Signaller error Kings Cross Sunday 8th June 2025

[choochoochoo]

So, given that the train was approaching on the slow lines (according to the OP), if the train was signalled into p9, 10 or 11 (eight car platforms) and the driver accepted that...?

platform 11 has not been a thing at kings cross since it was remodelled a few years back.

 

[357]

So, given that the train was approaching on the slow lines (according to the OP), if the train was signalled into p9, 10 or 11 (eight car platforms) and the driver accepted that...?

The driver did not accept a route into 9 or 10. 11 does not exist.

The driver stopped at the last signal before the station, that could have also sent the train into 7 or 8.

It is unfortunately impossible to see platform occupation from the other end of two sets of tunnels when being routed onto a line that could legitimately put a train into multiple platforms where it would be fully accommodated.

Please take a look here and then read the first post alongside the track diagram to get a better idea of where the train was and what's happened: https://www.opentraintimes.com/maps/signalling/ecm1#T_KNGX

 

[choochoochoo]

Am I right in thinking the train occupying platform 8 was not leaving anytime soon ?

Would it not have been possible to signal the train on platform 8 up to 1017 signal and then let the train that had to go wrong road go into platform 8 ? Or was the train that was in platform 8 too long that it's back end wouldn't have cleared the points if brought to a stand at 1017 signal ?

Similarly was the train in platform 7 not leaving anytime soon either ? Could they not vacate the platform via D route and signal 1021, thus leaving a platform 7 free for the inbound train ?

 

[Horizon22]

Am I right in thinking the train occupying platform 8 was not leaving anytime soon ?

Correct (nor was 7). I listed the timings here earlier in the thread

 

[357]

Am I right in thinking the train occupying platform 8 was not leaving anytime soon ?

Would it not have been possible to signal the train on platform 8 up to 1017 signal and then let the train that had to go wrong road go into platform 8 ? Or was the train that was in platform 8 too long that it's back end wouldn't have cleared the points if brought to a stand at 1017 signal ?

Similarly was the train in platform 7 not leaving anytime soon either ? Could they not vacate the platform via D route and signal 1021, thus leaving a platform 7 free for the inbound train ?

It's highly likely there was no driver available to shunt the trains.

 

[Steve Harris]

It is rather amusing reading all the different hypothesise of the above situation when in fact it's clearly a limitation of ARS (ie it can not look ahead to see if the booked platform is/will become available).

 

[Spartacus]

Really it's because the HT was longer than booked so had to be replatformed, but really 1P83 shouldn't have been allowed to get so far that it needed to be reversed to get to a platform so long after the Hull arrived.

 

[357]

Really it's because the HT was longer than booked so had to be replatformed, but really 1P83 shouldn't have been allowed to get so far that it needed to be reversed to get to a platform so long after the Hull arrived.

I'm not sure I'd personally blame Hull Trains for it - there are industry systems that advise staff of train lengths.

However, I've always found in my career that a courtesy call to the signaller goes a long way (both on the day and to improve the working relationship!)

My calls normally start after introductions, with something along the lines of "I know you guys are normally the last to be told anything, and I don't know if it affects your plans... Just checking you know I'm longer than normal today"

 

[Spartacus]

I'm not sure I'd personally blame Hull Trains for it - there are industry systems that advise staff of train lengths.

That assumes they were updated at the time...

 

[357]

That assumes they were updated at the time...

I was going to say how unlikely that is... But now I've decided to refrain from commenting

Drivers reading this - don't be afraid to call the signaller if you're longer than normal

 

[Steve Harris]

Really it's because the HT was longer than booked so had to be replatformed, but really 1P83 shouldn't have been allowed to get so far that it needed to be reversed to get to a platform so long after the Hull arrived.

Indeed, but ARS wasn't updated hence why I said "ie it can not look ahead to see if the booked platform is/will become available.

Computer systems are only as good as the data inputed into them, thereby they have limitations !

It would be nice to know what was inputed into the delay attribution system. Do they have a code for ARS is rubbish??

 

[norbitonflyer]

Even assuming the driver remembered how long his train was (and some modern practices such as the "ceremony of the doors" suggest the powers that be think they might not always do so), would he be able to stop if the first indication he gets of the platform he is routed into is at the last (yellow) signal before the buffers.

Assuming he does pass that signal and drives into a terminal platform that is too short, so that his tail is blocking the exit to one of more other platforms (and, as a potential outgoing train, would be beyond the platform starting signal), what would happen next? Could the driver change ends and drive back out? Presumably the section would be clear of other trains as it is partially occupied by the errant one.

