NMOS of CMFB in linear region

[armity]

I meet a full differential amplifier.there is a NMOS of CMFB in linear region always. the other simulation's result is right, I don't know that if it ok.who would help me? thanks

 

[suria3]

Re: about CMFB

CMFB being in linear (triode) region is something not good. Your error amplifer won't be able to provide a constant gain as not in saturation, thus this will create a offset in your dc reference voltage which will finally effect your common mode dc voltage. So, try to make your CMFB always in saturation region.

 

[armity]

Re: about CMFB

does it make the amplifier unstable?
thanks!

 

[suria3]

Re: about CMFB

does it make the amplifier unstable?
thanks!

Regarding the stability matter, it doesn't mean that when your transistor is in linear region your circuit is unstable, it will still give you good phase margin as long your loop gain is high, but if many of your transistor goes into linear region, than you won't be getting enough loop gain in which finally make the circuit become unstable upon reaching negative gain.

 

[sachinagg77]

Re: about CMFB

Sometimes transistors in linear region that act as RESISTORS are used in the CMFB loop. Such CMFB circuits do not use an external reference.

For such CMFB implementations:
1. The loop gain is normally poor
2. Circuit component parameters decide the reference common mode level

Because of the above two limitations, there might be lot of variations in the output common mode level depending on process corners. If this is acceptable, you can use this form of implementation.

 

[zyyang]

Re: about CMFB

hi
you mean your sensing circuit in CMFB circuit? if true it can work in triode region.
because the sensing circuit won't contribute gain.