Signal issues

Hi all;

I have some intriguing problems with my signals.
I am building a large route, and have set up some single track with passing tracks.
I have placed signals at each end of the passing tracks, and intermediate signals between the interlockings on the single track.
I have included a track diagram below.

The signals I am using are:
Safetran Clight Signal 08 Interlocking for the interlockings
Safetran Clight Signal 05 for the intermediates.

Problem #1

In a particular stretch of single track that lies between interlockings, I have 4 intermediate signals, and 3 industry sidings located close to one of the interlockings.
The problem I have is that when there are two trains heading towards each other from the interlockings onto single track, it is possible for both trains to enter the single track at the same time.


Note that the diagrams below are captured from Train Dispatcher 3, a CTC simulation
I have not aligned the interlocking at CP10, because I wanted to show what happens in TS when both interlockings are lined to the same section of single track.
This CTC simulation will not allow me to do this, thus I cannot actually show it in the diagram.





Fig 1
Train at CP22 (right) is going to get the railroad while train at CP10 (left) is going to wait until the southbound clears the interlocking.







Fig 2

Train at CP22 has proceeded onto the single track, but at this point, the train waiting at CP10 still has a green light (switch is lined for that train onto the single track as well. This is not allowed on a CTC system, but it is allowed in Trainz.
The northbound signal at CP10 will not change to red until the southbound train is south of the switch for industry G1.




Fig 3.
Southbound train has passed the siding at G1, and northbound signal at CP10 turns red.


I understand why this happens.
TS signals search for trains approaching, but only until they reach a junction. When that happens without finding a train approaching, the search ends, and the signal initiating the search displays according to the next signal.
The approaching southbound is still not detected by CP10 northbound because the junction G1, G2, and G3 (actually, it's the first junction that really matters) lies between the starting signal and the train.

If the scenario is reversed, where the northbound train gets the railroad and the southbound is waiting at CP22, things are different.
As soon as the northbound train goes north of the CP10 interlocking, the southbound signal at CP22 turns red.

This happens because the block up to the first industry, G1 is now considered occupied, and since there are no signals between that point and CP22, the home signal is also red. If there were one more signal between those points, then the home signal at CP22 woudl be yellow (approach).

What I need is a way to tell TS that the industry junctions should be ignored, and the search should continue until it reaches the next interlocking, or home signal (or it finds an approaching train).

I tried adding direction markers to each siding, pointing out, thinking that the search would hit those and ignore the junction, but that did not work.
Short of modifying the signal’s script, I really don’t know what to do.
I don’t know whether modifying the Safetran’s script would even help, since I have also tried this scenario with TS built-in USA 01 (unscripted) signals, and got the same results.
--------------------------------------------------------------------------
Problem #2

There is yet another strange problem, and this one may explain why some TS players have problems when AI is running on signaled track.

If there is a train approaching a home signal, that signal will always be red, to prevent another train from occupying the same track in opposite direction.
If that approaching train is stopped, then reversed, the home signal will change, reflecting the fact that the train ahead is no longer approaching, but moving away.

When I first started my signals testing on a separate route I built for this purpose, I found that sometimes the home signal would change (from stop) to a less restrictive aspect after the approaching train had come to a stop.
I could not figure out what was happening at first, but then I realized that because I was running in DCC mode, the train’s brakes were not being applied properly, and when the train came to a stop, the train would “bounce” backwards just a hair because the train’s brakes were off.
This happens in DCC, because it is a simple way of stopping the train.
The problem is that this is the way that AI runs trains. I confirmed this by setting up a couple of track markers, and added wait commands to have trains stop, then start again along the way.
The result was that sometimes the back-bounce would occur.

When I switched to cab mode, and controlled the train myself, using proper braking procedures, I prevented this backwards bounce, and thus the signal issue.

Anyone have any ideas on problem #1?

