Jump to content

Early BR computers memories and history


Recommended Posts

I have obtained a copy of a book called "Railways as a Career" which was published in the late 1950's and makes fascinating reading about the major changes British Railways were undergoing. As someone who worked in computing for BR in the mid 1970;s one line hit me and that was the BR had installed an Analogue Computer for working out timetables, I was going to test the collective knowledge of the RMWeb community to see if any more was known about this. However I did an internet search and this came up:

 

https://www.ebay.co.uk/itm/333445126554

 

which partially answers my question. 

 

However the question still stands - does anyone know anything more about this?

Edited by MyRule1
Change title
  • Informative/Useful 1
Link to post
Share on other sites

7 minutes ago, MyRule1 said:

I have obtained a copy of a book called "Railways as a Career" which was published in the late 1950's and makes fascinating reading about the major changes British Railways were undergoing. As someone who worked in computing for BR in the mid 1970;s one line hit me and that was the BR had installed an Analogue Computer for working out timetables, I was going to test the collective knowledge of the RMWeb community to see if any more was known about this. However I did an internet search and this came up:

 

https://www.ebay.co.uk/itm/333445126554

 

which partially answers my question. 

 

However the question still stands - does anyone know anything more about this?

ISTR an article in Modern Railways about this. Let me have a gander through them and I'll get back to you.

  • Thanks 1
Link to post
Share on other sites

  • RMweb Gold

The earliest application train planning type of application I know off which 'sort of worked' was the computer installed in Great Northern House at Kings Cross - which you could see from the pavement outside - in, I think, the early 1960s.  It didn't last for more than a year or two.

 

I suspect the one in the article linked above was used. or tried out, by what we used to call 'the Derby computer people' who originally did their train performance calculations the old fashioned way but subsequently moved to electronic computation using a computer.  What is described is very obviously not a timetable computer but is clearly some sort of train performance calculator which would have been used to derive point-to-point running times for different traction types with different loads - a science which, in some respects, went back to the early 20th century.  By the late 1980s the Dertby Train Performance group had got things down to an extremely sophisticated level and the figures they produced for Class 60s were found to be very accurate in the real world.  However the 'raw' times which they produced, even that late, had to be adjusted/rounded to the nearest minute or half minute in order to use them for timetabling purposes.

 

Using computers for timetabling work came in two forms - the simplest being what amounted to an adding machine which was then developed/upgraded to be able to draw graphs.  The latest BR systems had got to that standard by the late 1980s and had been developed for use on micro (i.e. desk top) computers.  The principle system initially in use being TBM (Timing By Micro) which worked but was cumbersome.  it was succeeded by a far better system - which had the graphing facility - in the late '80s - and which was very advanced for its time in comparison with timetabling software used in many other countries such as the exceedingly cumbersome Thor system still being used by SNCF in the mid/late 1990s.

 

But all these system were 'dumb' in that effectively they were no more than adding machines which could draw a graph; they could not achieve the holy grail of computerised timetabling - conflict resolution.  The later requires very sophisticated relational databases and a lot of conmputing power especially for a complex network such as Britain's.   There was a private company working to develop that sort of thing in the 1990s and their general timing program was quite impressive but it always came with 'jam tomorrow' capabilities which dissuaded me from buying it for the TOC I then worked for.  Eventually by about 1996/97 they achieved some form of conflict resoluton but it was pretty basic but it did work reasonably well for single lines railways and they sold it to Israeli Railways.  Eventually I think Railtrack bought their software although at that time - in the late '90s it still couldn't do proper conflict resolution on anything approaching a complex railway and RT were, I suspect, over ambitious in what they thought it would be able to do for them.

 

My involvement which such things finished 20 years ago and computing moves rapidly and with ever increasing sophistication plus the development of AI so no doubt more is possible nowadays.

  • Informative/Useful 8
  • Interesting/Thought-provoking 1
Link to post
Share on other sites

  • RMweb Gold
9 hours ago, The Stationmaster said:

.........some sort of train performance calculator which would have been used to derive point-to-point running times for different traction types with different loads - a science which, in some respects, went back to the early 20th century.  

