The point source pixel design I came up with last fall has a lot of advantages: Small size, low power & good color mixing in a 10mm package. The hardest part was finding a way to chain multiple boards together. I'd done lots of tests by hand and found that building the wiring harness often took much, much longer than the actual board assembly.
http://www.response-box.com/rgblights/2007point.shtml
Life's too short to spend days and days building a wiring harness.
So I redid the design... It ended up costing a few pennies more in components, but the time saved making wiring harnesses more than compensates.
Since each pixel has an on-board voltage regulator - and since the LED is driven from the processor directly - voltage drop within the cable is less of an issue than otherwise expected.
It's based on a 'backplane' of 10-conductor ribbon cable. 4 conductors each for power and ground, plus a pair for DMX data.
It's easy to lay out the cable, mark it every 6" or 12" and install IDC female headers as needed.
Since I was concerned that the high-speed data would be corrupted over long lengths of flat cable, I also designed a tiny DMX repeater / splitter board. It has connections for DMX in & through, plus an RS-485 receiver. Then, two RS-485 transmitters feed data to two separate 10-pin headers.
Each header connects to a female ribbon cable jack. So in essence, a string of lights can be driven from the center.
I expect to drive 16-18 pixels on each side of the 'T,' with each arm being 8-10 feet long.
Each splitter board will be connected to the control equipment two cables: a shielded, twisted pair for data and a heavy gauge pair for power.
Pictures attached. One is the splitter / driver board, One is an assembled pixel and one is the same pixel covered in heatshrink tubing.
~600 of these are being assembled by a shop here in town... Will post some video clips when it's all installed.
EDIT: After taking the picture of the board + heat shrink tubing, I used a razor blade to trim the tubing back to the junction between the LED's wider 'neck' and main bulb. This way most of the 10mm bulb is visible but the water resistant seal remains reasonably sound.
These are being mounted in the eaves, under the raingutter and near the soffit. It's not designed to be fully submersible, just resistant to the occasional sideways-blowing blizzard.
JEC wrote:The point source pixel design I came up with last fall has a lot of advantages: Small size, low power & good color mixing in a 10mm package. The hardest part was finding a way to chain multiple boards together. I'd done lots of tests by hand and found that building the wiring harness often took much, much longer than the actual board assembly.http://www.response-box.com/rgblights/2007point.shtml
Life's too short to spend days and days building a wiring harness.
So I redid the design... It ended up costing a few pennies more in components, but the time saved making wiring harnesses more than compensates.
Since each pixel has an on-board voltage regulator - and since the LED is driven from the processor directly - voltage drop within the cable is less of an issue than otherwise expected.
It's based on a 'backplane' of 10-conductor ribbon cable. 4 conductors each for power and ground, plus a pair for DMX data.
It's easy to lay out the cable, mark it every 6" or 12" and install IDC female headers as needed.
Since I was concerned that the high-speed data would be corrupted over long lengths of flat cable, I also designed a tiny DMX repeater / splitter board. It has connections for DMX in & through, plus an RS-485 receiver. Then, two RS-485 transmitters feed data to two separate 10-pin headers.
Each header connects to a female ribbon cable jack. So in essence, a string of lights can be driven from the center.
I expect to drive 16-18 pixels on each side of the 'T,' with each arm being 8-10 feet long.
Each splitter board will be connected to the control equipment two cables: a shielded, twisted pair for data and a heavy gauge pair for power.
Pictures attached. One is the splitter / driver board, One is an assembled pixel and one is the same pixel covered in heatshrink tubing.
~600 of these are being assembled by a shop here in town... Will post some video clips when it's all installed.
Yup, ribbon cable is a good idea. I worked for a company that had a high voltage 8 bit bus for individual access of locks. We strung a lot of ribbon cable for that.
One question, how are you handling one node to the next??? do you cut a few wires on the ribbon at each node???
Nice use of shrink tubing too , by the way.
Tony M.
As usual John, a clever design..
Me thinks i'll run up several hundred next year.
A Marchini wrote:Yup, ribbon cable is a good idea. I worked for a company that had a high voltage 8 bit bus for individual access of locks. We strung a lot of ribbon cable for that.One question, how are you handling one node to the next??? do you cut a few wires on the ribbon at each node???
Nice use of shrink tubing too , by the way.
Tony M.
Thanks.
I'm not quit sure what you mean by cutting wires. All ten wires are common to all 16-18 boards on the bus.
Since it's RS-485, the individual boards just tap the data signal as needed. Each pixel has a discrete DMX address, so it only responds to a particular section of the datastream.
EDIT: However, each 'arm' of the T uses a separate ribbon cable, which is driven by its own RS485 transmitter. So yes, there is a complete break in the cable, right at the driver junction. Power and ground are common to both sides, of course.
An interesting side effect is that the main DMX controller / Vixen Computer / Lighting Console can now feed a daisy chain of 32 point source drivers. And each point source driver can in theory handle 64 (32 * 2) point source pixels.
JEC wrote:Thanks.I'm not quit sure what you mean by cutting wires. All ten wires are common to all 16-18 boards on the bus.
Sorry, thinking loop in and out daisy chains. Forgot each device grabs its own part of the frame (or 3 parts I suppose in this case).
Tony M.



