Class 345 progress

[SeaKing]

Did I just see 064 on 9P18 or were my eyes playing tricks.

 

[JonathanH]

345064 has been converted back to 7-car formation and is currently working 9P18 0705 Reading to London Paddington.

That takes the compliment of available 7-car units to twelve (006/07/13/15/17/22/44/49/51/52/57/64) for eleven diagrams.

Presumably one more to come based on the posting by 'JN114' at the weekend.

 

[JonathanH]

Did I just see 064 on 9P18 or were my eyes playing tricks.

Yes, you did but it has been converted to 7-car formation - worked 5G54/5G55 test trip last night.

 

[kevin_roche]

Does anyone know why they have to remove two cars to operate without the ETCS on board?

Is it something to do with the braking profile for TPWS and 9 cars?

Or is it just that particular configuration is approved and 9 cars without ETCS is not?

 

[JN114]

It’s just software versions.

AFAIK there isn’t an approved version of the train management software for 9 cars without all the ETCS and other bells and whistles enabled (at least not signed off for use on passenger services) - remember the software also holds things like the SDO database, so you can’t just run the 7 car software on a 9 car train.

If you’re going down the road of writing a new version of the software for 9 car non-ETCS, then you’re already doing more work than the fix involves.

Removing coaches is fairly trivial, software to run 7s already exists and is approved for use; and as soon as it’s done it’s just a quick shake down run to check it’s all working as expected and you have another unit good to go.

 

[Ex31Rigger]

Units 047 & 063 currently in process of being converted to RLU

 

[kevin_roche]

It’s just software versions.

AFAIK there isn’t an approved version of the train management software for 9 cars without all the ETCS and other bells and whistles enabled (at least not signed off for use on passenger services) - remember the software also holds things like the SDO database, so you can’t just run the 7 car software on a 9 car train.

If you’re going down the road of writing a new version of the software for 9 car non-ETCS, then you’re already doing more work than the fix involves.

Removing coaches is fairly trivial, software to run 7s already exists and is approved for use; and as soon as it’s done it’s just a quick shake down run to check it’s all working as expected and you have another unit good to go.

Thank you. That is what I suspected.

 

[JonathanH]

Units 047 & 063 currently in process of being converted to RLU

345063 out on the 5G50 / 5G51 test run to Reading and back today.

5G50 1146 Old Oak Depot to Reading https://www.realtimetrains.co.uk/train/P62297/2020-06-27/detailed
5G51 1303 Reading to Old Oak Depot https://www.realtimetrains.co.uk/train/33992/2020-06-27/detailed

 

[paulravuslante]

Looking like there will be some moves in / out of the central section tonight to OOC: 5Y70 leaves Westbourne Park at 2232; followed by 5Y71 at 2302.
Return workings run pretty swiftly back as 5T14 and 5T15 from OOC at 2318 and 2338 respectively.

Tomorrow morning 5T16 and 5T17 also leave OOC for WBP at 0059 and 0130.

 

[JonathanH]

Looking like there will be some moves in / out of the central section tonight to OOC: 5Y70 leaves Westbourne Park at 2232; followed by 5Y71 at 2302.
Return workings run pretty swiftly back as 5T14 and 5T15 from OOC at 2318 and 2338 respectively.

Tomorrow morning 5T16 and 5T17 also leave OOC for WBP at 0059 and 0130.

No sign of 5Y70, 5Y71, 5T14 or 5T15 running - seems strange to add STP schedules and not use them - still no signalled move to berth 0092 or from 0091 or 0099 since 8 June.

I'll go to sleep and see whether there is a sign of 5T16 and 5T17 having happened in the morning.

 

[JonathanH]

As predicted by 'Ex31Rigger' up thread, and based on information here, 345047 appears to have run a test trip to Hayes & Harlington this evening so will presumably appear in service in the next day or so taking the current west-side 7-car fleet to fourteen units - 006/07/13/15/17/22/44/47/49/51/52/57/63/64. 345047 last worked as a 7-car at the start of May so it would appear that it was converted from 7-car to 9-car and back to 7-car without venturing out of the depot.

No sign of signalled movements to the lines feeding the tunnel last night. There is a VSTP schedule for another move tonight.
https://www.realtimetrains.co.uk/train/35386/2020-07-02/detailed

EDIT: 5T13 2240 Old Oak Depot to Paddington Crossrail, the VSTP move, is running.

