Crossrail: Breaking Down the Crisis

On Thursday 26th July 2018 a meeting took place at City Hall. The subject was Crossrail – the Elizabeth line – and whether it would open on time. Up front were Simon Wright, Crossrail’s then-CEO, and Sir Terry Morgan, its Chairman. To the assembled senior officials from TfL and City Hall, their presentation would have left little room for doubt: Crossrail would not be ready to open in December.

For some in the room, this may well have come as something of a surprise. Present were Mike Brown, Commissioner of Transport for London and head of TfL (in effect, Crossrail’s lead sponsor), Mark Wild (then of TfL, now Crossrail CEO) and a selection of senior officials from City Hall – David Bellamy (Chief of Staff), Heidi Alexander (the Deputy Mayor for Transport) and, most importantly, Sadiq Khan, Mayor of London.

The presentation by Wright and Morgan laid the facts bare. Siemens (the signalling contractor on the central section and subcontractor on the train) and Bombardier (TfL’s direct contractor for the train), should have been able to begin testing on the central section by now – that is, test the trains and signalling together in increasingly complex ways to prove it would all work in passenger service. This required the ‘routeway’ – in essence the bits that Crossrail had to deliver – to be finished by now. Despite previous assurances from Crossrail, however, it was not.

The Crossrail presentation was bleak. It offered three different scenarios for when testing was likely to start. There was a 10% chance that it would begin in November, pushing trial running and actual Elizabeth line operations back to somewhere between February and May. This assumed that early testing (and software iteration) by Bombardier and Siemens was possible and went to plan.

If this didn’t happen, there was a still a 50% chance that with some parallel working the Elizabeth line could open between April and June. What was most likely though, if work on the routeway continued to be delayed,was that the line could open between May and August 2019, although that would depend on how reliable the trains and signalling proved to be.

Slide 6, presentation to Mayor by Crossrail, 26th July 2018

None of this was likely to be news to Heidi Alexander and David Bellamy. A week earlier they had been briefed by TfL on the state of the project as known – and this briefing had included almost exactly the same slides.

Slide 2, presentation to Alexander and Bellamy, 20th July 2018

Similarly, this also wasn’t news to Mike Brown. Regardless of whether TfL had been made aware of Crossrail’s build issues by the company themselves before that point, the Transport Commissioner had been tipped off to the fact that there were launch problems several months before. According to LR sources, Siemens had written to him directly in early 2018. In that letter they had warned that, from their perspective, the train programme (which they were heavily involved in signals integration for) was seriously behind schedule. That a delay of 18 months to the trains programme existed was also alluded to by Terry Morgan at the beginning of December on Radio 4’s PM programme. Siemens had warned Mike Brown that even at that stage, as things stood, it was unlikely that they would be able to commission the signalling and pass the trains through testing in time for a December opening.

Even this shouldn’t have been news to him. Crossrail themselves had issued a notice to TfL and the DfT (Crossrail’s other sponsor) as early as July 2017 warning about the challenge of software development for the train and meeting key dates.

Indeed the only person in the room for whom all of this may well have been unknown territory was the Mayor himself.

Up to this point, the Mayor had rarely interacted directly with Crossrail themselves as, rightly, it was through TfL (as a project sponsor) that he received the majority of project updates. It is therefore possible that the status of the routeway and build problems had to this point been entirely unknown to him (although the question remains as to whether it was unknown to Crossrail’s senior management themselves). It is possible, however, that he may have been aware that there were looming issues with the trains and signalling, as this was directly within the remit of TfL.

Either way, this was likely the first point at which the Mayor became fully aware that the Elizabeth line could not open in its planned configuration in December 2018.

This left one question to be answered – if it couldn’t open as planned, then could at least part of it open instead?

Unfortunately, the next part of the Crossrail presentation answered this question pretty definitively:

“2 options for partial or sectional opening in December have been analysed.” The presentation stated.“ They are judged as not feasible.”

Slide 7, presentation to Mayor by Crossrail, 26th July 2018

On the 26th July 2018, to those who were collectively charged with its construction, operation and opening, one thing was now unavoidably clear: The Elizabeth line would not open on time.

A month later in August, after the necessary mechanisms of accountability and market notification had ground through, the Mayor was finally in a position to announce the delay to both the public at large and the London Assembly. Indeed in front of the Assembly his anger was palpable as he insisted he had been unaware as to the true state of the engineering crisis that had been developing ostensibly on his watch.

Questions about just what he had known, when continued – particularly from the London Assembly who were (and clearly still feel) aggrieved that they were not informed, even unofficially, before the news was publicly announced.

This ultimately culminated in a heated confrontation at Mayor’s Question Time in October 2018 between AM Keith Prince and the Mayor, when Prince asked the Mayor if ‘the buck stops with you?’

“I’m the Chair of the TfL Board, yeah.” Khan responded.

Prince questioned whether that mean Khan had known about the delayed opening earlier than he had previously let on – suggesting the Crossrail Board meeting on 19th July as a possible reason for this.

“Towards the end of July there was concern expressed by the TfL Board – and others – about the pressures and challenge that Crossrail were talking about. Crossrail still thought they could meet the opening of the central section by December 2018.” The Mayor replied, tersely. Before explaining that towards the end of July some concern had been expressed, triggering TfL to commission a major review. It was that review, he insisted, that represented the point at which Heidi Alexander and the Mayor himself learned that the opening would definitely be delayed until August 2019 at the earliest.

“So are you saying then Mr Mayor that you were aware that there could be a delay earlier than the date you initially claimed you were aware?” Prince pushed.

“I met with – put aside the TfL Board for a second – I met with Crossrail Ltd. people.” The Mayor continued. “That includes Terry Morgan, and Heidi may have been there, at the end of July, where they talked about challenges and pressures.”

“As is often the case, there are scenarios.” He said. “What was not told to us then was that there would be a delay until August 2019”

Again, the Mayor insisted it was only after the TfL review that he became aware that a delay would be likely. Prince pushed again.

“The question is, where you aware that there was the potential for any kind of delay prior to that 29th [August] date?”

“I think it would be fair to say one of the scenarios – and there are often different scenarios – probably was delay. But nobody talked about… Crossrail Ltd were not saying that scenario was going to happen, and Crossrail Ltd were not saying that that scenario would lead to a delay from December until August.”

Prince pushed again.

“So you were not aware of any potential delay, at all?”

“One of the scenarios given was probably some delay.” Khan answered. “But that wasn’t the scenario that Crossrail envisioned would happen… Crossrail Ltd. were confident, and have been confident throughout, that the central section of Crossrail will open in December 2018. And the only time that has moved back to August 2019 is in the late August meeting I referred to.”

Prince reiterated that he wasn’t challenging the Mayor’s word that this was the first time he knew that there would be a nine month delay, but again he asked:

“[Were] you aware that there was the potential for a delay, but didn’t know how long it would be?”

“It’s even further than that.” Khan answered tersely. “Crossrail Ltd. weren’t saying there was going to be a delay, even in July. What I’m saying to you is that one of the scenarios is always – I’ve not got the document here – but would have been that there’s the possibility of there being a delay.”

“That’s not what Crossrail Ltd. were saying.” He stressed.

Even using the most charitable of interpretations, given the slides we have available for this meeting – which include the 10% chance, in bold red text, for an opening in February and the clear statement that no December opening option is available, it is almost impossible to reconcile the Mayor’s account with the factual documents that exist for that meeting.

In the course of five minutes of very specific questioning the Mayor had, at the very least ‘in spirit’, lied multiple times to the London Assembly.

In doing so he created a political crisis, which continues to unfold, of the same scale as the engineering one that City Hall, TfL and Crossrail now collectively face.

Breaking down the blame game

At this stage, it is important not to speculate too heavily about where the problems with Crossrail’s delivery lie. TfL’s own final review into the project, undertaken by KPMG (the one referenced by the Mayor) has yet to be publicly released, although they have promised they will do so.

We already know, however, that its conclusions have lead to an emergency funding package and agreement between the DfT and City Hall that will cover an estimated £1.6bn – £2bn project overrun. Most of this London will need to pay back, drawing on the existing Business Subsidy (money that it was hoped would, within that timescale, be available to Crossrail 2) as well as other financial commitments from City Hall.

We also know that this that has questioned whether even an autumn opening is optimistic, although critically we do not know how much of this is down to issues with building works or problems with the train. There is also a National Audit Office review of the project underway.

What we can say for certain, based off information already released and other sources, is that three interrelated crises have resulted in the situation London finds itself in today:

  1. Crossrail Ltd. did not complete their element of tunnel fit-out or station construction on time.
  2. The trains, contracted for by TfL from Bombardier, still require extensive testing and software updates to work with the signalling in places and may not have been capable of running an effective service even had Crossrail Ltd. completed work on time.
  3. The Mayor – and senior officials (current and former) at both TfL and Crossrail – have created a political blame crisis that, unchecked, threatens to impact on ongoing efforts to solve the first two problems.

The issues at Crossrail

It is worth reiterating that it is only with the release of reviews to come that the full scale of the issues at Crossrail – along with the causes – become clear.

Already, however, there is sufficient evidence to be able to build up a broad picture of what went wrong. On a fundamental level, this is that multiple strands of construction and fitout work, across the project became individually delayed. Each of these delays in themselves may have been recoverable, but cumulatively they eventually formed a package of recovery work that simply wasn’t practical – or logistically achievable – within either the existing project deadline or the risk period.

The key ‘tipping point’ in the chain of events leading to this situation appears to have been the November 2017 ‘explosion’ of a voltage transformer at Pudding Mill Lane substation, caused by substandard electrical work by a contractor. We covered this incident in detail in December 2017 (you can read the full piece here ), but the headline effect was that it seems to have caused a cascade of work delays and issues – including, critically, to the energisation of traction power in the central section and the start of testing.

Identifying the issues (and again, there are likely to be more that emerge) is only half the question, however. Just as importantly, what needs to be identified is why the sheer scale of the issues remained unidentified (or undisclosed) for so long.

The first clue to that can be found in another document released as part of a Mayoral Crossrail document dump on 10th December 2018. This is the Jacobs Report, an assessment of the state of Crossrail as a project after Crossrail Ltd. first revealed they were likely to run over budget in May 2018. Jacobs had also been appointed by the DfT and TfL as the Project Representative at the start of the programme in 2009 to undertake independent assurance (a role they continue to undertake).

The existence of the Jacobs report is something that provokes its own, related, immediate question – why did an independent review in May not reveal problems on such a grand scale that became so obvious to all just three months later?

The answer to that question gives us our first real clue as to what went wrong. In its terms and assumptions, the Jacobs Report makes it clear that it is a very surface-level review, due to the limited time available. As a result, much of its conclusions – broadly, that Crossrail is fine if a little over budget – have been made on a combination of substantive and subjective evidence.

That subjective evidence is based on the level of project completion that Jacobs were able to determine from questioning the senior Crossrail project executives in detail.

Handily, the report even includes a pie chart, indicating how much of the report is based on subjective opinions Crossrail’s senior leadership held about the projects progress, rather than physical evidence.

Subjective vs Supported evidence balance, Jacobs Report p26

With the benefit of hindsight, the above pie chart makes for interesting reading. It shows that almost half of the evidence Jacobs had to rely on was the informed opinion of those running the Crossrail Project. This isn’t entirely unusual for a review of this nature, but it also suggests that Crossrail fell foul of a problem that seems to repeatedly plague the railways – most recently with regards to the May Timetable rollout – the ‘Thermocline of Truth’

Back in September, after the initial delay was announced, we questioned whether this might prove to have been an issue. You can read our full piece here, but a basic description of the Thermocline of Truth from that piece is included below:

It is the principle that bad news tends to accrue at a lower management level, because no one wants to be the person who moves a project risk marker from ‘yellow’ to ‘red’ on a RAG chart.

As a result, pessimism and a belief that the project will overrun ‘bubbles up’ to a certain decision-making level but never beyond, as if hitting the thermal layer that exists in the ocean.

Eventually, the issues reach critical mass and force their way through, leaving senior management wonder why everything ‘suddenly’ went wrong, when in fact the signs that the project was troubled existed at a lower level for some time.

In the last few weeks, additional information about the current problems has emerged which suggests that not only was the above a problem, but that it was aggravated by a further issue as well – many of Crossrail’s key work packages were functioning as silos. So it wasn’t just that information wasn’t filtering up, it wasn’t filtering sideways either. This meant that where contractors were slipping, they looked to available contingency in the future as a way to get back on track so reported that they could still deliver. What they were failing to consider – and what no one in the centre of the project seems to have been set up to spot until it was too late – was that contractors were essentially ‘double-booking’ the same blocks of contingency.

Some support for this state of affairs can be seen in comments by Mike Brown to TfL’s Programme and Investment Committee this week, where it was admitted that only now was Crossrail really adopting an integrated project approach.

Where all this leaves Crossrail now – and in terms of going forward – is something we will tackle in detail in a future piece. Broadly though, the challenge for Mark Wild (as new CEO) and Crossrail Ltd. themselves will first be establishing what state of play the ‘critical path’ really is in. That will then need unpicking into something that is genuinely deliverable, with a degree of confidence – something that the KPMG summaries suggest is still unknown.

One positive, however, did emerge at the same meeting. Wild asserted that he was confident enough to announce that testing would begin on 14th January 2019. Testing will run five days in every seven, although just how long that will take to reach a point where a deliverable service exists remains to be seen.

The issues at TfL

Whereas the issues at Crossrail are becoming increasingly clear, perhaps the biggest issue with the project from a TfL perspective right now is that the exact opposite situation seems to exist: little information seems to be emerging as to what the scale of problems are with the train.
In this regard, the absence of information is as concerning as its presence. Some facts, however, are clear.

