1. Why do we need them ? They're expensive ! On the road network, drivers are expected to drive such as to be able to stop within the distance they can see to be clear. They are aided by traffic lights at junctions and by various warning signs, but "within the distance they can see to be clear" is the general principle.
Why won't that approach work on railways ?
1. Because of the lower coefficient of friction between steel wheels and steel rails compared with that between rubber tyres and roads, trains take a longer distance to stop from a given speed.
2. We (or at least the customers) want the trains to go fast nevertheless.
We'll compare the stopping distance from 30 mph. This converts to 30/.621 = 48.3 km/hr, or 48.3 x 1000/3600 = 13.4 m s-1.
The Highway Code suggests the braking distance for a car (as opposed to the "thinking distance") is 45 feet (13.7 metres) from that speed.
For the train (suppose it weighs m kilogrammes) the contact force with the rails will be mg newtons, so, if the coefficient of friction is 0.25 (about the best we can expect) the maximum braking force is 0.25mg. Mr. Newton told us that force = mass times acceleration, so the acceleration (actually deceleration in this case) is 0.25mg/m = 0.25g = (-) 2.45 m s-2. We can now apply the (?) well-known relation 2as = v2 - u2.
-2 x 2.45 x s = 0 - 13.42 and s = 13.42/(2 x 2.45) = 36.6 metres .. a lot further than the car.
How far will the train take to stop from 100 mph (45.7 m s-1) ?
How far will it take if "leaves on the line" have reduced the coefficient of friction to 0.1 ?
So it does look as if we do need signals.
2. Some historical background The original idea was to divide the line into Sections (which is still done). Each section had an official called a Policeman (!) at each end whose job was to allow the train to enter a section only if it was deemed to be safe. The principle was that a train could be allowed to enter provided that the previous one had gone in at least ten minutes previously (often a bit longer if a faster train was following a slower one). This arrangement was known as the Time Interval system. It was much better than just allowing trains to proceed in a totally uncontrolled manner but it assumed that the train ahead had managed to keep moving. Unfortunately trains do break down (and did so more often then) and rear-end collisions were quite common. A subsidiary problem was that drivers were obliged to slow right down in case the policeman was going to signal them to stop. Initially the policemen used flags and lamps for signalling but fixed signals (often of varied design initially) soon began to appear.
Developments through the nineteenth century:
1. Arrangements for communication with the policeman (who progressively became a signalman) at the far end of the section so that it could be definitely established that the previous train had indeed left the section (the Absolute Block System). The arrangements came to include:
a) Electric telegraph and eventually telephone.
b) Block instrument showing Line Clear, Train on Line and Line Blocked.
c) The electric bell.
2. Signals, including the distant signal some distance before the "Stop" or "Section" signal which warned the driver in advance if the "Stop" signal was at danger.
3. Interlocking of points and signals to prevent signals from being set to Clear incorrectly.
Unfortunately these developments mainly occurred as a result of a large number of fatal and injurious accidents and, in many cases, pressure from the Railway Inspectorate (founded 1840). In those days the inspectors were serving Royal Engineers officers (which situation continued to the 1970s) and their contribution to railway safety in Britain cannot be overemphasised.
Some "accidents with lessons":
1. Brandon 1853 - a goods train driver ignored a policeman's red lamp signal and crashed into the back of a cattle train whose locomotive had broken down and was being repaired by its driver, watched by the two guards who should have walked down the line to warn approaching trains. An interesting conversation happened subsequently (see account of accident). Signals (and their observance) - and absolute block - clearly needed !
2. Menheniot 1874 - home signals but no starters. Two trains in the loop on the single-line route. Absolute block in use. The signaller shouted "Right away, Dick" to allow one train to start but the other one's guard was also called Dick and he signalled his driver to start. A head-on collision resulted.
3. Norwich Thorpe 1874 - single line with signals and absolute block but no other precautions. Because of confusion and lax observance of block telegraph regulations, the Mail train and another passenger train were allowed to enter the Norwich-Brundall single-line section simultaneously from opposite ends. Another head-on resulted.
