Signalling problems on the Portsmouth Direct Line

[D6130]

I was speaking on the 'phone with a friend in Petersfield yesterday and he was telling me that, for the past week or so, Portsmouth Direct Line services have been sustaining delays of between 5 and 10 minutes in each direction because of signalling problems between Petersfield and Rowlands Castle. Apparently signals HT9, 11 and 13 on the Down Line and signals HT10, 8 and 6 on the Up Line are being held at danger until trains come to a stand at each one and the driver has communicated with the Havant signaller. My first thought was that there were maybe multiple track circuit failures in the area, caused by the recent bad weather; but then I remembered that this section of line is controlled by axle counters. Does anyone here have any idea what's going on?

 

[zwk500]

Faulty Axle Counters? Or possibly a track defect rather than actual signalling issue. IIRC those signals are around the general area of a tunnel, so perhaps the recent heavy rain has caused issues with drainage or something.

 

[D6130]

Faulty Axle Counters?

That's what I was thinking....but on both lines?

Or possibly a track defect rather than actual signalling issue.

Yes....or maybe a trackbed/formation issue caused by the exceptionally hot dry summer followed by the very wet winter, as happened on the Salisbury-Exeter line and - more recently with the Hook landslip.

IIRC those signals are around the general area of a tunnel, so perhaps the recent heavy rain has caused issues with drainage or something.

THe affected section of line is some miles to the South of Buriton Tunnel and runs through open chalk downland countryside via a mix of shallow cuttings and low embankments....which could well be the source of the problem.

I'm hoping that @pompeyfan may have more information.

 

[High Dyke]

Defective signalling cable between the two locations, for the last few days. An unrelated cable fault affecting services today between Havant and Portcreek Jn.

 

[D6130]

Defective signalling cable between the two locations, for the last few days. An unrelated cable fault affecting services today between Havant and Portcreek Jn.

Many thanks! I'll pass that on.

 

[infobleep]

The problem keeps reoccurring. And today they say services will be disrupted until tomorrow (Sunday 5 Feb).

This is what Network Rail said in a series of Tweets

The system in use in the area is run from Havant and has some unique characteristics that make it unusually disruptive when things go wrong.

This can be summed up by saying it's very fail-safe.
......
So far we've found issues with signal heads and cable faults, fixed them, found it has worked for day or two, and then the signals go red again and we have to reset again.

It's fair to say this is incredibly annoying and disruptive for passengers and is soaking up huge amounts of signal technician time and experts from across the region too.

We continue to work through every aspect of the kit in the Rowlands Castle area to bottom this out.

You will, however, be pleased to hear that designs are being drawn up for the eventual replacement of the system, albeit not for a good few years yet.

We hope that the work we're doing now will mean this works much better until then.

You can read the full series of Tweets at

https://twitter.com/NetworkRailWssx/status/1618294179821682689?t=pBgggrWI-aAtYHhkhrHdig&s=19

This is what Network Rail posted in a recent Tweet (captured from a image thanks to Google Lens):

At Rowlands Castle, we have a repeat fault which is causing the system to automatically shut down and impact three signals. We've carried out extensive tests and involved a number of independent experts, including the manufacturer. We've installed a number of extra 'loggers' which record how the system is working to help us understand exactly what happens when the fault occurs. Unfortunately most of this work can only be completed when trains are not running as it involves shutting down a wider area of the signalling system.

Once again I apologise for the ongoing disruption this is causing to our customers and assure you we are doing everything we can to resolve both of these issues.

Mark Killick
Route Director (Wessex)
Network Rail

https://twitter.com/NetworkRailWssx/status/1620387929494519809?t=YLU7aDR8pROypTacfQUvig&s=19

 

[pompeyfan]

The whole of Havant box needs ripping out and starting again, various signals with temporary approach control on because particular signal heads can’t display a green without the whole system going pop and going into a “hard” fail safe. Regular code 2 AWS issues (Horn vice bell), there’s been repeat failures at Bedhampton (although this wasn’t actually the fault of the signalling or its design), and Portcreek Jn. Various off indicators and banner repeaters also regularly bagged up because they’ve failed, which in some cases causes multiple signals to revert to red.

