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!
Clearing out my garage yesterday, I came across a box containing some interesting documents:
1. Course material from the training many of us received at Reading signalling drawing office for the ACEC GW-ATP system, when it was being rolled out.
2. Material from a similar course I attended in the mid-90s detailing the SJ (Swedish Railways) signalling system and their Ebicab ATP system (also interfaced via signal feeds in trackside equipment cupboards).
Some exciting bedtime reading there! I'll post some impressions and comparisons when I've digested them, maybe sharing some useful illustrations from the courses. I recall the Swedish system uses deferred speed limits for junctions beyond the transponder at the signal, e.g. 60km/h in 400m. If the reduction from line speed needed to commence before the protecting signal, then messages at approach infill loops or previous signals would enforce that.
Well ETCS transparent balise controllers that interpret existing circuits using current transformers are approved products in Europe. Siemens in particular makes a whole range of them under the 'Trainguard Trackside' range. They even come with things like integrated flashing detection on the lamp circuits.
Thousands have been deployed in Switzerland and elsewhere.
Just because they are approved for use with various continental administrations' signal heads and circuits does not mean that they are suitable, let alone approved, for use with UK-standard versions.
Conceptually no, since this is already integrated into the ETCS specification.
Instead of the training having a "phone" call to the radio block controller, it dials the number of the next radio interface unit provided it to be a balise on the track, possibly the same balise group that provided the previous signal aspect.
That's the magic of GSM-R after all, any end point can connect to any endpoint in the numbering plan.
Of course knowing the mess that is the ETCS rollout, some trains may require software modifications because they didn't bother to implement that part of the specification for whatever reason.
It is highly unlikely that they will have implemented interfaces and software that is not needed. Let alone interfaces and software for a non-standard L1 concept that hasn't even been developed yet. How will they have tested that the trains are compatible with something that doesn't exist?
Why wouldn't it?
If the signal is set for a restrictive aspect or for a slow diverging route, the ETCS equipment knows it and can impose the relevant speed restriction, and impose it continuously until the actual divergence point.
You are assuming that there is only one speed restriction ahead. In practice, in a complex station area, there may be a whole sequence of restrictions ahead. My experience with programming Eurobalises is that they can only provide a very limited amount of speed restriction information. This was requiring compromises on a fairly moderate-speed metro line, with trains being unnecessarily restricted. On a high speed line where there could well be many speed restrictions within the potential braking curve of a train, I suspect it may be much more of an issue.
ETCS Level 1 can be implemented, using semi continuous infill like radio infill units or euroloops, in a manner allowing the elimination of trackside signals, should that be desired.
If the balises obtain their information from current transformers connected in series with the signal head, then if you remove the signal they no longer have any current to measure. Yes, you could provide current loads in the location case that look like a signal. But what you have ended up with is a location case full of unnecessary equipment, all of which has to be maintained for the ETCS to carry on working. The whole idea of getting rid of signals is not just to get rid of the signals themselves, but also all the equipment and cabling that goes with them. You would be increasing the amount of trackside equipment needing to be maintained, not reducing it, making the railway inherently less reliable.
== Doublepost prevention - post automatically merged: ==
No, not all signal heads have 2BA terminals and even some that do, do not have space for current transformers or additional cables.
Many existing Western Region large size cupboards had plenty of room.
There are now additional problems, as the railway has moved away from using 2BA terminals because they have exposed live parts that can be touched. Now WAGO or other DIN rail modern terminals are used (for staff safety reasons).
And while traditional location cases used wooden backboards, on which additional equipment could be screwed into any suitable spare space, modern location cases use stainless steel bars, making it much harder to fit additional equipment.
Having said all that, the easiest way to add ETCS level 1 is still by using current transformers fitted in the existing location cupboards. It's just that some new wiring would be needed. By only adding a new (longer) wire that replaces an existing wire (the new longer wire goes via a current transformer), it makes the testing of the existing signal far quicker and easier than any other alteration.
And this work can be done by Network Rail's own signalling maintenance staff (a small amount of training would be required). Don't need expensive and hard to find installers and principle testers.
Any such alteration to an existing location case would need that case correlated before the alteration was designed. Having the current transformer that is supposed to be monitoring the red aspect in fact monitoring the green, for example, could have serious consequences.
And how long would it take to do if you were to do it with NR's own signalling maintenance staff? Do they really have nothing else to do? If you drag them all off for a decade or more to install this at every signal, what is going to happen to maintenance in the interim?
And having spent billions of pounds and a decade or more rolling this all out, you will have ended up with a railway that is not significantly safer than now, but has even more trackside equipment to maintain and go wrong than now, and still be having accidents that L2 would have avoided.
Also to add to the "can be done by Network Rail's own signalling maintenance staff" bit, in addition to the with what time (don't forget the teams are smaller now) not if it's a functional change it can't. Maintenance are allowed to do a like for like change, plus a bit (things like changing Filament Lamps Signals to LED signalheads) and testing. This is the qualification the teams will have under the Signal Maintenance Testing Handbook (SMTH). If a change is required it needs Signal Works Testing to the Signal Works Testing Handbook (SWTH), witch they won't have the competence for. There is an intermediate level testing standard but I've not had dealings with it and would have to look up what can be done with it. I suspect changing train protection radically will go straight to SWTH (I'm fairly sure but don't like to state for certain without checking)
Just because they are approved for use with various continental administrations' signal heads and circuits does not mean that they are suitable, let alone approved, for use with UK-standard versions.
But it demonstrates that the creation of such equipment is certainly compatible with a high standard of railway safety.
Is there any great reason to believe that UK railway signalling, based on fundamentally the same ideas of lighting up lamps, is going to be impossible to fit with such equipment?
Given the railway is a safety case industry, doesn't everything related to this sort of thing require safety approval?
However, there is established precedent, in the UK, for controlling safety grade train protection equipment using current transformers in the lamp loops - as noted with Great Western ATP.
There is established precedent, in continental europe, for using switchable balises to transmit safety-grade control messages in trains based on lineside signal indications
Use without lineside signals, through infill, is apparently an envisaged use case for ETCS Level 1, given that it is mentioned on the websites of a major sponsor organisation
This specific application using the selected specific pieces of equipment will require approval as a matter of course, but nothing about it is particularly groundbreaking.
