Signalling

[Tomnick]

The 'mainline' practice of having of several home signals within 'station limits' differs slightly from the LUL practice. The purpose of the latter is clearly to get a second train into a platform behind a departing previous train as quickly as possible - and this is generally, if I'm not mistaken, found on ordinary plain line sections with auto signals.

I think you're referring to closing-up signals

Geoff - sorry for the confusion. I was referring to semaphore signalling practice, as I believe that's where the discussion originated. Thanks for the insight into closing-up signals though!

[JasonM]

Well, I did broach the idea to them when I visited, but they weren't interested. So I asked whether the engine would be accessible in the same way that MS Flight Simulator is accessible to 3rd party software (the various ATC add-ons being a relevant example). Again, they weren't interested - but I don't know what the end result was, whether it is accessible at all.

Geoff M.

I had a feeling you might of asked them, so only time will tell.

Don't worry it will be possible in Zusi 3.

[AndiS]

I think you're referring to closing-up signals. These can be found at London Bridge and East Croydon, amongst other places. There are three signals involved - let's call them L1 for the platform starter, L2 for the signal in rear, which is very close to the platform if not actually on it, and L3 in rear of that. With a train in the station (and thus standing at L1), a second train will be held at L3 because the 1st train is occupying the overlap of L2. With me so far?!

Next, the 1st train starts moving and hits the first track circuit beyond L1. With a typical train, the overlap for L2 will become clear very soon after that. L3 thus clears, first to single yellow, then to double yellow if the 2nd driver is a little slow off the mark. The driver of the 2nd train thus gets a clear run into the platform - and saves 20-30 seconds in the process. Thus by adding an extra section in, which isn't "used" as such in the usual way, gets trains through faster than without it. I hope that makes sense.

Whether KRS can replicate this is another matter.

How would the logic differ from the four-aspect signals included with KRS?
They work fairly well, I guess, in particular when there are no switches. They should clear from red to yellow when the end of the train passes the next signal in advance, changing those in rear accordingly.

There is one problem, though: All UK signals default to clear, and I cannot see a solid and reliable interlocking system anywhere, so you may find a signal change from any aspect to a "worse" one without warning, if some switch in advance is changed.

I should add that I have not tested them, just had a look at the Lua scripts.

[ordan77]

I think you're referring to closing-up signals. These can be found at London Bridge and East Croydon, amongst other places. There are three signals involved - let's call them L1 for the platform starter, L2 for the signal in rear, which is very close to the platform if not actually on it, and L3 in rear of that. With a train in the station (and thus standing at L1), a second train will be held at L3 because the 1st train is occupying the overlap of L2. With me so far?!

Next, the 1st train starts moving and hits the first track circuit beyond L1. With a typical train, the overlap for L2 will become clear very soon after that. L3 thus clears, first to single yellow, then to double yellow if the 2nd driver is a little slow off the mark. The driver of the 2nd train thus gets a clear run into the platform - and saves 20-30 seconds in the process. Thus by adding an extra section in, which isn't "used" as such in the usual way, gets trains through faster than without it. I hope that makes sense.

Whether KRS can replicate this is another matter.

How would the logic differ from the four-aspect signals included with KRS?
They work fairly well, I guess, in particular when there are no switches. They should clear from red to yellow when the end of the train passes the next signal in advance, changing those in rear accordingly.

The situation Geoff describes sounds much like a standard colour light setup, but with a much shortened headway around the platform itself, so the logic is essentially the same. The situation would probably be used in an area where the expected speeds are quite low, due to the shortened distance between signals, and possibly shortened overlaps. (Can you confirm this Geoff?)

This does sort of bring us back to an earlier part of the discussion where the desire to be able to set overlap distances is discussed.

There is one problem, though: All UK signals default to clear, and I cannot see a solid and reliable interlocking system anywhere, so you may find a signal change from any aspect to a "worse" one without warning, if some switch in advance is changed.

Equally bad, there are no two-aspect Y / G colour lights at all - I had a good look through the editor last night, and there are many variants of G / R two-aspect colour lights, but the colour-light distants are missing. This makes creating many routes North of Perth and Dundee (amongst others) impossible until new assets are created, as all (or all but one) semaphore distant have been replaced with the colour-light equivalent, even where the remainder of the signals are still semaphore.

A question: Has anyone tried mixing semaphore and colour-light signals yet? If so, does the logic (such as it is) flow correctly between them?

Another question: Is there any indication of being able to 'fix' a signal? In particular, I'm thinking of semaphore distants, where on Light Railways or where running speeds are low, the distants are allowed to remain 'on' at all times.

