Token block token exchange question

[Gloster]

This would have been a slight challenge on those lines with automatic token exchangers, changed at speed, for the train would be well into the following section before the driver could retrieve the token, check it, and then pull up if wrong. On Taunton to Barnstaple, where the GWR installed automatic exchangers in the 1930s, the only problem that really occurred was the auto exchanged was fluffed due to a range of issues, loco catcher out of alignment, token jammed in the catcher, catcher broken off, etc.

If an auto exchange was to be done the distant signal was pulled off. I don't think there was any special locking of this - it was of course dependent on the starter also being off, which follows on from the discussion above.

I think that this was interpreted to mean that in these circumstances it would be accepted as long as the driver stopped immediately. It would be primarily for situations where a token had not been issued: if the driver just missed the token pick-up there wouldn’t be an oncoming train and if there wasn’t a token to be picked up then the driver should be stopping anyway. In almost all situations the token in an instrument can only be the one for the next section, as the driver had the one for the previous section. The only situation would be at a junction where a token for the wrong branch had been issued, but this involves a mass of cack-ups. One oddity was on the Class 120 DMUs used on the Aberdeen-Inverness where the tablet pick-up was on the guard’s door: here the guard removed the tablet, checked it and used the buzzer to indicate that all was in order.

The phrase in the GA was intended to underline just how important it was to check that you had the right token when leaving an exchange point. But, if I understand right, for no other offence was the possibility of dismissal so clearly stated.

 

[Tetragon213]

Just out of interest, what would happen if, say, the token was somehow "lost" and/or bent so badly out of shape that it couldn't go back in the machine?

 

[Gloster]

Just out of interest, what would happen if, say, the token was somehow "lost" and/or bent so badly out of shape that it couldn't go back in the machine?

Then you would have to put some sort of Pilot Working into force until the S&T have come along and put everything back in order. I think that this mainly involved in putting the instruments back in phase by removing another token. Losing a token was a serious matter and a lot of effort would expended in trying to find it: as the whole basis of token working is that only one token can be out at any one time, having a loose token is definitely not wanted.

 

[Taunton]

Just out of interest, what would happen if, say, the token was somehow "lost" and/or bent so badly out of shape that it couldn't go back in the machine?

As ever on the railway there was a procedure for it. One Barnstaple line token went flying in an automatic exchange, and was found in the lineside undergrowth quite a while later by the ganger. Another bounced back under the train and was run over by the wheels. Meanwhile one driver just lost it from the footplate one night, and was nicknamed 'Dropper' for the rest of his career!

The GWR automatic exchange installation was the "Whitaker Patent", same as on the S&D, and known at Taunton as the "Scissors", as that is what the loco catcher looked like. It was manufactured by an outside signalling supplier rather than by the GWR themselves. Here's an extended description of it:

https://www.lmssociety.org.uk/topics/singleLineWorking3.php

Scottish lines used the "Manson Patent", named after the onetime G&SWR loco superintendent who devised it, right up to the units on the Class 120 dmus described above,.

 

[bleeder4]

I'm not a signalling expert, so forgive me if this is a silly question. I did the West Highland line at the weekend which uses electronic tokens, rather than physical ones. It's all controlled from the box at Banavie. The driver still has to stop at the beginning of each section to get the token, but it's a digital one transmitted straight into the cab, rather than a physical one. It seems to be a very sensible upgrade to the physical token system and avoids all these potential issues like tokens getting bent out of shape and what not. I wonder why it's not more widespread?

 

[Annetts key]

Just out of interest, what would happen if, say, the token was somehow "lost" and/or bent so badly out of shape that it couldn't go back in the machine?

Pilot working at first if trains need to run. As others have said, losing a token is considered a very serious matter. Tokens are not small, so there is no excuse for misplacing one. Similarly staff should take good care of them.
See also the quote in my earlier post, #25..

