Token block token exchange question

[Meerkat]

I assume that drivers are not allowed to swap tokens directly between themselves at a crossing point(as the signaller is in charge of giving permission to enter the section), but does the signaller have to return the token to the machine and then take it out again, or can they take it off the driver leaving the section and give it straight to the one about to enter it? Possibly seems pointless but otherwise the signaller at the other end wouldn't know the original train had arrived and another was on its way to them.
Triggered by watching a cab ride through St Bees* and the token exchange took a long time with the signaller going over the footbridge, meaning the level crossing was shut from before the first train arriving until after the exchange and the trains had left (one was a freight that was blocking the crossing during token exchange).

* Ben Elias Carnforth to Carlisle on the NMT - crystal clear on a beautiful late afternoon with the fells looking resplendent and all those semaphores. Well recommended for cab ride video enthusiasts.

 

[Gloster]

Although it is forty years since I worked signal boxes, I would say absolutely not. Firstly the whole integrity of Token working requires everybody to do a simple job thoroughly and to follow the rules, which is even more important on single lines than others. However, the Section Signal is almost always, probably now just always, released by the block, so if the token doesn’t go through the instrument, the signal can’t be cleared. Additionally, although there may no longer be any examples in use, with some single line instruments it was not possible to return a token to the instrument it came from: it had to go through the section. You also have the need to use the bell signals (and the needle indicator) when returning and removing tokens to inform the other signalman as to what is happening, though I suppose you could illegally get around this.

 

[Tomnick]

The train ready to enter the section has to be offered to, and accepted by, the box at the other end of the section (and "train out of section" first given for the one that's just left). There are reasons for which they might "refuse" the train, e.g. clearing point obstructed, or large/slow vehicle given permission to cross a level crossing in section. The token is restored to and withdrawn from the instrument as part of that process. The latter also, crucially, releases the section signal (for one pull).

 

[Elecman]

I assume that drivers are not allowed to swap tokens directly between themselves at a crossing point(as the signaller is in charge of giving permission to enter the section), but does the signaller have to return the token to the machine and then take it out again, or can they take it off the driver leaving the section and give it straight to the one about to enter it? Possibly seems pointless but otherwise the signaller at the other end wouldn't know the original train had arrived and another was on its way to them.
Triggered by watching a cab ride through St Bees* and the token exchange took a long time with the signaller going over the footbridge, meaning the level crossing was shut from before the first train arriving until after the exchange and the trains had left (one was a freight that was blocking the crossing during token exchange).

* Ben Elias Carnforth to Carlisle on the NMT - crystal clear on a beautiful late afternoon with the fells looking resplendent and all those semaphores. Well recommended for cab ride video enthusiasts.

No that shoddy method of working is fraught with peril and has led to fatal accidents in the past. They should always be returned to the token instrument and a new token obtained by cooperation with the other signaller

 

[74A]

The token has to be returned to the signal box put into the machine and another token taken out.

Once a train has departed the signal will be returned to danger. It is not possible to clear the signal again until a token is returned and another taken out.

If the tokens are simply swapped over then the signal into the next section will still be at danger and the train cannot proceed.

 

[Annetts key]

In the three token systems I have worked on, the token HAS to be returned to the instrument. 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.

The rules are different if pilot working is in place.

 

[matchmaker]

No that shoddy method of working is fraught with peril and has led to fatal accidents in the past. They should always be returned to the token instrument and a new token obtained by cooperation with the other signaller

Abermule being the classic example of what happens if the (very simple) procedure is not followed.

 

[Meerkat]

Thank you for your answers. I had forgotten that the signalmen still offer trains to each other (mixing it with NSTR in my head.....).
So is this right for trains crossing at B between two sections A-B and B-C......
Signaller has to get token A-B from the first train and return it to the instrument, before taking it out to give to the other train to go B-A (and collect token B-C from them), and then going back to clear the signal B-A*, put token B-C into its instrument, take it out again, give it to the train going B-C, then back to clear the signal for B-C?
No wonder the crossing at St Bees was closed for so long, with the signaller going over the footbridge twice!

* Or can the signaller clear the section before giving the driver the token for that section , with the slight risk the driver goes on the signal, without the token?

 

[Gloster]

Thank you for your answers. I had forgotten that the signalmen still offer trains to each other (mixing it with NSTR in my head.....).
So is this right for trains crossing at B between two sections A-B and B-C......
Signaller has to get token A-B from the first train and return it to the instrument, before taking it out to give to the other train to go B-A (and collect token B-C from them), and then going back to clear the signal B-A*, put token B-C into its instrument, take it out again, give it to the train going B-C, then back to clear the signal for B-C?
No wonder the crossing at St Bees was closed for so long, with the signaller going over the footbridge twice!

