CobKits
Well-Known Member
how'd you mount 'em?So I have my prototype water cooled light up and running.
how'd you mount 'em?So I have my prototype water cooled light up and running.
Love it. Can't wait until I can DIY my own liquid cooled LED.
Thanks bassman999 I appreciate it.
They arent bad sinks, but just make sure what wattage they can dissipate, and adding fans on the back will do wonders
It takes 1 BTU to heat 1cc of water 1 degree C°. So find your volume in cc's and then find out how much it rises and it should be close to how many BTU's you pulled out of the light. Not exact because you need a calorimeter that is calibrated for that but should be fairly accurate.So I have my prototype water cooled light up and running. Just tested it today, let it run for a couple of hours to see what was what, and so far it's staying cool as a cucumber, or at least the heat sink.
I put a small fountain pump (185 g/hr) in a 5-gal bucket and circulated it through the fixture then fired her up, and I must add here that you don't want to look directly at these lights for any amount of time, talk about blinded by the light, lol, even just the light hitting my peripheral vision started making me nauseous, which I found weird and uncomfortable. I finally put on my gas welding goggles, and it STILL was too bright to see well.
View attachment 3802649
After 3 hours the COB temp had reached nearly 49C and the water in the bucket was noticeably warmer (still cool, just not as cold as when it came out of the tap). Not sure yet how to extrapolate what that means regarding sizing the system when/if I get the other 7 fixtures built, but I think it's definitely promising, especially since this was strictly passive, as in the water itself wasn't cooled in any manner. With a radiator and fan installed it should do better, but more testing needs to be done to get more useful data.
View attachment 3802650
Btw, I'm officially christening this the Russbilt Sunbar 2000, lol, copyright pending
I have seen pc's run in fish tanks filled with oil and work fineAnd you can substitute mineral oil for water. If it ever springs a leak, nothing will short out. It also has good thermal transfer characteristics. Some of the really serious pc overclockers run it in their cooling system. And it doesn't evaporate as easliy as water, and a higher boiling point.
You need to worry about the heat watts, that is based on total watts and efficiency of cobs.Im trying to figure out if my heatsink can dissipate enough heat from my cobs..
CitiCalc says I have 72 "heat watts" by running 12 citizens @ 1050ma. So here are the calculations from my bar.
View attachment 3802821
Can someone enlighten my how much wattage I need to dissipate? Is it total wattage at the wall or just "heat watts" (as displayed in the Citi LED calculator)
This is the photo I took with rough measurements
Can anyone interpret?
I used that thermal epoxy, took a while to decide how to do it. I thought about just epoxying the COB to the bar, but then thought I might want to change them in the future, so I took the Ideal holder, had to file off the small plastic nubs that protruded a bit on the bottom, and applied epoxy to the area under the plastic terminals.how'd you mount 'em?
NastyN8t posted this earlier on this thread, quite the getup, like COBs meet Frankenstein's monster, lol.anyone ever thought of using some old pc water blocks and maybe a heater core from a car.. especially on those 100V bastards..or... a Peltier module?
Thanks Ichabod, I'll remember that when I do another run. One question, since I only want to account for the heat dissipated from the COBs, should the water temperature be equal to the ambient air temp or close to it?It takes 1 BTU to heat 1cc of water 1 degree C°. So find your volume in cc's and then find out how much it rises and it should be close to how many BTU's you pulled out of the light. Not exact because you need a calorimeter that is calibrated for that but should be fairly accurate.
Did you at least put any thermal paste under the chips themselves? Once I saw $12 Citizen chips preforming as well as $40 Cree chips, I figured that destroying the chip to replace it would be affordable.I used that thermal epoxy, took a while to decide how to do it. I thought about just epoxying the COB to the bar, but then thought I might want to change them in the future, so I took the Ideal holder, had to file off the small plastic nubs that protruded a bit on the bottom, and applied epoxy to the area under the plastic terminals.
Then I used large 8" hose clamps and 3/4" PVC coupling laid across but supported by shims on either side of the COB to apply pressure, ugly and very fiddly to get right, but it did the trick more or less, lol (it was hard to keep them from drifting ;?). If I do decide to change out the COBs for whatever reason, I'll have to use a grinder and destroy the holder, but I should be able to save the leds, but not the best system, I realize. Any suggestions on how I could have done this differently would be greatly appreciated
Since this is 1/4" aluminum I toyed with the idea of drilling some shallow holes and using some 1/4" self-taping screws, but since I plan on going with 1/8" tubing for any future fixtures, I had to use something that would work on those as well. All in all it's not going to win any beauty contests that's for sure
Oh yeah, I forgot to mention that, I did use some paste, something called Thermaltake TG-7, says it has diamond powder in it (rated 3.3 W/m-k).Did you at least put any thermal paste under the chips themselves? Once I saw $12 Citizen chips preforming as well as $40 Cree chips, I figured that destroying the chip to replace it would be affordable.
If you insulate the water lines and containers it should not matter. A calorimeter is nothing but a container that is insulated and has a known amount of heat lose for calculations.Thanks Ichabod, I'll remember that when I do another run. One question, since I only want to account for the heat dissipated from the COBs, should the water temperature be equal to the ambient air temp or close to it?
I run cold water, about 55F. I didn't do it based on calculus but the seat of my pants, lolIf you insulate the water lines and containers it should not matter. A calorimeter is nothing but a container that is insulated and has a known amount of heat lose for calculations.
But if you run the water at room temp you will not have any heat dissipate to the roon but your cooling will be less efficient at the chips. Colder water will take the heat quicker.
I am not a expert in all this. I just know about the calorimeter from when I was a chem lab assistant in college.
If you really want to go down the rabbit hole you can calculate how much heat is lose to the room with some calculus. But I dont remember all that. It was calc III and I missed over a month of class. About 5 weeks out of 13 weeks because I had a microplasma infection on my lungs. Still got a 3 point but that shot my perfect 4 point in math.
Sorry the calculus was for heat lose to the room. Colder water will cool the cobs faster that warmer water. Didnt mean to imply that.I run cold water, about 55F. I didn't do it based on calculus but the seat of my pants, lol
It is very effective nonetheless.
My intent was not malicious in any way, but merely to point out that good science and engineering can still be done even without a complete grasp of the underlying math.Sorry the calculus was for heat lose to the room. Colder water will cool the cobs faster that warmer water. Didnt mean to imply that.
You can calculate the pattern of heat dissipation in the room to figure out how much heat was given off from the light. Was cool when I was doing it in college. But the next semester I crashed my immune system and had to drop out. That is how I started growing. Had to pay the bills and take a year off to recoup. So I started growing.
Hey researching,I am considering these for a build. I would like to go with the smaller profile for cost/weight. I plan to use a 30" length with 10 3w leds run off a 20w 650ma driver. Would that profile be adequate (passively cooled) My second option is the larger profile (passively cooled)
1st choice:
2nd choice: