Our new ticketing site is now live! Using either this or the original site (both powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!
No update on what will happen to descoped parts (District south of Fulham area, west of Stamford Brook), and Ealing depot (sometimes called Ealing radio island, as it allowed testing before exiting depot).
Metropolitan will be completed as its signalling equipment is older, and patching it up for years is likely to cost more than finishing the replacement.
Eventually the descoped sections will end up with equipment which has become non standard (as it will be relatively small part of network), and probably be harder to get spares and maintain. But for time being seems to be in the defer for next 5, 10, 15, 20 years (or whenever) category.
Thanks, this brings sense to some things that had me wondering:
SMA12's Wimbledon Depot interactions certainly cause unique headaches at East Putney; however
SMA10's Overground interactions beyond Stamford Brook don't seem inherently different to SMA9's & SMA13's Chiltern interactions (Tripcock-fitted Class 165s notwithstanding)
SMA11's Piccadilly interactions appear identical to SMA14's Piccadilly interactions
Age and serviceability of existing equipment definitely makes the difference. Perhaps there is also the long-rumoured prospect of swapping the Piccadilly from Uxbridge to Ealing Broadway, causing further deferment of SMA11 behind the scenes.
The key difference between SMA9/13 and SMA10/12 is that the Met is wholly owned and operated by LUL, and Chiltern trains run over LUL infrastructure. Therefore, CBTC can provide the primary interlocking (‘underlay’) which drives the lineside signals. Technically, this is relatively simple to achieve and something LUL have total control over.
For the Wimbledon and Richmond branches this obviously isn’t the case: signalling is provided by Network Rail, and so a CBTC ‘overlay’ is (was) required that is derived from the existing signalling. Given the nature of these assets - and the more limited abilities to access them - this is a much more drawn out process, and there were also some contractual issues between LUL, NR and Hitachi/Thales, all of which was for relatively little overall gain.
The complexity of the Acton Town area was the main reason for descoping SMA11 - whilst technically the same as SMAs 9, 13 and 14 (CBTC underlay) it becomes of limited additional value given the restrictions imposed by ‘legacy’ (Piccadilly) rolling stock, and the fact that the S7 fleet will still require tripcocks to be fitted and maintained given the descoping of Wimbledon and Richmond. It is, however, simply pushing the problem down the line - unless the Picc receives the same flavour of CBTC (unlikely) there will almost certainly be further issues on the same scale as seen by the Wembley Park/Neasden mixed mode area where two/three different signalling systems all have to interface. What has been implemented can best be described as a bodge.
Finally, as already mentioned, the legacy kit in these areas isn’t as life expired as that on the Met. The main pusher is to allow Earl’s Court control room to completely close, removing some significant unreliability there. As mentioned, Wimbledon and Richmond would have retained the existing interlocking anyway, and most of the unreliability of the Acton Town area was resolved by the Piccadilly Interim Control Upgrade (PICU) project a few years ago, which introduced a new control room and signalling interface to the Piccadilly and west end of the District.
And all of this overlay complexity etc. as you've described brings about the biggest advantage of Invensys / Siemens DTG-R, compared with the other two systems, in that it is a "Protection and Authority" system, that can sit atop of, realistically, any interlocking system. (But works best with the likes of Westrace and FS2550 onwards).
All of the lessons learned on VLU were to be carried onto the SSL, but alas, TfL wanted to prove how amazing at programme management it is, and we all know how well that went.
VLU = Victoria Line Upgrade
SSL = Sub Surface Lines
FS2550 = Track circuit / train detection system.
The key difference between SMA9/13 and SMA10/12 is that the Met is wholly owned and operated by LUL, and Chiltern trains run over LUL infrastructure. Therefore, CBTC can provide the primary interlocking (‘underlay’) which drives the lineside signals. Technically, this is relatively simple to achieve and something LUL have total control over.
For the Wimbledon and Richmond branches this obviously isn’t the case: signalling is provided by Network Rail, and so a CBTC ‘overlay’ is (was) required that is derived from the existing signalling. Given the nature of these assets - and the more limited abilities to access them - this is a much more drawn out process, and there were also some contractual issues between LUL, NR and Hitachi/Thales, all of which was for relatively little overall gain.
