Factors influencing what signals are replaced when

[brad465]

If I was in charge of replacing railway signals, my natural instinct would be to replace all the oldest signals on the network possible within available budgets, then repeat this for each budget period. Ultimately this is not always possible, especially as major infrastructure changes that change track/station layouts ultimately come with a need for new signals, while I suspect you also get the odd permanent failure or accident (like with the South Bermondsey gantry collision years ago).

However what I don't understand is why I've seen signals that date from the early 2000s or later have been replaced with the latest forms of LED signals, when there are still signals from the 1980s or perhaps earlier working elsewhere on the network, including on some of our core mainlines? I'm not saying this is necessarily wrong, I just don't know the reasons why this is what appears to be done.

Examples I'm thinking of in particular are in my local county Kent, where all the oldest NK division signals date back to a major resignalling programme in the early 2000s (I can't remember the year as it was early school years for me back then). In the last few years some of them have been randomly replaced it appears (ignoring replacements linked to Abbey Wood and Gravesend redevelopments), in contrast to many of the nearby VS division signals still going, which maybe from the 1980s at least (Victoria station has new LEDs, but not most of the Chatham Mainline within the same division).

 

[Joseph_Locke]

Because the signal is all you are looking at. Behind it are changes in control centres, interlockings, technical progress, standards, operational track layouts, etc., all of which could cause the little twinkly light to have to be replaced.

Anyway, at least you've got colour light signals; major routes in the North (you know, that space between Watford Gap and Edinburgh) still have signalling equipment from the 1880s.

 

[Belperpete]

If I was in charge of replacing railway signals, my natural instinct would be to replace all the oldest signals on the network possible within available budgets, then repeat this for each budget period. Ultimately this is not always possible, especially as major infrastructure changes that change track/station layouts ultimately come with a need for new signals, while I suspect you also get the odd permanent failure or accident (like with the South Bermondsey gantry collision years ago).

However what I don't understand is why I've seen signals that date from the early 2000s or later have been replaced with the latest forms of LED signals, when there are still signals from the 1980s or perhaps earlier working elsewhere on the network, including on some of our core mainlines? I'm not saying this is necessarily wrong, I just don't know the reasons why this is what appears to be done.

Examples I'm thinking of in particular are in my local county Kent, where all the oldest NK division signals date back to a major resignalling programme in the early 2000s (I can't remember the year as it was early school years for me back then). In the last few years some of them have been randomly replaced it appears (ignoring replacements linked to Abbey Wood and Gravesend redevelopments), in contrast to many of the nearby VS division signals still going, which maybe from the 1980s at least (Victoria station has new LEDs, but not most of the Chatham Mainline within the same division).

I believe it is NR's intention to replace all their filament signals with LED signals. It would not make sense to replace signals in an area that is slated for resignalling in the forseeable future. That might explain why some of the newer signals are being replaced by LED, when some of the much older are not.

However, that assumes some kind of logic. We used to have a saying that the painters turning up to repaint your box was a sign that it was due to be abolished.

 

[Tio Terry]

Back in 1964, soon after I started, I was taken to Brundall Junction by new boss for a bit of induction. I can remember looking at the junction signal with it’s two arms and saying something like “I suppose that will be replaced with a colour light within a few years”. His reply was along the lines of “Son, there will still be semaphores when you retire”. I retired, after 53 years service, in 2017, that signal was still standing then and is only now about to be replaced as part of the Wherry Lines resignalling project. Mechanically there’s nothing wrong with it, it could probably last another fifty years, it’s just being overtaken by technology.

Replacing incandescent signal lamps with LED’s is very much a reliability issue. In todays delay attribution driven railways every effort is made to reduce the minutes of delays to trains. High intensity train services quickly rack up the delay minutes when something goes wrong so those lines with highest intensity of services are likely to benefit before those with one or two trains an hour.

 

[brad465]

Just been through the Hither Green area from Crayford and discovered an ongoing resignalling project, which when complete will see LEDs replace LEDs. Some of the existing LEDs are no more than a few years old, so I wonder what the point of them being fitted was if this larger resignalling project was always planned?

 

[edwin_m]

As well as the reliability benefit, LEDs have a safety benefit as there is less need to send someone into a dangerous trackside environment to change a bulb. And they can replace filament lamps with only small changes to nearby wiring, and attach to the same fixings, so it's quite an easy job and may be worth doing even if the whole signal will be replaced a few years down the line. If that happens the signal heads may be kept in store for use elsewhere.

There are also LEDs available that just replace the lamps themselves inside the existing signal head, so the signal still appears to be "old" but has in fact been upgraded. Just about the only way to tell from the outside would be to see if the lamps come on and off instantaneously rather than getting brighter and dimmer over a fraction of a second as filaments do.

 

[hwl]

Just been through the Hither Green area from Crayford and discovered an ongoing resignalling project, which when complete will see LEDs replace LEDs. Some of the existing LEDs are no more than a few years old, so I wonder what the point of them being fitted was if this larger resignalling project was always planned?

Reliability (especially during London Bridge rebuild) and huge reduction in maintenance in the mean time. It also slightly reduces power consumption which can be useful in some circumstances.
Signal bulbs are dual filament, when the first one fails the second takes over and the bulb is listed for replacement. Changing bulbs on busy routes can be a bit of a logistical challenge.

The current work is part of the recontrol and shift to Three Bridges ROC

 

[Llama]

As well as the reliability benefit, LEDs have a safety benefit as there is less need to send someone into a dangerous trackside environment to change a bulb. And they can replace filament lamps with only small changes to nearby wiring, and attach to the same fixings, so it's quite an easy job and may be worth doing even if the whole signal will be replaced a few years down the line. If that happens the signal heads may be kept in store for use elsewhere.

There are also LEDs available that just replace the lamps themselves inside the existing signal head, so the signal still appears to be "old" but has in fact been upgraded. Just about the only way to tell from the outside would be to see if the lamps come on and off instantaneously rather than getting brighter and dimmer over a fraction of a second as filaments do.

The colour of the light from LED 'light engines' (if that's what they're still called) is also a giveaway, especially greens.
Most drivers I know hate Dorman LED signals, and the newer VMS type. Far too bright. I've found myself squinting to try to cut out the glare from one of the new VMS signals showing DY even though it was early July and low summer sun about 8.30pm was shining from behind into the face of the signal. If they're too bright in that instance they're too bright for every other instance.

 

[AlastairFraser]

I suppose what would be sensible is to replace the whole lot with ECTMS gradually to save on infrastructure costs and increase line speeds. The only pitfall arethat this would probably require a new LTE/5G replacement for GPRS to support the new system which would cost quite a lot and legacy locomotives would be have to retrofitted at great cost,though this could mitigated by having both systems active for a certain length of a transition period.

 

[londonmidland]

In some areas, particularly where new track/layouts are laid, you will have the ‘newer style’/lightweight LED signals replace existing ‘older’ LED signals.

This is the case on the MML where new electrification and track is being installed.

 

[hooverboy]

I suppose what would be sensible is to replace the whole lot with ECTMS gradually to save on infrastructure costs and increase line speeds. The only pitfall arethat this would probably require a new LTE/5G replacement for GPRS to support the new system which would cost quite a lot and legacy locomotives would be have to retrofitted at great cost,though this could mitigated by having both systems active for a certain length of a transition period.

yes it needs doing, but as a complimentary feature rather than a replacement.

the in-cab signalling will serve as a backup rather than a primary indicator.
it will certainly have its uses in inclement weather conditions,or where speeds/track alignments mean there is only a short time to observe and react.

such a system though would be capable of relaying distant signal aspects and upcoming speed/points changes direct to cab much further in advance.

you could probably do away with banner repeaters and so on, but not the visual signal itself.
the need for extensive route knowledge woud be reduced a bit, allowing faster training,and easier cross-signing of routes.

in the event of a systems failure, you still need the backup of "proceed on sight"..aka the present signal system

 

[edwin_m]

yes it needs doing, but as a complimentary feature rather than a replacement.

the in-cab signalling will serve as a backup rather than a primary indicator.
it will certainly have its uses in inclement weather conditions,or where speeds/track alignments mean there is only a short time to observe and react.

such a system though would be capable of relaying distant signal aspects and upcoming speed/points changes direct to cab much further in advance.

you could probably do away with banner repeaters and so on, but not the visual signal itself.
the need for extensive route knowledge woud be reduced a bit, allowing faster training,and easier cross-signing of routes.

in the event of a systems failure, you still need the backup of "proceed on sight"..aka the present signal system

Eliminating the lineside signals is exactly what is intended, at least on routes where all the trains will have the necessary equipment. That is what has happened on the Cambrian pilot scheme.

