Let's talk DIY lights

Hi everyone!

Just thought I’d start a new thread for us to further the diy light discussion. I don’t sell lights, lighting components, or work in the lighting industry. I have worked professionally with a few of the components, but in the green energy field. I got involved with led grow lights repairing/modifying basic amazon lights a few years back. Since, I’ve kind of become a local 420 friendly electrical support guy. I have access to a handful of different grow rooms utilizing different lights I have built or assisted building. That being said, I’m no expert and definitely don’t have working knowledge of every manufacturer or component out there.

Let’s try to keep this fun and informative. It’s not about bashing on budget lights or different types of lights. In fact this doesn’t even have to be limited to led tech. If you’ve custom made a cfl reflector and want to share, post it up! No question will be to basic, and no stone unturned. So please ask, or offer up the info you have. I’m going to tag the people I know have shown interest, but feel free to add anyone. I thought I read that a couple of the members were working on a retail light, and I definitely don’t want to interfere with that. If you guys are out there reading, please feel free to add or abstain if it’s in your best interest.

@ktreez420 @Countryboyjvd1971 @bob31 @Myfriendis410 @MAXHeadRoom


I like it brother thanks for starting a new thread on it
We all have a budget to work with and we all work in our means so this is a good way for people to learn a bit more :+1::v::cowboy_hat_face:
Nice job I’ll be following along and will add to discussion when I feel I have a good point or question for sure
Thanks again for the tag @dbrn32 :v::+1::cowboy_hat_face:CB


Great idea! As this tech is evolving it’s nice to have a venue to discuss pros and cons of the different types available and coming. Thanks for including me @dbrn32!


WOW this is great thank you. I will go first. As you all may know when I first started out I purchased 4 no name lights from ebay. of varying wattage from 600 to 1200 watts. They boast a double chip design but I believe its only a 5 watt diode. The actual wattage is 78 to 118 respectively. Anyway can I change the drivers on these lights and make them better or is it a waste of money


:joy:well I found an old 4-bulb vanity of my girlfriends, and I took it apart and cut the reflector some and flattened it down with a ballpene hammer. Then I ordered four 18w led bulbs from eBay and hung it in my tent. I also have 3 other fixtures with CFLs in them. I feel like it is ghetto rigged for sure, but it seems to be working. Those lights were on sale for like $1.59 from China so I ordered 6 more of them. Here is a picture of the inside of my tent.


i like it when a plan comes together…


This is awesome @dbrn32, thank you for starting this thread! I will be following it closely!

@MAXHeadRoom GREAT question dude! I would love to know if, and how, to upgrade the LED’s I already have!

I have tons of questions for you DIY guys!



@MAXHeadRoom you could change the driver, or the lights most likely. It will really come down to how much information you can get about the respective driver and lights. If you can find out exactly which led’s they are, we should be able to track down some info. We would be mostly looking for forward voltage and max current of the led and voltage range, current, and type of driver.

We can gamble and assume things like a 5 watt diode is going to have a forward voltage of around 3v at max current of about 1500ma. Just like a 10 watt diode will have fv from about 3.5v-3.8v and a max current of about 3000ma, or 3 amps.

Those are just the industry standards, and exceptions always apply. That’s why I said it’s kind of a gamble moving forward without all the info. The more info we have, the more confident we can be moving forward. A minor tweak like changing a 500ma driver to a 700ma driver is fairly easy, cheap, and worry free boost in output. Going big can sometimes end up in failure.

Something else to consider, is modifying existing case and heatsinks into a cob build. I’ve done that a couple times.


@Sharkslayer I dig it!

I have done almost the exact same thing using a piece of rolled 4" hvac duct.


For any of you that don’t know, the major hang up with a lot of commercial lights is in the advertising. They are good lights, as long as you understand what you’re getting. We calculate watts volts x amps, very simple. In this application, there’s always some efficiency loss. Getting very minor with the right stuff, but there is always some there.

