Water cooled led system

[kenneth wolfe]

Hello fellow reefers.. I am in the process of a new build I like to try and be as natural as possible ..My thoughts on a led multichip high wattage cooling system .I thought with the cost of having to light and heat the system can be quit high and also ugly with the large heat sinks and noisy fans led me to re think the system ...hmm..Why not use a stainless steal heat exchanger and a high temp. refuge aka dsb ect. ect. My thought is with the remote high temp. you use that to stabilize the large dst and have a larger biodiversity with the remote high temp. growing things that don't grow at 78 degree F. So in theory the remote heater /"what ever" is "set at the "working temp" lets say 100 degree f." of the fixture and its full of goodies and when the display needs heat you pump from it.. This is an idea that I am running threw my head ...I like to put multiple functions into things I take the time to build..I see a slim fixture being water cooled ....total water volume =1000g ,500 gdt With all that being said any thoughts on my idea good bad what ever..I have good fabrication skills with access to good tools..he he he ..

[gbru316]

1. If you're planning on running saltwater through the heat exchanger, I'd think twice. Stainless will corrode in saltwater. Titanium would be your best bet, but it's not cheap.

2. Using your light fixture to heat water is going to be problematic at night. You're still going to need heaters.

3. LED fixtures don't need active cooling if heat sink has enough surface area. This does NOT mean they need to be large. I have 36 emitters on a passively cooled 8x12x1 heat sink and it doesn't get any hotter than 90 degrees. LEDs are also more efficient when running them at less than 100%.

4. You'll most likely have difficulty changing the temperature of that volume of water with the method you describe. See #3

5. What happens when the tank doesn't need heat but the LED's need cooled? Periodic active cooling of the LED fixture is going to be difficult to implement if it's directly tied to the aquarium temperature. You'd need a heat exchanger that can cool the fixture without water being pumped through it, or you'd need a valve to switch the cooling water from a separate, fan cooled heat sink (when the tank doesn't need heat) to a heat exchanger for tank water.


From an engineering standpoint, I find that following the KISS principle is often one of the best strategies. With the method you describe, you have multiple critical system components relying on 1 subcomponent. If the heat exchanger pump dies, you lose heat and light. If the lights fail, you lose heat. The setup you describe has a reliability issue that could be easily mitigated by keeping the heating and lighting separate. Although it sounds like a cool idea, my professional experience says don't do it.

[asudavew]

+ We usually try to keep the heat from our lights out our tanks. Not in.

[mcgyvr]

yeah... bad idea..

[marspeed]

I think it is a good idea

[karimwassef]

Liquid cooling works, but salt water is a poor choice. Use a fresh water loop with an evaporator.

[gbru316]

Liquid cooling works

Yes. However, you might still need a radiator (with fan) to cool the heated water down before it travels through the water block that the emitters are fastened to. Depends on cooling system volume. As the water heats up, the cooling capacity decreases.

Although, with the heat levels produced by LEDs (minimal, comparatively), I'm not sure something like this requires this much work. A plain aluminum heat sink (note, an actual finned heat sink) with some thermal grease accomplishes the job with much less effort, and you don't have to worry about leaks or pump malfunctions.

[Indymann99]

+ We usually try to keep the heat from our lights out our tanks. Not in.

The OP lives in Colorado... I bet 5mo out of the year he would benefit from heat from the lights going INTO the tank... Prob not much of an issue for you in TX...

That said I would think the complexity and potential for leaks would outweigh the benefits of liquid cooling LEDs

[sfsuphysics]

if your fans are noisy you are using the wrong fans and/or running them way too fast.

[kenneth wolfe]

a couple of things to consider..I am in the protective coatings business ( High grade stainless and a good coating should not fail) , when I say high wattage I meen say 1500 watts or more . the remote fuge will be large over 150 gallon ..its not a cure for not needing a heater but with that amount of reserve it should hold over night ...I also feel the high temp in the fuge is a great bonus to the system..how many fans and heatsinks do you imagine for 1500 w. that's a large noisy fixture..I see a fixture no more than an inch thick total..stainless steel exchanger inside a aluminum housing ..my water cooled welding rig is 250amp and cools with 1/8" inch feed and return lines.

