Return of the Backdated North London Line

[JustRight]

Hi there Colin,

Mate, I have been mucking about with your scenario "(CG) Freight to Stratford".

Using the first copy of 'Version 2' of Xavier's route, lets call it version 2.0, I was able to successfully pass the troublesome signal. The signal correctly changed to green, and showed a left hand feather.

However, using the second copy of 'Version 2', lets call that one version 2.1, I cannot get the system to path the player train through that junction successfully. As you approach the signal is at Red, and as others have said, it stays that way. When looking at the 2D map you can see that the dispatcher has no path for the player train, the blue highlight on the active track continues only a short distance beyond the signal. But, as others have found, if you press on regardless I pass the "Imported Coal to Yorkshire" which continues on it's merry way.

If I revert to the version 2.0 instance of the route, then arriving at Camden Road I can see that the dispatcher has a path for the player train through and beyond the offending signal.

I have messed about with the scenario itself and also with the signals on the route, but I just cannot see what is different between version 2.0 and version 2.1
It's a really strange one

Cheers,
Trevor

[CeeGee]

xguerra is looking at the scenario after making a small change to the signalling.
I have been doing more testing and the "problem" appears to be any AI travelling along the Freight line to Camden Road.
I will wait for xguerra to get back to me and if he has no good news then I will release an alternative version in the next 48 hours by altering some AI timings and sending them to Finsbury Park instead of through Camden Road.

[xguerra]

Right, I'll put this here instead of a PM so Mike can read this as well.
I have done a little experiment and found something very strange. The good news is that it is not the signals nor the track that is the issue.

It's the Kuju class 37.

I conducted a test scenario where two freight trains pass at Camden Road. Service1 (player) starts at Camden Road junction, and Service2 starts at Caledonian Road on the westbound freight line.
Both are the Kuju 37 and identical container loads. They are both marked as express freight.

Test 1: let them both go.
Result: You come across the issue in Colin's scenario where both are in a stalemate around the junction, where Service1 is held at signal NL1213 and Service2 just before the junction.

Test 2: change service2 to low-speed freight
Result: Same issue with the signals

Test 3: return both to express freight and substitute Service2 loco for a Kuju class 47.
Result: Service2 proceeds through the signals at green and service1 is held at NL1213 until Service2 has cleared the junction and then NL1213 becomes green and the train proceeds.

Test 4: change service2 to low-speed freight
Result: Same as above. Service2 proceeds first and service1 is held until the junction is clear.

Test 5: return both to express freight and make both locos a Kuju class47
Result: service1 is allowed to proceed first and service2 is held at the junction.


So what I can suspect is that there is something in the class 37 scripting or somewhere that doesn't seem to work well in AI for scenarios. From this, I would also anticipate that if one of the two offending services in Colin's scenario were changed from a class 37 to a class 47 or another loco, it would run fine. But there is no physical explanation for why the class 37 exerts this behaviour.

I can say that what changed between what Mike is calling 2.0 and 2.1 is the track properties for the first crossover after Camden Road heading eastbound were changed to main-line bi-directional like the rest of the surrounding track, as this issue had prevented one of my scenarios from working, stopping the player's train from being pathed into the correct platform at Camden Road. I suspect then that the reason the scenario worked on 2.0 was because the coal train couldn't reach Camden Road and so was held anyway. Nobody apart from those on this thread who have played this exact scenario have come to me experiencing issues, and I know of no other scenarios published that use the Kuju 37 as the player train. Therefore, it can be rest assured that the route is perfectly functional, and the issues lies with the use of the Kuju 37 in that scenario.

[CeeGee]

I'm totally gobsmacked!

I always said it was not the route at fault - although I was beginning to waiver in my opinion.
I was not expecting to be told that I had used the wrong loco. I quite like the old Kuju 37. Oh well, off to the scrapyard with it.
I will start changing some stock - probably to the RSC Class37Pack01, and then see what happens

Thanks for that brilliant detective work.

[gptech]

Do you get the same results if you change the wagons Xavier?
With container loads being *interactive* there's another source of scripting/geeky mumbo-jumbo in the equation.

[xguerra]

I'm totally gobsmacked!

I always said it was not the route at fault - although I was beginning to waiver in my opinion.
I was not expecting to be told that I had used the wrong loco. I quite like the old Kuju 37. Oh well, off to the scrapyard with it.
I will start changing some stock - probably to the RSC Class37Pack01, and then see what happens

Thanks for that brilliant detective work.

Or alternatively, do you have the BR Blue Diesel Loco Pack?
That has a blue split headcode 37 that might work instead.

[CeeGee]

Or alternatively, do you have the BR Blue Diesel Loco Pack?
That has a blue split headcode 37 that might work instead.

No - not yet. The RSC Class37Pack01 has done the trick. I have run the scenario once and everything worked as it should. I just need to run it a couple more times just to be sure and make an odd time adjustment to some AI before I can release it again. I'll also have to amend the "read me" to make sure nobody tries to use the Kuju model if they don't have the RSC model.

I'm good at sorting out problems in TS201*, but I don't think I would ever have thought that the Kuju 37 was to blame. Amazing detective work, Holmes.

[phil17]

Thanks for your findings xguerra. I ran the cloned scenario with driven train and coal train changed to class 45s and I believe the scenario ran as CeeGee intended, no problems at Camden or anywhere else.

I then went back to the original with the 37s. I did notice that driving with the HUD, the ETA was getting further away (at the very start ETA was 18.42). By Gospel Oak it was 00.33 the following morning. I then quit. It did the same yesterday too, not sure if it's relevant or not.

