LED array design, 1000 pcs.

Chargehead · Sep 26, 2003

Hi guys. I have a design problem for you all...

I was recently reading about some NASA experiments about using LED lighting as a light source for horticultural lighting. I.E. growing plants.

Always one to be interested in the new, the unusual, and often the highly impractical... I'm trying to figure out if such a thing is feasible to an individual without the billions of dollars.

Supposedly, each of these LED array's cost around 12,000 dollars. Obviously this would not be something a hobbyist would attempt (at least in his right mind /ubbthreads/images/graemlins/tongue.gif /ubbthreads/images/graemlins/ooo.gif /ubbthreads/images/graemlins/icon15.gif ). But with the falling costs of LED lighting, perhaps that cost assessment is somewhat inflated?

So, theoretically, with the laudable goal of improving my diet this winter, with some nice lettuces and tomatoes (or whatever), I try to accomplish this on a reasonable price basis achievable to a single person.

Being an amateur gardener, from time to time, I know that there are many other light sources suitable to indoor plant growth; at least in the off chance that I'm not going into space anytime soon. But for the sake of curiosity, I'm looking around.

I have currently found a wholesale source of 660nm red LED lights. According to a NASA document, this spectrum is well suited for the photosynthesis process (though less suited than the 680nm or 690nm range). They are rated at 1.85 volts, and 20mA. Light output is 4000mcd (millicandela) with a beam width of 30 degrees.

Overall quite suitable. Though less powerful than some others, we are looking for the greatest total output, with fair efficiency, at the greatest value. These LED's are available at about 14 cents a piece in volume.

1000 of these units is therefore quite affordable.

My electrical knowledge is quite lacking, however. I obviously need a reliable DC power supply (I have figured an output of 33 watts is neccisary, unless my math is wrong), connected to groups of these LED's connected in serial. However, I have found that it is apparently neccisary to maintain a regulated amperage across each of them...

This seems fairly difficult to acomplish, at least to a layman. I shortly investigated maxim brand IC controllers, but what little I understood seems ill suited to the task of amperage regulating 1000 LED's simultaneously, expecially considering the variations between diodes inherent in such things.

Anyway, any insight into this issue would be appreciated. I suppose I may be coming at the issue from the wrong direction...

-Chargehead /ubbthreads/images/graemlins/buttrock.gif

paulr · Sep 26, 2003

You don't need fancy IC controllers. Just a high current supply (a computer p/s with 3.3 volt regulated output would do nicely) and a current limiting resistor for each LED.

Really though, an incandescent or fluorescent (grow lamp) bulb will be a lot simpler and work just fine.

Chargehead · Sep 27, 2003

Well, the theory is that an almost completely usable spectrum of light is produced by the nearly monochromatic nature of LED's. Also, the intrinsic efficiency of them make it far more productive, per watt. Besides, they're LED (Yay! LED! Cool!).

When it comes right down to it, it sounds fun to build an array of 1000 LED's, but unless I did something, ummm, illegal with it... (It's surprising what comes up for a search for 'LED array'), I don't have much of a practical application.

So I don't see myself actually doing this. But for curiosity's sake, I do wonder if it is economically and technically feasible (To me, no less).

Part of this means not using resistive loads for current control, because the heat production and power drain of such a method lowers the efficiency of the system so as to have no advantage over other lighting systems, such as a HID lamp. Anyway, trying to avoid using the resistor method.

-Chargehead /ubbthreads/images/graemlins/buttrock.gif

Chargehead · Sep 28, 2003

Okay. After some further reading, I have realized there are no appreciable benefits from an LED array for this purpose. I've paged through probably 30 or 40 websites on LED properties and applications, and found a few systems that would probably work for the voltage and current supply, without the need for resistive loads. Unfortunately I don't believe they would fit into my budget (less than 100 bucks, about).

Upon even further reading (including Don Klipsteins' font of lighting knowledge; I can't believe I haven't run into that site before), I've realized that this 'magical efficiency', isn't really that. The LED that I was considering apparently approaches 10 lumens per watt, by rough mathematical estimate. This is hardly comparable to either fluorescent or HID's efficiencies, at about 80 and 110 lumens per watt, respectively.

Though I'm still interested in the potential of low-heat, monochromatic, high-efficiency, long-lasting, *somewhat* economical LED horticultural lighting, I think the technology still has some way to go yet... Lumileds has apparently produced a high output LED which is capable of 45 lumens per watt, however this is at the less-than-desirable 635nm spectrum. If anything like this comes out in the 660-690 nm range, and is also affordable at quantities of about 100, well, maybe I'll check out the idea again.

...But if anybody has anything to add, please do. I'll probably check out this thread for a while. Oh, and if anybody knows the peak current (and duty-cycle) for max efficiency on a pulse-driven GaAlAsP LED, let me know. /ubbthreads/images/graemlins/smile.gif

-Chargehead /ubbthreads/images/graemlins/buttrock.gif

LED-FX · Sep 28, 2003

As far as I remember you need down to UVA at other end as well to stimulate vegitative growth, all red would result in very spindly plants.

High pressure sodium ,white SON, is popular choice, SON Agro comes in bite size 400w.Trying to replace the sun is power sapping.

Adam

Chargehead · Sep 29, 2003

No, pure red was not intended. For space purposes, 8-10% blues light output was used, with blue LED's. Due to low efficiency, and high cost (over a buck a piece for 3000 mcd 470nm lights), I was not really considering this method. Rather, a 'daylight' white compact fluorescent was my blue spectrum source of choice.

An article I saw mentioned laser lighting as a light source. That's out there. Apparently very bad. The very limited range or wavelength, or possibly it's coherent-ness, made it very badly suited for growth.

SON Agro... I've heard of that. Kind of like the Hortilux, isn't it? Like HPS with doped gas for blue spectrum output.

Well, this is just for fun. If I actually start some sort of indoor garden, I'd probably use an HID light source, or a fluorescent. LED isn't quite up to it.

I did read that plants grown under nearly monochramatic lights develop with very dark, almost BLACK leaves. That's weird... Come out of the kitchen 'Here you go, Caesar Salad!' and put down a bowl of black leaves. Hehehehe. Ewww. /ubbthreads/images/graemlins/icon15.gif
Almost worth it just for that novelty. Perhaps that aspect is a little overstated, I cant picture black leaves, and I've not seen many mentions of them elsewhere.

P.S. I'm wondering aloud, here, if the coherent light of a laser doesn't work because due to the orientation of the chlorophyll molecule? And if you found out which way the molecule was usually pointed, you could use them efficently? This is making less and less sense as I type, so nevermind. If it is purely a matter of direction, the coherent beam would end up hitting as many molecules from the right direction as a random light beam, I would think.

Well, these kind of ideas would probably be better off discussed in some kind of gardening (or molecular biology) forum, I suppose. I'm kind of OT'ing my own thread here. Any idea's for array design are welcomed and appreciated. Even if plants are left uninvolved, a few hundred LED's in a pulse-driven array would be fun.

~Chargehead /ubbthreads/images/graemlins/buttrock.gif