The independent warning of problems with the train programme, by Siemens, stands as a serious marker that things haven’t progressed as planned. The issuance of a sponsor’s notice by Crossrail on the state of the Bombardier train software does the same.

This is important – and largely separate to any issues that Crossrail themselves are having. Traditionally, people tend to think of signalling as something that happens at trackside and interacts with the train, but this isn’t how the most modern signalling – such as that which will operate the Elizabeth line – works. Modern signalling functionality is as much on the train as off, heavily interwoven into the train’s software. The rollout of new software on a line such as this is as big a systems integration project on the train as it is at trackside.

In this instance, the warnings received don’t necessarily suggest that something has gone wrong on a management information level to the same degree as on the ‘build side’ of the project. But they are – and should have been – a cause for concern in their own right.

There is physical evidence as well, however, that the train and signalling integration is still a work in progress, even in places where Crossrail Ltd’s own failings aren’t a factor. Most notably, on the service out to Heathrow via the Heathrow tunnels.

Back in February we looked in detail at the (then) current issues trying to get the Bombardier rolling stock working in the Heathrow tunnels (you can read the full article here). This was meant to have been completed by May 2018 when TfL took over the operation of various services to Heathrow.

Indeed the new TfL Business Plan makes a big thing of pointing out that this is one of the few ‘green ticks’ on the current Crossrail schedule. However, in doing so they are committing a sin of omission: because, at time of writing, it is still our understanding that not a single Bombardier 345 has been successfully tested in the Heathrow Tunnel.

Elizabeth line opening plan diagram, TfL Business Plan

TfL’s green tick is not incorrect, but it is profoundly misleading. It exists because of the ingenuity and forward planning of the Elizabeth line’s Operations Director Howard Smith, who foresaw this issue and created a backstop solution involving leased rolling stock. Not because the underlying issue with the track-to-train setup, and onboard software on the Bombardier 345s, is delivering to spec.

All of this paints a picture which suggests that significant issues remain to be solved with the Bombardier rolling stock and how it interfaces with signalling, both at track and train level. It is perfectly understandable and reasonable for TfL to blame some of this on Crossrail Ltd. And to a certain extent, whether the Bombardier rolling stock would have been ready to run in service in December had Crossrail Ltd. delivered (thus allowing TfL the testing time required) is entirely academic as they were not given the opportunity to prove that either way.

What isn’t academic, however, is that there have been – and still are – fundamental issues to solve with the trains before they can enter service. And, if Wild’s assertion that testing can start in January is true, then at that point any such issues which remain will be as large a problem for the delivery of the project as anything Crossrail Ltd. have done so far.

Bluntly, TfL’s current attitude that it is ‘all about the build’ is something that they themselves should be wary of, as should the Assembly, the Mayor and the wider London public. It risks missing problems – or the very least the opportunity to accept and own them – with the train that could have been diagnosed much earlier. In effect, they may be replicating the mistake that Crossrail themselves have clearly made with the build.

From a TfL organisational perspective, it should also raise a further red flag. For train issues are the one thing that TfL cannot claim to have lacked prior knowledge of. Given the warnings issued to Mike Brown directly and TfL much earlier in the project, it seems likely that the Mayor would regard any lack of communication over issues here with the same level of anger aimed at Crossrail Ltd. before. What has already cost Sir Terry Morgan a job could easily do the same for the Transport Commissioner, should the Mayor feel – correctly or incorrectly – that he has been misled.

The issues at City Hall

Whilst the clear issues at Crossrail and the (not-so-clear) issues at TfL relate to the quality of work being delivered by their respective contractors, the issues at City Hall are firmly of the Mayor’s own making.

Their origins are laid out clearly at the head of this article, and the consequences are increasingly obvious. The Mayor, and his team, seem to have reacted to the unfolding engineering crisis at Crossrail not by asserting ownership of it – contrary to the Mayor’s own words, at times, about his role as ostensible head of TfL – but by attempting to divest himself of any blame.

It’s a very ‘Westminster’ reaction, and one that harks back to Khan’s arrival in office and his reaction to criticism for the ill-advised Fare Freeze promise. This was, at least in part, caused by a mix-up between TfL and Khan, as candidate, over what was genuinely practical when it came to costs. Khan’s reaction to that crisis was to lean further into his argument that it was possible and distribute the blame for any mix up elsewhere.

Here too, a similar state of affairs seems to be unfolding. In August, the Mayor insisted that he had been as in the dark as everyone else. Despite increasing evidence to the contrary, he continued to lean into this strategy.

The net result has been the creation of an entirely unnecessary political crisis. City Hall’s efforts to defend and deflect from that crisis increasingly risk causing real damage to the very thing the Mayor needs to happen the most: the completion and opening of the Elizabeth line.

This is evidenced by the concerted effort in the media this week to confuse and obfuscate over what the Mayor knew, when. The Mayoral document dump on the 10th of December ‘helpfully’ summarises all the ways the dates and documents it includes confirm the Mayor’s interpretation of events. Read the actual documents included for the events last July, however, and they clearly contradict that narrative.

For the Mayor not to have known that Crossrail was delayed after his meeting with TfL on the 26th of July the following premises must be accepted:

  1. The combined heads of Crossrail Ltd. (its CEO and Chairman) must have failed to point out that the project was now delayed in any way – something the released slide deck makes clear they did not.
  2. The Mayor must have failed to understand the clearly marked diagrams that deck contained, which show both dates of testing, percentage-based chances of meeting those targets and text clearly stating that there were no options for delivery in December.
  3. Everyone else in the room who didn’t work for Crossrail – including the Transport Commissioner, his Deputy Mayor for Transport and the man he has now appointed to lead Crossrail – must have failed to have comment on, or explain to him, what this all meant.

It’s a combination of events that is laughably unlikely. The idea that the Mayor left that meeting without understanding that Crossrail would not open in December fails the most basic principle of the ‘duck test’ – if it looks like a duck, swims like a duck and quacks like a duck then it is probably a duck.

To use a cinematic metaphor, it would be as if Jack, whilst clinging to Rose’s flotsam raft, kept insisting to her that there was still technically a chance the Titanic could make it to New York on time.

The denial of this knowledge is remarkable precisely because it was so unnecessary. It was entirely appropriate for the Mayor to know in July that a delay was coming, and it was entirely appropriate for him not to reveal that same knowledge to others until August. By denying his own role in that process, however, the Mayor has created an environment in which the politics – and the organisational efforts necessary within TfL and a much-leaner Crossrail Ltd. necessary to deal with them – are quickly spiralling out of control.

The chances of any crisis – political or engineering – damaging Khan’s chances of re-election were always small. At worst, they may provide a brief opportunity for Shaun Bailey to remind the world that politically he still exists. Instead the Mayor’s actions threaten to create an environment in which new Crossrail Ltd. CEO Mark Wild will spend as much time answering questions about who knew what, when as he does trying to oversee completion, testing and opening of the railway. They also leave TfL in a tricky position – to take a bunker mentality themselves over any issues with the train and remain silent, or risk being on the receiving end of the Mayor’s wrath themselves.

Finding a way forward

Emerging from the triple crisis that is affecting the Crossrail Project is going to require significant effort on the part of all those involved. There is still no question that Crossrail will be delivered. There is also still no question that, in time, the project will be regarded as a significant success and a vital addition to London.

What is critical now, however, is that Crossrail, TfL and City Hall all accept their own element of blame and move on and deliver.

All three would do well to note the behaviour – at least in public – of the project’s other sponsor, the DfT. Rather than get involved in the blame game, the Department have focused on putting in place the mechanisms – including the financial deals – necessary to put the project back on track.

In addition, for Crossrail this means focusing on pulling apart the tangle of clashing deliverables and getting them out the door. For TfL, it means focusing on making sure that when that work is complete, the train is ready to go and completes testing as quickly as possible. Already, information has begun to emerge as to how both organisations are trying to do that and we will investigate those next steps in a future article.

For the Mayor though, it means accepting that he has embroiled himself in a political crisis of his own making, and that his current strategy for handling both Crossrail and TfL risks hindering the work of both. He has made few genuine, major missteps so far in his time in office, but this is one.

How he handles this crisis now and how he owns his own highly misleading, if not outright untruthful, comments may not have a material impact on his chances of getting re-elected, but they will have an enormous impact on just how quickly the Elizabeth line opens.

And in London, a Mayor’s legacy isn’t formed from what they knew, when. It is drawn in the colours that they spread across Harry Beck’s map.

Like transport and a good natter? Join us for drinks tonight near London Bridge! LR is also a community-funded website. You can also help us continue to write by backing us on Patreon or visiting our shop.

148 comments

  1. I don’t see that the delay of the project has anything to do with Mayor at all. I don’t understand the emphasis placed on the aspect of a one month delay in the Mayor being informed or informing others as of the slightest significance.
    It’s a computer project that is not being completed on time. All signalling projects nowadays are computer projects. These are always late and often fail.

  2. I don’t see that the delay of the project has anything to do with Mayor at all. I don’t understand the emphasis placed on the aspect of a one month delay in the Mayor being informed or informing others as of the slightest significance.

    Which is my point about the political side of this.

    It was, and is, utterly unnecessary. Yet the real risk now – which is very much visible on the ground – is a growing blame culture within Crossrail and loss of transparency within TfL about what is happening with the train.

    You can’t separate politics form engineering or technology on a large project like this. We all wish that they happened in pure, logical isolation but they don’t. The politics will now affect the delivery and needs to be considered in the same way as the actions of both TfL and Crossrail Ltd, because at a senior level it influences both.

  3. Re Howard,

    TfL has tradeable debt so all statements concerning its financial position have to be handled in the same way as FTSE/AIM listed PLCs market sensitive information. Hence the Mayor has to be very careful not to do an “Elon Musk ” and mislead the markets. His knowledge and timing of acquiring that knowledge are therefore key to whether the he is guilty of misleading the markets which can come with stiff penalties.

    Given TfL is very near its borrowing limits any additional borrowing over the summer is highly likely to have been mispriced (in interest rates terms) due to statements about Crossrail being more positive then they should have been…

  4. It’s very likely with the untold problems of 345 interacting with so many stuff on the line that the delivery time will slip further and further.

  5. Thanks for this.

    Flippantly, I’m sure there weren’t these sort of problems when new lines were built for steam trains with semaphore signalling, but bringing a new line with new trains and new signalling all at the same time does seem to be beyond the capability of someone to manage properly. Separately the projects were probably deliverable, but the integration appears to have failed, so far.

    With so much reliant now on software development, a subject I am all too familiar with, I’m not sure that any publicised schedule is achievable.

  6. It’s worth noting that the Chief Executive of Crossrail left in May 2018 to go to BAe Systems. The Head of the Joint Sponsor Team left some time before to go onto Crossrail 2.

    Maybe too many key people have left before the project is operational?

  7. Maybe too many key people have left before the project is operational?

    This is something Pedantic will pick up in his piece on the technical delivery side of things. Short answer is: yes, and Mark Wild says the same.

  8. Thanks for the comprehensive update.

    Surely some of the contractors, especially Bombardier, should be on the hook for not delivering their trains on time? Obviously the fact that the central core isn’t ready is a bit mitigation for them, but it’s very clear that the trains are the really big problem

  9. Timing is everything. If Crossrail is opened before the next election, 2020, then the blame game will be a side show. What is required now is delivery. The blame game will remain a distraction but in the long term nothing more.

  10. TINITD: the CEO seems to have changed too, there’s a throwaway line in Para. 2 – was there an announcement?

  11. @AlisonW Depressingly, from experience I agree with you about delivery of software integration projects. As ever binary milestones are needed (done or not done) not the “80% complete”

    Looks like we’re now in stage 4 of the project lifecycle – “Search for the guilty”
    Next up stage 5 – “Persecution of the innocent”
    As JB says “a Mayor’s legacy isn’t formed from what they knew, when. It is drawn in the colours that they spread across Harry Beck’s map” so all set for stage 6 – Praise and honour for the nonparticipants

    http://wikibon.org/blog/this-project-management-joke-is-often-reality/

  12. It must also be significant the the class 345 units are still not operating other than a few services between Liverpool St and Shenfield which must point the finger towards Bombardier rather than Siemens.

  13. In my view the Quality Assurance people are not doing their jobs properly – here it is Jacobs and on Thameslink it was Chris Green and support team. Both appear to be taking management assurances at face value.
    Overall are the people at the top not ensuring quality in those undertaking Quality Assurance?
    The Quality Assessors in my view should be taking a project wide view and identifying where contingency time is “double booked”.
    Or is hindsight a wonderful tool?

  14. Re Mike G and Herned

    But Bombardier have delivered the trains pretty much on schedule and they work with the existing specified surface signalling AWS, TPWS (+) and the future ETCS (in open air e.g. Old Dalby Test track).
    The signalling issues lie with them not working with the Siemens CBTC in the Core tunnels (the swap over to/from CBTC especially on the move at Westbourne Park was always going to be technically challenging and ETCS in the Heathrow tunnels , the latter due to using the same frequency as GW-ATP and GW-ATP not being EMC friendly and even more so in the confined spaces of a tunnel. (In signalling terms GW-ATP is the great uncle of the ETCS train /track comms interface and the risks using related technology weren’t understood or may be even recognised)

    With ETCS all other signalling systems are dealt with by fitting modules supplied by signalling equipment manufacturers on the trains and the ETCS computer (EVC) interfacing with all the modules. This means the traditional train / track interface which was also the signalling equipment interface has now moved to with the on train wiring and software. Hence the trains are a massive systems integration project with lots of signalling equipment now on train from Siemens and Mors Smit amongst others that means the train manufacturer is not fully in control but reliant on external partners sometimes not of their choosing and with whom there is no direct contractual relationship like a normal supplier e.g. Siemens who are Crossrail responsibility not Bombardiers.

