Report of collision Bedford (Elstow) - 19/06/26

[Horizon22]

It's interesting the signaller received the first report of an accident from a passenger using a signal post telephone. Wonder if the passenger was (ex?) Train crew or operational staff.

They said there were 8 staff on board too, so maybe it was off-duty staff (although I would have expected them to be named as such).

Getting the horn instead of the bell at green signals happens more than you think, but as it's a right side failure it's not usually a big deal

Yes it is one of the various exception codes of which I think there are about 8 but I couldn’t number them all.

== Doublepost prevention - post automatically merged: Thursday at 18:43 ==

It's also notable that the information in this report is all attributable to what's been electronically logged. That gives us a chronology, but very little in the way of explanation

Which I think is fair as an interim report. I imagine 6-9 months before the full thing. RAIB caseload is also presumably quite high at this point with 2 derailments - not to mention smaller incidents - to investigate as well.

 

[Rail Ranger]

Do we know if West Hampstead signalling centre would have been aware of the SPAD at WH154 in terms of an alarm or would the signaller have seen the train entering the next track circuit? Presumably not the latter if the next track circuit was already occupied?

 

[tram21]

Do we know if West Hampstead signalling centre would have been aware of the SPAD at WH154 in terms of an alarm or would the signaller have seen the train entering the next track circuit? Presumably not the latter if the next track circuit was already occupied?

The report stated there was no alert system fitted and the signalling centre weren't aware.

 

[Horizon22]

Do we know if West Hampstead signalling centre would have been aware of the SPAD at WH154 in terms of an alarm or would the signaller have seen the train entering the next track circuit? Presumably not the latter if the next track circuit was already occupied?

Point 20 of the report:

There is no signal passed at danger (SPAD) alarm associated with signal WH154. SPAD alarms produce a unique tone in the signal box to draw the signaller’s attention to a SPAD, should one occur. In some circumstances, SPAD alarms can allow a signaller to intervene, for example, by making a GSM-R (Global System for Mobile Communications - Railway) railway emergency call (REC) instructing all trains in the area to stop.)

The only indication of an alert is a loss of power, which presumably came through the ECRO as well.

 

[driver9000]

Thanks for taking the trouble to set this up for us. What are the two repeaters at 1.01.06 and why is one on and one off?

And in the approach to various signals you can hear pips before the AWS bell, are they TPWS or something else?

It's a Splitting Banner repeater. The higher mounted banner is for the straight route but if the route was set to diverge to the left the lower banner would be off and ther higher one would be on. The pips you refer to are the Vigilance alarm built into the DSD. On the newer TPWS unit the SPAD/Overspeed alerts are given by a verbal announcement with a flashing lamp whereas the original unit just has a flashing brake demand lamp.

 

[sharpener]

It's a Splitting Banner repeater.

Thanks. I am sure the drivers are used to it but from the HF point of view it's a bit of a mess. I would have thought having the two heads side by side on the same post at the same height would be more logical. Why would you have this arrangement rather than a signal with a feather which is self-explanatory, conversely what are the indications for using a repeater at all?

 

[DadimF]

"However, if the rails are renewed (say old badly worn replaced with brand new), that makes a big difference. Especially if a different specification of rail is installed (RT60 / UIC60 replacing BS113A)."

As a Signalling design CRE that has never thought of this and over many years of renewals & can't recall others mentioning it either (assuming because it's more of a track and construction interface item) i think I'll be highlighting this as something to discuss with design, construction and NR assurance...

 

[mathstrains19]

Thanks. I am sure the drivers are used to it but from the HF point of view it's a bit of a mess. I would have thought having the two heads side by side on the same post at the same height would be more logical. Why would you have this arrangement rather than a signal with a feather which is self-explanatory, conversely what are the indications for using a repeater at all?

The repeaters give advance warning of what route is set, which I believe can allow a faster diverging route. I presume that the reason for having different heights makes it clearer which one is which - the higher one is for the main route, and the lower one is for the diverging route.

