News:

Simutrans Tools
Know our tools that can help you to create add-ons, install and customize Simutrans.

Swedish signals and signs.

Started by Vladki, April 08, 2015, 08:16:52 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

jamespetts

On the subject of T-signals, I have found an interesting reference in "The London Tilbury & Southend Railway; a history of the company and the line" (volume 2) at p. 139 suggesting that these might have been used in the U. K. at some point. The text is as follows,

"Double line sections, which always formed the majority of the LT&S system, were originally worked on time-interval. No use was made for signalling purposes of the Electric Telegraph which had been installed throughout. Each station had the usual double-armed 'Station Semaphore' with one arm for each direction, showing horizontal (red light) for stop, 45 degrees (green light) for caution, and vertical (hidden inside the slotted post) (white light) for clear."

What is not clear is whether those were signals for trains entering the station or for leaving it, and, if for entering it, where the trains would be held if not at the signal. It may be that I shall need to look more closely into the history of signalling (although I am a little reluctant to embark upon a whole slew of 1850s slotted semaphores for Pak128.Britain-Ex).

Can anyone tell me a little more about how the operation of these Swedish signals worked, and in particular where trains would stop outside the station, and how trains were signalled to leave the station?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves


Quote from: jamespetts on October 26, 2015, 11:50:24 PM
On the subject of T-signals, I have found an interesting reference in "The London Tilbury & Southend Railway; a history of the company and the line" (volume 2) at p. 139 suggesting that these might have been used in the U. K. at some point. The text is as follows,

"Double line sections, which always formed the majority of the LT&S system, were originally worked on time-interval. No use was made for signalling purposes of the Electric Telegraph which had been installed throughout. Each station had the usual double-armed 'Station Semaphore' with one arm for each direction, showing horizontal (red light) for stop, 45 degrees (green light) for caution, and vertical (hidden inside the slotted post) (white light) for clear."
That very much exactly describe the Swedish T-signal. Not strange, since much Swedish old railway are inspired (and imported) from England.

Quote
What is not clear is whether those were signals for trains entering the station or for leaving it, and, if for entering it, where the trains would be held if not at the signal.
Can anyone tell me a little more about how the operation of these Swedish signals worked, and in particular where trains would stop outside the station, and how trains were signalled to leave the station?

The signals where definitively for trains entering the station, and the driver was supposed to hold 100 meters outside the outermost junction.
I have read that the train gives departure whistle with the horn, only then the signalmen may change the signal status to caution and the train depart. That suggests that the signals where in fact used, possibly from behind the train driver, to deliver the message for departure. However, I haven't yet found that this definitely is the case.
There are also described hand signals to be used by signalmen, but currently Im not sure whether this is for normal train departure or just cases where there doesn't at all exist signal.
Further researching......

jamespetts

I think that I shall have to research the time interval method further.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

#73
Finally, I found some more filling information.
http://www.ekeving.se/b/gdg/regl_mm/BJ_tjgregl_1876.pdf page 11.
Here is described that usually departure signal was given by a signalman, holding his arm horizontal to the tracks, meaning "clear". At dark he would have a white lamp moving in circles.
If another train just has left the station in that direction, he can (at earliest five minutes after the departure of the former train) show the "caution" signal with a green flag and the train may drive with a speed limit.

jamespetts

May I ask whether this is describing the absolute block or time interval method of signalling, or whether this is not clear from the text?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

Quote from: jamespetts on November 04, 2015, 07:13:33 PM
May I ask whether this is describing the absolute block or time interval method of signalling, or whether this is not clear from the text?
Im not sure wether it best fit to time signal or absolute block signal. Those terms are not mentioned in that document.
As many early railways where only single track, the time interval method doesnt really work (as I have understood time interval method cant be used on bi-directional lines?).
Its maybe more like absolute block with only one train between stations, with the aditions that trains may enter an occupied block in the same direction as the first under some circumstances (eg if the former was behind schedule) with the rules earlier described.

Afaik, manual blockage of the tracks via Siemens telegraph mashines where used. The telegraph has been used since the beginning to protect singletrack (bi-directional) and doubletrack.

jamespetts

Ahh, I see - a token block system?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

Well, yes but without any physical objects. I tried to describe the "System M" earlier, but that might have been quite unclear. Hee.. I barely understand myself what I wrote.... :-)
Would you like me to try to elaborate how it works, or do you know how now?

Incidentally, are you considering T-signaling?

jamespetts

I am looking into "station signals" as part of time interval signalling (which is why I asked about that); I am waiting to hear back from somebody at the Signal Record Society who has told me that he will provide me with some more information as to the working of time interval signals at junctions and stations presently (the current implementation in the code I do not think deals with stations and junctions correctly).

