More Money Than Brains - a water cooled cab build

Any updates on this?

If my math is correct, you have about 370watts of LED's being cooled by a single 120mm radiator and fan with less than a gallon of water? That's pretty impressive. No issues with the thermal adhesive or the cheap water blocks so far?

More like 310 watts: three 4kK Vero-29 @ 1.88A (~240W), one 4kK '18 @ 1.67A, and one 4kK '18 @ 1.12A. The radiator is copper and double thickness at 120mm, so it's almost as good as two 120mm radiators (with two rads, both receive ambient temp air instead of the back half receiving air warmed up by the front half.) There's about half a gallon of water. The reservoir size just determines how long it takes to heat up before reaching equilibrium. I'm only using the thermal epoxy on the '18's. The '29's are just big enough that the holes clear the blocks so they are clamped on with T-Nuts and regular paste.

No issues at all so far with the cooling system. No leaks or anything like that. Its virtually silent, which I like. I haven't been running the flowering lights apart from the testing I did, however, it shouldn't make a difference. In a few days I'm going to start some flowering, I need to get some circulation fans and fix the doors on that compartment.

I made a schematic of the cooling system for folks that know nothing about PC water cooling.

So your water pump isn't self priming? Do you have a link to the pump that you're using by chance? What kind of flow/pressure is it capable of?

Props sir, it's a very nice system. I love the fact that it's very modular, easily expandable, and actually carries away the heat out of the room instead of venting it in the room.

Right now I'm using a small aquarium pump, an Eheim 1046. It's not self priming, none of the pumps used for water cooling usually are. That's one of the reasons why PC WC setups have a reservoir and are available as pump/res combos. The res makes filling easier and keeps the pump from running dry if there is a bunch of air dislodged or introduced through inevietable evaporation.

The Eheim says it only has 4' of head pressure but it seems to be working fine against 5' of head, five blocks, the radiator, and 40' of tubing. I received a new controller board for the Aquacomputer pump I have, which is based on the Eheim 1046. I haven't installed the board or tested the new pump yet. The new pump is basically the same thing but has an integrated water temperature sensor, fan output, and controller to regulate the fan and pump speed (the 1046 just runs flat out all the time.)

Also, as SupraSPL has been trying to drill into our heads, it's not so much the total wattage but the dissipation wattage that matters when sizing the cooling system. You could have the same wattage at the wall but if it is spread over twice as many of the same emitters it will dissipate less heat while generating more light. If you use less watts to get the same light you get a lot less heat. My system isn't running hard though it's not a paragon of efficiency either: the flowering emitters are running slightly soft at max (dunno how much I'll be turning them down yet), the mother emitter is average to soft, but the vegging emitter is driven fairly hard.

It's like any other engineering exercise, there are many tradeoffs: do you want to pay now or later? is heat your biggest concern? total energy consumption? fixture thickness? plant height? reflector availability? shock/fire hazards? labor costs? lots of ways to skin this cat!

There are always trade-offs with any design. The water cooling system is nice because it actually removes the heat from the chamber.

Any reason for running just 2700k emitters? Seems like a lot of folks like to supplement them with 4kK or 5kK emitters.

Looking forward to your results.

O&R, good job! I too, was wondering when someone
would create this type of build. Watching from afar.
Stay safe.

I'm running all 4000K Veros. Mr. Flux recommended them for flowering because the extra blue should help with stretching in my short cab. IIRC, his calculations show Vero to be somewhat more red than CXA so the 4000K is very much like 3500K in CXA. I'm using 4kK in the veg chamber because I'm also planning on using it for pre-flower when the flowering chamber is backed up. I used it also for the mom chamber because of inertia. By the time I actually ordered I was extremely burned out from 3+ months of intense research and was ready to just get started. I may put both 4kK '18s in veg because I think veg could use more light and more spread then get a 5kK one for mothers.

I'm not convinced that spectrum is as important as some folks make it out to be. Look at how well HPS does with its spectrum. It seems to me that intensity is more important overall and most of the poor results from LED come from simply not using enough of it (e.g. weak little UFOs and such.) COBs finally make it easy/cheap enough to match the intensity of HID.

BTW, the water carries most of the heat out but not all of it. Without any ventilation the lights heated up the flowering compartment by a bunch. The light itself is somewhat warm, you can definitely feel it when you put your hand six inches away. The block radiates some heat and so does the surface of the COB itself.

Have you tried running the flowering chamber with ventilation yet? Do how much, and how rapidly did they heat up the cab exactly? Do you know what your ambient temps are in the flower chamber without the lights on and the temps of the room your cab is in?

Sorry for all the questions, but it's all very useful info.

