Bletchley museum treasures vintage tech

3 of 11
  • Eight-inch floppy disk drive

    Eight-inch floppy disk drive

    The National Museum of Computing has many fine examples of historical technology, some of which dot the shelves that line the corridors in Bletchley Park's Block H. On a recent visit, ZDNet UK was unable to resist the chance to take pictures of various iconic storage and communication devices, including some that may be familiar to the more mature viewer.

    8-inch floppy disk drive
    The charmingly mechanical insides of an 8-inch floppy disk drive. Introduced in 1971 by IBM as a read-only disk with 80KB of storage, a disk could hold 1.2MB by the time the technology stopped being developed in 1977. Used on mainframes and minicomputers, the drive format was only occasionally found on desktop systems; its unsuitability for personal computing and word processing led to the ubiquitous 5.25-inch drives that dominated  in the late 1970s and early 1980s.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Acoustic coupler

    Acoustic coupler

    Acoustic coupler
    Prior to market liberalisation in the early 1980s, it was illegal to plug unapproved equipment into the phone system — if you could find a plug — and very difficult to get equipment approved in the first place. A common compromise was the acoustic coupler, which played data tones into the microphone and picked them up from the earpiece of a standard telephone handset.

    This was suitable for very slow speeds — 300bps or less — and was also very sensitive to ambient noise, being knocked, poor-quality microphones and many other ills. However, it was very flexible: combined with a Tandy Model 100 portable computer, it could give portable dial-up access from phone boxes, thus creating the very first generation of mobile data access.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Anamartic wafer memory

    Anamartic wafer memory

    Anamartic wafer memory
    One of the ironies of computer memory is that multiple chips are made on one silicon wafer, expensively cut up and packaged, and then used next to each other in large numbers. Why not wire the chips together on the wafer, linking past ones that don't work, and just use that? That was the thinking by Ivor Catt, a British inventor, who sold his idea to Sinclair Research — thence a company called Anamartic, which got the results into production.

    Unfortunately, by the time the product hit the market in 1989, the cost of individual chips from the Far East was so low that there was no advantage in using what had proved to be a reasonably expensive way of doing things and Anamartic closed three years later.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • BT Merlin Tonto

    BT Merlin Tonto

    Tonto
    The Tonto was a rebadged ICL OPD — One Per Desk — an unusual project that saw UK mainframe company ICL take the internals of a Sinclair QL and rebuild them in a desktop computer that included telephony and a 1200/75bps modem. The idea was that this would be a universal office PC, with built-in Psion XChange software — word processor, database, drawing and spreadsheet functions — and microdrive data storage.

    The end result was surprisingly usable and on paper had a very good chance of establishing itself as a useful system. However, poor marketing and general bafflement saw it relegated to history's footnotes. Perhaps its most endearing feature was a voice synthesiser designed for answering machine messages; it had a vocabulary of a couple of hundred words with an office theme, but it was possible to make it say mildly racy things — "I am having my secretary under the table. Please call back".  

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Bubble memory

    Bubble memory

    Bubble memory
    For a couple of years in the mid-1970s, bubble memory seemed to be the way ahead. It works by shuffling tiny magnetic domains around a sheet of orthoferrite — an iron/rare earth/oxygen compound — by putting current through the sheet. It was denser than other memory systems, very robust, reasonably low-power and cost-effective. The bubble memory shown here is Intel's 7110, a 1Mb device, that saw use in a few laptops and embedded systems in the early 1980s. The unit here was made in 1984, towards the end of the technology's relevance.

    Ordinary solid-state memory developed faster than bubble memory and soon saw it off; it was always much faster, but soon became denser, cheaper and lower-power. 

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Ferranti Pegasus Delay Line Memory

    Ferranti Pegasus Delay Line Memory

    Ferranti Pegasus Delay Line Memory
    Hailing from around 1957 and using three valves, this unit from the British designed and built Ferranti Pegasus computer stores the equivalent of five bytes of data in a long nickel delay line that's coiled up below the paxolin panel seen here. The Pegasus actually used 40-bit words as its basic unit of data, one of which could be stored here. The line worked through data being fed in at one end by the equivalent of a speaker, and taken out again on the other end after it had spent some time travelling along the delay line as sound waves.

