self-replicating factories?

I was very excited to discover the RepRap last week. Since then I've been trying to investigate (online) the potential of self-replicating digital fabrication machinery. My question pertains to a scale of fabrication a bit larger than that of personal fabrication - the neighborhood or the small town.

I curious if anyone could comment on the technical feasibility (ignoring public support, market limitations, raw material supply chain) of building fully self-replicating factories at the aforementioned scale. If it is not feasible then to what extent could a factory be self-replicated on a small scale? Could specialized factories be built at a city or county level to complete the replication process? What stands in the way, technically, of achieving this?

Other open source hardware projects that I've since come across that also stimulated my thoughts on this are:
Rapatan – metal 3D printer
MultiMachine – all purpose machine-tool
Lasersaur – laser cutter
OSLoom – Jacquard style loom
Torch Table – plasma cutter

... the time-table for fully self-replicating factories schould be something NASA and ESA (and some other space agencies too) are working on - they have plans for 'seeding' robot driven bases on Moon and Mars ... if we manage to beat them on this, this will be the real fun

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Are there any assembly lines that produce any finished goods, of any kind, without involving a human somewhere between the receiving and loading of raw materials, and the moving/shipping of filled crates? How about even a proof of concept of any kind?

We're a ways away yet.

Lots of goods are produced without any direct human interaction. Lightbulbs, microchips, sprigs, cereal, bottled soda, and crayons to name just a few.

From a purely theoretical perspective, most factories are 'self replicating' in the respect they are almost entirely manufactured by similar machines. Assembly is still done by humans on most CNC products, mostly due to the low volume and the high cost of programming/acquiring robots. Even small changes in the design can cost a fortune in programming.

Having a factory that completely reproduces itself is entirely technically feasible. The constraints are almost entirely based on everything we chose to ignore

Personally It would be a good idea for a community place in general to be available for scheduled times so that 1 off items could be made. Would something like this need donated time, or tech for hire to work on the machines? Have any reprap clubs done this?

Smaller is the way to go, rather than larger.

There was a fascinating book I read awhile back, it was basically a taxonomy of self-replicators. An in-depth examination of the design space of self-replicating systems, and the various variables that define that design space.

It looked at theoretical proposals as well as existing systems (biological). One of the interesting observations it made is that size seems to directly correlate to speed of replication. Small systems reproduce faster than large ones. There was a multi-ton deflection based seperator design that took a projected 75 years to replicate. It was space-based, and used amount of deflection in a beam of ions to seprate raw materials, and the same techniques to deposit the raw materials into new parts.

There was another moon based replicator that was (again) huge, that was projected to take 50 years (I think) and then you compare elephants, mice, and bacteria.

Anyway, the size=replication time observation seems to hold across all known existing and proposed self-replicating system. Kinda cool, really. I just wish I could remember the name of that book. It was available for free online, IIRC.

It does mean that smaller is better for our purposes! I think we mainly need to work on degree of self replication, simplicity of assembly, and self-assembly. The electron beam metal depositing thing sounds like a great plan for doing the same thing in metals that we're currently doing in plastics. They might have the opposite problem when in comes to fabricating electronics - how do you deposit an insulator?

Having said that, have you seen the factor-e farm system? It's a proposal to do exactly what you're talking about on a small community farm scale. Their tractor design was kind of interesting, and I'm thinking I'll need to take a closer look at the compressed earth block technology for building myself a house. Probably not for the next 10 years or so though.

---

--
I'm building it with Baling Wire