Upgrading a V5 layout to V7

[simonjackson1964]

I think I may have found a bug?

This is an old V5 layout that I thought might work better at V7. It does, and so far I've managed to solve every conundrum it as thrown at me, except this one... The layout is paused for clarity, but I can record video of this happening if you wish.

As ypu can see there is a train on the junction. it has passed the signal to the next route and so the route it is leaving is now no longer selected, but is still blocked.

Trying to select that route gives a "blocked" message, as it should.

The yellow route is the anti-clockwise route crossing the track with the train on, highlighted blue with a pencil in the route sidebar. Logically if I select that route it should go "Blocked", right?

Except it doesn't. The route selects (and due to a deliberate event, also selects the next route down, as it's clear). Clearly this should not happen! As I said, I can easily provide video of this "signalling failure" that has caused hundreds of pretend little model people to be killed, when the trains hit each other and (fortunately) pass through.

It looks to me like the system is not recognising the coaches on the back of the loco. Except it is or the route it's on would be selectable...?

Help, please?

Thanks

Simon

[Goetz]

vor 6 Stunden schrieb simonjackson1964:

Except it doesn't.

Because the two routes don't have a single common element.
The studio cannot distinguish between a crossing bridge or a track crossing at the same level.

You may resolve this issue by selecting a single piece crossing:

This is recognised as being part of one active route, blocking another (crossing) route as a consequence.

Kind regards
Goetz

 

[simonjackson1964]

Hi Goetz

But they do. I even deleted the original crossing from V5 and put it back in at V7 to see if that might make a difference.

It could be something wrong with that particular track object, but I've used the same object in all the other places where two tracks cross each other and I've not seen it do the same thing, but that could be because I'm not watching for it.

As I said, the one track blocks while the the other is "locked", but as soon as it "unlocks" - when the train reaches the contact for the next block - the crossing route clears, while the train is still on the crossover.

 

[Goetz]

vor 3 Minuten schrieb simonjackson1964:

As I said, the one track blocks while the the other is "locked", but as soon as it "unlocks" - when the train reaches the contact for the next block - the crossing route clears

You did say that and I missed it. My apologies.

Of course a crossing is unlocked when the train using a route passes this section.
The route of this train stays locked to continue protecting the stretch in front of the train.
But what's behind the train, is no longer considered blocked.

That's the desired behaviour.

[simonjackson1964]

I just checked and it does it at other places where that crossing is used. I think it's a problem with the track itself.

Unfortunately I was too slow on the pause button, but there's no way the hopper train could have got that far without hitting the last few tanks.

2 minutes ago, Goetz said:

That's the desired behaviour.

Not on any railway I've ever heard about! The whole point of signals is to prevent trains hitting each other, not cause them to...!

[Goetz]

no, there's something else wrong

But I can't see, what.

It's nothing to do with the model. It's in your system

[Goetz]

vor 4 Minuten schrieb simonjackson1964:

too slow on the pause button

try the Spacebar next time ;-)

[simonjackson1964]

Restarted from the save point and managed to capture the whole event this time...

Hopper train is held at the light, and the point is showing to go the other way

Tank train comes around the corner. (By the way in case you're wondering, this part of the layout is supposed to be hidden from view, but while upgrading I wanted to see it.)

With the tanks still on the crossing, the signalman has a brainfart and changes the point and the signal. The driver of the hopper train is either blind or insane, because he starts moving forward!

 

KABOOM!

[Goetz]

It's not that I don't believe the sequence of events, Simon.
But your description doesn't help me to identify the cause.

Regardless, how colourful (literally as well as figurative)

Sorry.

Bearbeitet 4. September 2022 von Goetz

[simonjackson1964]

Sorry, I was having a bit of fun with it.

I have tried replacing the Arnold Spur N 1854 with a 1853 to the same effect. Now trying it with "|1435| K138/10 Kreuzung 10°" from the model oriented set, which also meant replacing the points so it fits together... sort-of. I'm not sure what I'm hoping to prove, because if I still get a collision then it is something else weird, but if I don't then I am going to have tp re-lay all the track which defeats the object of upgrading the layout..... Anyway I will let you know what the outcome is.

 

[Goetz]

Gimme an ID so I can check what's wrong, Simon.
Please!

Anything else has me guessing 'til the end of days.

Why's that so difficult for everyone to fathom?

[simonjackson1964]

Not the track...

The ground signal is still linked to the Arnold point so hasn't changed.

I'm going to see if I can reproduce this on a test bed, and if I can, I'll upload it for you.

[Goetz]

vor 16 Minuten schrieb simonjackson1964:

signal is still linked to the Arnold point

Why???
Why would you link a signal to a track point? That would be very wrong!

here's a simple test layout for you: CCDBCF02-97A9-41E9-A442-711DEC2EDA94
I used the Arnold crossing 1854 to build it.

