I'm sure if this was viable, it would have already been done.
I agree with or concede everything you said above except this. Lots of viable/feasible/better things have not been done.
And not because they haven't been thought of but because no one with the combination of idea,time and interest wanted to do it.
A better PC cooling device test rig (better than what Tom's been using/better than my thought experiments in these posts) wouldn't be harder than the awesome article up on Tom's right now about creating a better longer lasting mouse switch!
The hotplate would have more exposed surface area than a CPU would, by far.
Key to idea is that the plate would be tiny and ideally mated/soldered to a standard CPU lid.(without significant heatsink) Ideally mounted on a mother board in a standard tower.
This link is pretty close to what I am thinking but he has a large copper base block that would not be comparable to a CPU in heat sink value.
http://overclocking.guide/thermal-paste-and-cpu-cooler-test-bench/