[gptech]

If the default Kuju class 37 is solely to blame, then

Using the first copy of 'Version 2' of Xavier's route, lets call it version 2.0, I was able to successfully pass the troublesome signal. The signal correctly changed to green, and showed a left hand feather.

However, using the second copy of 'Version 2', lets call that one version 2.1, I cannot get the system to path the player train through that junction successfully. As you approach the signal is at Red, and as others have said, it stays that way. When looking at the 2D map you can see that the dispatcher has no path for the player train

surely couldn't have happened, nor could Colin have wrote the scenario. The loco hasn't changed so any *fault* in it would also affect the running of the scenario in the earlier edition of V2 of the route.

Does anybody else have the earlier V2 in which to try Xavier's test scenario?

[chrisreb]

Not sure if anyone else had this but in a scenario where I left Broad St in an EMU, another EMU was entering the station on an adjoining line there was insufficient clearance between the tracks and I collided. Had to try again and leave a minute late.

[xguerra]

If the default Kuju class 37 is solely to blame, then

Using the first copy of 'Version 2' of Xavier's route, lets call it version 2.0, I was able to successfully pass the troublesome signal. The signal correctly changed to green, and showed a left hand feather.

However, using the second copy of 'Version 2', lets call that one version 2.1, I cannot get the system to path the player train through that junction successfully. As you approach the signal is at Red, and as others have said, it stays that way. When looking at the 2D map you can see that the dispatcher has no path for the player train

surely couldn't have happened, nor could Colin have wrote the scenario. The loco hasn't changed so any *fault* in it would also affect the running of the scenario in the earlier edition of V2 of the route.

Does anybody else have the earlier V2 in which to try Xavier's test scenario?

In the previous versions where the scenario worked, the crossover that the coal train uses to cross from the eastbound to the westbound line just before Camden Road station was marked as "passenger" track type rather than "main line". This meant that with the service type as "express freight", the train would have avoided that crossover if an alternative path was available. This path is wrong line through Camden Road platform 2 and to Gospel Oak where the train performs a reversal over the crossover there and back down to Camden Road platform 1. But, crucially, the player's train occupies that section of track up to the end of the junction at Maiden Lane. Therefore, AI would dictate that the coal train wait at the last signal before the junction and allow the player's train to pass before proceeding. Or alternatively, the coal train would have just sat at the signal until the scenario was over.

Now I changed the track properties to alleviate the same issue, which I had found in my 3rd scenario "One For The Scrap Heap" where the player train (the BR Blue Loco Pack class 37 and scrap wagons) is wrongly pathed into Camden Road platform 2 instead of platform 1 because the "passenger" track property does not allow certain train service type to traverse that bit of track. Upon changing the property to main line the train was able to proceed on the correct path.

Granted the dots don't join perfectly together but nothing changed between the two versions apart from that one crossover and what I have said above is exactly what having the wrong track properties does. It fixed my scenario and avoided issues with others. I can't explain why changing the loco from the Kuju 37 has fixed Colin's scenario either but if it's worked for two people so far, that's promising.

[gptech]

In the previous versions where the scenario worked, the crossover that the coal train uses to cross from the eastbound to the westbound line just before Camden Road station was marked as "passenger" track type rather than "main line". This meant that with the service type as "express freight", the train would have avoided that crossover if an alternative path was available.

Would that be this one under the second wagon of the player train, and should it use the one circled ?:



Surely signal NL1205 would/should hold the player train if that's the case (which it doesn't) and as NL1213 is after the junction it can't.

I've just tried a version of your test piece, and swapping the 37 for a 47 made no difference.

nothing changed between the two versions apart from that one crossover

Exactly my point, the class 37 hasn't changed between when it worked and when it didn't, so the obvious place to start looking is where the change has been.

Just to confuse things even more though, I've junked the original test and made a new one but with the AI service starting 500 yards further East---this time NL1205 did hold the player train until the AI had crossed on to the *right* track.

[xguerra]

In that case it's just pure fluke.

Fundamentally, the signalling has not changed and is as the original DTG route albeit the newer LEDs changed for 3 or 4 aspect signals that perform the same function.
If changing the track properties has made no difference overall, fair enough, but it's made things easier in other cases.

I doubt I'll ever get to the bottom of why this phenomenon occurs but if changing the locomotive fixes the issue for some, however improbable it sounds, that satisfies me.

[gptech]

I doubt I'll ever get to the bottom of why this phenomenon occurs but if changing the locomotive fixes the issue for some, however improbable it sounds, that satisfies me.

Maybe, what if the player really wants to drive a class whatever along the route? Changing a loco is pretty easy, but there'll always be somebody who doesn't know how to, and as doubts have been cast on it's effectiveness (did you exit the game after making the changes or just drive straight from the scenario editor?) it's at best a 'here's something to try and see if it works' ploy rather than a solution.

Fundamentally, the signalling has not changed and is as the original DTG route albeit the newer LEDs changed for 3 or 4 aspect signals that perform the same function

But do they achieve that function in the same way?...there could be a very fundamental change 'under the hood' that we're all unaware of.

Isn't the easiest solution in cases like this to use scenario specific markers to *force* routing?

[JustRight]

Great detective work guys. I for one would never have thought of changing the Loco type. I'll give that a go myself and see how I get on.
I'm sorry that I missed the conversation, but being on the other side of the world doesn't help.

Thanks to both CeeGee for the scenario's and Xguerra, for a fabulously rendered version of the NLL.

Cheers,
Trevor