Sharing some project planning phase: A (digital) ELECTRO-MECHANICAL Network

[RJSV]

This photo shows, everything is non-attached, so all comes apart, by simple pulling, and wiggling a bit.
The two main, interim walls, separate the 3 compartments, of the 3 Pole switch.

[RJSV]

...the color shots just add a little, to the grey on grey color scheme (of the blank wood).

[RJSV]

This photo shows more detail, how the mechanical switch 'toggles', there showing receptor wheel on left will be getting contact / rotation.
The signal is 'dry switched', meaning that there is no actual rotary signal, until later with a different cycle.

[RJSV]

...and this view, toggle seen causing contact with the receptor wheel, on right side of photo.  That side is the so-called 'chain' side, having (usually) a rotary signal in and rotary signal (chain) output.
   The other side, of switch is the 'SELECTED' or 'ready' state.

[RJSV]

...and a little more detail doesn't hurt:
The compartment shown, is the 'C' channel, situated the lowest of the 3 segments (A,B, and C).
Not shown here, but each of those rectangular compartments is further split, into INPUT, OUTPUT sub-compartments.
Generally, if nothing else, this arrangement gives some structure, to work with while solving basic problems.
   By the way: those other shafts seen poking through, are for the other switch segments, and have to be carefully checked, for clearance.

[RJSV]

   So:. Is this, actually, A LADDER NETWORK ?? (Without realizing it).
   At this juncture, here are some problems, dealing with mechanical switching:
   Stages need self-switching to transparent mode (after current stage, or station, does some local, custom rotary output session).  One method involves a 'split-ganged' layout, having left and right sides. By having right side finish, with control having moved on, to next station, a signal to clear the current left-side can be sent 'backward' from, for example, station #17 backwards to clear the left side on station #16.
   This way there can always be a clear path, for the chained signals (A, B, C) going through all stations #1 through and up to station #18.

[RJSV]

...Actually, a more efficient method, for putting each station into transparent mode, has the signal looping back, but done right there in local station (switch box).
The signal to do this 'self cancel' after box has done, is simply the first signal after transparent mode has been put into place. That's going to be a 'SELECT' that is destined to pass through the following switch box, to act on the box 'twice removed' which is box #18 in the case of sending from box #16. It's a signal pipelined setup, sometimes hard to wrap mind around, as things involving timing often do.
   The difficult part involves how to make such action only happen once (the first time).  Otherwise that extra resistance will be present for everything, and from each station, along the serial line...

[RJSV]

   Another problem area, is when mixing two different signals:  it's probably necessary to prevent 'BACKFEED' of an active signal, down a path that is (only) supposed to be one-way, from the other source.
There are various ways to deal with this. (In electronics that's called 'WIRED OR')
   One partial solution involves using helix or worm-gear drives, as they cannot be driven from reverse direction, that is that while the rotary direction can be either way, CW or CCW, that rotary 'signal' can only go one way, through a worm gear. Plus, worm gear will hold position (of toggle lever, in this case).

[RJSV]

   A CHALLENGE.
   So, NO, I haven't got it to work, yet, but have defined a couple problems to work on.
   As to why design (current) using 'toggle' commutators, when final design might require some other form: The whole package is so new enough, as to be difficult to imagine, as a whole.  Then, typically, many other aspects and problems, of design can be tackled. The substitution, later, of a helix worm gear, turns out to be relatively minor, so the whole project can progress forward

[RJSV]

...also assessing problems with the toggling 'tip wheel' that will do selecting between one output wheel or other. Along with a detent action, is probably needed a bit of 'springy' force between the tip driver and the receptor wheel.  Some use of very springy rubber (wheel) itself might be enough
Also, the direction of the tip drive wheel rotation probably affects the toggle lever-bracket so some testing needed to determine if there is a 'disengage' force to contend with.
   My methodology, (perhaps a bit 'Drunken Warrior's method, lol), is to fabricate, best I can, cross fingers, try it, even with known questionable results. (Some backwards or unplanned signal paths are harmless dissapation effects. But keeping friction down is critical, when so many (10 to 30) units can be involved, in serial string.
   Along those line, consider a switch box having 90% efficiency, (90 % of rotary force gets through).   The percent declines, going down the series network, such that after about 10 stations, the rotary signal is diminished to about 10% or less, of original rotary signal coming from a BASE control unit (with regular motors).
   Probably, even that is too optimistic...
But, I can find out these things, with relatively minor 'workshop' fabs. I certainly would not expect the 'usual' customer, for a KIT FORM product, to go to these lengths.
   Mechanically speaking, I probably will get more skill, out of development, than any hypothetical customer.

   THOMAS EDISON, I've heard, had a WHOLE CREW to go off, and build this / build that.
Makes me wonder; What stories would EDISON's lab staff have to tell ?

