VO.75 D ‘The value of Cand «,, can be estimated by comparing the characteristic equation of the system with standard form of second order characteristic equation. ot), 9+ 260,8+0, 1163 |----+ On comparing we get, 1 N AN ALA ‘abt | 2o,=1 cs e,=tradisee | =sb=t-085 ° "20, 2 ae o 1,=3 628 sec \ Meximum overshoot, M, =e" =e=95" - 0163 Fig 1 : Response of under damped system. % Maximum overshoot, %M, = M, x 100 = 0.163 x 100 = 16.3% z x x Peak time, t, = — === = 3.628 sec Os oto? tx Vi_05* “The response of the system is underdamped and itis shown infig 1. RESULT Transientresponse of the system, c(t) = [0.11 85sin(VO75 t)+cos(/075 » ‘% Maximum peak overshoot, %M, = 163% Peak time, t, 3.628 sec EXAMPLE 2.8 ‘ Aunty feedback control system has an amplifier with gain K, = 10 and gain ratio, G(s) = 1/s(s +2) in the feed forward path, A derivative feedback, H(s) = sk, isintroduced as a minor loop around G(s). Determine the derivative feedback constant K, so that the system damping factoris 0.6. SOLUTION The given system can be represented by the block diagram shown in fig 1. Ris), Here, K,=10; Gls) = =a and H(s)=sK, +) 1K 5) ‘The closed loop transfer function of the system can be obtained by block diagram reduction techniques. cs)2.30 ‘Step 1: Reducing the inner feedback loop. RG), Gs) ce) 14+ G(s) H(s) és lel ¥ 1 Qs) FQ) 1 1 = 1 1+G(s) HS) 4,7 sk, S8+2)48K, 87+2548K, $°+(2+K,)S s(st2) Ris), 1 cs) +) ia 7a (2+K)S ~ ‘Step 2 : Combining blocks in cascade I) © cs) wens Lo Step 3 : Reducing the unity feedback path 10 f 1 R@),| s=2+K)s_ | OS) Ris), 10 cls) >| > Yaskysst0 | 0 874 (2+K)s ts) _ 10 The closed loop transfer function, == Rs) s?+(2+K,) $410 “The given systemis.a second order system. The value of K, can be determined by comparing the system transfer function with standard form of second ordor transfer function given below. Standard form of Cs) o2 (2) Second order transfer function) Fis)" Salas res (On comparing equation (1) & (2) we got, @2=10 24K, =2o, 2 @, = ¥10 = 3162 rad/sec |. K, =260,-2 =2x0.6x3.162-2=1.7044 RESULT The value of constant, K, = 1.7944 EXAMPLE 2.9 ‘Aunity feedback control system has an open loop transfer function, G(s) = 10/s(s*2). Find the rise time, percontage overshoot, peak time and settling time for a step input of 12 units. SOLUTION | Note: The formulae for rise time, percontage overshoot and peak time remains same for unit step and step input.2.31 CGaritrol systems Engineering > The unity feedback system is shown in fig 1. Gs) __ Gs) Ris). Cis) ‘The ciosed loop transfer function, FS) Tas te) The closed loop transfer function, Given that, G(s) = 10/s (5+2) Fig 1 : Unity feedback system 10 . Cs) _s(s+2) 40 10 wat Ris) 4, 10 5s (62)710 S74 28110 3(8+2) The values of damping ratio ¢ and natural frequency of oscillation @, are obtained by comparing the system transfe® function with standard form of second order transfer function. ‘Standard form of Os) _ oe Second order transfer function{ Ris) s+ 2f0,8+02 On comparing equation (1) & (2) we get, 02 =10 2oq=2 + ©, = 10 = 3.162 rad sec aVi-0316% Tg 1248 rad y= ony1-C? = 3162V1-0316" =3 rad/ sec Rise time, t 249 0.93 sec “a 03166 aa ae Percentage overshoot, %M, = e¥* x100=e!F 99 x 100 = 0.3512%100= 35.12% Peak overshoot = ee x12 units = 4.2144 units Peak time, === 1047 sec oe 3 i Time constant, T=—L- = {sec Go, 0316x3162 :. For5% error, Settling time, 1, = 37 =3 sec For2% error, Setlling time, t, =4T=4 sec RESULT Risetime, 0.63sec Percentage overshoot, %M, = 35.12% Peak overshoot 4.2144 units, (fora input of 12 units) Peaktime, t, 1.047 sec Settling time, t, 3secfor 5% error 4 sector 2% errorEXAMPLE 2.10 fe gt 64e Where cis the displacement of the output shaft, ris the displacement of the input shaft and e = undamped natural frequency, damping ratio and percentage maximum overshoot for unit step input SOLUTION ‘The mathematical equations governing the system are, ‘Aclosed loop servo is represented by the differential equation ~¢, Determine oe ) Spek ate este (2) Pute=r-c in equation (1), o E55 euro) soeed (3) Let £{c}= C(s) and £47} = R(s) (On taking Laplace transform of equation (3) we get, 8? C(s) + 85 C(s) = 64 IR(S) -C(s)] <. S#C(s) +85 C(s) + 64 C(s)= 64 Ris) (s? + 8s + 64) C(s) = 64 R(s) 64 8?+88+64 (4) The ratio C{s)/R(s) is the closed loop transfer fnction of the system. On comparing the system transfer function with standard form of second order transfer function, we can estimate the values of Cando, Standard form of Xs) CH ‘Second order transfer function! R(s) 5° + 2o,s+02 On comparing equation (1) & (2) we get, Percentage peak overshoot, %M, = €' «100 =16.3% RESULT , ‘Undamped natural frequency of oscillation.a, = 8 rad/sec ‘Damping ratio, ¢ = 05 Percantage peak overshoot, %M, = 183%2.33 29 TYPE NUMBER OF CONTROL SYSTEMS ‘The type number is specified for loop transfer function G(s) H(s). The number of poles of the loop transfer function lying at the origin decides the type number of the system. In general, if N is the number of poles at the origin then the type number is N. The loop transfer function can be expressed as a ratio of two polynomials in s. x PO. pt Gm) Gm) G =K— =k + OHO BOG Sep) +P.) G+ PD) are zeros of transfer function are poles of transfer function see (2.48 where, 2), Z) Zp Py Pp K = Constant N= Number of poles at the origin The value of N in the denominator polynomial of loop transfer function shown in equation (2.49) decides the type number of the system. IFN = 0, then the system is type — 0 system IfN = |, then the system is type ~ 1 system If N = 2, then the system is type — 2 system If N = 3, then the system is type — 3 system and so on. 2.10 STEADY STATE ERROR The steady state error is the value of error signal e(¢), when t tends to infinity. The steady state error is a measure of system accuracy. These errors arise from the nature of inputs, type of system and from non linearity of system components. The steady state performance of a stable control system is generally judged by its steady state error to step, ramp and parabolic inputs. Consider a closed loop system shown in fig 2.15. Let, R(s) = Input signal RO AF) OS) E(s) = Error signal C{s) H(s) = Feedback signal cone Ll Cs) = Output signal or response Fig 2.15. The error signal, E(s) = R(s) - C(s) H(s) The output signal, C(s) = E(s) G(s) On substituting for C(s) from equation (2.51) in equation (2.50) we get, EG) = R® - E®) G9] E(s) + E(s) G(s) His) = R@) BG) [1 + GG) HO] = RO) = E2—_2O__ smi 252 28) GG) HO) ‘ Let, e(t) = errorsignal in time domain.211 2.34 cette CYriop= ot} R_| 42.53 = CPO frag HO 259) Let, e,, = steady state error. ‘The steady state error is defined as the value of e(t) when t tends to infinity. Lt e(t) (2.54) The final value theorem of Laplace transform states that, If, FG)=L{f(} then, Lt f® = Les kG) wn d255) Using final value theorem, ‘The steady state error, ¢,,= Lt. e(t)= Lt sE(s)= Lt Sis) 2.56) +0 1+G(s) HG) STATIC ERROR CONSTANTS: When a control system is excited with standard input signal, the steady state error may be zero, constant of infinity. The value of steady state error depends on the type number and the input signal. ‘Type-0 system will have a. constant steady state error when the input is step signal. Type-1 system will have a constant steady state error when the input is ramp signal or velocity'signal. Type-2 system will have a constant steady state error when the input is parabolic signal or acceleration signal. For the three cases mentioned above the steady state error is associated with one of the constants defined as follows, Positional error constant, Ky = Lt, G(s) H(s) 2.57) Velocity error constant, = K, = Lt s G(s) H(s) A258) Acceleration error constant, K, = Lt, G(s) H{s) (2.59) ‘The K,, K, and K, are in general called static error constants. 2.12 STEADY STATE ERROR WHEN THE INPUT IS UNIT STEP SIGNAL sR) Steady stat ry Og = Lt) ly State eIF0r, ess = 1, 1+G(s) H@) When the input is unit step, R(s) =1/s L SS 1 1 ee ee (2.60) 0 TFGOHO 90 1+GG)HO 1+ Lt G)HG) 1+K, where, K, = Lt G(s) H(s) so The constant K, is called positional error constant. Type-0 system _ ae RSPAS HE) SH) Kp= LOW) HO) i Seep (s+ Po) (S+Ps ne, = constant 1 1+K, Hence in type-0 systems when the input is unit step there will be a constant steady state error.2.35 Type-1 system (S424) (+22) (8+ 25). K, = G(s) H(s) = Le KE fF) = SS Gop) @#P2) Ps) 14 1K, l+o In systems with type number 1 and above, for unit step input the value of K, is infinity and so the steady state error is zero. 2.13 STEADY STATE ERROR WHENTHE INPUT IS UNIT RAMP SIGNAL ' _ 56) Steady state error, ¢,, = Lt TG@He When the input is unit ramp, R(s) 3 1 si : 1 - i 1 se (2.61) 1,2 Lt ——S = Lt ———__ - SO TFGE)H(S) 99 sHSG@HG) “Lt sG@H®” K where, K, = & s G(s) H(s) The constant K, is called velocity error constant. Type-0 system (s+ 2) (St 22) (8+ 25). (S+P1) (+P2) 6+Ps). K, = Lt G(s) H() = Lt sk. es so ey =1/K,=1/0=% Hence in type-0 systems when the input is unit ramp, the steady state error is infinity. Type-1 system = (S42) HZ) SF Ap) CL K, = Lt sG(s) H(3) = Lt f Neen) (s*pr ) ps). 1.e,,=1/K, = constant Hence in type-1 systems when the input is unit ramp there will be a constant steady state error. Type-2 system 2 (6+4) (8+%) (S42 Be HO) Oe Fp.) GPa) GBs) =UK,=1/o=0 In systems with type number 2 and above, for unit ramp input, the value of K, is infinity so the steady state error is zero.2.14 STEADY STATE ERROR WHENTHE INPUT IS UNIT PARABOLIC SIGNAL 2.36 sR(s) Steady state error, e,,= Lt Faro When the input is unit parabola, R(s) => st ness lt —S =u = 4 et 0 14+G(s) H(s) 59 s +s° G(s) H(s) Lt s’G(s) His). K, where, K, = Lt s°G(s) H(s) The constant K, is called acceleration error constant. Type-0 system K,= Lt 260) H@)= Lt %x Cw G*%) G2). 