 

[DelW]

From NR's statement that "Unfortunately most of this work can only be completed when trains are not running as it involves shutting down a wider area of the signalling system" (taken from Infobleep's second quote above), then wouldn't it have been worth extending the intended closure in a fortnight's time to include the problem area? There is a full seven day closure (11th - 17th inc) from Guildford to Haslemere or Petersfield - once pax are put onto buses, then extending the bus bridge to Havant isn't much more disruptive or much slower than going to Haslemere or Petersfield, although presumably needing more buses. Since it's half term that oughtn't to be insuperable.

 

[zwk500]

From NR's statement that "Unfortunately most of this work can only be completed when trains are not running as it involves shutting down a wider area of the signalling system" (taken from Infobleep's second quote above), then wouldn't it have been worth extending the intended closure in a fortnight's time to include the problem area? There is a full seven day closure (11th - 17th inc) from Guildford to Haslemere or Petersfield - once pax are put onto buses, then extending the bus bridge to Havant isn't much more disruptive or much slower than going to Haslemere or Petersfield, although presumably needing more buses. Since it's half term that oughtn't to be insuperable.

There's a limit to how quickly work like that could be organised - the Farncombe to Petersfield resignalling will be fringing with Havant, so if you disconnect the signalling to do other fault finding you may impact on the testing regime for the planned blockade (I'm not sure if this is the final phase of the work, but I think so because they've had one blockade already). There's also only so many S&T people in one area, and Wessex has been having a rough time of it lately so they may already be allocated to jobs.

 

[DelW]

There's a limit to how quickly work like that could be organised - the Farncombe to Petersfield resignalling will be fringing with Havant, so if you disconnect the signalling to do other fault finding you may impact on the testing regime for the planned blockade (I'm not sure if this is the final phase of the work, but I think so because they've had one blockade already). There's also only so many S&T people in one area, and Wessex has been having a rough time of it lately so they may already be allocated to jobs.

I understand the issues of time taken to mobilise resources, but the tweet I quoted from suggested that they've already had boots (or screwdrivers?) on the ground for a while, but have been hampered by limited possession times. I was hoping that they might be used more productively without daytime trains running.

However, if those resources are also involved in the programmed blockade works, that would make my suggestion unworkable. I didn't know what type of works were scheduled during the closure.

 

[TSG]

The whole of Havant box needs ripping out and starting again, various signals with temporary approach control on because particular signal heads can’t display a green without the whole system going pop and going into a “hard” fail safe. Regular code 2 AWS issues (Horn vice bell), there’s been repeat failures at Bedhampton (although this wasn’t actually the fault of the signalling or its design), and Portcreek Jn. Various off indicators and banner repeaters also regularly bagged up because they’ve failed, which in some cases causes multiple signals to revert to red.

The interlocking and control system at Havant do need replacing. Unfortunately the majority of the lineside signalling and cables across that area also need replacing because much of is intrinsically linked to that particular type of interlocking, and incompatible with others. The only good news is that the re-signalling project has now got rolling.

 

[zwk500]

However, if those resources are also involved in the programmed blockade works, that would make my suggestion unworkable. I didn't know what type of works were scheduled during the closure.

IIRC: the Farncombe to Petersfield resignalling involves moving control from Farncombe, Haslemere, and Petersfield boxes to Basingstoke ROC and includes several level crossings, new or upgraded signals, and I possibly renewal of the S&C. It also includes swapping shunt signals for main aspects at Haslemere and Petersfield down platforms to allow turnback moves without shunting across. I'm not sure which parts of the work were done in the first blockade a year ago. I also can't remember if it's 1 or 2 panels but think it does have a separate LX workstation. HT1 is straight after Petersfield platform starter, and HT2 is the second signal in rear of the Petersfield Up platform starting signal so would be part of the aspect squences being tested with the changes to the crossover.

