Lets create the most effective cob light!

Yes, i might be using it/them.

Edit: These Cobs are now built Jump to post #78 for final reult.

The stronger Cobs are very efficient lumen/watt-wise, but a downside is that you loose some of that efficiency due to the extra distance needed compared to multi 3w grow lights.

My first thought was to use run several smaller cobs, that also are pretty efficient, but then reading a thread on another forum i'm realizing that running CXA 3070 on a very low current gives tremendous l/w-outputs.


If you run the 5000k version on 350mA you get an output of 166 lumen per watt! It will only be 12 watts but still give you almost 2000 lumen. If you use an 18w driver you could also throw in a couple of 630, 660's or whatever you think could be beneficial for your grow situation.. Personally i would go with 630's cause its the easiest way to up the par, and also has proven to work well for a lot of people in real life.

Or maybe going with the 3000 k version and get outputs of 155 w is a better option?

Pros:

#Super efficient, pays back on electricity bill, environment and karma

#Several smaller lights gives better spread and utilization of light emitted.

#less heat and need for active cooling.

#longer (eternal?) lifetime of cob.

Cons

#Expensive to buy parts

#Bigger effort to build, and more hassle to mount if built on several heatsinks, which is necessary to get the advantages of better spread - unless you have a very big heatsink

I'm planning to build several units, with the advantage of having more flexibility, more even footprint, and having lights closer to plants then fewer stronger lights would let me do.

I would like to use a constant current driver with a little bit of extra volts so that i have the ability to tweak the spectrum with a few 3-watters.

I need a bit of help though, since my electrical skills are very rudimentary...

Would this driver work?
http://www.ebay.com/itm/321367204370?_trksid=p2055119.m1438.l2649&ssPageName=STRK:MEBIDX:IT

I hope this project could interest anyone. And i also hope that i've written understandable english, cause it's 3 in the morning and english is not my native language.

All the best!

edit: Most efficient is what i'm looking for. They should be effective though, if you only use the right amount of them!

damn, I am pulled in now!!!

I thought the XM-L2 did a good job in flower / veg.. but these..... HOLY COW


problem being pricing.. these are much more expensive then near any solution. that light would be one hell of an expensive light, and tell me I know...... I think the underestimation .. which could be made cheap, but I don't roll that way is you think, man this is going to be cheap... didn't add in costs of drivers, wiring, heatsink, fan, and all sorts of misc. crap.

thanks coco / supra for the chart, as the thing I never was able to do was calculate power / lumen per watt and such... so thank you amigos... I'll be lurking, and maybe trading in the XM-l2 bars... all 8 !

just think what the future holds for running veg tents, I'm running 18.5 watts a square foot with EXCELLENT growth rate and node spacing and such.. I couldn't imagine like 8 watts a square foot, with excellent growth rates and super fast vegging.... I've done tests once the plant is established to pull down watts, but haven't done enough to talk on it yet as to what power level and what growth I'm getting. well here we come, the future is looking bright as hell





driver looks like it will work, but tired as hell, and I don't have the chart to see what levels it's going to actual draw, that's how I always did it anyways, looking at the cree pamphlet to see the mA / voltage chart.. But I would ell you straight up, if I built this, it's going to be using meanwell dc drivers, or those tiny little mini drivers... there cool as hell, easy to replace, super efficient, AC-DC driver is very cheap for how much power they can output and very very future flexible in my eyes to the ease of the small driver and switch out is so simple.. also the driver on these will most likely be cooled by fan ( some models don't ) and the driver box is very small in comparison of watts.. I run a setup with two bars and I find it very nice.. was going to switch all my bars to these type of driver setup, but never followed through. I like playing around with all the new tech, and can tell you I don't know why everyone hasn't switched to these yet... basically follow the aquarium users. .they all switched over, so there always ahead of us, and smarter in a way.. well technology and math way.. not relaxed way..

Thank you for taking a look at this habeeb! I really lack some of the electrical knowledge needed for this kind of projekt! I know It's borderline stupid to make this kind of expensive investments considering the pace that leds and cobs are developing right now, but i always found a huge satisfaction in optimizing stuff.

If money didn't matter there also is some new bins 3070 ab/bb that gives you around 10000 L @1925. They are very hard ti get your hands on, but would probably give around 180 l/w if driven at 350mA!

A cheeper way to go could be to build this multi cob setups with the older and cheeper cxa2520-2530 cobs. the higher bin 5000 k give you up to 150 l/w at @550mA, and probably even higher driven at 350mA. I haven't found any exact numbers of that though.

PetFlora said:

I am very interested on COBs but smaller ones (~ 30w) with parabolic reflectors, and run multiples to maximize canopy coverage/penetration

Currently using a SPYDR 600 which employes a ton of 3w diodes. You can select your own spectral blend

Now in 44th day of flower

Pretty awesome

Come by to see pics

Click to expand...

