Is this what you call a SPAD?

[zwk500]

Is any of that applicable to a train that only needs to go back a little bit along a line it just came along legitimately?

Yes. Again, it is not like waving a mate out of a driveway.

 

[endofthetunnel]

One thing that’s been missed is that the MOM who comes to help doesn’t necessarily need to set the train back -at all. In cases where the train has taken the signal and therefore locked the route, but not actually passed the points in question, the MOM can instead manually pump the points into the relevant position and then the train can proceed in the correct direction with no need for reversing (with relevant rules followed, signaller authority etc etc)

 

[Meerkat]

Yes. Again, it is not like waving a mate out of a driveway.

Well no, because you don't have a bod controlling traffic lights stopping people driving by whilst you wave your mate out.

 

[43066]

That would be a risk assesment for the signaller/controller. "Do we need to risk a MOM trackside, or can we safely use a driver who is about to pass by the other way?"

They won’t necessarily be familiar with the location on the ground…

You’ve been told several times by experienced and knowledgable posters why your suggestion won’t work, so why do you keep arguing?

 

[Surreytraveller]

Well no, because you don't have a bod controlling traffic lights stopping people driving by whilst you wave your mate out.

Let's run the railway your way, and see how many people we can kill

 

[Trothy]

I can think of several locations where if a wrong route were offered, the Driver approaching it would have insufficient time/distance to bring their train to a stand at the signal upon first being made aware of the routing beyond it

The delay attribution guide was ammended for this - it now reads that if the driver is unable to to stop at the junction signal safely it's only is he then passes another signal on the incorrect route that the incident would be split between the signaller and the driver.

In the Manchester area we have definitely had trains approaching Ordsall Lane heading towards Manchester wrong routed by ARS and not be able stop at the junction in time.

 

[Efini92]

The delay attribution guide was ammended for this - it now reads that if the driver is unable to to stop at the junction signal safely it's only is he then passes another signal on the incorrect route that the incident would be split between the signaller and the driver.

In the Manchester area we have definitely had trains approaching Ordsall Lane heading towards Manchester wrong routed by ARS and not be able stop at the junction in time.

It’s somewhat ironic that PRI’s were installed around ordsall lane, apart from where the approach speed was highest.

 

[najaB]

That would be a risk assesment for the signaller/controller. "Do we need to risk a MOM trackside, or can we safely use a driver who is about to pass by the other way?"

A risk assessment done by someone who isn't on site. Who is responsible if things go wrong - the signaller/controller or the driver(s) who gave them the description of the situation?

 

[Stephen42]

Anyone able to provide a summary of this incident?

This is from publicly available information and has some gaps as a result. The freight train was from London Gateway to Hams Hall with signalling headcode 4M47.

The usual route uses the Barking to Gospel Oak line which wasn't available due to engineering works. There were two schedules in place both using that headcode:

• An altered schedule from the usual process, RealTimeTrains link which was pathed at Canonbury to go to Finsbury Park and onwards up the ECML
• A very short term plan schedule (typically within 48 hours) RealTimeTrains link which was pathed at Canonbury onwards to Camden Road and up the WCML

The altered schedule was cancelled in TRUST probably around 9am when it was activated. The very short term plan schedule was the one in use. The train was around 5 minutes late when going through Canonbury behind a London Overground service at roughly 12:50.

From what's posted on here Automatic Route Setting (ARS) equipment may have called the route for the altered schedule. ARS is relatively simplistic, typically acting based on the headcodes and scheduled route. (If the ARS doesn't connect to either TRUST or the very short term plan schedules it could have had just the one schedule for the Finsbury Park route.)

The train went far enough past the signal to trigger the berth step in the signalling system and lock the route without it being able to reset by the signaller. It's unclear exactly how far it went past though the photo in the original post suggests not very. After the issue was resolved the train then continued with an hour delay along the very short term plan route.

 

[Mcr Warrior]

@Stephen42. Many thanks for your concise summary.

