LED Grow lights

Science
#1

Tonight I am beginning my DIY grow light projects. This is for fun, for specific plants in my house, and to practice a little electronics and machining. Eventually, I’d like to turn it into a very functional and aesthetically pleasing project to help the bio group with their lighting goals. I’ll be at the space tonight checking out the Arcade Restoration and Race Car. Email me and I’ll send you my number to coordinate.

Here are topics that I’m interesting in meeting with some people and discussing tonight:

  • sourcing affordable LEDs somewhere other than ebay (preferably locally)
  • sourcing aluminum (preferably locally)
  • thoughts on color spectrum ratios
  • adding in extra LED colors to offset the purplish glow
  • DIY LED drivers
  • Making discrete, possibly hidden lights for smaller plants capable of running off rechargeable batteries for 30 days at a time
  • Making functional, yet barely noticeable lights for indoor trees (I own a lemon tree that must come inside during the winter)
Equipment for electronics lab
Equipment for electronics lab
#2

I’m interested. I have a prior thread on that. What you want to build depends on your use. If you strictly want maximum PAR (photosynthetically active radiation) then your build will be different from something that looks nice (some sort of “white” light). I have a pretty good idea of what LED spectrum and parts you need to maximumize PAR lumens per watt. I have searched Digikey and talked with their lighting consultants.

Ebay has a variety of products but you will likely run into fakes and products that are rated at nowhere near what they put out.

#3

A few weeks ago, I salvaged a bunch of electronics parts from a projection TV that I found in an ally. One of the nicest things in it were the aluminum heat sinks. You might find some good heat sinks from salvaged electronics.

I’ve wondered for a while if it would be practical for us to collect aluminum cans and melt them to produce our own components.

#4

Good idea. Proper heat management is critical for longevity and preventing loss of output with aging lighting LEDs.

We should also save any aluminum shavings from the milling machines. Making heat sinks would be a great application for the CNC milling machine.

#5

I think the milling machine shavings already are saved by the machine shop committee, and used to help fund their committee.

#6

How many lumens are you needing? I have built some high power LED based lights and while you can get cheaper they often are not as good as a finished product. Also one mistake and there goes your $20-$50 LED.

LED Floodlight/Spotlights
LED Growlights
Ebay 400w growlight

#7

@Lampy Those were interesting setups in the links.

I haven’t decided yet on the Lumens. It will be dictated more by my final structures than need, as I am trying to cram as much into the structure of the light as possible while being able to maintain thermals and aesthetic style. I am imagining that I’ll end up providing more than needed.

I am okay with my personal use lights not being quite as good as a finished project. I’m not willing to pay $400 on one grow light; however, I am willing to eventually spend $400 over the next year 2 years playing, experimenting and learning as I go.

The key to being successful with the cheaper LEDs will be to find/acquire/borrow a spectrophotometer. This way as I (or we) purchase cheaper LEDs, we can send them back if they are not correct. I would also want to verify more expensive name brand LEDs as well.

Now, if their is interest in the Bio group having a DIY LED grow light(s), I’m willing to go the route of better products…I think having a nice looking, stylish, finished looking setup will help draw more interest to the area, but i still think it should be DIY.

1 Like
#8

Forgot to add, in my opinion a passively cooled design is a better option even though the heatsinks will need to be bigger. Had a device die due to a failed fan.

#9

I concur on passive cooling when possible.

#10

LED Floodlight/Spotlights2
LED Growlights1
Ebay 400w growlight1

Those lights work out to from 85 to 100 lumens per watt according to the advertisement and they are, I presume, full spectrum. I suspect that they are overstating the lumens. 100 lm/w is about the max you might find in one super efficient full spectrum LED unit and I don’t think I have ever actually seen one. Most are well under 50 lm/w. And if they are lying about that, then they are probably skimping in other areas like heat sinking or LED longevity (may be overdriving their LEDs).

