Instructables

DIY LED Growlights from LED Xmas bulb strings

I just got an Aerogarden for Xmas and it is a pretty neat little unit, but it has some annoying and expensive problems. The biggest one is that it uses CF bulbs that are on 16 hours a day, and they say that you need to replace the bulbs every 6 months, because the spectrum of the bulbs change as they age, making the plants grow less well. I did some research and found that the Weed industry has created a solution: LED arrays with a combination of red and blue LEDs (the color frequencies plants thrive on). These arrays are very long lived and use less than 1/2 the power of the CF bulbs in the Aerogarden. The problem is that even the cheap ones (200 LEDs) are $50, and you would need 2 for complete coverage in the Aerogarden. Then I had a brainstorm! LED Xmas Lights. They are already wired, ready to mount on a panel, and you can get Red and Blue strings of 100 on ebay for about $10 incl shipping. I don't know if they are the right Red and Blue, but it is an interesting experiment. Even if it does not work, you are set for Xmas lights for next year.

mr_man3 years ago
nice idea, but I doubt they're bright enough.

I'm an electrical engineer experienced with lighting and high-brightness LED applications.

For good results, specific wavelengths of light are needed, using high-brightness LEDs drawing 1 watt or more, with proper heat-sinks.

This is a good reference:
US Patent 6921182
http://www.google.com/patents?id=jhAUAAAAEBAJ&zoom=4&pg=PA1#v=onepage&q&f=false

The biggest advantages of LED grow lights are that they will reduce by several fold the electric power consumption, and reduce the waste heat produced. The size will be similar to a CFL setup. The biggest disadvantage is that it will cost at least 5 times more money for the lamp, and this is only if you build it yourself (free labor) with a good set of free plans, and a preassembled parts kit.

I have a nearly completed design that I'd like to make available.
I could probably make a parts kits available on ebay for about $80. Or I could just provide the circuit board for $5 plus shipping. This design would put out nearly 8 watts of radiant light, and use about 26 watts of electrical power. I estimate it would roughly compare to a 54 watt CFL. This is a much bigger investment than an equivalent florescent lamp costing only $10 or $20. However, the savings in electricity would pay for the LED lamp in just 2 or 3 years, depending on your electric rate and how many hours/day the lamp is on. The higher these are, the faster it will pay for itself.

Do you think I should make an 'ible on this project?
Would anybody be interested in building one of these?
Yes, I would be very interested! I am looking for exactly this! Please make the 'ible.
I started the 'ible, but never finished it or the project.   http://www.instructables.com/id/High-Brightness-LED-Grow-Light/
Since then I think the supplier stopped carrying the LED I specified, although there are newer, cheaper, better LEDs coming out all the time.  Available for instance from mouser.com

Also my original design called for ordering a custom circuit board.  While that would give a higher quality result, now I'm thinking that might be unecessarily complicated, or expensive, or not in the DIY spirit.
It may be possible to simply solder the LED to a couple of strips of copper sheet metal, (e.g. roofing flashing), or even a couple of solid copper coins, such as a pre-1981 U.S. penny, of which there are plenty still in circulation.  Total cost, 2 cents!  If it's not flat enough, then it can be sanded down a bit.  The copper serves as both the electrodes and the heat spreader, which is then coupled to the larger heat sink.  If the LED is the type where the heat sink pad in the middle bottom of the LED is electrically connected to one of the electrodes, then you just have to make sure that the external connections match the internal ones so that the circuit isn't shorted out. 
}{itch6 years ago
what might be a better idea is to splash out on some luxeon star LED's, i'm pretty sure that they'd be better in the cost to amount-of-light-ratio.

i've tried messin around with small numbers of LED's for growing stuff:
*blatant plug of own instructable* ;)
http://www.instructables.com/id/Growing-Plants-With-LED-Lights/

i'm hopefully gonna give it a go with some 3W LED's somtime soon.
}{itch }{itch6 years ago
sorry for the double post, just realised $10 for 100? thats like 5quid in the queens England. blatantly a better idea! i think most red and blue LEDs emit light on near enough the same wavelength (about 660nm for red and 470nm for blue) so i reckons they should be ok and not differ so much form the commercial ones. hope it works out.
I WAS WONDERING IF YOUR EXPERIMENT WITH XMASS TREE LIGHTS WORKED OUT? DID THE RED AND BLUE TURN OUT TO BE THE PROPER WAVE LENGTHS? sorry about the caps... you can email me at laydbkleo@hughes.net thanks
you could get a cheap set of lights on sail right now and then take out the other color and replace it with one working strip.