4. Abermule 1924 - another single-line situation. The Tyer token system was in use but confusion at Abermule station allowed a train driver going towards Newtown to be given both the Montgomery-Abermule token and a green signal. The driver did not examine the token and duly set off. Precautions have to be applied to work !
5. Welwyn 1935 - East Coast main line - caused by a confused signaller. He allowed a second train to enter the section whilst the first one was still in it, causing a rear-end collision. It produced "Welwyn Control" which prevents signallers from accepting a following train unless the one in front has occupied and cleared the track circuit at the home signal.
6. Castlecary 1936 - another signaller's error (arguaby much worse) compounded by two drivers travelling faster than usual to recover time. Would you as a signaller allow a following train to enter your section if the previous one had just apparently travelled through your signals at danger and probably crashed into another train ahead ?
7. Winsford 1 1948 - a signaller thought a train had passed when it had not and consequently accepted a following train. A rear-end collision resulted.
8. Cowden 1985 - a driver passed a red signal controlling access to a "tokenless block" single-line section. The line and train had AWS (see later) but it is likely to have been inoperative in the leading cab of the train. A head-on collision resulted.
9. Bellgrove () - almost certainly a driver's error but could it have been prevented ?
10. Clapham Junction 1987 - a rare case of a "wrong-side" equipment failure caused by incorrect wiring. Also a possible "hole" in the rule book ?
3. Traditional signalling - Absolute Block system - One end of a block section (only L-R line arrangements shown)
Previous section to here
D = Distant signal, H = Home Signal, S = Starter (Section) signal.
The Distant can only show Caution or Clear (the train driver can pass it at Caution but must slow down ready to stop at the other signals if necessary)
The Home and Starter signals can show Danger (train must stop before the signal) or Clear.
A train will only be allowed to enter the section on the left if the previous one has cleared the overlap beyond the Home signal (often interpreted as the distance between the Home and the Starter). If the previous train has not cleared the overlap beyond the Starter, the Home signal cannot be cleared at all. If the previous train has cleared that overlap but has not cleared the one beyond the next Home signal, the Home signal is only cleared as the train approaches it at slow speed, indicating to the driver that the starter is likely to be at danger.
The signals can either be of the Semaphore type or be colour lights (resembling traffic lights). The latter are now normally used in new or replacement installations but some semaphores are still expected to be about for the foreseeable future.
Sometimes local circumstances make it necessary for two or more slow-moving trains to occupy the same section (or "Station Limits"). In that instance a special device called a Position-Light Signal is used, normally located on the same post as a stop signal. If semaphore, it has a smaller arm than the main signals. It only applies if the main signal is at danger and it gives the message that a train may pass it but must proceed cautiously as the section or station limits may not be clear. It is sometimes also used in situations such as major stations where the main signal allows access to a number of routes some of which are dead-end platforms, in which case the "Proceed" signal will be the position-light signal in conjunction with a Red (and probably an alphanumeric indicator showing which platform !). Other similar "ground" signals are used for controlling movements to and from sidings.
Where high levels of traffic have to be handled, one of the stop signals is often dispensed with and the distant signal for the next block is placed on the same post as the remaining stop signal.
4. A more modern approach Main line signalling is now normally done using four-aspect colour-light signals (red-yellow-green-yellow working upwards) or three-aspect (red-yellow-green).
Red = Stop
One Yellow ("Single Yellow") = be prepared to stop at the next signal
Both Yellows ("Double Yellow") = be prepared to stop at the next signal but one
Green = Clear (train may travel at up to the appropriate speed limit).
Flashing aspects are sometimes used to indicate that the train is to take a diverging route ahead.
The signals are usually spaced at even intervals; the distance depends on the route speed limit and the types of train using it but 1200 yards is typical for four-aspect on fast main lines.
