Upgrading the Chiltern Line

[Zomboid]

Is that a bad thing? Are dynamic passing loops particularly difficult to operate on the network?

Just that platform loops for overtakes aren't compatible with how the modern railway is operated, and so it'd need reasonable lengths of slow lines in each direction, which means lots of reconstruction of bridges and widening of earthworks.

 

[squingo44]

Just that platform loops for overtakes aren't compatible with how the modern railway is operated, and so it'd need reasonable lengths of slow lines in each direction, which means lots of reconstruction of bridges and widening of earthworks.

Why is it incompatible? Don't networks in other countries do it?

 

[Zomboid]

Platform loops are capacity killers, as I understand it. Yes there are some, and some are even used, but they're not what would be designed now.

 

[Autspectorder]

Imho everywhere with enough clearance should be electrified.

 

[Virginfan123]

H

6 TPH from Princes Risborough to Elizabeth Line core, all stops
2 TPH from Oxford to Marylebone, all stops until Prince's Risborough
2 TPH from Banbury to Marylebone, all stops until Prince's Risborough
1 TPH from Stratford-upon-Avon to Marylebone, all stops until Banbury
1 TPH from Birmingham Snow Hill, all stops until Banbury

== Doublepost prevention - post automatically merged: 10 Jun 2026 ==

From a report on the learnings from Crossrail:

For an hourly service from Stratford, you would need to re-install the second track between Wilmcote and Hatton or drop the frequency of the Worcester-Birmingham-Dorridge-Stratford Train to 1p2h. As someone who lives in South Warwickshire, I just can't see Stratford gaining a train at the expense of Moor Street and especially Solihull + Dorridge, which have lots of commuters travelling to London.

Potentially, we could see services join together and Leamington Spa, e.g. a 2 car unit from Stratford and 4 car from Moor Street, but this adds alot of operational constraints especially on the quite busy through platforms at Leamington Spa which are quite limited in layout and operational flexibility.

 

[CyrusWuff]

West Ruislip has the advantage over South Ruislip of walking distance to the Metropolitan and Piccadilly at Ickenham. That isn't a massive advantage, given that developing one of the Sudbury stations more would be better for the Piccadilly line, and changing at Harrow-on-the-Hill or walking from Baker Street works for passengers between the Met and Marylebone.

West Ruislip also has a car park (owned by Chiltern), which South Ruislip doesn't, along with step-free access to the (reversible) Platform 4.

 

[zwk500]

Why is it incompatible?

Incompatible is perhaps not the right word. But Static passing loops are certainly much less efficient and leave the passengers on the slow train with significant additional journey time.

Platform loops are capacity killers, as I understand it.

Correct, because the loop itself will usually have a lower turnout speed than if it was just a plain line, there is a high risk of the 2nd train being 'checked' on the approach waiting for the 1st train to clear the line, and then the 1st train must dwell for a period of between 1.5-2x the planning headway to allow the 2nd train green signals, and the 1st train will then take time to clear the loop and accelerate.

On Chiltern you are looking at 5 minutes more dwell plus half a minute extra deceleration and acceleration.

It is also means you can't put anything between the 1st and 2nd train (before or after the loop) unless you have more loops or convenient junctions facing the correct way.

Dynamic loops only need to be 2-3 stations long, but do need higher speed junctions to maximise efficiency.

 

[squingo44]

Incompatible is perhaps not the right word. But Static passing loops are certainly much less efficient and leave the passengers on the slow train with significant additional journey time.

Correct, because the loop itself will usually have a lower turnout speed than if it was just a plain line, there is a high risk of the 2nd train being 'checked' on the approach waiting for the 1st train to clear the line, and then the 1st train must dwell for a period of between 1.5-2x the planning headway to allow the 2nd train green signals, and the 1st train will then take time to clear the loop and accelerate.

On Chiltern you are looking at 5 minutes more dwell plus half a minute extra deceleration and acceleration.

It is also means you can't put anything between the 1st and 2nd train (before or after the loop) unless you have more loops or convenient junctions facing the correct way.

Dynamic loops only need to be 2-3 stations long, but do need higher speed junctions to maximise efficiency.

Ok so what about putting the passing loop at the terminus of the slow services? Then each slow service can be made to wait for ~10 minutes or so anyway, giving plenty of time for the fast service to overtake.

And then a shoulder passing loop a station or two beforehand could be used to as a recovery tool in the case of late running - a situation in which the loss of time is less important, since the trains are already running behind schedule, and allows for recovery of the schedule.

