Another open source pixel controller project

J

jstjohnz

Apprentice elf
Joined
Sep 19, 2010
Messages
97
Location
Indianapolis, IN
I have developed 3 boards that I will be releasing as open source.

#1 is a pixel driver, 1-4 universes of DMX512 in via 2 RJ-45 connectors, 16 pixel string outputs via 4-pin headers, same connector and pinout as MPH's boards. Supports 2801 and 6803 pixels. Can be a mix of 5V and 12V strings, a mix of 2801 and 6803 strings, various lengths, etc. >200hz refresh rate of the strings.

#2 is a converter to convert between E1.31 (DMX over ethernet) and standard DMX512. 4 Universes in and out. The current version is for E1.31 in and 4xDMX512 out, but the same board, with an IC swap and a couple of jumper changes, and some new software, can go the other way.

The original plan was to use boards 1 and 2 together to drive pixel strings from E1.31, but I was eventually able to get both functions on one board, so #3 is a pixel driver with E1.31 DMX over ethernet in, and 16 pixel string outputs.

All boards are 2.5"x3.8". I have working prototypes of all 3 boards. If there is interest I will post some pictures and additional specs.
 
Very very insteresting.
Some more details ......
I will have to manage around 2000 ws2801 pixels....
So all ideas are welcome ;)

Cheers.... Henri
 
I don't think I have a "concentrated" need for a board like #3,

but both #1 and #2 are of great interest.

I'll watch for pics!
 
Docx file and pdf file attached with some more detailed specifications.
 

Attachments

  • Pixel Driver Specs.doc
    14.6 KB · Views: 55
  • Pixel Driver Specs.pdf
    18.9 KB · Views: 73
I'm testing with 8 pixel strings now. Light Show Pro appears to work, at least via the controller test mode, still learning how to sequence with LSP. Hope to have a demo video soon.
 
riri7707 said:
Very very insteresting.
Some more details ......
I will have to manage around 2000 ws2801 pixels....
So all ideas are welcome ;)

Cheers.... Henri
Click to expand...

That's pretty much what I'm shooting for this year, so I'll let you know how it goes.
 
Very overdue for an update, sorry about that, family and all sometimes gets in the way of the fun stuff......

I did do many small tweaks to the PC board layout for the E1.31 pixel controller. The 3 new boards are in, 1 is built up and tested and working fine, and I hope to have a total of 5 boards assembled and tested by this weekend. These will be the boards for my own display this year.

The changes were pretty minor. The voltage regulator now lies flat on the board instead of being mounted upright. Looks better and the PC board copper helps sink heat.

I eliminated the (expensive) connectors for the ethernet module and have gone to soldering it directly to the PCB. This required relocating a handful of resistors that were previously mounted under the ethernet module.

Slightly smaller terminal blocks for pixel power in, but still rated at 24 amps. Two sets for 48 amps total which is about right for 16 strings of 50 pixels fully on.

If you are using 5V pixel power (typical for pixel strings), you can now power the board logic from that just by installing a dip jumper. If you are only using 12V pixel power (typical for pixel strips) you need an external 5-7VDC power source for the controller.

Board size is still 2.5" x 3.8". Not much new as far as software work other than adding several built-in test patterns (an idea from MPH's board), and several built-in configuration sets that can be loaded via the on-board programming pushbutton.

I've seen some discussions re using point-to-point or multicast E1.31. My board will only do multicast at this time. To me it's a much more hassle-free plug and play setup compared to point to point. The only thing you have to program is the universe offset (via a pushbutton on the board), which basically tells the board which group of 4 universes to look for. No networking setup is required. Neither the sender nor the receiver needs to know the IP address of the other end.

We are getting to the time where I need to start focusing on my own display setup. I am hoping to go virtually 100% RGB pixels this year. I should have enough pixels on hand by next week to put together a full 1200-pixel megatree controlled by 2 of my E1.31 pixel controllers. THAT should be a good torture test of the boards, the power supply (5V @ 70 amps) and light show pro! I am **REALLY** looking forward to being able to replace my previous setup of 4 16-channel LOR 120-volt controllers, 120 strings of 100 incandescent minis (24 each of 5 colors), and 2 20A circuits to power it, with 24 strings of fifty RGB pixels and a controller/power supply combo that will measure about 5x5x10 inches.

-jim-
 
Great stuff Jim.

Have you done any practical work with LSP to see how it behaves with RGB channel counts of that magnitude?

If not, might be worth satisfying yourself that the software can support the hardware at scale before you get much further! My sense is that LSP has come a long way (and as a customer I really try to support the development) ... but still not far enough for this scale yet..... There's so many 'little things' that add up to making it hard to use in this context...

I know it 'works' now. But its still a pretty unpredictable proposition for what I want it to do.

I'm hoping to get back to using LSP ... next year.

Meanwhile... back to Vixen for me.. (And I'm running < 1000 pixels)

Tim
 
TimW said:
.. but still not far enough for this scale yet..... There's so many 'little things' that add up to making it hard to use in this context...

I know it 'works' now. But its still a pretty unpredictable proposition for what I want it to do.
Click to expand...

Tim - is there something specific we as LSP users should be aware of when working with, in my case, 4k channels?
 
Sorry David, wasn't trying to start an LSP war... just sounding a note of caution to Jim from my limited pixel experience so far.

I know LSP basically works with the technology. Its really a question of whether it works for you the way you need it to a month from Christmas 2010..... particularly if sequencing is starting now!

I have been playing with a grid of 320 pixels and an ECG this year. I really liked the idea of transitions, matrix animations all the good stuff that it promises.

But First we had a few issues with ECG to get over.... took a while to fix that...

Then I had a few issues with matrix animation.... That got fixed as well....

Then I started looking at sequencing. I am sequencing in a production environment... I like to look at the actual lights. Every time I want to run the sequence I'm working on LSP wants to reoptimise... that takes time with transitions.... I went out and bought a high end new PC.... it still takes lots of time....

sometimes it outputs the first time you select output... sometimes it doesnt... (just keep clicking or maintenance rebuilding the sequence?)

If you try to output at section of a sequence it might work... or it might not... I just can't predict it.

(I can't figure out how to loop a section of output reliably... but it all works fine in the visualiser...)

The Channel tester is flaky. Sometimes it works, sometimes the low channels don't turn on...

When I select intervals in the sequencer the software often picks up the previous interval (sometimes... not always)

I read you can't trust the visualizer for accurate interval setting ???!!!

I really really wanted to use LSP this year. I like the product and the support has been good. But at some point I have to admit defeat until next year. I've reached that point , with regret.

But specifics? YMMV! Looking forward to seeing others' success and trying to learn from it...

What I have learned so far is that you don't always get to just plug it in and it works . But that's DIY for you :)

Tim
 
I have become so focused on the hardware that I haven't really spent the time with LSP that I need to, and I know there are a lot of reports of issues with LSP. I have been playing with the Madrix demo lately, that program is really something. But major $$$$, and as far as I can tell it will only program one light matrix, not multiple display units.

So, yes I am well aware that things probably won't go smoothly with LSP, and I will most likely end up with fewer songs in the show this year. That is if there is a show, I just got in 800 pixels from Ray Wu and he sent me 12 foot strings instead of the 16 footers that I had gotten before.... It never ends.
 
You must log in or register to reply here.
Back
Top