Thanks

FW

are you using my latest scripted signals for this test? reason i ask is because obviously no other signals work this way (i double checked, mine would not be affected by junctions in the way). any regular TS signal is going to allow trains onto a single track if there are any intermediate signals at all. i think i gave you the test ones i was making, they do not allow this and the signal will give you a 'conflicting train movement on single track' message if a train is coming towards the absolute signal protecting single track. there is a limit to length and number of intermediate signals searched however. it will only look ahead by 10 intermediate signals, or 10 miles, whichever one comes first (this could be extended i think but ive just never needed it). this search is not affected by turnouts unless they are switched the opposite way of the path. that isnt a problem though and siting it as one would be unfair to the signals, since you wouldnt clear them ahead anyhow without the path being set.
the only exception i can think of with mine is a situation where the train would reverse and move away then come back, this would give the opposing signal the wrong idea about where the train was heading, and let the other train past because according to that signal, the other train is now moving away... but you cant program the signals to know where the train occupying the track will be going in the future. i just think thats pretty unlikely anyway.


#2 - sounds like you used an always controlled signal here. the train will always come to a stop before it can cross these.

Where are your scripted signals?

norfolksouthern37 said:

are you using my latest scripted signals for this test? reason i ask is because obviously no other signals work this way (i double checked, mine would not be affected by junctions in the way). any regular TS signal is going to allow trains onto a single track if there are any intermediate signals at all. i think i gave you the test ones i was making, they do not allow this and the signal will give you a 'conflicting train movement on single track' message if a train is coming towards the absolute signal protecting single track. there is a limit to length and number of intermediate signals searched however. it will only look ahead by 10 intermediate signals, or 10 miles, whichever one comes first (this could be extended i think but ive just never needed it). this search is not affected by turnouts unless they are switched the opposite way of the path. that isnt a problem though and siting it as one would be unfair to the signals, since you wouldnt clear them ahead anyhow without the path being set.
the only exception i can think of with mine is a situation where the train would reverse and move away then come back, this would give the opposing signal the wrong idea about where the train was heading, and let the other train past because according to that signal, the other train is now moving away... but you cant program the signals to know where the train occupying the track will be going in the future. i just think thats pretty unlikely anyway.


#2 - sounds like you used an always controlled signal here. the train will always come to a stop before it can cross these.

Click to expand...

Hi Justin;
Yes, I'm using the latest scripts you e-mailed to me.
I also tried the same scenario with the older ones (DL from Jointed Rail) and the same thing happens.

There is one thing I have never tried: Adding the positions of the distant junctions between home signals to the setup of the 08 signals at CP10.
I have done this for CP22, but never for the southern home, which is about 11 miles from the first junction.
I will try this, and see if that fixes the situation.

FW

norfolksouthern37 said:

are you using my latest scripted signals for this test? reason i ask is because obviously no other signals work this way (i double checked, mine would not be affected by junctions in the way). any regular TS signal is going to allow trains onto a single track if there are any intermediate signals at all. i think i gave you the test ones i was making, they do not allow this and the signal will give you a 'conflicting train movement on single track' message if a train is coming towards the absolute signal protecting single track. there is a limit to length and number of intermediate signals searched however. it will only look ahead by 10 intermediate signals, or 10 miles, whichever one comes first (this could be extended i think but ive just never needed it). this search is not affected by turnouts unless they are switched the opposite way of the path. that isnt a problem though and siting it as one would be unfair to the signals, since you wouldnt clear them ahead anyhow without the path being set.
the only exception i can think of with mine is a situation where the train would reverse and move away then come back, this would give the opposing signal the wrong idea about where the train was heading, and let the other train past because according to that signal, the other train is now moving away... but you cant program the signals to know where the train occupying the track will be going in the future. i just think thats pretty unlikely anyway.


#2 - sounds like you used an always controlled signal here. the train will always come to a stop before it can cross these.

Click to expand...

Tried adding setup for industry switches for both 08 signals. This had no effect.
First, I set up opposing junctions as L for right diverging, R for left diverging, but that made absolutely no difference in signal operation, so I tried using the F direction for opposing junctions. All this did was to change the clear signals to limited clear.
I then re-aligned the junction CP10 for the southbound train to route onto another track, so there is no conflicting movement, but this did not produce a clear signal at CP22 south. Signal was still limited clear.
It appears that the junctions need to be set up L or R, not F.

Anyway, all of this made no difference in operation.
The junctions between the two home signals are causing the problem.
If there were no intermediate signals, there would be no issue.

