"TOPS" loco class classification - when did this actually start?

[eldomtom2]

The modern system of classifying locos and multiple units with two- or three-digit numbers is often called "TOPS" classification. But I've found a November 1969 accident investigation (of an accident that occurred in May) that refers to "Class 20" and "Class 50" locomotives, and from some quick research BR didn't even decide to purchase TOPS until months after this. So what's the actual history of these classifications?

 

[Magdalia]

The new locomotive classification was introduced in the summer of 1968, roughly coinciding with the end of steam traction. In Railway Observer it was set out and first used from the September 1968 issue.

The association with TOPS comes from the use of the classification in the fleet renumbering scheme that came when TOPS was introduced.

There was a recent discussion of the TOPS renumbering scheme here:

TOPS renumbering process

Other than taking off the old stickers and putting on new ones, what was involved in the renumbering scheme, and why did it take so long to renumber classes? And secondly, why did some classes never get renumbered to TOPS? I'm aware TOPS was the backend system and the number was just the asset...

www.railforums.co.uk

 

[snowball]

The new locomotive classification was introduced in the summer of 1968, roughly coinciding with the end of steam traction. In Railway Observer it was set out and first used from the September 1968 issue.

Others will know much better than I do, but I think it was several years later that individual locos got 5-digit numbers with the class number as the first two digits.

 

[randyrippley]

TOPS was well in use at Westbury for monitoring freight wagons in 1970/71 (school trip to the signal box there) so you can assume that took 2-3 years from purchase for it to be working then. That knocks back the purchase decision to ~ 1968 or earlier, and any preparative work such as class numbers even earlier
The intended loco class numbers were known by then, but renumbering was a couple of years away - a recent thread had details

 

[Magdalia]

TOPS was well in use at Westbury for monitoring freight wagons in 1970/71

This can't be correct.

The first references to TOPS in BR files at Kew is in 1970, the investment was approved early in 1971, and the first local TOPS office was at Exeter in May 1972.

 

[Gloster]

The 1973 Ian Allan combined volume has twenty-three of the 195 AC electrics and fifteen Class 76 with their new TOPS numbers: the date for LMR information is 23-12-1972. I think they completed the LMR electrics before going on to the diesels.

TOPS was well in use at Westbury for monitoring freight wagons in 1970/71 (school trip to the signal box there) so you can assume that took 2-3 years from purchase for it to be working then. That knocks back the purchase decision to ~ 1968 or earlier, and any preparative work such as class numbers even earlier
The intended loco class numbers were known by then, but renumbering was a couple of years away - a recent thread had details

When I was a numbertaker in 1978-1979, they were still completing the work on wagons and only just starting parcels, etc. stock.

 

[randyrippley]

This can't be correct.

The first references to TOPS in BR files at Kew is in 1970, the investment was approved early in 1971, and the first local TOPS office was at Exeter in May 1972.

Unless they were telling us porkies.
It was my first year at secondary school, which was 1970-71. Latest it could have been was early summer 71, and they showed us the local computer terminals they were using. Maybe they had a test/demo version running?

 

[Magdalia]

It was my first year at secondary school, which was 1970-71. Latest it could have been was early summer 71, and they showed us the local computer terminals they were using. Maybe they had a test/demo version running?

Are you sure that you are not out by a year?

The roll out was regional so it is quite possible that Westbury was up and running before the 1972 school summer holidays.

 

[Gloster]

Westbury might have been a location for a fairly early installation of the system because of the importance of the growing stone traffic, which might have been a good trial application.

 

[randyrippley]

Are you sure that you are not out by a year?

The roll out was regional so it is quite possible that Westbury was up and running before the 1972 school summer holidays.