There is a benefit in improved safety, but since TPWS and other measures have largely eliminated SPAD-related accidents this benefit is small. There is a capacity benefit from not having to worry about sighting when positioning the block markers that replace the signals, and by permitting much shorter blocks which wouldn't be possible with lineside signals because they can only show four aspects. None of this is possible if the signals are retained.

The same interlocking drives the ERTMS system as drives the lineside signals, so if the signals were kept then most failures would affect both systems.

 

[DarloRich]

yes it needs doing, but as a complimentary feature rather than a replacement.

the in-cab signalling will serve as a backup rather than a primary indicator.
it will certainly have its uses in inclement weather conditions,or where speeds/track alignments mean there is only a short time to observe and react.

such a system though would be capable of relaying distant signal aspects and upcoming speed/points changes direct to cab much further in advance.

you could probably do away with banner repeaters and so on, but not the visual signal itself.
the need for extensive route knowledge woud be reduced a bit, allowing faster training,and easier cross-signing of routes.

in the event of a systems failure, you still need the backup of "proceed on sight"..aka the present signal system

I am not sure you quite understand the topic under discussion

@edwin_m has it spot on

 

[brad465]

Reliability (especially during London Bridge rebuild) and huge reduction in maintenance in the mean time. It also slightly reduces power consumption which can be useful in some circumstances.
Signal bulbs are dual filament, when the first one fails the second takes over and the bulb is listed for replacement. Changing bulbs on busy routes can be a bit of a logistical challenge.

The current work is part of the recontrol and shift to Three Bridges ROC

I suspected the Hither Green area works are related to the Three Bridges shift, the only thing I can't understand is why some of the existing LED signals in the area (Mottingham-HGR and a few others towards Lewisham) replaced 4-aspect non-LEDs around 3-5 years ago at most, when the Three Bridges shift must at least have been planned, so could have held out a bit longer without what is essentially a short "stop-gap" replacement? (if this makes sense)

Edit: Apologies I might have misread your first sentence as a generic advantage to LED signals, if this was your answer to the specific project in question.

 

[hwl]

I suspected the Hither Green area works are related to the Three Bridges shift, the only thing I can't understand is why some of the existing LED signals in the area (Mottingham-HGR and a few others towards Lewisham) replaced 4-aspect non-LEDs around 3-5 years ago at most, when the Three Bridges shift must at least have been planned, so could have held out a bit longer without what is essentially a short "stop-gap" replacement? (if this makes sense)

Edit: Apologies I might have misread your first sentence as a generic advantage to LED signals, if this was your answer to the specific project in question.

Hither Green - Yes they are.

Both generic advantage but even several years can be worth while especially in high traffic areas with large reactionary delay and often difficult access.
With signalling reliabilty in the area falling it is also a simple change to improve overall reliability.

 

[brad465]

Hither Green - Yes they are.

Both generic advantage but even several years can be worth while especially in high traffic areas with large reactionary delay and often difficult access.
With signalling reliabilty in the area falling it is also a simple change to improve overall reliability.

Considering the old signals in the area must date back to at least the 1980s (not sure exactly how old they are) I can see why reliability has been an issue there.

I don't know if Network Rail carry out "signal recycling" but I imagine if it was possible for LED signals like the ones in this example, they could probably go somewhere else on the network, being only a few years old.

 

[hwl]

Considering the old signals in the area must date back to at least the 1980s (not sure exactly how old they are) I can see why reliability has been an issue there.

I don't know if Network Rail carry out "signal recycling" but I imagine if it was possible for LED signals like the ones in this example, they could probably go somewhere else on the network, being only a few years old.

Early 70's not 80's.
A few of the early LED head units you would not want to reuse but most should be reusable an issue is how much retesting before reuse is just makes it cheaper to go for new.

 

[AlastairFraser]

Eliminating the lineside signals is exactly what is intended, at least on routes where all the trains will have the necessary equipment. That is what has happened on the Cambrian pilot scheme.

There is a benefit in improved safety, but since TPWS and other measures have largely eliminated SPAD-related accidents this benefit is small. There is a capacity benefit from not having to worry about sighting when positioning the block markers that replace the signals, and by permitting much shorter blocks which wouldn't be possible with lineside signals because they can only show four aspects. None of this is possible if the signals are retained.

