hi there,

for my fully-differentail fold-cascode OTA with SC-CMFB, the clock frequency, say, can be up to 400kHz. Currently, I set CMFB clock with 200kHz with the same sampling frequency which I will use later.

For my OTA differential inputs, they can be set up to 100kHz.

I am wondering that, in this case, whether I can use QPSS, instead of PSS, to do periodic steady-state simulation, since I thought the inputs can be various fundamental frequencies by name, besides the different names and frequencies for CMFB clocks.

Am I right? please advise.

Hi bunny,

Thank you for your suggestion.

[quote author=bunny  link=1120552636/0#1 date=1120564707]hi ethan,
I too was bugged by the same problem with switched capacitor CMFB. Why dont u try using simple transient
[/quote]

I have tried it with conventional transient analysis with SC CMFB, it also works for common mode output voltage and Vctrl node. But when I added a small sine wave in the inputs, the output peak-to-peak swing was small, not as large as expected.

[quote author=bunny  link=1120552636/0#1 date=1120564707]
analysis followed by ac analysis (The ac analysis should use the operating point calculated by transient analysis at some particular instance). It worked for me.
[/quote]

I have tried conventional AC analysis either. But I found it was hard to control the dc operating biasing since there two half clock phases for CMFB with different DC biasing.

Would you mind describing a little bit more detailed about how you did control that "ac analysis should use the operating point calculated by transient analysis at some particular instance"?


[quote author=bunny  link=1120552636/0#1 date=1120564707]
>Of course PSS and PAC also worked for me.  but the first method makes the things very simple
[/quote]

I guess I have not fully understood PSS and PAC yet.

ethan