 

[GB]

Thanks. I am sure the drivers are used to it but from the HF point of view it's a bit of a mess. I would have thought having the two heads side by side on the same post at the same height would be more logical. Why would you have this arrangement rather than a signal with a feather which is self-explanatory, conversely what are the indications for using a repeater at all?

A banner repeater (or a splitting banner repeater) has a specific function. They are provided where the view of a main aspect in compromised. A route indicator in still used on the main aspect but as the post above says, splitting banners are able to offer extra information to the driver with regards to which route is set at an earlier point.

 

[DadimF]

Banner repeaters are usually (but not exclusively) provided on approach to signals that have inadequate sighting at the required reading distance. So for example on a tight bend with vegetation in the inside or a bridge or structure that blocks the view of the signal until much closer than required, you would place the banner further back to give advanced knowledge of the signal aheads aspect. Usually it just shows either on "-" so red, or off "/" so Y or G if it's a 2 state banner. 3 state banners can also show off as "/" with the background green.

Splitting banners as the one in the video are positioned as such (free to be corrected) i think as a legacy of the historical arrangements of semaphore signals which also had the main route above and diverging below and to the side.

On the point of not exclusively being provided for inadequate sighting, they can also be used to increase operational performance in certain circumstances as i have designed one in for that reason before but that's not the usual use.

 

[Russel]

There were trains on both lines approaching the area very rapidly at the time of the collision. Had it not been for the heroic actions of the Train Manager of 1B67 putting the track clips down, the accident could have been much, much worse.

It's always been good to see both EMR & RMT commending the train crew of 1B67.

For clarity, what are track clips?

 

[Kilopylae]

For clarity, what are track clips?

The signalling system runs a mild current through the rails to detect the presence of a train, because the wheel and axels connect the circuit across the tracks. Track clips are placed across the tracks to short the circuit and tell the signalling system something is there, to set nearby signals to danger.

 

[stuving]

That’s interesting, because the Carmont (the last but one) interim report gave clear causes. As did the previous Margam report. Maybe the policy changed between 2021 and 2025.

I'd expect to see an interim report only if the investigation is long, and it would contain conclusions so far and what further work is still needed. That would appear roughly when a final report might be expected - say 9-12 months.

This one is more like a second, expanded, preliminary report or statement. The preliminary statement (a news story on June 24th) was rather early by RAIB standards, and they had not then even read all the OTDR and CCTV recordings. This second one covers that, and the analysis of them, and other evidence. It does also contain some investigation results, though not much by way of conclusions.

A preliminary statement needs to contain the evidence gathered immediately after the incident, with analysis where appropriate, and describe how the investigation has been set up and why. RAIB investigations don't vary a lot; the words describing them and their objectives vary even less.

So I don't think it is a change of policy; the timing of investigation steps can vary a lot between different incidents.

 

[WAO]

The signalling system runs a mild current through the rails to detect the presence of a train, because the wheel and axels connect the circuit across the tracks. Track clips are placed across the tracks to short the circuit and tell the signalling system something is there, to set nearby signals to danger.

Axle counters?

WAO

 

[Annetts key]

The RAIB report does not state whether any train prior to 1B67 had reported the AWS Fault (Code 2) at WH154 signal, although there is mention of a subsequent report. Which presumably indicates that although the magnet was too low, its effectiveness was only impaired for a tiny fraction of the number of trains approaching the signal displaying green, which makes the accident an even more tragic event.

AWS code 2 faults (horn sounding when a bell expected) are by far the most common reported AWS fault. Back when I was an S&T fault team leader, across the area we covered, it was not unusual to have around two code 2 faults per week. Granted, some weeks we had none, but some weeks we had more.