:offtopic:

In our railway careers we all probably came across some bit of historical treasure we wish we could retrieve now. One of mine was a proposed train performance graph for the Coronation Scot from Euston to Cheddington. It had been produced manually probably based on some dynamometer car data. The times were calculated per quarter mile in minutes to 3 decimal places, thus 1/4 mile at 70mph was shown as 0.214 minutes

  • Like 3
Link to post
Share on other sites

  • RMweb Gold

Anecdotally the GN House computer, which was intended, with a fanfare of public trumpets, to write the new WTT for the Deltics on ECML, actually got the first train as far as Finsbury Park and that was that. By the time I had an office on the 7th floor there in 1990, there was no sign, although I don't recall much use of the ground-floor room in which the computer had been housed.

  • Like 3
  • Informative/Useful 1
Link to post
Share on other sites

  • MyRule1 changed the title to Early BR computers memories and history

I have changed the title of this thread to accommodate the interesting memories already posted and hope for more.

 

Looks like @Oldddudders might have inherited the office we had in the 1970's when I worked of a state of the art ICL 2904 located there.

Link to post
Share on other sites

  • RMweb Gold

For other computer / Solid State electronics applications BR was fairly well advanced. I remember our payslips going over from hand-written to computer printed just aftrr I started in 1966. 

Solid state was being used for TDM and FDM remote controls and transistor logic was being trialled for interlockings. I actually removed an experimental one of the latter which had been ghosting a working relay interlocking in 1970.

  • Agree 2
Link to post
Share on other sites

I remember around 1992 getting a quote for a timetable graph from a company of consultants and was horrified at the price so wrote a computer program in Turbo Pascal (it was a long time ago) which did the same thing. Retiring at the start of 2021 I rewrote it using Visual Studio and used it for some work on the extended Borders Railway. Started on the timetable conflict resolution but other things got in the way!  Inputs were from a train simulation package which used the unit characteristics, grades speed limits etc.  Currently using it for my proposed WHL model.

Extended Borders Timetable Graph.png

  • Like 8
  • Interesting/Thought-provoking 2
Link to post
Share on other sites

  • RMweb Gold

One of the lesser known aspects of BR's computng was what was known as the 'iron ring' which was a dedicated data link connecting various main centres with the BR mainframe computing centre in Nottingham and it was an early conversion to fibre optic cable.  But overall BR seems to have been far better and more efficient at quickly electronically moving large quantities of data and even the excellent Micromail system could move data in very large files long before commercially available systems such as email.  When Railtrack converted from Micromail to email in the latter half of the 1990s it was no longer possible to transfer the working timetable by electronic transfer because the files were too large and it had to be moved around as a dat-tape which meant sending a person to collect it, or to meet the motorcycle courier who was carrying it,  instead of sitting at their desk and downloading from the mainframe. 

 

Another little known fact was that traditional BR line connections were way ahead of BT in that the baud rate used by BR was much higher than that used by BT.  When we had some fax machines installed as part of TOPS trial site introduction in South Wales in 1973 it was found that BT lines could not handle them and BR circuits had to be extended along  what was by then a colliery branch line to enable the fax machines to be used.

 

As far as signalling was concerned the WR installed a trial/experimental solid state interlocking at Henley-On-Thames in 1961 although it used discrete components - no ICs etc back then.  It worked with a standard WR Integra domino panel  and controlled relays for the ground functions and the circuit controllers on FPL levers as the points remained being worked from the VT lever frame;  Mullard supplied the electronics and - from what I can recall from 'outside the fence'  - it took at least three weeks to get it all working properly.

  • Like 2
  • Informative/Useful 7
Link to post
Share on other sites

  • RMweb Gold

The use of BR's data system was, as Mike has said, quite intensive. ISTR an app called Telegram Sam being used for large data packets. By the mid-80s, Board HQ was running an investment programme called IMPS, which was in effect a distributed database, written by staff in the Financial Planning Dept, in which Sub-Sectors could enter their data and it would be added in for analysis at HQ. 

