• Our booking engine at tickets.railforums.co.uk (powered by TrainSplit) helps support the running of the forum with every ticket purchase! Find out more and ask any questions/give us feedback in this thread!

Rail crash driver died after telling signalman train had passed

Status
Not open for further replies.

Mintona

Established Member
Joined
8 Jan 2006
Messages
3,592
Location
South West
The crossing in question here is crossed less than a minute after exiting Sapperton short and long tunnels (with a small gap in between). I’m not sure how reliable GPS would be with a train in a tunnel for about 2 minutes and outside for around 30-40 seconds before passing the crossing.
 
Sponsor Post - registered members do not see these adverts; click here to register, or click here to log in
R

RailUK Forums

Pete_uk

Established Member
Joined
28 Jan 2017
Messages
1,250
Location
Stroud, Glos
I dont understand, why would the signaller rely on the word on an untrained member of public? That's an accident waiting to happen.

Yes, as one of the relatives said, it's strange in this day and age that the signalers need to be told of where the train is.

There are so many little crossings up the valley from Stroud to Sapperton without any warning tones or lights, although I know of one which has been done up with both and use it regularly.
 

Mintona

Established Member
Joined
8 Jan 2006
Messages
3,592
Location
South West
Yes, as one of the relatives said, it's strange in this day and age that the signalers need to be told of where the train is.

There are so many little crossings up the valley from Stroud to Sapperton without any warning tones or lights, although I know of one which has been done up with both and use it regularly.

There are a lot of foot crossings but I believe this is the only one which can be used by a vehicle (discounting the manned St Mary’s).
 

DerekC

Established Member
Joined
26 Oct 2015
Messages
2,107
Location
Hampshire (nearly a Hog)
RAIB doesn't seem to be investigating this one - I find that a bit surprising, although I know they don't cover everything. Anyone know why?
 

Llanigraham

Established Member
Joined
23 Mar 2013
Messages
6,074
Location
Powys
RAIB doesn't seem to be investigating this one - I find that a bit surprising, although I know they don't cover everything. Anyone know why?
I suspect that from their initial investigation they found that the signalled had done nothing wrong and this was just an unfortunate accident.

I do wonder whether members here realise just how many of these crossings there are in use. Machynlleth SCC controls 100! If you have a Garmin sat nav you can download a file that will show every level crossing of all types; the number is astounding!
 

sw1ller

Established Member
Joined
4 Jan 2013
Messages
1,567
GSMR has its own network with “100%” coverage, this is inherently more reliable than GPS and my radio cuts out all over the network! Sod GPS for safety critical movements. I use two sections of single line track that are 40mph and 60mph both directions and full of blind bends. If you told me that was now to be controlled by the use of GPS, I’d refuse to drive the train over it. Simple as that. GPS isn’t fail safe, so no, it can’t be used for safety critical movements.

Tell me how to make it 100% reliable AND fail safe, and I’ll shut up.
 

Mathew S

Established Member
Joined
7 Aug 2017
Messages
2,167
GSMR has its own network with “100%” coverage, this is inherently more reliable than GPS and my radio cuts out all over the network! Sod GPS for safety critical movements. I use two sections of single line track that are 40mph and 60mph both directions and full of blind bends. If you told me that was now to be controlled by the use of GPS, I’d refuse to drive the train over it. Simple as that. GPS isn’t fail safe, so no, it can’t be used for safety critical movements.

Tell me how to make it 100% reliable AND fail safe, and I’ll shut up.
No system can ever be 100% safe, and I'm sure you know that. Failsafing GPS, though, is straightforward. Multiple receivers to eliminate precision errors combined with track side beacons to correct any error that creeps in. Technology wise, it's a piece of cake; I could design a system to make it work. The expense, though, would be eye watering.
 

Peter Fox

Member
Joined
1 Mar 2018
Messages
60
The position of a train is only half the problem. The real issue for a crosser is the time available. Drive a car, negotiate the bends with my tractor and trailer, herd a flock of sheep across? That needs cooperation and experience.
 

alxndr

Established Member
Joined
3 Apr 2015
Messages
1,467
Regional differences strike again then. It was the case when the EBI Gate crossings were first introduced (in Scotland at least) no caution was imposed when they went into dark mode because the crossing was deemed to have ‘reverted’ back to a simple UWC and the expectation was that a member of the public coming across one in dark mode would understand this and contact the signaller as per the signage.

