To do list:
Tune the debounce. It's almost perfect, but I believe I detected a couple of rough spots so I will work on that. It could be related to the encoder I'm prototyping with though.
Add the x10 multiplier logic. The encoder I'm using in the prototype has 24 detents per revolution with four transitions between each detent. That's 96 transitions per revolution. It's not appropriate at all for this application, but was good enough to test with. The one on the BOM in an earlier post has no detents and 24 total transitions per revolution. It should work way better for accuracy, but the x10 feature will be necessary to move quickly through the input values if you want to get from one value to a another distant one.
Add a "loop-around" for the channel select. What I mean by that is if I'm rolling down, when I get to channel 1, the next transition down will get me to 512. I actually have an "all" selection just above 512, so we can dim all 512 channels together. The "loop-around" feature would give the channel selection range of 1...512.ALL.1...512.ALL.1...512.ALL. and would continue to loop around as long as you keep turning the encoder in the same direction. Looping will work in either direction.
Look at driving a transmitter component as an option. I'm sure the Lynx transmitter will work fine, but it would be nice to have wireless inherent in the same device. Maybe having dual output or a selection switch to choose. What do you think?
ICSP port.
PCB design. We've got a couple of people on this already. I think Mike is already pretty close to done with the current prototype design.
Once we get all this stuff done, we'll have a pretty handy little device. Given that the ST485BN works, I put the cost for non-bulk pricing at roughly $23 to build this prototype (without PCB and case). Some of that cost is for the power plugs and power port to supply 6v to the the Lynx transmitter. Eliminating them would lower the cost more yet.
Dave