Not impossible but I would have thought unlikely

 

[norbitonflyer]

I am not even certain that the renumbering was required for TOPS, or was an unrelated project. Although known as "TOPS" numbers the actual class numbers were allocated several years before TOPS was introduced - indeed, there were many class numbers allocated that were never applied as new 5-digit numbers because the classes were withdrawn before renumbering commenced. (Classes 04, 10, 11, 12, 14, 15, 16, 17, 21, 22, 23, 29, 30, 41, 42, 43, 48, 53, 70, 77 and 80 - as well as the specual cases of classes 35 and 52)

And cCertainly the TOPS computer system was introduced before all locomotives had been renumbered, which would not be until 1978. This didn't seem to prevent TOPS being used to trace wagons- its primary purpose. Indeed it would not be until the 1980s before duplication of carriage and loco numbers was addressed, (certainly after withdrawal of the Deltics, whose numbers duplicated the single-unit DMUs) and TOPS was certainly installed before then.

The first locomotives to enter service with TOPS numbers from new were the Class 87s, which had been allocated numbers in the E32xx series (if my ian Allen Volume of 1973 is to be believed), but never carried them. And of course the Hymeks and Westerns were never renumbered, presumably because replacing their cast numbers (single casting in the Westerns, individual characters for the Hymeks) would have been too expensive, and simply removing them would have left unslightly holes.

The first "TOPS" number I saw was in July 1972 1973 - 45102, at Exeter Central of all places, and I had no idea what it was - beyond the obvious fact it was a "45". The Peaks and 86s started to be renumbered earlier than the rest, as they went through works for modifications - fitting electric heating and revised suspension respectively, but because this took well over a year, they were also two of the last classes to be completed.

[EDIT - year corrected]

 

[Dunfanaghy Rd]

One definite date that I have found is 6th October 1971 - the date of the letter of approval from the Minister of Transport.
Pat

 

[Falcon1200]

According to the January 1972 Railway Magazine the first loco to be renumbered was 76050. There is also, in the same issue, an article about TOPS (for wagons) stating there would be a pilot scheme in Devon and Cornwall to be operational by the end of 1973.

 

[Magdalia]

The first "TOPS" number I saw was in July 1972 - 45102, at Exeter Central of all places, and I had no idea what it was - beyond the obvious fact it was a "45"

The class 45 Peaks were the first main line diesel locos to be renumbered, but that was in the spring of 1973. In 1972 the only renumbered locos were electrics.

 

[D7666]

It was my first year at secondary school, which was 1970-71. Latest it could have been was early summer 71, and they showed us the local computer terminals they were using. Maybe they had a test/demo version running?

My first year at secondary school level in Salisbury was also 1970-1971.

I agree that Westbury had some form of computer wagon control working by early 1972 which means almost certain to have been a 1971 install.

I well remember the early 1972 re-launch of the much increased Mendips stone traffic which occurred when major stone flows for M2/M23/M25 construction started to Merstham, Gatwick, and Fareham as well as more to Botley.

My memory tells me at least the HTV and MCO wagons then used were already under some kind of computer control. I had always assumed it was TOPS.

Seen through Salisbury, three D800s (814 825 829) had been reinstated early 1972, nameplateless, as well as the other 10 that survived the 10/71 holocaust were routinely deployed on this work.

From some time 03/1972 onwards, we started to see 5-7 D800s a day on stone, after a gap of almost no D800s at all since they ceased Waterloo Exeter weekend 2-3/10/71. This start up was some time between the 02/72 spring school term half term (when there were minimal stone trains) and the 04/72 Easter school break.

This early 1972 lot of trains, and here were as many as 8-10 a day - ran as 6Z-- reporting numbers; they got 6O-- and 6V-- in the May 1972 WTT.

D800s worked HTVs as class 6.

D1000s worked MCO as class 8.

The HTV and MCO wagons were marshalled into sets with the 4 digit set number stencilled on the side of each wagon. The MCO sets ran with a stencilled brake van at each end; gradually the vans gave way to fitted heads, about 3-4 MCV displaced the same numer of MCO at each end as well as the van. I am pretty sure those wagon set numbers were computer data related as they had no 8 or 9 in the number range from what I saw which suggests they were octal numbers.

