Re: [CAD_CAM_EDM_DRO] Re: Hexapod question

Tim wrote:

>Hello everyone,
>Can someone answer this stupid question for me????
>What would be the need for a hexapod machine anyway?
>I think it is cool & would probably build one myself if I had the
>time, but can't seem to figure out a use for one??? haha!!
>
>

A true implementation of the Stewart platform gives you a machine
with a workspace volume almost as large as the machine itself,
the ability to work on 5 sides of the workpiece without moving
the workpiece, a much lighter machine for a given size and
stiffness, the ability to move very quickly as it doesn't have
huge masses being moved around, etc.

The disadvantages are two, I think. One is that it requires 6 motors
and associated precision motion hardware (leadscrews, encoders, etc.)

Two, it requires intricate movements of all six motors to even move
in straight line. The computation is already taken care of, but it
requires accurate knowledge of the precise locations of the anchor
points of each end of the 6 struts. This is a difficult problem for an
experienced tool and die maker to do on a larger machine, ie. measuring
the 3D coordinates of the fixed anchors to a few thousandths of an inch
over a span of several feet. Starting with a precision ground plate and
using ball mounts that all have precise heights above the flange would be
a start.

If you don't have problem 2 really pinned down, I don't think the machine
will ever work well. There is probably a way to reverse-engineer these
positions by moving the machine to known positions and measuring the
strut lengths, thus inferring the location of the anchor points. But, that
sounds like a LONG trial and error session, with slowly improving
accuracy and stiffness.

Note that EMC still doesn't give a full Stewart platform implementation,
as it doesn't handle rotations of the platform, just translation.

Jon