Why is Thameslink central section linespeed so slow?

[JN114]

Utterly delusional.

Things go wrong. Recovery means that a robust timetable can exist in a world where outside influences mean the railway doesn’t run perfectly - Passengers can plan journeys they actually have a chance of completing as advertised.

They shouldn’t overdo it, that sentiment I’ll agree to, but you’re in cloud cuckoo land if you think a very busy timetable will work without recovery time.

 

[tsr]

I think it's pretty clear - but in case you're genuinely unsure and not a disingenuous apologist (and I apologize if the former): Timetable according to optimized journey times, and honest arrival times and dwells. Less 'recovery' which is BS for padding, framed as operational and reliability driven, but is a mechanism against delay penalties first and foremost.

How about aiming for less delays overall, and tougher standard to aspire to - aka eating less rather than loosening a belt - same as induced demand for motorway lanes. Pad and that space will be comfortably filled by laxer ops.

Padding or recovery, as you call it, often refers to timing allowances or running times which are basically contingency plans... such as giving trains more time for speed restrictions or for the typical time to get everyone on and off at a busy station.

On a network (as a national whole) with so many different types of services and flat junctions and differential speed limits and all the rest of it, you need these things to effectively synchronise the service.

There are some schedules where trains need these allowances almost every single day. There are others where they only need them occasionally, but if they weren't in place, a slight delay could cause massive knock-on effects if they were not recovered from in good time. Some of the biggest delay tallies I see on a daily basis, excluding major incidents, stem from a “high peak” train taking one or two minutes too long in a signal section or at a station. Good contingency measures in the timetable can often reduce the impact of this relatively easily, and it’s the sensible thing to do when you consider how many passengers will otherwise lose minutes from their day in reaction.

Dispatch procedures and professional driving policies do not become more lax if more time is available. In fact you can often wait at a station on a good day for a couple of minutes, whilst doing nothing at all, because dispatch will take just as long whether you have arrived three minutes early or 30 seconds. The dispatch procedure will not be made more relaxed and slower by arriving early. Everyone sticks to the same procedures and rules whether they often arrive early, on time or late on a particular schedule. You don’t rush and you don’t dawdle.

 

[AM9]

I think it's pretty clear - but in case you're genuinely unsure and not a disingenuous apologist (and I apologize if the former): Timetable according to optimized journey times, and honest arrival times and dwells. Less 'recovery' which is BS for padding, framed as operational and reliability driven, but is a mechanism against delay penalties first and foremost.

How about aiming for less delays overall, and tougher standard to aspire to - aka eating less rather than loosening a belt - same as induced demand for motorway lanes. Pad and that space will be comfortably filled by laxer ops.

By describing somebody as a 'disingenuous apologist' without knowing anything about me, you have rather set yourself up as not wanting to consider the issues of operating a railway under normal constraints. Before I continue, you must understand that I have no relationaship with any part of the railway (NR,TOCs, RoSCos, DfT) other than as a fare paying passenger and via this forum, a technical/aesthetic interest in rail operations. I don't know what your involvement with railways is so I will assume that you have similar knowledge to me.
I see that TSR has covered most of the issues that I was going to mention, so I sum it up as: it's naïve to forget the old adage, 'more haste less speed' which applies just as much to running a railway as making a car journey. The paying passenger is happier with a service that runs to time 90% of the time compared with one that is supposed to be 5% faster but half of the time is 20% slower.

 

[cle]

I did say if... so were you not teeing me up then?

To the other two posters, I understand the operational reasons why it is the way it is. I'm not questioning what it exists for, but challenging the fact that we accept it rather than making the investments necessary to negate the need for such padding in future. Look at historical journey times, even on routes which have not had frequency explosions - and they have all crept in line with privatization. This is a more crowded railway, but it is naive to say that it isn't to circumvent penalty, now it is commercial venture.

A little contingency, sure, as human, technical and other factors for delay can be unforeseen - but these should be the exceptions, whereas we are building in the leeway into making them the rule.

 

[big all]

also worth noting you have 4 tracks or more worth off trains onto two track and back to 4 tracks or more so closer headway gives greater capacity

 

[Bald Rick]

Please explain - this seems counter to logic. The higher the speed the train can accelerate to as it moves off the platform, the sooner it clears the platform block and the sooner the next arriving train can enter. There is of course a plain line headway calculation to go with this, but at present we are talking about platform reoccupation (run in-run out) time.

The stations in the core have very small signalling sections - some of them as short as 150 metres with the signals, less than half that under ETCS / ATO. Braking distance at 30mph is less than 100metres (on level track). Because of the speed the overlaps can be short, which means that trains can get much closer to the train in front. Fro example, southbound at Blackfriars their can be a train waiting almost in the platform whilst one is actually in the platform. With a higher linespeed, this would not be possible, as the signalling overlaps (clear section beyond a signal held at danger) must be longer. With ETCS/ATO in full use, a train accelerating away from a core station will clear the rearmost signalling section at about the same time it reaches linespeed (13-14 seconds), and the following train can then be given authority to into the rear of the platform, and follow the departing train at linespeed.