 

BR had also twigged that the fibre links installed for signalling schemes could be enhanced to have dark fibre available for commercial exploitation, as well as adding considerable resilience to the BR telecoms system.

 

My first video conference was in 1993 with a team from offices near Kings Cross talking to the Nottingham team. 

 

And back in the 1930s, Southern Railway had installed a new telephone exchange at Waterloo that was so advanced that GPO engineers came on official visits to inspect it. 

  • Like 5
  • Informative/Useful 2
Link to post
Share on other sites

My memory of BR computing? Not being paid on my first payday (September 1989) because the payroll tape had gone awry in some way. I had a huge credit card bill from a summer of bliss spent InterRailing around Europe.

 

Second memory - dot matrix printers. My office had which would spew out things like control logs and the like. Night shift was 2100-0700. At about 0620 each morning the dot matrix would churn out the previous day's control log with a cacophany. Quite convenient, actually, as it didn't look good to be asleep when one's colleague on the day shift arrived....

  • Like 4
  • Friendly/supportive 1
Link to post
Share on other sites

I remember the payroll tape going astray ! The story we were told was that someone had sent the backup tape to the bank instead of the real one and it had been (correctly) rejected. 

 

The 'iron ring' was, I believe, the BR High Speed Data Ring - ring User Control quoting the LNIATA number but first take the Comms lead out of the back of whatever wasn't working and blow on it of course. It was remarkable how often that cleared the fault ! 

  • Like 4
Link to post
Share on other sites

  • RMweb Gold

BR was quite advanced on the comms side. When I started in the mid 1960s we had the STRAD (Signal Transmitting, Receiving and Distributing) teleprinter network which started on the LMR c1962. The only other installations of this system were military. The main computer recorded incoming messages then distributed them to the appropriate receiving stations around the system. 

Our Extension Trunk Dialling telephone network was only exceeded by the MOD as far as private phone networks with automatic exchanges were concerned. 

We were early users of Fax technology, co-ax and fibre optic transmission systems for data transfer. In the mid 1980s we were using remote terminals for CAD design in outbased S&T offices and were early users of the Lotus Notes email system introduced c1989/90, about six years prior to the introduction of Hotmail on the internet.

  • Like 3
  • Agree 1
Link to post
Share on other sites

  • RMweb Gold

Talking of computerised central payroll arrangements on the WR the second stage of that was introduced in 1968 when I was working in the London Diviion HQ at Reading.   The first stage back in 1966 seems to have gone pretty smoothly apart from folk doing timesheets who had to change over to the new way of working and some didn't like it one bit.  But in 1968 the data processing was centralised including all the deduction for the entire BR network and I might still have somewhere a copy of the code list we all received so we could check our deductions.  

 

Anyway the first new paybill at Reading many of us had a strange code and a small deduction allocated to it and on checking against the list we found that we were all contributing to a former LNER saving/pension fund for Hull and Grimsby Dock Staff.   It turned out that 'somebody' entering the codes had allocated us to that instead of to the LNER Superannuation Fund; always pays to read the small print

  • Like 1
Link to post
Share on other sites

I was surprised recently when this came into my hands

 

tops.jpg.e8e8172725d88548d0fd8c85407f24f4.jpg

 

It's a tops print out for the "Loco's available at Temple Mills at 20:29 on 29.03.74", the picture above is taken on a mobile phone  but I will be scanning it fairly soon.

 

The loco's listed are:

 

 

 

25/1 25025 Tl

25/2 7522 WN

31/1 5506 SF

31/1 5510 SF

31/1 31002 SF

31/1 31008 SF

31/1 31015 SF

31/2 31102 MR

31/2 31107 MR

31/2 31148 MR

31/2 31151 MR

31/2 311GU MR

31/2 31185 FP

31/2 31197 MR

31/2 31202 FP

31/2 31204 MR

31/2 31214 FP

31/2 31312 MR

33/1 33002 EH

33/1 33020 EH

33/1 33037 HG

33/1 33040 HG

33/1 33045 HG

33/1 33052 HG

33/2 33111 EH

33/3 Uknown

33/3 53211 HG

35/0 7028 00

35/0 7193 OC

37/0 6967 SF

37/0 37017 MR

37/0 37028 MR

37/0 37073 MR

37/0 37078 MR

37/0 37090 MR

40/0 400870 YK

47/0 47150 SF

47/0 47173 IM

47/0 47447 ?

