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

[MarkyT]

From this video:

I estimate signal WH152 comes into view about 700m on approach, just after passing over the A421 bridge. Mounted high up on the gantry, the green displayed would have likely reinforced 1H46's misconception they had movement authority until, coming around the long right-hand curve, the rear of 1B67 came into view a few coach lengths before the signal. This might help explain the very late braking. The loss of rear facing CCTV just before impact is interesting as it suggests a power loss, possibly an attempted reboot of all systems on the 810. If that was the case, did the rear tail lights extinguish? That would leave the stranded train rather invisible with no yellow ends or other prominent markings.

 

[Snow1964]

I don't see how the AWS fault on 1B67 has anything to do with the spad by 1H46.

There might not have been a fault on the train as says track electromagnet was installed out of tolerance. Clearly harder to read a unit not out of position at high speed. But the permanant magnet was read.

The spad isn't a direct result, but possibly that signal is never normally red and driver 'read though' without registering it. Thats for final report to decide

 

[gravitystorm]

The loss of rear facing CCTV just before impact is interesting as it suggests a power loss

My understanding from reading the report is not that the CCTV failed before the impact, it's just that not all of the images were written to storage before the power was lost as a consequence of the impact.

 

[tram21]

I'm sure it'll all come out in a final report, and this is by no means slating the work of the RAIB etc., but it surprises me there isn't more information about 1B67 and what happened, because surely the drivers were interviewed etc. and can give direct information as to what was happening, for example in terms of the train fault etc? It all seems like they are trying to work things out, that the drivers could tell them directly, from dodgy data recordings etc?

 

[josh-j]

I'm sure it'll all come out in a final report, and this is by no means slating the work of the RAIB etc., but it surprises me there isn't more information about 1B67 and what happened, because surely the drivers were interviewed etc. and can give direct information as to what was happening, for example in terms of the train fault etc? It all seems like they are trying to work things out, that the drivers could tell them directly, from dodgy data recordings etc?

They mention witness evidence in that section, my assumption was that basically meant the driver and anyone else in the cab, since who else would be a witness to the AWS warning.

 

[NSEWonderer]

There are Two versions of GSMR units available.

One has a back up battery, these are fitted to the less reliable old diesel trains. So in the event the train loses electrical power the battery in the GSMR will supply power to it.

Then there's the version without fitted to most trains. These trains are very vulnerable as although their electrical systems are more reliable, the trains batteries are usually on one of the middle carriages, so if the carriages separate for any reason then all power including to the GSMR is lost.

I can only assume the GSMR units with back up batteries cost more.

Quite a few trains also have onboard backup batteries that power emergency lighting, GSMR and any safety critical low consumption systems, id have arguably thought that was the norm especially with the newer 810. Not sure about the 360 Desiro but 450s and 444s if the line light goes and no power is coming to the train, the batteries and emergency lighting kick in not long after.

I assume the reason they might not have is potentially the collision severed or damaged cables or the batteries themselves?

 

[321362]

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.

According to the report:

"From saloon and platform CCTV footage, RAIB has so far identified 100 people as being on board (1B67) , including eight EMR staff."

so very likely it was an offduty member of staff.

 

[NSEWonderer]

I'm sure it'll all come out in a final report, and this is by no means slating the work of the RAIB etc., but it surprises me there isn't more information about 1B67 and what happened, because surely the drivers were interviewed etc. and can give direct information as to what was happening, for example in terms of the train fault etc? It all seems like they are trying to work things out, that the drivers could tell them directly, from dodgy data recordings etc?

It was described quite well i thought?

OTDR data and witness evidence showed that the vellow AWS button started flashina on the enhanced TPWS control panel once the brake application was triggered. Witness evidence was that the train control and management system (TCMS) generated a major fault alarm because of the emergency brake application. Data downloaded from the TCMS shows that the TCMS acknowledgement button was pressed by the driver after the brake application generated the alarm.

The driver reported to the signaller via their GSM-R radio that there was an AWS fault with the train. The recording of this call includes a loud alarm sounding continuously in the cab in the background. Hitachi has confirmed to RAIB that this was the AWS warning indication and that a loudspeaker above the driver's head provides the audible alerts from the AWS/TPWS system.

Witness evidence was that the alarm was coming from the loudspeaker above the driver's head and that neither the driver nor driving instructor recognised this sound as being an AWS warning indication. Witness evidence, supported by OTDR evidence, is that the AWS acknowledgement button which would have silenced this alarm was not pressed

On 9 July 2026, another class 810 unit approaching signal WH154 incorrectly received an AWS warning indication while this signal was displaying a green aspect. In this case. the driver acknowledaed the incorrect warning indication within the 2.7 second response window. They reported the matter to the signaller immediately, and the train was able to continue its journey.