They needed 8-10 trains a day as HTV were 21 t coal hoppers but overall is low capacity per wagon, and increasing train length beyond D800 brake capacity.

Like I said above,my memory tells me at least the HTV and MCO wagons were under some kind of computer control and I had always assumed it was TOPS.

There is no reason why Exeter office could not have been responsible for Westbury. If May 1972 TOPS open is correct, they almost certain were trialling before then, a pilot like this they'd have been doing so in 1971. Which aligns with the earlier post.

Not co-locating a TOPS office at a traffic centre or traffic source was done in the early days of TOPS implementation. Again, Salisbury for example, which was a traffic source, had no TOPS office, and never did while the concept of separate TOPS offcies existed. The local office was actually at Romsey station, which itself generated no freight traffic at all even then.

Further, there was a BR developed computer system that existed before TOPS. Around about 1966 or so there is a Learned Society (I.Mech.E.(Loco) or possibly I.Loco.E.) paper refering to this. A copy of said paper is (well certainly was to 1982) resident in the University of Warwick library stacks, and, one of the BR personnel who had worked on that was later one of the engineering teaching staff during my degree time there.

I can't comment on what is or is not at Kew, but will say that any record that is missing from there will lead to incorrect assumptions by readers.

But I am sure at least 03/72 Westbury stone traffic wagons were under computer control, be it TOPS or something else.

 

[Magdalia]

I have now found my copy of the history of TOPS by Robert Arnott.

This confirms that Exeter was the first trial site.

It also clearly states that the first cutover for live use of TOPS was at St Blazey and Plymouth in August 1973. This should have included Exeter but that was delayed by 4 weeks because of some local unofficial industrial action. These match the dates that are on file at Kew.

There is a short report in the September 1973 issue of Railway Magazine, page 472.

Whatever was being used at Westbury in 1971 was not TOPS. It was probably a version of a Fortran program that had been developed at Derby and was first used to manage the Aire Valley MGR trains in 1970.

 

[AdamWW]

I have heard many times that the "TOPS" class numbers were necessary as TOPS couldn't cope with the previous numbering system.

From previous discussions on this, the true situation appears to be that the class numbers were thought up before TOPS, but TOPS introduction gave a justification for renumbering locomotives to include the new class numbers.

Of course some locomotives ran into the 70's without TOPS numbers and it seems that it happily coped with this.

Various statements are made about what TOPS can and cannot cope with (such as numbering starts at 001 because it couldn't handle 000) but I think it turns out that the locomotive number is just effectively a free text field and anything can go in it, and numbers start at 001 because non mathematicians generally don't like starting to count things from zero.

 

[D7666]

Well as Fortran was the lingo for the computer part of my Warwick engineering degree, that does align with the engineering lecturer ex Derby.

So all what I type is not rhubarb.

I have heard many times that the "TOPS" class numbers were necessary as TOPS couldn't cope with the previous numbering system.

From previous discussions on this, the true situation appears to be that the class numbers were thought up before TOPS, but TOPS introduction gave a justification for renumbering locomotives to include the new class numbers.

Of course some locomotives ran into the 70's without TOPS numbers and it seems that it happily coped with this.

Various statements are made about what TOPS can and cannot cope with (such as numbering starts at 001 because it couldn't handle 000) but I think it turns out that the locomotive number is just effectively a free text field and anything can go in it, and numbers start at 001 because non mathematicians generally don't like starting to count things from zero.

Those early restrictions are really due to the cost of hardware.

For example having a field that is 000 (i.e. class 61 number 61000), any nit wit can write some code to make the system work with 0 in that field.

But, when you do add that code to be able to recognise that 000 is not null but a valid item, and all the dotted 1's and crossed 0's checks involved, and all the code to do that, and all the other checks for other things, your memory space and CPU capability starts to get swallowed up.