The signals are often operated automatically by devices called Track Circuits (Axle Counters are sometimes used instead, e.g. on the upgraded West Coast line and in the Severn Tunnel where dampness would interfere with the track circuit operation.). A Track Circuit detects whether a train is in a particular section or not. Idea No. 1 might be like this:
If there is no train, the battery voltage will reach X, but a train will produce a short-circuit between the rails and cause the voltage to drop across the resistance and not appear at X. The arrangement is reasonably fail-safe as a break in the circuit or a flat battery will also cause the voltage at X to disappear. A drawback with such a simple circuit is that it can suffer from interference if the line is electrified, as the track is used as the return path on electrified lines. A number of ingenious circuits are used to overcome this and other problems; one common approach with dc electrification is to use an alternating source instead of a d.c. one and to use a.c. vane-type relays which do not respond to direct voltages and currents (see below). Another track circuit problem is in separating the sections in c.w.r. Special joints do exist but they represent a weakness in the rail. Another solution is to use a.c. track circuits as i) alternating track-circuit voltages become smaller with distance from the source because of the inductance of the steel rails and b) adjacent sections can use different frequencies. High-voltage pulsed track circuits are used where rusty or contaminated rails can prevent vehicle axles from shorting across the rails with the normal low voltage.
The components which control the signals in response to the track circuit currents and voltages are still relays in many installations, though solid-state systems (Solid-State Interlockings or SSI) are now permitted and used. A relay is an electromechanical device which switches the current in one circuit in response to the presence or absence of a current in another one.
Current switched in this circuit
M is a piece of magnetic material which is attracted by the magnetic field of the coil when the coil is energised. In this circuit, the switch is open when there is no coil current (a "normally-open" relay) but relays are available in which the switch is normally closed and opens when the operating current flows ("normally-closed"). So if the red lamp of a particular signal had to be illuminated whenever a train was either in the following section or in the overlap following it, and track circuits covered both overlaps and the rest of the section, we could operate the signal lamp like this (the relays are all assumed to be normally-closed types).
Track Circuit First overlap
Track Circuit Second overlap
Track Circuit Section
If any of the track circuit voltages are absent, the red lamp needs to be illuminated - and it will be by this arrangement.
Other relays which may be encountered are polarised relays (using permanent magnets, so currents of either polarity can switch them in different directions), slow-acting and slow-releasing relays, relays incorporating delays, and bistable relays (otherwise known as latching relays - they stay in the state an applied current last put them into until a current is applied in the opposite direction).
Practical rail-industry relays often have both normally-open and normally-closed contacts. From the way they are arranged on relays placed on a shelf in a relay room, they are known as "Front" contacts (n/o) and "Back" contacts (n/c).
More on relays in general on
Ingenious arrangements are used to minimise the risks presented by colour-light signals failing through a failure of a bulb.
The latest ideas are moving towards dispensing with lineside signals altogether and providing an in-cab display for the driver. Balises like the ATP ones (see later) are used to convey the necessary information to the cab system. This approach is usual on TGV routes in Europe and it is used on the new "Channel Tunnel Link".
5. Junctions The situation becomes more complicated when trains need to be allowed to proceed from one track to another. The signal engineer needs to provide arrangements for operating the points ("turnouts" or "switches") and for informing drivers of the route set for them at junctions - and interlocking arrangements such that signals cannot be cleared dangerously by mistake. Steps are also taken to ensure that the points are properly set for the intended route and that they cannot move as a train is passing over them. They are:
Detectors - which prove the points to be properly "home".
Facing point locks - which prevent unintended point movements.
So a signaller in a manual box would have to do all the following operations before being able to clear the signals for the train to proceed:
All relevant signals to danger/caution
Unset the relevant facing point locks (levers)
Set the relevant junction points
Reset the facing point locks
Clear the appropriate signals
(S)he would be prevented by mechanical interlocking from moving any points if the relevant signals were not at danger/caution and from clearing any signals if the relevant points and facing-point locks were not correctly set or another train had been signalled over a conflicting route. (When last seen, you could actually try this at weekends at the Peak Park Warden Centre in the former signal box at Hartington, which has now no railway or signals attached but it did still have the complete lever frame and interlocking !)
