Mixed Signals: A Small SSR Press Release With Big Implications


On 24th March 2015 there was a surprise announcement from the Mayor and TfL about the re-signalling of the Subsurface Railway (SSR). It is largely a re-announcement of what has already been announced numerous times, and it is far from free of doublespeak, but it does also contain some genuinely new information. Unfortunately, despite appearances, most of that information is not really good news for the travelling public of London.

The difficulties of resignalling

Resignalling tends to be a fraught process. It is generally a lot simpler to start from scratch. In London we have the problem of not building sufficient numbers of new railway lines and a rapidly expanding population. Improving capacity therefore leaves only a few options – increase train capacity, increase train frequency or both.

When it comes to London Underground, there are severe limits as to how train capacity can be increased and unfortunately most of the existing options have already been taken up. That really only leaves increasing frequency and to do that one needs to push to the limit of, or even beyond, current signalling technology. Consequently from the building of the Victoria Line onward it has generally been a real challenge to provide a signalling system that is truly fit for purpose in London. The exception is stage 1 of the Jubilee Line, where signal engineers had the rare luxury of being able to install tried-and-tested state-of-the-art equipment and being very confident it would work and provide the necessary capacity.

20th century resignalling

When opened the Victoria Line was supposed to be able to handle 30tph but at best managed 27tph on a good day until it was resignalled in the 21st century – and this for a line that could hardly be simpler to operate. The Jubilee Line Extension, meanwhile, was supposed to have a signalling system capable of running 36tph on opening. Not only did that prove impossible, but the intended system had to be ripped out and conventional 24tph signalling installed in order to beat the December 31st 1999 deadline for the full opening of the extension. Whilst past results are never a reliable indicator of future performance, the late 20th century omens for reliable 21st century developments in signalling were not good.

21st century resignalling

As one century moved into another, the Central Line automatic train resignalling was completed as the final part of a package to replace trains, signals and power supplies. The combination of the three meant many years before the system really settled down and worked properly. Meanwhile the second generation of automatic signalling on the DLR initially fared little better.

Having moved fully into the 21st century, we finally saw a resignalling that was reasonably successful – that of the Victoria Line – although at roughly the same time Thales were struggling to get their implementation working on a resignalled (again!) Jubilee Line. This was the era of the Public-Private Partnerships (PPP), described with considerable justification as “one of the scandals of the decade”, and as such Thales had little incentive to develop the system properly off-line. Instead the Jubilee Line was effectively used as a test bed to develop their new signalling system – good for Thales, but extremely bad for Londoners as this was the primary reason for its frequent closure on weekends. This experience led to London Underground making noises at the time about never again buying a signalling system that hadn’t been properly tested before being installed.

Thales – bad boy made good

The surprising thing for those who despaired about the Jubilee Line is that once Thales did get their system working it actually worked quite well. Indeed, when they installed it on the much more complex Northern Line, installation and commissioning went better than anyone had dared hope.

To cancel one signalling contract …

It is fair to say that so far the story on the Sub-surface Railway – the District, Metropolitan, Hammersmith & City and Circle – has been more tortuous. Originally the work was to be carried out through the PPP mechanism. When that collapsed TfL took the opportunity to cancel the signalling contract entirely. They had long felt it contained some serious weaknesses but under the PPP regime they had not had the power to intervene.

The contract was re-tendered and won by Bombardier. The system proposed appeared to be well-established and TfL must have thought they were taking the safe option. Issues, however, soon emerged. One problem, not picked up at the time, was that the metro systems on which it had already been successfully installed weren’t actually that complex. TfL also later claimed that Bombardier had overstated what it could do – claims they had felt at the time that, given Bombardier’s status as a well-established train manufacturer, they could rely on.

We have reported at length on the fracas that led to the decision to pull the Bombardier contract. As the sadly now-dormant Railway Eye website put it at the time:

To lose one signalling supplier, Mr. Brown, may be regarded as a misfortune; to lose two looks like carelessness.

With hindsight, it is hard to criticise the decision to cancel the Bombardier contract as, once work had begun, it did look hopelessly optimistic as to what could be achieved. The project failed to meet the most basic of predetermined milestones in the early stages of implementation, which was always going to cause considerable alarm in an organisation still bearing the battle-scars of the Jubilee Line. That said, the decision to award it in the first place is something that, in retrospect, did seem to involve a level of naivety and misguided trust.

Brave Face

In trying to be as positive as possible about what had happened, TfL expressed optimism that another supplier could be found and the original deadline met (or at least only narrowly missed). This was inevitably treated with much scepticism by the knowledgeable railway press, such as Roger Ford in Modern Railways or in Railway Gazette. With only one potential supplier (Thales) a serious contender it wasn’t hard to see where the balance of negotiating power really lay. Brave talk of financial penalties for failing to meet deadlines seemed a nonsense. What firm was going to sign up for that?


