M2-3S Active Filter (Part II) • • • • • Biquadratic function filters Positive feedback active filter: VCVS Negative feedback filter: IGMF Butterworth Response Chebyshev Response Active Filter (Part 2) 1 Biquadratic function filters s 2 Z s 2 Z s cs d QZ H ( s) 2 K P s as b 2 2 s s P QP 2 Realised by: (I) Positive feedback (II) Negative feedback Active Filter (Part 2) (III) Band Pass Biquadratic functions s s K s 2 as b s 2 P s P2 QP (IV) Band Stop (I) Low Pass H ( s) K H ( s) K 1 1 K s 2 as b s 2 P s P2 QP s2 b s 2 Z2 H ( s) K 2 K s as b s 2 P s P2 QP (V) All Pass (II) High Pass 2 2 s s H ( s) K 2 K s as b s 2 P s P2 QP s2 Z s Z2 s 2 as b QZ H ( s) K 2 K s as b s 2 P s P2 QP Active Filter (Part 2) Low-Pass Filter Voltage Gain (dB) Voltage Gain Qp = 1.5 1.5 Qp = 1.5 Qp = 1 0 K = 1, p = 1 Qp = 1 1 Qp = Qp = 1 2 -10 2 Qp = 0.5 -20 p Q = Qp = 0 .5 0.5 Qp 0 1 0 =0 -30 K = 1, p = 1 .1 1 2 3 4 0 .1 0 1 2 Frequency Frequency H s 1 s 2 s 1 Qp Active Filter (Part 2) 3 4 High-Pass Filter Voltage Gain Voltage Gain (dB) 5 Qp = 1.5 1.5 Qp = 1.5 K = 1, p = 1 0 Qp = 1 1 Qp = Qp = 1 1 Qp = 0.5 -5 2 Qp = Qp = 0 .5 2 -10 0.5 Qp = Qp = 0 0 1 0 1 -15 .1 2 0 .1 K = 1, p = 1 3 4 -20 0 Frequency 1 2 Frequency s2 H s s s2 1 Qp Active Filter (Part 2) 3 4 Band-Pass Filter Voltage Gain (dB) Voltage Gain 5 1.5 Qp = 1.5 K = 1, p = 1 0 -5 Qp = 1 1 Qp = 1 -10 2 -15 0.5 -30 2 Qp = 0.1 -25 Qp = 0.1 0 Qp = 0.5 -20 Qp = 0.5 0 Qp = 1.5 Qp = 1 1 Qp = 2 4 Frequency 6 8 K = 1, p = 1 0 2 4 Frequency H s s s s2 1 Qp Active Filter (Part 2) 6 8 Band-Stop Filter Voltage Gain Voltage Gain (dB) 5 10 Qp = 1.5 K = 1, p = 1, z = 2 Qp = 1.5 4 0 3 Qp = -10 2 0 .1 0.5 Qp = -20 1 2 Qp 1 1 Qp = Qp = Voltage Gain Qp = 1 2 K = 1, p = 1, z = 2 .1 =0 0 0 2 4 Frequency 6 -30 8 0 2 s 2 H s s 2 s 1 Qp 2 2 Active Filter (Part 2) 4 Frequency 6 8 Voltage Controlled Votage Source (VCVS) Positive Feedback Active Filter (Sallen-Key) Z3 if ii ia Z1 Vi By KCL at Va: where, Z4 Va ii i f ia 0 V Va ii i Z1 if Vo Va Z3 Va ia Z2 Z4 K Z2 Vo Therefore, we get V i V a Vo V a Va 0 Z1 Z3 Z2 Z4 Re-arrange into voltage group gives: 1 Vi Vo 1 1 Va 0 Z1 Z3 Z1 Z3 Z2 Z4 Active Filter (Part 2) (1) But, Vo Ki a Z 4 K Z 4Va Z2 Z4 (2) Substitute (2) into (1) gives Vi Vo Vo Z2 Z4 1 1 1 0 Z1 Z3 KZ4 Z1 Z3 Z2 Z4 or H Vo K ZZ 1 Vi Z1 1 K 1 2 1 Z1 Z 2 Z3 Z3Z 4 Z4 (3) In admittance form: H K 1 1 Y3 Y3Y4 1 Y4 1 K Y1Y2 Y1 Y2 Y1 (4) * This configuration is often used as a low-pass filter, so a specific example will be considered. Active Filter (Part 2) VCVS Low Pass Filter 1 H (s ) K 2 s as b In order to obtain the above response, we let: Z 1 R1 Z 2 R2 Z3 R1 R2 1 1 jC 3 sC 3 Z4 1 1 jC 4 sC 4 C3 C4 Then the transfer function (3) becomes: H ( s) K K' 1 sC4 R1 R2 sR1C3 1 K s 2 R1 R2C3C4 s 2 P s 2 P QP Active Filter (Part 2) (5) we continue from equation (5), K H ( s) 2 s R1 R2 C3C4 sC4 R1 R2 sR1C3 1 K 1 1 R1 R2 C3C 4 H ( s) C 4 R1 R2 R1C 3 1 K 1 2 s s R R C C R R C C 1 2 3 4 1 2 3 4 K Equating the coefficient from equations (6) and (5), it gives: P 1 R1 R2 C3C4 1 1 R1C3 R2 C4 QP 1 RC R1C4 R2 C4 1 K 1 3 R2 C3 R1C3 R2 C4 Now, K=1, equation (5) will then become, H (s ) 1 1 sC 4 R1 R 2 s 2 R 1R 2C 3C 4 Active Filter (Part 2) Simplified Design (VCVS filter) Z1 mR H ( s) Z2 R Z3 1 1 j (nC ) snC 1 1 sRC m 1 s 2 nmR 2C 2 mR Comparing with the low-pass response: 1 H (s) K s 2 P s P2 QP It gives the following: p 1 RC nm mn Qp m 1 Active Filter (Part 2) Z4 1 1 jC sC R nC C Example (VCVS low pass filter) To design a low-pass filter with f O 512Hz and Q 1 2 nC Let m = 1 mn 1 n n 1 QP m 1 11 2 2 mR R + vin - C n=2 1 1 P 2 (512Hz ) RC mn RC 1 2 RC 2 vo 1 Choose C 100nF 2k2 200nF 2k2 + Then R 2,198 ~ 2.2k vin What happen if n = 1? Active Filter (Part 2) 100nF - vo VCVS High Pass Filter C1 C2 R3 + vin R4 - vo KS 2 K'S2 H ( s) P 2 2 1 1 1 1 2 s s 1 K s s P Q P R4C1 R4C2 R3C1 R3 R4C1C2 Active Filter (Part 2) VCVS Band Pass Filter R1 vin C2 + R4 C5 K H ( s) R3 - s R1C5 s 1 1 1 C R1 R3 s 1 K 5 C5 R1 R4 R3 R4C2 R1 R3 R4C2C5 2 Active Filter (Part 2) vo K'S s2 P QP s P2 Infinite-Gain Multiple-Feedback (IGMF) Negative Feedback Active Filter Z 4 Z1 V i vi 0 vi 0 By KCL at Vx , Z 5 Z 3 V x Z 2 Vo Z5 Z Vx Vx 3 Vo Z3 Z5 Vi Vx Vx Vx Vx Vo Z1 Z 2 Z3 Z4 V o Note: because no current flows into v+, v- terminals of op-amp. Therefore, from KCL at node v- : Vo/Z5+ Vx/Z3 = 0 (1) (2) substitute (1) into (2) gives Vi Z3 Z3 Vo Z3 Vo V V V Z1 Z1Z5 o Z 5Z 2 o Z 5 Z 4 Z 5 o Z 4 Active Filter (Part 2) (3) rearranging equation (3), it gives, 1 V Z1Z 3 H o Vi 1 1 1 1 1 1 Z5 Z1 Z 2 Z 3 Z 4 Z 3Z 4 Or in admittance form: H Vo Y1Y3 Vi Y5 Y1 Y2 Y3 Y4 Y3Y4 Filter Value LP Z1 Z2 Z3 Z4 Z5 R1 C2 R3 R4 C5 HP C1 R2 C3 C4 R5 BP R1 R2 C3 C4 R5 Active Filter (Part 2) IGMF Band-Pass Filter s Band-pass: H ( s ) K 2 s as b To