Revealed: wrong track layout is key to Paddington rail disaster

By Christian Wolmar

By Christian Wolmar

17 October 1999

INVESTIGATORS studying the Paddington rail crash are examining potential flaws in the track layout which they believe may have been one of the principal causes of the disaster.

They are concerned that a vital set of points were positioned in such a way as to allow the Thames Train to go into the path of the express rather than being switched out of harm's way. A crucial turn-off, which would have led the Thames train to safety, was unaccountably closed, ensuring that once the red light was passed the collision was inevitable.

A former senior BR manager said: "There is something very wrong here. It goes against all the basic principles of signalling and track layout. I suspect that is why the Health and Safety Executive's initial report talked about a systems failure."

The front data recorder of the Thames train was destroyed in the impact and its data lost, so we will never know if the normal warnings sounded in driver Michael Hodder's cab. However, following analysis of tapes from the recorder at the back of the Thames train, a clear sequence of events of the train's last few minutes has emerged.

Hodder, who had driven "impeccably" on his first outing to Reading and back that day, accelerated to about 40mph out of Paddington station and then coasted to 38mph, driving through two signals at caution, one set at double yellow and the second at yellow. Then, after passing the red signal SN109 still coasting, he accelerated for much of the 700 yards between it and the junction where the collision occurred, exceeding 50mph, suggesting an combined impact speed of more than 120mph.

He clearly never realised that, for the last 300 yards, he was on track supposed to be used only by London-bound trains.

The points that he should have taken to reach the fast line to Reading were, correctly, closed against him, and he carried straight on. But then he also passed a second set of points, which, had they been open, would have led him on to the slow line to Reading, which was clear; and signallers, realising his mistake, would have ordered him by radio to stop his train.

Hodder only realised something was amiss just 2.5 seconds before the crash, barely time enough for the brakes to be fully activated. The collision occurred at the junction, with the two drivers' cabs smashing into one another at a slight angle.

The key areas of the investigation are: a) whether Hodder saw signal SN109; b) the history of troubles with the signal; and c) the track layout.

The fact that Hodder accelerated suggests that either he failed to notice SN109 at all, which seems unlikely given that he had been along the route 15 times since qualifying as a driver, and dozens of times before that, or for some reason considered that it was set at double yellow. Double yellow is an extra warning indicating that two signals ahead there is a red. If he had seen it as red, he should have stopped; if he had thought it was yellow, which shows the next signal is red, he should have slowed down.

If the signal had been green, he would have heard a bell in his cab, rather than the horn that sounds for every signal other than a green. The most likely explanation is that he saw a "phantom" double-yellow signal caused by reflected sunlight, but it may be impossible to test this. The day after the accident, an attempt was made to check the signal at the same time of day. While there did not appear to be a phantom aspect, the quality of the video recording was too poor to use as evidence. One of the investigators told the Independent on Sunday: "Because the angle is slightly different every day, we will never be able to reproduce precisely the same conditions as at 8.10 am on 5 October 1999. Next year will not be quite the same, and anyway it might not be sunny."

The troubled history of the signal, which had been passed at red eight times in six years, led to a series of meetings chaired by Railtrack, which is responsible for safety of the track and infrastructure.

However, Railtrack's Great Western Zone has been in turmoil virtually since privatisation in 1996, with a poor performance record in punctuality and reliability and frequent changes of senior staff. Earlier this year the zone was publicly reprimanded by the Health and Safety Executive for broken rails in the Severn tunnel after Railtrack abandoned the long-established policy of changing the rails every six years in the damp tunnel and began allowing them to last nine years. It was only good fortune that the series of broken rails that ensued did not lead to a major disaster.

There are concerns that Railtrack did not take the problem of SN109 seriously enough. The meetings, in Swindon, were chaired by the production manager, not the zone director, and no reference was made to head office in London. As a rail industry insider put it: "Nobody took control of this problem. Nobody put two and two together, realising that not only was there a problem with this signal but also the potential consequences were disastrous." It is rare in the railways for a signal passed at danger to result in an immediate collision between trains going in opposite directions. A rail manager said: "Murphy and his law played a blinder to cause this accident."

The former BR manager said that, under BR's unified system, decisive action might well have been taken about SN109. He said that in the 1980s there had been a problem with a signal at London Bridge which, if passed even by a few yards, could have caused a collision: "BR installed detonators on the points so that if a train passed on red the driver immediately knew about it."

However, the biggest mystery, apart from why driver Hodder passed the red light, is about the track layout and why the junction was not protected by the points that would have put the train back on the slow line to Reading. This is likely to be the main focus of the official report.