Hi, all. I'm designing a telescopic amplifier with a switched-capacitor CMFB circuit and just found a strange phenomenon in Spectre PAC analysis.

I made my amp to have a gain of 50, but the PAC results showed me only gain of 4. When I run normal AC simulation by connecting 10Gohm resistors to SC CMFB's capacitors in parallel, I was able to verify that my circuit's gain is 50.

PSS waveforms show similar voltage values to DC analysis. So, I'm really confused which AC analysis is a correct one (PAC or normal AC)?

+) It seems that the below post discusses similar problems as me. But, there's no satisfying solution...

http://www.designers-guide.org/Forum/YaBB.pl?num=1251692728

Hmm... maybe my explanation on this issue was not enough. The circuit is a double-cascoded telescopic amplifier with nmos input pair. I used a switched-capacitor common-mode feedback circuit and run PSS/PAC simulation to check the dc gain.

However, the dc gain was much lower than I expected and using the ideal CMFB, I was able to get the dc gain as expected.

Actually, I've just found the reason. There was a reset switch that short the output node to ac ground. It was my mistake....

Thanks.