before I start ordering...

[Rahz]

Before I start ordering parts for my diy led panel I thought I would ask whether anyone would do things differently... either less wattage or more wattage is the basic question. They say 50 watts per square foot, though that number might relate to LEDs a little differently.

For reference the new A51 190 watt lamp at a conservative footprint of 2'x3' provides 32 watts per foot.

My design will put out 51 watts per foot. This is my first LED venture so I don't really have a point of reference.

Thoughts?

 

[GreenSanta]

51 watts per square foot with A51 lamps will give you really awesome results, you wont have to worry that the light might not be good enough, you will be able to focus on your grow techniques.

 

[budbro18]

Should be good. Only thing i can say is consider using multichips. Or cobs. Same thing. Dont have to only use them but maybe put 4 in the center down the middle and then surround it by smaller ones. Either way you should be good if youre using 51w/sqft

 

[Rahz]

Cool, just want to make sure I'm not overshooting my mark. I'm going to use cobs. The plan is to run them at 40 watts spaced out evenly over the whole grow area and I'm hoping to keep the panel 12" above the canopy with this configuration.

 

[budbro18]

Nice! check out my grow journal. I was gonna build my own but then decided to grab some panels that werent too expensive that use cobs.

 

[Rahz]

Nice to see some manufacturers using cobs.

I didn't even consider a diy project until I learned about cobs. Soldering several hundred stars isn't appealing. My plan calls for 18 lamps. The emitters have plug style connectors and I could make the other connections with wing nuts and a crimper. I'll probably end up soldering the driver leads but it's not a necessity.

 

[budbro18]

Nice to see some manufacturers using cobs.

I didn't even consider a diy project until I learned about cobs. Soldering several hundred stars isn't appealing. My plan calls for 18 lamps. The emitters have plug style connectors and I could make the other connections with wing nuts and a crimper. I'll probably end up soldering the driver leads but it's not a necessity.

Agreed. Ive always wanted to make a DIY led panel but like you said fuck wiring that many stars.

I wish i knew more about electronics because there has to be a way to cut down on # of power supplies im just not electrically talented haha.

I then switched to a COB/multichip idea but after a little research for best spectrum i found growevolution.com and they were about 100 bucks over my expected panel costs without factoring in time/labor/stress. So i said fuck it. Rather pay for 1 good panel than 1 home made panel.

 

[Spiko]

Agreed. Ive always wanted to make a DIY led panel but like you said fuck wiring that many stars.

I wish i knew more about electronics because there has to be a way to cut down on # of power supplies im just not electrically talented haha.

I then switched to a COB/multichip idea but after a little research for best spectrum i found growevolution.com and they were about 100 bucks over my expected panel costs without factoring in time/labor/stress. So i said fuck it. Rather pay for 1 good panel than 1 home made panel.

I had a similar experience and decided that the price difference was so negligible that I may as well just buy something good. Also it seems to me it's hard to source quality diodes at reasonable prices. It seems impossible to verify if what you're buying is even real.

 

[SNEAKYp]

I just go to the LED manufacturers website and find all the authorized sellers and start there. They are usually reputable
Rahz! Glad to see your getting started on the DIY panel. Good luck dude, not that you need it. Those plug connectors on the Vero's are pretty fucking cool. Solder-less sounds like a piece of cake when it comes to installation.

 

[budbro18]

I had a similar experience and decided that the price difference was so negligible that I may as well just buy something good. Also it seems to me it's hard to source quality diodes at reasonable prices. It seems impossible to verify if what you're buying is even real.

For real. The difference in price was made up by a 5 year anything goes warranty (besides physical damage caused by my dumb ass) and along with a solid case, assembled by people making them regularly, less risk of fire due to my dumb ass wiring. The cons started quickly outweighing the pros and every con had $$$$ attached to it. haha

But for someone with enough electrical knowledge its probably one of the easiest things ever. Series, parallel, amperage, voltage. Its all pretty simple math and shit ive just never been one to delve into electrical workings. To much of a risk of fire, shock, shorting things, etc...

And after enough research i found an LED manufacturer who didnt want your first born for a couple solid panels. hahahaha

 

[Spiko]

But for someone with enough electrical knowledge its probably one of the easiest things ever.

LEDs are about as basic as it gets. Building one is well within my technical ability it's just I'm not convinced it's actually cheaper. I found myself wanting to copy area 51 and it quickly became apparent I should just buy their fixture. If I could build something similar for cheaper I would have.

 

[Rahz]

The prime reason I decided to go LED is the light spread. I would rather have a blanket of even par light than a small panel doing the same basic job a HPS bulb does, emitting light from what in practical terms is a single point with a hot spot at the center.