* Or can the signaller clear the section before giving the driver the token for that section , with the slight risk the driver goes on the signal, without the token?

Unless it has changed, then the signalman can normally clear the Section Signal to A as soon as he has taken out the token for B-A. He can reasonably presume that the driver will not leave without the B-A token and possibly with the C-B token. If the driver does leave without the correct token, his remaining career on the footplate is likely to be very short. A driver leaving entering a token section without the correct token was, I believe, the only offence where the General Appendix said ‘…a driver will render himself liable to dismissal...’ Nowhere else was the point hammered home with such force. Drivers are supposed to check the token is the correct one before the leave.

 

[King Lazy]

The token instruments are electrically locked with the signals so the token going through the instrument is necessary.

At the north end of the St Bees-Bransty section the token instrument is remote from the signalbox. When a driver collects the token on a southbound run they ask the signaller by telephone to release it. You can then hear the bells down the phone as the signallers at Bransty and St Bees exchange bell codes. The token can only be released if both signallers hold the plunger in on their instrument during token release. The signaller tells the driver when to take the token. The signal can then be cleared. Sometimes the signal won't clear and the driver will replace the token and start again all whilst on the phone under the signaller's instruction.

The Bransty machine was temperamental and would require a technique of turn 90 degrees - slight pause - turn 90 degrees to release the key or it would jam and another attempt would have to be made.

Drivers have been known to leave without the token. Usually due to a distraction between releasing it and returning to the cab.

It would be technically possible to simply hand the token over but the signal would have to be passed at danger and in today's world would require the TPWS train stop override to be used which would mean an OTDR record of bad practice which would soon be flagged and investigated. It is absolutely not allowed under the rules. (I'm not sure but don't believe there is any data recording of tokens passing through the instruments).

As an aside. I always thought St Bees made for an interesting debate regarding DRA. The signal can be cleared but in itself doesn't form your authority to leave until the signaller completes their long walk over the bridge to the cab.

 

[Meerkat]

Thanks all, very educational.
I do feel sorry for the signallers/drivers when they have to do the token exchanges in winter and/or hammering rain.

Another amateur question - what happens when the driver has to leave the cab for a token, can you take the key out of the train without powering down?

 

[Gloster]

Turning the token when removing it from the instrument was a knack. It was two quarter turns with the faintest pause at the half-way point: too short or too long a pause and the relays wouldn’t pick up, so you would have to start again.

I suspect that there is a machine in Bransty box, even if a token is rarely removed from it: possibly only for testing or the odd engineers train. Otherwise you would have to have a special apparatus in the box with a plunger and needle, etc.: much easier just to have the standard arrangement with a normal instrument in the box and an auxiliary one on the platform. At one place I worked all ‘coming to‘ tokens were returned to the instrument in the box and most ‘going from’ were taken out of the the platform auxiliary.

 

[Annetts key]

The Bransty machine was temperamental and would require a technique of turn 90 degrees - slight pause - turn 90 degrees to release the key or it would jam and another attempt would have to be made.

There’s a technical reason why the token has to be turned half way, then a pause, followed by the final turn. This issue also applies to some other token systems.

I forget the exact details of why. But IIRC, before a token can be extracted, the token system checks any other instruments in the system. This check only starts when the token is turned in the instrument where someone is trying to release it.

 

[King Lazy]

Thanks, I never knew that. I just remember being told there was a knack to it or it would jam.

To answer another question. You can remove the key from the cab desk without 'powering down'.

The key on most trains generally just locks/unlocks the direction selector so the engines would still run. But to remove the key requires the direction selector being turned to OFF which powers down the GSM-R. You would need to remember to save the registration on the GSM-R or you'd have to re-register which doesn't really take long.

Usually when exchanging a token at a station and needing to leave the cab to do so, the train would be placed in neutral with the DRA on and the cab door locked/secured when leaving. Usually there is an additional back up at stations where token exchange occurs in that the train doors being released locks the brakes on and if those doors are worked by a guard they wont be closing them until the signal is cleared. If I were leaving the train at a groundframe or to exchange a token I would take the key out and retain the GSM-R registration just in case some unauthorised person with knowledge and malicious intent were to enter the cab while I'm gone. Even then if they knew how to use a key and had one there wouldn't be much you could do!

