100 year anniversary of disaster at Abermule

[30907]

At Nairn, which was (I think) the last place with two boxes and instruments in the booking office, they were provided with a bicycle. Makes a change: cycling at work, rather than to work. I wonder if they were provided with company bicycle clips.

The sequence of operations of there was an electrical release on the starters must gave been complicated - or perhaps one wasn't provided?

 

[Dr Hoo]

The sequence of operations of there was an electrical release on the starters must gave been complicated - or perhaps one wasn't provided?

NOTE; I am not a signalling engineer!

In logical terms the sequence at somewhere like Nairn seems quite straightforward to me. The interlocking is blind to people taking tokens from drivers, walking back to the instruments, cycling along the platform to change points and so forth. I assume that the sequence was like (say):

[Edited version below]
Train en route from Forres, in possession of token, starting signal at Nairn obviously locked, instruments for section 'out of phase';
Train arrives at Nairn, surrenders token, replaced in instrument, now 'in phase';
Nairn can now seek permission to withdraw token for train in opposite direction (towards Forres);
Co-operative release, token removed and instrument placed 'out of phase', starting signal unlocked for 'one pull';
Token handed to driver of eastbound train, cycle to end of platform, re-set points, clear starting signal;
Train despatched.

I look forward to this broad outline being refined by an expert.

 

[30907]

NOTE; I am not a signalling engineer!

In logical terms the sequence at somewhere like Nairn seems quite straightforward to me. The interlocking is blind to people taking tokens from drivers, walking back to the instruments, cycling along the platform to change points and so forth. I assume that the sequence was like (say):

Train en route from Forres, in possession of token, starting signal at Nairn obviously locked, instruments for section 'out of phase';
Train arrives at Forres, surrenders token, replaced in instrument, now 'in phase';
Nairn can now seek permission to withdraw token for train in opposite direction;
Co-operative release, token removed and instrument placed 'out of phase', starting signal unlocked for 'one pull';
Token handed to driver of eastbound train, cycle to end of platform, re-set points, clear starting signal;
Train despatched.

I look forward to this broad outline being refined by an expert.

Nor am I!
For Forres in para 2 read Nairn.
Now add in the arrival at Nairn, and the departure from Nairn, from the Inverness direction... that's where I got lost.

 

[Dr Hoo]

Nor am I!
For Forres in para 2 read Nairn.
Now add in the arrival at Nairn, and the departure from Nairn, from the Inverness direction... that's where I got lost.

Sorry, original done in haste, post now edited. Hopefully more meaningful.

There has to be an even number of tokens in the two instruments taken together for them to be 'in phase' and able to release one token to allow a train to run. That is why, if a token is lost or damaged the S&T technician has to take another token out to put the instruments back 'in phase'.

 

[Gloster]

There has to be an even number of tokens in the two instruments taken together for them to be 'in phase' and able to release one token to allow a train to run. That is why, if a token is lost or damaged the S&T technician has to take another token out to put the instruments back 'in phase'.

And why, when moving tokens from one end of a section to the other on a line with more trains in one direction than the other, it is always an even number of tokens that is moved. The same applies when moving tokens between an instrument in a signal box and an auxiliary instrument, or between auxiliary instruments.

 

[GusB]

This is way off-topic as regards Abermule, but at Nairn the token instrument was contained within the station building - I think the actual "boxes" at the end of the platform were there solely to house the levers for controlling points etc.

The topic has been discussed here before, but it's far too late/early in the day.

 

[30907]

This is way off-topic as regards Abermule, but at Nairn the token instrument was contained within the station building - I think the actual "boxes" at the end of the platform were there solely to house the levers for controlling points etc.

Yes, that's how we got there from Abermule via Barrhill
Dr Hoo has described the process to and from Forres; exactly the same was needed to and from the Inverness direction - so I make it 2 cycle rides to and from each of the 2 boxes when trains crossed, there was no short cut!
But I see you are rather more local than either of us....