[RJSV]

   Here's more thoughts, regarding HOW TO shut off a (rotary signal) switch, when the energy for doing that is supplied by...(You guessed it); supplied by, or sourced,  by the double-pole switch itself.
   Imagine the moving toggle; when the toggle, in standard electric switch leaves the 'pole' contact, that energy is stopped, and same thing occurs with mechanical commutator switches; that is, the source that causes switch movement, in the first place, is no longer there. A small flywheel, perhaps, or some kind of spring, could maintain power, as it winds down; pushing the switch to complete arc of travel.  By using split paths, the arrangement can maintain power, in a portion, (actually literally in the next station), and the 'turning off', or de-selection, can occur, crossways, from right side, over to go transparent mode on left side portion. The right side, in current station, has been put into transparent mode, already, and with the actual right side 'active' state actually being in the following station (confusing!).
   That is a situation where the 'active' status is 'smeared' a little, across a couple of stations simultaneously.
   A FLYWHEEL ?? Man, this sh~π's getting out of hand (lol)...What's next: ...I can't even think of anything, that this design process can't top...
   How about: Requirement for a FLAGMAN and two flares, for warning folks: "Warning, Mechanical Switch."
...
   At any rate, like I said, these are interesting questions, I want to know, and PLUS, besides, there could be other projects that come up, for experimenting with Mechanical Logic options.

[RJSV]

Some more improvements:
   Picture shows 'Back Face' panel, make from 'plexiglass' or similar soft-flexible, so that can hold parts aligned and for testing. As with other internal parts, the face plate has no attaching screws etc.
Actually, for light weight duties, the little (1/8) wood shafts do well holding the plate steady.
Red color portions, are 'A' logical segment, of the 3-part switch box. (Think 3 - Contactor mechanical positioned switch assembly).

[RJSV]

One new part built, is the moveable toggle lever.
It has a 'Tip Wheel', driven by belt from pivot area (has same axis).  The tip wheel has pulley, while belt driver pulley rests or is supported by the stationary pivot shaft. For each of the 3 switches, the 'chain' input is what drives the pivot mounted, and thus the tip wheel, for selecting 1 of 2.
   Photo shows pivot lever or 'toggle arm', omitting the tip and drive wheels (in this partial mock-up).

[RJSV]

   Arrangement for SET and RESET:
   In this picture, it helps to view as 'planes', where the wood panel is 'outputs' compartment.  So going towards viewer's foreground, the several different planes are shown spread out, for clarity.
   First, see larger yellow pulley wheel, meant to match with the blue color drive pulley, seen diag down / left.
That is 'SET', causing the toggle lever to move.
Alternately, logically, the RED color pulley, is meant to match, with the white color wheel, diagonally right nearby.
   Actual 'data' being switched, or commutated, is, yet another 'plane' compartment; that's the two white colored wheels, at the more 'picture foreground' end.

   For orientation; the two faces, front and back, resemble a wall clock, where the clock front face is the 'outputs', and back face has some inputs, or even a 'winder'.
   I like to consider switch box having NORTH-SOUTH axis, from back to front (outputs).  Thus, the sides, of the box, are EAST-WEST, orientation.

[RJSV]

   PROBLEMS WITH SELF-SWITCHING:
   The switch box has a combination of generic components and some customized interconnections,
(for implementing this particular network).

[RJSV]

Photo shows:
   It seemed pretty inconvenient, to have 3 shafts from the other switch segments, penetrating the 'C' segment volume (rectangular compartment).
But, seen in upper left, both the 'A' local output (to do reset), and the 'B' input, are needed, and used right there. Extremely convenient.

[RJSV]

   Photo is to show, generally, the path of that reset, to the 'C' segment of switch, going against the grain, so to speak; The reset coming from the 'A' local output takes the path usually reserved for input, as there isn't any input process. That is the other side (local). The main side is for chaining signal, input to chain output (via the toggle switch with tip wheel).

   This view is from front, towards the back. Altogether, there are 18 layers: 6 for each segment of the ABC 3 section contactor.

[RJSV]

   Jumping ahead, for a second: Here is some outline for a Base controller.
   Currently, have identified 2 key issues, details for implementing the basic path switch, or commutation switch assembly.  One issue, involves mixing mechanical (rotary) signal 'data', inadvertantly causing 'backwards direction signal, into other components.
Other current issue involves how to actually turn off or reset all channels in a given switch box,  meanwhile the preceding boxes have all gone to 'transparent' mode, and so cannot effectively send out local controls.