== Doublepost prevention - post automatically merged: 11 Jun 2026 ==

For an hourly service from Stratford, you would need to re-install the second track between Wilmcote and Hatton or drop the frequency of the Worcester-Birmingham-Dorridge-Stratford Train to 1p2h. As someone who lives in South Warwickshire, I just can't see Stratford gaining a train at the expense of Moor Street and especially Solihull + Dorridge, which have lots of commuters travelling to London.

Potentially, we could see services join together and Leamington Spa, e.g. a 2 car unit from Stratford and 4 car from Moor Street, but this adds alot of operational constraints especially on the quite busy through platforms at Leamington Spa which are quite limited in layout and operational flexibility.

I see. I was simply using 1tph as my minimum unit of frequency, I didn’t realise that there’s already less frequency than that. I guess that’s another infrastructure intervention we can add to the list.

 

[The exile]

Platform loops are capacity killers, as I understand it. Yes there are some, and some are even used, but they're not what would be designed now.

They are for freight (which you basically want to keep moving) - not so sure about stoppers being overtaken by expresses (provided the stoppers’ stops are timed long enough). If it s “stopper” being overtaken by a limited stop train at one of the latter’s stops that’s precisely what you want anyway.

 

[The Planner]

They are for freight (which you basically want to keep moving) - not so sure about stoppers being overtaken by expresses (provided the stoppers’ stops are timed long enough). If it s “stopper” being overtaken by a limited stop train at one of the latter’s stops that’s precisely what you want anyway.

Yes, but you need to consider the journey time implications that has. Make the stoppers too long people either won't use it, or just drive to a station with a quicker service or just drive full stop.

 

[The exile]

Yes, but you need to consider the journey time implications that has. Make the stoppers too long people either won't use it, or just drive to a station with a quicker service or just drive full stop.

The point being that the only people who stay on the stoppers beyond one of those passing points are those who want one of the intermediates before the next one (ie the stopper is a succession of feeders to the semi-fast rather than a through service to the terminus). As you say, though, it comes with downsides and is not a universally transferable model.

 

[zwk500]

The point being that the only people who stay on the stoppers beyond one of those passing points are those who want one of the intermediates before the next one (ie the stopper is a succession of feeders to the semi-fast rather than a through service to the terminus). As you say, though, it comes with downsides and is not a universally transferable model.

This assumes that the express would stop at the looping point as well, which requires additional journey time penalties to the stations after the overtaking point not called at by the express.

Ok so what about putting the passing loop at the terminus of the slow services? Then each slow service can be made to wait for ~10 minutes or so anyway, giving plenty of time for the fast service to overtake.

Then it's not being overtaken, so you only need a bay platform. It also means you are much more likely to want to stop the Express there as well (to provide the onwards connectivity), so it changes the infrastructure requirements.

And then a shoulder passing loop a station or two beforehand could be used to as a recovery tool in the case of late running - a situation in which the loss of time is less important, since the trains are already running behind schedule, and allows for recovery of the schedule.

That's lovely if you have the room, but unfortunately (and regrettably, IMO) the way rail funding works is that 'nice to haves' are generally not going to be approved unless they are negligible to the overall cost, risk, and schedule.

 

[The exile]

This assumes that the express would stop at the looping point as well, which requires additional journey time penalties to the stations after the overtaking point not called at by the express.

That was precisely the scenario I originally described and the reason I wrote “semi-fast” not “express”.

 

[Zomboid]

The Chiltern main line doesn't have any stations that I can think of where there's more than 1 through platform in each direction, to provide overtake/interchange opportunities the relevant stations are likely to need to be flattened and rebuilt from scratch.

Something like a set of slow lines between Beaconsfield and Gerrard's Cross may well allow one or more fast trains to overtake a stopper, but we're talking big money and some significant station work for that.

 

[zwk500]

That was precisely the scenario I originally described and the reason I wrote “semi-fast” not “express”.

Apologies for having missed that, although my point still stands about the fact that you are further extending the journey times to provide the interchange.

Normal stop in stopping train costs c.2 minutes.
Overtaken by non-stop costs c.7/8 minutes (2 minutes + 5-6 minutes additional dwell)
Overtaken by train also stopping: c.10-12 minutes. (7-8 minutes extended by 3-4 minute penalty to fast train now stopping)

All of those times dependent on headway, line speeds, and relative speed differences of trains.

 

[The exile]

Apologies for having missed that, although my point still stands about the fact that you are further extending the journey times to provide the interchange.

Yes - but if the alternative is to halve the frequency of the stopper, most front door to front door journeys will still be quicker.

 

[squingo44]

How about terminating at High Wycombe, where there’s space for four tracks and platform faces, and running 2-4 semifasts that also stop at High Wycombe and 2-4 expresses that don’t stop there?