    Hence Crossrails’s failure to realised their role as systems integrator rather than a Super PMO is key.

  15. Interesting article, but nothing new really, just a case of rummaging around in the various bin bags dumped in the wake of the Crossrail crisis, before finding a pair of trousers that fits 🙂

    What we all know is Crossrail was miserably mismanaged and its progenitors came up with a morass of lies claiming the line was on time, budget, and would be opening in December 2018.

    I wrote a year ago that Crossrail didnt look like it would open this month! At least I am happy its workers, managers, and bosses got their massively inflated wages and bonuses and they can fart about safe in the knowledge their jobs are secure until 2020. Crossrail as someone pointed out, is again getting more money far in excess of the amount of cuts in benefits and other things screwing a lot of people in this country, forcing people to take massive wages cuts and making many of us destitute – myself included.

    By now it should be realised any one of us who gets these reassurances from railway managers or companies of any sort should know their words are empty rhetoric. This is the new future of our railways. Its basically ‘we’ve never had it so good’ even when the evidence points to the entire experience being hugely problematic. Its a new consciousness this country’s acquiring and its one which ignores the realities because overall its all about shoving numerous people into dead end sidings pretending their experiences do not count. Its extremely dangerous.

  16. Would be interesting to compare the signalling experience with Thameslink, where the signalling contractor (also Siemens) have installed an ETCS Level 2 + ATO overlay system. From my limited perspective as a passenger, it seems to work very well.

    Key difference here is that the Class 700s were also Siemens built – could this have made a difference?

    Same signalling contractor but installing a different signalling system and interfacing with a different rolling stock manufacturer. I know Thameslink has not been at all straightforward, but I am not aware of the signalling presenting issues as big as we are seeing with Crossrail.

  17. @ NGH

    Thanks for the detail, however the article refers repeatedly to delays with Bombardier’s on-board software, which they are also struggling with for the Goblin trains. Clearly the blame can be spread quite widely

    As a wider point, my own view is that the complete failure to manage expectations which is the worst thing about this situation. Why they didn’t use the obvious opportunity of the transformer explosion to announce a delay is beyond me

  18. No mention here but perhaps will be in the technical article is the challenge of stopping the trains in the correct place to line up with the PEDs. This is particularly challenging with the high braking rates on the approach to meet the TPH targets. Analysis of stopping accuracy on the Victoria Line showed that the stopping positions depend on both the TPH & the number of hours operated. Hopefully the stopping position is part of the non-safety critical software which makes minor adjustments relatively easy. If it is built into the Safety Software as in some of the ex-UK Siemens products the need to go through a V&V cycle for each change makes adjustments quite slow.

    Misalignment at stopping and the requirement to make ‘small’ forward or reverse movements will wreck the TPH. Reverse movements are normally more difficult due to the impact on the safety of the signalling system.

  19. Thank you for another well written and organised article – even if its’ contents are miserably sad and say a lot about our political system, management ethics and QA quality. Having spent 40 years in IT of various shapes and sizes the complexity of this project is mind-boggling and to an ill-informed outsider has always looked like the biggest risk. Perhaps all the separate bits work on their own but integration is the key in this sort of effort and this appears to be dangerously absent.

  20. Anonymous @ 12:52

    “Would be interesting to compare the signalling experience with Thameslink, where the signalling contractor (also Siemens) have installed an ETCS Level 2 + ATO overlay system. From my limited perspective as a passenger, it seems to work very well”.

    That would be because it isn’t yet operational and the most reliable interface is the one you haven’t got: ATO isn’t considered essential until the timetable steps up to 20tph hence isn’t currently in use, although some testing has been carried out.

  21. An interesting article as always. There are a few worrying observations in the article and in the released papers. The apparent lack of evidenced knowledge about physical progress is one and the other is the integration point. I struggle to understand how so little seems to be known about what is going on and I’m almost bamboozled by the failure to properly understand the integration point. If the assurance process has been properly planned then this should have made the need for really effective systems integration obvious.

    Would I be unfair in suggesting that Bombardier’s struggles with both class 345 reliability and signalling system integration have placed enormous pressures on its software development resources? And even more unfair to suggest that the progress (or lack thereof) with the class 710 Aventras is related to Bombardier’s need to get the class 345 sorted? In short there’s only so much software development resource available and it’s been prioritised for Crossrail?

    As for the political fall out well we’ll see how the Mayor and TfL cope with that. I expect the Transport Cttee meeting on 21 December is going to be a fascinating watch. I remain gobsmacked that TfL and City Hall were apparently so unaware given Mr Wild and two other TfL Board members were/are non execs on the Crossrail Board. They’ve been there for a fair while now – were they not allowed to report back to their sponsor (TfL)? I assume the DfT’s nominees on the Crossrail board did report back.

  22. Re Anon @1252,

    Possibly but Siemens had massive software issues with the 700s and they needed wheels turning experience (which the 345s haven’t really be able to do yet with certain systems) to help find and solve those issues. With the 345s the ETCS L2 works outside the Heathrow tunnels. The 700s didn’t have to content with getting ETCS working with ATP in the background or CBTC so 2 extra big hurdles to clear.

    The ETCS equipment on the comms side is the same in both the 345s and 700s but on the 345s it can’t cope with GW-ATP using the same frequencies and at much higher power level (and from pre EMC reg. era).

    It took about 18 moths to get the 700s software settled and the 345s will probably be similar for the signalling issues.

  23. Re John M,

    And Knorr Bremse and been having a fair few issues with the PEDS too.

    This is just the first article…

  24. I get the issues with the signalling and trains, at least I can understand the kind of challenges they faced and how they could lead to delays. What I don’t really understand is why the completion of the stations in the central section is so far behind schedule.

    And as an aside it would be nice if they could get round to finishing the work at Harold Wood – it’s been a building site for about two years and there’s been no activity for months.

  25. Some typos

    Prince questioned whether that mean(t) Khan had known

    “The question is, w/h/ere you aware

    We also know that this /that/ has

    It was very noticeable that all Khan’s media interviews had the backdrop looking down on a fleet of idle 345s in West London, after Sir Terry’s train comments.

    On testing given only one 345 has run in tunnels, I wonder why Mr Wild has so definitively committed to the middle of January. We’ll see in 4 weeks which would make it not so far off track in global terms. The only change has been the securing of funding to commence testing.

    @Andrew Adams – those are termed TfL’s unfunded On Network Works. Same story at all legacy stations. Platform facilities are considered nice to have and not critical to operations despite being ‘promised’ three years ago. TfL will continue deferring until their budget has a surplus for investment. I have concerns about the reduction in operating capacity.

  26. @AlisonW
    With steam trains and semaphore signalling projects that didn’t work on time, trains crashed and people died. Norwich and, to an extent, Abermule as examples.

  27. Another problem arising from the incompatibility is there’s no easy way to globally switch off all the GW ATP equipment in the Heathrow tunnels in order to test the new trains on ETCS overnight in the wee small hours. You’d have to go through the tunnels accessing distributed equipment housings to manual remove many individual fuses to power down, then meticulously go round again after the very short night’s work to restore ATP to full operation again in each cabinet before commencement of the next days services. Try repeatedly doing that every night for months of testing and there’s bound to be the odd fuse left out or module that doesn’t power up again properly and then normal service could be disrupted, quite apart from the inefficiencies of very short shifts in the no service window and the time taken to prepare and restore. I’m not even sure any useful extent of dis- and re-connections would be feasible in the time available. Complicating matters further, the existing Heathrow Express units (cl.332) cannot be modified for ETCS. They do not even have TPWS capability because of their limited operating range and the presence of GW-ATP throughout that route, so a temporary fitment of enhanced TPWS track equipment in the tunnels to replace ATP does not help. Class 387s can’t take over HEX early because they won’t have GW-ATP and it would not be practical to equip them with that obsolete system for such a short time, as is the case with the much larger 345 fleet. Therefore the most compelling option is a ‘big bang’ approach where the infrastructure is switched over to ETCS at the same time the new express rolling stock is introduced to service. Any time after that, following suitable testing, the new Elizabeth cl.345 service can also be introduced to Heathrow.

    I recall Thameslink integration specialists stating they had, after all, managed to achieve the kind of stopping accuracy required for PEDs using their newly developed ETCS-based ATO, but this ability was an unknown at the time Crossrail decided upon their core signalling technology, so with an eye to future capacity as well, stated to be only achievable using a moving block system, they sensibly played safe and opted for a proven, off-the-shelf metro CBTC product known as Trainguard MT, originally developed by Matra Transport International and more recently supplied by Siemens Mobility.

  28. @Marek

    I agree. There was the proverbial brown stuff flying about the JLE (think senior LT people getting grilled by the DCMS Select Committee) in 1999 but that’s forgotten now.

  29. The other consideration is that Bombardier are presumably still integrating the (3rd !) ATO system with S Stock. Little has been said about the problems with the latest system except that I was told that the first section to be commissioned earlier in the year had to be rolled back. There is a completion date ( 2023 ?) in the latest TfL business plan .

  30. What few seem to have realised in all the furore is that the £15.9bn budget being quoted is in the original 2007 prices. I’m not aware of any other infrastructure project where the prices have not been updated in 11 years.

  31. @ Mark T – while I understand the system and rolling stock issues you outline re the Heathrow tunnel I remain completely bemused as to why Crossrail were reporting that if Version 7.3 of the signalling software was delivered by Bombardier that they felt that full 345 service to Heathrow could be achieved in October 2018. This is clearly shown on the weekly update reports produced by Crossrail. The date has now slipped to Feb 2019 but why the insistence that there is a fix to allow 345s to run into the tunnel that contradicts what you and NGH have set out repeatedly on this blog? I’m no signalling person but is there a “magic” solution whereby the 345s can work with ETCS and “ignore” the old ATP system’s interference? That’s the only thing I can think of that would allow the apparent Crossrail strategy to proceed with removes the dependency on the removal of class 332s, ATP and introduction of 387s.

  32. @JOHNM – At least on the subsurface job, the S stock can continue using the historic signalling if a stage commissioning is delayed, so while costs may rack up, service shouldn’t suffer, although promised reliability and capacity improvements will be deferred clearly.

  33. We saw videos etc of test-runs of trains going through the new tunnels. Can the same sort of runs now be applied, with passengers, perhaps Mon-Fri daytimes with testing of software done overnight and on weekends?

  34. @WW – From what I’ve read here and elsewhere I too can’t see that the two systems could be operated reliably alongside each other, so as far as I can see without any insider knowledge, the changeover to ETCS must occur first, as soon as possible once GWR are ready to replace the 332s with 387s. Clearly the new express trains will require some final integration testing as well, but I believe all or most of the new ETCS trackside kit is already in place awaiting commissioning. That appears to be the key dependency for this part of the work. Once the ATP was powered down, ideally there would be no going back, although there could be a backstop contingency plan to revert to ATP with 332s on the expresses if an unexpected hurdle appears during final 387 testing, but that would then delay vacation of the HEX Old Oak Common depot which might then have implications on HS2’s commencement of work in that area. The GWR depot has already closed. Once ETCS changeover has taken place successfully, then routine night testing of 345s can begin and hopefully the train can be approved swiftly for passenger service in the tunnels.

  35. The fault for the delay almost certainly does not lie with Khan but that is not the same as saying he has no responsibility for this. In his various roles it is his job to know, just as the captain of the Titanic should have known. If Khan fails to monitor and ask and challenge, what is the point of him at all – it can’t just be to grace the posters he puts around London at taxpayers’ expense.

  36. The article says: “Up to this point, the Mayor had rarely interacted directly with Crossrail themselves as, rightly, it was through TfL … that he received the majority of project updates. “. Surely the project is subject to a series of Independent Project Reviews (IPRs) which I would have thought were addressed to TfL but copied to the mayor.

    That’s my experience of far smaller projects. Isn’t the same true for TfL sponsored projects.

  37. @ C Butler – I suspect that Crossrail is not subject to standard TfL practices. It has its own governance which was put in place a long time ago. One of the first things the Mayor said after this furore broke was that he was critical of the governance arrangements and would not have put the current ones in place if he had been responsible. AIUI the oversight arrangements inside TfL have changed several times and matured over the years but I doubt they’ve applied to Crossrail.

    There is another question in all of this – who is the intelligent client? The problem with big construction based projects is the project develops a life of its own believing it is accountable only to itself. Pesky client teams asking questions are an irritation that might stop “stuff getting done”. I saw this in spades with the JLE. I was a technical support resource to the JLE Client team on ticketing matters and did a lot of work for them. I spent far too much of my time having my work challenged and reworked by the JLE project team. It was tiresome and unnecessary, especially when they graciously concluded my calculations were “correct”. I’ve no idea how TfL have run their “client team” vis a vis Crossrail and what level of control, if any, it had. The one aspect where there does seem to be some “power” is the operator element where there is a clear willingness to say “no” when it needs to be said and to preserve the programme time for test running / familiarisation / operational readiness.

  38. I agree that it is very hard to see any circumstances where Shaun Bailey beats Sadiq Khan but there must be a chance that at least one of them won’t be on the ballot. If the Crossrail fiasco rumbles on and Sadiq Khan is looking vulnerable presumably the Tories will find a way to replace Shaun Bailey with someone who might actually have a chance of winning. And if there is an early general election Sadiq Khan might be tempted to do a Boris, cut his losses and head back to Westminster.

  39. London’s first automated railway was the Victoria line and for this London Transport developed and tested the automated train control system *before* committing to it.

    As some readers will know, this included trials between two west London stations plus a District line passenger train and then full scale trials with several trains on the Central line between Hainault and Woodford. For all these trials they used trains which already existed – after all, the idea was to test the new technology – not the trains!