Bear in mind 1960s hardware both memory and CPU were enormously expensive; even in early 1990s when your home AT PC could do the entire BR fleet in one excel sheet, if you did want to increase your RAM or update to a higher CPU you ended up significantly more ££££ cost.

So BR said no XX000 and saved a fortune on hardware capital cost.

More or less - not exactly - the same less hardware needed reasons apply that aircraft squawk codes are 0000-7777 octal and why LT continued with 000-777 train numbers into computer days long after the original 3-set relay reason no longer applied.

So BR dumbed down the spec, and said no 000; there was probably still a check on operator typos though.

Back in 1968 UIC brought in the 6+1 digit numbering system, with several countries, even eastern bloc, going over to it. BR did not. Popular rumour was "we are not connected to Europe so not needed" but that is not the reason at all - UIC needs a 6 digit field + check digit = 7; BR by using only 5 digits no check allowed much capital cost saving.

 

[AdamWW]

The early restrictions are really due to the cost of hardware.

For example having a field that is 000 (i.e. class 61 number 61000) any nit wit can write some code to make the system work with 0 in that field.

But, when you add that code, with all its dotted 1's and both 0's, and all the other checks for other things, your memory space and cpu capability starts to get swallowed up. Bear in mind 1960s hardware bot memory and cpu were enormously expensive; even in early 1990s when you home PC could do the entire BR fleet in one excel sheet, if you did want to increase your RAM or update to a higher CPU you ended up significantly more.

So BR dumbed down the spec, and said no 000; there was probably still a check on operator typos though.

Back in 1968 UIC brought in the 6+1 digit numbering system, with several countries, even eastern bloc, going over to it. BR did not. Popular rumour was "we are not connected to Europe so not needed" but that is not the reason at all - UIC needs a 6 digit field + check digit = 7; BR by usng only 5 digits no check allowed much capital cost sabing.

Do you have evidence for this or is it supposition?

I don't actually understand why 000 would automatically require extra code, though you might choose to reserve it to describe a null entry or some such.

Also it's rather hard for me to imagine a significant capital cost from keeping storage of locomotive numbers down to 5 digits while wagons (far more of these than locomotives) had longer numbers.

Anyway in a previous thread (which may or not be easy to find) there was much discussion along these lines, culminating in someone apparently digging out the specs and finding out that in fact TOPS didn't handle the locomotive file as a 2 digit class + 3 digit number and in fact anything could go in there.

As might be surmised from the fact that Westerns and Warships went in with their old "D" numbers (albiet I think without the D).

 

[norbitonflyer]

The class 45 Peaks were the first main line diesel locos to be renumbered, but that was in the spring of 1973. In 1972 the only renumbered locos were electrics.

My mistake - it was 1973

 

[43096]

The class 45 Peaks were the first main line diesel locos to be renumbered, but that was in the spring of 1973. In 1972 the only renumbered locos were electrics.

41001/2 were delivered as such in 1972. They were of course not renumbered, but new vehicles carrying TOPS numbers.

 

[D7666]

Do you have evidence for this or is it supposition?

I don't actually understand why 000 would automatically require extra code, though you might choose to reserve it to describe a null entry or some such.

Also it's rather hard for me to imagine a significant capital cost from keeping storage of locomotive numbers down to 5 digits while wagons (far more of these than locomotives) had longer numbers.

Anyway in a previous thread (which may or not be easy to find) there was much discussion along these lines, culminating in someone apparently digging out the specs and finding out that in fact TOPS didn't handle the locomotive file as a 2 digit class + 3 digit number and in fact anything could go in there.

As might be surmised from the fact that Westerns and Warships went in with their old "D" numbers (albiet I think without the D).

TOPS software could always do 2+3 anything.

TOPS hardware at first implementation could not.

Back in 1970 hardware cost significant.

