Let's talk DIY lights


@1BigFella your statement isn’t completely accurate. In almost all of the current cobs the higher cri peak deeper into the red spectrum. So while you are correct about what cri means, there are definitely some horticulture applications for some of the higher cri leds.

If you wanted to argue that the benefits of the spectrum don’t cover the difference in efficacy, that’s a different story. I’ve seen side by side grows that could sway you either way. In this application however, I could definitely see the benefit. In fact, I could easily suggest mixing eight 90 cri at 3000k with four 80 cri at 4000k.


I did think about that, but I want this light for a full grow and was worried about a weak veg cycle and too much stretch. It is alot of money to invest in a “test” lol


That depends on what you’re trying to target for watts. 12 cobs at 50 watts each provides solid canopy coverage in a 4x4 with reflective walls. Your space is a little larger, and I’m not sure if you have the help of reflection or not?

The hlg-320-2100 will run four of the 36v nominal (d version) Vero 29’s for a total of about 310 watts. 3 of them would put you at about 930 watts of led. I think the drivers are about 93% efficient, so you would probably see close to 1000 watts at the wall. Personally, I don’t think you need that much light. However, you don’t really want to go with less cobs due to the size of your area either.

Long story short, I think you can do a little better. I’m just not necessarily sure if better will be cheaper? As far as watt per dollar, the hlg-200 series comes in a little cheaper. You can usually get them under $50, meaning 800 watts for less than $200. As to where the last I seen the 320 was pushing $100 each, and 240 almost $70 each. However, if you’re controller will handle the load, that would put you at choices of 600 watts, 800 watts, or 1000 watts. That’s assuming we match cobs to fill drivers. Perhaps using 1 hlg 320 and 2 hlg 200’s could be a solution, putting you around 700 watts.

I guess your target would really dictate how you go about it. If you want the 900+ watts, then solution above will work.


I’m definitely not going to attempt to change your mind about which color Temps you choose. You have a good layout. My mentioning that was simply because the higher cri also produce more blue wavelengths in addition to peaking further in the red. But the spectrum provided differs by manufacturer. If you were going with a citizen cob, I would definitely recommend more of the 90 cri. But the bridgelux produces less blue, and peaks lower on the reds. Staying with them, I would say you’re layout is pretty good.


@1BigFella, here is the graph for citizen and Cree 90 cri. Compare them to any of the 80 cri and you should see what @Daddy and I were talking about. The increase does come at the cost of efficacy though, so you have to keep that in mind. In my opinion, I’d rather have spectrum and I’ll drive my led’s softer to keep the efficacy up. Even if it means adding a few extra cobs.


What would be the difference in running the series B (52v) vs running the series D (36v), and no I am in an open space without reflective wall (just flat white). And yes my controller can handle 1Kw x 8.


A mix or 3000s and 4000s would certainly hit both the short and long chlorophyll absorption peeks. Hard to beat that.

My experience with CRI is in video projectors, where we wanted to use 5000 degrees or so color temperature to get good Color Rendering. Using a MH bulb did that, while using an HPS bulb looked like absolute crap. You would pick out a MH bulb by Color Rendering Index: 90 was not bad. 95 was better. It was all about how the colors would look on the screen when people observed them. I say CRI doesn’t matter in this application because a higher CRI may not be better for the plant. It will give you lovely photos, though.

Funny that white LEDs are really fluorescent lamps. In standard fluorescent tubes they use mercury vapor to get ultraviolet light to stimulate a mixture of phosphors that glow “white” (really a bunch of colors). In white LEDs, they use blue light to stimulate phosphors to emit the other colors.


@1BigFella I wish I could grow outside year round, no one is going to win growing theirs under lights vs the sun, so until they invent a mini sun for us growers, then the best we can do is cover as much of the light spectrum as possible and live with the results lol. I will have enough room to add UV and IR in the future if I think its worth the work.


The 52 volt cob is more efficient, but not by a lot. Mostly your driver cob and cob matching comes down to witch cobs fill driver the best. If you’re target is say 100 watts, you can get around there with two 52 v cobs running 1000ma, four 52v cobs running 500ma one 72v cob running 1400ma, three 36v cobs running 1000ma, four 36v cobs running 700ma.

It seems difficult, but there’s just a handful of things to take into consideration. Target wattage, budget, led efficiency, and coverage. 4 cobs driven softer will provide more even coverage and higher efficiency than 1 cob driven harder. It will also drive your build cost up. Budget is usually the biggest factor, so you try to find a happy medium that works for you.

In your case, we’ll say you want three 300 watt bars. Your options per bar are 3 cobs at 100 watt each, 4 cobs at 75 watts each, 5 cobs at 60 watts, and we’ll stop there because we can all do that math. Each time you add a cob the build cost will go up, but the coverage and efficiency will get better too. Once you decide where you’re at with that, you look for driver/cob combinations that will work without leaving a lot of excess voltage available. As you originally stated, the 36v Vero and hlg-320-2100 give you that. I’m not sure that there is a better option in the Vero line using meanwell driver and series wiring. Possibly the 52v and hlg-320-1400, but I’d have to pull the data sheets and check voltages for compatibility.


@dbrn32 Thanks brother, I think I have what I need, you dont need you to do any more research for me lol. 12 series D cobs and 3 320-2100ma drivers. I really appreciate all the time and info.


