Stuart Stevenson wrote:
> Gentlemen,
>    These questions require a little set up. Here goes.
> 
>    1. The orthogonality of the cinci is not perfect (duh)
>        A. in 14 inches of Z motion the X axis walks .004 in.
>        B. in 14 inches of Z motion the Y axis walks .003 in.
>        C. in 120 inches of X motion the Y axis walk .006 in.
> 
>    2. I have the linear compensation collected and implemented
> 
>    3. The rotary axis compensation has not been collected.
> 
>    4. The centerline of the spindle is not coincident with the
> centerline of the A axis rotation
> 
>    5. The centerline of the spindle is not coincident with the
> centerline of the B axis rotation
> 
>    6. The centerline of rotation of the A axis is not coincident with
> the centerline of the B axis rotation.
> 
>    7. I can collect an ordered set of points [commanded position X, Y,
> Z, A, B - actual position X, Y, Z, A, B] for positions in 3D space.
> 
>    8. question - are partial differentials the answer for 3D compensation?

Not sure - my college math is a bit rusty.

>    9. question - is it possible to do this in real time?

Probably - it depends on how complex the errors are.  A simple linear 
shift as the axis travels should be easy, if it wallows back and forth 
like a drunken sailor, not so easy.

>    10. question - whether real time or not does anyone have the time
> (desire) to develop (or have developed) the equations to calculate
> compensation values
>                           (inputs would be the ordered set of point)
>                           (outputs would be the comp values for X Y Z A B)
>                           These comp values could be for use in
> existing comp files or in dynamic comp.

The existing comp is NOT sufficient.  The current comp applies 
corrections to X base on the position of X only, ditto Y corrections are 
based on the position of Y only.  In your case, you want to apply 
corrections to Y based on the position of X, and corrections to X based 
on the position of Z.

Cross-axis corrections would be done in Kins I think.  The corrections 
could either be added to your cinci-kins module, or put in their own 
maybe more generic module and somehow added "in series" with the cinci lins.

>    11. question - is this a realistic goal? If not I will pursue
> creating a function in my kinematics to correct the orthogonal errors.
> 

It seems doable.  I'm not sure how general the solution can be, there 
will probably always be some machine specific issues.  I suspect trying 
to make it completely general will make the job quite a bit harder.

Regards,

John Kasunich

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to