Infinite loop: the Sinclair ZX Microdrive story
The rise and fall of a 'revolutionary' storage technology
Archaeologic They would, Clive Sinclair claimed on 23 April 1982, revolutionise home computer storage. Significantly cheaper than the established 5.25-inch and emerging 3.5-inch floppy drives of the time - though not as capacious or as fast to serve up files - ‘Uncle’ Clive’s new toy would “change the face of personal computing”, Sinclair Research’s advertising puffed.
Yet this “remarkable breakthrough at a remarkable price” would take more than 18 months more to come to market. In the meantime, it would become a byword for delays and disappointment - and this in an era when almost every promised product arrived late.
Sinclair’s revolutionary product was the ZX Microdrive. This is its story.
At the launch of the Spectrum, held on the first day of the Earl’s Court Computer Fair, Sinclair presented a prototype Microdrive to assembled journalists and provided a brief overview of the device’s promised capabilities. Many of the gathered hacks were impressed.
“Perhaps the biggest rabbit that Clive pulled out of his magician’s hat was the ZX Microdrive,” wrote Popular Computing Weekly’s Nick Hampshire for the paper’s 6 May issue. “This is a very tiny disk drive using two quarter-inch diskettes, with each diskette capable of holding 100KB, and a transfer rate of 16KB per second. You will be able to connect up to eight of these drives to the ZX Spectrum... The price: £50.”
That was indeed a “remarkable price”. Commodore’s single-disk 5.25-inch drive for the Vic-20, launched around the same time, was priced at just under £400 - eight times the price of the Sinclair offering. Even Sony’s aggressive launch price for the 3.5-inch drive it brought to the UK in November 1982 was still, at £235, much higher than that of the Microdrive.
Sure, the Vic-1540’s disk capacity was greater - 174KB per disk rather than the 85KB the Microdrive would eventually offer, and the Sony diskette could hold even more - 218KB - but who cared when the up-front cost of the Sinclair drive was so much less?
There was no question, though, that the Microdrive wouldn’t take some months to come to market. At the launch of the Spectrum, Sinclair admitted the Microdrive wouldn’t go on sale until later in the year. Over the coming months, the trials and tribulations of getting the Spectrum into production and shipped out to buyers - the company wouldn’t even begin to get over the order backlog until well into the autumn of 1982 - before the focus of complaints about broken Sinclair promises would shift to the Microdrive.
“Our Microdrive, when it comes out, will revolutionise mass storage thinking,” said Sinclair engineer Martin Brennan in the 18 November 1982 issue of Popular Computing Weekly after being asked about the new Sony drive. His take: yes, 3.5-inch unit was faster, but it’s the price that really matters. Of course, what Brennan couldn’t then answer was the question posed by his own comment: when would the Microdrive actually come out for people to buy?
As 1982 came to an end, the section within Sinclair’s Spectrum advertising describing the Microdrive began to change. Ads from 1982 had described the product as a “a single interchangeable microfloppy”. But by the end of the year, that text was changed to read “a single interchangeable storage medium”. The release window went from “later this year ” to “the early part of 1983”.
By the Spring of that year there was still no sign of the promised product. At least Sinclair wasn’t yet taking orders. The continuing no-show prompted Sinclair Research computer division chief Nigel Searle to issue an update: “The design has been finalised and we are now waiting for custom-made semiconductor chips - being manufactured in the UK - to arrive,” he told the press.
“The delay on the Microdrives has been the result of mechanical difficulties we had not foreseen,” he added. “These have now been solved along with an improvement in the performance of the drives. They are now much more reliable than we had hoped to achieve... The designers would like to go on and on making improvements. But a line has now been drawn.”
By now it was being rumoured that the Microdrive was not based on rotating disk technology, as the early advertising’s reference to a “microfloppy” had implied. Indeed, most attendees left the Spectrum launch assuming from what they’d heard that the Microdrive used some kind of diskette. Instead, it would use a “high speed tape loop”, as Your Computer put it at the time. In fact, hints that the Microdrive was not a disk system had begun to slip out much earlier, toward the end of 1982. Martin Brennan’s aforementioned comments on the introduction of the Sony 3.5-inch disk noted that “a floppy disk based system will be much faster than the Sinclair Microdrive” - a sure sign that the Sinclair offering did not, after all, use diskettes.