 

[rebmcr]

As predicted by 'Ex31Rigger' up thread, and based on information here, 345047 appears to have run a test trip to Hayes & Harlington this evening so will presumably appear in service in the next day or so taking the current west-side 7-car fleet to fourteen units - 006/07/13/15/17/22/44/47/49/51/52/57/63/64. 345047 last worked as a 7-car at the start of May so it would appear that it was converted from 7-car to 9-car and back to 7-car without venturing out of the depot.

That's a lot of effort so we can surmise the AWS software issue doesn't have a quick fix, which in turn indicates either a deeper architectural problem within the software (requiring a wait for a new version), or even a hardware problem (hopefully not, as a non-quick-fix hardware issue probably involves replacing control desk equipment and/or ETCS Onboard Computer cards).

Were the 9-car 345s still running version P_D+12 or had they moved on?

 

[Snow1964]

It’s just software versions.

AFAIK there isn’t an approved version of the train management software for 9 cars without all the ETCS and other bells and whistles enabled (at least not signed off for use on passenger services) - remember the software also holds things like the SDO database, so you can’t just run the 7 car software on a 9 car train.

If you’re going down the road of writing a new version of the software for 9 car non-ETCS, then you’re already doing more work than the fix involves.

Removing coaches is fairly trivial, software to run 7s already exists and is approved for use; and as soon as it’s done it’s just a quick shake down run to check it’s all working as expected and you have another unit good to go.

It’s an easy solution for Operators, but of course is putting operational ease ahead of the purpose of having full length trains (to move passengers) as you decrease passenger capacity by about 22%

Of course you could retain capacity by increasing a 4 trains per hour to 5 per hour, but ’the Railway’ doesn’t give a stuff about maintaining capacity for its revenue supply

 

[JN114]

It’s an easy solution for Operators, but of course is putting operational ease ahead of the purpose of having full length trains (to move passengers) as you decrease passenger capacity by about 22%

Of course you could retain capacity by increasing a 4 trains per hour to 5 per hour, but ’the Railway’ doesn’t give a stuff about maintaining capacity for its revenue supply

The 9 car trains, with the current version of train management software, have a serious flaw in a critical safety system. As such they rightfully cannot operate them in passenger service - there is zero ifs or buts to be had here - even ECS moves are very strictly limited.

Its not operational convenience, until they are fixed it’s 7s or nothing. And while 7s do unfortunately represent a 22% capacity decrease, nothing represents a 100% capacity decrease.

As for just running an extra train per hour, where are the paths coming from? Crossrail don’t have exclusive use of the Relief lines on the GWML, there are GWR services to mingle with, plus it’s already 8tph Crossrail between Hayes and Paddington in the peaks - plus GWR plus freight. Should those operators suffer reduced capacity because Crossrail are running slightly shorter trains? And if so why is it ok for them to have reduced capacity but Crossrail run at 100%? There isn’t the capacity to run more on top of the current timetable.

 

[Siemens Staines]

Obviously I don't know the details, but from the description of the problem relating to the integration of AWS/TPWS and ETCS then at the very least I would have thought the locations where those two systems interface (e.g Airport Jct?) would require special attention when any software fix is identified. In turn I doubt that you can't really run those sort of tests on safety critical systems when there are other railway services about, so nigh-time possessions may be required. That in itself will affect other operators, and won't be able to occur at the drop of a hat.

 

[Snow1964]

The 9 car trains, with the current version of train management software, have a serious flaw in a critical safety system. As such they rightfully cannot operate them in passenger service - there is zero ifs or buts to be had here - even ECS moves are very strictly limited.

Its not operational convenience, until they are fixed it’s 7s or nothing. And while 7s do unfortunately represent a 22% capacity decrease, nothing represents a 100% capacity decrease.

As for just running an extra train per hour, where are the paths coming from? Crossrail don’t have exclusive use of the Relief lines on the GWML, there are GWR services to mingle with, plus it’s already 8tph Crossrail between Hayes and Paddington in the peaks - plus GWR plus freight. Should those operators suffer reduced capacity because Crossrail are running slightly shorter trains? And if so why is it ok for them to have reduced capacity but Crossrail run at 100%? There isn’t the capacity to run more on top of the current timetable.

