John
The first question that needs answering is "does the hand sculptured model define all elements of the finished article"
If it does then you can go the scanning to nc route or casting to make an electrode.
Problem 1 if you go to the casting route - electrodes need spark gaps, plastic has shrinkage and so does any metal. Therefore any direct casting will need some messing about to get the right size bits and bobs - rapid soon becomes a long time unless you strike it lucky.
Problem 2 - split lines. you will have to try to cast the split line as well unless its pretty straight forward.
If you go through the route of scanning - is it a simple open and closed tool, if so great(ish). If not the route may need something like a complete 360 scan ie on a digibot ( jeffk@digibotics.com )and then split the point cloud accordingly and process through some software that machines point clouds directly - Matra Datavision Styler for example.
If it is a simple open and closed tool, scan it on a patch scanner, say a Renishaw Cyclone (or our type - if interested call), take a 2D profile at the same time, turn the job over and scasn the other side. The control system of the cyclone is called TraceCut, this takes point data captured on a patch scanner, mixes all the different patches together and produce roughing, finishing, electrodes male to females etc. etc. The toolpaths a really good, you can take account of shrinkages , spark gaps etc. Its is a very successful way of doing the job - a lot of hand carved novelty toys out of China are tooled this way. If there are undercuts needing side actions the time to run the hole process can really run on - patch scanning is brilliant for open and closed tooling - but a real pain as the number of faces increases.
If there are internals and you need to check other components, then and only then model the thing up - best one (my opinion) is Styler from Matra Datavision. The best type of scan to use here is a complete, whole body scan ie you mount the job once and let the system get on with it (Digibot II, CT Scan, CGI). If you use the patch route here you can have problems composing the data in the modelling software unless you know the tricks of the trade. Dont try to do it with a solid modeller if its sculptured - no matter what the salesman may say you'll be there for ever.
Some systems automatically fit surface data to point clouds - be warned some CNC systems have a real problem machining some of these surfaces - I've seen a single surface at 250+MB (then my system died through lack of disk space!)
The sculptured piece, what goes with it and the clients budget will dictate the route. The Cylone/TraceCut route would, from your mail message, be the best route by far unless you have undercuts etc. If you can't find anyone in the states give me a call and I'll see if I can track someone down - if not it can always be shipped here.
Best Regards
John Wright
Response Associates Ltd.
email: johnw@ra-ltd.co.uk
Tel: 44 1747 822404
Fax: 44 1747 822545
> We have been asked by one of our customers to
> build a plastics injection mold. That's our business, so
> normally this wouldn't present a problem. The catch,
> here, is that the only data available is a hand-carved,
> full-scale model of the part. The object is of a sculptured,
> multi-faceted, free-form geometry that does not lend
> itself to modeling in a cad package such as Pro/E.
>
> Casting directly from the model might be an option, however,
> the cavity blocks would be approximately 11" x 7" x 2", so
> we have ruled out Kel-Tool.
>
> The best option for us would be to have the model scanned
> into some sort of a surface model that we could output to our
> machining center.
>
> Can someone on this list tell me the best way to approach
> this; or, better yet, scan the object into a useable surface
> model (for a fee, of course).
>
> Thanks for your help.
>
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 - 22:45:26 EEST