UMC018 Monte Carlo simulations

[Braski]

Hi all.

I am trying to run MC in ADE XL 6.1.4.

My PDK provides the models, but the procedure to use, as given in some pdf in the docs, is very strange.
Has anyone used the monte carlo models before?

I included a .lib.scs which has the mc section, and i defined the required sigma parameter. But this isn't enough. In fact, i managed to run mc, but all the runs are equal. Please help

 

[antaryami.mt.er09]

Well i did it with TSMC .18um in Virtuso ADE .
U have to select the no of runs , then proces only( which is for unCox , w/l etc) can be thought of as device level, or
Mismatch ( is between two identical device with same Vgs , mismatch in current or with identical Ids but mismatch in Vgs) as ckt level or both
Then u go for simulations and define co-relations, then select on schematic to select the devices ( select all for process variations) and define the
coefficient which should be less than 1 (1 means all the devices are matched) , if u select 0.3 that corresponds to 30% .

Or u can sweep the parameters and do the same way...
I have done this way.

 

[toldo83]

Hi Guys,
can you give more detailed info about monte carlo simulation with ADE XL, please?
However, antaryami.mt I don't get your point "then u go... to 30%", how do you define co-relations?
Braski, have you managed with ADE XL?

 

[Braski]

Yes i have managed to do my MC runs. However, i have just given some tries to get a clue about them. Not in depth. however, I have found the MC models into my model directories, added them to the model files in the test editor. Then, you can do what antaryami told us. BAsically, you should select MonteCarlo from the menu instead of single run, be sure of include your mc models, with the desired option and then launch it. It works quite fine. However, i have still to properly understand all the options.