Heatsinks for DIY LED lamps

[sforza]

@sforza If you used a thick copper plate, it would quickly spread the heat away from the COB, but it needs surface area to dissipate that heat. So the heat has to pass through the thermal interface between your plate and the aluminum heatsink surface to get to the fins. The quality of your thermal interface would affect how much resistance was created. The resistance could cause a slight thermal gradient, which would lead to a slightly higher junction temp in the COB and therefore more temp droop (light loss)

So in practice if your thermal interface was very good, you might reduce your temp droop from 2% to 1.5% or something along those lines (with a hard running CXA3590). But if the thermal interface was poor for any reason, it could increase your temp droop from 2% to 5% for example. If you are running your CXA3590s soft (~25-50W), you are probably getting less than 1% temp droop, maybe even less than .5%. In that case there are no significant gains to be had by investing in more cooling.

The heatsinks we use are normally a nice 1 pound chunk of aluminum/copper. I have tested temp droop of CXA3070s of different heatsinks at high dissipation levels (100W) to see which perform best. So far the winner has been the large heat pipe style OCZ Vanquisher. Second place the blow through style with copper core, like this, although the mounting surface is too small for CXA3590 or Vero29. The Rosewill RCX-Z1 and RCX-Z200 use copper plates mounted to aluminum and the performance is similar to the Alpine11 which is solid aluminum.

All at 2.47A:
OCZ Vanquisher (large heat pipe style cooler) with CXA3070 3K AB = 2.6%
blow through style Intel with CXA3070 3K Z4 = 3% temp droop
Rosewill RCX-Z1 with CXA3070 3K AB = 4.5%
Alpine 11 with CXA3070 3K AB = 3.5-4.5%

Rosewill RCX-Z1 with Vero 29 3K = 2% temp droop

CXA3590 so far untested

Great information, SupraSPL. I don't know how you find the time to help out us less technically savvy newbies and still run your own garden. You are a one man army and doing a great deal to power the transformation in lighting to LEDs. Have you considered setting up a consulting business where you could charge people to answer questions? I know I would be happy to pay for the information you provide because otherwise I would be spending money on the wrong equipment.

I am going to use the Ideal holders for the CXA3590s, so using the plate will give me the surface area I need to mount four of them. I plan to use the thermal paste you recommend, I think I have the P3 version and already used it on one Chinese COB. So there will be paste between the COB and the plate and then paste between the plate and the heatsink/fan combination. Since there is a plate in between the surface of the heatsink/fan and the surface of the COB do not need to match them up in size.

Since the OCZ Vanquisher worked well, I was thinking of mounting four of these on the plate and have fans blowing through the fins. It is a Dell Y1851 heatsink and they are cheap used.

I have an infrared thermometer, so I can check the temperatures of the cobs after I hook everything up and see what temperatures everything is running at. If it runs too hot, I will switch to the finned heat sink. I will post a picture to the thread if and when I get it set up and working and report the temperatures I find.

Thanks for the help and detailed information SupraSPL.

 

[I Hate Usernames]

I finally understand that you're referring to page 1 of this thread lol. You know, stoners... That's where it came from (and the discrepancy). Poor @SupraSPL had to calculate all the perimeters while they were in that pdf the whole time.
Just a small detail - perimeter is in inches, area in inch^2 (since we multiply inch*inch).

Thank you, I've been trying not to ask to many questions as I want to learn rather than just be told but I didn't know if my math was wrong or what was going on.

 

[Dloomis514]

I finally understand that you're referring to page 1 of this thread lol. You know, stoners... That's where it came from (and the discrepancy). Poor @SupraSPL had to calculate all the perimeters while they were in that pdf the whole time.
Just a small detail - perimeter is in inches, area in inch^2 (since we multiply inch*inch).

The perimeter is calculated first, then the length is multiplied to get area, his figures are per inch of heat sink, so the area is the same as the perimeter, if you consider 36.4 in and 36.4 in^2 to be equal.

 

[SupraSPL]

I finally understand that you're referring to page 1 of this thread lol. You know, stoners... That's where it came from (and the discrepancy). Poor @SupraSPL had to calculate all the perimeters while they were in that pdf the whole time.
Just a small detail - perimeter is in inches, area in inch^2 (since we multiply inch*inch).

My perimeter results disagree with Heatsink USAs PDF. I checked my work and it seems to be OK, not a stoner mistake. I admit I am very prone to careless mistakes which is why I need spreadsheets.

@Dloomis is correct, when I post them, the spreadsheets are set at 1 unit of length, so the perimeter almost matches the surface area. BUT, I realized later on that that the surface area of the cut ends was not being accounted for, so I added that to the equation and you can see that in my post above. The surface area of the cut ends can be significant in the case of a wide profile being cut to a short length, as in @churchhaze setup. That additional surface area would become less and less significant as you extend the length of the cut.

