Economical multi LED Chip Projects for Growing

[DIY-HP-LED]

A Pair of Veg Tube Lights
Here are a pair of 5000K veg lights made from LED street lamp arrays and drivers. The driver stacks are located in the ATX power supply boxes along with the fan power supplies and terminal strip connecter. I'm using the ATX box power supply fans to cool the drivers, but I'm going to try and disconnect one and see what the temperature of the drivers is like with out cooling and if it works out, I might remove the fans from the electrical boxes on both lamps. The 3" x 1" x 30" aluminum tubes are cooled by 30cuft/min fans drawing air up through the tubes and they run about 2.5C above ambient. Each tube lamp uses two 10" X 3" on each end and two 5" X 3" LED street lamp arrays in the center. The driver stack is composed of a 44 watt driver on the bottom and a 88 watt driver on the top, the 44 watt one drives the two center arrays and the top 88 watt one drivers the two outside arrays. They are held together and mounted in the ATX box with nylon stand offs and nuts purchased as an assortment kit on ebay. Each of the two tube lamps uses 132 watts of power for a total of 264 watts used by the pair. The 10" arrays running at 525 ma have an efficacy of 88 Lm/W X 88W = 7,600 Lumins, the 5" arrays running at 525ma have an efficacy 90 Lm/W X44 W = 3,750 Lm, for a total light out put per veg lamp of 11,350 Lumins, so the pair of lamps output 22,700 Lumins.

 

[DIY-HP-LED]

I've been continuing with my cheap eBay LED experiments and coming up with new designs for grow lamps. Some of the double chip designs made with kitchen canisters and old PC power supply boxes ran hot and eventually damaged the cheap leds. A CPU heat sink bolted to an aluminum plate was too much for the heat sink to handle and spacing the LEDs with lenses close together caused hot spots on the canopy. A single led running at 50 or even 75 watts mounted directly to the heat sink works better for this type of design.

I've recently built lamps out of old RF amplifier housings bought at the local scarp yard that work quite well as double chip lamps that run very cool and with a chip spacing of about 10.5". This provides a more even and larger distribution of light on the canopy 14" X 24" @ 14", high using lenses and reflectors with a luminous flux of around 50 to 60K lux. This lamp uses 2 generic warm white 100W LEDs running at 50 watts each in parallel, off a 100 watt driver. Lens and reflectors help to focus almost all the light the LEDs produce into two over lapping intense cones of light at the correct distance of around 14" from the canopy

My latest designs use 4"W X 2"H X 16" long aluminum channel with a chip spacing of 11". The top of the channel is covered by a custom bent piece of aluminum flashing, to make a 4" square tube 16" long with fans in both ends. Heat sinks are mounted over the LEDs on each end of the lamp and the air flows through in one direction. This means I can run the 3" fans slower and if a fan fails, then there is a backup, in addition these lights can be ducted together with hose to extract heat from the grow if required. I feel that this design offers coverage, economy and flexibility. With the addition of side lighting it will grow one plant for about 100 watts of power and is perfect for a closet or small tent grow.

I think I'll start a new thread outlining my adventures in building cheap LED lights and throw out a few builds to give folks some new ideas and ways of looking at led plant lighting.

 

[DIY-HP-LED]

NEW WATER COOLED DESIGNS

It's been awhile since I've posted anything here, but I've been busy building lights, growing and learning. My light designs have gone through a couple of generations and I've recently completed a large air cooled light (500 actual watts) with a flow through design and ducting to remove heat. I wasn't satisfied with the result, so I dropped plans for a larger 4' long version. I'm using over 1000 watts in a regular sized room and the heat build up is too much and air cooling the LED lights wasn't working out too well. So I decided to try a radically different approach, water cooling. I think I might start a new thread on this subject since I figure it deserves more attention. I've got a couple of economical designs in the works that should prove very interesting to any DIY grow light builder.



A few days ago I made a prototype light to test a few ideas and shatter some assumptions. The ease of building, cost and efficiency, convinced me that this is the way to go for anybody who wants to put lots of LED power into a small space with no increase in heat. Your carbon filter and blower are for reducing odor, not cooling your lights, with water cooling you can reduce your blower speed to a fraction. A sealed room and CO2 enrichment is easily possible with water cooling and the only noise you'll hear is the sound of an oscillating fan.

The most common arguments against water cooling are: expense, difficulty and it's not necessary for most small growers. What if I told you that you could build a water cooled light bar cheaper than you could build the equivalent LED grow light on a heatsink and for less bother and with no screw tapping. If your LED junctions run at 80 C with air cooling, then with ambient temperature water cooling (24 C) they will run at 35 to 40 C, cool the water down to 15 C and your juction temps will be in the 25 C range. This means your COBs will last over 200,000 hrs with very little diminishment over time, you can drive them harder, they will produce more light, especially the reds and use less power. Even Cree and Vero fans will find this useful, since if yer paying hundreds or thousands of bucks for COBs you'll want the best possible cooling, and water is it. Stick around and I'll post some pictures, plans, builds, data, ideas and info sources on economically water cooling COBs. All the materials for a water cooling system can be obtained locally and are cheap, the only tools required are an electric drill and a 3/8 drill bit.

