Re: [CAD_CAM_EDM_DRO] Re: Training a paint arm robot

Bill:

>>time of the pulses from the encoders, one should be able to control

ramp and velocity without having to program it. Of course you would
have to move the arm no faster then the motors.

These files would be large to do it this way. At a read frequency of 45Khz
to get good smooth responce, you would get 45000 12 bit words of encoder
data per second. You could program out the non-movement lag spaces, but even
with those gone, your taking around 80Kbytes per second of movment. If you
job takes 20 minutes to complete, the file would be approximately 108Meg
long for the 20 minute job. Possible to do for sure, I guess thats not too
large these days. ...

If I may suggest an alternative solution...

Probably better to use the final jog coordinates and use a cubic spline
code to join them into smooth contour movements. This would mean you jog all
axis to the next control "knot" and press a button to indicate that it is a
control position. A 6-axis cubic spline is then created from the sequence of
moves to create a contour path which is then G-Coded by telling a cam
program to follow the path to create its G-Code. I think I could get you
the capability of getting the control "knots" put into a file, is it
possible to generate a cubic contour from that information? Any Math wizards
out there?

With this technique, the movement would be smooth and fluid from one knot
to the next and so on. In fact, by altering the constraints of the cubic
spline, the movements would smooth out further and further so the end
program would have a "fluidity" adjustment variable which would equate to
how muscular the painter was for that cut (or how much torque you wish to
apply to the motors during the run.). Hmmm, makes me wish I had more time.
This technique would "simulate" the way a real painter paints without
actually having to rely on getting smooth training movement which can also
be difficult. A knot at each end of the painters stroke would give a nice
smooth back and forth movement with overshoot and speed at end of stroke
being variable by varying the elastisity and magnetism of the knots....

Interesting ...(Stop it. I'm too busy...)
Art
www.artofcnc.ca