calculating Tc by temp droop/voltage

since Tc is such a hassle to setup and measure, critique this method for a given heatsink/chip combo to determine its limits:

-turn on chip at given current with regulated power supply
-measure instantaneous PAR and voltage
-wait 10 minutes or until PAR and voltage reach steady state, record
-back-calculate Tc and/or compare droop with your 'acceptable limit'

what would be the measure of success of successful cooling? 55C Tc? 85C Tc? 5% droop? 10% droop?

seems that all the LM data is done at 85C so if youre not starving for efficiency seems the chips wont mind it too much..

i think most of us do start for efficiency though


for me 65c and 5% would be acceptable

I'm all for the relatively small one time cost of doing it well for a permanent increase in par output. Both active and passive cooling benefit from thicker bases, more surface area and increased airflow.

thanks rahz, its an effort to put some substance behind the age old "what size heatsink do i need for X chip at Y wattage" seems like a 3x3 matrix of LES and heatsinks would give good data. if nothing else provide a metric for users like "ok i can run my chip out to 95W on a 140 x 80 with 10% droop without sacrificing life"

Yes it would be good to know, with two or three potential sinks for instance how many more watts you can get at the same efficiency with the bigger versions and make an informed decision based on preference. The datasheets provide enough data to describe droop, efficiency change and output, but of course many customers won't be studying datasheets.