The same interlocking drives the ERTMS system as drives the lineside signals, so if the signals were kept then most failures would affect both systems.

Thanks for sharing your industry knowledge? I have a couple questions. Firstly- isn't traditional signalling partly very vulnerable to theft because the power and signal cables to the signals are vulnerable to weather/theft? I know there can be glitches in computer software that would affect in-cab signalling,but this would be a lot more reliable than traditional signalling in that way yes?
Secondly,the problem as I highlighted with having in-cab signalling only is that the equipment is expensive to install to every train and a network-wide installation of LTE/5G would be incredibly expensive,almost prohibitive to start with. So would it be possible to reap some of the benefits of in-cab signalling without converting fully to just in-cab signalling over a say 5 year transition period? Once again,thanks for your help. I know I'm a serial thread derailed,so I'll get back to the point. In my view,in-cab signalling would reduce the redundancy in physical kit that often gets replaced only half way through its design life because an area is being remodelled or new tech is being introduced which is wasteful. It seems we need to design one in-cab technology which is reliable,efficient and flexible,so you don't have to install new signalling kit everytime we change the railway layout or upgrade it etc.

 

[edwin_m]

Thanks for sharing your industry knowledge? I have a couple questions. Firstly- isn't traditional signalling partly very vulnerable to theft because the power and signal cables to the signals are vulnerable to weather/theft? I know there can be glitches in computer software that would affect in-cab signalling,but this would be a lot more reliable than traditional signalling in that way yes?
Secondly,the problem as I highlighted with having in-cab signalling only is that the equipment is expensive to install to every train and a network-wide installation of LTE/5G would be incredibly expensive,almost prohibitive to start with. So would it be possible to reap some of the benefits of in-cab signalling without converting fully to just in-cab signalling over a say 5 year transition period? Once again,thanks for your help. I know I'm a serial thread derailed,so I'll get back to the point. In my view,in-cab signalling would reduce the redundancy in physical kit that often gets replaced only half way through its design life because an area is being remodelled or new tech is being introduced which is wasteful. It seems we need to design one in-cab technology which is reliable,efficient and flexible,so you don't have to install new signalling kit everytime we change the railway layout or upgrade it etc.

Cab signaling will eliminate the signals and the associated cables but they will still be needed for points and for train detection. Any future ERTMS Level 3 (which doesn't yet exist) would also eliminate much of the train detection hardware. Any cable removed is one less item to fail, but there is also the risk of glitches in radio communications (hopefully not in the software!) which might cause everything to stop. So the radio itself needs to be highly reliable and have enough bandwidth for all the communications needed, which the current version of GSM-R doesn't.

More recent train deliveries include the necessary equipment for ERTMS, or at least are designed to make it easy to install that equipment at a later date. The strategy for transition to ERTMS takes account of where those newer trains run. For example the southern part of the ECML is a priority, because the 1970s signaling is due for replacement but also because nearly all the trains using it (80x, 700, 717) are "ERTMS ready". There is a government funded programme to fit freight locomotives, which need to be able to go virtually anywhere. It's possible that conventional signals will be retained in "islands" where non-fitted trains will still need to use the route, such as around Peterborough.

The most important bit of the signaling that you don't see is the interlocking. Originally mechanical, this is the safety logic that prevents unsafe events such as trains being signaled into collision or points being moved under trains. Most of the re-signaling done before the late 1980s has used large interlockings consisting of hundreds of electrical relays housed in a brick shed. These are hardwired to the signals, points and track circuits so need a lot of cables. Modifying the function of this signaling to take account of layout changes is difficult as the wiring within the interlocking has to be changed while effectively still in use. The Clapham Junction collision was a result of bad practice while modifying signaling, and the extra precautions introduced afterwards made such modification more difficult and expensive. Also the wires used between the relays in some of these schemes has proved not to be durable, to the extent that the insulation may crumble away if anyone touches it, with obvious safety risks. So practice for this type of signaling has tended to be to replace everything in one go, and make any layout changes at the same time. However peripheral changes are possible, like replacing a filament signal head with a LED version that shows the same aspects under the same conditions.

Liverpool Street and the northern ECML in the late 1980s were the first examples of software-based and data-driven signaling. The trackside equipment was the same as before, but now connected via modules which are linked to the interlocking by a serial data link, so there are fewer cables to be damaged or stolen and there is also a backup data link that takes a different route and can take over if the primary one is damaged. The functionality of these systems is changed by uploading amended configuration data, so the only wiring changes are where the trackside equipment itself has changed. Needless to say there are still a lot of checks and tests to ensure this happens correctly, but it's still easier than changing the older type of signaling.

