Line pairings on a 4 track railway

[GTR fail]

I noticed the other day that whilst the 4 track sections of the WCML, GWML, MML, and most of the GEML are separated into fast and slow (i.e. DS, US, DF, UF), the ECML is paired by direction (DS, DF, UF, US). Why is this? And what are the pros and cons of this?

 

[gabrielhj07]

One advantage of having lines paired by running direction is that a train may cross from the slow to the fast (or vice-versa) without impeding traffic in the opposite direction.

 

[DerekC]

Having the slows on the outside (like ECML) makes intermediate stations where only stopping services call easier to build - you just need a platform each side of the line and changing tracks easier (as @gabrielhj07 says). Pairing by function (like WCML) is better when you get to a terminus because you can have your long distance/fast platforms separate from the short distance/slows, and the slows don't have to cross both fast tracks during turnback. Some routes have both, with a flyover/diveunder at the change approaching the terminus (for example SWML at Wimbledon and ECML just north of Copenhagen Tunnel on the approach to Kings Cross).

 

[swt_passenger]

I noticed the other day that whilst the 4 track sections of the WCML, GWML, MML, and most of the GEML are separated into fast and slow (i.e. DS, US, DF, UF), the ECML is paired by direction (DS, DF, UF, US). Why is this? And what are the pros and cons of this?

Adding to above replies it’s generally considered that pairing by use is best at a terminus, hence examples like the flyover at Wimbledon to arrange the paired by direction SWML into paired by use approaching London, or the Holloway flyover to sort the ECML approaches to Kings Cross.

Whether pairing by direction is better elsewhere depends a lot on where the branches are, ie which side of the main route they are on. It allows overtaking and weaving between fast and slow, and is probably the best solution where most branches are accessed by grade separation.

There was a similar discussion a few years back, also addresses the pros and cons regarding line closures for maintenance:

Paired by use or paired by direction

On the West Coast route, between London* and where the Northampton route diverges at Roade, there are 4 tracks. These are paired by use so from East to West you have up slow, down slow, up fast and down fast. North of Rugby it is paired by direction, so from east to west you have down slow...

www.railforums.co.uk

 

[gabrielhj07]

hence examples like the flyover at Wimbledon to arrange the paired by direction SWML into paired by use approaching London,

This brings up another benefit, which is that it allows cross-platform interchange from the Up Main Slow to the Down Main Fast at Clapham Junction.

 

[Magdalia]

I noticed the other day that whilst the 4 track sections of the WCML, GWML, MML, and most of the GEML are separated into fast and slow (i.e. DS, US, DF, UF), the ECML is paired by direction (DS, DF, UF, US). Why is this? And what are the pros and cons of this?

The GN end of the ECML is paired by direction because it has 2 track bottlenecks, and used to have a lot more of them. The key 2 track bottlenecks are now Welwyn Viaduct and Holme Fen. Since the 1950s others have been eliminated notably New Barnet-Potters Bar, Arlesey, Sandy and Huntingdon (last of these was 3 tracks).

But that doesn't explain York-Northallerton.

the Holloway flyover to sort the ECML approaches to Kings Cross.

That only applied after the 1970s remodelling and electrification. Prior to that Holloway flyover was only for goods to get into Kings Cross Yard.

 

[mr_jrt]

The other lesser-seen variation is having the fasts on the outside and the slows in the middle, so you only need a single island platform at smaller stations. This was used on the GCML and on the Met, though where the two ran together the GC "fast" lines were on one side. The Met's slow island stops are now the Jubilee line. The only unique problem with this system is that it means you need more land for the fast lines to splay out to bypass the stations without excessive curvature. It obviously shares the need for grade separation with the more common form of paired by direction.

 

[Taunton]

There was indeed a whole article on exactly this subject in an about 1960 edition of Modern Railways' predecessor, Trains Illustrated, with diagrams, written by none other than Gerry Fiennes himself, when he was ER Traffic Manager. In typical forthright Fiennes style he described the Kings Cross throat as "a mess". I wonder if he used it in a board proposal. Shame it took another 15 years to get done.

