Hi,

In all analysis I have seen of S/H amps, the kT/C noise is calculated for the input sampling cap.  What about for the feedback capacitor that is wrapped around the op amp?  

Imagine this feedback capacitor, Cf, is reset by a parallel switch connected between its terminals so ideally it has no charge.  When this reset switch turns off, is there now a kT/C noise voltage across this capacitor and now appearing at the ADC output?  Why or why not?  I am very confused about this.

My thought is that yes, there is a kT/C voltage on the feedback capacitor between the virtual ground of the op amp and its output in this case.  But, it doesn't matter because it is divided by the op amp gain when it is referred to the input so it's value is practically nothing compared to the kT/C noise of the sampling cap.

Do you all agree?

Carl