They might still be good lights and maybe even a good value, but something is not right with those numbers.

#11

http://www.ebay.com/itm/MarsHydro-300W-LED-Grow-Light-Full-Spectrum-Veg-Flower-Hydroponic-Lamp-Panel-/390567591853?pt=US_Hydroponics&hash=item5aefa47fad

Here is a brand I have been looking at on ebay. The reviews look good for this company’s lights. The numbers look believable - 7150 lumens for 300 watts or 23 lumens per watt. I was thinking about buying one of the smaller models and evaluating.

#12

Man, LEDs are really dim compared to HID lighting! My little 400W MH apparently puts out something like 36,000 lumens (with a fresh bulb). Of course, that’s a very broad spectrum, I imagine if we only counted the PAR bands it would be considerably lower, and all that excess light is waste. Makes me really curious how they actually stack up to each other in real life growing situations. Looking forward to seeing how these turn out!

#13

Oh, I’m sure you are correct. I remember the days when they spec’d car amps as peak to peak. 400w was really 20w.

#14

@rice81 @Lampy @Opcode @SRichardson3D @Afloyd

I hear that this weekend the committees are meeting with the board for expenditure requests. Is the Bio committee area interested in purchasing or requesting a spectrophotometer for grow lights (mainly for evaluating LEDs)?

#15

I wouldn’t think the wavelength would be that critical.

#16

I have found similar thoughts as @rice81’s earlier comment. Having analyzer equipment would allow us to better design the lighting.

That does,of course, assume y’all are interested in using lights I want to build :wink: If we are able to purchase the LEDs at the significantly less costs, validate the good ones, send back the no-so-good ones, then there is a potential for very excellent, customized lighting for the area.

#17

Wouldn’t a spectrophotometer just show a red blue and green spike for LED lighting and no colors in between? Those values would only be interesting in relative magnitude to each other, which should be available from the datasheets for LEDs used.

#18

LED spectral width (point at which lower power values are less than 50% of peak value) is typically 50 to more than 100 nm. That is, you won’t really see a ‘line’ rather a distribution centered on the stated LED frequency. (in contrast to a laser or laser diode that has a spectral width of maybe less than one nm).

Thus, when you mix a few LEDs of close frequency, you need a spectrophotometer to really see the net spectral power. You could sort of calculate the spectum, knowing all the LEDs and the powers and the angles, etc, but with the uncertainty of all the variables and whether you are really getting what you think you are getting, the photometer gives you the flux at the level of the exposure area.

I think it is probably wishful thinking to actually get such an instrument. They are really expensive and the money would probably be better spent on other things. Unless we found one on ebay.

But if we got serious and developed a product and wanted to make a bunch, it would be cool to have.

#19

I started googling it, and you’re right. It’s not nearly as spiky as I figured it’d be.

#20

Gentlemen,

I have recently used a spectrophotometer in my Biology lab and it is sourced from an educational products company which offers it at a relatively attractive price point and will give you a graphical wavelength distribution.

http://www.vernier.com/products/sensors/spectrometers/visible-range/svis-pl/

It runs about $470 and for measuring light emissions spectra, you also need an optical fiber adapter for $70. (+ tax and shipping)

The slight catch is that you also need a LabQuest interface and the LoggerPro software, (another $320 - $575 depending on the interface) but I already own both as I have been using Vernier sensor array products for over 12 years.

Let me know if you are interested.

380 nm–950 nm Wavelength reporting interval ~1 nm Optical resolution* 4.0 nm (at 656 nm) 25 nm (at 486 nm)

Wavelength accuracy† ± 3.0 nm (at 650 nm) ± 7.0 nm (at 450 nm)
Photometric accuracy‡ ± 13.0%
Typical scan time ~ 2 s Operating temperature 15–35ºC

BTW, they do have an even more accurate model that is specifically designed for light emissions, but it runs about $230 extra.

-JAG