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

joshj00 · 19 Jun 2026

If a TCB fails its designed to be fail safe? How can 2 trains be in one section together? It would appear that TCOC clips have been applied to the track, perhaps after the incident. But the question remains, was this wrong routing from the the slow to the up fast? or was this a signal and detection failure allowing the trains into the same section ? thoughts? Had the stopped train cleared the previous signal but not passed far enough? Surely a call to the signaller would warrant an emergency broadcast to stop all trains.... the signals are also showing as solid yellow caution aspect... Been far too many incidents on the Bedford section in the last 12 months

RailUK Forums

AlterEgo · 19 Jun 2026

tpfx89 said:
OTDRs can help fill in the gaps on this however, manner of driving prior, speed of AWS acknowledgements, control inputs and the manner they've been used etc

They can fill in some gaps, but not conclusively.

joshj00 said:
If a TCB fails its designed to be fail safe? How can 2 trains be in one section together? It would appear that TCB clips have been applied to the track, perhaps after the incident. But the question remains, was this wrong routing from the the slow to the up fast? or was this a signal and detection failure allowing the trains into the same section ? thoughts?

Far too early to speculate, to be fair, and the information from which to speculate is far too thin on the ground to make any speculation worthwhile. What we can say is that the train from Corby is booked to cross at that junction onto the fasts and that would have been an expected move for the driver.

Watershed · 19 Jun 2026

Very tragic news.

bengley said:
The signal after Bedford South Junction on the up fast is automatic and is not fitted with TPWS, so the train would only have stopped in this situation (if the signal was at danger) if the driver applied the brakes himself.

I would imagine this will be an important factor in the investigation. I think there are legitimate safety questions to be asked about signals not fitted with TPWS. At the moment the industry's response seems to be that ETCS will address these risks, but I don't think there's any current timeline for ETCS rollout on the MML.

357 · 19 Jun 2026

tpfx89 said:
OTDRs can help fill in the gaps on this however, manner of driving prior, speed of AWS acknowledgements, control inputs and the manner they've been used etc

To an extent, yes, however there are a lot of units around the age of the 360s that really do not record much at all. Many just record "power", "brake" etc without any further detail

name_required · 19 Jun 2026

As a reminder of how safe UK railways have become in recent years: in the 20 years up to the year 2000 there were 23 fatal train collisions leading to a total of 111 deaths, whereas in the past 20 years (including today) there have only been two fatal train collisions (the other being at Talerddig in 2024) leading to two deaths.

EDIT: a couple of replies have mentioned derailments, so I'll add the numbers for those. In the 20 years up to 2000 there were seven fatal derailments (not leading to a collision with another train) that led to 23 deaths, whereas in the past 20 years there have been two fatal derailments (Grayrigg and Stonehaven), leading to four deaths.

joshj00 · 19 Jun 2026

If a TCB fails its designed to be fail safe? How can 2 trains be in one section together? It would appear that TCOC clips have been applied to the track, perhaps after the incident. But the question remains, was this wrong routing from the the slow to the up fast? or was this a signal and detection failure allowing the trains into the same section ? thoughts? Had the stopped train cleared the previous signal but not passed far enough? Surely a call to the signaller would warrant an emergency broadcast to stop all trains.... the signals are also showing as solid yellow caution aspect... Been far too many incidents on the Bedford section in the last 12 months

Royston Vasey · 19 Jun 2026

name_required said:
As a reminder of how safe UK railways have become in recent years: in the 20 years up to the year 2000 there were 23 fatal train collisions leading to a total of 111 deaths, whereas in the past 20 years (including today) there have only been two fatal train collisions (the other being at Talerddig in 2024) leading to two deaths.

There was also Stonehaven.

Nicholas Lewis · 19 Jun 2026

Watershed said:
Very tragic news.

I would imagine this will be an important factor in the investigation. I think there are legitimate safety questions to be asked about signals not fitted with TPWS. At the moment the industry's response seems to be that ETCS will address these risks, but I don't think there's any current timeline for ETCS rollout on the MML.

Majority of autos arent fitted with TPWS and its relevant question to be asked given ETCS roll out is a decades long project now if ever delivered but TPWS is a relatively easy overlay on existing signalling. The only rub is there are c 10k+ signals around the network.

357 · 19 Jun 2026

name_required said:
As a reminder of how safe UK railways have become in recent years: in the 20 years up to the year 2000 there were 23 fatal train collisions leading to a total of 111 deaths, whereas in the past 20 years (including today) there have only been two fatal train collisions (the other being at Talerddig in 2024) leading to two deaths.

