Re: [CAD_CAM_EDM_DRO] Encoder/Servo question

>-----Original Message-----
>From: Bill Vance [mailto:ccq@...]
>Sent: Friday, July 16, 2004 4:35 AM
>To: 2cc@...
>Subject: [CAD_CAM_EDM_DRO] Encoder/Servo question
>
>Howdy;
>
>I've noted various references to 1000 and 2000 line encoders, but are there
>any servos that will actually turn one or two thousandths of a turn per
>step/line?
>
>
>

Absolutely! The closed-loop system DEMANDS that the encoder be turned
that small fraaction of a thousandth of a turn. I'm using 1000 line
encoders
on my system, which give 4000 quadrature counts per revolution. I can
generally hold the encoder within +/- 2 counts of the null point. So, the
ENCODER is being held within a very tight bound of about 1/1000 of
a turn. But, that has no bearing on table position! It is up to the
leadscrew
to force the table to be within the equivalent position, and that may be
beyond the capabilities of the leadscrew, it's bearings, and the nut.
I see about .0015" backlash in my 3 axes, which is 30 times the encoder
resolution. This is NOT a failure of the servo loop to maintain encoder
position, but a failure of the leadscrew to constrain the table to an
equivalent precision. Much of it is due to flex in the mechanical
mounting, compression of the leadscrew itself (it is .895" dia),
backlash in the ballnuts and rocking of the table. If the ways are
perfectly flat, this rocking won't happen, but my 1938 Bridgeport
is quite worn. The table and saddle are all worn in odd convex and
concave shapes.

Jon