Hi,
did you ever get this sorted?
I'm interested in how you got on.
My guess is that as the home switch activates Mach backs off until the switch returns to in-active, from your description you have index homing in action, the servo
drives until the index signal is active and sets the machine coordinate to 1) zero, if no home offset is active or 2) home offset for that axis when active.
If for whatever reason the position accumulator in your servo drive was at 34567 say at the instant that the machine coordinate was reset. If however Mach sets
the machine coordinate to your home offset, say 20mm, with steps per of 500 say, then the position accumulator would suddenly be presented with a position
change of 20 x 500 =1000, ie the accumulator would jump to 34567 + 1000 =35567. That would constitute an instaneous jump of 1000 counts. If your following
error is set less than that, almost certain, then it would attempt to accelerate to the new position but fault 'following error' almost immediately.
What intrigues me is that my understanding of home offset is that it sets the machine coordinates to the required offset but does not issue a command to go to
that location. Therefore I would not expect the position accumulator to experience the suuden jump in position. Whats going on?
One experiment I would like to suggest is to temporarily mount one home switch somewhere in the middle of travel of one axis. Disable the index homing for that axis
and try it again. If my understanding is correct index homing is enacted by the motion controller not Mach. I do not use index homing nor even home offsets but such
experimenting I have done with home offsets suggests that Mach is performing as I expected. This experiment would allow you to rule the motion controller
in or out.
I think also that if the Machine Diagnostics page were open when you homed that axis you might see the machine coordinate entry change which may in turn give you some
insight as to whats going on.
Craig