This is why the forum is so great, you get ideas you didn't think of.



Great possibility - a much simplified turn table - that could serve 2 tracks in with an exit track out.




Agreed - the turnout solution is dead simple reliable - no moving parts since the turnouts can always be held in the only correct position needed by the springs. No electronics there - just the uncoupler (and recouple) problem. The only downside is width - it requires 2 more tracks (exit tracks is 3 there) than a transfer table style approach (only 1 exit track). Its all a trade off of width and length. And complexity.



Brilliant... this may be the way to go for uncoupling. Indeed, every train that enters will always need the loco uncoupled, so a solution that can reliably uncouple the loco and doesn't move would be perfect.

Now that this though has come to mind, I can see a mechanically pivoting possibility that pushes up once the loco is in a specific position. The slider may be useful in pushing up the uncoupler based on forces needed. And if the locos are about the same length, that may even block the following cars from moving onto the uncoupler (which can be reset by a servo, after recoupling so that acts as a blocking assist for the re-coupling action taking place after the drive-round).



This may work as the passenger locomotives driving in would be nearly all the same length: 37576 (21.9cm), 37013 (18.9cm), 39414 (18cm), so if the lever action isn't driven by the slider, I could instead have the lever positioned to a side of the slider and triggered by an installed bump I place on the locomotives to ensure it is always the exact same distance from the back end of the train.. thus always uncoupling the locomotives. Will need to experiment. Certainly better idea than a bunch of electrically driven uncoupler tracks.

Yes, this would need lots of feedback on the positions of the ends of the cars and loco... but if its all inside the station, I can put sensors up higher (light barriers) that wont' be seen to assist. No doubt this is the engineering challenge of my layout plan. It saves me space if I can figure out how to do it, more other useful stuff on the layout. If that Germany guy got his revolver shadow station working, I figure I've got a decent chance at this as there is WAY less moving mechanical interaction.



Thats not prototypical nor would it be reliable. Imagine an InterCity passenger train being pushed from München up to Hamburg through Germany.



Great idea on exiting from the engine yard - the station building will most certainly have to be a hand built building portion or kit-bashing effort to custom fit the building in width and height (and 2 bldgs if the exit is out of a separate engine yard).


Its a chicken and egg problem for me.. because the solution I end up with will determine how much space I have to and thus may influence the overall layout design in a major way, so I need to figure this out before I do the layout and begin building it. On the plus side, this experimentation just takes some table space a few tracks and some fun engineering.

Another option with switches only - its going to come down to what real-estate will it actually take with real track and turnouts.

The one region is under the roof of the station (so you won't see it unless you look under or remove the roof (for maintenance designed to), while the side return track comes out of a separate building of a train shed, so the locos from the outer two platforms get returned out of the shed, while the inner two get returned by the other inner track (so it requires that track remain clear - so both center tracks both blocked for an incoming train until after the loco has been maneuvered out.

Again, the switches are always in one position held by the springs.



Summary
From this brainstorming, idea suggestion, and a Stummi thread from a few years back that describes the operations at the two end stations I want to have (though not mimic in size/scope/operation necessarily) there are several forms of prototypical operation:

1. Train enters w loco at front, loco decouples (while passengers exit and enter), different loco attaches to back, train departs with new pulling loco, freed up loco goes to BW (this will require purchasing more locos <darn>)
   
2. Train enters w loco at front, (passengers exit), train pushes out to staging station/track in the nearby area for re-arrangement, and is then pushed back in by the loco that's been relocated to the proper end (passengers enter),same loco pulls train out on new run
   
3. Train enters w loco at front, (passengers exit), loco uncoupled, switching loco pulls the train out, freed loco drives out, switching loco puts cars back by platform, uncouples and goes away, main loco reattaches (passengers enter) and train pulled out by same loco. YouTube - München Operation
   
4. Train enters w loco at front,( passengers exit), loco uncoupled and loco then exits out via switches through a siding track to drive around the cars, (passengers enter), same loco pulls the train out
   

So from the prototypical operations stand point, all varieties exist.

From the building my self based model that has such a station at the end (two actually, one at each end, just differing in scenery/backdrop/decoration).. I need to implement some form of operation in the space limits, seeking to automate (read: computer control) as much of this as possible (all ideally). Decoupling apparently works pretty reliably if you do it right (sequence of loco movements on the uncoupler), the re-coupling is the trickier bit... will have to test things.

What I'm then left with is deciding which operations are most reliably automated in the space I have, where I can (under the hidden area of the station under the removable roof) implement cheats to simplify/shorten the operations.

• A transfer table lets me avoid having to have the switching track area before the station to facilitate the rearranging of the loco and cars, but adds the complexity of a moving component that has to be orchestrated - so engineering/sensing/control complexity added to reduce space requirement. But the complexity of moving the train around in the switching area before the station is eliminated (which would be more uncoupling/recoupling)
  
• Using switches and a loop-around possibly, eliminates the transfer table complexity but may add more space needed in the hidden part of the station area.
  

In all cases the uncoupling-recoupling problem has to be solved. In all cases, the train coming in to the station and then exiting again will be a complex control system dance of rolling stock, loco(s), switches, and possibly other equipment to make it happen. Automation will be a challenge, but is theoretically attainable, if I can get recoupling to happen reliably. Otherwise automated trains arriving will get stacked at red lights coming into the station as I have to manually take control to recouple locos for exiting trains. On the plus side, the testing of "did it couple" should be easy to instrument for so the control software can try it, test for success, try again, and maybe after 2-3 fails throw an alert to the HUMAN to come help.

Clearly not a layout for one to sit there and watch the trains go round-and-round. Maybe the solution that requires the least amount of uncoupling/recoupling is the best.