A visit to the Barclays Bank archive to meet an old acquaintance.

No, not the lady in the picture below; the machine she is leaning on.

A Burroughs TC500 minicomputer publicity shot from the late 1960s

I consider myself lucky to have started my IT career about the time IT 'took off', and I grew with it. I'd joined a bank on leaving school but almost immediately knew that boring environment wasn't for me, and after talking to some Burroughs computer engineers who were installing equipment in the bank I answered an advert by Burroughs for trainee Field Engineers. After a rigorous selection process initially at the Midland Hotel and then at Burroughs Manchester offices I was one of a few of probably over a hundred applicants offered a job and joined the Burroughs Manchester office, assigned to the Banking division in view of my previous job. After initial training I found myself fixing mechanical accounting machines at banks in the city and I wondered if I'd made the right move. But something I didn’t know about was happening which would put my career on the track it was to follow for the rest of my working life.

Barclays, Nat West, and Midland banks had recently placed massive orders with Burroughs for terminal computers to be installed in every branch of those banks, and I was to be trained on this new machine, the TC500.

To put this in context, it was 1970 and few people knew what a computer was back then. They had vague ideas of electronic brains in large air conditioned rooms, costing a fortune and attended by propeller-head boffins in white coats. But through sheer luck, I was in the right place at the right time. The computer revolution was about to take off, and I had a front row seat.

I got to know Southend quite well over the following months as I attended a series of courses on basic computing, data communications, and the TC500 itself at the Burroughs training centre. The TC500 was one of the first office mini-computers, the size of a large desk. It had a full length mechanical keyboard built into the front of the desk, and a golf-ball print head traversing up and down a platen about four feet long on the top of the machine. There was a mechanical paper tape reader for loading programmes, and inside the cases was a rack of printed circuit boards with RTL chips (Resistor Transistor Logic, the forerunner of the later TTL, Transistor Transistor Logic) chips and a head-per-track fixed hard disc spinning at six thousand RPM which not only stored the programmes, but also acted as the machine’s working memory.

The golf-ball print head mounted on its carriage, showing the lead screw shaft which when rotated caused the print carriage to traverse left and right 

To bring the machine to life one pressed the start button and the disc started to spin up, with an ever rising whine like an airliner jet engine starting. After thirty seconds the disc was up to speed and almost inaudible and the TC500 ‘initialised’; that is, the gong sounded, all the indicator lights came on, the golf ball print head slowly traversed to the right hand end of the platen until it hit the end stop, then whizzed quickly back to the left, counting in about half a second the three hundred or so character spaces which would bring it to rest at the left hand end of the platen. The indicator lights extinguished except for those over the four programme select keys, and the machine was ready for use.

The TC500 could operate as a stand alone small office computer, but in the configuration ordered by the UK banks a TC500 in each bank branch was linked to a mainframe IBM 360 computer in the banks’ computer centre. That is why we were trained in data communications. The link comprised a dedicated 4-wire telephone line from a modem in the bank branch to one in the computer centre with a standby 2-wire dial-up link for use if the main link failed. 4-wire line speed was 1,200 bits per second, reducing to 300 BPS for the 2-wire standby link. Amazing to compare that to today’s home broadband (over copper pair, not fibre) giving, say, 8 Megabits per second, or 8,000,000 bits per second. That’s well over six and a half thousand times faster than that dedicated 4-wire link, twenty seven thousand times faster than the 2-wire link of 1970. So today’s broadband speeds come to us over the same 2-wire phone line we used in 1970, but twenty seven thousand times faster.

And those modems – back then they comprised grey tin boxes about eighteen inches square by eight inches deep, with ‘GPO’ enamel badges on them. Today modems have been replaced by modem/routers, but the last modem I saw for use in a laptop PC was smaller than a credit card.

