In article , raden
writes
In message , Andy Hall
writes
On Tue, 05 Oct 2004 20:44:42 GMT, "Malc"
wrote:




Because it was a crap design basically. They used a capacitive dropper from
each side of the output relay (i.e. from the NC and the NO contact but not
the common) to provide power for the stat (both feeding into a common
rectifier, zener and smoothing cap). In some modes of wiring a CH circuit it
is possible for the power to arrive at the stat from the NC side of the
relay, as I'm sure you know. In this case the power had to charge up the
smoothing capacitor which kept power on the relay as it changed over. If the
relay moved too quickly then there wasn't enough charge in the capacitor to
power the relay across to the NO contact and it would oscillate and burn out
the zener.
If the circuit had been provided with bigger droppers and also a higher
wattage zener then it would have been fine but in the interests of cheapness
things were pared to the limit. As it was my job to analyse field returns I
got to know the various failure modes by sight without having to test the
stat. At one point we were running at 10% failure rate.

Didn't anybody ever apply a brain to this?

Surely the cost of the returns in dealing with them would start to
compare with the cost of a better design.......

Andy you're on dangerous ground here, applying logic and such


IIRC Ford in the USofA had a car, the Pinto, that almost blew up in a
rear end collision, apparently some bolts would tear the fuel tank
apart.

It seemed at the time it was cheaper for Ford not to modify production
but to argue the accident claims.

This was of course before litigation became the business it is today..
--
Tony Sayer