    The designer of the delay line, John Fairclough, was a junior engineer at the time: he finished his career as Sir John Fairclough, chief scientific adviser to Margaret Thatcher.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Ferrite core memory

    Ferrite core memory

    Ferrite core memory
    Invented in 1955, ferrite core memory was the mainstay random access storage medium for mainframe and minicomputers throughout the sixties. It works by hundreds or thousands of tiny O-shaped ferrite rings — ferrite being a iron-containing ceramic — that can be magnetised when current passes through two wires threaded through their centre.

    This magnetism can be read back by another wire — the sense wire, which also passes through the core. However, to get this signal, the ferrite has to be demagnetised by a pulse of electricity through the first wires. Thus, once read, the ferrite has to be reset back to its original state. Density is low, with one bit per ferrite, but planes of core memory.

    When not being read, however, ferrites were non-volatile; you could switch a computer off overnight and turn it back on with memory contents intact. The first solid-state memory replacements for core store had huge batteries to try and replicate this behaviour, until it was realised that reloading off tape was almost always more sensible.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Post Office modem

    Post Office modem

    Post Office modem
    Something of a mystery object, this is probably one of the first data modems used in the UK. With separate data modulation and demodulation units, together with a control unit and a power supply, this modem is as one would expect an FSK (Frequency Shift Keying) design, with two oscillators supporting two channels, one for data zero and one for data one.

    If this unit was designed to work over ordinary telephone lines, it will be no faster than 300bps. That seems plausible, as it seems to have an auto answer mode, which wouldn't have been needed on a leased line.

    However, any further information will be gratefully received. Data Modulator No. 9A has secrets yet to plumb.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Psion Protea Prototype

    Psion Protea Prototype

    Psion Protea Prototype
    There's a huge story to tell about Psion, one of the UK's most innovative consumer electronics companies and one of the few that could outdesign anyone on the planet. Protea was the codename for the Series 5 PDA, which remains one of the benchmarks for elegance, functionality and plain fit-for-useness.

    Running an ARM chip from two AA batteries that lasted up to 20 hours, it had a half-VGA resolution backlit greyscale LCD and an utterly charming slide-out keyboard that, in the opinion of many, has yet to be bettered in anything this size. The operating system, EPOC32, became Symbian, and came with built-in spreadsheet, word processing, database, diary and contact management.

    It was never actually produced in yellow. Shame.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • Psion Series 5

    Psion Series 5

    Psion Series 5
    Another variant of the Series 5 which was never commercially available — this time with a translucent case that brought your correspondent close to actual aggravated theft.

    The Series 5 was developed in under two years, including new silicon, software and casework. Psion was especially proud of its software: the word processor could embed graphics and spreadsheet, had a spell checking, outliner, multiple format options and was near instant in operation — and it took 20KB of memory, less than an empty Microsoft Word document. There's an excellent article on this and more by Andrew Orlowski of The Register.

    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

  • WS2000 modem

    WS2000 modem

    WS2000 modem
    The favoured modem of the early British hacker, the WS2000 was an approved modem from the early 1980s that packed just about every contemporary standard into one box. It was nevertheless quite a simple design, based on the AM7910 World Chip single chip modem from AMD (yes, that AMD), but was clever in the way it allowed the host computer to control the modem's functions, including autodial and answer, by pulsing various control lines.

    Although UK telecom approval forbade the use of the Bell American modem standards, the WS 2000 had them — and got away with it by having an end-stop on the rotary switch that prevented their selection. It was the work of a moment to remove that stop.

    Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.


    Photo by: Rupert Goodwins

    Caption by: Rupert Goodwins

3 of 11

Anamartic wafer memory

Anamartic wafer memory
One of the ironies of computer memory is that multiple chips are made on one silicon wafer, expensively cut up and packaged, and then used next to each other in large numbers. Why not wire the chips together on the wafer, linking past ones that don't work, and just use that? That was the thinking by Ivor Catt, a British inventor, who sold his idea to Sinclair Research — thence a company called Anamartic, which got the results into production.

Unfortunately, by the time the product hit the market in 1989, the cost of individual chips from the Far East was so low that there was no advantage in using what had proved to be a reasonably expensive way of doing things and Anamartic closed three years later.

Caption by: Rupert Goodwins

3 of 11

Related Topics:

Innovation Hardware Reviews

Related Galleries

    • 1 of 3