Recent signalling failures.

[harz99]

Just recently there have been a significant amount of very disruptive and major signal failures in different areas of the network.

AFAIK most of the outages are in areas with recent modern signalling installations, which sets me wondering about the vulnerability of digital and other modern systems to hacking from the likes of Russia or protest groups and the like?

Could that be the case, or is the network just being very unlucky, suffering from hot weather and so on.

 

[The Planner]

Just recently there have been a significant amount of very disruptive and major signal failures in different areas of the network.

AFAIK most of the outages are in areas with recent modern signalling installations, which sets me wondering about the vulnerability of digital and other modern systems to hacking from the likes of Russia or protest groups and the like?

Could that be the case, or is the network just being very unlucky, suffering from hot weather and so on.

They aren't being hacked. If that was the case you wouldn't just cause disruption in one place for a limited time. It isn't like that god awful series on tv a little while back where the train was hacked.

 

[harz99]

They aren't being hacked. If that was the case you wouldn't just cause disruption in one place for a limited time. It isn't like that god awful series on tv a little while back where the train was hacked.

Thanks, I wasn't even thinking about the TV series. Is it just bad luck or are there bigger issues?

 

[Horizon22]

I'm sure someone with the performance statistics will be able to tell us objectively whether this is getting worse or not, as opposed to anecodtes.

Certain lines are suffering more with the signal-related issues it seems to be, including the GWML and WCML, but again the stats would be the actual confirmation of this. Generally there are lots of reports that NR maintenance budgets are stretched very thing, which may mean trying to "sweat" assets for longer and potentially repeat failures.

As with anything, some people depending on where they live, their origin and destination and when they travel on a given route might have vastly different experiences on the same line.

 

[Peter Mugridge]

Some signal failures, such as at Waterloo last night and again mid morning today will by their very nature get more coverage and cause a lot more disruption than a whole week of failures on some small branch line somewhere won't they?

 

[Annetts key]

Just recently there have been a significant amount of very disruptive and major signal failures in different areas of the network.

AFAIK most of the outages are in areas with recent modern signalling installations, which sets me wondering about the vulnerability of digital and other modern systems to hacking from the likes of Russia or protest groups and the like?

Could that be the case, or is the network just being very unlucky, suffering from hot weather and so on.

The vital parts of the signalling system are not connected to, or able to be influenced by anyone using the internet or world wide web.

The only way to interfere is by physical access to the railway infrastructure.

I'm sure someone with the performance statistics will be able to tell us objectively whether this is getting worse or not, as opposed to anecodtes.

Certain lines are suffering more with the signal-related issues it seems to be, including the GWML and WCML, but again the stats would be the actual confirmation of this. Generally there are lots of reports that NR maintenance budgets are stretched very thing, which may mean trying to "sweat" assets for longer and potentially repeat failures.

As with anything, some people depending on where they live, their origin and destination and when they travel on a given route might have vastly different experiences on the same line.

This is an almost impossible to answer question, as there are so many variables, depending on the time frame, area, how you classify the failure etc.

However, I will say that unfortunately, the modern systems suffer from "the bigger they are, the harder they fall". This is a result of increased centralisation and the computer based technology used.

A computer based central interlocking HAS to be working and so does the communication links between it and the remote equipment. For large systems, there is a lot of equipment controlled by a centralised signalling system such as a modern signalling centre or ROC. So if a failure does affect these (despite the design trying to prevent a single failure from causing this, hence duplicated data links), it can absolutely cripple the train service, as there is no other back-up system and insufficient staff to set up ticket / long block working.

This was less of a problem with the older relay based interlocking, as the work was divided between different smaller local interlockings. Most faults caused only one or a small number of functions to fail. The main vulnerability with these was power and damage to multicore lineside cables.

The remote control / communication system between the PSB and the local interlocking was also a weak point, but most places had either duplicated remote control systems (so if one failed, the other could take over) and / or a fall back to a through-routes / selective overrides or equivalent system.

But during summer, the main causes of failures is the hot weather (including wide temperature swings). Yes, the equipment is designed for these conditions, but if a component is starting to go out of specification, either hot or cold weather will most likely cause the equipment it is part of to fail.

