Reply-To: "fred bartoli"
From: "fred bartoli"
Subject: Re: SPICE simulation goes awry?
Date: Thu, 12 Dec 2002 14:15:19 +0100
X-Newsreader: Microsoft Outlook Express 5.00.2314.1300
Organization: Guest of ProXad - France
NNTP-Posting-Date: 12 Dec 2002 14:16:04 MET
Paul Burridge a écrit dans le message :
> On Tue, 10 Dec 2002 13:22:20 -0800, John Larkin
> opined thusly:
> >A dumb enough transistor model will do this. Clearly the collector
> >voltage must be swinging far negative, which isn't realistic.
> Yes, it does. At lease initially. But if you let the simulation run
> and run, it *eventually* starts going positive again. The overall
> effect is exactly like an AM/RF modulation envelope that's had the top
> side cut off at the 1v level.
> >And the 1v rail does look low. Try and see if it oscillates at Vcc=0!
> You may be joking, but I did, and it *still* worked! With zero Vcc,
> the output signal went as far as 9v negative before starting to creep
> up again (the envelope I mean). But during all this time of course,
> it's still oscillating at 7 meg or thereabouts.
> Question is, is this just a well-known shortcoming of SPICE in
> general, or something no one else can induce, tending to indicate a
> bug in the particular version I'm using here? No oscillator ought to
> work at zero volts supply.
that is in no way a shortcoming of spice. Neither it is normal, in any
I'm pretty sure it's not a bug of the version you use as it will make it
I'd bet this is a big error in one macromodel you use. But without a
> "What is now proved was once only imagin'd"
> - William Blake, 1793