Help: I please need to know my LEFT from RIGHT in terms of junction/signal logic.

JonMyrlennBailey

Well-known member
It helps to understand which side of a RR switch is LEFT or RIGHT from the point of view of Trainz.

When defining L for "left" or "R" right in the path configuration of a signal, I realize that getting LEFT or RIGHT
confused may get AI drivers confused also if the path is incorrectly defined in terms of LEFT or RIGHT
at a junction block.

Which way should a HUMAN Trainz editor in Surveyor face the switch to determine the LEFT or RIGHT
diverging part of the switch? How does the signal consider what is the LEFT or RIGHT?


All along, in terms of defining LEFT and RIGHT, I think I may have been wrongfully facing the switch in the direction where it branches out instead
of the direction where two tracks converge at common point.

signal%20logic%20left%20or%20right_zpsun8t4pqm.jpg
 
Last edited:
Facing the switch of course! I don't suppose you have any choice in which direction to go if you're travelling down a trailing switch do you?

Signals do not tell left or right by themselves. You have LD and RD versions for left diverging and right diverging. If you put an LD in front of your example the signal will think the straight-ahead is the diverging path.
 
So, if I want a train to travel from the red segment through the yellow segment in my example, I would specify R for RIGHT in the signal properties?

How should L and R be used when defining a path through switches?


I have been having a lot of problem with AI test drives through a complex block of switches lately.

The AI driver comes to a signal and gets brain cramps. AI starts fiddling with the switches, changing the settings on some of them, back and
forth repeatedly, but sometimes can't figure out how to negotiate them. Does AI need to be able to tell LEFT from RIGHT?

Does AI look at the switch settings or does AI strictly follow the colors of the signal lights in deciding what to do? Is it the signal's job to read the switch settings?

I am under the suspicion that I may be using L instead of R or vice-versa in error when defining a path through a block.
 
Last edited:
You do not need to name junctions, they all can be labled ` (the lowercase key of the KB symbol ~)

AI will throw all your turnouts named ` automatically, by "Drive Via" or "Navigate Via", a named trackmarker

Looking from the straight section (Points) of switch, toward the frog ... the R is in the R, the L is on the L

Depending on which direction you lay your track, a switch lever, a signal, or frog, will be reversed from the normal position (the urban legend says it makes no difference, but I always lay my track in the intended direction of travel)
 
Last edited:
Thanks, Cascade. Yo have my question answered down to a T. Looking from the switch lever toward the frog, in other words, the red segment toward the yellow or blue one, for that matter.


I am trying to prevent conflicts at junction blocks between two or more trains, coming from opposing directions, all wanting to occupy the same place at the same time.

I am also trying to prevent signals from being stuck on red forever and prevent AI from indecisively throwing switches in a block back and forth repeatedly like its mind is totally lost.

During my testing, I have seen AI throw switches back and forth while stuck stopped at a junction block even when I manually set all the switches to clear a path for the train by hand and there were no other trains present near the junction block. Sometimes an AI train will throw one or more switches, the signal light will change from red to green for a second or two and back to red again. This may repeat a couple of times. The light turns back to red again after the AI train throws a switch or two when it shouldn't. The train may even budge a foot then immediately stops again without ever making it past the signal. Then the exclamation point is displayed by the driver's picture (avatar in the command bar) meaning AI has hopelessly given up trying to negotiate this block of switches.

Why AI acts this weird sometimes, I don't know to save my soul. I have had some luck improving this by altering my track at junction blocks. Deleting junctions to make the junction block less complicated with fewer switches. Placing junctions at different locations and fiddling with signal placements as well. AI does not like to see a bunch of switches bunched up close together in front of the train as which often occurs whent he main line meets the yard or when on train line of multiple tracks merges with another line.
 
Last edited:
Opposite direction of the arrow, from the red track to the blue and yellow. Since the yellow track is the curved one you're looking at a "right diverge". Not necessarily always the case, if a spur comes off the curve on the mainline the through route is the curve and the straight would be considered the diverge since you're DIVERGING from the main line by going straight.

With Trainz it doesn't matter that much, if a switch is part of a double crossover for example the double head will show red over green for either straight or diverge regardless of which signal you use. That's why I use lighted red/green switchstands, if I want to know which way the switch ahead is thrown a double head signal is useless in some track configurations.

I don't know what the max number of switches ahead the AI train will attempt to set in order to find a path through a complex junction, so I tend to use dwarf signals and extra track markers to guide them through. For an example see Western Avenue Interlocking on the Chicago Metro route;

Chicago Metro 3,<kuid:522774:100015>

Main route session for that one;

chimet 3 40th 3020,<kuid:522774:100014>

Look at the schedule library in that session for how the AI traffic works. Trigger Multiple Signals is also needed at Western due to the crossovers and double slips, Trainz signals don't actually do anything for crossovers without the triggers.
 
Back
Top