AWS code 2 faults have numinous causes, but typicality fit into one of these categories:

• Fault found to be with the track mounted equipment, the cables or connections to this equipment or the control circuit from the lineside location cupboard, or relay room (or interlocking). Sometimes multiple trains will report the same signal/AWS installation.
  .
• Fault found to be a problem with a particular train/unit/loco/cab. Often identified because the same train reports multiple signals/AWS installations as faulty but no other trains do. And no problems are found when the track mounted equipment is checked and tested.

But there are also reports where no problems are found with the track mounted or track side equipment (or not at first) and nothing to indicate that it's a on train fault.

Sometimes the operation of an AWS electro-magnet is marginal. It works okay for most trains, but occasionally, a train will experience a code 2 fault. Some of the possible causes include (but not limited to): the electro-magnet height being low, the voltage to the coils of the electro-magnet being low (faulty power unit usually), poor or intermittent connections in the circuit to the coils of the electro-magnet or in the control circuit, one of the coils of the electro-magnet being open circuit (when it’s being operated in the 12V configuration) or where it’s been incorrectly wired.

The problem is made worse by not every driver reporting code 2 faults when they occur. This is a problem that the railway does not appear to notice. One of my standard questions that I used to ask the signaller was “how many drivers have reported this AWS as faulty?” to which the answer was normally “only one”. And in some cases, it could be two to six hours since the original report.

I know this because in the past, an S&T technician has accidentality left an AWS electro-magnet control circuit disconnected (so EVERY train that had a green aspect would get a horn), but despite lots and lots of trains travelling over the installation, only one, or occasionally, a couple of drivers reported the code 2 fault… And this has happened more than once… One installation had missing wires for the control circuit for over ten years (although this was on a freight line with the relevant signal protecting the junction where it joined the main lines, so I presume the drivers just assumed a horn at this signal was normal for the rare times they actually got a green).

Changing the height of an AWS magnet is a fairly rare thing that is done. It’s not a trivial thing to do. Normally the height is set when the AWS is installed, or when removed and refitted in connection with either track renewal work or renewal or replacement of the AWS magnets. You have to undo some large fixings (screws or bolts) and either turn the mounting plates the other way up, renew/replace them, or in very rare cases, fit packing under the mounting plates.

So if the height was found to be out of specification by that much (4mm is a lot for an AWS), it’s very unlikely to have changed that much over a fairly short time period. It has likely been incorrect for some time.

 

[Falcon1200]

So if the height was found to be out of specification by that much (4mm is a lot for an AWS), it’s very unlikely to have changed that much over a fairly short time period. It has likely been incorrect for some time.

Indeed, but presumably despite that almost every train still received the correct AWS indication?

Personally, as an Operations Controller, I was concerned if we had multiple Code 2 faults reported which for whatever reason could not or had not been resolved, due to the risk of Drivers becoming so used to cancelling an incorrect warning at a location that eventually a correct warning would be cancelled in error. Not that that happened at Elstow, or that, as far as we know, the Code 2 fault had been reported before the collision.

 

[Annetts key]

As a Signalling design CRE that has never thought of this and over many years of renewals & can't recall others mentioning it either (assuming because it's more of a track and construction interface item) i think I'll be highlighting this as something to discuss with design, construction and NR assurance...

There was a time (back in BR days) when it was normal for the civil engineering team to describe in detail with the S&T what work was going to be done. Then the (experienced) S&T could interactively point out during this meeting potential affects on S&T equipment.

Of course in this particular installation, I suspect it's an error in the fitting of the electro-magnet, as the permanent magnet appears to be within specification. The report says that the electro-magnet:

was positioned lower than the associated permanent magnet.

For clarity, what are track clips?

Officially called "Track Circuit Operating Clip". There's a photo of one on this web site.

The signalling system runs a mild current through the rails to detect the presence of a train, because the wheel and axels connect the circuit across the tracks. Track clips are placed across the tracks to short the circuit and tell the signalling system something is there, to set nearby signals to danger.

