Owen Duffy wrote:
. . .
Anyway, to your original question of how much increase in power rating
with oil immersion. In the long term, it is determined by the ability of
the load to dissipate heat from the outside for a tolerable rise in
temperature of the oil and the resistor element. In the short term, it
comes down to the resistor element, and different constructions will
behave differently (eg metal film vs composition).

Yes, I'd worry a little about grossly overdriving carbon composition
resistors because of the thermal resistance between the heat-dissipating
composition material and the oil. There's going to be a temperature
gradient across the phenolic case that might be considerable -- in other
words, the resistance part could be much hotter than the oil. Resistors
which have the actual dissipating portion very close to the outside,
like film or (noninductive) wire wound resistors would be much better in
this regard. If you overstress carbon comp resistors you'll typically
get a permanent change in resistance. If the overstress is extreme,
they'll crack, explode, or catch fire. Make sure your oil and container
won't present a hazard if this happens.

Roy Lewallen, W7EL