re: IJK interp was Re: G code Examples / Cutviewer

Jon,

that's a little scary! I can see why quadrants might have been wanted,
the X & Y's can change direction at the 90 degree points. Actually, the
slopes flip on octants, 45 degree points. In alternating octants, one
axis will move considerably further then the other. You want to make
the "slower" axis motion dependent on the "faster" axis (I could use
some better/proper terms here). I kinda LIKE the "flavor" I posted, I
can understand it! And I believe I have it working correctly, for ANY
angle range specified. Was it because the old Allen-Bradley controls
were limited in processing power? A lot easier now with high-level
languages with floating point and trig support.

I've been threatening to implement 'R' G02/G03's also, I just need to
know the reasonable limits to put on the input arcs. Once I can
calculate the Starting/Ending angles of the arcs, the code I have
written can handle the rest.

I thought you posted (other Jon?) test code for arcs a short while ago
(for Art). Couldn't find it here, I'm sure I SAVED it, so maybe I left
it in my other machine!
What I've been trying to draw (for test code), is a set of various sized
arcs, in alternating directions. Different radii and Starting/Ending
angles would give my code a good test. OK, maybe you'd have trouble
machining it, but the arcs "ending up" at the requested points (starting
points for next arc) would tend to show up any errors! The arcs should
be numerous enough to start/end in all combinations of quadrants,
quadrant boundaries etc. Ah testing!

Alan KM6VV

Jon Elson wrote:

>
> Alan Marconett KM6VV wrote:
>
> > Ballendo,
> >
> > Can you give Manufacturer's names to these "flavors" of G02/G03? The
> > one I first "understood" (?) must be "TO" the center?
> >
> > G00 X 0.191 Y 0.462
> > G03 X 0.191 Y 0.462 I -0.191 J -0.462 (#1 CCW 360 circle @ 67.5)
>
> Old Allen-Bradley controls required the IJK to be unsigned. A negative
> IJK word would cause the most hysterical movements I've seen, such as
> running in an expanding spiral about 10 times and then heading off to
> extreme -X at high speed! (The machinist got a bit hysterical, too!)
> This was on the A-B 7320 series, from the late 70's. Arcs could only
> be specified in quadrants, if you asked it to do more than 90 degrees,
> other strange stuff happened.
>
> Jon