Discussion of proposed nanofactory designs

Back: technology level III

Gem gum factory design as proposed in this wiki

Differences to the other proposals:

• Fewer assembly levels where chosen than proposed in "Nanosystems".
  
• Slightly bigger assembly level steps where chosen than in the "productive nanosystems" video.
  
• A sensible far term target is the focus. Exploratory engineering.
  Not an supposedly easier supposedly more early (via the direct path) achievable design like Chris Phoenix (2013) approach.

Fewer assembly levels that each bridge a bigger size step give:

• much more design freedom (less design restrictions) for products that are to be made by the factory
  at the cost of a bit of speed
• an actual possibility to intuitively understand the size-scales involved. See: Magnification theme-park.

Gem-gum factory design as proposed in the Book "Nanosystems"

Page 422 – "Table 14.1. Manufacturing system paramaters. See section 14.4 for description."

There are a lot of assembly levels. This should make the factory fast but goes at the cost of

• more complex design of the factory, which is ok since it is a far term target after all.
• more design constraints for every product single product that this will be designed to nebe made with this factory, which could be problematic.

It seems hard to decipher the motivations behind all the decisions.

Gem-gum factory design as shown in the "productive nanosystems" video

Three assembly levels are shown. (wiki-TODO: work out the sized and list them here)
There is quite a lot attention to detail in there that will go unnoticed by casual viewers.
...

See main Article: Productive Nanosystems From molecules to superproducts

Primitive Nanofactory Design by Chris Phoenix - October 2003

Sources: pdf-file: (source1 source2), on-citeseerx, html-text

TODO: Include new insights from this sci-nanotech thread, specific sub-post, deadlink

This work is assuming the direct path. It is intentionally close to highly compactly self replicating molecular assemblers tacked onto a chip surface. Plus higher assembly levels added atop. So this system would be quite inefficient and would not feature the performance advantages that advanced far-term target nanofactories will feature.

Meaning of primitive

Instead of incremental technology improvement over technology levels a direct step to diamondoid APM is assumed. (See: Two types of DME design). Thus the presented nanofactory does represent a design that's supposed to be easy to build when probe based mechanosynthetic capabilities are assumed and does not represent a "final" goal near optimal design where we want to end up. Some hints how to get from the "easy to built" one to the "near optimal" one are given in for Section "4.6. Improving the design".

Simple forms of mechanosynthesis and exclusive use of bulk diamond are assumed (graphite and polyyine rods are mentioned later).

The existence of general purpouse mechanosynthetic devices capable of production of 200nm sidelength nanoblocks is assumed (see "Meaning of primitive" above). They are called fabricators.

Bottommost levels

A qualitative distinction of the bottommost assembly levels similar to the one presented here is made further in the document. It is first noted that the factory's organization changes at the bottommost levels. Later in Chapter "5.1. Levels of design" six levels similar but not quite the same as assembly levels are presented.

• The mentioned (1) nanoparts and (2) nanomachines correspond to DMEs and conglomerates of DMEs.
• The (3) nanoblocks correspond to microcomponents with 0.2µm sidelength they are assumed to be rather small thus the nano in the name.
• And three further levels are mentioned: (4) patterns (5) fill regions (6) folds

Convergent assembly

Convergent assembly is a central topic in this proposal nonetheless it isn't summarized in chapter "9. Conclusion and discussion" and you have to carefully read through the whole document to get the picture. Thus a summarization is given here.

A set of eight plus one redundant fabricators (in a single layer) at the bottommost levels is a lot slower then the first mergement stage (atom by atom mechanosynthesis vs simple snap together) The first mergement stage is thus greatly underchallenged and operated way below its potential. A stack of multiple layers of those fabricator sets (or usage of mill systems) could provide the speed but is not chosen to avoid the need of transport through fabricator layers.

A strictly ordered stratified design with octal branching/(stage merging) (plus one redundant) and octal assembly/(nanoblock merging) follows upwards for four convergent assembly stages. (A 2D fractal structur with iterations extruded in the third dimension). The ratio eight to eight leaves the assembly time constant instead of doubling it (as in the case of a not used four to eight ratio) what would be natural according to the scaling law of frequency.

• The excessive idle time inherited from the base of fabricator nine tupels shrinks every stage upward to idle_time/2^4 = idle_time/16
• A 3x3 square turns into a (2x2)x2 block => after four stages empty space builds up (3/2)^4 ~ 5

Above the fourth stage a local microcomputer is situated that steers this production module without feedback control and almost completely reversible. The hole structure of fabricators plus 4 stages plus computer plus logistics is called a production module Continuing upward those modules are assembled in a 3D fractal fashion (adhering scaling law ? - to check).

• This design uses convergent assembly in a non n^2/m^3 ratio of hand-up to mergement. See: level throughput balancing

Vacuum

[Todo: sum up what is sayed about vacuum; unfolding; dripstone cave shaped products].

Related

• General metrics that seem useful for nanofactory design are getting collected at the page about design of gem-gum on-chip factories.