Heatsinks for DIY LED lamps

SupraSPL

SupraSPL

Well-Known Member
Hey BT, it would be no problem to run the Vero29 at 2.1A on the Alpine11 even though the fan is relatively low rpm, 5V creates enough pressure to move the air between fins without overdoing it. So in this case low RPM works out in our favor because 5V adapters are plentiful, cheap and easier to load match. Approx power consumption of the Alpine11 fan:
5V = .3W (60mA)
6V = .45W (75mA)
7.5V = .8W (105mA)

You could run it at 5V and get ~2.5% temp droop. At 6V, ~2% temp droop. 7.5V bumps into the law of diminishing returns with no improvement in temp droop.

From every test I have done so far, the only time I have seen results that would justify running the Alpine11 at 7.5V was when testing CXA3070s at 2.5A. At 5V temp droop was ~4.5-5.5% and at 7.5V it was ~3.5-4.5%.
 
Last edited:
SupraSPL

SupraSPL

Well-Known Member
Yes it is a good match for (2) CXA3070s at 1.45A as long as you mean active cooled. On the other hand you might get better results if they were spread a further apart like 4.9"X14" or 4.23" X 18"
 
C

churchhaze

Well-Known Member
It's hard to go by surface area alone when picking a passive heat sink. In general, more, wider fins, spaced far apart, is better than long channels with thin fins, spaced closer together. That makes it hard to come up with a rule of thumb or equations to calculate the perfect size heat sink for passive.

A cxa3070 at 1.4mA requires a pretty big ass heat sink to run passively at temperatures of around 50C, ideally one of the widest profiles.

 
Last edited:
A

Andrew1981

Member
Where are you guys measuring the temp? I'm not sure if I am measuring right I have very 18 on a large heats ink 40 inches long with one 120 fan in the middle obviously the sides are a little hotter. If I measure on the top where the fins are I'm getting like 27 c but I measure the light itself it goes wayyyyyyy up.not sure where to measure to make sure I am within good range thanks all
 
JavaCo

JavaCo

Well-Known Member
Andrew1981 said:
Where are you guys measuring the temp? I'm not sure if I am measuring right I have very 18 on a large heats ink 40 inches long with one 120 fan in the middle obviously the sides are a little hotter. If I measure on the top where the fins are I'm getting like 27 c but I measure the light itself it goes wayyyyyyy up.not sure where to measure to make sure I am within good range thanks all
Click to expand...

TJ temp is measured where the red circle is by using a thermo coupler.
 
SupraSPL

SupraSPL

Well-Known Member
churchhaze said:
It's hard to go by surface area alone when picking a passive heat sink. In general, more, wider fins, spaced far apart, is better than long channels with thin fins, spaced closer together. That makes it hard to come up with a rule of thumb or equations to calculate the perfect size heat sink for passive.

A cxa3070 at 1.4mA requires a pretty big ass heat sink to run passively at temperatures of around 50C, ideally one of the widest profiles.

Click to expand...
I think those formulas are assuming vertical orientation, which is much better for passive cooling than horizontal. Some of my modules that are used for side lighting are slightly tilted and it does allow the air to convect through the channels better, resulting in a noticeably cooler heatsink. Even for horizontal applications the shorter fin lengths might help slightly, but in that case getting the widest shortest fins makes use of the slight breeze and allows for more convection.
 
C

churchhaze

Well-Known Member
While vertical orientation would be better (for passive), and the formulas are for vertical orientation, I think shorter fin channels are still better for horizontal orientation because it means more air intakes. Longer channels/fins means the middle of the channels will have much warmer air flowing over it than the outsides of the channels that are closer to the source of ambient air. This means the average air temperature between the channels is lower with shorter channels.

Longer channels for active cooling is actually more of a cost saver, as you require fewer fans, but you could theoretically have 2 fans for twice as many channels, and run them at a lower speed and get the same cooling performance.

SupraSPL said:
I think those formulas are assuming vertical orientation, which is much better for passive cooling than horizontal. Some of my modules that are used for side lighting are slightly tilted and it does allow the air to convect through the channels better, resulting in a noticeably cooler heatsink. Even for horizontal applications the shorter fin lengths might help slightly, but in that case getting the widest shortest fins makes use of the slight breeze and allows for more convection.
Click to expand...
 
Last edited:
SupraSPL

SupraSPL

Well-Known Member
Actual measurement of junction temp (Tj) is not possible for us it has to be done is a lab, but we can "try" to measure case temp or TC, the point that Java mentioned. I have found it nearly impossible to get repeatable accurate results measuring Tc because the emitted photons interfere with the readings, whether using thermocouple, mercury or infrared. So I have resorted to measure the hottest part of the heatsink that I can without letting the photons interfere and I use that as a guide. It has proven to be reasonably repeatable and it correlates with temp droop well. If I can get between the heatsink fins I look for the hottest part I can find between the fins on the base plate and the hottest part I can find on the front of the base plate without the light interfering. Generally I try to keep it below 32C but 28C is very good.

Another way you can gauge Tj is by monitoring the Vf drop from a 25C start as it warms up, because from a cold start we know for a fact that Tj is the same as ambient. So if a CXA3070 decreases by .15Vf and is running at the same current, you can translate that into 15 degrees C. So then you would add the case temp + 15 to get junction temp.
 
Last edited:
A

Andrew1981

Member
Ah ok Ty for response. So the vero 18 as long as I stay under 50 c I'm ok. I measured all around on top near the cob everywhere I'm like maxing 30 c but mostly 27 so then my setup is ok I. Assuming?
 
C

churchhaze

Well-Known Member
Here's a heat sink with some extra doodles attached to show what I mean visually.

Fingerpainting simulation time. This is why I think it's better for passive sinks to have short channels/fins. The heat from convection causes air to be pulled in from the sides.

KX3HS.JPG

Pure-Aluminum-font-b-heat-b-font-font-b-sink-b-font-good-cooling-font-b.jpg
 
Last edited:
JavaCo

JavaCo

Well-Known Member
Andrew1981 said:
Ah ok Ty for response. So the vero 18 as long as I stay under 50 c I'm ok. I measured all around on top near the cob everywhere I'm like maxing 30 c but mostly 27 so then my setup is ok I. Assuming?
Click to expand...
You can run the chips hotter then 50c and they will be fine they wont burn up. The main reason you want them as cool as possible is the cooler you run them the more light they put out. If you look at the data sheet they test them @ 25c and 85c. You get a pretty big lumen drop when the chips are at a 85c TJ temp. So 30c is pretty good unless your ambient temps are really low.
 
I

Imptspd

Well-Known Member
SupraSPL said:
Nice work lmpt! Is that a 3000K for flowering?

CH, love the lamp! What kind of heatsink temps are you getting?
Click to expand...
I have (2) 3k ab's (1) 4k and (1) 5k but I can't remember off hand what bin they are. I built this light in attempt to have a multi stage light for flower and veg, but also as an experiment to see how the different spectrum may be of any benefit or not. I plan on buying more ab 3k's later on and will most likely replace the 4k and 5k.
 
You must log in or register to reply here.
Top