Arduino based LED lighting?

Did you see a new gadget or toy and start brimming with ideas? See something that has tons of potential? Discuss these thoughts here.

Arduino based LED lighting?

Postby neltnerb » June 23rd, 2012, 10:24 pm

I am building a new arduino system, and am trying to decide what features to include.

The core is an Arduino Leonardo for sure, since I want the free UART for other communication. It has attached 15W of LED light as RGB+white. This is pretty darn bright, and I'm pretty excited about it. I found that LEDEngin has a fairly affordable four color LED and so with any luck I can build like 100 of these lights and make some crazy arrays. The question is how much or how little to build in once I've gone that far down the dark road of a major art project.

Basically, I am considering adding on DMX control, a Xbee/RN-XV pinout for wireless communications and wireless programming, and a microphone with a frequency analyzer (MSGEQ7) so that people can do cool standalone interactive art. Each little part in itself isn't all that expensive, but they all add up. I would love to try to include them all, or at least include pads for them all so that if I change my mind I can add on the other features, but as far as I know I'm the only one interested in lights that automatically respond to sounds or which automatically create patterns over ethernet cable based DMX controlled over wifi.

Do you think this is worth trying to make? It obviously costs a lot to make prototypes, and I am trying to make this so that I can sell them to others for under $75 each... that might be impossible though. I'm happy with budgeting a few $k towards putting this together if people agree that it's not a waste of time.

Which features are most important to you all in a LED light based around the arduino? Cost? Lumen output? I am thinking currently to just use a triple pole dual throw switch so that you can switch between wifi, DMX, and user-added UART, along with connecting the pins to GPIOs so that you can do things like send data to one light over wifi and have it act as a DMX master to control others. Or just the ability to use the single light as a USB -> DMX controller (which seem to be well over $100 by themselves for some reason).

Any encouragement or advice would be very welcome =) Right now I'm excited because it seems that I can do 16-bit PWM on each of the RGB channels and 10-bit on the white, which should let me make extremely, extremely nice colors. I also want to control it all over OSC like with the Saiko5 fixture, but for a lot lower cost. If I can manage to make these so that I can spend under $75 per system I'd love to be able to provide them to other hackers too and get the prices even lower. Manufacturing seems to be quite pricey at these kinds of small quantities =(
neltnerb
 
Posts: 3
Joined: June 23rd, 2012, 10:18 pm

Re: Arduino based LED lighting?

Postby Quin » June 24th, 2012, 8:41 pm

If I were looking for an array of LEDs for some display, paying $75 for 15W seems like massive overkill. I've done interactive art projects before, and that much per LED controller would be a no-go from the start.

You really want to create a lamp that you can sell, only have the XBee as an option; keep something like IR control for single bulb projects. Might be cheaper to look for a dedicated PWM chip controlled by 1 or 2 digital lines, instead of using a massive ATMEGA. And the microphone/FFT, while it sounds cool, is more parts and more cost to your unknown target audience. Have one bulb that has a major CPU and does that computation, and leave the other bulbs as slaves. Consider the cost it adds to each board, and whether each board even needs it if you are going to slave a bunch of them together into a large display array.

Now, for something like DMX, I have to assume you mean some form of studio lighting. Consider that a 12 fixture, 192 channel controller goes for just under $200 according to google. And you want your target buyers to spring for between 50 and 100 a bulb? Just to save on wiring the controls together, but with the risk that noise on the XBee channel might nullify their timed display?

Not trying to be a total downer, and if you are going to use these yourself, I see nothing wrong with the plan if you need all those features. But if you are going to sell it, really think about your target audience.
Quin
 
Posts: 106
Joined: February 24th, 2012, 9:17 am

Re: Arduino based LED lighting?

Postby neltnerb » June 24th, 2012, 9:05 pm

Cool Quin,

Thanks for the advice!

I've also been thinking about doing something closer to what you're describing, with a "mother board" that has all the various communication options, and "daughter boards" with just the LED and some minor power electronics. Definitely an option.

Would your opinion change if it went from 15W to 60W? That's pretty easy to accomplish, but adds about $35 to the manufacturing costs.

Just throwing some ideas around, I've talked to a lot of people about this and there seems to be a ton of varying opinions on what would be interesting to have.

The vision I have for this is less of a LED bulb and more of a fancy arduino with a lot of easy to add features. The incremental cost for each feature really isn't all that much. For instance, the DMX setup adds $2 of manufacturing cost. The mic and audio analysis another $3. The wireless header another $1 (I don't plan to include the module, that I agree seems like overkill). The big costs are the LED (~$9 for RGB+white), the power controllers to drive the four channels ($3), and the atmel ($4).

The wifi is nice because you can control it from far away without a pile of wires. I've found this to be extremely valuable for my own projects, and since you'd be able to program it over wifi you could do things like install it permanently in a ceiling and reprogram the arduino over wifi. That's pretty neat (to me).

