Clive Sinclair (1941-2021) Part three — resurgam

In part one we saw how the young Clive Sinclair created Sinclair Radionics — twice — and built it into a successful business that launched hi-fi products in the 1960s and the first ever pocket calculator in 1972. In part two, we investigated the genesis of the Radionics microcomputer. Read on to learn what occurred when Radionics finally went under, and how the ZX80 came about…

Clive Sinclair in 1981

Let’s imagine that we are Clive Sinclair. It is early 1979. We know what the full-year results for Sinclair Radionics, the company we founded in 1961 but which now no longer control, are going to say and it is not good.

At Radionics we have a microcomputer project which has sketched out a compact Z80-based machine that is capable of being battery powered, albeit briefly. It has an integrated single-line display. It is a sophisticated, advanced design, but that means it will be expensive to make and sell, and there is a growing amount of very strong competition in the market for £500-plus personal microcomputers. These are mostly intended for business or academic use, where budgets are higher, rather than for the person in the street, who has much less to spend. The market for sub-£500 is much more open, and that below £200 even more so. The hobby oriented Nascom 1, which launched at the very end of 1977, had a good 1978, racking up 10,000 sales its manufacturer claims. It’s just a naked circuit board, not a packaged, consumer-friendly product, and it costs £200, but the cost goes up as you build on its very basic capabilities. The Radionics design could be cheaper to produce than the Nascom, but it requires a hefty up-front investment in custom chip design and purchase commitment to be feasible.

Can we cost-engineer the design down to something that could sell for £100-150 and undercut the competition?

We have a second company, Science of Cambridge, which has been producing for about a year now a basic microprocessor kit, called the MK14, and it hasn’t sold at all badly. Unfortunately, Chris Curry, the man we put in charge of Science of Cambridge and who launched the MK14, is now about to leave. Why? Because he has independently developed a follow-up on machine that is a more sophisticated. It isn’t aimed at the same hobby market, but at industrial firms who need a means to integrate microprocessor control into existing and new products. It has been borne out of the consultancy work that came Curry’s way through the MK14 and he and his business partner Hermann Hauser’s links to Cambridge University. It’s not really intended to be the MK14 2.0, but it could be sold that way too. In any case, with Curry departing, we should come up with something Science of Cambridge can sell next.

Might our cost-engineered Radionics design become the natural successor to the MK14? It might have to be offered by Science of Cambridge, especially if Radionics’ managers decide to go ahead with the existing design.

The logical course: start developing the low-cost model. The Radionics microcomputer engineers clearly have the skill to do so, and we’ve always ensured there was never clear blue water between Radionics and Science of Cambridge. They can work up a lower cost micro, and Science of Cambridge can release it if Radionics doesn’t want to.
What did Clive Sinclair do? Time, ailing memories, the death of some of the key participants and a paucity of official documentation mean we will probably never know, but we can reconstruct a sequence of events from individual details given in interviews the half a decade after 1979 with a number of those who were present during these events. Some of these details contradict others, so we have to tread cautiously.

The Acorn Microcomputer was developed during the latter part of 1978 and went on sale early in 1979. The Radionics microcomputer was also developed during 1978, but may not have reached the same degree of completion and certainly did not go on sale at any time that year. Instead, it was held in limbo while Radionics bosses, Sinclair among them, dithered over whether they could afford to bring the machine to market.
The back and forth was complicated by Radionics’ declining fortunes, which probably also give Sinclair the motivation to consider a post-Radionics future more closely than he had at any time since the middle of 1976 when he asked Chris Curry to set up the company that eventually became Science of Cambridge. It was clear to Sinclair that the company needed a new product if it was to continue, and since he might very well be leaving Radionics, he had better make sure it did go on. Science of Cambridge was now firmly associated with the growing market of low-cost personal computer hardware designed for enthusiasts, so coming up with a successor to the MK14 was the ‘low hanging fruit’ Sinclair can’t have failed to have spotted. Any update would require a better processor, more memory — all now cheaper thanks to ever-falling chip prices — in fact, the full computer that the MK14 could be if you bought all the optional extras: Qwerty keyboard support, video output and cassette tape storage. A glance at any of the established electronics titles or the new personal computing magazines will show you those are minimum requirements now. It should have the Basic language built in too, not rely on machine code input via a monitor program. And it must not be offered as a clunky aggregation of circuit boards — that’s not the Sinclair way at all. No, we need a case with a keyboard to wrap it in, something that will make it look more like a real computer than that calculator like thing Curry is trying to sell. It also needs to be small. That too is the Sinclair way.

