The happy route builder,
Rick
Rick
That's interesting. The setup there is
almost exactly what I use for passing loops too, but with one important difference.
Assuming a US layout, I make X a 2-headed signal like the default USA-02.
I also include
A,B,C,D and they are all the 1-headed USA-04 (facing towards the trains as they exit either branch of the loops). I can't say if
A,B,C,D are necessary or not, but I add them because it looks like you're controlling the junction from every possible angle of approach.
The difference with my understanding of signalling is the signals at
Z. I was told (I read somewhere)
never to put any "hard" signal at
Z . By "hard", I mean a signal capable of showing a stop aspect. USA02 and 04 are both hard stop signals.
The reason is that if you are running trains in both directions, 2 trains travelling towards each other will eventually be permitted to enter opposite ends of the same single track between loops. If one of them has been stopped by the
Z signal, you will have a stalemate (or worse).
You can only put signals at
Z if they are of the
permissive type. These signals simply relay the status of the next signal further down the line and are informational only. I don't think they can stop a train. USA-05 is an example of a permissive.
If there are no signals, or only permissives at
Z, then the blocks (which are really defined by the junctions) are controlled only by the signals at
X (and possibly at
A,B,C,D). This setup of bi-directional passing loops has worked well for me, apart from the artificial hiccups like high-speed SPADs or too many trains for the AI to compute.
If you want to try hard signals at
Z, I'd be interested to know how it goes over an extended period of running time, just to see if that theory of stalemates is right.
~ D