Re: To many SB; and a slightly OT question.

From: Priit Kull (tre@kodu.ee)
Date: Thu Jun 04 1998 - 21:52:19 EEST


Hi list

I was away from my bench for a while therefore a delayd responce on the
topic of too many SBs.

I do not know the situation in US but while visiting Hannover Messe, which
is the biggest industrial show in Europe, I could not spot a single SB or RP
machinery provider there. Several moulders and casters had RP patterns on
display showing how advanced they are, but if asked, they said that they get
their patterns from SBs, no inhouse equipment. For me it spells that in
Europe the SBs are so busy working instead of looking for new opportunities
that there should be plenty of space for more.

Since the exhibition grounds are huge I could have missed them, but I asked
specifically for rapid prototyping from the information booths.

Priit Kull
tre@kodu.ee

Slightly OT question

Another thing I tried to spot in Hannover Messe, but failed, was a
measuring/sensing solution for my problem.

My main bread is making injection moulds for thermoplastics, but I also
provide turnkey engineering solutions for manufacturing various
stamped/bent/drilled.... parts with the followup engineering servicing of
the production. This includes quality control procedures and gadgets. One of
the parts I am supporting is a nasty piece of metal which starts from a
plain 6mm thick metal sheet but ends up in a very bent 3D part with many
holes in it. Approx 20 production steps that all can go wrong. Measuring it
is a real pain since there are no proper base surfaces left. I do a
measuring report on a ZEISS measuring machine every once in a while, but I
need to be sure that every part is OK. For that purpose I built a mechanical
cage with probes to go through the holes, which every single part has to
pass. Fine so far, but this involves human interference, the probes ware
out, it does not generate any statistisal reports about where the production
is going, I am not happy. What are my options? Several spring into mind. CMM
every part, 3D scan them, video, machine vision, compressed air gauging etc.
All the options have some of the weak points, they are either too expencive,
slow or not accurate enough.

I tried thinking out of the box. A single light diode costs almost nothing.
Computational power costs near to nothing and dropping. There should be
somewhere a manufacturer of some sort of sensors that are sold by meter
(foot). It looks like a Scotch tape and costs maybe 1$ per meter (OK 100$
per meter). I will make a box of appropriate dimensions. I will tape some
sensors in the strategic locations. There will be a fixture inside the box
for rough orienting of the part. I will plug the tapes into a computer and
the software will do the magic. This way it should be possible to have some
couple of hundreds of thousands (millions) individual sensors all busy
measuring my part. It should be possible to generate some useful and
accurate data out of this information. I do not care actually about what
type of sensors they are, optical, ultrasonic, infrared, x-ray,
magnetoresistive, capacitive, whatever, as long as they come in bulk tape
and do not cost a fortune. It is no good for taking absolute measurements,
but it should be excellent for taking relative measurements against a known
good part.

Has anybody heard about a remotely similar solution somewhere? The idea of
spilling this question to the RP community was triggered by the vivid
responce to some question asking information about 3D scanners. This is not
directly RP and also not directly reverse engineering but this is a very 3D
and data handling question and I hope that you may have some interesting
ideas.

We can develop this concept a bit more. Let us make the tape a bit wider and
print it with nice flowery pattern. Now we can use it as a wallpaper and we
can get a perfect 3D surveilance system in any room, and nobody will notice
it.

Once again
Priit Kull
tre@kodu.ee

For more information about the rp-ml, see http://ltk.hut.fi/rp-ml/



This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:45:50 EEST