    Only when they were sure that everything would work did they commit to automation on the Victoria line. In addition, prior to use on the Victoria line, all of its new trains were first tested on the automated section of the Central line to ensure that everything worked as intended.

    Re: Crossrail, it is perhaps too late now to do likewise. Or is it?
    Maybe a lesson should really have been learnt from the Jubilee line extension which, it may be recalled, prior to opening faced such significant signalling issues that (as an emergency stop gap) a different (albeit lower capacity) signalling system was installed to ensure that the new section of railway opened in time.

    Once Crossrail line 1 has opened and it becomes time to commit the full story to paper, then in addition to pointing fingers at the politicians, managers, different (and rival) commercial organisations whose trains and signalling do not seem to interact as hoped, it might be worthwhile pondering whether testing everything in a small scale way before committing to those technologies would have resulted in fewer problems. Yes this might have added a little to the overall project duration, but at least it would have been more likely that the new railway would have opened by the advertised opening date!

    I accept that Crossrail is using brand new trains (Class 345) of an unproven design, but as I pointed out earlier, with the Victoria line the original testing was with trains (1960 tube stock) that had already been built and had a known technical quality. This ensured that the blending of technologies was tested – not the trains.

    Maybe a lesson could be learnt from the issues now being faced and when it come to other railway schemes (HS2, Crossrail 2, 3, etc) the planned technologies should be tested somewhere (and kinks ironed out) in advance of installation and a sudden ‘oh sh*t’ moment when it is discovered that products from a mix of suppliers might work well on paper – but when installed on a real railway do not interact as expected.

  40. @SIMON – While this doesn’t help the current crisis very much, I think longer term that ETCS/ERTMS will be a saviour, as the idea and practise of emulating numerous custom and legacy signalling systems within the standard on board EVC (european vital computer) and DMI (driver machine interface) using STM (specific transmission module) techniques will have become completely mainstream in the next generation of train and infrastructure projects. Systems like GW-ATP, that do not conform to modern EMC standards at all, will have have been consigned to the history books by then and most systems, even custom applications like TASS (tilt and speed supervision – for Pendolinos on on WCML) and Thameslink style ATO (automatic train operation) will be able to be constructed from standard ETCS components with centrally registered and non conflicting transmission standards and messaging protocols. Although never personally involved, through friends and ex colleagues I remember hearing of the developing debacle of the JLE signalling, and around the same time the industry under Railtrack walked straight into the similar WCML signalling disaster, trying to purchase a vaporware moving block train control system that still, twenty years later, has not been achieved on any main line railway anywhere in the world.

  41. Re Mark T @0256

    Agreed EVC +STM approach will be the saviour but it is this generation of trains it will be happening on with rapid changes over the next few years. And this is the point it need to be made to work.

    TASS effectively use ETCS equipment designed for additional local functionality anyway – packet 44 and Thameslink uses packet 68 for ASDO.

    Re Simon and Mark,

    Part of the issues are location specific hence the test it elsewhere approach doesn’t always work. e.g. the non standard GW-ATP installation in the Heathrow Tunnels might cause hiccups and separately for EMC issues the in tunnel and out of tunnel will change thing given the frequencies used for GW-ATP and ETCS (27MHz – a bad choice for rail tunnel geometry).

    The Belgian approach (decided in 2012) when faced with the same problem was to remove and replace TBL1 (GW-ATP uses slightly modified TBL1 equipment in disguise) to enable ETCS roll out.
    The DfT approach (also in 2012) was to order a huge last batch of GW-ATP (TBL1) equipment for the IEPs before production effectively ended.

  42. Jake – good point: I was so focussed on getting Shaun Bailey’s name right I wasn’t paying attention to the Mayor’s name. Maybe one of the mod’s can help me save face and edit my original comment.

    [Done. Malcolm]

  43. Re WW and Mark T,

    1. The Heathrow express 332s are in a very poor state mechanically and at least 2 units will probably already never move under their own power again so retrofitting with TPWS is very very unlikely.

    2. The Central Link North – South Corridor in Hong Kong also announced massive delays earlier in 2018 (and much embarrassment for MTR) – a Rail project interfacing Trainguard MT and legacy signalling systems in similar ways to Crossrail.

    3. A software solution to the Heathrow tunnels could be possible, the first step would involve taking very detailed electrical and magnetic field measurements in the Heathrow tunnels with ATP and ETCS on, to enable analysis to be undertaken to see if (software) filtering is theoretically possible and if so then to develop (modify) the filter (during which time further testing would probably be wasted effort). What you are trying to filter in open air will be different to what you are trying to filter out in the tunnels hence the need for real in tunnel runs.

    It therefore appears that there is the belief that software filtering can work and the filter has been developed but is awaiting testing. Various bits of software will be bundled up to reduce integration and regression testing requirements so a non filter related part fo the software may be the hold up.

    4. The new STM for AWS and TPWS (as fitted to 345s) also uses software based filtering compared to a mix on previous AWS /TPWS equipment so the software only approach isn’t a novel one.

  44. @NGH
    (possibly answered in your forthcoming article)
    Is it just in the Heathrow tunnels where GW-ATP is interfering with ETCS? Do the GW reliefs have a different signalling system, or is it just that in the open the RF isn’t bouncing around off everything and causing havoc?

    Is the signalling in the Heathrow tunnels capable of running trains bi-directionally? I would assume there were contingency plans to allow continued running in the event that one of the tunnels was blocked: the scissors crossover at Heathrow Tunnel Jn suggests so.

    If Heathrow Connect (or whatever it’s called now) was bustituted between Hayes & Harlington and Heathrow for a couple of weeks, the timetable looks just-about feasible to run HEx in just the ‘up’ tunnel, the ‘down’ tunnel having the GW-ATP turned off once for a long blockade to do testing; if the testing is successful then it can be used for 345s or 387s and the section from Heathrow Central P1 through to Heathrow T4 switched over to ETCS while HEx uses Heathrow Central P2 both ways and goes onto T5. Once a ‘bedding in’ period is complete, the rest of the tunnels are converted and HEx dumps the 332s for 387s as planned.

  45. @ Moosealot

    99% sure that ATP is only fitted to the main lines, and the reliefs have TPWS. I believe ATP only exists on the tracks that have a speed limit of 125 mph… although I expect Mark T or NGH will be along to correct me any minute!

  46. @NGH And guess who is the system integrator for the Intercity Express Programme with its Electro Magnetic Compatibility issues? Those masters of Hi Tech – the Department for Transport.

  47. The Belgian TBL 1+ upgrade was very interesting, as they were able to remove all their legacy trackside equipment in favour of standard Eurobalises, while keeping the cab equipment behaviour on older trains exactly the same as before. This approach also allows a future painless upgrade, using the same track transponders, to a fully compliant Level 1 LS (limited supervision) system once all trains have EVCs and DMIs. It was an approach pioneered by Swiss SBB, with strong support from other national networks, who are now unique in having already achieved complete nationwide ETCS fitment thoughout their standard gauge network via their Euro Zub/Signum project.

    French KVB is a similar balise-based limited supervision system that should be comparatively easy to migrate to L1 LS, now incorporated as a standard mode in the international standards. Germany is planning a migration strategy to replace their legacy PZB (Indusi) trackside equipment using digital eurobalises in LS mode, and Italy is implementing a similar project for its busy international corridors, also using Euroloop infill for earlier ‘better aspect’ notification. All these national networks are also using standard Level 2 ETCS, but it is confined to high speed and very high traffic corridors initially, where the additional capacity benefits of continuous radio-based movement authority are particularly beneficial and worth the additional cost and difficulty. A big advantage of L1 LS is that it is not dependent on resignalling, so route-based rollout can be decoupled from the signal asset renewal programme. Clearly the GWML and Heathrow tunnels would not be an ideal long term candidate for LS techniques, being one of the busiest sections of the UK national network, but a similar approach might have been appropriate as part of migration towards the higher levels in an alternative history.

    Here’s a small presentation about the Swiss approach from 2011:
    http://www.irse.org/knowledge/publicdocuments/8B_IRSE_London_Zuercher.pdf

  48. @HERNED – The Reliefs are equipped with GW-ATP as far as Heathrow Airport Junctions while it is fitted on the Mains alone beyond that. That is because HEX services can be diverted via the Reliefs and the cl.332s do not have TPWS.

  49. Re Herned,

    Spot on – No need to correct!

    GW-ATP Fitted:
    Paddington – Reading Fast lines only
    Reading – Bristol, both routes
    Reading – Newbury
    And later Heathrow Branch

    Scotrail is having issues with their newly arrived HST power cars as turning ATP off turned out not be that easy…

  50. Re Captain Deltic,

    I didn’t need to guess! I liked your recent graphs on the IEP subject – the core issue the Hitachi deciding not to use the lesson learnt from the original Electrostars (which all the other manufacturers including Hitachi had learned from up till now e.g. 900 not 750Hz) . I recently did some work for a 3 phase Drive manufacturer and one of the foci was around why their customers didn’t buy (decent*) filters and didn’t understand the need for them…

    *Typically made by German Mittelstand firms

  51. Re Moosealot,

    1. The GW-ATP installation in the tunnels is non-standard, someone decided to double up on all the track beacons to improve reliability

    2. Tunnels ” is it just that in the open the RF isn’t bouncing around off everything and causing havoc?”
    Yes but even worse than that, the frequency used by both GW-ATP and ETCS is above the waveguide threshold for the tunnel size and geometry so not only does it bounce it keeps bouncing, propagating like light down a fibre optic strand too… (Barnes Wallis would be proud of the amount of bouncing going on!)

    3. The 345s are running every day on GW Reliefs over ATP equipment, with both the Paddington – H&H services and the Maidenhead test runs without issue.

  52. Re Mark,

    Indeed I should have been a bit more careful in my description of the “Heathrow”, added later for Heathrow in reliefs.

    There Phases of GW-ATP installation:
    1. original pilot Swindon – Bristol Parkway
    2. Paddington – Swindon (Main /fast lines only), Parkway – Temple Meads, Swindon – Temple Meads (via Bath), Reading – Newbury
    3. Heathrow inc reliefs to just east of West Drayton (a few signals west of Airport Jn)

  53. Epping Blogger
    Not helped by the, erm, let’s call it “misinformation” shall we, highlighted in the article.

    Herned
    Re. Transformer explosion.
    Yes. The perfect opportunity to announce a delay – but they blew it.

    Simon
    Re: Crossrail, it is perhaps too late now to do likewise. Or is it?
    I wonder: Abbey Wood – Whitechapel ???

    NGH
    Re. GW-ATP: Sophisticated “updated” filtering is all very well in theory, but the time & expense of going through that, when it may not work anyway, coupled with what you said here: The GW-ATP installation in the tunnels is non-standard … strongly suggests that the “probable-best” solution is to wait to the appropriate moment & “simply” rip it all out & replace with ETCS eurobalises, or whatever is most suitable?
    Oh yes, apart from Paddington-Heathrow, how much of the “old” GW-ATP equipment is still in regular use, now?

  54. Re Greg,

    “Rip out and replace” – the replace bit is already done, switch off and remove at a convenient time later is all that is required. DfT probably see this as the way forward given the deal they sorted with GWR and HEx for GWR to take over and Bombardier to retrofit ETCS to the 387s (designed for easy installation of ETCS as were 379, 377/6 &/7).
    (the GN 387s are getting retrofitted at some point shortly after too.)
    The problem at the moment is that GWR need the 769s to release 387s (should be sorted longer term with the 6 387s ex c2c.

    Software filtering isn’t easy…

    “how much of the “old” GW-ATP equipment is still in regular use, now?”
    All of it, as it is safer overall than TPWS or TPWS+ hence ORR won’t allow the safety level to be reduced hence ETCS as the future. TPWS+ can match – ATP safety levels most of the time without too much effort but not in all locations.

  55. I wonder if a temporary remote isolation system could be installed in the tunnel ATP system so large areas could be switched off and back on again quickly to enable night time testing on ETCS to commence before the 387s take over.

  56. ngh 14 December 2018 at 10:09
    “The Central Link North – South Corridor in Hong Kong also announced massive delays earlier in 2018 (and much embarrassment for MTR) …”

    That is, of course, the same MTR which was awarded a concession in 2014 to operate Crossrail for eight years (currently branded TfL Rail) and which has a 30% stake in the joint venture operating the South Western Railway franchise.
    They should be used to it.

  57. “The problem at the moment is that GWR need the 769s to release 387s”
    Crossrail / tfl surely now have more class 345s that they need for the H&H to Paddington shuttle service? Given the delays in getting the class 769s into service, could the temporary use of some 345s to release GWR387s free up the logjam?
    I’m guessing I’ve missed something that is preventing this being done – what?

  58. NGH et al, ‘ one of the foci was around why their customers didn’t buy (decent*) filters ‘ one of the reasons is that ‘decent’ filters are usually heavier and require more space so there is a lifetime penalty in carrying them about. The Piccadilly replacement rolling stock has ended up with a larger filter to ensure they can run on the existing (125Hz) signalling system. However the space required resulted in problems in fitting the other equipment under the cars.

  59. @Island Dweller
    “could the temporary use of some 345s to release GWR387s free up the logjam?”

    For some of the same reasons that releasing emus (or even dmus) for the Goblin, this is probably a non-starter.
    1. GWR don’t own the 345s so they would have to be sub-leased
    2. They certainly aren’t intended to go beyond Reading so they would be limited in which GWR services they could operate – indeed have they yet been cleared to work west of Airport Junction?

