Problem with SCS2006 rule taking wrong path

[anand99]

Hi,

I am just starting to play around with the SCS2006 rule. I have setup a basic piece of track as shown in the screenshot. I setup a track mark "final destination" beyond the loop as my destination and added the following commands

Train Control Type SCS
Path Set
TPR Stop At Trackmark

Basically the Path Set is to take junctions Anand1 and Anand2. The Stop at Trackmark is to stop at the "final destination" trackmark. For some reason the problem as you can see from the screenshot is that SCS is making the train go down the loop rather than continuing on the main line. There is really no reason to take the loop. The documentation for SCS says that the path set will usually take the route with minimum set of junctions. In this case both the loop and the main line have the same number of junctions.

Can anyone tell me what I am doing wrong and how I can make SCS take the line straight ahead to reach the trackmark? Link to picture below.

 

[Euphod]

Trainz will try to travel the shortest route as well. Your signaling confuses me though. In Trainz you need only place a signal on ONE line when two are about to converge into one. You can signal both lines before a convergence into one line, but it's only crucial to stop one train to prevent a collision. Using a prototypical signal set up will only confuse the AI.

 

[colourlight]

Being a professional signaller I have to disagree with Euphod. While what he says is true, in the real world you would have a signal on both lines to protect the junction. The whole idea of Trainz, to me, is to make it look as realistic as possible. To work around your problem I would place a track marker on the line you wish the train to go on and use the "drive via" command.

 

[Euphod]

You may place all the signals in a prototypical fashion that you like, but that tends to confuse the AI, cause problems with commands and complicate session building. Dummy signals would solve this problem, for those that want to "make it look as realistic as possible".

 

[anand99]

I don't think Drive Via is an option when using the SCS2006 rule. The SCS2006 pretty much has its own library of commands, and also picks its own path to the destination. The only way you can influence the path is by taking permits for specific juntions. However, in this case I the first junction I take a permit for is itself causing the path to go the wrong way.

 

[Kennilworth]

Hi anand99

I would use 2 Path Set commands in this situation

Path Set Anand1 Straight.. Anand1 Left.. Sig Name
Path Set Anand2 Right... Anand2 straight..Sig Name
TPR Stop at Trackmark

Incidentally if you set the third parameter of the TPR Stop at Trackmark command to 0.015 and the fourth parameter to 8 it should stop on the trackmark without the "inching up" that you often get with the default figures.

Regards

Brian

 

[anand99]

I didn't realize that the entry and exit junctions for the Path Set could be the same. Wow, I guess this does give granular control. I changed things around to how you described and things work great.

Yes, I did notice that issue and found your response to the same on the TPR forum. I tried it messing with it and got things to work at 0.03 and 5 for parameter 3 and 4 for my loco. It looks like for different locos these two parameters have to be played with a bit to find the right balance?

 

[Kennilworth]

Hi again

Yes the parameters can be adjusted for individual locos. The figures that I use are for UK steam and early diesel locos and I find that they generally stop the train on the trackmark without any problems. I ran trials with each class of loco that I use and adjusted their max-accel and max-decel figures in their engine spec files to give a slightly different performance for each class. This shows up in the timetabled sessions that I have programmed as a lower class loco on a train will take a little longer to run the same route with the same tonnage.

I use the Add Random Consist command to add all of my consists so the session always looks different each time it runs. My main interest is the early 1960s on British Railways and combining this command with a randomly generated variable it is possible to simulate such things as engine failure with for example a 1 in 5 chance of a steam loco appearing on a train scheduled to be diesel hauled.

The things that can be done with SCS2006 are only limited by your imagination. It sometimes takes a while but I can usually find a way to do what I want to do, even if only by trial and error.

Regards

Brian

 

[anand99]

Thanks again Brian. I am just only beginning to get a handle on SCS2006. Still have ways to go. But, I am noticing that if there is a junction and some track after the junction with just signals and a train1 is on that track Path Set still obtains a path

-----train2----S1-----J1--------------S2----------train1------trackmark1----railend

In the configuration above, train1 has just reached and is sitting at trackmark1 on a straight piece of track. It is aboutto reverse and head back in the direction of S2. Once train2 arrives at the location shown above I request a path set for J1. S1 shows red, but train2 just ploughs through the red signal. There is a diverging line in the direction of train2 from J1 that train1 is supposed to take which train2 is supposed to wait for at S1.

What I am not sure is why is train2 just not waiting at S1 since the track aghead is not clear? I have not set follow on on Path Set for train1.

 

[Kennilworth]

Hi

Do you have the follow on set to yes for train2 ?

Has train1 set its path back through J1 before train2 tries to set its path ?

Is there sufficient distance before S1 for train2 to stop if necessary when it sets its path through J1 ?

Does S2 face J1 or the rail end ?

There is no reason for train2 to pass a red signal as it should obey the signal irrespective of whether it has the permit for the track ahead.

There are other possible solutions to this including a token system using a variable for the section of track between J1 and the rail end but as we have the family coming today I won't have a chance to look at it until later this evening.

Regards

Brian

 

[anand99]

answers below

Hi

Do you have the follow on set to yes for train2 ?

No. Follow on is set to No for both trains.

Has train1 set its path back through J1 before train2 tries to set its path ?

No looking at the debug window it looks like Train2 set its path through J1 before Train1

Is there sufficient distance before S1 for train2 to stop if necessary when it sets its path through J1 ?

There is but once it crosses J1 it keeps going and couples to Train1, Train1 also starts heading back and they are going toward each other.

Does S2 face J1 or the rail end ?

It faces J1

There is no reason for train2 to pass a red signal as it should obey the signal irrespective of whether it has the permit for the track ahead.

When I set the paths, I assumed that since there is a train occupying the track ahead of J1, Train2 will not be able to get the permit for J1, even though it requests it before Train1, and will sit at the S1 signal. Train1 will get the path through J1 and the loop and proceed, then Train 2 will have a clear track ahead. But that's not how its working out.

There are other possible solutions to this including a token system using a variable for the section of track between J1 and the rail end but as we have the family coming today I won't have a chance to look at it until later this evening.

Regards

Brian

 

[Kennilworth]

Hi again

From your replies my first suggestion would be to either remove signal S2 or move it so that it is between trackmark1 and the rail end. Your diagram suggests that train1 has completely passed signal S2 in which case S2 would be the signal protecting the rear of the train and this may be the reason that train2 passes S1.

Does train2 run randomly? If not the solution could be as simple as delaying it by a few seconds to allow train1 to grab the junction first. This would be the simple answer but if it runs randomly then we may need to look at the token system that I mentioned earlier.

Regards

Brian

 

[anand99]

Brian,

Thanks very much for your help with this. I will look into the signals.