73/0 71001 AF

71/1 71010 AF

71/1 71012 AF

73/2 73117 SL

73/2 73131 SL

73/2 73141 SL

 

Note the two class 71 electric loco's!

 

I wonder how many other early TOPS reports still exist?

  • Like 2
  • Informative/Useful 1
Link to post
Share on other sites

  • RMweb Gold

I recall from my brief time in Crewe in 1988 a team there was developing a system called TBM (Timetable by Microcomputer).

 

Does anyone know how that project developed?

Link to post
Share on other sites

I went into TOPS in 1981, just at the back end of card TOPS using the small cards, that superseded the original IBM cards. When I was training at Webb House, (another infamous BR location as anyone who went there will testify) we were given a stern warning not to fool about with the card chads from the punch hopper, and not to try playing cassettes in the programme player on the top of the terminal.  

 

The Ventek terminals we used had a screen roughly the size of a paperback book, and seemed to be the cause of the majority of TOPS clerks needing to wear spectacles, at some stage of their career. Remarkably, the terminals were subject to monthly inspection and servicing by Ventek engineers, our office at Dover was always a popular job, as the task included a visit to the office at Dunkerque to service the terminals there.

 

The Stationmaster mentions Baud rates, it was always noticable the speeds, I seem to remember a rate of 500 or 600 was normal speed, when we had a line failure S&T would be summoned to set up an acoustic coupler to get us back on line, the rate on that was around 200, printing a train list wou,d take an eternity.

 

Another strange device we had was a Mufax, single terminal fax machine between us and the Town yard. Used to make a hell of a racket, as it printed out via a blade onto some truly evil wet paper soaked in a mixture that smelt like a cross of parafin and meths.  One shunter was caught short one Sunday in the yard and grabbed a roll to use as toilet paper, 30 minutes later he was on his way to hospital, unable to sit down with severe chemical burns to his backside.

 

 

  • Like 7
Link to post
Share on other sites

  • RMweb Gold
19 hours ago, Colin_McLeod said:

I recall from my brief time in Crewe in 1988 a team there was developing a system called TBM (Timetable by Microcomputer).

 

Does anyone know how that project developed?

It became a standard BR timing system - not bad actually although i don't think it was developed at Crewe as most timetable related systems were under the control of a group based at York with mainframe development work done at Nottingham.  There was national representative group which dealt with all train planning technological and systems matters which was chaired by the man who headed the group in York and who was responsible for TSDB (Train Service Database) which held all the timetable information for the whole of BR and from which other systems bled off whatever information they needed.  Stupidly I can't remember the name of that representative group despite being a member of it from 1989 t0 1994:scratchhead:.

 

I was member of BR's Train planning training Group also from 1989 to 1994 (latterly being the TLF rep on the group) and we organised one week long train planning training courses at Webb House plus arrnging the necessary speakers and training material so I was well familiar with Webb House as mentioned above by Simon Lee.

 

Oddly also referring to his post above I was assistant AM at Radyr in 1973 when it became a TOPS trial site and was one of the (if not the) first site(s) to be equipped with Ventek machines on trial for wider use instead of te cumbersome IBM machines.  Our terminals arrived direct from Texas and one of them had to be immediately reprogrammed because it had originally been set-up fora hospital in the USA so was configured for TOPS.   They also had an amusing 'keyboard training' feature which allowed a couple of simple games to be played on them.