This in turn means that signaling upgrades can now be more piecemeal than the all-or-nothing approach in the past. I don't know the details of London Bridge but for the Reading upgrade the first stage was to replace the relay interlockings with data-driven systems but keep the existing trackside equipment and functionality. The various stages of upgrading the layout could then be done by changing the interlocking data rather than having to re-wire a relay interlocking.

 

[30740]

I'm told by someone working in the industry that nowadays the lifetime of a signalling system is 20-30 years, beyond which it becomes unmaintainable. The electronic hardware becomes life-expired and irreplaceable. Compare with an older system based on relays, which you can keep repairing for ever.

 

[Bald Rick]

I'm told by someone working in the industry that nowadays the lifetime of a signalling system is 20-30 years, beyond which it becomes unmaintainable. The electronic hardware becomes life-expired and irreplaceable. Compare with an older system based on relays, which you can keep repairing for ever.

Not so.

Relays in interlockings wear out. They are regularly replaced, sometimes individually, sometimes in whole batches. It is a lot, lot more complicated, time consuming and expensive to replace a whole load of relays than a couple of racks of processors in a Computers Based Interlocking cubicle; it is (in simple terms) power down, old racks out, new racks in, power up, load interlocking data (if not pre loaded), run checks, go live. Takes about an hour or so.

Besides, the point is that modern signalling systems consist of 4 main components:

Control system (VDU or Panel based) - the method this tthe signaller instructs the system what to do
Interlocking - the safety critical ‘brain’ of the system
Ground equipment - the signals themselves, plus train detection, points operating equipment, level crossing equipment etc
Transmission - the communications from the control system to the interlocking, and interlocking to the ground equipment.

It is quite feasible to do any one of these items on their own. And this regularly happens. Replacing parts of the ground equipment - such as signal heads - is a routine activity, and is not considered as ‘resignalling’

A full resignalling, however, replaces all of this equipment. A ‘recontrol’ will replace only the control equipment and possibly the transmission. A ‘relock’ replaces only the interlocking and possibly the transmission.

In so far as an LED signal head is concerned, the cost of installing a new signal head is around 1% of the cost of a complete resignalling of one signal (ie the costs of a resignalling divided by the number of signals and points). However it will only be done if it is needed, for example if the old kit is in the wrong position to meet new standards that apply for resignalling, or if it is to be standardised to reduce the training and competence requirements of the local maintainers (and also reduce the parts required to be held in stock). Or, if the transmission system is being changed then it may be cheaper to replace the signal heads than to install a new interface between the new transmission and old signal heads.

Please be assured that the people planning these jobs have a good reason for delivering what they do.

 

[edwin_m]

Relays in interlockings wear out. They are regularly replaced, sometimes individually, sometimes in whole batches. It is a lot, lot more complicated, time consuming and expensive to replace a whole load of relays than a couple of racks of processors in a Computers Based Interlocking cubicle; it is (in simple terms) power down, old racks out, new racks in, power up, load interlocking data (if not pre loaded), run checks, go live. Takes about an hour or so.

And that's for relays that are plug-in components. As I posted above, if the wiring that connects them together is shot then that's a lot more difficult to deal with.

 

[Bald Rick]

And that's for relays that are plug-in components. As I posted above, if the wiring that connects them together is shot then that's a lot more difficult to deal with.

Yes, agreed (and apologies I thought I had tagged you to say that everything you said earlier was an excellent response).

I well remember being shown wire degradation in a relay room of a rather important ‘box on the WCML, shortly followed by a signal reversion and some embarrassed faces all round.

 

[AlastairFraser]

Cab signaling will eliminate the signals and the associated cables but they will still be needed for points and for train detection. Any future ERTMS Level 3 (which doesn't yet exist) would also eliminate much of the train detection hardware. Any cable removed is one less item to fail, but there is also the risk of glitches in radio communications (hopefully not in the software!) which might cause everything to stop. So the radio itself needs to be highly reliable and have enough bandwidth for all the communications needed, which the current version of GSM-R doesn't.