  60. @ City Worker – well yes it is the same MTR but of course they are not building Crossrail nor are they responsible for fitting and testing the signalling / control system for Crossrail. They are simply the operator and Crossrail have not handed anything over for operation yet so MTR aren’t playing on the core section. If they cock up the actual operation of Crossrail then I am sure due criticism and derision will be heaped on their heads at that time. As someone who uses London Overground I miss the influence of MTR in that operation. It is decidedly worse now it is Arriva on their own.

    By way of a small diversion I thought the major problem on the Shatin – Central link was to do with huge problems with construction of the tunnels and some of the stations. This is partly to do with archeologically important sites being found once construction started and also ground conditions forcing redesign in several places. If we add on potentially fraudulent practices by construction firms not building some stations properly and MTR seemingly not spotting this at the time then you have an almighty mess. Quite how MTR have got themselves into such a state is a bit beyond me given their past record has been very good in terms of new line construction and getting them into service. I was not aware that signalling was yet posing a big problem for them on the Shatin – Central link (SCL). I do know that there has been extensive upgrading work on the Ma On Shan line (which links to the SCL at Tai Wai). This has included train lengthening, platform extensions, fitment of platform edge gates and I assume resignalling too. It seems to have progressed reasonably well from what I’ve read. West Rail has long had signalling issues and that eventually will link to the SCL at Hung Hom while East Rail (the old Kowloon Canton Railway) will be diverted into a new cross harbour tunnel to Admiralty. That new tunnel is a second phase of the overall SCL project (it’s all a bit involved given lines get built in one form and are then reconfigured in a second phase).

  61. Other problems with using the 345s on the current GWR services are the way in which the off peak timetable diagramming is set up Reading – Paddington – Didcot – Paddington – Reading. There doesn’t appear to be platform capacity for longer turnarounds at Paddington and something needs to use platform 14 that won’t take 9-car 345s.

    So, the timetable would need to be rewritten in some way.

    Not a quick fix.

  62. NVRees 13 December 2018 at 19:09

    ” videos etc of test-runs of trains going through the new tunnels. ”

    I think you’ll find that was just one train; not 15 trains per hour.

  63. Timbeau & Jonathanh. Thanks for explanations on why 345s won’t come to the rescue.

  64. @GregT
    Transformer explosion.
    “Yes. The perfect opportunity to announce a delay – but they blew it.”

    I see what you did there 🙂

    @NVRees/AG ” test-runs – I think you’ll find that was just one train”

    https://www.railmagazine.com/news/network/morgan-crossrail-awaits-train-fit-for-testing

    “Crossrail does not have a train fully fit for testing yet, chairman Sir Terry Morgan told the London Assembly’s Transport Committee on September 12. Testing of trains in the central section began in February this year. Morgan added that at no point in testing has more than one train at a time run through the tunnels.”

    Feb 2018 run – first 345 to be driven from Abbey Wood. It entered the new tunnels at Plumstead Portal and travelled under the Thames heading for Connaught Tunnel before returning to Abbey Wood station. https://www.youtube.com/watch?v=KEGyP3A2PmE

    May 2018 – first footage reaching Canary Wharf from Abbey Wood reverse running in Eastbound tunnel
    https://www.youtube.com/watch?v=mw5CR8yEsD4

    Those may be ‘tunnels’ but ‘central section’ !!

  65. Talk seems to be that if TfL aren’t ready for through service to Reading as planned, they could start it from current Paddington for fares revenue.

  66. This London Reconnections article is quoted in an excoriating piece on Sadiq Khan in the Sunday Times today by Andrew Gillian.

    The article is behind a paywall unfortunately.

  67. This London Reconnections article is quoted in an excoriating piece on Sadiq Khan in the Sunday Times today by Andrew Gilligan.

    ‘When historians come to write about the mayor of London, Sadiq Khan — assuming any would bother — they could do worse than look at the month of June 2018.

    We now know this was when a 51-page report from the consultants Jacobs was completed, detailing cost overruns on Crossrail, the capital’s new east-west railway, and speaking for the first time of the risk that it might “open late”: problems that have since grown to a year or more’s delay, and a deficit of more than £2bn.

    He pledged to “both freeze . . . fares and invest record amounts modernising London’s transport infrastructure”. You cannot, of course, do both, though in fact Khan has done neither.

    Travelcard prices for commuters have continued to rise, but the freeze in other fares, and now the Crossrail crisis, has forced Transport for London (TfL) to cancel upgrades to the Tube and suspend even routine road maintenance. The son of a bus driver is making the first serious cuts in 30 years to London’s bus network.

    … public transport use — again, for the first time in decades — is down.

    It is not all his fault. As he always points out, government grants for TfL and the police have been cut, though many of the reductions happened well before he came into office, …

    On Crossrail, Khan continues to insist he learnt of the delay only on August 29, two days before the public. But new papers released last week show he was told on July 26 that the original opening date was “not feasible” and “not possible”, confirming reports by The Sunday Times and claims by the former Crossrail chairman, Sir Terry Morgan.

    As the transport website London Reconnections has said, it was perfectly appropriate for the mayor to know this before he shared the knowledge with the world, and he has created an “entirely unnecessary political crisis” by denying it.

    [This was, in essence, a political piece about much more than Crossrail. I have been very kind and spent time editing it to leave in the relevant bits about Crossrail. We don’t like overt one-sided political opinion. Moreover, we don’t generally take kindly to long cut and pasted comments which are only partially relevant. Cutting and pasting is quick and easy. Editing it into an acceptable form for this website takes a lot of our time. What remains is considerably shorter than what was originally posted. PoP]

  68. There are three other relevant pieces of coverage about Crossrail problems, available in online links, all going back some months previous to June 2018, and one as far back as 2015:

    https://www.dailymail.co.uk/news/article-2905641/Crossrail-facing-major-signal-problems-prevent-trains-running-2018-15billlion-line-open.html
    Interestingly old major signalling issues article in Daily Mail 11 Jan 2015, when Terry Morgan commented on it. So the topic is NOT a new one, and Crossrail management knew at least some of the issues then.

    https://www.standard.co.uk/news/transport/crossrail-chiefs-warn-that-14bn-elizabeth-line-could-blow-its-budget-and-open-late-a3754356.html
    Evening Standard 31 January 2018 saying transformer explosion has raised major timescale issues – and that Mayor Khan knows it.

    http://www.railtechnologymagazine.com/Rail-News/elizabeth-line-hit-with-major-challenges-after-electrical-explosion-pushes-back-testing
    Story in Rail Technology Magazine on the same date (31 Jan 2018) – with Mark Wild jointly talking with Terry Morgan to Sadiq Khan and being reported as saying we can still do it by December 2018.

    The RTM report remarks: “Other issues had also pinned back some progress of the project, although both Wild and Morgan were confident that the planned timelines were possible. However the London Underground MD added: ‘We can still do it but it’s very, very hard and complex and it brings with it cost pressures as well.’… Khan called an end to further discussion of the specific problems but said it was for ‘commercially sensitive reasons’ rather than to hide details.” Perhaps it is a shame that there wasn’t further open discussion then, to achieve earlier public clarity about progress and continuing issues.

    Mark Wild’s views have changed considerably, if you contrast those reported January 2018 comments above with his presentation last Tuesday at the 11/12/18 TfL Programmes & Investment Committee meeting. Following his initial reviews as CEO, there was now “no confidence” in opening anything in 2019. After another 2-3 months assessment, it would then be a range of possible dates.

  69. PoP- Thank’s for taking the time to ‘sex down’ the Gilligan piece. It doesn’t really say anything new on the subject as far as I can see. Just goes to show the dangers of political office relating to large infrastructure projects. If things had gone to plan, Khan would have been able to bask in the mid term reflected glory. Now it looks like scheme completion may be bumping up to the end of his term, so how it goes from here could have serious implications on his re-electability.

    If timetabling issues can be sorted out quickly, taking over Paddington – Reading stopping services as quickly as possible seems a very good idea to start getting some income in, even if the frequency offered is a reduction from what had been planned originally if through running had been possible to and from the core. The big advantage of this route is it only requires bog standard TPWS and AWS signalling functionality to be available on board.

  70. @MILTON CLEVEDON – That old Mail article is interesting but seems to slightly miss target as to the particular signalling problems they’re having now. It suggests there will be ‘no problem’ with the orthodox metro system used in the central tunnels and the problems will occur at the immediate interface with the surface railway at each end. I believe that was addressed by the ‘plan B’ approach of opting to use TPWS/AWS (enhanced to an extent where neccessary) on the GWML proper, as also always envisaged for the GER segment. The Heathrow tunnels were always going to be a slightly more difficult challenge, largely because the existing HEX trains do not have TPWS and are impractical and uneconomic to convert to ETCS. Even if the cl.332s were TPWS enabled it would have been unlikely that system even if applied in its enhanced form with extra loops beyond typical surface main line standards, would have been considered sufficiently safe for such a ‘metro-like’ application. ‘Metro-like’ railways have always historically required a very robust and safe train protection system due to the possible consequences of tunnel collisions and derailments. Hence early mechanical train stop and trip cock systems. The GW-ATP pilot scheme system was seized on by the HEX project as a modern (at the time) solution to this requirement, but that has meant the 332s never needed TPWS as all the lines they run on have the ATP equipment.

  71. @MT 14:45
    Thank you for that. Your comment would suggest that there was then no anticipation in early 2015 that new train-signalling specifications would raise issues within the central tunnel section between the different technologies specified (as opposed to interfaces with On-Network systems and in the Heathrow tunnels). May I ask NGH to cover that point when he writes up his intended detailed technical commentary?

  72. @ Mark T – given your remarks about Crossrail’s opening running up against the end of the Mayor’s current term two things spring to mind. First is the extent to which the Mayor will “suggest” (being kind) that it would be a “good idea” that Crossrail is open before the hustings start. You don’t need much imagination to think of a different tone in which that point might be being made right now. The second point is that once the official campaign is in place then purdah descends preventing any sort of big event from which politicians seeking re-election might benefit. Furthermore TfL effectively shuts down, in terms of comms and PR, during the purdah period. We had precisely this issue with the opening of the East London extension from Dalston Junction southwards. I suspect those issues are forming part of the consideration of the revised programme.

    I also think that Mr Wild’s apparent conversion from optimist to pessimist reflects the need to under promise and over deliver so as to “pull the magic rabbit from the hat”. It might also be because he’s taken fright at what has to be done but I’m sceptical about that *if* he is as skilled and experienced as is said. Based on that he must have been aware for a long time that things were not deliverable given he attended board meetings, attended special briefing sessions etc where I’d have expected him to be asking a load of informed and challenging questions about what was really happening. Surely that was his role as a TfL nominee non exec director on the Crossrail board? The same applies to the other two TfL Board members who are also non execs on the Crossrail board.

  73. I always disliked RAG. First I wasn’t sure I understood it well. Then I noticed that different project managers seemed to have different definitions of what actually constitutes red/amber/green. Now reading about the Thermocline of Truth I wonder if it simply unfit and whether a more grainy system is required. However any new system would also need to be easy to understand, so maybe school marks could be used: A (very good) would be progress clearly ahead/savings realised, B (good) would be on time/costs, C (average) would be within planned tolerances, D (acceptable) would be behind schedule/over cost, but both known, E would be behind schedule/over cost with further schedule/costs unknown, and F would be project output at risks. (Or something similar, I’m making these up.)

  74. Milton Clevedon,

    From a non-technical perspective my understanding is that initially new train-signalling specifications and interfaces were a great concern. Indeed, much of the early debate about which signalling system to go for included concern about interfaces. At the end of the day, then, it was decided that ETCS was just too immature for the busy central section with platform edge doors and ETCS would take to long to reach maturity due to the fact it was a standard that needed to be ratified. So multiple interfaces it was then.

    The initial concern was tempered by the fact that railways in different parts of the world seemed to be successfully transitioning from one signalling system to another and often one of the systems was ETCS. So it was a known issue that had not yet been solved in the case of Crossrail but was thought it would be solvable when the time came to implement it. So the Daily Mail back in 2015 was correct but probably overstated the level of concern at that time.

  75. Christian: Altering the grading system is unlikely to make any difference to the “thermocline of truth”, in my view. The phenomenon occurs because people reporting up are always expected to meet two oft-conflicting standards. One is to report accurately, and the other is to bring good news. When the only news available is not-so-good, there is a natural tendency to blur things. A policy of not-blurring is selected against, because messengers bringing bad news tend to get shot, or at least they fear that they might.

  76. @ Christian – having had to report on RAG “traffic lights” on a whole load of things it all boils down to people knowing what is *really* going on, those people telling the truth and those above them being willing to receive bad news and act on it. Sadly there are a thousand reasons, some valid others not, as to why people cannot keep up with events, why people lie and why superiors don’t want bad news. I have certainly struggled to be certain about what was going on some things I was reporting on and also met resistance if the message was not good. I didn’t lie because that’s a stupid thing to do (IMO) but I was certainly pressured not to be so pessimistic. There was also the occasional threat of “career death” if the wrong status was reported to a Director level meeting. Only had that threat once and a remark that I was quite prepared to be harrangued by the then MD for my supposed failings was not well received. I always thought that was odd – I thought you were supposed to take responsibility for things. It doesn’t matter what grading system you use it is down to people and their behaviours as Malcolm says. That is what you need to tackle.

    JB has said that Crossrail clearly suffered from “hubris” in how they viewed their progress and achievements. It was as if nothing could stop them achieving their magic deadline of December. Sadly reality stopped them achieving their deadline. Even now I am somewhat astonished that this happened as I’ve always viewed Terry Morgan as a “no nonsense” person who would dig out problems and ensure they were fixed. Clearly though that no nonsense approach did not sink very far down the management structure nor did a desire to really know what was happening or a robust process of challenge to find the truth.