Eventually, Sinclair Research’s development work was done and, in July 1983 - after some grumbles by the Advertising Standards Authority - the company announced it would start shipping drives, along with the ZX Interface 1 add-on required to hook them up to a Spectrum, the following September. The first 1000 units would be offered specifically to the folk who had first ordered Spectrums, spun as compensation for having had to wait so long for their new computers back in 1982. It was confirmed that each 43 x 30 x 5mm Microdrive cartridge contained not a disk but a loop of tape, 2mm wide and claimed by Sinclair to be made of same same material as high-quality videotape and not what you’d find in an ordinary audio cassette.
By all accounts, the Microdrives were always intended to be tape-based technology. So how did they come about?
Sinclair had an early introduction to the notion of tape-loop storage in the summer of 1974. A young engineer called Andrew Grillet, who had indirectly done some work for Sinclair Radionics through a stint at one of its sub-contractors, was interviewed for a job with the company. According to Grillet, when told that Sinclair was “going to build a computer” and asked what he ideas might bring to the project, he proposed a data-storage system based on the eight-track music cartridge popular in the early 1970s before Philips’ Compact Cassette format took hold.
“This would be an ideal thing,” Grillet recalls telling the Sinclair staffers, “because you could do roll-out, roll-in swapping. You’d need to have the tape shortened so you had two 64KB memory images on the track only, so you dump one and read the other, or at least if you had to wait for it to go round and switch to another track you wouldn’t have to wait too long.
“And that’s basically what the Microdrive was, except the eight-track cartridges were enormous and they shrunk the cartridge down a bit.”
Speaking to The Register nearly 40 years on, Grillet doesn’t remember who he was interviewed by at Sinclair, but they were impressed by his abilities and ideas: he was subsequently offered a job, “but Xerox offered me twice as much money, so I went to work for Xerox”.
Microdrive precursor: the Extratron Stringy Floppy
He thought no more about the tape storage system until April 1982 when Sinclair Research announced the Microdrive. In his 1985 book, The Sinclair Story, Rodney Dale, a one-time Sinclair Research employee, claims Sinclair product head, Jim Westwood, and Sinclair’s Chief Engineer, David Southward, jointly conceived the Microdrive in 1982. That was eight years after Grillet’s interview. Were these men his two interviewers, and had his notion remained hidden in the back of one or the other’s mind until a new need to deliver a better-than-tape, cheaper-than-diskette format arrived? We will probably never know.
It’s interesting, however, to note that Brian Flint, a long-time Sinclair electronics designer who worked on both the ZX80 and ZX81, mentioned in a recent interview with the website Polymath Perspective that “I think Jim Westwood was working on a tape drive so you could download software from an audio cassette tape” during the ZX80 development process. Of the ZX81, he added, “they did have a tape drive to go with it so you could load in programs”.
Both computers used regular cassette decks for storage. Was Westwood working on a Sinclair-branded version, or is this a reference to a unit not unlike what Grillet claims was his idea? Flint stopped working for Sinclair Research before the announcement of the Spectrum’s ZX Microdrive, but other former Sinclair staff don’t recall any effort to build an ordinary tape player at any stage.
Grillet wasn’t the only one to come up with the notion of looped tape storage. In 1979, Sunnyvale, California-based Exatron began pitching what it called the “Stringy Floppy”, a $100 device which took its own endless-tape cartridges, which Exatron called “Wafers”. Each wafer could hold 70KB of data on a loop of 1/60-inch tape. The company claimed 16KB programs would load “in less than 20 seconds”. Exatron launched the Stringy Floppy system for Tandy’s TRS-80 micro.
Like the Microdrive, the Stringy Floppy wasn’t a random access system, but by moving the tape sufficiently quickly it was possible to make it appear to offer a kind of pseudo random access. At best the file you were seeking was just ahead of the read head’s postion, at worst the device would have to spool right through the tape to reach the requested item.