 

[Class15]

Caledonian Road and Barnsbury is another weird one, with freight tracks without platforms on the outside. I always think that the fast tracks are the freight ones as they don’t have platforms but now I think about it it’s really just a regular goods loop.

 

[FGW_DID]

I noticed the other day that whilst the 4 track sections of the WCML, GWML, MML, and most of the GEML are separated into fast and slow (i.e. DS, US, DF, UF), the ECML is paired by direction (DS, DF, UF, US). Why is this? And what are the pros and cons of this?

GWML uses ‘Reliefs’ & ‘Mains‘ not Slow & Fast - Up Relief, Down Relief, Up Main, Down Main

 

[CAF397]

One major pinch point is Slade Lane Jn a few miles south of Manchester.

Leaving Manchester Piccadilly lines are paired by speed Up Fast - Down Fast - Up Slow - Down Slow.

At Slade Lane Jn they change to being paired by direction Up Slow - Up Fast - Down Fast - Down Slow.

 

[swt_passenger]

The other lesser-seen variation is having the fasts on the outside and the slows in the middle, so you only need a single island platform at smaller stations. This was used on the GCML and on the Met, though where the two ran together the GC "fast" lines were on one side. The Met's slow island stops are now the Jubilee line. The only unique problem with this system is that it means you need more land for the fast lines to splay out to bypass the stations without excessive curvature. It obviously shares the need for grade separation with the more common form of paired by direction.

There’s a relatively short section of that “inside out” layout through Southampton and Millbrook of course. Allows the central island at Southampton to be used predominantly for terminating services inside the through lines. Millbrook is then an example of a central island for local stoppers.

 

[Taunton]

The main reason pretty much all 4-track arrangements arose is they just evolved from various widenings and such like, which steadily got joined together. This is not the New York Subway, who built many of their 4-track sections as such from scratch. Beyond the expansion of loops into long sections there have been various rearrangements over time as well. One significant change is the end of slow freight trains on the "slow" lines, and these progressively having capacity absorbed by more stopping passenger services.

The onetime Midland Railway provided considerable 4-tracking, but the slow lines were for goods trains only, and signalled to the lesser standard thus needed, with all passenger services on the main lines. It was much the style that each major railway did their own thing. How has the WCML ended up today paired by use south of Crewe and paired by direction north of Crewe? Just how each element developed.

 

[edwin_m]

But that doesn't explain York-Northallerton.

Probably something to do with the former flying junctions at Northallerton where the Ripon line came in, part of which provided grade separation for trains diverging towards Tees-side (if they didn't stop at the station). However the flat transition between down-up-down-up and down-down-up-up at Skelton Bridge (where only three tracks cross the river) is a notable bottleneck, and there have been discussions about grade separating this as well. The whole section could be made paired by use, with the slows to the west linking to the existing diveunder at Northallerton, but would involve re-building Thirsk and probably cause issues because there's quite a bit of switching between fasts and slows to handle the mix of fast trains, stopping trains and freights.

 

[AM9]

The main reason pretty much all 4-track arrangements arose is they just evolved from various widenings and such like, which steadily got joined together. This is not the New York Subway, who built many of their 4-track sections as such from scratch. Beyond the expansion of loops into long sections there have been various rearrangements over time as well. One significant change is the end of slow freight trains on the "slow" lines, and these progressively having capacity absorbed by more stopping passenger services.

The onetime Midland Railway provided considerable 4-tracking, but the slow lines were for goods trains only, and signalled to the lesser standard thus needed, with all passenger services on the main lines. It was much the style that each major railway did their own thing. How has the WCML ended up today paired by use south of Crewe and paired by direction north of Crewe? Just how each element developed.