0 590 (8+ Py) (S*P2) (S+Ps Hence in type-0 systems for unit parabolic input, the steady state error is infinity. Type-1 system ke soy 1g 2 2D £200 HO Ly OR Ere pa) 1 =o 1 2 (0 Hence in type-1 systems for unit parabolic input, the steady state error is infinity. Type-2 system K, = 14969) H@)= Lt s ta) BFA) E Tow (8+ p,) (S+P2) (s+ P3). constant 1 <= constant Hence in type-2 system when the input is unit parabolic signal there will be a constant steady State error. Type-3 system orgie = 11 ex tBG+e) +2). B= POO) HO” 14, PRS Grp rps) GP) In systems with type number 3 and above for unit parabolic input the value of K, is infinity and so the steady state error is zero.2.37 ‘TABLE-2.2: Static Error Constant for ‘TABLE-2.3 : Steady State Error for Various Type Number of Systems Various Types of Inputs vor ‘Type number of system Tnput ‘Type number of system constant [0 T 23 Signal 0 T 2 3 K, [constant |__© fe 7 Unit $1 Tel 0 x ; nt Step | TK o]o q constant | 0 [ao 7 unit Rs ane k, 0 0 [constant | —_— “| K ° I Unit Parabolit oO 2 x 0 XK 2.15 GENERALIZED ERROR COEFFICIENT ‘The drawback in static error cocfficients is that it does not show the variation of error with time and input should be a standard input. The generalized error coefficients gives the steady state error as a function of time. Also using the generalized error coefficients, the steady state error can be found for any type of input. The error signal in s-domain, E(s) can be expressed as a product of two s-domain functions. -—_ 29 ____ | _egy- 20) SEG HD SOHO RO =FORO nu (2.63) where, F(s)=-——-_— T+G@) 16) Let, e(t) = £*{E(6)} (error signal in time domain) * f(t) = £F(S)} r(t) = £°{R(6)} (input signal in time domain) The convolution theorem of Laplace transform states that the Laplace transform of the convolution of two time domain signals is equal to the product of their individual Laplace transform. ie, LEA) + (Y= Fs) RG) where + is the symbol for.convolution operation LUA) RD} = AO « 0 (2.64) From equation (2.63) & (2.64) we can write, e(t) = ft) # re) Mathematically the convolution of f(t) and r(t) is defined as, u f(t) « r() = J £(1) (t-T) a ; where T is a dummy variable ” e(t)= J f(T) (t-T) aT It is assumed that the input signal starts only at t = 0 and does not exist before t=0. Also we are interested in finding error signal at any time t after t = 0 (i.e, for t> 0). Hence in the above equation the limit of integral can be changed as 0 to t. ett) =f £(1) (t—T) aT 22.38 Using Taylor's series expansion the signal rT) can be expressed as, r(t-T)=r(t) ~ r+ xy-E (t). = 1* derivative of r(t) i(t)=2™4 derivative of r(t) where, i( 1()=n* derivative of r(t) On substituting the Taylor's series expansion of r(t —T), the error e(t) can be written as, i wy Pan Pa a. ey= f(T) [r0-rH9.E ‘w-T FORD r(t)...| aT } ! 7 nl A ‘ ' 2 e(t) “J £(T) 1) dT - J £(T) T e(t) dT + J > i(t) aT bP ' wT J A> ror] £1) OS FO aT. Since r(t), i), #(0,....r(0) are constants when the integration is done with respect to T, the error signal can be written as, p 5 ay t =n) fH) at-iXy J THD a+ HOP TET) dt 0 0 “0 0 f OF oe oat TPE(T) dtt...+(-1) a TAT) dt... ‘ £ Let, Cy= +f f(T) eT Cy=-] PT) ar > 9 5 Gf TE(T) dT 3 ‘ ‘ nsf Te(T) dT C, crf T'E(T) dT e(t)=r(t) Cy + F(t) Crt® or a 2+ Fy ah +t & = Co t(t)+C, F(t) +2 + Ft) +f (2.65) The equation (2.65) is the general equation for error signal, e(t). ‘The coefficients C,, C,, C, ,.C, are called the generalized error coefficients or dynamic error coefficients. The steady state error ¢, is obtained by taking limit t + « on e(t). a(t) Say a & TF) te et 1 (t)aee ssee2.66) -. Steady state ertor, ¢,, = Lt us] no Co + F(t) C, +70) & +#() & = Cyr(+G io+Z + HO+2.39 2.16 EVALUATION OF GENERALIZED ERROR COEFFICIENTS The generalized error coefficient is given by, ef oo. . - 1 cC.=(-)' J 7°41) aT: where F(s) anne We know that £{£(T)} ~ F(s), hence by the definition of Laplace transform, F(s) =| £(1) ear sweel2.67) 3 On taking Lt on both sides of equation (2.67) we get, 0) ‘ Lt F(s)— Lt [ f(T) eta wrongs [ms =| FC) Lt etar=f £(T) dT=C, ° ° ef EER (2.68) On differentiating equation (2.68) with respect to s we get, a af = gro ad (1) eT aT Tem 3 ear af 5 “| fer) © ear “J #1) (CT) Tat t =-f tee Tar senel2.69) 0 On taking |“ on both sides of equation (2.69) we get, d = i ST, Lt ZFO= Lt - J TET) TT : ‘ --J TET) Lt, ear=-] THT) aT=C, ° 3 — | Sey 2.70) 2G 1g FO] (2.70) On differentiating equation (2.68) on both sides with respect to s we get, afd al st & [goo]-<[- J THT) € a ¢ ‘ a ‘ -|_ oo Lay et ez F(s) | TET) & (c | J TE(T) (-T) dT PEO). fpr et SS | aT) eTaT soe QT) as? }Applying the limits > 0 on both sides of the equation (2.71) we get, Lt cane = Ut fe fT) aT ss0ds sod £ 1 -Jr 0) Lt e*ar= fr RT) AaT=C, ° a sn(Q.72) Similarly it can be shown that, weed.) C,= So (2.73) 2.17 CORRELATION BETWEEN STATIC AND DYNAMIC ERROR COEFFICIENTS The values of dynamic error coefficients can be used to calculate static error coefficients. The following expressions shows the relationship between them. | 1 C0 BPO" BT GOH L + It, G) HG) 14K, 2.18 ALTERNATE METHOD FOR GENERALIZED ERROR COEFFICIENTS er . ___ RG) The enor signal in s-domain, Fis)= SO Fe) ee (2.77) “RE GOH) The equation (2.77) can be expressed as a power series of s as shown in equation (2.78). BO a pees Stet See... sne(2.78) R(s) 1+G(s) H(s) 2 3 ©. E(s) = Cy R(s)+ osro+ Se Rose R6sann saonl2.79) On taking inverse Laplace transform of equation (2.79) we get, 0) = Coste Cin(t)+ BP Ve a(t}. ri 3 (2.80) The equation (2.80) is same as that of equation (2.65) in section 2.14.This method will be useful to find the generalized error coefficients without using differentiation, but