When you refer to working "without any physical objects", are you referring to something similar to "tokenless block"; were the signals interlocked with the telegraph to prevent more than one train entering the single line section in opposite directions?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

QuoteWhen you refer to working "without any physical objects", are you referring to something similar to "tokenless block"; were the signals interlocked with the telegraph to prevent more than one train entering the single line section in opposite directions?
As far as I have understood it, the interlocking was a human with the key to the signals. In theory, a station chief could wrongly send a train (and set points and signals to be clear) in the head direction of an approaching train. The "interlocking" is the telegram he got from the machine.

QuoteI am looking into "station signals" as part of time interval signalling (which is why I asked about that); I am waiting to hear back from somebody at the Signal Record Society who has told me that he will provide me with some more information as to the working of time interval signals at junctions and stations presently (the current implementation in the code I do not think deals with stations and junctions correctly).
Interresting with station signals! Are they supposed to be placed like the T-semaphore or what kind of logic will they have?

jamespetts

Quote from: Ves on November 04, 2015, 10:31:52 PM
As far as I have understood it, the interlocking was a human with the key to the signals. In theory, a station chief could wrongly send a train (and set points and signals to be clear) in the head direction of an approaching train. The "interlocking" is the telegram he got from the machine.

Interesting, although in practice for Simutrans purposes this may make little difference from standard token block operation. (Does the token block working method seem to behave correctly for this situation with the exception of the details of T-signal working?)

QuoteInterresting with station signals! Are they supposed to be placed like the T-semaphore or what kind of logic will they have?

I have yet to complete my research on this, and am waiting for Mr. Stirling to reply with his information on the early workings of time interval signalling.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

After I played around with it in Simutrans, I think it would work. With the exception that they should not have to stop outside the signal (to collect the token). Maybe an option in the datfile virtual_tokens=1 or something similar to make different of a token machine (generating physical tokens) and just "token type" of signalling?
Thought could be given if the possibilities to drive, even if there is a train on the line heading the same direction, should be added.

jamespetts

The not stopping at the signal to collect the token would not be too hard, but allowing a train to proceed in the same direction would be so different to how token block signalling operates that it would require a whole new working method (the verification that there is no train coming in the opposite direction would require a directional reservation, not used with token block signalling at present). Also, sending multiple trains in the same direction: were there signals along the track, or was this in effect permissive or even time interval working in this respect? How early were these signals, with the block telegraph instruments, in use? Also, do you know whether these signals were normal clear or normal danger?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

QuoteAlso, sending multiple trains in the same direction: were there signals along the track, or was this in effect permissive or even time interval working in this respect? How early were these signals, with the block telegraph instruments, in use?
As I know, there where no signals along the track, except with remote sidings, protect streetcrossings etc.
I guess its a kind of unintended time interval working (trains where not supposed to go before previous train had arrived to next station). From the very beginning, they used telegraph (Sweden started around 1858 with railways) I read somewhere.
The signals themself where the T-semaphore, and a kind of plate-signal (like your vane signal-graphics), later more kinds of semaphores.

first at 1920's automatic blocking of lines where used
http://www.ekeving.se/lb/aut/

QuoteAlso, do you know whether these signals were normal clear or normal danger?
All signals where, to my knowing, normally showing danger  (except some signals which we do not aim for in Simutrans)

jamespetts

Interesting. May I ask how it was known that the train had reached the next station, and whether the next station was always one with a passing place (in which case, this is just token block and there is no question of sending multiple trains down a block section at once) or whether this could include just a single track halt with a signal (in which case, this gets more interesting, but could possibly just be represented in Simutrans by 2 aspect track circuit block signals, albeit with some adaptation for the T-signal form discussed)?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

The receiving station would call back and tell that the train had arrived.
First by telegraph codes (machine by Siemens)
later by normal phone.

As I have understood, all halts had two tracks or more. Otherwise, I think such a station would be considered as "ungarded", which means the dispatchers at each side of that station would be communicating to each other (dispatchers sitting in "guarded" stations), instead of the singletrack'ed "station", however im not sure.

jamespetts

Hmm - so is this not identical to token block working as far as Simutrans is concerned, except for the T-signals? I am having trouble reconciling what you wrote about sending two trains down the line in the same direction with what you also write about a train having to have arrived at the next station (which would also be a passing place) before the next train could be sent.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

I gues you are right that it more or less is identical to token block signaling between stations.