The BML lights are very convenient (mix&match) and probably also effectiv/efficient because of their awesome spread.

But now i'm wanting to build my own small cob panel and reach 150-160 l/w with a good spectrum. It might be more expensive and will of course be more hassle.

Petflora, the reflectors will let you direct the light in a certain direction, but if you could, wouldn't it be better to find out the perfect density of cobs and let the spots overlap each other?

I'm very interested in your thoughts about cobs! Iv'e been thinking about running 40-50 watt cobs @ 12,5- 25 watts for optimal efficiency. And using 4-5000k with additional smaller red diodes because of the high efficiency of the cooler leds. But pretty unsure about this and still considering warmer spectrums.

What cobs would you opt for in the 30 watt region?

sounds like a great plan, cocogrower! I have about 150 usd which I gonna use to small scale test some leds soon. I will get either multichip boards with cree leds (can be found in 10-30-60-100W size) or cob leds (but not sure if white will be as efficient as red blue spectrum..). I guess I need to wait a bit more to see what you guys get.
Anyway, I took a seat! Good luck!

cocogrower said:

The BML lights are very convenient (mix&match) and probably also effectiv/efficient because of their awesome spread.

But now i'm wanting to build my own small cob panel and reach 150-160 l/w with a good spectrum. It might be more expensive and will of course be more hassle.

Petflora, the reflectors will let you direct the light in a certain direction, but if you could, wouldn't it be better to find out the perfect density of cobs and let the spots overlap each other?

I'm very interested in your thoughts about cobs! Iv'e been thinking about running 40-50 watt cobs @ 12,5- 25 watts for optimal efficiency. And using 4-5000k with additional smaller red diodes because of the high efficiency of the cooler leds. But pretty unsure about this and still considering warmer spectrums.

What cobs would you opt for in the 30 watt region?

Click to expand...

Lumens per watt is not the be-all you seem to think

I recommend you going to BML site and reading the pdf

As I am not skilled to build my own, I am unable to answer

You have to understand the umoles/m^/s parameters as well as the spectral graph of the specific cob and build accordingly, as it fits into your grow space

I promise you petflora, that i have spent more time trying different spektrums on bml's HomePage than any sane person would actualy do. But there is two ways to go when it comes to cobs. More lumen or better spectrum.. The 93-95 cri ones in 2700-3500 looks pretty sweet, but with a signifacant drop in lumens. The 5000 75-80 cri ones on The other hand are very efficiant even in The lower bins. They also five you a good anount of 440-460. If you just ad The right amount of 630-660 you will get very good par-values. Not so much green, but i do believe that plants need very little of that..

If you run The cobs on a 350-700 amp you could easely find a driver that let you ad few efficient 3 watters. But it might be wiser to start with a 4000k cob. It seems as you will loose a bit of 440nm by doing that though. I guess that you would suggest a higher cri 2700-3000 k high cri cob and accept the loss in total lumens? Actually there are to many options, but I think they all would work out pretty well if you just keep par and red/blue ratio in mind to a certain degree.

If you took your time to visit cree and look at different cobs in the 2520, 2530, 3050 and 3070 series, I'm sure that you could come with very valuable input.. Or maybe you already did?

Explozive, how big is your space? Red-blue is ugly, and the white LEDs are very efficient now a days. I'm sure petflora can explain more about why white is good or maybe refer you to a thread that goes deeper in to that discussion. I'm already convinced, even if I think that a bit of extra red can be beneficial when it's time to flower. White and red on two different dimmable drivers would be pretty cool!

cocogrower said:

I promise you petflora, that i have spent more time trying different spektrums on bml's HomePage than any sane person would actualy do. But there is two ways to go when it comes to cobs. More lumen or better spectrum.. The 93-95 cri ones in 2700-3500 looks pretty sweet, but with a signifacant drop in lumens. The 5000 75-80 cri ones on The other hand are very efficiant even in The lower bins. They also five you a good anount of 440-460. If you just ad The right amount of 630-660 you will get very good par-values. Not so much green, but i do believe that plants need very little of that..

If you run The cobs on a 350-700 amp you could easely find a driver that let you ad few efficient 3 watters. But it might be wiser to start with a 4000k cob. It seems as you will loose a bit of 440nm by doing that though. I guess that you would suggest a higher cri 2700-3000 k high cri cob and accept the loss in total lumens? Actually there are to many options, but I think they all would work out pretty well if you just keep par and red/blue ratio in mind to a certain degree.

If you took your time to visit cree and look at different cobs in the 2520, 2530, 3050 and 3070 series, I'm sure that you could come with very valuable input.. Or maybe you already did?

Click to expand...

