DIY LED 220w Cree XTE + Philips ES

Not sure if you explained it anywhere in here but I thought I might add something. The reasoning behind polishing the heatsink. The more reflective the metal the more it'll relect the heat from lighting. IE it's helping to keep the heatsink from getting hot from the actual light and not just operating heat. Also of course it spreads any light back better! Nice touch!!!

Originally KNNA recommended 100cm2 of heatsink per dissipation watt and we were expecting max heatsink temps of 50c. That was when we were using the 26% lamp. Those numbers worked as expected so now I am aiming for much lower heatsink temps ~35c using the same amount of surface area 100cm2/watt.

For the new lamps I estimated and averaged the efficiency the LEDs ~41% (@50c with all LEDs operating at 700mA).
old lamp: 1 watt = .26w photons .74w heat
new lamp: 1 watt = .41w photons .59w heat
59/74= 20% less heat so I assumed we can use 20% less heatsink to achieve the same temps or 20% more watts on a given heatsink.

With the new lamps at 100cm2/watt I have measured max heatsink temps ~37c (in practice as low as 32c thanks to circulation fan). That is good evidence that the radiometric efficiency has improved.

Yes I run the XTEs at 700mA. Originally I planned on running them at 1050mA but after looking at the numbers it seemed like 700mA was a good compromise. Running them hard saves you on the up front cost of LEDs but XTEs are relatively cheap. Running them softer saves you money on electricity, heatsinks and drivers.

If I had better access to adjustable drivers I would probably go with 600mA for the XTE but 700mA drivers are cheap and easy to find. Other than the ones I got from KNNA, the only adjustable driver that caught my eye is the inventronics which is 3X the price. Maybe next lamp I will give it a shot.

Cree XTE R3 (3200K 7A tint)
350mA 41.5%
500mA 38.3%
600mA 36.8%
700mA 34.6%
1050mA 28.6%
1500mA 25.5%

The polished heatsink does reflect some stray light into the canopy and it did look very nice until I marred it up with kapton tape but looks get sacrificed for utility in this case.

Lapping the heatsinks was a good bit of work but I was motivated to do it to improve heat transfer between the stars and heatsink surface. There are tons of machine marks running across all heatsinks that can be smoothed away. While blue/white LEDs lose efficiency as current rises, reds lose efficiency as junction temp rises. The faster you can get the heat away from the emitters the more bud you will get.

To keep the thermal path as thin and efficient as possible I used a top quality thermal paste Prolimatech PK2 rather than an adhesive. With a thin layer of PK2 the stars stick to the heatsink very well but the kapton tape is there to make sure they stay put for the life of the lamp.

This thermal path is greatly improved over the copper/kapton tape system we were using. Not long before his health problems began troubling him KNNA determined that out copper/kapton thermal resistance was much higher than he expected but I never got the chance to ask him about it. I was able to find super thin 1mil kapton tape designed for thermal transfer but it is not readily available.

Ah shit didn't know Knna had such health issues.....do you still communicate with him?

hey Supra,

thanks for sharing bro, and thanks for the very useful experiences, very nice project!

Supra,

Yes it could be a good idea to use only one royal blue/ 2XTE / 4 oslon 660 or 1RB/ 3XTE / 3 oslon 660

• Maybe it would be best to use the XTE to provide the 630nm rather than the oslon JS. Based on the datasheet the JS bin is a minimum of 23.75% efficient @700mA @60c versus the 2T 660 and the XTE R2 which are both about 33-34% efficient under the same conditions.​
  
  
  
  

Click to expand...

how can you know that the 625nm JS is 23,75% efficiency@700mA??? the flux is in lm on the datasheet, how can you convert lumens to mW?
same question for the warmwhite XTE, how can you convert to mW and find the efficiency? this is the question i'm trying to resolve since 18 mounths . Maybe you are using a software? from knna??

• Originally KNNA recommended 100cm2 of heatsink per dissipation watt and we were expecting max heatsink temps of 50c. That was when we were using the 26% lamp. Those numbers worked as expected so now I am aiming for much lower heatsink temps ~35c using the same amount of surface area 100cm2/watt.​
  
  
  
  

Click to expand...

It is a very empirical method to find the size of the heatsink, and in most cases it works, especially with U aluminium.
But cooling is related to heatsink shape too, heatsink position, and if it is raw aluminium or black anodised.

For raw anodised heatsink, fins shine the one on the other one, and this phenomenon is less present with black anodised heatsink, so black heatsinks are more efficient with passive cooling. If you add a small fan on a raw aluminium heatsink it is going to improve a lot the cooling. With a fan, black anodisation or oxydation does not play anymore its role.

There is a simple method to measure the junction temperatures to verify the heatsink, with a multimeter and a small temperature probe, but you need a multimeter with temperature function.
Measure the Tcase of the LED, here is the calculation.



Maybe you know all this, it is just my experience feedback, maybe it could be useful for somebody.

In order to estimate the radiometric efficiency at operating conditions I use the data provided by the manufacturer spec sheets. They give an output at 350mA or 700mA and I extrapolate the output at other currents based on the output graphs. They also provide how much the vF will shift at 50c and how much the output will fall off at 50c, which I take into account. I use a spreadsheet to make the calculations easier and quicker.

