South ECML ARS after re-control from King's Cross PSB to York ROC

[notverydeep]

An observation of a questionable routing choice last night prompts me to ask whether the Automated Route Setting (ARS) on the Southern part of the ECML between King's Cross and Stevenage has been altered as part of the switch over of signalling control from King's Cross Power Signal Box (PSB) to York Regional Control Centre...

The observation concerns 9S56 1912 Brighton to Cambridge. This is one of three Brighton to Cambridge Thameslink trains each weekday evening that call at Welwyn Garden City (WGC), Welwyn North and Knebworth in addition to the normal calling pattern for such trains which omits these stations. This is despite the trains being operated by 700/1 12 car sets (700151 yesterday) and all three stations having only 8 car platforms. The planned route appears to be on the slow line from Finsbury Park to WGC, although the first time I took this train from Blackfriars to WGC, it was routed on the fast line between Finsbury Park and Potters Bar. I assume that it cannot use the fast line all the way to WGC because the rear of a 12 car 700/1 stopped in platform 3 at WGC would still be fowling the crossover from the Down Fast Line to the Down Slow Line at the South end of WGC station.

Yesterday, the train had recovered from 19 minutes late passing Bricklayers Arms Junction on the approach to London Bridge, to 12 minutes late departing Finsbury Park with some smart dwell time management by the driver. However it was for nothing. Following the booked route from Finsbury Park, it had caught 2V53 2027 Moorgate to WGC all stations by Alexandra Palace and a tedious slow run, with a 25 minute late arrival at WGC. This was pulled back to 22 minutes late by Baldock and on arrival at Cambridge. Yet this extra delay could have easily been avoided by being routed via the Down Fast line to Potters Bar, enabling 9S56 to overtake 2V53. No northbound fast trains passed Finsbury Park on the fast line until 7 minutes later when the following 9J64 1925 Horsham to Peterborough (next stop Stevenage) departed. At 9 1/2 minutes after 9S56 departed Finsbury Park (its booked time to passing Potters Bar), 9J64 had only reached Hornsey, so there would have been no chance of 9S56 delaying 9J64.

It is quite common for the 2CXX (King's Cross - Cambridge) and 9YXX (Sevenoaks - WGC) services to be impeded by slower Moorgate trains, but 9S56 is one of very few trains that run non stop Finsbury Park to WGC - the first such trains at useable times for some years, though the 2 x 317s were previously booked 14 minutes using the fast line to Potters Bar rather than 16 minutes for the slow line path 9S56 gets! Yesterdays incident does seem to have a chance to avoid Delay Repay payments not taken.

Does anyone know if the ARS can automatically optimise routing over this section, or can opportunities to make up time only be taken by a signaller (I understand that they won't be able to watch every train) and has the likelihood of ARS varying the route taken changed with the move of control to York ROC?

 

[lineclear]

An observation of a questionable routing choice last night prompts me to ask whether the Automated Route Setting (ARS) on the Southern part of the ECML between King's Cross and Stevenage has been altered as part of the switch over of signalling control from King's Cross Power Signal Box (PSB) to York Regional Control Centre...

The observation concerns 9S56 1912 Brighton to Cambridge. This is one of three Brighton to Cambridge Thameslink trains each weekday evening that call at Welwyn Garden City (WGC), Welwyn North and Knebworth in addition to the normal calling pattern for such trains which omits these stations. This is despite the trains being operated by 700/1 12 car sets (700151 yesterday) and all three stations having only 8 car platforms. The planned route appears to be on the slow line from Finsbury Park to WGC, although the first time I took this train from Blackfriars to WGC, it was routed on the fast line between Finsbury Park and Potters Bar. I assume that it cannot use the fast line all the way to WGC because the rear of a 12 car 700/1 stopped in platform 3 at WGC would still be fowling the crossover from the Down Fast Line to the Down Slow Line at the South end of WGC station.

Yesterday, the train had recovered from 19 minutes late passing Bricklayers Arms Junction on the approach to London Bridge, to 12 minutes late departing Finsbury Park with some smart dwell time management by the driver. However it was for nothing. Following the booked route from Finsbury Park, it had caught 2V53 2027 Moorgate to WGC all stations by Alexandra Palace and a tedious slow run, with a 25 minute late arrival at WGC. This was pulled back to 22 minutes late by Baldock and on arrival at Cambridge. Yet this extra delay could have easily been avoided by being routed via the Down Fast line to Potters Bar, enabling 9S56 to overtake 2V53. No northbound fast trains passed Finsbury Park on the fast line until 7 minutes later when the following 9J64 1925 Horsham to Peterborough (next stop Stevenage) departed. At 9 1/2 minutes after 9S56 departed Finsbury Park (its booked time to passing Potters Bar), 9J64 had only reached Hornsey, so there would have been no chance of 9S56 delaying 9J64.