I was aware of the reasoning, and of course a 22% cut is not as bad as a 100% cut. But that rather misses the point. The point I am trying to make is that passengers are getting a poor deal.

Of course we all know that is a working set of software takes 10 years to write, it would have been sensible to have started the day after Boris pressed the start button on a piling machine at Canary Wharf. But that never happened and now we have trains, tunnels with track and electrificatio, but no trains with perfect software. The defence of the railway really ought to start with whichever bright spark forgot to get software writing started early enough

 

[Domh245]

The point I am trying to make is that passengers are getting a poor deal.

A poor deal, yes, but also the least worst deal.

The defence of the railway really ought to start with whichever bright spark forgot to get software writing started early enough

All well and good having software written in advance, but if you can't find the bugs until you start running it in anger on the infrastructure it's supposed to be running on....

 

[Energy]

With complex software comes many small bugs which are hard to find unless you actually use the software with everything it is designed for.

 

[coppercapped]

I was aware of the reasoning, and of course a 22% cut is not as bad as a 100% cut. But that rather misses the point. The point I am trying to make is that passengers are getting a poor deal.

Of course we all know that is a working set of software takes 10 years to write, it would have been sensible to have started the day after Boris pressed the start button on a piling machine at Canary Wharf. But that never happened and now we have trains, tunnels with track and electrificatio, but no trains with perfect software. The defence of the railway really ought to start with whichever bright spark forgot to get software writing started early enough

I am not so sure that the passengers are getting a 'poor deal' - they are simply not getting, at the moment, quite such a good deal[1] as was hoped.

What is already running, even in seven coach form, is a considerable improvement over the recent service offered by 3 coach Class 165s and 166s.

For what its worth - how do you know that software development didn't start early enough? My take is rather that there has been a fundamental change in the way that trains are now controlled. Previously some functions in what was essentially hardware were software controlled - now the 'train operating system' is supported by various bits of hardware. It's very similar to the change in aircraft engineering which has seen the complete replacement of rods and cables by 'fly by wire'. This change also took a lot longer to complete than was previously the case - and for a similar reason - it required a complete change in philosophy and attitude in those affected before it became the standard way of doing things.

The Class 345s and similar generation products from all the other manufacturers are the first fully 'fly-by-wire' trains. They will get there - but it will be more difficult than anyone considered when they started down this path.

[1] For some value of 'good deal'. The 345s have a lot of standing space and therefore a reduced number of seats. Some of these are longitudinal and there are no toilets. It is not immediately obvious to me that the creature comforts are on a par with those these trains have replaced.

 

[hwl]

I was aware of the reasoning, and of course a 22% cut is not as bad as a 100% cut. But that rather misses the point. The point I am trying to make is that passengers are getting a poor deal.

Of course we all know that is a working set of software takes 10 years to write, it would have been sensible to have started the day after Boris pressed the start button on a piling machine at Canary Wharf. But that never happened and now we have trains, tunnels with track and electrification, but no trains with perfect software. The defence of the railway really ought to start with whichever bright spark forgot to get software writing started early enough

The signalling system for the Core was only selected 6 years ago and Siemens only bought the firm providing that technology 9 years ago. The version 2 of the AWS/TPWS STM* module (supplied by Mors Smit) only became available in 2016 and was first used in the Crossrail units. New in the last few years rolling stock also has to have the V4 driver interface for AWS which hasn't been problem free either (Stadler also having issues here)

* V2 being the sorted for mass rollout sole AWS & TPWS to ETCS STM for the UK.

What it has shown is that lab testing alone won't do.

 

[Peter Sarf]

I was aware of the reasoning, and of course a 22% cut is not as bad as a 100% cut. But that rather misses the point. The point I am trying to make is that passengers are getting a poor deal.

Of course we all know that is a working set of software takes 10 years to write, it would have been sensible to have started the day after Boris pressed the start button on a piling machine at Canary Wharf. But that never happened and now we have trains, tunnels with track and electrificatio, but no trains with perfect software. The defence of the railway really ought to start with whichever bright spark forgot to get software writing started early enough

It is already a big improvement over the services that were 3 and 5 car formations of 165s and 166s. The most overcrowded services in the UK.

