Points failure...

[najaB]

A question for the infrastructure guys - when there are delays caused by points failure, what is the most common actual problem? Physical failure of the switch, the points motor deciding it is a modernist sculpture, facing locks not locking or a detection problem? Just curious how they seem to be able to clear them so quickly? (Not that I'm in any way complaining!)

 

[EM2]

In my experience (a long time ago) it was generally detection, often caused by debris between the switch and stock, or the 'throw' of the drive needing adjustment. Duff microswitches are also a pretty common cause.

 

[headshot119]

A question for the infrastructure guys - when there are delays caused by points failure, what is the most common actual problem? Physical failure of the switch, the points motor deciding it is a modernist sculpture, facing locks not locking or a detection problem? Just curious how they seem to be able to clear them so quickly? (Not that I'm in any way complaining!)

I can't give you an actual answer as I don't know. But I can give some suggestions as to why they can clear quickly.

Debris in the point blades, might get moved sufficiently after a train passes to clear the fault.

Axle counter failure causing points to be locked one way or the other, can be reset after the line has been examined.

 

[Tio Terry]

A question for the infrastructure guys - when there are delays caused by points failure, what is the most common actual problem? Physical failure of the switch, the points motor deciding it is a modernist sculpture, facing locks not locking or a detection problem? Just curious how they seem to be able to clear them so quickly? (Not that I'm in any way complaining!)

Without doubt, as EM2 says, something in the points that prevents detection being achieved.

 

[carriageline]

Low or lack of oil in clamplocks, a tight lock (hopefully one of the S&T bods can explain that please!) ballast or debris caught in them, voiding under the points can cause detection flicking, the point. motors can die too, clutch failures etc.

Sometimes you can get detection after moving them on the IPS (individual point switch) a couple of times, and I generally leave them in that position once I have got detection. Last thing you want to do is move them again, and lose it all.

 

[Bald Rick]

More often than a fault in the detection elements. Usually by being out of adjustment, which is caused by repeated heavy traffic / use, or ground movement, or big changes in temperature. Tolerances can be as low as 2mm. Obstructions are maybe 30% of the detection issues, and it is almost always ballast, litter, or a broken bolt.

Other common faults are loss of hydraulic oil in clamplocks - almost always due to a hose that has failed, often due to damage rather than age; seized lock bodies (the bit that drives the locking element of clamplocks) - usually due to chemical issues causing rust. Rarely used points will sometimes fail just because they are a little bit rusty or need lubrication.

That covers about three quarters of failures. Most of the rest is power related, and a very few are a catastrophic failure of point motors themselves.

 

[MarkyT]

. . . a tight lock (hopefully one of the S&T bods can explain that please!)

That's where the lock mechanism isn't engaging properly or easily, usually because something else has moved, broken or gone out of adjustment. The lock mechanism is proved fully engaged in the detection. Point machines have an internal bolt that engages one of the two slots in a moving drive bar that correspond to the correct switch rail positions. Clamp lock points have a rising hook-like clamp mechanism that clamps the closed switch rail directly to the appropriate stock rail. In order for detection to be successful, the clamp must also be fully engaged. Almost all point failures first manifest themselves as a detection failure either in the position they are lying in already or in the new position when an attempt to move them is made. Points in the direct line of route and other points providing flank protection must all be detected in correspondence before a signal will show a proceed aspect.

 

[Yabbadabba]

At this time of year it will often be a piece of ballast being jet washed into the blades from passing water jets of the RHTTs, even though they are not supposed to spray through points.

 

[TheEdge]

Its detection. Its always detection!

 

[SpacePhoenix]

When points do fail, will NR generally manually clamp and lock them to one direction and then fix them overnight?

 

[High Dyke]

When points do fail, will NR generally manually clamp and lock them to one direction and then fix them overnight?

That depends on the location of the points and any traffic that may require their movement. I've been involved in a points failure situation where all bar one train could be diverted by an unaffected route. There was a reluctance for maintenance teams to attend that day until it was pointed out that one train could not be diverted and needed the affected route. Hey presto! Response staff turned up within an hour...

There have also been cases where points that have been clamped and padlocked in the required position have still lost detection.

 

[TheEdge]

When points do fail, will NR generally manually clamp and lock them to one direction and then fix them overnight?

When there is a failure around Stratford (London) that does generally seem to be the done thing. I guess because taking a block there during service is a damn sight more disruptive than just locking one route out of use.

 

[alxndr]

Insufficient lubrication is a common one around here, although I do suspect there's a bigger underlying cause too.

I don't think I've ever seen our points team work a night other than for a half switch change, we're fortunate enough to be able to get out and resolve things in between trains. If that's not possible it does tend to be clipped and scotched (if required) to wait until later in the day/weekend, occasionally with someone going out to wind them over if absolutely necessary.

 

[talldave]

Could rarely used points be regularly cycled to prevent seizing (like most domestic heating pumps are nowadays)?

Has any work ever been done to predict failures, based on usage, age, past reliability, environmental criteria, etc? Or is it just a case of fire-fighting?

 

[alxndr]

Could rarely used points be regularly cycled to prevent seizing (like most domestic heating pumps are nowadays)?

Even points rarely used operationally will be maintained and Facing Point Lock tested on a regular basis, after which they should be swung over and back, on power, at least once, as well as being manually wound during the testing. So in effect they are "cycled", even though it's not the aim of FPL testing.

 

[Bald Rick]

Has any work ever been done to predict failures, based on usage, age, past reliability, environmental criteria, etc? Or is it just a case of fire-fighting?

