Economical multi LED Chip Projects for Growing

DIY-HP-LED said:
I tried to, but felt the readings were pretty meaningless with the light from the COB and an infrared thermometer. I'm gonna try some new batteries in my multimeter and see if I can get the "K" type thermocouple probe to work, that should allow more measurements including LES surface temps.
Click to expand...


It works it just takes the right distance and a little longer to get an average reading.

I know I was getting 30C-35C around the light then 40C-55C face temps depending on mA's so it wasn't a huge difference.

I do know once I put the driver to 1750 mA ( 250w/4 cobs ) temps would climb over 60C on the face.
 
Shugglet said:
Its more efficient than the competition, the rest of the claims though are easily debatable. But even the efficiency could probably be debatable (air is a lot easier to move than water is).

I mean, you really think its easier to build a custom water cooled heatsink for your lights than adding passive heatsinks which seem to be extremely popular? These passive designs also create no noise.

Bare in mind Im all about water cooling and its perks. But Im also willing to admit that its typically more of a niche for individuals where a traditional air cooled set up simply will not work.
Click to expand...
Look further up in the thread, I've already built a cheaper water cooled light than an air equivalent (and I've built a few of those too). In the up coming 400 watt water cooled design, I'll post the costs of the cooling system, what somebody puts on the business side of the lamp in terms of LEDs is their concern. I'm just trying to come up with economical grow light designs that some one could build in an apartment for very little cost or trouble. If you want to put 400+ watts in a sealed tent with an alcohol lamp for CO2 enrichment and have your carbon filter recirculating air in the tent with hardly a whisper, then this is it.
BTW 12 V pumps are cheaper than fans and certainly less expensive than blowers, vinyl hose only costs about 25 cents a foot and I just got a car heater core at the scrap yard for $10 to act as a chiller for the big water cooled rig that's in the works.
20160904_184718.jpg
 
DIY-HP-LED said:
For 175 watts to raise the temperature of 1 liter of water 1 degree C takes 39.5 seconds. Here's how I found out
Since I'm moving 2.5 liters/min of water through the system, I could literally cover the bar with 100 watts chips driven flat out with no heat problem at all and if I had any issue with heat, I'd just get a better pump and up the flow rate or lower the temp of the water.
Click to expand...

Few things Ive noticed. If you use that site for calculations be sure to convert fractions of a minute correctly to seconds (easy oversight but has a pretty big impact). Case in point the answer you were looking for isnt 40 seconds but closer to 25 and thats a pretty big discrepancy.

Also, upping the flow rate would not help because you are in a closed system, you have no choice but to lower the temp of the water by removing heat from the system.

The numbers you need to look at are how fast youre adding heat to the system with pumps and lights vs how fast youre removing the heat from the water via a chiller or radiator.

You should consider the same reasons water is great at taking in heat from the lights makes it difficult to transfer that heat from water back into the air. Granted the radiator for the car will be much more efficient at removing heat from your system but you still might find its tougher to scale it up.
 
Shugglet said:
Few things Ive noticed. If you use that site for calculations be sure to convert fractions of a minute correctly to seconds (easy oversight but has a pretty big impact). Case in point the answer you were looking for isnt 40 seconds but closer to 25 and thats a pretty big discrepancy.

Also, upping the flow rate would not help because you are in a closed system, you have no choice but to lower the temp of the water by removing heat from the system.

The numbers you need to look at are how fast youre adding heat to the system with pumps and lights vs how fast youre removing the heat from the water via a chiller or radiator.

You should consider the same reasons water is great at taking in heat from the lights makes it difficult to transfer that heat from water back into the air. Granted the radiator for the car will be much more efficient at removing heat from your system but you still might find its tougher to scale it up.
Click to expand...