obtain the band-pass response, we let Z1 R1 Z 2 R2 Z3 1 1 jC3 sC3 Z4 sC 3 R1 H (s ) C C4 1 1 1 s 2C 3C 4 s 3 R5 R 5 R1 R 2 *This filter prototype has a very low sensitivity to component tolerance when compared with other prototypes. Active Filter (Part 2) 1 1 jC4 sC 4 R1 R2 C4 C3 Z 5 R5 R5 Simplified design (IGMF filter) sC R1 H (s) 1 2C s s 2C 2 R1 R5 R5 R1 C R5 C Comparing with the band-pass response s H ( s) K s2 P QP s P2 Its gives, p 1 C R1 R5 Qp 1 R5 2 R1 K- 1 CR1 Active Filter (Part 2) H j p 2Q 2 Example (IGMF band pass filter) To design a band-pass filter with P f O 512Hz and Q 10 1 2 (512Hz ) C R1 R5 C 100 nF R1R5 9,662,741 2 1 R5 QP 10 2 R1 R1 155.4 100nF 62,170 155.4 R5 62,170 vin 100nF With similar analysis, we can choose the following values: C 10 nF R1 1,554 and R5 621,700 Active Filter (Part 2) + vo Butterworth Response (Maximally flat) H j 1 1 o 2n where n is the order Normalize to o = 1rad/s Hˆ ( j ) 1 1 2n Butterworth polynomials: 1 Hˆ ( j ) Bn ( j ) Butterworth polynomials B1 s s 1 B2 s s 2 2 s 1 B3 s s 3 2s 2 2s 1 s 1 s 2 s 1 B4 s s 4 2.61s 3 3.41s 2 2.61s 1 s 2 0.77 s 1 s 2 1.85s 1 B5 s s 5 3.24 s 4 5.24s 3 5.24s 2 3.24s 1 s 1 s 2 0.62s 1 s 2 1.62s 1 Active Filter (Part 2) Butterworth Response Active Filter (Part 2) Second order Butterworth response Started from the low-pass biquadratic function H ( s ) K For p 1 K 1 Q 1 2 1 s2 P QP 1 (second order butterwoth polynomial ) 2 s 2s 1 1 H ( j ) 2 2 j 1 1 H ( j ) 2 2 1 2 2 H (s) H ( j ) H ( j ) H ( j ) 1 1 2 2 4 2 2 1 1 4 1 1 22 1 1 2n Active Filter (Part 2) s P2 Bode plot (n-th order Butterworth) x 1 1 2n In dB form : Hˆ ( j ) 20 log 1 st -20 1 -60 d or de o rd r er r h or -80 de r de -100 or r suppose 1 3r 5t h Hˆ ( j ) 20 log( 1 2 n ) 2nd -40 o rd e 4t 1 2n voltage gain (dB) Hˆ ( j ) x Hˆ ( j ) 20n log( ) Butterworth response For decade condition, 10 x The Butterwort h filter wou ld have (-20n)dB/d ecade For 2nd order : - 40 dB/decade For 3rd order : - 60 dB/decade For n - th order : - 20n dB/.decade Active Filter (Part 2) Second order Butterworth filter K K' 1 sC4 R1 R2 sR1C3 1 K s 2 R1 R2C3C4 s 2 P s 2 P Q 1 P R1C4 R2C4 RC 1 K 1 3 R2C3 R1C3 R2C4 C3 R R H ( s) QP 1 2 + Setting R1= R2 and C1 = C2 1 1 1 QP 1 1 1 K 1 2 1 K 3 K C4 vin - RB vo RA Now K = 1 + RB/ RA QP 1 3 K 1 R 3 1 B RA 1 R 2 B RA For Butterworth response: 1 QP 2 1 1 QP 2 2 RB RA Therefore, we have 2 RB 2 