Yes Level 1 is broadly the same operating principle as BR ATP.
There will be some interesting conversations where a driver has a SPAD having not engaged ETCS when it was available, they were trained and able to do so. I know this has happened.
-
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!
First ETCS train operates on the East Coast Mainline
- Thread starter najaB
- Start date
Yes Level 1 is broadly the same operating principle as BR ATP.
There will be some interesting conversations where a driver has a SPAD having not engaged ETCS when it was available, they were trained and able to do so. I know this has happened.
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R
RailUK Forums
LNW-GW Joint
Veteran Member
That's one of the reasons the WCRM project cost £8 billion or so, and it didn't fix the crossings issue north of, I think, Weaver Jn, which were not in scope.
The business case for ETCS is that "normal" signalling renewal is now unaffordable.
So NR wants to put its limited money into ETCS to reduce the renewal and ongoing operational costs.
Hi,
Given that any driver involved in such an incident (with or without the SPAD actually occurring) has probably caused the train operator to contravene the Railway Safety Regulations 1999 then I should imagine the conversation would be very brief and to the point.
Scores of level crossings were removed north of Weaver Junction; i know the people who built the footbridges and went round to various farms with chequebooks (during Foot and Mouth, natch). What werent removed were the (very few) highway crossings, but that is because they didnt need to be. Although some were upgraded from AHB to CCTV.
The highway crossings that were removed - Banbury Lane on the Weedon line, Canley, Tile Hill and Berkswell on the Coventry to Birmingham Line, and Hademore on the Trent Valley, were all removed because of the increase in traffic on those sections making it impractial to keep them - the barriers would rarely be up.
Tipton Owen St was also closed, although this was more to do with the local authority requirements than railway.
bahnause
Member
A few reasons why I'm not keen on it:
ETCS is all sorts of things, but it is not standardised or simple. Every manufacturer and every infrastructure manager has their own ideas about how it should be implemented.
Neither the hardware nor the software is particularly durable, which has a massive impact on costs. The promised cost savings have not materialised, and route capacity has also fallen. Vehicle equipment that is not even 25 years old already has to be replaced. The braking curves in Level 1LS are ridiculous. The network is now a patchwork, where braking curves and release speeds are handled differently depending on the location and the whim of the infrastructure operator. This is of course unmanageable for the train driver, who must therefore always assume the most restrictive case. Work is therefore being carried out on a network-wide adaptation of ETCS L1LS (steeper braking curves and higher release speeds -> higher risks) in order to keep the current timetable operational with the increased introduction of vehicles with Baseline 3.
I now have operating manuals with 250 pages each for Alstom, Siemens (twice, as different versions are in circulation) and Stadler. There is also a 50-page document explaining the specialities on the network and the differences in the interaction with the various versions of the trackside equipment. And of course another document which tries to explain the differences between vehicles with software version 3.5 and 3.6. With little success, I might add.
The whole operation and view of the DMI looks a bit like an afterthought and differs significantly between the manufacturers. This is particularly noticeable in the L1LS, where the information on the display can contradict the signal at first glance. No UX experts seem to have been involved in the design.
And last but not least, SPADs are also possible in ETCS. Both in Level 2 and Level 1, and not only due to human error. Errors in vehicle localisation have already led to dangerous situations which only did not result in an incident thanks to the attentiveness of the train driver. Of course, this risk has now also been shifted to the train driver and signaller by means of a process...
It works perfectly fine if you want to drive 300km/h on a high speed railway with little or noch stations. As soon as bigger and more complicated stations are involved, things can get messy.
Belperpete
Established Member
- Joined
- 17 Aug 2018
- Messages
- 2,406
I initially thought you were talking about UK experience on the Cambrian lines but presumably you are talking about experience in Switzerland?
Of course, most of this is a result of being an early adopter. As with a lot of technology, as the market matures it will inevitably become more standardised and costs will fall.
bahnause
Member
Yes, but history repeats itself. The same problems also arise in other countries, only ten years later. Our ETCS training was route-specific at the beginning. What does this has to do with "standardised"? In the meantime, we are making do with "generic" processes that look and work the same everywhere (even if there are differences depending on the line equipment, which can be ignored). New country-specific requirements are constantly being implemented or functions adapted, as each country wants to introduce its own specific requirements. After all, nobody wants to reinvent "their" railway, but rather emulate the previous system and regulations as well as possible with ETCS. For example, a route indication that is displayed on the DMI.
They are not. Especially not if an oboard unit cannot be used for the entire service life of the vehicle, but has to be replaced. And because the specifications change constantly over the course of a lifetime. The trackside equipment will not be spared updates either.
ETCS is a bad implementation of a good idea if no changes are being made.
Switzerland is the biggest user of L1LS, and rolled it out very early too. I expect some of the difficulties are from being an early adopter, like our Cambrian L2 which needed a significant upgrade. ETCS is still a young system in signalling terms, but it is maturing now and can only become more mainstream.