Reason a train could be sent down a line before the former train had arrived to the next station is, if either the former train was behind the schedule, or this train is an "extra train" (meaning its not figuring on the time schedule).

I do think that they planned their time schedules so no train had to be sent before the former had arrived to the next station. However, when delays occured or there where this "extra train", this could be done in order to prevent further delayment of the schedule.

Junna

Quote from: Ves on November 06, 2015, 03:46:08 PM
I gues you are right that it more or less is identical to token block signaling between stations.

Reason a train could be sent down a line before the former train had arrived to the next station is, if either the former train was behind the schedule, or this train is an "extra train" (meaning its not figuring on the time schedule).

I do think that they planned their time schedules so no train had to be sent before the former had arrived to the next station. However, when delays occured or there where this "extra train", this could be done in order to prevent further delayment of the schedule.

Two trains going the same way would generally be done using a manual time-interval set-up where a certain number of minutes had to have lapsed before the second would be permitted to go, and it required that the next station have a certain ability to handle it. I don't think this was done often, nor is it feasible to do it in Simutrans (certainly without signals) as it would just end up for more opportunity for things to get stuck.

Ves

I could agree that the function to send more trains in the same direction might be too complicated for simutrans.

Vladki

I was trying to assemble all the bits here to what would be feasible to do. I have this proposal:

Early signalling for double track - time interval - 1858-1877 ?
Normal signal - a man with blue hat and flag (red-green-white)
Choose signal - (half of) T semaphore with 3-aspects, no choose aspect.

Early signalling for single track - token block - 1858-19xx ?
Normal signal - a man with red hat and flag (red-green)
Choose signal - (half of) T semaphore with 2-aspects, no choose aspect.

Classic lower quadrant signals - absolute block (may be automatic upgrade from time interval ?) - 1877 - 1907
Normal signal - one arm
Choose signal - two arms
Distant signal - none

Classic upper quadrant signals - absolute block (automatic upgrade from lower quadrant) - 1907 - 1947 (?)
Normal signal - one arm
Choose signal - two arms
Distant signal - green disc
Optionally Combined signal - three arms: signal (upper arm) + presignal (middle arm) + choose aspect (lower arm)
This will make a possibility for new aspect: top and bottom arms at 45° and middle arm vertical, meaning go clear for alt route: 40 + clear. Introduction of distant and combined signal may be different, I just put them here for consistency. Distant signal is at least from 1913: http://www.ekeving.se/b/swb/regl_mm/TjgrEJ_1879-1913.pdf

Color light signals - circuit block, default danger, non permissive, max speed 80 km/h - 1920(?)-now
Normal signal - two lights
Choose signal - three lights
Distant signal - three lights (lower green not used)
Optionally Combined signal - five lights (white not used) - same aspects as upper quadrant combined signal

ATC signals - cab signalling, default clear, non permissive (?), max speed 200 km/h  - 1960/1980(?) - now
Normal signal - 4 lights (3-aspect)
Choose signal - 5 lights (3-aspects+choose) - clear + choose  could be shown as green+white+green to distinguish this signal from combined choose signal above. Top two lights would be always the main signal (green/red), 3rd and 4th presignal (green/white) and bottom light always the choose aspect (green/none).

Do you have any idea what sign could be used for early drive-by-sight signal?
What sign is usual on gates over railway (gates of depot and such) ?

I found something about T-semaphores here: http://www.signalbox.org/overseas/sweden/signals.htm
They say they were working in the same way as in UK.

Also here is quite some info - but in swedish. Maybe you could use it to improve my guesses. http://www.ekeving.se/
Could you find more about the signal on the left picture here: http://www.ekeving.se/si/mek/index.html?
It is also at 1877 section here: http://www.ekeving.se/hi/sj-regl-1858.html

jamespetts

Thank you very much for the link to the article in English on the Swedish signals: that is very useful, especially this part,

Quote
As with early British signals, any driver approaching such a signal showing danger would be expected to bring his train to a stand before any points, rather than stop at the signal itself.

This resolves one specific part of the puzzle of how to implement such signals: the train would have to stop one tile before the last set of points before the signal. There are other difficulties with this type of signal that will have to be considered carefully before I decide whether I am able to implement it. I shall in any event have to do some more research into time interval signals, the operation of which at stations and junctions are difficult to pin down. If there is anything in the Swedish documents that assists with how this worked in Sweden, a translation/summary would be very helpful, as one can reasonably infer that it was very similar there to here (and in any event, the code would be useful for the Swedish pakset).