So Mars, viparspectra, hydro galaxy, or pretty much any of them build a light. Let’s say they use cree xm-l2, they don’t but we’ll say they do. It’s a 10 watt diode, fv of about 3.5 with max drive current of 3000ma. Doing the math it’s a tad over 10 watts, but we’ll stick with the 10. So then they call their guy in China and design a pcb with 100 xm-l2 diodes. A hundred 10 watt diodes is easy math right, 1000 watts!

But we know from the cree xm-l2 data sheet that running anywhere near 3000 ma on this diode will take excessive cooling and still shorten the life of the diode. So let’s run it at 1000ma, put a fan on it and ship them out. There’s still 100, 10 watt diodes so let’s call it 1000w. Somewhere in the description we’ll put that it draws 350 watts and call it good. 310 watts of led light and heat loss, 20 watts of cooling fan, and 20 watts of efficiency loss from the driver. And I’m being real generous with the driver and fan loss.

Now, nobody in their right mind can you they don’t grow. Look at these grow journals, proof is in the puddin so to say. But let’s be honest, how many of us thought we were getting a 300 watt light for $70??? I for one! And to tell the truth, I’d buy another under the right circumstances. But I would understand exactly what I’m getting now.


The main difference between old school and current tech is the shift from colored to white diodes. The diodes of 15+ years ago would never compete with the photon density of the hps lights. So instead, grow light designers targeted peaks in the McCree curve. Enter your blues and reds. They were pretty successful, so for some time making these better was the area of focus. As the technology grew, so did the ability to create photons while maintaining efficiency of led power. Now we create more efficacy with white lights that can cover the McCree curve, which is exactly what brings us together on this thread. There are plenty of manufacturers that we can buy kick ass lights from, at a major cost. But you know we all want more for less… so let’s build them!


It looks like my thermal tape will be delayed, so I’m a few days out on building anything. Sorry!

Since it seems to be a big topic of discussion, let’s throw the cxb 3590 out there…

Before we discuss the actual cob, let’s make sure that everyone understands the diode binning. To the best of my knowledge, diode binning started as a way to separate differently performing diodes from the same line. If one production run proved to be brighter that another, they put it in a separate “bin” or box or shelf for that matter. When another outperformed all the others, it went to the top and so on. So within the industry the top bin were always the most sought after. As the led business became more profitable, the manufacturers were able to increase their research budgets, leading to upgrades in technology. Rather than release a new line every year or sooner, they began to just change the bin assignment for subtle changes to an existing line.

I’m not even sure how many bin assignments are available for the cxb right now, but it’s safe to say several. This becomes important when shopping, because the most efficient bin will be in the highest demand and carry the highest price. Right now the db bin of the cxb 3590 is the highest for most of the temp ranges, and the price is around $40 ish per cob. Yet the same supplier may have a lower bin for $10 less. The difference in $10 is probably about 2000 lumens at test current. Both are good, both grow, difference is performance vs price.

When looking at the efficiency chart for the 3590, you’ll see a % of luminous Flux going up and the different drive currents along the bottom. When reading the fine print it will tell you that it’s a 72 volt cob ran at 1400 ma. The bottom of the chart will tell you it only produces 60% of claimed lumens at 700ma. Yet when you double the current it’s only a 40% increase (rated luminous Flux at binning current). Notice how the luminous Flux doesn’t double when you double the power consumption? This is led efficiency, and you will typically see it referred to as lumens per watt. If you’re still reading, higher or more recent bin almost always means more lumens per watt. The only exception to this that I know of is an increase in cri (color rendering index)


So much info! Keep talking lol!

Tell me about these Bridgelux chips/LEDs if you don’t mind. I’ve only ever really heard about the CXB3590’s and building with them. What are these new quantum boards, and what’s different with them compared to the COBs? I’m going to start asking a ton of questions, so I hope you’re ready lol!