Back at Sinclair, David Southward, who oversaw Sinclair Research’s work on peripheral devices, took overall control of the Microdrive project in 1982, and put the analogue electronics work in the hands of Ben Cheese, an Electronic Design Engineer. Ben passed away in 2001. He was just 46 years old, but beyond the Microdrive and other early work at Sinclair, he co-founded Flare Technology to develop a prototype games consoles for Sinclair and Amstrad, and he worked at Argonaut Software to create the Super FX graphics chip used in SNES games, including Star Fox.
“Ben was a quiet, intense guy, who was as comfortable with analogue design as digital. Most of us digital guys would not touch analogue stuff. There was a neatness and elegance to his work that was a delight," says John Mathieson, one of Cheese's colleagues at Sinclair.
Inside the Microdrive
Both Southward and Cheese were based at Sinclair Research’s facility in The Mill, St Ives, Cambridgeshire. In a Sinclair staff guide of 1982, Cheese is described as “currently designing the Micro Floppy electronics”. Meanwhile, over at Sinclair’s King’s Parade office in Cambridge, the company’s industrial designer, Rick Dickinson, worked on the look of the drives and the cartridges.
Dickinson, who in 1986 left Sinclair to establish his own design agency, which he is still running today, says the look of the Microdrive came directly from the Spectrum itself, the unit replicating the computer’s plastic casing with its distinctive raised rear section - there to make room for the TV modulator in the case of the computer - and even featuring a black-painted aluminium upper faceplate, as per the Spectrum. There, the metal sheet covered the rubber intra-key areas of the keyboard; on the Microdrive it was there simply for the look. As for the cartridges, says Dickinson, “once you’ve got a place for the exposed tape, a grip for when it’s being removed from the drive, a place for a label and a small case to keep it in, it’s pretty much designed itself”.
Ben Cheese in Sinclair's 1982 Staff List
The ZX Interface 1, which would connect the Microdrive to the Spectrum, was engineered by Martin Brennan, who would go on to work on Atari’s Jaguar console and, in the past decade, to design and market the Brennan JB7 digital music storage and player system. Brennan designed the Interface 1’s electronics and produced the unit’s Rom chip, writing its networking code himself. The mechanical design was handled by John Williams.
Brennan came to Sinclair Research in late 1982, having been offered an open-ended engineering role with an emphasis on artificial intelligence design by Nigel Searle on the back of his work on games software. Brennan today recalls becoming involved in the Microdrive project because, on his arrival at Sinclair, he was keen to see one of the promised ‘revolutionary’ drives and asked if he could. This was seven or eight months after the drive had been announced. It became immediately clear to him that there was still a lot of work to do. “The mechanical design had been done, as had the industrial design, by Rick Dickinson,” he remembers. “But almost nothing had been done on the electronics side and nothing on the software.” Brennan took on these tasks.
Ian Logan, a Lincolnshire-based freelance coder, author and medical doctor, was commissioned to write the extra Basic commands Spectrum users would enter to operate the Microdrive. These were built into the Interface 1’s 8KB Rom chip and patched onto the Spectrum Basic Rom when the Interface was plugged in. Bug fixes for the Spectrum Rom were added too.
Wanted: Spectrum Rom expert
Logan’s involvement arose from his work disassembling the Spectrum Rom, an effort he detailed in a hugely selling book of the time, The Complete Spectrum Rom Disassembly, which he co-authored with Frank O'Hara. Logan went through the computer’s firmware, detailing routines, jump table locations and such. The result was a reference not only much in demand among would-be machine code writers - Sinclair didn’t offer such a detailed reference itself - but with Sinclair staffers too, Martin Brennan recalls.
The Spectrum Rom had been developed by Steve Vickers on behalf of Nine Tiles, the software developed commissioned by Sinclair to produce the Spectrum’s Basic interpreter and operating system. Vickers left Nine Tiles in May 1982. Spectrum hardware designer and Sinclair employee Richard Altwasser had gone too, to join Vickers to develop a computer of their own, the Jupiter Ace. According to Brennan, they left behind all the machine code print-outs and source code listings you could want, but without them there was no one at Sinclair who knew the software intimately - the sort of people who could be called on to help tie together the Spectrum’s code to that needed by the Interface 1 and the Microdrive.
