Hello
I am designing a 4th order switched-capacitor delta-sigma modulator. During mismatch analysis (transient) I observed that the modulator was  power-on  unstable for the default gmin value (1pS) . Decreasing  gmin to 100fS the modulator appears to be stable for the power-on conditions.  
Though I have provided a control signal to reset  the integrators - however, I would still like to understand the cause and implications of power-on instability for the default gmin value.  Also, gmin value appears of consequence only when mismatches are introduced in the modulator- the mismatch free modulator is power-on stable.

Thanks
Bupesh