Further investigation following this second incident found that the electromagnet within the AWS track equipment (which is energised onlv when a green aspect is shown, and interfaces with the AWS equipment on the train to produce the AWS clear indication) was positioned lower than the associated permanent magnet. It was also positioned 16 mm below railhead level, outside the required tolerance of t12 mm. Work is continuing to understand the interface between the AWS electromagnet at signal WH154 and the AWS equipment on train 1B67 which led to the driver receiving a warning indication with the signal displaying a green aspect. RAIB has found no evidence that the AWS electromagnet was not, correctly energised,

 

[LYuen]

I don't see how the AWS fault on 1B67 has anything to do with the spad by 1H46.

Completely agree. The fault on 1B67 correctly triggered the danger signal at WH152, which should have protected both 1B67 and 1H46. That's the end of matter with 1B67.
It was a fault but it did not cause the accident.

The direct cause of the accident is why SPAD could happen. And why SPAD is an AWS warning can be overridden to remove all protections.

 

[Haywain]

Had it not been for the heroic actions of the Train Manager of 1B67 putting the track clips down,

Train Manager AND driver. And I think "heroic" overplays it a touch.

 

[NSEWonderer]

Completely agree. The fault on 1B67 correctly triggered the danger signal at WH152, which should have protected both 1B67 and 1H46. That's the end of matter with 1B67.
It was a fault but it did not cause the accident.

The direct cause of the accident is why SPAD could happen. And why the AWS warning can be overridden to remove all protections.

It was a link in a chain that caused the collision, the right side failure had it been acknowledged, said train carries on, no collision. That isnt blaming anyone its just a fact, same as had that same train(1B67) not been late, again the same event, no collision. Investigators won't just look at one thing and say thats the reason, they look at all possible links in the chain including those major and make recommendations based on those to prevent the next as every link can hold the opportunity to deal with an issue that may cause the next unfortunate event.

 

[Haywain]

it surprises me there isn't more information about 1B67 and what happened, because surely the drivers were interviewed etc. and can give direct information as to what was happening, for example in terms of the train fault etc?

Or perhaps RAIB feel that it's not appropriate to go into such detail at this stage.

 

[Farnborough]

The driver reported to the signaller via their GSM-R radio that there was an AWS fault with the train. The recording of this call includes a loud alarm sounding continuously in the cab in the background. Hitachi has confirmed to RAIB that this was the AWS warning indication and that a loudspeaker above the driver's head provides the audible alerts from the AWS/TPWS system.

Witness evidence was that the alarm was coming from the loudspeaker above the driver's head and that neither the driver nor driving instructor recognised this sound as being an AWS warning indication. Witness evidence, supported by OTDR evidence, is that the AWS acknowledgement button which would have silenced this alarm was not pressed

I can understand a driver under training not knowing everything (that's the point of training), but the instructor not knowing that this sound as being an AWS warning indication is a touch concerning?

 

[baz962]

I can understand a driver under training not knowing everything (that's the point of training), but the instructor not knowing that this sound as being an AWS warning indication is a touch concerning?

To be fair to them , the aws in an 810 sounds or did sound very different to any other traction I have driven , including 222's and 360's , plus others at previous TOCs. All the sounds in the cab were different and quite awful really. Although I personally don't note the unit numbers and can't say for sure about 810015 .

 

[tram21]

Train Manager AND driver. And I think "heroic" overplays it a touch.

With respect, I do know more about exactly what happened here, and it was certainly heroic.

== Doublepost prevention - post automatically merged: Thursday at 12:40 ==

Or perhaps RAIB feel that it's not appropriate to go into such detail at this stage.

Yes I'm sure

 

[CalderRail]

There might not have been a fault on the train as says track electromagnet was installed out of tolerance. Clearly harder to read a unit not out of position at high speed. But the permanant magnet was read.

I doubt it was installed out of tolerance. It is more likely to have shifted to an out of tolerance position over time, due to environmental, wear or vibration factors.

 

[MarkyT]

My understanding from reading the report is not that the CCTV failed before the impact, it's just that not all of the images were written to storage before the power was lost as a consequence of the impact.

The report states:

The available rear-facing CCTV from train 1B67 ends before this train came to a stand, so images of the approaching train 1H46 are not captured.

1B67 came to a stand at 17:11:31, and the collision occurred at ~17:13:35. That suggests images were not captured for over two minutes before any collision damage could have happened. If that's representative of the latency of the storage process, it seems a very long time for a judicial system recording fast-changing operational events.

 

[SteveT]

Train Manager AND driver. And I think "heroic" overplays it a touch.

Indeed. Aren't track clips the modern equivalent of detonators?

 

[Dave W]

Would an AWS warning at a clear signal not have generated confusion? - but whatever action the driver of 1B67 took they would have fully expected to be - and were - protected by the signals in rear.

From memory, there are examples on Railways Archive of reports into collisions immediately after junctions where an overtaken train joins the line which the overtaking train has passed and then unexpectedly stopped. Swiss cheese indeed.

 

[Haywain]

IB67 came to a stand at 17:11:31, and the collision occurred at ~17:13:35. That suggests images were not captured for over two minutes before any collision damage could have happened.