[drjimi]

Andi, the difference in the LUL implementation is the 'outer homes' are testing the occupancy of several track circuits within the platform length. Refer to the diagram at the TubePrune link I posted in the other thread - http://www.trainweb.org/tubeprune/signalling1.htm#Multi%20Home%20Signals. They are also 2-aspect signals, each thus allowing a draw-up (usually at reduced speed) toward the next signal. Idea is to reduce run-in time as the train ahead gradually clears the platform.

A good example on the District is the EB slow road from Ealing Common to Acton Town. The first outer home (A) is speed sensing and clears when/if the train is at or below 20mph. The next two (B, C) clear as the train ahead clears the track sections covering about the first and second thirds of the platform. The final (the usual 'home') clears once the train ahead clears the platform and starter overlap. Thus the approaching train can safely close up on the train ahead and reduce its run-in time. TubePrune describes it with actual photos of the area here: http://www.trainweb.org/tubeprune/Speed%20Control%20Photos.htm

But they depend on reading the state of track sections ahead of the train (in the platform area), and not those normally associated with a section starting at the signal. From what I've read here about RS (and what MSTS does) is limit the 'block occupancy' test for a signal to its 'own block'. Thus my idea (as yet untested) of having fake hidden 'info type' signal objects (and their interactives) part ways along the platform that have a state defined by testing their block occupancy. Thus a real signal in rear of them could test these 'platform track circuit signals' for state and set their own state/aspect accordingly.

My hope with RS was that a signal actually tested a track circuit (default it's own immediately ahead) but could also test a circuit of specified ID ahead of itself. The current implementation (as with MSTS) appears to use what I think of as implied track circuits or blocks. As for the apparent RS implementation of "Oh - a train just passed my signal pole, thus my block must be occupied"... BTW how does the sim know how/when the block behind a signal is no longer occupied i.e. train completely past? And regarding the debate about how to represent overlaps (move the sig pole and interactive apart), how far can this be done? Not that I feel that's quite the right solution, as the overlap needs to be 'coupled' to the block ahead.

Just curious at this point, and awaiting the US release when I can research and tinker...

EDIT: Just read Ordan's post, which appeared as I was writing. No yellow/green? Argh - I need those too (repeaters in LUL parlance - yellow warns of a red semi or auto ahead). I was also checking a few of my LUL signal diagrams -- overlap distances on starters are typically around 70m (assumed speed accelerating from a stand being about 25mph at that point), and full-speed overlaps on signals approaching platforms could be 200m given approach speeds ranging from 30-45mph.

I'm itching to get RS and look at all this. I know how to code signal logic (having built the LUL signal set), but if the underlying features and capabilities (e.g. test block occupancy, test route set, etc.) are implemented in 'an unusual manner' (flawed assumptions and underpinnings?), this could all be tricky at best.

Thanks, Jimi

[gmayo]

The situation Geoff describes sounds much like a standard colour light setup, but with a much shortened headway around the platform itself, so the logic is essentially the same. The situation would probably be used in an area where the expected speeds are quite low, due to the shortened distance between signals, and possibly shortened overlaps. (Can you confirm this Geoff?)

Oops, slight .-up in the description. The 1st train standing in the platform will usually be clear of the overlap of signal L2, thus allowing L3 to clear (long trains excluded). *However*, L3 does *not* clear until the 1st train occupies the first track in L1's overlap. So the conditions for L3 to clear actually depend, amongst other things, on a track becoming *occupied*, not clear.

Obviously if there's no train in the station then L3 will clear in the usual way.

Here's a shameless plug - a screenshot of SimSig Croydon (currently in beta) which shows the sequence. The first picture shows a route set from what I called L3 earlier. Note the overlap extends to the middle of the platform.


Next, the 1st train comes along and gets a route into the station with double yellow.


With train 1 stopped at the station, it is clear of L2's overlap, so in *theory* L3 should clear. But using the closing up controls, it doesn't.


Now train 1 has started moving and occupied the next track circuit. Note that L3 is now off - because of something that happened 2 block sections ahead.


Geoff M.

[ordan77]

The situation Geoff describes sounds much like a standard colour light setup, but with a much shortened headway around the platform itself, so the logic is essentially the same. The situation would probably be used in an area where the expected speeds are quite low, due to the shortened distance between signals, and possibly shortened overlaps. (Can you confirm this Geoff?)

Oops, slight .-up in the description. The 1st train standing in the platform will usually be clear of the overlap of signal L2, thus allowing L3 to clear (long trains excluded). *However*, L3 does *not* clear until the 1st train occupies the first track in L1's overlap. So the conditions for L3 to clear actually depend, amongst other things, on a track becoming *occupied*, not clear.

Obviously if there's no train in the station then L3 will clear in the usual way.
<pics snipped>

Thanks Geoff, I understand the sequence now.

