Microcomponent maintenance microbot

Diamondoid AP products made out of microcomponents can potentially be taken apart again to do self repair or system upgrades. If the exchange of microcomponents is done while the system is not active it's microcomponents can be disassembled by the upper assembly levels of a compatible nanofactory or microcomponent recomposer device but if the exchange shall be done while the system is running or without disassembling the whole system mechanisms for transport of microcomponents must be present.

Microcomponent maintainance units - small diamondoid AP products out of a single or multiple microcomponents - featuring legged mobility would be a possibility to provide this mobility. They could e.g. hold onto a simple cubic crystal lattice with eight arms and move from interstitial point to interstitial point by arm extension or shrinkage. Further manipulators would allow them to unlock remove and transport microcomponents. For high throughputs a fractal channel design will be necessary but for self repair of the usually low damage rates a simple cartesian channel system will suffice in most practical cases.

Note that microcomponent maintainance units do not have to have such high "intelligence" (fluid dynamics emulation) like (speculative) utility fog has to have. Also they have no means for self replication like the outdated concept of molecular assemblers.

Operation via a tree like access channels system

• This is likely ok self repair with microcomponent maintenance units.
• This is a very bad design for advanced productive nanosystems thus gem-gum factories do not feature such topology.

Why a tree like access channels system is not good for the initial assembly of a product.

One naive (and bad) strategy for production would be to:

• (1) Fill up a volume completely with nano to microscale production devices .
• (2) Let the production devices produce the product that is thereby (a) displacing the production devices (b) forming a tree like channel system
• (3) retract the displaced and now in-the-way production devices through the newly formed tree like channel system (branches to root)
• (4) Note: In this approach resource supply also needs to pass through the tree structure (root to branches)

This approach leads to a number of problems:

Molecular assembler scaffold (bad):
Filling up a volume completely with molecular assemblers makes an inefficient under-performant system. Mainly due to lack of specialization on standard part production. Such a system would run hot slow and would be more difficult to design than other alternatives.

To have enough space for nanoscale factory like specialized assembly lines for standard parts we want to outsourcing the mechanosynthesis of crystolecules onto a chip's surface. We have factored out a thin crystolecule pre-producing chip.

Microcomponent assembler scaffold (bad):
If we are still completely filling up the hole volume above the with microcomponent maintenance units that performs assembly of these preproduced crystolecules and microcomponents (on second and third assembly level) then that would lead to a ridiculous over-performance.

• (1) A volume completely filled with "efficient" nanomachinery would lead to exorbitant levels of nominal throughput
  See: Higher throughput of smaller machinery.
  Sidenote: "efficient" above means low energy turnover which excludes the fist assembly level with its mechanosynthesis that rips and froms almost every singe bond in the product.

"Ridiculous over-performance" may sound nice but that over-performance potential unfortunately cannot in the slightest be tapped into due to a massive bottleneck.

• (2) Channels narrow down due to product buildup (eventually to zero if no channels are to be left) that leads to ...
• (3) Dendridic channel structures are subject to the "fractal growth speedup limit".

Summing it up:
So for an initial product assembly process we wants not only to:

• outsource mechanosynthesis by molecular assemblers done in bulk volume to mechanosynthesis by mechanosynthesis cores in on-chip nanofactories

but we also want to:

• outsource the thereafter following assembly of crystolecules and microcomponents from being done in bulk volume to being done in on-chip nanofactories

This leaves no more in place production in the volume at all.
All of the assembly happens on-chip. And the final product gets extruded rather than "grown" inside out from a scaffold.
All being on a chip also allows for convergent assembly layers all the way up to the macroscale.

Why a tree like access channels system is ok for repair purposes

In the case of self repair the situation is a bit different though.
For most products the resupply and waste-removal turnover rates to expect are very very low.
That is: Many orders of magnitude lower than in a reasonably speedy initial product assembly.
Thus a tree like channel system is likely sufficient for this very slow paced self repair process.

In case of rather high damage rates
(e.g. due to operation in a high radiation environment or operation at extreme temperature)
some other design needs to be chosen.

Related

A macroscopic device with the same restrictions on capabilities is the microcomponent recomposer device.
It also cannot do mechanosynthesis. But the much more space in the macroscopic device still gives the potential to greater capabilities.
Multidevice microcomponent recomposer device systems should come a little closer in capability the macroscopic microcomponent maintainence units.

• Self repair
• Mobile nanoscale robotic device
• Legged mobility
• Fractal growth speedup limit
• Testing of gemstone based nanomachinery

(wiki-TODO: Eventually split-off discussion of nanorobotic devices for testing for product functionality status.)