Heatsinks for DIY LED lamps

bicit

bicit

Well-Known Member
Abiqua said:
I would also think that @ HSUSA, you would be able to replicate the Rapid Led sink. They offer machine work now, so you could even offer up a custom pattern for machined holes.

I have no idea on price though.....:peace:
Click to expand...
It's expensive, double to quadruple the price of the heat sink itself. It's an option though, just need a drawing.

Yes they can do metric threads.
 
Last edited:
bicit

bicit

Well-Known Member
SupraSPL said:
Cool setup @kamikaza , thanks for sharing that info :joint:

@Abiqua Back in the day I used 150cm²/W but the LEDs were about 26% efficient. 110cm²/W should be good for LED running at 35-40% efficiency. I am running my CXAs at 48% efficiency minimum and I use about 95cm²/W to get heatsink temps of 27C, almost as good as active cooled performance. So if you are running the Vero 10s at 37% typical, you could use anything from 90-110cm² depending on your goals.
Click to expand...
Any chance you've been able to run any tests on the 5.886 profile? Hungry minds and all :P
 
AquariusPanta

AquariusPanta

Well-Known Member
bicit said:
Any chance you've been able to run any tests on the 5.886 profile? Hungry minds and all :P
Click to expand...
@SupraSPL

Other then the materials, like a HS, COB, etc., what sort of tools do you use when measuring the variables from these tests, such as temperatures and light?

I'd like to think that we, as a community, could collectively work together on this passive & active cooling adventure; you as the leader.

As of now, I personally believe this to be the most intriguing and dynamic thread on the LED forum. It's like the next level of COB DIY threading. We started with COBS and drivers, now were sort of shifting towards the thermodynamics behind H/S and COBs.

We need more data!!!
 
R

rockett899

New Member
SupraSPL said:
Cool I did not know that and I learned a new term.


Yep that should work. So you need a total of 130vF. If you split them in 2 strings, the longer one would need 78Vf. This one should work but it would probably be only 270mA. I will look around to see what else I have that would work. If you want high dimming you could use the HLG-60H-C350A
Click to expand...
This is totally off topic to whatever you guys are talking about but would you guys know how important the heat sink is for this model of LED- http://au.mouser.com/ProductDetail/Cree-Inc/CXA3070-0000-000N00Z230F/?qs=sGAEpiMZZMu4Prknbu83y0bObCPMgyC/l6eQby5L/DGQDb9TViwpOg==
I have just bought these chinese heat sinks and the lights look like they would fit perfect but im worried these heats sinks arent big enough. Dont want to screw them up from over heating...Do they even get hot?? Im waiting for drivers so i havent had the chance to turn them on :/
does anyone have advise i'm new to this. - http://www.ebay.com.au/itm/55mm-Cooling-Fan-Heatsink-Cooler-for-PC-Computer-Laptop-CPU-VGA-Video-Card-OG-/271588731336?pt=AU_Components&hash=item3f3bf329c8

Cheers
 
SupraSPL

SupraSPL

Well-Known Member
Thanks to @foreverflyhi , I hope you don't mind I used your 5.88" module for this quick test :leaf:
DSC07719a.jpg DSC07718a.jpg

Ambient temp 23C
5.88" X 12" = 3120cm²
heatsink surface was flattened and polished to 1000 grit - (this may not be worth the effort for large builds)
TIM = Prolimatech PK3 paste
CXA3070 AB X 2 @ 1.45A = 105W dissipation
30cm²/W

high RPM 140mm fan running at 5V (.123A * 4.9V = .6W)
heatsink temp stabilized at 27C
light lost to temp droop = ~3%

increased fan to 9V (.237A * 9.15V = 2.17W)
heatsink temp stabilized at 25.5C
light lost to temp droop = ~2%

passive cooled at 90cm²/W
heatsink temp ~31C
light loss to temp droop = 2.88%

passive cooled at 60cm²/W
heatsink temp ~33C
light loss to temp droop = 6.98%

passive cooled at 30cm²/W
heatsink temp ranges from 36-40C (may have gotten hotter with more time, but not much hotter)
light loss to temp droop = 8.29%

So if we call baseline efficiency of the COB 42%, this is how they stack up when we take the electrical consumption of the fan into account:
Active cooled 5V 30cm²/W = 40.56%
Active cooled 9V 30cm²/W = 40.2%
Passive 90cm²/W = 40.89%
Passive 60cm²/W = 39.16%
Passive 30cm²/W = 38.6%

Once again the 5V outperformed 9V (I should try 7.5V). The passive cooling numbers are impressive, confirming the results from the 10.08" profile (page 5), although 90cm²/W is an expensive proposition. In my test there was very little air movement and the heatsink was not in an optimal position for convection, so in practice I expect the passive cooling could have performed even better.
 
