One question is about CMFB in folded-cascode OPAMP.

[skymusic]

Dear all,
A folded-cascode OPAMP is shown in fig. When the folded-cascode OPAMP is with CMFB, all MOS transistor operate in saturation region, and gain is 80dB. Use tt corner to simulate the folded-cascode OPAMP without CMFB, all of MOS operate in saturation region, and the gain is about 80dB. But in ss corner, M3 and M4 operate in triode region, in ff corner, M7 and M8 in triode region, so the gain is down to about 40dB.
So I ask whether I must adjust W and L of circuit to let all of MOS to operate in saturation region, in folded-cascode OPAMP without CMFB in any corner, and whether it affect the yield and robust of circuit.
Thanks.

 

[tsb_nph]

Hi,
In general, it is a bad idea to fabricate a high impedance node (in this case Vout) without any means to stabilize its DC operating point. Though some simulation may show proper operation without the CMFB, try applying some mismatches in the transistors, then you will see that the cirucit will not operate properly.
According to my experience, your suggestion of using no CMFB for a high impedance output node is going to give a very bad yield - You would be better off with a CMFB circuit.

Bharath

 

[dwilliam]

Hi,
a fully differential opamp wil simply not work correctly without a CMFB circuit. The process corners, temperature and mismatches will saturate the outputs of your opamp. If you find difficult to adjust the operating point of the cascodes, them maybe you have too much saturation margin on the current sources, try to reduce them, you will gain margin on the cascodes...

 

[avlsi]

Hi,

In a fully differential amplifier ,we need a CMFB to control the common mode output voltage.So,gain will decrease as we are using a negative feedaback here. And more over the PMOS transistors of the folded cascode in the common gate node must operate in saturation in order to give you a swing ,when u apply a sinosoidal signal.it is next to impossible to get a swing in linear region. Please check the bias points at all corners and make changes.

 

[survivor]

In addition to all process corner, temperature and voltage variation, remember to consider bias current variation.

So the simulation of total combination should be;
5(process corner)x2(-20C,100C)x2(3V,3.6V)x2(bias variation)=40.

That will guarantee your yield !

 

[skymusic]

I know folded-cascode OPAMP must have CMFB. If not all MOS transistors operate saturation region, when a folded-cascode OPAMP is without CMFB in ss and ff corner. But the folded-cascode OPAMP is with CMFB, all MOS transistors are in saturation region, and her gain is 80dB. Is this folded-casecode OPAMP OK?

 

[fxxjssc]

hi, all
SC CMFB used in my design, but you know it can not performe ac analysis. do you have good solution besides simulation with Spectre. Thanks

 

[winsonpku]

To run an AC sim on a SC circuit, you first run a transient sim, let it
settle to the desired operation condition, save the operation condition,
and run the AC sim using the operation condition.

hi, all
SC CMFB used in my design, but you know it can not performe ac analysis. do you have good solution besides simulation with Spectre. Thanks

 

[sixth]

hi, all
SC CMFB used in my design, but you know it can not performe ac analysis. do you have good solution besides simulation with Spectre. Thanks

Hi,
If you use candence IC deisgn tools, you can use spectreRF to simulate the ac analysis. First step is to do PSS analysis and then do PAC analysis, and thus you can get what you want.
Good luck!
sixth

 

[tsb_nph]

hi, all
SC CMFB used in my design, but you know it can not performe ac analysis. do you have good solution besides simulation with Spectre. Thanks

Hi,
Here is a very good document on how to perform AC simulation on a switched capacitor filter.

https://www.designers-guide.org/Analysis/sc-filters.pdf

 

[jgarciarivera]

Hello,
All amplifiers with a high impedance output require a common-mode feedback to control the common-mode output voltage. I simulated my amplifier using AMI mosis models and specifically one simulation corner (I mean, ff,ss) my amplifier failed to amplify. Then I used a Complementary Common-mode feedback which is like the complementary differential pair. When one CMFB turns on, the other is off and viceversa.

 

[fxxjssc]

thanks all,
I have got solution and simulation result
EDAboard is full of kind hearted man