How long before all mechanical signalling is replaced on Network Rail

[Bald Rick]

Increasingly, with the latest signalling systems, “Resignalling” is becoming less popular. There are essentially three parts of the signalling system:

Control System
Interlocking
Ground equipment

These parts are joined by a transmission system (effectively a telecoms link).

New control systems now are, effectively, data run on something little different to a home computer.
New interlockings now are, effectively, data run on something little different to a home computer.
Ground equipment is broadly the same regardless of what the interlocking / control system is.

Each of these three components can be, and are, renewed independently depending on their condition / obsolescence.

The big cost in big resignalling schemes is all the design, data prep, testing, and so on, which uses very specialist skills (eg there’s only a couple of hungpdred signalling principles testers in the country). Testers in particular tend to get used at awkward times of the day / week / year - not many of them ever have a Christmas Day off for example, but my word they are compensated for that. Another big cost is all the cabling and cable routeing, however with modern IP based signalling the cabling this is reducing.

As @MarkyT says, big resignalling is often driven by major layout changes, and that means all the ground equipment needs to go whether it is life expired or not; often the same applies to the control equipment and interlocking, ie it is cheaper, easier and safer (see Waterloo) to build new than amend old.

So, in theory, headline major resignallings will reduce in number over the years as the relevant components are renewed when necessary, rather than doing the:all in one go.


Regarding mechincal / small boxes; they will, eventually, all go. I don’t have the numbers, but I reckon around 30-40% have gone in the last 10 years or so. Certainly in Anglia there have been nearly 40 small boxes closed in the last decade, including the Wherry Lines (some were just Level crossing gate boxes), and there’s now only about 30 left.

 

[MarkyT]

Increasingly....
New control systems now are, effectively, data run on something little different to a home computer.
New interlockings now are, effectively, data run on something little different to a home computer.
Ground equipment is broadly the same regardless of what the interlocking / control system is.
Each of these three components can be, and are, renewed independently depending on their condition / obsolescence.

A Good summary. I'd just like to reassure readers that the primary processor-based 'safety-critical' subsystems embedded in or closely connected to the trackside infrastructure such as interlocking and axle counters do not employ Microsoft operating systems! 'Safety-related' systems such as the workstation computers at control centres may use conventional commercial operating systems, although some experts recommend that specialist Unix family products are used where they can be supported. This is common ground with other safety related sectors, e.g. Military, Emergency Services dispatch, Medical machinery.

 

[Shaw S Hunter]

I remember seeing a document a few years ago that listed all re-signalling plans with projected dates. If I remember correctly, there were a lot of 2050 dates, which was intended to mean 'some time *after* 2050'. So there is (probably) currently no year by which mechanical signalling is even provisionally planned to be gone by.

I remember seeing such a list too. IIRC the last "big" jobs were Worcester and Shrewsbury, the latter somewhat ironic given the nearby Cambrian ETCS. With so many resignalling schemes having been deferred in recent times I suspect @The Planner's forecast of 25 years to the end of mechanical signalling will prove to be optimistic by at least a decade though I am unlikely to be around long enough to verify that claim!

 

[BRblue]

Just a couple of observations...
Not all mechanical boxes are absolute block, many are now track circuit block and can handle the same number of trains as a more modern installation.
Having worked an 80's era panel previously, I now work an 1880's era track circuited mechanical box, I can honestly say I have far fewer problems now with the old kit than I did with the newer electronic 80's panel with a similar number of trains/movements.
Whilst it is true that locking fitters are as rare as a white rhino these days, I did hear a rumour that a few apprentice locking fitters had been taken on not so long ago.
I'm not stuck in 1955 or a nostalgic enthusiast... I use mechanical signalling every day with a little bit of assistance from electric point motors and a couple of lights on sticks/gpl's, and it is invariably the electronics that cause a problem not the old mechanical kit.
What's the betting after saying all this, that it all goes pear shaped on my next turn!

 

[Ken H]

You have no idea what you are talking about. Mechanical signalling cannot he repaired or maintained by "any old fred" whatever that means. It requires some highly skilled and rapidly aging blokes to keep things going.

It also requires numerous one off pieces of equipment manufacturing at a very high cost and at long lead times. That is if anyone will make the parts at all.

