The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Strange problem with Monte Carlo sim. in Spectre
https://designers-guide.org/forum/YaBB.pl?num=1137412750

Message started by Roberto on Jan 16^th, 2006, 4:08am

Title: Strange problem with Monte Carlo sim. in Spectre
Post by Roberto on Jan 16^th, 2006, 4:08am

Hello everyone,
I am using the Monte Carlo analysis within Spectre for transient simulations as well as PSS simulations. I found that for some of my schematics, the simulator is not able to correctly print the different result of each iteration. It actually presents all the iterations results as they were identical (for example while plotting family curves, all iterations are superimose to eachother). For some other schematics, it works properly, showing different results for each iteration. I did a lot of trials, I found that the only difference between the working schematic and the others are that in the working one the definition of the global source "VSS!" (my ground) is done by means of the Assura interface (Setup->Stimuli->global sources), while in the failling ones the VSS! is defined on the schematic by means of DC generator.
I tried to change the definition of my VSS! in the schematics where it was defined with the generator, and the problem get solved in some case, but not in all of them.

This looks pretty much like a bug, anyone has found that before?

(In the cases where results are identical for each run, the histograms of the various parameters shown that the statistical choice of the parameters value is running properly, then the bug involve only the plotting of the results I suppose)

Thank you very much for your kind help,
Roberto, University of Udine

Title: Re: Strange problem with Monte Carlo sim. in Spect
Post by Andrew Beckett on Jan 17^th, 2006, 9:24pm

There are a couple of other explanations (which are probably more likely):

You're asking for mismatch variation, when the models are not set up to do mismatch (the models need to be in subckts; see http://www.designers-guide.org/Modeling/montecarlo.pdf
The expression you're calculating is referring to a fixed result directory (i.e. it has the ?resultsDir arg in the data access part of the expression) - and so is not seeing the monte carlo results.

Andrew.