It's been discussed here before, pairing by direction on the four tracking of the MML southern end would greatly ease the ducking and weaving that is needed to run fast Thameslink services in between IC paths, certainly as far as Luton. Particularly when the up EMR trains arrive out of their path window, the timetable starts to fall apart. Even removing the sluggish acceleration of the HSTs hasn't completely removed the problem.
Rearranging the formation including grade separated crossovers would be a major task and as @Bald Rick has said, it is very unlikely to happen.

 

[Taunton]

It's been discussed here before, pairing by direction on the four tracking of the MML southern end would greatly ease the ducking and weaving that is needed to run fast Thameslink services in between IC paths, certainly as far as Luton. Particularly when the up EMR trains arrive out of their path window, the timetable starts to fall apart. Even removing the sluggish acceleration of the HSTs hasn't completely removed the problem.

Indeed, upsides and downsides, doing the Midland 'by direction' would mean of course that those terminating at St Albans, Luton and Bedford, every Thameslink service, would need to shunt right across the four tracks of the layout to get from being Down to Up services, which would introduce its own difficulties, instead of sticking to their "own side" as at present.

 

[John Webb]

The main reason pretty much all 4-track arrangements arose is they just evolved from various widenings and such like, which steadily got joined together. This is not the New York Subway, who built many of their 4-track sections as such from scratch. Beyond the expansion of loops into long sections there have been various rearrangements over time as well. One significant change is the end of slow freight trains on the "slow" lines, and these progressively having capacity absorbed by more stopping passenger services.

The onetime Midland Railway provided considerable 4-tracking, but the slow lines were for goods trains only, and signalled to the lesser standard thus needed, with all passenger services on the main lines. It was much the style that each major railway did their own thing. How has the WCML ended up today paired by use south of Crewe and paired by direction north of Crewe? Just how each element developed.

Brief bit of history - when doing the 'London Extension' the Midland railway could buy the land (cheap) but not the steel rails for four tracks all at once. So as they came south from Bedford there were the pair of passenger lines all the way. They then started with an up goods line and gradually extended that southwards as getting loaded freight trains - particularly ones carrying coal as that was where they made money - towards London.
By the 1890s there was sufficient money to expand to four lines. For a while, until the opening of the second Elstree tunnel, the new lines were the Goods lines. After the tunnel opened in 1895, the goods lines became the slow lines from south of St Albans, the passenger lines became the fast lines.
In 1906 the goods lines were converted to passenger lines to just north of Harpenden and became part of the slow lines. It was only with the introduction of DMUs by BR that the goods lines north of Harpenden were converted to being the slow lines in 1959.

 

[AM9]

Indeed, upsides and downsides, doing the Midland 'by direction' would mean of course that those terminating at St Albans, Luton and Bedford, every Thameslink service, would need to shunt right across the four tracks of the layout to get from being Down to Up services, which would introduce its own difficulties, instead of sticking to their "own side" as at present.

That's if the current pattern of slows were adopted with a paired by direction formation. The track layout and service [pattern are mutually dependent.
Different termination stations might have produced a solution where either crossing the mainline was easier, - maybe Luton, or there was a 'Welwyn Garden City' arrangement where a single track diveunder/flyover gave the necessary separation.

Brief bit of history - when doing the 'London Extension' the Midland railway could buy the land (cheap) but not the steel rails for four tracks all at once. So as they came south from Bedford there were the pair of passenger lines all the way. They then started with an up goods line and gradually extended that southwards as getting loaded freight trains - particularly ones carrying coal as that was where they made money - towards London.
By the 1890s there was sufficient money to expand to four lines. For a while, until the opening of the second Elstree tunnel, the new lines were the Goods lines. After the tunnel opened in 1895, the goods lines became the slow lines from south of St Albans, the passenger lines became the fast lines.
In 1906 the goods lines were converted to passenger lines to just north of Harpenden and became part of the slow lines. It was only with the introduction of DMUs by BR that the goods lines north of Harpenden were converted to being the slow lines in 1959.