Enter, then, Ian Logan, whose work had given him that knowledge. Logan pinpointed three memory addresses Brennan could use as a jumping off point to intercept the Spectrum’s Basic interpreter and redirect the computer’s program counter to the Interface 1 Rom and the Basic commands it added.
Speaking to Sinclair User magazine after the launch of the Microdrive, Logan said of the peripheral: “It was developed in a large crate with the ULA at the centre. There is very little to the insides of the Microdrive. There is one ULA and the dual heads which read the tape. The Microdrive program was developed on EPROM. If there were corrections to the program I would go to Cambridge with the alterations and we would blow a new EPROM. In the end, Martin Brennan, who was in charge of the project, said ‘Right, that’s the end’. I have no doubt that he added two extra things by the next morning.”
The ULA was designed by Ben Cheese, the internal mechanics by John Williams. Jim Westwood provided guidance on how the Microdrive should be handled in Basic.
Logan revealed that the initial 100-odd Microdrives all went out with EPROM chips, to allow them to be updated with new code because, as he put it, “Sinclair Research wants to redesign the circuit board at some stage”. Burning a final Rom would take place, he hoped, when at least two bugs in the then-current code had been squashed.
It quickly emerged that there were other problems too. While many reviewers were generally positive about the Microdrive - the speed was the main issue; it wasn’t really quick enough to be a floppy killer - only regular usage would reveal the design’s weak spots.
At a stretch
Like all media based on very narrow, thin magnetic tape, the Microdrives accumulated particles of the magnetic oxide material, a problem exacerbated by the need to pull the 2mm-wide tape out of the centre of the storage reel - from where it was fed in a loop over the read head and back to the outer edges of the spool - and the extra friction it induced. The tape was moved by a rubber-wrapped wheel in the Microdrive, pinching the tape between it and a plastic wheel inside the cartridge. Firing up the drive, taking the tape’s speed at the head from zero to 750mm per second, would give the tape a big tug, leading to stretching at the point where the tape emerges from the centre of the storage wheel. Later versions of the drive incorporated a 22µF capacitor to allow the motor to come up to full operating speed more smoothly.
Even Sinclair, in the Microdrive manual, had to admit: “Microdrive cartridges will not last forever, and will eventually need to be replaced. The symptom of an ageing cartridge is that the computer will take longer and longer to find a program or file before loading it. So it is a good idea to keep back-up copies of important programs and files on another cartridge, or on a cassette.”
The Microdrive head mechanism schematic
While the drives were, at £49.95 a throw, considered cheap, replacement cartridges, which cost £4.95 each, were not. That’s about three times the price of a 5.25-inch floppy disk at the time. Storage capacity, initially said to be 100KB, had by launch become “no less than 85KB”, to allow for capacity lost to tiny differences in the length of the tape in different cartridges, motors running at slightly different speeds in different drives and, to a lesser extent, bad sectors on the tape. Each tape stored up to 200 512B sectors spread over two parallel tracks, the drive’s electronics automatically ensuring the correct track was read.
Each cartridge could hold no more than 50 files, and information had to be read into memory, changed, the original erased and the new version written onto the tape. There was no way to modify the files directly, a consequence of the drive’s lack of true random access.
The drives themselves had flaws. Adding extra drives seemed easy: just join them with a ribbon cable that clipped into the side of each unit. But each drive also had to be screwed together using a special bracket - clearly, Sinclair Research was worried that an inadvertently knocked drive would break its connection, crashing the system and potentially losing a user’s data. Up to eight drives could be lined up alongside each other this way.
It was later found that the mounting for the microswitch used to sense whether a cartridge’s write-protect tab had been removed was placed in a way that overly vigorous insertion of the cartridge into the drive could cause the mount to bend, “causing incorrect switch operation”, as a later, October 1985 Microdrive service manual reveals.
