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

Me too!

>I've replaced a few of these over the years for NOT turning the radiator fan
>ON but never had one of these sensor/switches fail in the ON position.  Has
>anyone else?  Do you think the heavy rain had anything to do with the
>failure or is that just a coincidence?  I can't quite see the rain as a
>factor but.it's a Vanagon!

This is fascinating.  If you'd be willing to send me the failed
switch to examine I'd be happy to pay the postage.  At this moment
I'm unable to suggest a failure mechanism not involving lightning,
EMP (anything you top-lefters haven't been telling us???), 1000+ gee
impact or possibly massive corrosion.  Or very tiny brass-loving
exothermic demons.

The actual switch mechanism is a pair of bimetallic snap-disks,* one
with a hole in the center.  They are stacked (suitably spaced) in the
working end of the sender and each pushes on a small ceramic rod,
hence the hole in the inner disk.  When the disk snaps the rod shoves
on a springy contact and forces it closed.

This is not a sensitive circuit where a drop of rain will false-trigger it.

* These are shallow convex discs stamped and formed out of thin
sheets made of two layers of (different) metal completely bonded
together by some sort of impact welding I believe.  The layers have
different coefficients of expansion, and the layer with the larger
rate is on the concave side.  At the calibrated temperature the
expansion of the inner layer is sufficient to overcome the tension
needed to snap the disk from concave to convex, and it does.  If it's
sitting on a table at this point it will leap off the table.  The
disk will now remain in the new form until it has cooled sufficiently
to again overcome the spring tension and snap back to the original
form.  This hysteresis gives positive operation and also avoids the
fan "short-cycling" which would be hard on the switch.

Yours,
David