Re: 5.25 card into IIGS

["Michael J. Mahon" <mjmahon@aol.com>]

Roger Johnstone wrote:
> In <1150795845.520962.52220@r2g2000cwb.googlegroups.com> mdj wrote:
>
> > Michael J. Mahon wrote:
> >
> >
> > > FWIW, I've never seen a decoupling capacitor on a card fail, unless
> > > it was because its leads shorted or broke.  Undoubtedly it can happen,
> > > but it is a very unusual event.
> >
> > Seems like it's almost impossible, given that they're connected across
> > the rails anyway. You'd need a fairly nasty current spike to pop one 
> > of those.
>
>
> It does happen though. I know, because my IIGS bears the scar to prove 
> it. This is a repost of the story I told a few years ago:
>
> A tiny surface mount chip capacitor on the _underside_ of the 
> motherboard shorted and burnt up. 
> It started with a tiny stream of smoke coming from the top of my IIgs, I 
> immediately turned it off and opened it up. Close examination of the 
> board 
> and the cards revealed no obvious source of the smoke, so I left the 
> cards 
> out and turned it back on, with the top off of course. Immediately I 
> could 
> see, through one of the holes the motherboard has in front of each slot, 
> a 
> disturbingly bright orange glow reflecting off the shiny metal base pan! 
> The 
> capacitor was a bypass from the +5V rail to ground, it had shorted and 
> the 
> power supply was doing its best to clear the short by burning a hole 
> through 
> my motherboard! 
>
> Fortunately no major damage was done, once I dug out the remains of the 
> capacitor and some of the charred board everything worked fine. It's 
> missing 
> a bypass capacitor on slot 4 but since nothing's in that slot it doesn't 
> cause a problem.

This reminds me of something that happened in 1981.

A friend (Allen Baum, for you Apple II history buffs ;-) at work was
in the final stages of constructing a mainframe prototype--mostly
wire-wrapped.  It was a very fast (for the time) ECL-based machine,
and had large, high-current power supplies.  It also had *lots* of
0.1uF decoupling capacitors *everywhere* on the cards and backplane,
to minimize switching glitches on the power rails.

It turned out that one batch (of several thousand) capacitors was
marginally rated, with the result that when the prototype had been
powered up for a day or two, caps began to short out.

Now, when a cap shorts a 400 amp supply, it becomes a small cloud
of vapor in less than a millisecond!  The resulting *bang* was not
quite as loud as a gun, but *much* louder than popcorn.  ;-)

The failure rate was such that they would go off like a Geiger
counter on steroids, at a rate of about one "count" every half
hour or so--with, of course, statistical fluctuations.  ;-)

After a week or so, the entire lab was in on the joke--and
Allen was really annoyed.  There was no practical way to
replace thousands of capacitors, except as they died, and
as it became clear that it wasn't "just a few weak ones",
we all realized that the prototype would continue to go
*bang* every so often for its practical life!

It's hard to give management a demo of a fine new prototype
when it's punctuating the presentation with loud noises
and flashes of light!

Eventually the prototype was retired, but it never stopped
popping whenever it was powered up.  It was the perfect
4th of July machine.  ;-)


-michael

Parallel computing for 8-bit Apple II's!
Home page:  http://members.aol.com/MJMahon/

"The wastebasket is our most important design
tool--and it is seriously underused."