Where will the last semaphore signals be on the national network?

[MadMac]

Ayr - Stranraer?
Girvan to Stranraer is the last mainline use of tablet working and Glenwhilly uses at least one distant semaphore.

Glenwhilly has the last operational (in theory) semaphore distant in Scotland. Rather bizarrely, we were trying to get rid of the tablet instruments south of Girvan some 30 years ago due to reliability issues!

 

[matchmaker]

Fort William Junction may be around for a while yet. The only semaphores on all of the West Highland lines!

 

[Highlandspring]

Fort William Junction may be around for a while yet. The only semaphores on all of the West Highland lines!

Apart from the previously mentioned Pass of Brander stone signals, which are also likely to remain in use for decades to come following the abject failure of the trial acoustic monitoring system a few years ago.

 

[Ashley Hill]

When one of our Block Instruments went wrong spares came from the Severn Valley Rlwy

This is not unknown. I've seen similar items bought at railwayana auctions for use on NR. It's often cheaper to buy them here and pay to have them refurbished than to purchase "new" from NRS.

Is there still a signal with a half timber post there? I remember seeing it 30 years ago!

No it's been replaced by an upper quadrant. The wooden post signal has been preserved at Yeovil Steam Centre.

 

[PudseyBearHST]

It depends on the type of signalling system in use, but yes in certain circumstances a reflective board ‘distant’ can be used as a distant for a colour light signal.

They cannot however be used if there is a signal in rear (on approach to) where it is required to prove the signals ahead are lit…

Could you please explain this further? I’m having a blank moment and don’t understand it properly. Thanks

 

[Ken H]

Could you please explain this further? I’m having a blank moment and don’t understand it properly. Thanks

if signals are close, and there is not braking distance in the short block, the stopping signal to the rear must be held at red until the next signal is clear.

 

[Annetts key]

Can you explain what OCS and MES are please, and the implications- ie is it decontrolling semaphores, or just a cheaper way of replacing them with colour lights?
thanks.

A OCS or One Control Switch panel is a small panel where there is one switch for each signal on plain line, or where lines converge. If there is a signal that controls a diverging junction, then there is one switch per route from this signal. So if there are two signalled routes, there will be two switches. One for the main route (main line) and one for the diverging route (line). Indications are provided for the colour light signals, and for the track circuits.

The points also use one switch per set of points. And they also have indications.

These panels are much less complex compared to the bigger panels found in power signal boxes. The small ones can fit on a office desk.

MES - minimum electrical signalling, normally means as cheap as possible two aspect colour light signalling with track circuits and point machines. It is designed to be as inexpensive as possible with just the bare minimum facilities needed. And is normally controlled by a OCS panel or other low cost control system.

However, if a mechanical signal box locking frame still has years of service, sometimes the cheaper solution is to replace stuff piecemeal (e.g. replace a motorised semaphore signal with a LED two aspect colour light) and continue to use the locking frame for the interlocking rather than provide electrical interlocking.

The real question is, what is the likely cost / benefit analysis of keeping the existing system (with it’s operating and maintenance costs) compared to the capital costs of removing the old and providing a new system. Especially at a time when the railways in this country are being told by government to cut the running costs and not to overspend on capital works…

 

[Tomnick]

if signals are close, and there is not braking distance in the short block, the stopping signal to the rear must be held at red until the next signal is clear.

Just...no. Although it's true that a signal might be approach released from red when the distance to the next signal is particularly short (and that next signal is at danger), or indeed for other reasons. In any case, if the signal section is shorter than braking distance, what use would you have for a fixed distant in the middle of it?

 

[Annetts key]

Could you please explain this further? I’m having a blank moment and don’t understand it properly. Thanks

So I will give two examples.

Example one - fixed distant sign / board not permitted

On a passenger line that is a track circuit block line (including where axle counters are used instead of, or as well as track circuits) that has colour light stop signals, any ‘distant’ signal has to be proved lit (alight). The distant signal (or the signal acting as the distance signal) must me sufficiently visible both in daylight and at night, because it is this signal that tells the driver that the next signal is at red (danger). The distant signal is positioned at (a minimum of) service braking distance from the next signal. If a driver misses the distant signal, it is very likely that the train will pass the next signal at red and obviously there is then a danger that the train may hit another train ahead or fowl a junction or level crossing.

The proving circuits that detect the distance signal is lit, pass this information on to the control system for the previous signal (the signal in rear, or to put it another way, the signal that allows access to the track circuit block section that leads up to the distant signal.

