Trying to pick the best COB. CXB5690 or CXM-22 or what?

[Klompen]

Cool, thanks for the code! I will try to do some more research about drivers. I've spent the last year working on 18650 battery banks, so I'm sure I can figure all this out given enough time.

I know its a smidge more money, but wouldn't this ballast work?

https://www.mouser.com/ProductDetail/MEAN-WELL/XBG-160-AB?qs=vLWxofP3U2wxzy8WUBTWDA==

I like the idea that I could affix that to the heatsink like Optic LED and Migro lights do..... plus IP67 rating

There are less expensive meanwell drivers outside of the HLG series that work quite well. I'd suggest a XLG-100-H-A dimmable driver for one QB96, which are less than $30 each. Also when you order the QB96 from HLG don't forget the "diy10" coupon code for 10% off.

 

[indagroove]

Cool, thanks for the code! I will try to do some more research about drivers. I've spent the last year working on 18650 battery banks, so I'm sure I can figure all this out given enough time.

I know its a smidge more money, but wouldn't this ballast work?

https://www.mouser.com/ProductDetail/MEAN-WELL/XBG-160-AB?qs=vLWxofP3U2wxzy8WUBTWDA==

I like the idea that I could affix that to the heatsink like Optic LED and Migro lights do..... plus IP67 rating

It would work, but I wouldn't suggest connecting it to the heatsink. You want to keep your led board as cool as possible for best efficiency. I hang my drivers a foot or two away from the lights, and just run a couple of wires in between.

 

[EcoNepenthe]

Great Thread

Great Thread

Great and interesting thread. Thanks. Nice to see ya grow op and how much thought towards improvement. DIY Cree COB user myself...newer LED tech became popular a lil later. Again, great thread, thanks.
Cheers. Eco
Stay safe!!!
Success w/ya grow!!!

 

[Klompen]

That's a beautiful grow there EcoNepenthe. Having a whole room dedicated to it like that would be really nice in a lot of ways. We're currently trying to go for greater stealth though so its all cabinets for now. It should be fun to have parallel chambers where I can run veg, autos, and photoperiod plants all in the same closet. Its going to cost a pretty penny to get this upgraded like I need, but our power bills during this last auto grow seriously undermined our benefit from growing it. We definitely came out ahead, but we're smoking through it a lot faster because of the relatively low quality and density of most of it. Some of that has to be genetic(I hadn't run any of those strains before), because one of the plants is noticeably more potent than the others and it only got more light than the others at the very top of its main cola. That said, I think proper non-blurple light should up my game. I have some pretty nice photoperiod strains lined up, with a couple in particular that I am excited about, but I really want to do them right.

Those Cree COBs are really nice. I do have to wonder why Cree just sort of gave up and let other companies take the lead while they rested on their laurels with the CXB3590. It was top dog when it hit the market. At ~2 µmoles/joule at a moderate drive power, its definitely in a class way above cheapo lights like mine(then again, my COBs cost $1.70 each lol). I have been doing a lot of research and unless HLG is lying about their stats, the QB96 V2 is in a league of its own. It seems shocking more people aren't using these when we see folks raving about them so much. One guy I was reading a diary from said that it even changed one of his lanky, low-yielding Trainwreck strains into a solid plant with considerably better yield. I grew TW back in the day and that thing was a floppy mess of weak branches, but still fairly special stuff. The point is though: it seems like everyone absolutely adores this light after using it. HLG seems to be the subject of a great deal of hype, but the QB96 style of light might be the perfect blend of the qualities of a COB with the qualities of a "quantum" style board. I plan to use mine with heavy side-lighting to maximize top-to-bottom bud growth. Once I am done here there should be 3 QB96V2 overhead lights and 12 Q strips or Bridgelux strips(unless anyone shows me a better option). I want no shadows, top or bottom.

 

[SuperBadGrower]

Whatever you do, if you do it right you are going to see like double the performance or maybe even more. I'm not sure if you grasp just how inefficient those driverless cobs are. Don't half-ass it & you are in for a treat brother.
I'm jealous haha If I make some upgrades I won't see that big of an improvement. if you do a good job you're going from the back to pole position in one move
So take the time you need

 

[Klompen]

Thanks man. I made do with those driverless COBs but I'm very aware of how terribly inefficient they are. Man the cooling I had to put on them is just nuts. There is easily 100 or more watts of heat coming off my 150W fixture. Its like running a pair of server CPUs. The griddle draws about 275W and probably has 20 pounds of aluminum and still takes a 12" box fan to cool it. The driverless COBs kicked ass at the last parts of veg and the first parts of flower but the spectrum is probably missing something. I wonder what ever became of Cree's driverless COB chips. They were the ones that invented the idea I think.

I really have to wonder who they sell so many of these Chinese COBs to. It seems unlikely that tons of Americans are growing with them, and it also seems unlikely that professional horticulturalists would use it when there are other more efficient options. As far as I can tell, my grow with them may be the most successful pot grow using them ever. I seriously can't find many examples of people pulling off a whole grow. Someone must be buying them though because you can barely bring up Ali Express or Ebay without seeing them....