Incidentally, having signals of one working method automatically upgrade to those of another working method will not work: the player may well end up with signals of a mixed type on the same line connected to an incompatible signalbox. Further, if you allow time interval signals without a signalbox, and then have them upgrading to absolute block signals (which cannot function in the game without a signalbox), things are liable to go very wrong. I should probably simply not allow upgrading to signals of a different working method at all in the code to prevent this sort of thing happening. May I suggest that somebody amends the reference post relating to signal parameters to reflect this?
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

#92
Ok, I have read alot about signals these last days, so I have summarized underneath what I found.

QuoteCould you find more about the signal on the left picture here: http://www.ekeving.se/si/mek/index.html?
That is a so called "plate signal" in the lack of a better translation. I have made a shortshort description underneath

QuoteDo you have any idea what sign could be used for early drive-by-sight signal?
Maybe fig. 17 (Grön bansignaltavla och dess placering)  on this link: http://www.ekeving.se/t/sao41/sif.html?
The "one way sign" could be fig. 10 (Spårspärrskärm) on that link?
"End of choose" could be fig. fig.9 b2.

QuoteWhat sign is usual on gates over railway (gates of depot and such) ?
I dont understand which sign you mean.

Firstly some comments:
Combined signal was to my knowledge never used with semaphores. First with track circuit blocks, a signal could show the state of the next signal.
The maxspeed should be quite higher than 80kmh for the pre ATC signals, like 160kmh as its in pakbritain.
Time interval I nowhere found where ever used (only in the exceptions I have descibed in earlier posts), so I think that this working method should not be used.
Token block was also only very very rarely used and therefore I think should not be included. Although the one train staff also where very rarely used, I think that could be very usefull in Simutrans.

The telegraph has been used since the beginning, using a system called Tam which best describe as both absolute block signaling and later track circuit. Everywhere I read, Tam is what has been used. Therefore I think the working methods of Pak.Sweden should be:
Absolute block
Track circuit block
one train staff (very handy)
cab signal
Moving block


I have some corrections which I will try to explain via the underneath timeline:

-------


*Dawn of railways* Communication station to station via telegraph (absolute block)
All signals has to be set manually = signalboxes have very short range
*Dawn of railways* Signalman with flags (stop/clear, maybe also a choose version (with no aditional choosegraphics))

1858 T-semaphore (stop/clear, maybe also a choose version (with no aditional choosegraphics))
1858 Platesignal (stop/clear, maybe also a choose version (with no aditional choosegraphics))
Semaphores where mainly signalling wether a train where allowed to enter the station.
When they wasnt a T-semaphore, they where placed around 50 meters outside the furthermost point.
Platesignals where painted red and only facing one direction, and could (instead of a semaphore) be placed on either side of a station (outside the furtherst point).
T-Signal meanings:
Left wing horizontal: Stop
Left wing pointing down: Drive
Both wings pointing down: Points are locked to main track and is "ungarded" (mostly used when the station is closed)
Also, the left wing could hang vertically down which would mean Drive and proceed, however that was seldom used at stations (and was not allowed on tracks with junctions)

Platesignal meanings:
Red plate facing approaching train: Stop
Turned along the track: Drive

1880 Presignals (clear/caution)
A green plate facing the train would mean caution, no green plate would mean clear.

1888 "Hävstångsställverk" (absolute block)
1890 "Ställbock" and "Vevställverk" (absolute block)
All three are mechanical signalboxes of different types and sizes.
"Ställbock" is a very small signal"box", or rather a switchgear, often bolted to a wall or where standing directly on the platform. Usually only manouver a couple of points and signals.
"Vevställverk" is a signalbox where the mechanism to change tracks and signals where based on wheels. Length up to 350 meter
"Hävstångsställverk" is a signalbox that uses big arms to change the states of signals and tracks. Length up to 550 meters

1890 Twowinged semaphore (choose signal: stop/clear-main/clear-div)
The two winged semaphore had the logic that when the upper wing alone pointing 45 degres down, the train will enter the main track. If both wings are pointing 45 degres down, then the train will enter a sidetrack on the station.

1906 All semaphores now points upwards

1907 Threewinged semaphore (choose: stop/clear-main/clear-div)
When clearance is given with three wings, it means sidetrack/shorted track. This signal could in principle replace the twowinged semaphore.
Later, when Simutrans trains can reassemble, three wings would be the "call on" aspect.

1910 Electric Signalbox (absolute block)
Could manouver signals at mutch bigger distance, however only signals inside his sighting distance. When track circuit blocks becomes available, the distance could be much higher.

