How unique are identifiers on the railway.

[plugwash]

How unique are things like headcodes and signal numbers (including the alphabetic prefixes)? Are they unique across the whole network or only unique within a limited area?

 

[ls2270]

Only unique within a limited area. In the case of headcodes, these can even be duplicated within one TOC, eg SWR has 2Rxx and 2Sxx services both in the Hounslow and Romsey areas. Signal box prefixes are also duplicated, eg W is used for both Wimbledon and Westbury.

 

[norbitonflyer]

Only unique within a limited area. In the case of headcodes, these can even be duplicated within one TOC, eg SWR has 2Rxx and 2Sxx services both in the Hounslow and Romsey areas. Signal box prefixes are also duplicated, eg W is used for both Wimbledon and Westbury.

On some busy routes the headcode may even be duplicated on the same day: If you run 3tph for 18 hours a day, you need 54 headcodes in each direction First "up" train might be 2A01, then 2A03, 2A05 etc, to 2A99, then starts again at 2A01.

 

[The Planner]

You shouldnt have duplicate signal numbers on the same control area. I am fairly sure there are a couple that slipped through the net around Stafford, will double check.

 

[Ianno87]

Signal box codes can be duplicated.

DJ off memory is used for:

-Dalton Jn (near Barrow)
-Deansgate Jn (Navigation Road)
-Diggle Jn

L is/was used for:
-Liverpool Street IECC
-London Bridge

 

[4F89]

You shouldnt have duplicate signal numbers on the same control area. I am fairly sure there are a couple that slipped through the net around Stafford, will double check.

Correct. When speaking to the signaller at LS1301 and LS1303 ground position lights, you must specify Stafford, for which additional notes are made on the signal. The other LS ground position lights are at Litchfield.

 

[Horizon22]

Headcodes (officially train IDs) can be duplicated and often are around the country. But I believe the same headcode shouldn't run in the same region within 8-12 hours. This necessitates 9xxx headcodes at GTR, London Overground and TfL Rail / Crossrail as otherwise there would be too many potential conflicts.

 

[edwin_m]

The combination of an Engineers' Line Reference (ELR) and a milepost distance can identify uniquely any spot on the network, although a few ELRs aren't unique as they signify formerly through routes that have been part closed (such as DCL). ELR plus structure number uniquely identifies any structure.

 

[gimmea50anyday]

TPE has 2 1M91
06:13 EDB-MIA and
16:11 EDB-MIA

There is also a 1M91 16:09 Brighton-Victoria

Looking at Genius however the headcodes aren't 4 digits which probably helps identify them on modern computer and tracking systems, although we still use the 4 digits as standard

 

[sw1ller]

Yeah, Chester sees multiple instances of the same head code in a day. Off the top of my head, 1G00 (TFW 0422 Chester to Birmingham new street) 1G00 (2135 to Birmingham new street *edit* only runs to Crewe now because of Covid) Think that one starts at Bangor and is run by Avanti.

 

[The Planner]

Correct. When speaking to the signaller at LS1301 and LS1303 ground position lights, you must specify Stafford, for which additional notes are made on the signal. The other LS ground position lights are at Litchfield.

Thanks, knew I wasn't going senile.

 

[Surreytraveller]

Headcodes (officially train IDs) can be duplicated and often are around the country. But I believe the same headcode shouldn't run in the same region within 8-12 hours. This necessitates 9xxx headcodes at GTR, London Overground and TfL Rail / Crossrail as otherwise there would be too many potential conflicts.

I've heard of cases of two trains running at the same time on the same route. One began before midnight, the other afterwards

 

[DerekC]

I had to explain UK headcodes to a bunch of German engineers a few years ago. They couldn't believe that we have a system in which you can have duplicates! They asked why we don't have a couple more digits. I guess that's because the headcode had to be fitted into the four boxes on the front of a diesel loco way back when, and be big enough for the signaller to read.