Your views in modern signalling are bunkum. It isn't a finite capacity with modern signalling. It is a rapidly changing technology that leads to change.

'Old Fred' is a Roger Ford-ism. Its how somewhere is kept going by a bloke who knows how to do something just right but its never documented. The he retires and the job goes to bits. Its how he explained the fact that new EE507 traction motors suddenly became troublesome.

 

[The dogbox]

As a driver, as nice as absolute block looks in a photograph, it’s a pain in the rear to drive under. Less capacity, having to count signals and semaphores can be bloody hard to find in fog, bright sunlight, dusk, and at night in a built up area with light pollution. Add in some of the more complicated bits of the rule book (passing signals at danger under your own authority anyone?) and I will not be sad to see them go! Just my thoughts though, lots of my colleagues love it.

 

[DarloRich]

'Old Fred' is a Roger Ford-ism. Its how somewhere is kept going by a bloke who knows how to do something just right but its never documented. The he retires and the job goes to bits. Its how he explained the fact that new EE507 traction motors suddenly became troublesome.

It is all recorded! The problem is persuading young people to sign up to train in a really specialised skill is very hard. There is no apparent longevity and a chance that an apprentice today could find themselves without the skills needed to propser in the modern railway

 

[Ken H]

It is all recorded! The problem is persuading young people to sign up to train in a really specialised skill is very hard. There is no apparent longevity and a chance that an apprentice today could find themselves without the skills needed to propser in the modern railway

Spose its same as IT. Masses of legacy systems with old farts maintaining them. Where I work its code that was written in the mid 1990s - Some even written in COBOL, maintained by an ageing workforce. What to do when we retire is worrying management right now. (I am over 60 BTW). Some well known companies affected.

 

[DarloRich]

Spose its same as IT. Masses of legacy systems with old farts maintaining them. Where I work its code that was written in the mid 1990s - Some even written in COBOL, maintained by an ageing workforce. What to do when we retire is worrying management right now. (I am over 60 BTW). Some well known companies affected.

Actually that is a really good comparison. The key is timely upgrade and on going refresh and service support but that often gets missed off the purchase spec.

( I also do IT btw)

 

[Llanigraham]

mechanical signalling is very long lasting, and is easily fixed and kept going. Parts can be manufactured by any metalworking shop.
Newer signalling doesnt seem to be so easily kept going. Not least because the bits get very difficult to source.

The "bigger" bits I'd agree with you, but when one of our block instruments went wrong there were no spare parts available and I understand they ended up getting parts from the Severn Valley.

 

[Llanigraham]

Nope

But some mechanical boxes are very very old. I bet the gear in them is like my brush - that has had 5 new heads and 4 new handles.

Is the problem the mechanical bits, or the electrical stuff?

Both, especially where the mechanical bits are also switching electrical bits!

As I mentioned above, perhaps looking inside a block instrument at the complexity of the brass and bakelite turnings, cams and switches would show that it is very much NOT a simple instrument, and the maintenance and adjustment of them couldn't be learnt from the back of a fag packet.
And perhaps then look at the complexity of the below floor locking frame and the micrometer adjustment that is sometimes needed to take up the wear of pulling levers.

 

[Meerkat]

How complicated is a block instrument in terms of making a plug in electronic replacement -ie more expensive development but then maintainable by modern skills?

 

[Tomnick]

How complicated is a block instrument in terms of making a plug in electronic replacement -ie more expensive development but then maintainable by modern skills?

Not very complicated, I’d imagine - there are already numerous examples of block instruments incorporated into panels, with lights instead of a needle to indicate the state of the sections, and sometimes with buttons (or else a modern rotary switch) in place of the commutator.

 

[Ken H]

How complicated is a block instrument in terms of making a plug in electronic replacement -ie more expensive development but then maintainable by modern skills?

cant be beyond the wit of man to make a program that runs on a pc to simulate a set of block instruments.
But surely there is a shedload of these things taken from closed signalboxes that can be reused.

 

[edwin_m]

How complicated is a block instrument in terms of making a plug in electronic replacement -ie more expensive development but then maintainable by modern skills?

cant be beyond the wit of man to make a program that runs on a pc to simulate a set of block instruments.
But surely there is a shedload of these things taken from closed signalboxes that can be reused.