Rush to production
“I believe Clive rushed the Microdrive into production too soon, and the variability of the drive heads and the tapes meant reliability was quite spotty. If you went through a few drives, you could generally find a good one,” remembers John Mathieson, in the early 1980s a Sinclair engineer working alongside Ben Cheese and Martin Brennan, though not directly involved on the Microdrive project. He designed the ZX Interface 2, the add-on that would allow the Spectrum to accept Rom cartridges.
As Brennan and Cheese ploughed on to complete the electronics and the software, others under David Southward refined the drive’s mechanical engineering. Brennan recalls a hectic pace - “coming in to work on Saturdays and Sundays; working late most evenings” - but no specific deadline to which they all needed to reach.
Not that there weren’t very difficult challenges to overcome. The promised 100KB capacity was one. Using an analogue storage medium, the tape loop inside the cartridge, meant Cheese had to deal with a frequency response that was not only not flat but varied with the speed at which the tape passed over the read head. Speeding up the tape improves the response, but at the cost of storage capacity for a given length of the material. Getting an even frequency response is essential if the system is to accurately pick up the changes in frequency which encode the 1s and 0s of binary data.
Combine that with the tolerances accepted in the tape length, motor speed and such, and Cheese and Brennan saw considerable difficulty ahead if they were forced to meet the 100KB target. Cheese, backed by Brennan, argued that Sinclair needed to play it safe with capacity, and they were able to convince Southward and Searle that the capacity should be 85KB, with the potential for a higher value if a given tape and drive were in perfect alignment. That’s the capacity that the Microdrive system promised when it finally went on sale in September 1983.
Promoting the ZX Microdrive
The Microdrive soon had competition. US company Astec introduced what it called the Wafadrive in the summer of 1984 - the product was sold in the UK by Rotronics - which comprised a dual-drive hardware unit which took 16KB, 64KB or 128KB tape cartridges. The drives were rather more expensive than the Sinclair offering - £130 for the Spectrum version, or £160 for a unit that connected to the Commodore 64 - but the tapes were cheaper: just £3.95 for the highest capacity. The drives also had Centronics and RS232 ports on the back. Again, each cartridge contained a 5m loop of tape to simulate random access, though the speed of access fell as the capacity increased, as there was more tape to spool through. Sinclair User found the system to be rather more resilient than the Microdrive - mostly thanks to the larger, more robust tape cartridges - but slower.
Of course, what neither system could replace was the audio cassette. For almost all users of Microdrives, the tech became a place to store their own programs and data. Few companies were able to offer software on Microdrive cartridge, at least until the Sinclair QL, which also used the Microdrive, began to sell well. The QL hardware was developed by David Karlin, who sat close to Brennan and was able to get regular updates on the progress of the Microdrives while the machine, originally codenamed the ‘ZX83’, was under development. Likewise Tony Tebby who worked on the QL’s system design and driver software.
Decline and fall
The QL was launched in 1984 and fitted with a pair of modified Microdrives and no cassette tape interface. Mass duplication kit was non-existent, and the Microdrives themselves were too flimsy for running out the dozens and dozens of copies even a small commercial software vendor would need to churn out. And that’s before they took into consideration the cost of the cartridges, then only available from one supplier: Sinclair.
There were other problems too. The QL Microdrive motors were set to run 25 per cent faster than those in the Spectrum add-ons. This made duplication more difficult and reading copies likewise because greater speed meant greater friction and this caused more tape wear. It also made the QL drives less able to deal with dust and dirt in the mechanism and on the tape itself, and all the other issues Ben Cheese had had to figure out ways around. According to Rodney Dale, who was working at Sinclair at the time, “the shipment of QLs was painfully slow, as each Microdrive cartridge was tested on its host machine to make sure that it worked properly”.
Sinclair only had itself to blame. According to Rick Dickinson, “the Sinclair way” of doing things meant that rather than redesign the Microdrive internal chassis for the new computer, the Spectrum drives’ were simply put into the QL case. “Some of the failures of the Microdrive were really down to pushing production tolerances and materials too far so variability came in,” he remembers.