A track circuit uses the two rails as conductors. Electrical current (power) at a low voltage (less than 20V) which is electrically limited is fed to the two rails at one end of a track circuit. At the other end of the track circuit, a receiver detects the voltage across the rails, if the voltage is above a threshold, the track circuit is considered to be clear. Meaning there is no train "in section" so main aspect signals can show a proceed aspect.

If however, a train is "in section" on the track circuit, it's steel axles will short the electrical current so that little or none makes it to the receiver. With no voltage or rather, at a voltage below the threshold, the track circuit is considered occupied. So no signals can show a main proceed aspect.

Note that at both ends of the track circuit section, there are methods used to isolate each track circuit section from the next track circuit section. Traditionally, this is done by using insulated rail joints (IRJ).

The Track Circuit Operating Clip shorts the electrical current just like a train does, so the track circuit shows occupied and hence the signals that protect that section go back to red (if they were showing proceed) and will stay red while the Track Circuit Operating Clip is present.

 

[sharpener]

Fault found to be a problem with a particular train/unit/loco/cab. Often identified because the same train reports multiple signals/AWS installations as faulty but no other trains do. And no problems are found when the track mounted equipment is checked and tested.

I imagine this might be because the train-borne detection coils are for whatever reason mounted a bit high in relation to the track.

In the present instance you could well have a tolerance buildup where it is a bit of the track and a bit of the train, where neither is sufficiently out of tolerance (actual operational not "as specified") to give rise to a Code 2 fault on its own.

How big is the air gap, nominally, compared with the 4 mm discrepancy? Is it affected by the suspension travel or are the coils rigidly mounted on the bogies?

 

[norbitonflyer]

It's a Splitting Banner repeater. The higher mounted banner is for the straight route but if the route was set to diverge to the left the lower banner would be off and ther higher one would be on.

Is it? There is no diverging route at the next signal. Are they not banners for the tracks either side of the pair? (up fast and down slow - which is actually reversible at this point). The aspects on the banners match those on the corresponding signals. I don't see a banner for the up slow but maybe sighting distance for that one is good enough ewitghout one.

 

[louis97]

Is it? There is no diverging route at the next signal.

There is, 1.01.15 clearly shows a route indicator for a route to the left, for crossing to the Up Slow at Bedford South.

 

[MarkyT]

There is, 1.01.15 clearly shows a route indicator for a route to the left, for crossing to the Up Slow at Bedford South.

The splitting banner is for WH156, which has a diverging route to the up slow line.

 

[Annetts key]

I imagine this might be bc the train-borne detection coils are for whatever reason mounted a bit high in relation to the track.

In the present instance you could well have a tolerance buildup where it is a bit of the track and a bit of the train, where neither is sufficiently out of tolerance (actual operational not "as specified") to give rise to a Code 2 fault on its own.

How big is the air gap, nominally, compared with the 4 mm discrepancy? Is it affected by the suspension travel or are the coils rigidly mounted on the bogies?

I can't help with most of your questions, as I was only ever involved with the signalling (i.e. the track mounted and lineside) equipment.

The train fitters handled the train equipment faults and often, the only feedback we got was if the signaller was informed (and who then told us) or (infrastructure) fault control told us.

I do know however that the height is important. For fixed AWS installations, the height was very rarely an issue.

However, it was a different story for the "portable" permanent magnets used by the permanent way (track maintenance) staff. For temporary or emergency speed restrictions, they used these "portable" permanent magnets to warn drivers of the TSR or ESR boards, or the emergency indicator ("Dalek").

Sometimes they did not fit them correctly or the fixings were broken on them, and they were then not at the correct height (too low). That then resulted in drivers reporting AWS code 7 faults (no indication instead of horn) which are classed as wrong side (critical safety hazard) failures.

A small distance made the difference between no fault reports and multiple fault reports.

What the exact distances are, I don't know. I do know that magnetic field strength falls off rapidly as the distance is increased.

