Individually addressable RGB LEDs like Neopixels, WS2812s, and WS2811s are the defacto standard for making blinkey glowey projects. To build a very bright display, you need a lot of them, relegating very bright RGB displays to those of us who can afford the hardware and figure out how to drive that many LEDs. For his Hackaday Prize entry, [AJ Reynolds] is cranking these tiny RGB LEDs up a notch by building an individually addressable 10 Watt RGB floodlight.
Instead of building an RGB LED floodlight from scratch, [AJ] is leveraging the most mediocre of what China has to offer. He found 10 Watt RGBs for a dollar a piece and a few floodlight cases that cost about $5 a piece. By dispensing with the white LED in the floodlight case and replacing it with a 10 Watt RGB LED and some custom circuitry, [AJ] can build a powerful RGB floodlight with a BOM cost of under $15.
While there are big RGB floodlights out there, controlling them either means a custom proprietary protocol or messing around with DMX. A floodlight that speaks the same language as a WS2811 leverages an enormous amount of work from the world of Arduino and a lot of projects from around the Internet, making this a great entry for really bright blinkies and an excellent entry for The Hackaday Prize.
“…controlling them … means … messing around with DMX” – I don’t understand what is so hard about this. DMX is just 250 Kbps UART with an extended break and a RS485 line transceiver. Extremely Arduino friendly. Way more so than the WS281x protocol. Of course commercial DMX lights are way more $$, but that wasn’t the stated objection.
Maybe he’s complaining about the comparatively expensive XLR connectors for “real” DMX?
I just used RJ45 and cat5, which maybe makes me a bad person. At least, an incompatible person.
Actually if I recall correctly, DMX over cat5 is supported in the official standard, and is actually decently common
I was gonna say that. What’s wrong with “messing with DMX”? It’s too standard? It might be too easy to reuse? Too well documented? Too easy to integrate with other systems? Too easy to pull off?
WS2811 isn’t exactly ideal… Proprietary protocol which you end up having to bit-bang usually. Oh well.
Either ways, it’s typically the thermal management that’s the most interesting part in projects like these. If you don’t cool them well enough they die a quick death. Providing power and controlling them is the easy part.
Um…the LEDs are $13 each. How is the BOM “under $15” when the LED alone costs nearly that?
Because the LEDs are actually $1.30 each.
I agree it’s a very nice build but seen before except for the great PCB. But like Alan said, I don’t think the WS281x protocol is a plus. It’s a sensitive protocol and because every LED has a case I don’t think they are spaced withing 5cm of each other. DMX is for more robust and libraries are available as well. And if it’s a fixed light (because I don’t spot a connector on the project) you can still just use CAT5 cable and not use XLR although the cheap 3-pin version is also almost the facto standard…)
But in the build log he also talks about wireless (nRF24 with Komby Protocol). But most of those casings are aluminum. Reception inside isn’t easy….
I run tiny custom ws2811 boards longer than 5cm no problem. Including like this exact project funnily. Just have a resistor on each’s input signal.
Actually, the purported distance between WS2811 chips is up to 10m, with correct decoupling caps. IIRC it’s in the datasheet.
The WS2811 protocol allows the Floodlight to be incorporated into an existing RGB display. Since most RGB displays (Christmas, Halloween) are now moving to the WS2811 protocol in support of being able to control each individual light, it allows the floodlight to be placed within the 20ft end of any existing string. Additionally they can be daisy chained together which would appear as a nothing unique to existing controllers Controlling the LED is accomplished using one of the various software sequencers available such as Vixen3, XLights, etc..
..and of course you can run WS2811 protocol over RS422 physical layer if you need distance.
Could you tell me more about this? I have been in search of a way to run longer distances between WS2811 installations without using null pixels, and havent come up with anything simple/elegant. Thanks.
I remeber that floodlight, have a video about them:
https://www.youtube.com/watch?v=vbt2ojkXPuo (Ebay LED floodlight scam warning)
Or you could just take the bare WS2811 Chip and make it switch a set of FETs for powerleds instead…
Just realized that that’s exactly what he does…
Yep. Note that the ws2811 is open collector N so you need pull up resistors and P fets or inverters for more substantial N fets
Don’t need a pull-up resistor; just use the open-drain NFET output with a PNP to make a weird-shaped pseudo Sziklai pair.
The 2811 is open collector, you’d risk floating noise
Hello everyone
I made mine no longer SMD version. This is a test card before make the final board. Should be able to control 4 RGB LEDs (12 channels)
Picture : http://img15.hostingpics.net/pics/967243IMG2140.jpg
CLMNT
Individually addressable RGB LEDs like Neopixels, WS2812s, and WS2811s are the defacto standard for making blinkey glowey projects.