A total nube's DIY LED build 🛠

[nube]

I recently built new LED flowering lights and wanted to detail it in its own thread. I’m a total electrical newb and I have no workshop or tools other than a drill, but I DIY’d some lights before and figured I’d do it again.

First, a little background. Starting in the mid 90s, I’d grown strictly personal under MH and HPS in DWC or hand-watered rockwool using GH nutes. In the late 2000s I made some big life changes, quitting smoking cigs after nearly 20 years, changed my diet, got back in shape and took some time off growing & smoking weed.

When I got back into growing, I did a lot of research before jumping back in, and decided to update my grow style. I decided to switch to organic recycled soil using two Cree 3590 COB LED fixtures I built. I couldn’t be happier with the results.

Having sold the bigger fixture when I moved, I’ve been using the 400w 8-COB fixture for the last year or two. While it worked well and pulled lots of good harvests, it wasn’t quite powerful enough to flower a 4x4 to my tastes.

I’ve kept track of lighting research this whole time. Here are a few great resources on the subject:

LEDGardner.com this website and forum have great DIY guides detailing the parts lists and how to wire strip builds.

Cheap LED Strips : A Viable Alternative Overgrow’s very own resource for LED strip builds.

scholar.google.com for lighting and agricultural research

And Bruce Bugbee’s famous youtube videos where he uses a NASA-funded, FDA approved university research lab to study the effects of lighting and environment on cannabis. Here are a few of his seminal videos:

Far-red: The Forgotten Photons

Toward an Optimal Spectral Quality for Plant Growth and Development

PAR, PPF, PPFD, and PFD Explained

USU Cannabis Lab Tour with Dr Bruce Bugbee

Cannabis Grow Lighting Myths and FAQs with Dr. Bruce Bugbee

It’s worth reading the research and watching those videos, and then re-reading and re-watching that stuff. That stuff is gold.

[nube]

After researching this stuff for the last several years since building my COB fixtures, I realized in 2018 that lower powered strip lights were the best technology for growlights. You could get to 60% efficiency with great spectrum for about $500, so I wanted to build a new light. Problem is, I’m really poor.

I know firsthand that LEDs work great, but they are more up front cost than bulb lights - it costs about 2x to DIY a 4x4 LED fixture compared to a decent brand 1000w HPS. But the LED brings a better spectrum, a lot of cost savings on electricity and AC and no bulb changes every 6mo because modern LEDs have little loss of output over 10 years, so the ROI is less than 2yrs in most places. There’s no question, I’m sticking with LED.

Bridgelux has been making the best bang for your buck LED lights for a few years, and current research indicates the best canopy spread and highest efficiency comes from midpower white LED strips run at lower power for high efficiency. Their EB2 line of strips was the bang-for-buck champ the last couple years, so much so that other LED strip companies think they must be selling at a loss somehow.

Although Emerson & Lewis found in their landmark 1943 paper that far red (700-800nm) increases the efficiency of photosynthesis (this is greatly simplified) resulting in the famous Emerson Effect, with LEDs, researchers are finally able to study the effects of specific wavelengths of light on various types of plants, including cannabis. Recently it’s been proven that far red from 700-740 has a really big positive impact on plants when combined with full spectrum (400-700nm) light, more at the lower end of that far red range and less at the higher end, with little/no photosynthetic benefit from about 750nm and above. (2018 Zhen et al, and 2020 Zhen & Bugbee).

HPS has very little far red, and it’s not normally found in most light sources.

But, in 2019 Bridgelux came out with the EB3 line of strips that increased efficiency and offered 90cri options, which took an already great strip and gave it more far red output while keeping prices way lower than any competing strips. Here’s what an independent LED light manufacturer found when he tested the 2700k EB3 90cri “Slim” strips:

A real-world whole system efficiency of 2.67umol/j is very good. That’s what finally convinced me it was time to build my new light. But what? I have veg covered, with my clone and seedling tub using a few of the EB2 strips, and I can reuse my old COB light for veg.

I needed to build an inexpensive flowering light with full spectrum white and some far red, with enough power and coverage to do a 4x4 no problem, in a form factor that fits in a tent. And I’m a total newbie, as the name implies, so it had to be both easy to build and easy to use.

