You absolutely get better thermal transfer with a thermal interface, unless you get like micro-honed surfaces. But shitty thermal tape may very well be better than just raw metal. I would use a thermal compound myself, let the bolts hold it on, but good thermal tape or non adhesive thermal pads is just the same, if not better… but I’m not sure the strips need that much heat sinking. Allegedly they run with no additional heat sinking with I think… 2” air space? Those numbers maybe false memories; don’t hold me to that. I do own an contactless thermometer and thermocouple, plus might be adding automation to the build… Let’s just say expect something cool next year sometime.
So getting back to the build, after sketching it out and deciding the build was solid, then seeing Windy and ChemicalDependent have success building out my design, my metal was delayed because the Metal Supermarket moved and couldn’t get to my order for a few weeks. Then I got a puppy and lost 6wks to lack of sleep and puppy interruptions. I literally got nothing done those 6wks but it was all worth it cuz the pup is great. Enough with the excuses, though, right?
When I finally got around to doing the work, I laid out everything on the floor to make sure I had it all. This includes everything used in the build except for the 2x 18awg PC power cords I had laying around and repurposed for the power cords of my two drivers. Also, it includes 1/2" grommets that I didn’t talk about before.
So I then got to work laying out the strips and marking hole locations. I did this all on my kitchen table underneath a skylight.
This was a low tech build, and all I did was lay the strips down starting from one end, use no-residue masking tape to keep them in place, use 1" masking tape to separate the strips, and then a Sharpie to mark hole locations and strip edges. This got me an approximation of what I needed, and allowed for the space in the middle for the driver standoff thru-hole marks.
You can’t see it but I marked three holes on each strip where I was going to fasten it down. Each end and one in the middle. This was the most critical placement of the whole thing.
I don’t have a workshop, so to drill and sand and deburr the plates, I took them outside and put them on milk crates on my patio. I used a center punch to create a starting point for my drill bits so they wouldn’t walk all over the place, and I used old drill bits that I had laying around.
Shockingly, I didn’t break any bits but I did misuse some pretty severely by using them in ways significantly different than their intended purposes. The strip fastener holes are 9/64th which fit the M3 plastic fasteners. If you use metal fasteners to attach the strips you can short them out. I used 1/4" for the eyebolt holes and the driver standoff holes.
Then I had to figure out how to drill 1/2" holes for the grommets. Those holes are where my wiring is going, and the grommets are to prevent the wire jacketing from being damaged or rubbed raw. I ended up using the 1/4" bit to drill a starter hole, and then used a 1/2" woodworking spade bit to make the finished holes. I’m shocked this worked out, but it did.
Like I said, this is super low tech. I used a drywall joint knife to chisel some of the biggest burrs, and I then deburred all the holes with sandpaper. That’s about it. Nothing fancy. I don’t have lots of tools and didn’t want to buy any because I’m poor. I did leave the tape separating each strip to help me guide the thermal tape when I finally applied it.
I love those bruce bugbee videos, he’s a champion of cannabis researching justice. I definitely need some more IR in my LED room. Can’t wait to see your build hit first light!
Hey man, IR isn’t super useful to plants other than as heat. Maybe you meant far red (FR)?
lol the irony is I was heating the room up the water temps dropped to 16C, yah you’re right I need some far red but I am obsessing over IR.
On the heat. I kinda planned for running them at lower levels.
They barely get warm @ 600ma.
Currently running them at .380 in late flower. 315 CMH in the middle.
My aluminum frames were way overkill.
Might be in my blood to build light. Lol
Nice write up man, and good info as always. I like the straight forward and simplistic build style. You gotta work with what ya got! Your plants are looking great under your new light, are you gonna retire your COBs? Strips only?
That’s pretty cold, so yeah, maybe you do want IR! lol The Bugbee videos should be required for everyone building a DIY LED light.
Can’t argue with those low currents. Hopefully they stay happy and output remains consistent for many years to come! The EB2 strips are a supreme value at their blowout prices. Amen to DIY!
No reason to retire the COBs yet. When/if I can expand with a veg tent, they’ll be put to use in there. They’re still great and I wouldn’t want them to goto waste, but I’ll only build strips for the foreseeable future - they make a lot more sense than anything else. However, the prebuilt bar lights from Meiju are looking super sexy for the money… Thanks for the kind words.
Continuing the build, after deburring all my through-holes and sanding down the edges of the plate so they weren’t razor blades against my tent walls, it was time to mount the strips. For each strip I put down the tape and then the strip immediately after so there was no dog hair or contamination between the tape and the strip.
