How do train desciber berths work?

[plugwash]

Talk of messages triggered by train describer berths in another thread got me wondering, how do they actually work. In particular how does a headcode move from one berth to the next. Does the signaller have to manually enter the headcode in the next birth and remove it from the previous one? do they press a button to move a headcode from one berth to the next? is it automatically triggered by train detection systems? does it vary depending on the type of singalling in use?

 

[WestRiding]

Generally works in the signal box by the track circuit occupation moving along past the relevant signals when they are clear (TCB). They can however be inputted manually by the signaller, but only to 4 didgit/letter combinations.

 

[Sunset route]

I know in the Eastbourne area before it was resignalled and where there were no suitable track circuits in the Absolute Block areas, they used a combination of new lever position detection (lever bands) within the signalbox with track side treadles to step the train description along.

 

[norbitonflyer]

Generally works in the signal box by the track circuit occupation moving along past the relevant signals when they are clear (TCB). They can however be inputted manually by the signaller, but only to 4 didgit/letter combinations.

An interesting explanation of how they work (or should work) is given in the report on the accident at Hyndland in 1980. (see aparas 2 and 24)

DoT_Hyndland1980.pdf (railwaysarchive.co.uk)

An underlying bug was revealed by the more intense service operating after the Argyle Line had opened the previous year (all trains through both Queen Street and Central Low Level stations have to pass through Hyndland) which caused some train descriptions to be lost if three trains were occupying successive track circuits. The result, after a chain of misunderstandings, was that a train was authorised to pass a red signal and ran into a "missing" train (technically a head on collision as the driver of the stationary train had changed ends in readiness to enter Hyndland depot through a trailing connection).
The driver of the stationary train was using the signal-post telephone at the time of the collision. The driver of the other train retreated into the passenger area and warned the passengers of the impending collision. (Interesting in view of recent discussions of the Salisbury accident)
Both units, then almost-new 314s, were repaired. However, 11 years later the leading carriage of one of them was damaged beyond repair in the Newton accident (it was replaced by a redundant class 507 carriage).

 

[Annetts key]

Talk of messages triggered by train describer berths in another thread got me wondering, how do they actually work. In particular how does a headcode move from one berth to the next. Does the signaller have to manually enter the headcode in the next birth and remove it from the previous one? do they press a button to move a headcode from one berth to the next? is it automatically triggered by train detection systems? does it vary depending on the type of singalling in use?

It depends on if the area is controlled by an electrical/electronic signalling system (e.g. track circuit block such as controlled by a PSB, ROC) or it’s a mechanical/semaphore signal box.

With an electrical/electronic signalling system, normally the only time the signaller has to do anything, is when a train exits a depot or sidings, a train starts it’s booked service at a station, terminates at a station, joins or divides at a station (or in sidings). This is when the signaller manually enters the train number/description/head code into the train describer (TD) system. With computer based signalling systems, it may be possible for the head code to be automatically entered.

The TD system receives “stepping” information from the signalling system. Every time a train passes a signal, the signalling system tells the TD. It then “steps” the description (head code) to the next signal “berth”. Normally all main signals that can show a stop aspect have a TD berth.

In a mechanical/semaphore signal box that is on a line where the next “box” has an electrical/electronic signalling system, that mechanical/semaphore signal box will normally be provided with a “fringe” TD unit. Descriptions for trains from the direction of the electrical/electronic signalling system will automatically arrive. But the signaller has to enter the head code manually for trains going in the other direction.

Other mechanical/semaphore signal boxes are not normally provided with any TD equipment.

 

[Springs Branch]

I'm not sure if this is 100% correct, but as I understand it when Manchester Piccadilly PSB's area was extended along the Bolton line in 1990, Piccadilly's Train Describer system could not "talk" to the T.D's at the 1970s-era PSBs at Warrington & Preston.

This meant fringe boxes had to be retained for a couple of decades at Blackrod, Crow Nest Jn and Bromley Cross (?) - so the signallers could (among other tasks) take the headcode received from Warrington / Preston and key it in to Piccadilly, and vice versa.