 

[alesh]

My perimeter results disagree with Heatsink USAs PDF. I checked my work and it seems to be OK, not a stoner mistake. I admit I am very prone to careless mistakes which is why I need spreadsheets.

@Dloomis is correct, when I post them, the spreadsheets are set at 1 unit of length, so the perimeter almost matches the surface area. BUT, I realized later on that that the surface area of the cut ends was not being accounted for, so I added that to the equation and you can see that in my post above. The surface area of the cut ends can be significant in the case of a wide profile being cut to a short length, as in @churchhaze setup. That additional surface area would become less and less significant as you extend the length of the cut.

I'm sure that your math is correct. I'm calculating the perimeter the same way. But since the values it that PDF seem to be very precise, I'd think that they might be actually valid. Not sure where the differences came from.

 

[SupraSPL]

Thnx @sforza RIU is a great place for this. It seems like anything we share we ending learning even more from others. It will be interesting to see what you think of those heatsinks.

I have some Dynatron P66 heatsinks that remind me of that. Small fan, loud but nice hunk of copper

 

[Positivity]

Dynatron coolers are nice but the fans are too small...

I have two that i'll end up only using for lower currents with the fan on low

 

[taproot]

Guys, I'm trying to build my first light and need some help with the heat sinks. My original post in AP's thread led me here:

https://www.rollitup.org/t/all-things-vero.851330/page-43#post-11371387

I'm going to run 6 vero 18's ( in series ) @ 500 or 700 mah with the "HLG-185H-C1050". I'm thinking 700 is nice but the driver can go up to 1050mA should I ever upgrade lamps. I'd additionally like to plan my heat sinks out to give me a little room to upgrade should I ever need to. I don't think I'll ever need more than 200W and my target is to flower enough for my medical needs with 150W. Also, I figured that splitting up the sinks and putting 3 lamps per strip would even the light distribution out more and hopefully allow me to passively cool for the most part with the air that is pulled across the sinks being via the exhaust fan. So, around 75/100 W per strip. I found the following sink which might work, I was going to do 10 or 12" strips. The flow in the 2x2x4 cab should be around 125/150 cfm with the fan and filter. Can you guys help me to get the heat issues dialed in as it's usually pretty hot in the summer and the inside temp is normally 78 - 80F to keep AC bills in check.

http://www.heatsinkusa.com/4-600-serrated-fin/

I'm planning on using the EZ harness with the vero's with 20 AWG wire and these wagos to wire it all together.

http://www.amazon.com/WAGO-224-201-LIGHTING-SERVICE-CONNECTOR/dp/B00LQRXCLA/ref=pd_sim_sbs_indust_1?ie=UTF8&refRID=1CPC745KHP6HTRBYSE05

Thanks for all the hard work you guys put into this , and for this topic and the drivers ( Supra ), as you're helping a lot of people who need to get started to understand and get things correct.

Thank you!!

 

[SupraSPL]

I suspect the 4.6" serrated is not the most passive cooling friendly profile. Since you are considering 10-12" lengths you could go @churchhaze style and use a short length of the 10" profile which is especially passive cooling friendly. For longer lengths I like the 3.5" 4.23" 4.9" 5.88".

 

[taproot]

Thanks Supra, I appreciate your input and help. I saw his build, which is freaking sweet, and thought that would probably work for me. I've been looking for a metal plate of sorts, which is what it looks like to me, but could not find anything locally. I'll try to find out what 10" profile is exactly as I'm not familiar with the term yet. Hopefully I can find it and get it cut locally or it not shipped for not too much.

Thanks again!

 

[SupraSPL]

http://www.heatsinkusa.com/10-000/

The idea is that the fin channels are shorter air can convect more easily, but you still get some spread between the COBs.

 

[taproot]

Hmm, ok...so if I understand you correctly you saying the serrated one would most like create too much turbulence or disruption in air flow at least for a passive setup? I really like Curchhaze setup and am thinking of going with one large one instead of two to keep it more simplistic but am debating the merits of both. Thank you!!

 

[SupraSPL]

Ya the 4.6" profile has tall fins and relatively thin base as well. @getsoutalive do you know if the C/W/3" data they give is based on passive cooling in the vertical orientation? Apparently the C/W/3" figures are based on a temp rise of 75C. Sine we are dealing with a temp rise of less than 10C, the thermal resistance numbers would be a lot higher.

http://www.aavid.com/thermal-tools/temperature-correct

Good info here, examining fin spacing
http://www.thermalsoftware.com/vert_vs_horz_sink.pdf

 

[getsoutalive]

Sorry, no. Unfortunately, HUSA offers nothing more than the rating itself and no details about how they got there.

Lots of good info in those links. thx

Remember that while its nice to pretend that we have solid data with extreme precision to two decimal points, thermal management is a very complex science. At best we are ballparking the true numbers. Factors like those in your 2nd link, fin orientation, air flow through and around the fins, temp gradient with ambient and of course, adding any heat load to a room will raise the ambient we are starting from. We can provide decent estimates, but in application, things are a bit different. It is always best to buy more than you think you need rather than less.