1910 Presignals can alert for more wings on the forthcoming signal (clear/caution/(cant be simulated now, but caution for sideway))
By adding a wing to the presignal, it could alert the traindriver if the main signal to the station would send him to main track or side track

1920 Track Circuit Block
1920 Electric Signalbox (extended range, track circuit block))
1920 Track circuit block signals (stop/clear/caution)
These first track circuit block signals had these states:
Red solid light = stop
Blinking green = caution
Green solid light = clear
(http://www.ekeving.se/lb/aut/index.html)

1922 Modern signals
two light: stop/clear (solid-red/solid-green)
three light: Choose signal: stop/clear (solid-red/2xsolid-green)
four light: Stop/clear/caution (solid-red/solid-green+blinking-white/solid-green+blinking-green)
five light: Choose signal: Stop/clear-main/clear-div/caution-main/caution-div (solid-red/solid-green+blinking-white/2xsolid-green/solid-green+blinking-green/3xsolid-green)
When Simutrans trains may reassemble, the 3xsolid-green should become the "Call On" aspect.
presignal, two lights: clear/caution (blinking-white/blinking-green)
An optional permissive signal could be:
Five light: three aspect permissive signal: Stop/clear/caution/Callon (solid-red/solid-green+blinking-white/solid-green+blinking-green/3xsolid-green)
Do note that some of the very early modern signals could have some variations (eg yellow light, signalling the caution for sideway) but I find it very dificult to find good documentation

1930 Modern presignals get an extern yellow lamp to alert for sideway entry (clear/caution/(cant be simulated now, but caution for sideway))

1975 Modern presignals get three lights
A green light are added to the signal itself (now has two green and one white), removing the yellow "caution for sideway" lamp.
States:
clear/caution/(cant be simulated now, but caution for sideway))


1980 ATC (cabsignalling)
As track circuit signals but speed limit is now raized to 250kmh
All signals should have a yellow plate between the rails (like on pictures of this google search: https://www.google.dk/search?q=atc+baliser&espv=2&biw=1786&bih=901&source=lnms&tbm=isch&sa=X&ved=0CAYQ_AUoAWoVChMIo6uDpYSbyQIVARUsCh2QwAW2

2010 ERTMS (Level 2)
No signals needed

Future ERTMS (Level 3)
No signals needed
------

Questions:
As to the five aspect choose signal I have a question: will the "Clear-main" be showed also when the train is stopping at the station, or is there a way to make this state only show when when the train drive through the station (using the main route)?
Regarding the 3 light choose signal, is that needed along with the 5 light choose signal?
Will there be implemented a signaltype that works as a T-signal?

jamespetts

That is a splendidly comprehensive description - I shall look forward to seeing this emerge. It is excellent to see people making signalling systems for other paksets.


Quote from: Ves on November 18, 2015, 10:38:52 PM
...

Questions:
As to the five aspect choose signal I have a question: will the "Clear-main" be showed also when the train is stopping at the station, or is there a way to make this state only show when when the train drive through the station (using the main route)?

In the current implementation, clear-main is shown when a train is stopping at its designated platform, or passing through on its designated route. I am considering changing this as requested in another thread to show only in the latter case, and to show caution-main in the former case.
Quote
Regarding the 3 light choose signal, is that needed along with the 5 light choose signal?

Is this a question to me about the code, or to another Pak128.Sweden contributor about what signals that you need to make? If the former, I am afraid that I do not understand. If the latter, my apologies for possibly confusing it with the former.

Quote
Will there be implemented a signaltype that works as a T-signal?

I have been giving this some careful consideration and am still not decided. This would be quite tricky to implement: it would require a signal being assigned to a station, and then facing in one direction but nonetheless controlling the movement of trains in both directions (or in theory all four directions if a station was a + shape, which is possible in principle), add code for that single signal to govern all departures from a station independently in each direction (for time interval, which is the method in which that signal was used in the UK, storing independently for each of the possible four directions the time at which the train left), and cause the trains to stop at the right point outside the station rather than, as at present, immediately beside the signal, on the same tile as it. Further, logic would have to allow it to control both entry to and exit from the station and to know in each case which it is doing. This is all made much more difficult by the way in which signalling is implemented, in which trains try to reserve tiles ahead on their route one by one, and, when they reach a signal, engage complex logic as to both whether and to where the train may then proceed from where it is now, and what the state of that signal should be. As may be appreciated, this system is particularly difficult to implement with signals controlling multiple tracks at the same time. I have still not worked out a clear idea of how this all could be implemented in the code, nor do I even understand how the real signals worked to a full level of detail.
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Ves

#94
QuoteIn the current implementation, clear-main is shown when a train is stopping at its designated platform, or passing through on its designated route. I am considering changing this as requested in another thread to show only in the latter case, and to show caution-main in the former case.
I would very much support that change! Maybe make it optional?