Thanks John. The GEML had similar issues between Liverpool St and Shenfield where Southend and some outer suburban services switched between the Mains (fasts) and Electrics (slows), and the GE majored on passenger traffic well before the end of the 19th century. Four tracks to Shenfield were finally completed in the '30s and electrification then brought the Ilford flyover, but the alternative could have been to switch the slows to the outside, - indeed there are bypass tracks at the flyover to allow f-s/s-f transfers but mainly for ECS purposes. The TfL takeover of the Electrics has rendered any change unnecessary now as the track pairs are run as two separate lines now from Bow to just before Shenfield.

 

[Benjwri]

I would say the configuration depends the intended use when the line was built, and space. If using a configuration where they are paired by speed, you have to build middle platforms, and a platform on at least one side, and therefore more separation between sides and space is needed. With pairing by direction only platforms on each side are needed, and so less separation and space. The lack of separation also makes approaches to stations easier to manage without a speed penalty being incurred by one pair of lines having to bend around a middle platform.

A system paired by direction is also better where the railway is forced to slim down to just two lines in areas, as each direction can merge without impeding opposing traffic.

 

[Bletchleyite]

The other advantage of paired by direction is that fast to slow interchange is always cross platform with two islands.

 

[AM9]

The other advantage of paired by direction is that fast to slow interchange is always cross platform with two islands.

... and in respect of recent times, there isn't the issue of incidents on the narmally unused platform facing one of the fast lines

 

[Horizon22]

The other advantage of paired by direction is that fast to slow interchange is always cross platform with two islands.

However there are certainly some disadvantages as well - line blocks can't easily be done in isolation if your slow lines are on the "outer" lines and your fast lines are the "inner" lines. You can't easily go down to a two-track railway or maybe you can on the fast lines, but a lot of passenger stops will be missed if there's no fast line platforms - parts of the SWML and Brighton Main Line (to London Bridge) come to mind here.

 

[jfollows]

One major pinch point is Slade Lane Jn a few miles south of Manchester.

Leaving Manchester Piccadilly lines are paired by speed Up Fast - Down Fast - Up Slow - Down Slow.

At Slade Lane Jn they change to being paired by direction Up Slow - Up Fast - Down Fast - Down Slow.

This was changed to the current arrangement in 1960, prior to then the lines continued paired by direction all the way to Manchester London Road, so DS/DF/UF/US from west to east just continued all the way to London Road station.
There was a more recent proposal, which was apparently taken seriously for a while, to alter the lines between Adswood Road (the southern end of the four-track railway between Stockport and Cheadle Hulme) and Slade Lane to be paired by use to match the changed Piccadilly-Slade Lane layout. This would have made stopping trains at Levenshulme and Heaton Chapel use the up fast which would have been silly, and would have messed up interchange at Stockport to the detriment of passengers. Fortunately the idea eventually got killed.
Attached are before/after diagrams from the 1960 changes from "Britain's New Railway, O.S.Nock, Ian Allan 1966”.
I always feel the problems caused by the current layout at Slade Lane Junction are overstated, it seems to work well generally and seems to work better today than it used to in the 1970s in my experience. Yes, it's a constraint, but it was done for a reason and it seems to deliver on that.

EDIT My guess, from a passenger perspective based on the relatively large number of times I used to be stopped on a train wanting to pass through Slade Lane Junction in the 1970s compared with today when I rarely find that I am stopped is that the operating improvement came when London Road signal box was replaced by Manchester Piccadilly PSB in October 1988. For whatever reasons, the more modern box seems better suited to operate the layout.

 

[swt_passenger]

The other advantage of paired by direction is that fast to slow interchange is always cross platform with two islands.

But it isn’t always laid out as two islands? There are numerous instances where there are two side platforms and a central island.