 

[306024]

If you look at a typical schedule in Real Time Trains there are various identifiers in the header. This one for example is the 16.30 Liverpool St to Norwich on a Saturday:

• UID P06298,
• identity 1P44
• TSC 21781002
• headcode 0123
• TRUST ID 521P44MT07

The UID is allocated by NR ops planning and is unique for every schedule.
The identity is what you are all familiar with, but has two further letters in planning systems which are suppressed in downstream systems. That’s how you can have more than one 1P44 if necessary.
The TSC is the financial group the train belongs to for all sorts of cost allocation and billing.
The headcode isn’t really a headcode, just uses an old field in planning systems for seat reservations.
The TRUST ID is for monitoring and recording data in TRUST, one of the computer operating systems.

 

[DorkingMain]

Headcodes (officially train IDs) can be duplicated and often are around the country. But I believe the same headcode shouldn't run in the same region within 8-12 hours. This necessitates 9xxx headcodes at GTR, London Overground and TfL Rail / Crossrail as otherwise there would be too many potential conflicts.

I believe 9 headcodes are used to denote trains with special signalling requirements. 9 headcodes are used at SWR when a 442 is running to denote to the signaller the train has limited route availability, while 9s on GTR / LO are used to denote trains that are cleared into the relevant core sections, to limit the possibility of a train being wrong-routed.

 

[Elecman]

And the 9 headcodes on WCML are forv5ge Anglo Scots routed via Birmingham rather than Trent Valley

 

[edwin_m]

I had to explain UK headcodes to a bunch of German engineers a few years ago. They couldn't believe that we have a system in which you can have duplicates! They asked why we don't have a couple more digits. I guess that's because the headcode had to be fitted into the four boxes on the front of a diesel loco way back when, and be big enough for the signaller to read.

The train describers in hardware panels have four-character displays which would need major modification to show more characters. Some of the older ones may still be based on early electronics which could only handle a four-character description, as may the protocols by which descriptions are transmitted.

Even screen-based systems such as TRUST and VDU signaling centres only have four character spaces for each berth. Changing these would cost a lot of money as data changes always do, and would probably need many other things to be re-positioned to make way for the longer displays. This in turn would probably trigger an extra screen in some workstations, which probably can't be accommodated due to either lack of space or excessive workload.

 

[webbfan]

Headcodes (officially train IDs) can be duplicated and often are around the country. But I believe the same headcode shouldn't run in the same region within 8-12 hours. This necessitates 9xxx headcodes at GTR, London Overground and TfL Rail / Crossrail as otherwise there would be too many potential conflicts.

Isn't that within the same signal area rather than region.

The UID is allocated by NR ops planning ....

Thought this was a Darwin generated identifier, hence associated with a TOC. Although they call it a RID by adding date onto it so making it unique for 24 hours.
Is a while since had to understand these (headcodes and UID's) so may not be exact.

 

[DorkingMain]

Only unique within a limited area. In the case of headcodes, these can even be duplicated within one TOC, eg SWR has 2Rxx and 2Sxx services both in the Hounslow and Romsey areas. Signal box prefixes are also duplicated, eg W is used for both Wimbledon and Westbury.

SWR also duplicates 2J headcodes in fairly close proximity (Hampton Court and Farnham-Guildford shuttles). Has caused some confusion as a couple of depots sign both routes.

 

[Eloise]

You shouldnt have duplicate signal numbers on the same control area. I am fairly sure there are a couple that slipped through the net around Stafford, will double check.

You may want someone to check Crewe in Dec:

1K02 arrives from Llandudno Jn at 05:59
1K02 depart Crewe for Derby at 06:58

TfW and EMR appear to be all over each other in the 1Kxx series. 1K07 are less than an hour apart.

 

[jopsuk]

At St Pancras, a 1Fxx is either a Southeastern service to or from Faversham or an EMR service to Sheffield, because HS1 and the MML are separate regions with seperate signalling control. Eg, on a weekday the 1131 departure to Sheffield is 1F27, then the 1140 arrival from Faversham is 1F27.

