I doubt its moving that far I think it may be out but he obviously has not RTFM and in other threads has claimed similar which is where we got to auto tuning as simpler than trying to get information required to do calculations for him. Over several different threads its obvious he doesn't quite understand how Mach works when told its definitely tuned right I tried to point out margin of error that would be required so used 2540% to emphasize enormity rather than 25.4x to attain the results claimed.
I would imagine set z Axis zero to top of table rather than top of material as his sole piece of G-code never goes negative or below 1.5mm, just difficult to get information to help would guess all being done from his phone with the minimal responses.