  77. Walthamstow Writer,

    Even now I am somewhat astonished that this happened as I’ve always viewed Terry Morgan as a “no nonsense” person who would dig out problems and ensure they were fixed.

    But surely that’s the problem? You have a lifetime of experiencing this and you know from past experience these problems can be got around if you focus on them and address them. So, you just assume that the same will apply this time. In my not-so-humble opinion, it is exactly the same problem GTR were suckered into. It takes a lot of uncomfortable objective thinking to realise that ‘this time it is different’.

  78. @ Pedantic of Purley.
    In the City, the phrase ‘this time it is different’ rings warning bells. For Crossrail and GTR the killer has become ‘This time is no different’.

    In my specialist subject, the new generation of software enabled trains with their Ethernet spinal chords and multiple interfaces are simply not responding to tried and tested processes for commissioning and reliability growth.

  79. Re PoP & Captain Deltic,

    Everytime is different, the question is how much, just because you’ve succeed the last X times doesn’t mean you will this time.

    Agree with Captain Deltic that traditional commissioning approaches aren’t working* and I’be suggest that a more software style approach is taken as Siemens eventually learnt with the 700s (see my graph in the GTR article) and Bombardier planned with the Aventras based on the wider groups aviation experience e.g. batching up problems / solutions for testing /implementation. The side effect of this is that commissioning takes far longer. A possible solution is that a couple of prototype /pre series units are produced well ahead of the production run to allow the software issues to have for time to be worked on pre “planned” large scale service entry.

    *Still works for the more mechanical shake down elements

  80. Re WW, Christian, Malc0lm,

    As someone who does have much fear of red shifting things (the antithesis of Crossrail Green shifting things?) I share Christian’s dislike of RAG as it strips to much information out to allow properly informed judgement is a delay because that task is having big problems or is it just delayed because it started late because of delays on completing a previous contingent task, the former and the later need to be treated very carefully.

    Another bugbear of mine is not leaving enough contingency built in earlier in projects and programmes to act as a bit of a proper fire break, the current tendency is all to add it later on the “more flexibility to replan” approach when infact the opposite is true as the flexibility tends actually involve everyone getting in each other ways for a considerable overlap period with a huge productivity drop and moving goal posts for the later tasks

  81. ngh,

    And it is interesting to look at SSR resignalling in this context. They failed to implement the first live section because of software problems but don’t seem to be panicking because the whole idea of earlier implementation was precisely to identify these problems early on.

    Meanwhile, they are still going ahead and putting in the hardware for future phases. They also have the software in place to fully test the junction capacity under the new system but it isn’t yet good enough to allow passenger-carrying service.

    One of two things will happen:
    – either SSR resignalling will ultimately be late and we will have the usual post-mortems and recriminations – or –
    – it will be on time (according to its latest schedule) because they tested the software early on and will be held as a model example of how this sort of thing should be done.

  82. One thing I don’t understand re this whole ‘thermocline of truth’ issue is how that can happen with the actual physical building of the stations and infrastructure. I can see how it can happen with software and training, but with tangible stuff then senior management could surely see with their own eyes that the build wasn’t going as well as planned.

    You don’t need to know very much about construction to tell the difference between a year’s work and 3 month’s work…

  83. Re PoP,

    Or a third: several months late and no one will worry too much… as it was “expected”

    Especially because LU has much more freedom on timetable change dates on the SSR and passenger growth has slowed.

    The parallel with Crossrail is that the trains are still being built and delivered (circa 50 of the 70 so far) but on the SSR signalling they have left themselves much more room to tackle problems and the way it is phased the possibility to catch up and are also much more open about progress (or lack of).

    You need the hardware in place to do most of the software testing and find the problems.

  84. Another factor de-risking SSR is that the modern CBTC signalling techniques employed allow almost all the equipment to be installed and set to work in parallel with the old signalling, even including parallel train detection, because unlike the existing, the new system uses no track circuits, relying instead on balises and odometry for fine positioning and axle counters for section clear proving. The only trackside objects that are not possible to connect up in final configuration are the point machines, which clearly must remain attached to the existing signalling until the formal changeover (although they might have some properly designed temporary test switching attached in circuit so the new system can rehearse the points operation during possessions before then). If a particular commissioning stage is missed, there should be no significant effect on service as operators can simply continue using the old signalling until the new date. There are far fewer dependencies than on a massive multidiscipline new railway project like Crossrail. SSR is also only attempting to grapple with one new signalling technology on these lines, and one that is, albeit in a slightly earlier form, now widely used across TfL, and is thus familiar to its technical staff.

  85. NGH
    A possible solution is that a couple of prototype /pre series units are produced well ahead of the production run …
    Of course, sometimes, this was done, even as far back as the days of steam traction ….
    This reminds me, oddly enough of the “production” models of DP2, ( Class 50 ) where the prototype was brilliant, so they built a production-run … with all sorts of (by the standards of the time) fancy bells & whistles added … which then promptly fell over in a messy pile of unreliability.

  86. Greg Tingey,

    Which is also what they did with Victoria line 2009 stock. The plan was to upgrade the two prototypes at the end of the production run but they turned out to be so different (so many lessons learnt) that I believe they ended up scrapping them and building two new trains.

    So the last two 2009 trains to be built were train 001 and 002.

  87. @ PoP – I see what you’re saying and that it can be plausible. However on an enormous project like Crossrail you’d hope (well I would anyway) that there was enough expertise in place over many years for the risks to be properly and fully understood and also managed / mitigated / removed. I’d also expect that expertise to be able to detect when reports were not matching reality. Perhaps I just expect too much of people? Circumstances would suggest I am wrong.

    @ Ngh – I understand your point. I think, though, we’re agreeing. At some point in the process of project planning, implementation, progress review meetings etc etc someone has to be sufficiently tough and robust to force the truth (detail) out of people. If that means they or key people who they trust have to go and see things for themselves or dig around on site or in planning offices or with suppliers then so be it. It shouldn’t be necessary but as I said not everyone tells the truth, not everyone is as competent as they might believe, circumstances can overwhelm people etc etc. It comes back to intelligent clients and effective project managers being appropriately robust and challenging with each other even if it risks “upset” in the relationship. The point is that the project comes in on time and to cost (or close to these parameters) so that everyone has a good result rather than sitting in the mire. I say this from having been a client on many projects but never a project manager. I recognised I probably didn’t have the right temperament to be a good PM.

  88. Walthamstow Writer,

    But isn’t part of the problem that Crossrail saw the project as an engineering project with a railway at the end of it. So the emphasis was on construction which is notorious for going well until towards the end when problems emerge (e.g. fire doors at Bloomberg entrance for Bank) and there is no time available to fix them. And people always concentrate on what they feel comfortable with and understand (one of Parkinson’s laws). So, with very little direct input from TfL due to the way the company was structured, they primarily looked at the engineering issues and decided all was well.

  89. PoP
    But … isn’t software & its application supposed to be an “engineering” type of subject, anyway?
    Especially when there is so much interaction with “physical” engineering in something like the crossrail project?
    Time to quote the late, great Prof. J E Gordon again?
    One can recognise a certain atmosphere of Gadarene inevitability about the whole procedure. Under the pressure of pride & jealousy & ambition & political rivalry, attention is concentrated on the day-to-day details. The broad judgments, the generalship of engineering, end by being impossible. The whole thing becomes unstoppable & slides to disaster before one’s eyes. …
    Thus are the purposes of Zeus accomplished. People do not become immune from the classical or theological human weaknesses merely because they are operating in a technical situation, & several of these catastrophes have the much of the drama & inevitability of Greek tragedy. It may be that some of our textbooks ought to be written by people like Aeschylus or Sophocles – these people were not humanists.

  90. Over on thread “London Underground: The Forgotten Digital Railway” @Dstock7080 updates that another software dependent project – 4 Lines Modernisation – continues to slip right.

    Interesting to notice in an earlier comment on that thread 24 June 2018 at 18:22 that “Mark Wild … has been at confidently telling various committees that it has all been up and running and working and it would go live by the end of June [2018] “.
    Over-optimism seems to be endemic .

  91. What no one mentions here or elsewhere is why ETCS is still being used.

    When the decision to shelve the project to install ETCS on the GW lines out of Paddington, surely it was a no-brainer to install TPWS instead of ETCS on the Heathrow Branch, so eliminating this short ETCS island and another ‘Point of Failure’ in Elizabeth Line operations (i.e. transition from one signalling system to another).

    If a decision had been made to install TPWS on the branch, it most probably would have been completed and in use by the TfL Rail Heathrow services operated by the class 345s, instead of the existing class 360s.

    It could not have been foreseen at the time, but the now takeover of HEX operations by GWR using class 387s instead of the existing class 332s could have been implemented much easier.

    If there was a case where the acronym KISS (Keep It Simple, Stupid) applies, this is it.

  92. Re W5er,

    “What no one mentions here or elsewhere is why ETCS is still being used”

    I have many time on many threads.

    Because it is the future, Upgrading GWML from TPWS to TPWS+ required comparatively few modifications and is a temporary solution as they are looking to ETCS the inner GWML within a few years.

    Fitting TPWS(+) in the tunnels would have been more work than the rest of the GWML mods, remember it has to be as safe or safer that ATP to be approved, a tough bar to clear.

    If you look further up the comments you’ll find I have commented on the ETCS Heathrow situation and pondered why it didn’t raise any alarm bells. Like you I’m still surprised that it didn’t get a mention in the risk assessments. You find I also pondered the same question last January and 87 months before that too…

  93. I don’t know how to upload pics here but I note the heavy duty tape that was on the EL signs on P1 / 2 National Rail platforms of Abbey Wood has been removed and signage and working CIS is in place on the EL platform.

  94. Herned @ 17 December 2018 at 12:34

    “You don’t need to know very much about construction to tell the difference between a year’s work and 3 month’s work… ”

    When I visited Canary Wharf, to platform level, I didn’t overhear anyone commenting on the amount of work still to be done. Must have been many present with much more knowledge of construction than I.

  95. Rob @ 18 December 2018 at 16:27

    Probably an instruction to do that on Saturday 8 December 2018 was never cancelled.

  96. The latest train management systems from all manufacturers are designed to be ‘natively’ ETCS capable, supporting legacy or other custom warning systems like TPWS by emulation through their standard ETCS on-board systems, the EVC (european vital computer) and DMI (driver-machine interface – i.e cab control screens, switches and sounders), with extra plug-in hardware modules, known as STMs (specific transmission modules) as required and custom software (think device drivers and utilities for a home pc equivalent). The ideas behind ETCS are a very long standing UIC technical project, predating the EU and even the EEC to promote interoperability and common standards. The EU nevertheless adopted and funded the programme enthusistically, as it coincided with their wider political and economic aims. Prior to this, railway administrations had always been locked in to their own unique national legacy protection systems and often to monopoly suppliers of that kit who could were free to charge what they liked and impose their own large-scale upgrade requirements when they unilaterally chose to cease supporting a particular vintage system or component.

    As well as the train builders, all mainstream signalling manufacturers and system integrators are skilled up to deliver ETCS-based schemes today. There is simply no other viable show in town. Main line signalling is converging on generic, open, communications based digital solutions, and is doing so according to open ETCS/ERTMS standards at the track-train interface. This trend extends far beyond Europe and any attempts to follow an alternative closed proprietary path again would be foolhardy (IMHO). However I fully understand why Crossrail decided to go with a proprietary metro product for the central section, given the uncertainties over what ETCS could deliver at the time the decision had to be made. In the light of what has since been achieved on Thameslink and elsewhere, it’s possible a different decision might have been made today, but we are where we are and the project must make Trainguard MT work, along with the additional systems the Elizabeth Line and other traffic on the GWML will share on the GWML. I have no doubt this will be achieved and there will be very many lessons learnt from the experience.

  97. @ Alan G – When I visited Canary Wharf on the last open day that was held there I was somewhat bemused that the ceilings didn’t seem finished, nor the lighting nor CCTV and other customer facing systems. I did ask how far things had got but no one would give a clear answer. That led me to conclude that something else wasn’t quite right in terms of power supplies or fire protection systems that still required access to areas that would normally be behind finishes. There were still issues with the platform edge doors too and no wayfinding anywhere in the station. There’s no denying that C Wharf was the most “complete” location but even then it was not complete to my eyes.

  98. @ Alan Griffiths

    But you weren’t (AFAIK) part of the project management team who should have been paying attention to that sort of thing, that’s what I was getting at.

    Besides, Canary Wharf was more or less complete several years ago

  99. I’m just catching up with weekend newspapers. Tucked away in the weekend FT (15/16 Dec) is this interesting observation.
    The FT reports that, according to McKinsey, only 2% of projects worth over $1 billion come in on time and budget. Average overrun of 80% and 2 years.
    That said, I’m agreeing with commentators above that most of these Crossrail issues should have been flagged sooner.

  100. My impression of ERTMS is that it aims to give the impression of making a standardised system, but in practice there seems to be a lot of small details that honestly are rather crappy.

    Afaik there are no readily available STM interface testing tools. Therefore in this case it’s open for Siemens and Bombardier to blame each other as much as they want if the trains won’t work, and there are no technical way to show who is right or wrong.

    Even though ERTMS is a standard and you can buy equipment from anyone, it seems like a big vendor lock-in for a rather large area as soon as you start to buy any equipment, at least for anything more advanced than level 1 LS. Maybe you can buy the balisers from any manufacturer but you are in practice stuck with one manufacture for the electronic interlocking system covering a rather large area, and it seems common for that manufacturer to charge for changes (like adding a switch somewhere or similar), something that usually were done by the railways own signal technicials with the older relay based interlocks.

    With the older classic relay based interlocks, you can kind of buy parts from any manufacturer and you can combine different types of interlocks and block control systems as you like.

