Re: digitizing arm

If my memory serves me well today, it seems that Radio Shack, back
about 1983 or so had a device like this for the TRS-80 computers. It
had a serial interface and worked, sort of, but indeed was a
frustrating thing to use. I had one that I attempted to learn to use,
now that was an experience and it was no where near is nice the the
microscribe unit. I believe the thing used potentieometers for
sensors.


John Guenther

--- In CAD_CAM_EDM_DRO@egroups.com, Andrew Werby <drewid@l...> wrote:
> Message: 12
> Date: Thu, 16 Nov 2000 10:04:42 -0500
> From: "Patrick Huss" <patrick@h...>
> Subject: digitising arm Digest Number 889
>
> Has anyone here ever used microscribe? I have, and it was in the top
ten
> MOST frustrating things I have ever tried. It is capable, if I
recall
> correctly, of measuring to .011" accuracy. But you hold it in your
hand,
> which doesn't even come close to the kind of stability required to
keep the
> scribe steady. Using a probe on a cnc machine, or a laser range
finder (not
> interferometer), will always be IMO a more reasonable experience. As
> Ballendo pointed out in his response re: steppers vs. servos -
'Accuracy is
> a function of the system design.'
> If your hand is an integral part of the system design, accuracy is
> nonexistent. If anyone is seriously considering creating their own
> digitising arm, I highly recommend that you use one to digitise an
object,
> you may have to fight a strong urge to throw the device out the
window!
>
> Patrick
>
> [I hope you managed to resist! Actually I've got a Microscribe, and
I've
> come to like it (I even decided to start selling them). I went
through some
> of the learning curve you describe, until I got a lesson in how to
use the
> thing. While digitizing a stream of points may be fast, it has, as
you
> point out, the disadvantage of picking up every irregularity in
your
> movement. Instead of doing that, here's the correct procedure:
>
> Draw a grid on the object to be digitized. These don't have to be
perfectly
> straight lines, as long as all the lines intersect all the lines
going the
> cross-direction. Put more lines in the areas with the most detail-
smooth
> areas need fewer lines; places with a lot of texture need more. With
Rhino,
> (other programs support the arms and may work differently, but this
is the
> one I know about) you need to either construct a grid of open curves
(think
> of a fringed rug) or a combination of open curves crossing closed
ones
> (like a teepee, with the "poles" overlapping the closed curves (the
> "canvas") at the top and bottom). These configurations will make
surfaces,
> using the Surface from a Network of Curves function. The open/closed
> combination will result in surfaces that have holes at the top and
bottom
> ends, but these can be patched over later on.
>
> Once you've got it set up this way, then start making curves,
following the
> lines you've drawn. Use the Interpolated curve function, with all
snaps and
> constraints off. Don't try to digitize a stream of points, but
instead
> digitize POINT by POINT, paying particular attention to the
intersections
> of your drawn lines. While it's admittedly tedious to click the
footpedal
> for each point, you eliminate the problem of hand-shake, since the
stylus
> is at rest when a point is chosen. Adding more points along the
lines will
> increase detail, if you want it. Since fewer points are used this
way, the
> curves are much less dense, and the file-size won't overwhelm your
> computer. Try it- you may even like it!]
>
> Andrew Werby
> http://www.computersculpture.com
>
>
>
>
> Andrew Werby - United Artworks
> Sculpture, Jewelry, and Other Art Stuff
> http://unitedartworks.com