And then the regular version of that same EB3 strip hit the market at $11.29 per 4ft strip earlier this year. They can be mounted to aluminum as both the frame and heatsink for your light. My mind was made up.

[nube]

I scraped together some cash and put in an order for the 2700k 90cri Bridgelux EB3 strips, then pieced out the rest of the parts. The safest build uses parallel wiring for low voltage so I don’t run afoul of any local electrical codes. The easiest build uses 4ft strips so I have to do the least amount of parallel wiring. The most positional flexibility in a 4x4 using 4ft strips requires 2 fixtures in that space, so I can have shorter plants on one side and larger plants on the other, if needed, then angle each of those fixtures to keep plants of varying sizes the same distance from the lights. Two fixtures means I need to have two drivers for cleanest wiring. Two drivers allows me to choose the power level and price I want to shoot for.

So, after that thought experiment, I decided on two fixtures using 4ft strips wired in parallel, each fixture housing 10 strips (because I had to by 20 of the 4ft EB3 strips at a time) run on a Meanwell HLG-320H-42a driver (HLG series because of the 7yr warranty, having had drivers die at 5 years old before, and because XLG only goes up to 240w). Those drivers were cheaper per watt than anything else in the HLG line, and they gave a lot of flexibility with their internal dimming pots. Plus, at around 320w each, they gave me the ability to really max out my space while also feeding the strips just barely above nominal current (760ma at 39v per strip), or whatever I wanted to dim to below that.

Here’s the full bill of materials and approximate delivered cost from where I bought them:

$270ish - 20x Bridgelux EB3 90cri 2700k 4ft strips BXEB-L1120Z-27G4000-C-C3
$150ish - 2x MeanWell HLG-320H-42a drivers
$130 - 20x pieces of 46" 1x1 square corner 6061 angle aluminum to act as the frame, the mounting surface and heatsinks for each strip, from the local Metal Supermarket
$0 (have) - 50ft roll of 18awg solid core wire (must be solid core)
$0 (have) - 10ft of 16awg lamp cord for WAGO-to-WAGO
$0 (have) - 10x 5-position WAGO connectors and a couple 2 or 3pos
$0 (have) - 2x 6ft 18awg PC power supply cables
$10 - 2x glans connectors for power cable to driver
$25 - Double-sided 1" tape to attach strips to angle alum heatsinks
$0 (have) - ratchet hangers
$10 1/4-20 x 4" eye bolts with nuts and washers
$15 - m3 plastic screws & nuts
$0 (have) - small nuts and bolts to connect the frame pieces
$0 (have) - drill & drill bits

Total: $600ish delivered for up to 640-680w (depending on variance in the drivers) of very high efficiency strips with good far red & deep red output.

Next step was buying all the parts and building it. Problem was, the stupid US tarriffs on aluminum meant that angle aluminum I was gonna use now cost $250 no matter where I sourced it, which put it out of my price range. So I ended up redesigning the whole thing and using two slabs of 44.25" x 20" x 1/8" 5052 plate aluminum from the same Metal Supermarkets company. That cost me $135 cut and deburred, so I was back in business.

[Chronickyle]

Just wanted to verify if this was in Canadian dollars. I believe it is, but would love to make sure.

[Gpaw]

Great post @nube!!

Nice job on the intro and excellent idea linking in Bruce’s video’s (…I should have thought of that).
Basically we built very similar lights. I used the Gen2s and HLG-320H-42B ('cause I like twiddling with knobs… ).

Excellent presentation.

Are you running the drivers external to the grow environment?

Cheers
G

[nube]

@Chronickyle

Just wanted to verify if this was in Canadian dollars.

No sir, this is all USD. I’m in Colorado, USA.

@Gpaw

Basically we built very similar lights.

Yep, and thanks for the nice comments! I got the idea from LEDGardener’s DIY guides in 2018, and have been wanting to build one for a long time. I convinced @Torontoke to build 3x of them with the EB2 and the 480H-42a drivers for his spaces last year, and this year advised @Vagabond_Windy to build the same thing I’m talking about…before I had built mine! He was the proof of concept beta tester, and it’s been working out great for him!

@Gpaw

Are you running the drivers external to the grow environment?