Example two - ‘unsignalled’ line, token or train staff controlled line leads up to a colour light track circuit block line.

Because an alternative method of controlling train movements is being used, there is no signal in rear. Hence a fixed distant sign / board can be used to provide the driver with a visual reference of when to start breaking for the stop signal ahead. Often this signal will always be red when the train approaches. Especially if the train has to stop to place the token in a token machine / instrument. If the train were to overrun the stop signal, often there is a relatively long amount of track before a point of collision, mainly because the line speed is often relatively slow in these locations.

If the line is a branch line provided only for freight, in addition there is normally a trap point so that it’s less likely that the train will make far enough to fowl the main line. In this instance the stop signal may be a position light mounted on a post (exactly like a ground position light, just post mounted for better visibility.

Note that often for example two, the train is (nearly) always slowing to a stop, or almost to a stop every time. Hence the risk of passing the red signal is low.

 

[PudseyBearHST]

So I will give two examples.

Example one - fixed distant sign / board not permitted

On a passenger line that is a track circuit block line (including where axle counters are used instead of, or as well as track circuits) that has colour light stop signals, any ‘distant’ signal has to be proved lit (alight). The distant signal (or the signal acting as the distance signal) must me sufficiently visible both in daylight and at night, because it is this signal that tells the driver that the next signal is at red (danger). The distant signal is positioned at (a minimum of) service braking distance from the next signal. If a driver misses the distant signal, it is very likely that the train will pass the next signal at red and obviously there is then a danger that the train may hit another train ahead or fowl a junction or level crossing.

The proving circuits that detect the distance signal is lit, pass this information on to the control system for the previous signal (the signal in rear, or to put it another way, the signal that allows access to the track circuit block section that leads up to the distant signal.

Example two - ‘unsignalled’ line, token or train staff controlled line leads up to a colour light track circuit block line.

Because an alternative method of controlling train movements is being used, there is no signal in rear. Hence a fixed distant sign / board can be used to provide the driver with a visual reference of when to start breaking for the stop signal ahead. Often this signal will always be red when the train approaches. Especially if the train has to stop to place the token in a token machine / instrument. If the train were to overrun the stop signal, often there is a relatively long amount of track before a point of collision, mainly because the line speed is often relatively slow in these locations.

If the line is a branch line provided only for freight, in addition there is normally a trap point so that it’s less likely that the train will make far enough to fowl the main line. In this instance the stop signal may be a position light mounted on a post (exactly like a ground position light, just post mounted for better visibility.

Note that often for example two, the train is (nearly) always slowing to a stop, or almost to a stop every time. Hence the risk of passing the red signal is low.

Absolutely fantastic, thanks!
For the first example, would that be allowed under Absolute block then? Is there still some sort of proving circuit for the distant having to be lit before the section signal can clear?

 

[Annetts key]

Absolutely fantastic, thanks!
For the first example, would that be allowed under Absolute block then? Is there still some sort of proving circuit for the distant having to be lit before the section signal can clear?

I’m far from an expert on Absolute block, because it’s been many, many years since I last had anything to do with this system. So I may not get this bit completely correct!

As semaphore signals originally did not have any light proving, then I suspect a fixed distant sign / board may be permitted to replace either a fixed semaphore distant signal or a worked semaphore distant signal. Even if the next signal (the home signal) is a colour light type.

The signal in rear (which will be the starter signal controlled from the previous signal box) will be controlled by signaller with the lever being interlocked with the block instrument. Again this could be a semaphore or a colour light type.

If however the distant signal was replaced with a colour light type, then if I remember correctly, this would be proved alight in the block circuitry.

When considering the provision of distant signals, especially fixed distants, the line speed, signal spacing and the required train service must all be considered.

 

[High Dyke]

IIRC on the original timescale of elimination of former Western Region boxes Tondu was scheduled to close somewhere around 2050!

Swinderby was never on the plan to close whatsoever, but its closure was announced about a year or so ago. However, that has just been put back again.

 

[PudseyBearHST]

I’m far from an expert on Absolute block, because it’s been many, many years since I last had anything to do with this system. So I may not get this bit completely correct!

As semaphore signals originally did not have any light proving, then I suspect a fixed distant sign / board may be permitted to replace either a fixed semaphore distant signal or a worked semaphore distant signal. Even if the next signal (the home signal) is a colour light type.

The signal in rear (which will be the starter signal controlled from the previous signal box) will be controlled by signaller with the lever being interlocked with the block instrument. Again this could be a semaphore or a colour light type.

