Automatic vs. manual semaphore signals

[ccityplanner12]

On my travels around the country, I made some recordings of semaphore signals changing aspect. I noted that those at Worcestershire tended to change aspect immediately after the passage of a train, while those in Cornwall often took quite a while. I concluded that the Cornish signals were manual & the Worcestershire ones were automatic.

Automatic signals, as I understand it, are signals capable of detecting a train passing & changing aspect in response. Manual signals are signals which require operator input for all changes in aspect. Automatic signals became standard for new installations before coloured-light signals did, so I doubt there are any manual coloured-light signals. Automatic semaphores, on the other hand, were a thing at some point, but are they still a thing & how widespread are they? Are most surviving semaphores manual or automatic?

 

[Gloster]

No. There have been ‘automatic semaphores’, I actually worked semi-automatic ones forty-five years ago, but there are few, if any, left. (Automatic signals operate by the passage of the trains, both clearing and returning to Danger as required; semi-automatic (in this context) ones return to Danger automatically, but need to be cleared each time by the signalman.) Almost all semaphore signals have to be cleared by the signalman and then returned to Danger by him or her afterwards.

The differences between the time taken for the signal to return to danger depends on a number of factors: the need to make another movement, local regulations, how good a view the signalman has of the train’s position, whether the signalman is doing something else at that moment, local habits, the speed the signalman works at, etc.

There are many, many manual colour light signals, usually replacing semaphores, but also in new installations. Most Distant signals in manual areas are colour light.

As a starting point, I suggest you read the Wikipedia article on Railway Signalling.

 

[LowLevel]

The "automatic" semaphores are not automatic in the sense of a colour light signal, they're just motor worked and can return themselves to danger when the train has passed. The signaller stills instructs them to clear, generally by working a lever, and then returns the lever after the move is completed, even if the signal has already returned itself to the on position.

It's usually signals that are a decent distance from a box or have infrastructure that makes a wire run difficult.

Boston is an example where the semaphore signals are quite close to the box, but the ones around West Street and Broadfield Lane crossings are generally motor worked.

There are exceptions - Immingham Reception for example has a Great Central Railway electro pneumatic slide frame and the slides can pull themselves back into the frame which is quite interesting to watch.

 

[John Webb]

To the best of my knowledge there are no automatic semaphore signals in England. I suspect that the prompt replacing at Worcester is due to the view of trains from the nearby box and the need to keep other traffic moving.
In more remote areas, replacement may depend on the observation of the occupation and clearing of track circuits.

 

[edwin_m]

There are many, many manual colour light signals, usually replacing semaphores, but also in new installations.

To clarify: the term "Automatic signal" has the specific meaning that the signal both clears and returns to danger automatically according to the state of the track circuits in the section beyond it up to the next signal plus a distance known as the "overlap" which provides a safety margin. "Automatic" these days applies to colour lights although there were automatic semaphores in the past - the LSWR main line had long stretches of them between stations.

An automatic signal is only used where the section it protects has no points in it. If there are points they have to be set appropriately for each train, so the protecting signals are referred to as "controlled". They can be cleared by the signaller or by an automatic route setting system, or some can be set to run in "auto mode" to work like automatics if there are several following trains taking the same route. Both controlled and automatic colour light signals return to danger when there is a train occupying the block section they protect. Automatic signals may only return to danger a short time after the train passes, because the track circuit that controls them usually only starts at the end of the overlap. Controlled signals usually return to danger as soon as the front of the train passes them, but a few have "last wheel replacement" which delays this until the whole train passes.

Controlled and some automatic signals can also be returned to danger by the signaller, but this would only happen during out of course working or emergency. If the signaller returns a controlled signal to danger the interlocking prevents the immediate setting of a conflicting route in case there is a train approaching the signal too fast to stop.

And as mentioned by other posters, semaphore signals are normally both cleared and returned to danger by the signaller operating a lever. There are protections in the system to ensure they have been returned to danger before another train can approach them and potentially get an incorrect "proceed" signal.

 

[Belperpete]

Where a semaphore signal is worked electrically by a signal machine, it is common for the signal to be replaced automatically when a train is detected passing the signal.

Where a semaphore signal is mechanically wire-worked from the lever-frame, then it usually won't be replaced until the signalman replaces the lever. However, I have seen descriptions of semaphore signal replacers in old signalling text books. The replacer automatically replaced the semaphore to danger when a train operated a treadle or suchlike beyond the signal. It effectively disengaged the mechanical link between the balance weight and the arm, so the arm returned to danger. The link would be re engaged when the signaller replaced the lever. I am not aware of any still in use.

