Maybe silly question about first-order sigma-delta modulator

I'm designing a simple first-order sigma-delta modulator with RC integrator.

Now I want to define the minimume and maximum sampling frequency of this modulator. Of course, minimume fs can be found by investigating the integrator output voltage --- if the fs is too low, the output voltage will be too high to implement.

However, the maximum fs is hard to achieve. Will the system be unstable when fs is too high? Is the maximum sampling frequency related to the bandwidth or gain of the integrator? And How to predict the frequency?

I'm a rookie in this field and maybe it's a silly question:)

But still hope to see your opinions. Thanks.

Reply to
kanjun.shi
Loading thread data ...

Walk around the loop and add up the delays. Don't count the integrator. When the delay is such that the news that the output of the flip-flop changed doesn't get back around to the input with enough setup time, the modulator will stop working right.

Reply to
MooseFET

Thanks for your help!

I added some delay into my D flip flop and found that SNR started to decrease when the delay is larger than the sampling period.

I'm not sure what I did is proper for testing the effect of the delay. And also, I haven't seen any unstable things happen in the modulator, the only change is that SNR dropped about 10dB.

Does it prove that the maximum sampling frequency is just (1/delay time)?

Look forward to your reply.

Reply to
kanjun.shi

t

The real situation is alway more complex than the simple model. The delay in logic circuits an comparators is not really fixed times. Comparators have less delay at larger overdrive levels. When the noise increases, it means that the comparator is getting a bigger signal so its delay decreases.

Most of the very fast comparators are really just very high gain amplifiers. They don't contain any positive feedback because adding it would slow them down. This means that the rise time of the output depends on the overdrive like the delay time does. The logic that is connected to this will not really have a simple setup time. A slower signal will need more setup than a faster one because the input section also looks like an amplifier.

Reply to
MooseFET

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.