   For controller aspects, photo showing toy 'robot dog's for wired control, where a clever switch setup helps for sending dual polarity to motors, for doing both directions, CW and CCW (that's counter-clockwise.)
   The toy hand controls will issue approx 3.4 volts DC of either polarity direction, to each Base motor / gear drive. For the Base controller  outputs, there are 3 'chained' signals (channels 'A', 'B', 'C'), with an additional 'local' output, using 'B' to send 'Select' to next station switch box.
   Various other Base Station features mostly manual controls, but with plans towards eventual controller, (Arduino likely).  The rotary switches seen are 2 pole, and 6 positions, for various options while testing / experimenting with series networks of switches.

[RJSV]

Those TOY DOGS, well, they provide some function examples that I could use in my project
   The motor has a two directional separator that puts any clockwise rotation to one output, in this case the dog wags tail, and barks, via small audio chip with little 1 inch speaker. The little PC board and connected motor respond to one POLARITY direction, diode isolated.  The other polarity output direction, of the two cell battery pack gets isolated, mechanically, and provides the 'walking' motions of that dog.

   Anyhow, being more software experienced I need that kind of fairly simple TOY mechanism, for implementing actual progress, towards a working 3Pole - 2 throw switch (3PDT).

   (It's been couple years since I've open one, but have forgotten how that toy doggy does the separation of the two motor directions, into 2 destinations, according to motor polarity.

[RJSV]

   Here is a view, I've been wanting to show: This diagram view, included, shows how there is a sort-of 'DUALITY' in the mechanical commutation switch.  In the view, simplified disks pressed together act as switch, for selecting one or the other. That would be the chain output, or the 'local' output.
   Now, any switch is going to have a TOPOLOGY (no expert; I can barely spell that), where it's a 1 in and 2 (signals) out.  When the topology, or connection layout is extended, as in my series connected, the layout, itself, suggests a 'fourth' terminal, rather than just the 3 used in one stage.
   The match, of the 'B' segment, of 3PDT switch, going out to the next station 'B' switch segment, suggests that a fourth location be assigned as 'B local input'.
This creates a nice, common sense result, putting each local output to be sent into next station's local input.  Formally, that doesn't show much meaning, but in the (series) network, in the bigger view that 'local' input makes sense, and preserves the need, to avoid any spatial offsets, from each station box to the next.
(Otherwise, there would be a 'drift', misaligning each next box, that either needs correction, or simply doing station box placements that aren't in a straight line.
   ...Confusing, YES I know. (Thanks).

[RJSV]

   Wanted to divert, for a minute, to outline the TOY hand control, having a clever polarity reversing scheme
   Both push buttons have a terminal, both normally connected to the (2 cell) batteries minus (-) terminal.
So, by pushing the RED button, for example, the toy 'red' wire gets connected to battery plus (+). For the opposite; pressing the BLUE button,  the TOY 'blue' wire gets connected, to battery plus (+).
   Since in either case, the TOY does get voltage supplied, via it's two wires, you get a 'reversable' or bipolar motor activation.
   (Picture showing the dog bone shaped TOY hand control.)

[RJSV]

   This second picture shows how the plus and minus battery voltage are brought out to two little tabs, for contact when either button is pushed.
Please see little screwdriver tip is just above those two little stationary tabs, for contact when moveable tab(s) on each pushbutton happen to be pushed (button has usual springs and captive housing.)

[RJSV]

   This TOY DOG mechanism shows the scheme for separating out rotation direction. Clock-wise coming in from motor, ends up same, clockwise or CW for short.
   With motor sending counter-clockwise, (CCW), the what I'm calling 'Square shaft', which is the main axel for turning, on both right and left sides.
   The main point is to show, a bit hazy, but in middle of photo, notice the gear has small gear integral, and the small gear now, drives another little gear. That can toggle back and forth, due to to rotation itself causing a rotary force (or reactive).
Result is kind of like a 'Rotary Diode'. The mechanism gears the motor down a bit, then positions that little moveable 'inter-gear', so it's effective path switch.
   Also, though, this is also resistant to backwards applied rotary force, at least in the one direction. It's like it 'slips', when wrong direction trys to drive it.

[RJSV]

This view s bit closer, of the little moveable gear.

[RJSV]

   Just stumbled upon a basic function I need anyway, and that is for some 'slip' or other factor that lets me send a bit 'too much' rotation, to reposition that toggle arm.
This set-up, shows first, emphasis on the movable bracket with inter-gear. That whole thing, when toggle is is in 'drift space' between either output gear, that whole toggle bracket etc is basically a 'sculpture', a 'blob'... until that encounters one or other output gear.
   I estimate, a swing of about 30 degrees, over time of a quarter second (250 mSec), and some thing like base motor at 1100 rpm est.
When that interim switching state finishes, that little gear impacts into the gear edge, the output gear to be driven.
   A big gob of green grease was seen, might help make that little captive gear from turning, during that initial free flight rotation, to impact.