Stephenj37826
Well-Known Member
Do you think alpine 64 would give similar performane to artic 11? I ordered 16 cxb3590 3500k cd bin and want to use the holders.
Heatsinks for DIY LED lamps
- Thread starter SupraSPL
- Start date
EfficientWatt
Well-Known Member
Hi all,
I was looking at some HS options, and been messing around with numbers, as often ... and thinking :
Efficiency is climbing on CXBs, and the 75cm²/W rule of thumb gives us :
1) 3000K CXB 3070 AB @49.83W : 3737cm² (and gives off 27.19W of heat)
2) 3500K CXB 3590 CD @48.85W : 3664cm² (but gives off only 21.33W of heat !)
Using efficiencies, to work out actual heat dissipation % of 3590 3.5K to 3070 3k :
100*( (1-45.43) / (1-56.34) ) = 80.29% heat per watt of dissipation
So wouldn't 75cm²*0.8029 = 60cm²/W for case 2 be just as effective as 75cm²/W for case 1 ?
(In fact slightly better because the 3590 has a bigger surface² to heatsink than 3070's ?)
@SupraSPL Does this sound right to you ? or is there something I'm missing that makes this calculation non linear ??
Tell me what you guys think about this. How low can we reasonably go with 56%+ efficiency ?
... maybe someone's experience with extra high efficiency veg builds can give a definitive answer
I was looking at some HS options, and been messing around with numbers, as often ... and thinking :
Efficiency is climbing on CXBs, and the 75cm²/W rule of thumb gives us :
1) 3000K CXB 3070 AB @49.83W : 3737cm² (and gives off 27.19W of heat)
2) 3500K CXB 3590 CD @48.85W : 3664cm² (but gives off only 21.33W of heat !)
Using efficiencies, to work out actual heat dissipation % of 3590 3.5K to 3070 3k :
100*( (1-45.43) / (1-56.34) ) = 80.29% heat per watt of dissipation
So wouldn't 75cm²*0.8029 = 60cm²/W for case 2 be just as effective as 75cm²/W for case 1 ?
(In fact slightly better because the 3590 has a bigger surface² to heatsink than 3070's ?)
@SupraSPL Does this sound right to you ? or is there something I'm missing that makes this calculation non linear ??
Tell me what you guys think about this. How low can we reasonably go with 56%+ efficiency ?
... maybe someone's experience with extra high efficiency veg builds can give a definitive answer
Last edited:
HydroDC
Well-Known Member
All of the CPU coolers will work from a heat dissipation perspective. Making sure it fits the holder is the more important issue and, as you have found out, these will accommodate the Ideal holders for the CXB3590s. Supra and others have pointed out many times that they're running the fans at less than 12V (even in some cases to 5V) showing just how much excess capacity these coolers have for our use. Good luck with your project.
SupraSPL
Well-Known Member
@EfficientWatt you are correct, since efficiency has risen so much recently the newest rule of thumb is already outdated. For very high efficiency builds you could use a figure like 60cm²/W. For those who want a bit more accuracy we can look at surface area/ W of heat.
For heatsinks that are passive cooling friendly in pasive setups, 120cm²/W of heat
For heatsinks that are active cooling friendly in active setups, 40cm²/W of heat
These are only guidelines based on KNNAs original recommendations, my limited experiments and what I think are reasonable temp droop goals. If you have a lot of direct airflow getting into the heatsink fins of passive cooled setups, you could use less. If you have a very high fan speed maybe you could use less for the active cooled setups. If you don't care as much about temp droop and really want to minimize the heatsink weight/cost for some reason, you could use less and accept a higher Tj and temp droop with no undue damage to the COB.
For heatsinks that are passive cooling friendly in pasive setups, 120cm²/W of heat
For heatsinks that are active cooling friendly in active setups, 40cm²/W of heat
These are only guidelines based on KNNAs original recommendations, my limited experiments and what I think are reasonable temp droop goals. If you have a lot of direct airflow getting into the heatsink fins of passive cooled setups, you could use less. If you have a very high fan speed maybe you could use less for the active cooled setups. If you don't care as much about temp droop and really want to minimize the heatsink weight/cost for some reason, you could use less and accept a higher Tj and temp droop with no undue damage to the COB.
