Why do so many trains have incompatible coupling systems?

[CBlue]

But that is precisely where the major delays etc that make it into the media are happening, where one unit fails and is incapable of even emergency rescue just with coupling/air brake, by loco, other unit, or whatever.

Is it though? The King's Cross debacle on Friday that kicked this discussion off wasn't due to coupler incompatibility, more to do with an EMU tangled up in the overhead wires. Coupling adapters exist anyway for that sort of eventuality.

Sometimes it feels like this forum wants some kind of Schrödinger's railway. Everything is too monotonous and boring with attempts to standardise fleets in particular regions - yet at the same time there's apparently too much variety because a IET can't couple to a 700 or something.

 

[Failed Unit]

We have gone backwards a little but even under BR they had quirks.

The DMUs were fairly standard on BSI. This even continued onto the Turbo's (170)

However, the NSE although using BSI wired theirs slightly different that they couldn't couple to a Regional Railways service. In reality this wasn't and issue as the only places they really met were Worcester - Hereford and the Chiltern extensions north of Banbury. (I don't think these started appearing until privatisation). I understand with these differences a RR DMU could drag and NSE DMU (Empty) and vice versa, but electrical connectivity wasn't working. (The brake had to be removed manually and the good unit relied on its own)

On NSE it was pretty much tightlocks everywhere. But not as many quirks.

Then along came all bets are off after privatisation.

 

[ComUtoR]

I am always reminded of XKCD https://xkcd.com/927/ in times like this. (linky is regarding competing standards)

Chuck in history/time and you can see why there are so many variants. With trains running around that are 20+yrs old. It is inevitaable that you would get a more modern variant. Increasing safety standards and better and more reliable technology will also contribute to changing of standards over time.

Do we really expect units to be running around with buffer and chain coupling :/

 

[hwl]

Why on earth were so many trains designed to have non-compatible couplings ( agreed not the reason for most of the delay in untangling the knitting). Do other railway systems across the world have the same issue - or is it just a byproduct of the way our privatisation worked?

Isn’t the “Schaku” or Dellner/Scharfenberg becoming the default mechanical coupling for new stock anyway? When RSSB did an explanatory paper about this around ten years ago, they basically saw the fleets dividing into existing regional DMUs all being BSI, and new EMUs and high speed DEMUs all going in the the Dellner/Scharfenberg direction.

And pretty much everything new since then has been Dellner/Scharfenberg/clone albeit at there are 2 mounting heights used (Due to UIC etc. trying to be helpful and "standardise" for high speed in a way that did work for GB in general).

The main issue had nothing to do with coupling, but it might be an idea if locos on thunderbird contracts had a coupling adapter on board or they get fitted with lifting variable height Dellner (improving on the 57s, what ROG does and is specing for the 93). DB should be upping their thunderbird provision.

Freight on the continent is about to standardise on a single auto coupler design which may see some change here too.

Do we really expect units thunderbird locos to be running around with buffer and chain coupling :/

Suitably edited for the real issue

 

[dosxuk]

But that is precisely where the major delays etc that make it into the media are happening, where one unit fails and is incapable of even emergency rescue just with coupling/air brake, by loco, other unit, or whatever.

How much of that though is the press just going along with the popular viewpoint that all the railway needs to do when a train breaks down is get something else to hook up and dump the broken one in a siding somewhere so everyone else can carry on as normal? A view that falls apart when it meets reality.


Likewise with the Lewisham self-evacuating - loads of bad press that they couldn't move the trains causing people to get out, but think what would have happened if they hadn't turned the power off and someone got hit or electrocuted.

The press writes stories the public want to read. If you've been stuck for hours because of a breakdown, you don't want to hear that everyone is doing everything they can, but it's complicated and difficult. You want to hear someone is going to get the sack for bringing the wrong bit of kit, which if they'd had everything would have been back to normal after 20 minutes.

 

[matchmaker]

I'm sure Triang and Hornby Dublo never had this problem

They did. You needed an adaptor wagon to join the two types!

 

[PHILIPE]

It's been proved today that a Class 67 can couple to a Class 769 unit (see TFW 769 Thread)

 

[Irascible]

I would imagine that incompatibility is only an issue in rare circumstances.