 

[Taunton]

Surprisingly not only the token machine but also the block bells were in the station instrument room, not in the signalbox.

I still don't get bits of this. The token is just a further safeguard. One still offers and accepts trains in the normal way. Yet it seems the train at Abermule was not offered to, or accepted by, Newtown, who would of course have picked up that Out of Section had not been received for the express coming the other way.

 

[LSWR Cavalier]

Don't quite understand why there has to be an even number of tokens
And what happens if another route is closed so that several diverted trains all going the same way 'exhaust' the tokens?

 

[ChiefPlanner]

Surprisingly not only the token machine but also the block bells were in the station instrument room, not in the signalbox.

I still don't get bits of this. The token is just a further safeguard. One still offers and accepts trains in the normal way. Yet it seems the train at Abermule was not offered to, or accepted by, Newtown, who would of course have picked up that Out of Section had not been received for the express coming the other way.

Briefly - the express was offered and accepted by Abermule , (not by the correct member of staff) , and the train entering section was not acknowledged by Abermule (the various staff were out on the platforms)

The stopping train went forward illegally into the occupied section as various people , assumed the express was not coming forward , and that "someone" had made suitable arrangements to offer the local train forward and draw the proper tablet.

It was only when after the local had departed , and it was planned to offer train entering section for the latter , that the awful realisation was made that there were 2 trains in section. A panicky phone call to Newtown indicated that the express had left some mins earlier.

Don't quite understand why there has to be an even number of tokens
And what happens if another route is closed so that several diverted trains all going the same way 'exhaust' the tokens?

The Signal and Telegraph (in those days) , would attend and arrange to redistribute the tablets / tokens. A fairly common task.

 

[Taunton]

The stopping train went forward illegally into the occupied section as various people , assumed the express was not coming forward ,

Yes I get all of that from the report. But the train was never offered, or acceptance received, which to me is as much of a gross omission as giving the wrong token. Yet someone then pulled off the starter. The token is a second safety device, over and above these normal procedures.

 

[ChiefPlanner]

Yes I get all of that from the report. But the train was never offered, or acceptance received, which to me is as much of a gross omission as giving the wrong token. Yet someone then pulled off the starter. The token is a second safety device, over and above these normal procedures.

They are all contributary factors for the disaster. The starting signal was (then) not interlocked with the "block" system , so this catastrophic event happened. No token could have been legitimately issued as we know - the "block" was locked with the withdrawn Newtown - Abermule tablet "out".

The rule book always made it clear that a driver leaving a station / block post with the incorrect token was liable to dismissal and he was ultimately responsible for the final , critical , safety check.

(as we know on that fateful morning , they got it horrible wrong - but haste was another important factor - a delay to the most important train of the day - the Express - was certainly on everyone's minds !) -hence the old adage of "stop and think" when something relatively out of the order is about to take place. The changed crossing patterns had happened in the past , so it was not totally unusual.

 

[edwin_m]

Surprisingly not only the token machine but also the block bells were in the station instrument room, not in the signalbox.

I still don't get bits of this. The token is just a further safeguard. One still offers and accepts trains in the normal way. Yet it seems the train at Abermule was not offered to, or accepted by, Newtown, who would of course have picked up that Out of Section had not been received for the express coming the other way.

I guess it would be a remnant from the original assumption that the stationmaster was in charge of train movements and would instruct "policemen" to work individual points and later signals from hand levers nearby. That would have evolved into having ground frames at each end of the station, but the next logical step of bringing everything under the control of the signalman hadn't yet happened (at Abermule at least).

In some ways it's rather surprising that signalling evolved that way, with boxes at one end of most stations and at both ends of many, at least in the days when the workable length of mechanical rodding limited how far points could be from the box. Putting the "signalbox" in the station building might have allowed it to work points at both ends, although limited visibility might have made it been more difficult to observe train movements and tail lamps. I think this was quite common for small stations on the Continent but rare in the UK.