Sure, the specific equipment will have to be designed, tested, qualified and approved. But the cost of such development is probably small than the costs inherent in any sort of large rollout of any train protection or signalling technology. Let alone a large ETCS Level 2 deployment programme.
It is highly unlikely that they will have implemented interfaces and software that is not needed. Let alone interfaces and software for a non-standard L1 concept that hasn't even been developed yet. How will they have tested that the trains are compatible with something that doesn't exist?
If they don't implement the entire interface then there is no "standard" ETCS installation or interface which has already been exhaustively tested and new ones are going to require safety approval anyway as a matter of course, surely?
You are assuming that there is only one speed restriction ahead. In practice, in a complex station area, there may be a whole sequence of restrictions ahead. My experience with programming Eurobalises is that they can only provide a very limited amount of speed restriction information. This was requiring compromises on a fairly moderate-speed metro line, with trains being unnecessarily restricted. On a high speed line where there could well be many speed restrictions within the potential braking curve of a train, I suspect it may be much more of an issue.
Well the same balise vendors that provide standard transparent balise equipment also allow the switched balises to "daisy chain", using the same control feeds. This allows multiple balises to be used in coordination to send longer messages.
Each balise can transmit ~820 bits of usable payload data. Obviously the precise figures would have to be determined by a proper development programme - the acronyms and shortened values in the ETCS documentation are giving me a headache! But an inspection suggests that each speed restriction only requires around 32 bits of payload, assuming a speed profile packet is being transmitted regardless.
It seems likely that a great many speed limit steps could be included before the packet length becomes a serious issue. It only has to provide information that the train needs to safely reach the next signal without braking unneessarily after all.
If the balises obtain their information from current transformers connected in series with the signal head, then if you remove the signal they no longer have any current to measure. Yes, you could provide current loads in the location case that look like a signal. But what you have ended up with is a location case full of unnecessary equipment, all of which has to be maintained for the ETCS to carry on working. The whole idea of getting rid of signals is not just to get rid of the signals themselves, but also all the equipment and cabling that goes with them. You would be increasing the amount of trackside equipment needing to be maintained, not reducing it, making the railway inherently less reliable.
In the very short term, yes. Just as their would be in an ETCS Level 2 overlay situation where the entire original trackside signalling apparatus must remain in place.
However, once the cab signalling system is in place, equipment can be removed and bypassed as and when convenient, depending on the complexity of the arrangements necessary to remove it.
For example, if bypassing the proving on any particular piece of equipment is simple, they can be removed rapidly. Equipment with proving circuits that are harder to subvert would remain until more extensive modifications are required or undertaken on the system.
The railway still benefits from a reduction in training burden on drivers from reduced diversity of signalling systems (from pursuing a faster rollout) and the installation of modern train protection.
As it stands, I seem very unlikely to be alive long enough to see ETCS Level 2 become the dominant signalling system on the UK railway.
Indeed, I don't expect to be alive to see the last signal box go into an ROC (is it still Morpeth circa 2070?) - i'm too old and too fat!
Sure, the equipment would have to be designed, tested, qualified and approved. But the cost of doing this is probably still going to be less than the billions of pounds necessary for a large scale ETCS Level 2 retrofit programme.
Don't forget a lot of the effort is the onboard and the integration, Signalling upgrades are not infrastructure only with ETCS, the trains (and a few more things) are involved as well. No point in having level 1 fitted to the signals and no fitted trains
You are assuming that there is only one speed restriction ahead. In practice, in a complex station area, there may be a whole sequence of restrictions ahead. My experience with programming Eurobalises is that they can only provide a very limited amount of speed restriction information. This was requiring compromises on a fairly moderate-speed metro line, with trains being unnecessarily restricted. On a high speed line where there could well be many speed restrictions within the potential braking curve of a train, I suspect it may be much more of an issue.
Is it possible to have more than one Eurobalise, which each providing additional speed restriction data? The GWML ATP can have an additional beacon if a single signal beacon does not have sufficient capacity within the data transmission for all the required speed restriction data.
And while traditional location cases used wooden backboards, on which additional equipment could be screwed into any suitable spare space, modern location cases use stainless steel bars, making it much harder to fit additional equipment.
Yes, but that's always a possibility, just like with any signalling alteration. When either AWS or TPWS (I forget which) was added to a line where it was not originally fitted, space had to be found for an additional TFMs (in BRB style cupboards) as there were insufficient outputs from the existing TFMs (TFM - Trackside Functional Module - an SSI module, BRB - British Railways Board).
Any such alteration to an existing location case would need that case correlated before the alteration was designed. Having the current transformer that is supposed to be monitoring the red aspect in fact monitoring the green, for example, could have serious consequences.
And how long would it take to do if you were to do it with NR's own signalling maintenance staff? Do they really have nothing else to do? If you drag them all off for a decade or more to install this at every signal, what is going to happen to maintenance in the interim?
All valid points. However, I never suggested that the existing staff would no longer carry out their normal duties.
Network Rail has it's own "Works Delivery" staff (one of a number of names for staff that normally do minor renewal work) that are often assisted by maintenance staff working overtime (something that has happened in my former depot and I know it happened elsewhere as well). And no, these alone would not be enough, but combined with actual installers/contract company staff, that would make a difference in how quickly the work could be done across an area.
The point that I was trying to make, is that this work would be less onerous compared to fitting something like TPWS (as required by current practices and standards) to all signals which requires significant alterations to the existing circuitry to cut the proving contacts into the existing control signals.
And of course there would have to be suitable testing to ensure that the current transformers and their outputs were wired and connected correctly. That would likely be done at a later stage by different staff when the ETCS was tested and certainly before it was commissioned / brought into use.
And having spent billions of pounds and a decade or more rolling this all out, you will have ended up with a railway that is not significantly safer than now, but has even more trackside equipment to maintain and go wrong than now, and still be having accidents that L2 would have avoided.
And AWS and TPWS didn’t add additional equipment? The reliability of some TPWS equipment is poor adding to the maintenance workload. In one installation, the TPWS modules were changed three times in as many months, the third pair of modules having been fitted after changing the backplate assembly. As I pointed out to my manager and tech support, having different metals for the module contacts and the spades in the plugboards on the backplate assembly, when fitted in locs where the air may become damp was asking for problems for low current/low voltage signals.