The ATO system which is going to be used in the core, will it be fixed block or moving block?

Fixed block, albeit many of the blocks are rather small.

 

[carriageline]

I did say if... so were you not teeing me up then?

To the other two posters, I understand the operational reasons why it is the way it is. I'm not questioning what it exists for, but challenging the fact that we accept it rather than making the investments necessary to negate the need for such padding in future. Look at historical journey times, even on routes which have not had frequency explosions - and they have all crept in line with privatization. This is a more crowded railway, but it is naive to say that it isn't to circumvent penalty, now it is commercial venture.

A little contingency, sure, as human, technical and other factors for delay can be unforeseen - but these should be the exceptions, whereas we are building in the leeway into making them the rule.

I agree with you to an extent, we shouldn’t need padding.

But the truth is, we do. The majority of the network is so congested, a train running just 2-3 minutes late can cause massive delays (I’ve seen some hit thousands of minutes). For example, watch Lewisham when you have a few trains late, but within PPM.

And 2-3 minutes delay can very easily be caused outside of the railways control. Long dwell time at station, passcoms, kids playing with doors, call for aids in toilets, weather etc etc.

Whilst padding does make journeys longer, wouldn’t you prefer on a more reliable and punctual train service, that is less fragile?

It seems padding is becoming smaller and smaller, which does explain why some knock on delays are so big now.

 

[Bald Rick]

A little contingency, sure, as human, technical and other factors for delay can be unforeseen - but these should be the exceptions, whereas we are building in the leeway into making them the rule.

But that’s exactly what recovery time is. Engineering allowance to allow for planned speed restrictions, which are relative predictable on a long route, but not precisely where; pathing allowance to allow all the train paths on this increasingly busy railway fit together; and performance allowance to allow for very minor delays that often happen - a passenger requiring assistance at a station with a 30 second dwell time for example; a driver having to brake more cautiously due to rain.

 

[cle]

I know, I do get it. I'm just saying we should aspire to lessening it. Break a few eggs for a better network. With solid Thameslink and Crossrail, changing at termini to final destinations is easier.

For instance, reduce overlapping (the South London network is terrible for this and the duplication of termini) - so force through the Sutton loop changes to end in Blackfriars for instance - isolate branches - and increase frequencies by a lot.

Get Rainham and Sutton off the Thameslink map, replace with lower impact services such as some through East Croydon or via Catford. Swap with other lines - e.g some Kent trains from Victoria to run via Elephant. Do the figure of eight Sutton mix with Herne Hill and Brixton/Victoria to compensate. The line via Greenwich line should only see Cannon Street service etc etc etc - see the Hertford Loop for how this might work.

This is how the network in Japan is so well run (plus other things) - routes are clearly delineated (with a variety of termini maintained through grade separation or at extremities).

 

[atraindriver]

<deleted>

 

[Bald Rick]

As a matter of interest, if you download the tube Working Timetables from the TfL website and then compare the train running numbers (shown on the front of the train) on the trains you travel on against the working timetable, you will find that it's rare for a tube train to be anything less than 4 minutes late.
That level of delay would, I believe, cause major problems at the interfaces between the TL central section and the "rest of the railway" either inwards or outwards, which is why that padding is required: to provide the contingency to ensure that the TL trains present in the right order at the constrained central section. It's all about ensuring overall robustness.

The Swiss understand that very well: it's not about top speed, or even a high average speed. It's about getting the trains in the right places at the right times and doing so every day, and if that means padding then so be it; the Swiss have long built padding into their Taktfahrplan, and nobody complains about it.


If you speak to JR managers, they will tell you that they could not run to Japanese levels of punctuality on the UK infrastructure in part because the infrastructure is so constrained and in part because the national cultures are so different that they simply couldn't run stations in the way they do in Japan; British passengers simply don't (and won't) behave in the same way as Japanese passengers do.

I've worked on the railway for well over 30 years. It is a fact that my trains now spend roughly twice as long at stations(*) as they did then because people simply take much longer to board nowadays. It's partly a change in the type of people travelling and partly that people now expect the world to accommodate them rather than them accommodating the world. If they want to dawdle on, to stand in doors stopping other people passing or try to wedge three pushchairs in one doorway - then they will.
It all slows the trains down, and it also means that a realistic timetable has to allow longer station dwells than in the past; the old Southern Region 20 second dwells simply aren't possible, and even the standard 45 second sliding door dwells (created by having 30 seconds at one station then 1 minute at the next) are getting more and more difficult to work to.

The longer I spend on this railway, and the more time I spend travelling around the world and observing their railways with an insiders eye, the more I understand that when foreign operators tell me how impressed they are with the UK railway, they're actually telling the truth. UK passengers are conditioned to believe that the UK railway is automatically so much worse than anywhere else in the world, but all railways worldwide are curate's eggs: they're good in parts.

