Limits of in place printing and in place mechanosysnthesis
Overhangs or "standoffs"
FFF/FDM printing:
Usually it is better to keep overhangs at a limit of ~45°
Otherwise one needs support material that then needs removal.
If full automation is an eventual goal (See: ReMec)
then removal of support material (and brims) poses a huge additional hassle.
Yes there are various tricks, but they
either need more fancy machine kinematics (more DOFs than three)
or are finicky and not very reliable and reproducible as they constitute thin wide high aspect ratio standoffs.
Sometimes these go with the name "wavefront overhangs".
Mechanosynthesis:
As at nanoscale gravity is massively overpowered by thermal excitations and vdW forces "overhang" is a bit of a misnomer here.
But thin long high aspect ration "standoffs" pose very similar problems.
Due to severe capability limits of SPM systems early mechanosynthesis will want to go layer-by-layer with very flat top.
Much like most macroscale 3D printers do.
A standoff that curls up ad deep cryo (like lHe 4K) makes top surfaces non flat and problematic.
Worse at higher temperatures it may even moves around but mechanosynthesis at higher temperatures
is a concern for later more advances capabilities anyways.
Support material for mechanosynthesis is difficult nontrivial if even possible.
There are no magically vanishing support carbon atoms with exactly the same bonding distance and lattice constant.
Anything put below will add significant distortions laterally and to the top surface.
Especially the latter causing some amount of challenges.
Anisotropic strength
FFF/FDM printing:
Many but not all 3D printing materials ave a significantly weaker in strength
vertically across the printing layers compared to horizontally.
Especially the easier to print ones (first and foremost PLA)
that do not need a heated chamber and extensive drying.
Mechanosynthesis:
Mostly a non problem here.
Vertical covalent bonding has no difference to lateral covalent bonding
aside the natural crystal anisotropy. See page: Cleavage planes
Clearances, backlash, smooth running
FFF/FDM printing:
In place printed moving parts (gears & rollers) are typically not as good as with post assembly.
Surface divets (or even some stringing) from jumps
(beyond the surface divets form unavoidable seam that is better randomly distributed)
makes the surfaces not as smooth as they could be.
Clarence between parts must be
★ big enough to prevent fusing of the parts but also
★ small enough to minimize clearance and backlash
zero clearance an backlash is not possible
Post assembly in contrast allows for pushing things together
to zero clearance and zero backlash with some pre-tenstion
Mechanosynthesis:
FFF printing like surface defects from seams and stringing are not a thinh for mechanosynthesis.
Mechanosynthesis of on non-bonded interfaces in compressed state is rather questionable.
Likely a better approach will be to use bending and pushing jigs in post assembly.
Throwing out failed parts
In place printing does prevent throwing out bad sub-parts.
Due to no way to throw out bad sub-parts if one part is broken the whole part is.
FFF/FDM printing:
Big and complex in-place-printed structures for serious applications are for the most part not done
as these are typically prohibitively difficult to design and then under-performing.
The error rate in 3D printing is reasonably low (and different nature to atom-by-atom).
So larger systems can be in place printed despite no way to throw out bad sub-parts.
But still if one part is broken the whole part is.
This becomes mostly a factor in increasing the part design challenge difficulty
rather than a statistical likelihood for print failure.
In case of post assembly:
Easy. Just use your eyes and hands to detect and throw out bad parts.
Mechanosynthesis:
Big and complex in-place-mechanosysntesized structures are problematic
Particularly in semi early systems as error rate in early primitive mechanosyntheis is expected to still be high.
Same as with FFF printing larger systems are likely to be increasingly difficult to in-place mechanosynthesizable
(and then under-performing). But higher error rates make it more severe for mechanosynthesis.
In case of post assembly:
Still hard. No eyes just touch test and other probing techniques.
Robotic post assembly system needed to throw it out
Design difficulty from incapability of throwing out of sub-parts is one point adding to other difficulties in the
the outdated concept of a proto-assembler. Beside the difficulty …
★ from debugging of deeply buried systems
★ from demanding everything to fit int certain limit of atom count
… which is a strong forcing factor for in place mechanosynthesis as post assembly hugely increases atom count,
… particularly when sensibly balanced to not have the mechanosysntehsis remain as a massive throughput bottleneck).