You're right that UK rail has become incredibly safe since the late 90s, but this specific statistic is quite cherry-picked. By explicitly using the word 'collisions' (train-on-train), it excludes fatal derailments - most notably the Stonehaven crash in 2020 that killed three people, and the Ufton Nervet level crossing crash in 2004 (7 deaths). Also, starting the clock exactly 20 years ago conveniently cuts out the cluster of major disasters from 2000–2002, like Great Heck and Potters Bar. The railways are vastly safer today, but those numbers erase several real tragedies through technicalities.

Darandio · 19 Jun 2026

name_required said:
As a reminder of how safe UK railways have become in recent years: in the 20 years up to the year 2000 there were 23 fatal train collisions leading to a total of 111 deaths, whereas in the past 20 years (including today) there have only been two fatal train collisions (the other being at Talerddig in 2024) leading to two deaths.

There was also Grayrigg. Still, remarkable safety record.

Trackman · 19 Jun 2026

My thoughts and prayers to everyone involved.

And a big thanks to our emergency services, second to none.
I've been on emergency exercises, these pose questions like 'How do I get through this 12ft steel fence, in an emergency?'
Today it was a case of cutting through hedges, so I've read.

hexagon789 · 19 Jun 2026

Royston Vasey said:
There was also Stonehaven.

Darandio said:
There was also Grayrigg. Still, remarkable safety record.

Strictly speaking, neither of those were collisions.

Grayrigg was a derailment on a crossover due to poor maintenance.

Carmont was a derailment caused by weather-related conditions, exacerbated by inadequate oversight of contractor built drainage and lack of potentially mitigating blanket speed restriction being imposed.

tpfx89 · 19 Jun 2026

357 said:
To an extent, yes, however there are a lot of units around the age of the 360s that really do not record much at all. Many just record "power", "brake" etc without any further detail

Certainly there have been massive increases in the granularity of the data available, a comparison between both units involved in this incident will reveal a gulf in the data which is available between the two. Hopefully however the data available from a class 360 will be similar in nature to that from 185s, another Desiro of a similar era, which can certainly deliver information related to the amounts of breaking and power in use. I've never seen a 360 OTDR but I'm relatively familiar with those of 185s, and lets hope their PBC position can be read in the same way.

AlterEgo · 19 Jun 2026

Nicholas Lewis said:
Majority of autos arent fitted with TPWS and its relevant question to be asked given ETCS roll out is a decades long project now if ever delivered but TPWS is a relatively easy overlay on existing signalling. The only rub is there are c 10k+ signals around the network.

I'm surprised this one doesn't have TPWS. It's straight after a junction and may well be a risk signal for simply notching up on "autopilot" expecting clear greens. Still, the report when it eventually comes will give some clarity.

CalderRail · 19 Jun 2026

A friend who works for the London emergency response teams tells me that the London HEMS unit has been sent north, and London emergency services are running part of East of England (presumably Essex) so that EofE can focus on Bedford. All of this is stuff they have pre-prepared plans for ofc.

yorkie · 19 Jun 2026

This thread will shortly be locked overnight.

Also, until further notice, please can we only submit posts that contain actual updates and/or new information.

Omnishambles · 19 Jun 2026

QSK19 said:
A very, very good point. Those people are almost always disregarded when considering who fits within the definition of “those affected”. They are very much within that definition and it’ll be a very difficult time for them.

Fully applauded, from a Control have been involved in incidents that have been declared major and 2 where staff members were killed on duty and can attest you fall into the “affected” category. Thoughts with all of those with dealings in this incident in any capacity

Strathclyder · 19 Jun 2026

Had a numb, sinking feeling in my gut when I came across the BBC News live feed for this eariler (much like I had when news of Carmont first started emerging), which has only deepened the more photos/videos & details have emerged.

Truly horrible incident. My heartfelt condolences to all those affected, in and out of the railway family and to the family, friends and close colleagues of the deceased. Not much else I can or should say really.

TT-ONR-NRN · 19 Jun 2026

Horrible news to come home to. Very sad.

Steve Harris · 19 Jun 2026

BBC 1 News coverage has just confirmed that the fatality was one of the train drivers (who's family have been informed).

I can't give a link as yet, as their website is behind their presenter Jane Hill.

AlterEgo · 19 Jun 2026

Steve Harris said:
BBC 1 News coverage has just confirmed that the fatality was one of the train drivers (who's family have been informed).

I can't give a link as yet, as their website is behind their presenter Jane Hill.

ITV confirming it was the driver who died.

https://www.itv.com/news/2026-06-19/trains-between-bedford-and-luton-cancelled-as-authorities-respond-to-incident

A train driver has been killed and 33 people have been seriously injured after two trains collided near Bedford.

