(no subject)

From: Gregory Pettengill (Cote Art & Engineering)
To: Jon Hunwick (Delcam  International PLC)
Date: Saturday, July 30, 1994
Forwarded to RP-ML by Greg, 94 08 03
Jon,
I am enjoying this very much. So much infact...<SNIP>
Last you wrote;
>By re-entrant features I mean those that are not mouldable in
>line-of-draw of the tool itself, e.g. undercuts, side pockets etc. for
>which a multi-part tool with side cores is necessary.

With "Flexible Rubber Molds" these re-entrant features are not needed.
Most machine parts can be cast from 1 piece "open face" molds.
It would be very rare to find a part that was made in a RP machine that 
would require more than a 2 piece mold to cast it in.

>Yes, I know that it is possible (and not too onerous) to design the
>FFFF model so it looks **exactly** like the conventional tool.

What I am suggesting is that we (design engineers) look at RP systems
as "Flexible (both bendy & versatile) Mold Making Machines".  I think 
that this is a viable way to get RP systems to turn into "Automatic 
Manufacturing Systems".  This may mean that we have to design a **Genesus** 
of conventional rubber molds.  Ones that will **not** look or act like their 
conventional counterparts.

> We (advert coming up), that is Delcam, produce a CADCAM system 
>called DUCT. This is the UK market leader for the tool and mould making
>industries, and has specific features for finding parting lines round
>objects with complex geometry (see earlier definition of complex).
>Once the part line has been generated the shut-out faces of the mould
>can be surfaced quickly and simply, to ensure correct registration (see
>I'm learning!). The part can then be NC machined or FFFF'd, whichever
>takes your fancy.

Is there any way that I can see a demo, short of going to Small Heath, 
Birmingham, England? :-)
> The problem is that CAD time is still fairly
>expensive for the sub-contractor, compared to the time on an NC
>machine, and it is only by turning big bits of metal into smaller ones
>that toolmakers make money. 

Hmmm...
CAD time here costs about $20/hr, most NC machine time starts at
$50/hr.  Is there  any way that we can make money trading CAD & NC machine time?
How much could I make as a CAD operator over there?
How much would European companies pay for US machine time?

>Another consideration is that most subcontractors these days are sent
>data via IGES, and busting one of those mothers open and finding out
>what is **really** going on can take a while.

Must be nice. All I get are paper drawings that have to be redrawn in CAD.

> If a prototype part is
>wanted the destructive FFFF route can be really attractive,

I agree. For the most part...

> and has
>the added benefit that the mouldmaker ends up with a physical part
>which he can look at.

Nice, but seldom necessary. They can always "look" at the CAD model.

>Tool design is a task which requires a great
>degree of skill if sensible, efficient tools are to be produced. Any
>lunatic with a CAD system could design a tool, but without an intimate
>knowledge of what actually goes on it would be worse than useless.

True, unless by "lunatic with a CAD system" you mean me. :-)
No offense taken.

>that, of course, is the danger - just because it's been done in CAD
>doesn't mean it's right. 

I find that most aberrations to the process occur at the "shop floor"
level rather than the "up front" level.

>Lets also make it clear that we are, in all these cases, talking about
>**prototype** parts only, and usually one-off at that. Because it's a
>prototype, the chances are it will be modified, meaning that most (if
>not all) work done on the tool design will be scrapped, or need
>drastic modification later.

Yes, but I believe that most of the time "1 of a kind parts"  are seldom 
desired. Often only 1 part can be afforded in terms of time/money.