Re: stepper vs. servo - Allegro and ignoring the edges....

--- In CAD_CAM_EDM_DRO@yahoogroups.com, "Mariss Freimanis"
<mariss92705@y...> wrote:

> That is the point I disagree with. Though a motor may have a +/-3%
> non-accumulative tolerance, it does not mean the steps adjacent to
> the current one will vary by +/-3%. In practice they vary by much
> much less than that. The error is cyclic, i.e. over the span of a
> full revolution, a "wobble" if you will.
> Mariss
>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, "turbulatordude"
> <davemucha@j...> wrote:
> >
> >
> >
> > As I read that, it seems they are saying that even though
> > the calculation says the current should be 98.1% of the fla,
> > the motor error is 3% and this it is not worth chasing.

I think the talk about motor error has muddied the water.
They are talking sin/cos currents. They do 16x microstepping,
so each step is 5.625 electrical degrees. So the step
currents would be:

uStep Angle Sin Cos
0 0.000 0.000 1.000
1 5.625 0.098 0.995
2 11.250 0.195 0.981
3 16.875 0.290 0.957
4 22.500 0.383 0.924
etc....

The electrical angle is arctan(sin/cos). If they
use 1.000 for the cos term for the steps closest
to zero, the result will be as follows:

uStep Sin Cos Angle Error
0 0.000 1.000 0.000 0.000
1 0.098 1.000 5.597 0.027
2 0.195 1.000 11.024 0.216
3 0.290 1.000 16.172 0.703
4 0.327 1.000 20.957 1.543

The errors are in electrical degrees. Since there are
50 sin/cos cycles per rev (200 step motor), the mechanical
error is 50 times smaller.

Those errors aren't very big for the first couple steps.
Here they are as percent of a microstep and percent of
a full step:

uStep Error Micro Full
0 0.000 0.00% 0.00%
1 0.027 0.48% 0.03%
2 0.216 3.84% 0.24%
3 0.703 12.49% 0.78%

So if Allegro wants to use 1.000 as the cos term for the
first two steps, I'm not gonna complain...

John Kasunich