(It's not that I'm not trying to help. But you make it very difficult.)

[simonjackson1964]

The ground signal, the one that shows which way the point is ... pointing. In that last picture, it's still showing the curve where the point is set straight...

Okay...

3B7194FC-3D4A-4B2F-910C-0508F4FB9078

My simple example.

Select route 4, which will select/lock, then select route 6. Route 6 will show blocked. Start train 2 moving and it will stop at the signal. Start train 1 moving and it will go through the junction, and stop at the end signal.

When it does it releases route 4. Route 6 then selects/locks and allows train 2 to pass through the cars still sitting on the junction...

I'm not trying to be difficult, I just want to make sure I haven't done something stupid first.

 

[Goetz]

vor 9 Minuten schrieb simonjackson1964:

The ground signal,

Silly me. So sorry!

[simonjackson1964]

No worries. It's cosmetic anyway.

I tried your example, and the trains are shorter than the gap between the crossing and the end of the block. I shortened one of the blocks and the trains both hit the crossing at the same time almost ignoring the signals!

Have you tried my example yet? it shows clearly what the issue is.

[Goetz]

vor 17 Minuten schrieb simonjackson1964:

the trains are shorter than the gap between the crossing and the end of the block.

That's the ticket.

In your example, the signals after the crossings are too close to be an appropriate end point for a route.

You want to extend your routes past that signal to the next destination.
(And maybe question if that signal serves any purpose at all)

Please understand that the system only detects obstacles on the same route. 
Detecting obstacles on crossing routes would require a very different (and much more resource hungry) approach.

That's why it is up to you to lock a crossing until it is safe to unlock.

Bearbeitet 4. September 2022 von Goetz
for additional remarks

[simonjackson1964]

I did this originally, and had a different issue:

The single route exiting the junction leads to an 8 track fiddle-yard. (The layout pre-dates virtual depots). When two trains arrived at the junction heading the same way ut from different directions, the one that went first correctly selected a route, the one follow did not. I finally fixed this, but one of the things I tried, before realising what was actually causing it, was to split the routes after the junction.

Re-combining them would totally destroy the system I put in that works...

But really, that isn't the point. It shouldn't matter how long the route is nor how long the train is. As I "colourfully" pointed out already, if this were the prototype, there would be hundreds dead and a massive lawsuit pending!

I think the fault does in fact lie within the system, because the light on route 6 should never go green while there is even one wagon still on that crossing from route 4 and vice versa.

I am going to put a "manual" check in place to prevent this, but I think the team should consider a fix for this.

[Goetz]

vor 47 Minuten schrieb simonjackson1964:

I'm not trying to be difficult

I know, Simon.

But you show what you consider is well and proper.
You try to prove you didn't do anything wrong.

But you showing what's right is of no use, when one is looking for the cause of something going wrong

You don't know where to look. Because you don't make deliberate mistakes.
How would you know where the fault lies and not know how to fix it?

That's why I myself need to have a look. 
Prefereably at the entire situation.

Bearbeitet 4. September 2022 von Goetz

[Goetz]

Hello Simon,

vor 5 Minuten schrieb simonjackson1964:

the light on route 6 should never go green while there is even one wagon still on that crossing

This amendment to my previous posting probably came after you wrote your response.
I repeat it (only to ensure it's not overlooked):

vor 26 Minuten schrieb Goetz:

Please understand that the system only detects obstacles on the same route. 
Detecting obstacles on crossing routes would require a very different (and much more resource hungry) approach.

That's why it is up to you to lock a crossing until it is safe to unlock.

 

Greets
Goetz

[simonjackson1964]

Got it, and fixed with a couple of events that work for all crossings in both directions. Of course the other option is to replace all crossings with double slips, because being points, they have the interlocking in their crossing routes!

Bearbeitet 4. September 2022 von simonjackson1964
Typo

[simonjackson1964]

In fact, I'm going to replace all the crossings with single slips, edited to remove the slip route. So much simpler!

[simonjackson1964]

So... that worked and was a lot easier that what had done. Which begs the question:

Rather than only using slips, and removing the slip routes, could we not have switch positions 0 and 1 on the crossings? Is this something that would be easy to implement and not require any change to the way the route system works? I mean yes it requires someone to go through all the crossings in the catalogue and add the switches to them, but at the end of the day it could easily justify the time spent in the simple fact that you no longer have to protect crossings from two trains being on them together!

[Goetz]

vor 11 Minuten schrieb simonjackson1964:

it requires someone to go through all the crossings

... to solve a very rare problem, which is easily solved as is by the user who requires it.
And you can use the track editor to change the properties yourself. New properties are stored with the layout.

[simonjackson1964]

Fair enough. It was just a thought that it would be a permanent solution, rather than having to re-solve it whenever it crops up...

What I might do is see if I can save the edited crossings in "My Models", so I've got them when I want them.