How did the experience change when systems switched from Euro SIGNUM/ZUB to L1LS. Are the braking curves imposed very different?
That's very unfortunate. Needs some more standardisation across the fleets. More software changes!
SPADs are unlikely alone to lead to a collision as long as the approach speed is already successfully limited, the event triggers automatic emergency braking and there's distance clear ahead of the movement limit to come to a stand before any plausible obstruction. L1 & L2 retain the concept of release speed developed in historic ATP systems, recognising a run past is possible, then having a last-ditch intervention available to arrest a train before danger. That's why ETCS needs overlaps.
Last edited:
Belperpete
Established Member
- Joined
- 17 Aug 2018
- Messages
- 2,406
It is a common misapprehension that ETCS prevents SPADs. In order to allow the driver to drive up to a signal without intervening, the system can inherently only intervene when it looks like the train is going to SPAD. By that time it may be too late to prevent the SPAD. But the emergency braking curve should still stop the train before it reaches a point of conflict, which is the important thing. At the end of the day, it is the driver who is driving the train.
In addition, there's always 'train stop' functionality at the signal/blockmarker with a safe overrun distance clear ahead at release speed protected by the interlocking.
12LDA28C
Established Member
- Joined
- 14 Oct 2022
- Messages
- 5,058
This principle is no different to any other safety system, be it ATP, TPWS or whatever.
Belperpete
Established Member
- Joined
- 17 Aug 2018
- Messages
- 2,406
Agreed, but for some reason it does seem to be a fairly common misapprehension around ETCS.
I'm not in the industry, but I'd be very surprised if the cost the onboard equipment isn't significantly lower now than it was in the early days as that is the way that things go with electronics - even specialist as opposed to consumer goods. Take something quite specialist like IFE equipment for airliners, where modern installations are a lot cheaper (though by no means cheap due to having to meet regulatory requirements) than early ones with significantly more functionality.
Applies to the electronic kit in trackside cabinets too. The track balises are actually very resilient, far more so than UK TPWS grids for example, while being extremely light compared to traditional AWS equipment, although newer lightweight components are available for that now using strong rare earth magnets.
While the ETCS standard concerns itself primarily with the interfaces between track and train, there's little standardisation for interconnection between trackside, interlocking and control centre, so while you may not be locked into a particular train or lineside supplier in general terms, its often not easily possible to mix and match equipment from different suppliers within the trackside and control centre ecosystem as those interfaces are often proprietary. A national infrastructure owner in the EU can still have local requirements to address these concerns for projects in their territory, in addition to the standard ETCS specs for track to train. ETCS provides an open framework within which engineers have been able to create disparate custom systems such as TASS and Thameslink ATO in the UK. The difference from the past is that the functionality in terms of message exchanges and protocols is fully declared rather than proprietary.
Last edited:
12LDA28C
Established Member
- Joined
- 14 Oct 2022
- Messages
- 5,058
Maybe not the cost of the onboard equipment itself but in the case of installing ETCS into existing vehicles the cost is eye-watering, particularly when it involves some structural rebuilding of the cab, movement of dials and gauges etc to accommodate the screens and so on. That's when the cost really mounts up, all paid for by NR.
Retrofit may not be a good idea for many classes, some of which might better be allowed to retire to avoid the need before their routes are equipped with ETCS. It depends partly on the size and age of a particular fleet whether its worth considering within remaining likely working life. Switzerland designed bolt-in replacements for older stock in their Euro Signum/Zub project that led to L1LS. The equipment replaced the previous legacy kit taking up no more space in a design approach echoing original TPWS rollout in the UK using equipment designed to slot into spaces occupied by AWS previously and provide both TPWS and AWS functionality in one unit coupled to the same simple traction and brake interfaces used by the original AWS.
New fleets should come at least 'ETCS ready' which means additional screens already in place or the space identified and reserved.
bahnause
Member
The annoying little niggles at the beginning of the L1LS introduction did happen and affected reliability. But on the whole, the system works as designed. And therein lies the problem, L1LS has many more limitations compared to conventional signalling to increase safety in an already very safe system. Even worse: this was known years in advance, but the warning voices from the operational sector were ignored. Now attempts are being made to increase the network capacity again by tinkering with the braking curves and release speeds. In some places, this is also reducing security below the level that was achieved with Signaum and ZUB. However, as long as these adjustments are not implemented network-wide, it won't help much. As a train driver, I cannot know in which track which release speed will be active from which location under which conditions. In practice, the current timetable would not be possible if all vehicles were equipped with L1LS.
ZUB was at least predictable. Now, in addition to the signals, you still have to keep an eye on the screen to be able to observe all restrictions.
Only works if the hardware ist compatible.
ETCS is a SIL4 train protection system. SIL 4 is never cheap. The onbard system is not just some kind of display with some kind of computer. Honeywell doesn't sell you their Doppler-Radar used for the odometry for cheap. Even on a "ETCS-ready" vehicle, fitting ETCS will cost you around 300'000 Euros. In some cases, it is no longer even financially justifiable to retrofit active vehicles (S-Bahn Stuttgart). As every vehicle fleet has its own specialities, ETCS hardware is still not a mass product.