1.414 RA We define Damping Factor (DF) as: Active Filter (Part 2) DF R 1 2 B 1.414 QP RA Damping Factor (DF) • The value of the damping factor required to produce desire response characteristic depends on the order of the filter. • The DF is determined by the negative feedback network of the filter circuit. • Because of its maximally flat response, the Butterworth characteristic is the most widely used. • We will limit our converge to the Butterworth response to illustrate basic filter concepts. Active Filter (Part 2) Values for the Butterworth response B1 s s 1 B2 s s 2 2 s 1 2.61s 1 s B3 s s 3 2 s 2 2 s 1 s 1 s 2 s 1 B4 s s 4 2.61s 3 3.41s 2 B5 s s 5 3.24 s 4 5.24 s 3 5.24 s 2 3.24 s 1 s 1s 2 0.77 s 1 s 2 1.85s 1 Order Roll-off dB/decade 1st stage poles DF 1 -20 1 optional 2 -40 2 1.414 3 -60 2 4 -80 5 6 2 2nd stage poles DF 1.000 1 1.000 2 1.848 2 0.765 -100 2 1.000 2 -120 2 1.932 2 Active Filter (Part 2) 0.62 s 1 s 2 1.62 s 1 3rd stage poles DF 1.618 1 0.618 1.414 2 0.518 Forth order Butterworth Filter R3 R4 , C3 C 4 R1 R2 , C1 C2 C1 0.01 F R1 8.2 k C3 0.01 F +15 V R2 8.2 k R3 8.2 k + R4 8.2 k -741C C2 0.01 F -15 V RB 1.5 k Vout RB 27 k -15 V RA 22 k RB 1.848 RA RB 0.152 RA + - 741C C4 0.01 F RA 10 k DF 2 +15 V DF 2 RB 0.765 RA RB 1.235 RA RB 0.152R A 0.15210k RB 1.235R A 1.23522k RB 1.52k RB 27.17 k Active Filter (Part 2) Chebyshev Response (Equal-ripple) H j 1 1 2Cn2 Where determines the ripple and Cn2 is the Chebyshev cosine polynomial defined as Active Filter (Part 2) Chebyshev Cosine Polynomials C0 1 C1 C2 2 2 1 C3 4 3 3 C4 8 4 8 2 1 C5 16 5 20 3 5 Cn 2Cn 1 Cn 2 Active Filter (Part 2) Second order Chebychev Response H j 1 1 2Cn2 Example: 0.969dB ripple gives = 0.5, C2 2 2 1 H 22 1 1 2Cn2 1 2 1 0.52 2 1 1 4 2 1.25 H 22 ( ) s / f Roots: s2 1 1 5 1 j 2 2 H 2 s H 2 s 1 2 1 2 1 s j s j 2 2 H 2 s H 2 s 1 Note: 20 log 10 ( s s 1.25 4 2 Active Filter (Part 2) 1 1 2 ) 0.969 Roots Roots of first bracketed term Roots of second bracketed term 1 0 4 j 2 s 2 1 j 2 1 0 4 2 s 2 1 j 2 1 5 1 1 5 1 j 2 4 2 2 4 2 0.566 j 0.899 1 j 1 5 1 1 5 1 j 2 4 2 2 4 2 0.566 j 0.899 Ce j C e j 2 C (cos( ) j sin( )) 2 2 1 1 H 2 s H 2 s . s 0.556 j 0.899s 0.556 j 0.899 s 0.566 j 0.899s 0.556 j 0.899 Note: or H 2 s a bj a 2 b 2 e j tan 1 (b / a ) s 0.556 j 0.899s 0.556 j 0.899 1 1 s 0.5562 0.899 2 s 2 0.112s 1.117 Active Filter (Part 2)
© Copyright 2026 Paperzz