Let's talk DIY lights


I like the idea of a thread just on the conversation re: PAR, PAR watts, ppf, and pffd. That thread could compare different light modes as a baseline. Then we could use those baselines to discuss the different ways you could build a light.

I think the vast majority of us are growing in 2x2, 2x4, 3x3 or 4x4 spaces. So I think from a “Lighting 101” perspective it would be cool to have a “minimum required ppf/pffd” kind of spreadsheet for those sizes, then we can plug in different options to get to our ppf/pffd requirements for the space.

Then you could have a more specific “build” thread. It would almost be cool to go over some of the more common options/configurations for each typical size of growing area in it’s own thread…

My thinking is that a lot of people come to this forum like I did - with a simple question (mine was wondering if I was in week 3 or 4 of flower because I messed up my journaling) but then find out they might not have enough light. We see it a lot. So if we had a baseline that said, “hey, this is what you need to provide your plants in terms of light for X space,” we could then tackle how to get said amount of light for that person.


Right now most people just throw out the “watts per square foot” number but that doesn’t really provide the best answer, IMHO.


Sure! I think the only issue we run into there is having access to par meters. I have access to one sometimes, but even there are limited leds at my disposal when I can typically get it. Which leaves us with having to extrapolate data from other sources, some of which obviously more accurate than others.

It’s a good topic though. I see regularly things like lux referenced. It’s a number, and has some value. If you want to locate the issue with lux readings though it’s not difficult. By definition, lux is simply lumens per square meter. Like I said, it’s a value with a value. The problem we run into is while a lux meter will read some of the wavelengths plants use, its heavily weighted to a wavelength that isn’t really used at all, right on the edge of yellow and green.

So a light that’s heavy in red and blue wavelengths and/or produces considerable intensity outside of the 400-700 range isn’t measured properly via lux meter can have considerably high par levels.


Yeah - clearly not having a lab and the ability to test things ourselves is a problem, but there are people out there doing it in some capacity, so where we can I think we validate data or cross reference it, or caveat it as necessary, we can do that. So far we only really have one good scientific paper I can find on the light requirements for cannabis as well… hopefully as it becomes more medically and socially acceptable we’ll get more studies and science we can glean information from as well. The rollitup folks have done a nice job, for example, coming up with a spreadsheet for COBs that answers some of these questions, but in their case there are a lot of assumptions made (IMHO) as well.


How many times can I end a sentence with “as well” today I wonder…


I agree trust me! I guess from my standpoint, I think the thread and conversation is a good one. I’m just not sure how comfortable I feel building a spreadsheet off of someone else’s data I guess. If that makes sense? One, because I just don’t like doing spreadsheets haha. And more importantly, the aspect of putting my name on someone else’s work and calling it matter of fact.

While it be time consuming, I like being able to reference it when need be and not only taking with grain of salt, being able to suggest others do the same.


As many times as you like here lol! I don’t care. I have a few things I catch myself using a lot, as well!


I hear you. Right now you can get enough data without your own testing equipment for a DIY light to make you feel almost confident, and confidence is super important when growing, IMHO. :slight_smile:


But the math is easy, buy as much of the 2nd or 3rd best lighting that you can afford then save up for more. Right? Lol


Sure! That second or third best though is really subjective. Hell, the best is subjective lol. I think there is a lot that rides on each scenario.

If you don’t have an electrical or electronic background, a kit from hlg is hands down the easiest way to go in my opinion. You’re tier 2-3 in cost, but tier one tech and level one build difficulty.

I would never go as far as to say the cxb is the best cob bar none, but if you asked 100 people familiar that’s the answer you’ll get. It’s also the most expensive, at double some of its closest competitors. That’s a tier one led but you’re also talking the highest cost.

Bridgelux, citizen, luminous, Phillips all have offerings that stack up nicely depending on needs and goals of individual builder.


I think I am a good example. I am just growing for myself. I don’t need to get a gram per watt out of my grows. I just need to sustain myself, which typically means an ounce a month on a “heavy” month. So if I can get 6oz a grow I am good. I am just trying to figure out how I can do that the most efficiently. And if that means I get 8-9oz then the more the merrier; it means I can be more selective about my grows, when I start grows, how I plan them, etc. Right now I am riding the edge as far as bud on hand and when my current grow will finish. But outside of that consideration I don’t need to get a pound each grow. Half a pound would be awesome. That’s why once I finish the complete COB DIY fixture I’ll be happy for awhile (like, for this year anyway).