It is quite common for the 2CXX (King's Cross - Cambridge) and 9YXX (Sevenoaks - WGC) services to be impeded by slower Moorgate trains, but 9S56 is one of very few trains that run non stop Finsbury Park to WGC - the first such trains at useable times for some years, though the 2 x 317s were previously booked 14 minutes using the fast line to Potters Bar rather than 16 minutes for the slow line path 9S56 gets! Yesterdays incident does seem to have a chance to avoid Delay Repay payments not taken.

Does anyone know if the ARS can automatically optimise routing over this section, or can opportunities to make up time only be taken by a signaller (I understand that they won't be able to watch every train) and has the likelihood of ARS varying the route taken changed with the move of control to York ROC?

Kings Cross PSB was a power box, and so did not have ARS (the Moorgate branch had some kind of automatic working though) and the workstations covering the area described above do not have ARS.

 

[Bald Rick]

Does anyone know if the ARS can automatically optimise routing over this section, or can opportunities to make up time only be taken by a signaller (I understand that they won't be able to watch every train) and has the likelihood of ARS varying the route taken changed with the move of control to York ROC?

Kings Cross didn’t have ARS, and even if the relevant workstations at York did have ARS (I don’t know if they do or not), ARS is not smart enough to make the move you proposed.

 

[Geeves]

ARS is definitely not smart. It will route the train to which ever line it's been programed to do whether it's behind a stopper or not.

Unless control rings to tell the box otherwise they won't mess with it manually controlled too. That's the way of the modern railway

 

[Horizon22]

Indeed - ARS isn’t “smart” computing but is more the “repetitive” type of computing - read the timetable and make all the moves as per the timetable.

In fact during disrupted working, it is often best switched off as it causes trains to unnecessarily wait. For instance, this is often the case during late running and approaching terminals when an adjacent platform/line is available next to the booked one and no manual intervention is made.

Other traffic management tools might “suggest” or highlight a clash but again that still needs a good signaller to notice and intervene.

 

[bramling]

An observation of a questionable routing choice last night prompts me to ask whether the Automated Route Setting (ARS) on the Southern part of the ECML between King's Cross and Stevenage has been altered as part of the switch over of signalling control from King's Cross Power Signal Box (PSB) to York Regional Control Centre...

The observation concerns 9S56 1912 Brighton to Cambridge. This is one of three Brighton to Cambridge Thameslink trains each weekday evening that call at Welwyn Garden City (WGC), Welwyn North and Knebworth in addition to the normal calling pattern for such trains which omits these stations. This is despite the trains being operated by 700/1 12 car sets (700151 yesterday) and all three stations having only 8 car platforms. The planned route appears to be on the slow line from Finsbury Park to WGC, although the first time I took this train from Blackfriars to WGC, it was routed on the fast line between Finsbury Park and Potters Bar. I assume that it cannot use the fast line all the way to WGC because the rear of a 12 car 700/1 stopped in platform 3 at WGC would still be fowling the crossover from the Down Fast Line to the Down Slow Line at the South end of WGC station.

Yesterday, the train had recovered from 19 minutes late passing Bricklayers Arms Junction on the approach to London Bridge, to 12 minutes late departing Finsbury Park with some smart dwell time management by the driver. However it was for nothing. Following the booked route from Finsbury Park, it had caught 2V53 2027 Moorgate to WGC all stations by Alexandra Palace and a tedious slow run, with a 25 minute late arrival at WGC. This was pulled back to 22 minutes late by Baldock and on arrival at Cambridge. Yet this extra delay could have easily been avoided by being routed via the Down Fast line to Potters Bar, enabling 9S56 to overtake 2V53. No northbound fast trains passed Finsbury Park on the fast line until 7 minutes later when the following 9J64 1925 Horsham to Peterborough (next stop Stevenage) departed. At 9 1/2 minutes after 9S56 departed Finsbury Park (its booked time to passing Potters Bar), 9J64 had only reached Hornsey, so there would have been no chance of 9S56 delaying 9J64.

It is quite common for the 2CXX (King's Cross - Cambridge) and 9YXX (Sevenoaks - WGC) services to be impeded by slower Moorgate trains, but 9S56 is one of very few trains that run non stop Finsbury Park to WGC - the first such trains at useable times for some years, though the 2 x 317s were previously booked 14 minutes using the fast line to Potters Bar rather than 16 minutes for the slow line path 9S56 gets! Yesterdays incident does seem to have a chance to avoid Delay Repay payments not taken.

Does anyone know if the ARS can automatically optimise routing over this section, or can opportunities to make up time only be taken by a signaller (I understand that they won't be able to watch every train) and has the likelihood of ARS varying the route taken changed with the move of control to York ROC?