I would love for L2 to be rolled out faster and across all the main lines, but realistically that's not going to happen on the vast majority of lines anytime soon. The expense and resources for replacing all the existing interlockings that are not suitable or compatible with L2 makes it currently unpractical.
If all stock/units/trains on an area had ETCS and all the lines in that area were fitted, the existing AWS and TPWS would no longer be required, reducing the maintenance workload.
As I have said before, if the issue was easy to solve, it would have already been done by now. But it's not easy to solve.
Also to add to the "can be done by Network Rail's own signalling maintenance staff" bit, in addition to the with what time (don't forget the teams are smaller now) not if it's a functional change it can't. Maintenance are allowed to do a like for like change, plus a bit (things like changing Filament Lamps Signals to LED signalheads) and testing. This is the qualification the teams will have under the Signal Maintenance Testing Handbook (SMTH). If a change is required it needs Signal Works Testing to the Signal Works Testing Handbook (SWTH), witch they won't have the competence for. There is an intermediate level testing standard but I've not had dealings with it and would have to look up what can be done with it. I suspect changing train protection radically will go straight to SWTH (I'm fairly sure but don't like to state for certain without checking)
Is there any reason we could not make a replacement signal head with the lineside equipment unit, rail infill unit and similar bits integrated into it? A lot of LED signal heads seem much smaller than the ones they replace.
It would function as a regular signal head until the high level signalling people appear to install the balise and programme it?
EDIT:
An incandescent signal head for a relay or earlier interlocking might pull 300mA @ 110Vac. Since a signal head should always have at least one aspect lit, conceptually there should be more than enough power to power the electronics in the head, whilst still allowing enough for an LED lamp. You would need a fairly fancy power supply to do this though, it would have a separate switched mode power supply for each aspect and also have variable resistors to ensure enough current is always drawn to keep a lamp proving relay from dropping out.
Is there any reason we could not make a replacement signal head with the lineside equipment unit, rail infill unit and similar bits integrated into it? A lot of LED signal heads seem much smaller than the ones they replace.
It would function as a regular signal head until the high level signalling people appear to install the balise and programme it?
There is no reason you couldn't make any unusual combination you want but you'd still need the SWTH staff to test it. And sticking all that up a post wouldn't be helpful, if stuff you need to maintain then needs working and such. Also the head still needs to talk to the existing interlocking and location cases. I worked on a job that had to add disconnection boxes over about 10 miles of railway to signals and such for a re-lock and the amount of different configurations was high.
*edit to add*
So you need a head that can talk to all the different cable and voltage combinations as someone said upthread.
Also to add to the "can be done by Network Rail's own signalling maintenance staff" bit, in addition to the with what time (don't forget the teams are smaller now) not if it's a functional change it can't. Maintenance are allowed to do a like for like change, plus a bit (things like changing Filament Lamps Signals to LED signalheads) and testing. This is the qualification the teams will have under the Signal Maintenance Testing Handbook (SMTH). If a change is required it needs Signal Works Testing to the Signal Works Testing Handbook (SWTH), witch they won't have the competence for. There is an intermediate level testing standard but I've not had dealings with it and would have to look up what can be done with it. I suspect changing train protection radically will go straight to SWTH (I'm fairly sure but don't like to state for certain without checking)
The work that I suggested that the existing Network Rail signalling maintenance (and Works Delivery) staff would do would be to SMTH, replace a wire with a wire, going from the exact same terminal to the exact same terminal. Each wire done one at a time. The difference only being that the new wire would be longer and go through a current transformer.
This is no different to how some monitoring/data logger systems were added to existing relay interlocking systems.
The rest of the wiring for the ETCS system would be done separately and likely by installation staff (likely contract company staff). As it would not be disruptive to the operation of the existing signal and signalling, it could be carried out without having to have the line closed (assuming there is a suitable access point that does not require the staff to cross any lines). And it would be tested to SWTH before the ETCS was brought into use.
Now, if the design calls for a different way of connecting, maybe by using a current transformer that requires the primary circuit to have the existing signal lamp feed connected to terminals, then what I have described would not apply. However, I'm sure a suitable method of carrying out the alteration could be devised whereby a full SWTH tester would not be needed (for the work on the existing signalling) and a suitably qualified SMTH tester could receive additional training so that said work could be done. In the same way that a suitably qualified SMTH tester could receive additional training to perform other alterations as per G110 / signalling non-conceptual works.
Is it possible to have more than one Eurobalise, which each providing additional speed restriction data? The GWML ATP can have an additional beacon if a single signal beacon does not have sufficient capacity within the data transmission for all the required speed restriction data.
Yes, you can add extra balises. So you can add an extra balise after the end of a PSR with information about further PSRs ahead. But the problem, as I recall, is that each balise must contain all the relevant PSR information up to the end of the braking curve. If there are too many PSRs ahead, you have to start simplifying the information, which inevitably means more restrictive information.
Yes, but that's always a possibility, just like with any signalling alteration. When either AWS or TPWS (I forget which) was added to a line where it was not originally fitted, space had to be found for an additional TFMs (in BRB style cupboards) as there were insufficient outputs from the existing TFMs (TFM - Trackside Functional Module - an SSI module, BRB - British Railways Board).
Fitting extra TFM bars into an existing case would be almost impossible. If you are lucky, there would be space on the existing bars for an extra TFM. If there weren't, then you would almost certainly need to provide an additional case, and all that involves.
The point that I was trying to make, is that this work would be less onerous compared to fitting something like TPWS (as required by current practices and standards) to all signals which requires significant alterations to the existing circuitry to cut the proving contacts into the existing control signals.
They certainly did. However, in those cases, the costs and risks that they added are significantly outweighed by the risks that they have mitigated, with a significant reduction in SPADs and accidents. However, their provision has now made the railway so safe that ETCS is hard to justify, as it will only give a very marginal improvement in safety. There are very few accidents nowadays that ETCS would prevent. Which is why ETCS is being promoted as much for its cost-saving and operational benefits, as for safety.
A method of implementation that both adds extra costs and adds extra risks of its own would therefore be very hard to justify.
I would love for L2 to be rolled out faster and across all the main lines, but realistically that's not going to happen on the vast majority of lines anytime soon. The expense and resources for replacing all the existing interlockings that are not suitable or compatible with L2 makes it currently unpractical.