[gbru316]

Sounds like your mind is made up. Have fun, take pics.

[kenneth wolfe]

no no ..not at all ..just trying to find some alterative thought..the titanium would be nice but not in my budget..if we are talking about 1500 watts that's not a lot of heat to process for a system that large . and the rooms are air conditioned .so its possible that just running in to the system would be fine..(the system has a chiller)I like the closed loop idea also ..my remote cooler for my welder would do the job for less money than a large heat sink and fans..not its hard to find large heatsinks cheaply with out welding them together or some extra fab work.vrs welding a heat exchanger.I have been threw most of the led build on here and haven't seen many water cooled threads ..its also a lot to do with the ability to have a hot bed for bio culture that's not present in a " cooler environment " the reword might be worth more than the risk ...

[karimwassef]

I run 200W on a 4" wide Aluminum finned strip (8ft long). I use small quiet 3" fans.

I can probably double the wattage and still have margin.

With 4x the Aluminum surface area, you can do 1600W.

<a href="http://s1062.photobucket.com/user/karimwassef/media/4AE21775-3C3E-4F8E-89EA-72051279323F_zpsyl6vd2nu.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/4AE21775-3C3E-4F8E-89EA-72051279323F_zpsyl6vd2nu.jpg" border="0" alt=" photo 4AE21775-3C3E-4F8E-89EA-72051279323F_zpsyl6vd2nu.jpg"/></a>

[karimwassef]

<a href="http://s1062.photobucket.com/user/karimwassef/media/AA01D0DB-9EE2-4117-8803-1766CFCA9D94_zpsja6sjenk.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/AA01D0DB-9EE2-4117-8803-1766CFCA9D94_zpsja6sjenk.jpg" border="0" alt=" photo AA01D0DB-9EE2-4117-8803-1766CFCA9D94_zpsja6sjenk.jpg"/></a>

[kenneth wolfe]

What temp. does that fixture run at if you don't mind Im trying to get an area to wattage formula ..yours is 384 sq. inch/200 1.92 per watt ..sound wright ??

[kenneth wolfe]

I follow your thread on the rock wall...wish you live closer..not that many close to me here..

[karimwassef]

Thanks. 200W/384in2 ~ 0.52W/in2 but I think you can double it.

I don't think it runs higher than 40C. I can touch it and it's warm but not hot.

[kenneth wolfe]

is your math backwards or mine..lol... 0.52 is small do you think??

[karimwassef]

I measure W/in2. You're measuring in2/W. We're saying the same thing.

[kenneth wolfe]

I see that now after I read it again..lol..My opening is 35" x15" for lighting a 48" x 72" round tank. My thought was with the roughly 475 in2 ..Im going to have to fill it to the max.. lets say 1500w/475 in2 is.. lets say 3 w in2 ...and possibly more maybe 2500 watts for 5/w gallon ..the depth of the tank is my concern ..I know I can put optics on the drivers ..if I need more penetration but I don't like the "cone effect" and with the fixture being so close to the water I need the spread..

[kenneth wolfe]

<a href="http://s44.photobucket.com/user/wolfesperformancecoatings/media/1409644493_zps8db62418.jpg.html" target="_blank"><img src="http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/1409644493_zps8db62418.jpg" border="0" alt=" photo 1409644493_zps8db62418.jpg"/></a>

[karimwassef]

How is it mounted (vertical?)? What are you trying to optimize? Why not use MHs instead?

[kenneth wolfe]

its on its side ..it had a t5 fixture on it before..the water level is about 1.5 inches from bottom of the opening..I plan on lowering the stand for more over the top viewing "with out a step ladder..lol..I am open for all options and thoghts my mind is not made up at all but I am not afraid of better ideas that's why I joined this forum I have been just reeding for years and never posted until now..I have other diy led experience on smaller tanks but this is going to be a show piece in my office. Also all my years of experience and craftsmanship , the light fixture has to be amazing ..and the envy of all led light fixtures..lol..

[kenneth wolfe]

its rolled over on its side in the picture so I could climb in and out to clean it..

[kenneth wolfe]

one of the reasons not halide is the lack of coolness ..you know dusk to dawn with the pop.and all the fine adjustments threw my reef angel.you just cant get it with mh in tiny fixture.."the wow factor"

[karimwassef]

Sure, but why LED?