I suspect that the reason for the difference noted by the OP was probably as explained by Gloster. A busy Worcester signalman would be replacing signals directly after the passage of each train, whereas a signalman at a lightly trafficked Cornish box would be doing it "drectly".

 

[Ashley Hill]

We have two automatic semaphores on the SDR, but only when the controlling box is switched out and the levers reversed.
One is track circuit operated. When a train occupies the uni-directional track circuit approaching the signal as long as the detection circuits on the line ahead are made and all TCs ahead clear then signal will clear and return when the TC is clear again.
The other is treadle operated but works in a similar fashion.
@MarkyT could probably give a more technical description.

 

[Harpo]

Would the distance to the clearing point also be a factor?

 

[Gloster]

Would the distance to the clearing point also be a factor?

Not really. The Clearing Point is the point beyond the outermost Home signal to which the line must be clear before the signalman accepts a train from the previous box. The aspect of the signal does not affect the Clearing Point, although it should be at Danger: I think that there is nowadays a Home Normal Contact everywhere. I suppose that it would be possible to have a shortened distance from the Home signal to the Clearing Point and then put the signal back quickly so that as soon as the rear of the train passes the Clearing Point Train Out of Section and a new Line Clear could be given. However, this would be unusual as shortened distances to the Clearing Point were avoided where possible even in my long-past days.

 

[edwin_m]

Would the distance to the clearing point also be a factor?

Not really. The Clearing Point is the point beyond the outermost Home signal to which the line must be clear before the signalman accepts a train from the previous box. The aspect of the signal does not affect the Clearing Point, although it should be at Danger: I think that there is nowadays a Home Normal Contact everywhere. I suppose that it would be possible to have a shortened distance from the Home signal to the Clearing Point and then put the signal back quickly so that as soon as the rear of the train passes the Clearing Point Train Out of Section and a new Line Clear could be given. However, this would be unusual as shortened distances to the Clearing Point were avoided where possible even in my long-past days.

Not directly relevant, but IIRC the clearing point can be reduced to 200yds (vs 440yds normally) if the distant is a colour light.

 

[High Dyke]

There are also motor worked distant signals. These return to caution when the passing train activates the berth circuit of the home signal (there may be some variations on this).

 

[Pokelet]

On my travels around the country, I made some recordings of semaphore signals changing aspect. I noted that those at Worcestershire tended to change aspect immediately after the passage of a train, while those in Cornwall often took quite a while. I concluded that the Cornish signals were manual & the Worcestershire ones were automatic.

Automatic signals, as I understand it, are signals capable of detecting a train passing & changing aspect in response. Manual signals are signals which require operator input for all changes in aspect. Automatic signals became standard for new installations before coloured-light signals did, so I doubt there are any manual coloured-light signals. Automatic semaphores, on the other hand, were a thing at some point, but are they still a thing & how widespread are they? Are most surviving semaphores manual or automatic?

Which Worcestershire boxes in particular?

 

[Krokodil]

If the signals are interlocked with facing points they should be kept off until those points are clear, to prevent the bobby inadvertantly swinging them under a train.

 

[PM77]

Controlled and some automatic signals can also be returned to danger by the signaller, but this would only happen during out of course working or emergency. If the signaller returns a controlled signal to danger the interlocking prevents the immediate setting of a conflicting route in case there is a train approaching the signal too fast to stop.

An automatic signal that can be returned to danger by the signaller is generally known as a semi-automatic signal. Fully automatic signals usually don't have any indication on the panel/workstation if in a non CBI area and most of these don't have any indications on the interlocking at the signalling centre end.

 

[edwin_m]

An automatic signal that can be returned to danger by the signaller is generally known as a semi-automatic signal. Fully automatic signals usually don't have any indication on the panel/workstation if in a non CBI area and most of these don't have any indications on the interlocking at the signalling centre end.

I believe semi-automatics have dropped out of use in more recent schemes - with CBIs all autos have an emergency replacement.

 

[bleeder4]

To the best of my knowledge there are no automatic semaphore signals in England. I suspect that the prompt replacing at Worcester is due to the view of trains from the nearby box and the need to keep other traffic moving.
In more remote areas, replacement may depend on the observation of the occupation and clearing of track circuits.

Spot on. Shrub Hill is my local station and many times I have stood on the signal box end of the platform as a train is leaving, and have seen through the windows of the signal box the signaller watching the train pass and then moving the levers to reset the signal as soon as the train has cleared it.

 

[zwk500]

An automatic signal that can be returned to danger by the signaller is generally known as a semi-automatic signal. Fully automatic signals usually don't have any indication on the panel/workstation if in a non CBI area and most of these don't have any indications on the interlocking at the signalling centre end.