The DMX is nice because you can connect it up with existing DMX systems, which is pretty neat (to me). It also lets you use just one master system control the rest of them with a single wifi module sending very high level effect commands, which is extremely neat (to me). That kind of thing could let me do really neat stuff like place them around the perimeter of a room with wires between them, and send commands from my phone to change how they are behaving.

Do you think it would make more sense to just produce a board that takes TTL level PWM signals as an input and has a single 15W four-channel LED? That's something I could probably produce for under $20 each, but of course it would need a substantial heat sink too if it's a small board in particular.

I plan to basically make these for my own use anyway since I want something autonomous for an art piece and the audio synchronizer chips I've used in the past don't offer enough configurability for me. I figure that if I want autonomous, reasonably bright lights with really nice colors (particularly at lower saturations where I can mix with white instead of using mixed RGB) with wifi and DMX for daisy chaining, other people might be interested to. So I suppose that I am the target market ;-) People who want high hackability in a fairly low cost light that they can use in a wide variety of applications.

With specific regard to the mother/daughter board topology, my thinking was more along the lines of a shared PWM line for each daughter board so that all the daughters are behaving identically. But of course I could also just daisy chain together some TLC5940s (I love that chip! I've been using it for almost six years now and it's fantastic -- just for any readers who might be interested in it.) But that chip doesn't sink enough current for the LED to reach full brightness. I suppose I could use two per daughter board, but then it's starting to get a little silly. In my vision, each daughter board would connect via a right angle header to the board before it, so you can just take one controller and connect up as many daughters as you like to reach the light intensity level you want in your fixture. I will probably also do that project, but I try to do only one thing at a time. Again, that's using me as the target market -- I've always wanted a light that throws out 1000W of LED light so that I can make clouds change colors... in the country at least.

Again, thanks for the feedback!
neltnerb
 
Posts: 3
Joined: June 23rd, 2012, 10:18 pm

Re: Arduino based LED lighting?

Postby Quin » June 25th, 2012, 7:36 pm

You will always be your own target market, that's what makes it dangerous. What works for you is what you want, exactly. What there is room in the market for, however, is always different. I was just commenting on your statement that you were prepared to drop a few grand into this and were wondering how you could sell it. And since you are still asking what prices seem reasonable, I'll continue that line.

Let's say you get a whole LED kit put together, DMX and mic/FFT for audio sync, and run the price under $100. For a concert hall to be interested, you need to sell them on why they should use yours instead of any other kit; or the harder sale of why they should upgrade to your kit. So if your LEDs burn out (it will happen eventually) do they need to buy a whole new microcontroller+sockets+LEDs (another $100 per fixture), or can they just use a drop in LED board? Is it an easy to replace board they can order or that you can supply in quantity, or is it something they can only get from a supplier that mostly sells to developers like you and not to the end user?

On the separate scale you mentioned, the walls of these lights presents the opposite problem. Why should I buy lights and controllers from you for $75 each when what I need is a wall of lights and just a few controllers to sync and drive them. Syncing a wall of separate controllers is a lot harder than syncing a few and letting them each control a few dozen LEDs or LED boards with PWM or any more complex data bus,

You might consider that what you have is two or three or more projects all wrapped up in one. A drop in, quality LED board would be very useful in lots of places; I'd even consider them in 12v for motorcycle lighting. Standardizing a socket could be the biggest problem of this step. Separate that from your controller logic, and make your controller a second project. In your wifi/DMX daisy chain example, why would each light need the full processing power of the controller? It isn't much cost, but it is cost that the end user has to foot. Think over whether it's needed. It may be in your case; and it may be that no one else needs exactly the same thing thereby limiting your sale audience.
Quin
 
Posts: 106
Joined: February 24th, 2012, 9:17 am

Re: Arduino based LED lighting?

Postby jimmy1974 » August 16th, 2014, 7:47 am

hi,friend I am a little late,It doesn't matter you are finish your idea.I want to say something for your good idea.
My views are as follows:

MCU: the arduino(module) does not facilitate production,You might consider using other mcu, such as Cortex-MX or Cortex-AX.
15W RGB LED:As for the costs of 15W LED You can refer to the(http://www.ledlightmake.com),Here are some informat of dmx512 and High power led modules costs.
jimmy1974
 
Posts: 2
Joined: August 16th, 2014, 1:43 am

Re: Arduino based LED lighting?

Postby neltnerb » August 16th, 2014, 1:29 pm

Oh, hi Jimmy!

I actually completed this project some time ago and it went great. Just never followed up here =)

http://blog.saikoled.com has about 100 posts about my adventures and the stuff I built in.

But I do completely agree with the comment about Atmel, I am making my next goal to build a light that is based on ARM. A friend actually ported my Arduino code so that it works also now on the ARM here:

https://mbed.org/users/frankvnk/code/HSI2RGBW_PWM/

which lets you use Hue-Saturation-Intensity color space to control a 40W or more RGBW LED to get perfect 48-bit colors plus an additional white channel! I hope you like it =)
neltnerb
 
Posts: 3
Joined: June 23rd, 2012, 10:18 pm


Return to The new gadget brainstorm pit

Who is online

Users browsing this forum: Bing [Bot] and 1 guest