There ought to be space for 4-tracking between Beaconsfield and Seer Green and Jordan’s, even if there isn’t space to 4 track all the way between Beaconsfield and High Wycombe.

== Doublepost prevention - post automatically merged: 11 Jun 2026 ==

Apologies for having missed that, although my point still stands about the fact that you are further extending the journey times to provide the interchange.

Normal stop in stopping train costs c.2 minutes.
Overtaken by non-stop costs c.7/8 minutes (2 minutes + 5-6 minutes additional dwell)
Overtaken by train also stopping: c.10-12 minutes. (7-8 minutes extended by 3-4 minute penalty to fast train now stopping)

All of those times dependent on headway, line speeds, and relative speed differences of trains.

why is the additional dwell time so long? Why can’t the normal stop time be included into the overtaking time?

But tbh that additional dwell is worth it if it provides easy interchange between the stopper and the semi fast

 

[The Planner]

How about terminating at High Wycombe, where there’s space for four tracks and platform faces, and running 2-4 semifasts that also stop at High Wycombe and 2-4 expresses that don’t stop there?

There ought to be space for 4-tracking between Beaconsfield and Seer Green and Jordan’s, even if there isn’t space to 4 track all the way between Beaconsfield and High Wycombe.

You have to consider whether that is the right place for the 4 tracking based on the quantum and timetable you want, or you fix the timetable based on that 4 track section. Just because you can do it, doesn't mean its right.

== Doublepost prevention - post automatically merged: 11 Jun 2026 ==


why is the additional dwell time so long? Why can’t the normal stop time be included into the overtaking time?

But tbh that additional dwell is worth it if it provides easy interchange between the stopper and the semi fast

Based on a 3 minute planning headway, which the Chiltern line is up to Princes Risborough. Your first train arrives in the loop at xx.00, therefore a fast cannot pass until xx.03 (it may be xx.04 depending on how restrictive the loop is, for example if the turnout was at the end of the platform, the overlap will extend out onto the mainline meaning it needs to time out), the earliest you could get the stopper out is xx.05 depending on what the fast train is doing next. If its off into the distance and not stopping anywhere close you can go at a reduced headway as the fast train pulls away. However, if it was stopping somewhere close afterwards, chances are you are not departing that stopper until xx.06

 

[zwk500]

why is the additional dwell time so long? Why can’t the normal stop time be included into the overtaking time?

Normal stop: 1 minute stationary (1 minute for slowing down/speeding up:
Overtaken stop: 6 minutes stationary (2nd train is 3 behind, follow 3 minutes after, if a 4 minute headway you might get 4 before/2 after).

+5 minutes penalty.

But tbh that additional dwell is worth it if it provides easy interchange between the stopper and the semi fast

For an 8-10 minute penalty to be worth it will depend on having a very particular demand profile. You need the demand to the further stations to be proportionally much higher to justify the overtake, but you also need the demand to the intermediate station to be high enough to make the alternative option of simply running a different service pattern to not be acceptable, while not too high as to justify the larger investment in a dynamic loop.

If interchange to the faster service is going to be the predominant option taking then simply terminating the stopper at the semi-fast's stopping station is likely going to produce a more cost-effective infrastructure and resourcing requirement (and more robust service offer).

 

[squingo44]

You have to consider whether that is the right place for the 4 tracking based on the quantum and timetable you want, or you fix the timetable based on that 4 track section. Just because you can do it, doesn't mean its right.

If the loop is intended for recovery only, does that mean there’s flexibility to plan operations around the location of the passing loop rather than the usual determining the location of the loop from the timetable?

 

[Mgameing123]

The section from West Hampstead to Harrow where the lines are adjacent but separate is awkward but not that difficult.

The section between Harrow and Amersham is more difficult. You have 4 basic options:
- Install 25KV overhead above the DC electrification. Very, very expensive as mixing AC and DC return paths causes very quick corrosion.
- Get dual-voltage third rail units and bond the LU 4th rail as is done elsewhere. Requires changes to the Met to accomodate and potentially power supply upgrades but is a prove solution and OLE/3rd Rail dual-voltage units are common enough in the UK.
- Get dual-voltage 4th rail units - least impact to LU, may still require a power supply upgrade. May cause awkwardness for the train design as TPWS, AWS, and ETCS equipment is mounted centrally under the unit where the 4th rail pickup would be (but the pickup could be moved about to clear the space).
- Use battery trains from Amersham to Harrow. Means using more charge but allows a potentially simpler 25KV AC BEMU.