If all stock/units/trains on an area had ETCS and all the lines in that area were fitted, the existing AWS and TPWS would no longer be required, reducing the maintenance workload.
As I have said before, if the issue was easy to solve, it would have already been done by now. But it's not easy to solve.
Again, I don't disagree with this sentiment. NR's master plan for ETCS has already slipped badly. I may still have a copy of their roll-out dates somewhere, which should be good for a laugh.
To build up the workforce, both in-house and contractors, there needs to be a firm, long-term commitment to a rolling and steadily-increasing programme of work, such as happened in the 1950s. I just don't see any such commitment at present. It is still one small isolated contract here, another isolated contract there.
NR recognised that they would need to make massive up-front investments to get the ball rolling, replacing interlockings early and retro-fitting stock early, both to build up the expertise, but also so that they don't end up with an unmanageable bow-wave of work later on. Again, that just isn't happening.
It all still seems to being justified on a short-term, project-by-project basis. This kind of short-term thinking is likely to cost big in the long-term. In the same kind of way that happened with the roll-out of electrification.
== Doublepost prevention - post automatically merged: ==
Is there any reason we could not make a replacement signal head with the lineside equipment unit, rail infill unit and similar bits integrated into it? A lot of LED signal heads seem much smaller than the ones they replace.
It would function as a regular signal head until the high level signalling people appear to install the balise and programme it?
EDIT:
An incandescent signal head for a relay or earlier interlocking might pull 300mA @ 110Vac. Since a signal head should always have at least one aspect lit, conceptually there should be more than enough power to power the electronics in the head, whilst still allowing enough for an LED lamp. You would need a fairly fancy power supply to do this though, it would have a separate switched mode power supply for each aspect and also have variable resistors to ensure enough current is always drawn to keep a lamp proving relay from dropping out.
And how long do you think this would take to develop, let alone get safety approval? Not only would you need versions for all the different types of main signal head, but how would you deal with signals that also have subsidiaries, route and junction indicators, etc that all affect the information that needs to be transmitted to the train?
Feeding the unit from the existing lamp supplies might be possible, but if the unit is going to be feeding several balises, infill unit, GSM-R radio transmitter, etc etc how much power is it going to need? The problem I see would not be preventing the lamp-proving relay from dropping out, but ensuring that the lamp-proving relay does still drop out when the lamp fails, but all the extra gubbins are still drawing power.
When the existing cable is disconnected from the old head, it would need to be megger-tested before connecting it to the new head. You would need to alllow contingency in each case for the cable failing the megger-test, and having to be replaced. That is not going to be a quick job for signals on a gantry, for example. Even if the cable passes, you would still have had to make the allowance in case it had failed.
A nice idea, but impractical. I agree with MarkyT and others, that if you were going for a L1 retrofit, the best way to do it would be with current transformers in the existing loc case. But I don't think we should be going for a L1 retro-fit in the first place.
The work that I suggested that the existing Network Rail signalling maintenance (and Works Delivery) staff would do would be to SMTH, replace a wire with a wire, going from the exact same terminal to the exact same terminal. Each wire done one at a time. The difference only being that the new wire would be longer and go through a current transformer.
This is no different to how some monitoring/data logger systems were added to existing relay interlocking systems.
The rest of the wiring for the ETCS system would be done separately and likely by installation staff (likely contract company staff). As it would not be disruptive to the operation of the existing signal and signalling, it could be carried out without having to have the line closed (assuming there is a suitable access point that does not require the staff to cross any lines). And it would be tested to SWTH before the ETCS was brought into use.
Now, if the design calls for a different way of connecting, maybe by using a current transformer that requires the primary circuit to have the existing signal lamp feed connected to terminals, then what I have described would not apply. However, I'm sure a suitable method of carrying out the alteration could be devised whereby a full SWTH tester would not be needed (for the work on the existing signalling) and a suitably qualified SMTH tester could receive additional training so that said work could be done. In the same way that a suitably qualified SMTH tester could receive additional training to perform other alterations as per G110 / signalling non-conceptual works.
I see what you are saying, and you could do that, but to what benefit? The signalling contractors to put in the ETCS will need to go in there several times (correlate drawings if it has not been very recently worked on, install & test). Using sparce maintenance to do what would be part of a circuit seems a bit of a waste of resource to me
== Doublepost prevention - post automatically merged: ==
Again, I don't disagree with this sentiment. NR's master plan for ETCS has already slipped badly. I may still have a copy of their roll-out dates somewhere, which should be good for a laugh.
To build up the workforce, both in-house and contractors, there needs to be a firm, long-term commitment to a rolling and steadily-increasing programme of work, such as happened in the 1950s. I just don't see any such commitment at present. It is still one small isolated contract here, another isolated contract there.
NR recognised that they would need to make massive up-front investments to get the ball rolling, replacing interlockings early and retro-fitting stock early, both to build up the expertise, but also so that they don't end up with an unmanageable bow-wave of work later on. Again, that just isn't happening.
It all still seems to being justified on a short-term, project-by-project basis. This kind of short-term thinking is likely to cost big in the long-term. In the same kind of way that happened with the roll-out of electrification.
Maybe we will see a more coordinated way of working under GBR (once it gets going, I'm not suggesting day 1) with Infrastructure & TOC being under one roof. Though Open Access and FOCs are still a sticking point potentially
But it demonstrates that the creation of such equipment is certainly compatible with a high standard of railway safety.
Is there any great reason to believe that UK railway signalling, based on fundamentally the same ideas of lighting up lamps, is going to be impossible to fit with such equipment?
I am not saying that it is impossible, but if you are developing a new concept from scratch, then it will take a significant time to both develop and more importantly get safety approval. Whereas L2 is already safety-approved and ready to roll.
If they don't implement the entire interface then there is no "standard" ETCS installation or interface which has already been exhaustively tested and new ones are going to require safety approval anyway as a matter of course, surely?
However, once the cab signalling system is in place, equipment can be removed and bypassed as and when convenient, depending on the complexity of the arrangements necessary to remove it.
If your whole ETCS L1 installation is based on trackside equipment, then how are you going to upgrade it to L2? L2 would require either a relock or resignalling, in which case all the trackside signalling equipment and all the L1 equipment that you have just installed would become totally redundant. There is no point in bypassing it.