Apologies, what does CBI stand for (I can make a guess, but don't want to mislead!)

 

[Railsigns]

An automatic signal that can be returned to danger by the signaller is generally known as a semi-automatic signal.

A replacement facility in the signal box does not by itself turn an automatic signal into a semi-automatic signal. A semi-automatic signal was one controlled from a location that could open or close (usually a ground frame), and was plated as semi-automatic so that drivers knew to treat it as an automatic signal when that location was closed and as a controlled signal when it was open.

I believe semi-automatics have dropped out of use in more recent schemes - with CBIs all autos have an emergency replacement.

New schemes have neither automatic nor semi-automatic signals. Controlled (intermediate) signals are provided instead of automatic signals.

Apologies, what does CBI stand for (I can make a guess, but don't want to mislead!)

Computer-Based Interlocking

 

[Belperpete]

A replacement facility in the signal box does not by itself turn an automatic signal into a semi-automatic signal. A semi-automatic signal was one controlled from a location that could open or close (usually a ground frame), and was plated as semi-automatic so that drivers knew to treat it as an automatic signal when that location was closed and as a controlled signal when it was open.

Semi automatics had a similar plate to an automatic, but with the word SEMI underneath the bar. As you say, they were provided for colour-light signals protecting ground-frames in otherwise plain line areas. Also where a signal box could switch out, and the signal worked automatically while the box was switched out.

Some very early signalling schemes had long stretches of plain line where the signals all worked automatically, with the signaller having no control over them, or any indication. However, by the 1970s, BR required that certain signals should have a signaller replacement facility, eg those protecting tunnels, and that at least one in every five autos should have a replacement facility. These were still plated and called auto signals. Later on (it may have been following the Clapham disaster), it was changed to require ALL automatic signals on new schemes should have a signaller replacement. Again, they were still plated and called auto signals.

There were two types of replacement. Signals with what was termed an emergency replacement (E button) would have a red indication on the panel, that would light when the replacement had been operated, but didn't actually prove the signal at red and alight. Signals with an R button I think proved that the signal was actually at red. I seem to recall that the later type were used for auto signals protecting level crossings, so that the signaller could be sure the crossing was protected before allowing an abnormal load over the crossing.

With relay interlockings, automatic signals saved money in not having to provide control and indication circuits to and from the signal. Likewise they saved having to provide buttons and indications on the panel. However, with CBI, all signals are controlled by the central interlocking, so there is no saving. Likewise with VDU based control systems, there is no saving. Which is why automatic signals have fallen out of favour. It is better to just provide a standard controlled signal with an automatic-working A "button".

 

[Gloster]

Back when I started, Semi-Automatics were also used on lines with long sections of plain line. At least every fifth signal would have an Emergency Replacement Switch in the Panel Signal Box: it would normally operate as an Automatic, but could be put to Danger in an emergency.

 

[Belperpete]

Back when I started, Semi-Automatics were also used on lines with long sections of plain line. At least every fifth signal would have an Emergency Replacement Switch in the Panel Signal Box: it would normally operate as an Automatic, but could be put to Danger in an emergency.

As I pointed out in my previous post, that is NOT a semi-automatic, as officially defined in BR SSP11. Automatic signals with a replacement facility were still designated and plated as automatics. Some signals with entrance-exit route setting buttons were even designated as autos!

SSP11 defined semi-automatic signals as "A signal is designated as semi-automatic where it protects ground frame or ground switch panel operated equipment". At one time it had been practice to plate controlled signals that worked automatically when the signalbox switched out as semi-autos, but this practice was stopped (but many existing signals so plated remained plated that way for a long time).

 

[Gloster]

As I pointed out in my previous post, that is NOT a semi-automatic, as officially defined in BR SSP11. Automatic signals with a replacement facility were still designated and plated as automatics. Some signals with entrance-exit route setting buttons were even designated as autos!

SSP11 defined semi-automatic signals as "A signal is designated as semi-automatic where it protects ground frame or ground switch panel operated equipment". At one time it had been practice to plate controlled signals that worked automatically when the signalbox switched out as semi-autos, but this practice was stopped (but many existing signals so plated remained plated that way for a long time).

That may be the situation now, but I said at the beginning that that was when I started as a signalman: 1979. It was definitely referred to by all that I came across (management, signalman, S&T and others) as a Semi-Automatic; I am fairly sure that I saw it used on an official document about a signal that did not and never had protected a Ground Frame.