FW

well the problem is probably using the 08 signals, they are made to guard turnouts that are diverging away from them not to guard what would be a single track mainline. your signal setup does not make sense using the 08s this way. using the term 'home' signal is also misleading, as that is an entirely different function. i dont know why it wouldnt work still, except that you say its 12 miles, and the signals are set for 10 miles. they only stop looking when they find an absolute or run past that length. if they cant see the next absolute signal they function like the default signals and will probably cause a 'cornfield meet' they totally ignore the 05 signals and the junctions unless the junctions change the path on which they search. i would also NOT set them up for the spur junctions. again the 08s are mainline speed routing signals anyway, and dont fit the application you are trying to assign them.

the correct way to signal your setup there would be to 04 absolute signals guarding the singe track entries at CP22 and CP10 and 05s anywhere in between. 08 or 06s would guard entry into the sidings. thats proto, and should work just fine. but like i said, the 08s share the properties of 04s and 06s in that they are absolute signals so its should still work.

i think i might have included this with the signals when i sent them.



using them this way is how ive been doing it since i scripted them in, and ive had a totally ai controlled line running for 6 hours with trains in both directions with no problems at all. i think you are making the signal problems more complex by trying to set the junctions for them and using them in the wrong place to begin with. :/

OK. I'll make the changes you suggested.
As far as setting up signals for the turnouts, I originally did not do that. I was just trying different things; groping in the dark

FW

Made the changes, and a bit more experimenting. No change.
I now have the following:

CP10 southbound, single to multi track: 08
CP10 northbound, multi to single track: 04
CP22 northbound, single to 2 track: 06
CP22 southbound, 2 track to single: 04
Intermediates (4 signals): 05

I also removed the southern most industry track, to test whether it was actually the 10 mile point from CP10.
That only caused the predictable result: Southbound train now trips CP10 northbound signal to stop when it passes the 2nd industry track (which is about 3/4mi north of the one I removed).

If the 10 mile limit were what is causing this, then the northbound train should not trip the southbound signal at CP22 until it reaches MP 13 (CP22 is at MP 23), but this happens at MP 11, just as the train enters single track.

I do not have the industry tracks signaled. I only added the signals in TB3 because it would not compile the layout unless I had them there, and I wanted to use the TD3 diagram, not the TB3 with its grid lines, for clarity.

If I remove the intermediate signals, then all will be well. But I need them there.

I would not be concerned about this at all if I were only going to run trainz myself, not using AI. But I am trying to create a route where AI will work.

The Safetran signals are the best available for Trainz, so I will consider my time well spent if I can come up with a solution where only these signals will work.

Well, I will be spending 4 much needed days away from my computer, in Killington Vermont, starting Sunday. Maybe when I'm away from the game, I will think of a solution<g>

FW

p.s. Not a ski trip. A bit too early here in the northeast. It's a camping trip.
I will be alone, so no one to keep my mind from wandering<g>

you are not thinking with the same logic as the signals...

let me explain exactly how these work. the absolutes will look ahead for a train moving toward the signal from the protected side. they see right through permissive signals as if they are not there when doing this, thus the ability to have two way signalling on a single track. they can only search for 10 miles ahead OR they will stop searching if they encounter 10 signals on the way.

now if they find a train moving towards them from the protected side in that way you get the stop message with 'conflicting train on single track' message.

those turnouts in between will not do anything to upset this. i have tested it over and over.

that being said, maybe i should increase that search distance, but thats what the testing was for.


as for default signals, those are a totally different story and they wont work at all with intermediates, so thats the answer there.

I will do some more testing on my test route, which does not include the ATLS grade crossing triggers. The test route is 15 miles long, and the interlockings are the same distance apart as they are in my real route.

I can then e-mail or upload to DLS my test route for you and anyone else to experiment with.

FW

I've returned from my camping trip to Vermont, and have been testing the simple route with Safetran 04, 05, and 08 signals.

I have shortened the route, so that the two interlockings are only 5 miles apart (within the 6 mile limit the signals will search for trains).
I now have one siding, located 2 miles from one, and 3 miles from the other interlocking.

I am using Safetran 04 signals to protect single track at each interlocking, and 08 signals to protect the double track at the interlockings.
I have two sets of intermediate signals, each is a Safetran 05.
My siding track is un-terminated, and unsignaled.

For this test, both interlockings are lined track 1 to single track, which would allow each train to proceed onto single track towards each other.

Here's what now happens when I run two trains, one from each interlocking onto single track towards each other.
Actually, I am running only one train, with the other standing on the track 1 at one interlocking.

