Posted for Gary Casey <glcasey@adelphia.net>:

 I'll take a shot at answering this:  The alternator is a current- limited
device and the battery is essentially a voltage device that  acts like a large
capacitor.  It has a very low internal resistance,  so if the voltage
regulator shorted, sending full output voltage into  the field the alternator
current (assuming the rpm is high enough)  will go to its maximum, say 60
amps.  If the battery was already  fully charged the voltage will rise a
little, maybe half a volt.  No  problem, so far - in short, the alternator,
when connected to a good  battery, is incapable of producing a "very brief
high-voltage  transient."  As I understand it the problem is that overcharging
the  battery will generate gas on the surface of the plates.  The gas is
 non-conductive so the internal resistance will rise, allowing the  voltage to
go up.  This will end up in a runaway condition that is a  concern.  The real
question is, how long does this take to happen?   Some will say in
milliseconds, but I don't believe that.  It would be  interesting to run the
test and find out for sure, but I haven't  gotten around tuit, as they say.
 The real problem with over-voltage  transients is not the shorted alternator
field that everyone worries  about, but an intermittent connection at the
battery.  The alternator  is producing current and the battery, the sink for
that current,  suddenly goes open.  The voltage will immediately rise to
whatever  the alternator will produce - and that could easily be over 100
 volts.  In the marine business the scenario that happens all too  often is on
the first nice day of summer the owner would find his  boat battery dead
(disconnect it for the winter?  Store it inside?   Charge it before wanting to
use it?  That would take effort).   Anyway, he pulls it out of the boat
(jumper cables won't reach),  charges it for a few minutes, sticks it back in
the boat and pushes  the cables on (no wrench, family waiting) and starts the
engine.  The  alternator is now pumping out maximum current to charge the
still-low  battery.  With normal vibration and shock loading the cable
 connection becomes intermittent and the whole electrical system is  exposed
to a continuous series of severe voltage spikes.  That's as  bad as it gets
and we tested all electrical system components under  those conditions.  Is an
aircraft system that bad?  I don't think so,  but make sure the battery
connections are secure - that's the best  protection for your electrical
components.
 Gary Casey