 

[Gloster]

Don't quite understand why there has to be an even number of tokens
And what happens if another route is closed so that several diverted trains all going the same way 'exhaust' the tokens?

Put simply, if there is build up of tokens at B because of more trains running from A to B than B to A, then the S&T will go to B, open the instrument with a special key (other methods are/were available), take out a number of tokens from the instrument there and transfer them to the instrument at A. For technical reasons, which someone more gifted than me will have to explain, the number of tokens transferred must always be an even one. The explanation is something like: when you remove one token, it puts the system’s electrics out of phase and unable to release another token (as it should be) until that token is returned, which puts the electrics back in phase and able to release a token again. When transferring an even number of tokens, the first one removed from the instrument put it out of phase, the second one put it back in, etc.

 

[Taunton]

When transferring an even number of tokens, the first one removed from the instrument put it out of phase, the second one put it back in, etc.

It was quite able to handle an odd number of tokens. When the driver managed to let the token fall off the loco one dark evening on Taunton to Barnstaple (he was thereafter known as "Dropper" to all until the day he retired), although a pilotman procedure was applied for a few hours, when Exeter S&T turned up the lineman was able to reset the instruments with an odd number of tokens.

The ganger found it some days later in the drainage ditch. Another loss on that line was a token that fell out of the scissors-like GWR automatic exchanger on the loco and went under the wheels, being flattened. Presumably the same applied.

 

[Gloster]

It was quite able to handle an odd number of tokens. When the driver managed to let the token fall off the loco one dark evening on Taunton to Barnstaple (he was thereafter known as "Dropper" to all until the day he retired), although a pilotman procedure was applied for a few hours, when Exeter S&T turned up the lineman was able to reset the instruments with an odd number of tokens.

The ganger found it some days later in the drainage ditch. Another loss on that line was a token that fell out of the scissors-like GWR automatic exchanger on the loco and went under the wheels, being flattened. Presumably the same applied.

I think that when he reset it he effectively cancelled the withdrawal. It doesn’t matter how many tokens there are in circulation, just that one being taken out puts it out of phase and locks it, which prevents another being removed.

I was never S&T, but one box I worked at had a regular transfer of tokens due to more Down than Up trains.

 

[LSWR Cavalier]

What if there were no trains going the other way? Talerddig to Carno, say, is a short cycle ride, but some sections were much longer
Was there a work round, a way to avoiding having to transfer tokens? How many tokens were there for each section?

 

[Dr Hoo]

The token transfer process really isn't a big deal. Although precise methods may vary a common practice was for an S&T technician to do a 'tidy up' with a portable 'magazine' once per week. Instruments commonly had four 'slots' and could easily hold at least a dozen tablets or tokens at either end of the section.

 

[Gloster]

What if there were no trains going the other way? Talerddig to Carno, say, is a short cycle ride, but some sections were much longer
Was there a work round, a way to avoiding having to transfer tokens? How many tokens were there for each section?

Yes, in emergency someone might have to walk, cycle or ride between one signal box to get the token to the other end, but unless this was the S&T doing a transfer, it would be one token at a time. The whole essence of the token system is that, except in very limited circumstances involving the S&T, only one token for each section can be out at a time: Traffic Department staff did/should not have any opportunity to take a second token out. There was no work around other than getting the S&T in: any fiddle by other staff would be seen as a very serious disciplinary offence.

As to how many tokens were available for each section, this would have been a decision made when installing the system. It would be based on expected traffic levels and inequalities in the timetable. Remember, most lines with token working had a fairly limited and straightforward service, so generally the numbers in each magazine would only go up or down a little with the ebb and flow of traffic. Large numbers of diversions in one direction were unusual, but in emergency get the S&T out and, if need be, lend them an engine. I have a vague recollection that tokens supplied with the instruments from the makers generally came in multiples of twelve.

 

[Taunton]

How many tokens were there for each section?

The WR normally made 20 for each section unless there was a good reason for more. Installed 10 at each end.

