CMFB of S/H cant work!

[urian]

Hi,there.
Recently I am designing a S/H using charge transfering technolog, whose op includes a two-stage amp with a traditional SC CMFB. The VDD is 3V and choosing 1.5V as output CM voltage. When finishing the amp and correct its bias, I simulate the whole S/H and find that the output CM voltage is around 1V and the SC CMFB seems not working. I have checked the transient OP and find all transistors are at saturation region. Then I use a ideal CMFB and find the result is so odd, for the Vcmfb is 100+ voltage. I dont know why, any comment would be greatly appreciated! And below is the amp.

Regards
urian

 

[FvM]

Where did you get this strange amplifier topology?

For reasonable gain and common mode range, I would expect p channel transistors in place of M1 and M2 and the output stage driven from the top (M10/M12) rather than from the bottom (M5/M6).

 

[urian]

Hi,FvM
for this two-stage topology there is no need for compensation.

Do you mean the cm voltage for the output stage input is a bit low, so you choose the top p transistors (M10/M12) to be driven?

 

[FvM]

for this two-stage topology there is no need for compensation

Because the input stage gain is absorbed by M1/M2 in common drain circuit? This doesn't look like an advantage to me.

 

[Junus2012]

Hello
I think you want to implement folded op-amp, but you have done a mistake in your circuit , plz look at again

 

[tompham]

Hi all

I dont think urian make a mistake here. He wants to design a 2 stages fully differential amplifier but he need one more circuit for CMFB.

 

[Junus2012]

strange two stage

 

[urian]

Hi
yes,this is a two-stage amp without special compensation.And its CMFB is a traditional SC CMFB circuit not shown.
I wonder should I bias the output op of the amp to the desire output CM voltage or just let the cmfb circuit to do it.

 

[FvM]

I wonder should I bias the output op of the amp to the desire output CM voltage or just let the cmfb circuit to do it.

The CMFB circuit will do - if the amplifier circuit can reach the intended operation point at all and the CMFB gain has correct polarity.

100 V CMFB control voltage with low common mode output suggests wrong polarity, because the control voltage has inverting action in your circuit.

 

[urian]

Though I think CMFB can do these, I find the output CM is about 400mv lower than 1.5V with all transistor working active region.
And the polarity is not wrong, because when I changing the polarity, the CMFB control voltage gose even higher, around 300V.

 

[FvM]

Wrong polarity was just a guess, but amplifier not working correctly has been stated by all contributors to your thread.

 

[urian]

yes,i think so,there must be some problems with the amp
For the output CM voltage is about 500mv lower than the desired 1.5V. When connecting as negative feedback in SH, during sample phase the amp is unity buffer and its input CM equals the output CM, and is around 900mv-1V. Through now I can raise the W/L ratio of (M1/M0) to raise the output CM to come close to the desired value with cmfb active, the input transistors can easily go into triode region. The interaction of input and output cause is very hard to adjust, I think.