DiY LED - Cree CXA3070

CobLedz

Member
Happy
Cool design... Are those cobs you have there, Chinese? They look like the ones I used to use for my business. If they are, you might want to check if they have the right color. I thought mine were 5000K (i bought them as 5000K) but at a qualitytest it appeared they were 7000K. Chinese quality, they said that it was normal. No Chinese leds for me anymore.
Happy, thanks for the comment. The COB LEDs are the cheap Chinese, Ebay items. How do you check, verify the color? I have been importing other hydro items from China for years and know that specs need to be double/triple checked. I do have a stellernet spectrometer that reads 200-1200nm. I have tested the 3000k and the spectrum "looks" great when compared to the chlorophyll absorption spectrum graphs. I intend to get all the cheap 100w ebay cob leds that I can get my hands on and test them all.

And yes, I can hear all the comments already, what a dummie I am, as I could have purchased a truck load of high end Cree products for what I spent with stellernet. There is a story too long to tell, maybe later. With the spectrometer tests I am interested in seeing if the cheap cn cob led match any of the spectrums of the most popular HID ie, Hortilux 1000 HPS and the SolisTek 1000k. I live in Denver and commercial growers all use the Hortilux or Plantmax exclusively. I am attempting to find off the shelf products that can compete.
 

happy75

Well-Known Member
Happy

Happy, thanks for the comment. The COB LEDs are the cheap Chinese, Ebay items. How do you check, verify the color? I have been importing other hydro items from China for years and know that specs need to be double/triple checked. I do have a stellernet spectrometer that reads 200-1200nm. I have tested the 3000k and the spectrum "looks" great when compared to the chlorophyll absorption spectrum graphs. I intend to get all the cheap 100w ebay cob leds that I can get my hands on and test them all.

And yes, I can hear all the comments already, what a dummie I am, as I could have purchased a truck load of high end Cree products for what I spent with stellernet. There is a story too long to tell, maybe later. With the spectrometer tests I am interested in seeing if the cheap cn cob led match any of the spectrums of the most popular HID ie, Hortilux 1000 HPS and the SolisTek 1000k. I live in Denver and commercial growers all use the Hortilux or Plantmax exclusively. I am attempting to find off the shelf products that can compete.
Cool...your own spectrometer, I would also want one to play with. I recently bought a spectrometer of public labs, low cost and hopefully it works a bit (i did not received him). If you find an efficient Chinese led with the right spectrum en enough par, you can use it for sure. I am curious if they match the Vero's or Cree's on par or efficiency and price per watt. Would be nice if you made your own topic and informed the rest of the interested people on your quest to find the best (off the shelf products). I will follow for sure.
 

SupraSPL

Well-Known Member
Bueno Time, awesome build ! Can't wait to see the buds. And thank you for the kill a watt numbers. Glad to verify that the LPC-60-1400 is .6 pf. If it were power factor corrected that would be handy for very large builds, good to know.

COBLEDz, great design I love it! ! Of course you know I am a huge fan of the Cree and I believe they are actually equal or cheaper than the Chinese COBs, but what youve got there will grow some bud :joint: If you are trying to estimate the color temp you can compare it to a warm white fluoro. If it looks more like a cool white fluoro then it would be way off spec for a 3000K
 
Last edited:

SupraSPL

Well-Known Member
Happy75, I made a comparison of the Cree versus Chinese COB on HolyGhosts thread. Something along these lines:

"The light you are looking at admits it is only 80lm/W in the 3000K. If you build your own you can get 128lm/W and the light will last indefinitely with no lumen depreciation.
CXA3070 $45
1.4A driver $11
fan/heatsink $10
total $66 for 52W

It may seem like we are just spending huge piles of cash on fancy LEDs but in truth the diy LED guys that are running at 1.4A are getting the best bang for the buck possible.

HolyGhost, if you went with the $40 setup, you would get 4000 lumens (allegedly) for $40. So that is 100lumens/$ initially (it may fade fast due to heat or even burn out). Also, I am suspicious whether it will actually output 4000 lumens initially, we have not seen a reliable spec sheet.

If you go with the $66 Cree setup, you get 6656 lumens (guaranteed minimum) so that is also 100lumens/$. It will last indefinitely and will not fade. It will also save some $ in your electric bill."
 

SupraSPL

Well-Known Member
It is worse than I thought (3200lumens/59W). Here is a PDF for the Epistar 50W COB. When I calculate the same way I do for the Cree, I get 54 lumens/W when driven at 1750mA (59 Watts dissipation), which is about 17% efficient. BUT that is at Ta 25C so who knows what kind of cooling they are assuming, it could be massive laboratory cooling. And finally, the PDF says "Tolerance of measurement of luminous intensity±15%"

It suffers very significant current droop, I would not run it harder than 700mA when used as a grow light. It would only dissipate about 17W at 700mA and would be about 109lm/W and 33% efficient.

If I understand correctly the Epistar cost $19 so if you run it at 700mA (33% efficient), that is $3.39/PAR W. For comparison, the Cree CXA3070 Z4 run at 2.1A (33.8% efficient) you pay $1.60/PAR W. So the Cree is less than half the price of the Epistar.
 
Last edited:

Bueno Time

Well-Known Member
Figured out a bit better system for hanging the modules in the tent.

