At 01:52 PM 6/1/2010, David Beierl wrote:
>At 01:22 PM 6/1/2010, Andrew Martin wrote:
>>Problem solved but I'm curious about this type of sensor failure.
>
>Me too!

Dear Andrew -- sender came yesterday and I opened it up
today.  Pictures at
<http://picasaweb.google.com/dbeierl/VanagonRadiatorSwitchThatFailedShortInTheRain?feat=directlink>

This was not the snap-disk type I described, but rather a wax capsule
like a thermostat, pushing an actuator for two over-center snap
switches with different over-center points.  It clearly had gotten a
bit of water inside at some point, and I'm pretty sure it came past
the seal between the brass and plastic.  However your particular
failure was a freak -- chip of metal jammed between the high-speed
contacts.  The later ones with the disks have a better sealing
arrangement as well as a slightly lower switching temp.

I don't understand the reason for the horrible condition of the
low-temp contacts.  The metal transfer from pit to corresponding
mountain is no big deal, and there was still a fair depth of silver
left.  But the nasty crud all around the contact is a mystery to
me.  The high-speed one only carries drive current for a relay so it
ought to last pretty much forever, but it also had a light coating of
unburned crud on it.  The stuff on the bottom of the sender flaked
off with moderate fingernail pressure.

Very interesting, thanks.

Yours,
David