 

[ChiefPlanner]

A good friend - relief SM "somewhere in the south of England" got called to a SB - where the honest signalmen admitted that somehow , 2 tokens had been issued independantly one each at both ends of the line. There was no dodgy practices involved.

So he told the other end to lock it in a drawer , set up working by pilotman (i.e himself) - locked the token in a drawer and retained the key and escalated the issue to the S&T. They sorted it out. I must ask him when I see him next - probably in a years time with COVID.

 

[Dr Hoo]

'Mass migration' of tokens wasn't uncommon. I can immediately think of three cases:

On the Grain Branch in Kent. Back in the days of the oil refinery 'double' sets of empty tanks were often sent down from Hoo Junction with two locomotives that then came back as two separate trains. So there was a tendency for tokens to run short at Grain and accumulate at the train crew worked token hut near Cliffe.

In the North West a single line had a section of double track at one end onto which trains would be 'dropped' waiting for a move in the opposite direction to clear. The signaller obviously had to wait to retrieve the token from the incoming working. The train crew of the 'dropped' train would then be able to obtain a token from an auxiliary instrument, the act of extracting it turning the colour light signal green presuming that the points had been re-set. No tokens were ever replaced in the auxiliary instrument in normal working. So it was a regular 'tidy up' by the S&T.

In South West Scotland when coal traffic from Knockshinnoch and Greenburn was busy it was often worked on a 'one way' basis via Mauchline-Annbank-Newton on Ayr-Barassie-Kilmarnock-Mauchline. So the tokens were constantly depleted from Mauchline.

Local S&T staff just dealt with things as a normal part of their work. I don't immediately recall ever 'running out' of tokens.

 

[Gloster]

The accident at Abermule was described by the Inspecting Officer as ‘unique’, seemingly because it was a combination of so much slack working, casual behaviour and lack of checking. There will have been many instances of members of staff acting in a slipshod manner, but the checks and balances in railway rules have meant that it is almost impossible for the slapdash behaviour of one person, or even two, to cause an accident. At Abermule all four of the station staff and the driver of the local acted carelessly.

The accident can be regarded as having remained unique for nearly a hundred years until an incident on the Romney, Hythe & Dymchurch Railway in 2019. The RAIB accident summary on this goes so far as to mention Abermule.

 

[Merle Haggard]

I think the previous accident was almost exactly a year before the Abermule tragedy. It involved two good trains colliding at slow speed, on a snowy night, near Oswestry, in the Oswestry - Ellesmere section. The Cambrian management were baffled at how it could have happened, and invited the Chief S&T Engineer of the Midland Railway to investigate. The cause, as I recall, was snow on the pole route wires bringing some of them into contact, allowing a false feed to one of the tablet instruments permitting a tablet to be withdrawn when that should have been impossible as one was out for the opposite direction. Matters were not helped by one of the signalmen failing to enter one of the trains, or the correct times (I forget which), in his register. I think one of the trains was an additional working that had been arranged at short notice.

John Prytherch.

I've read the accident report, available on Railways Archive, and there did seem to be a suggestion that there were suspicions that couldn't be proved and the wires explanation was a neutral one to close down the enquiry. .
The two signalmen each gave an account of events flatly contradicting that of the other. Furthermore, one signalman, having given a token to the driver of one of the trains, then left his box, against the rules, to go to the collision site, and on arrival (he was the first person to arrive from elsewhere at the scene) he asked that driver whether he had a token. Seems an odd first question; the driver is given in the report as receiving serious injuries.
It was a possibility, in my opinion, that there was a suspicion that the signalmen did not use the token instrument but kept the tokens out, possibly because of the unreliability of the electrical system between boxes - which would explain why the two signalmen could not give a uniform account of their use of the tablet machines
It's an interesting read; maybe I'm reading too much into it, but old accident reports sometimes seem to be constructed in a way that hints at a (possibly controversial) explanation by subtle language. in this case, perhaps the Ministry had suspicions about lax operations on the Cambrian, but until Abermule there was no concrete evidence.