Re: [Help-gnucap] Convergence failure with BC547C

[Al Davis]

On Friday 30 January 2004 11:49 am, Karel Kulhavý wrote:
>  am trying to simulate a differential 2-transistor stage with BC547C.
> With rstray, it fails to converge in one cycle (on second edge, the
> edges are always difficult)
> With norstray, it goes through all 8 cycles programmed.
>
> ngspice doesn't seem to have problems with BC547C because AFAIK it
> doesn't have any rstray feature.
>
> Is it possible to enhance somehow the convergence so I can make it
> converge even with rstray? Tinker with some convergence options?
> Which options are relevant in this case?

dampstrategy, dampmax, dampmin, vmax, vmin


> Does it mean that the circuit will oscillate due to parasitic
> feedback caused by stray circuit elements? Is convergence failure and
> tendency to oscillate tightly coupled?

yes.

Actually, most convergence problems I see have nothing to do with what 
seems most likely.  Most I see come from the circuit not being what you 
think it is, or strange values of model parameters.

Spice always does rstray.  You can't turn it off.

If you send the circuit, I will look at it.  What version are you using?