There are soooooooo many things wrong with that paper.... Whoever wrote it does not understand what he's working with. Perhaps most glaring:
"The low-cost machine has been set-up to use 1/4 stepping for a positional accuracy of 0.0000625 inches (about 1.6 microns). It should be noted that in all cases, the accuracy achieved is dependent on other considerations such as manufacture and assembly processes employed. In addition, spindle run-out is responsible for much of the inaccuracy. Typical spindles have inherent run-out in the region of several microns. However, the accuracies are improving. The low-cost spindle which is being used in this application, shown in figure 3 below, has a low run-out; approximately 0.00024 inches (about 6 microns)."
He confuses resolution with accuracy (two TOTALLY different things!), and makes no allowance for the inaccuracy of the leadscrews, and other purely mechanical factors, such as backlash (which is virtually NEVER zero).
Regards,
Ray L.