The driver is informed of the line her/his train is to take at a junction by one of a number of methods.
Semaphore signals - "splitting distant" (official title "Directing Distant") signals in which the signal for the diverging route is on a shorter post at the appropriate side of the post carrying the signal for the "straight ahead" route. A similar arrangement with the home signals normally follows. A third post and signal are sometimes used if there is a further possible route.
Colour-light signals - a direction head showing a diagonal and/or horizontal row of five white lights (known to drivers as "the feathers") above the colour-light signal. There may be more than one such indicator if there is more than one diverging route. In some cases flashing yellow signal aspects are used at a previous signal in addition but these are not universal. A modern version of "directing distants" and - the latest idea - illuminated arrows are also used.
Both signal types - an alphanumeric light-matrix indicator display (sometimes referred to as a "theatre-type indicator") is often used in situations where several routes are possible. It operates by means of a matrix of white lights (traditional type) or fibre-optic cables. It is normally placed below semaphore signals but can be above, below or at the side of colour-lights.
A three-aspect colour light signal with an alphanumeric route indicator - and a position-light signal underneath.
6. Operating the points and signals Traditional arrangement - signalboxes in which the points and (semaphore) signals were (are) operated by levers. The levers were quite long because they had to provide sufficient force to move the points and signals directly via rodding (points) or cables (signals). The signals were (are) weighted towards the "Danger/Caution" position so that they go there if the cable should break. Sometimes manual-frame boxes operate systems in which some points are power-operated (see below) and some signals are colour-lights; this is the arrangement on much of the Sheffield-Manchester "Hope Valley" route where the signals near the box are semaphores but those further away, which would be more difficult to operate reliably by cables if semaphore, are colour-lights.
Later arrangement - operated by switches on a mimic panel diagram on which the tracks, points and signals are represented. The signals are normally colour-lights and the points are power operated - generally electrically but pneumatics and even hydraulics are sometimes used. Often the route is set automatically (if it safely can be) by just operating switches at each end of the intended route on the panel. The interlocking is electromechanical (relay logic).
Latest arrangement - like the "mimic panel" one but the panel is implemented on the screen of a computer and the operations are performed by means of a mouse or (more usually) tracker-ball. The interlocking is now often solid-state, though it may be a relay system to which the computer is interfaced. Until the 1980s the final interlocking of points and signals was obliged by the Regulations to be either mechanical or electromechanical (relays) as solid-state devices could not be designed to fail safe reliably; they might either end up short-circuit or open-circuit, as opposed to relays which almost always fail stuck in their "normal" position. Solid-state interlockings were developed in which the vital safety elements are triplicated with a high-reliability "majority voter" network giving an output logic level equal to that at the majority of the inputs - and producing a fault indication if one input is different ! I have found it difficult to find complete information but this appears to be how an interlocking would be arranged:
The "twisted pair" is an inexpensive method of wiring the modules together with a minimum risk of electromagnetic intereference either from received r.f. signals or from stray magnetic fields (both may arise from passing electric or diesel-electric trains !). The central interlocking processor sends a message (known as a "Data Telegram") to each TFM in turn, giving it "instructions" on how to set its points or signals and asking it to report the states of its inputs. It sends a "data telegram" back reporting those states (all of this happens at least once per second for each TFM).
Single lines Single-line railways have an increased accident potential in that head-on collisions are possible as well as rear collisions. Head-on collisions are often the most destructive of railway accidents and it was realised early on in Britain that special arrangements were required to prevent them. The result was that only three head-on collisions on single-line railways (as opposed to junctions where trains travelled in opposite directions on the same tracks) happened in Britain in the whole of the twentieth century, two caused by gross breaches of operating rules and one apparently by a driver passing a signal at danger. The USA and much of continental Europe did not use such arrangements and they have consequently suffered a greater number of head-on impacts. The measures were and are (though 1 and 2 are now rare):
1. Having a token of some kind which the crew of any train requiring to enter any section must have before being allowed to enter it. This was very safe but it made it difficult to have two successive trains travelling in the same direction.