In terms of making one that works, pretty easy. In terms of demonstrating that it will never fail in an unsafe manner, much less so.

In much the same vein however, Park Signalling has recently produced a modern version of the classic Tyers token machine, communicating via a data stream over the internet rather than being hardwired. There's an article in a recent Rail Engineer magazine but unfortunately it doesn't seem to be on their website.

 

[MarkyT]

In terms of making one that works, pretty easy. In terms of demonstrating that it will never fail in an unsafe manner, much less so.

In much the same vein however, Park Signalling has recently produced a modern version of the classic Tyers token machine, communicating via a data stream over the internet rather than being hardwired. There's an article in a recent Rail Engineer magazine but unfortunately it doesn't seem to be on their website.

Yes, I saw that at the Swindon Panel Society opening day last year at Didcot Railway Centre. The red painted units are similar in size, but very minimalist in appearance compared to the originals, and the company have faithfully reproduced the curved top to the casing of the original, perhaps so you can't place a cup of tea on them!

 

[underbank]

But if you have an existing signalling system that works and is easily maintained, and there's no compelling reason to change it, why change. It's all well and good saving the cost of a signalman, but only if the savings aren't wiped out by resignalling every forty years.

But it's not just the cost of a signalman. There's also the disruption and inconvenience (and cost) when that signalman fails to turn up for work, calls in sick, or goes sick when on duty. It's not that long ago that the Furness line was out of action for a couple of hours when a signalman didn't turn up, and on the Bentham line when a signalman went off sick during a shift. These were remote places, so getting a replacement took time in terms of travel etc.

You also need other staff to cover for holidays, breaks, shifts etc. And cost of a worker is far more than their wage - there's training, management, employers NIC, pension costs, uniform, etc etc.

Having central control from regional control centres means that it's quicker, easier and cheaper to have cover in place for unforeseen eventualities and you need fewer staff.

 

[Ken H]

But it's not just the cost of a signalman. There's also the disruption and inconvenience (and cost) when that signalman fails to turn up for work, calls in sick, or goes sick when on duty. It's not that long ago that the Furness line was out of action for a couple of hours when a signalman didn't turn up, and on the Bentham line when a signalman went off sick during a shift. These were remote places, so getting a replacement took time in terms of travel etc.

You also need other staff to cover for holidays, breaks, shifts etc. And cost of a worker is far more than their wage - there's training, management, employers NIC, pension costs, uniform, etc etc.

Having central control from regional control centres means that it's quicker, easier and cheaper to have cover in place for unforeseen eventualities and you need fewer staff.

The signalling costs fall on network rail. The reputational damage falls on the TOC.
Thats the problem with separating infrastructure from the retailer. Same as Openreach and ISP's

 

[edwin_m]

But it's not just the cost of a signalman. There's also the disruption and inconvenience (and cost) when that signalman fails to turn up for work, calls in sick, or goes sick when on duty. It's not that long ago that the Furness line was out of action for a couple of hours when a signalman didn't turn up, and on the Bentham line when a signalman went off sick during a shift. These were remote places, so getting a replacement took time in terms of travel etc.

You also need other staff to cover for holidays, breaks, shifts etc. And cost of a worker is far more than their wage - there's training, management, employers NIC, pension costs, uniform, etc etc.

Having central control from regional control centres means that it's quicker, easier and cheaper to have cover in place for unforeseen eventualities and you need fewer staff.

Is there any other industry that permits 12-hour shifts working alone at an isolated location with no formal breaks?

 

[Tio Terry]

Is there any other industry that permits 12-hour shifts working alone at an isolated location with no formal breaks?

Fairly common in the Security Industry.

 

[The Planner]

Millom. 2 shifts, 45 movements yesterday.

Add up all the signallers and crossing keepers required for Arnside, Grange over Sands, Ulverston, Dalton Jn, Barrow, Park South, Askam, Foxfield, Skelly Cragge L.C., Millom LC, Kirkstanton L.C., Limestone Hall LC, Silecroft, Bootle, Saltcoats LC, Drigg Station, Sellafield, St Bees, Whitehaven Bransty, Workington No2, Workington No3, Maryport and Wigton. Someone somewhere is doing the maths for that already believe me.

