Quick follow-up, the prototype system I have looks like this (the screen looks a lot better in real life):
The "real" system's electronics are almost complete:
A few more mods and that's all for the hardware (<$40 in parts).
Software is shaping up nicely. The core dealing with the rower pulses is one page. Most of the work goes into the screen driver and the whole font rendering and layout... I can make the code available in a few days, right now I'm reorganizing some stuff... The off-line processing allows me to chart stuff like:
Edit: I forgot to add that I'm somewhat tempted to add an S load cell into the pull rope to measure pulling force. At that point I'd forego tracking the water wheel revolutions and essentially write a physical simulation of a rowing shell based on force input, or more likely, just integrate the force to get power and track raw power. You could probably do the same, although you'd need two load cells. Overall it's more challenging 'cause beyond a proof of concept which could use wires to hook up the load cells one would need a little BLE transmitter in the handle / paddle stick. (The set-up I have already tracks heart rate via BLE, so the RX side is pretty much covered.) Dunno whether I'll go there 'cause the rower is not me and she's not that deep into tracking performance...
For direct msgs you can use my screen name here at voneicken.com