From: Mike Monett
X-Mailer: Mozilla 2.02 (Win16; I)
Subject: Re: SPICE model of XTAL
Date: Tue, 10 Dec 2002 06:05:40 -0800
NNTP-Posting-Date: Tue, 10 Dec 2002 06:05:39 EST
Organization: Bell Sympatico
John Woodgate wrote:
> I read in sci.electronics.design that Kevin Aylward
> wrote (in li.net>) about 'SPICE model of XTAL', on Tue, 10 Dec 2002:
> >In real life, simulation truly adds to ones understanding,
> >despite the fact that people don't like the idea.
> I don't think that most people don't like simulation; they don't like it
> being used as a substitute for understanding how a circuit is supposed,
> in principle, to work. For example, it would be wrong to use simulation
> to design an audio treble and bass tone control circuit, without knowing
> that you want the 1 kHz gain to stay as constant as possible at all
> control settings and that you need something between 12 dB and 20 dB
> boos and cut at 20 Hz and 20 kHz. You also need to decide how much
> distortion is tolerable, and, separately, how much noise.
> Regards, John Woodgate
You could step the component values and repeat the analysis. This would give a
very clear picture of how the response changes at 1KHz. You can also insert
noise at various points to see the sensitivity and how much appears at the
Paper calculations would be tedious and may not give as much insight into how
the circuit really works.
I think the people who are against simulation worry about inaccuracy, poor
models, or the limitations of SPICE. Any of these can give false and
misleading results. Similar problems can occur in breadboarding, so it's
necessary to verify the results. But SPICE analysis and breadboarding can work
together and produce far better results than either one used alone.