Since May 18, the timetable has been arranged so that virtually all of the 2Cxx stopping services stay on the slow lines, and as such the signallers tend to leave things alone, especially as there's elements of recovery time built in to the schedules, not least very long turnround times at each end. The "old school" King's Cross PSB of the 1990s would certainly have made attempts to make full use of the infrastructure, however this isn't really how things are done any more. That said, I've just come off a service which *was* turned out on to the up fast at Potters Bar, however since May 18 this is very much the exception, and no doubt this will have been encouraged by it being first thing in the morning.

Why the industry pays people good money to do a job, and then seems not to trust the people to actually do that job to the fullest and most worthwhile extent, is another discussion!

 

[edwin_m]

Even in the early days when I was involved with it, ARS was capable of running trains in something other than timetable order and could include specific strategies for diverting them onto alternative lines. However, like all computer systems it's only as good as what is put into it, and I suspect the problem is more to do with these rules being incorrect or allowed to get out of date.

 

[louis97]

Why the industry pays people good money to do a job, and then seems not to trust the people to actually do that job to the fullest and most worthwhile extent, is another discussion!

Unfortunately, with the increasing roll out of ARS on new workstations as a result of closure of various boxes, in some circumstances this has been used to reduce the number of signallers for a given area. The result is busier signallers that don't have the time to regulate the train service in the same way. Combine this with poor training for signallers, where they are taught how to signal and not necessarily how to regulate and unfortunately it is only the passenger that suffers.

Even in the early days when I was involved with it, ARS was capable of running trains in something other than timetable order and could include specific strategies for diverting them onto alternative lines. However, like all computer systems it's only as good as what is put into it, and I suspect the problem is more to do with these rules being incorrect or allowed to get out of date.

I suspect no one wanted to pay for someone to keep them up to date, as such the system sticks to booked route.

 

[SPADTrap]

The standard of signalling on the ECML and indeed generally has become absolutely diabolical. I understand the reasons why but York ROC are something else.

 

[notverydeep]

Since May 18, the timetable has been arranged so that virtually all of the 2Cxx stopping services stay on the slow lines, and as such the signallers tend to leave things alone, especially as there's elements of recovery time built in to the schedules, not least very long turnaround times at each end. The "old school" King's Cross PSB of the 1990s would certainly have made attempts to make full use of the infrastructure, however this isn't really how things are done any more. That said, I've just come off a service which *was* turned out on to the up fast at Potters Bar, however since May 18 this is very much the exception, and no doubt this will have been encouraged by it being first thing in the morning.

Indeed - I am used to this and to be fair at peak times and I have not been across Marsh Moor crossover from the Up Slow to the Up Fast in years! The stopping pattern of a late running 2CXX would often hold things up with the volume of fast line trains - although as you say, there is the occasional fast run between Potters Bar and Finsbury Park. However these four specific trains: 9S56 1912 Brighton to Cambridge, 9S60 2012 Brighton to Cambridge, 9S64 2112 Brighton to Cambridge and 9S68 2212 Brighton to Cambridge are unusual being non-stop Finsbury Park to WGC. They don't date from the original May 18 version of the timetable (which had now withdrawn 2CXX trains roughly occupying similar paths) - that had only one very early SB Peterborough - Horsham and one very late NB Horsham - Peterborough call at WGC. Strangely the SB service does not have corresponding stops at Knebworth, Welwyn North or WGC on Cambridge to Brighton trains. The 16 minute start to stop timing from Finsbury Park to WGC means that they potentially loose much more time following a Moorgate service than the equivalent 2CXX would as these have 23 minutes plus a bit of recovery at WGC.

Kings Cross PSB was a power box, and so did not have ARS (the Moorgate branch had some kind of automatic working though) and the workstations covering the area described above do not have ARS.

Sorry, my memory must be playing tricks as I thought I had read a Roger Ford article in Modern Railways that suggested ARS of some type was in use, in the context of upgrades to it being demonstrated - I think he referred to it as ARS+. This probably explains the low level of fast line use to recover time, even where my reading of the Open Train Times map suggests it would not have impacted other services, as the opportunity would have to be noticed by the signaller and they would have to be confident that fast line running would not cause a further delay.

 

[SPADTrap]

Since May 18, the timetable has been arranged so that virtually all of the 2Cxx stopping services stay on the slow lines, and as such the signallers tend to leave things alone, especially as there's elements of recovery time built in to the schedules, not least very long turnround times at each end. The "old school" King's Cross PSB of the 1990s would certainly have made attempts to make full use of the infrastructure, however this isn't really how things are done any more. That said, I've just come off a service which *was* turned out on to the up fast at Potters Bar, however since May 18 this is very much the exception, and no doubt this will have been encouraged by it being first thing in the morning.

