DIY LED: The Definitive Discussion

[Aristoned]

Good morning, OG!

I’ve decided to start a thread to discuss all things LED (COB, strip, star, SMD, etc.) and how we can benefit from the high-emission and low-heat option of illumination.

We all know there is a giant leap in fundamental photon production, yet there is always a trade-off when we convert energy from one form to another. The biggest difference between the old HID’s and the LED’s is radiant heat.

This heating of the grow-space is necessary, to a point, which is why we see so many LED operators opting for much higher temperatures than their HID counterparts, saving on air-conditioning which is a huge expense.

Another reason LED’s have become so popular is the longevity, efficacy and customization. With a HID we are allowed only a single point of light where the LED’s can be spread to the absolute extreme’s of the grow-space.

Now, we have covered the features and benefits let’s focus on the building of the luminaires.

When I think LED I think Bridgelux.

No, this is not a sponsorship or a paid advertisement. I purchased Gen 7 SE Vero’s about a decade ago and I was pleasantly surprised by the power @ 1.05A on passive heat-sinks, which work a lot better when actively-cooled. Since then, I’ve built two more luminaires using the Bridgelux EB Slim Gen 3 strips. Each luminaire used 4000K @ 80 CRI for vegetation. The 340mm luminaire is 100W with the XLG-100-L and the 590mm luminaire is 150W with the XLG-150-L while the Gen 8 SE Vero’s that I’ve recently upgraded to are using the XLG-240-L.

Meanwell makes an excellent Constant Power driver. This means, the driver will adjust voltage and amperage to achieve a set wattage instead of a set current, with a set maximum current. I’ve found these drivers to be plenty efficient on 120V, much better on 240V (everything is better @ 240V, everything). Not only that, I don’t have to worry about leaving any “power on the table”. With the HLG drivers I have a set current and my array will determine the over-all wattage. It always annoyed me having a 150W HLG-120H but only being able to use ~140W.

The Constant Power drivers have allowed me to be able to “hot-swap” drivers on the large vegetation luminaire. I’ve tested it @ 150W/200W/240W with amazing results at each level.

Also, the Bridgelux EB strips do not require heatsinks to operate. I’m using 1/16” thick aluminum as backing and a fan above blowing downwards. The strips will operate at higher temperatures, but if we keep them near or under 45° C then we maintain a 90%+ emission factor.

Same issue we have with the drivers. All drivers will generate heat, if we keep enough airflow this will not be a problem. The XLG-100 has an issue losing efficiency at 100% power that causes an increase in heat generation decreasing driver efficiency. This isn’t so much of a problem on the meter-side, but the thermocouple will trigger a shut-down until the driver cools down. Fair warning.

There are three more luminaires I will be building in the near future using the EB strips. One will be a remake of the 150W and the other two will be identical builds using 240W for flowering. Bridgelux also produces EB Boards, which are just like the QB’s, but I cannot find them for sale anywhere. Once I get my hands on a few boards I will test them rigorously and share the results for the Cannabis Community.

Happy OverGrowing!

Cheers!

[sfzombie13]

you should merge yours with the one already going on. almost 3k posts on that one

[Aristoned]

There are a few other threads about LED, but the discussion was more focused on the newer builders instead of a discussion among the more experienced builders. There was an opening in the forum where builders can discuss more intricate ideas with a focus on fundamentals and theories. It is good to have threads that will help the beginner become the novice and eventually a master builder, but we also need a place where people that don’t need to learn about wire gauge and “where can a find a potentiometer for my HLG?” posts.

The posts asking simple questions are valid and they have a place, what we don’t have is a place where we can discuss ideas that haven’t been discussed yet.

Cheers!

[sfzombie13]

they like to merge posts around here all the time and i was just pointing out how similar this one was to the other one. there are many of them, but these two are nearly identical in terms of things discussed. presonally i hate the merging of them since you have to wade through thousands of posts to get to the fresh stuff if you just start, but it’s the way of the forum and i just accept it. good luck with this one.

