Dave's Crazy-Ass 80-channel EL Sequencer Project Page

Photos

updated: 22 July 2008

OK, I finally have some photos and a video of the Rev B (80 channel) board in action. First, here's the board:

Eighty channels in a 4.5 x 4.5 inch footprint, not too bad. This board can do 80 channels of MOSFETs (to control LEDs), or 80 channels of triacs to control EL wire, and in theory it could have a mix of the two but I haven't tested that yet. Before we get to the video, let me provide some context. The purpose of this lighting conroller was to control 80 channels of LED rope light for the flying saucer mutant vehicle - with 80 completely independent channels under software control, I'd be able to create some interesting chase sequences to create the illusion of spin and so on. Once the 80 channel PCBs were built, I wanted to have some sort of desktop representation of the flying saucer so I could write animation sequences in my office, instead of having to write them down in the garage, or out in the alley, or wherever it is that the flying saucer is parked. So I ordered up a bunch of LEDs, and wired 72 of them up in a circle on a piece of sheet metal. Each LED is supposed to represent a piece of LED rope light. Here are some photos of this partially-assembled test fixture:

I built the test fixture before I built the saucer itself, and at that time I thought the saucer was gonna have 72 "spokes" radiating out from the center. That number looks like it's going to change, but whatever. Anyway, the first part of this video shows a couple simple sequences on this 72-LED test fixture. At the time I shot this I had a few misbehaving LEDs but hopefully you get the general idea. Once it seemed clear the board was working, I ordered up crapload of LED rope light (seriously, 300+ feet of the stuff) and when that I arrived, I chopped up a dozen or so segments and wired those up. That's what's in the second part of the video.

Sorry about the crappy video quality, but with that caveat, here is the first video of the 80 channel sequencer in action.

The prototype

Here's the two of the photo boards, one raw and one populated with parts. This is the top view.

And here's the bottom. One of the limitations of Pad2Pad is that you can't specify different trace widths. I would have preferred to have wider traces for the high voltage AC signals.

This is the Jackrabbit BL1810 single board computer. This version of the SBC (a $69 value!) is included with the $139 Jackrabbit development kit. I'll probably get a few more Jackrabbits to have as backup units, but I'll go with the $49 BL1820, which doesn't have the lithium battery (which is the big honking red-and-white tumor on top.) You can see the differences between the various Jackrabbit flavors here.)

The development kit also comes with this prototyping board. It's just a bunch of buttons and LEDs, and there's a large space to do some breadboarding if you like. The Jackrabbit mates to this board with two 40-pin 2mm connectors. The proto board will not be part of the finished design, but I found it quite useful to have for prototyping with the Jackrabbit.

The same signals on the Jackrabbit's 2mm connectors are presented again on the prototyping board using slightly larger .100" 40 pin connectors (it's the exact same connector used by IDE hard drives). I picked signals that are all accessible from one of these connectors, so a single IDE cable connects the Jackrabbit prototyping board to my board, including power and ground. This is everything all hooked up. Pretty clean, huh?

Well, this is the least pretty part of the prototype. Since I'm building an EL-wire sequencer, I need EL wire somewhere, right? I had a big long piece of EL wire on my bike, and at some point it got pinched and stopped working. That happens with EL wire. It's flexible, but only to a point. If you kink it too hard, you'll break one of the internal wires, and then you're screwed. So when the wire died, I just threw it into a box figuring I'd find a use for it. Well, this was the use - I chopped it up into 18" segments and used gaffer's tape to attach them to two pieces of cardboard. Pretty low tech, but it does the job. And I never did figure out where the kink in the EL wire was. Once I chopped it up it, all the segments worked. If one of these segments stops working, I'll have to remember not to panic. "Oh yeah, one of those segments should have a kink in it somewhere...."

Video

A short video of the sequencer prototype in action can be found here. It actually looks way better in person than it does on video. The high speed stuff looks particularly cool, but the EL changes state at a much faster rate than the video frame rate, so it doesn't look like much on the screen. Trust me, it's cool. You'll just have to see it for yourself on the playa. :)

I was worried that the serial interface required by the 4094 would slow things down too much, but now that I've seen it in action, I'm not worried. I think the Jackrabbit is fast enough to handle hundreds of channels without any problem.