I'm not a signalling expert, so forgive me if this is a silly question. I did the West Highland line at the weekend which uses electronic tokens, rather than physical ones. It's all controlled from the box at Banavie. The driver still has to stop at the beginning of each section to get the token, but it's a digital one transmitted straight into the cab, rather than a physical one. It seems to be a very sensible upgrade to the physical token system and avoids all these potential issues like tokens getting bent out of shape and what not. I wonder why it's not more widespread?

Money, or rather the lack of it.
Not helped with some of the electronic systems that B.R. introduced to replace traditional token systems being obsolete, so not available for new installations.

I don't know what the current plan is, but ETCS/ERTMS was the intended solution by Network Rail. But the introduction on the British network is taking a lot longer than originally intended.

Axle counters and computer based interlocking combined with LED signals could also replace some token systems/sections (track circuit block - TCB, but without needing lots of track circuits).

On branch lines that have no loops or junctions and which only go to a terminal station, One Train Working (OTW) systems are possible. These work like TCB but don't need track circuits along the whole line.

 

[Tetragon213]

Then you would have to put some sort of Pilot Working into force until the S&T have come along and put everything back in order. I think that this mainly involved in putting the instruments back in phase by removing another token. Losing a token was a serious matter and a lot of effort would expended in trying to find it: as the whole basis of token working is that only one token can be out at any one time, having a loose token is definitely not wanted.

As ever on the railway there was a procedure for it. One Barnstaple line token went flying in an automatic exchange, and was found in the lineside undergrowth quite a while later by the ganger. Another bounced back under the train and was run over by the wheels. Meanwhile one driver just lost it from the footplate one night, and was nicknamed 'Dropper' for the rest of his career!

The GWR automatic exchange installation was the "Whitaker Patent", same as on the S&D, and known at Taunton as the "Scissors", as that is what the loco catcher looked like. It was manufactured by an outside signalling supplier rather than by the GWR themselves. Here's an extended description of it:

https://www.lmssociety.org.uk/topics/singleLineWorking3.php

Scottish lines used the "Manson Patent", named after the onetime G&SWR loco superintendent who devised it, right up to the units on the Class 120 dmus described above,.

Pilot working at first if trains need to run. As others have said, losing a token is considered a very serious matter. Tokens are not small, so there is no excuse for misplacing one. Similarly staff should take good care of them.
See also the quote in my earlier post.

Thanks for explaining!

 

[MarkyT]

Pilot working at first if trains need to run. As others have said, losing a token is considered a very serious matter. Tokens are not small, so there is no excuse for misplacing one. Similarly staff should take good care of them.
See also the quote in my earlier post.

When in possession of train crew they're usually contained within a leather pouch or other carrier which adds to their bulk with a large hoop attached for easier exchange. If a pouch, it will include an opening through which the engraving on the token can be read without removing it so crew can ensure they have the correct one.

Money, or rather the lack of it.
Not helped with some of the electronic systems that B.R. introduced to replace traditional token systems being obsolete, so not available for new installations. I don't know what the current plan is, but ETCS/ERTMS was the intended solution by Network Rail. But the introduction on the British network is taking a lot longer than originally intended.

NR replaced all the radio elements of the original Scottish systems as the frequencies were being reallocated, and retrofitted TPWS using a new module that switches the correct transponders on and off (and their indicators) by monitoring the radio traffic. It's all fairly modern with upgraded SSI equipment on which RETB interlockings are based. ETCS seems comparatively expensive at the moment for such rural applications, as was found on Cambrian. We really need a cut down regional variant that uses intermittent radio transmission of MA, like RETB, and sends commands from control centre to trackside (e.g. point movement) via the train radio system. RETB has had so much bolted onto it over the years I doubt it's really any cheaper than modern rural TCB today.

Axle counters and computer based interlocking combined with LED signals could also replace some token systems/sections (track circuit block - TCB, but without needing lots of track circuits).

Often fairly cheap on rural lines where there aren't many signals and those that exist can be simple integrated lightweight ones not on major structures. Axle counters dramatically minimise the amount of equipment between signalled locations as they only have sensors at the extremities of the logical block. Individual track circuits have length limitations so to cover a 10 mile section of single line there might have been up to 10 separate TCs logically summed up in the block. Each TC would require equipment housings, power, cabling etc. In these applications, axle counters represented major saving even in the 1980s when some early schemes using the (then novel and expensive) tech were completed.