The Thameslink and Eurostar services all use 9xxx, but I don't think they share letter codes, as services to Paris are 9Oxx and those to/via Brussels 9Ixx, letters I think that are usually avoided

 

[Spartacus]

I had to explain UK headcodes to a bunch of German engineers a few years ago. They couldn't believe that we have a system in which you can have duplicates! They asked why we don't have a couple more digits. I guess that's because the headcode had to be fitted into the four boxes on the front of a diesel loco way back when, and be big enough for the signaller to read.

Technically they are longer, 10 digits, for instance today's 09:16 Euston - Lime Street is 721F11ME08, but 1F11 suffices for most things.

 

[The Planner]

You may want someone to check Crewe in Dec:

1K02 arrives from Llandudno Jn at 05:59
1K02 depart Crewe for Derby at 06:58

TfW and EMR appear to be all over each other in the 1Kxx series. 1K07 are less than an hour apart.

I was on about signal numbers, we all know headcodes slip through the cracks. I expect someone has spotted those.

 

[pdeaves]

You shouldnt have duplicate signal numbers on the same control area. I am fairly sure there are a couple that slipped through the net around Stafford, will double check.

Yes, plated 'Trent Valley LS' and 'Stafford LS'.

 

[alangla]

Yes, plated 'Trent Valley LS' and 'Stafford LS'.

Daft question, but why weren’t one set either re-lettered or, say, 4000 not added to the signal numbers?

 

[The Planner]

Daft question, but why weren’t one set either re-lettered or, say, 4000 not added to the signal numbers?

Re-lettering would go against the geographical area, number change would make sense but would be pretty expensive to resolve for what are not main aspects.

 

[Surreytraveller]

At St Pancras, a 1Fxx is either a Southeastern service to or from Faversham or an EMR service to Sheffield, because HS1 and the MML are separate regions with seperate signalling control. Eg, on a weekday the 1131 departure to Sheffield is 1F27, then the 1140 arrival from Faversham is 1F27.

The Thameslink and Eurostar services all use 9xxx, but I don't think they share letter codes, as services to Paris are 9Oxx and those to/via Brussels 9Ixx, letters I think that are usually avoided

Thameslink use 9O - St Albans to Sutton / Sutton to St Albans via Wimbledon

 

[SargeNpton]

Thought this was a Darwin generated identifier, hence associated with a TOC. Although they call it a RID by adding date onto it so making it unique for 24 hours.
Is a while since had to understand these (headcodes and UID's) so may not be exact.

The UID is allocated within Network Rail's Train Planning System, and has to be unique on any one day. It is exported on the CIF timetable data files and Darwin is one of the downstream systems that makes use of it.

 

[Mollman]

I believe 9 headcodes are used to denote trains with special signalling requirements. 9 headcodes are used at SWR when a 442 is running to denote to the signaller the train has limited route availability, while 9s on GTR / LO are used to denote trains that are cleared into the relevant core sections, to limit the possibility of a train being wrong-routed.

9 headcodes are one of those railway oddities. Used to denote slow freight trains (uncoupled I think). When Eurostar came along 9 was adopted to fit in with the French train code system (9Oxx and 9Ixx UK become 90xx and 91xx respectively in France). Since then 9 headcodes became used for a wider variety of services including those that don't take the 'traditional' or direct route hence Euston - Scotland via West Midlands (rather than Trent Valley) and Liverpool - Edinburgh via ECML (rather than WCML) trains carry these.

 

[LAX54]

TPE has 2 1M91
06:13 EDB-MIA and
16:11 EDB-MIA

There is also a 1M91 16:09 Brighton-Victoria

Looking at Genius however the headcodes aren't 4 digits which probably helps identify them on modern computer and tracking systems, although we still use the 4 digits as standard

We have the first train as 2F01, but by 2200 we are back to 2F01 once more, there are also S-ARS issues with the same train, when a train is platformed overnight, if you enter the headcode anytime before about 0400, even if the train has run the previous day, it will see it as 'late' and set a route