 

[Railsigns]

It was definitely referred to by all that I came across (management, signalman, S&T and others) as a Semi-Automatic

But was it plated as semi-automatic ("semi" above a black horizontal band)?

 

[Annetts key]

Semi automatics had a similar plate to an automatic, but with the word SEMI underneath the bar. As you say, they were provided for colour-light signals protecting ground-frames in otherwise plain line areas. Also where a signal box could switch out, and the signal worked automatically while the box was switched out.

Some very early signalling schemes had long stretches of plain line where the signals all worked automatically, with the signaller having no control over them, or any indication. However, by the 1970s, BR required that certain signals should have a signaller replacement facility, eg those protecting tunnels, and that at least one in every five autos should have a replacement facility. These were still plated and called auto signals. Later on (it may have been following the Clapham disaster), it was changed to require ALL automatic signals on new schemes should have a signaller replacement. Again, they were still plated and called auto signals.

There were two types of replacement. Signals with what was termed an emergency replacement (E button) would have a red indication on the panel, that would light when the replacement had been operated, but didn't actually prove the signal at red and alight. Signals with an R button I think proved that the signal was actually at red. I seem to recall that the later type were used for auto signals protecting level crossings, so that the signaller could be sure the crossing was protected before allowing an abnormal load over the crossing.

With relay interlockings, automatic signals saved money in not having to provide control and indication circuits to and from the signal. Likewise they saved having to provide buttons and indications on the panel. However, with CBI, all signals are controlled by the central interlocking, so there is no saving. Likewise with VDU based control systems, there is no saving. Which is why automatic signals have fallen out of favour. It is better to just provide a standard controlled signal with an automatic-working A "button".

If a signal has electrical controls, if it is a colour light, it's normal practice for the interlocking controls to return it to its most restrictive aspect where track circuits or axle counters are provided once a train has passed it. This may apply to some semaphore signals that have an electric motor.

When it changes aspect due to track circuit or axle counter controls depends on where these are. Some overlaps are over 300 yards. And the distance between the signal and the T/C or A/C for a distant signal may be even greater.

This does not apply to certain shunting signals. It does not apply to purely mechanical signals.

Main aspect colour light signals generally fall into one of these classifications:

• Colour light repeater/distant signal (including associated with a mechanical signal box)
• Controlled signal
• Automatic signal - are fitted with an automatic sign/plate
• Semi-automatic signal - are fitted with a semi-automatic sign/plate.

Notice that the word manual is not used.

A controlled signal will only clear if, either the signaller operates the control or an automatic route setting system is in operation. This control is a one shot control. Once a train passes the signal and it returns to red, the signal cannot reclear on its own. The signaller has to cancel the route/return the control to the normal position and then re -operate it to get the signal to clear again.

Some controlled signals may have an automatic switch/control. When this is operated, the signal will work automatically with no further input required from the signaller. This function may only be available for one route (e.g. the main line route). To an observer on the ground, it will look similar but not exactly the same as an automatic signal.

During a failure of the main remote control system, a back-up system, when used, may cause some controlled signals to operate in an automatic mode.

An automatic signal works completely automatically and the displayed aspect is only dependent on the Interlocking controls (signal ahead proved lit, state of track circuits/axle counters etc.). Some may have additional controls (may have a direction control if on a bi-directional line, may need a controlled signal routed up to it before it will clear, or the route set for the controlled signal that forms the exit from the bi-directional section and a track circuit or axle counter detecting a train approaching it).

If there are no signaller emergency controls, the signaller cannot put it to red/danger.

As described above, there are various emergency controls that were provided in the past.

• Signal post replacement switch - disadvantage being someone had to go to the signal to operate this. But providing it was cheap.
• Emergency replacement control operated by the signaller, non-proved type. Because no circuitry or indication was provided, the signaller had no way of knowing if the signal was lit or what aspect it was showing. Hence it could not be relied on.
• Emergency replacement control operated by the signaller, proved type. With this type, circuitry and an indication was provided. When the indication showed red, the signaller could rely on the signal being lit, being red and being commanded to be red (as opposed to just happening to be red).

Semi-automatic signal - a signal that operates like an automatic signal when the GF points that it protects are proved to be in the normal position and the GF is closed (not in use).

There are some oddities.

Signals that are not block post signals / don't form part of the aspect sequence. These either show green all the time (no matter where trains are) or only operate as a colour light distant (but may show double yellow as well as single yellow and green). They only show red when operated by the individual emergency switch/control available to the signaller, when a special group emergency switch/control is operated or when some other special emergency system is operated.

Former automatic signals but which have had their automatic sign (plate) removed. Done for operating reasons before the rule that allowed a driver to proceed under their own authority was changed. They still operate as automatic signals.