 

[bleeder4]

I was on a steam charter last year in which we were signalled into a platform at Bristol Temple Meads for which the steam loco was out of gauge. Luckily, the footplate crew realised and stopped the train. Permission was then given for the steam loco to propel the train backwards back across the points so the signaller could re-route us into a more suitable platform.

I've also been on a West Midlands Trains service between Worcester and Birmingham where we we wrongly routed at Droitwich onto the Stoke Works line instead of the Snow Hill line. In that scenario we we just continued on that route to Birmingham New St and passengers for Snow Hill line stations were instructed to get off at New St and walk across to Snow Hill for trains.

 

[Spartacus]

Indeed, but ARS wasn't updated hence why I said "ie it can not look ahead to see if the booked platform is/will become available.

Computer systems are only as good as the data inputed into them, thereby they have limitations !

It would be nice to know what was inputed into the delay attribution system. Do they have a code for ARS is rubbish??

Yeas, but this wasn't an ARS issue as such. More that one issue was compounded by ARS being left to do it's own thing when it shouldn't have.

== Doublepost prevention - post automatically merged: 12 Jun 2025 ==

Even assuming the driver remembered how long his train was (and some modern practices such as the "ceremony of the doors" suggest the powers that be think they might not always do so), would he be able to stop if the first indication he gets of the platform he is routed into is at the last (yellow) signal before the buffers.

Assuming he does pass that signal and drives into a terminal platform that is too short, so that his tail is blocking the exit to one of more other platforms (and, as a potential outgoing train, would be beyond the platform starting signal), what would happen next? Could the driver change ends and drive back out? Presumably the section would be clear of other trains as it is partially occupied by the errant one.

Of course he or she can stop before they go into the platform, that's pretty much what happened here. What happens next if they take the signal into a platform that's too short all depends on the circumstances, depending on track circuits, door positions, conflicting services etc

 

[43066]

Even assuming the driver remembered how long his train was (and some modern practices such as the "ceremony of the doors" suggest the powers that be think they might not always do so), would he be able to stop if the first indication he gets of the platform he is routed into is at the last (yellow) signal before the buffers.

Correct. Drivers are human and mistakes can be made. The “ceremony of the doors” is an odd way of describing safety critical procedures that reflect this.

Yes, a driver will be able to stop if they’re signalled into the wrong platform.

 

[LCC106]

Thanks for all the explanations.

 

[357]

would he be able to stop if the first indication he gets of the platform he is routed into is at the last (yellow) signal before the buffers.

In this location - I would say yes. The signal is very well lit and positioned, and is in a tunnel. So very easy to see.

 

[norbitonflyer]

In this location - I would say yes. The signal is very well lit and positioned, and is in a tunnel. So very easy to see.

Not sure i explained my question. If the signal is showing a clear aspect, the previous one will have been at double yellow or green and the driver will not have been expecting to stop at it, and may not be able to do so, however good the sighting is. That is, after all, the point of having yellow and double yellow aspects. So if, after passing a signal showing he is clear to the buffer stops, he then sees the next signal with an incorrect routing, what happens next?

So when 1P83 was routed to signal 2068 as booked it could only access platforms 7-10 if moving forwards, all of which were occupied and as zwk500 points out above only 7 and 8 were usable.

I note also from the diagram on OTT that signal 2068 also controls the junction into the Canal Tunnels. Assuming it could stop before that signal, could the train not have been diverted into St Pancras Low level? Or did the problem only become apparent when it reached the next signal (2046) - which is beyond the canal Tunnel junction and the last before the bufferstops?

 

[357]

Not sure i explained my question. If the signal is showing a clear aspect, the previous one will have been at double yellow or green and the driver will not have been expecting to stop at it, and may not be able to do so, however good the sighting is. That is, after all, the point of having yellow and double yellow aspects. So if, after passing a signal showing he is clear to the buffer stops, he then sees the next signal with an incorrect routing, what happens next?

Sorry - I did understand you but I think you misunderstood my reply!

The speed limit here is 40 at the start of the tunnel and 20 at the end of the tunnel. I find myself approaching the signal at about 30ish most of the time and can read the platform number from well before the magnet (200m from the signal). Any unit should be able to stop before the signal comfortably.

== Doublepost prevention - post automatically merged: 12 Jun 2025 ==

I note also from the diagram on OTT that signal 2068 also controls the junction into the Canal Tunnels. Assuming it could stop before that signal, could the train not have been diverted into St Pancras Low level? Or did the problem only become apparent when it reached the next signal (2046) - which is beyond the canal Tunnel junction and the last before the bufferstops?

The problem would only become apparent at the final signal.

 

[norbitonflyer]

The problem would only become apparent at the final signal.