[Foreigner]

The technology has progressed to the point where the cost benefit of building your own doesn’t jive.

You’re likely better off doing good research and finding what already out there.

Several years ago I did the math. I almost built one. But I bought one instead.

But there’s no question you could build your own quality fixture in just not sure if you should.

[Aristoned]

The 100W Bridgelux EB Slim luminaire cost me less than $100.

The 150W Bridgelux EB Slim luminaire cost me less than $150.

The 210W Bridgelux Vero SE luminaire cost me $220.

The 240W Bridgelux EB Slim luminaires cost me less than $240 each.

As it stands, I can build a better luminaire than I can purchase. Perhaps your math was off somewhere, I’m getting outstanding results.

Now, if you have something to add to the discussion that isn’t “don’t do it because I couldn’t do it” then the Cannabis Community would like you to share. Otherwise, we would prefer to not scroll through thousands of posts telling us not to do something we have been wildly successful with.

Cheers!

[Foreigner]

@Aristoned

I’m getting outstanding results.

Well good. Carry on.

[BigF]

Call me a diode snob, but dont want my girls under anything but Samsung 301B or 301EVO. It just doesnt seem like the Bridgelux chips deliver the same quality end product. However, my environment is pretty tight, so the differences between chips may be more readily apparent.
Am quite interested in the DIY aspect. I have a Spider Farmer 8000 with a sagging power supply, had to take it down. Would you have a recommendation on where to find an 800w driver replacement?

[Bobbles420]

@BigF

Would you have a recommendation on where to find an 800w driver replacement?

Try rapid led, they sell all the parts to build your own lights and carry meanwell drivers in a few sizes and options. I built quite a few lights from them for my reef tanks back in the day. Link should take you straight to drivers but you’ll have to filter for dimmable cuz I assume that’s what you want.

All Drivers

[Pigeonman]

@Aristoned

As it stands, I can build a better luminaire than I can purchase. Perhaps your math was off somewhere

Are you in the @Aristoned ?

@Foreigner and I are in , so often our issue is the the math we are obligated to use is marked up at a premium by comparison to distributors.

This is what I found when I looked into it as well.

Additionally, when I do this kind of math I also account for my time as part of the value. So with parts and labour I found very much the same issue with it being an overall better value to purchase my fixtures outright than build them.

However there are solid OG’s here like @Mr.Sparkle and @JohnnyPotseed who run with DIY fixtures and have killer plants.

My take-away is if it’s cost effective and safe to DIY then it is a better option that paying for someone else work and mark-ups.

[TeddyNuggets]

@BigF

Call me a diode snob, but dont want my girls under anything but Samsung 301B or 301EVO. It just doesnt seem like the Bridgelux chips deliver the same quality end product. However, my environment is pretty tight, so the differences between chips may be more readily apparent.

I was so excited to try a LED light with the new Samsung EVO diodes. I got one of the new AC infinity lights with the evo’s but I haven’t been that impressed so far. Part of my issue may just be that my environment isn’t dialed in well enough for the intensity of the EVO’s … ?

I know Samsung diodes are widely considered to be the best but So far my favorite diodes have been the Osram and LUMILED diodes that my photontek lights use. I get the best quality and results with the least amount of light stress using those lights vs my other LED lights.

I don’t know if I’ve tried any LED’s with bridgelux diodes.

[Aristoned]

Samsung uses more blue light in their LM301H EVO diodes, doing that will increase PAR. Bridgelux 4000K strips make ~10% more PAR than the 3000K if I recall correctly.

Samsung uses the 5000K blue chip for their efficacy, which isn’t entirely honest since I will never use the 5000K in my garden.

There is absolutely no magic here, they just play with numbers. If you pay attention to the Bridgelux 4000K strips you will see gobs of red light, which our plants use a ton more red than blue.

[Aristoned]

There are cost differences due to International Banking, we all have different situations and I’m paying tariffs on a few things. I get it.

The point of the thread was to discuss DIY LED at a deeper level than “where can I find a driver/potentiometer” or “how do I solder” or “what is series and parallel”.

The idea here is for high level builders to discuss fundamental operation and design to improve upon what we already have at-hand.

For instance, AC Infinity has a 2x4 LED luminaire for $310 @ 280W. I’m building two 240W luminaires for less than $300 to cover the same space with an increase of 200W over the Samsung diodes as well as all of the light is in the 3000K spectrum.

[Aristoned]

I’m using Vero SE COB’s, 2700K @ 97 CRI, and EB3 Slim’s, 3000K/4000K @ 80 CRI. The 4000K EB3 Slims have the best vegetation spectrum I’ve seen available, @ 80 CRI they are producing well over 200 lm/W which translates to high PAR numbers as well.

I tested my (x12) 590mm EB3 Slim luminaire @ 240W producing over 1,000 ppfd @ 18”. Now, those number sound great, but there isn’t enough red to make a good flowering option, although I’d imagine the bud would be excellent quality it would just be a bit light.

The Vero SE chips I’m using will make 1,000 ppfd @ 12”. There is a trade-off for light-accuracy.

There are Samsung LM301H EVO diodes in the humi-dome, they make ~400 ppfd @ 100%, but they only have two red diodes per bar.

[Mr.Sparkle]

Couple things to point out, Relative intensity graphs only show how a spectrum is distributed vs what is the max intensity, which is different for each color temp, so one could have a 5000k graph look similar to say a 3500k with less blue but in reality alot more blue is passed over than what the graph may suggest, also the further blue you are the more the PAR efficiency decreases compared to the red side like 2.8-3.2 par efficiency but at the same time you also see higher light outputs on the blue side because the underlying chip is blue just with different phosphor coatings ratios so usually the PAR efficiency vs actual output kinda becomes very similar.

Chip wise its actually the 301H evo “mint white” chips that are the top atm, reason being is they have shifted the blue wavelength creation to be slightly more into the PAR range “cyan” for better efficiency vs the standard “blue” 470ish nm region. Higher end fixtures use this with high bin numbers and augment with red diodes to get really high PAR efficiencies

Also cost for building a fixture one has to figure in shipping costs, materials for any frame or construction, and ones time, im not saying one can’t make equal or better fixtures for the same cost as commercial options, its just that with light sales and for the average person buying a prebuilt fixture is just easier and worth a couple extra $20’s.

The benefit of DIY route is if one has a custom space that your trying to optimize, if its a tent well commercial options are a heck of alot easier.

Also look into the Bridgelux F90’s

And slims vs reg if the same gen ill take reg any day just due to greater ability for heat dissipation, and no need to pack them in so tight. as quoted below your still using a heat sink cause the slims do run hotter cause of there size.

@Aristoned

Also, the Bridgelux EB strips do not require heatsinks to operate. I’m using 1/16” thick aluminum as backing and a fan above blowing downwards. The strips will operate at higher temperatures, but if we keep them near or under 45° C then we maintain a 90%+ emission factor.

[Aristoned]

LED diodes are coated with phosphorus to filter out blue light. The more phosphorus the less blue light and reduced PAR. Samsung uses their 5000K chip for their efficacy numbers, which doesn’t tell the truth when we look at the 3000K, since the numbers are lower and the 3000K is what we want for cannabis.

This means, a 5000K won’t have a similar graph to a 3500K unless the 3500K has a color temperature shift above nominal currents.

The “mint white” diodes have a worse spectrum than the 5000K.

The EB3 Slim’s have no problem living at 20W on the 590mm. Also, I don’t use any heatsinks on my Bridgelux EB Slims builds.

I repeat, I don’t not use heatsinks on my Bridgelux EB3 Slim builds running @ ~1.05A.

This is what it looks like under Bridgelux Gen 8 Vero SE COB’s @ 2700K/97 CRI. There is a big hit to PAR due to the phosphorus coating, but I’ve never seen an LED so red before these.