 

[D6130]

and HT2 is the second signal in advance of the Petersfield Up platform starting signal

HT2 is the second signal in rear of the Petersfield Up platform starting signal.

 

[zwk500]

HT2 is the second signal in rear of the Petersfield Up platform starting signal.

Edited! Apologies for the confusion.

 

[MadMac]

Isn’t this the job that Siemens made an utter bollocks of some years ago?

 

[D6130]

Isn’t this the job that Siemens made an utter bollocks of some years ago?

Yep! (allegedly).

Wasn't it the first double track main line resignalling using axle counters instead of track circuits?

 

[pompeyfan]

Isn’t this the job that Siemens made an utter bollocks of some years ago?

It was back in 2006, was meant to take 3 months but took closer to a year. I think there was also a spate of cable thefts which didn’t help.

Bournemouth ASC was done to a similar spec at about the same time, and seems to be reasonably reliable. I’m not sure why there is such a difference in the end product between the 2 schemes.

 

[DelW]

IIRC: the Farncombe to Petersfield resignalling involves moving control from Farncombe, Haslemere, and Petersfield boxes to Basingstoke ROC and includes several level crossings, new or upgraded signals, and I possibly renewal of the S&C. It also includes swapping shunt signals for main aspects at Haslemere and Petersfield down platforms to allow turnback moves without shunting across. I'm not sure which parts of the work were done in the first blockade a year ago. I also can't remember if it's 1 or 2 panels but think it does have a separate LX workstation. HT1 is straight after Petersfield platform starter, and HT2 is the second signal in rear of the Petersfield Up platform starting signal so would be part of the aspect squences being tested with the changes to the crossover.

Thanks for the info, I think I knew that was due to happen sometime but I hadn't linked it with the imminent blockade.

Isn’t this the job that Siemens made an utter bollocks of some years ago?

Was that the one in the early 2000s when all services in and out of Pompey were reduced to pilotman working, with ensuing problems that went on for several months?

 

[zwk500]

Thanks for the info, I think I knew that was due to happen sometime but I hadn't linked it with the imminent blockade.

NR website says project continues to late 2024 so this probably isn't the transfer of the signalling fringe.
NR's website says this closure is for renewal of the points at Haslemere, along with some other bits.

Engineers will install a new and improved set of switches and crossings in Haslemere, an essential part of the Portsmouth Direct Upgrade that will allow the line speed to be increased in 2024.

More work as part of the F2P project will be carried out along with a large package of general maintenance.

 

[MadMac]

Thanks for the info, I think I knew that was due to happen sometime but I hadn't linked it with the imminent blockade.

Was that the one in the early 2000s when all services in and out of Pompey were reduced to pilotman working, with ensuing problems that went on for several months?

If memory serves, they ended up reopening (after a fashion) the old box at Portsmouth Harbour that had closed in the 60s!

 

[TSG]

Isn’t this the job that Siemens made an utter bollocks of some years ago?

They had an awful lot of help from Railtrack. Railtrack were excessively desperate to bring new players into the signalling market for commercial reasons. That the product was not ready or was wholly unsuitable mattered not, Horsham (EBILock?) never even got commissioned. Manchester South (Ansaldo) wasn't a great success I understand. Bournemouth had enough problems that they should have known not to do Portsmouth, but I believe they were stuck in the contract. Railtrack were also keen to do things there own way rather than the way the supplier suggested and had succeeded with elsewhere (e.g. the Siemens interlocking is designed to talk to a German 3 phase drive and detection point machine. Railtrack insisted on standard British machines using dc motors and single phase detection, which needed a complex interface to be designed. Extra money, extra maintenance, more to go wrong).

Yep! (allegedly).

Wasn't it the first double track main line resignalling using axle counters instead of track circuits?

No, but the pilot at Bournemouth was I think.

It was back in 2006, was meant to take 3 months but took closer to a year. I think there was also a spate of cable thefts which didn’t help.

Bournemouth ASC was done to a similar spec at about the same time, and seems to be reasonably reliable. I’m not sure why there is such a difference in the end product between the 2 schemes.

