Agreeing with Rocket Soul,
In my experience it's
NOT just the thickness of the aluminum and surface area - it's
total mass also.
You have to be able to maintain the temperature of any sinking device or you can experience thermal runaway. Feeling the frames on these QB288's or the heatsinks I've added to the QB132's is telling.
Boatguy,
I've been dabbling with electronics for the better part of forty years including a tenure at a FCC certified communications electronics repair shop.
I've found putting the sinks on a plate will indeed cool better and even have had discussions about electronic theory and what actually occurred in real time tests.
While it's true a thinner plate will draw and cool faster - once the mass has reached it's thermal limitations, the "full line" if you will it can no longer keep the temperature in line without runaway.
(Also the reason you see poorly designed "active cooling" on many asian light fixtures.)
If you were to take a Slate 2 single and make an copy of the extrusion with the same X&Y dimensions and increase the (Z axis) plate thickness to say, 10mm and even maintain the same fin height (but would also readily benefit from taller fins [mass & surface area] you would see an increase in dissipation) and more power handling.
I certainly agree on getting better transfer from a TIM though I'm not a fan of thermal pads myself - they do work decent and are tons less messy than a good paste which I used on a sunboard strip build and it's under the meanwell drivers remotely mounted to larger extrusions bolted to the joists which are barely above ambient without adding any heatsinks to the cases.
I'd be interested in where your findings have originated though.
I once helped a student with a final project of a very large stereo amplifier and he preached the formulas and theory of the teachings adamantly yet after returning from the lab he couldn't explain why the power output was so much higher than what the charts, graphs and formulas (and myself) told him.
He had problems with thermals in the bias circuitry and IDK if he ever solved them but I know he warped a voice coil in a midrange driver in one of my speakers when we were testing the power output it "wasn't" capable of.lol
He did get an excellent grade though!