Took a piece of scrap wood from the garage and cut it to length. Drilled 3 1/4" holes in each end and 6 1/4" holes in the center, these were drilled to run zip ties through the 1/4" holes. Then I measured and marked off from 4-12" in 1/2" increments from each end and drilled a pilot hole at the center of each 1/2" line. These are the markings and holes I can use to adjust the distance of the COB modules from each other and from the sides of the tent. The wood mounting rail (we will call it) is centered in the tent front to back.

Using small screw in hooks I can move to adjust the COBs side to side and I had picked up some #16 jack chain that is easily made to length and adjusted for height adjustments and also I could tilt the COBs at an angle if I ever needed to for any reason by having the chains shorter or longer on one side or the other. Pretty simple.

Still wanting to route the wiring better. Thinking of an idea for that now.



 

SupraSPL

Well-Known Member
Awesome rig BT! I use a similar idea. 2X3's running across the top of the tent with hooks screwed in at various intervals. Makes for cheap, sturdy and quickly adjustable connections. I make a lot of adjustments when bringing a new batch of ladies in and during the first few weeks of stretch. Because I mix so many varieties every run, for me it is easier to adjust the heatsinks to the canopy rather than worry about canopy management on the plants, other than a little pinching and bending.
 

SupraSPL

Well-Known Member
Final numbers on the 1st all LED batch .9 gr/W. That is closely trimmed, dried til they snap off the branch. I do expect that number to improve, but I am very happy with that result considering I was getting an average of .4gr/W with HPS. The bud quality was off the chart and my friends stampeded. The Purple Ace was more purple than usual and much more dense. The Kandy Kush actually came out more dank, which I did not know was possible.

I am planning on being less stingy with my soil and ferts, and upgrading power to the vegging side, since I now have more height to work with in the flower tents. The reason I was being stingy with my soil is because last year I was getting plants that were finishing too green, leafy and getting some foxtailing (with HPS). So I blamed it on overfert and reduced the soil from 19L/bucket to 14L/bucket, which did help.

Now with COBs, the ladies fade too early so I can add soil along with vegging them up bigger. I also bought a 5 gallon bucket of organic fish hydrolysate 2-4-1 for a boost. I can now expand the canopy thanks to the reduced heat and the flexibility of the LED modules.
 
Last edited:

SupraSPL

Well-Known Member
My g/kWh is probably unfair because I jump through hoops to keep power consumption low. I open and close windows depending on humidity and temp to reduce AC dehumidifier usage to zero most of the year. All four tents are connected by large passive airways so the whole system can be vented by one 6" Max Fan (60W) when temps are low. If temps rise I can increase power to 70 or 80W. If temps still to high I can kick on a second Max Fan (+60W). The downside is that large passive airways take up lots of space in the tent and even thought they are in a light trap shape, it allows a very slight glow from the room next to it to bounce off the floor, so you have to have herm resistant ladies.

Circulation fans in flower rooms are large ceiling fans on timers (55W). Circulation in veg is 19W fan on a timer and only runs a few times a day. Vegging room is only 70W LED and runs 20/4. Cloning box is 1W.

Your system beats the crap out of mine in terms of driver losses. Most of my drivers are 85% efficient and a bunch are 88-89% whereas yours is 93.5%. So the LEDs dissipate about 1075W and driver losses are about 150W. Active cooling cost me about 10W. Calls for a spreadsheet to calculate g/kWh
 
Last edited:
This is likely very true, not because of the hoops you jump through, but due to my novice gardening skills. We've all seen those girls of yours :bigjoint:

I'll likely be running into the 8 kwh/g barrier come harvest in the next week or two, but I knew this would happen - it was just a single abused auto and an in between with old CFLs I had laying around to tide me over until I have time to get the CXA arrays up. Not good, as hydro just went up here again. Can't put a price on growing your own, though :)

I suppose switching drivers wouldn't be a positive ROI from simply a hydro perspective at the present moment. Looks like 4 years to recoup the costs with some napkin math, assuming you average out to about .15$/kwh with delivery. I wonder if the extra umols without much consequence (running at say 1.05 vs .9) given HLGs increased efficiency would make a switch a positive economy?
Haha, I'm such a sadist, giving you yet another spreadsheet to compile :fire: LOL
 

SupraSPL

Well-Known Member
LMAO, yes if I had to redesign the system I would probably opt for the HLG-185H-C1050A. And thank you for bringing my attention to that driver in the first place! I have already redesigned and swapped out a bunch of LEDs and drivers once, so if I run out of projects to work on, it could happen again.

Bummer that costs are going up for you again. May come in handy having those dimmers. One thing is for sure, you are going to have some serious buds very shortly and you will be able to hang in there longer than anyone. Can't wait to see !
 
Last edited:

stardustsailor

Well-Known Member
Supra,I need your opinion/ 'verdict' ,please ...
Do you think any ..you know ...' damage ' ..is done in some dies inside that => chip #3 <=..
I drove it up to 2000mA ,but Tc once went over 45°C ....
But look at the nice ....discoloration of the thermal paste ....
Too much heat ...
What do you think ?
First thoughts,there ?
 

SupraSPL

Well-Known Member
The only color temps available are 3K 4K and 5K. We have mid bin and top bin in 3K, but only the lowest bins in 4K and 5K. They list a 3500K but none have been available.

The Cree XML2 4C 4500K U2 has better performance and a similar cost so if you are looking for a higher color temp for vegging I recommend those. I think there is a 4200K T6 bin also.
 
Top