    The interoperability also seems questionable. Maybe it’s different in different countries / on different railways, but at least in some places it seems like every combination of on-board equipment and fixed infrastructure has to be tested to get approval for useage. I get the fear of any problems causing accidents but really there should be a once-and-for-all approval for any train installation type and any infrastructure, and after that you should be able to take a train running in one part of Europe and run it in another part of Europe (assuming being compatible otherwise (loading gauge, power system e.t.c)).

    I might be biased from the experiences of ERTMS here in Sweden where I live. There were quite a few problems when it were first introduced. Maybe that itself was a result of it being introduced on a newly build railway in the northern parts of Sweden – much like if it were used on a hypotethical new railway in Scotland in the UK, rather than on any core section in a large city. The recently buildt “crossrail” tunnels through Malmö and Stockholm here in Sweden still uses the existing old ATC system because it seemed to expensive to switch over to ERTMS where a large number of trains run.

    Btw if the GW-ATP system is a relative of the french system, afaik it’s then also a relative of the ATC system used in Sweden and Norway. Back in the days 27MHz might had seemed like a good choice. Afail the swedish system started to be installed about four decades ago, and it provides more or less all of the wanted functionality of an ERTMS level 1 LS system, except that developement has been stopped due to political decisions and therefore there are no good way of using it for speeds above 200km/h.

    Btw I’ve read about railfans listening on a 27MHz CB radio to be alerted of the noise from the train-baliser communication of the ATC system to be ready to spot the train they want to see 🙂

  101. I still wonder, if the Mayor elections were in 2019 not 2020, whether things would have been resolved in time for that.

    The issue seems to revolve aroound the signalling, both the complex new type and also the frankly excessive number of different systems in a single journey, changing systems (and thus their suppliers) every few minutes. Having an adequate test track with ALL these installed should have been envisaged years ago so the issues there were resolved at an early stage. Instead it was decided to wait till construction was complete, with just months before opening, and then use the new line itself as the test track.

    The inability to just run the key core section Paddington to Abbey Wood on its own with just one signalling system, because there is no depot on that part, and interfacing is needed to get to any depot, is surely a poor bit of overall design. Why was the signalling change at Paddington not put beyond the Old Oak depot? I suspect this one will come back round to bite them many times in the years ahead.

    I still don’t understand what the additional billions are actually going to be spent on. The various contractors were to finish their work and get paid, if they are late they should surely be getting the original money only when they finish what they were originally asked to do. There has been no increase in scope for them. There is some loss in that the trained operational staff will need to be paid without getting the revenue back, but that is not £2bn of cost.

    Quite frankly the “what did the Mayor know” is an irrelevant sideshow and does nothing to advance the opening date.

  102. Re Mr Beckton,

    Test Track – Most of (not Siemens CBTC) it is installed at the test track at Old Dalby / Ashfordby, hence they know the units are happy on ETCS and can swap between ETCS and TPWS(/+/AWS). You can’t simulate the core or local environment of the Heathrow tunnels though. What works in Leicestershire doesn’t always work in London…

    The siding at Plumstead will provide stabling for 8 units with 19 need to run 16tph Paddington – Abbey Wood Service which makes the depot (OOC) issues less critical. With trains coming into or out of service through the Westbourne Park swap over area they can come to a halt and take time to swap over so it isn’t as problematic as it sounds.

    Given that they were still installing the signalling equipment in the swap-over area just over a month ago is much more indicative of the delays and indicates that a December start wan’t going to realistically happen a very long time ago (unless they believe in cross fingers as a solution)

    There were cost over run and variations which CR appeared to handle well earlier on because they were spending less on some things due to delays, the cash was used elsewhere to cover the extras accumulating in other areas so all looked well. The bills for the previously delay work are still coming in but the budget has already been spent.

    Take the example of the PEDs. Many stations were behind schedule so weren’t ready for them to be fitted. The supplier has all the labour contracted to turn up on date A to start fitting but can only start on date B when they then find the previous work hasn’t been completed to the agreed designs so the doors can’t be fitted everywhere, they return on date C get a bit more done than another iteration. The suppliers labour costs (and assumed efficiencies from having a clear run) are way off so as per contact they put a claim in with Crossrail. The PEDs apparently completed 6 months after the planned date recently so a large bill will be sitting on a desk in CR’s finance department at the moment. Multiply this by a huge number of contracts. Major structural design changes at 3 station due to geology not being as instructed by Crossrail is also going to see several large bills.

  103. Re MIAM,

    Largely agree, that experience isn’t unique to Sweden we had a similar trial scheme on a quiet rural line in Wales (Cambrian) and there were lots of problems (the biggest being Ansaldo).

    “Btw if the GW-ATP system is a relative of the french system, afaik it’s then also a relative of the ATC system used in Sweden and Norway. ”

    GW-ATP isn’t, it is based on the Belgian TBL 1 system.

  104. @MIAM
    Clearly either the politics or the technology has moved on a little as the third Swedish “crossrail” (Västlänken in Gothenburg) will be constructed with ERTMS.

  105. @NGH
    “GW-ATP isn’t [a relative of the French system], it is based on the Belgian TBL 1 system”

    And yet trains seem to run between Paris and Brussels (and indeed Coquelles and Brussels) seamlessly.

    (Not all frontier crossings within Schengen go smoothly though – I have been delayed for over an hour at the Belgium/Germany border, and again on a French TGV in Switzerland, by power supply problems).

  106. @MIAM: Wikipedia seems to indicate the ERTMS in Malmo is fine, but the problems were with the trains (especially the Danish ones):

    Due to delays in equipment installation for integration between ERTMS and the older Swedish and (especially) Danish systems, there is a temporary arrangement using ATC. In 2012 (tentatively), when all trains using the City Tunnel are retrofitted, ATC will be removed and ERTMS deployed. Since ERTMS is already installed and tested, the changeover can be done overnight.[7]

  107. @ Timbeau

    They do now. But how much testing, infrastructure tinkering and rewriting software did it take to get to that stage?

  108. @MIAM – The legacy ATP system in Sweden is not limited supervision. It is a classic 1970s/80s full supervision system with full driving information available in the cab derived from intermittent track balises and loops at and around signals. LS is more limited, usually ‘silent’ in normal operation and monitoring in the background, although may indicate persistent reminders of past aspects or speed codes. It is there primarily to intervene if a driver does something wrong in response to the lineside signal aspects observed and interventions when they do occur are often less subtle than in full supervision; an emergency brake application to a stand rather than a small speed adjustment. The Swedish system was often held up as the gold standard of what ATP could achieve as I believe full (or near full) network fitment was achieved very early compared to most other railway administrations worldwide. It is analogous to a level 1 full supervion system in ETCS terms, as is the GW-ATP. TBL1 was employed as a limited supervision system in Belgium so the equipment’s use in full supervision mode in UK was a development that perhaps Belgian state railways shunned in favour of migration to a eurobalise based solution through TBL1+.

  109. With reference to @PoP’s response to Greg Tingey,

    One major reason for scrapping the first two train bodyshells of 09TS was that the cab was bigger on the later trains. I also think there was some sort of structural problem with the body shells of the first two units that would not have been easy to fix.

    The first two trains had new bodyshells but most of the rest of the components had already seen service on other trains so they are a ‘bitza’. Hence their stopping position did not follow the age related changes on the other trains ( see my previous posings about predictability of stopping positions).

  110. @Timbeau – TBL1 is extinct in Belgium today, having been superseded by the eurobalise-based TBL1+. On high speed lines, TBL2/3 is actually the same thing as French TVM430, using track circuit transmitted speed codes rather than balises, track loops or radio.

  111. Re Timbeau and Mark T,

    Just because the Belgians use the same set of initials for most ATP type signalling systems does meant they are related technologically, the number is the key to technology. e.g. families 1 & 1+, 2 and 3.
    Most lines were never fitted with TBL hence the decision to move on and also upgrade on a pathway to to full supervision (Most were fitted with Memor (similar to Croc.))

    The Belgium – France issues were sort by installing TBL2 so the Belgian were using the “same*” system as the french.
    *Almost the same…
    The Netherland – Belgium issues have been address with ETCS L2 on the HS lines.
    And the French are modifying the LGV signalling to move on from 430 to ETCS (Including HS1). Nothing to do with the firm that developed 430 being bough by Ansaldo and later Hitachi so ETCS is now the most “French” option.

  112. Re Mark,

    “TBL1 is extinct in Belgium today,” three weeks before its physcial extinction in Belgium, DfT decided to fit IEPs with GW-ATP and pay lots to get some more equipment built… 6 years after Belgium started upgrading 1 to 1+.

  113. @NGH – just a note for those who don’t know, ‘Croc’ or ‘Crocodile’ as used in MEMOR was an old (very old) French developed ‘AWS’ type system based on physical electrical contact between a wiggly metal track ramp and a wire brush under the train. Like UK AWS, it was only found at or approaching distant signals, but unlike the similar vintage GWR ATC it was not failsafe so if the trackside power supply failed the train would not receive its caution indication, which was historically given by automatically blowing the steam whistle.
    https://en.wikipedia.org/wiki/Crocodile_(train_protection_system)

  114. On that Jacobs supported / subjective pie, is that info broken down by type of work anywhere? Because as many other shave commented, for station building there really shouldn’t be any ‘subjective’ even if there is only very little time to write the report…

    On contingency time, I feel having bits of it at each stage can lead to contractors/stage PMs seeing the end of the contingency as their real deadline. So I have on occasions told different contractors/stage PMs different deadlines to hide the fact that I had put in a bit of extra contingency into the project plan here and there.

  115. It makes depressing reading. I believed after the major issues with the JLE and recognition that large scale British railway engineering projects seemed to always be delivered over budget and late, that Crossrail was a new dawn -the £6bn railway to be delivered to time and budget. But as we now know this is not the case. This also makes it me think it’ll be far harder (politically) to see other big projects getting off the ground such as Crossrail 2.

    Having experienced all the software issues with TL and our Class 700s, I also fear that the 345s are going to (continue to) have major issues. In this case Bombardier and Siemens having to integrate systems (usually a direct competitor I note) for the benefit of Crossrail spells a long process to me and one of the project’s biggest risks (my unlearned perspective).

  116. Having sat through around 3 hours of Transport Cttee webcast it seems the strategy from the Mayor is to take a legal timeline basis for his defence and to state that Sir Terry has “misrembered” what was said in the July 26 meeting.

    A lot of discussion about obvious concerns building over the year but “having to take the word of the Crossrail team”. Sadly as the room was full of politicians we got the inevitable attempts to make the Mayor out as a liar and the Mayor accusing opposition politicians of smears. We also had Shaun Bailey effectively asking the Mayor for tips on how to improve Crossrail’s governance so he could nick them for his manifesto. Sigh!

    The most bemusing aspect of the discussion was Mike Brown more than once suggesting that he might not be around for very much longer. Hard to know if we was joking or if he’s reached the point where he has had enough – especially after multiple questions based on the length of time he’s worked for LT / TfL and his expertise and why it wasn’t “obvious” that Crossrail was going to be late.

    Clearly the Transport Committee are still not happy and the session next year when Sir Terry attends is unlikely (IMO) to get us much further, especially if Sir Terry sticks to what he said to the media. That will leave a clear discrepancy between the two accounts of what was said by whom and who knew what when. More fun and games.

    Still, more intensive train testing is apparently scheduled to start on 14 Jan 2019 in the core section. There was an absolute refusal to commit to any sort of timeline for when the core may open. It’s clear there are too many uncertainties. It was also interesting that stations may open in an incomplete but safe state with work continuing to complete things. This is no change from what Crossrail were previously considering.

    Stage 4 (Shenfield line to core) would follow a few months after the core – provided the service had stabilised. Also confirmation that TfL are seeking to take on Stage 5 services (i.e to Reading) but from Paddington (surface) to the original timescale of Dec 2019. This is subject to assurances around the initial timetable and then the later transition to running into the core from Network Rail. This is all around avoiding any more “timetable meltdowns”.

  117. Walthamstow Writer,

    It could be they are threatening to make Mike Brown a scapegoat. But, by his own admission, he is not particularly a transport man and could certainly walk into a load of unrelated jobs – lets not forget he did a stint at Heathrow Airport before coming back to TfL to run the Underground.

    The trouble is that, assuming he is replaced internally, then that leaves TfL a bit thin on the ground with senior staff with a lot of experience. Plus, despite all that has happened, they have a good reputation so end up in New York, Australia etc. if there is nothing to keep them here.

    That said, Mike Brown seems to have adapted well to the commissioner’s role (partly because he is not ‘just railways’) , so he may want to stay and the Mayor may decide that he is better off with the devil you know.

    You do start to feel that Crossrail being delayed is going to have all sorts of long-term ramifications, some of which we haven’t yet anticipated.

  118. @ PoP – Well politicians only seem to be happy when they have “a head on a plate” provided it’s not their own. I note Ms Pidgeon has put out the inevitable press release saying that the Transport Cttee won’t stop until they get the truth. I suspect they’re in for a very long haul as I don’t see either side changing their position. If people won’t change their view then I don’t know what the Cttee can do – put them in court under oath??

    I don’t disagree with your assessment of MB’s ability to walk into another job outside of TfL. He’ll have a handsome pension provision given his role and length of service at TfL if he decided he just wanted to put his feet up. I actually don’t see him going soon but perhaps he is cogniscent that change may be unavoidable post May 2020 irrespective of who wins the Mayoralty. I have a feeling that if Khan wins a second term he will want a new face, likely female, at the helm of TfL. If the Tories win the Mayoralty then will want an axe-man in charge of TfL who has no great association with the organisation.