The method was pretty simple. I first cleaned the whole plate with alcohol twice. I wanted to make sure they were spotless. Also, I should have said something before but this plate comes with one side covered in a plastic coating to protect it. Remove that and then clean the plate! After cleaning, I applied the thermal tape to the aluminum plate close to the area outlined for the strip. I say close because the tape is about 1-2mm wider than the strips. The most critical thing was the placement of the strip’s holes in alignment with the holes I drilled. Like I said, I’m no engineer and I just eyeballed the drill holes after marking them.
To get the screw holes to line up, I poked holes through the tape at each of the 3 fastener holes. Then I put a plastic M3 x 12mm screw thru a hole on each end of the strip. The strips are pretty flexible since they’re so long, so if you don’t get the hole placement just right you’re kinda screwed. To prevent them from sticking to the tape before they were aligned, I put another M3 screw in the middle to keep the sag from sticking to the tape. Then I just lined up the screws on the ends with the holes and then lowered the whole strip to the tape. It’s hard to explain but if you build one, you’ll see. They turned out pretty good.
The tape is extra sticky. This was the 3M 8810 Thermally Conductive Adhesive Transfer Tape, 1" width x 36yd length (I found a super cheap deal on Amazon, normally $220 per roll lol!), and though it says no cure needed, everywhere I read advises a 72hr cure before any heat is applied, so that’s what I did. The M3 x 12mm screws and matching M3 nuts are to make sure it stays put both before and after it’s cured. Lots of people have issues with the thermal tape sticking properly, and it’s almost always because they don’t allow the 72hr cure before using.
Anyway, the next step was wiring. Because the push connectors on these boards are so fragile, I wanted to make the wire harnesses before plugging everything in. You can tell many quantum board manufacturers have realized how fragile those connectors are because they’ve now started putting big plastic block protectors over all of them. Strips don’t have that luxury.
With parallel builds, you want the wires to each individual strip to be approximately the same distance from the driver leads. Again, I’m no electrician or engineer, but as I understand it, this is to prevent voltage and current sag due to more distance in some rather than others. Well, it was pretty easy to do. I just cut 20x 12" lengths of my 2-conductor jacketed 18awg solid core thermostat wire.
Then I stripped off the outer jacket on each end, then stripped about 1/2" of each of the conductors. The reason I used 2-conductor wire was to keep the wiring fairly clean, and the outer jacket added an extra layer of protection if the grommets came out of the wire thru-holes.
I decided to use WAGO connectors as terminals for my wiring harnesses because I had them. Terminal strips with bus wires between each terminal would have been cleaner, but would have cost more money. If you’re not familiar, this is what WAGOs look like (they’re so so so much nicer than wire nuts):
The wiring pigtails I made are pretty simple. On each side of the fixture, 5x 18awg positive and 5x 18awg negative wires get pushed into 4 of the terminals in a 5-position WAGO, with the remaining terminal left for 16awg lamp cord that on each side go to a 3-position WAGO which combines all the positive and negative wires into their own single wire to connect to a final WAGO that connect the strips to the driver output leads. This is what those wiring harnesses looked like when done along with the 20 year old super basic wire stripper and wire cutting pliers I have.
I marked each WAGO as either + or - so I wouldn’t confuse them. It looks like a jumbled mess of wires here, but the next pics of it all wired up should make it clearer.
Thanks for those documents. Unfortunately, they’re either misleading sales pitches or they’re out of date already. This is why people should watch Bugbee’s videos and the other ones I mentioned, and read the latest peer-reviewed research on light sources and spectrums.
I don’t mean to trash you or what you posted, but people will be led astray if they take that as the final word in lighting. They’re full of old or wrong conclusions and speculation, and some are just sales mumbo jumbo.
This is the problem with applying early and evolving research to cannabis - it’s not directly applicable unless the published results are from a direct study of cannabis or from biologists who study cannabis directly. And then the conclusions have to be reached by other researchers without major dissent. I don’t ever trust white papers put out by people or companies trying to sell a product - I’d recommend to only trust the real published, peer-reviewed research that’s later validated by others.
Take, for instance, that very first PDF. The one from Osram horticultural lighting says only blue and red primarily contribute to photosynthesis, but we now know that’s not true and that all the wavelengths between about 380-750ish help with photosynthesis, and nearly equally but in different ways. Plus their info on far red (730nm) isn’t accurate. Bugbee and Zhen’s research below proves it pretty conclusively. Also, Osram bought Fluence and this year they commissioned a dozen plant biologist studies to evaluate their spectrum and designs. None have yet been published.
And then the Energy Neil Yorio PDF is a sales pitch from BIOS lighting company for 7 year old LED technology (they claim 1.7umole/j is the highest efficiency LED prior to their 1.85umole/j lights, but current highest efficiency of lights you’d actually use for growing weed is 3.0um/j or slightly higher). Unfortunately, it’s not much help here either.