I think Bournemouth has less cables going shorter distances running a less complex piece of railway (less crossings, less junctions, less stations, less trains etc etc). The type and amount of cable does not help, nor does its sensitivity to cable problems. It's an absolutely bonkers architecture when used outside of a equipment dense yard or station area.

 

[MadMac]

They had an awful lot of help from Railtrack. Railtrack were excessively desperate to bring new players into the signalling market for commercial reasons. That the product was not ready or was wholly unsuitable mattered not, Horsham (EBILock?) never even got commissioned. Manchester South (Ansaldo) wasn't a great success I understand. Bournemouth had enough problems that they should have known not to do Portsmouth, but I believe they were stuck in the contract. Railtrack were also keen to do things there own way rather than the way the supplier suggested and had succeeded with elsewhere (e.g. the Siemens interlocking is designed to talk to a German 3 phase drive and detection point machine. Railtrack insisted on standard British machines using dc motors and single phase detection, which needed a complex interface to be designed. Extra money, extra maintenance, more to go wrong).

No, but the pilot at Bournemouth was I think.

I think Bournemouth has less cables going shorter distances running a less complex piece of railway (less crossings, less junctions, less stations, less trains etc etc). The type and amount of cable does not help, nor does its sensitivity to cable problems. It's an absolutely bonkers architecture when used outside of a equipment dense yard or station area.

Around this time, I was Signal Engineer at LA Metro. Siemens wanted to come round and give us a demo of their "S7" interlocking product. They showed up, and as soon as they heard I was ex-BR, they knew the game was up. I did make a point of mentioning that I had heard that they were working at Portsmouth at the time.....

 

[pompeyfan]

They had an awful lot of help from Railtrack. Railtrack were excessively desperate to bring new players into the signalling market for commercial reasons. That the product was not ready or was wholly unsuitable mattered not, Horsham (EBILock?) never even got commissioned. Manchester South (Ansaldo) wasn't a great success I understand. Bournemouth had enough problems that they should have known not to do Portsmouth, but I believe they were stuck in the contract. Railtrack were also keen to do things there own way rather than the way the supplier suggested and had succeeded with elsewhere (e.g. the Siemens interlocking is designed to talk to a German 3 phase drive and detection point machine. Railtrack insisted on standard British machines using dc motors and single phase detection, which needed a complex interface to be designed. Extra money, extra maintenance, more to go wrong).

No, but the pilot at Bournemouth was I think.

I think Bournemouth has less cables going shorter distances running a less complex piece of railway (less crossings, less junctions, less stations, less trains etc etc). The type and amount of cable does not help, nor does its sensitivity to cable problems. It's an absolutely bonkers architecture when used outside of a equipment dense yard or station area.

I hadn’t considered that outside of Bournemouth station and Branksome, it’s a straight forward 2 track railway over a geographically small area.

Like you say Havant Jn, Havant station, the “Triangle”, and the bi-di from Fratton - Harbour, and then throw on 3 L/X on the top makes it a significantly more complex railway over a much larger area.

Is someone able to confirm if plans are actually afoot to replace Havant box as well as Petersfield - Farncombe?

 

[jfollows]

Manchester South (Ansaldo) wasn't a great success I understand.

It failed completely 16 to 20 September 2020 (https://www.railforums.co.uk/threads/crewe-delays-16-sep.209188/), with the inference that if it hadn't been a unique installation there might have been know-how to fix a failure much more quickly.

 

[zwk500]

Is someone able to confirm if plans are actually afoot to replace Havant box as well as Petersfield - Farncombe?

Other than NR's overall plan to move into the ROCs, I don't believe there's any firm work being done on Havant at the moment. The Wessex Route is likely to want to do Eastleigh and Salisbury before Havant, as they're older Relay interlockings and Eastleigh in particular is running out of capacity in the interlocking while the Southampton area has huge pent-up demand for rail services.

 

[TSG]

Other than NR's overall plan to move into the ROCs, I don't believe there's any firm work being done on Havant at the moment. The Wessex Route is likely to want to do Eastleigh and Salisbury before Havant, as they're older Relay interlockings and Eastleigh in particular is running out of capacity in the interlocking while the Southampton area has huge pent-up demand for rail services.