 

[Class15]

Until very recently the North London/Mildmay line took the ‘award’.

Constant problems with the signalling, including an infamous incident where the entire system used to control the whole line east of Camden Road completely failed. Passenger services nonexistent for both peaks, freight trains delayed because drivers didn’t sign via South Tottenham route etc etc

Signalling problems seemed to be particularly common on the South Acton to Richmond section, which even got its own thread!

Something has now presumably been done about it; I don’t know what exactly, but the line has become much more reliable now.

 

[LNW-GW Joint]

Certain lines are suffering more with the signal-related issues it seems to be, including the GWML and WCML, but again the stats would be the actual confirmation of this. Generally there are lots of reports that NR maintenance budgets are stretched very thing, which may mean trying to "sweat" assets for longer and potentially repeat failures.

Fortunes have been spent recently on renewing and upgrading signalling on the GWML and WCML, with the ECML currently in progress north from London.
These are not obsolescent signalling installations, or have been victims of NR budgeting, far from it.
Granted the WCML signalling north of Weaver Jn is 50 years old and well overdue for replacement (with ETCS?).

 

[Horizon22]

The vital parts of the signalling system are not connected to, or able to be influenced by anyone using the internet or world wide web.

The only way to interfere is by physical access to the railway infrastructure.



This is an almost impossible to answer question, as there are so many variables, depending on the time frame, area, how you classify the failure etc.

However, I will say that unfortunately, the modern systems suffer from "the bigger they are, the harder they fall". This is a result of increased centralisation and the computer based technology used.

A computer based central interlocking HAS to be working and so does the communication links between it and the remote equipment. For large systems, there is a lot of equipment controlled by a centralised signalling system such as a modern signalling centre or ROC. So if a failure does affect these (despite the design trying to prevent a single failure from causing this, hence duplicated data links), it can absolutely cripple the train service, as there is no other back-up system and insufficient staff to set up ticket / long block working.

This was less of a problem with the older relay based interlocking, as the work was divided between different smaller local interlockings. Most faults caused only one or a small number of functions to fail. The main vulnerability with these was power and damage to multicore lineside cables.

The remote control / communication system between the PSB and the local interlocking was also a weak point, but most places had either duplicated remote control systems (so if one failed, the other could take over) and / or a fall back to a through-routes / selective overrides or equivalent system.

But during summer, the main causes of failures is the hot weather (including wide temperature swings). Yes, the equipment is designed for these conditions, but if a component is starting to go out of specification, either hot or cold weather will most likely cause the equipment it is part of to fail.

I meant more delay attribution that had been coded under signalling issues.

 

[Bald Rick]

What is described as a “signalling failure” can be many things.

The recent heat has had a big impact.

 

[Nicholas Lewis]

Fortunes have been spent recently on renewing and upgrading signalling on the GWML and WCML, with the ECML currently in progress north from London.
These are not obsolescent signalling installations, or have been victims of NR budgeting, far from it.
Granted the WCML signalling north of Weaver Jn is 50 years old and well overdue for replacement (with ETCS?).

You never hear of Warrington, Preston or Carlisle falling over like we've seen with more recently installed signalling systems despite their age. Signalling installations commissioned nearly decade after these are being replaced aka Cambridge ASC because the wiring is an issue with degraded insulation. Then we have a few isolated mechanical areas like Shrewsbury to which NR have no immediate plans to replace. In some respects in a low speed area mechanical signalling with some additional aids like TPWS are now seen as enduring and lets be honest with the vast cost (£375m on Feltham which fell over yesterday evening for example) of resignalling its the right solution.

 

[harz99]

@Annetts key thank you for the detailed explanation.

 

[High Dyke]

Whilst not a cause of failures, the backlog in signal maintenance along some routes, some of which is caused by a lack of trained technicians doesn't help. Just last week a lad, with experience of heritage railways, was helping assess the state of some rodding for level crossing gates at the location i was working..

Some signals may require specific parts, which may not be readily available. Even the more modern systems can be affected.

When Andrew Haines started at Network Rail he expressed some concern about "putting all your eggs in one basket", when talking about ROC's. He mused "How can I justify disruption to passengers at King's Cross caused by a signalling fault at York."

