A few months back I lucked into some old CPU heatsinks. I found them in the proverbial University Science department hallway box. I think they are the old iteration of the Cooler Master X Dream K640. I say old because they're copper not aluminum.
The heatsinks in question
They have a luxuriously spacious surface for mounting to (54mm x 67mm).
The fan is a 12V 0.83A and according to the spec sheet it can do 800 ~ 2800 RPM with an airflow of 69.59CFM (Max). The fan however is controlled by a thermistor. I'm not sure how that will factor into using the heatsinks for lights.
Spec sheet:
http://www.coolermaster.com/service/sup ... -KIF-L9E1/
I am hoping to use these heatsinks to build a COB fixture and therefore need to work backwards to figure out what I can reasonably run on them. Since this is a CPU heatsink there does not seem to be a thermal resistance figure available for it from the manufacturer so I've take a stab at figuring it out for myself (please chime in if I have made any errors).
Doing some back of the envelope math (please ignore the volumetric calculations I did before I remembered its not the same thing as surface area) I have broken up the heat sink into four rectangles. A-C being the fins and D the base. It is an awkward thing to measure so where things were tricky I did my best to air on underestimating the measurement.
SA of a rectangular prism = 2ab+2bc+2ac
Fins (fin thickness = 0.33mm)
A. 2(39mm*17mm) + 2(17mm*0.33mm) + 2(39mm*0.33mm) = 1362.96mm2
B. 2(30mm*14mm) + 2(14mm*0.33mm) + 2(30mm*0.33mm) = 869.04mm2
C 2(5mm*19mm) + 2(19mm*0.33mm) + 2(5mm*0.33mm) = 205.84mm2
Total = 1362.96mm + 869.04mm + 205.84mm = 2437.84 * 57 fins = 138956.88mm2
Base
D. 2(54mm*6mm) + 2(6mm*67mm) + 2(67mm*54mm) = 8688mm2
SA fins + base = 138956.88mm + 8688mm = 147644.88mm2
Now using the Bridgelux AN30 Thermal management paper's rule of thumb, 64.5cm2 / 1 hWatt.
Pulling from http://ledgardener.com/find-heat-sink-cob/
147644.88mm2/645mm2/hWatt = 228.9 Watts
Or Growmau5's choosing your heatsink video part 4/7 of 110cm2 for passive or 40cm2 for active.
147644.88mm2/1100mm2/hWatt = 134.22 Watts
147644.88mm2/400mm2/hWatt = 369.11 Watts
These numbers seem to large to me. Where am I going wrong?
Cooler master X dream K640
- Jolly Green Giant
- LED Wizard
- Reactions:
- Posts: 835
- Joined: Thu Jun 15, 2017 1:35 am
- Location: somewhere in NY
I'm not a cpu guy.. and always have trouble figuring thermal math...I don't know if those numbers are high or if math slightly off. either way I went another route.
http://www.cpu-world.com/CPUs/K8/AMD-At ... ZWOF).html
that's the biggest cpu I believe that heat sink was designed for... max running temps are roughly same as we want for efficiency (try not to get above 50-60C) and it's TDP ( I had to see what TDP ment http://www.cpu-world.com/Glossary/T/The ... (TDP).html ) is 125watts....
I'm assuming you could use the same rough values for cobs... you probably could get away with almost 150w max each... if you only plan 50-75w on each you should be fine... if you want to try to push 150w on them.... start at half lower and watch temp over a few hours and work your way up.
http://www.cpu-world.com/CPUs/K8/AMD-At ... ZWOF).html
that's the biggest cpu I believe that heat sink was designed for... max running temps are roughly same as we want for efficiency (try not to get above 50-60C) and it's TDP ( I had to see what TDP ment http://www.cpu-world.com/Glossary/T/The ... (TDP).html ) is 125watts....
I'm assuming you could use the same rough values for cobs... you probably could get away with almost 150w max each... if you only plan 50-75w on each you should be fine... if you want to try to push 150w on them.... start at half lower and watch temp over a few hours and work your way up.
Keep in mind that the thermal conductivity coefficient of copper is twice better than that of aluminum.
Copper pots, besides the gorgeous looks, are better than others (e.g., steel, aluminum) for some purposes precisely because they give a better spreading of the heat.
With such a heatsink you can probably load quite a bit, but I'm not about to get number-y
Copper pots, besides the gorgeous looks, are better than others (e.g., steel, aluminum) for some purposes precisely because they give a better spreading of the heat.
With such a heatsink you can probably load quite a bit, but I'm not about to get number-y
@Jolly Green Giant - Some more tech savvy friends who don't have experience with lighting but lots with computers have said similar. My brain just really wants to be sure I know how far I can safely push it.
@majorana - Exactly! I'm surprised they were being thrown away. Copper heatsinks are hard to come by these days never mind free ones. I suspect the rules of thumb I used up top are probably geared towards steel/aluminum since they're more common these days. I'm just not super confident in my math. If I could be sure of that I could probably dig for more math to actually calculate a more precise number based on them being copper.
Same lab was also throwing these ones away: Not sure what I'm going to use them for. Crunching the numbers on something with heat pipes is a little above my mathematical pay grade.
@majorana - Exactly! I'm surprised they were being thrown away. Copper heatsinks are hard to come by these days never mind free ones. I suspect the rules of thumb I used up top are probably geared towards steel/aluminum since they're more common these days. I'm just not super confident in my math. If I could be sure of that I could probably dig for more math to actually calculate a more precise number based on them being copper.
Same lab was also throwing these ones away: Not sure what I'm going to use them for. Crunching the numbers on something with heat pipes is a little above my mathematical pay grade.