DiY LED - Cree CXA3070

I found a YouTube video which explains everything. Check it out. Best feature of all - no soldering to the COB. The holders take care of the electrical connection.

They sell for about $1.65.

No worries about damaging your new COB from soldering, and you can use real thermal paste instead of thermal adhesive. I'm sold.

EDIT: I'm re-reading this thread from the beginning. ellydee mentioned the Ideal COB holders way back at post #65. The first time I read this thread I blew right on past that post. Here I was goin' on like I'd discovered Eldorado and elly'd already brought it up a week ago...

It can be tricky keeping it all straight at first but it will become second nature. You are correct I was referring to the retail availability of the 2700K in the CXA3070.

I had the same concerns as oXCoffee but thanks for posting that vid the cob holder is very cool!

Cree claims 276 lumens/watt in a 4000K 350mA lab test. Apparently LEDs will continue to push the envelope.

http://www.ledsmagazine.com/articles/2013/02/cree-announces-a-new-laboratory-led-efficacy-milestone-at-sil.html

The new batch at 4 weeks. Not much interesting to see yet. 7 of these are new varieties so I have no idea what to expect from them. So far the Darkstar seems to have the biggest buds and the OG18 is seeming underwhelming.


This batch is HPS bud finishing under LED. My grows always look like a trainwreck at this stage because they start to flop over and I bend them to try and keep an efficient canopy.

SupraSPL said:

My grows always look like a trainwreck at this stage...

Click to expand...

Aesthetics are secondary. You're experimenting.

Geez, those heatsinks are massive! I know you mounted them securely, but I'd still want a hardhat any time I was in the drop zone!

Such colorful splendor there Supra. Can't wait to see your chop-day photos!

Gaius said:

For my CXA3070s I just ordered some of these Meanwell LPC-60-1400 constant current drivers based on a recommendation from Guod.

Click to expand...

I'm not questioning your judgement by any means, just confused and wondering what I'm missing. According to the Jameco link you provided, those MW LPC-60-1400's output 9V DC. That's what it looks like to me.

The Cree .pdf for the CXA3070 (page 2) states typical voltage is 38.5V (@ 1925mA, Tj=85 degrees C).

So to this newb it looks like those drivers weren't the right choice.

OK, EDIT, in the second column on the Jameco site it says something about "Additional Output Voltages - 10 to 42". The spec sheet also indicates voltage is variable. I don't get it.

Constant output current, but variable output voltage of 9-42V. See Guod's graphs he linked earlier for a visual representation of this. The drivers with the 3070s should hover pretty close to their ideal 38.5V.

Hard had would not be a bad idea, if for no other reason than a good potential for bumping my head on them once in awhile. They are hung from 2X4' with stainless hooks and zinc chain. If for some reason a hook or chain let go, the other side should keep it from falling.

Mello I am not sure why it says primary voltage is 9V but the voltage range is 9-42V so it should be good to go

Tnx gaius, so far I can say the LEDs definitely stepped up the frosting in this batch. I have run these ladies many times (G13 Sour Jack, SSSDH Shire, GH Lemon Skunk, Qleaner, CC Buddha Tahoe OG) and they have never looked better in terms of frosting. Tahoe OG already got the chop

OK, thanks, it's important to clear up confusion ASAP or I start losing it. Really quite amazing to me that the driver takes care of supplying proper voltage on its own.

Gaius, I did go back to the other thread that you linked to, but of course that was after I posted here l appreciate hearing from you guys directly because that's more reassuring.

SupraSPL said:

Cree claims 276 lumens/watt in a 4000K 350mA lab test. Apparently LEDs will continue to push the envelope.

http://www.ledsmagazine.com/articles/2013/02/cree-announces-a-new-laboratory-led-efficacy-milestone-at-sil.html

Click to expand...

And whats that, like high 70's low 80's in efficacy %? Hello 3+ GPW in the future ..only takes 10 to 15 years at this rate to get those buggers to the market..

Great post SupraSPL; Thanks for all the information. So currently I've been growing(veg and flower) out of a small space using a 1000w digital ballast that I've dialed down to 60%. I've recently cleared out a 5ft x 9ft space and want to expand my setup to using the small space for veg and make the larger space for flower. With energy cost's being what they are I really want to do LEDs for the larger space but still have a lot to learn about LED setups so I'm hoping someone on this thread can help me out. Since the cxa3590 still is unattainable unless you want to buy at least 40 I'm thinking that your setup of cxa3070s will be a good place to start getting familiar with these LED setups. I've been looking at the datasheet for the cxa3070 and have a couple of questions:
1. the datasheet keeps referencing 1925mA - is this the optimal input amps at 38.5V?
2. how does wattage figure in; is it just a case of heat dissipation (it is not mentioned in the datasheet)?
3. what would be a good flower setup for a 25sqft space?

