On Sun, 11 Sep 2016 10:19:06 -0700, Jeff Liebermann
wrote:

Duh. I thought C106D was a capacitor or typo error. It's a real part
and most definitely an SCR:
http://www.onsemi.com/pub_link/Collateral/C106-D.PDF

Certainly is. Been around a while too.

I could calculate the amount of heating from the DC present if you
would kindly risk your life and obtain a scope waveform of what is
present across the coil while energized.

Unlikely. Solenoids are cheap, I only have one life to give and it
isn't going to be for this cause, however noble from the scientific
research pov.

I don't see any sign of heat damage around the solenoids in your
photos.

The original pair were showing darkened areas on the brown stuff, but
that was two decades worth of higher utilisation.

(I don't like using the gas line as an electrical ground, but
that has nothing to do with the problem).

There is quite a bit of earth bonding in there, excessive in my view,
but I guess paranoia never killed anyone.

The white wires connected
to the spade lug connectors would have turned brown by now if there
had been any overheating of the terminals. Therefore, my guess(tm) is
that the solenoid terminals are NOT getting hot, which would certainly
happen if they were melting the internal solder connections.

Typical oven operating temp is 200C. Some occasions are 220/230 for
30 minutes. Coils have limited prospects for ventilating airflow, and
self-heating to boot, so they would reach ???

Measuring the terminal temperatures during operation would offer a
clue. If they are anywhere near the melting point of RoHS solder of
about 220C, it might be the culprit. However, since RoHS solder melts
about 40C higher in temperature than leaded solder, I just don't see
it. If you need some entertainment, try heating the terminals of one
of the failed solenoid coils with a really hot soldering iron and see
if you can reflow the solder inside the solenoid.

In other words, I don't think it's a solder problem.
(snip)

A high voltage spike could produce an arc between turns,
followed by a shorted turn. Shorted turns really increase the current
drain of an AC solenoid, which could be sufficient to fuse the #34
wire (5A).

When I have one of these failures, need to remember the
characteristics. They go O/C when hot, with normal program resuming
as they cool (at say 125C oven as a WAG). What I then do, after the
mandatory outburst of expletives, is grab my diagnostic lead which I
recall referring to earlier. This interposes a small naked 6V
incandescent lamp in series with each coil. Then I fire the oven up
again and wait for one or the other lamp to extinguish as the coil
reaches the failure temp again. I do NOT recall the current spec of
the lamps but suspect it is 50mA. They glow with decidedly equal
brightness until one is interrupted, so any inter-layer etc shorts
must also be transient.

If my guess(tm) is correct, then the problem will persist with almost
any solenoid that uses thin #34 gauge wire. So, something must be
done to prevent arcing between windings.

If it *is* arcing, it reoccurs with regularity on these post-failure
test runs - or not at all - as it would possibly take out the lamp or
at least show substantially increased illumination prior to the return
of O/C mode.

Since the pulses are at a much higher frequency than 50Hz,
(snip)

I'm not supporting that hypothesis at this stage.

Note: This was written before taking my morning vasodilator pill and
is therefore presumed to be relatively coherent and sane.