Looks like I'll go take some pictures with my phone for anybody who's interested, if you are, post a reply. I'll post the build info and data on the test light here and perhaps the big 1000 (actual) watt water cooled rig I'm planning. I'm starting another water cooled light (perhaps today) with some scrap aluminum pieces and recycled drivers, so perhaps I'll document the build.

If I carry this forward on a new water cooling thread, I'll link to it here.

 

[DIY-HP-LED]

PICTURES AND INFO ON A WATER COOLED TEST LAMP
Ok time for show and tell, below are some photos of my water cooled test lamp. It was built to answer a few questions foremost of which were: how to seal the ends of the tube, apply the chips and water tight hose fittings. I sealed the ends of the 20" X2" X 1" rectangular aluminum tube with 1/2" thick plastic plugs sealed with marine grade silicone adhesive and siliconed an aluminum plate with an overlap over that. I used 1/2" pipe threaded barbed brass fittings threaded into the aluminum using the fittings to tap the threads in the soft aluminum. There are 5 100 watt warm white LEDs on the rig 4, aluminum backed cheapos and the middle LED is a better quality copper backed type. The crappy chips are held on with cheap black heatsink tape that's not even rated for CPUs! The middle chip is held on with nylon electrical tie wraps and uses heatsink paste.




This is a test rig made from parts I happened to have laying around and was put together in a day and bench tested the next. It's purpose was to prove a few ideas, test water cooling, collect data and gain real world experience before committing to a larger water cooled rig.

 

[DIY-HP-LED]

Here is how I'm cooling the test rig, the system was on for about 8 hours (ambient temp 22 C outside the grow room) when these photos were taken. I'm using a small plastic tote as a 24L reservoir, a 10' 3/8" copper coil and a small fan set on medium to cool everything. I have it arranged so air flows around and under the tote for added cooling. The tote is also at about the same height as the lamp, cause I'm using about the cheapest 12V pump on ebay (this is built from stuff I had laying around).

Some shots of reservoir and lamp inlet and outlet temps, the flat black spray paint is to get accurate readings from aluminum surfaces.

 

[DIY-HP-LED]

Here is how I'm protecting the lamp, so far. I'm using a 60 C thermal cut off switch that I happen to have. For the big rig I'll use a 40 C thermal cut off and a 65 c thermal fuse in series with the lamp power supply as a final fail safe. Primary thermal cut off will be via a digital thermostat set at 35 C


The center chip is tie wrapped on and it's cold to the touch after hours of operation as is the tie wrap. Note the 60 c thermal cut off switch under the tie wrap with cheap thermal tape.

 

[DIY-HP-LED]

THE TEST LAMP AND THE MATH OF WATER COOLING

The lamp is built using a piece of 20"X2"X1" aluminum tube like this.

You can get this tubing cut to length from any metal supply store or machine shop that builds or fixes trailers, fishing or agricultural equipment. You can easily cut it using a circular saw or table saw with a carbide tipped blade (use safety glasses). Depending on the size, it ranges from $1 to $4 a foot.

I filed the ends flat and beveled them a bit inside and out, then I roughed up the ends, inside and out with coarse sandpaper for about an inch in to help the silicone stick better. Just before gluing the plugs in and the ends on, everything was washed with alcohol and a paper towel to remove any dust or oil.

 

[DIY-HP-LED]

I have five 100 watt warm white leds on the tube driven at 50 watts each by 2 x 100 watt and 1x 50 watt drivers for a total actual power draw of 250 watts. About 70 percent of that power is in the form of heat mostly emitted from the backs of the LEDs and 30% of the energy is light with almost no heat emitted from the front of the LED. Of that 250 watts 175 watts is heat that must be carried away from the chips and dissipated. The volume of water each tube holds is 2.8" x 1.8" x 20" = 100.8 cubic inches = 1.65 liters (use Google to figure this out). My cheap ebay pump can deliver about 4 liters a minute with no head or tubing resistance, in practice, with the way it's set up now, it moves 2.5 L/min of water through the system. It's not how much water your tubes hold that counts for heat removal, but rather the water flow rate and temperature.

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. With this lamp you can put your finger on the edges of the aluminum or copper LED chip backs and they feel cold even after hours of operation. You can hold your finger right on the LED and it feels quite warm from the light emitted, try that with an air cooled fixture and you'll leave roasting finger flesh on the LES.

 

[DIY-HP-LED]

Here is some info on water cooling compared to air and why water cooled high powered grow light bars run at or below ambient temps. This means you can attach the chips with nylon tie wraps, CPU tape, cheap silicone heatsink plaster or thermal epoxy. The temperature ranges of the coolant are either cold or ambient as are the light bars. Since the system is unpressurized and operates at or below ambient temps, think in terms of aquarium construction, many modern large glass aquariums are sealed and even held together with silicone rubber adhesives. You're not going to hook these light bars up to city water pressure (50 psi), just pump low/no pressure water through them.