 

[Annetts key]

The train went far enough past the signal to trigger the berth step in the signalling system and lock the route without it being able to reset by the signaller. It's unclear exactly how far it went past though the photo in the original post suggests not very. After the issue was resolved the train then continued with an hour delay along the very short term plan route.

Depending on the train detection system used (track circuit type or axle counter) for a controlled signal, the distance between the signal and the first axle of a train causing the train detection system to “show occupied” can be as little as 10cm. Obviously it could also be further, sometimes as much as a couple of metres.

Of course, in reality the first axle is not right at the front to the train, so the distance the train can travel is always slightly greater.

Assuming the route was called for the signal to clear to a proceed aspect, as soon as the train detection system detects the train passing the signal, the route will automatically become locked (this is the route locking, which the signaller does not directly control). And the train describer (TD) will step the description (head code) forward.

Even if the signaller cancels the route, in most interlocking systems, the route locking will remain effective until the train detection system detects that the section is clear of trains.

If the route locking is effective, no points in the route where the route locking is effective can be moved on power. Similarly any points directly locked by the train detection system also cannot be moved on power.

Hence there are three options:

1. Allow the train to go forward as per the route that was set,
2. Arrange for the train to be moved in the opposite direction, either until the front is in rear of the signal it passed, and hence allow the route locking to become free. Or for the train to travel further so that it can use a crossover elsewhere.
3. Manually operate points and then the signaller can give the driver instructions to proceed.

Obviously all relevant rules and procedures must be correctly followed.

It should also be mentioned that the above are the usual arrangements. It’s the railway, so it’s perfectly possible that there are differences with some signalling systems/interlockings.

 

[Falcon1200]

The booked WTT path for this service is via South Tottenham, then Willesden and WCML. Gospel Oak-Barking was closed yesterday so the schedule was amended to run NLL to Canonbury, then up the ECML. The amended schedule was then cancelled and a STP schedule created for NLL to Willesden, then WCML. So with 3 different schedules going 3 different routes, it is not surprising that confusion arose!

The usual route uses the Barking to Gospel Oak line which wasn't available due to engineering works. There were two schedules in place both using that headcode:

• An altered schedule from the usual process, RealTimeTrains link which was pathed at Canonbury to go to Finsbury Park and onwards up the ECML
• A very short term plan schedule (typically within 48 hours) RealTimeTrains link which was pathed at Canonbury onwards to Camden Road and up the WCML

So presumably it was realised late on that the booked diversion via the ECML could not be used (Driver route knowledge issues, other possessions elsewhere?) so the NR Control responsible for the train's origin location created a VSTP schedule for a rediversion via the WCML, however for whatever reason this was not picked up by ARS, or maybe communicated to the Signallers involved?

 

[norbitonflyer]

Depending on the train detection system used (track circuit type or axle counter) for a controlled signal, the distance between the signal and the first axle of a train causing the train detection system to “show occupied” can be as little as 10cm. Obviously it could also be further, sometimes as much as a couple of metres.

Of course, in reality the first axle is not right at the front to the train, so the distance the train can travel is always slightly greater.

Does that mean that drivers of locos where the cab is considerably more than 10cm, and maybe even 2 metres, behind the leading axle (37s etc, not to mention steam locomotives), see signals going to red in front of their eyes?

 

[zwk500]

Does that mean that drivers of locos where the cab is considerably more than 10cm, and maybe even 2 metres, behind the leading axle (37s etc, not to mention steam locomotives), see signals going to red in front of their eyes?

No, because the signals are not visible from the side.

 

[norbitonflyer]

No, because the signals are not visible from the side.

How close do you get to a signal before it can't be seen from a cab?

 

[zwk500]

How close do you get to a signal before it can't be seen from a cab?

I'm not sure on the exact numbers, but the usual standback distance is 20m. It will vary depending on where the signal is position relative to the drivers eyeline (Gantry, left/right etc) and what the visibility from the cab is like.