Is this something eventual transition to ETCS could help with, or would it ultimately need the branches transferring to LUL, being resignalled to CBTC and then providing lineside signals for non-CBTC fitted trains?
That is not how I interpreted Bluegoblin's post. #446 mentions 10, 11 and 12, apparently in response to post #445 enquiring which SMA are going to be indefinitely deferred.
Is this something eventual transition to ETCS could help with, or would it ultimately need the branches transferring to LUL, being resignalled to CBTC and then providing lineside signals for non-CBTC fitted trains?
There are two ways this can ultimately progress, one is an trackside overlay, effectively a system that copies the signalling into a form the onboard CBTC can read, even if it is fixed blocks. Ultimately all it needs to know is section ahead clear, occupied, or caution so need to slow down ahead of next stop point.
The other way is the trains get ETCS with a converter, so they can pick up ETCS signals and convert them into CBTC
Either way, there are the two extremes; stop as in red signal, clear to proceed at line speed as in green signal. The difference is in the number of intermediate steps between these.
The Waterloo suburban area was resignalled at end of 1980s, so ETCS is unlikely to come to the area for many years. And likelihood of all trains in the area using ETCS and line side signals removed seems unlikely before 2040s at earliest, so first solution seems most likely.
Not particularly - because then you’d need to fit ETCS to the S stock, for which I suspect there isn’t the physical space. Seltrac is an ‘all in one system’ and really isn’t designed to interface with other systems, so ultimately there would be the same challenges if you had anything other than a Seltrac underlay.
Given the Wimbledon branch is owned, almost in its entirety, by LUL it isn’t beyond the realms of feasibility that when the existing signalling becomes life expired a wholesale resignalling to Seltrac happens, with added colour lights as per the Met for SWR trains. The same isn’t true for Richmond, but given the chronic unreliability of the Gunnersbury area I know this is on TfL’s radar for a more permanent solution.
In both cases it will ultimately come down to the cost difference between retaining and maintaining train-borne tripcocks on a permanent basis, and all the associated circuitry (the various latch relays were only ever intended for a temporary transition and cause a reasonable amount of the various ‘defective train’ situations) and providing an ATO system, of whatever flavour, when resignalling is needed.
ETCS and commonality have a lot of benefits, but for a (largely) closed network with trains of almost identical characteristics a proprietary ‘complete control’ product still makes a lot of sense. A former senior manager at LUL once likened Seltrac to “buying a complete railway” rather than just a signalling system. It isn’t quite accurate but it’s not a bad analogy for the lay person.
== Doublepost prevention - post automatically merged: ==
That is not how I interpreted Bluegoblin's post. #446 mentions 10, 11 and 12, apparently in response to post #445 enquiring which SMA are going to be indefinitely deferred.
Different ways of presenting the same information - I listed only the boundaries, beyond which have been descoped (ie, a permanent change - deferred suggests they’re happening later. They might, but not as part of 4LM!). @Dstock7080’s post is the ‘other side’, I.e. what will be commissioned.
That is not how I interpreted Bluegoblin's post. #446 mentions 10, 11 and 12, apparently in response to post #445 enquiring which SMA are going to be indefinitely deferred.
If I remember correctly, and can't look it up at moment, one of the areas, and I think it is SMA12 has a partial implementation, part of route is going ahead, the end that adjoins another SMA area, but about 60-75% is deferred. This might be the confusion.
The Waterloo suburban area was resignalled at end of 1980s, so ETCS is unlikely to come to the area for many years. And likelihood of all trains in the area using ETCS and line side signals removed seems unlikely before 2040s at earliest, so first solution seems most likely.
Correct, although Waterloo approach is an area that would benefit strongly from the capacity gains of ETCS so may be done at the front end of life-expiry rather than being strung out.
Not particularly - because then you’d need to fit ETCS to the S stock, for which I suspect there isn’t the physical space. Seltrac is an ‘all in one system’ and really isn’t designed to interface with other systems, so ultimately there would be the same challenges if you had anything other than a Seltrac underlay.