The third PDF is a published journal article about the effect of blue and red light on water and nute use of lettuce (no idea if it was peer-reviewed). It’s correct in concluding that spectrum matters to the nutrient uptake of plants, but that’s not super groundbreaking. Their test of only blue and red in varying concentrations is pretty limited, as shown by dozens of other plant biologists since. Blue and red (aka blurple), while definitely being the major photosynthetic pathways, isn’t the best lighting spectrum for just about any plants. Plants evolved under the sun we have, not a blurple sun, and as such, most of the spectrum put out by the sun is photosynthetically useful in some way or another to most plants.
The 4th PDF is a 2014 University of Utah Dr. Bruce Bugbee cost analysis of bulb vs. LED lighting in greenhouses. Modern LEDs are more than 2-3x more efficient now than they were then, and about 1/4 the cost per watt that they were back then, so their cost analysis and conclusion isn’t really relevant anymore. In fact, I believe Bugbee now says LEDs are the best option for new greenhouse construction because the ROI is so much faster than it was 6 years ago when he did that analysis.
The last PDF is from 2017 and analyzed the harvest weight of fruit from HPS, HPS-LED, and LED-LED supplemental greenhouse light sources in Nordic (cold) regions. Unfortunately, they didn’t normalize the light sources to PPF output like they should have, nor did they control the environments to have the same temps, and they were using very old LED technology, so it’s not really a robust methodological paper. Their LED-LED treatment had nearly half the output as the HPS and HPS-LED treatments. LOL! Being plant biologists and researchers, they should have known better. As such, I’d say their outcomes don’t have much bearing on reality if you controlled for the things I mentioned. Of course the lights with half the photon output and a lot colder temps produced less harvest.
Again, I appreciate you posting those because it’s important to have a robust discussion from lots of viewpoints, and I’m not trashing you by mentioning this stuff. Reasonable people can disagree reasonably. It’s just that those are old and/or insufficient papers, and some are just outright sales pitch white papers with a profit motive.
Here are two new peer reviewed papers worth reading about the effects of Far Red (I referenced them earlier):
I don’t take what any of those so called certified electricians has to say seriously after being told that you can safely touch a high amperage fixture which is the farthest thing from the truth. None of my circuits exceed 300v @.7a. I’d trust that much more than 50a 40v. Considering it only takes .3a across your heart to stop it.
As far as the frame it flexes some but that was only while stuffing it in a tent. I didn’t run angeles for each strip because mine are so close it would have liked like a ribbed plate lol I did the perimeter and 6 cross bars and mounted the strips to them. When I get home I’ll take some pics and post them up. I’m thinking of adding another 8 strips and another 240w driver and doubling my space to a 4x8. My last run starcube loved the 840w, bubba kush and gold leaf not so much. I ended up killing off all my gold leaf clones just because I didn’t like all the foxtailing in late flower. I did notice that I’m having to use more calmag since swapping from hps
its cool man. most of the stuff i have is research from years ago. ill check out those two
No worries, thanks for jumping in! The videos in the first post do a great job at summarizing the current 2019/2020 state of spectrum research on cannabis.
After building the wiring harnesses and letting the thermal tape cure for 72hrs, I got to work wiring them all up. I first installed the eyebolts so they would act as legs to keep the fixtures up off the work surface on both sides. I worked off the edge of the kitchen table so I could poke the wires up through the grommets. Each grommet has one wire through it except for one per side of the fixture that has two. Plugging in the wire’s straightforward, but I do have one piece of advice:
Be really careful that you pre-bend all the solid core wires when working with these strips, as you’ll break the connectors if you try to plug them in and then bend the wire.
This is how they turned out.
Then I got to work wiring the driver to my 18awg PC power cords. I cut the ends off the power cords, then mated them to the driver’s AC inputs via a 3-prong gland connector, really just to make a clean connection that’s waterproof. Here’s what that looks like.
Once that was all done, this is what the wiring looks like and all the plastic nuts and bolts keeping the strips in place. I feel like it turned out pretty simple, just how I like it.
Here’s the first fixture’s smoke test with the driver dimmed all the way down. This pic has automatic white balance from my phone, and it’s not really what they look like. They’re much yellower than that.
And then the second smoke test with both of them turned up all the way. I let them burn in for an hour or two just to make sure I didn’t have any issues with any of the components. This pic is more realistic color.
I also got out my knockoff Kill-A-Watt meter and looked at the total draw. The MeanWell HLG-320H-42A drivers are 320w rated, and dimmed all the way down they drew only about 120w each, which was below what I was expecting. I thought the dimmers only went down to 50%.
