Skip to main content

RC-CR transformation with examples

Short answer

RC-CR transformation converts a low-pass to high-pass filter by replacing each resistor Ri with a capacitor Ci'=1/Ri, and each capacitor Cj with a resistor Rj'=1/Cj, leaving gain-setting resistors unchanged. The transfer function becomes THP(s)=TLP(1/s). Examples: first-order LP becomes HP; Sallen-Key LP becomes HP with same Q and reciprocal normalized ωo.

Explain RC-CR transformation with suitable examples.

The RC-CR transformation converts an active RC low-pass filter into a high-pass filter (and vice versa) by interchanging resistors and capacitors.

Rule (normalized values, ωo=1\omega_o = 1):

  • Each resistor RiR_i becomes a capacitor Ci′=1/RiC_i' = 1/R_i.
  • Each capacitor CjC_j becomes a resistor Rj′=1/CjR_j' = 1/C_j.
  • Resistors that only set a dimensionless gain ratio (op-amp gain resistors RA,RBR_A, R_B) are not changed.

Why it works: a resistor's admittance 1/Ri1/R_i becomes sCi′=s/RisC_i' = s/R_i, and a capacitor's admittance sCjsC_j becomes 1/Rj′=Cj1/R_j' = C_j. In both cases the new admittance is s Y(1/s)s\,Y(1/s), where Y(s)Y(s) is the old one. The transfer function depends only on ratios of admittances, so the extra factor ss cancels and only the replacement s→1/ss \to 1/s remains:

THP(s)=TLP(1s)T_{HP}(s) = T_{LP}\left(\frac{1}{s}\right)

Every pole pp moves to 1/p1/p. So QQ is unchanged, and ωo\omega_o becomes 1/ωo1/\omega_o (for a normalized filter with ωo=1\omega_o = 1 it stays 1). After the transformation, frequency scaling moves the response to the wanted frequency.

Example 1: first-order RC low-pass. Series R=2 ΩR = 2\ \Omega, shunt C=1C = 1 F:

TLP=11+sRC=0.5s+0.5T_{LP} = \frac{1}{1 + sRC} = \frac{0.5}{s + 0.5}

Transformation: the series resistor becomes a series capacitor C′=1/R=0.5C' = 1/R = 0.5 F, and the shunt capacitor becomes a shunt resistor R′=1/C=1 ΩR' = 1/C = 1\ \Omega:

THP=sR′C′1+sR′C′=0.5s1+0.5s=ss+2=TLP(1s)T_{HP} = \frac{sR'C'}{1 + sR'C'} = \frac{0.5s}{1 + 0.5s} = \frac{s}{s + 2} = T_{LP}\left(\frac{1}{s}\right)

The pole moves from −0.5-0.5 to −2-2, the reciprocal of its magnitude. For the normalized case R=C=1R = C = 1 the result is 1/(s+1)→s/(s+1)1/(s+1) \to s/(s+1).

Example 2: Sallen-Key low-pass to high-pass

LP:  Vin-R1-+-R2-+-->(+)       C1 from node to Vo
            |    C2 to ground

HP:  Vin-C1'-+-C2'-+-->(+)     R1' from node to Vo
             |     R2' to ground

For the normalized circuit:

Ks2+1Qs+1  ⇒  Ks2s2+1Qs+1\frac{K}{s^2 + \frac{1}{Q}s + 1} \;\Rightarrow\; \frac{K s^2}{s^2 + \frac{1}{Q}s + 1}

With equal-element values (R=C=1R = C = 1, K=3−1/QK = 3 - 1/Q), the HP circuit also has C′=R′=1C' = R' = 1 and the same KK. Frequency scaling then gives Ks2s2+(ωo/Q)s+ωo2\dfrac{Ks^2}{s^2 + (\omega_o/Q)s + \omega_o^2}.

Example 3: band-pass. For TBP(s)=(ωo/Q)ss2+(ωo/Q)s+ωo2T_{BP}(s) = \dfrac{(\omega_o/Q)s}{s^2 + (\omega_o/Q)s + \omega_o^2}, replacing ss by 1/s1/s gives

1Qωo ss2+1Qωos+1ωo2\frac{\frac{1}{Q\omega_o}\,s}{s^2 + \frac{1}{Q\omega_o}s + \frac{1}{\omega_o^2}}

This is another band-pass response with the same QQ and the same peak gain, but its centre frequency is 1/ωo1/\omega_o. For a normalized circuit (ωo=1\omega_o = 1) the response is unchanged, so a band-pass circuit maps into another band-pass circuit.

✓ Answer checked by Claude Sonnet 5.5

All Filter Design chapter 6 questions →