I ran the same test with my external sensor attached to the cooler side (away from the power in pin "40 amps"). As expected that side did not heat very fast. The temp stabilized at 59 deg. C in the same 15 minuets. I was expecting this side to keep heating for a longer time.

So my internal temp sensor is probably accurate with the designed value put back into my software.

I ordered some thermally conductive potting epoxy and will post that result when I get one built with that. But I'm wondering if that will gain anything as far as better results. I'm guessing it will reduce the high side temp and increase the low side temp. So not sure what that buys me.