>
>
> I would have to agree; the thermocouple in the burner box is basically
> an "on off" type device that serves only to keep the shut off valve
> open. But......
>

The thermocouple output is proportional to temp difference between the two
junctions.  But the output is either enough to hold the gas valve open, or
not.  Certainly an eroded one is on the way to failure.

> Maybe a lower voltage from the thermocouple would only pull the valve
> part way open?
>

No, it can't pull the valve open at all, only hold it open.  The
thermocouple is basically two pieces of wire, so it has ridiculously low
impedance.  That means that its tens of millivolts output can push IIRC
better than one ampere through the magnet in the gas valve, which is only a
few turns of heavy wire with pole pieces that are machined flat.  The
actuator piece is also machined very flat, so that it makes extremely good
physical contact with the pole pieces.  Because of the close contact the
magnetic circuit is very good and the valve can be held open with
considerable force.  But that force vanishes as soon as the two pieces
break contact, so the valve has to be opened by you "pushing the pickle".

>
> I *think* that the "silver" narrow tube going from the thermostat to
> the cooling unit in fridge is, in part, responsible for sensing if
> more cooling is needed hence a larger flame is required. As I'm sure
> you know, there is a small white silicone or rubber bit that keeps
> that tube pushed up against the aluminum cooling unit.
>

The thermostat is a snap-action device operated by that capillary tube.
When it calls for cooling, it opens a clear channel for gas to pass through
and be limited by the orifice.  When it does not, the gas for the pilot
flame is metered through an extremely small orifice inside the gas valve.
The flame has only two levels; cooling and pilot.

Yrs,
d