using laplace transform.2.41 ear EXAMPLE 2.11 10(s+2) Fora unity feedback contro! system the open loop transfer function, G(s) = ie 3 -Find a) the position, velocity and acceleration error constants, b) the steady state error when the inputis R(s), where R(s) = 2-5 + @ SOLUTION a) To find static error constants Fora unity feedback system, H(s)=1 i . = 1p 100842) Postion enor constant, Ky = Lt GIs)HUS}= LE GIS) = Lay = . _ wie <10(8+2) _ Velocity error constant, K,= Lis Gis}4s)= Lbs Gis) = Lt s eT == Acceleration error constant, K, = Lt, S7G(S)H(s) = Lt s*G(s) 2 Lt 52 10(8*2) 10% 2 _ 99 so” sis) 7 b) To find steady state error Method-I ‘Steady state error for non-standard inputis obtained using generalized error series, given below. The error signal, e(t) =1(9C, +f(1}C, +0 Ayn... 3.2.4 Given that, Ris)= 2-4 ay ; , 3.2.4 Input signal in time domain, r(t) = .-"{R(s)} = of-3 + at = ie e = 3-meLE = 3a a a 1 t 1i(t)= Zitt)= 2+ t= -245 @ a 1 F)= we = Ally wy d. 7)= Sy = Siq)=0 ‘The derivatives of (i) is zero after second derivative. Hence we have to evaluate only three constants C,.C, and C,. ‘The generalized error constants are given by, . deg. . & Soe HF): MEA: = Laer) Fe)- —1__-_1_- 1 ssf) si+s? ~TSGQ)KS) 14G(s) , 10/8+2) s%(s+7)+10(s+2) so +s7+10s+20 e+ Co= LyFis)= Lt +8? +108 +20sit Saal eevee 20)(95? + 25)- (38 +8°)(86? +25 +1 = lt (s+s?+ 10s +20)" C= Lt tors) = es =u 3s° + 284 +36" +2s° + 308° + 208" + 60s* + 40s - 35° - 2s* - 10s? - 3s" - 25° - 108? (s°+s?+108+20)" | 208° +70s? + 40s T3020 4nm on” +4 (5° +8? + 108 + 20) 20s* + 70s? + 40s | =k or (s) af, F( 3 oe (9 Leas| ae" sa mee (2° +8? +105 +20) (60s? +1408 + 40) eu ~{208* + 708* + 40s) 2x(s? +8? + 10s + 20) (35? + 28+10) _ 207x401 om (o? +5? + 105+ 20) © 2040 Error signal, e() =F + i0}C +0) S2 = 1 Steady state = LL elt utit ty sate error, es = Lt e(t)= Lt 35 = a5 . Method - I! The error signal in s-domain, E(s) = 7S Sons 3.2 1 10(s + 2) awentiat Re)=2- Fash soy ES: Herat 32,4 3.2, . _s 83s) = 8 s* 38" $9)" 40622) ~ Hex) 105-2) 3*(s+1) s*(s+1) 23) s+) s%(s+1) 1 s(s+1) izes ea | # S| s*(5+7)+10(5+2) +353] Se 7)+10(54 The steady state error e,, can be obtained from final value theorem. Steady state error, €,, = Lt e(t)= Lt 5 E(5) ects fal " eee4 2f_ s+ 1 (+1) = 330 § iss (+1) +10(s+2) |- s*| s(6+7)=10(s+2) “35 (e+) + 1082) 357(8+0) 2is+1) (+t) zo-0+t. = Ly S(S+)+10(s+2) (s+) +10(8+2) 3s7(5 +1) +30 (s+ 2) 1 “oO2.43 Method - il Ris) 1+ G(HE) Es) _ 4 “Ris) 1 GG)H(s) 10(8+2) . s*(s+1) * Error signal in s-domain, E(s) = Given that, G(s) =. His)=1 Es), 1 s*(s+1) “RO 4, 10S+2) s%(s+1) + 10(s+2) s*(s+1) si+s? S4s?+10s+20 20: 10s+s*+8° 24g? 20° & sd otys) +154 Zo HS) + GES PIs) ss? e)-nel oS | ‘Ontaking inverse Laplace transform of the above equation we get, | 2.1 +a eRe et) = £YR(S)} = of 2 aay. 1 t (t) = Ort) =-24+ 22t=-2+4 HO) =F) = 24 B= 245 Error signal in time domain, e(t) = hi =a! = sz Steady state error, e,, = Lt eft) = RESULT @ Position errorconstant, =n Velocity error constant, Kee ‘Acceleration error constant, K=20 (b) When, R(s)= 120+10s+s?+5 EXAMPLE 2.12 For servomechanisms with open loop transfer function given below explain what type of input signal give rise toa constant steady state error and calculate their values. 20(5+2) | oer ers) 10 5) Gis) “was a) Gis)= c) G(s) 10 s%(5+ 16+ 2)apr Ten RE prise ysis) a4 SOLUTION 20(s +2) s(S+1(s+3) Letus assume unity feedback system, -.H(s)=1 The open loop system has. pole t origin. Hence itis 8 type-1 system. In systems with type number-1, the velocity (ramp) input will givea constant steady state error. a) Gs)= 1 The steady state error with unit velocity input, e, Velocity error constant, K, = Lt_ s G(s) H(s)=_Lt ‘s G(s) soo stb 20(8+2) _20x2_ 40 so ® Se=fer3) 1x3 8 Steady state error, ess = 10 9 ayes Lotus assume unity feedback system, <. His) ‘The open loop system has no pole at origin. Hence itis a type-O system. In systerns with type number-O, the step input will givea constant steady state error. The steady state error with unit step input, @,. = Postion error constant, K, = Lt, GlS)H(s)= Lt Gis) = U1, @ Steady state error, e:. = ®) Gls)- a Letus assume unity feedback system, -.H(s)=1 The open loop system has two poles atorigin. Hence itis a type-2 system. In systems with type number-2, the acceleration (parabolic) input will give a constant steady state error. ‘The steady state error with unit acceleration input, e,. = 10 10 s%(s+(s+2) 1x2 Acceleration error constant, K, = & 8* G{s)H(s)= Lt s? G(s) = uy s? Steady state error, 6 =—2 - 02 RESULT 1. Insystem (a) with unit velocity input, Steady state error = 0.075 2. I system (b) with unit step input, Steady state error = 0.375 3.in system (0) with unit acceleration input, Steady state error = 0.22.45 EXAMPLE 2.13 ‘The open loop transferfunction of servo system with unity feedback is G(s) = 10/s(0.1s+1). Evaluate the static error constants of the system. Obtain the steady state error of the system, when subjected to an input given by the polynominel, a He) =a rate Se, SOLUTION To find static error constant Forunity feedback system, H(s)=1 =. Loop transfer function, G(s) H(s) = G(s) a ‘The static error constants are K,, K, and, +i - w Position error constant, Ky = 14, GI6)= Ub ase = . _ : 10 Velocity enor constant, K, = Lt sG(s)= Lt s=o7 a =10 10 (018+) ‘Acceleration error constant, K, = Lt, 8°G(s) = Li s* To find steady state error Method - | ‘Steady state error for non-standard inputis obtained using generalized error series, given below. The error signal, e{t) =r(}C, +H()C, +102 Given that, r(t)=a, +ad+ oe 2 F)= uty = (10 The steady state error e,, can be obtained from final value theorem. Steady state eror, ¢,,= Lt (= Lt s E{s) fa,{_ s(o1s+1) aa 9 ae Fa '? ne s(01s+1) }-2{ s(0.1s+1) } [3(018+ +10 | "S| (0.15 +1)+10 | = rp | 208(Ots+) | a(ate+t) 2f0+1) | 9. ow $59] (01641) 10” S(0.18+ 7-10 * s[s(ors+ 1+ 10) 70 Method - iI . - =—RS)__. ES) Ewer signalin s-domain, Ete) == Es: Ae) TSS 10 Given that, G(s) Hotsey Ot) 1 E(s) 1 -_018+1) __O1s?+s_s+0ts? _ ss? + “RE 4, 10 S(01S+)+10 Ot? +5470 T0+s10% 10 1000 soy se | Dividing numerator poly: | /E(8)= 35 RO) gh | by denominator polynomial, _|2.47 On taking inverse Laplace transform, 3 1, 4 10 7600 ate 1 gays... = 79" to00 10+8+01s? | s+0.1s? a. | st st Given that, r(t)=a) +a +e | Sto t100 22102100 d f= GD =a, baat = $10= a, amy oe A= GHD =0 : ‘i =iet(a,+ - Error signal in ime domain, e(t)= + >ftt)= = (a; + act) Steadystaieert, e,= Lte(= It we +231) #00 RESULT (@) Position error constant, K=< (©) Velocity error constant, K,=10 (©) Acceleration error constant, kK=0 2 (@. When input, r(t) = 2, +a +e Steady state error, @,, =< EXAMPLE 2.14 Consider a unity feedback system with a closed loop transfer function 28) - KS*° _ RS) s+as+b (@-k) Determine open loop transfer function G(s). Show that steady state error with unit ramp inputis given by SOLUTION For unity feedbacksystem, H(s)=1 The closed loop transfer function, M(s) “ = wets =F So. Gis) “T+ Gts) =Ms) On cross multiplication of the above equation we get, Gls)= M(s)I1+G{s)] = Ms) + MB) GIs) Ks+b s*+as+b ++ G{S)-M{s) Gis) =M(s)_ => GIS)I1-M{s)]=Mis)_ > MI) ~. Open loop transfer function, Ks+b Mis) as 75 Ks+b Gis)= =—S4as ib TMs) 7. _Ke+b ‘as +b)- (KS +b) s+as+b } Ks+b Ks+b Ks+b S4as+b-Ks-b ss(a-Ks ss+eK)]248 Ks+b b §st@-K] a-K Velocity error constant, Ky = Lis G(s)H(s)= Lbs G(s) = Lt s With velocity input, Steady stato’error, e,, = RESULT Open loop transfer function, G=—*h - s[s+(a-k)] With veloctiy input, Steady state error, @,. = ack EXAMPLE 2.15 A.unity feedback system has the forward transfer function Gs) - (28+) _ when the inputrit)= 16t, s (58+ I)(1+s)’ etermine the minimum value of K, so that the steady erroris less than 0.1 SOLUTION Given that, input r(t)=1+6t Oj taking laplace transform of rt) we get R(s). 2. Rls) = £{nt} = {1+ 8} = ie 4 ‘The error signal in s-domain E(s) is given by, 1,6 1,6 um(e)=_ POs ss [Here His) . 1+GQ)H(s),_-K(2s+1) s (58+1) (1+5)°+K(25+1) 958+ 1t+s)" 5 (5s+1) (I+5)° 8| 85+ (1-5 +K(250)| =| y at sGsi)(tis’ “| 6 s(6s+1)(1+8) | 5 (65+ (ts) +Ki(2s+ | The steady state error e,,. can be obtained from final value theorem. Gs = Lt elt) Lt, SEIS) af sfSs+t1+s)” | ‘| 9(5+1)(14 5)” }| “us tptetoa "| 60 9)? +28 +1) | | 9(58-+ 1ift+s)? (5s + N(1+s)? | = 0-8 6 © | 568+ (14s) +K,(2s+1) sSs+1(1+s) +K(25+9) KK o1- 8 =f. Given that, e,, < 04 O= ef Kin gyn 60 RESULT For steady state error, e,, < 0.1, the value of K, should be greater than 60.2.9 Central 2.19 COMPONENTS OF AUTOMATIC CONTROL SYSTEM The basic components of an automatic control system are Error detector, Amplifier and Controller, Actuator (Power actuator), Plant and Sensor or Feedback system. The block diagram of an automatic control system is shown in fig 2.16. Automatic contre Sree oA Reference Error detector a 39) input Amplifier and | Co" \ Bane Output Controle P| Actuator [>| Plant (input Error signal signal) Sensor or Feedback signal Feedback system [* Fig 2.16: Block diagram of automatic control system. ‘The plant is the open loop system whose output is automatically controlled by closed loop system. The combined unit of error detector, amplifier and controller is called automatic controller, because without this unit the system becomes open loop system. In automatic control systems the reference input will be an input signal proportional to desired output. The feedback signal is a signal proportional to current output of the system. The error detector compares the reference input and feedback signal and if there is a difference it produces an error signal. An amplifier can be used to amplify the error signal and the controller modifies the error signal for better control action. The actuator amplifies the controller output and converts to the required form of energy that is acceptable for the plant. Depending on the input to the plant, the output will change. This process continues as long as there is a difference between reference input and feedback signal. If the difference is zero, then there is no error signal and the output settles at the desired value. Generally, the error signal will be a weak signal and so it has to be amplified and then modified for better control action. In most of the system the controller itself amplifies the error signal and integrates or differentiates to produce a control signal (ie., modified error signal). The different types of controllers are P, PI, PD and PID controllers. 