DIY LED: Mixing COB and singles?

Playing tetris with parts-hunting on a budget, while woefully under-prepared for this endeavor, has indeed been frustrating... but thanks to you guys (especially one in particular), i'm almost there (to a sufficient starting point). Someday, i hope to build something elaborately tuned to my ideals... but i've got a LONG way to go before i should even attempt such a thing.

For now, i'm starting w/ just 2 cobs and some "ninjenuity." I will be red-heavy, and i hope i can wrangle the crop to avoid excess stretch... but i'm not that worried about it. I'm generally pretty good at improvising with things i can manipulate w/ my hands... (which unfortunately doesn't include electricity! lol)

I'm not worried about soldering; i've done it a few times (successfully), but since then, i've read quite a lot more about it, so my understanding of it is better than it was the last time i actually did it. I'm super meticulous when it comes to anything potentially dangerous or delicate. If i have to spend a whole day preparing to solder 4 points correctly, that's what i'll do. Getting important things right the first time, is often worth the effort. "Do it right or do it twice." "Measure twice; cut once." Some things don't provide retries. I like to have my whole plan loaded into my "RAM," and "cached," before i do anything. That's part of why i ask questions of apparently knowledgeable people, well in advance of executing any action. I might need to cross several bridges before arriving at the need for any particular answer, but i like to see as far ahead as possible, and have a plan, or multiple variations or alternatives, ready to go, long before i arrive upon the need for physically implementing any solution. Time-sensitive improvisation is chaotic and stressful for me, so i like to avoid it when/where possible, unless i know it's "just one of those things," in which you have to actually touch the problem before you can understand and implement the best available solution.

I'm pretty sure i'm going to add a 3rd (and/or 4th) "custom" circuit at some point, for spectrum supplement/enhancement... but as of now: "i'll cross that bridge when i get there."


Also: i really wish i could go all out on a massive heatsink... but shipping is a killer, so that's another bridge i'll have to cross later. And i'll be much closer to 100w than 200w.

I am looking for a definitive cree led light build youtube video. The one's I have found are not very comprehensive or are unrelated to our purposes. Any good links?

DarthBlazeAnthony said:

I am looking for a definitive cree led light build youtube video. The one's I have found are not very comprehensive or are unrelated to our purposes. Any good links?

Click to expand...

No video's like that specifically. You have to kinda extrapolate the info from the reef/flashlight builders. Is there something specific giving you troubles?

bicit said:

No video's like that specifically. You have to kinda extrapolate the info from the reef/flashlight builders. Is there something specific giving you troubles?

Click to expand...

everything...this is a daunting,challenging, and exciting project.How many watts would a cree light require to cover a 4x4 footprint?

I'm wondering if the OP found a source for only a few of the crees, I am,wanting around 50-100w as well but you need to place big orders at most places

DarthBlazeAnthony said:

everything...this is a daunting,challenging, and exciting project.How many watts would a cree light require to cover a 4x4 footprint?

Click to expand...

As per my research and the copious advice i've received, i'm betting 1x cxa3070 per square foot would be fantastic.

4x4 is roughly (over) 1m... 4x cxa3070 could be ~160-200+ watts (depending on how hard you drive them), of one of the most cost and energy efficient lamps available, and even 1+ gram per watt potential. With the right factors converging, i would not be surprised to see 4x cxa3070s pull a pound from a square meter. Haven't seen it happen yet, but IMO, it's a matter of time. In fact, i wouldn't be surprised to see a single 3070 pull a qp from a single plant.

vitamin_green_inc said:

I'm wondering if the OP found a source for only a few of the crees, I am,wanting around 50-100w as well but you need to place big orders at most places

Click to expand...

The OP did indeed find a couple Z4s.

Since then, i've seen at least one other source pop up for ABs, yesterday or so... keep your eyes peeled. ^^


I would still like to develop a custom spectrum approach, based on these cree cobs, but that will have to wait a while.

Dang, I need to! Just look on eBay or what?

