[Help-gnucap] Re: Time step control

[a r]

I've just noticed that for larger circuits itl4=4 is prohibitive
(gnucap tries to decrease timestep all the time).

What I am using in practical simulation is ether:
.option method=trap dtmin=1e-16 abstol=1e-13 vntol=1e-7 chgtol=1e-15
reltol=1e-4 itl3=3 itl4=6
or:
.option method=euler dtmin=1e-16 abstol=1e-13 vntol=1e-7 chgtol=1e-15
reltol=1e-4 itl3=3 itl4=10

-r.

On Dec 1, 2007 5:17 PM, a r <address@hidden> wrote:
> I think default settings for the time-step control are too aggressive.
>
> I am getting best results with the following settings:
>
> .option method=euler dtmin=1e-16 reltol=1e-4 itl3=3 itl4=4
>
> .tran 1n 20n trace alltime > inverter_test.dat
>
> If I do not specify low thresholds for itl3, itl4 gnucap has a
> tendency to use maximum allowed time steps (here 1n) and "jump over"
> all interesting transistions.
>
> The test circuit is attached. Note that without setting itl3, itl4 the
> results are not readable (although they might be considered
> "correct").
>
> If the time step is set to 0.001 (and itl3, itl4 are not set) the
> output is good but the time step is effectively constant which is not
> good for simulating sampled circuits.
>
> The question is: are the defaults sensible?