Whatever the technical issues were, these apparently got resolved by 2013 and Blackrod & Crow Nest Jn boxes were eliminated and their signalling taken over by Piccadilly - and IIRC Bromley Cross survives just as a level crossing box.

 

[jamesst]

Something as a driver I've always wondered.
How is a train passed from one signal box to another? I'm meaning on a track circuited line from one modern box to another.

 

[Ken H]

Another question. How do the describers cope with trains dividing/joining en route? Or even as assisting loco being added to a train?

 

[Stampy]

Something as a driver I've always wondered.
How is a train passed from one signal box to another? I'm meaning on a track circuited line from one modern box to another.

Indeed, as well as having to pass on the information of signal aspects I presume too...

Take for example my (current) local PSB, Peterborough......

All southbound trains from Grantham will be under the control of Doncaster PSB, until they pass through Stoke Tunnel - Then they come under the control of Peterborough..

The Peterborough first signal they see, could have a "feather" on, directing their train onto the Up Slow - or they could get a "flashing green"

So Doncaster's LAST signal southbound (just before Stoke Tunnel) would either be Green, or a restricted aspect for the turnout....

 

[edwin_m]

Something as a driver I've always wondered.
How is a train passed from one signal box to another? I'm meaning on a track circuited line from one modern box to another.

Indeed, as well as having to pass on the information of signal aspects I presume too...

Take for example my (current) local PSB, Peterborough......

All southbound trains from Grantham will be under the control of Doncaster PSB, until they pass through Stoke Tunnel - Then they come under the control of Peterborough..

The Peterborough first signal they see, could have a "feather" on, directing their train onto the Up Slow - or they could get a "flashing green"

So Doncaster's LAST signal southbound (just before Stoke Tunnel) would either be Green, or a restricted aspect for the turnout....

Usually the boundary between power boxes is located on a stretch of plain line with automatic signals. This simplifies the technical and operational interfaces.

Signal aspects need to be handled by high-integrity systems to avoid wrong side failures. If the tracks are uni-directional the information the last signal in one power box area needs to "know" from the next one is the state of the track circuit(s) up to the overlap of the next signal, and the aspect and lamp proving status of that signal. Depending on the technology used, these could be hardwired directly to the relay cabinet next to the signal, or supplied via outputs from a computer based interlocking module, wired to inputs for the other power box's system if that is CBI too. These days they may well pass it over data networks, with suitable encoding to prevent a corrupted message causing a wrong side failure.

If the track at the boundary is bi-directional it gets a bit more complicated, as the two power boxes have to "agree" which direction it is being used in, so vital control information needs to pass both ways for each direction of train travel. On reasonably modern installation there is also an emergency alarm between the centres.

Train descriptions are not of the same high integrity and I think traditionally went over serial data links, with the format given by a BR specification whose number I don't recall. These days all train description steps are reported over data networks to feed things like TRUST and passenger information systems, so neighbouring power boxes may just use this.

 

[swt_passenger]

Usually the boundary between power boxes is located on a stretch of plain line with automatic signals. This simplifies the technical and operational interfaces.

I noticed (via opentraintimes) that the boundary between Eastleigh and Brockenhurst is also staggered (or overlapping) along the line, by two berths - I had previously assumed that’s so that the signals approaching the various crossovers and junctions are all controlled by the best location for the particular direction?

 

[MarkyT]

On reasonably modern installation there is also an emergency alarm between the centres.

I recall 1960s/70s era WR panels (now mostly gone) usually had traditional block communication bells hidden inside the consoles for emergency use in case of TD failure, with a button to operate them on the panel face!

Generally, at junction signals where trains have a routing choice, as well as the track circuit status used for triggering, additional route set inputs to the TD system are monitored to help the system determine which way to step the description.

 

[SilentGrade]

Another question. How do the describers cope with trains dividing/joining en route? Or even as assisting loco being added to a train?