No sir, they’re in the grow tent right now. Since LED grows do best in the 80-85F temp range at the canopy, I wanted the extra heat in the space during the winter. Our house stays about 68-70F during the day, and about 59-60F during the night. I have the lights on from 7am - 7pm so I don’t have to deal with any light leaks in the cheap tent, and temps stay 84-86F and humidity is 50-60%, although I’ve since lowered the humidity a bit since I’m late in flower. In the summer, I can remote the drivers pretty easily.

[Skiball]

I put together a 12 x 288 lm516c 1800w system for 900$ two years ago.

[cannabissequoia]

Thanks man. You’ve made next year’s project much easier for me & probably others.

[nube]

@Skiball

I put together a 12 x 288 lm516c 1800w system for 900$ two years ago.

Very cool! Yeah, those inexpensive QBs from China are pretty sweet, or when HLG has a blowout sale on their prior gen boards. I wanted to get a little higher efficiency and a bit better spread with strips while not having to use heatsinks, but the big 3ft lm516c QBs from Meiju are freakin awesome for the money.

For me, LEDs are about getting the most photons in the best spectrum delivered to the plants for the cheapest total cost of ownership. I want to run the lights as low as possible over the canopy so I don’t lose efficiency to the inverse square law or wall losses, which means I need to have very low heat and low intensity from each point source. As such, my decision to use 20 of these strips means I have 20x224 diodes in my 4x4 tent. That’s 4480 diodes in my space, currently running at 550w total, at 6" above the top of the canopy, but I’ll get to that in a future post.

The price of angle aluminum really threw a wrench in my plans. While places like HLG had been supplying big slabs of aluminum and/or heatinks with their light fixtures for a few years, I really preferred the angle aluminum builds. Or the builds using 8020 T-nut extrusions. But the new aluminum tariff price put those out of reach, so I was stuck with slabs for my build. The trick was to design it in such a way that they’re not unwieldy and don’t flex too much. I’m no engineer or master craftsman, so I just had to do the best I could.

First, I sketched out the layout. Because the strips are 44" long and 1" wide, I figured laying out 10 per 44.25" x 20" slab would be easy. Well, it was harder than I thought because I’m no artist or designer either. This is what I came up with lol.

As you can see, there was a lot of erasing and rewriting because I’m terrible at this shit. The strips are the shaded parts. And it’s not to scale. But there’s two very cool things about this design I didn’t realize till drawing it out.

First, with the 3" space in the middle it prevents too much of a hotspot, from either temps or light spread, especially with both fixtures side-by-side in a 4x4 with a 3" gap between them. Secondly, that space in the middle of each fixture allows me to drill standoffs for the driver so I can mount it on top of the fixture with four screws and a few nuts. This allows it to maintain electrical ground to the fixture without having to run a wire.

But it’s not all peaches and cream. The spacing and placement of the eye bolt holes was a theoretical attempt to make sure the fixture wouldn’t sag in the middle. It was just a guess based on weight distribution. In practice, it didn’t quite work out as theorized.

One thing that was brought up by @ChemicalDependant when he helped Windy build his fixtures was having handles of some sort built into the design, because moving big 4x2 slabs of metal can be slightly difficult and somewhat dangerous if you dropped them on your foot. Luckily, they’re not too heavy (about 12-13lbs each fixture with everything installed) and once you install them, you won’t need to do much maneuvering other than raising and lowering. However, I did consider mounting some drawer pulls to them just to make them easier to handle, but didn’t end up doing it because I’m a slacker and I didn’t have anything on hand.

The more I thought about it, the more this general layout made sense to me since I was forced to use the slabs of aluminum plate. This was back in August. The only thing to do now was build it.

[ChemicalDependant]

I should add, it’s not the weight that makes me not want to drop one of those sheets on my toe… actually chipped my bone dripping one of those rigid plastic Lego base mat things. You have essentially a big, dull blade. Rope would do the trick even, just want to make sure it doesn’t slip out of your hands.

I tried to design something better than Nube’s for a while, can’t find a way yet. If you can get around the aluminum plate it would help a lot, but aluminum is kind the ideal for this… Honestly friends have asked me about LED grow lights, I’ve outright just shown them Nube’s builds. Everyone who has inputted here has come together to make pretty damn near as good as it gets. As I get the cash I’m 100% going with what he’s suggested, it’s literally what I would do after doing my own research.

