Dimlow: Thanks.
I do have a bit of Rigol equipment, mostly due to the bang-for-buck. It was really impossible to beat the feature/price of the DS4024 vs. a 4 channel Agilent 3000 series. The Agilent might be better, but I couldn't justify the price. Same thing for the power supplies. As for the 6.5 digit meters, I love the Agilent 34410A (on the left). When it came time for another meter, I go the Rigol DM3068 for its flexibility and additional features (also a lot cheaper).
I was using the 16F886, but it just really wasn't enough. So, I moved up to the 18FxxK50, but the C18 libraries are crap in v3.45, so I settled on the K22 series.
The K22 probably is more than needed for this, but it seemed to simplify placement a bit, due to dual SPI and dual USART. Also, the K22 works a lot better with the ICD3 debugger than any of the 16F chips did.
The last, and most complicated, part of this project will be figuring out a bootloader and the ability to upgrade the software using a USB stick.
Then it will be time to layout the board, with lots of lightning protection, fused MOVs monitored by opto-isolators, etc.
Edit: attached shot of the bench. Doesn't show everything... the pegboard panels on the ends of the bench slide out on rails, with pegboard and bins hanging on both sides.