TDs will be manually inputted in these cases by the controlling box. Any TD berth can usually have a description entered into it by the signaller.

 

[edwin_m]

I noticed (via opentraintimes) that the boundary between Eastleigh and Brockenhurst is also staggered (or overlapping) along the line, by two berths - I had previously assumed that’s so that the signals approaching the various crossovers and junctions are all controlled by the best location for the particular direction?

In the past there have been some quite long stretches of plain line where the Up and the Down were controlled by different power boxes, so that the signallers had the best picture of trains approaching even if they couldn't do anything about them until they got to the first junction within the area. This doesn't seem to happen any more, though some of the previous ones remain. I guess the reason is for safety - best to have the same person looking after both lines in case there's an incident on one of them. The signallers now have TRUST, which gives them a much better idea of what is heading towards them.

TDs will be manually inputted in these cases by the controlling box. Any TD berth can usually have a description entered into it by the signaller.

If it's a timetabled split/join, an automatic route setting system can probably do this where one is fitted.

 

[Annetts key]

Something as a driver I've always wondered.
How is a train passed from one signal box to another? I'm meaning on a track circuited line from one modern box to another.

Although the technology has changed, the principles are not really any different than between different control (‘interlocking’) areas of the same PSB or ROC. What is used depends on the technologies of the adjoining signalling systems. Often interface circuits using safety relays are used.

Another question. How do the describers cope with trains dividing/joining en route? Or even as assisting loco being added to a train?

With most TD systems, the signaller can ‘interpose’ a new description into any TD berth. Be it empty/blank or to replace an existing description. Most TD berths on ‘plain line’ are not specified to be able to automatically cope with trains dividing/joining. So the signaller will manually update as needed.

If it’s a location where this is planned to happen regularly (e.g. station) the signalling scheme may include provision for more than one TD berth, or provide hidden TD berths. This allows more than one description to be stored at that point on that line, at that signal. Exit signal berths on permissive goods lines like loops are a good example, some can store up to three descriptions.

I recall 1960s/70s era WR panels (now mostly gone) usually had traditional block communication bells hidden inside the consoles for emergency use in case of TD failure, with a button to operate them on the panel face!

The older generation of signallers that I knew much preferred the ‘block bell’ and hated the emergency alarm. They also found it useful to ‘alert’ the signaller at the ‘next box’ if they were slow to respond to the telephone…

The S&T technicians also preferred the ‘block bell’ as it generally was more reliable and easier to fault find on.

Back to TD systems. To connect and communicate descriptions between signal boxes/PSBs/ROCs/Control Centres, modems or data links are used. Normally routed via telecoms infrastructure. Occasionally via private BT wires.

If the TD systems in the adjacent signal boxes are the same type, or are compatible with one another, this keeps it simple. Otherwise a special interface unit may be provided to translate the messages.

On the signallers display, they will be a special train approaching TD display. This shows the description of the train that is approaching the end of the control area of the adjacent signal box. Some systems can ‘look back’ multiple signal sections, so as to transmit the description as early as possible to the next signal box. If a number of trains are following one another, the system will send each description in turn, one after the other, but only when the train approaching TD display in the next signal box is empty. This display becomes empty when the train actually passes onto the area of control of this signal box. It then gets stepped into the TD berth of the first signal on that line controlled by that box.

Some older designs of TD systems could do this for two trains, so the train approaching displays are labelled ‘1st’ (train) and ‘2nd’ (train) approaching.

It’s actually more complex to try to explain, and watching it in action is far easier!

 

[Stampy]

I noticed (via opentraintimes) that the boundary between Eastleigh and Brockenhurst is also staggered (or overlapping) along the line, by two berths - I had previously assumed that’s so that the signals approaching the various crossovers and junctions are all controlled by the best location for the particular direction?

And furthering my example from earlier - using OpenTrainTimes, when a train is passing Southbound through Grantham - you can normally see it's reporting code (i.e 1A72) showing in the 1st berth of Peterborough's area..

