Ian its noticeable when sending down a tool ie threading to friction fit tolerance, not altering anything it will either take a spring cut or need up to 0.08mm feed in to take a slither off. Next time send down loads come off.
Not a problem if just doing a thread, idea above is friction fit thread tolerance ie nil, by time its anodised its sloppy and not the sort of thing want on 0.75mm pitch holding 300bar back.
Its a physical movement of the Y axis magnetic 5 micron scale. Some years back did a test of cheap vernier type, C type Sino glass and the magnetic all in situ with remarkable results which lead to errors. Worst was using the supplied bracket with the vernier type think was up to 0.3mm out by time it started to move.
I can only treat the magnetic as some where near to within 0.15mm on diameter. If take off 0.075 it may well be still oversize +0.06 to well under by 0.1+ well out of synch to the stated measuring accuracy.
Glass Sino are ok to 0.05 repeatability about 3 times better.