 

And I'nm fairly sure that our Mufax machines were some of the first to be installed on BR 

  • Like 2
  • Informative/Useful 1
Link to post
Share on other sites

25 minutes ago, The Stationmaster said:

It became a standard BR timing system - not bad actually although i don't think it was developed at Crewe as most timetable related systems were under the control of a group based at York with mainframe development work done at Nottingham.  There was national representative group which dealt with all train planning technological and systems matters which was chaired by the man who headed the group in York and who was responsible for TSDB (Train Service Database) which held all the timetable information for the whole of BR and from which other systems bled off whatever information they needed.  Stupidly I can't remember the name of that representative group despite being a member of it from 1989 t0 1994:scratchhead:.

 

I was member of BR's Train planning training Group also from 1989 to 1994 (latterly being the TLF rep on the group) and we organised one week long train planning training courses at Webb House plus arrnging the necessary speakers and training material so I was well familiar with Webb House as mentioned above by Simon Lee.

 

Oddly also referring to his post above I was assistant AM at Radyr in 1973 when it became a TOPS trial site and was one of the (if not the) first site(s) to be equipped with Ventek machines on trial for wider use instead of te cumbersome IBM machines.  Our terminals arrived direct from Texas and one of them had to be immediately reprogrammed because it had originally been set-up fora hospital in the USA so was configured for TOPS.   They also had an amusing 'keyboard training' feature which allowed a couple of simple games to be played on them.

 

And I'nm fairly sure that our Mufax machines were some of the first to be installed on BR 

Hi Mike,

Did you come across a chap from York called David Leonard? I seem to recall he was often dashing to either Nottingham and/or London for TOPS and Train Planning type meetings.

Link to post
Share on other sites

I remember in the late 70s/early 80s attending a line fault in the Doncaster Decoy TOPS office. At one end of the office they had MUFAX/NEFAX machines, and at the other end of the office, what was then the latest in printer technology, the Philips Copy 80 printer. Such a wonderful mix of technologies on the BR(E) Telecoms network at that time, from open-wired pole routes still on some areas, to fibre-optic cables and digital transmission systems being installed on other areas - and technology-wise, everything else in between these two extremes keeping the railway running on a daily basis.

 

Happy times.

  • Like 3
  • Agree 1
  • Interesting/Thought-provoking 1
Link to post
Share on other sites

On 21/08/2021 at 23:09, Wheatley said:

I remember the payroll tape going astray ! The story we were told was that someone had sent the backup tape to the bank instead of the real one and it had been (correctly) rejected. 

 

The 'iron ring' was, I believe, the BR High Speed Data Ring - ring User Control quoting the LNIATA number but first take the Comms lead out of the back of whatever wasn't working and blow on it of course. It was remarkable how often that cleared the fault ! 

With the HSDR soon to be superseded, perhaps not too surprisingly, by VHSDR - Very High Speed Data Ring. 

  • Informative/Useful 2
Link to post
Share on other sites

Being quite a newby I started using Protim and DIADS way back in 1989 at Swindon.... 

 

Planning computer systems have moved on a bit since then now with working conflict detection and traincrew optimisation.

 

Neil

Link to post
Share on other sites

My grandfather worked for ICL and did some work installing computers for BR (as well as removing redundant kit).

 

At least some of the scrap equipment he removed formed the basis of an electronic quiz game he and my uncle constructed for a family friend.

 

 

My father accidentally caused a BR computer to malfunction. He was doing some insulator washing experiments for CEGB at Carrington substation to find out how pure the water needed to be before it caused flashover. The experiment was carried out by gradually increasing the level of impurities until the flashover occurred. When it did, it caused a BR computer to trip out and set all the signals to red between Manchester and Crewe! (The experiment was carried out at night). BR were actually quite thankful as it highlighted a weakness in their systems!

 

  • Like 4
  • Interesting/Thought-provoking 1
Link to post
Share on other sites

32 minutes ago, Neils WRX said:

Being quite a newby I started using Protim and DIADS way back in 1989 at Swindon.... 

 

Planning computer systems have moved on a bit since then now with working conflict detection and traincrew optimisation.

 

Neil

I started in timetabling in Swindon in 2002 - and we were still using Protim for much of the work.  It was looking very dated by then.  Trainplan was the successor, with better train graphs and a better interface (in my opinion).

Will

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...