World

What Went Wrong Before the Elstow Train Collision

Elena MarquezPublished 3d ago3 min readBased on 5 sources
Reading level
What Went Wrong Before the Elstow Train Collision
source:www.gov.uk

A southbound East Midlands Railway express from Corby hit the rear of a stationary Nottingham to London service near Elstow, Bedfordshire, at about 5.13pm on 19 June 2026. That is the account in Interim Report 01/2026, published 1 October 2026. The driver of the moving train was killed. One hundred and sixty people were injured. The Guardian

The moving train was driven by 60-year-old Shaun Burton. Investigators found he acknowledged the AWS warning after passing the signal at red. AWS is the in-cab alarm that sounds at a restrictive signal and must be cancelled by the driver. Cameras recorded the signal showing a red aspect, meaning stop, both on approach and at the moment of passing. RAIB That signal had not displayed red at any point in 2026 before the incident.

After that SPAD, the term for passing a signal at danger, the Corby service did not slow. It continued to accelerate to 76mph. Emergency braking then intervened. Speed at impact was 49mph.

The train ahead had stopped after a false AWS horn. It had passed a green aspect, meaning go. Neither the driver nor the driving instructor on board recognised the sound as a warning, in part because the cab layout was unfamiliar. AWS requires acknowledgement within 2.7 seconds. Without that input, it commands an emergency brake application. The Nottingham-London service stopped on the running line under that brake demand.

That leading train was a Class 810 Aurora built by Hitachi. The type had entered service with East Midlands Railway within the previous six months. In the weeks after the collision, a second Class 810 received an incorrect AWS warning at the same signal while it displayed green.

Tributes were paid in the week after the collision to Burton, an East Midlands Railway driver. BBC

The broader context here is a sequence with two separate AWS events. The first created the obstruction, when a new fleet type stopped in section after a false horn and a missed acknowledgement window. The second involved a SPAD with the AWS acknowledged in the cab, followed by acceleration to line speed.

Looking at what this means for the investigation, the focus falls on human-machine interface and fleet introduction as much as on signal observance. A signal that had not shown red all year changes what a driver expects on approach. A cab where crews do not identify the horn changes the response in the seconds after a green. A repeated false indication at the same signal points investigators toward infrastructure and rolling-stock interaction. The interim report, detailed in Interim Report 01/2026, leaves those threads open pending the full investigation.