On branch lines that have no loops or junctions and which only go to a terminal station, One Train Working (OTW) systems are possible. These work like TCB but don't need track circuits along the whole line.

If classed as OTNS (one train no train staff), with modern tech, such stub branch lines can be fully track circuited by means of an axle counter sensor at the junction end only, with no equipment beyond except a distant signal or board approaching the junction protection signal and a fixed distant board approaching the terminus bufferstops. I expect buffer stop TPWS would also be required today, but that might be powered by a local supply and be just permanently powered without any interlocking control or monitoring. If the same line was classed as TCB instead, then a starting signal and addition train detection sections are required at the terminus platform for turning back.

 

[Taunton]

When in possession of train crew they're usually contained within a leather pouch or other carrier which adds to their bulk with a large hoop attached for easier exchange. If a pouch, it will include an opening through which the engraving on the token can be read without removing it so crew can ensure they have the correct one.

By the time dmus came along to branch services on the Western Region, it was typically just the token that was handed over, left on the driving desk between stations.

 

[edwin_m]

I'm not a signalling expert, so forgive me if this is a silly question. I did the West Highland line at the weekend which uses electronic tokens, rather than physical ones. It's all controlled from the box at Banavie. The driver still has to stop at the beginning of each section to get the token, but it's a digital one transmitted straight into the cab, rather than a physical one. It seems to be a very sensible upgrade to the physical token system and avoids all these potential issues like tokens getting bent out of shape and what not. I wonder why it's not more widespread?

Further to the reasons already quoted, the train has to stop for token exchange, which can take a couple of minutes. It's possible to include "long section" tokens to allow the train to run non-stop through the intermediate token exchange station, but this requires the full long section to be available at the time it is issued. So although it can theoretically be applied to double track routes, it's only really suitable for single tracks where all trains normally stop at each loop.

 

[zwk500]

We really need a cut down regional variant that uses intermittent radio transmission of MA, like RETB, and sends commands from control centre to trackside (e.g. point movement) via the train radio system.

I thought the whole point of ETCS Level 2 is that everything was transmitted back to the RBC over GSM-R radio? Out of interest, how are American radio-controlled points powered, or do they not have the same locking and proving requirements as us as it's mainly freight?

 

[MarkyT]

Further to the reasons already quoted, the train has to stop for token exchange, which can take a couple of minutes. It's possible to include "long section" tokens to allow the train to run non-stop through the intermediate token exchange station, but this requires the full long section to be available at the time it is issued. So although it can theoretically be applied to double track routes, it's only really suitable for single tracks where all trains normally stop at each loop.

Indeed. RETB was removed on the Ipswich - Lowestoft East Suffolk Line which has some long double track sections, or dynamic loops as they might be termed today. All Replaced by modern TCB using axle counters and LED signals.

 

[Ashley Hill]

By the time dmus came along to branch services on the Western Region, it was typically just the token that was handed over, left on the driving desk between stations.

With DMUs there was the possibility that a token in a hoop could swing round and smash the first passenger window behind the cab. That could be why hoops fell out of favour for DMUs.

 

[MarkyT]

I thought the whole point of ETCS Level 2 is that everything was transmitted back to the RBC over GSM-R radio? Out of interest, how are American radio-controlled points powered, or do they not have the same locking and proving requirements as us as it's mainly freight?

Bombardier developed a regional variant using the techniques I describe a long time ago, but little work has been reported since. Unfortunatey passenger operations on the pilot scheme line shut down straight after as the local operator couldn't justify the expense of retrofitting the tech to their trains! The MA was still carried by the radio but its continuous refreshment would not be required in the same way as L2. L2 also does not carry any infrastructure commands over the train radio, rather requiring a conventional lineside infrastructure apart from the signals. There's some new enthusiasm for a low cost variant in Italy and Poland. I don't know about US radio controlled points but I'll do some research into that later.

