Maker Pro
Maker Pro

Pspice MOSFET models

S

Subhajit Sen

Jan 1, 1970
0
I am experimenting with Pspice (v.10) for analog VLSI
circuit design using MOSFET models available from
the web(from modeling companies/foundries etc.)
and some old level-3 HSPICE level-3
models. I am aware than Pspice does
not support 'binning' that is characteristic
of VLSI MOSFET models but prepared to do the necessary
work-arounds for this (write separate models etc.).

So far I have been able to get Pspice to simulate
by modifying the models file e.g. comment out
parameters e.g. ACM (which specifices diode cap.)
etc. However, I am not getting expected results
e.g. the gain values I am computing differ from
what Pspice gives.

My question is: has anybody
had successful experience in using Pspice as
with models published in the web (from foundries,
modelling companies etc.) for
analog VLSI design and obtained consistent results.
Also does anybody know of of foundry supplied models
that work well with Pspice?

Regards,
Subhajit
 
J

Jim Thompson

Jan 1, 1970
0
I am experimenting with Pspice (v.10) for analog VLSI
circuit design using MOSFET models available from
the web(from modeling companies/foundries etc.)
and some old level-3 HSPICE level-3
models. I am aware than Pspice does
not support 'binning' that is characteristic
of VLSI MOSFET models but prepared to do the necessary
work-arounds for this (write separate models etc.).

So far I have been able to get Pspice to simulate
by modifying the models file e.g. comment out
parameters e.g. ACM (which specifices diode cap.)
etc. However, I am not getting expected results
e.g. the gain values I am computing differ from
what Pspice gives.

My question is: has anybody
had successful experience in using Pspice as
with models published in the web (from foundries,
modelling companies etc.) for
analog VLSI design and obtained consistent results.
Also does anybody know of of foundry supplied models
that work well with Pspice?

Regards,
Subhajit

(1) Foundry-provided models work just fine with PSpice and correlate
virtually exactly with simulations on HSpice, Cadence, etc.,
PROVIDED...

(2) Use a high level model. Level=3 is amateur land.

(3) If you simulate a different gain than what you calculate, your
calculation is inaccurate.

...Jim Thompson
 
I

iQbal

Jan 1, 1970
0
I had recently used Orcad PSpice modeller to perform level 49
simulations that matched exaclty as HSPICE.

What you need to do is convert the foundry data using PSPICE model
convertor

iQbal
 
S

Subhajit Sen

Jan 1, 1970
0
I had recently used Orcad PSpice modeller to perform level 49
simulations that matched exaclty as HSPICE.

What you need to do is convert the foundry data using PSPICE model
convertor

Is the model converter part of Orcad or Pspice GUI's or a
separate utility? Is it part of model editor? How does one access it?
Also is there support for the various BSIM models going upto BSIM4.x?

(Also to Jim's comment that Level-3 is "amateur land": that
may well be the case but one of my tests for a good simulator when I
try out
a new one is to see if it "works" with levels 1-3 IC MOS models.
Isn't it true that these are some of the first IC MOS models developed
for Berkeley Spice versions around which most modern Spice simulators
(including Pspice) are built?
Also if you are a just begining learning analog IC design and use some
standard
text books such as Gray & Meyer, Laker & Sansen, Johns & Martin the
examples given are mostly levels 1-3.)

Subhajit
 
J

Jim Thompson

Jan 1, 1970
0
Is the model converter part of Orcad or Pspice GUI's or a
separate utility? Is it part of model editor? How does one access it?
Also is there support for the various BSIM models going upto BSIM4.x?

It was posted for awhile on the Cadence site, then withdrawn due to
bugs... likewise was on my site for awhile. Brian Hirasuna, of
Cadence, promised it will be refined later this year.
(Also to Jim's comment that Level-3 is "amateur land": that
may well be the case but one of my tests for a good simulator when I
try out
a new one is to see if it "works" with levels 1-3 IC MOS models.
Isn't it true that these are some of the first IC MOS models developed
for Berkeley Spice versions around which most modern Spice simulators
(including Pspice) are built?
Also if you are a just begining learning analog IC design and use some
standard
text books such as Gray & Meyer, Laker & Sansen, Johns & Martin the
examples given are mostly levels 1-3.)

Subhajit

Textbooks are rarely up-to-date. Levels 1-3 are used in textbooks
since the equations are simple-minded... higher level models are for
computers ;-)

I shared the podium with Willy Sansen in 1987 in Melbourne, Australia,
at the Royal Melbourne Institute of Technology. He taught CMOS analog
design, I taught bipolar.

Or download "Binning.exe" from my website (written by my son Aaron).

See "HSpice2PSpiceNewEdit.pdf" on the S.E.D/Schematics Page of my
website.

You **should not** be seeing differences between simulations done with
PSpice and HSpice.

I have **1522** folders of foundry-provided Spice models that work
quite nicely with PSpice ;-)

(Before anyone asks... NO, I CAN NOT SHARE THEM... I am bound by
NDAs.)

...Jim Thompson
 
Top