At stations where trains are tight up on the signals additional 'OFF' indicators are provided, or very rarely Co-Acting/repeater signals (IIRC Waterloo has a set of mini aspects at 90 degrees to the main signal on some platforms.) Example from Penzance here: https://www.railsigns.uk/photos/p_clsig2.html#pic_mcus

 

[norbitonflyer]

I'm not sure on the exact numbers, but the usual standback distance is 20m. It will vary depending on where the signal is position relative to the drivers eyeline (Gantry, left/right etc) and what the visibility from the cab is like.

Just googled the Flying Scotsman, which is apparently 70 feet long (21.6m) - which I think includes the tender, so 20m should be enough!

 

[zwk500]

Just googled the Flying Scotsman, which is apparently 70 feet long (21.6m) - which I think includes the tender, so 20m should be enough!

Yes, that length includes the tender (Flying Scotsman is rather snug on a 70' turntable).

 

[Efini92]

How close do you get to a signal before it can't be seen from a cab?

Depends on the type of signal, its location and weather conditions.
The newer LED type signals you can see up until you are beside them.

Just googled the Flying Scotsman, which is apparently 70 feet long (21.6m) - which I think includes the tender, so 20m should be enough!

Some signals probably would go back before the cab reached the signal but it would only be noticeable at slow speed, and by that point the focus wouldn’t be on that signal anymore.

You see the opposite on the mid platform signals at Birmingham new street. They will still be clear when 5 cars have passed it.

 

[Meerkat]

They won’t necessarily be familiar with the location on the ground…

You’ve been told several times by experienced and knowledgable posters why your suggestion won’t work, so why do you keep arguing?

Because I am interested why it wont work, rather than just a "wont work".
Its not like the railway doesn't have a habit for 'because we have always done it this way' and 'its safer' without an actual calculation of whether it is actually safer enough to justify it.

 

[zwk500]

Because I am interested why it wont work, rather than just a "wont work".

But when people give you a list of the reasons why it won't work, you then just repeat you questions.

Its not like the railway doesn't have a habit for 'because we have always done it this way' and 'its safer' without an actual calculation of whether it is actually safer enough to justify it.

The railway has a very clear valuation of how much safety is worth. For sure, the railway is guilty of 'we've always done it this way', but quite often the proposals either create new risks or cost a large amount that doesn't justfiy the safety benefit.

 

[43066]

Because I am interested why it wont work, rather than just a "wont work".

The reasons have been clearly explained, yet you refuse to accept them. That suggests you aren’t interested in the answers at all, and just want to argue for the sake of it.

Its not like the railway doesn't have a habit for 'because we have always done it this way' and 'its safer' without an actual calculation of whether it is actually safer enough to justify it.

How would you know, have you ever worked on it? That sounds like your own bias talking.

For sure, the railway is guilty of 'we've always done it this way', but quite often the proposals either create new risks or cost a large amount that doesn't justfiy the safety benefit.

Indeed. I’m not convinced the railway is any worse than many other industries in this respect. There’s often a very good reason why long established practices are as they are, safety is paramount, and cost underpins everything. Armchair experts always know better, however!

 

[Chris M]

The railway has a very clear valuation of how much safety is worth. For sure, the railway is guilty of 'we've always done it this way', but quite often the proposals either create new risks or cost a large amount that doesn't justfiy the safety benefit.