I think the temperature side of the temp/humidity gadget I got is off by 9 degrees! So my cabinet is not nearly as hot as I thought it was. When turned up to 1.9A, the '29's do put out some heat but backing them off a bit really keeps them cool. I recommend going for 1.4A on the Vero-29's when designing your system.

I still need to play with the ventilation, to see how much I really need. The true test will be once the flowering chamber is filled with big buds! I'll bet I'll need even more flow then.

Four out of the five Timewrecks turned out to be male! The damned "mothers" in veg didn't show sex until I made some cuts and flipped them, of course. I just killed all the males. A couple of days ago I made a cut of the only mother which will go into a big pot to veg under a scrog net next week if it shows roots. The flowering cut I'll just let go it its dixie cup since it's been under 12/12 for over a week.

Now my cab is feeling empty so I'm also going to pop some more beans, this time five Mr. Nice Critical Widow.

Observe & Report said:

I'm running all 4000K Veros. Mr. Flux recommended them for flowering because the extra blue should help with stretching in my short cab. IIRC, his calculations show Vero to be somewhat more red than CXA so the 4000K is very much like 3500K in CXA. I'm using 4kK in the veg chamber because I'm also planning on using it for pre-flower when the flowering chamber is backed up. I used it also for the mom chamber because of inertia. By the time I actually ordered I was extremely burned out from 3+ months of intense research and was ready to just get started. I may put both 4kK '18s in veg because I think veg could use more light and more spread then get a 5kK one for mothers.

I'm not convinced that spectrum is as important as some folks make it out to be. Look at how well HPS does with its spectrum. It seems to me that intensity is more important overall and most of the poor results from LED come from simply not using enough of it (e.g. weak little UFOs and such.) COBs finally make it easy/cheap enough to match the intensity of HID.

Click to expand...

http://www.licor.com/env/pdf/photosynthesis/AppNote5.pdf

I bet increasing flow through the blocks will help with heat when I turn up the '29's all the way. With everything on you can definitely feel a difference in temp between the water going into the cab and the water coming out if you put your hand on the metal fittings. I still need to fix the fancy pump and try that, it is supposed to have much more pressure and flow than the stock pump. I also should remove the COBs and check the TIM, I probably have too much glooped on and not evenly enough which would prevent good contact with the block.

OK, I've been exploring my ventilation situation and I've figured out some issues.

1. The intakes are sucking in hot air from the drivers. The 90W drivers put out a bunch of heat behind the cabinet. There is a bunch of ducting mashed up against the wall and my fresh air flow is not setup right and I will have to reroute some ducting. I'll probably need to rig some cardboard ducting around the side of the cabinet so fresh air from outside the closet gets into the cab. I will also move the drivers away from the cabinet, probably outside of the closet which was my original plan. I'll probably have to turn them down to keep from bleaching my plants but getting them out of the closet is probably a good idea anyway.

2. I could cut another intake on the flowering chamber. When I open the door I can totally feel more flow out of the carbon filter. However, with the drivers relocated and fresh air rerouted I might not need the high flow and can turn the fan down to the lowest setting again. It will depend on humidity levels when filled with buds.

nice build

Hi mr "observe and report" LOL,

Well I don't know where to begin, Great build with nice and precise indications. Don't say you are not handy because this is one of the cleanest setup I saw.
I will do pretty much the same wardrobe than you, and while I was planing everything and wondering some stuff about watercooling I decided to post here.
Abiqua send me on your thread. When I discovered it I became mad ^^. This is the perfect guide

Please let me ask some questions and criticism:

Gaius said:

Bravo! Can't wait for the report

Velcro, Genius. Hopefully they stay cool enough so that the Velcro adhesive doesn't fail. Some staples might help it stay put though.

Click to expand...

Velcro is the shit, I didn't think of it but now, it makes sense... Does the glue melt or does it work perfectly ? That was just a GENIUS IDEA!

I wouldn't have choose the parallel venting, because the air from flowering is perfect for vegetative phase: hot and humid air...

Why don't you put the watercooling-heatsink in the very front of your the exhaust directly ? a PC fan just add noise and is less powerful than your exhaust.

I am sure you are already aware but there are cations and anions exchange when the water passes through the system. You choose Aluminum joints but copper waterblocks and your heatsink might be copper protected another metal (in general it is the case). I think with time the water will "eat" your metallic part so you should put a bit of additives or car's cooling liquid.

I have an eheim pump too, the best according to me (also loving aquarium ), but I think your one is better, as it's ceramic one. You said it is 4' high but how much L/hour ?

