+ Post New Thread
Results 1 to 2 of 2
  1. #1
    Newbie level 4
    Points: 1,868, Level: 10

    Join Date
    May 2006
    Posts
    6
    Helped
    0 / 0
    Points
    1,868
    Level
    10

    Standard deviation of threshold voltage

    Hi,
    I want to do the variability analysis on my design and hence I need sigma of vth and u0 (standard deviation of threshold voltage and mobility) for monte carlo simulations. I have the 90nm PDK as well, but I dont know where these values are written. Can you please tell me how to get these values?

    [P.S: I can access PVT corners (fast, typical, slow) but I do not want to run worst case analysis. Rather I want to perform variability analysis].

    •   AltAdvertisment

        
       

  2. #2
    Advanced Member level 5
    Points: 37,575, Level: 47

    Join Date
    Mar 2008
    Location
    USA
    Posts
    6,071
    Helped
    1758 / 1758
    Points
    37,575
    Level
    47

    Re: Standard deviation of threshold voltage

    If your foundry says anything about what corners represent,
    you can probably fake some statistics. But foundries' WAT
    limits vs capability vs model over-coverage is always an
    internal negotiation and outcomes vary. Your WAT range
    could be nominal +/- 3 sigma or could be +/- 5 sigma with
    the models sandbagged to +/- 6 sigma.

    Now going by simple WAT limits you will gain no insight to
    correlation; this is important (Tox tracks 1:1 pretty much
    between N and P; VT to a lesser extent, but somewhat).
    If you were serious and funded, you might ask the foundry
    for raw WAT dumps and crunch the correlations yourself.
    I'd bet, however, that this material exists pre-chewed and
    it's more a matter of getting to the right person and
    convincing them you deserve the data.



--[[ ]]--