 

[endofthetunnel]

Cable theft is hugely on the rise at the moment again, which often is publicly called ‘signal fault’ or ‘signalling issues’ as that’s how it presents, and until it’s proven as theft it won’t be described as that.

 

[merry]

Then we have a few isolated mechanical areas like Shrewsbury to which NR have no immediate plans to replace. In some respects in a low speed area mechanical signalling with some additional aids like TPWS are now seen as enduring

I have heard that a large part of the Shrewsbury hold out includes there being no way (yet found) to resignal the layout, or an acceptable revision, according to current uk signalling principles. E.g. no room for overlaps, with signals very close to switches carrying conflicting traffic, etc.
And if not Shrewsbury, then not the Marches, Crewe and Chester lines. At the moment the high cost of modernisation is also perhaps making the staffing costs of mechanical look less expensive than in the past.

 

[Annetts key]

You never hear of Warrington, Preston or Carlisle falling over like we've seen with more recently installed signalling systems despite their age. Signalling installations commissioned nearly decade after these are being replaced aka Cambridge ASC because the wiring is an issue with degraded insulation.

Vital British railway signalling wiring and cabling uses PCP rubber insulation. Some batches used during the 1980s have been found to have a rather shorter life than expected, suffering from the inner insulation degrading and becoming brittle, then changing to go a fine powder. Meanwhile the outer sheath also cracks and splits. Heat accelerates the process (despite what it says in the linked to web site).

This has all been known about for many, many years. In many places, the powers that be decided not to replace the wiring until they were ready to spend money on a new signalling scheme.

 

[MadMac]

You never hear of Warrington, Preston or Carlisle falling over like we've seen with more recently installed signalling systems despite their age. Signalling installations commissioned nearly decade after these are being replaced aka Cambridge ASC because the wiring is an issue with degraded insulation. Then we have a few isolated mechanical areas like Shrewsbury to which NR have no immediate plans to replace. In some respects in a low speed area mechanical signalling with some additional aids like TPWS are now seen as enduring and lets be honest with the vast cost (£375m on Feltham which fell over yesterday evening for example) of resignalling its the right solution.

Haymarket and Carstairs (both 1970s installations) had dry wire degradation issues.

 

[Harpo]

When Andrew Haines started at Network Rail he expressed some concern about "putting all your eggs in one basket", when talking about ROC's. He mused "How can I justify disruption to passengers at King's Cross caused by a signalling fault at York."

He could have just asked Robin Gisby as the Operating Strategy was his creation while at NR.

== Doublepost prevention - post automatically merged: 20 Jul 2025 ==

At the moment the high cost of modernisation is also perhaps making the staffing costs of mechanical look less expensive than in the past.

It has not been possible to make a business case for resignalling based just on staff savings for a very long time. Hence why robust mechanical boxes have survived for 100-150 years.

 

[harz99]

Waterloo again this morning, with a "do not travel" warning...

 

[D365]

Granted the WCML signalling north of Weaver Jn is 50 years old and well overdue for replacement (with ETCS?).

Correct. The 'TriLink' project is being mobilised to deploy ETCS Level 2 to WCML.

 

[Deepgreen]

Waterloo again this morning, with a "do not travel" warning...

Yes - yet again! P1-14 OOU, and the severe disruption stated as continuing until 1600, but p1-14 were back in use again some time ago.

 

[bahnause]

The vital parts of the signalling system are not connected to, or able to be influenced by anyone using the internet or world wide web.

The only way to interfere is by physical access to the railway infrastructure.

Siemens now offers interlocking via the cloud with "Simis Interlocking as a Service" (SIaaS). It has been tested by Appenzellerbahnen among others and, as far as I know, is currently in use by the Gornergratbahn.

 

[MarkyT]

It has not been possible to make a business case for resignalling based just on staff savings for a very long time. Hence why robust mechanical boxes have survived for 100-150 years.