More recent train deliveries include the necessary equipment for ERTMS, or at least are designed to make it easy to install that equipment at a later date. The strategy for transition to ERTMS takes account of where those newer trains run. For example the southern part of the ECML is a priority, because the 1970s signaling is due for replacement but also because nearly all the trains using it (80x, 700, 717) are "ERTMS ready". There is a government funded programme to fit freight locomotives, which need to be able to go virtually anywhere. It's possible that conventional signals will be retained in "islands" where non-fitted trains will still need to use the route, such as around Peterborough.

The most important bit of the signaling that you don't see is the interlocking. Originally mechanical, this is the safety logic that prevents unsafe events such as trains being signaled into collision or points being moved under trains. Most of the re-signaling done before the late 1980s has used large interlockings consisting of hundreds of electrical relays housed in a brick shed. These are hardwired to the signals, points and track circuits so need a lot of cables. Modifying the function of this signaling to take account of layout changes is difficult as the wiring within the interlocking has to be changed while effectively still in use. The Clapham Junction collision was a result of bad practice while modifying signaling, and the extra precautions introduced afterwards made such modification more difficult and expensive. Also the wires used between the relays in some of these schemes has proved not to be durable, to the extent that the insulation may crumble away if anyone touches it, with obvious safety risks. So practice for this type of signaling has tended to be to replace everything in one go, and make any layout changes at the same time. However peripheral changes are possible, like replacing a filament signal head with a LED version that shows the same aspects under the same conditions.

Liverpool Street and the northern ECML in the late 1980s were the first examples of software-based and data-driven signaling. The trackside equipment was the same as before, but now connected via modules which are linked to the interlocking by a serial data link, so there are fewer cables to be damaged or stolen and there is also a backup data link that takes a different route and can take over if the primary one is damaged. The functionality of these systems is changed by uploading amended configuration data, so the only wiring changes are where the trackside equipment itself has changed. Needless to say there are still a lot of checks and tests to ensure this happens correctly, but it's still easier than changing the older type of signaling.

This in turn means that signaling upgrades can now be more piecemeal than the all-or-nothing approach in the past. I don't know the details of London Bridge but for the Reading upgrade the first stage was to replace the relay interlockings with data-driven systems but keep the existing trackside equipment and functionality. The various stages of upgrading the layout could then be done by changing the interlocking data rather than having to re-wire a relay interlocking.

Thanks for the detailed explanation. I forgot about tie movement of points,but could this not be done by transponders to receive the signal from signalling centres via relay transmitters and small substation to power the points. My idea is aiming to reduce the amount of physical kit needed to replace at every resignalling or remodelling to reduce waste and needless expenditure by the railway. On the subject of railway radio networks,that's why I commented that we need a LTE/5G replacement for GSM-R to support the new system,which would be very expensive but hopefully a long-term investment,able to be used for at least half a century. So,in essence,if you replaced the wire-based interlocking with a data-based interlocking,it's a lot easier to upgrade to fully in-cab signalling when the time comes.
So are all new interlockings data-based?

 

[edwin_m]

Thanks for the detailed explanation. I forgot about tie movement of points,but could this not be done by transponders to receive the signal from signalling centres via relay transmitters and small substation to power the points. My idea is aiming to reduce the amount of physical kit needed to replace at every resignalling or remodelling to reduce waste and needless expenditure by the railway. On the subject of railway radio networks,that's why I commented that we need a LTE/5G replacement for GSM-R to support the new system,which would be very expensive but hopefully a long-term investment,able to be used for at least half a century. So,in essence,if you replaced the wire-based interlocking with a data-based interlocking,it's a lot easier to upgrade to fully in-cab signalling when the time comes.
So are all new interlockings data-based?

I guess there's no reason why the points couldn't be actuated by radio - the data stream to the trackside modules is encoded to prevent corruption and to ensure that if it happened to end up at the wrong place it wouldn't be acted on. It can even be sent over shared data networks. It would however have to be a near-continuous data stream as the interlocking will assume the points have lost detection if there is no report back for a second or so. And there would still need to be some cables for power supply and for the control and detection circuits.

I think all major re-signaling schemes use computer-based (data-driven) interlockings but there may be minor schemes where relays are used instead. I read something recently about introduction of the first level crossing without any relays in the control circuits.

 

[MarkyT]

And that's for relays that are plug-in components. As I posted above, if the wiring that connects them together is shot then that's a lot more difficult to deal with.

