Re: [rp-ml] FW: Geometric Accuracy of RP parts

From: ben halford (
Date: Tue Jan 04 2005 - 10:24:21 EET

I agree with Markus. We're building in layers so scanning / analysing each xy border is best done with something like a camera (100) or PC desk top scanner (35) in semi-realtime. How to grab z information (CMM) and then do anything about existing errors other than beam / bead compensation of subsequent layers is a another matter. As always, keeping it simple and off the shelf sounds like the best bet.

Markus Hitter <> wrote:
Am 28.12.2004 um 15:15 schrieb Good, Jim:

> Additionally, we would want real-time inspection to verify the
> part-build, as each layer is produced. Is this feasible? Can we make
> it work? Is it over-kill?

Depends on the fabrication process.

Think about stereolithography, where parts are transparent and the same
color as the surrounding build material: You'll have a hard time to
measure anything using some optical measuring system.

Think about the Z-corp process: you could use black glue to build each
layer in white powder. Think about the FDM process: each layer can be
surrounded by support material of a different color. Both times, a
simple snapshot with a camera, evaluated by some approriate software
will tell you quickly and accurately how the build system perfroms. At
least horizontally.

Each time you do real time inspection, you'll have some differences
between the just-built layer and the final part: Inaccuracy of the
layer border (-> side surface roughness), shrinkage/expansion of the
whole part due to post processing, temparature differences etc.

You have to consider more global build failures: layers don't bond
together, etc.

my $0.02

P.P.: no need to post to "".
"" is fully sufficient.

- - - - - - - - - - - - - - - - - - -
Dipl. Ing. Markus Hitter

 ALL-NEW Yahoo! Messenger - all new features - even more fun!

This archive was generated by hypermail 2.1.7 : Mon Jan 02 2006 - 08:09:05 EET