Hi,

I read Andrew Beckett's explanation of how BSIM 3v3 defines MOSFET capacitance, but I am still confused in practice.  I did a DC simulation in Eldo where I swept Vgs and W of the NMOS device and then got the BSIM states for the different capacitors (Cgs, Cgd, Cgg) using .PLOT DC statements.  I also did a single point simulation in Eldo to sanity check and back annotated the results into Cadence Composer.  My problem is the Cadence back annotation says that the Cgg is 2.24fF for my particular device which is 2.4X the value I get from .PLOT DC S(M0->CGG).  All the other BSIM values I checked, such as Id, Vth, gm, etc, came out spot on.  

The BSIM values give crazy results for calculating ft, but the back annotated values from Cadence are in the ballpark.  What am I doing wrong?  Are there other caps I need to add into the BSIM analysis?  I want to characterize this device in a script so I can't just go off the back annotation value.  Besides, I want to make sure I understand what is going on with the model.

Carl

OK, so this feels a bit like an echo chamber, but I sorted the issue out and I'm going to update everyone in case someone else gets a new job where Eldo is used.  Basically, the state syntax .PLOT DC S(M0->CGG) only gives a portion of CGG.  In fact, according to the Eldo .chi file CGG = CGG + CGDO + CGSO.  This is confusing enough that they use CGG to mean two different things (namely, intrinsic gate capacitance and total gate capacitance), but CGDO and CGSO are not available using the state syntax.  So, basically, what I was trying to do would not work.

Luckily, there is another way in Eldo.  You can use the .PLOT command to get a large number of device parameters directly.  For example, .PLOT DC CGG(M0) gives the total gate capacitance.  Seems so simple, now, but that 1000 page Eldo manual is formidable!  

Carl