“With different materials it would have been more reliable. So, for example, I wanted to start off with a brand new chassis in a different material and build the two microdrives on that, but they insisted that now we’d already tooled up a chassis, we’d just have to use that. And of course what would happen is that, because you’re going into a whole new case with a whole new set of parameters, you’d screw the microdrive chassis in and the torque reaction from the screws would twist all the boxes slightly and just slew the whole chassis - both of them - off at one degree.”
Promoting the ZX Microdrive
Come early 1984 and Sinclair User itself was claiming fewer than 1000 Microdrives had been sold. It reported on two applications that had been adapted to make use of the Microdrive for data storage - Campbell Systems’ Masterfile and Richard Shepherd Software’s Cash Controller - but added that only “several companies have shown interest in putting software onto Microdrive cartridge”. For most games developers, the technology was too pricey: the young gamers who were their customers had pocket money enough for tapes, but not for Sinclair’s new storage system. Of course, this helped limit the number of games made available for the QL, which Sinclair was determined would be a serious machine for serious users, but it limited the number of sales to be had from folk upgrading from the ZX81 and the Spectrum.
By June 1984, Popular Computing Weekly’s Editor, David Kelly, felt compelled to complain: “By using Microdrives in the QL, Sinclair is taking a risk that the machine may never receive proper software support. Anyone who writes a brilliant program for the QL cannot simply trot off to the nearest duplication plant, run off a few thousand copies and start selling them because Sinclair keeps sole control of Microdrive manufacture and duplication to itself.”
Sinclair certainly had duplication facilities: the Microdrive shipped with a cartridge of demo software, and the QL came packed with four cartridges containing the business software that had been commissioned for the machine from Psion.
Sinclair MD Nigel Searle was not unsympathetic. “The [software] houses we have so far signed contracts with are mostly writing packages which will sell for around £40, so the price of the disk is less significant,” he told Popular. “I wouldn’t disagree that the current cartridge price is too high - the software houses all think so. But we have to balance supply and demand.” He revealed that Sinclair was then punching out 100,000 cartridges a month but the goal was to increase that figure to 40 million a year, at which point the price would come down.
Indeed, early in 1985 Sinclair chopped the price of cartridges to £1.99 to bring them into line with floppy disk prices. “Microdrives have truly come of age now and we anticipate that this technology, exclusive to Sinclair, will continue to be the preferred method of data storage for most Spectrum Plus and QL owners... and with the new price, software houses and members of the public will be able to exploit this versatile medium to the full,” Sir Clive Sinclair said in the announcement press release.
But it wasn’t to be. Microdrive software did come in greater numbers thanks to the cartridge price cut, but not in a volume to come close to the number of Spectrum cassettes on sale.
Sinclair suspended QL production in 1985 to save money - the company reported a £18.3 million loss for the year to 31 March 1985 - and by early 1986 was telling journalists some 150,000 had sold to date. But when Amstrad acquired the ailing company in 1986 for £5 million, it formally knocked the computer and the Microdrive platform on the head. Amstrad was keen on the 3-inch diskette format it had bought in from Asia for its PCW8256 word processor, and wanted to spread to other machines, the Spectrum and its own CPC among them, to achieve lower drive prices through greater purchase volumes, and to grow the user base. So out went the controversial Microdrive.
Amstrad’s favoured 3-inch diskette format would eventually go the way of the Microdrive, as would the old 5.25-inch floppy: the industry would shift to the 3.5-inch format, the chosen mechanism of new, 16-bit home machines like the Commodore Amiga and the Atari ST, and of business and professional computers, from the Mac to the ever increasing number of IBM PC compatibles. The growing use and capacity of hard drives would eventually limit even floppies to software installation and data exchange. Cassette storage passed into history too - tape storage eventually found a niche in back-up systems, which didn’t require high access speed.
While the Sinclair Microdrive may have been, then, something of a dead end in the evolution of storage, that shouldn’t detract from the novelty of the technology, nor of the efforts of Ben Cheese, John Williams, Martin Brennan, Rick Dickinson and Jim Westwood to turn them into a viable commercial storage system for home micros. ®