In binary, for every number string that is one bit longer than the previous one, your memory needs double. Work that back through every line of code it will add up.

 

[norbitonflyer]

As might be surmised from the fact that Westerns and Warships went in with their old "D" numbers (albiet I think without the D).

Also the Hymeks and the Falcon

Hopefully there was never any confusion of 01001 or 01002 (isolated on the Holyhead breakwater) with "Western Pathfinder" and "Western Explorer", or the first 14 Hymeks with the Southampton Dock shunters (07001-07014).

Although in the 80s a huge amount of coaching stock, including DMUs, were renumbered to avoid duplication with classes 25, 26, 50, 56, 81, 85 and 86, and ex-SR emu vehicles in the 14xxx, 15xxx and 16xxx ranges.

 

[AdamWW]

Also the Hymeks and the Falcon

Hopefully there was never any confusion of 01001 or 01002 (isolated on the Holyhead breakwater) with "Western Pathfinder" and "Western Explorer", or the first 14 Hymeks with the Southampton Dock shunters (07001-07014).

I think the initial zero for the TOPS numbers would distinguish then from 4 digit numbers if it was actually needed.

Although in the 80s a huge amount of coaching stock, including DMUs, were renumbered to avoid duplication with classes 25, 26, 50, 56, 81, 85 and 86, and ex-SR emu vehicles in the 14xxx, 15xxx and 16xxx ranges.

And that's interesting because I'm pretty sure that TOPS stores locomotive numbers in a different database (the "locomotive file"?) to coaching stock and wagons.

So it seems to have been to reduce the risk of confusion not because TOPS couldn't actually cope with the duplication - and indeed until regional prefixes were dropped the coach numbers would be distinct from the locomotive numbers anyway.

I expect someone who actually knows about this stuff rather than just remembers what they've read on web forums will be along in due course to put us all right.

In binary, for every number string that is one bit longer than the previous one, your memory needs double. Work that back through every line of code it will add up.

Or to put it another way, to double your memory capacity you need to double your memory capacity.

 

[D7666]

Or to put it another way, to double your memory capacity you need to double your memory capacity.

Well put.

By not tripping over the threshold of expansion demanding a doubling, you save code => save cost. BR was very good at that sort of thing.

 

[Taunton]

When you entered the loco number on TOPS, were the class number and the 3-digit serial number separate fields?

If the loco was a shunter class 08, did you have to enter the leading zero?

 

[Gloster]

My extremely dog-eared copy 1974 edition of the Ian Allan Locoshed book (LMR dates 16-03-1974 and ER 23-03-1974) has only 45 001-012 and 101-121 with their new numbers. The rest of the 45 are shown with their old numbers and had to be filled in as they were dealt with out of order. There were also twenty-one 86 still to be given their new (out of order) numbers and they were doing various things with the 76. There were also straightforward additions to Classes 27/2, 31/4, 47/4 and 47/6.

 

[AdamWW]

When you entered the loco number on TOPS, were the class number and the 3-digit serial number separate fields?

If the loco was a shunter class 08, did you have to enter the leading zero?

I believe you just typed in what was painted on the side of the loco, as is. (Likewise for cast iron plates...)

4 or 5 digits, with any leading zero.

 

[Taunton]

I believe you just typed in what was painted on the side of the loco, as is. (Likewise for cast iron plates...)

4 or 5 digits, with any leading zero.

Thank you. Then it must have been able to cope with 47000. Or even D1000.

I didn't imagine it would be changed from what Southern Pacific had, who had locos eg 3000, or with an A or B suffix, and never renumbered anything when they devised TOPS.

 

[43096]

I didn't imagine it would be changed from what Southern Pacific had, who had locos eg 3000, or with an A or B suffix, and never renumbered anything when they devised TOPS.

And also didn't have the concept of a class number followed by a serial number.

I have seen it suggested that the renumbering was not about TOPS but was to do with implementation of another computer system, RAVERS.