As it stands, I seem very unlikely to be alive long enough to see ETCS Level 2 become the dominant signalling system on the UK railway.
Indeed, I don't expect to be alive to see the last signal box go into an ROC (is it still Morpeth circa 2070?) - i'm too old and too fat!
I am not saying that it is impossible, but if you are developing a new concept from scratch, then it will take a significant time to both develop and more importantly get safety approval. Whereas L2 is already safety-approved and ready to roll.
It's safety approved in the equipment sense, but converting a relay interlocking into an SSI one that can be fitted for ETCS Level 2 is going to require a lot of specific work that will need highly complex safety approvals, surely.
As noted by other posters above, much of the documentation of relay interlockings built decades ago could be missing or incomplete and confirming that the information is accurate could prove to be a laborious task, especially if the notorious brittle wire problem surfaces. Perhaps my perception is distorted by the pictures of relay interlockings in things like the Hidden Report, but my own experience of troubleshooting wiring arrangements much less complex than those (and with lower stakes for undetected faults or mistakes!) is one of pain and tedium.
Developing new equipment to avoid going anywhere near as much of the installation at possible seems likely, at least based on my own (nuclear) industry's experience, to be a better choice than wading in and replacing the guts of the system.
If your whole ETCS L1 installation is based on trackside equipment, then how are you going to upgrade it to L2? L2 would require either a relock or resignalling, in which case all the trackside signalling equipment and all the L1 equipment that you have just installed would become totally redundant. There is no point in bypassing it.
I was referring more to equipment such as banner repeaters and (most, if not all) distant signals that would be unnecessary in the proposed system, since the RIU would permit the aspect in front to be read over the radio from the beginning of the block section. Some of that equipment should prove simple(ish) to remove, but I expect some will be wired into the proving arrangements in such a way as to make it hard to extract (like the earlier mentioned electromechanical banners), that equipment would be left in place until other major work was needed.
I would not expect a near term conversion to L2 until wholesale replacement of the interlocking is called for.
Which could be decades further into the future.
Because there are multiple problems including a lack of contract company staff who can work on existing live signalling. Many installers don't hold SMTH and the number of SWTH testers is not great either. Many of these staff (installers and testers) will be already tied up with existing signalling contracts.
Plus, because of the restrictions on what was called "red zone" working and the limited engineering access when work can be done on live in service signalling equipment.
Would it be efficient, I don't know, Most likely no, but that's difficult to say until the work plan is put together. It most certainly will be less efficient than in the past when BR had a larger (compared to now) department of installers and works testers and there were less difficulties getting access to the equipment.
This is the problem of ETCS fitment having been put back and back and back. It seems a very long time ago that I and the others in my depot were briefed that the GWML would be fitted with ETCS within the next ten years or so (they were describing L2 for the future but I can't remember if the intention at that time was to start with L1). That was some time before the electrification programme got going...
The technology exists to prevent (or at the very least substantially reduce the consequences) of a collision of trains on the same line. And the technology has existed for many, many years in one form or another. Yet, a collision that likely could have been prevented has occurred (although we need to wait for the RAIB to report before we know all the facts).
I can't tell what will happen in the future, and no one else knows either, but the risk of another collision (or derailment due to excessive speed at/over junctions, points or curves) that could result in a death or deaths and serious injuries remains substantially the same since the national TPWS programme concluded. With no substantial change to signals / locations that have not been fitted with TPWS at all. The only significant improvements have been in driver training, signal sighting and rolling stock improvements (better designs that can better cope with an impact) for new trains.
Look at how bad the U.K. railways looked to the general public after the Ladbroke Grove rail crash which occurred not long after the Southall rail crash.
So IMHO doing nothing is unwise. The L2 programme is too slow. And I can't see it covering the whole network within the next twenty years even if the rollout was ramped up. The last date that I heard about for any railway near me is 2035, that's the earliest).
So how else do we substantially reduce the risk of another collision or derailment caused by speeding?
L1 is far from perfect, but it is better IMHO than trying to fit TPWS to all the remaining signals on lines with a linespeed of 75MPH or greater.
Having said all that, if anyone has a better idea, I welcome hearing about it.
Because there are multiple problems including a lack of contract company staff who can work on existing live signalling. Many installers don't hold SMTH and the number of SWTH testers is not great either. Many of these staff (installers and testers) will be already tied up with existing signalling contracts.
Plus, because of the restrictions on what was called "red zone" working and the limited engineering access when work can be done on live in service signalling equipment.
Would it be efficient, I don't know, Most likely no, but that's difficult to say until the work plan is put together. It most certainly will be less efficient than in the past when BR had a larger (compared to now) department of installers and works testers and there were less difficulties getting access to the equipment.
This is the problem of ETCS fitment having been put back and back and back. It seems a very long time ago that I and the others in my depot were briefed that the GWML would be fitted with ETCS within the next ten years or so (they were describing L2 for the future but I can't remember if the intention at that time was to start with L1). That was some time before the electrification programme got going...
The technology exists to prevent (or at the very least substantially reduce the consequences) of a collision of trains on the same line. And the technology has existed for many, many years in one form or another. Yet, a collision that likely could have been prevented has occurred (although we need to wait for the RAIB to report before we know all the facts).
I can't tell what will happen in the future, and no one else knows either, but the risk of another collision (or derailment due to excessive speed at/over junctions, points or curves) that could result in a death or deaths and serious injuries remains substantially the same since the national TPWS programme concluded. With no substantial change to signals / locations that have not been fitted with TPWS at all. The only significant improvements have been in driver training, signal sighting and rolling stock improvements (better designs that can better cope with an impact) for new trains.
Look at how bad the U.K. railways looked to the general public after the Ladbroke Grove rail crash which occurred not long after the Southall rail crash.
So IMHO doing nothing is unwise. The L2 programme is too slow. And I can't see it covering the whole network within the next twenty years even if the rollout was ramped up. The last date that I heard about for any railway near me is 2035, that's the earliest).
So how else do we substantially reduce the risk of another collision or derailment caused by speeding?
L1 is far from perfect, but it is better IMHO than trying to fit TPWS to all the remaining signals on lines with a linespeed of 75MPH or greater.