2.20 CONTROLLERS A controller is a device introduced in the system to modify the error signal and to produce a control signal, The manner in which the controller produces the control signal is called the control action. The controller modifies the transient response of the system. The electronic controllers using operational amplifiers are presented in this section. The following six basic control actions are very common among industrial analog controllers. ‘Two-position or ON-OFF control action. Proportional control action. Integral control action. Proportional- plus- integral control action. Proportional-plus-derivative control action. Proportional-plus-integral-plus-derivative control action. ae e yeDepending on the control actions provided the controllers can be classified as follows. 1 2. 3. 4. 5. 6. Two position or ON-OFF controllers. Proportional controllers. Integral controllers. Proportional-plus-integral controllers. Proportional-plus-derivative controllers. Proportional-plus-integral-plus-derivative controllers. ON-OFF (OR) TWO POSITION CONTROLLER The ON-OFF or two position controller has only two fixed positions. They are either on or off. The on-off control system is very simple in construction and hence less expensive. For this reason, it is very widely used in both industrial and domestic control systems. The ON-OFF control action may be provided by a relay. There are different types of relay. The most popular one is electromagnetic relay. It is a device which has NO (Normally Open) and NC (Normally Closed) contacts, whose opening and closing are controlled by the relay coil. When the relay coil is excited, the relay operates and the contacts change their positions (j.e., NO -» NC and NC — NO). Let the output signal from the controller be u(t) and the actuating error signal be e({). In this controller, u(t) remains at either a maximum or minimum value. u(t) = uj; fore(t) <.0 =u): fore(t) > 0 R(s) EGS) PU UG) [> u— Feedback signal Fig 2.17 : Block diagram of on-off controller. F()=L{e()} + UG) = L{uts)} PROPORTIONAL CONTROLLER ( P - CONTROLLER ) The proportional controller is a device that produces a control signal, u(t) proportional to the input error signal, e(t). . In P-controller, u(t) «< e(t) u(t) = K, e(t) vevn(2.81) where, K, = Proportional gain or constant On taking Laplace transform of equation (2.81) we get, U(s) = K,E(s) (2.82) Transfer fimetion of P controller, ty K, eo(2.83) ‘The equation (2.82) gives the output of the P-controller for the input E(s) and equation (2.83) is the transfer function of the P-controller. The block diagram of the P-controller is shown in fig 2.18. RG) Ets) Us) K, | Feedback signal Fig 2.18 : Block diagram of proportional controller.2.51 - From the equation (2.82), we can conclude that the proportional controller ainplifies the error signal by an amount K,, Also the introduction of the controller on the system increases the loop gain by an amount K,. The increase in loop gain improves the steady state tracking accuracy, disturbance signal rejection and the relative stability and also makes the system less sensitive to parameter variations, But increasing the gain to very large values may lead to instability of the system. The drawback in P-controller is that it leads to a constant steady state error. EXAMPLE OF ELECTRONIC P-CONTROLLER The proportional controller can be realized by an amplifier with adjustable gain. Either the non- inverting operational amplifier or the inverting operational amplifier followed by sign changer will work as a proportional controller. The op-amp proportional controller is shown in fig 2.19 and 2.20. Ry R R, R a MN R wo a ~ c(t) a 44 u(t) Oo Inverting amplifier ¥ Sign changer Fig 2.19: Op-amp P-controller using Fig 2.20 : Op-amp P-controller using non-inverting amplifier. inverting amplifier. By deriving the transfer function of the controllers shown in fig 2.11 and 2.12 and comparing with the transfer function of P-controller defined by equation (2.83), it can be shown that they work as P-controllers, ANALYSIS OF P-CONTROLLER SHOWN IN FIG 2.19 In fig 2.19, the input o(t) is applied to positive input. By symmetry of op-amp the voltage of negative input is also e(t). Also we assume an ideal op-amp so that input current is zero. Based on the above assumptions the equivalent circuit of the controller is shown in fig 2.21. By voltage division rule, R R RR, ‘ O-Re ud); ult)= Ri 2. et) (2.84) + woe On taking Laplace transform of equation (2.84) we get, ett) | R,+R, ! YOR E(s) (2.85) Fig 2.21 : Equivalent circuit UG) RyRy of P-controller EG) Ry shown in fig 2.19, ‘The equation (2.86) is the transfer function of op-amp P-controller. On comparing equation (2.86) with equation (2.83) we get, sev(2.86) Proportional gain, Kp -Aue (2.87) "| Therefore by adjusting the values of R, and R, the value of gain, K, can be varied.aiapler 2 Tana Response Analyst 2.52 ANALYSIS OF P-CONTROLLER SHOWN IN FIG 2.20 The assumption made in op-amp circuit analysis are, 1. The voltages at both inputs are equal 2. The input current is zero. Based on the above assumptions, the equivalent circuit of op-amp amplifier and sign changer are shown in fig 2.22 and 2.23. i, ae aaa | Fig 2.22 + Equivalent circuit of amplifier Fig 2.23 : Equtvalent cireut of sign changer: From fig 2.22, e()=iRii =o (2.88) BOR (2.89) Substitute for i, from equation (2.88) in equation (2.89). : -Or 7 u(= Re (2.90) From fig 2.23, u(t)=-4R 3669 (2.91) BR (2.92), Substitute for i, from equation (2.91) in equation (2.92). 