Needed Parts

For assembly of the minimal prototype demo (as in the photos) you'll need:
– 2 nuts
– 2 screws 
– 2 clips
– 2 end-hullsegments
– 2 pins
– 2 mid-hullsegments (2 holes version suffices)
– 3 centering-wedge-plates
– 2 thin chain-segments
– 0 thick chain-segments
(17 parts in total)

This is designed aimed at printability without a brim for the sake of eventual automatability. Results in this regard are improvable. The clip part and the nut are a bit precarious to print brimless.

Note: With this short of a core-chain and no dedicated flexible elements core-chain tension is only rather crudely adjustable. From too loose to over-tensioned across just two teeth or so.

 

Project Details

★ Nanoscale context: http://apm.bplaced.net/w/index.php?title=ReChain_frame_systems
★ FFF printing context: https://reprap.org/wiki/ReChain_Frame_System
★ demonstrated principles include:
– tensioning
– self-centering
– positive locking …
 

External source file dependencies

★ https://github.com/mechadense/scad-lib-cyclogearprofiles

 

GOOD – Successfully shown

★ Many parts can be connected without one or more screws at every place a pair of parts is mating. => 
★ The few final locks can be positively locked without excessive volume and mass overhead.
– – – –
★ With the chosen geometry core-chain pre-tension can provide sufficient resilience against bending and torsional loads. Despite the rater shallow self centering grooves.
★ Self centering works great to give a defined highly accurate overall geometry and leaves no room for unexpected slips under nominal loads.
★ A brimlessly printable (and supportlress) design works. (improvable)
★ The chosen self centering structures are rather scale independent.

 

BAD – Needs fixing
 

★ This specific design (as it is) is not good for automated assembly. 
It should be hugely improvable though.
The problems: 
– chain-segments are  not well guided on insertion
– the centering-wedge-plated can slide sideways when de-tensioned and loosened so far that re-tensioning does not restore it's correct position.
– pins are unrestrained from falling out when de-tensioned
There were attempts from keeping the pins retained by clips but these where unsatisfactory in many regards. (Increasing assembly complexity, increasing mass overhead, decreasing available hull-segment surface modification potential, unsuitability of hard creeping FFF materials for springs leading to cumbersome geometry and wearout.)  Pin-retainment structures were removed again for the sake of clarity.

★ Clip mechanism shortcomings:
– not constraining to a straight slide-out
– prone to break (for the FFF printed case)
– not properly locking against complete removal 
– small printbed contact plus slanted print-orientation is a bit precarious

★ Core chain shortcomings:
– Functionalities of length-adjustment (low force high range) and tension-adjustment (high force low range) are conflated. This is undesirable due to reasons listed on the project detail pages. The tensioning mechanism can and should be factored out of the frame system.
– Tensioning from the side would be desirable as the ends are supposed to connect to spaceframe/spacetruss nodes eventually.

★ Lack of solutions for connecting to (not yet designed) frame system nodes.
This makes this demonstration of principle prototype not practically usable for pretty much anything even with manually assembly.

 

SO-SO

★ Troubles with micro-cracking in PLA/PHA blend plastic. See pics.
=> possible strategies: lower tension & better material
★ Positive locking works but the mechanism is improvable. Solution upcoming!
★ Chain-pins are …
– unnecessary additional parts (that need retainment)
– impose asymmetric assembly
Fir-tree-joints should be a better option. 
★ Pins and holes in chain-segments could have their edges broken for a funneling effect.
I didn't bother for this prototype though.

 

Misc notes

★ As this is just an early prototype the SCAD code is quite messy and not much cleaned up.
★ This assembly does absolutely nothing 😂 beside being a stiff structural member that is neither monolithic nor relying on many many unreliably friction-only-locked screws between all two mating part pairs.
★ Screw profile generation can be painfully slow here. Known issue.