In my experience that's the most common reasoning used for early interlocking renewal proposals. Large scale relay interlocking rewiring in situ is rarely practical except in the smallest installations, so various other approaches have been developed:

1. 'Relock' and (optional) 'Recontrol'. A modern recreation of the exact same layout using (largely) new internal and external SSI equipment. If a pre-existing layout is being maintained so-called 'grandfather rights' can allow certain standard compliance issues to be consciously avoided, subject to risk assessment. The solution can be very useful to enable future staged remodelling as at Reading, although that was primarily influenced by the need to completely flatten the signalbox building (including the main station area interlocking relay room) to make way for the early construction activity for the new relief side tracks and platforms. Hence a simultaneous recontrol to Thames Valley Signalling Centre (TVSC) was also necessary. Other remote relay interlockings formerly controlled by Reading PSB were retained initially, with TDM control transferred to TVSC.

2. 'Interfaced SSI (or modern equivalent processor-based interlocking)', where all the external equipment and cabling are retained and interfaced with new SSI trackside functional modules (or other more modern i/o devices) housed in the old relay room. The new interlocking can be tested comprehensively beforehand, and the final changeover is a fairly simple wire swap at the existing cable terminals, which can also be rehearsed beforehand in possessions leading up to the main event.

3. Similar to 2, a complete new relay interlocking might be provided inside the existing building instead. In many buildings, there will not be the space available for this however, although it has been done in some cases. Stoke Gifford (Bristol Parkway) was an example on the Western, where a brand new WR E10k relay interlocking was provided in the mid 1990s to solve a wiring condition problem AND cater for the revised track layout associated with the short-lived Royal Mail hub built there. The signalling has been renewed once again on this stretch in connection with electrification, and in that case all the external equipment was also replaced.

I well remember being shown wire degradation in a relay room of a rather important ‘box on the WCML, shortly followed by a signal reversion and some embarrassed faces all round.

At least it was a reversion... When the insulation starts to peel away from the wires due to chemical degredation in a large room with many thousands of wires bundled together tightly in conduits and trays there is a very real possibility of wrong side failures occurring too; something that definitely keeps signal engineers awake at night! The first step after the phenomenon becomes known or suspected at a particular site is to try and limit physical disturbance of the wiring in normal faulting and maintenance activity and instituting an enhanced inspection regime as well as starting to build the case for an appropriate accelerated renewals project. Eventually BR and its successors recognised trends in manufacturers, batches and materials used in wiring that helped identify higher risk locations. GEC interlockings, particularly those from the late 1970s and early 80s used a type of wire that was notorious in this respect. Earlier wiring from the same and other manufacturers installed in the 1960s and early 70s was usually far better and it's notable that many of those interlockings have outlived the troublesome later ones.

 

[MarkyT]

I guess there's no reason why the points couldn't be actuated by radio - the data stream to the trackside modules is encoded to prevent corruption and to ensure that if it happened to end up at the wrong place it wouldn't be acted on. It can even be sent over shared data networks. It would however have to be a near-continuous data stream as the interlocking will assume the points have lost detection if there is no report back for a second or so. And there would still need to be some cables for power supply and for the control and detection circuits.

An experimental ETCS 'Regional' implementation on a very rural railway in Sweden some years ago used islands of trackside equipment controlled through the train radio system that also carried the movement authority. The junctions in the field also retained conventional fixed train detection deadlocking through the points to ensure movement of the switches could never occur under a train but used virtual fixed block concepts for separation elsewhere. https://en.wikipedia.org/wiki/ERTMS_Regional

 

[Raul_Duke]

Are there that many people around who can still do the old mechanical interlocking for manual boxes?

I seem to recall reading in Rail Engineer a bit back about a mechanical interlocking technician retiring in Scotland which halved the number of people in his grade!

 

[MarkyT]

Are there that many people around who can still do the old mechanical interlocking for manual boxes?
I seem to recall reading in Rail Engineer a bit back about a mechanical interlocking technician retiring in Scotland which halved the number of people in his grade!

It's a rare skill today, mostly the preserve of older staff, who often have to travel all over the country to keep the remaining installations properly maintained, periodically tested and overhauled. Out of necessity, the skills can and have been learned by some younger employees. Obtaining spare parts, especially for some of the rarer types of frames, can also be a major challenge.

 

[Bald Rick]

I’m reasonably use there are a couple of locking fitter apprentices on the books.