    Interesting remark re long term consequences. The Mayor was keen to talk up the prospects for Crossrail 2 but I think that’s going to be the major casualty. Beyond that I’ve not really considered other consequences. I do think the stripping away of resource and institutional learning / knowledge at TfL is going to have massive consequences in the short to medium term. However that’s only a part consequence of Crossrail hammering TfL’s finances.

  119. As we have previously discussed, part of this & related crises are the extra expenses involved, & now two more have appeared.
    Firstly – A newspaper report that the Battersea extension is going to be delayed by increasing costs. Though admittedly some of these are not of TfL’s making, it still does not look good.

    The second – was brought home to me vividly yesterday. Anyone travelling anywhere between Lea Bridge & the new Angel Road stations should look out of the window & compare the structures supporting the “old” & the “new” OHLE. The ( to my eyes ) ridiculous over-engineering & massive supports being used for the single line of wires on the new alignment, compared to the pre-existing structures says it all. The simple volume & therefore mass of the new supports must be at least four times that of the perfectly adequate structures that have been there since the 1970’s – & this will surely be reflected in the costs?

  120. @ GT – The lineside OHLE supervisor [superstructure?] on eastern Crossrail gave me similar feelings regarding frustration at ripping out serviceable gear for the new contract.

    The NLX was a deal with the US for investment in a long time problem derelict site. Since then the inducement has yielded more intense housing and now agglomeration with the likes of Apple bringing in 1100 jobs. The station design from 15 years ago is no longer fit for the expected two way peak capacity so this is an opportunity to spend more on enlargement.

    Can anyone enlighten us on the spend here
    https://www.openstreetmap.org/#map=19/51.56841/0.13250&layers=HG
    it looks like a 180 year old agricultural right of way probably field boundaries lining up with the development of Eustace Road at Chadwell Heath.
    Crossrail have built the reversing siding here and built a new rail crossing between a locked gate into an embankment further down to avoid the new switches.
    https://www.google.co.uk/maps/@51.5686047,0.1336373,150m/data=!3m1!1e3
    What perpetual rights are being served here and by whom?

  121. @Aleks Your supposed crossing on the Google Maps link is in fact a Rail Access Point. Hence the locked gate on one side only.

  122. @Rayjayk – Thanks, I wondered if it was for trackside vehicles during the reversing installation but it seemed permanent and maybe preceded it.
    There was a similar feature at Mile End sidings which was more obviously for rail access.
    https://www.google.co.uk/maps/place/Mile+End+Station/@51.526726,-0.043351,59m/data=!3m1!1e3!4m5!3m4!1s0x48761d307372760b:0x1fb31adc1221f1f1!8m2!3d51.52512!4d-0.03361

    The supervisor was finishing his career with this project after 40 years on the line. It was a quiet afternoon and I got a tour. Couldn’t follow all the detail – the return side being removed was better quality than the replacement. Unlike the Anglia example where new installation was higher spec than neighbouring gantries, the Eastern had pre-existing heavy duty original DC line structures. I understand the line has effectively been rebuilt to a specification rather than incorporating previous facilities, new OHLE, ducting, signal cabling, switches. So no legacy points of failure.

  123. If anyone wants to see over engineered structures, look at the canopies over the approaches to and in the ‘temporary’ taxi rank at Dustin, provided by HS2.

  124. @ Greg – I think it is clear that whatever issues there were on the NL extension have been settled. A paper went to the recent Programmes and Investment Cttee which set out a series of steps / agreements to resolve the dispute over increased costs. The numbers and details are not there for reasons of commercial confidentialty but I doubt TfL have incurred much extra spend that has not been compensated for via the agreement. The “delay” to the opening is also probably nothing to get overly excited about. The developer of Battersea Power Stn forced a design change to the tube station. That took time to sort out. It will also have stopped some elements of the work thus endangering LU’s ability to meet the “in service” date set in the original agreement for the extension. LU will have demanded a revision to that date – quite reasonably in my view as it didn’t cause the delay. Why should LU be subject to penalties for late completion when it didn’t cause the delay?

    Being a tiny bit speculative we don’t know if the works to refurbish the power station are on time. We also don’t know how occupied it and the new housing will be. We do know that the housing market is softening in London especially at the higher end and that sentiment amongst Chinese and Malaysian investors is not what it once was. A lot of Malaysian money has gone into the Battersea Nine Elms development so their view of prospects is important. A delayed opening to the extension may be to the advantage of several parties but that does involve a view about the future that may end up being wrong.

    One other factor is the need to co-ordinate all changes on the Northern Line. There are the closures for the Bank works then the need to reflect the new run and dwell times at Bank once the new tunnel is in use. It makes sense not to have a series of short term timetable / roster / signalling system changes. Doing one change to reflect the Bank improvements and the opening of the extension makes a great deal of sense and will no doubt save time and money given the Northern Line is one of the more complex routes to timetable and run. I don’t see the delay as being symptomatic of wider issues at TfL. If I’m wrong I’m sure someone will pop up to correct me.

  125. The Morden branch is one of the most heavily used in London; when the Bank branch is closed (40 days northbound and 117 days southbound) those trains must go via Charing Cross. No-one is going to allocate some of those paths for the new Battersea branch, so it must open after Bank blockade is over. Any delays to the Bank works therefore delays the Battersea opening.

  126. Mark Townend: Thanks. I already knew about the Swedish system but now I know a bit more about ERTMS/ETCS.

    So Level 1 LS is basically kind of Level 1 without integration of distant signals in the ERTMS installation? That seems like a rather silly way of saving money.

    Southern Heights (Light Railway): The tunne l through Malmö (and probably Stockholm) might technically be classified as an ERTMS installation, but in practice the old Swedish ATC standard balisers is used, and all trains are only equiped with the older standard (and in Malmö many of them also have the Danish standard). So in practice there is an electronic interlock which is ERTMS compatible but it is installed with Swedish ATC balisers. And with Level 1 the interlock just sends some signals to the balisers and it’s up to the balisers to conform to ERTMS or whichever standard you want. So it’s ERMTS-ish, kind of, almost but not really. 🙂

  127. Betterbee…. Yes, Euston….I was victim once again of my tablet’s spellchecker.

  128. Greg Tingey 09:33 22nd December

    The date for the Battersea extension opening has always been very vague as it was dependent on other activities. It was also always appreciated, as pointed out by Taz, that if it wasn’t going to open well in advance of the Bank closure then the sensible thing would be to delay it until afterwards. So no big deal here.

    On the subject of premature replacement of assets, one is on dangerous ground suggesting this is wrong because the asset isn’t life expired. One also has to take into account maintenance ongoing maintenance costs and the cost of an additional closure of the railway to fix it at a later date instead of doing it during a current closure. As you don’t have the figures, I would argue that it is impossible for you or me to form a sensible judgment.

    The ultimate absurdity would be not to replace filament bulbs in signals with LEDs because the filament bulbs are not life-expired and have further life in them.

    On is reminded of the story of the cost of the unplanned replacing a £5,000 part on the East Coast Main Line. Cost of part: £5,000. Cost in compensation to the operator for closing the ECML to install part: £1,000,000.

  129. @Greg: At around 22 minutes in it encounters either a Turbostar or Electrostar. I wonder what it was doing there?

  130. @Sh(LR)

    It’s an Anglia Turbostar, presumably on the short-lived Basingstoke to Norwich service (via the Kew West curve) which ran from 2000 to 2002.

  131. @MIAM-
    LS techniques are commonly used to emulate a ‘Class B’ legacy national system such as UK AWS/TPWS. There is a bit of a sliding scale between the most basic and the most sophisticated traditional systems, but they all have one or more of the following kinds of transponder placed in association with fixed signals and signage:

    1. Warning, at (or approaching) the first distant aspect encountered at braking distance from the associated red, or at an advance warning board for a speed restriction.
    2. Overspeed trap, on final approach to a stop aspect or speed restriction commencement.
    3. Trainstop, at a stop aspect.

    The behaviour varies according to individual system. In the simplest UK AWS, the caution warning only requires acknowledgement by the driver within a short time to avoid a full service braking application, while in the German PZB, the train must perform a standard speed reduction from a caution within a measured distance to avoid a brake intervention. An overspeed transponder in UK TPWS merely checks if a train is going faster than a carefully calculated ‘set speed’ on final approach to a stop aspect or speed restriction and intervenes with an emergency brake if the set speed is exceeded. No further restriction is imposed on the train behaviour if the speed trap is passed successfully so there is nothing to prevent a driver accelerating again towards the red. In PZB by contrast, a similar speed trap check is carried out on final approach, but on-board equipment remembers a speed trap was passed and prevents re-acceleration within a standard measured distance.

    Different systems also have varying levels of cab indication. Some function entirely silently in the background like TPWS and only intervene when a particular rule is broken, usually by means of an emergency braking response. Others, including PZB, include visual cab reminders of the previous transponder passed.

    Legacy LS systems typically don’t have custom site specific braking envelopes calculated for each signal according to speed and gradient profiles, nor do they provide full driving information displayed in the cab as the train remains driven to lineside signalling aspects and operators’ route knowledge. They tend to have less subtle braking interventions, initiating full or emergency brake applications instead of gentle speed adjustments in the event of transgressions. Some systems such as the Swiss ZUB do have more detailed site specific braking curves encoded in the track to train messages however. The pre-existing Signum system, first installed in Switzerland during the 1930s, was closely related to the DB PZB (Indusi), but didn’t include overspeed magnets applied on final approach to a red as in Germany, so after an accident in the 1990s, ZUB was developed by SBB and Siemens as an overlay system on approach to selected stop signals where justified on safety grounds to ensure trains could always stop before a conflict point. Danish State Railways also adopted the same ZUB equipment in its legacy protection system, now being replaced by nationwide ETCS level 2.

    The distant signal is arguably the most important signal as it indicates a last braking commencement point if you are bowling along at speed and are going to have any chance of obeying a stop signal at red protecting a station or junction. That is why distant signals received the very earliest attention in driver assistance technology with the very early innovations of French crocodile and UK GWR ATC (not to be confused with the much later GW-ATP!). The LS approach as applied to ETCS definitely does have distant signal integration, but typically employs legacy generic responses to caution aspects rather than applying specific local braking curve supervision as featured in L1 FS (full supervision). However I would argue there is a sliding scale between ‘limited’ and ‘full’ supervision in the level 1 arena, and each legacy national system takes its own unique position along that scale.

    Once a national system is migrated as a Class B system into the ERTMS environment on board, whether using STMs and retaining the legacy transponders, or with those replaced by standard eurobalises, euroloops, radio infill units etc, the protective behaviour could be enhanced incrementally as desired using the EVC’s inbuilt ability to impose (say) an ongoing standard speed reduction or restriction within a measured distance for example. As a notional development, further fixed route data information could be introduced into balise messages to allow ETCS to calculate more detailed location specific behaviour where justified, just as already incorporated in Swiss ZUB and its ETCS-based ‘Eurozub’ emulation. In this way it could be possible for all LS systems to migrate towards the combined best features of all previous national legacy LS systems, most of which I believe are were already included in DB PZB and SBB Signum/ZUB systems.

    These legacy systems are not the only ones to be created using ETCS components and protocols. In Berlin, mechanical train stops on the S-Bahn are being replaced by Eurobalises in a new Class B train control system called ZBS. Unlike the national main line legacy system emulations, this is meant to be a long term solution, not a medium term migration path to a later ‘standard’ Class A ETCS level. Trackside and and on-board environments use standard ETCS components, but with a custom configuration and software package. S-Bahn trains could be modified at some future date to also be compatible with a Class A level if some shared running with other trains was ever desired, or a particular fleet of standard main line ETCS trains could be modified to also work on ZBS equipped lines, both using STM emulation techniques, and in this case not requiring any additional plug-in antenna units etc on board because both systems would already be using the same transponders.

    From your description of the Swedish examples, it seems STM methods are already in use on board to emulate the older Swedish ATC system in modern ‘native-ERTMS’ trains.

    ERTMS is an exciting field today as it represents far more than the standard series of Class A levels. It is now seen by engineers as a toolkit and framework within which there are practically no limits as to what kind of functionality could be built into custom trackside and on-board systems, all of which must neccessarily fully expose and document their track to train interfaces within standard message protocols and between standard hardware components: WCML TASS, Thameslink ATO, European LS initiatives, Berlin ZBS etc are all examples of this.

  132. Update on my post of 24th December …..
    Docklands & Stratford, 7 Sisters & on towards Palace Gates in about 1959
    With STEAM TRACTION ……
    As seen here

  133. @Timbeau I do miss that option. The Anglia Basingstoke to Norwich service which was really useful as it stopped at several places in North London.

  134. @Kevin Roche – but deadly slow- and hardly used at all as a through service

  135. GH: agreed, but very convenient if (as happened to me during its short period of operation) you were lucky enough to want to travel from Basingstoke to Stratford.

    KR: I recall that its eastern terminus was Ipswich rather than Norwich.

  136. @betterbee I did indeed use it once during a tube strike, to travel from Woking to Highbury (and thence to Moorgate) and sat solitary state in the 1st class sipping the free coffee whilst the NLL hordes pressed their noses against the window of the saloon. A pleasant run through some of the more interesting bits of Egham and West London. Took a couple of hours. Didn’t have the time to do it daily. Not a good use of line capacity…

    Yes it originally started back from Ipswich but as I recall was fairly quickly truncated to Chelmsford.

  137. I used the Anglia service just once, for a work related trip between Feltham and Basingstoke I think. It wasn’t planned at all. The train just appeared on the information screens while I was waiting and I elected to go that way instead of via Reading.

  138. I also used the Anglia service once, as a planned trip from Ipswich to Feltham, for Heathrow. It was an interesting tour but, as a journey to the airport, far less efficient than the “normal” Liverpool St, Paddington and Heathrow Express route I normally took.

Comments are closed.