I think you could improve the efficiency by firstly making the loop passing through the reservoir before cooling done the veg area (when the reservoir is still cool at least) and secondly puting your veg area above the flowering room. So that the height between them is only 2' instead of 5'. Then your pump could be pushing on less height so more water should flow and the led would be cooler. Btw what is the temperature of the water ? and the one of the cob junction ?

Last question: You have no reflector nor lenses on your cob and they are more than 3' above the young plant, does it stretch a lot ? do you think you have enough light from this height ?

I apologize for all this questions and remarks, but I think you are the closest to my set up and maybe what I said will help you improving the cab .
I am sure it will grow some big bud anyway
cheers

Thanks toze you and speedyganga!

The Velcro adhesive doesn't melt, the COB/block doesn't get that hot. However, the block is small and the lines going in and out on one side of it can put a lot of uneven force on it. That can make the whole thing not point straight down, but it's no biggie really because the lights aren't very directional. I'm intending to bolt the blocks to a thin piece of plastic that is 2-3X larger and velcro that to the ceiling.

Parallel venting seemed easier to rig up so thats what I went with. However, serial ensures you get the full flow through the flowering chamber. My real problem with ventilation is the carbon filter is restricting the flow down to very little. I don't want to increase the fan noise so I'm going to look into finding a different filter that is less restrictive (Green Gator?)

Using the exhaust fan to cool the radiator is a great idea! However, that would reduce flow from the exhaust and right now I'm trying to increase it without adding more noise if possible. Also, the air temperature would be elevated over ambient having been heated by the LEDs inside the cab so you would lose some cooling efficiency. In addition, I have the carbon filter outside the cab at the end of the exhaust so I would need to switch to a very expensive inline one. Finally, I have the exhaust and cooling system on different sides of the room! So this one probably isn't happening.

I am pretty sure all of the fittings are plated brass and not aluminum. They are certainly too heavy to be aluminum and are plated (nickel?) I probably could have some corrosion going on as there is some gunk in the lines. Some of the gunk is from the CuSO4 I added to keep algae from growing. Switching to some proper coolant might be a wise plan.

IMHO, Eheim is top quality and very few brands are up there (Royal Exclusiv and Tunze) I wouldn't trust anything else. I have no idea how much flow I'm getting but it's not a tremendous amount. It certainly could be more since I can feel a temp difference between input and output on the metal bulkheads on the back of the cab.

Putting the veg on top is another brilliant idea! I'm probably not going to rebuild the cabinet just yet, however. I did recently removed some excess tubing which will certainly help a small bit. I think elevating the pc-case/pump off the floor would also help some and I'm looking into that. On the advice of bcit, I rerouted the hot return from the cab to go straight into the radiator. I also shortened the lines a bit which will help flow slightly.

I raise the pots up when the plants are small so they get enough light and I use an Apogee quantum meter to ensure they are getting the right amount. In previous posts you can see I can get over 1000 umol/s/m in the center 12" away from the emitters, which is only 19" from the floor. So the plants won't be stretching from lack of light and also the 4000K Veros have plenty of blue. Veg could use another COB though because I want to have better spread and more power. I should get at least one grow out of this cab before I dump a bunch more money into reconfiguring it though.

Without reflectors the spread is pretty good over 35 x 18 with only three emitters but the spread would be better if I had more emitters at lower current while also having better efficiency. Reflectors would also require even more emitters to cover the whole space evenly but would be very efficient but with a very high initial capital and labor outlay.

So I've done a bunch of tweaking.

After some postings and discussion between the usual suspects (Star Dust Sailor, SupraSPL, Positivity, etc...) I redid the TIM on the flowering blocks (the veg ones are expoxied.) Sure enough, I put on too much and didn't make sure it completely covered the spreader on the back of the COB. So I cleaned it off and smeared on the thinnest layer of Arctic Silver 5 I could using a gloved finger. I really smeared it across and it was hard to completely cover it since the smearing would take off some previously applied bits. The tube was an old one I had laying around, I hope it is still good but I think it is only a couple years old.

before pic; you can see the TIM doesn't cover the heat spreader on the emitter all the way and is probably too thick. I just squirted on a bunch and hoped it would spread out. Nope, it's too thick. I wonder if that is corrosion on the water block from aluminum<>copper contact where there was no TIM. At least the thing appears flat and good contact can be had.



after pic; smeared on the thinnest layer I could all the way across the spreader, maybe with a heat gun and a tool I could get it thinner?

WRT heat, I moved the drivers off the back of the cab to a spot outside the closet. I also cut another hole in the flowering chamber and redid the intake ducting. I'm positive now my problem with not getting enough flow is the carbon filter is too restrictive. As I mentioned above, I'm going to see if maybe Green Gator filters are less restrictive since they're basically the only one that isn't a can full of granulated carbon with a screen around it.