However when VAMOS arrived the instruction that came with them was, for whatever reason, that if one of those went into dark mode then a caution was to be imposed. Obviously this was inconsistent and confusing given they are just two different proprietary ways of solving the same problem so the instruction became - in Scotland at least - that a caution must be imposed for any crossing with overlay equipment in dark mode and new Signalbox Special Instructions were issued.

Dark mode monitoring information is currently sent by text message to either the Flight Engineer or the Signal Section Manager (there’s a bunfight ongoing at the moment about who gets it) and there can indeed be a delay of several hours before this information arrives after a failure. All in all its a rather strange situation we’ve found ourselves in with these installations.

Definitely still a regional thing as our darkmode information eventually goes to the Flight Engineer (my mind drew a blank on anything other than our nickname for them last night) and they run trains at linespeed on the understanding that the phones are operational. The MOM sometimes gets sent to check the phones if S&T (Signalling & (once upon a time) Telecoms technicians) are tied up elsewhere.

It surprises me that they can deem it necessary to run at caution over an EBIgate in darkmode, and yet information on the status of the crossing is known to be susceptible to delay and isn't tested during maintenance. The risk is that a regular user will surmise that if the lights don't work then trains will be going slowly and not bother to ring at a time when the darkmode message hasn't been sent and trains aren't being cautioned. It's similar to my gripe about the system not showing a green if one road is failed, but still showing a red if a train strikes in on the other road. A member of the public witnessing the change from black aspect to red might quite reasonably come to the conclusion that the green light is broken and incorrectly assume that no light=safe. They shouldn't do either of these things, but the public isn't known for always obeying the rules.

As you might be able to guess, I'm really not a fan of the way these systems have been implemented.
 
Last edited:

Llanigraham

Established Member
Joined
23 Mar 2013
Messages
6,074
Location
Powys
The position of a train is only half the problem. The real issue for a crosser is the time available. Drive a car, negotiate the bends with my tractor and trailer, herd a flock of sheep across? That needs cooperation and experience.

Exactly!! See my reply at #4.
These crossing are not dangerous and do not need expensive solutions for a problem that isn't there. There are thousands of these crossings used hundreds of times each day, and if used properly there is not a problem.
 

Llanigraham

Established Member
Joined
23 Mar 2013
Messages
6,074
Location
Powys
Definitely still a regional thing as our darkmode information eventually goes to the Flight Engineer (my mind drew a blank on anything other than our nickname for them last night) and they run trains at linespeed on the understanding that the phones are operational. The MOM sometimes gets sent to check the phones if S&T are tied up elsewhere.

It surprises me that they can deem it necessary to run at caution over an EBIgate in darkmode, and yet information on the status of the crossing is known to be susceptible to delay and isn't tested during maintenance. The risk is that a regular user will surmise that if the lights don't work then trains will be going slowly and not bother to ring at a time when the darkmode message hasn't been sent and trains aren't being cautioned. It's similar to my gripe about the system not showing a green if one road is failed, but still showing a red if a train strikes in on the other road. A member of the public witnessing the change from black aspect to red might quite reasonably come to the conclusion that the green light is broken and incorrectly assume that no light=safe. They shouldn't do either of these things, but the public isn't known for always obeying the rules.

As you might be able to guess, I'm really not a fan of the way these systems have been implemented.

Could you please explain who this "Flight Engineer" is, as it certainly isn't a term we used on the Marches?
 

dgl

Established Member
Joined
5 Oct 2014
Messages
2,391
Surely a partial solution is the use of a baliste after the crossing(s) and GSMR to tell the signallers that the crossing has been passed. This doesn't have to be after every crossing but on sensible intervals.
It could even be done with a baliste on the train to save having to put expensive comms equipment on the train, just a passive baliste with trackside equipment reporting back to base.
 

Highlandspring

Established Member
Joined
14 Oct 2017
Messages
2,778
As you might be able to guess, I'm really not a fan of the way these systems have been implemented.
Me neither, there seems to be a lot that hasn’t been thought through properly but they’ve been hastily adopted as a relatively cheap (but not particularly cheerful) magic bullet to reduce ALCRM scores and get the FWI down.
 

alxndr

Established Member
Joined
3 Apr 2015
Messages
1,467
Could you please explain who this "Flight Engineer" is, as it certainly isn't a term we used on the Marches?