 

[Annetts key]

Thanks, I never knew that. I just remember being told there was a knack to it or it would jam.

Until the token system has performed tests to see the status of each token machine in the system, it will not allow the electric lock to release in the instrument where you are trying to get a token out. So it's not really a jam.

If an attempt to get a token out doesn't include a slight pause, the lock will not pick and it won't allow you to get a token out.

We had so many reported failures (with the vast majority being no fault found when investigated) on one system (all instruments were remote from the signal box, so it was drivers or shunters obtaining the token) that my manager arranged for detailed step by step instructions to be attached to the wall in the token hut where most problems occurred.

 

[Tomnick]

If I'm not mistaken, the integrity of the system is maintained simply by a circuit over a pair of wires between instruments having its polarity reversed whenever a token is withdrawn or restored at any instrument, so the "check" is quite a simple one. I understand that the pause partway through the process of withdrawing is because the "release" isn't energised until that point, to guard against the possibility of two tokens being withdrawn at exactly the same moment. It feels similar to how you sometimes have to pause briefly to allow the electric lock to pick when pulling a lever – if you try to keep pulling, the solenoid won't be able to overcome the force that you're applying to the lock.

 

[Annetts key]

If I'm not mistaken, the integrity of the system is maintained simply by a circuit over a pair of wires between instruments having its polarity reversed whenever a token is withdrawn or restored at any instrument, so the "check" is quite a simple one. I understand that the pause partway through the process of withdrawing is because the "release" isn't energised until that point, to guard against the possibility of two tokens being withdrawn at exactly the same moment. It feels similar to how you sometimes have to pause briefly to allow the electric lock to pick when pulling a lever – if you try to keep pulling, the solenoid won't be able to overcome the force that you're applying to the lock.

It’s rather more complex than that.

One of the token systems I worked used the railways own cables and was entirely battery powered.

The other two systems I worked on used commercial “private wire” telephone “lines” (or modern equivalent with BT terminating equipment).

There was a complex arrangement of relays and for the systems that used commercial ‘phone lines, various items of FDM (reed) communications equipment.

In the normal resting state, there is no power or signal between the different parts / instruments. This saves power for the token systems that are powered by primary cells (non rechargeable) that make up a battery bank.

When you turn the token half a turn, only then does the system perform the test to see if the system has an even or odd number of tokens in each instrument. Each instrument mechanism has two “rest” positions: even and odd. A signal is sent out to the rest of the system. Then the system waits for a response. Only after the correct response (all instruments in the even position OR all instruments in the odd position) will a relay pick that allows the electric lock to pick (the release button(s) also has to be still held depressed as well).

Keep in mind that the design has to protect against an inadvertent cross of the lines or against attempted misuse.

 

[Elecman]

Abermule being the classic example of what happens if the (very simple) procedure is not followed.

Quite

 

[MarkyT]

If I'm not mistaken, the integrity of the system is maintained simply by a circuit over a pair of wires between instruments having its polarity reversed whenever a token is withdrawn or restored at any instrument, so the "check" is quite a simple one. I understand that the pause partway through the process of withdrawing is because the "release" isn't energised until that point, to guard against the possibility of two tokens being withdrawn at exactly the same moment. It feels similar to how you sometimes have to pause briefly to allow the electric lock to pick when pulling a lever – if you try to keep pulling, the solenoid won't be able to overcome the force that you're applying to the lock.

Having the check position, where the lock dog is brought up to the face in the first quarter turn of the commutator, means the electric polar relay and solenoid lock of the instruments are only then connected and don't repeatedly pick and drop needlessly while bell codes are being exchanged before the accepting box holds the bell plunger in for the release. This helps to protect mechanical and electrical components from wear and contact sparking, extending their life. Although universal today, early block systems didn't all provide a line clear/token withdrawal release for the section signal. For token systems, that feature ensures tokens can't be easily reused without passing through an instrument. If a token has been withdrawn for a movement that does not take place, the signaller should replace the token in their instrument. Modern instruments cancel the electrical release for the section signal lever (if not already taken) in that case to avoid risks of a forgotten stored release. If the section signal is cleared but then replaced before departure, the token must be returned to the instrument and a new token obtained to release the section signal again, the so-called 'one pull' feature. Modern controls also prove the section signal at the receiving box is at danger with lever locked normal at the time a token withdrawal takes place. Detailed controls have been revised and improved many times over the years to resolve weaknesses revealed by accidents.

