+ Post New Thread
Results 1 to 3 of 3
  1. #1
    Newbie level 2
    Points: 766, Level: 6

    Join Date
    Jan 2010
    Location
    Austria
    Posts
    2
    Helped
    0 / 0
    Points
    766
    Level
    6

    MonteCarlo Design Input Parameter Variation

    Hello,

    I'm busy trying to find a easy way to vary parameters within my schematic during monte-carlo analysis by the use of veriloga-models. I tried to used the functions $random or $rdist_normal, but those seem not to vary throughout the different iterations, in a DC simulation I get a peak at a varying value but no standard deviation.

    I read that $arandom or newer versions of rdist_normal support this feature, but my version of mmsim does neither accept more than 3 parameters for rdist_normal nor recognize $arandom as a function name.

    If anybody could help me with this topic it would be great!

    Best Regards,
    Novaris

    •   AltAdvertisement

        
       

  2. #2
    Advanced Member level 5
    Points: 40,686, Level: 49

    Join Date
    Mar 2008
    Location
    USA
    Posts
    6,559
    Helped
    1928 / 1928
    Points
    40,686
    Level
    49

    MonteCarlo Design Input Parameter Variation

    Cadence wants to sell you a Monte Carlo license, so don't rule out
    malice....

    In cdsSpice we used to use gauss() and track() functions.
    Never peeled the lid off Spectre MC setup when I had the chance.



    •   AltAdvertisement

        
       

  3. #3
    Newbie level 2
    Points: 766, Level: 6

    Join Date
    Jan 2010
    Location
    Austria
    Posts
    2
    Helped
    0 / 0
    Points
    766
    Level
    6

    MonteCarlo Design Input Parameter Variation

    Well, I use the Monte-Carlo Interface but Spectre is only varying process and mismatch data within the models but no user-defined variables like a supply voltage or biasing current. You can select one parameter in the list but this ends up in a sweep, no gaussian variation.

    It would be nice to have the possibility to add a block/source, which is changing its value during a monta-carlo run every iteration, but not during the simulation (DC, AC, TRAN) itself.

    Added after 2 hours 54 minutes:

    Finally i found a solution by writing my own scs file. thanks anyway for your time :)



--[[ ]]--