I've been having some issues getting the expected temperature rises when using contact resistance to add an extra thermal layer between a heatsink & a PCB chip.

Essentially, I've applied a single contact resistance (in one feature) of 0.1 K/W to the 4x faces shown below. With a heat load of ~25W per chip, I'd expect to see a delta T of 2.5 deg C across the boundary. Instead I'm seeing much, much higher than that.

When I separate the single contact resistance into 4 sets (one per chip, same  value of 0.1 K/W), I get the anticapted 2.5 deg C temperature rise.

Should combining a contact resistance for multiple faces into one feature always cause issues like this? It seems pretty tedious to have to do this for anything more than a few faces; for example, having to edit 10+ features all from 0.1 K/W to 0.2 K/W.

Thanks.