pnoise timedomain

Hi Jason,

The thing in my mind was that in the case of SC integrator the circuit doesn't clear its memory, so noise at the output , that is at the end of phase 2 is not only dependent on what was sampled in the previous phase 1 but also on what was accumulated from previous integrations. However, what Andrew said still makes sense, I think, because although the circuit is accumulating noise samples the next block looks at the aggregate result at the end of phase2 or maybe at the next phase 1, depending on the arrangement.

Thanks

Svilen

Hi，Svilen:

Thanks for your input,in a sigma delta loop,the 1st integrator's output is sampled by the 2nd integrator,they work synchronously.According to your explanations ,t seems that i should insert an ideal sampler at end of phase1 instant,and at the output of 1st integrator.

An ideal track and hold switch(a switch and a cap),we all know the noise samples are uncorrelated in different cycles,but in a sigma delta loop,the noise samples are correlated in different cycles.

I just refer the output noise to the input by divide the opamp gain(the real integrator's gain is the opamp open loop gain),if not,how could i calculate the snr of the loop,i don't know if i'm right.

Thanks

Jason

Hi Jason,

An ideal track and hold switch doesn't have a memory in the sens that from sample to sample and given small enough time constant the noise sample is practically uncorrelated from the previous noise sample at the previous sample moment. Not the same with integrators, they keep memory of the previous noise sample that was put on the feedback capacitor.  Then whether you look at the output of the integrator in phase 1 or phase 2 is secondary and depends on when the integrator output has been sampled by the next stage.
In the case of integrators it is a bit tricky to refer noise to the input (input of integrator, not input of the OTA). There is a paper by Richard Schreier about calculating noises in switched capacitor circuits. It is worth reading. It also touches on the noise of the integrators. I think that the content of this paper appears as one of the appendices of his green book on sigma-delata converters.

Thanks

Svilen