Remote Condition Monitoring ring any bells perhaps? They've had various names and nicknames.

I'm not sure what their entire job entails, but they recieve "traces" graphing the current drawn from certain equipment, mostly points and track circuits. If there is a trend that suggests something isn't quite right they get the S&T to have a look. They're also the people who the darkmode/failure messages from EBIgate overlay crossings get sent to in (some!) areas, which is the pertinent part here.

I believe you were a signalman? If you ever had the S&T request an unplanned line blockage, with a fault number, to look at a set of points but they were working fine as far as you were aware then that would be where they were getting their information from. RCM/Flight Engineers would have noticed that the points were struggling to motor across and raised a fault for them to be looked at before they failed entirely.

EDIT: Just realised where your neck of the woods was, so probably not!
 

edwin_m

Veteran Member
Joined
21 Apr 2013
Messages
24,793
Location
Nottingham
Any information on whether the train was running late? I see on another thread there is a survey that claims a high proportion of professional (road) drivers use the timetable to work out if a train is due.

Asking the driver in question whether a train has passed is probably less hazardous than asking them at other crossings (by means of a sign) to stop look and listen and cross only if no train is approaching.

If you're going to the trouble and cost of providing hardware to detect the position of trains relative to these crossings, it's surely better as a principle to pass this information directly to road drivers via lights (or whatever), oddities and defects in the present equipment and processes notwithstanding. If done properly it eliminates a lot of workload and potential for errors with either safety or operational consequences, compared with still expecting the driver to phone the signaller who then interprets the indicators and gives permission or not. And if they can't give permission immediately then rather than repeating the same rigmarole a few minutes later they will probably just tell the driver to cross once the train has passed, so the hazard that started this thread still exists as long as there is a need for phone contact.
 

Spartacus

Established Member
Joined
25 Aug 2009
Messages
2,907
Any information on whether the train was running late? I see on another thread there is a survey that claims a high proportion of professional (road) drivers use the timetable to work out if a train is due.

Apparently it was a few minutes down, but no more than 5.
 

6Gman

Established Member
Joined
1 May 2012
Messages
8,383
No system can ever be 100% safe, and I'm sure you know that. Failsafing GPS, though, is straightforward. Multiple receivers to eliminate precision errors combined with track side beacons to correct any error that creeps in. Technology wise, it's a piece of cake; I could design a system to make it work. The expense, though, would be eye watering.

Which is the key. How many incidents of this nature have taken place in the last 20 years?

"Signalman? I'm waiting to cross at xxxx crossing. Is it safe?"
"I've got a train in section heading west. Once it's passed you it'll be safe."
"I've just seen it go."
"OK then - safe to cross."

But it hasn't, and it isn't.
 

Highlandspring

Established Member
Joined
14 Oct 2017
Messages
2,778
"Signalman? I'm waiting to cross at xxxx crossing. Is it safe?"
"I've got a train in section heading west. Once it's passed you it'll be safe."
"I've just seen it go."
"OK then - safe to cross."

I think you’d end up on a development plan for your safety critical comms with that conversation. When I was in the box I always used to work it like this -

“Signaller any signalbox.”
“Hello I want to cross at at XXX crossing.”
“What are you crossing with please?”
“A car.”
“And how long do you need to cross including the time to open and close the gates?”
“Two minutes should do.”
“Ok you want two minutes to cross with a car at XXX level crossing, I can confirm it’s safe to cross immediately.” *


Or...
“Sorry no there’s a train coming, please call me back once it has passed you.”

Then...

“Signaller any signalbox.”
“Hello that’s the train passed me.”
“Hello, can you just confirm you were looking for two minutes to cross with a car at XXX level crossing?”
“Yes that’s right.”
“And which way was the train going please?”
“Towards Anytown.”
“Ok it’s safe to cross immediately at XXX”


*For many years in Scotland signallers used to be instructed to “confine your answer to a statement as to how much time is available to cross” as the Signalbox Special Instructions always quaintly put it, apparently as a result of legal advice from the Lord Advocate’s office. So you’d say something like “you want two minutes to cross with a car at XXX crossing, I can confirm you have two minutes”. This changed to bring Scotland in line with the rest of Network Rail in about 2008ish but I don’t know whether a specific change in the law brought that about.
 
Last edited:

R G NOW.