5 =-20 R__, - ()=-TP R=—w) On equating the equations (2.90) and (2.93) we get, ul) 2 R; u(t) a (2.94) Hl On taking Laplace transform of equation (2.94) we get, UG)= GF EG) 095) 0) =k EQ) Rp al (2.96) The equation (2.96) is the transfer function of op-amp P-controller. On comparing equation (2.96) with equation (2.83) we get, Proportional gain, K, 42.97) Ry Therefore by adjusting the values of R, and R, the value of gain K, can be varied.2.53 INTEGRAL CONTROLLER (I-CONTROLLER) The integral controller is a device that produces a control signal u(t) which is proportional to integral of the input error signal, e(t). In -controller, u() a feo dt; .u= K few at 98) where, K, = Integral gain or constant. On taking Laplace transform of equation (2.98) with zero initial conditions we get, B uG)= «8 (2.99) <. Transfer function of 1+ controller, 2. af (2.100) The equation (2.99) gives the output of the I-controller for the input E(s) and equation (2.101) is the transfer function of the I-controller. The block diagram of I-controller is shown in fig 2.24. Ls Feedback signal Fig 2.24 : Block diagram of an integral controller. The integral controller removes or reduces the steady error without the need for manual reset. Hence the I-controller is sometimes called automatic reset. The drawback in integral controller is that it may lead 10 oscillatory response of increasing or decreasing amplitude which is undesirable and the system may become unstable. EXAMPLE OF ELECTRONIC I-CONTROLLER ‘The integral controller can be realized G by an integrator using op-amp followed by a sign changer as shown in fig 2.25. By deriving the transfer function of the a ~ controller shown in fig 2.25 and comparing with the transfer function of [-controller defined by equation(2.101), it can be shown + that it work as I-controller. Fig 2.25 : I-controller using op-amp. ANALYSTS OF I-CONTROLLER SHOWN IN FIG 2.25 et) Ry + uo” _L Sign changer The assumptions made in op-amp cirouit analysis are, 1. The voltages of both inputs are equal. 2. The input current is zero, Based on the above assumptions the equivalent circuit of op-amp integrator and sign changer are shown in fig 2.26 and 2.27. From fig2.26, e@=iR,3 Qa (2.101) se (2.102)1 (2.103) oe e “woe Qw Fig 2.26 + Equivalent circuit of integrator. Fig 2,27 : Equivalent circuit of sign changer. From fig 2.27, u()=-iR, © nw sa (2.108) ut) =1R 2.105) Substitute for i, from equation (2.106) in equation (2.107), Lae we re (2.108) On equating the equations (2.103) and (2.106) we get, changer as shown in fig 2.29. By deriving the transfer function of the 1 o ww controller shown in fig (2.29) and comparing with Integrator with gain I siee % the transfer function of Pl-controller defined by ay nen changer equation (2.114), it can be proved that the circuit Fig 2.29 : Pl-controller using op-amp. shown in fig 2.29, work as Pl-controller. ANALYSIS OF PI-CONTROLLER SHOWN IN FIG 2.29 The assumptions made in op-amp circuit analysis are, 1. The voltages at both inputs are equal. 2. The input current is zero. Based on the above assumptions the equivalent circuit of op-amp integrator and sign changer are shown in fig 2.30. and 2.31. ® Ry From fig 2.30, e(t)=iR, ; 7 eoe(2.114) u(0=-i,R, Substitute for i, from equation (2.114) in equation (2.115). elie ve(2118) ef 1 . y@-— Op, fea svoe(2116)2.56 i > Roi, A | + - ¥ Fig 2.30 : Equivalent circuit of integrator. Fig 2.31 : Equivalent circuit of sign changer. 2 wo 4) uo | From fig 2.31, wt)=-iR, + ip wo se Qui7 u()=iR sau(2118) Substitute for i, from equation (2.117) in equation (2.118), ‘ n w= 2r=-u sa (2.119) On equating the equations (2.116) and (2.119) we get, Op 1 eo =u(t)=-SR, -— [= at ui)= FPR, ole R L : Wr ORE feo dt swou(2-120) On taking Laplace transform of equatiori (2.120) with zero initial conditions we get, R; 1 BG) us) = &2 Bey + Zh) OR PORG s . UO Re 1 Els) Ry (+35) a(2.121) The equation (2.121) is the transfer function of op-amp Pl-controller. On comparing equation(2.121) with equation (2.113) we get, R, R Proportional gain, K, Integral time, T, = R,C, By varying the values of R, and R,, the value of gain K, and T, can be adjusted. PROPORTIONAL PLUS DERIVATIVE CONTROLLER (PD-CONTROLLER) The proportional plus derivative controller produces an output signal consisting of two terms : one proportional to error signal and the other proportional to the derivative of error signal. In PD- controller, u(t) a [- +f «| 3 +. Transfer function of PD - controller, Be =K(+R)) (2.124) s Le The equation (2.123) gives the output of the