 

[CalderRail]

What the exact distances are, I don't know. I do know that magnetic field strength falls off rapidly as the distance is increased.

It is - effectively - an inverse cube law.

 

[edwin_m]

The report stated there was no alert system fitted and the signalling centre weren't aware.

It's somewhat surprising if no system was fitted as they are relatively easy to retrofit.

However, such a system, whether fitted or not, may not have been effective at this signal. A SPAD monitoring system would be looking for a condition of signal showing red and track circuit beyond signal going from not occupied to occupied. We don't have details of track circuits at this site but the track circuit that would be used to detect the SPAD may be the same one that was still occupied by the 810. In which case, except in unusual fault situations, the signal would only ever show red when that track circuit was occupied anyway.

 

[Belperpete]

It's somewhat surprising if no system was fitted as they are relatively easy to retrofit.

However, such a system, whether fitted or not, may not have been effective at this signal. A SPAD monitoring system would be looking for a condition of signal showing red and track circuit beyond signal going from not occupied to occupied. We don't have details of track circuits at this site but the track circuit that would be used to detect the SPAD may be the same one that was still occupied by the 810. In which case, except in unusual fault situations, the signal would only ever show red when that track circuit was occupied anyway.

Is the signal an automatic? If so, there may not be any indication of what aspect (red or otherwise) that the signal is showing, for a SPAD monitoring system to use.

As you say, if the only time a signal is likely to show red is because the forward track-circuit is occupied, which will inhibit detection of a SPAD, then providing a detection system that is never going to detect would be pretty pointless.

 

[sharpener]

It is - effectively - an inverse cube law.

Not too sure about that, I think where the length of the magnets is large compared with the air gap it will be at most an inverse square law. Magnetic field strength round a long wire is plain inverse (power to -1) law IIRC so might even be that.

 

[AndrewE]

Not too sure about that, I think where the length of the magnets is large compared with the air gap it will be at most an inverse square law. Magnetic field strength round a long wire is plain inverse (power to -1) law IIRC so might even be that.

but the magnets look to be about 10cm by 10cm (square) at most, from the rust particles that you see stuck over the permanent magnet bit, so I guess the clearance will be the same or a bit bigger

 

[kkong]

CCTV generally records in 5 or 10 minute blocks, the missing video will simply be due to the file not having been completely written to disk.

That is definitely not the case with any of the professional video management systems I've used (in the upstream oil and gas industry).

It would be really poor if that is how it's implemented on modern rail vehicles.

There is no reason to have more than a few seconds of video cached before writing to disk.

 

[leezer3]

That is definitely not the case with any of the professional video management systems I've used (in the upstream oil and gas industry).

It would be really poor if that is how it's implemented on modern rail vehicles.

There is no reason to have more than a few seconds of video cached before writing to disk.

Isn't the problem here that rail vehicles are bouncing around at ~100mph?

I'd assume they're running spinning disks, not flash memory for bulk storage with the age of the vehicles.
Spinning disks certainly aren't going to like being constantly written to with the amount of vibrations and so on on a rail vehicle, so keeping a buffer in flash, then periodically purging to the bulk spinning disk would seem to me to have been sensible design when implemented?

 

[35B]

Isn't the problem here that rail vehicles are bouncing around at ~100mph?

I'd assume they're running spinning disks, not flash memory for bulk storage with the age of the vehicles.
Spinning disks certainly aren't going to like being constantly written to with the amount of vibrations and so on on a rail vehicle, so keeping a buffer in flash, then periodically purging to the bulk spinning disk would seem to me to have been sensible design when implemented?

If that's the case (and I'm not convinced), it is placing the technical constraints over the purpose of saving the data in the first place. If CCTV is there to support incident investigation, those last few seconds are some of the most important moments.

But I suspect that's one of several design decisions where the requirements specifications may prove to have been somewhat complacent.