I am thinking of eventually regearing my griddle light into a QB style veg fixture. The amount of aluminum on that thing is crazy and I should be able to take the fan off of it once I get a proper board there. I was thinking a blue spec board, but who knows. I'm not sure my 150W fixture is quite big enough or I'd put a QB96 on that instead. I hate to completely retire something I put so much work into. Its been a real crash course getting back into growing with totally different tech. It was all HID and now everything is LED. I really don't know what to do with the 400W CMH I have sitting in storage.

 

[SuperBadGrower]

Thanks man. I made do with those driverless COBs but I'm very aware of how terribly inefficient they are. Man the cooling I had to put on them is just nuts. There is easily 100 or more watts of heat coming off my 150W fixture. Its like running a pair of server CPUs.

Haha yeah it seemed like I could fry eggs on the heatsink I'd made for them

I really have to wonder who they sell so many of these Chinese COBs to. It seems unlikely that tons of Americans are growing with them, and it also seems unlikely that professional horticulturalists would use it when there are other more efficient options. As far as I can tell, my grow with them may be the most successful pot grow using them ever. I seriously can't find many examples of people pulling off a whole grow. Someone must be buying them though because you can barely bring up Ali Express or Ebay without seeing them....

Those driver-on-board things are intended for bulbs and flood lights (plug into the wall AC stuff). Ones that don't go to bulbs can be sold to tinkerers (just my theory). Not only weed growers are interested in purple light. There's people who grow succulents, microgreens or other plants and want to put some lights on them. The absolute entry level there is to screw in a purple bulb I guess. The paradigm "purple = plants" is still pretty prevalent. Of course purple is fine but if white lights are so cheap, and provide a full spectrum that lets you actually see the plants, it's an easy choice.

 

[Klompen]

That makes sense. I guess I often forget about all the non-pot indoor growing done on small scales all over the place. I know I just got them because I was homeless and 1.70 per 50watt COB was tempting. We were just about to move to a rural property and I got excited about the idea of being able to grow again and figured I would at least give them a shot. All things considered they really didn't do terrible. By the end of the grow I bumped it up to about 518W total in there and got nearly a pound. Its just many of them are leafy and not a strong as I would have liked. Only Three Bears OG x Toof Decay turned out to have a really good flavor and potency. I haven't run any of these before, but I am certain a light with far red and UV added would be superior. This light definitely has plenty of infrared though. No diodes, its just so hot that it is glowing in the IR spectrum. So there's that lol....

We also had a 24 hour power outage at about day 55 of flower and then lost power for 4 days at about day 75. There's no way that didn't affect them. It definitely makes it hard to evaluate the COBs, the genetics, or anything else in isolation.

Since they're about to be retired, I figured I would post up a couple pics of the results.

The chamber during power outage:



A couple shots of our best cola, from TB x TD:

 

[indagroove]

After a TON of reading and digging around, I discovered that the QB96 almost certainly has either Samsung LH508B diodes or perhaps Bridgelux(HLG doesn't say). This would place the lm/watt at about 180 tops(less if the IR/far red is accounted for). So the top-bin CXB3590 and the CXM series from Luminus actually are more efficient at some wattages. Does this translate to better? I really think the QB96 is too much for each stall. If only they made a QB48! I could run them dimmed, but that's a lot of money for what is basically overkill for the space....

Check out the data sheet for the CXM HortiLum White line.... it looks pretty amazing actually

https://download.luminus.com/datash...White_Horticulture_COBs_Product_Datasheet.pdf

Has anyone here used Luminus chips?


In many ways it would be nice to have something similar to this for side lights(but without a 100 part minimum order lol)

https://www.alibaba.com/product-det...spm=a2700.galleryofferlist.0.0.302514fffB1Lb1

I don't know how accurate this info if, but saw this post on another forum this week..

 

[Klompen]

I don't know how accurate this info if, but saw this post on another forum this week..

View Image

Looks like the most efficient of those 5050s are about 172lm/watt. They sure do look like the ones on the QB96 though.

 

[indagroove]

Looks like the most efficient of those 5050s are about 172lm/watt. They sure do look like the ones on the QB96 though.

Yeah, I'm kind of surprised that they would use something other than Samsung 5050 series diodes. Not sure where that guy got the info, but a call to HLG wouldn't hurt if you really want to know.

 

[Klompen]

My guess would be that the ppf efficacy was better on these diodes than their lumen efficacy. Also, they could be getting top-bin diodes that might get a 15-20% efficacy bump. I think they needed to use 5050s because of the high power handling these boards are expected to take. I mean, they say with heavy cooling it can handle a sustained 4200ma! Not very efficient at that drive level I am guessing, but it would shine like a nuclear fireball lol

I suspect they also bump up their PAR efficacy by throwing in all that extra red. I'm not sure how the density of the array factors in either, but I could imagine the overlap of the light from each diode might be involved somehow. I'm still learning here, but it is amazing how many different ways there are to make diodes, to deploy them, and to drive them.