So much Sinclair might have sketched out from even cursory research into the state of the market undertaken by reading a few back numbers of Personal Computer World and Practical Computing magazines. As for the hardware specification, he would almost certainly have knocked on the office door of Mike Wakefield, the electronics engineer working on the Radionics micro and perhaps asked what, if Wakefield were challenged to redesign the machine to get it into a lower price bracket, he might do. That might easily be followed by the suggestion, ‘well, why don’t you build one to see’. According to Sinclair, asked about the project in September 1980, work on the low-cost micro, which would eventually come to market as the ZX80, began at exactly this time, March 1979.

What about software? Enter John Grant, owner and chief programmer at Nine Tiles Information Handling Ltd, a software development consultancy based in Cambridge. Grant had been involved in the development of the software running on the Radionics microcomputer, so he too was an obvious person for Sinclair to talk to. Their conversation was obviously fruitful because in April 1979 — the same month, you’ll note, that Radionics would report its devastating and, as it turned out, final full-year financial report, for the 12 months between 1 April 1978 and 31 March 1979 — Sinclair commissioned Nine Tiles to write a Basic interpreter and operating system for a new micro. According to later comments made by Grant, it was already clear at this time that the low-cost micro would have limitations which his code would have to help mitigate. That’s surely a reference to the use of a cheap membrane keyboard of the kind used in the MK14, and which would make it very hard to type in long Basic programs. Ergo, what was needed was a smarter way to enter Basic programs than keying them in character by character. It’s very hard to imagine the sophisticated Radionics machine being thus constricted, so Nine Tiles’ subsequent decision to implement one-touch Basic keyword entry must surely have arisen out of work done for the low-cost micro. This technique not only allowed the user to tap a key once to have a complete Basic keyword appear on the screen, but the operating software was smart enough to know where in a line of program code the user was so that it could indicate to the user through the cursor character whether the next keypress would generate a line number, a keyword, the letter printed on the key, or the key’s shifted character.

August was set as the deadline for completion: Sinclair knew that his days at Radionics were numbered. “We thought about making the ZX80 in August [1979],” Sinclair said in 1981. “We needed a product and that was it. Clearly I had anticipated the success of the ZX80 because we ordered a 100,000 parts.” With Mike Wakefield working on the hardware and Nine Tails on the software, Sinclair could perhaps spend more time on the discussions taking place among the Radionics senior staff and the NEB about the future of the company, which was not looking good.

Radionics comes to an end

At the start of 1979, Britain’s National Enterprise Board (NEB) instructed Radionics’ management that it could not continue to fund Radionics in the company’s current form. In short, Radionics had better find a partner with it could merge and which had sufficient financial resources to continue funding Radionics’ various development projects. It’s not hard to see why the organisation reached this decision: by September 1978, the NEB had pumped £3.15 million into the company. This total would rise to £7 million by March 1979. Even with that investment, Radionics still lost £2.8 million between May 1977 and December 1978. The NEB was obliged by the terms of the Act of Parliament under which it had been founded to seek a return on any investment it was made. That clearly was not going to happen unless another company acquired Radionics as a whole, or the troubled electronics firm was sold off piece by piece. By March 1979, only the latter course remained. Overtures made in the earlier part of the year to Thorn EMI, GEC and other large British companies in the electrical and electronics industries had proved unproductive. Don’t forget that though Radionics’ full-year results would not be made public until April 1979, Radionics’ managers, the NEB and Sinclair already knew what the numbers were going to be.

No wonder then that this is the very moment at which Sinclair started to think about his low-cost micro. Possibly not yet as a post-Radionics plan, but perhaps as a strategy to save the company. Whatever Sinclair’s motivation, and he may have entertained both ideas, the NEB announced that it would immediately begin seeking buyers for Radionics’ key divisions and product lines: the microcomputer, the instruments business, the pocket TV business, the flat-tube R&D effort and, lastly, what was left of the calculator family.