Forgive me but I don't understand what you mean by 'running out of capacity'. Relay interlocking may reach a condition where you feel it is unsafe to make alterations to increase the capacity of the railway, or you may conclude that replacing it with a CBI is cheaper than making the level of alterations required. However, a relay interlocking is a parallel logic system with pretty much each function having it's own path; you can just keep adding new functions or cutting contacts in, even if you have to add an REB next to the relay room; I don't see how it can run out of capacity in the way that a CBI can i.e. too much logic to process in the time available.

Anyway, whatever Wessex Route's aspirations for Eastleigh and Salisbury may be, they might struggle to meet any demand in the Portsmouth area if they don't get on with doing Havant.

 

[zwk500]

Forgive me but I don't understand what you mean by 'running out of capacity'. Relay interlocking may reach a condition where you feel it is unsafe to make alterations to increase the capacity of the railway, or you may conclude that replacing it with a CBI is cheaper than making the level of alterations required. However, a relay interlocking is a parallel logic system with pretty much each function having it's own path; you can just keep adding new functions or cutting contacts in, even if you have to add an REB next to the relay room; I don't see how it can run out of capacity in the way that a CBI can i.e. too much logic to process in the time available.

Anyway, whatever Wessex Route's aspirations for Eastleigh and Salisbury may be, they might struggle to meet any demand in the Portsmouth area if they don't get on with doing Havant.

I'm not a signalling engineer so I don't know the ins and outs but when I was looking at some projects around Southampton a signalling engineer was very concerned about the capacity of the interlocking to take on much more function.

 

[Annetts key]

Yeah, I don’t understand either. Are there practical constraints with relay interlocking systems? Yes, for example there is a practical limit on the number of relay contacts in a circuit. But there are ways round that (use summation circuits).

So, as @TSG says, in terms of explaining, if the railway have the money, a relay interlocking can cover anything from a small simple junction to a very, very complex area. Indeed, conventional relay interlockings were quicker than BR SSI systems.

I suspect there is some misinformation or misinterpretation here. One of the big disadvantages of a relay interlocking, is that any significant alterations require much more work compared to a computer based system. Hence any such alterations are costly these days (unlike in the past).

So if operations want extra facilities (more points, extra running lines) or to reduce the spacing of the signals (say convert from two aspect to three, or convert from three aspect to four) to increase line capacity, it takes longer to do the work, plus as I say, it is expensive. And with all proposed schemes, the amount of money available is the limiting factor.

 

[MadMac]

Yeah, I don’t understand either. Are there practical constraints with relay interlocking systems? Yes, for example there is a practical limit on the number of relay contacts in a circuit. But there are ways round that (use summation circuits).

So, as @TSG says, in terms of explaining, if the railway have the money, a relay interlocking can cover anything from a small simple junction to a very, very complex area. Indeed, conventional relay interlockings were quicker than BR SSI systems.

I suspect there is some misinformation or misinterpretation here. One of the big disadvantages of a relay interlocking, is that any significant alterations require much more work compared to a computer based system. Hence any such alterations are costly these days (unlike in the past).

So if operations want extra facilities (more points, extra running lines) or to reduce the spacing of the signals (say convert from two aspect to three, or convert from three aspect to four) to increase line capacity, it takes longer to do the work, plus as I say, it is expensive. And with all proposed schemes, the amount of money available is the limiting factor.

Recently retired after nearly 44 years Signal Engineer here.

There comes a point with alterations to a relay interlocking where you start to think “Wouldn’t we be better just renewing it with something more modern?”. There are also other factors such as material availability/support, does the design expertise still exist, space availability. Each case needs to be looked at on its own merits.

 

[zwk500]

I suspect there is some misinformation or misinterpretation here. One of the big disadvantages of a relay interlocking, is that any significant alterations require much more work compared to a computer based system. Hence any such alterations are costly these days (unlike in the past).

Misinterpretation is by far the most likely.