Lead screw twist... some musings

If the screw does twist from the torque required to drive the table,
and the encoder is on the motor end, the table will now be out of position
by the amount of the twist (times the pitch of the screw). Only when the
load is removed, and the windup is taken back out of the system, will the
table be in the position the controller thinks it is in.
If the encoder is mounted on the end of the shaft opposite the
motor, (the end that is not twisted), it will more accurately measure the
actual position of the table, as the lead screw will not be under load from
the ball nut to the encoder. I understand the potential for resonances, and
other nasties with this system, but doesn't it have the inherent advantage
of more accurately telling the control where the table is under a heavily
loaded cut?
Of course, a table mounted encoder eliminates this source of error
completely.

Brad Heuver