Re: [rp-ml] New machines could turn homes into small factories]

From: Adrian Bowyer (
Date: Thu Mar 24 2005 - 14:19:15 EET

John Eric Voltin wrote:
> Maybe I'm misunderstanding the article, but it describes using RP
> machines to produce most of the components for new RP machines without
> incurring an enormous cost. Although the concept is theoretically very
> interesting, we have a long way to go before such a task is possible.
> To actually achieve such a goal, RP systems would have to be capable of
> producing parts from a very wide variety of materials at a
> reasonable/economical cost. Additionally, for such a project to be
> viable, the parts produced have to be completely functional for some
> minimum amount of time within the environment of the new RP machine. In
> other words, limited use parts would not be acceptable. The article
> mentions building a RP system for a few hundred pounds. I assume the
> cost of building a new RP machine would include both the material and
> operational costs of the machine used to make the new system. Its a bit
> hard to imagine building all of the necessary components in different
> materials, buying a variety of complex components such as ICs, and
> assembling a finished RP system for a few hundred pounds (roughly $500 US).
> For the record, I would be thrilled to see this goal achieved.


If I may presume to say so, you're asking the wrong question, vis: "Look at this
conventional RP machine. How do we get it to make another?"

Whereas another way to look at it is: "We have a machine that can make things in
two materials (say): an insulator and a conductor. How can we use that to make
a _similar_ machine that could make the majority of its components?"

Answers to that then lead to two subsequent questions:

   1. What are the minority of bits that are left?

To answer that, we're not going to try to make electric motors and
microcontrollers because they're too complicated and they're too cheap to bother.

   2. How do we maintain accuracy down the generations?

This is far more interesting. But it is a problem that has been completely
solved in conventional manufacturing by callibration. Lathes have not been
getting progressively less accurate since Whitworth. Quite the reverse. And as
soon as you have a computer in the loop, you can store all sorts of maps to turn
merely repeatable systems into accurate ones and so on.


Adrian Bowyer

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