Re: [CAD_CAM_EDM_DRO] Re: CNC Lathe question/ Linear encoders/backlash

vrsculptor@... wrote:

> --- In CAD_CAM_EDM_DRO@y..., Doug Fortune <pentam@h...> wrote:
> > vrsculptor@h... wrote:
> > >
> <snip>
> > There ya go, clear as mud.
> >
> > Doug Fortune
> > http://www.cncKITS.com
>
> Doug,
> Thanks a lot.... that actualy makes a lot of sense. I think a ball
> screw on the back of the lathe / servo / gecko would be the way to
> go.
>
> I've had a conceptual concern about long linear encoders, servos and
> backlash for a while. Servos while idling continually do a pretty
> quick +/- 1 count dance to keep track of where they are.

Not all servos do this. My velocity servos (with DC tach feedback) do
not hunt. There is some flicker, but it is very slow, and has to do
with mechanical spring and stick/slip friction of the system. My Z
axis does hunt at about 1 Hz because of the weight of the quill.
Now, this is over a distance of .00005" for X and Y, and .000025"
for Z.

> With a high
> count encoder mounted on a lead screw this dance doesn't amount to
> much. On my mill 1/4,000 of a revolution. If you used a linear
> encoder and there was any backlash at all the motor might be turning
> quite a few degrees to move the table an encoder count. This would be
> like a jack hammer on what ever is causing the backlash and I could
> imagine the whole setup exploding. Any thoughts on this?

Linear encoders with an Acme screw would be a horrible situation.
But, with ball screws properly set up, the hunting should be quite
small. A tach on the motor will greatly reduce hunting, as the servo
knows only the smallest correction is needed, so it will only allow the
slowest possible velocity for correction.

I will note that while the encoders detect the hunting on my system, you
can not feel any motion of the motors.

Jon