I am design a circuit with Jazz_semiconductor's PDK. during the extreme condition simulation, eg, power supply variarion, temperature and process variation, I found that the circuit has quite different performance. I just wondering whether or not there are some related documents talking about how to compensation this kinds of variarion?
Do you mean the PCB itself - or electrical design?
For each there are nomber of compensation schemes and design rools (for example - if you need strong independence on variance of passive elements in Filters - than some active solutions are better than others, and Bessel - will be better than Cheb. or Butt. ) Other time teransistors are used, again on other times reflect - diode and transistors are good also - current sources have strong stability over temperature ets...
If you have a specific issue - post it and I'll try to help.
The best is that you design a circuit that is insensitive to the PVT variation.
For a commercial design, you need to meet the specs at all the design corners.
There are some circuits not sensitive with power supply, some circuits not sensitive with temperature. You can find it in razavi's book or P.R. Gray's book.
If you still have problem, you can post it to tell us.
Hi Guys, thanks for your reply. What I am looking for are some common circuits that people usually use to compensate this PVT variation if the robustness is not enough to meet all the extremes.
I think it cannot exist such a general ans so specific document. It depends strictly on your design. For example if you're doing a gain such a gm/R you have to bias the transistor with PTAT current, but your design can be very complex and so it depends on it.
Usually the designer has some design equations and from these equations it's clear what are the parameters to compensate.
PVT is the way to make sure your design. Some SPEC is worst case spec, so when you design in nominal, you spec has to be better than SPEC. If nominal design, then 15% +/- is good enough for some design depending on company.