Having said all that, if anyone has a better idea, I welcome hearing about it.
Sorry I had got into the trying to keep sensible mindset, and not thinking about the actual topic of the thread (but also remembering there are still unanswered questions that are for RAIB and the like to find out if possible).
This is just my initial opinion and my opinion not anyone I work for or organisations I am a member of (as an aside - anyone know how to add signatures/footers on this forum-I have seen some but can't find the option/setting)
I don't think every signal should be fitted with TPWS, for example a plain bit of double track railway where you can see to the next signal would possibly be excessive for TPWS as we want to put funds toward ETCS but:
If we want a 'quick' solution:
For isolated areas away from ERTMS/ETCS lines I think changing the TPWS calculations and adding Loops, for example like a full aspect sequence after the Junction/crossover, but with ALARP I don't know how the cost would come out. My thinking here is that a full aspect sequence the driver would maybe be out of the "I always get greens as the train in front pulls away" on a new line mentality and get into the new aspects, but that is for a human factors expert to work out. This has the advantage that train fleets are fitted and lots of competent staff to design, install and test the trackside equipment which is simple to test.
Areas with a fleet that has ERTMS fitted you could put a small infill ETCS bit of either level 1 or 2 depending on the existing lineside. But that will be as ever no use if the trains don't have the onboard hence the above. And you can't just put ETCS in and test with a meter like with AWS/TPWS each little island still would need a lot of testing. ETCS is more of a system needs more testing with trains and other systems.
In an ideal world with no other funding issues the rail industry would get more funding for a national ETCS roll out, as, although there Signal works testers (and all the other bits you need like designers) are a small pool compared to the UK workforce overall there are ones leaving the industry as there is not enough projects to work on and a set plan you can increase workforce
The Radio Infill Unit doesn't really address anyone, the connection is made on the initiative of the train using information it obtained from a balise in the track. This would still be fundamentally an ETCS Level 1 system, not a level 2 one.
I am losing track slightly of what you are proposing here. At one point you seemed to be proposing a unit that communicates with the train via its L2 GSM-R, now you seem to be proposing a totally different L1 radio infill unit? Obviously, a train that has only been fitted for L2 won't have been fitted with such a receiver.
I could forsee a problem with the GSM-R approach, if every signal on the approach to say Paddington were fitted with its own totally independent GSM-R base station, with coverage out to its outermost balise. A lot of overlapping coverage, with no guarantee that the GSM-R receiver will latch onto the correct one. Similarly with a L1 radio infill unit, every signal is going to need its own radio transmitter, with broadcast coverage out to its outermost balise. There will be a significant number of totally-independent transmitters, with significantly overlapping coverage. Rejecting messages from the wrong infill unit would be the easy bit, establishing communication with the right one might be the hard bit.
If the train can't establish a connection the train just proceeds with the aspect it obtained from the balise at the start of the signal section.
Unless the signaller puts the next signal back against them after they are in section there is no safety issue, and that is obviously a.... not-ordinary safe operation in any case.
If that happens you are still no worse than with a visual signal which the driver can no longer see (because they've passed it).
The problem with having no infill is what happens when a signal steps up. The driver can see that the signal has gone to green, but the on-board ETCS is still commanding him to stop at it, based on the last information it received. Agreed this is not a safety issue, but it is a big performance issue. It is the main reason most ATP systems have some kind of infill.
== Doublepost prevention - post automatically merged: ==
It's safety approved in the equipment sense, but converting a relay interlocking into an SSI one that can be fitted for ETCS Level 2 is going to require a lot of specific work that will need highly complex safety approvals, surely.
Relock with SSI already has safety approval. The equipment, configurations and methodology is well documented and approved. You could start work on a relock tomorrow.
As noted by other posters above, much of the documentation of relay interlockings built decades ago could be missing or incomplete and confirming that the information is accurate could prove to be a laborious task, especially if the notorious brittle wire problem surfaces. Perhaps my perception is distorted by the pictures of relay interlockings in things like the Hidden Report, but my own experience of troubleshooting wiring arrangements much less complex than those (and with lower stakes for undetected faults or mistakes!) is one of pain and tedium.
Some of the documentation may be missing, particularly on site, but to say much of it would be going too far. Where site records are missing, they can usually be recreated from centrally-held records, and vice versa. The issue is more that small changes may have been made on site, such as moving a circuit from one cable core to another due to a faulty core. This kind of issue can affect both the trackside multi cores, and tail cables to things like signal heads. So a relock would require correlation of the central interlocking, in the dry in a safe environment. Whereas fitting L1 equipment trackside would require every affected equipment case to be correlated, working trackside with all the access, weather, safety, etc issues that would involve.
And as I said previously, you wouldn't start by backwards engineering from the existing. You would use automatic data prep tools to develop the interlocking data from scratch, to current standards. You would only use the existing documentation to find any quirks in the existing, to decide whether or not you want to incorporate them in the new.
There would surely be a massive outcry if an accident were to happen on a line that we had gone to all the time and expense to fit with ETCS, due to a deficiency inherited from a decades-old existing interlocking. Which would be a possibility with any retrofit approach.
Most early schemes would need to be dual signalled, to allow for stock that wasn't currently fitted. But as the rollout progresses, and more and more stock is fitted, then the need for dual signalling will become less and less, as recognised in the NR position papers. So as the rollout progresses, NR will increasingly want to go straight to signal-less ETCS, So with a relock there would be no need to provide interface circuitry for all the signals, only for the point and train detection. This would considerably reduce the amount of relock work required, allowing the rollout to accelerate. The existing outgoing signal circuits from the interlocking would simply be disconnected, fuses pulled, and the signals would be bagged, and recovered when convenient.
Whereas if you are doing a L1 retrofit, you inherently have to keep the signals. And subsequently going to signal-less L2 would require the interlocking replaced, either by relock or resignalling. Effectively, all you have done is postpone the work, at great expense.
Don't get me wrong, I am proposing relock as a quicker and cheaper alternative to achieving L2 than doing a full resignalling, as it wouldn't require all the trackside work that the latter would require.
I really don't see any benefit to going for an L1 retrofit. As I say, you would effectively be spending billions just to defer work that is still going to need doing to achieve signal-less railway. More importantly, you would be keeping any deficiencies in the existing signalling. And you would be making the existing signalling more expensive to maintain, and with more to fail, with the risks that imports. Whether you would make the railway any safer is questionable.
