Was the Cowden crash due to cost cutting?

[thedbdiboy]

Cowden and Paddington were probably two of the most pivotal accidents in terms of forcing a rethink on 'acceptable' signalling standards and SPADs. Whilst it is true that Cowden couldn't have happened if the line hadn't been singled, that point is not relevant to the need to signal single tracks with passing places safely (and if the Uckfield line hadn't been singled it would have closed - there was no money to renew the infrastructure to double track and the traffic didn't justify it).
The HSE had moved away from favouring trap points because they are designed to be run into and derail trains, which is itself not without risk. However, what neither they or the industry had sufficiently taken account of was the risk of a catastrophic SPAD - that is, not an overrun by a few hundred yards but a train simply being driven beyond a red as if it had a proceed aspect. Neither the layout at Cowden nor the one at Paddington provided for any mitigation in those scenarios. At Paddington, an overrun of signal SN109 could have been protected by a default route setting that would have created the risk of a side-swipe a few hundred yards beyond; but instead the default route led to a head on collision much further beyond, because the side swipe was considered too high a risk, whereas it was not believed probable that an overrun would result in a driver continuing to power his train a further mile or so. Furthermore, the need to have active mitigation of signal over-runs and proper train-to-signaller communication was proven essential.
In the case of Cowden, the signaller had to deal with the horror of seeing two trains converging for several minutes on a single line, with no radio or means of communicating with them. There was no TPWS protection on signals either - these two measures have been vital in addressing the safety gap that existed then. The Cowden loop still does not have trap points today but it is a much safer operation.

 

[D6130]

the signaller at Three Bridges

Excellent post....but I thought the line was controlled by the signaller at Oxted at that time?

 

[thedbdiboy]

Excellent post....but I thought the line was controlled by the signaller at Oxted at that time?

You're probably right, the key issue was that the signaller was aware of an impending collision for some minutes but with no means to remedy. I'll edit my post to avoid distracting from that point.

 

[D6130]

You're probably right, the key issue was that the signaller was aware of an impending collision for some minutes but with no means to remedy. I'll edit my post to avoid distracting from that point.

From what I've heard from former colleagues in the area, the emergency services were on their way to the approximate location of the collision before it actually occurred. As a guard at Brighton, I often worked with the Norwood driver of the down train a few years earlier on the Brighton-Manchester Inter-Cities and always found him to be an excellent driver and a sympathetic and sociable colleague - unlike some of his contemporaries, who looked down on guards and treated us with a degree of contempt.

 

[bramling]

Covered in dead flies I believe. In foggy weather. But the dead fly issue affecting the illumination seemed to be glossed over.

Yes very much glossed over. The Cowden report is notable for being one of the worst reports I’ve read, but in fairness the investigators didn’t have much to go on - all the key people died in the accident, there was virtually no technical evidence from the accident, and - I get the impression - some closing of ranks.

Personally, I’d put Cowden down to the yellow and red aspects at OD58R and OD58 being normal for a northbound journey up the branch, so the driver will have normalised the yellow aspect. Then with the extra factors of fog, a degraded illumination from the signal, and the distraction of a second person in the cab, the red was simply missed / forgotten.

If the guard really was driving then it may also have been a case that they both missed it, perhaps assuming the other was on top of things. But this is very circumstantial indeed, as there isn’t really anything to point to the guard driving apart from suspicions.

== Doublepost prevention - post automatically merged: 3 Nov 2021 ==

Excellent post....but I thought the line was controlled by the signaller at Oxted at that time?

Yes Oxted signal box, which was built as part of the singling / resignalling scheme. Still in use today.

 

[Father Jack]

In my early railway career on the Southern Region, I seem to remember that the Botley-Fareham section - singled in the early 1970s - was protected by trap points at both ends.

It isn't now according to satellite views.

I can definitely say that the trap points either end of the single line between Botley and Fareham are still there, well they were last night when I went over there.

 

[markbeech]

Singling three sections of the Uckfield line was a low-cost scheme designed to secure the future of the line; the colour light signalling + AWS which replaced the semaphore signals fulfilled all safety standards relevant at the time.

The DEMUs were deemed to be life-expired even at the end of the 1980s, this is why BR didn’t provide cab radios. As others have said the signaller at Oxted could see the trains moving towards eachother on the single line section and had no way of warning either of the drivers.

Roundabout end of 1995 BR fitted SPAD signals to additionally protect the single line sections and installed cab radios in the DEMUs (which soldiered on until 2004). Until the 171s came along drivers wouldn’t go down the Uckfield line if the cab radio had a fault.

Begs the question how much the singling actually saved. Just pleased the line’s fortunes have changed since then.

 

[thedbdiboy]

Singling three sections of the Uckfield line was a low-cost scheme designed to secure the future of the line; the colour light signalling + AWS which replaced the semaphore signals fulfilled all safety standards relevant at the time.

The DEMUs were deemed to be life-expired even at the end of the 1980s, this is why BR didn’t provide cab radios. As others have said the signaller at Oxted could see the trains moving towards eachother on the single line section and had no way of warning either of the drivers.

Roundabout end of 1995 BR fitted SPAD signals to additionally protect the single line sections and installed cab radios in the DEMUs (which soldiered on until 2004). Until the 171s came along drivers wouldn’t go down the Uckfield line if the cab radio had a fault.

Begs the question how much the singling actually saved. Just pleased the line’s fortunes have changed since then.

