Differential Switched CapacitorAmplifier Accuracy

Hi everyone,
I am currently trying to implement my very first switched-capacitor non inverting Amplifier that I described in attachment (fully differential scheme)
The principle of the circuit is described in many books:
- at the end of phase (a): the OTA has stored his own offset value over C2 (not shown on this picture)
- at the end of phase (b): the input signal is amplified with a quiet precise ratio given by C1/C2, cancelling the offset of the OTA with the charge stored in C2 during the precedent phase.
- in my case I try to realize a high gain with: C1=10pF, C2=100fF, C1/C2 = 40dB

Looking at the Single Ended topology I know that this circuit is quiet insensitive to the input capacitance of the OTA. (it is at least damped by its open loop gain)
But when I try to verify this principle with a fully differential circuit (cf. attachment) I find a significant offset shift at the end of phase (b) when the OTA input capacitances are unbalanced (on pin in+ and in-).

My question is:
- is it reasonable to expect such a circuit to achieve very low offset (of few micro volts) or is this circuit only a "study" case given by many books?
SC Amp are known for being ultrra accurate but I can´t imagine input capacitances to be completely matched on both sides of the OTA inputs even after a good layout.
- does anyone know any trick to achieve the promised offset cancellation feature of this circuit?

In advance many Thanks for sharing your experiences!

Cheers
Zedrck

zedrck said:

... (it is at least damped by its open loop gain)

Click to expand...

Don't think so: The damping-factor = Open loop gain / closed loop gain . With your 40dB closed loop gain, there might not be much damping factor left over. Try with smaller closed loop gain!