Would be good to document how we got to the conclusion that the accuracy of the flight path calibration is limited to 2-5 cm because of moderator geometry.
Probably adding it to the notebook itself that does the calibration on codeshelf would be a good place?