News of the plan leaked out, at least in part: in May 1979, the Financial Times reported that the NEB was seeking a buyer for Radionics’ pocket TV business. Undoubtedly it was at this point that that the NEB sold the Radionics microcomputer to Newbury Labs.

By July 1979, it was all gone. Since August 1976, the NEB had invested £5,350,000 in Sinclair Radionics and extended the company a further £2,443,000 in loans — £7,793,000 in total. The NEB’s investments in Sinclair Radionics were written off.
The NEB sold the second-generation Microvision and all existing stocks of pocket TVs to UK consumer electronics company Binatone for a reputed £1 million. Binatone took what was left of Radionics’ calculator business, for what little that was worth, too. Despite the positive attention received by a number of recent, premium models, including the Sovereign, inspired by the Queen’s 1977 Silver Jubilee, the division had been on the slide since the mid-1970s. Calculators were now cheap, reliable, able to run for ages on just one or two AA batteries, and more likely than not made in Japan. Radionics could no longer compete. The NEB had tried to interest the Japanese in the product, but none of the companies approached wanted it — they could do better themselves and had by now built up strong brands of their own. Further key Radionics personnel, among them David Southward, Michael Pye and Dave Chatten, resigned or were made redundant. They were soon asked to come and work for Science of Cambridge.

The sale of the TV and calculators to Binatone, and the microcomputer to Newbury Labs, left Radionics with its electronic instruments business, which had become a standalone operation in the June 1977 restructure put in place after the NEB made its second investment in the company. This business was worth keeping, largely because it had always made good, solid products that sold well — unlike so many other Radionics offerings. Too dull to attract Clive Sinclair’s interest and interference perhaps, it had become the only part of Radionics able to make a profit consistently. Two Radionics executives — Derek Holley, the Financial Director brought in by the NEB, and David Argent, the firm’s Production Director — had been planning since the beginning of the year to offer to buy the instruments divisions, and now they were able to do so simply by acquiring all that was left of Radionics. Joining them was Dennis Taylor, one of Radionics’ non-executive directors. Their holding company, Heavepalm, which had been formally founded in February, now bought up Radionics’ assets and then quickly renamed itself Sinclair Electronics. Early in 1980, the three owners formed Thandar Electronics — a combination of their initals: Taylor, Holley AND Argent — and it essentially took over Sinclair Electronics. Thandar continues to operate to this day, as Aim-TTi.

Sinclair at SoC

Sinclair himself stepped away in July 1979, taking with him a £10,000 payoff and the rights to the flat-tube television. He went straight round to Science of Cambridge. During July and August, a handful of Sinclair’s closest allies at Radionics and other staff, among them Jim Westwood, Brian Flint and Peter Maydew, came to join him. All three of them had been working on the TV tube when Radionics went down. Maydew was Flint’s assistant; Westwood was in charge of the tube’s engineering team. They formed the new core Science of Cambridge product development operation and took up residence on the top floor of the company’s office at 6 King’s Parade. Sinclair and his secretary, Molly Pearson, had offices on the first floor, at the front of the building; industrial designer John Pemberton, who had designed the Sovereign calculator for Radionics, the Wrist Calculator for Sinclair Instruments and was now also part of Science of Cambridge, had an office at the back, overlooking the goods yard of the Eagle pub round the corner in Bene’t Street.

Throughout this period, Mike Wakefield worked on prototype hardware. Nine Tiles continued to develop the software it would run. They were largely left to get on with it. Sinclair had described the price at which the microcomputer would be sold — the engineers’ job was to make sure that the tally of the machine’s components, both physical and virtual, would fit within that limit and leave Science of Cambridge with a decent profit margin.

Grant’s work involved creating the software which managed the computer’s hardware — to send data to and read it back from a connected cassette recorder, to poll the keyboard for taps and presses, and to maintain the data which the video circuitry would present on a connected TV. He also provided the software which interpreted and performed the instructions within those Basic programs. The machine’s component budged didn’t extend to a separate character generator chip, or to any silicon dedicated to the display, so he also had to cram in image data for all the characters the computer would be capable of presenting on screen. Grant finished the code some time in July 1979, only to find that it was bigger than the amount of Read-only Memory (ROM) that the machine had been allotted. His code was 5KB (5120 characters) in size; the target was 4KB (4096 characters). So he spent the following month editing the code down to fit.