If however the distant signal was replaced with a colour light type, then if I remember correctly, this would be proved alight in the block circuitry.

When considering the provision of distant signals, especially fixed distants, the line speed, signal spacing and the required train service must all be considered.

Appreciate the reply, thank you.

 

[pint]

Daft question time again from a lurker...
Is there any advantage to Semaphore arm signals against coloured lights, such as visibility? and can semaphore arm signals be controlled by a computer system as opposed to the old-fashioned lever signal cabin/ box?

 

[LowLevel]

Daft question time again from a lurker...
Is there any advantage to Semaphore arm signals against coloured lights, such as visibility? and can semaphore arm signals be controlled by a computer system as opposed to the old-fashioned lever signal cabin/ box?

1 - not really. They're simple things and don't often go wrong. But modern colour lights are very visible. Picking out the arm is still harder than a colour light even in daylight.

2 - yes, they can be electrical motor/hydraulically/pneumatically worked and controlled by switches, panels or pretty much anything else if need be.

 

[Meerkat]

A OCS or One Control Switch panel is a small panel where there is one switch for each signal on plain line, or where lines converge. If there is a signal that controls a diverging junction, then there is one switch per route from this signal. So if there are two signalled routes, there will be two switches. One for the main route (main line) and one for the diverging route (line). Indications are provided for the colour light signals, and for the track circuits.

The points also use one switch per set of points. And they also have indications.

These panels are much less complex compared to the bigger panels found in power signal boxes. The small ones can fit on a office desk.

MES - minimum electrical signalling, normally means as cheap as possible two aspect colour light signalling with track circuits and point machines. It is designed to be as inexpensive as possible with just the bare minimum facilities needed. And is normally controlled by a OCS panel or other low cost control system.

However, if a mechanical signal box locking frame still has years of service, sometimes the cheaper solution is to replace stuff piecemeal (e.g. replace a motorised semaphore signal with a LED two aspect colour light) and continue to use the locking frame for the interlocking rather than provide electrical interlocking.

The real question is, what is the likely cost / benefit analysis of keeping the existing system (with it’s operating and maintenance costs) compared to the capital costs of removing the old and providing a new system. Especially at a time when the railways in this country are being told by government to cut the running costs and not to overspend on capital works…

Thank you. So it still involves losing the semaphores, and makes that easier than moving it all to a big centre.

 

[DaveTM]

I suspect that semaphore will out-live coloured lights. Why? Because of maintainability.

If a mechanical interlocking breaks, it is simple enough that any mechanical engineer can fix it. And being ancient it is unencumbered by patents.

If an ETCS system is installed now, and the hardware is wearing out in 30 years time, the specification documents will still be available and open to all. And even if the silicon devices that originally implemented those specifications are no longer produced, some manufacturer somewhere will be producing hardware capable of running the software.

But what about the coloured light stuff in between? If a relay based interlocking is getting fragile now, we can't maintain it because there are no longer any manufacturers of relays. If a solid state interlocking is on its way out, those big transistors in cans are no longer produced. Unless some standards body decides to create some set of public domain interface specifications for coloured light signalling, as bits wear out it will become increasingly difficult to replace them on a like for like basis.

Mystic DaveTM's vision of the future is that Network Rail or Great British Railways or whoever (tm) will slowly replace coloured light signalling with ETCS in-cab signalling on a line by line basis across the network. In doing so they will have to come up with a way of dealing with an absolute block area bordering an ETCS area. Once that problem has been solved once, the same solution will be used in every location where the 1800s meet the 2000s. And thus we will be left with a werid steampunk ancient and modern intersection at places like Bognor or Littlehampton.

 

[DorkingMain]

Pretty much all of Cornwall is still semaphores (Liskeard onwards). Unsure if there are any plans to change that soon.

 

[Annetts key]

But what about the coloured light stuff in between? If a relay based interlocking is getting fragile now, we can't maintain it because there are no longer any manufacturers of relays.

The BR930 series of signalling relays are made to British Railways specifications, hence the BR930 bit. So any manufacturer that can make relays can make them. And despite solid state semiconductor devices displacing relays in a lot of applications, vast numbers of small and medium relays are still made and used.
Also, even new signalling schemes use BR930 relays. Most track circuits still use at least one relay per track circuit section. SSI uses relays for point detection if there is more than one detector needed. And SSI uses relays if the point machine(s) are anything other than a clamp lock (or similar). Relays are also used to interface between different interlocking systems. Even between different SSI systems. And relays are used to interface SSI systems to other signalling systems. For example the older generations of axle counter systems use BR930 relays to interface to SSI systems.