Thanks for the correction on the flow rate. As I mentioned this is just a test system I'm playing with and the cooling system is rudimentary to say the least! It's mostly for testing out sealing the ends of the tubing easily and cheaply and adhering the LEDS. I live on the east coast of Canada on an island at the end of the world and it's quite chilly here much of the year. Just placing the car heater core outside in a bucket of water/antifreeze will keep my coolant temps near or below 0 C for 3 or 4 months of the year. If I do this, I'd have to insulate the water jacket on the lamp with 1" wide weather striping to avoid condensation forming on the lamp, hoses and reservoir! Also drilling a couple of 9/16" holes in a basement window frame is pretty easy. I don't anticipate too many problems dumping heat out of the system, there are many economical solutions like car radiators and heater cores.

Right now, with 175 watts of heat, the test system runs about 5 degrees C above ambient using 24 liters of water with the crappy cooling system placed near the ceiling outside the grow room. I can do better on the first build and that should be done and posted in a week or sooner.

Also, I was just out to the thrift store (second hand) and scored a 60 liter tote for $1.25 and a 23 watt fountain pump with a 1/2" outlet that has a better head and 3 times the flow rate of the pump I'm using for about 50 cents, works like a top.
 
Last edited:
DIY-HP-LED said:
Thanks for the correction on the flow rate. As I mentioned this is just a test system I'm playing with and the cooling system is rudimentary to say the least! It's mostly for testing out sealing the ends of the tubing easily and cheaply and adhering the LEDS. I live on the east coast of Canada on an island at the end of the world and it's quite chilly here much of the year. Just placing the car heater core outside in a bucket of water/antifreeze will keep my coolant temps near or below 0 C for 3 or 4 months of the year. If I do this, I'd have to insulate the water jacket on the lamp with 1" wide weather striping to avoid condensation forming on the lamp! Also drilling a couple of 9/16" holes in a basement window frame is pretty easy. I don't anticipate too many problems dumping heat out of the system, there are many economical solutions like car radiators and heater cores.

Also, I was just out to the thrift store (second hand) and scored a 60 liter tote for $1.25 and a 23 watt fountain pump with a 1/2" outlet that has a better head and 3 times the flow rate of the pump I'm using for about 50 cents, works like a top.
Click to expand...

Sounds like a pretty ideal locale for water cooling. Keep in mind my comments are more steered toward the people who see what youve done and think it wouldnt be hard to incorporate in their grow, but your setup might not work well with their particular situation. In areas where ambient temps are much higher it would take enormous amounts of water to counteract the deficiency in removing the heat.

I love the way you hunt deals though, very smart and thrifty. The only thing is I would perhaps be wary of putting a used pump in the system as you have no idea how much life its got left (youre probably using it for testing which is smart, just might get a new one for the permanent fixture).

Also, just remember you are dealing with a closed system with water cooling so when youre doing calculations use all the water in the system and all the heat energy being input (dont forget submersible pumps put all their wattage as heat into the water).
From there you can calculate how much the system should be heating up, and the difference between that and the actual temp will be how much heat your radiator is removing from the system.

Hope this helps and look forward to following your progress.
 
A WATER COOLED MODULAR LED GROW LIGHT
This is what I'm going to build first, its a 400 watt grow light rig composed of 6 1"x1" x18" water cooled light bars attached to two pieces of 1"x1" aluminum angle. I got these particular thin walled tubes at the junk yard, but you could do the same thing using tubing from a metal supply or machine shop. These thin walled aluminum tubes were once legs for tables, I think, and have plates welded on each end with 1/4" tapped holes. One tube has an end missing and you can see the profile of the tube. I'll put a fitting in this end using silicone adhesive, just as you would if you were building a lamp from new material. The rest of the holes in the ends of the tubing will be drilled out to 1/2" and tapped with the brass 1/2" pipe thread of the barbed fitting. I'll screw the fitting in half way to tap the pipe threads in the tube end then remove, after cleaning things up, I'll screw them in 3/4 of the way using silicone or teflon tape on the fitting threads.
20160904_185233.jpg 20160904_185247.jpg
 
Here is what it will look like with the 6 18" long light bars laid out on the bench. It will be 3 feet long and is designed to cover half my grow table until the big water cooled rig is built. When I get the big rig done then I'll take this light apart and turn it into two five foot long light bars made of 3 modules each hooked together with short lengths of hose. This will then become part of a big water cooled light I'm gonna build with in the next month or two.