How often do you need/want to mix stock?

I understand the issue from the perspective of breakdown recovery but not from timetabling

Possibly people are looking at it from a view of "what happens if we move unit X to somewhere else". Down this way we've had local services made up of whatever combo of 80s DMU is available for quite a while ( changing to 90s DMU now - not 100% sure mix & match is over yet but no idea if you can jam a 158 & a 166 together anyway ). I'm a little surprised there's no spec for a TMS interface protocol, honestly - we've had considerable experience at that sort of thing in the last few decades & it is the sort of thing people like to devise.

But, in general with the way stock is purchased at the moment, it does seem to be mostly a non issue.

 

[Noddy]

Freight on the continent is about to standardise on a single auto coupler design which may see some change here too.

Yes this was mentioned on the class 69 thread recently. While I appreciate what the Digital Automatic Coupler (DAC) is trying to achieve (complete automation) the four trial systems all look pretty complex. I wonder what:

A) the cost implication will be for installing it on 20-30 year old wagons?

B) how serviceable will they be on a cold dark wet night when one has failed?

C) Given the system is digital what the digital lifespan of the system will be?

I know folk have mentioned it elsewhere but I can’t help thinking that wider adoption of the ‘Janney’/‘AAR’/‘whatever other name you want to give it’ semi automatic coupler might be a better. While it’s only semi-automatic its very reliable, simple and can cope with ridiculous loads.

 

[Irascible]

B) how serviceable will they be on a cold dark wet night when one has failed?

C) Given the system is digital what the digital lifespan of the system will be?

One thing about digital interfaces is at least you can make the electrical connectors simpler than using loads of analog circuits. Not to say it still fail anyway, but might be a little better than you might think at first. Wonder how well it degrades & if there's provision for emergency manual jumpers, given if a coupler fails electrically you've broken the bus for everything behind it too...

If the system itself is sensibly and robustly designed to handle future expansion in the data it stores, and changes in the hardware interface - we do this stuff all the time these days, we've learned a lot from trying to globally network *everything* - then it should be fine for as long as we process data the same way. That is unfortunately quite an "if".

ISTR quite a bit of detail of ETRMS was published years before anyome rolled it out, will have to dig into this too.

 

[AM9]

I'm curious what those aircraft electrical interconnections are. I can't think of anything comparable to a rail vehicle coupling that is standardised or in an aggressive environment

My post was describing interfacing of electronic systems, i.e. many of the systems installed in aircraft are from manufacturers with totally different design practices, however the system integration authority (usually the aircraft manufacturer defines standard interfaces that enable these systems to work together. Most modern aircraft have digital interfaces such as ARINC 629 data interfaces for data interchange. There are several 'fieldbus' implementations where near real-time communications can provide control and signalling facilities. For safety critical signals the gold standard on aircraft is to have discrete signals, i.e. hard wired interfaces that geniune real time. Such an application on land based transport might be for braking or interlock circuits. All of these interfaces are mandated to operate normally within the potentially agressive EMC environment found on aircraft.
The fact that the rolling stock fleet at any given point might consist of desgns spanning 20+ years, doesn't mean that steps cannot be taken to harmonise and eventually standardise their physical and electrical interfaces. The Dellner/Scharfenburg coupling is a good start on the mechanical side, what is then needed is a mechanical standard for the physical electrical connections, which would rapidly develop into a common electrical interface. Not all trains would usee all the discretes, and not all databus protocols would use the full range of signals, but it would be possible to specifiy that all new trains have the minimal requirements for full interoperability where needed.

 

[Julia]

Potentially useful? How about an operational and safety nightmare. How do you do door control from the VEP, including door interlock on the 455? That's before you get to different door close procedures and the risk of guards forgetting they had the other type of unit attached when dispatching.

And some of those nightmares only become apparent after a time in use. The Bedford-St Pancras hydraulic DMUs (class 127) were *in theory* fully compatible with other 1st generation units, but a Cravens coupled on the back of one went up in flames thanks to the absence of an audible clue to the driver that it needed to change gear.

