DiY LED - Cree CXA3070
- Thread starter SupraSPL
- Start date
What I love about the LED's is that the Vf is higher if it's cooler junction temp so you only need to worry about enough volts for the initial start!
SupraSPL
Well-Known Member
Good news retiring the HPS is paying off so far. In room #1 yield went up from 256g to 324g to 432g. These were all cuttings I have run many times and every one did much better than usual. Nug quality was off the chart. In the most recent 432g run it went like this
week1 - 600W HPS
week2 - 600W HPS
week3 - 1200W HPS
week4 - 1200W HPS
week5 - 1200W HPS
week6 - 600W HPS + 100W LED
week7 - 600W HPS + 200W LED
week8 - 200W LED
week9 - 200W LED
so average power was 744W or .58gr/W
week1 - 600W HPS
week2 - 600W HPS
week3 - 1200W HPS
week4 - 1200W HPS
week5 - 1200W HPS
week6 - 600W HPS + 100W LED
week7 - 600W HPS + 200W LED
week8 - 200W LED
week9 - 200W LED
so average power was 744W or .58gr/W
smokey the cat
Well-Known Member
I love that light cycle btw, that plant musta been shitting it's pants with terpenes at the constantly changing light stressors. I can totally see how hid might be the supplementary light of the future - maybe an hour or two at high noon?
Awesome to see you make so much progress in the garden with your lights. I've never thought of LEDs as "better" than HID before - it's always seemed like an obscenely expensive way of getting occasionally comparable performance. People are really going to start paying attention to this stuff I reckon.
Celebrate each tiny power bill with an epic session mate.
Awesome to see you make so much progress in the garden with your lights. I've never thought of LEDs as "better" than HID before - it's always seemed like an obscenely expensive way of getting occasionally comparable performance. People are really going to start paying attention to this stuff I reckon.
Celebrate each tiny power bill with an epic session mate.
SupraSPL
Well-Known Member
Finally got the Fasttech shipment with the 1A and 1.6A test drivers. I tested the 1A driver. The circuitry is setup differently than the 650mA driver. This one is regulated so from 22vF (min) to 36vF (max) it put outs a steady 890mA. If you go under it will just flicker, if you go over it will drop the mA to nothing very quickly. The max (36vF) is cutting it close for the CXA3070 (mine ran at 35.3vF) but other than that, the driver current is perfect for the CXA3070 and the efficiency of the driver is excellent 
21.75 vF
.89A
19.36W
21.5W in
90%
34.6vF
.887A
30.69W
34.5W in
90%
36vF
.887A
31.93W
35W in
91.2%
21.75 vF
.89A
19.36W
21.5W in
90%
34.6vF
.887A
30.69W
34.5W in
90%
36vF
.887A
31.93W
35W in
91.2%
smokey the cat
Well-Known Member
Those are very good numbers! Easily comparable performance with MeanWell.
What is your input voltage Supra? If I understand correctly (and that might be pushing it at times
) I think that drivers should be a bit more efficient on high voltage 220V than they will be on 120V.
What is your input voltage Supra? If I understand correctly (and that might be pushing it at times
) I think that drivers should be a bit more efficient on high voltage 220V than they will be on 120V.
SupraSPL
Well-Known Member
We get a pretty steady 120V. That is a good question, I have been under the impression that the closer the voltage of the source to the voltage of the LED string, the higher the efficiency. So for drivers with very high voltages I would expect 240V to be more efficient but I have not been able to test it because I don't have a precise way to measure the current on 240V. I always stick with low voltage drivers so I guess it wouldnt help in my case.
Organix420
Member
Thanks for all the info everyone - Supra I had noticed that chart in the datasheet but didn't understand it's meaning until you pointed it out. With the new driver I have two 3070s going at 35.3V at 800mA. Thanks again!
SupraSPL
Well-Known Member
Nice! Those will be pumping out some light at 800mA, glad you got it figured out.
updated numbers for the 3000K CXA3070 at 900mA:
4250 lumens
vF 35.72
32.15 Watts
132.2 lumens/watt
40.5% efficient
So considering those numbers, since my 700mA drivers are about 85% efficient and the 900mA drivers are 91% efficient, it is actually 1.4% more efficient to run at 900mA than 700mA.
updated numbers for the 3000K CXA3070 at 900mA:
4250 lumens
vF 35.72
32.15 Watts
132.2 lumens/watt
40.5% efficient
So considering those numbers, since my 700mA drivers are about 85% efficient and the 900mA drivers are 91% efficient, it is actually 1.4% more efficient to run at 900mA than 700mA.
smokey the cat
Well-Known Member
You're right. I had it completely backwards, ha! .
Here's a graph from dimming series LPF driver from MW.
Yours stack up for sure
I put in an order for 16 more
Mellodrama
Well-Known Member
Organix, could you please provide a link to that driver in Post #269? It looks like the driver is some sort of dual-output? I went back several pages and found some links you'd provided in Post #223, but neither one looks quite like the driver in your picture.
Oh, hey, I wanted to mention something to those of you who are experimenting with COBS. I've got some CPU coolers laying around, and have been trying to bolt CXA's on to them. These are "modern" CPU coolers with the copper heat tubes and copper base that stand off from the fins, not the older heatsinks that are just your basic slug of aluminum.
Anyway, I'd built one light that was functional but hadn't hung a fan on the cooler fins. I ran it for a few hours, hoping that passive cooling would be OK. One of the advantages to the heat tube style coolers is you can reach right in there, almost directly behind the COB, and put a finger on the heat tubes. They were hot to the touch. Not first degree burn hot, but it got my attention. Shut it down, strapped on a small PC fan, jury-rigged it to a cheap 6V wall wart, started it back up again. A few hours later the heat tubes felt inert. I couldn't detect any differential from ambient.
I'm using a high-quality thermal paste, and know enough about it from PC repair to apply a thin layer. You want just enough to fill in the imperfections, not create a blanket.
Anyway, moral of story in case you didn't already know: CPU coolers and passive cooling is probably not viable.
PetFlora
Well-Known Member
FYI Reflectors http://ledil.fi/node/2/?search=true&family=Lenina
SupraSPL
Well-Known Member
Ya CPU heatsinks passively might be good for a few watts but def no good for COB. I use on passively in my living room lamp but it is only dissipating 3 watts.
Thanks for the link PF. Interestingly the flat white reflector is 87% efficient. Easy enough to make one of those from a 2 liter bottle.
Thanks for the link PF. Interestingly the flat white reflector is 87% efficient. Easy enough to make one of those from a 2 liter bottle.
Organix420
Member
Mellodrama - those drivers that I linked to were input voltage of 200-240V so they really wouldn't work for me. So I found this MW MDR line and I'm using the 60w 48V one. I'm also using these 20-50w individual heatsinks.
Organix420
Member
Thanks Supra, I haven't run them for very long but I ran them this morning for a few minutes and one became detached from the heatsink so I turned it off and could feel that both the heatsink and led were very hot. I kind of figured I was going to need a fan since my driver is 60w and these heatsinks are made for 20-50w. Do you think a fan will be enough? I'm going to see if I can get one of those Temp guns today and do some experimenting. I'm still in the process of building my flower room so I have some time to 'dial in' the leds before I run them on a 12/12 cycle.
Mellodrama
Well-Known Member
Problem is they're expensive and hard to find. Other than that they're freakin' awesome.