And then there are the thousands and thousands of conventional relay interlocking systems that still need to be maintained. And some types of relay have to be renewed every ten years (track relays, magnetic latching relays, some special function relays). While some ordinary line relays can operate reliably for fifty years plus.

So I think it will be a very long time before the last relay gets removed from the mainland U.K. rail network.

In terms of how long mechanical / semaphore signalling lasts, well, as hinted by others above, most resignalling plans have been put firmly in the bottom of the filing cabinet due to the mainland U.K. railway not having the money or the will to do anything at the moment. So if there is no business case, the existing signalling will likely continue until either it is deemed unmaintainable, it becomes unreliable, spares can’t be obtained or the cost benefit shows capital costs of a new resignalling scheme (small or large) will save money due to lower running costs and the money can be found for the scheme.

So, yes, it’s likely that most current mechanical / semaphore signalling systems will continue for many years. Especially as the powers that be may decide there is no point resignalling using conventional colour light (LED) signals if the existing system can last until ETCS/ERTMS is installed.

Exactly the same will also likely apply to existing relay interlocking systems and SSI systems.

 

[Ken H]

The BR930 series of signalling relays are made to British Railways specifications, hence the BR930 bit. So any manufacturer that can make relays can make them. And despite solid state semiconductor devices displacing relays in a lot of applications, vast numbers of small and medium relays are still made and used.
Also, even new signalling schemes use BR930 relays. Most track circuits still use at least one relay per track circuit section. SSI uses relays for point detection if there is more than one detector needed. And SSI uses relays if the point machine(s) are anything other than a clamp lock (or similar). Relays are also used to interface between different interlocking systems. Even between different SSI systems. And relays are used to interface SSI systems to other signalling systems. For example the older generations of axle counter systems use BR930 relays to interface to SSI systems.

And then there are the thousands and thousands of conventional relay interlocking systems that still need to be maintained. And some types of relay have to be renewed every ten years (track relays, magnetic latching relays, some special function relays). While some ordinary line relays can operate reliably for fifty years plus.

So I think it will be a very long time before the last relay gets removed from the mainland U.K. rail network.

In terms of how long mechanical / semaphore signalling lasts, well, as hinted by others above, most resignalling plans have been put firmly in the bottom of the filing cabinet due to the mainland U.K. railway not having the money or the will to do anything at the moment. So if there is no business case, the existing signalling will likely continue until either it is deemed unmaintainable, it becomes unreliable, spares can’t be obtained or the cost benefit shows capital costs of a new resignalling scheme (small or large) will save money due to lower running costs and the money can be found for the scheme.

So, yes, it’s likely that most current mechanical / semaphore signalling systems will continue for many years. Especially as the powers that be may decide there is no point resignalling using conventional colour light (LED) signals if the existing system can last until ETCS/ERTMS is installed.

Exactly the same will also likely apply to existing relay interlocking systems and SSI systems.

Surely the problem going forward is that as we get more and more electronic signalling, which has a finite life, then more and more of the capital budget will have to be spent on replacing it at end of life. So i think mechanical signalling, be it signals, points worked by rodding or mechanical interlocking, will be here to stay. at lest for my lifetime. Maybe the same with relay stuff, if someone still makes relays in the future.

 

[Tomnick]

I suspect that semaphore will out-live coloured lights. Why? Because of maintainability.

If a mechanical interlocking breaks, it is simple enough that any mechanical engineer can fix it. And being ancient it is unencumbered by patents.

If an ETCS system is installed now, and the hardware is wearing out in 30 years time, the specification documents will still be available and open to all. And even if the silicon devices that originally implemented those specifications are no longer produced, some manufacturer somewhere will be producing hardware capable of running the software.

But what about the coloured light stuff in between? If a relay based interlocking is getting fragile now, we can't maintain it because there are no longer any manufacturers of relays. If a solid state interlocking is on its way out, those big transistors in cans are no longer produced. Unless some standards body decides to create some set of public domain interface specifications for coloured light signalling, as bits wear out it will become increasingly difficult to replace them on a like for like basis.

Mystic DaveTM's vision of the future is that Network Rail or Great British Railways or whoever (tm) will slowly replace coloured light signalling with ETCS in-cab signalling on a line by line basis across the network. In doing so they will have to come up with a way of dealing with an absolute block area bordering an ETCS area. Once that problem has been solved once, the same solution will be used in every location where the 1800s meet the 2000s. And thus we will be left with a werid steampunk ancient and modern intersection at places like Bognor or Littlehampton.

