Here are my ideas on "uniform heating" - several scenarios to toss around.

a. Cover 85% of base area. With Seri's spreading calculation, and the test case I used, the total resistance increased by 0.015 C/W for a thick base (.25"), and .023 C/W for a thin one (.125"), from 0.71 C/W with uniform heating, or just over 2% and just under 4%, respectively - well within the error you can expect from other sources.

b. Using a spreading calculation or other means, show that the reduced-size result is within x% of the uniform-heat result. For example if x=5%, the minimum heated/base area ratio in my test cases was 73% for a thick base (.25") and 80% for a thin base (.125"). The area ratio will depend on the base thickness and the effective heat transfer coefficient on the fin side.

c. "Uniform" is anything that gives less than x% error (closely related to "b"). Maybe the standard could contain design charts in addition to Seri's description of the calculation method. We will need to agree on what x should be.

I think we agreed in the last meeting that the "heat sink temperature" should be measured over the center of the heat source. My approach is that the temperature sensor should be embedded in the base within 1 mm of the surface. I recommend using milled thermocouple grooves with a hook shape at the end, to minimize temperature gradients in the wire near the sensor. If milling is not an option, a hole drilled from the fin side would be ok, as long as the wire temperature gradients are mimized by fastening it to the heat sink base surface for at least 10 wire diameters (K type) or 100 diameters (T type). I believe the number of wire diameters comes from a Bob Moffat recommendation.

Let me know how the discussion goes...
Cathy