A single MH bulb could deliver 1000W.

What are you trying to do better? More efficiency? Lower power use? Highest light density?

[kenneth wolfe]

I see a fixture that sits in the hole maybe 1 inch from side flush with the surface ..with no optics I should get the spread but not the penetration that's why I feel I need to have so many high power drivers and by using the larger multichip sets the uv and specialty drivers also it should be like making a large multichip out of multichips..there are so many to choose from.. less components into the fixture, with the higher wattage but also less space to work with..

[kenneth wolfe]

I don't want a large fixture over the tank with loads of blow out..Iwant to keep sps so I say density is what I am after..good point sir..

[kenneth wolfe]

Ive had fixtures when there that close to the top ..with lets say 3w have localized collar blead and blow out close to the top of the tank..even with out optics its still there the multichips seem to do better at blending the collar more evenly when at closer distances .That's one thing I can improve upon with a densely pacted fixture...

[kenneth wolfe]

1000w of mh does not match the penetration of 1000w of closely packed leds..no way even without optics ..the heat from that 1000w of tightly packed led array is what concerns me with my idea of small fixture. I believe there is a way ..

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410212808451_zps74fd650e.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410212589091_zps6c54e0cf.jpg

[kenneth wolfe]

here is my first draft "idea" this is total of 1700w there should be enough representation to tune to the desired color .any thoughts??I have seen using the 10w uvs that if you want to get any pop from them they must be High wattage to be seen threw with whites..in my 75g I have 40w and would like 3x more while the whites are on..with a total of 365w 40 of being 405nm 365 at 100 %blue 70% white d-120s and ct-light 85w 40w uv supplement ..I like the control ability of the multiple channels.

[Rognin]

I've built all my puters with liquid cooling loops and building a liquid cooled LED light has always appealed to my more computer orientated side. I've even thought of trying the AIO liquid coolers (eg: Corsair H80) on an Ecotech radion G3 Pro.

Couple things need to be know first.

DO NOT MIX METALS. Try and stick with copper for your cooling blocks and radiators. Mixing aluminum and copper or steel will result in catastrophic failure within two years due to dissimilar metals within an electrically conductive loop (think anode/cathod with alu being the anode). You'll also clog the blocks and even maybe the tubes with goop from the melting metal that was the weakest in the loop.


USE PURE WATER WITH A BIOCIDE. Yeah, well you should already have pure water with RODI. All you need is some type of copper sulfate to put in the loop. Look up Mayhems liquids, best guys in the world for that stuff. Or a strip of 99.99% pure silver or copper. Make sure you get the less chances possible for the loop to leak in the tank or you'll have nightmares for years.

LAW OF DIMINISHING RETURNS. At one point adding fans and rads just becomes redundant. Rule of thumb in liquid cooling is 80w TDP of power per 120X120mm radiator. Do the math and you should be pleased with the results. As a more precise measurement, 164w of power at 1gph flow will raise the water 1c after the block (1/2" ID tubing using a D5 or MCP355). It keeps rising if you have no way of exporting the heat (rads). Plan accordingly.

24H TESTING REQUIRED. Always run your loop with only the pump to see that there are no leaks. You're gonna run some expensive stuff and if it gets wet you may get mad... hehe.

Don't buy cheap, you're talking liquid and electronics (sometimes expensive). D5 pumps would work very well for what you're gonna do.

That said I saw nothing but your diode plans for the loop. You have anyother drawings of what you wanted to do with the cooling loop?

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140909_180213_zpsaa8aa543.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410285658102_zps837e914b.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410310720597_zps5e8945f3.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140909_193946_zps0e0c0e69.jpg

[kenneth wolfe]

the first picture is a aluminum receiver I made for my race car (its getting new cooling system as well) that will work good *** a supply holder with level, the second is a small radiator ($40.00 on ebay free ship, snow mobile radiator) that I modified (its new) it measures 15" x 8.5" w/fan ..the third is the material that I pick up today ($0.0 dollars) and the last pic. is me holding the aluminum plate (33" x 14") ..After considering all methods I believe the closed loop is the best idea to try first..I have a new Jegs brand constant duty fuel pump 110. gph at 10 psi. I will start welding up the exchanger after I figure out my layout in my head the paper .. the 1/5 " x 1" flat bar will be used as the edge seal most likely weld it from inside then machine a o-ring grove into it for a removable top also less overall welding I also considering just making a flat bar frame the screwing both sides to the frame with o-ring and some rtv. therres the wires from the diodes will have to pass threw some where so its time for the old thinking cap...any suggestions ??