The blue-backed SPAD signals were installed at a number of other high risk locations too. London Bridge (old) platform 6 got one, but the new layout doesn't use them. I think it's academic how much the singling 'saved' - it saved the line for sure at the time.

 

[etr221]

My own thought on the Cowden accident is that major cause was that you had the situation where a train could start from a regular stop - Ashhurst Station - accelerate towards line speed, and then encounter a stop signal - OD58, controlling entrance to the single line - without any warning of it being at danger (as the distant signal OD58R was before the station).

And that this was a bad practice from a signalling point of view, to be avoided if at all possible. And there are a lot of ways of avoiding it - if the stop and distant signals could not be moved, than an obvious one is to have a banner repeater as a station starter.

I recall reading at about that time of an accident in similar circumstance - train starting from a station stop encountering a stop signal without warning - at Western Avenue, in Chicago, USA

 

[markbeech]

Definitely that’s it’s not ideal but wouldn’t the positions of the signals be covered by route knowledge? And to my knowledge even if the driver approached OD58 at line speed, realised it was at danger and made an emergency brake application he/she could still stop clear of Blackham Jn (start of the single line).
Can we speculate whoever was sat at the contols of the up train misread/assumed a proceed aspect and wrongly continued onto the single line?

 

[bramling]

Definitely that’s it’s not ideal but wouldn’t the positions of the signals be covered by route knowledge? And to my knowledge even if the driver approached OD58 at line speed, realised it was at danger and made an emergency brake application he/she could still stop clear of Blackham Jn (start of the single line).
Can we speculate whoever was sat at the contols of the up train misread/assumed a proceed aspect and wrongly continued onto the single line?

Probably. Whilst the station was between the repeater and stop signals, OD58 was and is well sited. It's clearly visible from Ashurst platform, so a stopping train should have no difficulty seeing it. Indeed this is no different to how LU drivers drive.

But my own personal view is that this isn't quite the full story, as what this setup does do is introduce a psychological situation where the repeater is irrelevant to a stopping train - perhaps less so nowadays especially with aids like DRA, but almost certainly so in 1994. All of this was no problem until something else arose - which it did on the day of the accident in the form of several factors: thick fog, a second person in the cab, and possibly the sub-par illumination being emitted from the signal lens. So with the yellow aspect essentially irrelevant, and the irrelevance of that yellow being a perfectly normal scenario for driving a train up that branch, the only thing preventing a catastrophic accident was the red aspect, which was fine if it was amply visible (as was the case 99% of the time), or if the driver remembered to look for it. But looking specifically for red aspects is normally a task only carried out after passing a yellow aspect, yet this yellow aspect was irrelevant most of the time. So all you're left with is the AWS, which we know can quite easily be subconsciously acknowledged.

I hope I've explained that well. This is only a personal theory, but it's based on seeing what pitfalls people have fallen into elsewhere.

 

[Taunton]

Adrian Vaughn wrote a damning summary of the circumstances of the Cowden accident in his railway accidents book; if you have an interest in this particular event it's a good read.

I've never quite got the distaction element of any second person in the cab, it seems to be grabbed as a get-out-of-jail for investigators. For the first 150 years of railways there were always two persons in the cab and it was never given as an issue - the 1957 Lewisham accident report, for example, criticised the fireman for not assisting the driver, despite the need to fire on the uphill gradient. Do we, when driving our cars, have notably reduced alertness when we have a passenger with us?

 

[Ianno87]

. Do we, when driving our cars, have notably reduced alertness when we have a passenger with us?

I do find that personally (if my wife is being particularly "conversational" in a particularly complex road situation).

I guess the issue is not so much the second person, but being disciplined about the conversation in the cab.

E.g. in aviation casual conversation between the flight deck crew is not permitted during take-off and landing phases.

 

[bramling]

Adrian Vaughn wrote a damning summary of the circumstances of the Cowden accident in his railway accidents book; if you have an interest in this particular event it's a good read.

I've never quite got the distaction element of any second person in the cab, it seems to be grabbed as a get-out-of-jail for investigators. For the first 150 years of railways there were always two persons in the cab and it was never given as an issue - the 1957 Lewisham accident report, for example, criticised the fireman for not assisting the driver, despite the need to fire on the uphill gradient. Do we, when driving our cars, have notably reduced alertness when we have a passenger with us?

The answer is it depends. Some people's job roles (e.g. instructors) are used to having people in the cab with them. For others it will be a very rare event - especially nowadays where cab access is more strictly controlled than in the past. Then one has to consider who the second person is - a fellow driver should know when to pipe down, someone else might not.

So there's a few variables. The situation at Cowden probably tended towards the more undesirable corner of the spectrum - a comparatively inexperienced but not novice driver, a guard who was regarded as being rather extroverted, and the poor weather conditions.

Train driving is all about routines, and it's fair to say that for some people at least, having someone else present can disrupt this.

 

[edwin_m]

For the first 150 years of railways there were always two persons in the cab and it was never given as an issue - the 1957 Lewisham accident report, for example, criticised the fireman for not assisting the driver, despite the need to fire on the uphill gradient.

I think the difference there is that the fireman on a steam locomotive was kept busy with their own job (including observing signals that the driver couldn't see so easily), but in modern traction a second person in the cab has no real role (unless there for a specific purpose such as instructing) so may want to converse with the driver just to pass the time.