BSIM3v3 model file with statistical model block for Monte Ca

[katyal]

Does anyone know how to setup a statistical block in the bsim3v3 model
file with varying standard deviation (as a function of W & L of the
transistor). For fixed value of sigma I can do simulations but when I
try to introduce a function to take care of the sqrt of W*L of the
transistor, it doesn't work. I want to setup the bsim model file so
that when I change my size of the transistor, I don't need to go back
and change the sigma value.

Will be looking for a solution
Vipul

 

[Andrew Beckett]

On 16 May 2007 10:07:04 -0700, katyal <katyal@gmail.com> wrote:

Does anyone know how to setup a statistical block in the bsim3v3 model
file with varying standard deviation (as a function of W & L of the
transistor). For fixed value of sigma I can do simulations but when I
try to introduce a function to take care of the sqrt of W*L of the
transistor, it doesn't work. I want to setup the bsim model file so
that when I change my size of the transistor, I don't need to go back
and change the sigma value.

Will be looking for a solution
Vipul

It will be something along these lines:

parameters wlsigma=0

statistics {
process {
vary wlsigma dist=gauss std=0.1
}
}

inline subckt nch (d g s b)
parameters w=1u l=0.18u wlvary=sqrt(w*l)*wlsigma

model nchmod bsim4 ...

nch (d g s b) nchmod w=w+wlvar l=l+wlvary

ends nch

Don't get too hung up on the exact syntax - I just typed that in off the top of
my head and didn't test it.

The idea is that wlsigma is centred around 0, and is normalized. You're then
scaling it by sqrt(w*l) within the subckt wrapper around the model - and you can
do whatever relationships you want there.

Regards,

Andrew.



--
Andrew Beckett
Senior Solution Architect
Cadence Design Systems, UK.

 

[Svenn Are Bjerkem]

On 16 Mai, 19:07, katyal <kat...@gmail.com> wrote:

Does anyone know how to setup a statistical block in the bsim3v3 model
file with varying standard deviation (as a function of W & L of the
transistor). For fixed value of sigma I can do simulations but when I
try to introduce a function to take care of the sqrt of W*L of the
transistor, it doesn't work. I want to setup the bsim model file so
that when I change my size of the transistor, I don't need to go back
and change the sigma value.

Will be looking for a solution
Vipul

aatoolsuser.pdf in your cadence doc hierarchy will tell you more on
the format.
--
Svenn

 

[okguy]

You can find it in :
$CDSHOME/tools/dfII/samples/artist/monteCarlo/models/statsLib.scs