Given the Wimbledon branch is owned, almost in its entirety, by LUL it isn’t beyond the realms of feasibility that when the existing signalling becomes life expired a wholesale resignalling to Seltrac happens, with added colour lights as per the Met for SWR trains. The same isn’t true for Richmond, but given the chronic unreliability of the Gunnersbury area I know this is on TfL’s radar for a more permanent solution.
In both cases it will ultimately come down to the cost difference between retaining and maintaining train-borne tripcocks on a permanent basis, and all the associated circuitry (the various latch relays were only ever intended for a temporary transition and cause a reasonable amount of the various ‘defective train’ situations) and providing an ATO system, of whatever flavour, when resignalling is needed.
Given the low level of NR traffic on the Wimbledon Branch I can see the CBTC with signalling overlay being used there.
Richmond is trickier as there's a passenger service from both sides at reasonably high frequencies. As there's no connection at Richmond between the Twickenham Line and the LU side any more, if the branch was converted to Seltrac with Signalling overlay then there'd only be one transition point to manage for NR trains between South Acton and Gunnersbury.
I guess it'd be technically possible but extremely expensive to have Seltrac interlocking controlling a lineside signalling overlay fitted with ETCS L1 protection. Then LU wouldn't need to transition and NR would only need to transition from ETCS L2 to L1 on the connecting chord. That or send LO to Kew...
ETCS and commonality have a lot of benefits, but for a (largely) closed network with trains of almost identical characteristics a proprietary ‘complete control’ product still makes a lot of sense. A former senior manager at LUL once likened Seltrac to “buying a complete railway” rather than just a signalling system. It isn’t quite accurate but it’s not a bad analogy for the lay person.
Given the Wimbledon branch is owned, almost in its entirety, by LUL it isn’t beyond the realms of feasibility that when the existing signalling becomes life expired a wholesale resignalling to Seltrac happens, with added colour lights as per the Met for SWR trains. The same isn’t true for Richmond, but given the chronic unreliability of the Gunnersbury area I know this is on TfL’s radar for a more permanent solution.
In both cases it will ultimately come down to the cost difference between retaining and maintaining train-borne tripcocks on a permanent basis, and all the associated circuitry (the various latch relays were only ever intended for a temporary transition and cause a reasonable amount of the various ‘defective train’ situations) and providing an ATO system, of whatever flavour, when resignalling is needed.
ETCS and commonality have a lot of benefits, but for a (largely) closed network with trains of almost identical characteristics a proprietary ‘complete control’ product still makes a lot of sense. A former senior manager at LUL once likened Seltrac to “buying a complete railway” rather than just a signalling system. It isn’t quite accurate but it’s not a bad analogy for the lay person.
== Doublepost prevention - post automatically merged: ==
Different ways of presenting the same information - I listed only the boundaries, beyond which have been descoped (ie, a permanent change - deferred suggests they’re happening later. They might, but not as part of 4LM!). @Dstock7080’s post is the ‘other side’, I.e. what will be commissioned.
Thanks for the clarification. It wasn’t obvious (to me at least) that you meant only part of those SMAs are deferred/descoped.
I would hardly class the District as a largely closed network, as the majority of services start/end on NR territory. ETCS was designed to resolve exactly the kind of interface issues that LU has landed itself with by going for a non-standard signalling system that does not readily interface.
Is the Wimbledon branch owned almost entirely by LUL? I thought that the half between East Putney Jcn (where the chord from the Windsor lines joins) and Wimbledon was NR owned. It was certainly owned by BR when I drew the AWS scheme plan for it.
I see the Wimbledon branch as the bigger problem. I doubt that NR/SWR would be happy to give it over to TfL as things stand, due to the depot access issues. It is probably not feasible to retrofit all SWR trains that might need to access the depot with CBTC. It might be possible to separate the depot access from the Wimbledon branch, but this would require significant remodelling, which I doubt TfL would want to pay for. It might be possible to fit a CBTC overlay on the existing signalling, but as a previous poster has noted, in view of the age of the signalling and the complexity of the layout, this could be difficult and expensive. I suspect that this isn’t going to be solved until NR next want to resignal/remodel the area.