In a statement, British Transport Police said they responded to a call about a collision at 5.15pm, and they now know "a number of people have been injured, and one person has very sadly died."

Deputy Chief Constable Stuart Cundy said: "We've declared a major incident and a significant emergency service response is ongoing. We're working at pace to establish exactly what's happened and will provide further information as soon as we are able to."

Annetts key · 19 Jun 2026

Nicholas Lewis said:
Majority of autos arent fitted with TPWS and its relevant question to be asked given ETCS roll out is a decades long project now if ever delivered but TPWS is a relatively easy overlay on existing signalling. The only rub is there are c 10k+ signals around the network.

It should be noted that the now obsolete GWML ATP system is far easier to overlay on an existing signalling system than TPWS is for plain line signals like automatic signals.

AlterEgo said:
I'm surprised this one doesn't have TPWS. It's straight after a junction and may well be a risk signal for simply notching up on "autopilot" expecting clear greens. Still, the report when it eventually comes will give some clarity.

TPWS is normally only fitted to signals where there is a junction or points ahead (beyond) the signal. That's one of the reasons why it's cheaper than ATP systems. Hence by definition, an automatic signal will not normally be fitted.

Only where there is some other requirement (double blocking maybe) would an automatic signal be fitted.

Seagull2006 · 19 Jun 2026

Condolences to all affected by this tragedy.

rmHawk765 · 19 Jun 2026

Annetts key said:
It should be noted that the now obsolete GWML ATP system is far easier to overlay on an existing signalling system than TPWS is for plain line signals like automatic signals.

I imagine ETCS Limited Supervision is just as easy to roll out though and it would solve this issue of overspeeding under restrictive aspects definitively.

zwk500 · 19 Jun 2026

rmHawk765 said:
I imagine ETCS Limited Supervision is just as easy to roll out though and it would solve this issue of overspeeding under restrictive aspects definitively.

Level 1 Limited supervision would only eliminate the problem at fitted signals (by definition).
To eliminate the problem definitively would require L1 Full Supevision or L2 (overlay or no-signals).

Ive no idea how installing L1 LS compares to Western ATP but I suspect it's not as simple due to the fact NR havent rolled it out as a 'first step' and it wasn't mandated for the replacement of Chiltern ATP.

Annetts key · 19 Jun 2026

rmHawk765 said:
I imagine ETCS Limited Supervision is just as easy to roll out though and it would solve this issue of overspeeding under restrictive aspects definitively.

The GWML ATP uses small current transformers to replace the 2BA (or 0BA) links in the location cupboard in the aspect feed to the signal. That means that it's easy to add without any significant modifications to the existing signalling system.
I'll add details in a separate topic.

Edders23 · 19 Jun 2026

My sympathies to all those affected.

rmHawk765 · 19 Jun 2026

zwk500 said:
Level 1 Limited supervision would only eliminate the problem at fitted signals (by definition).

I am assuming here that it is actually fitted at all signals... it would of course share an issue with TPWS if not.

Now personally I am not an advocate for a halfway-done rollout again, like we saw with ATP. I'd rather L2 is just slowly rolled out everywhere but I fear we are decades away from even the beginnings of such a project.

Annetts key said:
The GWML ATP uses small current transformers to replace the 2BA (or 0BA) links in the location cupboard in the aspect feed to the signal. That means that it's easy to add without any significant modifications to the existing signalling system.

I see, didn't know about it at that level. I'm assuming ETCS L1 LS cannot be set up like this?

Flange Squeal · 19 Jun 2026

AlterEgo said:
ITV confirming it was the driver who died.

https://www.itv.com/news/2026-06-19/trains-between-bedford-and-luton-cancelled-as-authorities-respond-to-incident

As of 2331, they seem to have edited the page to say “One person has been killed” since your quote.

Javelin_55 · 19 Jun 2026

AlterEgo said:
The forward facing camera does not record why things happened, for example, if the driver was distracted, or became incapacitated.

Excellent point. Even with any camera footage, the worst outcome I can think of for the investigation is if all the data is reviewed and shows no signalling irregularities and the train being controlled appropriately to the conditions, up to the point of collision. Under those circumstances the only person who could explain the incident would be the driver themselves.

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

Member

RailUK Forums

Verified Rep - Wingin' It! Paul Lucas

Veteran Member

Established Member

Member

Member

Established Member

Established Member

Established Member

Established Member

Established Member

Veteran Member

Member

Verified Rep - Wingin' It! Paul Lucas

Member

Forum Staff

Member

Established Member

Veteran Member

Member

Verified Rep - Wingin' It! Paul Lucas

Established Member

Member

Member

Veteran Member

Established Member

Member

Member

Established Member

Member