 

[Rescars]

I suspect that there is a machine in Bransty box, even if a token is rarely removed from it: possibly only for testing or the odd engineers train. Otherwise you would have to have a special apparatus in the box with a plunger and needle, etc.: much easier just to have the standard arrangement with a normal instrument in the box and an auxiliary one on the platform. At one place I worked all ‘coming to‘ tokens were returned to the instrument in the box and most ‘going from’ were taken out of the the platform auxiliary.

Does this mean that you'd end up with an excess of tokens in the instrument in the box and a shortage in the platform auxiliary? Were there arrangements to restock the auxiliary instrument with tokens in bulk from time to time?

 

[Gloster]

Does this mean that you'd end up with an excess of tokens in the instrument in the box and a shortage in the platform auxiliary? Were there arrangements to restock the auxiliary instrument with tokens in bulk from time to time?

Yes. Periodically the S&T would turn up, unlock the instrument, remove an even number of tokens, carry them across to the other instrument and put them in it. The same would happen if you had an uneven service: as we had (theoretically) eight more Down trains a week than Up ones, they would remove most of the tokens at one box, drive to the other and replace them in the instrument there. There was a design of special carrier incorporated in some token machines to allow this, but our S&T had made a neat little carrier out of scrap wood and bonding wire.

 

[MarkyT]

Does this mean that you'd end up with an excess of tokens in the instrument in the box and a shortage in the platform auxiliary? Were there arrangements to restock the auxiliary instrument with tokens in bulk from time to time?

Yes. Tyer sold a transfer magazine system that affixed to the side of their instruments. Tokens to be transferred would be moved into the magazine of the source instrument, which could then be detached and reattached to the destination instrument. While detached, no tokens could be removed from the magazine or from either instrument via the special 'port' the magazine docked with on the side. so it was perfectly safe for operations personnel to use. They were installed where imbalance was expected routinely, such as auxiliary instruments at ends of goods loops. When rarely required elsewhere on non magazine equipped systems, after some weird traffic patterns for example where everything went through in one direction for a time, balancing had to be carried out by S&T technicians who would be able to unlock the instrument covers to manually pick the solenoid lock and take out tokens. For this method, it is always necessary to remove an even number of tokens from the source machine as that keeps the system in the same phase as found.

 

[Rescars]

Oh, clever, clever! There had to be a way, homemade or otherwise!

I wonder how tokens were transferred from one end of a section to the other before S&T had access to motor cars. Was it possible to remove a number of tokens, but still allow a train into section, so the excess tokens could be carried through to the other end. Did this require emergency pilotman working to be instigated perhaps?

 

[MarkyT]

Oh, clever, clever! There had to be a way, homemade or otherwise!

I wonder how tokens were transferred from one end of a section to the other before S&T had access to motor cars. Was it possible to remove a number of tokens, but still allow a train into section, so the excess tokens could be carried through to the other end. Did this require emergency pilotman working to be instigated perhaps?

As long as the technician takes out an even number of tokens and is satisfied the system is still working properly, they can lock the covers up and hand back the system into normal operational use then travel by train between the boxes concerned in custody of the temporarily withdrawn tokens. No pilotman required.

 

[Annetts key]

In some systems, e.g. where the single line covered a branch line that terminated, a token transfer was rarely needed. Indeed, one token system in the area I worked on was removed and replaced with a single train staff (which the driver removed from an instrument when the signaller pressed their release button).

But in other token systems, you could have separate entry and exit token instruments. Or if the train service pattern was uneven with more traffic in one direction than the other. So over time, the number of tokens in the entry instrument would diminish and the exit token instrument would fill up.

We had two token carriers/magazines. Each could hold about 20 tokens. However if you were careful, you could fit 21 or 22 tokens in. We then walked along the cess transporting them to the entry instrument.

On one of our systems, if the line was busy, this happened about every two to three weeks. If we S&T were busy with other work, so didn’t perform a transfer, eventually, a driver would inform the signaller, who would then report it as a fault.

Some years later, our manager negotiated with operations and the job of token transfers using the token carriers/magazines landed on the M.O.M.

When using the token carriers/magazines, because the tokens exited or entered the instrument via a separate side port, it did not matter if there was a train in section or not. Indeed, it was possible for a driver to release and extract a token in the normal way while a token carrier/magazine was attached to the instrument.