“This year anyway” lol. Ya, until the new stuff starts coming out!

In retrospect, your citi’s have gram per watt plus capability. As long as you’re aware that the guys pulling those numbers are usually growing from clone that’s proven to be high yielding, probably have multiple grows with that pheno under their belt, and using methods that produce quantity first.


Please also tag me if a new thread is created. I’m very interested in this topic, specifically related to the discussion of energy efficiency, but really all of it is fascinating to me. I’m considering embarking on a build just because it provides an interesting learning experience.


Will do partner


I just ran a very interesting experiment with my Vero29 5000K COBs, my lux meter, and some colored gels used for stage lighting. I set the lux meter on a flat surface about 5 feet below the lights and turned it on. Then I put the red, yellow, green, and blue gels on top of the light sensor.

none…13960 lux…energy in specturm
red…923 lux…30%
yellow…8871 lux…55%
green…1367 lux…55%
blue…1373 lux…50%

When you compare these to Bridgstone’s published spectra data for their different color temps, it actually looks pretty good. All the colors add up to 12534 lux which is reasonable (about 10% loss) for going through the filter sheet without dye. (They did not give me a clear one.)

If you divide the yellow value by 6.5 you get a reasonable approximation to what you would expect from the spectrum. The other three colors match what you would expect pretty well! I suspect this is because the lux meter is especially sensitive to yellow, but no problem: We don’t care about yellow much!

So for a $10 lux meter and $5 worth of color filters we do have a pretty good tool for measuring the red and blue content, because you can calibrate it against a published spectrum. Of course it does not give you a real spectrum with peaks and valleys to match against the Chlorophyll absorbtion peeks, but neither does a PAR meter. In fact a PAR meter would tell you you are fine with just a brilliant green and yellow source, when you are not. This way you can get some numbers in the blue and some numbers in the red.


Thanks for doing the work for us. I’m far too lazy for that kind of work lol @1BigFella


I think the most important thing is to know your light’s spectrum. Once you know that is a decent spectrum for plants, then a PAR meter would be ideal. They are calibrated for sunlight’s spectrum and there are correction factors of a sort for fluorescent, MH, HPS, etc. But the PAR guys are very reluctant to give out a correction factor for LED since the spectrum could be anything! White COBs with their known color temps could have correction factors for PAR, but once you go off into pure blue and pure red LEDs it gets pretty unknown. Because a lot of some colors would make a high PAR, even if you were completely missing some other essential colors. If you know the spectrum, you will know you are not doing that.


I’m not familiar with that published document, so if you don’t mind helping me understand.

You take the readings with specific filters, record the data, apply a correction and then you’re left with some numbers. What do the numbers represent?

Is it just the relationship of intensity of one wavelength to the next? Photometric units, radiometric units? Is it an energy level at all?

My understanding is that a couple hundred dollar par meter could be fooled in a way you describe, but not a Spectroradiometer based meter. It separates and corrects for those non useful wavelengths. Kind of how a lux meter does exactly the opposite.

I’m trying to understand more than anything, just not seeing the value of what the values mean. Those findings look what I’m figuring is identical to the spectral distribution chart located in data sheet. In which case, we can just go look at right? I don’t see any correlation to the amount of energy available to our plants. Certainly not enough to compare one cob to the next or one color temp to the next. All that it tells me is that when light spectrum is identical higher efficacy will provide more par on a watt for basis. You don’t need a meter or filters to determine that.


The document I am referring to is the Bridgelux Vero29 datasheet. It contains spectral graphs.
I am calculating lux within each color band, which is exactly what a spectrophotometer does. It’s accuracy is just higher.
Yes, if you had a real spectrophotometer, then you could get very accurate values at every wavelength, but those cost a lot of money. Just look at this as a poor man’s spectrophotometer with 100 nm resolution.
I wanted it so I could see what effect adding some red 660 nm LEDs would have to my total spectrum. With this technology I can do that by comparing the red in the white COBs to the red of the combined COBs and red LEDs. Since I know the red LEDs are all 660 nm, I can draw a new spectrum for the combined system. I could do that before just by knowing the spectra of both COBs and LEDs, but with this setup I can draw the 660 nm peak height correctly. That will tell me how many more red LEDs I need to add.


Then the difference between the a Spectroradiometer and a spectrophotometer is that the one I was referring to reads radiometric energy vs photometric energy? Which if I’m not mistaken is the most significant reason why lumen based measurements aren’t really any good.

In the case of being able to chart your light spectrum when it’s pieced together like that, that’s pretty cool!