Last edited:
NapalmD

NapalmD

Well-Known Member
SupraSPL said:
Thanks to @foreverflyhi , I hope you don't mind I used your 5.88" module for this quick test :leaf:
View attachment 3306793 View attachment 3306792

Ambient temp 23C
5.88" X 12" = 3120cm²
heatsink surface was flattened and polished to 1000 grit - Prolimatech PK3 paste
CXA3070 AB X 2 @ 1.45A = ~105W dissipation
30cm²/W
high RPM 140mm fan running at 5V (.123A * 4.9V = .6W)
heatsink temp stabilized at 27C
light lost to temp droop = ~3%

increased fan to 9V (.237A * 9.15V = 2.17W)
heatsink temp stabilized at 25.5
light lost to temp droop = ~2%
Click to expand...
Sweet! That's exactly what I want to set up.
 
Positivity

Positivity

Well-Known Member
Good info..

Could even do a thinner profile with those numbers. They don't take much to cool at the lower currents.

I cant even believe how cool my closet is when i walk in and all the lights are running lately. These leds are worth every penny..
 
bicit

bicit

Well-Known Member
SupraSPL said:
Thanks to @foreverflyhi , I hope you don't mind I used your 5.88" module for this quick test :leaf:
View attachment 3306793 View attachment 3306792

Ambient temp 23C
5.88" X 12" = 3120cm²
heatsink surface was flattened and polished to 1000 grit -
CXAs mounted with Prolimatech PK3 paste
CXA3070 AB X 2 @ 1.45A = 105W dissipation
30cm²/W
high RPM 140mm fan running at 5V (.123A * 4.9V = .6W)
heatsink temp stabilized at 27C
light lost to temp droop = ~3%

increased fan to 9V (.237A * 9.15V = 2.17W)
heatsink temp stabilized at 25.5
light lost to temp droop = ~2%
Click to expand...
Can you do a simulate a fan failure? Or what do the heatsink temps look like if the fan isn't used?
 
SupraSPL

SupraSPL

Well-Known Member
Good call bicit, updated the post with passive cooling results.

O crap I forgot the paste! JK it is under there. I cleaned the edges with ISO so might be hard to see the paste. I used the phrase "mounted with paste" but I should have said paste and tape.
DSC07719b.jpg
 
bicit

bicit

Well-Known Member
SupraSPL said:
passive cooled at 30cm²/W
heatsink temp ranges from 36-40C (may have gotten hotter with more time, but not much hotter)
light loss to temp droop = 8.29%

So if we call baseline efficiency 42%
Active cooled 5V 30cm²/W = 40.56%
Active cooled 9V 30cm²/W = 40.2%
Passive 90cm²/W = X%
Passive 60cm²/W = 39.16%
Passive 30cm²/W = 38.6%

Once again the 5V outperformed 9V (I should try 7.5V). Amazingly, the passive cooling numbers are not far off the mark. In my test there was very little air movement and the heatsink was not in a noptimal position for convection, so in practice I expect the passive cooling could have performed even better.
Click to expand...
That's pretty impressive passive performance all things considered. No real danger of frying the cobs if the fan does fail. The CPU coolers tend to act like a solid block of aluminum without a fan.
 
bicit

bicit

Well-Known Member
Crude math, just stinking.

1x vero 29, 70cri, 5000k+molex EZ-mate: $35
1x 2"x5.886" Heatsink: $3.50
1x Fasttech driver Link: $2.50
=$41+($41*0.15)= $48

Estimating about 61% efficient even after temp droop, 10.5w power draw.

A bit expensive but it would beat the pants off a CFL and would be fairly easy to assemble.
 
Last edited:
Dloomis514

Dloomis514

Well-Known Member
bicit said:
Crude math, just stinking.

1x vero 29, 70cri, 5000k+molex EZ-mate: $35
1x 2"x5.886" Heatsink: $3.50
1x Fasttech driver Link: $2.50
=$41*($41*0.15)= $48

Estimating about 61% efficient even after temp droop, 10.5w power draw.

A bit expensive but it would beat the pants off a CFL and would be fairly easy to assemble.
Click to expand...
How in the heck is that Fastech site searchable? http://www.fasttech.com/products/1612/10008079/1714800
 
SupraSPL

SupraSPL

Well-Known Member
bicit said:
Crude math, just stinking.

1x vero 29, 70cri, 5000k+molex EZ-mate: $35
1x 2"x5.886" Heatsink: $3.50
1x Fasttech driver Link: $2.50
=$41+($41*0.15)= $48

Estimating about 61% efficient even after temp droop, 10.5w power draw.

A bit expensive but it would beat the pants off a CFL and would be fairly easy to assemble.
Click to expand...
I had a similar idea and a pair of 5000K CXA3590s hanging around. Running them at 250mA and no temp droop to speak of, they are about 65% efficient or 210lm/W. The 5000K Vero 29 at 250mA should be about 58% efficient or 188lm/W. Awesome vegging lights! Cost is $5-$5.50 /PAR W, much less than I recently paid for my Oslon SSL deep reds LOL.