If a token carrier/magazine was not available or the token instrument was equipped for one, then strict rules applied:

5.2 Changing Electric Tokens
5.2.1 The technician must strictly observe the Block Regulations and other relevant instructions when dealing with electric token instruments of any kind.
5.2.2 Whenever it is necessary for the techinican to withdraw tokens or insert them in instruments, he must ensure that any section signals are not released. An entry stating the number of tokens dealt with must be made in the Train Register and in the technicians Token Register. These entries must be jointly signed by the Signalman and the technician concerned.
5.2.3 When a token is damaged, the technician must be in possession of the damaged token before putting the instruments into phase. When a token has been lost the technician must not put the instruments into phase without the written authority of the Area Manager or his representative. This person must be present when the instruments are put into phase after the damage to or loss of a token.

(This copy being from 1982).

As MarkyT says, after extracting the required number of tokens, the system could be used as normal. If a train was in section, it was not a good idea to start an extraction and in any case, the signaller would not grant permission if you asked. And you were required to ask.
Similarly the same applied at the other end.

Before S&T had road transport (the first vehicles S&T had were small car derived vans I believe), the options were to walk along the cess (back then cess paths with stone chippings were maintained), ride the bicycle that the company provided for you (again, along the cess), or travel on a suitable train (staff passes being issued for work purposes).

 

[Ediswan]

How do these traditional systems handle both ends requesting a token at exactly the same time ? (I'm sure they do, just wondering how.)

 

[MarkyT]

Here's an image and description of the magazine system from an old company booklet, appearing here: https://www.tyerandco.uk/ :


== Doublepost prevention - post automatically merged: 28 Mar 2025 ==

How do these traditional systems handle both ends requesting a token at exactly the same time ? (I'm sure they do, just wondering how.)

There's a bell code exchange beginning with a single call attention stroke so the direction of communication can be established. Normally the requesting box bells first to ask for line clear, then the accepting box repeats the code to indicate acceptance, holding the button in at the last beat to effect the release. Each bell plunger cuts off the feed momentarily for the bell at the same end via a normally closed contact, so if both signallers pressed at exactly the same time there would be no bell sounded at either end.

 

[Annetts key]

How do these traditional systems handle both ends requesting a token at exactly the same time ? (I'm sure they do, just wondering how.)

See also post #17.

There are contacts in each instrument to detect the position of the commutator. The electric lock that prevents the commutator from fully rotating 180 degrees to allow a token to be removed is electrically locked if any of the other instruments are not “in phase”. That is, their commutators in the normal position (no one attempting to remove a token).

The electric lock can only be energised in an instrument if the signaller is pressing their release button (if the instrument(s) are remote), the button on the instrument itself is being pushed, the token (and hence the commutator) has been turned 90 degrees and the system has done the tests I describe in post #17 to check the status of all the other instruments.

 

[Taunton]

A driver leaving entering a token section without the correct token was, I believe, the only offence where the General Appendix said ‘…a driver will render himself liable to dismissal...’ Nowhere else was the point hammered home with such force. Drivers are supposed to check the token is the correct one before the leave.

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 exchange was fluffed due to a range of issues, loco catcher out of alignment, token jammed in the catcher, catcher broken off in the exchange, 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.

The Barnstaple line did suffer from an uneven number of trains in summer, as assisting engines would be sent over double-heading, but come back on separate trains.

 

[MarkyT]

See also post #17.

There are contacts in each instrument to detect the position of the commutator. The electric lock that prevents the commutator from fully rotating 180 degrees to allow a token to be removed is electrically locked if any of the other instruments are not “in phase”. That is, their commutators in the normal position (no one attempting to remove a token).

The electric lock can only be energised in an instrument if the signaller is pressing their release button (if the instrument(s) are remote), the button on the instrument itself is being pushed, the token (and hence the commutator) has been turned 90 degrees and the system has done the tests I describe in post #17 to check the status of all the other instruments.

Further to that here's a schematic of part of the wiring of a GWR instrument. Current paths marked with the numbered bubbles are made only in the commutator positions shown.


Other contatcts in the machine are used in the block release circuit - here shown at a heritage site where there's a short and a long section machine. The 'NA' contact band in the circle on the top rung breaks when the section signal is moved from its normal position, destroying the release after one pull. The BNR relay drops out on reinsertion of a token after withdrawal to also cancel the signal release. The BSR and BCR don't need to be polar types but those were selected because they are faster acting than shelf types. There are two wired in series simply becasue they only have one contact each and two contacts are required for the circuit. The BSPR is the relay whose contacts appear in the electric lock controls for the section signal lever.