Andrew R Stewart 

icemilkcoffee- I agree about hand help calipers not being very repeatable in measured numbers.

When I started to figure out how I would do this I did consider using a dial on the cable pull measurement, the 2" dial was bought for this and I fitted a flat back to allow a closer positioning to the down tube, so the indicator point might hit the cable block tight to the cable. One issue with the cable block is that the cable will sag down just under the weight of the block. To add an off center pressure (the indicator point) would increase this and result in more deflection to the cable measurement. I had hoped that a longer cable block would reduce the sag (and as I write this I realize that rotating bike so the cable is vertical would eliminate this. Oh, how the creative process works...)

I mentioned technique. In the 5 tests I've done I saw some changes of values dependent on how I cycled through the shifting range. Quickly I decided to "standardize" my technique to have any real ability to compare or extrapolate from. The two initial test runs I discarded as being relevant to the system design and about learning how to work the set up. This will evolve. Andy