Same for the other way - trains will show up in the berth just before Stoke Tunnel, as that's Doncaster's 1st signal...

 

[TSG]

Depending on the technology used, these could be hardwired directly to the relay cabinet next to the signal, or supplied via outputs from a computer based interlocking module, wired to inputs for the other power box's system if that is CBI too. These days they may well pass it over data networks, with suitable encoding to prevent a corrupted message causing a wrong side failure.

There may be a system where they've cracked network transmission but I haven't seen it myself. I think most, if not all, fringes will be done with relays, for several good reasons:

1. While CBIs of the same type can usually communicate over internal data links at a signalling centre, these links won't always have a 'long line' facility to go over transmission to an interlocking many miles away.
2. Different CBIs can't usually talk directly to each other even when they're in the same room (yet)
3. The trackside modules of different (or even the same) CBIs often can't easily be plugged output to input. The various proving and self checking features of the modules' inputs and outputs would just play merry hell with each other. Everything can drive a relay. Everything can detect a relay contact.
4. When a CBI is first put in, often it will be interfacing to relay signalling so you already have a relay interface at the fringe. It's usually simplest to change one end of that to accommodate the new CBI.
5. Transmission from trackside back to a central interlocking, then over transmission again to another central interlocking on a different processing schedule, then over transmission again out to the trackside at the other side of the fringe, and then all that in reverse again can be a relatively long time. You'll likely have less latency and timing issues with a relay interface between systems (see Tollerton control).
6. If you're on site at the fringe, commissioning testing or fault diagnosis is easy when you have relays to look at and voltages to measure. With a 'network' fringe you may need a conference call between someone at each fringe trackside case, a control centre tech at each central interlocking and maybe even two signallers and the Network Management Centre to see what's happening. Good luck!

Train descriptions are not of the same high integrity and I think traditionally went over serial data links, with the format given by a BR specification whose number I don't recall. These days all train description steps are reported over data networks to feed things like TRUST and passenger information systems, so neighbouring power boxes may just use this.

BR1810, or PS9 in modern parlance. As far as I know, box to box TD transmission is still all over serial links even at the ROCs. Some smaller boxes send all their steps (i.e. even ones internal to their control area) over their serial link, because they don't have a link to SMART (the system that just about everything that isn't a TD or an ARS gets its TD steps from), and their larger neighbour forwards these steps along with their own.

If it's a timetabled split/join, an automatic route setting system can probably do this where one is fitted.

I'm not certain but I suspect some basic Automatic Code Insertion systems at power boxes predate ARS

If the TD systems in the adjacent signal boxes are the same type, or are compatible with one another, this keeps it simple. Otherwise a special interface unit may be provided to translate the messages.

I would have thought everything built since the 80's should be compatible. Anything older than that is doing very well to still be going! I did hear a tale of somebody interfacing an early computer TD to an older one built from GPO rotary pulse relays, the sort that were used in telephone exchanges. It worked for a little while, but they hadn't considered the pulse timing. It stopped when the relays shook themselves to bits....

 

[swt_passenger]

And furthering my example from earlier - using OpenTrainTimes, when a train is passing Southbound through Grantham - you can normally see it's reporting code (i.e 1A72) showing in the 1st berth of Peterborough's area..

Same for the other way - trains will show up in the berth just before Stoke Tunnel, as that's Doncaster's 1st signal...

The other odd thing about my SWML example is that there are two stations in the overlapping areas of control where the opposite platforms are in different signal control areas. That adds to @edwin_m‘s good point about the difficulties if there an incident.

 

[Annetts key]

I would have thought everything built since the 80's should be compatible. Anything older than that is doing very well to still be going!

Given that there are still 1970s TDM remote control systems still in use, I was being careful… Especially as before privatisation the different regions of BR did their own thing (the old 1988 WR SSTD design needed a Vaughan special interface unit at Westbury to interface to other TD systems for example).