You can get single AB and AD bin cxa3070 cobs.

your not going to get a lb from 4 cxa3070 sorry.

In a 4x4 tent i would go with 9 cobs for flowering.

purplegrower02 said:

You can get single AB and AD bin cxa3070 cobs.

your not going to get a lb from 4 cxa3070 sorry.

In a 4x4 tent i would go with 9 cobs for flowering.

Click to expand...

Well yeah, not if you just let 'em do their thing... but you could potentially fit 4 mainlined plants in a square meter, each in a large pot. "With the right factors converging," it could happen. But i would not expect such a thing without intensive training and essentially perfect conditions otherwise. I know the guy in the mainlining thread pulled 7+ zips from one plant (2 plants, 1 w/ 8 colas, 1 w/ 16... total was a lb, iirc). If you could fit 4 mainlined plants in a square meter (you could), and pull a qp from each... that would be a pound. As long as you don't have too much undergrowth getting insufficient light due to top canopy blockage, you "could" pull it off... but certainly not in a scenario without training, or any scenario which allows pests to slow growth and decrease yield. But sure, just tossing some things in a small space and letting them go, surely won't yield a pound! lol.

iirc, another well-known member here has done an experiment w/ 2x 3070s, and yielded ~1oz per plant, and over 1g/w. I don't think 112g is unheard of for a mainlined plant w/ multiple colas. 448g = 1lb, /4 = 112g. I bet you could fit 4 of those in a square meter (10.76 sq ft). You could give each one its own 15-20gal pot. A mainlined plant in its own large pot... can be quite the producer (assuming nothing goes wrong). Plus, w/ 4x 3070s, you could focus one directly on each plant, while positioning the colas toward the corners, so that the middle is more open, meaning each will still benefit from the other 3 cobs.

I'd love to be able to prove this is possible, but again, it will be a while before i'd be able to attempt it. But still, 4x 3070s, i think, could easily pull a qp from 1sqm, without much effort. Let's say 4 plants; that's a zip per plant, from having a dedicated 3070 for each, and "side light" from the other 3. And if you added extra veg time for training, you could potentially quadruple the production from each plant.

Sorry but you oversimplified it a bit.

4 3070s which is usually 208 watts and that is slightly running them hard will not produce 448grams.

I'm sorry but no one is going to pull 2.15gpw from any light source.

Here is about 1.25 to 1.5 per meter grown under Cycloptics luminare and Elite Agro bulbs done on a monthly perpetual rotation. Should do better with the CXA3070's.

purplegrower02 said:

Sorry but you oversimplified it a bit.

4 3070s which is usually 208 watts and that is slightly running them hard will not produce 448grams.

I'm sorry but no one is going to pull 2.15gpw from any light source.

Click to expand...

Light isn't the only factor.

Exhibit A: https://www.rollitup.org/t/the-main-lining-thread.542308/#post-7666123

2 plants, 24 colas, yielded just shy of 14oz. (didn't see whether he used a 400w hps, or whether these were outdoors...)

Pack 4 of those same plants into a square meter, using the same techniques... I'd bet money you'd yield a pound, even w/ only 4x cxa3070s. It's not just the lights though, it's also the extra 4-6 weeks of veg time for training. I'm sure you're correct that 4x 3070s will not yield a pound in a square meter, by simply throwing plants in a box and flipping the lights. However, if you were to Mainline 4 plants in that space, under 4x 3070s, i bet you could do it. But then we'd have to calculate how many extra kWh would be spent on veg, and whether it's "worth it." Maybe you could just make 2 lower yielding runs in about the same time it takes to once run mainlined 4 plants per meter.

This is all hypothetical theorizing, of course... but if you think about it, each of those 4 plants would be getting the majority of the total light, from 4x 3070s. (obviously, 9x would be much better!)

