2x Vero 29 COB + 1000x individual diode combination - Request advice/input

Hello RIU,
I'm beginning to put together equipment for my first grow. Probably not the best time to research experimental lighting, but screw it, I like electronics projects, so I'll keep on.

So far I'm leaning towards using a combination, of individual diodes and Xlamp or Vero COB lights, in order to strike a balance between cost and efficiency.

Parts I'm thinking of using:
-2x Vero 29 - 4k 80CRI - $36.99ea (haven't looked at drivers for these yet)
Datasheet link

-6 Channel Led Driver MC34845CEP (using several to power the individual diodes) ($1.38ea)
Datasheet link

-1000x 5mm - 624 to 630nm - led diodes $0.04ea ($40)
Datasheet link

Electrical and optical characteristics（Ta = 25℃）

Forward Voltage
VF IF=20mA ---- 2.0 2.4 V
Reverse Current
IR VR=5V ---- ---- 30 μA
Dominate Wavelength
IF=20mA ---- 624 ---- nm
Peak Wavelength
IF=20mA ---- 632 ---- nm
Spectral Line half-width
IF=20mA ---- 20 ---- nm
Luminous Intensity
IF=20mA 3000 - 5000 ---- mcd
Viewing Angle
IF=20mA ---- 12 ---- deg.

The 1000x 630nm diodes should draw about .0458W per ea (2.4V * 20ma), and make 5-8lumens per ea.
So 45.8W + PSU losses should give me 5000-8000 lumens from 624nm-632nm

The 2x Vero 29's (BXRC-40E10K0-L-03) should draw 173.6 (41.3V * 4200ma), and make 15210-19070 lumens ea.

So I have a total of 393W draw + power supply losses for 35420-46140 lumens.

The 1000x led diodes 624nm-632nm will make up, between 10.84% min and 22.59% max, of the total output. (Min= 8000lumens/35420lumens Max= 5000lumens/46140lumens)
The 4k COB's will make up the remaining 77%-89% of output.

How are things sounding so far?

So you're planning on really maxing out those vero's....
I wonder what that does in terms of effrciency.
Also, looking at the data sheet the forward voltage was 38.6 at 2.1a was wondering where you got 41.3..not that it really matters. I almost wonder if running three cobs a little softer instead of two at the max might give you more lumen per watt. not that I really know what I'm talking about, I'm in the middle of building my first like myself, ordered all my parts last night.
Good luck!

Oh Ouch: My best digikey find for COB drivers was this guy
http://www.digikey.com/product-detail/en/LDB200-048SW/633-1191-ND/3078550 24V-48V 4.25A $110.35

I'd need two of them at$110ea, that's the cost of everything else combined.
Anyone have experience designing high voltage, high wattage constant current sources? Or have a cheaper prefab source?

Maybe someone could help me take a PC PSU and step up the voltage/regulate output current on the 12V.

This one:
http://www.newegg.com/Product/Product.aspx?Item=N82E16817151119
is 80+ Gold (should be over 87% efficient at any load), single 12V rail (45A continuous on one output rail, decent brand so there should be low ripple)

Any thoughts?

Hi tightpockt,
I got those #'s on the COB from table 3 in the datasheet (Product Performance atCommonly Used Drive Currents):
It lists 41.3V@4200ma:
Typical Flux @25C Tj: 19070 Lumens
Typical DC Flux @85C Tj: 15210
Typical Efficacy @25C Tj: 110lm/W

They have 4 common drive current values, and typical performance values at each, for every bin.

You will need lots of cooling so add more for passive, half that for active.

One dude was doing a water block grow, but I haven't seen an update in a bit. You will need lots to get a 25C Tj @ 4.2 amps. Like ridiculous amounts. Even then...... Why, at least back them down a bit for some efficiency and hence less cooling maybe @ 3?

Obviously if you have cooling figured out... Grow on

As far as power supplies search for

picograv

1st to do a vero grow on here I believe,uses solid state benchtop power supplies [48v+] that could be an option.

How about something like this?
48V/7.36A SMPS
It says it is constant current limiting, but I don't see any current adjustment in the picture, only voltage.

Hooking up 2 vero 29s to that power supply in parallel would be an excellent way to destroy them.

I don't think it is worth running the veros as hard as you plan to 24/7. Cooling them would be no small feat and could possibly end catastrophically (just assuming the worst case, not implying it couldn't be done - albeit, at a relatively large cost both initially, and efficiency wise over the long run assuming you pay for hydro). This is not to mention the degradation of the LEDs. Check the derating curve in the vero datasheet...

