[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-gnucap] Bizarre behaviour -- possibly a bug.
From: |
Orestes Mas |
Subject: |
Re: [Help-gnucap] Bizarre behaviour -- possibly a bug. |
Date: |
Sun, 26 May 2013 13:18:43 +0200 |
User-agent: |
KMail/1.13.7 (Linux/3.2.0-4-amd64; KDE/4.8.4; x86_64; ; ) |
El Diumenge, 26 de maig de 2013, en/na al davis va escriure:
> On Saturday 25 May 2013, Orestes Mas wrote:
> > All elements are in series, so clearly I(Vg)=-I(L1)
> > But instead, the first output value of I(Vg) is -900kA, and
> > should be zero. Current through the inductor is OK.
> >
> > This does not happen if "UIC" is removed from TRAN command.
> > Why? Are voltage independent sources using initial
> > conditions internally?
>
> "It's Not A Bug, It's A Feature!"
>
> Spice avoids the problem by not printing anything for time=0
> when you do UIC. Gnucap plainly shows that UIC is screwed up,
> as it implements bug-for-bug compatibility.
>
> UIC does not give reasonable results, by design. It just makes
> all voltages zero, jam sets capacitor and inductors initial
> conditions without regard to consequences. Think of it as
> jamming a bunch of ideal voltage and current sources into the
> circuit.
>
> The purpose of UIC is so you can do a transient analysis on a
> circuit that does not have a stable DC operating point.
>
> In this case, with option short=10u, it is shorting the 9 volt
> voltage source, and correctly reporting a current of 900
> kiloamps.
I must confess I'm a bit lost here, due to the fact I'm not aware of some
simulator internals.
I know that to do transient simulations one must use transient sources,
perhaps a PWL 0,0 1n,8 in this case. But the students tend to use DC source +
UIC to simulate step responses as they seems to consider this way
"conceptually easier". This problem arose when I was reviewing a student's
exercise where she printed out the current through the DC source along with
the inductor's current.
I simply used UIC because I wanted to start with zero current in the inductor.
If I don't do that the simulator calculates the DC operating point on its own
and
But wait... Are you meaning that if one wants to do a transient simulation
should use only transient sources? perhaps a PWL 0,0 1n,8 in this case?
>
> > Ah, I nearly forgot it: I'm using Gnucap 2009.12.07 RCS
> > 26.136. Not tested with stable release.
>
> You might want to upgrade .. the latest snapshot is 2013.04.23
> RCS 26.138.
Yes, I know, and I'll do, but we've about 300 students here and definitely I
don't want to upgrade their simulator version (and teacher's one) until the
term is over.
--
Thanks
Orestes.