 

[edwin_m]

I thought the whole point of ETCS Level 2 is that everything was transmitted back to the RBC over GSM-R radio? Out of interest, how are American radio-controlled points powered, or do they not have the same locking and proving requirements as us as it's mainly freight?

That's true of the train but I don't think ETCS includes anything about how the interlocking controls and detects the remaining trackside equipment.

With DMUs there was the possibility that a token in a hoop could swing round and smash the first passenger window behind the cab. That could be why hoops fell out of favour for DMUs.

Hence why certain DMUs (Cravens 105s I think) had a panel in the prime target area just behind the cab window, some kind of plastic I think which was less prone to being damaged.

 

[zwk500]

That's true of the train but I don't think ETCS includes anything about how the interlocking controls and detects the remaining trackside equipment.

Fair point - ETCS presumably allows for conventional interlocking controls to remain if desired, but I thought at the very least it allowed for the Trackside control module (LEU?) to be on a GSM-R link to the Interlocking so the only cabling would be from the cabinet to the points motor. Probably needs a new thread.

 

[ac6000cw]

Out of interest, how are American radio-controlled points powered, or do they not have the same locking and proving requirements as us as it's mainly freight?

In more remote areas, where mains electricity might not be readily available, small remote interlockings (e.g. at the ends of passing loops in CTC-controlled areas) often have a small wind turbine plus solar panels, which trickle-charge batteries used to power the track circuits, signals (approach-lit, so normally dark) and move the switches occasionally. In earlier times it would have been just batteries, which were replaced with fresh/recharged ones on a regular basis by the S&T people.

Keep in mind that 30 trains a day (counting both directions) is a busy single line on a US freight railroad, so even at that train density you are not moving the switches very often. If it only carries a few trains a day it might not have any lineside signaling at all ('dark territory'), just movement authorities issued verbally over the radio by a 'dispatcher' with the switches operated by the train crew as needed.

 

[norbitonflyer]

In the three token systems I have worked on, the token HAS to be returned to the instrument [and cannot simply be passed from e.g an eastbound train to a westbound train]. The token is the authority from the signaller to the driver that they can enter the token section.

With installations that meet the current standards, the Interlocking for the section signal will not release the signal if the correct token sequence hasn't been completed.

Even if there is no section signal, the same procedure should be carried out.

The token instrument system knows if there is an overall even or odd number of tokens in the system and some systems have the ability to detect when a token is extracted.

In 1921 the token machines at Abermule were not interlocked with the signals, which was what made it possible for a train to leave in one direction whilst a train was already in the section coming the other way. Had anyone actually been into the room where the token machine was, they would have seen that the situation on the ground (a train leaving the station) did not match what the machine would have told them (a token had been released from the machine at the other end of the section, and was therefore on a train approaching the station).
But of course no-one had been into the machine room to exchange the tokens, and therefore the paradox went un-noticed.

(For those who are not familar with the details, a misunderstanding between the station staff led to the driver being given back the token for the section he had just left, and he failed to notice that it was not the correct token for the section he was about to enter, which was already occupied by an approaching train)

 

[satisnek]

Tangential question: when was it deemed OK to dispense with tokens (physical or virtual) and to make single line protection solely reliant on drivers obeying signals?

 

[Annetts key]

Tangential question: when was it deemed OK to dispense with tokens (physical or virtual) and to make single line protection solely reliant on drivers obeying signals?

As opposed to reliant on drivers obeying signals on multi-track bidirectional main lines or at converging junctions?

The entire system is reliant on drivers obeying signals.

 

[Harpo]

Bombardier developed a regional variant using the techniques I describe a long time ago

The one labelled ‘L3’ as it didn’t need fixed detection? I seem to recall it had no alternative means of doing ‘train complete’ but put it’s faith in train integrity, i.e. modern trains don’t lose bits on the way.

I might have felt the same until the incident west of Reading when a driver asked to stop & examine for a track circuit presumed to be SOWC or showing occupied when clear, came across a turbo left behind by its inadequately coupled front portion when they’d left the depot.

