Monte Carlo analysis- analysis variation

Hi can anyone pls tell me the difference between process analysis and mismatch analysis. I thought both were same. Nothing but varying the transistor dimensions to see during fabrication process if some errors causing slight change in dimensions of the transistors are caused, how well the circuit behaves. But then in monte carlo analysis-> in analysis variation why do we have process only, mismatch only and process and mismatch. Can someone pls tell whats the actual difference.


When simulation is run transistor dimensions are varied but say how can i determine in wat range is it varying. Suppose a transistor is 1u/1u (W/L).. Then how can i find out what is varied and of how much value. Becoz i think there is some tolerance right. Meaning we can choose right that variation is 5% or 10%. How can i do and find all that in cadence?

Finally i am looking to plot this graph for a 4Q multiplier.

Like i just give 100 runs in monte carlo and i do get a DC analysis output wer i can see the output varying from one another. But i still cant select each transistor and check how much it is varied and don't understand how to plot it as the above graph. Since i need a histogram in
Results->Plot-> Histogram i get the following

Can anyone pls explain me what is this par 1, par 2 etc. Becoz all seem to give a histogram but i am not sure which one is the right one to plot THD percent. I am not able to determine the X axis if its percent or what. Its just like X(). Would be really great if somebody could help!!!

Perhaps this tutorial can be helpful: View attachment Cadence_Monte_Carlo_simulation_tutorial.pdf

preethi19 said:

... in monte carlo analysis-> in analysis variation why do we have process only, mismatch only and process and mismatch. Can someone pls tell whats the actual difference.

Click to expand...

See p. 15 of the above tutorial.

Thank you for the link. Pls correct me if this understanding is wrong. so Mismatch is something that say if two transistors are matched then their matching is varied by varying W and L and the results are provided. This is within a single die. So process analysis is something were say in a whole wafer my design is in many bits (different dies) and even though the design is same for each die during fabrication process some variation may occur between each die and this result is provided in process analysis. Also is W and L the only factors varied?? I read somewhere doping and some other factors are varied too. I am guessing all the doping relates with process analysis right? Becoz mismatch only involves W and L parameter. Or is only W and L the two parameters always for both process and mismatch???

- - - Updated - - -

In short process analysis -> variation in W and L between two different chips.

Mismatch -> variation in W and L between transistors on the same chip.

But then what is the use of process analysis. Becoz say in my circuit i need two transistors to match so the circuit works fine and i need to be concerned about their matching. And same case with another chip having the same design. Concerned mismatch within each chip. What is the need to check mismatch between different chips? Anyway each chip will perform individually right??? Thanks a lot the help!!!

Too many questions in one single post, preethi!

Seems you didn't study the tutorial. A few answers here in short:

1. Mismatch doesn't only consider variations of W & L, there may be a lot more variables involved, e.g. tox, Vt ... check the "vary" parameters (p.13) .

2. "Mismatch only" MC analysis gives you the possible variations to be expected between two devices on the same chip from one single wafer.

3. "Process only" MC analysis results in the process dependent variations to be expected between two devices on the same chip from any wafer of a single wafer lot. This you need to know if you buy a lot of chips.

4. "Mismatch and process" MC analysis results in the superposition of both variation effects, this is what you have to consider for yield estimation on high volume orders.

You have explained really well and thanks a lot for helping me understand!!!