 

[wobman]

This causes issues for TFW with 175's are unable to couple to there other units, having 1 universal coupler is very beneficial for strengthening services. Like adding a 150 / 153 to 158's or vise versa makes unit allocation much easier.
Unfortunately the 230's can't couple to any other traction, that's one big reason they are all staying on the borderlands line

In the future 197's are only compatible to a 196 but not a 195 ! The electrical boxes are in different places, so in an emergency maybe mechanical couple could be a solution. It's baffling how the DFT don't stipulate 1 coupler for all traction, as part of a group standard.

 

[Mat17]

I guess many heritage railways that have acquired Pacers recently are going to come across this conundrum as well, can't sent a 108 out to rescue a failed 142.

 

[dosxuk]

My post was describing interfacing of electronic systems, i.e. many of the systems installed in aircraft are from manufacturers with totally different design practices, however the system integration authority (usually the aircraft manufacturer defines standard interfaces that enable these systems to work together. Most modern aircraft have digital interfaces such as ARINC 629 data interfaces for data interchange. There are several 'fieldbus' implementations where near real-time communications can provide control and signalling facilities. For safety critical signals the gold standard on aircraft is to have discrete signals, i.e. hard wired interfaces that geniune real time. Such an application on land based transport might be for braking or interlock circuits. All of these interfaces are mandated to operate normally within the potentially agressive EMC environment found on aircraft.
The fact that the rolling stock fleet at any given point might consist of desgns spanning 20+ years, doesn't mean that steps cannot be taken to harmonise and eventually standardise their physical and electrical interfaces. The Dellner/Scharfenburg coupling is a good start on the mechanical side, what is then needed is a mechanical standard for the physical electrical connections, which would rapidly develop into a common electrical interface. Not all trains would usee all the discretes, and not all databus protocols would use the full range of signals, but it would be possible to specifiy that all new trains have the minimal requirements for full interoperability where needed.

Can't think of many occasions when aircraft designed and built by different manufacturers to completely different specs and requirements have needed to interface such that multiple of them can be controlled from one flight deck with all systems operating as one.

 

[Noddy]

One thing about digital interfaces is at least you can make the electrical connectors simpler than using loads of analog circuits. Not to say it still fail anyway, but might be a little better than you might think at first. Wonder how well it degrades & if there's provision for emergency manual jumpers, given if a coupler fails electrically you've broken the bus for everything behind it too...

If the system itself is sensibly and robustly designed to handle future expansion in the data it stores, and changes in the hardware interface - we do this stuff all the time these days, we've learned a lot from trying to globally network *everything* - then it should be fine for as long as we process data the same way. That is unfortunately quite an "if".

It’s not just the digital interface/electrical connectors though! This is one of the four trial systems:

(NOT MY PHOTO)

How much of this could be fixed on a wet windy night at immingham for example?!! Compared to this:


(NOT MY PHOTO)

 

[thenorthern]

As others have mentioned coupling is only one part of it, the electrical systems of trains is a big part as trains may have the same coupling but different electrical systems.

 

[AM9]

Can't think of many occasions when aircraft designed and built by different manufacturers to completely different specs and requirements have needed to interface such that multiple of them can be controlled from one flight deck with all systems operating as one.

I think you've missed my point. I am talking about multiple electronic systems from different manufacturers requiring to work reliably in safety critical environments. Aircraft are an example where interface standardisation is expected and delivered. This is bread and butter for the avionics industry, yet there are those that seem to think that the equivalent on the railway is so much more complicated that nothing can be done. As I said, the physical mechanical coupler interface seems to be gravitating to a Dellner/Scharfenburger type, but the electrics (far more simple than electronics) their locations on the coupler, the pin configuration seem to be different depending on who specifies them and who manufactures them. It doesn't need to be this way, (but the UK railway and it's supplier industries seem content to perpetuate the mess for some self-defeating reason.

 

[dosxuk]

I think you've missed my point. I am talking about multiple electronic systems from different manufacturers requiring to work reliably in safety critical environments. Aircraft are an example where interface standardisation is expected and delivered. This is bread and butter for the avionics industry, yet there are those that seem to think that the equivalent on the railway is so much more complicated that nothing can be done. As I said, the physical mechanical coupler interface seems to be gravitating to a Dellner/Scharfenburger type, but the electrics (far more simple than electronics) their locations on the coupler, the pin configuration seem to be different depending on who specifies them and who manufactures them. It doesn't need to be this way, (but the UK railway and it's supplier industries seem content to perpetuate the mess for some self-defeating reason.

