Dick Adams wrote:

> The oven worked unreliably, and finally would not ignite.
> Examination revealed everything apparently in order, with a
> good pilot flame centered in the element which it heats.  But
> the "thermostat", which appears to be a sealed wire-wound
> potentiometer, did not feel right, and did not test consistently
> with an ohm meter.  Function was restored by drifting in TV
> tuner cleaner and exercising the control.  Which is to say that
> the oven now works like new.  When the oven is off, there is
> infinite resistance across the "thermostat" terminals.  On, there is
> a low variable resistance.


> My problem is this -- I do not have the faintest idea how this
> system works.  There are two skinny cylinders, which are
> connected to the main valve by narrow copper tubing, and
> the pilot also, which heats one of the cylinders.  The other cylinder senses oven temperature at the exhaust vent.  And
> there are the two wires to the above-mentioned potentiometer.
> Although the range has a light and a clock, there is no electrical
> power to the oven gas valve.


> I can imagine a resistive bridge circuit between the potentiometer
> and the heat sensor, assuming it is a thermocouple, maybe driven
> by power generated by the device heated by the pilot flame, assuming
> it is a thermocouple.  However, in my experience, such-looking things
> have sometimes turned out to be hydraulic devices containing mercury.
> In any case, I cannot begin to guess how the force to open and close
> the gas valve is generated.


What you have is most likely as you guessed, a thermocouple-driven valve, known in furnace controls circles as a "millivolt control" system. The cylinder heated by the pilot contains a number of t/c elements in series, generating enough voltage to operate a sensitive solemoid in the valve. The valve may also be pilot operated -not referring to the pilot flame, but to a very small valve that vents pressure into a diaphragm to open the main valve. I can only roughly visualize the oven sensor/temp control side of the equation, as it seems like it would be hard to get well-calibrated switch point, with controlled hysteresis from only a pot and a second thermocouple. Possibly that circuit is in series with the primary t/c and the pot, but with reversed polarity, so the sensor t/c bucks the primary, shutting off the gas valve. The pot would limit the current, making the setpoint. You didn't say, but I suspect the pot is at a minumum resistance at maximum temp setting, "exposing" the full current of the sensor t/c to the control loop. For that matter, are all three terminals of the pot present, or only two?

Have to think about it some more...

Dave