[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: Gnucap crash



The first one was sent by mistake, before I finished editing it.
Try again..

On Monday 09 April 2007 08:52, Patrick Doyle wrote:
> * gnetlist -g spice-sdb -o spice.netlist.wpd TwoStageAmp.sch
> .model 2N3904 NPN(IS=1E-14 VAF=100
> +  Bf=300 IKF=0.4 XTB=1.5 BR=4
> +  CJC=4E-12  CJE=8E-12 RB=20 RC=0.1 RE=0.1
> +  TR=250E-9  TF=350E-12 ITF=1 VTF=2 XTF=3 Vceo=40
> + Icrating=200m mfg=Philips)
> R5 Vin 1 10
> .OP
> RE1 0 Vem1 100
> Q1 Vcoll1 Vbase1 Vem1 2N3904
> R2 0 Vbase1 2K
> .end
>
> > > The crash/assertion failure occurs when I execute the
> > > "op" command at the Gnucap prompt.
> >
> > It didn't crash for me.
>
> Hmmm....
> It does crash for me... and it triggers an assertion failure
> when I build with debugging enabled -- the assertion is
> definately tied to the fact that 'precalc()' doesn't get
> called for RE1.

I see it now.  I ran it different than you did.

I know what happened ...  It has to do with the ability to 
change a circuit and continue.  What triggers the bug is 
putting ".op" (or any simulation command) in the middle of a 
Spice netlist.  The bug first appeared during work that allows 
you to change values in the middle of a run without losing 
data.  It works fine if there is an intervening command, but 
screws up when they are mixed like this.

This is supposed to work.

I will be unable to work on it for 2 weeks.  Until then ...

Don't put commands in the middle of the netlist in a file.

If you have a file that has commands in the middle, do something 
like "build" (and add nothing) or "delete" (delete nothing) and 
it will force a rebuild on the next simulation command.

This is a real bug.  I will fix it.  This hack is to get you by 
until I do.

> You've mentioned that one way I can help is by being a newbie
> doing newbie things to the simulator.

Yes.     This is very valuable.  Thank you.

> Based on what I saw in the code, I remain very confused that
> you don't see the same error, but that's life, I guess.

I saw it now. 

Gnucap optimizes, so sometimes things that seem like they should 
not matter make it take a different path.  Tracing can be 
tricky.  It does not use the same flow as Spice.  That is why 
it is so much faster on some large circuits.



_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user