4x mainlined plants with 8x colas each, is 32 colas; a potential 33% improvement over 24 (because 8 is 1/3rd of 24). Imagine each of those 32 colas, getting ~200w worth of 4x 3070 photons. Even at 1g/w, you're almost halfway there under Normal grow circumstances. What's 14oz + 14/3oz? (14oz plus one third of 14oz)

What if you could pull "a little over a half oz" from each cola? (14/24 = "a little over a half")

32 halves is what?

I never implied that 4x 3070s would magically create a pound from a sq meter by not doing anything else... i clearly stated "with the right factors converging..." meaning "it's not only about the lamps."

You can definitely fit an 8 cola mainlined plant in a 20"x20" space, and you can definitely fit 4 of those in a square meter.

I'd love to see an experiment with a single plant under a single 3070, with all other relevant factors maxed and controlled.

But i'm sure cree will be releasing something even better, at some point in the not too distant future.

I mainline my plants right now.

I'm telling you it will not happen. You do not have enough light.

you can't just say we'll 1 plant yields 5 ozs so cram 4 in there and I'll have 20ozs. No it does not work like that

It's doesn't matter if you have one plant in a Scrog. Or mainlined plants or a sog your going to yield pretty much the same given the same space and light.

well i realize space is one of the relevant factors, but i still think 8 colas can be constrained to 25% of a square meter, without blocking the other 3 plants (assuming all 4 are arranged the same way).

I'm not just arbitrarily adding numbers.

How much do you think 1 mainlined plant w/ 8 colas could yield, in a 20"x20" area, under 4x 3070s, if you flower them at the right time so they finish at about a meter tall? (assume no detrimental factors aside from the 20"x20" area)

Now, why would it be incorrect, to assume that if you had 4 of those same plants, and 4 of those 20"x20" areas, adjacent to each other... it would yield nearly 1:1 increase?

Any light that is not blocked by other plants, is used by each plant. I realize there are intensity curves, so maybe it would be more like having those 4 plants under only 3x 3070s... but still.

I would love to test it, but i only have about half the area and half the lights, so it would be more like 2 plants under 2 lights, which i would predict would potentially yield less than half as well as 4 plants under 4 lights (because regardless of the number of plants and space, any number of plants will only receive a maximum of 2x 3070s worth of light). If i meet or exceed 1g/w, i will pull about a qp. If i had twice as many plants under twice as much light... each plant would be receiving twice as much light! "Double" might be a bit overly optimistic, and while i would not necessarily "expect" such a thing, as i stated previously: "i would not be surprised." I would be quite pleased, however.

I'm just saying: it's possible to manipulate your space and light positioning in such a way as to maximize effectiveness and yield. "A pound in a cubic meter" would be pretty amazing... but it would take significantly more time, effort, and micromanagement, not to mention a lack of setbacks.

I suppose i should clarify that any surprise i would experience at such a yield, would be focused on the skillful manipulation of space, and lack of setbacks, rather than the lamps, in such a scenario; the cultivation methods and maintenance of a maximally conducive environment, not the cobs themselves. The cobs are already "surprisingly impressive" in their own right. What a skilled grower can do with them, is the real magical part.


Anyway, i don't intend for this thread to be an argument over whether 4x 3070s can yield 448gpm. Your experience likely trumps my hypotheses, but i prefer to think the null hypothesis has not yet been proven. The whole point of this thread was to discuss the minimum amount and type of additional lamps to most effectively supplement the spectrum of a cob (and more specifically, the 3070s).

I would like to combine 2x 3070s w/ 1x Lumeon M, some 405nm, some 430nm, some 660nm (and maybe some 670nm), and eventually, add some 730nm timed for the correct number of minutes at the end of each day cycle... and then make multiple units of that lamp (to create a spectrum/efficiency maximized single unit solution; all the stuff we want, without any "waste," such as an excess of greens), to the limit of effectiveness for a given space. But i sort of jumped the gun a bit, because that's just not in the budget at this time.

