Today was a (hopefully last) day home with this cold. Since I'm supposed to have a prototype for the 2012 Duckon blinkie ready for the concom meeting and hopefully then the boards fabbed and a dozen copies ready to go for Capricon, I figured I'd better, you know, actually get started on it.
I had the design done, but only in my head.
I use most of the software; AVR C and EagleCAD, only once a year for this purpose, so I have to spend about 4 hours re-learning it all again every time.
Anyway, at the end of the day I'd accomplished everything I wanted to. I have a breadboarded running copy, with the control changes I wanted to make in place. I've done away with the power switch and instead use the "power down" mode of the AVR chip, which according to my DVM is drawing about 1 microamp, which would take 17 years to kill the battery, so that's OK.
This design needs to be thinner than last year's since I intend it to be a pendant, so the CR2032 cell was out. I am using three LR44 cells instead, and I had to design and print my own battery holders, I bashed that out this afternoon as well.
I got up to speed (enough) in EagleCAD to lay the schematic and board out. I intended to use the RepRap with a pen holder installed to draw ink onto a blank board but the HPGL to GCode software is causing me trouble. I'll work on that later.
I finally realized that it makes more sense to create the Charlieplex scan bits in the pattern design spreadsheet and store that rather than putting bits in the data section that represent the pattern and making the microcontroller do the Charlieplexing work. More complex spreadsheet formulas, but the firmware code goes from 80 lines to 2 lines and should draw less power.
Blinkie prototyping:

New battery holder (there's a slot on the other side to make it possible to get the cells out):
