jeff5may & all,
I went down to my local Geek hang out last night (DorkBotPDX.org) and chatted with a few folks. One of them, to whom I gave a URL of our discussion, has shown an interest in what we are doing, and is interested in offering PCB design support for our project. Of great interest to me is that he has already done PCBs for Arduino shields, which opens up some very interesting possibilities.
So we're not necessarily locked in to existing shields.
The up-side is that it expands our design possibilities, and the down-side is that it reduces the off-the-shelf appeal of our project.
* * *
Please understand, I don't want to seem negative here, but rather speak as a 'Devil's Advocate', and ask questions regarding possible 'project creep'...
In considering NiHaoMike's comment about including two independent RS485 ports... Do we actually require these on our most basic controller? Are two RS485 vital to functioning? How will they be used? Can someone explain to me why one RS-232 would not be sufficient?
How vital is the LCD display? Does it's utility truly justify the resources required to run it? I don't question that enough clock cycles will be available, but how about I/O resources and memory consumed? Could we not get by with a few indicator lights and maybe an alarm beeper? Can we not connect yo the USB to change specific vartiables?
By the way, I previously used a data-logger from Spark-Fun called Log-O-Matic, I don't know if it is still sold, but one feature about it that was very interesting is that on start-up, it would read an ASCII file that was available in memory, and that file had specific variable values that the logger would read at start-up. It made it very easy for a new user to tweak those values, values like time between data-read cycles, etc. BTW, if that data file was missing or corrupt, it would create the original default file.
If we have sufficient resources to accomplish our aims, an LCD display is just great. I'm just thinking about complexity here...
Best,
-AC