(*) 45-60 seconds instead of 20-30 seconds. Doesn't sound much, but it adds up!

Great post.

 

[RichardGore]

The ATO system which is going to be used in the core, will it be fixed block or moving block?

It’s an overlay on fixed-block ETCS L2.

 

[bramling]

As a matter of interest, if you download the tube Working Timetables from the TfL website and then compare the train running numbers (shown on the front of the train) on the trains you travel on against the working timetable, you will find that it's rare for a tube train to be anything less than 4 minutes late.

Not quite sure how you’ve arrived at that.

Sure there’s a couple of lines which have a tendency to run late for various reasons (Piccadilly Line is a favourite, the city section of the Metropolitan is another, and the Bakerloo another thanks to the erratic nature of the Network Rail section), but elsewhere it’s more common than not to be right-time unless there’s been a disruption. Generally the peaks run a bit late, but this is then generally fairly promptly recovered naturally.

On something like the Northern individual trains can be “trapped” running late, for example a train doing Edgware-Kennington rounders won’t have the opportunity to make anything up as it will always get stuck behind something else, however this is then something which the signallers will put right by one of a number of methods. But apart from that, generally it’s common for the signaller to look at the train listing for the next 10-15 trains and see everything shaded in white, which isn’t bad considering they turn red at just two minutes late (or perhaps even one minute, I forget).

I think overall people might be surprised just how “right time” much of LU runs for much of the time.

 

[rob w]

Many years I remember reading in the railway press an article about the widened lines modernisation and it mentioned that due to clearances in the tunnels slab track was needed and even then the pantograph’s of the 317’s were fully depressed hence a 30 mph restriction was needed. Ive no idea if the modern rolling stock has the same clearance issues.

Lastly the Blackfriars to Kings X section is starts at 1:04 here:

 

[Deepgreen]

I wished the other night, when TL was badly disrupted, that my driver had been a little more purposeful, as we crawled into every station from City to Wimbledon and came to a gentle rest each time. I was counting every second as I had had to change my homeward journey from a concert at the Barbican via Farringdon, from via Redhill to via Wimbledon at the last moment. My TL train was timed (in the WTT) to arrive at Wimbledon at 2324 and my Dorking train to leave Wimbledon at 2325. I know this wasn't a connection, but I would have caught my (last) Dorking train if the 700 had been driven a little more purposefully! No-one to blame, just frustrating.

 

[Firestarter]

I wished the other night, when TL was badly disrupted, that my driver had been a little more purposeful, as we crawled into every station from City to Wimbledon and came to a gentle rest each time. I was counting every second as I had had to change my homeward journey from a concert at the Barbican via Farringdon, from via Redhill to via Wimbledon at the last moment. My TL train was timed (in the WTT) to arrive at Wimbledon at 2324 and my Dorking train to leave Wimbledon at 2325. I know this wasn't a connection, but I would have caught my (last) Dorking train if the 700 had been driven a little more purposefully! No-one to blame, just frustrating.

What do you mean by Purposeful?

 

[atraindriver]

<deleted>

 

[bramling]

By dint of recording all my tube journeys and cross-referencing against the relevant WTT.

Apologies if that spoils your fun.

I can only think that you’re experiencing particular sections at particular times. As I say, there are some bits which do run late on a regular basis. But equally there is a much larger proportion which doesn’t. I was thinking about this earlier looking at some train listings which in some cases were 100% right time!

To be fair, there have been timetables which have never really worked. The Picc was plagued by this in the late 90s and early 2000s, as was the Northern up until about 2008, no doubt there’s plenty of other examples. However LU has done a lot of work over the last couple of decades to produce timetables which aren’t inherently unworkable.

 

[atraindriver]

<deleted>

 

[Bald Rick]

That's good. Perhaps they'd like to give Network Rail and the TOCs a kick, show them how it's done?
Some of the schedules appearing on the "main line" are worryingly inept.

It’s a rather different prospect timetabling the Sussex commuter services compared to the Piccadilly line.

 

[atraindriver]

<deleted>

 

[cjp]

Lastly the Blackfriars to Kings X section is starts at 1:04 here:

Interesting video thank you. I expected to see or hear much more when the train changed from DC to AC

 

[swt_passenger]

Interesting video thank you. I expected to see or hear much more when the train changed from DC to AC

May have been edited out for time? Assuming (from the comments) it’s about a year old video, the northbound changeover should normally have happened at City Thameslink, which was a quick stop as shown...

 

[Need2]

May have been edited out for time? Assuming it’s a year old video, the northbound changeover should normally have happened at City Thameslink, which was a quick stop as shown...

Definately edited, normally takes about a minute (give or take a bit!) and you hear absolutely nothing.

 

[Bald Rick]

Definately edited, normally takes about a minute (give or take a bit!) and you hear absolutely nothing.

The station dwell time is one minute, but the voltage changeover only takes about 15-20 seconds. On the Electrostars it was about 70seconds.