Hi all,

I am designing a gain-boosted opamp by modifying a folded cascode opamp.

I read the paper by Bult & Geelen and followed the guidelines to get the doublet
out of the way. However, I see mysteriously slow settling components for certain  
cases (not where I am at the limits of my output swing either, but somewhere in the middle).

I check the settling of the individual gain-boosting amps, and they seem to have reached >60 dB settling, although they continue to settle a little. Along with the 60 dB offered by my normal opamp, I expect a 120 dB overall settling accuracy, but see something about 20-30 dB lower.

Is my understanding of the setup wrong? Is it not enough if the gain-boosting auxiliary amps reach 60 dB of gain level?

Please offer your suggestions. They will be greatly appreciated.

Thanks
Vivek

Hi Yawei,

I am actually simulating the settling behavior. The gain-boosted opamps are expected to have a gain > 60 dB.

When I look at the settling behavior of these opamps, I see that the settling error (as defined by Bult to be the ratio of the differential inputs @ virtual ground nodes to the differential output voltage of the opamp) keeps on falling with a constant slope until it flattens out at a level < - 60 dB.

So, the gain-boosting opamps have done their job, and I expect that the overall settling error will be < -120 dB because my main opamp has > 60 dB. But, I see that the overall settling error starts falling off correctly, but suddenly, it changes slope and then on, it settles very slowly. I do not understand this.

The gain-boost opamps are finished settling, and I don't expect a doublet there. The output is not beyond the allowed swing level of my opamp.

In fact, I see that the settling is fine for the largest output swing levels, where I get < -120 dB accuracy, but for some inputs less than this peak swing value, I am seeing the slow settling. For a folded cascodes, this appears to be a strange behavior.

Regards
Vivek

vivkr,

You are welcome. You have shared your experience with us.


Best regards,
Yawei