Mechanical through joint threading
| 50px | This page is part of the RepRec project. Short for Replicating Recomposer Systems. For an index of all pages of this project, see the category page RepRec. |
This page is about mechanically threading drive power motions through mechanical joints to drive motions further down a robotic kinematic chain. This applies to serial robotics or parallel robotics that still have some serial robotic aspects.
Design goals for RepRec systems include:
– No mobile motors, all motors as static as possible.
– No electric transmission across mobile joints.
– Chains rather than belts.
A goal is Factoring Out Actuators.
Thus there is a need for threading motion mechanically across mobile interfaces.
This can be quite challenging.
Since in RepRec systems all motors are supposed to be mounted statically (with some caveats — see No Moving Motors),
and since electrical threading through joints to moving parts is not allowed (no chaincableguides, no spiralcableguides, no sliding contacts),
all mechanical motion must be threaded through joints.
This strongly influencing the RepRec Base Geometry
This may makes 1DOF revolute joints (simple hinges) preferable over 2DOF joints (ball joints & co).
Contents
Macroscale
Why no mobile motors?
Why not electrical transmission across moving joints?
– Moving heavy (iron core) motors around
– long flexible electrical cables (and their small plugs) massively complicate automation of assembly
– adds complex wire management (cable chains) & repeatedly bent cables may wear (minor point)
– sliding contacts tend to corrode, wear and usually are a very bad idea with few exceptions
Why chains rather than belts?
– easier automated robotic handling of chain segments compared to belts
– chain segments can be printed/cast avoiding Replicator Vitamins (belts need fiberglass reinforcement or similar in rubber)
Influence on choice of robotics geometry
Parallel robotic mechanisms can:
– 🤩 reduce the amount of necessary mechanical through joint threading (the mechanism itself carries that info on)
– 🤩 distribute through joint motion threading across multiple pathways
– 🙂 As a boon they are also typically often stiffer than serial mechanisms.
– 🙁 As a downside some parallel mechanisms come with a much reduced range of motion, especially in rotation
(like e.g. the steward platform or delta-bots)
It's easier to mechanically thread motion across
– 1DOF hinge type revolute joints rather than across
– 2DOF+ ball joints (like e.g. the steward platform or delta-bots)
Given all that here's the result:
Parallel mechanisms with 1DOF hinge type revolute joints (that are not steward platforms or dela-bots)
essentially leaves parallel SCARA robotics as a natural self suggesting choice.
This can only serve as the RepRec positioner though. Not as the RepRec orienter.
Additionally needed for a RepRec orienter is a robot wrist with
a high range of reachable space angle (2pi or more).
Preferably in parallel robotics.
Examples
Parallel SCARA robot:
When the (two) arms move then the rotations transmitted through the arms
(for end effector rotation and gripping) change due to the hinge angles changing.
This need to be accounted for by subtraction of this motion.
Basically just a relative drive.
Cartesian Robot:
Use the difference of the main drive chain to a second parallel drive chain
two drive the motion on the gantry bridge (via bevel gears — to sketch).
Same from the gantry to the Reprec poser to tool flange.
Differential drive approach:
– H-Bot … may be interesting
– Core X-Y … the oddly bent drivechains may be an issue especially in assembly
There are two issues with cartesian robot mechanisms though:
– The build-space is always smaller than the build stage ~ this makes "self-replication" more complicated (not impossible).
– RepRec prismatic joints (of some designs) can be more challenging than revolute bearings.
Nanoscale
Why no mobile motors?
Why not electrical transmission across moving joints?
– Nanoscale electrostatic motors may be quite big compared to the driven mechanical structures
– A change of geometry in nano-electrical cables may lead to non-trivial changesin conductivity (ballistic electron transport)
– Bending graphene nanoribbons / nanotubes tin too narrow a radius may bead to non-continuous kinking
Why chains rather than belts?
– dissipation modes of straining an unstrapping material
– potential kinking of belts on narrow radii
– easier automated robotic handling of chain segments compared to belts
Misc
Rolling rather than sliding
Slide bearing solutions preferred over roller bearing solutions (See: RepRec prismatic joints)
due to these being possible in a much more compact fashion.
Particular examples:
- the RepRec sprocket stacks in the [[RepRec poser] of a ParallelSCARA based design
- the Nested axles robot wrist in the [[RepRec orienter] of a ParallelSCARA based design
But is reason to worry about wear for a macroscale implementation.
See: Macroscale Mass Overhead section Macroscale.