So the reference to signal 2068 should have been to 2046? (If OTT is correct 2068 gives access only to platforms 7-10 and the Thameslink core, 2046 only to platforms 7-10)

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

 

[357]

So the reference to signal 2068 should have been to 2046? (If OTT is correct 2068 gives access only to platforms 7-10 and the Thameslink core, 2046 only to platforms 7-10)

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

I'm not in a position to get my maps out at the moment - but you don't get any indication what platform you'll be going into until the final signal in Gasworks Tunnels on any of the lines.

At the signals before that in Copenhagen Tunnels, you want anything from A - F indicated to ensure you're going into the station, but you won't know what platform until the next signal.

 

[MCR247]

So the reference to signal 2068 should have been to 2046? (If OTT is correct 2068 gives access only to platforms 7-10 and the Thameslink core, 2046 only to platforms 7-10)

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

I’m not sure they’d want to divert it into the core - where would it terminate? What about it and the drivers return workings etc

 

[43066]

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

Trains for the core might be diverted into Kings Cross in certain circumstances (just as MML TL trains can be sent into the St Pancras upper level platforms), but almost certainly not the other way around, as there’s going to be no sensible way of terminating/turning back, and it would absolutely scupper the TL service.

 

[godfreycomplex]

I'm not sure I'd personally blame Hull Trains for it - there are industry systems that advise staff of train lengths.

However, I've always found in my career that a courtesy call to the signaller goes a long way (both on the day and to improve the working relationship!)

My calls normally start after introductions, with something along the lines of "I know you guys are normally the last to be told anything, and I don't know if it affects your plans... Just checking you know I'm longer than normal today"

Always welcome from the signaller’s point of view

Once got censured by a shift manager for saying in a similar situation “personal question drive, how long are you?”

But that’s another story….

 

[Steve Harris]

Yeas, but this wasn't an ARS issue as such. More that one issue was compounded by ARS being left to do it's own thing when it shouldn't have.

== Doublepost prevention - post automatically merged: 12 Jun 2025 ==

I'm afraid it is very much an ARS issue, because ARS can't do what a human can, as in, make decisions when something other than when the planned schedule/timetable happens. It will do what it's programmed to do, which it did, it set the route as per schedule, but because the booked platform wasn't available (due to an out of plan decision earlier) it couldn't do what it was programmed to do and signalled a train into the proverbial cul-de-sac (a bit like sat navs have done with vehicles in the past).

As for ARS doing it's own thing.... it did what it was programmed to do, obviously the situation didn't get noticed until it was a problem, which could of been down to a number of reasons (to much workload for the signaller being one of many).

== Doublepost prevention - post automatically merged: 12 Jun 2025 ==

So the reference to signal 2068 should have been to 2046? (If OTT is correct 2068 gives access only to platforms 7-10 and the Thameslink core, 2046 only to platforms 7-10)

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

No. If you look at the playback of the signalling system referenced earlier up thread (post #8) the train only got as far as 2068.

 

[43066]

I'm afraid it is very much an ARS issue, because ARS can't do what a human can, as in, make decisions when something other than when the planned schedule/timetable happens.

I suppose the point the poster you quoted was making, though, isn’t so much that ARS is “rubbish”, rather that it’s a system with known limitations which need to be worked around by humans proactively intervening when required.

Obviously there’s another question about whether ARS boxes have sufficient human supervision.

 

[LAX54]

It is rather amusing reading all the different hypothesise of the above situation when in fact it's clearly a limitation of ARS (ie it can not look ahead to see if the booked platform is/will become available).

It will sit the train at the last signal until the route bcomes available.

 

[357]

Incidentally something similar happened yesterday morning, however a dynamic solution was found that didn't involve any shunting, but it did cause a delay.

 

[Horizon22]

So the reference to signal 2068 should have been to 2046? (If OTT is correct 2068 gives access only to platforms 7-10 and the Thameslink core, 2046 only to platforms 7-10)

As we are told the unit was a 700, diversion into the core would have been possible if the problem had been realised before passing 2068 (subject to driver's route knowledge).

That might have stuffed up the Thameslink core considerably - okay you’ve diverted it but then what? You’d have to shunt back out of a two-track tunnel so delaying a whole bunch of Thameslink departures too.

That just moves the problem elsewhere and the signaller for the Thameslink core probably wouldn’t have accept it

 

[357]

That might have stuffed up the Thameslink core considerably - okay you’ve diverted it but then what? You’d have to shunt back out of a two-track tunnel so delaying a whole bunch of Thameslink departures too.

That just moves the problem elsewhere and the signaller for the Thameslink core probably wouldn’t have accept it

While we are at it, I can think of some freight flows that we can divert that way too. I'm sure a 66 won't cause any issues in the core