Skip to main content

Sallen-Key LPF for 1/(s²+0.765s+1)

Short answer

Sallen-Key LPF with C1 = C2 = 0.01 µF, equal feedback resistors RA = RB =10 kΩ (K=2). For ω0=1 rad/s: R1 =130.7 MΩ, R2 =76.5 MΩ. Scaled to ω0=10⁴ rad/s: R1=13.07 kΩ, R2=7.65 kΩ. To get unity dc gain, split R1 into two 26.1 kΩ resistors.

Derive the transfer function of Sallen-Key low pass filter. Design a filter for T(s) = 1/(s² + 0.765s + 1) using Sallen-Key biquad. In your final design the values of capacitors must be 0.01μF and feedback resistors should be equal.

Sallen-Key low-pass biquad

The Sallen-Key (positive-feedback, VCVS) biquad uses one op-amp as a non-inverting amplifier of gain KK with an RC network.

                 C1
          +------||------------+
          |                    |
Vin--R1---+A--R2--+B---(+)     |
                  |     OA >---+--- Vo
                 C2  +-(-)     |
                  |  |         RB
                 GND +---------+
                     |
                     RA
                     |
                    GND

Amplifier gain K=1+RB/RAK = 1 + R_B/R_A, so VB=Vo/KV_B = V_o/K.

Derivation. Node B: VA−VBR2=sC2VB⇒VA=VB(1+sR2C2)\dfrac{V_A - V_B}{R_2} = sC_2V_B \Rightarrow V_A = V_B(1 + sR_2C_2).

Node A:

Vin−VAR1=VA−VBR2+sC1(VA−Vo),VB=VoK\frac{V_{in} - V_A}{R_1} = \frac{V_A - V_B}{R_2} + sC_1(V_A - V_o), \quad V_B = \frac{V_o}{K}

Eliminating VAV_A and VBV_B:

T(s)=VoVin=KR1R2C1C2s2+s(1R1C1+1R2C1+1−KR2C2)+1R1R2C1C2T(s) = \frac{V_o}{V_{in}} = \frac{\dfrac{K}{R_1R_2C_1C_2}}{s^2 + s\left(\dfrac{1}{R_1C_1} + \dfrac{1}{R_2C_1} + \dfrac{1-K}{R_2C_2}\right) + \dfrac{1}{R_1R_2C_1C_2}}

so ω0=1R1R2C1C2\omega_0 = \dfrac{1}{\sqrt{R_1R_2C_1C_2}} and ω0Q=1R1C1+1R2C1+1−KR2C2\dfrac{\omega_0}{Q} = \dfrac{1}{R_1C_1} + \dfrac{1}{R_2C_1} + \dfrac{1-K}{R_2C_2}.

Design for T(s)=1s2+0.765s+1T(s) = \dfrac{1}{s^2 + 0.765s + 1}

Step 1: Parameters.

ω02=1⇒ω0=1 rad/s,ω0Q=0.765⇒Q=10.765=1.307\omega_0^2 = 1 \Rightarrow \omega_0 = 1\ \text{rad/s}, \qquad \frac{\omega_0}{Q} = 0.765 \Rightarrow Q = \frac{1}{0.765} = 1.307

Step 2: Meaning of the conditions. The capacitors are equal, C1=C2=C=0.01 μC_1 = C_2 = C = 0.01\ \muF. The "feedback resistors" are the gain-setting resistors RAR_A and RBR_B of the op-amp, so RA=RBR_A = R_B and

K=1+RBRA=2K = 1 + \frac{R_B}{R_A} = 2

Step 3: Find R1R_1 and R2R_2. With C1=C2=CC_1 = C_2 = C and K=2K = 2, the R2R_2 terms cancel in the ω0/Q\omega_0/Q expression:

ω0Q=1C(1R1+2−KR2)=1R1C,ω0=1CR1R2\frac{\omega_0}{Q} = \frac{1}{C}\left(\frac{1}{R_1} + \frac{2-K}{R_2}\right) = \frac{1}{R_1C}, \qquad \omega_0 = \frac{1}{C\sqrt{R_1R_2}} R1=Qω0C,R2=1Q ω0CR_1 = \frac{Q}{\omega_0 C}, \qquad R_2 = \frac{1}{Q\,\omega_0 C}

For ω0=1\omega_0 = 1 rad/s and C=0.01 μC = 0.01\ \muF =10−8= 10^{-8} F: R1=1.307/10−8=130.7 MΩR_1 = 1.307/10^{-8} = 130.7\ \text{M}\Omega and R2=0.765/10−8=76.5 MΩR_2 = 0.765/10^{-8} = 76.5\ \text{M}\Omega.

Because the given T(s)T(s) is normalized (ω0=1\omega_0 = 1 rad/s), these resistors are impractically large. Frequency scaling to a practical ω0=104\omega_0 = 10^4 rad/s (same capacitors, same KK) gives

R1=1.307104×10−8=13.07 kΩ,R2=0.765104×10−8=7.65 kΩ,RA=RB=10 kΩR_1 = \frac{1.307}{10^4\times10^{-8}} = 13.07\ \text{k}\Omega, \qquad R_2 = \frac{0.765}{10^4\times10^{-8}} = 7.65\ \text{k}\Omega, \qquad R_A = R_B = 10\ \text{k}\Omega

Step 4: Gain. The dc gain of this circuit is K=2K = 2, but the given T(s)T(s) has unity dc gain. The excess gain is removed by splitting R1R_1 into a divider with α=1/K=0.5\alpha = 1/K = 0.5: R1a=R1/α=2R1R_{1a} = R_1/\alpha = 2R_1 and R1b=R1/(1−α)=2R1R_{1b} = R_1/(1-\alpha) = 2R_1 (about 26.1 kΩ\Omega each for ω0=104\omega_0 = 10^4 rad/s). This keeps ω0\omega_0 and QQ unchanged.

Answer: C1=C2=0.01 μC_1 = C_2 = 0.01\ \muF, RA=RBR_A = R_B (K=2K = 2), R1=130.7R_1 = 130.7 MΩ\Omega, R2=76.5R_2 = 76.5 MΩ\Omega for ω0=1\omega_0 = 1 rad/s (13.07 kΩ\Omega and 7.65 kΩ\Omega if scaled to 10410^4 rad/s). For unity dc gain, replace R1R_1 by R1a=R1b=2R1R_{1a} = R_{1b} = 2R_1.

Alternative reading (equal R1=R2R_1 = R_2). If "equal resistors" means R1=R2=RR_1 = R_2 = R with C1=C2=CC_1 = C_2 = C, then ω0=1/RC\omega_0 = 1/RC and ω0/Q=(3−K)/RC\omega_0/Q = (3-K)/RC, so K=3−1/Q=3−0.765=2.235K = 3 - 1/Q = 3 - 0.765 = 2.235. This needs RB=1.235RAR_B = 1.235R_A (e.g. RA=10R_A = 10 kΩ\Omega, RB=12.35R_B = 12.35 kΩ\Omega) and R=1/(ω0C)=100 MΩR = 1/(\omega_0 C) = 100\ \text{M}\Omega for ω0=1\omega_0 = 1 rad/s (10 kΩ\Omega for ω0=104\omega_0 = 10^4 rad/s). For unity dc gain, α=1/2.235=0.4474\alpha = 1/2.235 = 0.4474, R1a=2.235R1R_{1a} = 2.235R_1 and R1b=1.810R1R_{1b} = 1.810R_1.

✓ Answer checked by Claude Sonnet 5.5

All Filter Design chapter 6 questions →