QuoteIs this a question to me about the code, or to another Pak128.Sweden contributor about what signals that you need to make? If the former, I am afraid that I do not understand. If the latter, my apologies for possibly confusing it with the former.
haha, I think that was a general question to wether it really would be necesary for a pakset to have two choose signals. "Why would one place the one over the other" would be my question, and Im kind of struggling to find an answer to that :-)

QuoteI have been giving this some careful consideration and am still not decided. This would be quite tricky to implement: it would require a signal being assigned to a station, and then facing in one direction but nonetheless controlling the movement of trains in both directions (or in theory all four directions if a station was a + shape, which is possible in principle), add code for that single signal to govern all departures from a station independently in each direction (for time interval, which is the method in which that signal was used in the UK, storing independently for each of the possible four directions the time at which the train left), and cause the trains to stop at the right point outside the station rather than, as at present, immediately beside the signal, on the same tile as it. Further, logic would have to allow it to control both entry to and exit from the station and to know in each case which it is doing. This is all made much more difficult by the way in which signalling is implemented, in which trains try to reserve tiles ahead on their route one by one, and, when they reach a signal, engage complex logic as to both whether and to where the train may then proceed from where it is now, and what the state of that signal should be. As may be appreciated, this system is particularly difficult to implement with signals controlling multiple tracks at the same time. I have still not worked out a clear idea of how this all could be implemented in the code, nor do I even understand how the real signals worked to a full level of detail.
Great, Im really looking forward to see the results (if it comes that far!) of this :-)
If I may throw in a proposal:
In reallity, all the points at that time where not necesarily securily locked. If you add a BIG speed penalty for "ungarded" points (simulating that the train driver would have to stop the train, run forward and check wether the point is set and the track is clear, then run back again), then a T-signal could "guard" all points in a specified radius (removing the speed limit)? Normal signals (in pre track circuit blocks) would also guard signals behind the signal in a certain distance, and when track circuit blocks become available, all points behind the signal up to next signal (unlimited distance) are considered guarded.
If the line is branching between two stations, a signalbox and a T-signal could be placed and that junction is considered guarded in all directions. Routes would always be booked THROUGH the T-signal, never up and until the T-signal.
Regarding multifacing T-signals: Scroll down on this page: http://www.ekeving.se/si/mek/T-sem/T-sem.html it shows for 4 directions!

jamespetts

Quote from: Ves on November 19, 2015, 12:33:17 AM
I would very much support that change! Maybe make it optional?

Whether an aspect less restrictive than caution is shown will have to depend on whether an onward reservation (i.e., beyond the station) is made. It may be that this can be made optional in the parameterisation of schedules which is the next phase of works after signalling, but I am not quite sure how this might work at this stage and how to prevent there being excessive options cluttering the GUI.

Quotehaha, I think that was a general question to wether it really would be necesary for a pakset to have two choose signals. "Why would one place the one over the other" would be my question, and Im kind of struggling to find an answer to that :-)

One good answer might be that the five aspect one might be considerably more expensive than the three aspect one, both to purchase and to maintain.

QuoteGreat, Im really looking forward to see the results (if it comes that far!) of this :-)
If I may throw in a proposal:
In reallity, all the points at that time where not necesarily securily locked. If you add a BIG speed penalty for "ungarded" points (simulating that the train driver would have to stop the train, run forward and check wether the point is set and the track is clear, then run back again), then a T-signal could "guard" all points in a specified radius? Normal signals (in pre track circuit blocks) would also guard signals behind the signal in a certain distance, and when track circuit blocks become available, all points behind the signal up to next signal (unlimited distance) are considered guarded.

Adding a speed penalty would not simulate truly manually operated points; one would actually have to simulate the train stopping then proceeding again. One would then have to find a way of dealing with multiple trains approaching the points at the same time to determine which train gets priority (which would involve deciding whether a reservation is used, and, if so, how far that that reservation should go (if too far, it might as well be absolute block, which rather defeats the object of these early systems), and then deal somehow with multiple sets of points so as to avoid deadlocks. This would actually become very complex indeed, and it is not clear whether any of this would represent real working practice.

QuoteIf the line is branching between two stations, a signalbox and a T-signal could be placed and that junction is considered guarded in all directions. Routes would always be booked THROUGH the T-signal, never up and until the T-signal.

