Your Filter Is Lying to Its Driver Here’s How to Make It Stop

Ever wonder why your driving amplifier occasionally misbehaves when you hook it up to a filter you designed? The problem might not be the amplifier it’s what your filter is presenting to it.

In a tee-configuration low-pass or high-pass filter, the input impedance in the stopband rises without any upper limit. That means the amplifier “sees” something wildly different from what you intended, and that can create stability problems. The filter thinks it’s doing its job. The amplifier thinks it’s driving a mystery load. Everyone’s unhappy.

John Dunn’s tutorial on EDN walks through a partial remedy that’s remarkably elegant: pair a low-pass filter with a high-pass filter, each feeding its respective load. The result is that input impedance in the passband and crucially, in the stopband now tends toward the load resistance values (chosen here as 50Ω), which makes the driving amplifier far happier.

The corner frequency null doesn’t disappear, but the wild impedance excursions above and below it are tamed. Think of it as teaching two filters to work as a team rather than each doing its own thing in isolation.

This technique isn’t new, but it illustrates a principle that’s easy to forget in an era of active filter ICs and digital correction: sometimes the most stable system is the one where every stage knows exactly what impedance it’s talking to.