But when it comes to relative performance, in similar environments, we can be more confident of the math between the different models.

 

[Doer]

...Nerds...

Geeks are nerds that get things done.

 

[Doer]

Sorry, no. Unfortunately, HUSA offers nothing more than the rating itself and no details about how they got there.

Lots of good info in those links. thx

Remember that while its nice to pretend that we have solid data with extreme precision to two decimal points, thermal management is a very complex science. At best we are ballparking the true numbers. Factors like those in your 2nd link, fin orientation, air flow through and around the fins, temp gradient with ambient and of course, adding any heat load to a room will raise the ambient we are starting from. We can provide decent estimates, but in application, things are a bit different. It is always best to buy more than you think you need rather than less.

But when it comes to relative performance, in similar environments, we can be more confident of the math between the different models.

This is why I am planning a cool tube set up so I can maintain ambient in the room with A/C and exhaust the LED heat.

 

[Scotch089]

Hey I thought I saw a step by step process of calculating the amount of sink we need for given watts of light... cant seem to find it and other sites suck. Supe would you mind reposting it? Im looking to (probably actively) cool a max of 106w and as low as 50. Going to run 3 cobs with 16 xpls (probably, this is all likely to change) thought the stars would be a nice addition to it if I put reflectors on em for "multi layer" coverage, but maybe unnecessary. dono, still rolling it over.

I feel this insane urge to fill empty space on a heatsink...

driver would be the hlg-120h-c700a

Thanks dude


Edit: see, told you all this was pending- probably ditching the xpls, , no reason with how narrow of a space it is and wide 115° angle of the 3070s. So I'll probably be shopping a hlg-80h-c700a. Max draw would just be about 74w. That leads me to wonder how much sink would I need to run that passively but stay super duper cool. Fans aren't a huge deal, the ones I have are dead silent anyways, jw

 

[Doer]

Think about it, the computers can't take over if we control the heat sinks.

 

[SupraSPL]

From the testing with COBs and modern LEDs, I lowered my recommendation to 25cm²/W for active cooling and 75cm²/W for passive cooling, assuming you are using a passive cooling friendly heatsink. These are just general guideline, you could allow the COBs to run a lot hotter and they would still perform well, especially Veros. With these recommendations I am aiming for very low temp droop, 2% and under for those running medium-soft.

I would like to incorporate the C/W/3" data by translating that data to cooling/1000cm² chunks. But as @getsoutalive pointed out, it is very complex in practice. So our best bet might just be to get a few chunks of different profile sizes and test them with our actual conditions (horizontal orientation, slight air movement and very small temp differential between heatsink and ambient).

Here are the figures for C/W/1000cm² chunks. The smallest numbers should be the most passive cooling friendly, BUT again these numbers are based on no air movement, vertical orientation and super hot heatsinks, much different conditions than we will be dealing with. Really, this calculation is not valid for the profiles with large changes in length and the assumptions made about how that affect heat flow, but I went ahead and did it anyway.

1.8" -> 1.4 C/W/1000cm²
2.08" -> 1.8 C/W/1000cm² (great for small COB passive vegging due to cost and spread)
3.5" -> 1 C/W/1000cm² (my fav for active flowering due to spread and large COB passive vegging)
3.945" -> 1.6 C/W/1000cm²
4.23" -> 1.1 C/W/1000cm²
4.6" -> 1.46 C/W/1000cm²
4.85" -> 1.37 C/W/1000cm²
4.9" -> 1.2 C/W/1000cm² (my fav for passive flowering due to spread)
5.375" -> 1.22 C/W/1000cm²
5.88" -> 1.4 C/W/1000cm²
7.28" -> 1.22 C/W/1000cm²
8.46" -> 1.345 C/W/1000cm²
10" -> 1.3 C/W/1000cm²
10.08" -> 2.22 C/W/1000cm²
12" -> 1.46 C/W/1000cm²

 

[Doer]

Ya the 4.6" profile has tall fins and relatively thin base as well. @getsoutalive do you know if the C/W/3" data they give is based on passive cooling in the vertical orientation? Apparently the C/W/3" figures are based on a temp rise of 75C. Sine we are dealing with a temp rise of less than 10C, the thermal resistance numbers would be a lot higher.

http://www.aavid.com/thermal-tools/temperature-correct

Good info here, examining fin spacing
http://www.thermalsoftware.com/vert_vs_horz_sink.pdf

Uuuuu....good point. My swag so far has been assuming horizontal, in still air. I've never seen anything to suggest they testing them vertically.

Good point #2.. What sort of correction factor to the C/W/3, can you guess might be appropriate with only 10C gradient?

Good point #WOW. 25 sq cm / watt Got it. That will really help my calculations, given the C/W/3 is higher.