Most of the time, troubleshooting millivolt circuits is pretty easy. Measure the milliovolts, and trust the results.
But sometimes it's not that easy.
One occasional symptom I find is a pilot generator that gives a nice read under no load, but drops to zero under any kind of load. You'll find a pilot that wont stay lit, measure the millivolts across the gas valve and get zero.
That leaves the possibility that there is a short in the gas valve or some other place that's causing the zero millivolt read. If you disconnect the pilot generator and measure the voltage directly, you'll get 500-600 millivolts, so it looks as though the gas valve or something else is grounded out. If you change out the gas valve though, you'll discover that the pilot still wont hold.
So--- how to test millivolt systems accurately? What is needed is a known and reliable voltage source. I've toyed with constructing a simple voltage divider using a battery and variable resistor to give me a good millivot source so I can operate the millivolt system without the pilot being lit, but I haven't gotten around to doing that.
What I do is test the pilot generator under load, by using a 12 ohm resistor. If you measure the reistance of the magnet circuit on a Robert Shaw Model 700 series gas valve it's 12 ohms. So I use a 12 ohm resistor with alligator clips on each end, disconnect one side of the pilot generator and clip on the resistor and the test leads for my multimeter. When I fire up the pilot, I get an honest read of the millivolts produced by the pilot generator under load.
In the example at the beginning of this post, I'd get 500 millivolts with no load and zero millivolts across the resistor ---then I know for sure that the pilot generator is defective.
Once I know that I'm getting real voltage, I can proceed to test the gas valve and other parts of the circuit for defects or grounds, and do so with confidence that I have a good voltage source for the test.
Anyone have a better method? Has anyone constructed their own battery source to test millivolt circuits, or know of a commercially constructed source that would save me the bother?