PAR numbers are great, my 4000K luminaire will pump out over 1,000 ppfd @ 240W, without heatsinks, but the spectrum isn’t conducive to flowering as well as the 3000K/2700K. However, the 4000K should produce very tasty bud but very light bud.

@Mr.Sparkle

[Aristoned]

It is interesting to note:

Samsung LM301H 3000K - 2.6 ųmol/j

Bridgelux EB3 Slim 2700K - 2.6 ųmol/j

Based on the spec sheets the Bridgelux chip is actually far superior to the Samsung. Not only do they match power out-put; the Bridgelux has a much more preferable spectrum for cannabis.

Samsung diodes make great illumination for all plants, Bridgelux makes great diodes for cannabis.

[Mr.Sparkle]

@Aristoned

LED diodes are coated with phosphorus to filter out blue light. The more phosphorus the less blue light and reduced PAR.

thats not exactly how phosphor coatings work, depending on the phosphor or more accurately the combo of phospors used they dont block the blue light they convert some of that blue light or energy into wavelengths else where in the visible spectrum based on the phosphor used as when excited by said base chip blue light or well energy they phase change the energy to emit into different wavelengths and the further away from that base blue or base wavelength used the more loss to the conversion due to the wavelength change. Easy to push a 100lb rock 1 ft, take more work to move it 100ft.

Also a sheet of aluminum in which your strips are mounted to being cooled by a fan is in definition a heatsink, and is what i ment by my previous comment.

The mint white chips aren’t really ment to be used alone as previously stated their benefit is when augmented with red leds providing higher PAR levels than what would be achieved with the blue base chips aka older 301B for example

@Aristoned

PAR numbers are great, my 4000K luminaire will pump out over 1,000 ppfd @ 240W

also spot brightness and distance from light source vs actual area brightness and whats actually usable such that like over 650-800ppfd unless augmenting with co2 in a closed environment your at a point of diminished returns.

also as a bridgleux fan and proponent for many many years, samsung actually has the better chip and its been that way for years, bridgelux’s chips are always a step behind in that 5-10% less range efficiency wise, its just for price/output metric bridgelux is better.

Not like there isn’t an entire thread that wasn’t devoted on the topic since 2019

Cheap LED Strips : A Viable Alternative

Well figured id start a thread just to share for anyone thinking and looking for a viable alternative to SIL “Screw In Light” Led arrays to rigid DIY options such as QuantumBoards, Solstrips, Meiji Lighting products, or any pre built options out there like marshydro, and the likes. I’ll leave this thread open to discussion and post ups of other cheap strips and product lines as well because tech is always changing, and getting better and cheaper everyday. So to get this started let me introduc…

[Aristoned]

The point was missed here.

Samsung’s claim to fame is a blue LED.

Cannabis plants want red LED’s.

Samsung’s 3000K is 2.6 ųmol/J and so is Bridgelux’s 2700K.

If you compare the 5000K Samsung to the 4000K Bridgelux you will quickly see which is the better spectrum, yet the blue light still makes more light with 80 CRI.

Dirty blue light.

The 240W 4000K is just ridiculous. I was laughing so hard when I saw the numbers. 1,000 ppfd from just twelve strips.

Now, when I stepped up to the 2700K/97 CRI there was a noticeable decrease in PAR… but the light is red like an HPS, or the rising/setting sun. Very nice.

I’m sure people have discussed inexpensive options to build, this is more about how hilarious it is to see how much we can pull from these LED’s without using any heatsinks.

I just can’t see how everyone missed this.

[Jonesfamily7715]

All for the new diy thread, seen a lot of your stuff on ic mag on rocketsoul thread, Diy is only more expensive from ready made kits and high profiting suppliers like digikey and mouser. You just need to know where to look. I can build a high quality light using all cree and osram diodes for 1/4 the price, and thats only cuz driver costs are a bitch atm.

What yall plannin to do when samsung drops out later this year? Overrated, Imo there are many better options. Hype can carry you only so far. Personaly I try to avoid them, but i do have some strip builds