610 metres that did not exist, and a wheel measurement that was never saved
What happened
On 21 October 2026 the Rail Accident Investigation Branch published findings on a signalling failure near Welshpool in Powys. At about 23:52 on 15 September 2026, an infrastructure monitoring train — a train that inspects the track rather than carrying passengers — was running under what is called ETCS Level 2 supervision. ETCS stands for European Train Control System. It is a digital signalling system in which the train is told how far it is permitted to travel, and the on-board computer continuously works out where the train is and how much of that permission is left. The train reached the marker that should have been the end of its permitted movement, and stopped there. But its on-board equipment still believed it had about 610 metres left to run. In railway language this is a "wrong side failure". A failure on the right side makes a system more cautious than it needs to be. A wrong side failure makes it less cautious than it should be — which is the kind that matters. Nobody was reported hurt. RAIB describes the event as a failure rather than an accident. The cause identified was that the on-board equipment was using incorrect wheel diameter information to calculate the train's position.
Why this matters
The wheel diameter detail is the part that deserves explaining, because it sounds trivial and is not. A train partly works out how far it has travelled by counting wheel rotations. To turn rotations into distance, the computer needs to know the circumference of the wheel — and therefore its diameter. Train wheels wear down over time and are periodically re-profiled, so the diameter changes and must be updated in the system. If the stored diameter is wrong, every distance calculation is wrong, and the error accumulates with every mile. That is how a train can end up believing it has 610 metres of permission it does not have. Investigators found that a component in the on-board equipment had prevented the correct wheel diameter from being stored after routine maintenance. So the maintenance was done; the new figure simply did not save. The more significant finding is about the safety nets. ETCS does not rely on wheel rotations alone. There is an independent speed measurement system, and there are balises — electronic beacons fixed to the track which tell a passing train exactly where it is, correcting any accumulated error. In this case neither corrected the distance error. This matters because ETCS is being rolled out across Britain's railway. The Cambrian line, where this happened, has had it the longest, which makes it the place where such lessons emerge first.
What the evidence actually says
It is important to be precise about what RAIB has and has not done here, because this type of report is easy to over-read. RAIB decided **not** to carry out a full investigation. What it did was a preliminary examination, after which it wrote to Network Rail about possible weaknesses in the protection arrangements within the on-board ETCS equipment, for Network Rail to consider in its own investigation. Copies went to the Office of Rail and Road, the industry's safety regulator, and to the operator of the monitoring train. So this is not a finished root-cause analysis with recommendations. It is investigators identifying a concern and formally handing it to the people responsible, while noting it publicly. That is a normal and reasonable use of their powers, and the openness is a feature of the system rather than a gap in it. Some context that should not be confused: RAIB has examined ETCS integrity on the Cambrian lines before, including a 2019 investigation into loss of safety-critical data. A recurring theme on the same route is a legitimate reason for closer attention. It is also worth being clear about what this was not. This was an equipment failure on a maintenance train at midnight, with no reported injuries, and the train did in fact stop at the correct marker. It is a near-miss category event studied precisely so that it stays that way. Readers should not connect it to unrelated incidents on Welsh railways; RAIB treats each event separately, and so should we. What cannot yet be said is how widespread the component problem is. That is for Network Rail's investigation to establish.
Practical advice
There is nothing here for a passenger to act on, and it would be wrong to suggest otherwise. But a few things are genuinely worth knowing. **This is the safety system working, not failing.** The point of publishing a wrong side failure that injured nobody is to fix the cause before it matters. An industry that reports these openly is safer than one that does not, and RAIB publishes its findings for anyone to read. **Britain's railway remains extremely safe.** Rail is consistently among the safest ways to travel in this country. A technical failure report is not a reason to change how you travel. **If you want to follow rail safety properly, read RAIB directly.** Its reports are published on GOV.UK, written plainly, and considerably more informative than coverage of them. **For journey disruption, check before you travel.** Engineering work and signalling faults are posted by National Rail and the operators. On rural lines with few services, a missed connection can mean hours rather than minutes. **If you travel by train with medication, carry more than the journey needs.** Delays on single-track rural routes can be long. Keep medicines in hand luggage in original labelled packaging, and if you take anything time-critical, build in a margin rather than planning to the minute.
What to know
At around 23:52 on 15 September 2026, near Welshpool in Powys, an infrastructure monitoring train running under ETCS Level 2 stopped correctly at its limit marker — but its on-board equipment still believed it had about 610 metres of permitted movement left. This is called a wrong side failure. No injuries were reported. The cause was incorrect wheel diameter data. Trains calculate distance partly by counting wheel rotations, so a wrong stored diameter corrupts every distance calculation. A component in the on-board equipment had prevented the correct figure being saved after routine maintenance. The more important finding is that two independent safety nets — the separate speed measurement system and the trackside balises that correct position errors — did not catch it. RAIB chose not to run a full investigation. It wrote to Network Rail flagging possible weaknesses in the on-board equipment's protection arrangements, copying the Office of Rail and Road and the train's operator. So this is a formally raised concern, not a completed root-cause report with recommendations. Nothing here requires passengers to do anything. Rail remains among the safest ways to travel in Britain, and publishing failures that hurt nobody is exactly how it stays that way. Sources: GOV.UK (Rail Accident Investigation Branch), "Wrong side failure near Welshpool", 21 October 2026, https://www.gov.uk/government/news/wrong-side-failure-near-welshpool. GOV.UK, "Rail Accident Investigation Branch", accessed 22 October 2026, https://www.gov.uk/government/organisations/rail-accident-investigation-branch. Network Rail, accessed 22 October 2026, https://www.networkrail.co.uk/. This article is for general information and does not replace advice from a doctor, pharmacist or other qualified healthcare professional.
Related articles
Serbia votes on Sunday. Carry cash and charge your phone
On 22 October 2026 the Foreign, Commonwealth and Development Office added new information to its Serbia travel advice about parliamentary elections. Serbia goe…
A football tournament, a closed border, and why an emergency passport is not enough
On 23 October 2026 the Foreign, Commonwealth and Development Office updated its entry requirements advice for Azerbaijan. Three things changed. The first is a…
No in-person help, and no promise of evacuation: the Niger advice
On 23 October 2026 the Foreign, Commonwealth and Development Office updated its travel advice for Niger. The update revised information about attacks near the…