LED Efficiency VS Temperature

GrowQs

Well-Known Member
How much more efficient will LEDs get if I liquid cool them down to sub-freezing temperatures?

Like, if I put COBs like the cree cxb3590 on a liquid cooled heat sync full of antifreeze and put the heat exchanger outside in the bitter freezing cold and pump sub-freezing water/antifreeze through it?
 
We have recorded improvements on our lighting with added cooling. As far as going that cold, it would be intriguing to test :cool:.

the driver is also important to keep cool.
 
Cobs are about 35mm or almost 1.4 inches so I think they would fit nicely onto 2"X2" steel square tubing..
Make a horseshoe out of the 2x2 square tubing with water fittings on the open ends, face the tubing where the LEDs would land, drill and tap holes where the mounts (Ideal holders?) would land, insert studs with JB weld used like loctite so you have a couple studs sticking out of the tubing to hold the COBs down with nuts.. Put a radiator on the other side of the wall outside, pump liquid through it..

the driver is also important to keep cool.

Mount it to the cold tubing too, or put it outside also and run the high voltage DC a little distance?

j2VMBKO.png


I think it would be a lot harder to make a heatsync like this for QBs because of their size/shape, but COBs would be easy..

I really don't need to do anything like this but it is interesting to think about..
 
Cree has the temperature performance charts on their website.

WMatil3.png


They definately do put out more lumens per mA the colder they get but the charts only go down to 25C or 77F.. I wonder what they would do around 0C..

Something like this might have condensation/frosting problems though..

You guys ever see how they overclock Pentium 4 CPUs with liquid nitrogen? Same concept..
 
Cobs are about 35mm or almost 1.4 inches so I think they would fit nicely onto 2"X2" steel square tubing..
Make a horseshoe out of the 2x2 square tubing with water fittings on the open ends, face the tubing where the LEDs would land, drill and tap holes where the mounts (Ideal holders?) would land, insert studs with JB weld used like loctite so you have a couple studs sticking out of the tubing to hold the COBs down with nuts.. Put a radiator on the other side of the wall outside, pump liquid through it..



Mount it to the cold tubing too, or put it outside also and run the high voltage DC a little distance?

j2VMBKO.png


I think it would be a lot harder to make a heatsync like this for QBs because of their size/shape, but COBs would be easy..

I really don't need to do anything like this but it is interesting..

Did you check the weight? The tubing itself will weigh over 21lb. The tubing will hold over a gallon of water which adds 9.25lbs. By the time this is finished, you will have a light weighing over 40lb to cover a 2' X 4' area. Now add in the water lines running to/from the fixture. I see a collapsed tent coming.
 
Did you check the weight? ... I see a collapsed tent coming.

True that but no tents for me.. My rooms have 7+ foot OSB ceilings..
Nice calculations man! That sounded legit..
Reminds me of rigging class and calculating flow rates/densities etc..

A lot of work to do for something that doesn't really need it....

Chalk this idea up as 90% mental masturbation ;) Just a fun concept/theory because that's what my brain does..
Besides, this would be extremely easy to fabricate compared to half of the crap we make for fun.. I make some pretty crazy truly custom stuff on the regular mostly including a lot of metalworking and plumbing..

And even with normal heatsyncs @ $10 each that's $80 and would still need to make a rack, I could make this for darn near free..
It may seem like a lot of work if this isn't the sort of thing you do, but this is the sort of thing I do..
Cut, weld, fab, mechanic, custom performance work etc.. I make shit..
 
Nice calculations man! That sounded legit..

I knew the weight of water per gallon, and went to an on-line metal retailer to find the weight of a linear foot of 2 X 2 square tubing with 1/8" walls. I figured on 7' total length based on the light I built, and voila - an answer.

Something else you may need to address is how to adjust their height. You may need hose control, and the standard ratchet hangers may not support the weight.
 
The other thing to think about is condensation. I imagine it would build up quick.
 
Mechanical refrigeration would cost more than any minor efficiency increases.

I absolutely agree, but it is 20 something degrees outside right now and will be dropping into the teens all week..
An old truck radiator out there with lines coming in would provide a whole lot of free cold..

With all the free cold in the world you could even make a dehumidifier setup with the same contraption that would use no more electricity than it takes to pump a bit of water..
I'm sure my dehumidifier pulls some serious juice when it's running..

LED lovers seem so concerned about about peak efficiency and saving on electricity, this could be another step towards that..

The other thing to think about is condensation. I imagine it would build up quick.

Could probably fix that by encasing the tubing in foam board insulation like trimming out beams in a house..
 
We have recorded improvements on our lighting with added cooling. As far as going that cold, it would be intriguing to test :cool:.

the driver is also important to keep cool.

Hold up..
Why don't you just take some of your lights outside and test them?
Surely you have all the equipment on hand and it is plenty cold outside where you are no?
Where about in Michigan are you at exactly, and do you do any brick and morter/in person business?
 
When I built my reservoir chiller I used Peltier chillers and aluminum water blocks made for CPU liquid cooling.

The water blocks would be about the right size for a COB, don't weigh much and would be easy to plumb into a parallel cooling circuit. I found inexpensive ones for about 9 bucks a piece.
51-uWn+TVqL._AC_.jpg
 
I can see using the water blocks instead of passive cooling, or fans to move the heat out of the growing area. An automotive heater core with a fan would probably suffice to dissipate the heat. It would add complexity, and cost as I'd want to add a quality pump, and fan. The cheap ones are too prone to failure. I'd also want to add thermal protection to the design, otherwise a cooling failure could easily take the COBs with it.

Instead of liquid cooling, even if a chiller were to be used, I'd use oversized heat sinks, and add some extra ventilation. This is the approach I took with my DIY eight COB array, and tent topper. Rather than taking air from the tent at only one point, I have three along the center line of the array. This has cooled the temperature of both the COBs and my tent by several degrees.

I think the cost of liquid cooling for the COBs/LEDs adds unnecessary complexity and cost to a grow light. The time required for a return on investment would probably exceed the life of the light.
 
Hold up..
Why don't you just take some of your lights outside and test them?
Surely you have all the equipment on hand and it is plenty cold outside where you are no?
Where about in Michigan are you at exactly, and do you do any brick and morter/in person business?

We have tested our lighting in multiple indoor environments. We can do an outdoor test and see the efficiency gains. It's getting cold here now. We are in Oakland County.

Currently, the brick and mortar is a goal for 2020. We try to keep our pricing low by running as lean as we can.
 
Back
Top Bottom