Some info on why water is better:

Specific Density of water and air at 20° C:

Water density is 998.2071 kg/m^3
Air density is 1.204 kg/m^3
Water is about 829 times more dense than air.

Heat capacity of water and air (J/°C g):
Water heat capacity – 4.18
Air heat capacity – 1.01
Water has about 4 times more heat capacity than air

At normal atmosphere air takes up 829 times more volume and has about 4 times less heat capacity.
829 x 4 = 3,316 times more air needed than water for the same cooling.

 

[DIY-HP-LED]

If you need any further convincing, here is a link to a pdf from a Korean company called Parus. The data seems accurate and reliable, take a look at some of the graphs and charts comparing water to air cooling . As a side note, look at the photo of the end of the light bar on the bottom center of page 3 and zoom in to enlarge. You'll see that this is an air cooled bar with an aluminum extrusion adaptor with a water channel. Note the poor fit of the fins and lack of heat sink paste between the extrusions. I believe the aluminum tubing design mentioned here would do a far better job of thermal management than these air cooled adapted light bars.

Here's a link to an informative pdf on water cooling
Parus

Light bars give the best coverage, but are hard to cool with air at high power densities, water cooling provides the answer.

 

[DIY-HP-LED]

Here is another reason to consider water cooling, if it can protect a balloon, it should do better for a COB.
Have a look at this 2 minute youTube video.

 

[DIY-HP-LED]

Here is someone using water cooled aluminum tubes with cheap chinese drivers for a large 3000 watt grow. This is about the only design I've seen that comes close to the one described here.
A 3:29 YouTube Video
An innovative cheap water cooled setup

 

[Shugglet]

Loving the setups, I did something on a much smaller scale in order to fit a grow under a bed...

But the thing is most growers going the COB route simply dont need water cooling. They run ~200w chips at typically less than ~25% capacity and for that, passive cooling is all thats needed.

There are certainly plenty of benefits of water cooling but in most cases its simply not needed.

Advantages and disadvantages to each and I will say, in some cases water cooling is the only way to achieve adequate cooling of lights. Cooling solutions are typically designed to meet an individual users needs and in most cases air cooling is more than sufficient.

If interested in the topic I recommend researching PC cooling, as theyve been studying this topic for decades.

 

[DIY-HP-LED]

Loving the setups, I did something on a much smaller scale in order to fit a grow under a bed...

But the thing is most growers going the COB route simply dont need water cooling. They run ~200w chips at typically less than ~25% capacity and for that, passive cooling is all thats needed.

There are certainly plenty of benefits of water cooling but in most cases its simply not needed.

Advantages and disadvantages to each and I will say, in some cases water cooling is the only way to achieve adequate cooling of lights. Cooling solutions are typically designed to meet an individual users needs and in most cases air cooling is more than sufficient.

If interested in the topic I recommend researching PC cooling, as they've been studying this topic for decades.

Thanks for the interest and I agree with most of what you have to say. These designs are cheaper and easier to build than the equivalent air cooled types, more effective, efficient and run quieter. In the next few days or week or so I'll prove this by building a 400 watt water cooled rig in addition to the test rig already built.and post the build pics and info here.

 

[Organic Miner]

Nicely done. You've piqued my interest in this water cooling stuff! I think there are several here that have been successful with water cooling their COBs. Good job.

 

[Shugglet]

These designs are cheaper and easier to build than the equivalent air cooled types, more effective, efficient and run quieter

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.

 

[Uberknot]

Here is how I'm cooling the test rig, the system was on for about 8 hours (ambient temp 22 C outside the grow room) when these photos were taken. I'm using a small plastic tote as a 24L reservoir, a 10' 3/8" copper coil and a small fan set on medium to cool everything. I have it arranged so air flows around and under the tote for added cooling. The tote is also at about the same height as the lamp, cause I'm using about the cheapest 12V pump on ebay (this is built from stuff I had laying around).
View attachment 3773152 View attachment 3773153 View attachment 3773154 View attachment 3773155
Some shots of reservoir and lamp inlet and outlet temps, the flat black spray paint is to get accurate readings from aluminum surfaces.


View attachment 3773226

Whats the face temps of the LED's?

Did you get any readings aiming right at the led after it was on for a few hours?

 

[Shugglet]

Ill be using copper based CPU water cooling blocks in my next design.

Ill also add that this method becomes a lot more attractive if you already have a chiller in your setup.

 

[DIY-HP-LED]

Whats the face temps of the LED's?

Did you get any readings aiming right at the led after it was on for a few hours?

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.

 

[Shugglet]

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.

What were they out of curiosity? Lofty seems to use that method and gets max temps of ~47c.