No, it's your analogy that doesn't fit. Nobody is complaining about there being problems with the multitude of systems on board an individual train, or class of trains, communicating or meeting manufacturers specifications. The issues are coming when people want to connect multiple vehicles from different eras, created for different purposes, and have them behave as one.

Connecting a Class 700 up to a Class 67 is much more similar to connecting a Boeing 737 up to an Airbus A400. Sure, you can probably make a physical attachment and tow the other around, but nobody will have put any thought into getting the autopilot of one able to control the autopilot of the other.

 

[Nym]

It’s not just the digital interface/electrical connectors though! This is one of the four trial systems:
View attachment 97399
(NOT MY PHOTO)

How much of this could be fixed on a wet windy night at immingham for example?!! Compared to this:

View attachment 97400
(NOT MY PHOTO)

I'd love to know how, for a freight consist they plan on effective autocoupler heating. Basically a requirement in winter.

 

[AM9]

No, it's your analogy that doesn't fit. Nobody is complaining about there being problems with the multitude of systems on board an individual train, or class of trains, communicating or meeting manufacturers specifications. The issues are coming when people want to connect multiple vehicles from different eras, created for different purposes, and have them behave as one.

Exactly as I said in post #41:

The fact that the rolling stock fleet at any given point might consist of desgns spanning 20+ years, doesn't mean that steps cannot be taken to harmonise and eventually standardise their physical and electrical interfaces.

Until something is done about that the railway will struggle on with its absurd silo engineering practices. Railway operations in many other countries seem to manage without this mess.

 

[Irascible]

It’s not just the digital interface/electrical connectors though! This is one of the four trial systems:
<was a photo>

How much of this could be fixed on a wet windy night at immingham for example?!! Compared to this:

Oh yeah, I didn't want to go near the pipework - to be fair it's a bit apples & oranges given the US coupler still needs the pipework connected manually - there's also a little slack which is presumably something they're trying to avoid too. As someone said, how the hell are they going to get that thing apart if it's frozen...

 

[43096]

As someone said, how the hell are they going to get that thing apart if it's frozen...

It’ll be when, not if!

Imagine the scene, Brennero, mid-January, 2am. Train comes in with fancy new auto couplers and needs a loco change from Austrian loco to an Italian one. Couplers are frozen... What they going to do, wait two months til it warms up a bit?

 

[Bletchleyite]

Different units definitely have them at different heights. The other issue is software of course- for example a 195 can rescue a 331 in an emergency, but the 331 has to be dead. If you can't get 2 units ordered at the same time from the same company for the same operator to be compatible, it's going to be an uphill struggle. Obviously it's a different situation in an emergency rescue to service work.

Cheap as

80x EMUs and bi-modes can work together, indeed that sort of "lash-up" is often used on LNER.

 

[O L Leigh]

This old chestnut again...? Well, OK.

On NSE it was pretty much tightlocks everywhere. But not as many quirks.

Oh, there were plenty of quirks. PEP derived units could only multi with Mk3 based units for ECS and assistance purposes only. The same applied to Networkers, so you had operators like FCC who had, in effect, three separate and self-contained fleets all of which had the same coupler. I believe that Cl357s which also have (had?) Tightlocks can only multi within class.

This forum overblows a lot of things, and coupling compatibility is one of the biggest. Since moving away from the 3-link coupler and vacuum pipe the railway has always had inconsistencies. But why does that matter? As long as it makes operational sense there's no reason why, say, a TPX Cl185 should be expected to be compatible with, say, a Southern Cl171. If a train sits down there will be another similar train not very far away which can go and deal with it. For those rare occasions when it's a proper knacker and cannot be rescued in the normal way, as in the KX dewirement, you get a loco, a coupling adapter, hook up the air pipe and drag it away like that. Simple. All the rest is just fluff and nonsense.