 

[Railwaysceptic]

Caledonian Road and Barnsbury is another weird one, with freight tracks without platforms on the outside. I always think that the fast tracks are the freight ones as they don’t have platforms but now I think about it it’s really just a regular goods loop.

It wasn't always that way. In the 1960s when I used to commute to Broad Street, the southern pair of tracks were used by the electric trains to and from Richmond and the northern pair were freight only west of Canonbury Junction. East of that junction, the northern pair of tracks were also used by rush hour trains via Finsbury Park to and from Broad Street.

 

[Ian Murray]

The real trick is to make sure that there isn't a single point of failure so that when one component fails, all 4 lines are not shut down. Paddington to Didcot is a great example of this with multiple examples like the single air-con unit in a relay room that stopped working causing the entire signalling system to shut down, or the unnoticed power spike that tripped all the electrical equipment in another relay room, 8 hours later power was lost, the signals in that area went red and again the trains stopped running. Note Paddington to Didcot had the redundancy removed by the DfT to save money.

 

[jfollows]

The real trick is to make sure that there isn't a single point of failure so that when one component fails, all 4 lines are not shut down.

Good point.
The lines into London Euston are paired by use until Camden Junction, where a flyover took the up fast over the rest of the formation and switched to paired by direction with the fast lines on the outside and the slow lines in the middle. Up fast trains then arrived alongside the up side carriage shed and signals EN65 (Up fast) and EN64 (Up slow) were the last signals before the platforms. In order to arrive in a high-numbered platform (12 or higher) there were then a number of flat crossings which tended to fail, and caused major problems. The area was significantly resignalled just before the turn of the century (from memory [EDIT work completed in October 2000 I believe after a little research] ) allowing up trains to high-numbered platforms to remain on the down side (west) of the formation, and the single points of failure were significantly eliminated.

EDIT "Then" and "now" diagrams attached, the former from my personal collection & the latter from "Railway Track Diagrams 4, Midlands & North West", 5th. edition TRACKmaps 2022 just recently published. Line "D" allows up trains to keep to the right of the formation now.

 

[Class15]

It wasn't always that way. In the 1960s when I used to commute to Broad Street, the southern pair of tracks were used by the electric trains to and from Richmond and the northern pair were freight only west of Canonbury Junction. East of that junction, the northern pair of tracks were also used by rush hour trains via Finsbury Park to and from Broad Street.

Yeah I know, in the old days the northern pair were freight only

 

[30907]

But it isn’t always laid out as two islands? There are numerous instances where there are two side platforms and a central island.

Quite - no doubt because typically the main station building was already there, to one side or the other of the double track, and a second platform on the other.

Plus it was generally easier to quadruple track on one side or other of the existing lines. The GNR through the Northern Heights was an exception, and so has twin islands (plus some LNER-era examples further north), but the longest section I can think of is Shortlands to Swanley (5 stations, but one is 1939 and one 1959 - and that changed from paired by use to paired by direction anyway!).

 

[Senex]

This was changed to the current arrangement in 1960, prior to then the lines continued paired by direction all the way to Manchester London Road, so DS/DF/UF/US from west to east just continued all the way to London Road station.
There was a more recent proposal, which was apparently taken seriously for a while, to alter the lines between Adswood Road (the southern end of the four-track railway between Stockport and Cheadle Hulme) and Slade Lane to be paired by use to match the changed Piccadilly-Slade Lane layout. This would have made stopping trains at Levenshulme and Heaton Chapel use the up fast which would have been silly, and would have messed up interchange at Stockport to the detriment of passengers. Fortunately the idea eventually got killed.

PicVic proposed a flyover here (as also at Edgeley No. 2), but we all know what happened to those plans.

As for those later curious plans for Slade Lane to Adswood Road, I never quite understood what the benefits were if not coupled with a completion of the quadding between Adswood Road and Cheadle Hulme to make the Stoke lines completely independent of the Crewe lines (given that IIRC there wasn't the regular M/c-Euston via Crewe service while all that was being planned).