• Here is a link to the full explanation: https://rollitup.org/t/welcome-back-did-you-try-turning-it-off-and-on-again.1104810/

Heatsinks for DIY LED lamps

SupraSPL

SupraSPL

Well-Known Member
Uh oh I fried an egg in the 1.05A passive test. I let it get much hotter than I normally would (hottest part of the heatsink ~60C+) and now the Vf has risen (permanently) from 37.3 to 41 at 1.45A. Along with the power consumption increase, I am also noticing a decrease in output of about 20-25%. Maybe this is why:
DSC07767A.jpg

The max Tc for 1.05A is 125C and I was nowhere near that, so I am surprised that it was damaged. One way to find out how much is too much, a sacrifice for science :joint:
 
Last edited:
Dloomis514

Dloomis514

Well-Known Member
SupraSPL said:
Here is every test with the Alpine 11 and CXA3070. 700mA, 800mA, 1.05A, 1.43A, 2.23A. The 1.05A and under tests were much easier and more accurate thanks to the Mean Well HLG-185H-C1050B. The efficiency numbers are based on typical figures and consistent with the actual figures, so they are directly comparable to any of the others in these spreadsheets.
Click to expand...
Conclusion. When in doubt go with 7.5 volts.
 
P

PurpleBuz

Well-Known Member
SupraSPL said:
700mA, 800mA and 1050mA tests on the Alpine 11 + CXA3070 AB coming up. One thing worth mentioning, at lower currents the heatsinks run so close to ambient temp that 5V is all you need. Increasing fan speed does not increase light output at all. But when I tried running it passive, it gets smokin hot even at 700mA.
Click to expand...
so the alpine 11 will cook your eggs when run passively at 700ma ?

For safety in case of fan failure, seems better to use a cooler that can handle the load passively at the upper limits of the cob.
 
PSUAGRO.

PSUAGRO.

Well-Known Member
PurpleBuz said:
so the alpine 11 will cook your eggs when run passively at 700ma ?

For safety in case of fan failure, seems better to use a cooler that can handle the load passively at the upper limits of the cob.
Click to expand...
A cheap cpu cooler is ======= a cheap cpu cooler:wink:...................granted modern computer fans have a very low failure rate. That's the risk you take when go with the affordable active cooling design.......wiring in a temp shut-off isn't a bad idea imo.
 
SupraSPL

SupraSPL

Well-Known Member
I wonder if the thermal switch will work on the DC circuit without messing up the Vf too much. I know it is bad to leave the driver unloaded, but not as bad as a fried golden egg. If they will work on the DC circuit, that would avoid (in my setup) running AC power into the grow space.

Also, if you added a normally open switch, it could cycle your fan on an off saving some power and keeping your heatsinks under say 30-35C.

http://www.ebay.com/itm/5pcs-KSD301-30C-86-F-Thermostat-Normal-Open-NO-Temperature-Control-Switch-Temp-/251478583470
 
Last edited:
T

Tazbud

Well-Known Member
When I started thinking 'round', search through a heap of these looking for deep with a large core (if not copper):

Screen Shot 2014-12-12 at 2.55.36 AM.png

Pairing with a good fan or outlet fan. Maybe still too hot it's sounding like.
 
Last edited:
B

bondoman

Well-Known Member
Just get a good PC power supply for the fans, they have both the 12v and 5v rails, and will pretty much last forever.
 
SupraSPL

SupraSPL

Well-Known Member
Yes those will do the job but AC-DC power supplies are most efficient when they are running at 60% or more of their max amperage, preferably 80% of max. So if we were to use a large PSU we could end up with as low as 10-20% efficiency because of the overhead. Even the high efficiency PC PSUs could perform poorly with a very small load. So that makes small adapters appealing we can size them to the number of fans and get 70-80% efficiency. They are also cheap, compact, swappable.
 
B

bondoman

Well-Known Member
yep it wouldn't make sense for a small setup, as with the unscientific test I did the PS drew about 8w without a load, and then whatever amps the fans were. But with a larger setup it makes a lot of sense, as the efficiency goes up with each fan added.
 
T

Tazbud

Well-Known Member
(just to add) this was the other option I had saved, passive cooling but almost come full circle as far as size weight and cost v's channel h/s of the same mass/cooling.

Screen Shot 2014-12-12 at 4.57.03 AM.png
 
Mechmike

Mechmike

Well-Known Member
PurpleBuz said:
so the alpine 11 will cook your eggs when run passively at 700ma ?

For safety in case of fan failure, seems better to use a cooler that can handle the load passively at the upper limits of the cob.
Click to expand...
That is how my main light is designed. A lot of heat sink with a little airflow. With no airflow the sink tops out at 42c. With both fans running at 6v it tops out at 28c. I am pretty confident that the light will survive if the cooling fails.DSCN0286.JPGDSCN0289.JPG
 

Attachments

mc130p

mc130p

Well-Known Member
SupraSPL said:
Uh oh I fried an egg in the 1.05A passive test. I let it get much hotter than I normally would (hottest part of the heatsink ~60C+) and now the Vf has risen (permanently) from 37.3 to 41 at 1.45A. Along with the power consumption increase, I am also noticing a decrease in output of about 20-25%. Maybe this is why:
View attachment 3311214

The max Tc for 1.05A is 125C and I was nowhere near that, so I am surprised that it was damaged. One way to find out how much is too much, a sacrifice for science :joint:
Click to expand...
Do you think that this cob actually burned at 60C, or is there a possibility that the temp at the cob was much higher than measured? If it really crapped out at 60C, then I feel like that particular cob must have been a dud....
 
medicinehuman

medicinehuman

Well-Known Member
I had some CXA 3070 Z4's run for several hours without a fan a couple of times on artic 11 units. They seemed hotter than :fire:when I found them. But there is no noticeable damage that you can see by the eye and they are still kicking out some good Bud. It has been about 6 months or so, is my first DIY. Hopefully yours was a dud I guess.:dunce:
 
You must log in or register to reply here.
Top