Most points on busy routes have remote monitoring fitted, which measures point swing time, pump pressures, voltages and temperature, amongst other things. Trends can be examined over time, and alarms set when a certain metric goes above (or below) a set limit. These alarms are sent to control and/or the relevant technician who can take action before failure. But it doesn't pick up everything in advance, particularly any catastrophic failure or obstructions.

 

[najaB]

Thanks to all who have contributed - so much to learn.

 

[QueensCurve]

I can't give you an actual answer as I don't know. But I can give some suggestions as to why they can clear quickly.

Debris in the point blades, might get moved sufficiently after a train passes to clear the fault.

Axle counter failure causing points to be locked one way or the other, can be reset after the line has been examined.

Is the need for the line to be examined a reason why signal failures seem to cause more delay in the axle counter age than they did when the British Railway was reliant on track circuits?

 

[MarkyT]

Is the need for the line to be examined a reason why signal failures seem to cause so much delay in the axle counter age than they did when the British Railway was reliant on track circuits?

Difficult to say conclusively. Some track circuit failures are intermittent and appear to fix themselves (those that are caused by failing track pad insulation or contaminated ballast), but many TC failures do not. Axle counter mis-counts and other failures never fix themselves, however. Although axle counters are considered more reliable generally in the first place (once new system bugs have been ironed out - the 'bathtub curve' phenomenon), when miscounts occur, they are usually more disruptive. There is a particular known problem where wheels stop over a split section sensor midway through a platform. In some recent schemes, axle counters have been deliberately avoided in those specific locations where split platform sections are needed for permissive splitting and joining and other platfrom sharing purposes, whilst elsewhere in those schemes axle counters have been provided throughout.

 

[Tomnick]

The line needs to be examined for a suspected track circuit failure anyway, so that makes no difference. If anything, the fact that axle counters can often then be restored remotely significantly reduces the subsequent delay.

 

[alxndr]

I'd say the line needs to be examined more for a track circuit failure than an axle counter failure. Faulting on axle counters only really concerns the heads, evaluator, and transmission, whereas there's the potential for a problem to be anywhere within the track of the entire section during a track circuit failure.

My understanding is that a "sweep train" needs to pass through an axle counter section to reset it, which could be the only time anyone looks at the line in the middle of the section. For a track circuit it's quite likely that the length of the section will be physically walked so that cables, bonding, IRJs and insulation, etc can be examined.

 

[najaB]

My understanding is that a "sweep train" needs to pass through an axle counter section to reset it, which could be the only time anyone looks at the line in the middle of the section. For a track circuit it's quite likely that the length of the section will be physically walked so that cables, bonding, IRJs and insulation, etc can be examined.

That makes sense - track circuit failure could occur anywhere along the length of the section, axle counter is pretty much just at fixed locations.

 

[QueensCurve]

I'd say the line needs to be examined more for a track circuit failure than an axle counter failure. Faulting on axle counters only really concerns the heads, evaluator, and transmission, whereas there's the potential for a problem to be anywhere within the track of the entire section during a track circuit failure.

My understanding is that a "sweep train" needs to pass through an axle counter section to reset it, which could be the only time anyone looks at the line in the middle of the section. For a track circuit it's quite likely that the length of the section will be physically walked so that cables, bonding, IRJs and insulation, etc can be examined.

Thank you all for your comprehensive responses.

Going back to the 1980s it seemed to be common for a signal to fail (possibly the track circuit) and to be cautioned through with minimal delay.

Nowadays when signal failures are reported by nationalrail.co.uk they seem to involve more distruption.

Any feeling for why this should be so?

 

[Tomnick]

A simple track circuit failure is unlikely to cause that much disruption. Usually it is relatively minor delays (though, on a busy line, they can soon add up), with the first one taking a little bit longer because it's examining. The more disruptive failures reported by National Rail will usually be complete loss of signalling, multiple TCFs or occasionally a particularly awkward TCF (holding a route across a busy junction, for example). I'm sure that it still happened in the 80s, but back then it wouldn't be so widely publicised - and the response on the ground would almost certainly be quicker too.

 

[TheNewNo2]

Over roughly five years of failures that are coded as "points failures" for London Underground Jubilee, Northern and Piccadilly lines, the primary causes are "Indication" (35%), "Locks" (25%) and "Mechanical Linkages" (14%).

However, I'm no engineer, so I can't tell you what any of that actually means.

 

[edwin_m]

I was part of the commissioning team for York IECC and on Day 1* we were there when there was some major disruption due to loss of detection on the points where the Harrogate line diverges at Skelton. Technician went to site and found that there was "an item of ladies' underwear" trapped in the points and the inability to shift them had caused a hydraulic hose to burst. These are high speed points and took some time to bar over by hand so they could be clipped in one position until repared.

*Not of York station itself, this was one of the stages of extending it to cover the wider area.

 

[najaB]

Technician went to site and found that there was "an item of ladies' underwear" trapped in the points and the inability to shift them had caused a hydraulic hose to burst. These are high speed points and took some time to bar over by hand so they could be clipped in one position until repared.

Well that was a bit pants!

 

[yorksrob]

I was part of the commissioning team for York IECC and on Day 1* we were there when there was some major disruption due to loss of detection on the points where the Harrogate line diverges at Skelton. Technician went to site and found that there was "an item of ladies' underwear" trapped in the points and the inability to shift them had caused a hydraulic hose to burst. These are high speed points and took some time to bar over by hand so they could be clipped in one position until repared.

Well, Skelton Junction has never struck me as a particularly romantic location, but I guess, each to his/her own !

 

[causton]

Flushed from a toilet onto the track I guess? Eurgh! :P

 

[talldave]

Thanks to all who have contributed - so much to learn.

I'd like to second that. Appreciate the time taken to post informative responses.