For the lens distance it sounds as though you have some low magnification lenses with a different par-focal distance. A set of lenses with the same parfocal distance are designed to all focus in the same place relative to the shoulder on the RMS thread. This makes objective turrets work on normal microscopes. The Openflexure Microscope is designed for 45mm parfocal distance which is the most common now, but 35mm is also a standard and I have older lenses with all sorts of distances.
Your feedback on calibration is something that should be relatively straightforward to address in the software, but may not fit in the timeframe of the Openflexure Server v3 release. Server v3 is targeted at the microscope v7.0.0 release, which will not include swappable objectives at first. For a motorised swappable version you would need to save and retrieve the camera-stage mapping as well as the flat field camera correction. A motorised version could actually know automatically as soon as you move a little bit, by matching the image change seen with the available stage mapping files 
.
For a better experience on manual operation, I would point you to the modified body that is available, linked from the thread Minimal manual version of the Openflexure microscope. The manual body should interchange with the main body in any of the other versions of the microscope, not just the webcam version shown on that thread. That version does not create the STL for a manual separate z-axis for the upright, but it could easily do so. Edit: The merge request now does include a manual version of the separate z actuator for the Upright version.