[kenneth wolfe]

my post got jumbled up for some reaso and lost some info...

[gbru316]

.any suggestions ??

Make sure to drive any nails with a 15 lb sledge instead of a hammer. ;)

[kenneth wolfe]

lice thing about large projects as there is plenty to do..I am taking a break from the buffer until the cold front comes threw its hot in that tank even w/the ac blowing into it..

[kenneth wolfe]

????..don't follow..

[gbru316]

????..don't follow..
Are you familiar with Tim "the tool man" Taylor?

[Gorgok]

I would weld the whole deal, not much point being able to open it up especially on a flat or large finned hot plate... If it had dirty water running through it, maybe, but you should be using basically engine coolant. You really don't want it coming out over the tank.

You also want to force the water to spend some time in the plate, so you really want to channel it. I would make an S type shape, input from one corner, loop back and forth then back out. Along each wall you can add pass-throughs, where appropriate, by substituting part of the wall with a tube, then drilling out the plate.

In the end i would expect something like this:
|------------| V-In
| ----O-----|
|-O------O- |
| ----O-----|
|------------| ^-Out

Size wise i would make the plate fairly thin, channels wide enough that its comfortable to weld. Welding the internal fins to the cover plate will be more 'fun' than welding them to the bottom (where you can weld on the inside). But i expect you know what must be done to weld the fins to the cover.

[gbru316]

I would omit the passthroughs. You want to keep the water moving and creating alternate paths could allow stagnation.

[kenneth wolfe]

I understand your point about tim the tool man ...ruf ruf ruf..lol..I have to admit I am a lot like that..you should see my hot rods he only wishes..lol..welding it up solid was my fist choice(trying to hide some of the welds ..welding it from the out side is like a ton of welding..jury still out on that..yes that's what I thought about the pass threw ... inside I will use some 1/2 x1/2 flat bar and weld the snake onto the plate..those can be spot welded...that amount of welding on the outside concerns me about warping ..

[kenneth wolfe]

the pass threw is for the wires ...correct??

[Gorgok]

Obviously, just using the tubes to make a double wall portion to the divider walls.

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410317163182_zpsfb55b695.jpg

[kenneth wolfe]

something like this???

[kenneth wolfe]

maybe wire boxes on both sides??

[kenneth wolfe]

after a quick measurement #6 will not fit between so possibly weld tube in ..or some thing else..

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410317347580_zps6a857b1c.jpg

[kenneth wolfe]

this is an canon box I made for a 30 x 30 with auto motive style lights from china the ones that are missing were the uvs I made out of 10mm chips . this has been dismantled for parts.. very, very , very, bright the stainless grid allowed for mouting and flow past the diodes ..this slides into a box with a fan only for cooling ..the diodes are mounted into a aluminum threaded housing with nuts..very good piece for the money 7 watts each.with built in optics ..around 300 watts at 100% but ran at 65% from 30" from water twice my d-120s

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410320021110_zpsed0e663e.jpg

[kenneth wolfe]

my pod farm.

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410336348711_zpse5609447.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410384956534_zpse940a8ed.jpg

[kenneth wolfe]

in the beginning there was just raw materials ....

[Legalreefing]

Not worth it

[karimwassef]

It is if the pleasure is in the making, not just the getting. :)

[kenneth wolfe]

I do love the pride i get from a nice piece , when I get threw with a project.To some it seems like a lot of work but its not that many hours into it, I am keeping track so.. we, will see .. here is a picture of the basic layout I need some food ,so http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410479108976_zpsd630a254.jpggoing to wait until later tonight to weld when its cool outside

[kenneth wolfe]

on my grill its exactly 6 min a side...just saying!!!http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410480014599_zps17fa0b10.jpg

[karimwassef]

LED backplane grilling?