What can I do with $200 for a DIY build @dbrn32? Or is $200 not enough for a DIY? When I priced out the CXB’s last year I think it was going to cost me around $400 to build a 3-4 COB fixture. I’m obviously like you, and looking for the best price for quality and efficiency. Really, I want lights that produce little, to no heat at all, and are good for plants lol. That’s it!



@dbrn32 so Is a mars hydro 300w worth it I was thinking about getting one for side lighting


For side lighting, definitely! I use 2 of the 300w MarsHydro’s in my flower tent.

I just recently purchased 2 Roleadro 300w LEDs though. I’m curious what @dbrn32 thinks of these lights:




The Mars 300 is absolutely worth the price! It’s great at its price point.


@ktreez420 I probably like them as much as any of the Amazon lights. They’re a good value, and I’m a bigger fan of white light than purple. On a personal note, I dislike how it says color temp is 3000k in one spot, and 3500k in another. And it’s a long stretch from a cxb 3070. However, this type of stuff is consistent among most of amazon light descriptions.

Personally, I would use a 3500k from start to finish. But only use a 3000k for flower. They will both go the distance, typically just a tad more stretch than I prefer with the 3000k. So knowing exactly what I had would be key here.


So back to the cxb 3590… I had to do some double checking on prices to attempt to be accurate. After visiting rapidled.com, let’s go through a build for a 2’x2’ room. We’ll target 50 watts sq/ft using about the most common cob build in use today. Also, I used rapid because they are typically competitive, have free shipping over $200 (I think) and their website is easy to use. Arrow, digikey, and others maybe a bit cheaper.

So 50 watts x 4 sq. ft. will make our target 200 watts. We determined 1400ma was a good balance of operating current, I prefer to keep voltage as low as possible while maintaining even coverage and reasonable cost (safety) so we’ll opt for the 36 volt version of the cxb.

36 volts x 1.4 amps (1400ma) = 50.4 watts. Convenient isn’t it! So 4 cobs will get us 200 watts, and spacing 12 in on center will be great even canopy coverage. Let’s buy parts!

4 x cxb 3590 with solderless holder = $162
4 x 140mm passive pin heatsinks = $76
1x hlg-185-1400 = $50
1 x 8’ 1x1x 1/8 aluminium angle $20ish
Misc nuts, bolts, wire, connectors $25
Thermal transfer paste, pad, tape $5

I think we’re looking at about $340. You crafty pack rats that save everything, we’ll say you’re all in for $300.

We’re talking $300+ for a 2x2, does anyone actually like that? Sounds fugly to me, you can kick butt with 2 Mars 300 for half that right?

You are absolutely correct! The gain will be in the light intensity. You can still scrog, top, and train. But it will be by choice, not so you can yeild. Assuming if you’re here, you’ve seen growmau5 and greengene videos. Those guys are yielding 1.5 to 2 grams per watt on half the veg time doing very little to nothing to their plants. Granted they are experienced which will always be a factor.

Are we happy with this, or do we want other options?


@ktreez420 I’m getting ready to call it a night. When you get time, post up some info for me if you dont mind, and we can all go through some options together. Let’s start with the size of your room, and whether you’re looking to veg, flower, or both. Then rate the following items in order of importance please.

Canopy coverage
Watts per foot
Efficiency (amount of light per watt used)

We’ll look at a couple different options and see if we can’t find a good fit for you. If nothing else, I’m sure for $200 ish we can get you going.


dang,they have gone back up…lol good lights!
i have decided not to diy myself some super lighting…
i’m saving up for a p9-xlm2 light from platinum…$999, i was thinking advanced lights,but these guys have me sold… unless advanced calls back,and makes the deal sweeter…lol i’m talking,and working with both companies…
i will use my others for side lights,or back-ups lol
besides the diy lights,if you had a grand for lights where do ya’ll think i could get my best value,on a quality light… @dbrn32 @ktreez420