Re: Pulse Gen
Posted by
Jon Elson
on 2000-12-07 12:35:32 UTC
PhilC wrote:
grid, and want to pass through every point along one axis in the path.
This is great for drawing solid lines, circles and ellipses, etc. on a
raster screen, but really doesn't apply well to a machine tool path.
In that, you may want to skip over intermediate 'pixels' at high speed,
or spend some time 'between' them when moving very slowly. The
Bresenham algorithms would be cumbersome for this kind of purpose.
Simple trig is all you need to compute smooth paths at any velocity.
When you project the trajectory to a certain point in time over the
path, you then round off to the nearest discrete position based on
machine resolution, and go there. Although this works best with
floating point computations, that is no big impediment as it once
was.
Jon
> ----- Original Message -----But, the Bresenham algorithm(s) are specifically optimized for a discrete
> From: "Wally K" <cnc4me@...>
> <snip>
> >
> > Thanks for the reply Phil. I do not think the external pulse circuit
> > has to do complicated things like circular interpolation. Why not
> > let the computer do that. Just have the external circuit generate
> > the stepper pulses.
> >
>
> Actually circular interpolation is not very hard at all. Ever heard of the
> Bresenham Circular Interpolation Algorithm? Just a few shifts and adds...
grid, and want to pass through every point along one axis in the path.
This is great for drawing solid lines, circles and ellipses, etc. on a
raster screen, but really doesn't apply well to a machine tool path.
In that, you may want to skip over intermediate 'pixels' at high speed,
or spend some time 'between' them when moving very slowly. The
Bresenham algorithms would be cumbersome for this kind of purpose.
Simple trig is all you need to compute smooth paths at any velocity.
When you project the trajectory to a certain point in time over the
path, you then round off to the nearest discrete position based on
machine resolution, and go there. Although this works best with
floating point computations, that is no big impediment as it once
was.
Jon
Discussion Thread
Alan Marconett KM6VV
2000-12-05 16:58:48 UTC
Pulse Gen
Wally K
2000-12-05 20:43:19 UTC
Re: Pulse Gen
Alan Marconett KM6VV
2000-12-05 21:56:36 UTC
Re: Pulse Gen
Wally K
2000-12-05 23:30:20 UTC
Re: Pulse Gen
Mariss Freimanis
2000-12-06 07:16:05 UTC
Re: Pulse Gen
Mariss Freimanis
2000-12-06 07:55:06 UTC
Re: Pulse Gen
Alan Marconett KM6VV
2000-12-06 11:42:09 UTC
Re: Pulse Gen
Wally K
2000-12-06 13:35:49 UTC
Re: Pulse Gen
Dan Mauch
2000-12-07 06:46:10 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Jon Elson
2000-12-07 12:00:37 UTC
Re: Pulse Gen
Jon Elson
2000-12-07 12:28:46 UTC
Re: Pulse Gen
Jon Elson
2000-12-07 12:35:32 UTC
Re: Pulse Gen
Jon Elson
2000-12-07 12:39:29 UTC
Re: Pulse Gen
Jeff Barlow
2000-12-07 12:46:49 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Doug Harrison
2000-12-07 13:55:36 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Mariss Freimanis
2000-12-07 15:17:36 UTC
Re: Pulse Gen
Jeff Barlow
2000-12-07 15:52:44 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Jon Elson
2000-12-07 15:54:08 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Jeff Barlow
2000-12-07 16:19:48 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Mariss Freimanis
2000-12-07 16:22:45 UTC
Re: Pulse Gen
Smoke
2000-12-07 16:32:55 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Jeff Barlow
2000-12-07 16:40:14 UTC
Re: [CAD_CAM_EDM_DRO] Re: Pulse Gen
Alan Marconett KM6VV
2000-12-07 20:41:50 UTC
Pulse Gen
Jeff Barlow
2000-12-07 20:57:34 UTC
Re: [CAD_CAM_EDM_DRO] Pulse Gen
ballendo@y...
2000-12-09 13:19:31 UTC
Re: Pulse Gen
ballendo@y...
2000-12-09 13:28:03 UTC
Re: Pulse Gen
ballendo@y...
2000-12-09 14:46:59 UTC
Re: Re: Pulse Gen