There is another issue with overlaying CBTC signalling on NR territory: who owns and maintains it? I doubt that NR would accept LU technicians or contractors unaccompanied on its territory. If NR technicians have to maintain it, then it would need NR maintenance and faulting technicians trained on it. Which LU would likely have to pay for, on an on-going basis. NR are also likely to insist that any new signalling installed on their lines must at least be "ETCS compatible" or "ETCS ready".
I see the Richmond branch as the easiest to resolve, as all services running on that line are under TfL control (either District or Overground). The ideal solution would be for TfL to control the branch from the District line control room, with CBTC signalling with colour-light overlay for the Overground trains. However, for the above reasons, this would probably requiring TfL taking over ownership of the line from NR, which would be outside the remit of a mere signalling scheme.
While the Richmond branch may be a lot easier to resolve, there is little point in spending a lot of money doing so unless the Wimbledon branch can also be resolved, as trains will still need tripcocks until BOTH branches are dealt with. Resolving both branches is likely to involve spending significant sums of money - and what will the public get for it? As it will be hard to justify this spend, I can see it being a long time before these branches get tackled. Probably not until the Wimbledon area comes up for resignalling and/or remodelling, which as Snow said, may well not be for a couple of decades.
Is the Wimbledon branch owned almost entirely by LUL? I thought that the half between East Putney Jcn (where the chord from the Windsor lines joins) and Wimbledon was NR owned. It was certainly owned by BR when I drew the AWS scheme plan for it.
There is a RAIB report within the last decade which addresses historic confusion about where the track maintenance boundary was. IIRC it's not the same location as the signalling boundary, nor the legal boundary, but all are quite a bit further south than the East Putney connection. SWR has a protected right to run trains on the LU bit.
Is the Wimbledon branch owned almost entirely by LUL? I thought that the half between East Putney Jcn (where the chord from the Windsor lines joins) and Wimbledon was NR owned. It was certainly owned by BR when I drew the AWS scheme plan for it.
The line and stations (except Wimbledon) were sold to LUL in 1994 as part of the process for privatisation. The arrangement itself is rather complex in that the civils and track is owned and maintained by LU, to LU standards, however much, but not all, of the signalling and power assets are owned and maintained by NR.
There is a RAIB report within the last decade which addresses historic confusion about where the track maintenance boundary was. IIRC it's not the same location as the signalling boundary, nor the legal boundary, but all are quite a bit further south than the East Putney connection. SWR has a protected right to run trains on the LU bit.
Isn't the Richmond Branch planned to be resignalled to Upminster IECC North London Line workstation along with Acton Wells Signal Box? So, wouldn't that be the best time to install ETCS for NR and TBTC for LUL?
Isn't the Richmond Branch planned to be resignalled to Upminster IECC North London Line workstation along with Acton Wells Signal Box? So, wouldn't that be the best time to install ETCS for NR and TBTC for LUL?
I am not sure how practical it would be to have both ETCS and CBTC operational on the same line, with the interlocking talking simultaneously on two different information flows. Any solution is likely to be a novel, custom-built one-off, with extremely high development and validation costs.
If NR are responsible for resignalling the Richmond branch, they will want to resignal it with the same type of signalling as the rest of the NLL, compatible with their equipment at Upminster, and either ready-fitted with ETCS or at least ETCS-ready. How easy it would be to overlay CBTC on that is the big question (see Bluegoblin's previous posts).
I cannot see NR being keen on equipping the Richmond branch with a signalling system based on CBTC, that would likely be incompatible with the existing signaller workstations at Upminster, and that its techs would be unfamiliar with.
I still see the best solution being to signal the branch with CBTC compatible with the rest of the District line, with the NLL Overground trains transitioning to and from CBTC. If the NLL trains are fitted with ETCS, it should hopefully be possible to fit their onboard ETCS with a CBTC national mode. However, signalling the Richmond branch with CBTC-based signalling would probably require TfL to take over the branch.
I am not sure how practical it would be to have both ETCS and CBTC operational on the same line, with the interlocking talking simultaneously on two different information flows. Any solution is likely to be a novel, custom-built one-off, with extremely high development and validation costs.