IMHO, high CRI is not the be all end all. High CRI means the spectrum is very tight, not what mj needs. IE 2700-3000 is likely better than 3000 or 2700 alone

As to specific nms, it depends on the ACTUAL manufacturers diode spectral graphs, not just what they say it is

Look at my recent pics, 44 days from first pistils. Do they look deficient to you?

I've understood that cri has to do with if the spectral output from a light show colors of objects the way the way they actually look. Like par value but for the human eye if you like. The funny thing though is that high cri lights, at least not the one from cree, don't mimic the spectrum of the sun. Instead they seem to have extra much in the blue and red spectrums - the ones most important for photosynthesis. Unfortunately/obviously they put out less lumen. Now you might say that it doesn't matter because lumen is what matters to humans. But by adding the right wavelengths to a hi-lumen-lower-par-light you can take advantage of the raw power/ high efficiency of the strong but less par/plant inclined spectrum.

I think that you actually already know all this, but for some reason want people to buy Bml's lights instead of building their own.

You are nor so helpful when it comes to creating the most efficient light so please stick to the company "we have been waiting for". I'm sure they are great, but do you have to market them in every DIY thread?

It's possible to buy 12w cob spots at alibaba for 13$. They are built using cree 1507 cobs and give you about 100 L per watt according to the seller. I'm aiming for something greater, but this could be enough for some people less interested in DIY and optimizing.

I like the idea of lowering the mA/volts of the driver to produce more efficient light, but I don't like the cost. $38 per chip and I would need 6 to replace the xte that I have now.
To get them to my door would be $275 chips only. Yikes.

Cree is already expensive enough so I can understand using cheaper drivers and maximizing the lifespan of the chip by running it at lower current, but for space? I think some people (mainly myself) would have a hard time fitting all that into the cabinet.

That said if you are actually just using 12 watts of power you would not need much of a heatsink. If a hot 50w chip needs to dissipate 60 degrees C then your cree would only need to dissipate less the 15 degrees c??? Even to cool 30 degrees c takes very little.

I do like the 4000k spectrum. I wished I had bought some when I bought the 3k and 5k cxa2530's.



check out the 4k spectrum on the High-Density Integrated Arrays, but very pricy.

I run a single cxa3070 at 96w to avoid running 4 cxa3070 at 350mA. That seems efficient to me.

Nonagon, those chips look great, I bet they would be super efficient if driven low. Very expensive and hard to find though. Another problem is that the drivers will be more expensive to.. I think the 3070 are a better options. High binned 2530 might be the best option if we/I would opt for a compromise between efficiency and price.

As you say, the heat sinks wouldn't have to be that big, especially if we cool them actively. I'm thinking about putting the sinks in a vent tube (inspiration from lovedemshroom and sammuz2646). Drnagons 3d ilustration shows how to do this with square vent tube, but they are hard to find in metal..

If we talk about the 3070 at 350 instead of 1400mA, they would generate 25 % of The heat but also about 25% more light. If we also consider that lights could be placed closer to plants and give a more even footprint, I guess that you could get away with 40 % less watts and 40% less heat sink. If you have expensive electricity and keep these lights for a longer period I do believe that you would get your invested money back!

3Ddream, Nothing wrong with that, probably very effective. I do think you would get more bang from 3 3070 driven at 350 amp. expensive and possible problems with coverage, not to mention the hassle, i would probably not change if i were in your shoes..

what k and binning are you using? Do you have a log somewhere, so that can look at your plants?

I started with the 3050 and killed it by plugging into a live driver. I am now using a cxa3070 3000k (z2 bin) in a micro grow. I have a thread in the micro grow section called "100w led stealthy stacked tub". Plants are very small though. I am looking for a nice mom in a tiny space. I'll have some nice pics this weekend. I found a purple panama and she has no problems showing color under the led.

Disclaimer: This might not be the most efficient coblights ever built.

My cabinet only measures 1,5x3 foot and i would need two or eight lamps to get an even footprint. For already mentioned reasons i will go with 8.

But since my electricity is included in the rent and i get more than enough lumens from cheeper cobs i can't justify the extra cost of cobs that would give more than 180 l/w. Well, it would be pure vanity, anyways.

My compromise will probably be to buy 8 CXA2530-0000-000N00T440F. They will give me 131 lumen per watt @ 800mA should give around 145-160 l/w if driven at 350mA

They are 4000 k and look like this:


The spectrum is slightly worse than the 4000k chip Dr nonagon found. The blue peak is closer to 450 than 440 and there seem to be slightly less red But i found thees chips for 16$ So i will have some spare money invest in a few strategical chosen 3-10 watters

Do you guys think that i should be happy with this spectrum for veg? And how do you think i should supplement it for flower?

Each cob will run at 12.95 w and emit around 2000 lumens (yes i could chase the exact par values, but i think lumen still say more to most people). That leaves me with 3478 lumen per square foot.

Any thoughts on this? Should i push the button?