The X and Y axis of a retrofitted machine will not home consistently between reboots. If it varies it is exactly 0.100" more, exactly one encoder revolution. We purchased this machine not running so I don't know if it ever homed properly. It is an old Monarch VMC75 that has a 1:2 gear train driving BEI EX116 1000 line quadrature encoders and a tachometer nearby for the analogue amps. We have rotated the encoder in 0.005" axis movement increments, clockwise and counter clockwise, trying to find the best home switching position relative to the marker pulse without improvement. It doesn't seem to make a difference if the home switch closed 0.020" or 0.080" before the marker pulse; sometimes it still goes around again another 0.100" to the next marker pulse. We have changed the homing limit switch that slows the axis for marker pulse reading with no improvement. We have reduced all the axis speeds via parameters that influence the homing routine with no improvement in repeatability. In the static position all voltages to the limit switches and encoders are correct. We were thinking of putting an oscilloscope on the encoder to check for a low voltage marker pulse or noise. Any comments or suggestions as to what else we should be doing or looking for?

The Yasnac J50M manual mentions something about the "Pitch error compensation and the leadscrew error compensation functions" and involves parameter #6006D1. Have you read about it?

Thanks for the tip on the 60006D1. It was 1 so we tried a 0 with no improvement. One of the lingering questions is why do we have no problem with the Z axis?

The encoders were at fault. Even though we picked up a marker pulse with the scope it was not long enough for the control to receive it as a definitive marker. The control could be a little fussy. Any how the inexpensive option was to change the encoder as we were not 100% sure if the problem was the control or the encoder.

Good to hear you worked it out, thanks for posting your solution.