 

[matchmaker]

As ever on the railway there was a procedure for it. One Barnstaple line token went flying in an automatic exchange, and was found in the lineside undergrowth quite a while later by the ganger. Another bounced back under the train and was run over by the wheels. Meanwhile one driver just lost it from the footplate one night, and was nicknamed 'Dropper' for the rest of his career!

The GWR automatic exchange installation was the "Whitaker Patent", same as on the S&D, and known at Taunton as the "Scissors", as that is what the loco catcher looked like. It was manufactured by an outside signalling supplier rather than by the GWR themselves. Here's an extended description of it:

https://www.lmssociety.org.uk/topics/singleLineWorking3.php

Scottish lines used the "Manson Patent", named after the onetime G&SWR loco superintendent who devised it, right up to the units on the Class 120 dmus described above,.

Manson however refused to patent the apparatus, as he considered it a valuable aid to workers safety. I'm fortunate enough to have seen it in operation, at Elgin in the early 1970s.

 

[zwk500]

Tangential question: when was it deemed OK to dispense with tokens (physical or virtual) and to make single line protection solely reliant on drivers obeying signals?

When the line has the occupation positively proved in the interlocking by train detection such as track circuits or axle counters. If the layout is simple with no points in the single line section, this will only need detection at the junction as the train can be 'counted in' and 'counted out'. If the line has a slightly more complex layout or could be used by successive trains in the same direction, it will likely need Track circuits or axle counters along it's length.

Both methods of operation link to the concept that to set a route, the interlocking must be satisfied no train is in the section. The token machines perform the same function as the train detection, just at a different 'resolution', so to speak.

 

[Rescars]

Manson however refused to patent the apparatus, as he considered it a valuable aid to workers safety. I'm fortunate enough to have seen it in operation, at Elgin in the early 1970s.

Somewhat curiously, the GSWR wasn't totally committed to their man Manson's apparatus and used the Bryson catcher instead. AIUI, this had the advantage that the equipment attached to the loco was portable and could be fitted as required, rather than the permanent fixtures required by Manson.

 

[Annetts key]

The one labelled ‘L3’ as it didn’t need fixed detection? I seem to recall it had no alternative means of doing ‘train complete’ but put it’s faith in train integrity, i.e. modern trains don’t lose bits on the way.

I might have felt the same until the incident west of Reading when a driver asked to stop & examine for a track circuit presumed to be SOWC or showing occupied when clear, came across a turbo left behind by its inadequately coupled front portion when they’d left the depot.

As well as DMUs deciding to divorce, the couplings on freight trains still occasionally break. Yes, most of the time the driver of the freight train will quickly notice a problem. But as demonstrated in various other incidents, if the train is topped and tailed and the braking system is not correctly configured, issues can still occur.

 

[satisnek]

As opposed to reliant on drivers obeying signals on multi-track bidirectional main lines or at converging junctions?

The entire system is reliant on drivers obeying signals.

OK, so let's invert the question - if the entire system is reliant on drivers obeying signals then why did we ever need tokens on single lines?

When the line has the occupation positively proved in the interlocking by train detection such as track circuits or axle counters. If the layout is simple with no points in the single line section, this will only need detection at the junction as the train can be 'counted in' and 'counted out'. If the line has a slightly more complex layout or could be used by successive trains in the same direction, it will likely need Track circuits or axle counters along it's length.

Both methods of operation link to the concept that to set a route, the interlocking must be satisfied no train is in the section. The token machines perform the same function as the train detection, just at a different 'resolution', so to speak.

Thanks, this answers the question above. I've never thought of single line tokens and their machines as being a form of train detection, only as a form of insurance. But when you do consider them in this manner it makes sense

 

[MarkyT]

As well as DMUs deciding to divorce, the couplings on freight trains still occasionally break. Yes, most of the time the driver of the freight train will quickly notice a problem. But as demonstrated in various other incidents, if the train is topped and tailed and the braking system is not correctly configured, issues can still occur.

