Re: Digitising/reverse engin

From: Bjorn DeBear (
Date: Thu Jun 19 1997 - 20:31:20 EEST

        Reply to: RE>Digitising/reverse engineering

Dear John,
        We here at Argonne National Laboratory take CT scans of parts that we
wish to reverse engineer. This allows us to capture internal geometries
without destroying the part. We are currently converting the CT scans into
SLC files which we can manipulate and then print on our FDM. It's not a
pretty process, but we have the capability.
         Some time in the future we will turn our CT scans into point clouds,
fit b-splines to the point clouds, and fit constructive solid geometries to
the b-splines, so that we can play with them in Pro E. That's going to take
us a while to figure out, though. This research is still in its beginning

Bjorn DeBear
Date: 6/19/97 8:05 AM
To: Bjorn DeBear
RP World

I'd like to know your experiences, products used and general feelings about
digitising and scanning and the re-use of the data with CAD systems.

We use touch trigger probe digitising although we have access to laser
scanning, non-contact capacitance probe scanning and contact scanning. The
data is output to SurfaceMaster for offset to true surface, or scaled ,
mirrored , male to female etc. We output ASCII point data or IGES/VDA Surface
data for use with our ANVIL-5000 system or we output an stl file directly
from the scanned data.

The main reason we use touch probe digitising is because it is universally
applicable ie we can cope with 95% of all jobs asked about. Laser scanning
would reduce this to 20% (unless there have been some recent innovations ).
Touch probing takes longer - max 20 points per second - but since its an
an-manned operation who cares.

Another product used is TraceCut which turns the captured data directly into
nc data again with male to female translations, scaling, offsetting etc . We
can for example take a piston, digitise the inside forge surface, transform
it , apply spark gaps and cut the die and electrodes - reverse engineering is
viable and usefull.

Our general feeling is this - if the shape can be defined by rational
geometry use the data in the cad system and build up the complete assembly,
if its an arty-farty shape ignore the cad system unless you want to, say, put
pcb inside and need to know the design envelope.

There's my starter - any body like to respond

John Wright

Response Associates Ltd

------------------ RFC822 Header Follows ------------------
Received: by with ADMIN;19 Jun 1997 08:03:59 +0100
Received: from major by with local (Exim 1.62 #1)
        id 0wegAX-00025d-00; Thu, 19 Jun 1997 15:18:29 +0300
Received: from []
        by with smtp (Exim 1.62 #1)
        id 0wegAT-00025S-00; Thu, 19 Jun 1997 15:18:25 +0300
Received: (from root@localhost)
          by (8.7.6/8.7.3/AOL-2.0.0)
          id IAA21654 for;
          Thu, 19 Jun 1997 08:17:37 -0400 (EDT)
Date: Thu, 19 Jun 1997 08:17:37 -0400 (EDT)
Message-ID: <>
Subject: Digitising/reverse engineering
Precedence: bulk

This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:39:45 EEST