As each train is sitting, each has a green signal.
Intermediate signals are:
Advance approach, followed by approach. The 08 interlocking signals are both red (R/R/R), which is all normal, since the trains are separated by 3 blocks.

When one train starts to move forward, the other train now has an advance approach signal.
When the moving train passes the 08 interlocking signal (signal is facing in wrong direction of course), the waiting train at the other interlocking gets an approach signal, and both of the intermediate signals are now red.

This condition continues, until the moving train passes the junction for the siding.
At that time, the waiting train's signal changes to red.

So, this test has shown me that the distance between interlockings is not what is causing the issue.

Norfolksouthern37; I do not understand why in your setup that your signals, which are the same as the ones I use, are functioning differently.
Are you also allowing both interlockings to be lined onto the main, with trains allowed to occupy the same single track?

If you are at all interested in this, I can send you my route.
Let me know. Send me a PM.

FW

i just think you dont use them correctly. go ahead and send the route over, i will take a look.

I agree with you. I have sent you my test route.

FW

G'day fwassner,

I am wondering whose CTC rules you are following here? I appreciate that there will be different variations all around the world but one of the basic tenets of these rules is to maintain the integrity of the main line. Toward which end, in the rules that we use here in the state of Victoria, Australia, there is NO provision for intermediate signals for the 'protection' of intermediate sidings (either for entry into or exit out of them). The only 'intermediate' signals aside from the absolute signals protecting the exit from the single line at either end are a permissive signal located a short distance in advance of the "Home Arrival" to give the driver an advance indication of the route for which that signal is set (via the main track or via the loop) and whether that route is through or not. Intermediate sidings are 'protected' by means of special locally sited "Intermediate Switch Locks", which are a device that acts as a release for the points to allow for movements into and out of the siding. In order for this to work, the Driver of the train must first stop within a designated distance from the points (in either direction), on a track circuit. The train Controller can then (and only then) provide the release, which sets in motion the actions that ultimately allows for the local movements (which may be as many as is needed into and out of the siding, including the possibility of the train being 'Locked Away' in the siding). Whilst the release is "off", the "Departure signals" at both ends of the single line section are 'locked' and cannot be placed to proceed under any circumstances. When the crew is finished with these local movements, the release is returned to the Train Controller and the movement of the train off the track circuit takes that release capabillity away from the Train controller. If the Train Crew inadvertently neglects to return the release when finished the work, the whole system stays set and special exceptions to the rules allow for the passage of trains until the 'problem' can be fixed.

I think your issue here will always be the 'intermediate signals'...

Jerker {}

well, i found the issue with the signals... the signal was finding the opposing train but it was the distance the train was reporting back to the signal. the train did not look the entire distance back to the head to tell it where it was, thus your signal staying clear and not detecting the train until it was near the spurs... it just so happened to be that the distance the train finally triggered the signal was near the spur, it didnt have anything to do with the spur tracks themselves. i have solved this problem by having the signal look as far ahead as it needs to in order to find the next absolute head so there is no more set search distance but is set by the distance to the next head plus a little bit of room to allow for the turnouts. it then remembers this distance and in return, the train reports back to the signal the same distance, so there is no more confusion of how close the train must be in order to trigger the opposing red signal. thing is when i was testing it it was on a much smaller run, but thats why i gave them to you to test. i will get a new batch to you in the morning.

Jerker: he is trying to follow fairly standard US practices, for which i have written my signals. there are usually intermediate signals along a single track line to allow traffic to move more quickly and safely over the line, just like shown in our diagrams. the intermediate signals are permissive, and are to keep trains traveling in the same direction spread apart enough to allow them to function freely.

norfolksouthern37 said:

well, i found the issue with the signals... the signal was finding the opposing train but it was the distance the train was reporting back to the signal. the train did not look the entire distance back to the head to tell it where it was, thus your signal staying clear and not detecting the train until it was near the spurs... it just so happened to be that the distance the train finally triggered the signal was near the spur, it didnt have anything to do with the spur tracks themselves. i have solved this problem by having the signal look as far ahead as it needs to in order to find the next absolute head so there is no more set search distance but is set by the distance to the next head plus a little bit of room to allow for the turnouts. it then remembers this distance and in return, the train reports back to the signal the same distance, so there is no more confusion of how close the train must be in order to trigger the opposing red signal. thing is when i was testing it it was on a much smaller run, but thats why i gave them to you to test. i will get a new batch to you in the morning.