I can understand why continental railways might go for such a retrofit approach, particularly where ETCS replaces existing ATP equipment. They are under political pressure to achieve interoperability, and with funding made available for it. But that is not the case in the UK. Here the pressure is to reduce costs and increase safety. An approach that increases costs and adds risks is going to be hard to justify.
I am losing track slightly of what you are proposing here. At one point you seemed to be proposing a unit that communicates with the train via its L2 GSM-R, now you seem to be proposing a totally different L1 radio infill unit? Obviously, a train that has only been fitted for L2 won't have been fitted with such a receiver.
I could forsee a problem with the GSM-R approach, if every signal on the approach to say Paddington were fitted with its own totally independent GSM-R base station, with coverage out to its outermost balise. A lot of overlapping coverage, with no guarantee that the GSM-R receiver will latch onto the correct one. Similarly with a L1 radio infill unit, every signal is going to need its own radio transmitter, with broadcast coverage out to its outermost balise. There will be a significant number of totally-independent transmitters, with significantly overlapping coverage. Rejecting messages from the wrong infill unit would be the easy bit, establishing communication with the right one might be the hard bit.
The problem with having no infill is what happens when a signal steps up. The driver can see that the signal has gone to green, but the on-board ETCS is still commanding him to stop at it, based on the last information it received. Agreed this is not a safety issue, but it is a big performance issue. It is the main reason most ATP systems have some kind of infill.
Agreeing with and adding to the above, ETCS in complex areas like a station throat of a major terminus would be either a normal level 2, or already have TPWS at all the critical signals anyway. That's why my personal opinion is that maybe a few additional TPWS grids on the exit of junctions/crossovers and then more full ETCS schemes will be the best way forward.
I am losing track slightly of what you are proposing here. At one point you seemed to be proposing a unit that communicates with the train via its L2 GSM-R, now you seem to be proposing a totally different L1 radio infill unit? Obviously, a train that has only been fitted for L2 won't have been fitted with such a receiver.
There is no hardware difference between a Level 1 and Level 2 GSM radio. The GSM-R data radio on the train is functionally a mobile phone, it can dial any number it wants in the ETCS numbering plan.
In Level 2 operation, the train maintains a circuit switched connection to the relevant radio block centre all the time, and receives the relevant information for many signals from the same location.
In Level 1 radio infill operation the train dials a different number for each signal, dropping each call and dialling the next, just as it would if it was moving between radio block centres. It just does it a lot more often.
I could forsee a problem with the GSM-R approach, if every signal on the approach to say Paddington were fitted with its own totally independent GSM-R base station, with coverage out to its outermost balise. A lot of overlapping coverage, with no guarantee that the GSM-R receiver will latch onto the correct one. Similarly with a L1 radio infill unit, every signal is going to need its own radio transmitter, with broadcast coverage out to its outermost balise. There will be a significant number of totally-independent transmitters, with significantly overlapping coverage. Rejecting messages from the wrong infill unit would be the easy bit, establishing communication with the right one might be the hard bit.
GSM-R is a mobile network (just like regular GSM), assuming both end points have a signal and are in the same numbering plan, they can talk to each other regardless of the radio conditions for a link directly between them. There aren't really independent radio connections.
The radio infill units are (just like the train radios) functionally similar to mobile phones which just give the signal information to callers who ask for it using the correct IDs (and obviously dial the right number).
EDIT: The radio infill number is provided to the train using an ETCS Packet 133 "Radio Infill Area information", this would be provided by the previous signal balise or by a fixed balise in the track if one is being provided for some other reason.
The problem with having no infill is what happens when a signal steps up. The driver can see that the signal has gone to green, but the on-board ETCS is still commanding him to stop at it, based on the last information it received. Agreed this is not a safety issue, but it is a big performance issue. It is the main reason most ATP systems have some kind of infill.
Yes, this would only be a problem if the GSM-R call fails for whatever reason. The ETCS level 1 spec requires a call to a radio infill unit that fails to be retried until a connection is achieved or until the train leaves the infill area/recieves new infill information. So if there is a GSM-R blackspot the train won't get infill information (but wouldn't get it in ETCS Level 2 operation either!) until it passes out of the black spot.
Some of the documentation may be missing, particularly on site, but to say much of it would be going too far. Where site records are missing, they can usually be recreated from centrally-held records, and vice versa. The issue is more that small changes may have been made on site, such as moving a circuit from one cable core to another due to a faulty core. This kind of issue can affect both the trackside multi cores, and tail cables to things like signal heads. So a relock would require correlation of the central interlocking, in the dry in a safe environment. Whereas fitting L1 equipment trackside would require every affected equipment case to be correlated, working trackside with all the access, weather, safety, etc issues that would involve.
And as I said previously, you wouldn't start by backwards engineering from the existing. You would use automatic data prep tools to develop the interlocking data from scratch, to current standards. You would only use the existing documentation to find any quirks in the existing, to decide whether or not you want to incorporate them in the new.
What are the auto data prep tools actually fed?
The original design documents/control tables that describe what the interlocking is required to achieve?
Are they in a standard, machine readable format, I'm imagining typewriter type-set documents like the government reports I read in my field from the 1980s.
There would surely be a massive outcry if an accident were to happen on a line that we had gone to all the time and expense to fit with ETCS, due to a deficiency inherited from a decades-old existing interlocking. Which would be a possibility with any retrofit approach.
Probably, but accidents caused by deficiency in old interlockings causing seem to be quite rare. Clapham Junction only occurred due to a mistake during resignalling.
When was the last time one occurred? I can't think of one if I'm honest.
I suppose you could make an argument for Ladbrooke Grove and the lack of flank protection, but ETCS train protection probably would have prevented the accident just as functioning ATP would have done.
It's a risk certainly, but I'm not sure how big of one it is.
Most early schemes would need to be dual signalled, to allow for stock that wasn't currently fitted. But as the rollout progresses, and more and more stock is fitted, then the need for dual signalling will become less and less, as recognised in the NR position papers. So as the rollout progresses, NR will increasingly want to go straight to signal-less ETCS, So with a relock there would be no need to provide interface circuitry for all the signals, only for the point and train detection. This would considerably reduce the amount of relock work required, allowing the rollout to accelerate. The existing outgoing signal circuits from the interlocking would simply be disconnected, fuses pulled, and the signals would be bagged, and recovered when convenient.