[Daytripr69]

When I built mine the price of aluminum angle was just to much so I used the wall angle for drop ceilings. Worked great and a fraction of the cost. My setup is probably overkill x2 as I’m running 32 3500k 4’ gen 2 strips and 4 meanwell xlg 240 drivers in a 4x4. Mine are wired series for cleaner wiring, where I’m at building codes for portable light fixtures are the least of my worries

[Vagabond_Windy]

I can add that as someone who’d also been using HPS and MH for 10 years,the most intimidating part of switching to LED were the big words that most home HPS growers don’t care or don’t know about.
@nube did everything except put it together. @ChemicalDependant did that lol.
Hands down this is the easiest build even for a procrastinator like myself. My build came out to a little more but I bought enough extra wiring and tools to not have to buy more for another should I choose to build one(or two lol).

Temps are consistently at 80F and 45% Humidity and my drivers are mounted to each slab as well.

[Tejas]

Don’t get caught up trying to layout the strips, lay them down and moved them until they look equally spaced and mark the location. As far as the bracing for the slab, use two angle or tube aluminum and “X” brace it, then you can attach your ratchet straps to the ends. One other thing, if you are drilling through the slab to install the strips, get a punch or big nail and hammer to nick the point where you are going to drill. This prevents the drill bit from wandering. Looking forward to your build! Good luck!

[ChemicalDependant]

Aluminum extrusions are absolutely wonderful, and machined extrusions are extremely useful as sort of engineering Lego. The cost of aluminum is the real crux here. I almost wonder if perhaps mild steel angle or galvanized steel sheet would frame up better. It increases weight and difficulty for your tools; but you to have gains.

The trick is almost always either get the item from the source, or look for the same item with a different name. I’d be curious about some rectangular ductwork used as a backing. You would probably have to use nuts and bolts, or a backer, as I don’t know that the strips will really mount into such little material, if you’ve ever screwed ducts together know what I mean. You also may require cross bracing so that the shape does not sag, but stays relatively square. But this would allow a few pluses as well. I haven’t looked fully into it yet, I’m not sure on the pricing at retail.

But that’s only one idea I had, composites! You only need the aluminum because it’s uniquely light, strong, and thermally conductive. If you use a thinner aluminum sheet, and use a backing material you could see similar results. My first thought is XPS foam as it’s actually just my favorite lightweight building material. That is much less effective for compression stress rather than tension, so you’d need to bolt the mounting points to the aluminum face.

I learned a fair bit on composites watching the video on those DML panels I mentioned to you. They don’t need to be conductive though, kind of changes things, but could still be worth a look if you go that way.

[Daytripr69]

The strips are so light you don’t need much. I built a 2x2 using an old street sign lol way heavier than it needed to be.

[GramTorino]

I also had fun doing a build.
Started with Aluminum framing, ended up with this.

CMH + Strips Coco Cavern

Going flyweight… [20200825_203530] It was fun building the alum frames, but not using them for this phase. Willing to risk adhesive failure in my lil hobby grow. I have screws if they screw with me.

[ChemicalDependant]

Your aluminum, is there any going to other direction? It looks like your relying on the strip structurally.

My bad, I see your fiberglass rods! I am tempted now to do similar, but drill blind holes to seat the rods.

Wow. I actually really like this, back to reading.

[GramTorino]

This was the aluminum version…

CMH + Strips Coco Cavern

Ready for strips. [20200817_215328] Changed my mind. F the thermal tape. It’s just overkill sticky. Strips will be gently bolted directly to the bare ass naked 1" aluminum. Better cooling than with tape anyway. Top frame drilled for #4 screws.

Yes, the foam and fiberglass rods.

I was split. The metal is heavy and conducts…the foam board is light weight, non-conductive, but is flammable. So keep that in mind. Safety!
Happy with my choice.

Also, going forward, moving the CMH out and going all strip.
I’m building another flyweight panel for the middle.

[ChemicalDependant]

That’s… giving me ideas. Do you have a temp reading aluminum backings of the strips? Running at recommended voltage?

I could easily see something working out from XPS and fiberglass driveway markers. Would just be worried about not only losing a lot of heatsink material but adding insulation to parts, you were bang on using small strips of foam. It is flammable yes, but only at high temps. So no flames, no sparks, all should be well. It’s also sure is easy to work, compared to aluminum.

