gearmotors,encoders

Are gearmotors more or less verboten as axes motors in CNC
applications? Seems like the answer has gotta be "yes" for the same
backlash reasons that have landed acme screws and nuts in scrap bins.

But a 90 deg gearmotor sure would make a neat X-axis installation on
a Bport-class machine. It'd be fine on Y too, and no messing with
wheels and belts.

I've never heard of a no backlash gearset, except when someone messes
up the shim pack rebuilding their diff ...

---------------------------------------------------------------------

Apologies in advance for burning bandwidth; I've asked about this
before--but I still don't get it. Why not put rotary encoder(s) on
the non-motor end of the screw rather than on the motor shaft? It
seems like one wants to know the position, direction and speed of the
screw rather than of the motor per se. So if it's no worse, possibly
even better, to have the encoder on the far end--then what would be
wrong with a slightly flexible coupling (e.g. Lovejoy) between the
motor and the screw? I can see that during accel and decel the motor
and screw could get out of phase, and maybe in some bitter state of
the world this could feed on itself, but couldn't this problem be
tweaked away with a little gain correction? I like slightly flexible
connections; they absorb vibration, correct for slight misalignment,
would put some allowance for a few thou of axial heat-growth of the
screw in the system, etc.

Thanks.

--Jack