I used to work the WA route which, at the time I left, had a service pattern that, in coupler terms at least, went Tightlock, Delner, Tightlock, Delner and so on. At no point did this inconsistency cause any problems in terms of dealing with a failed train.

 

[pdeaves]

There's probably an argument that if we do standardise on a connector it should just send everything through one or two bus with a published communication protocol rather than using individual pins for everything.

Or, save the physical electrical connection(s) and go wireless. I mean, wireless technology has come on in leaps and bounds recently, like many other computer-related technologies. I'm sure something akin to weak wi-fi to talk the short distance across the coupling could be made to work if required.

 

[Irascible]

Or, save the physical electrical connection(s) and go wireless. I mean, wireless technology has come on in leaps and bounds recently, like many other computer-related technologies. I'm sure something akin to weak wi-fi to talk the short distance across the coupling could be made to work if required.

If you *really* wanted you could probably just go straight to managing it over GSM ( with the advantage that you can centrally monitor in detail then ), I don't think we're really dealing with much data. Safety critical stuff is always going to want hard connections anyway though, including a couple of electrical ones if you have passengers to lock in & shout at so as they're there anyway...

 

[Taunton]

And some of those nightmares only become apparent after a time in use. The Bedford-St Pancras hydraulic DMUs (class 127) were *in theory* fully compatible with other 1st generation units, but a Cravens coupled on the back of one went up in flames thanks to the absence of an audible clue to the driver that it needed to change gear.

That wasn't anything particular with the couplings/control gear, it was just a nonsense that an automatic transmission set could be driven with a manual change set attached to the rear - that's real mechanical engineering 1.01. They were supposedly compatible as the auto sets had been specially fitted with a change gear button on the control desk, which did work, but inevitably a driver forgot about the set on the back. Separately it should have had an overspeed detector which disconnected the transmission.

This old chestnut again...? Well, OK.

If a train sits down there will be another similar train not very far away which can go and deal with it. For those rare occasions when it's a proper knacker and cannot be rescued in the normal way, as in the KX dewirement, you get a loco, a coupling adapter, hook up the air pipe and drag it away like that. Simple. All the rest is just fluff and nonsense.

I used to work the WA route which, at the time I left, had a service pattern that, in coupler terms at least, went Tightlock, Delner, Tightlock, Delner and so on. At no point did this inconsistency cause any problems in terms of dealing with a failed train.

That is exactly what happened in the Kentish Town incident. The "similar train not very far away" wasn't and took several hours; the coupling adapter procedure fell down; and because of the planned service pattern the train behind was a different coupler and unable to assist even though the recovery procedure was written that it could. Not fluff and nonsense at all to the RAIB in their highly critical report.

 

[matchmaker]

That wasn't anything particular with the couplings/control gear, it was just a nonsense that an automatic transmission set could be driven with a manual change set attached to the rear - that's real mechanical engineering 1.01. They were supposedly compatible as the auto sets had been specially fitted with a change gear button on the control desk, which did work, but inevitably a driver forgot about the set on the back. Separately it should have had an overspeed detector which disconnected the transmission.


That is exactly what happened in the Kentish Town incident. The "similar train not very far away" wasn't and took several hours; the coupling adapter procedure fell down; and because of the planned service pattern the train behind was a different coupler and unable to assist even though the recovery procedure was written that it could. Not fluff and nonsense at all to the RAIB in their highly critical report.

The 127s coupling code was originally Blue Square, same as the majority of 1st generation dmus. It was later changed to Red Triangle so as to prevent mixing with mechanical sets. The only difference between Blue Square and Red Triangle was the symbol - there was no change in the actual MU cabling.

 

[Julia]

The 127s coupling code was originally Blue Square, same as the majority of 1st generation dmus. It was later changed to Red Triangle so as to prevent mixing with mechanical sets. The only difference between Blue Square and Red Triangle was the symbol - there was no change in the actual MU cabling.

OK, I seem to have ended up making an argument *for* incompatibilities in coupling! Not allowing two units to be put together by having physically distinct mechanisms or cabling is certainly better than having symbols that have to be observed (the 127s as Red Triangle), which in turn is better than relying on drivers never making accidental errors on the move (the 127s as Blue Square). Of course, nowadays this is more likely to be done in software.