I am not quite sure what this would actually entail. Most particularly, it would be very hard for code to be written to determine precisely what tiles constitute a single junction and which tiles are outside that junction (it is this kind of problem).

QuoteRegarding multifacing T-signals: Scroll down on this page: http://www.ekeving.se/si/mek/T-sem/T-sem.html it shows for 4 directions!

Interesting! That at least shows an historical precedent for them, but does not make them any easier to code...
Download Simutrans-Extended.

Want to help with development? See here for things to do for coding, and here for information on how to make graphics/objects.

Follow Simutrans-Extended on Facebook.

Vladki

Quote from: Ves on November 18, 2015, 10:38:52 PM
Maybe fig. 17 (Grön bansignaltavla och dess placering)  on this link: http://www.ekeving.se/t/sao41/sif.html?
The "one way sign" could be fig. 10 (Spårspärrskärm) on that link?
"End of choose" could be fig. fig.9 b2.
I have already used these:
- one way sign - red sign (fig 14/15) or dwarf signal at 11a
- end of choose - green sign (fig 17/18) or dwarf signal at 11b
- end of signalling - dwarf signal at 11c - no sign yet.
if I understand correctly, 9b is just 9a turned by 90°. I would prefer something more like a sign.


By railway gate I mean a gate that is used e.g. when track enters factory grounds surrounded by wall or fence. I have made such a gate for pak128.CS here: http://forum.simutrans.com/index.php?topic=9551.msg102556#msg102556
Real life version: http://www.vlakregion.cz/trate/041/vlecky/sadova/sadova_vlecky_13.jpg

The blue diamond on those pictures is a Czechoslovak railway sign meaning "shunting forbidden" and is placed where movements are physically impossible - fixed version on dead end or closed gate, changeable version on derailers, turntables. Old mechanic shunting semaphore was also a blue diamond (otherwise similar to mechanical presignal - it turned upwards when cleared).

So, which sign is the swedish equivalent for this blue diamond? I guess fig 9a, 10 or 12.

Quote
Firstly some comments:
Combined signal was to my knowledge never used with semaphores. First with track circuit blocks, a signal could show the state of the next signal.
in that case I would just make the choose signal with two arms, and the 3-armed semaphore will not be used at all

Quote
The maxspeed should be quite higher than 80kmh for the pre ATC signals, like 160kmh as its in pakbritain.
Here https://en.wikipedia.org/wiki/Swedish_railway_signalling they write, that the max is 80 km/h without ATC.
Here https://en.wikipedia.org/wiki/High-speed_rail_in_Sweden they write, that current max with ATC is 200 km/h

Quote
I have some corrections which I will try to explain via the underneath timeline:

Some of the dates are in contradiction with your earlier post
- presignals
- two/three arm semaphores

Do you have some reference, which is correct?

Ves


Quote from: Vladki on November 20, 2015, 12:43:38 AM
I have already used these:
- one way sign - red sign (fig 14/15) or dwarf signal at 11a
- end of choose - green sign (fig 17/18) or dwarf signal at 11b
- end of signalling - dwarf signal at 11c - no sign yet.
if I understand correctly, 9b is just 9a turned by 90°. I would prefer something more like a sign.
I interpret them all as signs, 9a has a light inside the sign, 9b1 also has a light (unclear if the light is inside or in front) and 9b2 is a normal sign with no light. Fig 10 is also just a sign, just a different shape (but with the same meaning).

Quote
By railway gate I mean a gate that is used e.g. when track enters factory grounds surrounded by wall or fence. I have made such a gate for pak128.CS here: http://forum.simutrans.com/index.php?topic=9551.msg102556#msg102556
Real life version: http://www.vlakregion.cz/trate/041/vlecky/sadova/sadova_vlecky_13.jpg
I don't think they look much different in Sweden than any other place. I really like the look of your pak.cs "fences" over the track, and I think they would be perfect if you just remove the diamond!
Quote
The blue diamond on those pictures is a Czechoslovak railway sign meaning "shunting forbidden" and is placed where movements are physically impossible - fixed version on dead end or closed gate, changeable version on derailers, turntables. Old mechanic shunting semaphore was also a blue diamond (otherwise similar to mechanical presignal - it turned upwards when cleared).

