{I accidentally posted this in the General Mach discussion area. Sorry for the duplication. Is there a way to unpost something?}
I have found the following during testing:
The test involves running the router around the outer perimeter of my table (24 inch by 12 inch) and back to a dial indicator to check for lost steps. I start at 0,0 then go to 0.5,0 then 23.0,0 then 23.0,11.0 then 0.5,11.0 then 0.5, 0 and back to 0,0 and the dial indicator.
1) No steps are lost unless I pause the program by pressing Feed Hold followed by Cycle Start. If a Feed Hold is executed, it will frequently lose steps (anywhere from .005" to .075"), but only on the x-axis.
2) Normally, the G-code display highlights the block being executed, but after a Feed Hold, the highlight stops updating although the program does run to completion.
3) After a Feed Hold, a rhythmic pulsation appears during y-axis motion that did not exist prior to the Feed Hold. In spite of the pulsations, y-axis motion remains accurate with no apparent loss of steps.
4) If a I do a Feed Hold early in the first traverse of the x-axis and nowhere else, it often completes the circuit without loss of steps, although the afore-mentioned rhythmic pulsation during y-axis movement still appears.