If NR are responsible for resignalling the Richmond branch, they will want to resignal it with the same type of signalling as the rest of the NLL, compatible with their equipment at Upminster, and either ready-fitted with ETCS or at least ETCS-ready. How easy it would be to overlay Seltrac CBTC on that is the big question (see Bluegoblin's previous posts).
I cannot see NR being keen on equipping the Richmond branch with a signalling system based on Seltrac CBTC, that would likely be incompatible with the existing signaller workstations at Upminster, and that its techs would be unfamiliar with.
I still see the best solution being to signal the branch with CBTC compatible with the rest of the District line, with the NLL Overground trains transitioning to and from CBTC. If the NLL trains are fitted with ETCS, it should hopefully be possible to fit their onboard ETCS with a CBTC national mode. However, signalling the Richmond branch with CBTC-based signalling would probably require TfL to take over the branch.
The Richmond branch falls under Network Rail's remit for re-signalling, as it is not just signalled but also owned by Network Rail. Network Rail would like to do a straight conversion from colour light to ETCS on the North London Line in the 2030s-2040s. According to ' The PWI video on West London Orbital ', a signalling system has a design life span of at least 30 years. The rest of the North London Line was done in 2015; however, Acton Wells Signal Box (AW) and Richmond Signal Box (GB) were dropped from it, which puts you in a tight spot with two main options, which are to do a re-signal Acton Wells & Richmond with colour lights for 15 years of use or make it a priority for ETCS and fit that as soon as possible.
For ETCS mainline trains, CBTC is bolted onto the ETCS as part of the ETCS NTC system.
NR has permitted CBTC to be installed on their tracks, for example, between Pudding Mill Lane Junction and Stratford (London) on the Up & Down Electrics.
What CBTC system does the Elizabeth Line use? Suppose it is the same as the SSR. In that case, it may cause reliability issues with the transitions, like it does with the Elizabeth Line. It has caused a 20mph TSR on the Up Electric to improve Elizabeth Line trains switching into CBTC mode reliability.
It is likely possible to have ETCS & CBTC on the same track, as NR aims to convert the full mainline network over to the national standard of ETCS, which will require Elizabeth Line to operate on track with both systems simultaneously. NR is extending ETCS L2 on the GWML from Ealing Broadway to London Paddington (High Level).
NR has permitted CBTC to be installed on their tracks, for example, between Pudding Mill Lane Junction and Stratford (London) on the Up & Down Electrics.
What CBTC system does the Elizabeth Line use? Suppose it is the same as the SSR. In that case, it may cause reliability issues with the transitions, like it does with the Elizabeth Line. It has caused a 20mph TSR on the Up Electric to improve Elizabeth Line trains switching into CBTC mode reliability.
It is likely possible to have ETCS & CBTC on the same track, as NR aims to convert the full mainline network over to the national standard of ETCS, which will require Elizabeth Line to operate on track with both systems simultaneously. NR is extending ETCS L2 on the GWML from Ealing Broadway to London Paddington (High Level).
The Elizabeth line uses an Invensys CBTC system, whereas the District uses a Hitachi CBTC system.
The Pudding Mill lane to Stratford section is a relatively simple CBTC overlay on the existing Alstom Smartlock SSI-type interlocking. The interface at Paddington is likewise to an Alstom Smartlock interlocking. Smartlock interlockings are ETCS ready, and indeed the Heathrow branch and part of the line towards Paddington have already been fitted with ETCS, albeit as a relatively simple overlay.
No part of the Elizabeth line has both ETCS and CBTC signalling, or was ever intended to have. You will not have a train running on CBTC following a train running on ETCS, or v.v.
There should have been a direct ETCS to CBTC transition at Paddington, but this was dropped in order to get the Elizabeth line into service, as its testing programme slipped ever backwards. Seeing how long was needed to test and validate the CBTC after the line was ready to run trains, god knows what length closure is going to be needed to validate the CBTC - ETCS transition when it is decided to commission it.
Is there any reason ownership could not be transferred from NR to LU? That would seem to open up the simplest* way forward of having CBTC Interlocking with Signal Overlay retained and then fitting ETCS L1 to the signals for NR Trains. This eliminates transitions between systems at a fairly critical point but allows for full ATP coverage across both types of train and eliminates the tripcocks.