Like Trigger's broom, with almost every component replaced and often modernised and reconfigured several times over their lives, except for buildings and major frame metalwork. Another constraint is overall capital spend and workload. Heavily trafficked main line signalling, of whatever technology, has always worn out or reached capacity the fastest, so it has needed replacement a number of times during the life of the remaining oldest boxes, usually on secondary and rural routes, keeping signalling suppliers busy. It's unfortunate that some of the most marginal railways carry the highest operating costs per train mile because the industry can't afford the capital or capacity to modernise and reduce staffing. This has always been the case. Nonetheless, cost effective small scale schemes to consolidate two boxes, transfer supervision of a crossing to a neighboring box, saving an attendant etc, continue. Some small boxes have been completely modernised with colour lights, new interlocking, and a panel. Because of these small projects most remaining small boxes actually have quite wide control areas, often supervising many level crossings in rural areas. Moving such employees to a panel may not save that much workload so the staff savings of old where a panel position or workstation could take over many boxes on the mainline no longer apply. While staffing and management of small boxes is challenging, moving the same signallers to a nearby panel or ROC may not be so cost effective if they end up joining a staff pool at a higher grade than their old box.

== Doublepost prevention - post automatically merged: 21 Jul 2025 ==

Siemens now offers interlocking via the cloud with "Simis Interlocking as a Service" (SIaaS). It has been tested by Appenzellerbahnen among others and, as far as I know, is currently in use by the Gornergratbahn.

Do all the signals go to red if the monthly payment isn't made on time?

 

[321362]

Until very recently the North London/Mildmay line took the ‘award’.

Constant problems with the signalling, including an infamous incident where the entire system used to control the whole line east of Camden Road completely failed. Passenger services nonexistent for both peaks, freight trains delayed because drivers didn’t sign via South Tottenham route etc etc

Signalling problems seemed to be particularly common on the South Acton to Richmond section, which even got its own thread!

Something has now presumably been done about it; I don’t know what exactly, but the line has become much more reliable now.

It's still extremely bad imo, I actively avoid travelling on it and go out of my way to take alternative routes as it's SO unreliable

 

[Class15]

It's still extremely bad imo, I actively avoid travelling on it and go out of my way to take alternative routes as it's SO unreliable

What times is it still so bad? I use it in the peaks and it seems to be much better now. I don’t know about weekends as I rarely travel that way on weekends.

 

[Nicholas Lewis]

Siemens now offers interlocking via the cloud with "Simis Interlocking as a Service" (SIaaS). It has been tested by Appenzellerbahnen among others and, as far as I know, is currently in use by the Gornergratbahn.

Umm that sounds like a cyber security nightmare to me. A self contained processor system with no direct interaction with the outside world sounds the safest bet to me.

 

[Horizon22]

Cable theft is hugely on the rise at the moment again, which often is publicly called ‘signal fault’ or ‘signalling issues’ as that’s how it presents, and until it’s proven as theft it won’t be described as that.

Yes they can be. Which is why I suggested the delay attribution statistics and performance data - which doesn't care about about what reason is presented on the CIS - will be much more relevant to the question at hand.

 

[bahnause]

Umm that sounds like a cyber security nightmare to me. A self contained processor system with no direct interaction with the outside world sounds the safest bet to me.

With encrypted connections and proper identification of the sender and recipient, this is quite easy to implement securely. ETCS L2 also uses encrypted over-the-air connections.

 

[ac6000cw]

With encrypted connections and proper identification of the sender and recipient,

...and data error detection/correction/integrity checking (basically you assume the links are not 100% reliable and design the overall system/protocols to deal with that). All everyday stuff in modern electronic communication links, be they physical or 'over the ether'.

 

[voyagerdude220]

You never hear of Warrington, Preston or Carlisle falling over like we've seen with more recently installed signalling systems despite their age. Signalling installations commissioned nearly decade after these are being replaced aka Cambridge ASC because the wiring is an issue with degraded insulation. Then we have a few isolated mechanical areas like Shrewsbury to which NR have no immediate plans to replace. In some respects in a low speed area mechanical signalling with some additional aids like TPWS are now seen as enduring and lets be honest with the vast cost (£375m on Feltham which fell over yesterday evening for example) of resignalling its the right solution.

Signalling around Acton Bridge/Warrington failed last Friday morning for a couple of hours. Several Avanti services which were supposed to run via Warrington diverted via Manchester.