The International Simutrans Forum

 

Author Topic: Time interval with telegraph - directional reservation  (Read 273 times)

0 Members and 1 Guest are viewing this topic.

Offline Ves

  • Devotee
  • *
  • Posts: 1708
  • Languages: EN, SV, DK
Time interval with telegraph - directional reservation
« on: September 13, 2019, 01:22:49 AM »
Hi James and all,

I am updating the help text's that I wrote quite some years ago now, and now I have come to the "Time interval with telegraph" working method. After those years I have learnt most things about the signals, having tried them myself and read in the forums about how they are supposed to work.

However I am now a little stuck with the text for the Time interval with telegraph working method. It seems to me as there is a design choice that was not intentional, or at least made some unfortunate consequences. At least I cannot get it to work properly:

The scene is a single tracked line with passing loops. The trains are initially departed from stations at either ends of the line with station signals. As it is a single tracked line, trains will be facing each other.
Now the station signals creates very nice directional reservations up until the first passing loop.
On the passing loop there is a tiwt (time interval with telegraph) signal on each exit.
When trains departs from the passing loop, there is a very nice absolute block type reservation from the signal all the way to the signal on the next passing loop.
However, the trains will only drive in half speed, since there obviously will be junctions before the following signal, namely the entry and exit point between the line and the passing loops.
As it is a waste to let trians run in half speed, you would try to add tiwt signals along the line in order to make stretches where there are no junctions between the signals.
But that will fail, because now the trains from one end doesnt know if a train from the other end has departed, and so just departs and deadlocks it self.

The situation above forces me to use the station signals also for the exit points of the passing loops, which works flawless. Now I dont know how realistic it where in the real life with passing loops without a station, but the 'non' station signals loose quite a bit of value when you can not actally use them to properly signal a single tracked line. Also, it is very unintuitive to place a station signal not at a station....

Pardon me if I am wrong and have completely misunderstood how it should work, but perhaps I might come with some suggestions on how to make it more consistent:
Let all tiwt signals create directional reservations like the station signal does. The big advantage of them over the non telegraph version, is that it should be possible to signal single tracked lines, and the most significant factor on a single tracked line is in what direction it is currently being traversed.
Also I would suggest to dumb down the signals to only do that, even though there are junctions or whatnot in between, very much like the station signal behaves. To make up for the loss of absolute block functionality, I think that an actual absolute block signal should then be included in the early type signalbox to use within stations or otherwise when desired by the player.
This distinction, I think, would make it more simple for the player to actually understand how those particular signals work, and what advantages a signal has over another.

Note: I did see a directional reservation coming from a standard tiwt signal when using an absolute block double block signal (found in swedish pakset) at the platform signal, and then  a standard tiwt signal a bit out from the station. The line was directionally reserved perhaps 1 time out of 4, the rest being absolute block (due to later oncoming junctions).

Offline Vladki cz

  • Devotee
  • *
  • Posts: 2836
    • My addons, mostly roadsigns, pak128.cs
  • Languages: EN, CS
Re: Time interval with telegraph - directional reservation
« Reply #1 on: September 13, 2019, 06:36:32 AM »
I'm not sure but I thought that the telegraph version was supposed to be used only on passing loops. Try using non-telegraph signals on the line.

Offline Ves

  • Devotee
  • *
  • Posts: 1708
  • Languages: EN, SV, DK
Re: Time interval with telegraph - directional reservation
« Reply #2 on: September 13, 2019, 08:08:16 AM »
I'm not sure but I thought that the telegraph version was supposed to be used only on passing loops. Try using non-telegraph signals on the line.

No, that doesnt work. The behavior is the same as if you have the time interval with telegraph signals. Trains will be dispatched from either end with no checks on what is on the other side of the line, only up to the signal telegraph/ non-telegraph signal that is being put on the line.

Offline Freahk

  • *
  • Posts: 450
  • Languages: DE, EN
Re: Time interval with telegraph - directional reservation
« Reply #3 on: September 13, 2019, 10:12:40 AM »
"Also, it is very unintuitive to place a station signal not at a station...."
For me, this sounds totally intuitive. However, in Germany we have different types of "Bahnhof" (stations), where one of these is the"Betriebsbahnhof", which is simply passing loop or any other more complex setup of junctions, not being used for passenger or cargo exchange. It simply allows faster trains to pass or for shunting.
So I always used station signals for passing loops. It sounds logical to me, not knowing if it is realistic in the UK, an it works.

Offline Ves

  • Devotee
  • *
  • Posts: 1708
  • Languages: EN, SV, DK
Re: Time interval with telegraph - directional reservation
« Reply #4 on: September 13, 2019, 11:42:07 AM »
"Also, it is very unintuitive to place a station signal not at a station...."
For me, this sounds totally intuitive. However, in Germany we have different types of "Bahnhof" (stations), where one of these is the"Betriebsbahnhof", which is simply passing loop or any other more complex setup of junctions, not being used for passenger or cargo exchange. It simply allows faster trains to pass or for shunting.
So I always used station signals for passing loops. It sounds logical to me, not knowing if it is realistic in the UK, an it works.

Yes, I know station signals where used outside stations, but translating that to Simutrans, you would need platforms to make the station signal work for both tracks of the passing loop, not just the track it is positioned on. And that makes it essential a station in Simutrans terms.
As of now you need to have a passing loop with one station signal at either end. That signal will only signal the departing train, making the reversed wing of the signal redundant. What I find unintuitive in this context is that the signal per definition is bidirectional (many graphics versions of it in pakset has two wings also), but only the one way is used.