I read somewhere that if you provide at least 18" of free conductor in a j-box (prior to splicing it) that higher temperatures (caused by conductor crowding in a conduit) won't be passed on to any spliced wires downstream.

Is this true? Is this some "rule of thumb"?

Is there any language in the code that relates to such "cooling off" phenomena?

What I'm dealing with is a situation where an SO cord splices on to some Thhn in an industrial environnment. There are less than 20 conductors (but more than 9) in the conduit, and the wiring will probably be in an ambient temperature of 50 degree C (122 degrees F) during the summer time. The SO cord will drop down from the j-box as a pendant for some stationary equipment. I don't want to cook the SO cord when it is spliced to the Thhn. I'm assuming the conductors in the cord are rated for 60 degree C. If I start with a 90 degree allowable ampacity, and adjust for 50% due to bundling the wires in the conduit, and then correct (by multiplying by .82) that resulting figure for the high ambient temperature, I'll end up with a wire that works just fine at a reduced ampacity, but it will still possibly be 90 degree C when it enters the j-box. Once it is in free air(at the j-box and at the source panel), it starts to cool. I need to know how much wire I need to provide for that "cooling off" effect.

Of course, I could just work off the 60 degree C allowable ampacity figure, and end up having to extra-oversize my wire (I'm already using #10 in place of #12), but I'm trying to avoid that.