spectre bsim3v3 model parameters : ps & pd

[S. Badel]

these are perimeters of source and drain diffusions, well.

what should i put as value :
- 2*(width+length)
or
- 2*width + 1*length
considering that the gate-side of the junction doesn't
contribute to capacitance

?

thanks,
stephane

 

[Erik Wanta]

Stephane:
Why wouldn't it be ps=pd=2l+w as there is only one w is along the
gate? I could be wrong but I think spectre gives default as, ad, ps,
pd values if not given. I think hspice has a parameter called acm to
control these values.
---
Erik

"S. Badel" <stephane.badel@epfl.ch> wrote in message news:<3fa0e74f$1@epflnews.epfl.ch>...

these are perimeters of source and drain diffusions, well.

what should i put as value :
- 2*(width+length)
or
- 2*width + 1*length
considering that the gate-side of the junction doesn't
contribute to capacitance

?

thanks,
stephane

 

[S. Badel]

A colleague found the answer and it was indeed
the main source that caused different results between
hspice and spectre.
spectre wants the complete perimeter as parameter value,
that is, 2w+2l
while hspice is expecting only three sides, that is, 2w+l.
most probably spectre does substract w during processing.
thanks for your reply,
stéphane

"Erik Wanta" <erikwanta@starband.net> wrote in message
news:84018314.0311020720.1fd4ae09@posting.google.com...

Stephane:
Why wouldn't it be ps=pd=2l+w as there is only one w is along the
gate? I could be wrong but I think spectre gives default as, ad, ps,
pd values if not given. I think hspice has a parameter called acm to
control these values.
---
Erik

"S. Badel" <stephane.badel@epfl.ch> wrote in message
news:<3fa0e74f$1@epflnews.epfl.ch>...
these are perimeters of source and drain diffusions, well.

what should i put as value :
- 2*(width+length)
or
- 2*width + 1*length
considering that the gate-side of the junction doesn't
contribute to capacitance

?

thanks,
stephane