But if you must, consider 2 of the inventronics euv-200s048st(sv) to power them at 4.2 A.

All of this just IMHO.

One things for sure, they will be damn bright at max drive capacity. It's so ballsy I almost want to see it done, lol.

There is also the Mean Well HLG-185H-48A, 3.9A dimmable. But like abi and overthinker said, consider running softer, with more COBs to spread both the heat and the light.

The Vero 4000K is excellent by itself btw and doesn't really need a thousand little 5mm diodes.

I'll probably add more COB's in the future. I'm thinking of arranging the red led's into 4 250 led groups, each with a COB in the middle.
Obviously at first it will be 2 COB's + drivers short, but I can add them later as funds allow.

So basically four sections laid out like this (with only two containing COB's initially) :
LED=*



___________________
|
|*************************
|*************************
|*************************
|*************************
|*************************
| | COB|
|*************************
|*************************
|*************************
|*************************
|*************************
|___________________

Sorry I should have been more clear. I was considering 2 of the 350W meanwell units (41-56V)
I'd run 1 COB 41V/4.2A per each. They're $50ea and exceed my current requirements.
Is there a good way to add the constant current regulation to each?

Everyone has been so worked up over the 4.2A draw on the COBs that OP is proposing (which I don't think he is listening to...,) but I have not seen anyone mention anything about the 1000x630 red LEDs.

I have no maths to back me up on this but I am pretty sure there would be a more economical and effective way of doing the reds. $40 worth of 4c diodes is not going to be easy to work with, or provide anywhere near enough light output, provide the plants with any sort of light intensity if you combine it with the COBs. The COBs will need to be at least min 12" from the tops of the plant to prevent bleaching and at that distance (you are planning on somehow driving these COBs at MAX, so there will be a lot of light intensity and heat so 12" may even be too close!) these small red LEDs will be worthless.

I am pretty sure that the kind of red LEDs that OP wants to use are the same ones used on those very early china panels. They never really worked at all for anything more than vegging a tiny plant.

OP - I really think that you should actually read the advice that people here have provided you. You are planning on driving those vero COBs WAY too hard. The amount of money and effort you will need to keep that thing even close to cool is simply not worth it. I also believe that the red LEDs you have proposed are also a bad idea. While I appreciate the different approach you have taken, I don't think it will yield you a good result. You are on the right track with the COBs, but consider driving them at 50%. Really put a lot of thought into how you will cool these. The hotter they get the worse they perform. And lastly on the reds, Ill pass on some advice I was given recently as I am also building a new light: Check out these: http://shop.stevesleds.com/LUXEON-ES-3-Watt-LEDs_c27.htm.

There are many good DIY threads happening right now that are fantastic to learn from. I once again strongly recommend checking these out!

Good luck with whatever you decide to do with your project.

^ He'll solder about 10-maybe 50 and be sick of it. I was going to post something, but even if they all worked optimally, did you see the #'s, 5-7 lumens ea. x 1000.

Another $29 V29 would match it.

420Sm0keWeedEveryday said:

Hi tightpockt,
I got those #'s on the COB from table 3 in the datasheet (Product Performance atCommonly Used Drive Currents):
It lists 41.3V@4200ma:
Typical Flux @25C Tj: 19070 Lumens
Typical DC Flux @85C Tj: 15210
Typical Efficacy @25C Tj: 110lm/W

They have 4 common drive current values, and typical performance values at each, for every bin.

Click to expand...

You would be far better off using significantly cheaper vero 18's and driving them at 1.4 amps. It would be much cooler, more efficient, and actually give you more usable light than 2 29's driven at red line. It's going to be next to impossible to cool those 29's to a reasonable temp. Even a TJ of 85*c is optimistic at best. Take a look at the size of heatsinks supraPSL uses for passively cooling half the power. Those cost $6.40 per inch plus shipping several pounds of aluminum. You would need something like that plus a fan to cool each '29.

4 vero 29's driven at 2.1 amps(nominal) would draw less power, put out more light, offer better coverage, last longer, and run cooler. The cost to assemble would be comparable since you could cool it with a $10 CPU heatsink/fan from amazon.

1000 5mm diods will be a pain in the ass so assemble, run hot, and give you spectacularly shitty results. You'll have to mount them within inches of the plant making your COBS ineffective and largely wasted.

But do it, be a rebel, make your own path, and report back.

Thanks for the advice. I guess maybe I don't really want to add an extra 300w of heat to dissipate in my cabinet.

What kind of lumen output could I expect from 4 Vero 29 or 18's . Maybe I could watercool and them and keep a Tjmax of about 30c. I think for that I'd want a TDP of like 100w combined or less.