Re: [CAD_CAM_EDM_DRO] Renishaw, digitizers

Something to keep in mind when digitizing. If you plan to scale the part
down, any errors in the scanning will be reduced by the scale factor. I
have done some very small (microscpic) detail work with a Deckel
pantograph at 10:1 reduction. It's amazing how sloppy (relatively
speaking) one can get with the pattern and not be able to see the errors
in the finished part.

I like the concept of the Heim joint idea Pete suggested.
Since the one project I plan to setup scanning for will involve
conductive parts, I came up with this:

A rod is fitted to the Heim joint with clearance so that it just will
slide freely. Snap ring on the upper side retains rod. Top end of rod is
fitted with ball or polished full radius.
Heim joint is fixed inside end of tube. On top of the ball is a plug
closely fitted to the ID of the tube and spring loaded against the ball
with the ball sitting in a shallow tapered socket. A straight axial load
will allow the rod to slide up through the Heim joint, returned by
spring pressure. Radial deflections are allowed by the Heim joint, the
ball being returned to center by the tapered socket. At the measuring
end of the probe, different styli could be threaded in place, matching
the contour of the cutter that ultimately will be used to reproduce the
part. If the end product is to be reduced, the probe tip profile is
increased by the inverse of the ratio. By matching probe profile to
cutter profile, one does not need to worry about generating a true
surface from the point cloud.

Since I only need electrical contact, I simply insulate the part being
probed. One side of a circuit is connected to the probe, the other to
the part being digitized. What this allows is a fast approach until
electrical contact is made, reversal until circuit breaks + some small
amount, then approach for the second signal at a much slower rate.
I'm interested in a fair amount of speed with this. My Ducati engine is
out in the shed, but from memory, the clutch side case would require
probably an 8x10 envelope to digitize. Working on .020 centers, this is
200,000 data points. If it takes 5 seconds per point, this would be 277
hours! Of course, I would not need to scan the entire case at this
resolution, but digitizing for fine resolution can be time consuming.

I spoke to a fellow at Ahha recently about using Ahha to digitize. It
would be fairly easy to write a program that could do all this, with
variable fields to define the envelope being scanned, probing increment,
etc. However there is no function available to write axis locations to
disk. However, it turns out they do have something tucked away that
would allow this with a second PC. It would interface to the Ahha card
and monitor steps sent out for position (assuming of course that no
steps are lost anywhere), and writing axis positions to a file.
He says they are playing with it, and might consider bringing it to
market. No cost estimate yet.

Outside of this, I'm not aware of any programs for digitizing on PC
based controls except for MaxNC, and if one goes that route, one would
probably just use his probe anyway. If anyone here is planning on
writing a PC based digitizing routine, I'd like to hear what you think
of this approach. If it looks workable, I'd be happy to draw up the
probe as I envision it.
I've thought this out to answer my needs, which is to digitize
motorcycle parts to produce scaled miniatures, probably not workable for
someone wanting to digitize non-metallic parts.


Jon