2. A "ticket" system in which a train could enter the section if either its crew had the token or its crew were shown the token and issued with a form (the ticket) authorising them to enter. If two trains were to follow each other through, the first one would be given a ticket and the second one would receive the token itself (unless the next train expected was also travelling in the same direction). This is a safe system but it still has problems if service disruption means that the next train needing to traverse the section is not going in the expected direction.
3. An arrangement of electrically-connected token-issuing machines which, though they contain a number of tokens, only allow one to be out at any one time. In that way trains can follow each other through the section but only one can be in it at once. This arrangement was invented by Edward Tyer in the 19th century and it is still in use on a number of routes (e.g. Harrogate-York). It has the limitation that trains now have to stop to exchange the tokens if the operation is done manually (previously they were allowed to be exchanged on the move, though it could not be done manually at high speed). Various species of token-exchanging machines were produced to allow faster running on single lines though I think they have now been replaced by other methods.
4. "Tokenless Block" in which the signalling alone is relied on to provide the protection, with interlocking to prevent the signals from being cleared when a train is already in the section. This approach is used also at junctions where trains travel in opposite directions on the same track on lines which are otherwise double track and on double-track lines which are bi-directionally signalled (not many in Britain though usual in continental Europe).
5. Radio Electronic Tokenless Block (RETB) which is a microprocessor-based system with instruments both in the train cabs and in the signalbox (which covers several sections). A driver wishing to enter a section speaks to the signaller by radio (part of the equipment) stating where (s)he is and which section is to be entered. Both signaller and driver have to press buttons (or "depress plungers" in signal-engineer-speak) simultaneously to ensure that the "electronic token" is only issued to the correct driver. A different radio frequency is used for each section. The cab instrument has a display which reads "You have the token from xxx to yyy" or is blank and it communicates with the signaller's instrument via radio.
Helping to prevent trains from over-running signals Three systems are in use (four if Underground "trainstops" are included)..
Automatic Warning System (AWS)
This system uses a permanent magnet and an electro-magnet along with detection equipment on the train with a cab indication.
An AWS magnet (believed to be a double one for a bidirectional line, i.e. electromagnet:permanent magnet:electromagnet; in Sheffield Station Platform 2)
The train travels over the permanent magnet first, which induces a voltage in a coil carried on the train. The train then passes over the electromagnet which is only energised, in the opposite sense to the permanent magnet, when the relevant signal is at green. The train-borne equipment will therefore produce something like one of the following voltage-time graphs depending on whether the electromagnet is energised or not (the broken-line parts only occur if the electromagnet is energised).
I have not been able to find the exact circuits used to deal with these voltages but the following would work if the magnets were strong enough and the coil on the train was of sufficient area with enough turns.
"Block X" incorporates a timer which times out after a time greater than the time between the positive pulses above (shown by the double-ended arrow). The T (toggle)-type bistable switches its logic output between 0 and 1 whenever a pulse at logic 1 is applied to its T input, so, if it started with its output at logic 0, its output will go to logic 1 when the train passes over the permanent magnet and back to logic 0 when the train passes over the electromagnet if the latter is energised. Block X will do the following when the timer times out:
Bistable output still 1 - sound a horn and, if the driver does not cancel by "depressing a plunger" to reset the system, apply the train brakes after a few seconds.
Bistable output returned to 0 - sound a bell.
The recent versions have replaced the horn and bell by similar-sounding electronic bleeps, but the idea is the same.