[4Hummer]

Love the Idea. Used to build Liquid Cooled Computers back when I owned a Retail Computer Store


Just a thought Mineral Oil has a Better Cooling capacity than Water, and is Non conductive. (You can immerse electronics in it)

[am.gee]

in the beginning there was just raw materials ....

... and a dream :bum:

thats going to be a very interesting build, nice job so far :D
i am tagging along.

[am.gee]

I do love the pride i get from a nice piece , when I get threw with a project.To some it seems like a lot of work but its not that many hours into it....

so true, its so fantastic when finally everything comes together and works as expected... i was sitting hours in front of my own ( and very first) LED project during the test phase, watching the clouds animation going and going :)

[kenneth wolfe]

I checked out your led build its great ...any thoughts on mine ?? any regrets on yours ...I love the real time on you fixture from the Cayman islands I want to do the same with mine ..when I get there I will need a coach for sure..how many diodes do you think you would have needed with out optics?? and spread ??

[am.gee]

I checked out your led build its great ...any thoughts on mine ?? any regrets on yours ...I love the real time on you fixture from the Cayman islands I want to do the same with mine ..when I get there I will need a coach for sure..how many diodes do you think you would have needed with out optics?? and spread ??
As you want spread AND penetration i would suggest to use both, multichips with optics ( 90 degrees ) and chips withou optics for spread. Because i was missing spread on my fixture i added those 10 watt chips without optics on the sides. That was really a good thing, now i am very lucky. If i build such a fixture again, i would mix those chips all over the board so there is a chip without optics sitting next to a chip with optics. I think a 50/50 combination of the chips would be cool. I am still thinking about the addition of adressable LED RGB strips to create even more spread and to add some warmer colors. For special LEDs ( uv for example) i would always choose to use no optics.

[marspeed]

Your going to want to regulate the flow thru that thing and find a happy medium

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410737785945_zpsa14a8111.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/1410737848537_zpsfac659b0.jpg

[kenneth wolfe]

ok..I got some time to work on the led fixture..the first pic is the back after it was all welded it still needs the flow diverters fit and welded, but material not here yet..the second is the diode side it was warped from welding so I sanded it flat again so it can be put back into welding gig (good thing the .25 plate)

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140923_152308_zps579c305a.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140923_154721_zpscb2055b5.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140923_182521_zpscb8c2b70.jpg

[kenneth wolfe]

http://i44.photobucket.com/albums/f24/wolfesperformancecoatings/Mobile%20Uploads/20140923_185326_zps9a3bf827.jpg

[kenneth wolfe]

well all the hard work is done!!! yea haw.. time for a beer.. or 6...lol my abs is the wrong size but you can see my idea. to cover the wires..Its time to leak test and hook up the rest of the parts "radiator, pump, ect.." any ideas how to put "x amount of heat" into the plate ..I was thinking like a X watt plate...or something ..still thinking on how to test it..

[glweek]

so fun to see this unfold!

[Gorgok]

The plate won't be an issue for heat transfer, its the radiator that makes or break everything. If you can't get the liquid cooled then the plate will eventually fail at its job. You might get localized hotter spots especially if you cluster the big heaters, but you would get those even worse on a plain finned heat sink (so proper spacing is still important). I wouldn't bother testing how much it can suck up heat though, and a radiator should say what they hope it can dissipate... Keep in mind their ambient may not be yours.

[kenneth wolfe]

another texas resident !!

[karimwassef]

Once you plumb it up and get it working, measure the change in temperature between the plate and your ambient air temp.

You know your expected power dissipation and geometry.

http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/heatra.html

You can determine the temperature rise due to the LEDs and compare to the temperature reduction you measured for the plate.

[karimwassef]

So how did this turn out?

I'm going to make one for a unique application (thin profile LED strips 96" x 0.75") and I'd love to learn from your experience first.

[kenneth wolfe]

So how did this turn out?

I'm going to make one for a unique application (thin profile LED strips 96" x 0.75") and I'd love to learn from your experience first.Sorry for so long reply, the tank was sold , the light fixture, was running when it lefted, quickly working with it , it worked out as planned, never heard back from them , honesty, I completely forgot ...until now..

Sent from my SM-G892A using Tapatalk