ECParts Toolkit LogoECParts Toolkit

Engineering Reference

Active Filter Topologies and Approximation Reference

Lookup for Sallen-Key, multiple-feedback, twin-T notch, Butterworth, Bessel, Chebyshev, elliptic, cascaded-stage, and ADC anti-alias filter terminology.

Reading Time
12 min
Format
Topology lookup
Updated
September 29, 2026

Active Filter Topologies

Active filter topology comparison
TopologyStructureResponseIdeal design relationshipPractical boundary
Sallen-Key low-passVoltage-controlled voltage sourceLow-passNatural frequency and Q set by R/C ratios and amplifier gainOp-amp GBW, slew rate, output drive and component tolerance
Sallen-Key high-passVoltage-controlled voltage sourceHigh-passDual form with input capacitors and resistive networkBias-current path, low-frequency saturation and op-amp limits
Multiple-feedback band-passInverting multiple-feedbackBand-passCenter frequency, Q and gain are coupled through component ratiosInput impedance and op-amp dynamic limits
Twin-T notchPassive or actively buffered twin-TBand-stop / notchBalanced R and C ratios establish the ideal notchDepth is highly sensitive to ratio mismatch and loading
Cascaded second-order sectionsBiquad stagesLow-pass or high-passHigher order is realized as first- and second-order stagesStage Q, ordering, gain distribution and headroom matter
ADC anti-alias filterAnalog low-pass before ADCLow-passAttenuates interferers before samplingNyquist alone does not define sufficient stopband attenuation

Common Response Approximations

Filter approximation comparison
ApproximationPrimary propertyMagnitude behaviorTransitionTypical selection reason
ButterworthMaximally flat magnitudeNo passband rippleModerate transitionGeneral amplitude response when flatness is preferred
BesselFlat group-delay emphasisGentler magnitude transitionGood transient fidelityWaveform preservation and time-domain response
Chebyshev Type IEquiripple passbandSpecified passband rippleSteeper than Butterworth for same orderAmplitude tradeoff accepts ripple
Chebyshev Type IIMonotonic passband, ripple in stopbandFinite stopband zerosSteep transition near stopbandStopband behavior is explicitly controlled
EllipticRipple in passband and stopbandFinite transmission zerosSteepest transition for a given orderTolerances and phase behavior require care
Linkwitz-RileyCrossover-oriented summed responseOften cascaded Butterworth sectionsIn-phase acoustic/electrical sum targetSystem phase, driver and implementation context govern

Second-Order Section Terms

Second order active filter terms
TermRelationshipInterpretation
Natural frequencyω0 = 1 / √(R1R2C1C2)Pole-frequency parameter; not universally the -3 dB point
Quality factorQ = 1 / (2ζ)Higher Q reduces damping and can create peaking
Damping ratioζ = 1 / (2Q)Second-order transient and pole damping
Section gainKCan affect Q in Sallen-Key designs
Cascade orderN = sum of section ordersOdd orders include one first-order section
Asymptotic slope20N dB/decadeFar-from-cutoff ideal approximation

Canonical Calculation Anchors

Active filter calculation anchors
CaseResultInterpretation
Equal 10 kΩ / 10 nF Sallen-Key network1591.549 HzNatural frequency from the canonical four-component product
Q = 1/√2ζ = 0.707107Butterworth second-order damping
Fourth-order Butterworth at fc3.010300 dB attenuationCutoff is -3.0103 dB independent of order
Fourth-order Butterworth realization2 second-order sectionsEach section has its own pole Q
10 kΩ / 10 nF balanced twin-T1591.549 HzIdeal notch frequency; depth still depends on matching
48 kHz sample rate24 kHz NyquistPassband and stopband requirements still set filter design

Op-Amp and Implementation Boundaries

Active filter implementation checks
CheckWhy it mattersTypical failure
Gain-bandwidth productMaintains required loop gain around pole frequenciesFrequency and Q shift
Slew rateSupports large-signal output slopeAmplitude-dependent distortion
Input common-mode rangeKeeps both inputs within valid operationClipping or nonlinear behavior
Output swing and driveSupports load and internal reactive currentClipping, heating or Q error
Source and load impedanceChanges assumed network ratiosPole and gain error
Component tolerance and driftMoves poles and notch cancellationResponse spread between units
Noise and resistor scaleSets thermal and op-amp noise contributionsExcess output noise
Layout and decouplingControls parasitic coupling and stabilityUnintended peaking or oscillation

Common Errors

  • Treating natural frequency as universally equal to -3 dB cutoff.
  • Ignoring section Q when cascading a higher-order response.
  • Using one Q for every Butterworth stage.
  • Assuming Sallen-Key gain does not affect Q.
  • Ignoring source and load impedance.
  • Expecting ideal twin-T notch depth with loose component ratios.
  • Selecting an op amp by GBW alone.
  • Ignoring slew rate and output headroom.
  • Placing anti-alias cutoff exactly at Nyquist without a transition band.
  • Using asymptotic slope as exact near-cutoff attenuation.
  • Cascading high-Q stages without checking internal overload.
  • Skipping tolerance and Monte Carlo analysis for critical designs.

Support reference

FAQ

What is a Sallen-Key filter?

It is an active second-order topology using an op-amp as a voltage-controlled voltage source with an RC network that sets natural frequency and Q.

Is natural frequency always the -3 dB cutoff?

No. For a second-order response the relationship depends on Q and response shape. They coincide for a second-order Butterworth low-pass.

What is a Butterworth approximation?

It is a maximally flat magnitude approximation with no passband ripple. At cutoff its magnitude is down 3.0103 dB from the passband gain.

When is Bessel preferred?

Bessel-like responses are chosen when group-delay flatness and transient fidelity are more important than a steep magnitude transition.

What does filter order control?

Order sets the number of poles and the ultimate asymptotic slope. Each pole contributes about 20 dB per decade beyond the transition in the ideal asymptote.

Why is twin-T notch depth sensitive to tolerance?

The ideal cancellation depends on matched resistor and capacitor ratios. Ratio error, source/load impedance and op-amp behavior leave residual signal at the notch.

Why does an active filter need op-amp bandwidth margin?

The op-amp must provide loop gain at frequencies where the section requires gain and Q. Finite GBW shifts poles and reduces or reshapes Q.

Does Nyquist frequency define the anti-alias cutoff?

No. Nyquist is half the sample rate. The filter must also preserve the desired passband and provide enough attenuation over the available transition band.

Can second-order sections be cascaded in any order?

The ideal linear transfer product is commutative, but practical stage ordering changes internal signal level, overload risk and noise.

Do normalized tables replace simulation?

No. They establish ideal targets. Tolerance, op-amp models, source/load interaction, noise, headroom and layout still require verification.

Connected Engineering Content

Related Calculators

Related Guides

Related References