The arrangement is normally used at distant signals (which means every signal with modern signalling as they can all give a caution aspect). This is its main weakness - the driver will hear the horn at DY, SY or R and, if the line is busy and most signals are showing something other than green, there is a risk that the horn at a red will simply get cancelled (probably what happened at Ladbroke Grove).
An AWS permanent magnet is often provided at locations where a permanent or temporary speed restriction involves a considerable reduction in speed.
Train Protection and Warning System (TPWS).
This system is better in that it does stop trains if their drivers try to pass red signals, though possibly not in time to avoid passing the signal. It would, however, limit the over-run sufficiently to prevent most accidents caused by passing signals at danger.
The track equipment is grids carrying small high-frequency low-power radio transmitters (acording to the former "Railtrack" website, but they are referred to as "electronic loops" by Ref.2 below) which are arranged to transmit if the signal ahead is at Danger. When the train passes over a grid ('arming loop"), the radio signal received from the grid by the train starts a timer in the train-borne equipment. A subsequent grid ("control loop") also transmits a signal which stops the timer. The distance between the grids is arranged such that, if the train is travelling at a speed which would enable it to stop at the red signal, the timer will have timed out by the time the second grid is reached. If it has not, the train brakes are applied and the driver cannot release them for at least a minute after they are applied. A second pair of grids close enough together that the timer will not time out even if the train is going very slowly is provided beside the signal itself to apply the brakes in case of an actual overrun.
A TPWS grid (the double type immediately following the signal)
The system is also used at terminal stations to prevent trains from entering dead-end platforms too fast.
Automatic Train Protection (ATP)
This system ensures that trains cannot pass red signals (or exceed speed limits for very long) but it is very expensive. It depends on communicating a safe speed to the train by radio-based lineside "balises". If the driver tries to drive faster than that speed, the brakes are again applied in a manner which prevents their immediate release. This system is very effective but, because both of the expense and (I think) its incompatibility with emerging European systems, it has not been installed widely. It is in use on two routes - London Paddington to somewhere short of Bristol, and the Chiltern line from London Marylebone.
Level Crossings are also looked after by Signal Engineers. Traditionally they had gates which closed across the road and they were operated by crossing keepers located at the crossing. They were protected by signals on the railway which could not be cleared until the gates were in position across the road. The gates are often power-operated but some are hand-operated either directly or via a large wheel in the keeper's cabin (or signalbox). The later developments are:
Full-Barrier crossings - electrically-operated lifting barriers are used instead of gates. They may be operated locally or remotely using CCTV to monitor the road traffic, but a human operator is still involved. When last seen a crossing on the Sheffield-Doncaster line near Mexborough was operated from Sheffield "Powerbox" like this. Flashing road traffic signals are used to stop the road traffic before the barriers are lowered.
Half-barrier crossings - also lifting barriers but they only close half-way across the road. This is safer than full barriers in that road vehicles cannot be trapped on the crossing but more dangerous in that foolish drivers can zigzag round the barriers ! Half-barrier crossings are often automatically operated with a treadle detecting approaching trains and starting the road signals flashing. The barriers come down several seconds later and rise when the train has cleared the crossing, unless another train is coming.
"Open" crossings just have the flashing lights. They are either remotely monitored (a signaller or crossing keeper has CCTV or instruments which confirm that the lights are working) or locally monitored (the train driver sees a flashing white light on approaching the crossing to confirm that all is in order). They are only allowed to be used under suitable conditions of train speed and frequency and road traffic density and the line speed limit is 75 mph or lower. The locally-monitored species has a much lower speed limit.
Useful Sources of Information Modern Railway Signalling Handbook, Stanley Hall
1. http://www.hse.gov.uk/railways/liveissues/tps.htm (website - AWS, TPWS and ATP)
2. www.trainweb.org/railwaytechnical/sigtxt1.html (Website again - signalling principles in general)
3. http://www.davros.org/rail/signalling/bellcodes.html#CA (the bell codes and how they are used - good clear info on junction signal arrangements elsewhere on the same site).