The cri of an hps bulb is like 20 right? Which would perfectly explain why a mh bulb was better for your projector project. However, saying the possible benefits of higher cri doesn’t matter would essentially be disregarding the validity of every horticulture study I’ve seen. You looked at the charts I posted right? Look at the Cree chart specifically, it shows the 3000k 80 cri against the 90 cri. The 80 cri peaks just over 600nm, as opposed to the 90 cri providing 20% higher intensity in blue wavelengths and peaking around 640nm.

You don’t think that makes a difference for horticulture applications?


Ok. I just now seen you posted that sketch too, like 2 hours ago. Sorry!

The combo of color temps will be plenty good for seed to harvest grow. And the pots, three 100k pots will work if you want to dim individually. Just make sure they actually go up to 100k, and I’m assuming you are aware of installing resistor to keep from dimming to zero? The 320 may be rated for dim to zero, but you should double check. The smaller drivers are not, so resistor is typically installed to avoid any accidental shorting of the dimming leads.

Feel free to post if there’s anything else I can help with.


@dbrn32 thanks again brother.


Any time @Daddy ! If you have the time post up some pics so we can follow along when you get to build.

Also, shop around for pricing. Rapid has been great in my experience. But arrow electronics is usually pretty competitive and ships for free over $20. They won’t have sinks and framework, but could save you a little cash on drivers and cobs. Website is kind of difficult to use though.


It might be better to have a Cree 90 CRI COB than a Cree 80 CRI, but I don’t know. And even if it is better, you can’t generalize and say all 90 CRI light sources are better than 80 CRI sources. CRI is just a measure of how well a light source matches daylight for human perception, which may not be optimal for plants. I can show you a white LED source that uses pure violet, green, and red individual LEDs that has a very high CRI but misses all the chlorophyll peaks. Human retinas have violet, green, and red receptors that are very broad in wavelength response. So broad that we see violet with a little green as “blue” even though we have no blue receptors. A pure wavelength of blue will stimulate the violet and green receptors and we see blue. But a mix of violet and green narrow wavelengths will look to us exactly like that blue wavelength. (And miss the blue chlorophyll peak.)

I think CRI includes a lot of green light which humans respond to the most intensely. Plants like some green but don’t need a lot. So CRI is not a very good measure for how good a light source is for plants. PAR is much much better. PAR and CRI don’t have a very good correlation.

For example, if you have one of these typical COB spectrums with a certain CRI, you could improves the CRI by adding more green light. But that would not improve PAR.


Ya, maybe. I’m not arguing what color rendering index is. Nor am I arguing that cri is the end all be all. Bridgelux’s 90 cri aren’t as good as producing the same spectrum as the 2 I posted, so I would definitely agree that not all are created equal. I would also agree that not all light sources are the same. But even in Bridgelux case their 90 cri seems to provide a better photosynthesis based light spectrum than their 80 cri. Like your assimilation of led to fluorescent, just because they both use phosphor to make light white, they are far from the same. Led’s convert blue wavelengths to white through phosphor, and the phosphor tech is pretty much what separates one manufacturer from the other. Whoever has the cleanest and least amount wins. Obviously there’s more to it than that, but in a nutshell.

Here are the Vero charts 70/80 cri on top, high cri models on bottom. You know where peak absorbtion happens, so which do you feel is better? The difference between the Vero and the others, is simply weaker production of blue wavelengths to hit the same on reds.


It doesn’t matter which I feel is better. This isn’t about feelings. It’s about science. And to do science you have to take measurements and you have to use the right instrument. Otherwise you could be measuring something that has little to do with the thing of interest. We could get a PAR meter and measure both COBs or we could go even further and put clones of the same plant under different CRI COBs and see how much THC each produces.with the same PAR reading.

Oh, and N better be about 20 for each COB if we want statistically significant results. Then we could publish that study in The Journal of Cannabis Production. Somebody has probably already done this but I refuse to pay for scientific studies that were funded with taxpayer dollars and hide behind paywalls. Aaron Swartz was my hero.


If par levels were the only thing that mattered we’d all be growing under blue led right? Highest ųmol/joule of any available light source we have available. Using the scientific studies already available should clearly justify adding high cri to the more standard options. Unless you’re saying things like the McCree action spectrum and Emerson effect aren’t applicable to cannabis cultivation?

You also don’t need a par meter to see that efficacy is lower on higher cri models. Data sheets clearly point that out, as well I have mentioned at least a couple times in this thread. It’s easily offset by running more cobs and/or running them more efficiently. Some people are willing to pay a little more to target spectral performance, instead of trying to cram photons for as cheaply as possible. Where you fall on that is completely up to you, and your ideals. I get what you were saying about cri in general, and agree for the most part. But neither @Daddy or myself were pumping cri. It’s wavelengths provided by these particular high cri models we’re interested in. If you don’t see any value at all in that based off the scientific data already available, I’m not sure what to tell you.


Okay, here’s my first, basic question.

If someone were looking to build their own LED lights, where should they start? :slight_smile:


For some reason I didn’t see/notice that there were already 458 replies in this thread - I’ll read through it and maybe my basic question is already answered.


Good luck with the reading, lol! @Bogleg

Size of the area you need to light, how you plan on using said area, and your budget are biggest determining factors. But there are other factors. Dollar for watt isn’t always necessarily cheaper to build. But if you look at the ųmol per dollar you’ll almost always come out ahead with a diy system. The ability to source higher efficiency leds used in $1000+ fixtures and put together yourself can typically cut the cost in half.

For example, you have 4x4 tent that you prefer to run scrog in will have different ideal numbers from say throwing a plant or two in the middle of it and leaving room around the edges.