Resignalling projects continue to occur without even an ETCS overlay though, so I think its going to be a very long time until ETCS resignalling or relocks can occur without any signals. If we ever reach that position then I might agree that conversion to ETCS Level 2 might be best, but we are a very long way from that.
Whereas if you are doing a L1 retrofit, you inherently have to keep the signals. And subsequently going to signal-less L2 would require the interlocking replaced, either by relock or resignalling. Effectively, all you have done is postpone the work, at great expense.
Yes, you have postponed the work, but in many cases the interlockings will serve for many more decades. Future replacement costs would be discounted into the future and be less of a burden today.
It is essential that ETCS achieve a critical mass that allows the elimination of all other signalling systems in order to achieve its benefits. On that I think we agree.
There is no hardware difference between a Level 1 and Level 2 GSM radio. The GSM-R data radio on the train is functionally a mobile phone, it can dial any number it wants in the ETCS numbering plan.
In Level 2 operation, the train maintains a circuit switched connection to the relevant radio block centre all the time, and receives the relevant information for many signals from the same location.
In Level 1 radio infill operation the train dials a different number for each signal, dropping each call and dialling the next, just as it would if it was moving between radio block centres. It just does it a lot more often.
Are you sure that the RBC is just a number that the train dials? My understanding was that there needed to be a direct data connection between the RBC and the main GSM-R base stations, as the RBC isn't just a standard "handset".
What are the auto data prep tools actually fed?
The original design documents/control tables that describe what the interlocking is required to achieve?
Are they in a standard, machine readable format, I'm imagining typewriter type-set documents like the government reports I read in my field from the 1980s.
Each contractor has developed their own tools, so it is difficult to be specific here. But in general they will be based on a track plan. Ideally, on a new scheme, this is a much simplified extract of the CAD scheme plan, but if need be it can be a simple sketch drawn from scratch showing the relative positions of points, signals, etc. You enter some general parameters, such as what type of signal replacement, approach locking controls, etc. you want as standard. Then there will be a table listing for each route what particular controls are required for each route. This can be relatively quickly entered from scratch, based on information in the design spec or from the control tables. You need to enter the start and destination of each route (so that the tool can work out what points and track circuits are in that route), but most entries will be just yes or nos (yes, we want delayed replacement, no we don't want comprehensive approach locking), and the auto prep tool will then work out how to achieve the required controls.
The automatic data preparation tools are also used to produce automatic validation tools. These look for potential anomalies in the data, and run a suite of routiner type tests on the data, trying every sequence of signaller commands and making sure that a set of points won't move if the dead-locking track circuits are occupied for example. These tools have also been used to revalidate existing data, after a number of incidents where existing interlockings with manually produced data were found to have deficiencies.
Probably, but accidents caused by deficiency in old interlockings causing seem to be quite rare. Clapham Junction only occurred due to a mistake during resignalling.
When was the last time one occurred? I can't think of one if I'm honest.
There have been a number of very near misses, but fortunately no accidents that I am aware of. Generally, the older an interlocking, the more likely that deficiencies will have been found. But equally, older interlockings can contain practices that would be unacceptable today. For example, shunt signal controls used to be very lax. And overrun controls and flank protection may not be what would be deemed acceptable these days.
Resignalling projects continue to occur without even an ETCS overlay though, so I think its going to be a very long time until ETCS resignalling or relocks can occur without any signals.
Provided that the resignalling is done with an ETCS compatible interlocking, then the overlay can be added later. But I do agree, by not providing the overlay now, they are just creating a bigger and bigger bow wave of work to be done in the future.
Yes, you have postponed the work, but in many cases the interlockings will serve for many more decades. Future replacement costs would be discounted into the future and be less of a burden today.
But you will have spent billions of pounds up front, in order to postpone the work. And left yourself with a more complex, expensive and potentially more dangerous railway in the interim.
The critical mass is mainly in terms of stock fitment. Once that is achieved, and dual signalling is no longer needed, then it will be much quicker and cheaper to roll out. Unfortunately stock is still being procured without any on board ETCS.
My view is that any resignalling being done now should not just have an ETCS compatible interlocking, but it should also include all the data necessary to interface to the RBC. And likewise any stock being procured now should not just have provision for future ETCS fitment, but it should actually be fitted and test run with ETCS equipment for acceptance.
Are you sure that the RBC is just a number that the train dials? My understanding was that there needed to be a direct data connection between the RBC and the main GSM-R base stations, as the RBC isn't just a standard "handset".
If any ETCS people show up they'd have to confirm, but I think that this specialised data connection is more a result of the need for the RBC to sustain dozens or hundreds of simultaneous circuit-switched connections, rather than an inherent property of GSM-R. Otherwise you'd need rack after rack of radios in the building, and them all being in the same place and running nearly continuously could cause spectrum issues.
The analogy I'd suggest is a call centre, where everyone dials the same number but dozens of separate circuits are provided to the endpoint, hidden from the user.
If any ETCS people show up they'd have to confirm, but I think that this specialised data connection is more a result of the need for the RBC to sustain dozens or hundreds of simultaneous circuit-switched connections, rather than an inherent property of GSM-R. Otherwise you'd need rack after rack of radios in the building, and them all being in the same place and running nearly continuously could cause spectrum issues.
The analogy I'd suggest is a call centre, where everyone dials the same number but dozens of separate circuits are provided to the endpoint, hidden from the user.
Yep, it's getting to the limits of my knowledge too. I was a signal engineer, so my knowledge of how the comms side of it works is limited.
I recall reading that problems had been found on a continental railway with trains pairing with the wrong base station and so being unable to communicate with the RBC. Which is what makes me suspect that it might not be the same as a typical handset-type call that can communicate with any other handset anywhere.
If the trainborne ETCS can only communicate with the RBC physically linked to that base station (whether a GSM-R or ETCS constraint), then a system with effectively many small RBCs in very close proximity might not be possible.
We are aware of an issue with emails from the Forum to Microsoft-based email accounts (hotmail/outlook/live.com email addresses). This is being looked into currently, thanks for your patience meanwhile.
