Hi,

I met a wierd problem, in my fully differential all pole chebyshev active filter, when I change one of the matched resistors 0.2%, system phase and frequency respond totally changed, it comes up zero and stop the attenunation in the stop band, which makes my filter away from the specification. Does anybody know the reason? why such small mismatch can affect that much ? and how to solve this problem?

Before mismatch:

it would be informative if you took and did the pole placements and how they shift due to mismatch.

Some pole-zero constellation placements are  more robust (in spectral response) to shifts than others.

jerry