Re: someone told me... encoder resoultion or accuracy

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "Torsten" <torsten@g...> wrote:

> --- In CAD_CAM_EDM_DRO@yahoogroups.com, "Earl" <colin@v...> wrote:
> > I am looking at buying some linear motion units. The specs on them
> > are:
> >
> > 16" travel, 20mm pitch ballscrew, 200W Panasonic DC brush motor
> with
> > 500 C/T encoder.
> >
> > The seller claims positioning accuracy of 0.0005" with a Gecko
> drive.
> > I am confused by this. from what I understand is 20mm (per turn
> > ballscrew) divided into 500 positions for that turn is only 0.04mm,
> > or ~.00157" resolution.
> > I asked him and he replied with this:
> >
> > The 500c/t encorder. The Gecko will count 1 pulse x 4 = 2000 pulse
> > per turn 20 mm / 2000 counts = .01 mm ( .0003937" )
> >
> > something is not right. am I confused or is he.. or are we
> both!!? :)
> > Thanks for any insite yall. I have been running a stepper system
> and
> > am looming to jump into the servo world.
> >
> > Colin
>
>
> Colin, encoder resolution is usually spec'd as quadrant so a
> 500 quadrature count encoder will resolve 2000 positions per
> 360 degr. of rotation.
> 20mm = 0.787401575 inch per turn
> So the theoretical resolution would be
> 0.797401575/2000 = 0.000393701 inch per step or 0.01 mm.
> This is called the resolution of this setup.
> The accuracy will be the actual positioning precision after
> all electronic and mechanical deviations are counted in and
> are likely to be be much more then this.
> Good Luck

Resolution if the ability to see something in small details and
postioning accuracy is the ability to place something to some
postistion.

in addition, there is backlash aka hysteresis, the ability to put the
thing there when starting from either direction.

and there is repeatability. often generalized as lost steps, but also
could be from changes from thermal movement.

The screws will have errors over their lenght, so you may find that one
screw adds the error unit over the lenght. say it is spec'd for +/-
0.001" per foot. a 6 foot lenght could be 6 ft 0.006". or is may be
that one foot is puls 0.001 and the next is -.001 so the overall lenght
is exactly 6 ft 0.0000000 but in the middle things can be off.

Also, encoders may be able to report to 2,000 points per revolution,
but that does not mean they are perfect. in other instruments, we say
the accuracy is good to +/- one step.

What I'm saying is that we often talk about theoretical accuracy. If
you are doing metal work or other home stuff this should all work out
great.

only a few people require the attention it would take to get the
precision many of us think we have. Or as the saying goes, it cost
money to get precice, how precice can you afford to be ?

or, it takes 90% of the time to get close to spec's and another 90% of
the time to get there.

Dave