Heatsinks for DIY LED lamps

Imptspd

Well-Known Member
There is, I just wanted to help you set up your space in the safest way possible. Hopefully avoiding any catastrophic failures.
 

SupraSPL

Well-Known Member
So I've read through this entire thread and can't grasp some aspects for passive cooling. But from what I'm now thinking, from the graphs on page 1, the 10.00" profile from Hsu should effectively dissipate 4.25 watts per inch of cut length. Is this correct?
You have the right idea. Generally we go by surface area, but some heatsinks will be more suitable for passive cooling and some for active cooling. The ones with thicker plates will perform better and cost slightly more. Larger space between fins reduces the pressure needed to move the air "blanket" that is surrounding the fins. Passive cooling works best with the fins oriented vertically and most of our setups will have the fins running horizontally, but it can still get the job done by increasing surface area.
 

nomofatum

Well-Known Member
There is, I just wanted to help you set up your space in the safest way possible. Hopefully avoiding any catastrophic failures.
Yep, would have been an issue if I had been thinking just plain water. Thanks and good knowledge to add to the bank.
 

Imptspd

Well-Known Member
You have the right idea. Generally we go by surface area, but some heatsinks will be more suitable for passive cooling and some for active cooling. The ones with thicker plates will perform better and cost slightly more.
Thank you supra,
My idea for my setup is to size my heatsink to be able to run passively in the 1percent chance that a fan fails. Redundancy is what I'm shooting for. I'd hate to lose multiple $65 cxa3070 ab cobs because of a fan going out while im at work, but still aquire the benefit of running the cxa's cooler with a fan or two.
 

nomofatum

Well-Known Member
Thank you supra,
My idea for my setup is to size my heatsink to be able to run passively in the 1percent chance that a fan fails. Redundancy is what I'm shooting for. I'd hate to lose multiple $65 cxa3070 ab cobs because of a fan going out while im at work, but still aquire the benefit of running the cxa's cooler with a fan or two.
If you get long heat sinks make sure there are a few cuts so you can get good air flow without a fan. Long small channels don't support great air flow, so a cut in the fins creates enough opening to help get air in without a fan.
 

Imptspd

Well-Known Member
If you get long heat sinks make sure there are a few cuts so you can get good air flow without a fan. Long small channels don't support great air flow, so a cut in the fins creates enough opening to help get air in without a fan.
Thank you nomofatum,
I will keep this in mind with my next build.
 

Imptspd

Well-Known Member
It's not hard to make the cuts yourself if you haven't assembled yet. Just need the right saw/grinder bit.
Oh yes I know ha ha. The only light I have right now is 10 inches long and using 3w diodes, for that I don't see a point in messing with it. My next light I plan on making much bigger. I have the tools and knowledge for modifying soft metals like aluminum (worked 5 years as a 3axis cnc machinist). Thank you for the ideas though, it is appreciated. :)
 

SupraSPL

Well-Known Member
I am running some tests on the 8.46" profile, which is similar to the 1.8, 3.5, 4.23, 4.9, 5.9, 5.4, 7.3, 10 profiles. Those profiles seem like they would be optimal for passive cooling because they have thick bases, short fins and wide spacing between fins.
DSC07864a.jpg

So first off, passive cooling at 45cm²/W. A pair of CXA3070s running at 1.4A (42% efficient minimum). I am measuring approximately 7.4% temp droop with no air movement in the room. Heatsink temps between fins range from 33-35C. The temp gradient was very good even though I put the CXAs near the corners of the heatsink.

Then I added a low RPM 140mm fan running at 5V. It does not cover all the fins but it is moving a good amount of heat out and is dead silent. Fan power is only .4W, an insignificant % of the total dissipation ~100W. The results were as expected at 45cm² active cooled, 3.3% temp droop and heatsink temps between fins ranged from 26-28C.

For comparison, the Alpine 11 with CXA3070 AB at 1.4A measured 1.6% temp droop and at 2.23A, 3.5% temp droop, using about 1W of fan power.

Next I will check the passive cooling performance at 66 and 77cm²/W. Maybe a s point of diminishing returns will jump out at us so we can use less than 90cm²/W, but without air movement, 45cm²/W passive is not optimal IMO.
 
Last edited:

nomofatum

Well-Known Member
I am running some tests on the 8.46" profile, which is similar to the 1.8, 3.5, 4.23, 4.9, 5.9, 5.4, 7.3, 10 profiles. Those profiles seem like they would be optimal for passive cooling because they have thick bases, short fins and wide spacing between fins.
View attachment 3321075

So first off, passive cooling at 45cm²/W. A pair of CXA3070s running at 1.4A. I am measuring approximately 6.5% temp droop with no air movement in the room. Heatsink temps between fins range from 33-35C.

Then I added a low RPM 140mm fan running at 5V. It does not cover all the fins but it is moving a good amount of heat out and is dead silent. Fan dissipation is about .5W an insignificant % of the total dissipation ~100W.

