I like JohnM's suggestion. A piezo crystal mounted at every station
near the end of a rod could provide just the right amount, about a
tenth of a degree, of *rotation* (perpendicular to way travel), to
break stiction once per travel clockwise and counterclockwise. Not
quite as perfect as a rotating rod, but sooo much simpler. I think to
move the rod axially would just introduce random errors into way
travel.

Whadda say, is this one patentable? You could sell premade shafting
with piezo ends, in modular lengths. You would lose the infinite
flexibility in design of having yard long shafting that can be cut to
any length, but gain in the fight for smoothness. A simple controller
could be patched to every rod. A controller with capacity for sixteen
rods would be cover pretty much any machine tool situation. Various
other drives are possible, such as an electromagnetic rod end that
merely needs to be hooked to 110 VAC.

--Doug