Weird problem with lowpass filter...

[Luke123]

Hello
I usually don't have problem at all with opamps and active filters.
But I have a response problem with a 4th order low-pass filter.
It is a pair of multi-loop feedback type in cascade.
Stage 1 has F=1500 with Q of 1.35 .
Stage 2 has F=1300 with Q of 1.30 .
When measured from 1000Hz to 20KHz, the response falls off regularly with the exact desired response.
And with the desired roll-off at 24 dB/octave also.
But at around 8KHz, the response flattens and starts to "go-up" at a 6dB/octave rate above 10KHz..
Something that reminds me as a Cauer filter!
This should not happen as my HF response should extend down to the noise floor.
Actually, I had the same result in a FFT analyser, a HP spectrum analyser. Even my low-noise scope easily shows it.
So it is not a noise floor problem.
I tried different opamps. No success. Anyway, the one tested were OP275's .
I also tried the Sallen-key filters type or the Generalized immitance converters (GIC)...no success.
So I guess it's not a topology problem.
The problem occurs and is the slightly the same, for any load or source.
So it seems to me it has something to do with extra phase rotations in opamp.
And I'm sure solution will come from the Opamp dept!
Trying to apply frequency compensation on each opamp? I ever tried it.
Even on several topology. And sometimes different circuit impedance.
>From now on, I don't know anymore from where to start!
Under : FR of circuit and circuit tested.

What do you think?
Anybody can help?
Thanks

Luke123

[sawreyrw]

Luke123,

Bonnie Baker, of EDN magazine, addressed this type of issue recently. See:

http://www.edn.com/contents/images/6356063.pdf

Rick

[Luke123]

Thanks sawreyrw!
Good reading.

Luke123