Next I will check the passive cooling temp droop at 60, 70, and 80cm²/W.
Those do look very good for passive cooling, they would benefit a bit from the cuts as I was referring to though. Not nearly as much as a narrow channel one would benefit though. Pointing a wall fan to blow across your "passive" cooling works wonders too.
 

SupraSPL

Well-Known Member
The fan sticker on it says .1A 8cfm just tested using it on an old 12v supply for a Remington electric shaver that died on me a couple years ago, actually its the power supply for my DIY colloidal silver gen. Im sure its not the most efficient supply, I plan to use a MW LPV-35-12 for the 12v supply also not that efficient only 84% but should have MW reliability and only $15 is fairly reasonable.

Kill-A-Watt
One 50mm on Remington 12v supply .56PF 1.6w .02A
Two 50mm on same supply .60PF 3.2w .04A
Good info thnx BT. Will PM you about the CS generator. I have read a lot of conflicting data about it so I just bought some nano silver until I find out more solid info.
 

SupraSPL

Well-Known Member
Those do look very good for passive cooling, they would benefit a bit from the cuts as I was referring to though. Not nearly as much as a narrow channel one would benefit though. Pointing a wall fan to blow across your "passive" cooling works wonders too.
Ya the circulation fan can make a huge difference and it is "free" cooling since it would be running during lights on anyway. Unfortunately in my case the air from the circ fan does not get to all heatsinks equally. I should test the 45cm² in the actual grow space and see what heatsink temps I end up with.
 

getsoutalive

Well-Known Member
Ya the circulation fan can make a huge difference and it is "free" cooling since it would be running during lights on anyway. Unfortunately in my case the air from the circ fan does not get to all heatsinks equally. I should test the 45cm² in the actual grow space and see what heatsink temps I end up with.

Was gonna ask if you could have an external fan hit it and check again. Off axis, or even an oscillator. Doesn't take much to pull much of that heat away.

Just to confirm the math from before. In your example above, you said 2 x 3070 @ 1.4a. So about 104 watts of dissipation @ 42% efficiency means about 60 watts of heat.

The 8.46" model has a c/w/3" of 1.10. I come up with a 12" length based on the 45cm2/w, correct?

So to plug in the numbers.

104w total x 58% loss as heat = 60w of heat to dissipate.
60w x 1.10 c/w/3" = 66 degrees per 3 inches of sink
66 / 4 = 16.5 degrees C above ambient.

Sound like you did better than that, unless ambient was rather low.
 

SupraSPL

Well-Known Member
Lets see if we can account for the difference. You are correct, a 12" piece of the 8.46" profile. I figure about 4800cm² and 103W of dissipation. Ambient was about 22C. The efficiency of the COBs was estimated at 42% minimum assuming Tj of 50C. "Typical" efficiency is 43.3%. So that means 56.7W of heat or 15.6 degrees over ambient. That should be about 37.8C.

I was measuring between the fins and the hottest spot was directly behind the COB, about 35C. But the hottest part on the front of the heatsink may have been 37C+. It is very hard to measure there with an infrared thermometer because the output of the COBs interferes with the measurement (and blinds the measurer) so I have been going by the hottest parts in between the fins instead. That may account for the discrepancy. But it does demonstrate the usefulness of the C/W statistic, thank you for that :)

It may also be a way for us to work backwards and find the approximate efficiency of a COB. I am very curious about the efficiency of these 100W generic COBs that are selling for a dollar or two each. Maybe not useful as a growing light but sure could be useful for short term home lighting situations.
 

SupraSPL

Well-Known Member
DP, because that is a constant current driver, I think it would try to pump 300mA into the A11 fan. Normally that fan draws ~175mA at 12V. Maybe the driver would try to put out a higher voltage to bring the fan to 300mA? Not sure but if you try it be aware that the fan might die LOL. I have killed a few fans by accidentally hooking them up to CC drivers.
 

SupraSPL

Well-Known Member
Repeated the passive cooling test at 1050mA, 65cm²/W. COB efficiency should be 46.7% typical. This test may be more accurate than the last because I am using an adjustable driver and taking driver efficiency into account from cool to warm.

Ambient temp 22C
lux: 3920
COB dissipation 75.15W

heatsink thermally stable at 30-32C
lux: 3690
COB dissipation 74.2W (due to Vf shift, .45V ea COB)

Temp droop = 4.66%
 
Last edited:

SupraSPL

Well-Known Member
Results
1450mA (45cm²/W) - 7.44%
1450mA (45cm²/W) - 3.32% (active cooled with .38W of fan power)
1050mA (65cm²/W) - 4.67%
900mA (77cm²/W) - 3.3%
800mA (85cm²/W) - 3.3%

So looking at that, 65cm² seems like a good minimum compromise for COBs running at high efficiency.
 
Last edited:
Top