It was likely to have been the case that, privately, most people involved within London Underground were resigned to the SSR not being fully resignalled until 2019 or 2020. These were dates that seemed far more plausible and achievable. It therefore comes as quite a shock to find out that clearly Thales have played their hand rather strongly, with talk now of completion by 2022. Less surprising is that there has been an increase in costs, although the amount raises eyebrows – an almost staggering £1.3 billion, as far as one can tell all related to increased signalling costs. What would be really encouraging to learn is that the increase is partly accounted for by including major sections of the Piccadilly Line in this contract, as once envisaged. It seems highly unlikely though or it would have been mentioned as a way to sweeten the public pill.

Not definite yet

It is also slightly concerning that buried in the press release is the statement:

LU is in the final stages of negotiating a position with Thales

So a deal hasn’t actually been done yet. One hopes that there aren’t shades of the Thameslink Rolling Stock supplier (Siemens), where the announcement pre-dated the final agreement by a full two years. The SSR resignalling announcement seems to have been made when it was because:

The Mayor and London Underground (LU) today confirmed detailed plans, including timescales and budget, to ensure the delivery of the much needed modernisation of the next four of the network’s lines.

The problem they had is that we will soon enter purdah due to the general election, where such announcements are generally frowned upon as being electioneering motivated. So if an announcement were to be made it seems that it is better to do it now. This still leaves the issue of why it has been so boldly announced when, basically, it is not good news. Perhaps the goal was simply to get out in front of the inevitable story with a positive spin.

Surprised? We shouldn’t have been

In retrospect perhaps we shouldn’t have been so surprised at the announced change in completion date. Some clues were there. First of all, London Underground unexpectedly pressed ahead with improving the off-peak SSR timetable despite the fact that previous indications were that this would wait for the signalling upgrade. More recently there was the reported cost escalation of the Croxley Rail Link and the suggestion that included in the costs were the initial signalling and a subsequent resignalling as part of the SSR upgrade. With the rail link not now due to open until 2018, if it gets final approval, it would not have made sense to have to resignal it within a matter of months of opening when construction could have easily have been delayed by short period if that would have saved a lot of money.

Another clue was buried in the latest Commissioner’s Report and easily overlooked even if one does trouble to read this document in detail. In it, he refers to:

The work is also preparing the ground for a new high-speed ‘scissors’ crossover west of King’s Cross, due to be installed next year. … When complete, when we need to, we will be able to reverse more than 20 trains an hour east to west from the two King’s Cross platforms. This is more than double the capacity available today.

Like the Croxley Rail Link it would not have made much sense to go to the expense of installing and signalling this crossover next year if it were to be followed around a year later by the resignalling of that section of the SSR.

There also continues to be extreme vagueness as to when the SSR will be part of the night tube, and the February joint announcement by the Mayor and the Chancellor about extending the night tube to the East London Line and the DLR only referred to it being extended to the SSR when signalling upgrade is complete – with no hint as to when that would be. If it were good news one would have expected a tentative date.

The classic TfL Mayoral press release

The press release follows a standard formula. The SSR upgrade as a whole is discussed as if this is a promise for the future despite the fact that much of it – such as new trains, lengthened platforms and rebuilt depots have either been completed or are already close to completion. The 34tph on the Victoria Line gets its usual obligatory mention. Not only that, this is done in a very misleading way:

The plan uses an even further improved version of the system built by Thales, now successfully operating on the Jubilee and Northern lines, and also follows on from the Victoria line modernisation where, with 34 trains every hour in each direction, there now is one of the most frequent train services anywhere in Europe.

This almost implies that the Victoria Line has a system built by Thales. In fact the exact opposite is true and the line with “one of the most frequent train services anywhere in Europe” is not using a Thales signalling system. Had this had been made clear a very obvious question could be asked about the ideal signalling system for the SSR.

As well as the Victoria Line we get the inevitable statement about the rising population of London, although the normal comment about the population of London going up by the equivalent to a tube train full of passengers every three days is, refreshingly, avoided.

Even the increase in cost gets a positive spin, being compared favourably with the cost of resignalling under PPP.

the firm expectation is that the new price will be in line with, or below, the cost per kilometre of modernising the Northern line signalling

It’s not actually a statement that reassures. Whilst costs would be expected to be high in some of the central subsurface sections, with their many complex junctions, one would not expect this to be replicated on the line on the surface out to Upminster, Uxbridge or Amersham. There are an awful lot of rural miles there that should make it easy to bring the cost per kilometre down below what it was for the Northern Line.

The good bit

Whilst, as we shall see, this press release raises an lot of issues, there are some more encouraging aspects of this apparent near-agreement. The first is that it does seem to be more realistic than anything put forward before.

The press release refers to:

additional infrastructure works and costs identified as necessary to the modernisation following the termination of the Bombardier contract

It also recognises that the existing Thales system is not adequate for the complexity of the SSR as it states:

The plan uses an even further improved version of the system built by Thales.

A lesson learnt, no doubt, from the failure of the Bombardier offering. So, hopefully, fewer nasty surprises.

This realism extends to dates and perhaps a good argument could be made that once you miss the 2018 deadline then Crossrail will be open and the pressure is off to improve SSR services. It then makes more sense to get this right and take advantage of a relative lull in SSR demand – on the northern side of the Circle if not the southern side.