When the solution to Risk A creates a new Risk B you need to weigh up which risk is greater and then account for possible mitigations for Risk B, which in turn may introduce Risk C, and so that and its mitigations need to be factored in. Then you factor in that as Risk A no longer exists, a mitigation for that doesn't need to happen so Risk D can also eliminated, but then you need to check whether that mitigation is also mitigating against other risks and whether they still exist. All this can get very complicated very quickly, and all evaluation of risk involves some degree of subjectivity. For example, "the Fatal and Weighted Injuries Index gives a fatality 10 times the weight of a serious casualty, and a serious casualty 10 times the weight of a slight casualty" (according to the first source google found) so if procedure A has a risk of 1 serious casualty in a given time period and procedure B has a risk of 10 slight injuries in the same time period which is safer? If procedure A risks 1 broken arm and 1 broken leg in a 5 year period is that better or worse than procedure B which risks 2 broken arms but no broken legs in the same 5 years?
[note this is all hypothetical, I'm not suggesting this is the risk for any specific procedure discussed or proposed in this thread]
There have been plenty of occasions where risk assessments by industry professionals have missed or miscalculated a risk (IIRC there was a recentish RAIB report where something that was risk assessed as happening once every few years was actually happening multiple times a month), and they're in a far better position to judge than those of us who are outside the industry.
Risk assessments all cost money, and I expect there will be instances of a risk assessment getting part way through and concluding that it is unlikely the new procedure will be unlikely to be significantly different in terms of safety so there is no point continuing with the assessment unless there is some other business benefit to the new procedure.

 

[Malaxa]

So what's the worst case scenario in the Meerkat approach then?

 

[najaB]

So what's the worst case scenario in the Meerkat approach then?

Two "worst case" scenarios that come to mind: a derailment if points move under the train. It would be low speed, but could still do quite a bit of damage. Or someone could get crushed by the train if the driver in the assisting train can't see them from the cab.

 

[Annetts key]

Does that mean that drivers of locos where the cab is considerably more than 10cm, and maybe even 2 metres, behind the leading axle (37s etc, not to mention steam locomotives), see signals going to red in front of their eyes?

It takes time for the signal to change the aspect it is displaying.

How far does a train move in half a second? In most cases, the signal is no longer in view of the driver when it changes the displayed aspect.

And the driver is looking beyond the signal.

In station areas or other areas where slow speed shunting or propelling movements may be made, “last wheel replacement” may have been specified in the requirements.

 

[66701GBRF]

Two "worst case" scenarios that come to mind: a derailment if points move under the train. It would be low speed, but could still do quite a bit of damage. Or someone could get crushed by the train if the driver in the assisting train can't see them from the cab.

Points moving under the train is a non issue in respect of who is controlling the move at the back. If they can move, then they can move whether it’s another driver or someone else on the ground seeing back the move. Someone being crushed or struck if the rear driver hasn‘t got full view is fair, but it might surprise som people that propelling a train without anyone at the back is actually permitted in some circumstances.

 

[najaB]

Points moving under the train is a non issue in respect of who is controlling the move at the back. If they can move, then they can move whether it’s another driver or someone else on the ground seeing back the move.

That is true. However someone on the ground is more likely to see/hear something than a driver up in a cab, possibly over an engine running and limited view back down the line.

 

[Annetts key]

Two "worst case" scenarios that come to mind: a derailment if points move under the train. It would be low speed, but could still do quite a bit of damage.

With respect, it’s very unlikely that motor points or mechanical points controlled from a signal box that have electric lever locks and track circuits would move under a train. Even without a signal route being called, track circuits and axle counters directly lock and hence prevent such points from moving under a train.

However, there is still a risk that the points are not set to the correct position for the move. Which could result in a derailment or damage to the point operating equipment and the rails through the points.

Or the train could end up going somewhere it’s not supposed to go, such as into sidings or in the wrong direction at a junction.

A MOM may not be qualified as a shunter, but they should be trained as a route setting agent, and hence should be qualified to confirm that all points in the “route” have been set correctly. That the correct arrangements are in place at any level crossings, and that they have reminded the signaller about the dangers of any conflicting movements.

If another driver in their own train is involved, there is also a risk that they will be concentrating too much on the train that is “reversing” (setting back) and miss a signal or speed restriction sign and as a result have their own incident, either right their, or further along the line. Such as SPADing the next signal.

 

[berwicksfinest]

As some one who signs that exact location, there is a "double" banner repeater prior to this signal indicating what route is showing on main aspect.

Also East London Line "right" next to here has third rail, not the place for the MOM to be wandering about at the back of the train