a question about signals

[Bartleby]

Why is a one-way signal red when the next block is free but the exit signal ist still blocked by another train ?
(With exit signal I mean any signal that terminates the block, not the specific one)
Would it not be more convenient if any one-way signal shows green whenever the block is free no matter if "exit signals" are blocked as they are red anyway ?
The behavior as it is now leads to difficulties with path signals and stations, as a path is blocked even if the station is free cause a leaving train blocks the exit signal direct after the station. It also makes the signal gap +1 for reasons i do not understand.

That default behavior may be necesairy for two-way signals (not sure) but certainly not for one-way signals.

[planetmaker]

May I suggest to illustrate your issue with a screenshot or better even a savegame?

I've the feeling you didn't quite yet understand how entry-combo-exit signal blocks are supposed to work. Read the signal tutorials (1, 2, 3) how signal blocks work and how entry-, combo- and exit-signals help to extend their functionality. It's usually a bad ide to combine path signals with these block signals, they're different concepts.

[Bartleby]

Hulesch & Quenzel, 14-12-1960.png

(54.66 KiB) Downloaded 1 time

Sorry it was obviously a bad choice to say "exit signal".
I do not mean the exit signal type (yes i fully understand how those combinations work), I mean ANY signal that terminates a block.
ANY signal that terminates a block LOCKS the block if a train (better the last waggon of that train) stands on the signals tile.
My question was, why this is as it is, because it seems unnecesairy for me, as any train stops a tile BEFORE the red signal.
Thank you.

It seems that all signal-tiles that terminate a block count to the block. But i would rather expect that a signal that leads out of the block (therefore I said "exit signal") does NOT count to the block but to the next one.

[FLHerne]

Real signals also have an area beyond the signal that counts as part of the preceding block. This is in case of trains being unable to stop exactly at the signal. While this isn't 'needed' in OTTD (all trains can stop instantly) it would look a bit silly if there were no gaps at all between stopped trains .

[Bartleby]

Dunno if it would look silly

I guess it is the way it is coded:
If a train enters a signal tile all signals that are connected to the preceding block are switched to red. When a train leaves the signal-tile all this signals go green again. Easy and elegant.
But to change the behavior i preferred, would not make much difference. The switch to green had to occur when the tile preceding the signal is free again.

But thanks for the answers. I understand it now.

[Michi_cc]

It seems that all signal-tiles that terminate a block count to the block. But i would rather expect that a signal that leads out of the block (therefore I said "exit signal") does NOT count to the block but to the next one.

If you want to understand why it is the way it is, just consider what would happen if the top right train in your example picture would reverse (in exactly that spot). After reversing, the engine would already be past the signal so if the entry signal would already be green, the two trains could crash.

-- Michael Lutz

[Bartleby]

Well a train should not be able to turn around if part of it stands on a one-way signal ...

But I see the point.

Probably things would be different if signals were between tiles and not ON tiles ( as it is in simutrans).

My main reason for this thread, was the problem that arises when a train, that is entering a station, waits for a free path, despite platforms being free, because the platform it has choosen is free but the signal behind is blocked by another (leaving) train. But maybe there is some other solution for this.

[planetmaker]

But maybe there is some other solution for this.

Yes. Proper signaling of the station entrance so that trains always go to the free platform(s) instead of waiting for an occupied one.

Simplest proper signaling is ONE path signal before the tracks branch to the station tracks (and no further signals facing that direction before the station platforms themselves).

[Bartleby]

But maybe there is some other solution for this.

Yes. Proper signaling of the station entrance so that trains always go to the free platform(s) instead of waiting for an occupied one.

Simplest proper signaling is ONE path signal before the tracks branch to the station tracks (and no further signals facing that direction before the station platforms themselves).

And exactly with THAT layout it can happen that a train waits for a free path DESPITE there are enough free platforms.
In the case a train that leaves the station is blocking a signal behind the platform the incoming train selected.

[audigex]

I see his point - the following train has to wait a full train length behind the train going in, so that with one entrance track there is a limit to how many platforms can be gainfully employed. The more practical solution would be to just have two places to wait for the station, which would act like a buffer.

[planetmaker]

But maybe there is some other solution for this.

Yes. Proper signaling of the station entrance so that trains always go to the free platform(s) instead of waiting for an occupied one.

Simplest proper signaling is ONE path signal before the tracks branch to the station tracks (and no further signals facing that direction before the station platforms themselves).

And exactly with THAT layout it can happen that a train waits for a free path DESPITE there are enough free platforms.
In the case a train that leaves the station is blocking a signal behind the platform the incoming train selected.

Hm, yes, you're right. Indeed, it does. That's somewhat unexpected.

Btw, it would have been helpful, had you posted the savegame. When discussing path finders, a savegame explains things much better as settings there matter a lot - and in a certain way also the orders even in such simple setup.

[YNM]

Ok, I'm understand fully what is the problem, the main problem is that OpenTTD signals occupy one full tile, not a line, not a vector, not tile boundaries, but exactly, one full tile. Graphics don't have any effect on this, and well, through the train waiting at signals occupy around 1/4 of the signal tiles, it have no effect to this problem. It's simply that the behaviour keeped from TTO, TTD, until OTTD. It's a sort of a basic problem, if I can say.

Sorry to planetmaker, its not about pathfinder. Its all about how signals are done.

A signal between 2 tiles is also nice, through I don't know can it be applied.

[planetmaker]

Ok, I'm understand fully what is the problem
(...)
Sorry to planetmaker, its not about pathfinder. Its all about how signals are done.

I'm afraid you mis-understand the problem the OP hints at: trains don't enter a station even while there's a vialbe path to a free platform. That has nothing to do with signals occupying a tile or being between tiles. Just look at the attached savegame (and read Bartleby's last posting here, again, too). The same problem would persist, if signals were between tiles.

[YNM]

Thank you planetmaker. Just see what he was attach in the second post.

Then, yes - 100% pathfinder problems. But is that happens on all pathfinder (Original, NPF, YAPF) ? Or its only specific to one of them ?

[Bartleby]

The same problem would persist, if signals were between tiles.

Partly disagree.

With "Signals between tiles" a situation like this path problem cannot occur, if a signal is placed directly after the platform and before the first tile of outgoing track, because any train that has left the platform automaticly has passed the signal too.
However if the signal is placed later "the same problem would persist" yes.

So maybe the path algorithm could be changed, that a platform without a possible path is considered as "occupied".

My work-arround currently is to place the signals after the station in a manner, that no train can stop with its last car standing on the crucial signal. (Thou this is a bit tricky when using trains with different lenghts).

[mfb]

You can try twoway signals there, this could help in combination with the PBS entry signal and the setting "red twoway signal = end of line"