(*) - Yes I am aware 'simple' is doing a lot of heavy lifting there.
Is there any reason ownership could not be transferred from NR to LU? That would seem to open up the simplest* way forward of having CBTC Interlocking with Signal Overlay retained and then fitting ETCS L1 to the signals for NR Trains. This eliminates transitions between systems at a fairly critical point but allows for full ATP coverage across both types of train and eliminates the tripcocks.
(*) - Yes I am aware 'simple' is doing a lot of heavy lifting there.
Wouldn't it be best to go for ETCS L2, which doesn't require any lineside signals? Also, the ETCS blocks are tiny at around 20M, so couldn't you run CBTC off those tiny blocks, which is more or less a Fixed Block acting like a Moving Block?
You solve one issue with changing ownership; however, you gain a new issue, which is you have two junctions in very close proximity, and based on Wimbledon, it probably isn't a great idea. Also, the electrical supply for the Richmond branch comes from Raynes Park.
Having ETCS L2 and CBTC on the same section would require quite complex interlocking to make sure both systems were working properly. Having CBTC interlocking "drive" an overlay that's equipped with ETCS L1 as an ATP system is a simpler system architecture.
Also, the ETCS blocks are tiny at around 20M, so couldn't you run CBTC off those tiny blocks, which is more or less a Fixed Block acting like a Moving Block?
ETCS Blocks are normally as long as the track detection sections (although AIUI it is possible to have 'virtual blocks'). I have heard people who know more about it than me say ETCS can cope with blocks down to about 50m, but they are obviously normally longer than that. 50m blocks would essentially be moving block conceptually, but I'm fairly sure you'd need all trains to be running ETCS L2 to use it like moving block and the whole idea is to avoid LU trains needing to transition out of CBTC or even need ETCS equipment at all.
You solve one issue with changing ownership; however, you gain a new issue, which is you have two junctions in very close proximity, and based on Wimbledon, it probably isn't a great idea. Also, the electrical supply for the Richmond branch comes from Raynes Park.
You have two junctions in close proximity today with the ownership boundary being immediately the east side of Gunnersbury Jn. It would be tight and you'd need a couple of balise groups to manage the transition appropriately but it feels less awkward than trying to get ETCS Level 2 and the CBTC functioning over the same section simultaneously.
The electrical supply might be a bit more awkward, although not an insurmountable issue as the control structures presumably already exist to manage any switch-offs. The largest issue would presumably be contractual elements, at least until the power supply next needs renewing.
Wouldn't it be best to go for ETCS L2, which doesn't require any lineside signals? Also, the ETCS blocks are tiny at around 20M, so couldn't you run CBTC off those tiny blocks, which is more or less a Fixed Block acting like a Moving Block?
You solve one issue with changing ownership; however, you gain a new issue, which is you have two junctions in very close proximity, and based on Wimbledon, it probably isn't a great idea. Also, the electrical supply for the Richmond branch comes from Raynes Park.
Yes, it would be best to go for L2, but that would involve interfacing the Seltrac interlocking with NR's ETCS radio block controller (RBC). The RBC is the kit that converts the information from the interlocking about what routes are set and clear, into ETCS radio movement authorities to be sent to the trains. It also converts information sent back from the trains into interlocking controls, such as to permit the interlocking to cancel a route. Such a bespoke interface would likely be both expensive and time consuming to produce and validate.
Two junctions in close proximity isn't an insurmountable problem for an interface, look at how many junctions are in close proximity to the CBTC boundary at Paddington!
The big problem with Wimbledon is the various crossovers between the Wimbledon branch and the SWR mainline and depot. You really need both ends of a crossover to be controlled by the same interlocking, as moves over one end of a crossover need to set the other end for flank protection. If you try and control them from separate interlockings, then you would end up with a lot of slot or release controls, which would be technically challenging, and an operational headache. Which is undoubtedly why they left the signalling there under NR control when the branch was transferred to TfL. And if you are going to leave Wimbledon, you might as well leave East Putney too, even though that should be relatively easy to deal with.
RailUK was launched on 6th June 2005 - so we've hit 20 years being the UK's most popular railway community! Read more and celebrate this milestone with us in this thread!