That said, aluminum is still a great choice, not much more expensive. I’m curious now, what issues you’d see with aluminum vs foam. I have to build two panels at least, maybe I’ll make two versions. But that really could cut the cost of aluminum, if you can get some square stock big enough. Just drill holes for screws and rods.

I need to put up a second space to veg extra, maybe I’ll pick up a driver and a few strips. I was going to buy a low wattage HID, I have a spare ballast and it’s open air for the moment, so heat is not a concern. So I could maybe afford a 2x4 panel, less the aluminum cost.

[nube]

@ChemicalDependant

I tried to design something better than Nube’s for a while, can’t find a way yet.

Well, if you have the money, 1/8" thick square corner angle aluminum is a better option than plate or sheet. Repurposing other types of aluminum extrusions would be the next best bet if you can find them for cheap and they’re rigid enough. I’ve seen people use tool wall extrusions and other neat things. This build ended up clean but, like you said, it’s basically a big dull blade if you were to drop it on anything.

@Daytripr69

I used the wall angle for drop ceilings.

That’s a good idea. Does it flex much? One of the most important things when using strips is keeping them rigid, especially these fiberglass strips, because if they flex the copper traces can go bad. Even deflection of 1/2" over the entire 44" length can cause issues, I’ve heard. Did you end up doing a lot of cross bracing to make your frame more rigid?

Also, yeah that’s way overkill. LOL 600w is probably plenty for the EB2 strips in a 4x4, and 500w is prolly enough of the EB3. 1000w is way too much.

Series vs. parallel is mostly about safety and insurance coverage if there’s a fire, as verified by certified electricians in the LED strip thread linked above. You probably can’t get hurt by our low voltage fixtures wired in parallel because there isn’t enough current flowing through them, but the high voltage series fixtures have the potential to kill somebody. Also, if you’re using high voltage fixtures without a permit and you have a fire, even if the lights aren’t the cause, you could be denied an insurance claim and could be fined by the county/locality. That’s not something I’m willing to risk just for some weed. Also, I’m a renter, so I would never do that to somebody else’s property. But I’m not going to judge somebody for disagreeing.

@Vagabond_Windy

Hands down this is the easiest build even for a procrastinator like myself.

You’re less of a procrastinator than me! lol I’m just glad it’s working out well for you. Hopefully other folks find it easy as well! I think the angle aluminum builds are actually more complex, and I don’t like how they fit in a tent with negative pressure.

@Tejas

Don’t get caught up trying to layout the strips, lay them down and moved them until they look equally spaced and mark the location.

Thanks for the notes on laying it out and building it. I figured it out pretty well in the end. I didn’t need to brace anything, but I appreciate the helpful tips. I did use a center punch to drill all the holes.

@ChemicalDependant

The cost of aluminum is the real crux here.

Yeah, fucking useless tariffs. Aluminum is better for all the reasons you mentioned. Steel is possible but it’s just so not good as a heatsink and it’s fucking heavy. You don’t want much of any flex in your fixture.

@Daytripr69

I built a 2x2 using an old street sign lol way heavier than it needed to be.

That’s essentially what I did, but with fresh slabs of aluminum plate instead of street signs. I did think about “borrowing” some though! lol

@GramTorino

I also had fun doing a build.
Started with Aluminum framing, ended up with this.

It’s a cool build! Thanks for that. I considered not using a heatsink like your build, but in the end I decided that longevity of the strips was paramount, since I’m so poor, and want this to last at least 5 years. Not that bare strips necessarily will die or suffer failures before 5 years, but we just don’t know about that specific use case.

All electronics in general do better with heat dissipation, but I’m curious to see if there’s much reduction in PAR output of bare strips run at or below nominal after 5 years. I’d be curious how you ascertained that part about “Strips will be gently bolted directly to the bare ass naked 1” aluminum. Better cooling than with tape anyway." I don’t agree that bare strips are better cooled than strips mounted to a metal surface with a thermal interface. My understanding from years of building PCs, specifically with regards to the cooling of CPUs and GPUs, is that air cooled is never better than heatsinked even when using thermal interface material pads, and it’s not even close.