Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
like any other statistical method, you need enough samples to build confidence that you are capturing the behavior of the overall population. it can be 10. it can be 1000.
Thanks for you answer. But how can i know that i have enough samples. For example i simulate a operational amplifier and look on offset voltage. If i have 100 point for monte carlo simulation, i have correct result? Or i should take 1000/2000 to plot correct curve of distribution?
The number of points needed is determined by the statistical confidence level you want.
There is no "correct result" in statistics, only a good result.
For a better understanding, you need a basic understanding of statistics.
Looking at the mean and standard deviation of the results will give you an idea of how good your Monte-Carlo data is in determining the expected deviation in a production run.
100 points usually for me gave a pretty good standard deviation for an OPAmp offset voltage, sometimes 30 is enough if the design is in an initial state, not bad, but to see the contribution percentage of the devices to the offset an 1000 points are required at least. These are just kind of thumb rules, but above answers covered this topic.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