Last edited:
Stephenj37826
Well-Known Member
So for passive cooling what would you recommend for 16 36v cxb3590s 35k cd bin driven at 700ma in bars of 4 from heatsink usa
SupraSPL
Well-Known Member
23W * 16 COBs = 368 dissipation W @ 64% efficiency = 132.5W of heat
132.5W * 120 = 16000 cm² total or 4000 per bar.
So for passive cooling you could use:
3.5" X 32"
4.23" X 24"
4.9" X 18"
That would put 92W on each bar so to get a good spread you might want the longer option. For that same setup I use 4.9" X 14" bars with 2 COBs on each, overkill as far as surface area but I wanted about 10" between COBs. I could have used the 3.5" but designed it before these COBs came out.
132.5W * 120 = 16000 cm² total or 4000 per bar.
So for passive cooling you could use:
3.5" X 32"
4.23" X 24"
4.9" X 18"
That would put 92W on each bar so to get a good spread you might want the longer option. For that same setup I use 4.9" X 14" bars with 2 COBs on each, overkill as far as surface area but I wanted about 10" between COBs. I could have used the 3.5" but designed it before these COBs came out.
Stephenj37826
Well-Known Member
So I'm actually building 2 lights. Would 23" 7.280 cool 3 cxb3590 36v at 1.4. I think it would but want to be sure.
SupraSPL
Well-Known Member
196W @ 56.34 = 110 PAR W and 86 heat W. So that is 10320 cm².
23" of 7.28" is 8000cm² so you would need 7.28" X 30" length. You could use 10" X 22" for shorter fin channels but the 30" length should give you better spread. If you use lenses and have a lot of headroom, you could try to get an even spread that way but if not you could end up with a very bright hot spot and dim around the edges.
23" of 7.28" is 8000cm² so you would need 7.28" X 30" length. You could use 10" X 22" for shorter fin channels but the 30" length should give you better spread. If you use lenses and have a lot of headroom, you could try to get an even spread that way but if not you could end up with a very bright hot spot and dim around the edges.
Everyone give props to these heady folks like supra, alesh, and many others who answer without fail.
Stephenj37826
Well-Known Member
There would be 3 on 1 bar and the fourth would be on an 8" bar by itself. Yes I'm using lenses as well. 23x23
X X X
X ....X
X X X
80° lenses
Thanks for all your help. You are a kind person for helping everyone so much.
X X X
X ....X
X X X
80° lenses
Thanks for all your help. You are a kind person for helping everyone so much.
Michoacanna
New Member
Good info as always...I'm a newb to all this and still learning...active cooling a heat sink means blowing AC unit air onto it rite? And I read u guys saying there are heatsinks for active cooling which weighs less ? Sorry for the newb questions
Stephenj37826
Well-Known Member
Active means a fan blowing directly on the heat sink. Passive means no fan needed. Passive works through convection.
Michoacanna
New Member
Ok I see, thanx for the reply
bicit
Well-Known Member
Yup, but it's really easy to over do the fans. Fan power doesn't grow plants. The fans can only reduce the heatsink to ambient temp at best. Using 12w to do the same thing as 3w is in efficient. So keep it minimal.
just to confuse the issue .... a small amount of fan power goes a long way to reduce the amount of heatsink required as opposed to a passive setup.
I used to be a "passive" cool nutcase, until I saw the results from providing low powered active cooling instead of passive. highly efficient computer fans run at relatively low power are definitely worth it.
bicit
Well-Known Member
That's more or less the point I was trying to make regarding fans. Less ir more in other words. Small 80mm fans ran at 5 volts is about the most ideal. I haven't found a compelling reason to use bigger or more powerful fans. Too much fan power will decrease the overall efficiency of the system and reduces reliability
Im an advocate for passive now days. No extra wires or drivers, no moving parts, true solid state lighting. Its a wonderful thing
Stephenj37826
Well-Known Member
Maybe a dumb question but how much cooling capacity does say .30 aluminum plate have? Just wondering. My heat sinks will be mounted to a 24x24 piece of aluminum. Was just wondering if it will help further dissipate some heat. Also there will be a constant airflow around my "passive" setup. Hoping for really good tj numbers as well as simplicity.
Last edited:
I get the simplicity!
btw with the right setup, a larger fan (120 mm, 140 mm, 200mm ) can move more air more efficiently than a bunch of wee little buzzers. They are also a lot quieter.