Member
Joined
25 Jan 2019
Messages
418
Location
gloucester
Just a thought, but not having any knowledge of the location in question quite happy to have my theory adjusted. This is more of a warning system than a fail safe system

If working in an area where signalling is AB or Axle counter, could a track circuit block be installed around the crossing, linked to a local power supply, that whenever the train enters section with gates open sirens and beacons go off to warn the person that the train is near? I understand a lot of these locations are countryside, but for something that would get so little use, a solar/battery type set up would be possible, and being interlocked with the gates it would not go off unnecessarily? I guess the only real risk would be that it could make users of this type of crossing complacent and not get permission to cross.

This I believe is the only answer, Although I think the signalling along this line is absolute rubbish. It needs upgrading to three aspect, and especially now that a soon to be increase in trains. At these crossings or near to them, that there should be a placement red switch on the panel or a circle on a VDU (IECC) screen to place them to Danger. I think it would have been better if Network Rail had signalled all the way to standish and connected it to TVSC when the Kembe to Swindon line was doubled. I.E Signal G272 on TVSC and G217 Gloucesters first from the junction.

Today I noticed a crossing near Stroud has been upgraded with more signs and additional lights.

TVSC = THAMES VALLEY SIGNALLING CENTRE, DIDCOT.
 
Last edited:

6Gman

Established Member
Joined
1 May 2012
Messages
8,383
I don’t think you’d end up on a development plan for your safety critical comms with that conversation. When I was in the box I always used to work it like this -

“Signaller any signalbox.”
“Hello I want to cross at at XXX crossing.”
“What are you crossing with please?”
“A car.”
“And how long do you need to cross including the time to open and close the gates?”
“Two minutes should do.”
“Ok you want two minutes to cross with a car at XXX level crossing, I can confirm it’s safe to cross immediately.” *


Or...
“Sorry no there’s a train coming, please call me back once it has passed you.”

Then...

“Signaller any signalbox.”
“Hello that’s the train passed me.”
“Hello, can you just confirm you were looking for two minutes to cross with a car at XXX level crossing?”
“Yes that’s right.”
“And which way was the train going please?”
“Towards Anytown.”
“Ok it’s safe to cross immediately at XXX”


*For many years in Scotland signallers used to be instructed to “confine your answer to a statement as to how much time is available to cross” as the Signalbox Special Instructions always quaintly put it, apparently as a result of legal advice from the Lord Advocate’s office. So you’d say something like “you want two minutes to cross with a car at XXX crossing, I can confirm you have two minutes”. This changed to bring Scotland in line with the rest of Network Rail in about 2008ish but I don’t know whether a specific change in the law brought that about.

I was paraphrasing :D

Whatever the form of words used, if the user says the train has passed when it hasn't the outcome is unlikely to be good.
 

Highlandspring

Established Member
Joined
14 Oct 2017
Messages
2,778
Yes indeed, that’s why I always sought to clarify which direction the train was heading and, on occasion, how long ago they saw it. It wouldn’t be the first time someone has said “The train has just passed” when in actual fact they mean that they saw one 10 minutes previously...
 

Highlandspring

Established Member
Joined
14 Oct 2017
Messages
2,778
Mind you, you get to know the regular users and they get to know the drill..

Hello signaller I need 2 minutes to cross with a car at XXX crossing and the train going towards Anytown has just passed me” - perfect!
 

yorkie

Forum Staff
Staff Member
Administrator
Joined
6 Jun 2005
Messages
67,438
Location
Yorkshire
Just a gentle reminder of our forum rules requiring acronyms and codes to be defined the first time they are used. There are several posts in this thead with acronyms that may not be widely known, such as "VAMOS", ALCRM", "S&T" etc. Can members please edit their posts to include a definition where appropriate and bear this in mind for future posts that would be greatly appreciated. Many thanks in advance :)
 

Llanigraham

Established Member
Joined
23 Mar 2013
Messages
6,074
Location
Powys
I think you’d end up on a development plan for your safety critical comms with that conversation. When I was in the box I always used to work it like this -

“Signaller any signalbox.”
“Hello I want to cross at at XXX crossing.”
“What are you crossing with please?”
“A car.”
“And how long do you need to cross including the time to open and close the gates?”
“Two minutes should do.”
“Ok you want two minutes to cross with a car at XXX level crossing, I can confirm it’s safe to cross immediately.” *


Or...
“Sorry no there’s a train coming, please call me back once it has passed you.”

Then...