The protocol used over the link between bank branch and computer centre was Burroughs poll – select. The operator in the bank branch would input data on the TC500 in the form of debits and credits to accounts, or request data such as account balances. The central computer would poll round each of many hundreds of TC500s every few seconds to see if there was any data to transfer. If there was, the TC500 would be ‘selected’ by the mainframe, and the data transfer would take place.

TC500s in a bank branch. These two machines are 'concatinated', in that they share a common modem and data link. The GPO modem is the large box on the stand between the two machines.

I became a Field Engineering Group Leader in charge of a team of engineers, and while they looked after machines in the Manchester and Cheshire bank branches I supported them but spent a lot of my time at Barclay's Bank computer centre in Wythenshawe, and later at the Barclays management centre at Radbroke Hall near Knutsford. I and a few of my senior engineers were trained on later and more complex Burroughs computers which Barclays would evaluate at Radbroke Hall.

I stayed with Burroughs for several years before going to DEC (Digital Equipment Corporation) exchanging hexadecimal machine code for octal. DEC which was an amazing company to work for at the end of the '70s, like latching on to a rocket set fair to topple IBM from its no.1 slot in the industry. But by 1980 micro processor based computer circuitry was starting to de-skill the role of computer engineer and I decided on a change in career direction. I left DEC and joined SPL (Systems Programming Ltd) as a project engineer, progressing to Project leader and technical sales support in manufacturing and  process control systems. Through a series of take overs SPL was eventually absorbed into what became EDS (Electronic Data Systems) from where I retired in 2008 as a Solutions Architect on multi-million pound bids into large corporate clients such as Rolls Royce Aero Engines, and public sector departments such as Inland Revenue and DWP. But it had all started with the humble TC500 in the early '70s.

Through Facebook I joined the Barclays Bank Wythenshawe Computer Centre page for old time's sake, and discovered that Barclays have an archive in Wythenshawe (the computer centre having closed a long time ago), and in the archive is a TC500! I hadn't seen one since about 1979 and had assumed they all long been scrapped, so I made arrangements to go and see this first step in my IT career.

Possibly the last remaining TC500 in the Barclays Bank archive this morning

This morning the keeper of the archive, Maria, welcomed me and showed me to the machine and allowed me to take photographs. It certainly brought back memories to come face to face again with this relic of the early 1970s; the first stepping stone in my IT career.

A close up of the print carriage and lead screw. The golf ball head had (if I remember correctly) four positions of tilt and sixteen of rotate, to present the required character to the paper. The shaft to the left of the lead screw is the print shaft; once the golf ball had bee tilted and rotated to the correct character position, a dog clutch engaged and rotated the print shaft one revolution which caused the print head to be thrown against the paper and platen, to print the character. This forward movement also engaged mechanical detents to hold the golf ball accurately in position. To print the next character the lead screw rotated fractionally to move to the next character position, and the process repeated. The lead screw was also detented for each character print. Amazingly, the print speed was twenty characters per second!

The narrow metal tapes visible to the right of the lead screw are what positioned the golf ball in rotate and tilt. These are driven by the 'decoder', a set of electrically-actuated clutches running in an oil bath and operating tilt and rotate arms with pulleys on them that the tapes run round. The single tape to the right of the print shaft selects black or red ribbon, and is also driven by the decoder. The electrical inputs to the decoder clutches came from the machine's electronic print drive system. The decoder was the TC500's Achilles heel, suffering a high failure rate. We became expert at changing them for replacement units! 

A rear view, showing the two multi-pin sockets at the left. The left hand one of these was for the modem cable, the right hand one for the concatination cable if a second TC500 was connected to the modem. 

Barclays Wythenshawe computer centre as I remember it. Note the 'arrow slit' windows to make life easy for the air conditioning, and the SELNEC bus!

A scene at the opening ceremony of Barclay's Wythenshawe Computer Centre in 1971, with the 'arrow slit' windows in the background. As officialdom looks on probably understanding little of the then leading edge technology, an operator demonstrates a TC500. The enormous grey GPO modem occupies the side table.

.