Investigating High Frequency Feedthrough in a Sallen Key Low Pass Filter

A Sallen Key low pass filter can follow its expected response near cutoff and then provide less rejection than expected at much higher frequencies. The ideal transfer function assumes an ideal amplifier. Real output impedance and parasitic paths become important when that assumption no longer holds.

Recognize the nonideal path

The TI Active Low Pass Filter Design report explains how amplifier output impedance affects the high frequency behavior of a Sallen Key stage and shows an additional output RC network as one possible improvement. That example is a starting point for analysis, not a component prescription for every filter.

Mark the capacitive paths connecting input, amplifier output, and ground. Determine how their impedances compare at frequencies where amplifier feedback no longer maintains the ideal response.

Measure the relevant stopband

Sweep far enough to include the actual interference frequencies, not merely one decade above cutoff. Keep source amplitude and impedance controlled. Measure the input at the board as well as the output when cable loss or generator bandwidth could change the applied level.

Check the measurement floor by removing the intended signal path or using a suitable reference connection. Coupling directly between instrument cables can resemble filter feedthrough. Keep fixture geometry consistent so an apparent improvement is not caused by moving a cable.

Evaluate a correction with its load

If an output RC pole is considered, calculate its effect on the required passband and downstream settling. Include the receiving input impedance and any sampled input behavior. A passive correction that improves rejection can also increase source impedance seen by the next stage.

Assess amplifier stability with the proposed load connection. Do not simply attach a large capacitor directly to the output and assume the resulting circuit remains well behaved. Use a supported isolation arrangement where needed.

Check the complete operating envelope

Repeat the response measurement at representative signal amplitude, supply voltage, and output loading. Look for compression or distortion separately from small signal attenuation. A large unwanted input can create an in-band error through overload even if a small signal sweep looks satisfactory.

Save the measured frequency range, dynamic range, source and load conditions, and circuit revision. Compare the final rejection with the system’s actual interference requirement. The resulting claim should identify the verified stopband rather than extending an ideal mathematical slope indefinitely. This is especially useful when the filter precedes a converter or detector that remains sensitive well beyond the desired signal bandwidth.