3790's?? Do you mean 3070?

Thanks Mellodrama good catch. I 've updated the original post it was definitely a stoner moment.

Hey, no prob, I was upset that maybe you'd found some brand new COB's!!

I just ordered a coupla 3070's from Digikey today. The exact same ones as in Post #1. I'm staying between the lines, ordering stuff that either Supra or Gaius specified.

Except I also ordered a few of those Ideal COB CXA30 holders from Newark. They were gonna charge me way too much for shipping either way, so I upped the order from 6 to 10. That'll show 'em. Don't know what I'll do with ten.

I've got several heat-tube style CPU coolers laying around. Think I'll try those instead of buying aluminum. I think big aluminum heatsinks would be cooler (no pun intended), but gonna go with low-cost for right now.

SupraSPL said:

View attachment 2984399

Click to expand...

SupraSPL said:

View attachment 2983044

Click to expand...

You have two different efficiencies listed for the cxa 3070 in post 54 and 59.

Which one is your correct estimation based on your extrapolation?

I ask because I am ready to order my modules. If the efficiency is really that high at 350 ma I will run 32 on meanwell hlg .7s (.35 to .77 adjustable). If it is only marginally better as per the first excel capture, then I will run 20 modules on meanwell hlg 1.05s (.525 to 1.07 adjustable). I already have a pretty big gain in efficiency using the hlgs at 92%+, vs individual drivers at about 85%. I think given this I can lose a couple % from the COBs and keep the BOM down.

Still haven't dialed in the ratio, but will likely be 6:2 at .7 (4 groups of or 4:1 at 1.05 (4 groups of 5).

Appreciate it!

RE Organix420

1925ma is harder than I would suggest running the 3070 for our purposes (long hours of use). It is not that 1925mA would not grow bud, just that it may be a false economy due to a significant efficiency penalty (current droop). I run at 700mA but 1A is great and 1.4A is also popular. I am at a friends but later I will post a chart that shows the change in efficiency and wattage based on qhich current you choose to run at.

The lowest hanging fruit IMO is led for vegging. Separate vegging area is a great idea and will supercharge your yield over time. You could veg with as little as 30W LED. In a 5x5 flowering canopy Id use 600HPS + 300 LED or 600W LED.

The higher newer number is the more accurate. You are planning on 350ma? I appreciate that you are more grams\W "visionary" than me .
In the first chart I was just using the PDF to estimate in the second chart I loaded the chart into MS paint and zoomed in and counted the pixels.

Opinion on efficacy driving 3070's at 1800 mA? 3050 's driven at 1500 mA working great!

SupraSPL said:

The higher newer number is the more accurate. You are planning on 350ma? I appreciate that you are more grams\W "visionary" than me

Click to expand...

Still unsure, but it's very possible I found the COBs for 33$ a piece in multiples of 20, so that's tempting. Just to confirm the estimated efficiency for the 3K 3070 is about 47, 42, and 39 % at .35, .7, and 1.05 respectively?

Either way I would run from .35 to .7, or .525 to 1.05 with 32 or 20 emitters, increasing the output wattage slowly from the end of veg (400 W) until peaking at close to 800 watts closer to the end of flower. I like the idea of 1 driver per array per plant. I also like the idea of different temperatures on different drivers. The true cost of DIY flexibility seems to be indecision caused by the infinite possibilities!

I will run a 5 x 5 with 4 trees personal grow. It will also be my veg chamber. Each of the 4 arrays will be centered over a plant.

Time to whip up a quick spreadsheet to help reach my final decision!

I know you mentioned the KNNA values of 20% for blue during flower as potentially idea. This is important to me as I plan no other supplemental lighting aside from the 3070... Looking at the umols/s/nm for these emitters posted in the cre cxa analysis thread (thank you Flux) it would appear that I need a way to calculate the area under the curve. I believe I read that you mentioned the 3K has about 10% blue. How did you calculate this? I am attempting to calculate the 3K:5K (or even 3K:4K:5K) ratio I need to arrive at to fulfill the spectral requirements through veg and bloom based on the photosynthesis charts I have compiled.

Also, how is your build going? I agree that lapping is difficult given the size of these sinks and is probably unnecessary with the thermal mass / area you are using per watt. Were you lapping with the 3M paper tapped to a piece of glass? Also, I believe you mentioned polishing. I typically forgo this step, as it is my belief that these buffing compounds leave a residue that may actually impede thermal conductivity vs the freshly lapped surface itself.

Cheers!