Cantankerous Behaviour of AI Drivers.

[Ian_Coleman]

I am building a layout which includes an 8-track terminus.

I have installed points (switches for our American friends) just prior to the terminus which makes it possible for any train to enter or leave the terminus on any one of the eight tracks.

When testing, with AI controlled trains, everything went well, except for one train that leaves from track 8 kept on insisting to take track 4 instead of track 3, as I wanted.

The problem was that track 4 was a track which had a direction marker pointing towards it, beyond the junction.

I tried everything, including re-laying the track in case there was a fault that I could not see, but all to no avail. I literally spent hours struggling with this!

I finally solved the problem by putting a second junction just beyond the first one.

When I tested this, expecting the train to use the new junction, to my amazement, the train took the first (original junction), and not the new one.

I have seen some very peculiar behaviour from the AI drivers, but nothing quite like that before.

Although the problem is now solved, can anyone explain why it now decided that the original junction was ok?

Ian

 

[Veristek]

I am building a layout which includes an 8-track terminus.

I have installed points (switches for our American friends) just prior to the terminus which makes it possible for any train to enter or leave the terminus on any one of the eight tracks.

When testing, with AI controlled trains, everything went well, except for one train that leaves from track 8 kept on insisting to take track 4 instead of track 3, as I wanted.

The problem was that track 4 was a track which had a direction marker pointing towards it, beyond the junction.

I tried everything, including re-laying the track in case there was a fault that I could not see, but all to no avail. I literally spent hours struggling with this!

I finally solved the problem by putting a second junction just beyond the first one.

When I tested this, expecting the train to use the new junction, to my amazement, the train took the first (original junction), and not the new one.

I have seen some very peculiar behaviour from the AI drivers, but nothing quite like that before.

Although the problem is now solved, can anyone explain why it now decided that the original junction was ok?

Ian

Did you do signalling or something for the junctions?

 

[Ian_Coleman]

I tried with, and without signals, but only beyond the junctions.

Ian

 

[Veristek]

I tried with, and without signals, but only beyond the junctions.

Ian

Maybe you could try setting up signals *before* the junctions and see if they work. Lemme know if that works for ya.

 

[Ian_Coleman]

As I said, the addition of a second junction just beyond the original one has resulted in everything running correctly.

It was just curiosity that made me ask why trains should start using the original junction just because I had added a second one a short distance away from the original.

Ian

 

[Veristek]

As I said, the addition of a second junction just beyond the original one has resulted in everything running correctly.

It was just curiosity that made me ask why trains should start using the original junction just because I had added a second one a short distance away from the original.

Ian

Tell me a bit more about that second junction. Did the diverging track result in a dead-end track or did the track itself diverge back into the mainline?

 

[Ian_Coleman]

The second junction was exactly the same as the first junction, but a short distance further on (going away from the terminus).
Just in case you are wondering, all lines were laid with due regard as far as possiblefor the expected direction of travel. Obviously, the junctions would, in some cases carry traffic in borth directions.

Ian