I also knew they were slightly overclocked from the factory so they’d put out a little bit more than 320w at full blast, but I didn’t expect them to put out THIS much combined:
That’s 360w per driver, which is ridiculous overkill. Obviously that’s too much for a 4x4, but it’s good to know I have the headroom to upgrade to a 5x5 and legitimately flower it out with these two fixtures at some point in the future if I want.
I then took the plants out of the tent, removed the old COB light fixture, and installed the new ones. I used only two ratchet hangers per fixture. I didn’t remote mount the drivers because it’s winter and I keep my house cold, so I wanted the extra heat in there. I also didn’t do any cable management before taking this installed pic:
They were installed dimmed down to about 450w to start, which is about what the old fixture was putting out, though much less efficiently and from only 8 point sources instead of the 4480 point sources in this build. I didn’t want to shock the plants. After a few days, I turned it up to 550w total or about 275w per driver, and that’s where it’s stayed. This is what it looks like in practice:
Overall I’m pretty happy. I got quite a bit more than I’d hope for out of the build. It’s definitely overkill for my needs, but that overkill means it gives me a little flexibility in the future if I want to go bigger.
The final numbers:
Cost: around $600 delivered
Wattage: 720w max
Current per strip: 880ma at max, about 670ma in normal usage (550w)
Total PPF: 1840umols at max, around 1450 in normal usage
System efficiency: around 2.5umol/j at max, around 2.65umol/j in normal usage
Thanks for following along.
That space is ready to stack and pack. Nice!
Very happy with mine. Not as clean as your setup because you seem to have more patience than I do lol.
This will be a great guide for someone building this efficient and reasonably priced fixture.
One thing that could’ve helped tidy up is multi core wire.
I’ve got a bunch of 26g 4-core telephone wire. Color code makes it all a bit confusing though, but does reduce visible number of wires… if I didn’t already own a few hundred feet I would maybe just cable manage with zip ties, but Nube’s 2 core seems to work ok.
I think if I’m willing to use Pro-co 12-2 speaker cable for the power cable I have enough spare cable around to wire a panel or two. Just tempted to maybe not make it sketchy; since I actually plan on putting it on a 220v breaker in the future, ideally it would all be up to code and just left as a 220v outlet, mother’s house so not going to cause her issues if she ever decided to sell.
Although, I can always fix it later… I think I might have to do a “spare parts” build just to prove it could be done. Honestly I’m looking at a ROI on electricity in a few months so it’s really looking appealing. I was going to try this paycheck; but as per usual things came up. Primarily all I need to buy is the strips and driver- that’s the biggest chunk of change. I considered cutting from 10 to 5 strips; and adding more light later, but that doesn’t make sense economically. (Right?) I think you’ve pretty much scraped to the bottom of the cost cutting, Nube! I’m even considering buying some to diode swap and use for another project needing IR illumination, but the metal core PCBs are so damn expensive to an industry outsider. My boss works his second job as an electrical engineer, he said he’ll look out for scrap if they ever work with it again. If you can source MCPCB and are able to mill and solder it you’ve got your own panel to mount to. Anyone else interested in that much work? Potentially could order some boards at bulk price. That removes the wiring, adds heat sinking, and does add rigidity.
But in the end how many people want to do it? Probably just me, but thought I’d share. I get good ideas from other builds, hope to give something back!
I should add; I have researched DIY metal core PCB and it’s a fucking nightmare cost wise. Essentially the only economic solution for a simple circuit like LEDs is to literally make it. You need copper and aluminum plate, as well as a dielectric adhesive. There’s a German company that makes some good adhesives, but I could never figure out how to source it. Funny enough my friend works for them; just does safety inspections and such. Had no recommendations.
Fantastic build and an even-better write up. I do have a question: How long until you and the smart others who’ve built these lights collaborate to make an official OG LED build guide? Lol
Thank you for taking the time to not only describe what you built, but to include the thoughts behind the decisions you made. There is so much information out there - to your point, much of it is misleading - it becomes difficult to wade through the bs.
I’m hoping to design and build a solid light that is easy to replicate as needed. I love the Bridgelux strips but I also really like the SolStix X3. I’m drawn to the SolStix because the two types/temps of the LEDs are on separate circuits; allowing me to precisely dial in the light temp most favorable for any particular cultivar.
I realize my build won’t be especially cheap but I want them overbuilt and solid. @nube - I will likely be bugging you with questions. Apologies in advance.
Nube, where did you purchase the PA66 three wire connector?
Nice work BTW!
That’s always been an issue for me prior to building the exact same fixture as Nube. Too much bullshit and guys having piss contests. Nube cut straight to the chase and answered all my questions as best he could