Re: New convergence failure (was Re: [Help-gnucap] Several problems: convergence failure...)

[Rubén Gómez Antolí]

Hello:

Continue with test.

Rubén Gómez Antolí escribió:
> Hello Al:
>
> al davis escribió:
> > On Monday 01 March 2010, Rubén Gómez Antolí wrote:
> > > Sorry about repeat topic, but I'd fall in a nightmare with
> > >  this circuit  and his evolution.
> > >
> > > If is possible that there are a fail in design because I
> > >  haven't correct  output of buck chopper (1.5 V instead 1V,
> > >  Ngspice runs circuit :-/ ), but I can't found the "bug".
> > >
> > > I note that in Ngspice output there are too much "noise",
> > >  that I think  is no normal.
> >
> > I can't run it .. there are pieces missing.
>
> Perhaps I should attach it instead do inline. :-/
>
> > Spice time step control and convergence checking is often not strict 
> > enough.  Sometimes it accepts results that are wrong.
> >
> > Here are a few ideas ...
> >
> > Breakdown doesn't work with the native diode model.  Try the one in 
> > models-spice3f5 instead and see if that helps.
>
> Well, seems that I fall in troubles too with models:

(...)
> I'm viewing that I lost last snapshot tarball; I'll try to compile 
> gnucap and extra models tomorrow.

Ok, I compiled last snapshot (12/07) include models tarball and now 
loads well:

gnucap>  load /usr/share/gnucap/lib/models/spice3f5/dio.so
D: already installed, replacing
stashing as D:0
d: already installed, replacing
stashing as d:0

Using Spice3f5 model not helps:

gnucap> tran 0 15ms 1us > z
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 15
open circuit: internal node 20
open circuit: internal node 15
open circuit: internal node 20
non-recoverable convergence failure, reducing (itl4)
newtime=8.532783e-06  rejectedtime=8.532784e-06  oldtime=8.532783e-06 
using=8.532784e-06
tried everything, still doesn't work, giving up

> > (...)
>
> > Try adding a resistor across the diodes.

I'm add a resistor across diode in main circuit. Try with differents 
values, no helps.

> > Try adding a switch, with the control port across a pair of nodes that 
> > you think are causing trouble.  Leave the output (switched resistor) 
> > unconnected.

I don't understand very well this. You say that add a switch in this way?:

Sdebug 0 100 (0 80) SW

and node 100 is connected to "air".

How helps this? Sorry, I'm a bit lost.

(In this way fails too).

I try other "trick", a change in triangle wave from:

VG 10 0 pulse -2.5 5 ....

to

VG 10 0 pulse 5 -2.5 ...

produce that convergence failure is retrased from 8us to 33us, change 
model diode (spice or gnucap) no affect this.

Ever fails with triangle wave change from v- to v+, crossing 0V.

This fact should indicated us that the "guilty" should be diode, LM311 
or Mosfet IRF540, elements that change his state at these time.

What you think about this?

Thanks for your attention and help.

Salud y Revolución.

Lobo.
--
Libertad es poder elegir en cualquier momento. Ahora yo elijo GNU/Linux,
para no atar mis manos con las cadenas del soft propietario.
---------
Desde El Ejido, en Almería, usuario registrado Linux #294013
http://www.counter.li.org