PD-controller for the input E(s) and equation (2.124) is the transfer function of PD-controller. ‘The block diagram of PD-controller is shown in fig 2.32. Ts x UG) Ris) 5 E@) U(s) K,G+Ts) >> Ris) ES) = Feedback signal Feedback signal Fig 2,32 : Block diagram of PD- controller. The derivative control acts on rate of change of error and not on the actual error signal. The derivative control action is effective only during transient periods and so it does not produce corrective ‘measures for any constant error. Hence the derivative controller is never used alone, but it is employed in association with proportional and integral contiollers. The derivative controller does not affect the steady- state error directly but anticipates the error, initiates an early corrective action and tends to increase the stability of the system, While derivative control action has an advantage of being anticipatory it has the disadvantage that it amplifies noise signals and may cause a saturation effect in the actuator. The derivative control action is adjusted by varying the derivative time. The change in the value of K, affects both the proportional and derivative parts of control action. The derivative control is also called rate control, EXAMPLE OF ELECTRONIC PD-CONTROLLER The PD-controller can be realized by an i x op-amp differentiator with gain followed by asign | changer as shown in fig 2.33. LAK ~ lor By deriving the transfer function of the 4, tw - controller shown in fig 2.33 and comparing with + the transfer function of PD-controller defined by {fetes sien equation (2.124) it can be proved that the circuit with gain ee mert shown in fig 2.33 will work as PD-controller. Fig 2.33 : PD controller using op-amp ANALYSIS OF PD-CONTROLLER SHOWN IN FIG 2.33 The assumptions made in op-amp circuit analysis are, 1. The voltages at both inputs are equal. 2. The input current is zero. Based on the above assumptions the equivalent circuit of op-amp differentiator and sign changer. are shown in fig 2.34 and 2.35. (t) de From fig 2.34,» i, “Ere “oO so (2.125) - =a) iR,=-H,0. Ry On equating the equations (2.125) and (2.126) we get, R d (Beso + necro] cv (2.127)2.58 cs. ba i “a | i R i R R LR iy aN [& W\ FR 4GR, *iRT tiR- | | ame) SO Day woe ¥ R liso i=0 © ¥ | + + Fig 2,34 : Equivalent circuit of differentiator Fig 2.35 : Eiquivalent circuit of sign changer. ; ru. From fig 2.35, u(t)=—iR 3 9 b= (2.128) u()=iR .129) Substitute for i, from equation (2.128) in equation (2.129). 2 u()= MOR=-u) seal (2.130) On equating the equations (2.127) and (2.130) we get, -ut) = {Exo + cao) 1 R 4 su= re FRC Ol (2.131) On taking Laplace transform of equation (2.131) with zero initial conditions we get, ve 2: US)= xR EQS)+RCysES) (2.132) 1 _ Us) _RB, To Re, OTR) seen (2.133) ‘The equation (2.133) is the transfer function of op-amp PD-controller. On comparing equation (2.133) with equation (2.124) we get, Proportional gain, Ky = = 1 Derivative time, T,=R,C, By varying the values of R, and R,, the value of gain K, and T, can be adjusted. PROPORTIONAL PLUS INTEGRAL PLUS DERIVATIVE CONTROLLER (PID-CONTROLLER) ‘The PID-controller produces an output signal consisting of three terms : one proportional to error signal, another one proportional to integral of error signal and the third one proportional to derivative of error signal.2.59 InPID-controller, u(t) [a+ foro arta] Ky. d (2.1345 ut) = K, waa] e(tdt +K Ty qe « where, K, ~ Proportional gain T, = Integral time T, = Derivative time On taking Laplace transform of equation (2.134) with zero initial conditions we get, . K, EG U@)=K,F(s) e FO, TsEs) (2.135) . UG) 1 \ I Transfer function of PLD - controller, eat =K, ( wae ae The equation (2.135) gives the output of the PID-controller for the input F(s) and equation (2.136) is the transfer function of the PID-controller. The block diagram of PID-contoller is shown in fig 2.36. T UW _. R&) Le] ELH h! = Fig 2.36: Block diagram of PID- controller. ‘The combination of proportional control action, integral control action and derivative control action is called PID-control action. This combined action has the advantages of the each of the three individual control actions. Ris) EG) U(s) >| > (4 Ki[t+ i +7s| Feedback signal The proportional controller stabilizes the gain but produces a steady state error. The integral controller reduces or eliminates the steady state error. The derivative controller reduces the rate of change of error. EXAMPLE OF ELECTRONIC PID-CONTROLLER The PID-controller can be realized by op-amp amplifier with integral and derivative action followed by sign changer as shown in s bn | fig 2.37. I A + 20 y 2 pve v po a + ui By deriving the transfer function of the controller shown in fig (2.37) and comparing with Amplifier the transfer function of PID-controller defined by equation (2.136) it can be proved that the Fig 2.37: PID- controller using op-amp. circuit shown in fig 2.37 work as PID-controller. L. Sign changer ANALYSIS OF PID-CONTROLLER SHOWN IN FIG 2.37 The assumptions made in op-amp circuit analysis are. 1. The voltages of both inputs are equal. 2. The input current is zero.