Arduino-Based, High Powered LED Lighting Over Wi-Fi

Gibbs-Duhem writes "This awesome video was produced by some MIT engineers recently. They've started a fully open-source, open-hardware high power LED lighting project that they designed to be modular enough to control with the Arduino (or any other control system). Using their open-source firmware, you can set up the Arduino to connect to Wi-Fi and receive Open Sound Control packets. Then, they went further and released open-source software for PureData and Python to do music analysis and make the lights flash brilliantly in time with the music! A full Instructable was also posted in addition to the existing documentation for design and assembly on their website."

Okay

[Sylak]

EDMX/ETCNet is still going to remain the control standard for using Wifi to control lights, especially when DMX information is readily available online.

This method will probably only lend way to hobbiests who can't justify purchasing the equipment, when even DJs use DMX now. (Although Net2 touring nodes are relatively cheap now, and Net2 compatible light boards are popping up on eBay with the new generation of light boards replacing the old ones)

Re:Okay

[makkbe]

Art-Net (although completely worthless, just as ETCNet) is much more widespread than ETCNet looking outside the US, even over WiFi. But while you're building something like this, why not go for an open protocol that is really designed for controlling lights, such as BSR E1.31? Plain stupid in my eyes; I'll have to fix their firmware. On another note, using WiFi for lighting control is idiotic. WiFi is no way near stable enough to provide a decent transport for time critical protocols.

I've been working on something similar...

[allanw]

Shamelessly linking my own blog here:

http://electronoblog.allanw.org/2010/10/triple-high-brightness-led-driver-arduino-shield/ [allanw.org]

The board is in Arduino shield form factor and it can drive RGB 700mA LED's of any voltage up to 30V. It has an onboard micro that communicates with I2C that allows you to dim the channels without having to do the PWM on the host micro. This is optional, and there will be cheaper versions that just takes in PWM input signals directly.

It seems like all these high power LED stuff is way overpriced. I designed this after seeing Sparkfun selling a similar board for $50! And now this, which costs at least $290 for a bare board.

I haven't gotten around to finishing it yet but I intend to sell these boards for only $25 for the basic feature version.

Arduino? Yawn.

[Bassman59]

I'm so tired of non-engineers puffing up the lame Arduino platform. Why bother with Arduino when you can get a Silicon Labs 8051 board, with an excellent USB JTAG dongle, for a hundred bucks? You can't buy the debugger for the AVR for that.

Re:$800 ?!

[Gibbs-Duhem]

Dear Speare,

As the person who designed the device and the case, I have to question your ability to do a valid cost analysis, but if you'd be interested in helping out with bringing the cost down we'd love to have you. You can join our hardware mailing list at http://saikoled.com/mailinglists/ Wramsdel is completely correct, I have to pay the initial software developers who spent hundreds of hours developing the base system so that it was in a usable state, and I think it's not terribly fair to argue that the thousand plus hours I put into design, prototyping, assembly, and programming is not valuable. This is a one-man shop, and it's taken me three years to get from my original light fixture covered on slashdot in 2008 (http://led-artwork.com [led-artwork.com]) to this one, and I can assure you that the time investment is well worth it in quality. These things are good enough to compete with CK light fixtures (although I still need to finish my DMX -> OSC converter, I'm kind of totally out of money at the moment though).

Most of the cost issue is that in small quantities *any* machining is expensive. Just having the 8 holes drilled and tapped into the aluminum piping was $17, for instance. If I was casting 1000 cases, and hiring FoxConn employees to assemble them, and not paying the software or hardware developers anything for their time, sure maybe the price could get down to $300, but $150 is crazy. The raw parts for the wifi module are $60 alone, not including any case whatsoever, nor the included arduino, nor the power supply, and so on.

Also, Wramsdel is completely correct that I have to license patents from CK/Philips to sell these, and although the agreement is confidential, the amount is not trivial (and took lawyer time and my time to make happen). Additionally, I am trying to sell these to clubs as well so that I can offer cheaper versions to hobbyists (see below for how to get them cheaper). They are comparable to CK fixtures of similar price and brightness, are easier to install and use, and have free software to make them actually do things, as opposed to spending another $3k on control hardware and software.

In any case, my goal with this is to get the price down to the point where I would happily buy lots of them. Since I'm poor, the only way I can do that is to get orders for enough of them that I can drop the price. Suggesting, however, that the price should be the cost of the components alone is unreasonable for any sane business. I want to be able to do this for a living and as a community service, not as a community service alone.

In any case, since I want to get this off the ground so I can drop the price, I will point out a few things that you can do to drop the price.

1. Buy the kit and build it yourself. My time is valuable (believe it or not), and I happily designed the product offering to be such that you could buy *everything* that you need to build your own, minus a screwdriver, for $680. I have a (hard) day job, and I'll pay top dollar for your time in doing it yourself.
2. Use our hacker coupon code - hobb123. 5% discount. More than that I can't really support right now, since I need to get about $30k in order to build a full run of 100 lights.
3. If you're really strapped for cash, buy the bare boards, and use the bills of materials which I *made fully downloadable from the site*, upload those to digikey, and source your own parts. Use the instructions I wrote for how to lay down the solder paste, place the components, and trivially reflow solder them on a hot plate. You won't have a case, but it sounds like you can surely manufacture your own, buy a power supply, and buy an arduino or whatever control board you like for $50. Go for it. If you just want raw component costs, excluding arduino and power supply and case, you can probably do it for $200 ($33 each special) or maybe even under $150 if you build a lot of them. Even better, you can put on your own colors of LEDs! I used red, green, and royal blue... you cou

Re:Wake me up...

[Gibbs-Duhem]

As the designer, we tried to take people like you into account. If you're willing to solder your own boards and deal with heat sinking on your own, you can buy bare boards and use the Digikey BOM we made available on the documentation page to make your own for perhaps $100-$200 not including whatever you value your time at.

There's also a 5% discount using the coupon code hobb123.

Feel free to join the development lists if you want to get more ideas about how to do a less expensive system. I certainly want one too... but most of the cheap RGB fixtures out there are, in my opinion, unhackable garbage. RGB strips are cool though, if you just want mood lighting.

w/r/t the comment below about 800 lumens being as bright as a standard 60W light bulb, I should point out why, although true, that's not actually a sane comparison.

First, the lights here is colored to start with. Take a 60W incandescent and filter out everything but red, and you have a very dim light.

Second, it is 800 lumens because we used royal blue (extremely pretty color) instead of standard blue. The difference in perceptible brightness, in my opinion and the opinion of everyone who has seen a side by side comparison, is that there is no difference. However, the conversion from lumens to watts declares that it is 200 lumens dimmer. I think that there is a serious flaw in the way we calculate bright-adjusted conversion factors.

Third, the light is focused into a tight cone, meaning we lose almost no light due to light going "up" into the fixture. In a standard 60W, you're spending about half your light illuminating the ceiling. We don't do that, so the lux (lumens per square meter) is much higher. Try looking directly into one of these lights for more than a second, and you'll understand the difference.

There is a difference between 60W from a white incandescent and a 800 lumen LED fixture. Yes, technically the number of lumens thrown out are the same, and it would be dumb to use the LED light to produce white light, but for colored light it is at least an order of magnitude brighter.