Why the industry pays people good money to do a job, and then seems not to trust the people to actually do that job to the fullest and most worthwhile extent, is another discussion!

Can't put it much better than this!

 

[Class 170101]

ARS can be re-programmed to follow something else if told. Roger Ford did an article about it in Modern Railways in the last couple of years, however this was at Didcot TVSC rather than Kings Cross or York.

 

[Sunset route]

ARS can be re-programmed to follow something else if told. Roger Ford did an article about it in Modern Railways in the last couple of years, however this was at Didcot TVSC rather than Kings Cross or York.

That’s not ARS but TMS which in turn interacts with ARS.

 

[bramling]

ARS can be re-programmed to follow something else if told. Roger Ford did an article about it in Modern Railways in the last couple of years, however this was at Didcot TVSC rather than Kings Cross or York.

Shouldn’t the signallers be paying attention to what the ARS is doing, or going to do? ARS was always intended to be an aid, not a tool for only looking at the panel when there’s a problem.

 

[CAF397]

There is a blame culture within Network Rail. If the ARS delays the train, the minutes get recorded against the system. If the signaller intervenes and changes what ARS does, but causes a delay somewhere, the signaller becomes liable for the delay and has to explain.

The signallers will be watching the ARS, they will see a conflict, but will be reluctant to rectify it. This direct from a signaller I was talking to about the ARS around Ordsall Lane Jn in Manchester causing delays by running trains in booked order, stopping trains on time and holding them for several minutes.

 

[Bald Rick]

Shouldn’t the signallers be paying attention to what the ARS is doing, or going to do? ARS was always intended to be an aid, not a tool for only looking at the panel when there’s a problem.

That is part of their job, yes.
ARS was (and is) intended to reduce workload on the signaller for a given geography so that each workstation can therefore cover a wider area. The signaller may have 10-12 trains under his/her control at any one time, and may also be checking paperwork for possessions, checking simplifiers for movements not in ARS (of which there are many), speaking to someone on the ground (a shunter, carriage yard supervisor, holder of a line block), talking to control, speaking to their supervisor, training a new signaller, etc. etc. Obviously we don’t know what actually happened here, but it is very possible that the signaller was busy with some other activity and therefore did not see the potential for that specific regulating move for a train that is out of pattern.

 

[bramling]

That is part of their job, yes.
ARS was (and is) intended to reduce workload on the signaller for a given geography so that each workstation can therefore cover a wider area. The signaller may have 10-12 trains under his/her control at any one time, and may also be checking paperwork for possessions, checking simplifiers for movements not in ARS (of which there are many), speaking to someone on the ground (a shunter, carriage yard supervisor, holder of a line block), talking to control, speaking to their supervisor, training a new signaller, etc. etc. Obviously we don’t know what actually happened here, but it is very possible that the signaller was busy with some other activity and therefore did not see the potential for that specific regulating move for a train that is out of pattern.

That’s a bit of a stretch in this case though, as on paper there would have been four potential opportunities to turn the 700 out on to the down fast to overtake the 717 (Holloway, Finsbury Park, Alexandra Palace and New Barnet). That’s quite a long time for the signaller(s) to be engaged with other tasks.

 

[Bald Rick]

That’s a bit of a stretch in this case though, as on paper there would have been four potential opportunities to turn the 700 out on to the down fast to overtake the 717 (Holloway, Finsbury Park, Alexandra Palace and New Barnet). That’s quite a long time for the signaller(s) to be engaged with other tasks.

Not really. Fairly sure that’s across 2 workstations as well, although I may be corrected.

 

[Horizon22]

That’s a bit of a stretch in this case though, as on paper there would have been four potential opportunities to turn the 700 out on to the down fast to overtake the 717 (Holloway, Finsbury Park, Alexandra Palace and New Barnet). That’s quite a long time for the signaller(s) to be engaged with other tasks.

I think it has in some cases made signallers "lazy" in terms of there may be an option but they are happy to let ARS do its thing. I've seen some...questionable decisions (or lack thereof) made and the excuse is often something to do with ARS. I think it's a pyschological thing which has put certain individuals into a false sense of security. That is not to say that what @Bald Rick says isn't also correct and it has allowed panels to expand and that 95% of the time, ARS is a benefit for the every day timetable which would be difficult to manage otherwise.

 

[Bald Rick]

I think it has in some cases made signallers "lazy" in terms of there may be an option but they are happy to let ARS do its thing. I've seen some...questionable decisions (or lack thereof) made and the excuse is often something to do with ARS.

There’s no doubt about that. “Sorry Guv’ I was momentarily distracted”.

but to be fair that the same in areas without ARS.

 

[Class 170101]

That’s not ARS but TMS which in turn interacts with ARS.

Cheers for this, Roger Ford does write quite a bit though and clearly I've forgotten some of it.