“Signaller any signalbox.”
“Hello that’s the train passed me.”
“Hello, can you just confirm you were looking for two minutes to cross with a car at XXX level crossing?”
“Yes that’s right.”
“And which way was the train going please?”
“Towards Anytown.”
“Ok it’s safe to cross immediately at XXX”


*For many years in Scotland signallers used to be instructed to “confine your answer to a statement as to how much time is available to cross” as the Signalbox Special Instructions always quaintly put it, apparently as a result of legal advice from the Lord Advocate’s office. So you’d say something like “you want two minutes to cross with a car at XXX crossing, I can confirm you have two minutes”. This changed to bring Scotland in line with the rest of Network Rail in about 2008ish but I don’t know whether a specific change in the law brought that about.

We used very similar questions and statements on the Marches.
 

DanDaDriver

Member
Joined
5 May 2018
Messages
338
I’m sure there is a good one, but is there a reason the crossing couldn’t be it’s own track section within a section so to speak.

Circuited or having maybe even it’s own axle counters so the bobby knows if something has gone over it and passed complete to the other side?

Edit: I don’t mean signalled as much as the panel/screen whatever, shows the train either approaching, on, or passed over the crossing within the signal section. If that makes sense
 

philthetube

Established Member
Joined
5 Jan 2016
Messages
3,749
But you'd still have the same user driven manual system, but with an additional GPS-led check. In the instance above, it would only misinform in so far as the person taking the phone call wouldn't be able to confirm the train had passed and would have to seek other secondary confirmation.

It's a potential layer of confirmation, and better than only having someone's word to go off.

The problem with that is that people would soon, rightly or wrongly, rely on this check and when GPS did not pick up a train, which will probably happen far more frequently than a member of the public mis reporting the location of one, then you are in trouble.

If it is not 100% reliable it confirms nothing and is dangerous.
 

alxndr

Established Member
Joined
3 Apr 2015
Messages
1,467
I’m sure there is a good one, but is there a reason the crossing couldn’t be it’s own track section within a section so to speak.

Quite simply, cost. There's no technical reason why it couldn't be done, but it wouldn't be cheap at all. Any redesign of the signalling is incredibly expensive, plus the cost of the equipment itself.

This is why they've started introducing the overlay systems mentioned previously as they work independently of the existing signalling, but these aren't perfect.
 

Mathew S

Established Member
Joined
7 Aug 2017
Messages
2,167
If it is not 100% reliable it confirms nothing and is dangerous.
But, as I said earlier, nothing can ever be 100% reliable. If we required everything to be 100% safe we wouldn't be able to do anything, let alone run a railway.
 

Taunton

Established Member
Joined
1 Aug 2013
Messages
10,019
I know I'm a bit old fashioned, but I find it surprising that the signaller doesn't have a very good idea of when a train would have likely passed the crossing or not. In the "bad old days" you did Train Entering Section at say 12.34, and the boxes at both ends would then know only too well that an express would be at that crossing at 12.38, a local at 12.40, etc. Someone phones you up at 12.36 and saying that the train has passed them would be spotted pronto.
 
Last edited:

DerekC

Established Member
Joined
26 Oct 2015
Messages
2,107
Location
Hampshire (nearly a Hog)
But, as I said earlier, nothing can ever be 100% reliable. If we required everything to be 100% safe we wouldn't be able to do anything, let alone run a railway.

True in principle, but railways have to meet safety standards that demand an extremely low level of unsafe failures. IEC SIL-4 level (which is usually the one quoted for signalling systems) demands a probability of dangerous failure lower than 1 in 100,000.000 per operating hour - putting it the other way round that's at least 99.999999% reliable. Even if you applied SIL-2 to this situation you would be looking for 99.9999% reliable.

In this situation "dangerous" would mean a user being told it was safe to cross when there was a train in the approach likely to arrive within the crossing time, including opening and closing gates etc. GPS (which I think was where the discussion started) just doesn't cut it on its own, even with augmentation. However there are some initiatives which look promising. One company I am aware of is combining GPS with image analysis from the forward facing camera and a map, so the train works out for itself (to a high level of integrity) where it is as it goes along, and could be asked (via data radio if only NR had implemented it with GSM-R) its exact location by the control centre. If it replies with a valid location, the time to arrival at the crossing could be calculated and a "safe to cross" (or not) message passed to the user. If the train doesn't reply you would be back to the current situation.
 
Status
Not open for further replies.

Top