This QB96 is even 94V-O rated, which means its rated to never flame up more than 10 seconds even in worst case shorting situations. It also means anything that drips off of it will not spread fire. That's a rating most people might not recognize but its actually nice to see it on there. I'm always wary of hype, but everyone raves about HLG it seems. Even the aluminum on the substrate seems like its more of a plate than a sheet.

 

[indagroove]

My guess would be that the ppf efficacy was better on these diodes than their lumen efficacy. Also, they could be getting top-bin diodes that might get a 15-20% efficacy bump. I think they needed to use 5050s because of the high power handling these boards are expected to take. I mean, they say with heavy cooling it can handle a sustained 4200ma! Not very efficient at that drive level I am guessing, but it would shine like a nuclear fireball lol

I suspect they also bump up their PAR efficacy by throwing in all that extra red. I'm not sure how the density of the array factors in either, but I could imagine the overlap of the light from each diode might be involved somehow. I'm still learning here, but it is amazing how many different ways there are to make diodes, to deploy them, and to drive them.

This QB96 is even 94V-O rated, which means its rated to never flame up more than 10 seconds even in worst case shorting situations. It also means anything that drips off of it will not spread fire. That's a rating most people might not recognize but its actually nice to see it on there. I'm always wary of hype, but everyone raves about HLG it seems. Even the aluminum on the substrate seems like its more of a plate than a sheet.

HLG makes a prebuilt fixture using the QB96's, that they call a 315cmh equalivant @ 225watts: https://horticulturelightinggroup.com/collections/lamps/products/hlg-225

I'm not totally clear on the specs, but they drive it at 225watts (I'm not clear if that's watts at the board, or watts at the wall including driver loss), using an even larger pin-type heatsink. I'm not what the exact current/voltage they are running it at, but on HLG's page, they claim 2.0 ?mol/joule, which seems about right for 4100ma to me. Interestingly enough, their manual says it's 2.1 ?mol/joule.

The one thing to consider about the QB96 is that you will have to hang it a bit higher than your typical quantum board-type LED, because of the intensity of the individual diodes being much stronger. You aren't going to be able to put a QB96 running at 100watts 6-inches above your plants. You can prolly get away with 12-inches at that wattage.

 

[Koondense]

Between 2.1 and 2.0 is a 5% difference which can be contributed to standard factory deviations in product specs, so nothing wrong with it really.
I'm starting to seriously think of trying 4 of them, driven at 2100mA in a square meter space. Could be(probably is) a COB killer



Cheers

 

[noknees]

18" wide and about 28" deep and about 6 feet tall.

I'm aiming for about 10-20W per side and then a powerful COB overhead.

Budget is a huge issue so it all has to be DIY

DIY on the cheap in a goofy-sized space is a straight line to SIL.

you'd also be buying time on the cheap; i.e. by the time you've fully agonized over, spec'ed out, waited for the shipping of, then finally built your strip/cob amalgamation....you might have a SIL crop harvested.

on top of that, what if you get it wrong? with SIL, the outlay would be minimal, and the grow straightforward. anyways, G/L


https://www.icmag.com/ic/showpost.php?p=8993494&postcount=4481

 

[Klompen]

I do have a fixture for screw-in lights, but they're not even in the same league as COBs in terms of efficiency. They certainly can do good work, but I'd be shocked if they performed anywhere near the QB96. Your work with them looks good though. I think you could cut your power draw in half though with proper diodes.

 

[indagroove]

SILs aren't nearly efficient in terms of lumen/watt as QB's, so it adds a lot of heat to the space, and costs more money in terms of energy in the long run. You might save some bucks in the short run setting up a SIL grow, and grow some decent buds, but it's at a cost in the long run.

 

[Klompen]

While I realize I am still figuring out how I am going to pay for my new lighting, I have some new ambitions to add to it all. I want to get an Arduino microcomputer or something similar and use is to run temperature sensors on every single light I use, keep track of cabinet temperatures, and monitor if fans and lights are operating. Ideally, I want a temperature probe that can dim or cut power to any lights that are overheating. Basically, I want a grow control system. I'd really like to get the QB96 on a huge copper water block and cool it properly....

 

[Horselover Fat]

^ make one cab, use it and you'll be much wiser for the next ones. Your build is different than most so in the end you gotta figure it out

 

[SuperBadGrower]

I want to get an Arduino microcomputer or something similar and use is to run temperature sensors on every single light I use,

Ideally, I want a temperature probe that can dim or cut power to any lights that are overheating.

Overambitious
If you end up building a light, the moment you turn it on you'll realize you can run it 24/7 with no problem at all. The only reason "overheating" is on your mind right now is because of the driverless cobs.

In the same vein: if nobody in the world needs a huge copper water block, why would you?

If it's longevity and efficacy you're after, just buy more leds and run them softer. (= lower current)