So, which sign is the swedish equivalent for this blue diamond? I guess fig 9a, 10 or 12.
Yes it's some of those signs (fig that would achieve the same. Except, fig 12, which means stop for all electric vehicles (because the catenary stops)
Quote
in that case I would just make the choose signal with two arms, and the 3-armed semaphore will not be used at all
Ok, then save the 3 armed semaphore for the call on aspect when to reassemble trains! :-)
Quote
Here https://en.wikipedia.org/wiki/Swedish_railway_signalling they write, that the max is 80 km/h without ATC.
Here https://en.wikipedia.org/wiki/High-speed_rail_in_Sweden they write, that current max with ATC is 200 km/h
Well you could say there is a 'before ATC ' and an 'after ATC'. Before ATC, the single green light just meant clear without any speed restriction, as speed restrictions where given on signs along the tracks. After ATC, all trains NOT equipped with ATC got this speed limit of 80 kmh on ATC equipped lines. On lines without ATC, the speed signs along the track still gives the speeds (also over 80kmh) to all types of trains. Since equipping a loco with ATC is not implemented in simutrans, there is no point in simulating that.

Apparently the Wikipedia articles about Swedish ATC diverge a bit. On the Swedish Wikipedia, it says that ATC is designed to 250kmh, although the safety regulations from the government only allows 200kmh.
Quote
Some of the dates are in contradiction with your earlier post
- presignals
- two/three arm semaphores

Do you have some reference, which is correct?
I'm sorry that I'm contradicting myself. The dates from the latest post should be the one that is most correct as I have picked information from that Swedish ekeving-site that I find quite thrustworty.

Vladki


Ves

Quote from: Vladki on November 30, 2015, 08:26:20 PM
Railway Gate added to SVN
Sorry if I have missed something, but are there a SVN-server? I could not find anything on either the Github or your webpage..

Vladki

sorry I meant GIT. I was using SVN for a while and got confused...

Ves

Im supposing you mean github? Did you syncronize also, I cannot find it in the Github repository (latest commit nov 18)..?

Vladki

How about now. I still have problems with git/svn commands. svn commit and git commint is not the same...

Btw, is there any way to distinguish ATC signals from non-ATC other than balises? I would prefer some sign on signal pole, to keep the possibility to show signals on right. What about 3.1 or 3.2 from http://www.transportstyrelsen.se/globalassets/global/publikationer/jarnvag/jtf/3h_signaler_systemh_webb.pdf

What about using signal 5.4 for end of choose, and "Grön bansignaltavla" for end of signalling?

Ves

QuoteWhat about using signal 5.4 for end of choose, and "Grön bansignaltavla" for end of signalling
signal 5.4 (page 39) would be excelent for end of choose.
"Grön Bansignaltavla" (fig. 17 from http://www.ekeving.se/t/sao41/sif.html) is also fine as end of signalling.

As far as I know, there are no ways to differ an "ATC-"signal from a normal signal, other than the ballisers. I know there are signs for when the train enters and exits an ATC-line, and those signs along with other "ATC-signs" are at page 96 ("ATC börjar" and "ATC slutar"). However, I think it would be fine to use the triangle or the diamond you suggested.

This makes me remembering that normally on an "infartssignal" (=choose signal), there is a little yellow square sign telling wich station is ahead (page 17, sign with "HOA A1/2"). Would this be suitable to distinguish that this is a choose signal?

Vladki

Quote from: Ves on December 01, 2015, 12:26:13 AM
signal 5.4 (page 39) would be excelent for end of choose.
I was thinking also about sign at page 81 or 27 (2.1), just because it would be easy to do. But I have no idea what they mean.

Quote
As far as I know, there are no ways to differ an "ATC-"signal from a normal signal, other than the ballisers. I know there are signs for when the train enters and exits an ATC-line, and those signs along with other "ATC-signs" are at page 96 ("ATC börjar" and "ATC slutar"). However, I think it would be fine to use the triangle or the diamond you suggested. This makes me remembering that normally on an "infartssignal" (=choose signal), there is a little yellow square sign telling wich station is ahead (page 17, sign with "HOA A1/2"). Would this be suitable to distinguish that this is a choose signal?

Now the signals can be distinguished by number of lights:
2 - 2-aspect stop
3 - 2-aspect choose
4 - 3-aspect stop
5 - 3-aspect choose
I think it is enough. I would use the yellow diamond, or some other plate to mark ATC signals.

BTW do you have any idea if some signals in sweden are "permissive"?

I think about using dwarf-signals (either alone or combined with main signal), for token-block signalling. Even if it is not used in sweden, you might need it for playing a game.

A thought about speed signs - if I guess right it is the round yellow sign with black number at page 41. and the triangles at page 42 are their distant counterparts? What is the sign with triangle instead of numbers?
Is the fig 38 in http://www.ekeving.se/t/sao41/sif.html just an older version of speed sign? (also as 1.9 at page 114)

What is 6.3 (page 64) ? Hinderpale