Software is never fully bug free until it has had real time testing in the environment it is meant for. You cannot fully test signalling and train until all are ready (or you think they are ready). In this case there are three different signalling systems to cater for (AWS, TPWS sand ETCS iirc). The switch from one to another mid service has to work. Unlike the software I used to write this is safety critical so it is not a case of plodding on as best they can.

 

[kevin_roche]

Software is never fully bug free until it has had real time testing in the environment it is meant for. You cannot fully test signalling and train until all are ready (or you think they are ready). In this case there are three different signalling systems to cater for (AWS, TPWS sand ETCS iirc). The switch from one to another mid service has to work. Unlike the software I used to write this is safety critical so it is not a case of plodding on as best they can.

There is also the CBTC system (AKA Trainguard MT) used in the central section which is not being used yet.

 

[Peter Sarf]

There is also the CBTC system (AKA Trainguard MT) used in the central section which is not being used yet.

True. This is not a simple introduction of software by any means. Can say it is easy to forget it is a fairly new design of train as well !.

 

[paulravuslante]

Looks like a PAD - HAY shuttle has operated all day today with 345s. Previously Hayes services called at Acton Mainline but this stop has remained with the Heathrow services for now. Hanwell was also served by some Hayes trains in the peak.

 

[Philip Phlopp]

There was an article published on the Rail Engineer website at the end of last week concerning the ongoing signalling system testing for Crossrail, which may be of interest to some of you.

Crossrail's Signalling Challenge - Rail Engineer

Listen to this article Crossrail has many complex systems but providing the entire route with a new signalling system is one of the biggest challenges. It is a highly complex arrangement, with many interfaces between different new and legacy systems, although final testing is now underway. It is...

www.railengineer.co.uk

 

[itfcfan]

There was an article published on the Rail Engineer website at the end of last week concerning the ongoing signalling system testing for Crossrail, which may be of interest to some of you.

Crossrail's Signalling Challenge - Rail Engineer

Listen to this article Crossrail has many complex systems but providing the entire route with a new signalling system is one of the biggest challenges. It is a highly complex arrangement, with many interfaces between different new and legacy systems, although final testing is now underway. It is...

www.railengineer.co.uk

A very informative article - thank you for sharing.

 

[JonathanH]

If the current expectation is that the opening schedule is mid-late 2021 for Paddington to Abbey Wood and late 2022 for Paddington to Shenfield then complete through running does the TPWS-CBTC transition need to be perfected by 2021 to avoid the need for test running possessions or do they have until 2022 to get this right?

 

[kevin_roche]

If the current expectation is that the opening schedule is mid-late 2021 for Paddington to Abbey Wood and late 2022 for Paddington to Shenfield then complete through running does the TPWS-CBTC transition need to be perfected by 2021 to avoid the need for test running possessions or do they have until 2022 to get this right?

I went to a lecture at the IET in Portsmouth almost a year ago where it was said that the transitions were, in fact, working better than expected. The issues were said to be in other parts of the software. The speaker was not able to be specific. Another IET lecture later last year pointed out some interference issues between the OLE and axel counters being installed on the GWML, which had just been solved. Software is one of those things where one thing can be fixed and the fix causes 3 more problems in other places.

 

[kevin_roche]

There was an article published on the Rail Engineer website at the end of last week concerning the ongoing signalling system testing for Crossrail, which may be of interest to some of you.

Crossrail's Signalling Challenge - Rail Engineer

Listen to this article Crossrail has many complex systems but providing the entire route with a new signalling system is one of the biggest challenges. It is a highly complex arrangement, with many interfaces between different new and legacy systems, although final testing is now underway. It is...

www.railengineer.co.uk

Thank you for pointing it out.

 

[itfcfan]

That's a lot of effort so we can surmise the AWS software issue doesn't have a quick fix, which in turn indicates either a deeper architectural problem within the software (requiring a wait for a new version), or even a hardware problem (hopefully not, as a non-quick-fix hardware issue probably involves replacing control desk equipment and/or ETCS Onboard Computer cards).

Were the 9-car 345s still running version P_D+12 or had they moved on?

A post from Rail magazine on Twitter indicates this may take some time to resolve:

Confirmed that there is no current return to service date for @TfLRail 9-car Cl.345s. They’ve been removed from traffic following a defect that affects signalling systems. The 7-cars remain ok to operate @BombardierR_UK tells me.

https://twitter.com/Richard_rail/status/1280163958117273604