Then I calculate the dissipation wattage @ 50c to arrive at a lumens/watt figure. Next you have to estimate the LER. For warm whites I estimate the LER to be approx 310lm. Using the XT-e warm white as an example, at 700mA it is 107lm/watt. Assuming an LER of 310 we get a radiometric efficiency of 34.5%

The weak points in this estimation are that we are relying on the accuracy of the manufacturers graph (I always use the stated minimum for a given bin). Even weaker is the estimation of LER which varies based on the color temp and SPD. I always use the same method so at the very least we get a qualitative comparison to the old lamps and between warm whites.

Thank you for the heatsink info, very helpful. Does the anodized layer cause additional thermal resistance to the mounting surface? Worth polishing anyway so you can have your cake and eat it to. I have a fluke 115 but unfortunately it does not have the temp function.

thanks for the infos about the LER, i didn't know it.

No I don't think that the anodized layer will cause additional thermal resistance to the mounting surface because this layer is very very thin. I compared the temperatures (T solder point) for the same lamps, one black anodised and one raw aluminium, with no air flow, passive cooling only, the black anodisation can give 20°C less, with the same heatsink sizes, same mounting, same thermal interface...
With air flow, the difference of temperature is only 3-4°C.

With active cooling, the thinner the face of the heatsink the better. If the face of the heatsink is thick, the heat will not dissipate fast enough to cool the led even though they are being actively cooled.

This is best understood with computer processor chips. Take a look at the heatsinks used, they are paper thin but have many more cooling fins. Also, the airflow should be flowing from the top of the fins, into the heatsink and out the sides and not being sucked in through the side and out the top of the sink.

With passive cooling the thickness does matter and overall the more mass/surface area up to a point is going to cool your leds enough, but active cooling, if done properly will give you the best longevity for your led.

^^ Well, all bets are off now that CREE is introducing its' new 'single-die' LED said to have high lumen output at .... wait for it.... 85*s C! Oh snap! Could this be the end of active cooling? www.cree.com

Wonder if they will include some colors with that, or if it will be white only.

PetFlora I think you are referring to the XPG-2 and XTE. CREE managed to improve the efficiency significantly and decided to bin them at 85c @ 350mA. Philips decided to bin their Rebel ES at 25c @ 700mA. A more useful approach would be to bin them at 85c @ 700mA.

Despite the improvements in efficiency and changes in binning, the same old principles apply to blue/whites LEDs, their output diminishes significantly as the current rises whereas for red leds their output diminishes significantly as junction temp rises.

So for general lighting they can get away smaller heatsinks and higher junction temps but for grow lights that depend heavily on red emitters, junction temp is still critical to efficiency.

The new CREE MK-R LEDs are binned at 700mA @ 85c. They are 9% brighter at 50c. At 700mA they dissipate 8.2 watts. The have a more focused angle than the XT-E. Very similar efficiency as the XT-E and XM-L. I calculated the efficiency of the 3000K warm whites to be about 36% at 700mA @ 50c.

The SPD graph is more specific than the one provided for the XPG and XTE so we can better estimate the percentage of blue.


Cutter has the MKR in 7B (3000K) with top bin available G4 (36% efficiency) for $9. They dissipate 8 watts each versus the XTE 7A (3000k) top bin R3 (34%) which is $3.25 and dissipates 2 watts so you could say the MKR is a better deal. It would penetrate deeper for those who prefer taller plants.

You've been running the XTE for a while now Supra, no complaints? Would you choose them over the Oslon WWs? (both running at 700ma)

Does your answer to the above change if junction temperature isnt an issue (active cooled)?

Hello SB. As far as I know it was only recently that OSLON SSL became available in the Americas so I hadn't considered them. They are selling for $5 whereas we can source the XTE for $3.25 from Cutter. Also I cannot get a specific bin of OSLON. I would end up with a mystery bin which would make it even more difficult to estimate the efficiency and SPD. Cutter offers specific bins of the XTE (even though it is sometimes sketchy deciphering their typos/descriptions). Finally, the XTE comes professionally mounted on stars which is very convenient for hand mounting, especially the 20mm stars.

I did a quick study on the OSLON SSL warm white based on this datasheet. Is that the same model you are asking about? It includes 3 bins of 3000k warm white. If we assume you get the lowest bin it works out 20% minimum radiometric efficiency @700mA@50c. If we assume you got the top bin it is 24% minimum radiometric efficiency. Compare that with the XTE @ 35% and being cheaper we have a winner.

Before I answer your other question I should expand on the passively cooled design. Depending on the size of the module, active cooling can reduce the efficiency of the system by 10% because of the electrical consumption of the fans. Passively cooled units can benefit from circulation fans that are already running in the grow room. Passive units are silent and a simple design less prone to failure over the long term. It is for these reasons that KNNA has encouraged us to use passive cooling. The downsides being a higher initial cost and heavier modules.

So to answer your question, I do not consider passive cooling a limitation so it would not affect my choice of emitters. To go further I suspect that most active cooled LED modules will have higher junction temps than my passively cooled units because I oversized them for a target Tj of 50c.

So far no complaints about the XTE in practice. Very impressive results in veg. Jury is still out on the bloom side. Will get some pics up soon.

Yes, pics piCS PICS! I can't wait to see how your plants are responding to the light you made.

Any of you want a good laugh, and to see how awesome some of the forum users are, come check out my latest thread!

https://www.rollitup.org/seed-strain-reviews/631747-serious-seeds-ak47-regular-going.html

SupraSPL said:

fast finishing, icy, dense, easy to trim nugs
View attachment 2427506

Click to expand...

Supra, what strain is this photo that you posted in post #1.

Supra,

I've been looking for the r3 bin XTE on cutter, but can only find the r2 bin (7a tint as well). Am i just being a dumbass or are they gone?