The bad bits

Most telling of all in the press release is the statement that:

customers will start to see the benefits of the work on the Circle line in 2021, with customers experiencing the full benefits across all lines in 2022. Once these four lines have been completed, LU will then move on to buying new trains and control systems for the Piccadilly, Central, Bakerloo, and Waterloo & City lines.

The true ramifications of this should be plain to see. If true, this means that there will be a substantial further delay to New Tube for London (NtfL) programme with all the consequences that will bring. Currently, according to the latest published Fit for the Future plan, installation of the new Piccadilly signalling will commence in mid 2019. If TfL are only going to start buying new control systems once the SSR upgrade is complete then that looks like a three year delay, although time could be made up by the resignalling and delivery of new trains going hand in hand. Given the interfaces between the Piccadilly and the SSR it makes sense to hold off resignalling the Piccadilly until the SSR is complete. It will also make it easier for Thales to bid for the work. If nothing else, this fiasco has shown the benefit of having more than one serious contender in the bidding stakes with more than zero being an absolutely vital prerequisite.

Piccadilly Line: The old unrefurbished train for London

We have recently reported on the major work necessary to keep the Bakerloo Lines running. In a similar vein there is a two stage plan to first upgrade the Central and Waterloo & City Line trains before giving them yet a further revamp with new traction motors. What is missing is any kind of major refurbishment for the Piccadilly line trains despite, it seems now, the fact that all of them will need to remain in service for at least another seven years. It does seem surprising and slightly worrying that this has not had a mention anywhere. With seven or more years of useful life needed from the stock one would have thought some kind of refurbishment programme was highly desirable. It is notable that the District Line D78 stock got a final significant refurbishment, despite at the time only being expected to be in service for another seven years at most.

Further knock on consequences follow. The Mayoral promise to stop Piccadilly Line trains at Turnham Green throughout the day, dependent on Piccadilly Line resignalling, seems to be receding further into the future. Supporters of the Bakerloo Line to Hayes will find it increasingly difficult to plan for this being completed before 2030 and certainly not much earlier given that the replacement for the Bakerloo Line stock will follow the (seemingly further delayed) Piccadilly Line replacement stock.

As delays to the Tube programme mount up it is going to get harder and harder to mitigate against this. Of particular concern is the fact that the further delays seem to be at least partially prompted by finance and cash flow considerations. If this is the case then that further limits ones options to spend one’s way out of trouble.

The curious bits

One very strange part of the announcement was a remarkable increase in the frequency of the Circle Line. This could have been a misprint, but for it being mentioned twice.

The system is needed to allow more trains to run – more than double the number will run on the Circle line alone

Once completed most Circle line customers will see a train up to every 4 minutes instead of 10

This clearly suggests that the agreed plans are different to what was originally proposed. It is hard to see how the Circle Line can have such a frequency unless, with a bit of tinkering, it absorbed the Hammersmith & City Line – which would have its eastern terminus swapped with the Metropolitan Line. However, even this does not seem possible because the press release refers to the Hammersmith & City as a separate line after the upgrade, so it will presumably continue to exist.

So either there has been a complete rethink and something quite radical is proposed or someone writing a press release has got very confused as to what will happen. If the Circle Line did truly run on the south side of the Circle every 4 minutes then that would leave around only 15tph on the District Line to serve Ealing Broadway, Richmond and Wimbledon – something that just does not seem plausible.

Too many trains too soon

A further consequence of the delayed resignalling is that it would appear that there will be considerably more S Stock trains built than can initially be used – something that also happened on the Northern Line and has only very recently been rectified. This will amount to a considerable sum of money and trains will to some extent deteriorate, or at least become dated, whether used or not. No mention is made of this but one wonders if there is scope for rescheduling delivery of the final batch of trains. As well as delaying payment, not getting new trains for the signalling upgrade until they were needed would also mean that any relevant technological developments could be taken advantage of at a later time – maybe even taking advantage of early benefits of the NTfL programme.

The opportunities for delaying delivery of the last of the S Stock seem limited and yet it could work out. Currently deliveries of S7 stock have slowed to a trickle, or even a halt, as Bombardier concentrates on lengthening the class 378 stock for the London Overground. They also have the contract to build the Crossrail trains and there may well be others in the pipeline. So it could possibly suit both London Underground and Bombardier not to build the final S7 trains until the signalling is ready for it.

Of course, London Underground could take the attitude that Bombardier more than owe them a favour, especially as it was their failure to produce a workable signalling system that caused the problem in the first place – but business doesn’t work like that.

It can’t be as bad as all that – can it?

Perhaps in the very long term things are not too bad. The more the New Tube for London gets delayed, the better it should be when it finally arrives – and it does have what is currently a very challenging technical specification. And as previously discussed, the more NTfL gets delayed, the more it becomes a logical successor to Northern and Jubilee Line stock which will be at the point of being due for replacement by the current end of the NTfL programme.

If the Piccadilly Line trains survive in service until their replacement arrives and no other great disaster befalls London Underground then it could all work out very well in the long term – but it is fair to say it will be as much through luck as through forward planning.

Written by Pedantic of Purley