It's going to end up being cheaper than my original estimate. Cob, driver, heat sink and connector for a 40watt unit runs around $50 + shipping. Additional costs include aluminum for the frame, nuts and bolts, a spool of wire, a handful of terminal strips, a yard of shrink tubing, outlet cord, solder, hanging hardware, tap bit and handle, some screws, thermal paste, maybe a fuse block. It should end up being around $1.50 per watt. For that price I get the light spread I want and passive heat sinks. I'll order a couple extra cobs and drivers so I can do a quick repair if a unit fails.

All the other options at recommended usage would provide around 600 watts in my space. This includes the Neosol panels which I like due to the passive cooling. They're pretty expensive though. But that's the reason I wanted to ask about the wattage per foot. Nobody is trying to offer a 50 watt per foot unit. The disappointment would come when I realize the panel I've built is too powerful to keep 12" above the canopy. It wouldn't be a major problem, but it's something I want to consider before taking the plunge.

 

[Greengenes707]

LEDs are about as basic as it gets. Building one is well within my technical ability it's just I'm not convinced it's actually cheaper. I found myself wanting to copy area 51 and it quickly became apparent I should just buy their fixture. If I could build something similar for cheaper I would have.

Very true. Bulk discounts make a big difference. Even the largest DIY panels can't get the price break a company spending $20,000+ at time does. Plus all the other components too.

 

[MrFlux]

...50 watt per foot...

That would work, it's about 500 W/m2, assuming an efficiency of 35% it becomes 175W/m2 of pure light or about 800-900 umol. A low target would be 400 umol and a high target 800 umol so you would be aiming high.

So what COBs (brand, size, color temp) and drivers are on the order list?

 

[Rahz]

Thanks for the umol estimate. I've chosen the Bridgelux Vero 18, paired with Mean Well LPC-60-1400 which will drive them at just over 42 watts. I'm using 2:1 ratio 2700K:3000K. I've seen a 2700K only grow and was impressed by the lack of stretch so I'm not convinced I need to include cool whites.

Any thoughts you have about the cobs/drivers/color temp/spacing would be appreciated.

 

[wimpy69]

I've seen a 2700K only grow and was impressed by the lack of stretch so I'm not convinced I need to include cool whites.

Probably it's the bbspils grow you are referring to. Keep in mind that

1. He supplements with an UV-B reptile light (I think) which would compensate somewhat for the lack of blue;
2. He has Cree CXA which have a lot more blue (and less red) than the Bridgelux Vero at the same stated color temp;
3. Picograph had a plant under a 2700K Vero and it stretched.

So my thought would be to go a bit higher with the color temperature. I have some 4000K Vero 10's and they are quite yellow, it looks about like the sun when it's 10 degrees above the horizon. This 4000K Vero is actually quite a bit warmer than the 3700K Cree XT-E as used in the A51 panel.

One of the bigger problems will be heat dissipation. In your current design it's about half a kilowatt of heat and you want to do it passively. Before plunging all your money, why not start out smaller, see how things work and then scale it up.

 

[Rahz]

Yes, it's bbspils 100W led thread. That is certainly worth thinking about. I will reconsider my options.

I had considered starting out smaller, but there's no way to easily split up the light grid as I have designed it. I'm using 50 watt radial heat sinks, about 15 lbs worth of aluminum, and plan on using thermal paste as opposed to adhesive or pads. The space is well ventilated...

But I think you're right about testing the thermal dissipation. Instead of starting out small I'll just purchase 1 cob and driver and test it on one of the heat sinks before committing to a large order. Solid advice, thanks.

 

[SupraSPL]

3000K Vero18 @ 1.4A is about 42W as you said. Radiometric efficiency is about 30% @ Tj 50c. So that is about 115 PAR watts/meter2 or 550 umol using Mr Fluxs estimate. That is still well within a good range for nug size and definitely a good bang for the buck.

Here are some very cheap 1600mA drivers. The vF range seems like it would be perfect for Vero 18. The more vF you load onto these drivers, the less mA they put out.

 

[SupraSPL]

If you want to add some blue, maybe the Luxeon M would mix well with the Vero. If you drive it at 700mA it is 57.7% efficient and at 1050mA it is 53.3% and dissipates 12.2 watts. Cheapo driver

 

[Rahz]

Wow, that 50W driver is cheap! I could easily push the Vero 18s at 50 watts but I'm not sure the 50 watt heat sinks would be up to it. I like the idea of keeping the amps down anyway. I've considered running 23 units @ 31 watts rather than 18 @ 42. That would result in 60 watts less producing around 6300 additional lumens. It would bump the price up by $250 so I would never reclaim the cost in electrical expense, but the extra lumens would help justify the cost, and surely the 50watt heat sinks would handle 31 watts.

Those Luxeon Ms would mix well with the 2700K Veros.

Also, looking over the Cree-vs-Vero conversion charts MrFlux provided, the 4000K Vero is looking good. They might complement the 2700Ks better than the 3000Ks would.

I'm going to have to let this stew in my head for a day or two.