Well, to finish off this entertainment, I should mention my effort
this evening to replicate the effect I saw at the weekend.
I used resistors in what used to be called the 1/8 watt size,
that probably started life with the intention of
being used in precision applications (but think of them
as floor sweepings..)
For example I used resistors whose body color was blue, and they
carried this marking:
CGW
RN55D
1153
FJ
They measured 115 kohm, and I twisted two together in series (to allow a
little more space between meter probes.)
Each passage through the flame lifted the resistance so that after a number
of passes they registered about 500kohms.
Specimens all went open circuit sooner or later(unsurprisingly.)
It was difficult to find a pair which replicated the effect I had seen.
I didn't see it at all with just one resistor of either 115kohm or 1 Mohm.
My best guess at the cause of what I saw: cooking the resistors gradually
lifted their resistance, while providing a distinct temporary increase
while hot (as you would expect).
This increasing resistance level at some point allowed the indication
to float up to several millivolts and when cooled, the resistance decreased
enough to zero the indication again. It was possible there was also a
contribution from a varying thermocouple signal.
Ho hum. Disappointing. Back to the drawing board.
:-)
brian whatcott <inet@intellisys.net> Altus OK
Eureka!