 

[Ken H]

Add up all the signallers and crossing keepers required for Arnside, Grange over Sands, Ulverston, Dalton Jn, Barrow, Park South, Askam, Foxfield, Skelly Cragge L.C., Millom LC, Kirkstanton L.C., Limestone Hall LC, Silecroft, Bootle, Saltcoats LC, Drigg Station, Sellafield, St Bees, Whitehaven Bransty, Workington No2, Workington No3, Maryport and Wigton. Someone somewhere is doing the maths for that already believe me.

I am sure the costs are very high (dont know if any boxes are switched out for some shifts)
But is the cash and signalling resources there to fix this?
Could some costs be avoided by closing some boxes and replacing with intermediate block. or installing automated crossings, like the one at Bingham (Between Nottingham and Grantham)

 

[Bald Rick]

Add up all the signallers and crossing keepers required for Arnside, Grange over Sands, Ulverston, Dalton Jn, Barrow, Park South, Askam, Foxfield, Skelly Cragge L.C., Millom LC, Kirkstanton L.C., Limestone Hall LC, Silecroft, Bootle, Saltcoats LC, Drigg Station, Sellafield, St Bees, Whitehaven Bransty, Workington No2, Workington No3, Maryport and Wigton. Someone somewhere is doing the maths for that already believe me.

The maths will have been done a long while ago. The average 3 shift “single manned” box will cost the best part of half a million a year to operate when all costs are included.

 

[Ken H]

The maths will have been done a long while ago. The average 3 shift “single manned” box will cost the best part of half a million a year to operate when all costs are included.

Millom only sees trains between just before 0600 and just before 22:00. So 2 shifts will do it, with a bit of overtime.

And are any boxes switched out at all?

 

[LowLevel]

Replacing lumps of cast iron is rarely the major problem with mechanical signalling. They can break or fail but they can also be replaced (many of the signals that people get so worked up about are actually products of the 80s or newer) to one degree or another with little difficulty but varying expense. New semaphores are still being fabricated to replace life expired structures or fittings.

Wiring is your big killer and that also goes for the older larger PSB setups (that or early modems etc for 1980s systems being life expired with parts no longer available).

Wire degradation is rife in some of these installations and the cost of having them out of action while you totally rewire them often means it's more cost effective to get rid and replace them.

I've visited a fair number of signalboxes of varying types and more than a few have a total ban on touching the wiring at all because it will quite literally fall to bits.

One of the Stafford boxes for example had quite literally had part of a relay room roof collapse.

 

[trebor79]

The big cost in big resignalling schemes is all the design, data prep, testing, and so on, which uses very specialist skills (eg there’s only a couple of hungpdred signalling principles testers in the country). Testers in particular tend to get used at awkward times of the day / week / year - not many of them ever have a Christmas Day off for example, but my word they are compensated for that.

How does one train and find employment as a tester?
Could someone from outside the industry be seen as employable for such a role?

 

[hwl]

The big driver for any mass wave of change will presumably be ETCS due to reduced lineside infrastructure cost when:
1. Enough Staff are used to installing ETCS so it isn't novel but routine
2. ETCS only installation happens
3. The stock on route is already ETCS fitted or significant number of the class have already had it fitted (not just the first in class programme) or the stock is ETCS fitted or easily retrofittable from new.

So stating in about 10 years time when most of the sprinters are due for retirement?

 

[Bald Rick]

How does one train and find employment as a tester?
Could someone from outside the industry be seen as employable for such a role?

Well most of those I know have degrees in electrical / electronic engineering, then it’s several years on the job training. I suspect there is a route in to the profession without a full degree, but it will involve several years training. Start with a few of the companies that do it, eg Atkins, Siemens.

 

[Tio Terry]

How does one train and find employment as a tester?
Could someone from outside the industry be seen as employable for such a role?

Have a look at the IRSE website, it has details of the training necessary to reach the various grades. To reach the top will take a number of years.

 

[GrimShady]

Is there any other industry that permits 12-hour shifts working alone at an isolated location with no formal breaks?

Maritime.