Initial impression would be to wire West Hampstead - Bicester North, Neasden South Jn - Harrow, Great Missenden - Aylesbury Vale Parkway, and Bicester South Jn to Oxford. If money stretches, include Princes Risborough to Aylesbury.
______

Perhaps relevant to this discussion is this RAIL Article: https://www.railmagazine.com/news/network-rail-issues-etcs-fitment-notice-for-turbostar-fleet

Which says NR have started the tender process for procuring ETCS Fitment to the Turbostar fleet, including the 168s.


Interestingly, the tender says:


Of course, fitting the 168s doesn't mean they're intended to stay with Chiltern that entire time. It could be to prep them for cascading to an ETCS area.

The other solution always discussed is extending the Metropolitan Line to Aylesbury.

 

[150219]

If the loop is intended for recovery only, does that mean there’s flexibility to plan operations around the location of the passing loop rather than the usual determining the location of the loop from the timetable?

High Wycombe no longer has space for 4 tracks since the platforms were extended and realigned. You'd need significant remodelling to make room to do so and keep the through line speed, which is itself already restrictive. The former subway (extant but disused) was also unable to accomodate track slewing without being removed/filled in.

 

[The Planner]

If the loop is intended for recovery only, does that mean there’s flexibility to plan operations around the location of the passing loop rather than the usual determining the location of the loop from the timetable?

Then that loop wouldn't be there, you don't build loops specifically for recovery, they are built to get used.

 

[cle]

I think from the Ruislips to Gerrards Cross as a quad section is probably more useful than ad hoc loops - to enable more dynamic passing. But the tunnel creates a station loop I guess... A bay there might see a stopper removed from contention entirely - although ideally the stoppers cover Beaconsfield and HW.

 

[squingo44]

Then that loop wouldn't be there, you don't build loops specifically for recovery, they are built to get used.

Isn’t that how dynamic loops work though? Aren’t they built to get used in the recovery case?

== Doublepost prevention - post automatically merged: 12 Jun 2026 ==

I think from the Ruislips to Gerrards Cross as a quad section is probably more useful than ad hoc loops - to enable more dynamic passing. But the tunnel creates a station loop I guess... A bay there might see a stopper removed from contention entirely - although ideally the stoppers cover Beaconsfield and HW.

we would just need enough to support running additional stopping and express services, right? Would 4 tracking the Ruislips to Gerard’s cross be enough? Assuming the stoppers run to High Wycombe

 

[zwk500]

Isn’t that how dynamic loops work though? Aren’t they built to get used in the recovery case?

No, they are built to allow you to time trains to be overtaken without suffering more than minimal delay to the journey of either train. The railway can no longer justify to its funders (HMT, nominally on behalf of a Mr. Tax Payer) spending money on things it does not intend to use on a regularly planned basis.

we would just need enough to support running additional stopping and express services, right? Would 4 tracking the Ruislips to Gerard’s cross be enough? Assuming the stoppers run to High Wycombe

It depends what your differential speeds and variations in stopping patterns were.

 

[The Planner]

Isn’t that how dynamic loops work though? Aren’t they built to get used in the recovery case?

== Doublepost prevention - post automatically merged: 12 Jun 2026 ==


we would just need enough to support running additional stopping and express services, right? Would 4 tracking the Ruislips to Gerard’s cross be enough? Assuming the stoppers run to High Wycombe

No, as @zwk500 has said, they are built to be used to deliver a timetable or level of capacity.

 

[Zomboid]

I suppose the question is what service enhancements are necessary on the Chiltern line. Would there be a meaningful benefit to extending Crossrail services to High Wycombe, and what infrastructure changes would that drive?

 

[squingo44]

It depends what your differential speeds and variations in stopping patterns were.

What stopping patterns would this infrastructure support? Crossrail taking over stopping services on the Chiltern Main Line from South Ruislip to Denham/Gerrards Cross but not all the way to High Wycombe?

== Doublepost prevention - post automatically merged: 12 Jun 2026 ==

I suppose the question is what service enhancements are necessary on the Chiltern line. Would there be a meaningful benefit to extending Crossrail services to High Wycombe, and what infrastructure changes would that drive?

Isn't the benefit of connecting an underserved commuter line to a through-running core with spare capacity fairly self-evident?

 

[Zomboid]

Isn't the benefit of connecting an underserved commuter line to a through-running core with spare capacity fairly self-evident?

Depends. There'd be no point only taking XR to West Ruislip, but how far out do the benefits go? What needs to happen to extend XR to Wycombe at 4tph, if that's our straw man? Would it be easier to chop the Ruislip branch of the Central line off and convert those tracks and stations beyond North Acton to XR? (Probably not!)