Agreed but as fleets are renewed, more trains will likely become capable of reporting their forward block occupancy and rear block clearance to the radio block centre and interlocking. The L3-derived virtual fixed block concept will have most benefit on densely trafficked lines with many short block sections. In those applications, it could save a lot of individual Train Detection sections out in the field. On more rural lines, there's little benefit in splitting long single lines up into many small discrete blocks so you don't save that many TD sections. Some can be provided where expedient clearly, and more cheaply than if requiring axle counter sensors. Axle counters are still provided for deadlocking junction areas (a principle established in the Swedish regional pilot. You really don't want to be able to move points under a train if it reports clear a little early. Those sensors can also be used for the long blocks between junctions to prove all movements in one direction are clear before reversing the block direction. Freight, including infrastructure engineering traffic is the remaining problem, as length is variable and there's currently no easy way for the front cab equipment to automatically and accurately determine the length of a consist nor is brake pipe depressurisation immediate enough a notification of a train split for block clearance purposes even if initial length was known for sure. My solution would be an End Of Train Device capable of reading track balises. It would work via radio data link with the forward cab ETCS to determine when the train has cleared a particular balise the front reader had previously passed over.

 

[Gloster]

OK, so let's invert the question - if the entire system is reliant on drivers obeying signals then why did we ever need tokens on single lines?

People can make mistakes: drivers, signalmen or other designated staff can make them. A driver could forget to check the signal, a signalman could forget where a train was, another member of staff could give a confusing message, etc. Remember, most staff were ill-educated, badly paid and overworked, so having a system that was simple to understand and easy to use made it much less likely that there would be a combination of mistakes leading to an accident: Abermule was the exception.

The driver is expected to obey the signals and the signalman is supposed to only clear the signal when the line is clear, but either can make a mistake, so a token brings in an element that both must apply to their actions. It is a belt and braces approach, but that was the basis of railway rules for many years: simple rules interacting so that, as far as possible, no single mistake by one person can cause an accident.

 

[MarkyT]

OK, so let's invert the question - if the entire system is reliant on drivers obeying signals then why did we ever need tokens on single lines?

Tokens are used in the UK clearly, and similar systems were employed across the British Empire and in other countries where Britain had great influence. Outside that sphere of influence, they were rare though not unknown. Bidirectional single tracks were considered to be a particular risk. In the very early days before telegraphic block and tokens, all passenger railways had to be built with two unidirectional tracks so head-ons could not occur. The introduction of one train staff and eventually electric token systems was what allowed railways to construct their vast rural networks more affordably.

Thanks, this answers the question above. I've never thought of single line tokens and their machines as being a form of train detection, only as a form of insurance. But when you do consider them in this manner it makes sense

They survived because acceptable alternatives proved too expensive for rural railways. Before axle counters, you needed a lot equipment to fully track circuit a long single line section. That needed power, cabling, equipment housings every mile or so, and a lot of maintenance. Axle counters eliminate most of that with sensors only at the logical block extremities and some form of data comms link from each to an evaluator unit.

 

[Taunton]

Tokens are used in the UK clearly, and similar systems were employed across the British Empire and in other countries where Britain had great influence. Outside that sphere of influence, they were rare though not unknown. Bidirectional single tracks were considered to be a particular risk. In the very early days before telegraphic block and tokens, all passenger railways had to be built with two unidirectional tracks so head-ons could not occur. The introduction of one train staff and eventually electric token systems was what allowed railways to construct their vast rural networks more affordably.

They survived because acceptable alternatives proved too expensive for rural railways. Before axle counters, you needed a lot equipment to fully track circuit a long single line section. That needed power, cabling, equipment housings every mile or so, and a lot of maintenance. Axle counters eliminate most of that with sensors only at the logical block extremities and some form of data comms link from each to an evaluator unit.

I've long been surprised that Centralised Traffic Control (CTC) never took off on UK single lines. Developed in the USA well over 100 years ago, it was extensively installed on Colonial rail systems in Africa. It just uses the same lineside wire pair as block signalling does, and where in Britain a signalbox controlling say 100 miles of line is a major structure, the equivalent unit for CTC is often just a console for one operator in a handy back room at a station.