0
\$\begingroup\$

Can someone please help me find a datasheet or at least the thermal resistence for the following TO-220 heatsink?

Heatsink image Heatsink dimensions

My local supplier doesn't have more info about it. They just marked it with those numbers that I don't know what they mean: K6439 and 183001/15 (15 is the heatsink height in mm). It looks like it's made of aluminium.

\$\endgroup\$
  • 1
    \$\begingroup\$ As Brian says - comparison with the many heatsinks with data available will allow you a good guess. Factors include total surface area, access to air flow, vertical surfaces count much more than horizontal ones, thicker is better than thinner (as it gets heat to outer limbs better) and a very thin very wide heatsink can be worse than you may expect. Be sure to ensure good heat transfer from TO220 device to heatsink - either a thermally conductive rubber washer with known characteristics if electrical insulation required or direct contact with good thermal paste if electrical contact OK. \$\endgroup\$ – Russell McMahon Dec 8 '13 at 22:01
  • \$\begingroup\$ @RussellMcMahon you forget color. Black anodized vs. bare aluminum. \$\endgroup\$ – jippie Dec 9 '13 at 7:07
  • \$\begingroup\$ @Jippie : black looks cooler, but what does Stefan's law (fourth power law) really mean at (say) 10% above ambient, as opposed to dull red? I think Newtonian cooling is what matters. The other option is airflow, of course... \$\endgroup\$ – Brian Drummond Dec 9 '13 at 10:50
  • \$\begingroup\$ @BrianDrummond you'll have to explain that sentence in chat some time. \$\endgroup\$ – jippie Dec 9 '13 at 11:09
  • \$\begingroup\$ Basically, for small temperature differences, colour doesn't matter because conductive cooling to air predominates (proportional to temperature difference). Airflow helps by refreshing the local supply of cool air. For large temperature differences, radiation predominates(it's propoportional to the fourth power of dT) and THEN colour matters. What I suspect (without running the numbers) is that a well designed heatsink is mainly conductive cooling, where colour is unimportant. \$\endgroup\$ – Brian Drummond Dec 9 '13 at 11:19
1
\$\begingroup\$

Going to Farnell's heatsink search page and entering your dimensions and "TO220" as search parameters, I can't find the exact model, but there are bunches with thermal resistances in the order of 20-25C/Watt. Yours is a reasonable quality aluminium extrusion so better than some; I would use 20C/W as a reasonable figure unless you can find the actual data.

Which means at 7.5W from yesterday's question ... 150C expected temperature rise...

The other option is airflow, of course...

Theoretical approach: you could re-run those numbers (get a datasheet from a proper heatsink!) and see what airflow would make it work, and if a small fan (another datasheet) could push air that fast.

Experimental approach : attach a thermocouple (I got one with a £10/$15 multimeter - no longer sold but it came from Maplin) to the heatsink, and measure the temperature as you draw more current. (Start at 0.25A on one reg, given your other question. Point a CPU fan at it, add ducts, re-measure temperature. Repeat until satisfactory.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks! I'll probably lower the input voltage to around 15V or just use a switcher instead. \$\endgroup\$ – Ricardo Dec 9 '13 at 1:23
  • \$\begingroup\$ @Ricardo remember, the 15V should be at the 2A load, not un-loaded. \$\endgroup\$ – Passerby Dec 9 '13 at 11:25
  • \$\begingroup\$ @Passerby Right. That lesson I've already learnt. That's why I hate those cheap wall warts, they don't combine well with variable loads and linear voltage regulators in my boards. I'll probably just get a 12V regulated wall wart and drop the 12V regulators altogether. \$\endgroup\$ – Ricardo Dec 9 '13 at 11:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.