I only visited Worcester briefly, but I know they do have some semaphore signals that have electric motor control.

 

[Belperpete]

That may be the situation now, but I said at the beginning that that was when I started as a signalman: 1979. It was definitely referred to by all that I came across (management, signalman, S&T and others) as a Semi-Automatic; I am fairly sure that I saw it used on an official document about a signal that did not and never had protected a Ground Frame.

I was quoting an old British Rail standard signalling principle, not the situation now. It was how semi-automatics were officially defined when I joined BR S&T in 1975, and remained the situation until privatisation.

== Doublepost prevention - post automatically merged: 3 Jan 2025 ==

As described above, there are various emergency controls that were provided in the past.

• Signal post replacement switch - disadvantage being someone had to go to the signal to operate this. But providing it was cheap.
• Emergency replacement control operated by the signaller, non-proved type. Because no circuitry or indication was provided, the signaller had no way of knowing if the signal was lit or what aspect it was showing. Hence it could not be relied on.
• Emergency replacement control operated by the signaller, proved type. With this type, circuitry and an indication was provided. When the indication showed red, the signaller could rely on the signal being lit, being red and being commanded to be red (as opposed to just happening to be red).

On some regions, the non-proved type was termed emergency replacement and had an E button, while the proved type had an R for Replacement button. As you say, the proved type was used in situations where the signaller needed to positively know that the signal was at red, such as for auto signals protecting a level crossing.

There are some oddities.

Former automatic signals but which have had their automatic sign (plate) removed. Done for operating reasons before the rule that allowed a driver to proceed under their own authority was changed. They still operate as automatic signals.

The situation became very confused after privatisation. Not only did some auto signals have their auto plate removed, but some controlled signals were given auto plates! Finally the term "passable" was defined, to try and regularise the situation.

 

[Mr. SW]

It took this picture of the eastbound Distant west of Melton Mowbray on 05/05/2017 and I always thought this was a remarkable survivor. The signal appears to be motor-worked. Question: would this signal return to the most restrictive aspect when passed as touched on earlier?

 

[Belperpete]

View attachment 172139
It took this picture of the eastbound Distant west of Melton Mowbray on 05/05/2017 and I always thought this was a remarkable survivor. The signal appears to be motor-worked. Question: would this signal return to the most restrictive aspect when passed as touched on earlier?

Remarkable in what way? There is still quite a lot of semaphore signalling left, and semaphore distant signals.

It would probably be replaced automatically when a train occupies the berth track circuit of the home signal ahead - but that is unlikely to start immediately after the distant signal. I say probably, because there wasn't always a requirement to automatically replace distant signals. It isn't as simple as just proving the track circuit clear in the feed to the motor, as you need to ensure that the distant doesn't reclear once the train has passed clear of that track circuit, if the lever is still reverse.

 

[MarkyT]

We have two automatic semaphores on the SDR, but only when the controlling box is switched out and the levers reversed.
One is track circuit operated. When a train occupies the uni-directional track circuit approaching the signal as long as the detection circuits on the line ahead are made and all TCs ahead clear then signal will clear and return when the TC is clear again.
The other is treadle operated but works in a similar fashion.
@MarkyT could probably give a more technical description.

Down outer homes at Bishops Bridge (Staverton) and Buckfastleigh are electrically operated with motors. Neither is automatically replaced by passage of the train. They are the furthest worked signals from their respective boxes. Distants, even further out, are all fixed, but they have power for lighting. When Bishops Bridge is switched out with long section token working in force, its motor signal is electrically released to clear only when the next signal ahead is cleared by operation of the Staverton crossing ground frame slot, once the hand-operated gates are in place across the road.

 

[thesignalman]

View attachment 172139
It took this picture of the eastbound Distant west of Melton Mowbray on 05/05/2017 and I always thought this was a remarkable survivor. The signal appears to be motor-worked. Question: would this signal return to the most restrictive aspect when passed as touched on earlier?

That depends on what you mean by remarkable

• Semaphore distant arms are less common than they used to be, but are not particularly rare
• Motor operation is not unusual
• What is extremely unusual to me is the provision of a white sighting board, this is rare on yellow distant arms as they are usually distinct against a non-sky background anyway.

Best regards,

John

 

[Class 170101]

• What is extremely unusual to me is the provision of a white sighting board, this is rare on yellow distant arms as they are usually distinct against a non-sky background anyway.

Best regards,

John

I seem to remember seeing on that route have signals on the wrong side for the line they are signalling? ie signalling the westbound line but signal located adjacent to the eastbound line?