I would absolutely love to get my hands on some 90cri 3000k 3590s... (check out the spectral curve on those! similar w/ the 3070s!) and then fill out the spectrum by surrounding them w/ the highest available efficiency stars of deep red and deep blue, so that the output is "balanced" enough to be an all purpose lamp with predictable performance. I would like it to have minimal "footprint" and be actively cooled (which branches into the idea of dual purpose fans, which act as both filters and coolers). Build something like that and you could sell a truckload for a mint. ^^ (not to mention having maximized your own setup) You could potentially even build a water-cooled version, make the cobs more modular, and drive them harder. (grow lamp overclocking?? lol)

MR-X 2.14 grams per watt! And this was a long time age.
https://www.icmag.com/ic/showthread.php?t=215147

I'm running 480 watts in a 4x4 and i yeild just over a lb so i really don't care how much math and thinking you want to put into the equation I'm telling you it is not possible to yeild the same amount that i am in a smaller area with half the amount of lights. And this grow is using the latest in available led technology all top bins units and the most effecient drivers.

hypothesis and real world are two completely different animals.

Positivity said:

Sorry if I didn't expand but time is usually short. Part hunting isn't the fastest thing to do.

Seems you are getting complicated too quickly.

Just build a light based on the cxa 3070 is my suggestion. Never mind blues and violets and reds...save that for a future build which you will undoubtedly do if you get the first done successfully

200w? Alright, something to work with..

Easy..get a 200w driver. Constant current either meanwell or inventronics. There's other brands if you want to search and experiment.


Why 700ma? So you can put other leds into the mix. Otherwise it will be a cheaper build driving the cxas harder and leaving other colors out. But 700ma will be very efficient, nothing wrong with that

How much cxas will fit at 700ma? Each pulling about 30w..

You can put 6 and end up with 180 watts of led power. Hooked up to a killawatt to see the draw you would be running right around 200w because of driver inefficiencys.....perfect.

You can put 5 and end up at 150w cxa power. Leaving 30w for whatever other leds you feel like. Again ending up with max of 200w...perfect.

First light? I say just put all the power into 6 cxas. Probably work better than any other combo anyways. I use reds but I'm simply experimenting and it's about my 10th light I've built.

So....

Dimmable 200w inventronics driver...http://www.futurelightingsolutions.com/en/Technologies/Semiconductors/Lighting-Solutions/LED-Driver-Modules/Pages/7018238-EUC-200S070DT.aspx?ManufacturerName=INVENTRONICS&isFLS=true&IM=0

Not dimmable 200w inventronics and not in stock either..but order able.....http://www.futurelightingsolutions.com/en/Technologies/Semiconductors/Lighting-Solutions/LED-Driver-Modules/Pages/9009943-EUC-200S070ST.aspx?ManufacturerName=INVENTRONICS&isFLS=true&IM=0


Others are much better versed in meanwell driver specs and where to buy them. Also the sticky has lots of info...https://www.rollitup.org/t/diy-leds-how-to-power-them.801554/unread



Heatsink...I'd use at least a 2'L but 2.5'L would be even better for a 36" area.
http://www.heatsinkusa.com

Seperate power supply for fans...http://www.amazon.com/Universal-Regulated-Switching-Power-Supply/dp/B002FH54L6/ref=sr_1_3?ie=UTF8&qid=1406899855&sr=8-3&keywords=12v dc 2a


That's the meat of it

But then there's other things. If you don't use holders and your soldering isn't very good...there's a good chance you'll shock yourself. The 200w high voltage driver will sting.. So in that case...read up..practice on a simple build like my first post showed or one that others have shared in the various threads using low voltage drivers and CPU coolers.

One driver....one led....Build two or three.

But really it's about what you want. Find one you like, copy it and build it. You want exact answers on my drivers or parts in my lights just ask..

Click to expand...

Are you referring to running the COBs in series off of the one 200W supply?

Yes. Could parallel also but I like series.

Parallel would drive the current down pretty low, why would you want that?

Parallel with a constant voltage driver instead of a constant current..

But if one goes out..voltage jumps..then fuses or zeners. I like constant current so far