I get L1 needing release speeds but why would L2 need them?
As far as I’m aware, release speeds were required in BR–ATP because as one approaches a signal at danger, the system would not know whether that signal aspect had been changed until the signal itself, and its associated balise, had been passed - and the release speed allowed that to happen.
The same would apply to L1 which has similar operating principles.
But for L2, ETCS Movement Authorities are conveyed via GSM-R. So is it not the case, that as the train draws up to an “end of authority”, that the system would instantly know if the authority gets extended, thus meaning release speeds are not required?
As I understand it a release speed with a suitable overlap locked safely beyond can allow trains to approach the end of authority (EOA - usually a signal or block marker board) more swiftly. No release speed imposes a much more controlled slower approach. In ETCS, routes with different release speeds can be implemented over the same path, selectable by the signaller in a similar way that main or warning routes can be selected in UK TCB where provided.
In traditional ATP, systems often have 'infill' on approach to signal specifically to convey 'better aspect' updates to the onboard equipment sooner, before you actually pass the recently cleared signal. Infill can take the form of additional transponders or special leaky feeder loops. Either can be provided in L1. Without these you are correct release speed allows a driver to stop decelerating to a stop once under the release speed in anticipation of impending clearance. If he anticipates too much then the trainstop function and locked overlap will prevent a catastrophe if not a SPAD. With ETCS, if you have to supervise down to stop at the EOA, then you have to control speed further back for safety which means a slower, less confident approach. Release speeds allow faster run-ins to red signals or equivalent, with safe overruns managed by the interlocking.
Last edited:
Which is why I made the comparison to airline IFE equipment. It still costs a few million dollars to equip a wide-body aircraft, but that's down significantly from where it used to. In the early days you might spend that much on just equipping the first-class cabin.
12LDA28C
Established Member
- Joined
- 14 Oct 2022
- Messages
- 5,058
On the Chiltern lines, The on-board equipment would not know if a signal had changed until the train reached the ATP loop for that signal and unfortunately the system was installed 'on the cheap' with some signals' ATP loops being very short indeed, meaning drivers could be almost brought to a stand at a signal that had long since cleared. Of course if continuous loops had been installed from the start this problem would have not existed as when the signal cleared the on-board equipment would have been updated instantly.
So the Chiltern system did not have release speeds?
12LDA28C
Established Member
- Joined
- 14 Oct 2022
- Messages
- 5,058
Yes they did which just meant that you could potentially have a SPaD as long as you were travelling below the release speed. The onboard equipment however would not update the train with the signal aspect until it reached the loop in the four foot. Not much use if the ATP loop was only 50 yards long but you had sighting of the signal for half a mile.
Chiltern SelCab, related to SelTrac and LZB, and now all supported by Hitachi who recently acquired the Thales rail control business, uses a twin-wire data antenna along the middle of the four-foot, just like those related systems. Transposed periodically for distance measurement, the 'wiggly wire' is the sole data communications medium but is not continuous like SelTrac and LZB. The wires are only provided on approach to signals and can be between 25 and 1200m long. At most plain line signals, designers economised heavily, providing only very short loops. In most junction signalling cases the loop would have covered back to when any approach release from red routes should clear. The GWML ACEC scheme mostly uses transponders at signals with additional ones on approach for infill. That system had the option of a leaky feeder antenna loop over a distance approaching a signal for finer better aspect updates I think they used that at some critical junctions.
12LDA28C
Established Member
- Joined
- 14 Oct 2022
- Messages
- 5,058
Indeed, and some of the loops for signals on the Chilterns had to be extended after initial installation because it was becoming apparent that trains were suffering delay due to having to be brought almost to a stand before the on-train receiver would reach the loop and be updated that the signal had cleared. This was also the cause of some frustration for drivers as of course they could see well in advance that the signal had cleared but could not apply power until the train reached the loop. This system is no longer in use though, unlike the balise-based ATP system on the Western Region.
As a deliberate act? Or by selecting the wrong level in error?
Remind me: what is the cost of a trackside signalling equivalent unit (SEU)? Access to a train in a depot is far easier, cheaper and safer than to the track.
The answer is "I have no idea", but that has nothing to do with the point that I was making: the capital cost of replacing equipment before it's life expired can still represent a net savings if the replacement equipment is cheaper to acquire and operate and/or provides more functionality than the existing equipment.
Indeed, trains come home to the depot for maintenance rather than remaining distributed around the trackside. It's rational to concentrate complexity on the vehicles rather than the trackside. There'll always be some electromechanical and control equipment on the trackside to maintain even in a future virtual block era however, notably point machines, a few fixed train detection sections around junctions for deadlocking, and the radio network. The fixed elements of block markers and balises will require a level of maintenance support even if it's only replacing damaged ones, and inspecting and cleaning them periodically.