Re: [CAD_CAM_EDM_DRO] Re: machine inaccuracies fixed by software compensation

Jon Anderson wrote:

>
> Well, looking at my Oriental Motor catalog, they express accuracy of +/-
> 3 minutes, or .05 degrees. Quoting from the catalog:
>
> "STEP ANGLE ACCURACY
> Under no-load conditions the stepping motor can maintain a step angle
> accuracy within +/- 3 minutes (.05 deg). This slight error arises from
> difference in the mechanical precision of the stator and rotor teeth and
> variations in the electrical precision of the DC resistance of the
> stator coil.
>
> STOPPING ANGLE ACCURACY
> This refers to the difference between the rotor's theoretical stopping
> position and it's actual stopping position. A given rotor stopping point
> is taken as the starting point, then the stopping angle error is the
> difference between the maximum (+) value and the minimum (-) value when
> measuring each step of a full rotation.

<snip>

> The step angle error is +/- 3 minutes, but only under no load. In actual
> applications, there is always friction load. The angle precision in such
> cases is produced by the angular displacement caused by angle-torque
> characteristics based upon the frictional load. If frictional load is
> constant, the angle of displacement is constant for rotation in one
> direction. When operating from both forward and reverse, however, double
> the displacement angle is produced by the round trip. When stopping
> precision is required, always position from one direction only."
>
> So, that ought to help settle the debate on stepping accuracy, though I
> have to say the issue of doubling the displacement error in
> bi-directional applications has me a bit confused.

Great! Good to have absolute numbers, too, so we know what we're talking
about. The bi-directional thing is just like backlash. If you approach a
particular step position from the opposite direction, you've thrown another
uncertainty into the system, and so the error is larger than if you are always
going in the same direction. That is still a small error, though!

Jon