Questions about Opamp stability & linearity

[zedrck]

Opamp Stability

Hi evrybody,
I am currently trying to design a 1GHz low-pass filter with this circuit (cf question.pdf)

I have always worked with circuits running at lower frequencies; so in these
cases, a clean design would have remained to set the cut-off frequency of
the filter with Cs and Cf (like in Butterworth filter config...) and take a marge
of 20dB between the closed loop gain (0dB) and the open loop gain at this
cut-off frequency;
but in this case, it would mean that the opamp regulates the loop with 20dB at 1GHz... and then shows a GBW=10GHz!!!!!

So my question is:
- if my opamp is slower (GBW=3GHz!!!! ;-) ), what´s the matter if Cs = 0 F
and Cf=0 F.... ?
If the cut-off frequency is set by the opamp itself, should I expect any kind of
problem regarding stability, linearity...or whatever?

Thanks in advanced for sharing your minds!

 

[FvM]

Re: Opamp Stability

Hello,

there are different things to say. If you are talking of an amplifier with external feedback circuitry, the question is hypothetical to my opinion. You won't be able to operate it stable with external feedback when it has a GBW of 3 GHz. This may be able with on-chip feedback, possibly also with very small thinfilm hybrid circuits, but is unlikely with housed ICs. But you may try.

Apart from this, filter charateristics will be affected surely by limited GBW. Typically it's impossible to realize low Q filter characteristics as Bessel or Gauss, respectively additional poles are introduced. At best parameters are optimized empirically. Above 0,5 or 1 GHz, also electrical circuit delays must be considered, lumped element models becoming inaccurate.

Regards,
Frank