Re: RepRap (was RE: [CAD_CAM_EDM_DRO] Re: Fab@Home)

Pete Brown (YahooGroups) wrote:

>I don't get the self-replication aspect.
>
>
>
>I thought, perhaps incorrectly, that no machine can replicate itself due to
>work envelope restrictions and compounding errors.
>
>
>
>

This would only be true if the machine was a single monolithic block.
As long as it is
composed of smaller components, it should be possible. It may be the
case that one
piece, a base-plate, for instance, can't be completely made in one
fixturing.

>For example, if you have a machine with a 5x5 work envelope, the leadscrews
>or at least the ways would be larger than that, and couldn't be replicated
>in a single piece. If they were multi-piece, they'd be too inaccurate.
>
>
>
>And even if you get past that, the inaccuracies in the machine (only
>theoretical machines will ever be perfect) will be compounded much like
>copying from an audio or video tape which was copied from a tape which was
>copied from a tape leaves you with a noisy crappy tape :) This will be true
>whether the limitations are design limitations (like low res) or errors
>(like backlash).
>
>
>
>
>

Assuming that this is a machine that KNOWS how to build itself, that
could be true.
Comparing to well-known biological systems, there apparently are schemes to
correct for cumulative errors.

But, that's not what we are talking about, here. The program to cut or
SLA assemble
a machine's individual parts would not have to be passed through the
MACHINE itself.
They could be posted on the net, for instance, and stay unaltered. I
believe it would
be up to the person assembling each machine to align the components, and
possibly
carry out calibrations on the completed machine before making anything.

If leadscrews or linear slides were being made by a succession of
machines, then
there could be all manner of incremental error that would eventually
make the
machines immobile, not just innacurate. I haven't been following this
closely
enough to know whether they are trying to do anything like that. Unless
they are
of the Doug Goncz persuasion, where self-replication is a primary goal
of the project,
then there really is no point in trying to make small leadscrews and
linear slide
components yourself.

Jon