A very simple maneuver but impossible to realize with IT

vincentrh

Well-known member
Consider the plan bellow :





We have an IT managing the pathes and a loco driven by the player.
The movements 1 and 2 go well. But the junction 3 close to S3 firmly stays right oriented (and the signal S3 remains red)
The only 'solution' I found was to slowly drive my loco as far as the junction 3, then stop and go back, and then drive the loco for the junction. At last, it is correctly oriented to the left and the signal becomes green !

Did you meet this sort of difficulties and did you find a solution ?

Thank you.

Vincent
 
Hi Vincent

When I was experimenting with ITs on the ECML at King's Cross I found that the consist had to clear the release signal by about 10 meters before the path was released. If the loco doesn't travel that far past signal S3 then this may be the problem.

Regards

Brian
 
Since Vincent states the signals are approx 30m beyond the junctions, the path should have released.

Vincent, I have come across the same problem. I was using pguy's ITSetPathandWaitActive rule. Basically, my driver just sat there waiting indefinitely. Like you, I found that if I moved the train with (say) a TDCdrivetotrackmark shunting command, and then re-issued the ITSetPathandWaitActive command, the path would be set. Since my problem occurred towards the end of a very long session, I put it down to possible timeout issues associated with my modest (Mac) hardware, but it could be a bug.

I don't believe the problem lies with pguy's commands, but may lie with the IT code itself. However, Pierre is probably in the best position to try and investigate what is going on, since he has a thorough understanding of the IT scripts.

John
 
Thank you all.

In the particular case I submit, I' don't use the Pierre' commands. Thus, the problem lies with the IT script itself I suppose too. But John you are right : I am going to write to Pierre to ask him his opinion on the question.

Vincent
 
Hi John

Just to clarify, the consist needed to clear the release signal by 10 meters, not the junction. This may have changed in the latest update to T:ANE as I haven't tried it recently.

Regards

Brian
 
You are right. And so did I while driving my loco. Unfortunately it didn't modify anything.

Vincent

hi Vincent, Brian, John and Martin.

There is currently a bug in IT implementation when a train requests (either using automatic path setup or ITSetPath or choosing a manual path in the select path box) a path starting at an entry signal which is the same signal as the previous path exit signal, if the locomotive has changed direction after passing the exit signal in the reverse direction (locomotive running backwards) , and the new path is requested while the previous path is still in the PATH_CANCELLING processing (path cancelling process last about 5 sec. real time).

The bug is in fact a very vicious bug, as the previous path will start its cancellation processing when you leave the exit signal, but will be reactivated by the bugged code if you stop your train while your previous path is still in the PATH_CANCELLING state ... The only way to avoid the problem is not to stop a train with locomotive in the backward direction before 6 sec. after leaving the exit signal (to give the time for the previous path to return to the PATH_NOT_SET state) ...

From the description of operation from Vincent, I suspect that movement 1 is with locomotive in the frontward direction, movement 2 in the backward direction and movement 3 in the frontward direction. If it is the case, this may explain the problem you encounter just after movement 2. If you reverse your locomotive at the start, you should encounter the problem after movement 1 instead of movement 2, which does not solve the problem but may confirm you encounter the "reverse too soon while in backward direction" bug ...

As I think you will say that waiting 6 sec. before stopping is not practicable ... the other way to avoid the problem is to use an enhanced interlocking tower ( either my private version or TRC 3 invisible interlocking tower) for which I can send you a temporary fix for this problem.
For Vincent, the fix needs to be done in the TRC3 interlocking tower library. I will send you soon (today) a new beta version 7 of this asset so that you can see if it solves your problem ...

I am speaking of temporary fix as I will report soon the problem to N3V support and it will certainly be fixed in in future hot fix 4 - 5 - ... ? but the delay for the fix may be in months ...

Regards.

Pierre.
 
Last edited:
I've juste tested the TRC3 invisible IT with the Pierre's code Library in its 7th version. And my problem is completly resolute !
I hope Pierre will soon be abble to upload this version. It would be of real interest for the community (except for those who don't use AI...:hehe: but TANE is a game and each one does what he wishes).

Vincent
 
Many thanks for the explanation, Pierre. In my case, not stopping the train for 6 seconds would have been, shall we say, problematic since the train in question was close to the buffers of a head shunt:D. Hope N3V sorts the code out soon(TM).
 
Hi.

For your information, the new v7 version for TRC 3 InterlockingTower library with the temporary fix to avoid the "path reactivation while in PATH_CANCELLING state" bug is now available on DLS.

Regards.
Pierre.
 
@ Pierre - Could you possibly put forward a suggestion to N3V about the use of signals to release the path?

In addition to being able to nominate a signal as the instrument for releasing the path, I suggest that a Track Circuit Insulator should also be considered as a path terminator object. This would allow a path to be released earlier if the insulator is nearer to the junction than is the signal, and then a signal need not be placed if there is no need for one.

TCB's would be very useful in their own right and I sincerely hope that the shortcomings presently experienced with them are addressed in future.

Best Regards - Trevor
 
A belated thank you, Pierre. Now, finally my automated Ffestiniog &WHR summer 2015 session works from start to finish:D.
John
 
Back
Top