Jerker: he is trying to follow fairly standard US practices, for which i have written my signals. there are usually intermediate signals along a single track line to allow traffic to move more quickly and safely over the line, just like shown in our diagrams. the intermediate signals are permissive, and are to keep trains traveling in the same direction spread apart enough to allow them to function freely.

Click to expand...

I have to disagree with your statement that it has nothing to do with the turnout.
When I removed the junction lever at the turnout (which effectively removes the turnout), the signals performed properly.
I also moved the turnout to different points along the route between interlocking signals, and each time, the train triggered the distant 04 signal to turn red exactly as it passed the turnout.

If you are testing on a scaled down route, this may be changing your results.

Whatever may be happening on my route, I appreciate your working with me on this.

FW

Jerker said:

G'day fwassner,

I am wondering whose CTC rules you are following here? I appreciate that there will be different variations all around the world but one of the basic tenets of these rules is to maintain the integrity of the main line. Toward which end, in the rules that we use here in the state of Victoria, Australia, there is NO provision for intermediate signals for the 'protection' of intermediate sidings (either for entry into or exit out of them). The only 'intermediate' signals aside from the absolute signals protecting the exit from the single line at either end are a permissive signal located a short distance in advance of the "Home Arrival" to give the driver an advance indication of the route for which that signal is set (via the main track or via the loop) and whether that route is through or not. Intermediate sidings are 'protected' by means of special locally sited "Intermediate Switch Locks", which are a device that acts as a release for the points to allow for movements into and out of the siding. In order for this to work, the Driver of the train must first stop within a designated distance from the points (in either direction), on a track circuit. The train Controller can then (and only then) provide the release, which sets in motion the actions that ultimately allows for the local movements (which may be as many as is needed into and out of the siding, including the possibility of the train being 'Locked Away' in the siding). Whilst the release is "off", the "Departure signals" at both ends of the single line section are 'locked' and cannot be placed to proceed under any circumstances. When the crew is finished with these local movements, the release is returned to the Train Controller and the movement of the train off the track circuit takes that release capabillity away from the Train controller. If the Train Crew inadvertently neglects to return the release when finished the work, the whole system stays set and special exceptions to the rules allow for the passage of trains until the 'problem' can be fixed.

I think your issue here will always be the 'intermediate signals'...

Jerker {}

Click to expand...

Hello Jerker;
My intermediate signals have nothing at all to do with the industrial sidings. As Norfolksouthern37 said, these signals are intended only to keep trains moving between interlockings.
That said however, if someone were to throw the industrial siding over, the signals on each side of that turnout would turn red, warning an approaching train that it is about to be side-tracked into the turnout.
Trains are permitted to pass these signals in order to make their moves into or out of the sidings, or if there is a malfunction of the signal for unknown reason, they are allowed to proceed, usually at 10mph, as visibility permits.

These intermediate signals do not show up on the dispatcher's board. He/she has no control over them. CSX referrs to these permissives as Automatic signals.

FW

fwassner said:

I have to disagree with your statement that it has nothing to do with the turnout.
When I removed the junction lever at the turnout (which effectively removes the turnout), the signals performed properly.
I also moved the turnout to different points along the route between interlocking signals, and each time, the train triggered the distant 04 signal to turn red exactly as it passed the turnout.

If you are testing on a scaled down route, this may be changing your results.

Whatever may be happening on my route, I appreciate your working with me on this.

FW

Click to expand...

just trust me on this Frank, its not the turnout. im using your route btw. i will email you the signals soon.

norfolksouthern37 said:

just trust me on this Frank, its not the turnout. im using your route btw. i will email you the signals soon.

Click to expand...

OK Justin. You wrote the script, so I shouldn't be arguing with you.

Thanks for your attention to this.

FW

G'day fwassner (& norfolksouthern37),

It would seem that some misunderstanding between the text and the images provided has occurred from my end and for this I apologise. We also have such 'intermediate signals' (calling them as NS does "Automatic Signals") but don't usually use them under our CTC rules. This type of signal is most commonly used (for the same purpose, I might add) in locations where there are adjacent bi-directional single lines (in which case the rules are slightly modified - and called A. T. C.) and there are long distances between the shared Crossing Loops. I shall, henceforth, maintain silence...

Jerker {}