Mounted on Alpine 11s so they can be cranked up if necessary.
DSC07361a.jpg
 
Last edited:
nogod_

nogod_

Well-Known Member
Sexy. There is beauty in simplicity.

Repeat test @ .7a & 1.05a?

(For science)
:rolleyes:

SupraSPL said:
Thanks to @foreverflyhi , I hope you don't mind I used your 5.88" module for this quick test :leaf:
View attachment 3306793 View attachment 3306792

Ambient temp 23C
5.88" X 12" = 3120cm²
heatsink surface was flattened and polished to 1000 grit - (this may not be worth the effort for large builds)
TIM = Prolimatech PK3 paste
CXA3070 AB X 2 @ 1.45A = 105W dissipation
30cm²/W

high RPM 140mm fan running at 5V (.123A * 4.9V = .6W)
heatsink temp stabilized at 27C
light lost to temp droop = ~3%

increased fan to 9V (.237A * 9.15V = 2.17W)
heatsink temp stabilized at 25.5C
light lost to temp droop = ~2%

passive cooled at 90cm²/W
heatsink temp ~31C
light loss to temp droop = 2.88%

passive cooled at 60cm²/W
heatsink temp ~33C
light loss to temp droop = 6.98%

passive cooled at 30cm²/W
heatsink temp ranges from 36-40C (may have gotten hotter with more time, but not much hotter)
light loss to temp droop = 8.29%

So if we call baseline efficiency of the COB 42%, this is how they stack up when we take the electrical consumption of the fan into account:
Active cooled 5V 30cm²/W = 40.56%
Active cooled 9V 30cm²/W = 40.2%
Passive 90cm²/W = 40.89%
Passive 60cm²/W = 39.16%
Passive 30cm²/W = 38.6%

Once again the 5V outperformed 9V (I should try 7.5V). The passive cooling numbers are impressive, confirming the results from the 10.08" profile (page 5), although 90cm²/W is an expensive proposition. In my test there was very little air movement and the heatsink was not in an optimal position for convection, so in practice I expect the passive cooling could have performed even better.
Click to expand...
 
nogod_

nogod_

Well-Known Member
Do you guys paint the heatsink or the cob with paste?

If you are making a bunch, why not use a stencil? Or if youve got the tools a silkscreen?

SupraSPL said:
Good call bicit, updated the post with passive cooling results.

O crap I forgot the paste! JK it is under there. I cleaned the edges with ISO so might be hard to see the paste. I used the phrase "mounted with paste" but I should have said paste and tape.
View attachment 3306869
Click to expand...
 
epicfail

epicfail

Well-Known Member
SupraSPL said:
Thanks to @foreverflyhi

Ambient temp 23C
5.88" X 12" = 3120cm²
heatsink surface was flattened and polished to 1000 grit - (this may not be worth the effort for large builds)
TIM = Prolimatech PK3 paste
CXA3070 AB X 2 @ 1.45A = 105W dissipation
30cm²/W

So if we call baseline efficiency of the COB 42%, this is how they stack up when we take the electrical consumption of the fan into account:
Active cooled 5V 30cm²/W = 40.56%
Active cooled 9V 30cm²/W = 40.2%
Passive 90cm²/W = 40.89%
Passive 60cm²/W = 39.16%
Passive 30cm²/W = 38.6%
Click to expand...
So what you're saying is I went overboard, no surprise its what I'm good at.

My setup:
5.88" x 18" = 4680cm²
Sanded to 1000g and anodized black
CXA3070 AB X 2 @ 1050mA = 76W dissipation
61cm²/W so it should be just under 40% efficient if run passive.

The thing is I also have a 12V 140mm 3000 RPM PWM fan mounted on to it and its a power hog which I guess is actually bringing the overall efficiency down. Right now the ambient temps in my tent are 22.5C at this time of year. The hottest part of the heatsink is only 25.8C according my IR thermometer, I guess I could just turn the fans off but that means I would need to add the circulation fans again. I removed them because the PC fans move so much air they weren't needed, I should see what uses less wattage the 4x PC fans or 2x 6" clip fans.

I'm lazy and its probably just going to stay the way it is for a while at least because everything seems happy.
 
Abiqua

Abiqua

Well-Known Member
guod said:
1k ohm 1 amp. Linear Pot...never..... maybe 1kOhm 1Watt acording the size
1k is also way to high... i think more at 0.1k which is 100 Ohm
Click to expand...
Your right @guod, it is actually a .5 watt. But the switch rating is 1.0A @ 125vac, that's what I was rattling off....here are all the specs...
potent.jpg
 
You must log in or register to reply here.
Top