Further to yr reply, i installed latest dev snapshot gnucap-2008-12-03
(recently posted) and tried to test a small circuit with LF 353. I
used model from National. It uses JFET model. Program response is as
below. ( my net file is in the attachment.)
gnucap> get [1]nbfm.net
* Spice netlister for gnetlist
J1 5 2 4 JX
^ ? J: no match
J2 6 7 4 JX
^ ? J: no match
.MODEL JX PJF(BETA=1.25E-5 VTO=-2.00 IS=50E-12)
^ ? model: "PJF" no match
Is there any way to improve response? I think I need a writeup using
various model package S/W used in the snapshot...
Antoher observation is:
I tried to build stable gnucap-0.35 from tarball without success,
whereas it is installing well using 'apt-get install gnucap'. Why the
difference?
However, development snapshot installed without errors..
On 12/7/08, al davis <[2]ad151@xxxxxxxxxxxxxxxx> wrote:
On Sunday 07 December 2008, vsrk sarma wrote:
> I am facing difficulties in including device/sub-circuit
> models into gnucap simulation.
> Specifically for transistors & analog ICs such as opamps.
> It appears to me that application package does not have any
> and we have to create. Am I right?
Models for specific devices are not included. You need
to "Google for it", or make your own.
> Is it possible to use available spice models? If so how to
> make them compatible with gnucap.
> how to ensure pin compatibility of imported model & my
> circuit.
Gnucap mostly accepts spice models.
The reason I say "mostly" is that there is so much variation in
Spice models. Is a model written for HSpice? PSpice? Each
one is a little different. You need to try it. The one
closest to gnucap is HSpice.
The development snapshot has more features, so if you get a
model that is not compatible with gnucap-0.35, it is likely it
will work with the development snapshot.
Pin compatibility is not an issue. Gnucap uses the same syntax
and pin lists as Spice.
If you are using J-Fets, you must use the development snapshot.
> Apologies for very basic questions (I have not used Spice
> earlier )
If you are just getting started, it is usually better to make
your own models. You probably don't need the detail provided
by the published models. You will benefit if you understand
the models you use.
For an op-amp, you can usually use a voltage controlled voltage
source.
For transistors, usually you can just add a ".model" statement,
and specify "BF" (beta) and "IS" (saturation current).
For MOSFETS you can usually specify "level=1" for the simplest
model and specify "VTO" (threshold voltage) and "KP"
(transconductance parameter). You need to know this to design
your circuit anyway.
Sometimes you need to experiment with the parameters. Here's a
reference that might help with that:
[3]http://wiki.gnucap.org/dokuwiki/doku.php?id=gnucap:experimentall
y_finding_model_parameters
Everything I said here applies equally to Spice. The models are
the same.
_______________________________________________
geda-user mailing list
[4]geda-user@xxxxxxxxxxxxxx
[5]http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
--
vsrk sarma
References
1. http://nbfm.net/
2. mailto:ad151@xxxxxxxxxxxxxxxx
3. http://wiki.gnucap.org/dokuwiki/doku.php?id=gnucap:experimentally_finding_model_parameters
4. mailto:geda-user@xxxxxxxxxxxxxx
5. http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
Attachment:
nbfm.net
Description: Binary data
_______________________________________________ geda-user mailing list geda-user@xxxxxxxxxxxxxx http://www.seul.org/cgi-bin/mailman/listinfo/geda-user