The report states that the "available rear-facing CCTV from train 1B67 ends before this train came to a stand", so it's in excess of two minutes. And I though that having 20 seconds missing from the FFCCTV on 1H46 was bad!

== Doublepost prevention - post automatically merged: Thursday at 13:00 ==

With respect, I do know more about exactly what happened here, and it was certainly heroic.

And, with respect, I take my information from the publicly available interim report and will form an opinion based on that.

 

[Dogbox]

Would an AWS warning at a clear signal not have generated confusion? - but whatever action the driver of 1B67 took they would have fully expected to be - and were - protected by the signals in rear.

From memory, there are examples on Railways Archive of reports into collisions immediately after junctions where an overtaken train joins the line which the overtaking train has passed and then unexpectedly stopped. Swiss cheese indeed.

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

 

[800301]

Would an AWS warning at a clear signal not have generated confusion? - but whatever action the driver of 1B67 took they would have fully expected to be - and were - protected by the signals in rear.

From memory, there are examples on Railways Archive of reports into collisions immediately after junctions where an overtaken train joins the line which the overtaking train has passed and then unexpectedly stopped. Swiss cheese indeed.

If the driver unfamiliar with the AWS Horn sound, seeing a green signal and getting a Horn hesitated it’s likely would be too late to cancel and and as such the brake came in as it should do, the fault alarm coming in immediately afterward would likely startle the driver further and probably putting him into a false sense of security that there actually was a fault

 

[Dave W]

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

Interesting, thank you.

 

[D1537]

I suspect the fact that the report includes this sentence is a suggestion as to why the SPAD occurred, but of course there is no way of proving such.

However, it is likely that the signal is very rarely encountered in this state. RSSB analysis indicates that before the accident, no train encountered a red aspect at signal WH154 in 2026.

 

[aleggatta]

The report states that the "available rear-facing CCTV from train 1B67 ends before this train came to a stand", so it's in excess of two minutes. And I thought that having 20 seconds missing from the FFCCTV on 1H46 was bad!

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. It may be possible for the disk to be interrogated and the missing files repaired, but with a disk that has been overwritten numerous times the ability to rebuild the video file will be limited. It may be that this can be recovered with time, but it’s not like cctv being recorded to videotape where the recording would be continuous until the point of power failure.

On another point, I too am surprised the driver missed the AWS warning on the first train.

 

[PZ 08895]

SPTs are normally only in locked cases in area of vandalism. When a SPT is used, the Signal Post Telephone concentrator will clearly show which signal the call is coming from, e.g. WH152, which the caller should also confirm verbally, but the signaller will still know the origin of the call if the caller is not familiar with railway emergency call protocol. On all except single units (e.g. class 153s) and locomotives, the GSM-R equipment at each end of the train is completely independent of that at the other end, so if there is power failure at one end due to collision damage, it is should still work at the cab at the other end assuming that has not been damaged, though appreciate it would take to get there.

 

[bramling]

I suspect the fact that the report includes this sentence is a suggestion as to why the SPAD occurred, but of course there is no way of proving such.

One wonders if what the RAIB have written is a little unclear on this. I’d be surprised if no train crossed over with the signal initially displaying red. What they presumably mean is that on all occasions it had stepped up to a proceed aspect by the time the approaching train reached it.

This would be the only time it *remained* at danger, which is quite an important distinction.

It’s notable there is pretty much nothing on human factors relating to the 360 driver.

 

[millemille]

SPTs are normally only in locked cases in area of vandalism. When a SPT is used, the Signal Post Telephone concentrator will clearly show which signal the call is coming from, e.g. WH152, which the caller should also confirm verbally, but the signaller will still know the origin of the call if the caller is not familiar with railway emergency call protocol. On all except single units (e.g. class 153s) and locomotives, the GSM-R equipment at each end of the train is completely independent of that at the other end, so if there is power failure at one end due to collision damage, it is should still work at the cab at the other end assuming that has not been damaged, though appreciate it would take to get there.

Not if the GSM-R equipment does NOT have its own internal battery.

Multiple units are not typically fitted with GSM-R with internal batteries, they rely on the train batteries. If the collision damage has put +ve battery voltage carrying cables to negative or vehicle body then breakers will have tripped or fuses blown and the vehicle batteries will no longer be supplying anything across the whole train.

 

[35B]

One wonders if what the RAIB have written is a little unclear on this. I’d be surprised if no train crossed over with the signal initially displaying red. What they presumably mean is that on all occasions it had stepped up to a proceed aspect by the time the approaching train reached it.

This would be the only time it *remained* at danger, which is quite an important distinction.

It’s notable there is pretty much nothing on human factors relating to the 360 driver.

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

 

[MarkyT]

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. It may be possible for the disk to be interrogated and the missing files repaired, but with a disk that has been overwritten numerous times the ability to rebuild the video file will be limited. It may be that this can be recovered with time, but it’s not like cctv being recorded to videotape where the recording would be continuous until the point of power failure.

Are systems available where the recording server constructs the file from a camera live stream and can neatly conclude it at any time if the feed is interrupted?