20160904_191046.jpg
 

Attachments

  • 20160904_191117.jpg
    20160904_191117.jpg
    183.3 KB · Views: 5
Shugglet said:
Sounds like a pretty ideal locale for water cooling. Keep in mind my comments are more steered toward the people who see what youve done and think it wouldnt be hard to incorporate in their grow, but your setup might not work well with their particular situation. In areas where ambient temps are much higher it would take enormous amounts of water to counteract the deficiency in removing the heat.

I love the way you hunt deals though, very smart and thrifty. The only thing is I would perhaps be wary of putting a used pump in the system as you have no idea how much life its got left (youre probably using it for testing which is smart, just might get a new one for the permanent fixture).

Also, just remember you are dealing with a closed system with water cooling so when youre doing calculations use all the water in the system and all the heat energy being input (dont forget submersible pumps put all their wattage as heat into the water).
From there you can calculate how much the system should be heating up, and the difference between that and the actual temp will be how much heat your radiator is removing from the system.

Hope this helps and look forward to following your progress.
Click to expand...

Thanks for the interest and contributions, I hope to build a couple of water cooled rigs and share what I've learned. There doesn't seem to be much on DIY water cooling COBS online so it's always good to till fresh ground. I also hope to get some info and suggestions from others who are following along and I'll go into as much detail as anybody wants to get them through a grow light DIY build. In Canada it's legal to grow medically and it will probably be legal to grow recreationally in the spring. My focus is helping poor people to grow their own medicine with cheap DIY lights.
 
Since the goal is to produce as many photons as possible, using lower cost components often requires more of them to produce the same radiant power as the expensive brands. There is a price point where it becomes more effective to use either and the challenge is to find out what those are. A good example is the Citizen 1212s up against the Cree 3590s; you can get 3 or 4 1212s for the same cost of 1 Cree, which combined will put out more radiant power using the same electrical power.

Find the most cost effective driver(s) that fit within your power limit then with the remaining capital acquire as many emitters as possible that will run on that driver. By running the chips softer you are improving efficiency in producing radiant power and reducing thermal waste generated by each emitter. Downside is additional work required to research components and constructing a larger system.

Water cooling anything generally will improve efficiency. Distributing the radiant power generation across more emitters improves efficiency and also reduces thermal waste, probably at a faster return than the cost for the water cooling I would guess.
 
Just called my metals guy and ordered in a 20' X 3" X1" @$3.85 CDN/ft. I could get shorter lengths cut to order, but I want to convert my entire air cooled grow to water. I also have a buddy who was so impressed with the performance and ease of building of the test rig he wants two 3' long light bars as well so he can convert two recently built DIY air cooled bars.
If you live in the US I'm pretty sure you could get a better price on aluminum tubing.
 
I didn't do much work on the lamp it was 30 C here today, the garden was hot as Hell at 32C and the workshop was like a sauna. I did manage to get a hole drilled out and the fittings on one tube. I siliconed a large brass fitting into one end of the tube with an end plate missing and tapped the hole on the other end with a brass fitting. The large pipe thread 1/2" barbed fitting shown here would thread right into thicker walled 1" square tubing, just fill in the corners with marine grade silicone adhesive. 20160907_152035.jpg

After I removed the brass filling I replaced it with a cheaper nylon version of the same thing (I wanna test it).
20160907_200908.jpg

Here is what the LED layout for one bar looks like, The 3W 660nm reds will be in series strings of 5 driven off a 12 volt power supply through an adjustable buck converter. The warm white 10W leds are driven the same way and the 25W 4 band multi spectrum COBs will be driven by 30V 3.5A 100 watt drivers (7 COBs per driver@500ma) All chips will be applied with double sided heat sink tape.
20160907_200652.jpg
This bar has 4X25W 4Band + 3X15W 660nm + 2x10W WW
The real wattage = 60W 4 Band + 20 W 660 + 15W WW = 95 Watts total power. Some bars will have less power for now as I'll have more LEDS on the way and will they will be applied when these modules are incorporated into a larger array.
 