[AndiS]

Equally bad, there are no two-aspect Y / G colour lights at all - I had a good look through the editor last night, and there are many variants of G / R two-aspect colour lights, but the colour-light distants are missing. This makes creating many routes North of Perth and Dundee (amongst others) impossible until new assets are created, as all (or all but one) semaphore distant have been replaced with the colour-light equivalent, even where the remainder of the signals are still semaphore.

These are easy modifications based on the G/R signals, as long as you are content with the shape. Of course, we'd rather wait for the famous tools, but if you were poised to fix it now, you would most likely succeed do swap the light colours. Then you merge the code of the semaphore distant with that of light signals and you are there.

A question: Has anyone tried mixing semaphore and colour-light signals yet? If so, does the logic (such as it is) flow correctly between them?

3 aspect colour signals are internally the same as semaphores with distant arm on the same post, as far as I see, so I guess you could swap out each of them individually, as long as there are no junctions involved.

Another question: Is there any indication of being able to 'fix' a signal? In particular, I'm thinking of semaphore distants, where on Light Railways or where running speeds are low, the distants are allowed to remain 'on' at all times.

Easy to comment the one line in the semaphore distant where it reacts to the signal in advance being cleared.

Regarding track interactives: Signals which are not protecting junctions have only one "track link", and this is normally placed next to the signal post. However, there is nothing to stop you from creating one with two links, where the additional one would be placed wherever you want to locate the track circuit to clear it. I guess you could even draw on the existing logic of the junction signal to get what you want, at any rate it is not difficult.

[jamespetts]

...I asked whether the engine would be accessible in the same way that MS Flight Simulator is accessible to 3rd party software (the various ATC add-ons being a relevant example). Again, they weren't interested - but I don't know what the end result was, whether it is accessible at all.

Was any kind of explanation as to why they were not interested proffered? That seems to be a very bizarre thing not to be interested in.

[AndiS]

Looking back to the diagrams from SigSim Croydon, I must say that what you call "closing up controls" would be difficult to implement. You could artificially extending the overlap, or create a signal which only clears of (to double yellows) if two block are clear, i.e., you show red instead of single yellow.

What would not work, as far as I know it, is the red at the platform for the first train. I know of no way how the signals learn about the schedule. You could still make a signal which only clears after the train stopped before it. But that would make no difference, i.e., you could never pass through the station with an empty train or a light engine or whatever.

[gmayo]

...I asked whether the engine would be accessible in the same way that MS Flight Simulator is accessible to 3rd party software (the various ATC add-ons being a relevant example). Again, they weren't interested - but I don't know what the end result was, whether it is accessible at all.

Was any kind of explanation as to why they were not interested proffered? That seems to be a very bizarre thing not to be interested in.

Not really. I think they were a bit daunted by the whole prospect of signalling to even consider going down the road of "remote control". Perhaps the engine doesn't lend itself too well to outside manipulation, not necessarily because of anything they've [not] done, but because a train simulator is actually more complex than a flight simulator. Ok, the physics are more difficult, but essentially being in free-form 3d space like a plane is easier to implement than a fixed track 2d space with signals.

It's a bit of a shame as a lot of what is currently poor signalling could easily be cured by external software. And enhanced in many ways - think of multiplayer over the internet, a 3d lever frame signalbox where you pull the levers and watch your mates driving past.

Geoff M.

[JasonM]

It puzzles me why with the rush hour run on the Pad-Oxd route that if you do not leave the platform after a couple of minutes the signal is put back to red and the turbo sitting outside is given the route into the station across your path. Now there must be something that say's if there is no movement cancel the route and set it for another train waiting. Also when the turbo arrives in the platform to the left of you the signal then to the right of you clears but not the one for your platform, strange.

[ozziedriver]

To see if that situation is time based or signal based
I would move my train forward to past the signal that controls the junction than stop my train and wait and see if the other train tries to moves over to your line
Or move forward so you foul the junction and stop and see if the other train tries to come over to your line
That should give you an indication as to what is controlling the other train

If time based the train should try and come over to your line irrespective of were your train is
If signal based it should sit there until you clear the section after the second signal

[JasonM]

To see if that situation is time based or signal based
I would move my train forward to past the signal that controls the junction than stop my train and wait and see if the other train tries to moves over to your line
Or move forward so you foul the junction and stop and see if the other train tries to come over to your line
That should give you an indication as to what is controlling the other train

If time based the train should try and come over to your line irrespective of were your train is
If signal based it should sit there until you clear the section after the second signal

It is signal based, I moved until the signal went back to red then waited, the turbo just sat and waited outside on the home.
I have been looking for a command in one of the scripts which tells the route to be cancelled if no movements after a set time(this should not be happening unless you get a message first from the signalman asking you if the signal can be replaced to danger in front of you), maybe this is controlled in the tools?

[bigvern]

I hope you submitted your 3185 Jason!