The flaw in Mystic DaveTM's foretelling is that a mechanical interlocking rarely, if ever, exists in isolation. Even a simple 'textbook' block post will have a decent number of relays and other components (electric locks, acting on levers, for example) - controls and releases associated with the block instruments, track circuits, sequential locking, that sort of thing. A more complex layout with motor points, level crossings, etc., increases that significantly. The fringe between a mechanical box and a ROC (or whatever) will have quite a lot of relay-based locking!

Sutton Bridge Jn SB (Shrewsbury) currently fringes with the ERTMS installation at Machynlleth, so it's already been done. It's not really much different to a conventional AB-TCB fringe - the means of conveying movement authorities are very different, but the underlying principles in terms of train detection and interlocking are similar.

 

[InOban]

Apart from the previously mentioned Pass of Brander stone signals, which are also likely to remain in use for decades to come following the abject failure of the trial acoustic monitoring system a few years ago.

Mind you, the rock fall which derailed a 156 some years ago didn't trigger the wires because I think the rock had fallen from next the track.

 

[Ashley Hill]

Pretty much all of Cornwall is still semaphores (Liskeard onwards). Unsure if there are any plans to change that soon.

I think it was mentioned on another thread that Lostwithiel,Par and Truros days are numbered. Lostwithiel may be retained as a gate box. I will be delighted to be corrected as I don't want to see boxes closed.

 

[wallan]

Swinderby was never on the plan to close whatsoever, but its closure was announced about a year or so ago. However, that has just been put back again.

Till When ?

 

[Islineclear3_1]

Dorking seems to be quietly soldiering on, though does that have any semaphores or is it all colour light these days?

All colour light

There are still semaphores at Deal (Kent) and Littlehampton/Bognor Regis (West Sussex) but I reckon these will go sooner rather than later, and thus won't be the last on the national network

 

[matchmaker]

Apart from the previously mentioned Pass of Brander stone signals, which are also likely to remain in use for decades to come following the abject failure of the trial acoustic monitoring system a few years ago.

Considering my avatar, I should of course have mentioned them!

 

[rower40]

The North Staffs line from Derby to Crewe has to be a contender for the last semaphores ever.
It's always been a Cinderella line, as it links the two major centres of the LMS - Crewe on the LNWR and Derby on the Midland. As such, the line between them has been seen as "foreign" by both rival (but forced to talk to each other) management teams. So it'll get updated only when there's something that forces the issue - such as a catastrophic level crossing incident, heaven forbid.

 

[Tio Terry]

When I first started back in 1964 in Norwich, I was taken to Brundall by the then Divisional S&T Engineer, Hugh Murray, and given an introduction to signalling. I can remember saying something like "I suppose that signal (the junction home with two arms, one for Gt Yarmouth via Acle and one for the Lowestoft route) will be replaced with a colour light soon", he smiled and replied "There will still be semaphore signals when you retire".

Well, I retired over 53 years later and that particular signal was still in use. It's gone now of course but it did live for a full 50 years more than my youthful enthusiasm thought it would!

I don't doubt that semaphores could be kept working long past when I'm pushing up daisies, but can't think of a good reason why NR, or GBR, would want to keep them working!

 

[matchmaker]

1 - not really. They're simple things and don't often go wrong. But modern colour lights are very visible. Picking out the arm is still harder than a colour light even in daylight.

2 - yes, they can be electrical motor/hydraulically/pneumatically worked and controlled by switches, panels or pretty much anything else if need be.

When Stirling Middle was resignalled several years ago, all the semaphore and ground disc signals were replaced by LED signals and all the points were converted to power operation, but all are controlled by the original mechanical frame (apart from the Alloa line which has a separate Entrance/Exit panel).

 

[vlad]

The North Staffs line from Derby to Crewe has to be a contender for the last semaphores ever.
It's always been a Cinderella line, as it links the two major centres of the LMS - Crewe on the LNWR and Derby on the Midland. As such, the line between them has been seen as "foreign" by both rival (but forced to talk to each other) management teams. So it'll get updated only when there's something that forces the issue - such as a catastrophic level crossing incident, heaven forbid.

I wouldn't say it was catastrophic but something did happen on the A5132 crossing near Hilton back in 2019.

This crossing has manually operated gates - and a car drove into one whilst they were being closed. I understand the crossing keeper needed hospital attention - and the crossing was closed to traffic for a couple of weeks whilst the damaged gate was replaced. I think they've made a few safety improvements but it's otherwise as it was.