Last edited:
I'll get a price on new thick wall 1" square tubing when I go to pick up my 3"x1" rectangular tubing on Friday, perhaps I'll pick up a short length while I'm there to fool around with and try a brass fitting in. I figure it should run about $1 - $1.50 CDN. Anybody know what 1" square tubing costs a foot in the states? The brass and nylon fittings are from the local hardware store and since I'm using 12 of them on this particular project it can add up, which is why I think I'll take some of the brass 3/8 fittings back and use nylon barb fittings (I'll keep the used fitting for tapping the other holes). I'm trying to do this project on the cheap, not because I can't afford better, but so that I can demonstrate how easy and economical water cooling can be. I'm kinda making a game of it and having a bit of fun as well as saving money of course. The tubes I'm using were bought at the local scrap yard for a buck a pound, but if you bought new thick walled tubing it would only cost $10-$15CDN, brass fittings are in the $3 range and nylon ones about a buck or so. The 7/16" vinyl tubing to connect the tubes is about 35 cents a foot and the half inch tubing to connect the lamp to the cooling system is about 45 cents a foot. I just scored a great fountain pump (23.5W) with a half inch outlet, about 15 L/min flow rate and a good five foot head for about 50 cents at a thrift shop (I've got a 12V 350L/hr pump coming from ebay too for about $13CDN). All these prices are in Canadian dollars, US prices should be less.
 
Last edited:
SUPER CHILLING:blsmoke:
Here's a bone I'll throw out to those interested. Is super chilling the coolant worth the effort? To avoid condensation problems it would involve insulating the lamps water jacket(s) with 1" wide self adhesive closed cell foam weather striping, insulating the hoses and reservoir and running an antifreeze mix. I could easily do this in the winter by placing a car heater core in a container of antifreeze and water outside to chill the inlet coolant to the lamp. For about 4 months of the year I could run my coolant at below zero just by drilling a couple of holes through a basement window casing, it would even be on the same level as my lamps!

I figure it's not worth it and I'm gonna shoot for a coolant temp of about 15C in winter, this should allow the LED junctions to run at 25 C. I don't expect much of an increase in white light and the test I did with 5 lbs of ice in 16 liters of water during the bench test of the lamp confirmed this. I had my LUX meter under the lamp during the test and as the lamp water jacket approached 0C the LUX readings went up by less than 5%. The lamp sweated like crazy in the summer humidity of the east coast and if it persisted long enough I would have had electrical grounding issues. The test lamp was running for an hour and a half and the 16 liters of coolant was about 31 degrees when the ice went into the tote, it was gone in 10 or 15 minutes.

However I did notice that the warm white LEDS seemed a little more red with water cooling. I expect that running my coolant at 15 C will give me about a 50% increase on the red spectrum LEDs and some red parts of the 4 band LED's spectrum.

Here is the data I'm going by, note the chart on the heatsink effect on LEDs produced by the US DOE. It's on a pdf on water cooling that you can find HERE
The document is from a Korean grow light manufacturer and the data appears reliable.
 
Last edited:
Way to rock the DIY man! Thank you for taking the time to post, much appreciated. What type of monos are those? My eyesight is getting bad, lol.
 
Thanks for the like and comments, the 3 watt monos are red 660 nm and the 4 bands are 660,640, 440 and 460 I believe.

I'll be posting the upcoming builds in detail so that some one else can follow along. Assuming you have LEDs and a driver, it only takes a piece of rectangular aluminum tubing, a couple of hose barbed fittings with pipe threaded ends and a $4 ebay pump to try your own test rig. Figure out how to seal the ends with silicone adhesive and yer done. These tubes run unpressurized and at below or at room temps with no thermal expansion and contraction cycles. Think in terms of aquarium construction methods for sealing the tubes. The tubes run so cold and are so good at sucking heat out of the COBs that you can use a wide variety of methods to adhere or attach them to the bar, CPU tape is very effective and cheap.
 
You must log in or register to reply here.
Back
Top