convergence failure?

[Chris Cole-7]

Hey all, I've spent the last day or so messing around with gschem and
gnucap trying to simulate a half wave rectifier circuit. I'm attaching
what I have so far...the problem I'm seeing is with running a transient
analysis. If my sine wave voltage goes over 10 Volts, I get the
convergence error...under 10 Volts and I can run without a problem? I'm
sure it's something stupid. Hopefully somebody has seen something of
this nature before. All you should have to do to replicate the problem
is run 'make' inside the extracted folder. I have the sine voltage set
to 11 V, so the error should appear.

Thanks!

_______________________________________________
Help-gnucap mailing list
Help-gnucap@...
https://lists.gnu.org/mailman/listinfo/help-gnucap

[Felix Salfelder]

On Wed, May 16, 2012 at 08:47:02AM -0400, Chris Cole wrote:

> Hey all, I've spent the last day or so messing around with gschem
> and gnucap trying to simulate a half wave rectifier circuit. I'm
> attaching what I have so far...the problem I'm seeing is with
> running a transient analysis. If my sine wave voltage goes over 10
> Volts, I get the convergence error...under 10 Volts and I can run
> without a problem? I'm sure it's something stupid. Hopefully
> somebody has seen something of this nature before. All you should
> have to do to replicate the problem is run 'make' inside the
> extracted folder. I have the sine voltage set to 11 V, so the error
> should appear.

Hi Chris.

i never used anything beyond \pm 5V. my guess is voltage limiting in
COMPONENT::volts_limited in e_compon.cc.

this function cuts down voltages using _sim->_vmin and _sim->_vmax,
which i guess come from OPT::_vmin/max, see u_opt2.cc. try something like
"options vm{in,ax}=20" respectively. read the manual...?

regards
felix

_______________________________________________
Help-gnucap mailing list
Help-gnucap@...
https://lists.gnu.org/mailman/listinfo/help-gnucap

[al davis-13]

On Wednesday 16 May 2012, Chris Cole wrote:

> Hey all, I've spent the last day or so messing around with
> gschem and  gnucap trying to simulate a half wave rectifier
> circuit. I'm attaching what I have so far...the problem I'm
> seeing is with running a transient analysis. If my sine wave
> voltage goes over 10 Volts, I get the convergence
> error...under 10 Volts and I can run without a problem? I'm
> sure it's something stupid. Hopefully somebody has seen
> something of this nature before. All you should have to do
> to replicate the problem is run 'make' inside the extracted
> folder. I have the sine voltage set to 11 V, so the error
> should appear.

The list strips out attachments and holds HTML postings for
moderation.  Can you send us the files as plain text?

I am aware of a convergence problem with the diode.  The
"modelgen" models are planned for replacement.  

A couple of suggestions:

1. Simple .. try "option dampstrategy=2".  (Take a partial
Newton step in certain situations.)  

2. (more like the long term solution) ..  Try using the spice
diode model plugin instead.

_______________________________________________
Help-gnucap mailing list
Help-gnucap@...
https://lists.gnu.org/mailman/listinfo/help-gnucap