Clive Sinclair would later tacitly praise Grant’s work: “The ROM in our basic machine is just 4KB which contains everything — Basic, operating system, keyboard control and display IO. Now there is no way we could have done all that with an off-the-shelf Basic.”
The hardware limitations nonetheless allowed for some innovation. The decision to use a membrane keyboard, for instance, prompted the addition of single-key command entry into the system software. Programming inevitably requires a lot of text entry, but since the membrane keyboard made quick typing virtually impossible and touch-typing very definitely out of the question, a novel mechanism was devised to overcome this dilemma.

Of course, the addition of single-key entry meant that the ZX80 keyboard couldn’t simply present the standard array of letters and numbers, plus shift, space and carriage return keys, so industrial designer John Pemberton had to design the keyboard layout — printed on a layer of plastic set above the keys’ electric contact layers and below a clear, protective outer layer — to present not only all of the standard alphanumeric characters but also Basic keywords, the symbols available when the shift key had been pressed, arrows to guide the on-screen cursor and other keys, like space and, in place of a standard carriage return (a relic of the manual typewriter), ‘New Line’. In a similar vein, the typewriter’s backspace key was dropped in favour of a key marked ‘Rubout’. The limited size of the keyboard meant that there wasn’t room for a standard spacebar along the bottom row of keys; instead, the bottom right key would produce a space. Since no one would by using the ZX80’s keyboard typewriter-fashion, this was judged an acceptable compromise.

There wasn’t room for everything, and the ZX80 was forced to sport a sticker on its case just above the keyboard to list the eight functions users would have to type in character by character if they wished to include them in their programs. This is perhaps the ZX80 case’s one aesthetic lapse, and it’s a wonder Sinclair agreed to have it added. The keywords’ absence from the one-key entry system was possibly a result of John Grant’s need to pare his system software code back from 5KB to 4KB. Certainly when his next computer was designed, it would contain all the possible commands the computer could accept. The ZX80 keyboard listed 30 commands; the ZX81 65. The Spectrum would take this to an almost absurd level: it presented the user with 89 Basic commands and half-a-dozen system commands available at the now familiar flashing K prompt.

Hardware delays

Of course, neither the keyboard nor the Basic interpreter would be of any use without electronics to connect them together. And here was a problem for Sinclair: Mike Wakefield was unable to meet his deadline. In May or June 1979, Wakefield moved from Sinclair Radionics to the Radionics micro’s new owner, Newbury Labs. There was no requirement to work on a potential low-cost micro here. By 1985 various suggestions were being offered by insiders as to why Wakefield didn’t complete the design: he couldn’t complete it because of his new commitments, or Newbury Labs explicitly told him to cease working on the now rival machine. It was even suggested that he did finish but Sinclair was in some way unhappy with his efforts. Whatever the reason, Wakefield stopped work on the computer, and Jim Westwood took over the hardware design work. Westwood had been a Sinclair employee since 1963 and had always been close to his boss, but he was brought on board the micro project with some trepidation on Sinclair’s part.

“Clive was very worried and uncertain as to whether in fact someone who hadn’t designed computers before could get him out of this latest major fix,” said Westwood in 1989. “He actually didn’t realise I had made one at home anyway. So at the end of October [1979], I brought him up [to the top floor at King’s Parade] and showed him the prototype in the lab. I had carefully arranged it to demonstrate that the screen would work. I’d got numbers going down the side — one, two, three, four etc. — and in the middle of the screen I had a little message that said ‘Jim has done it.’ I think he even smiled.”

Maybe it was the smile that made the ZX80 so memorable for Westwood. A few years after the machine had been completed, he said: “Of all the products with which I have been involved I think the ZX80 is my favourite. It was a real breakthrough in the use of cheap components. It’s something which ought to be in the Ark by now but I am still proud of it.”

Westwood had help, of course. Brian Flint, for instance, drew up the circuit board layout and that of a 1-3KB RAM pack that would be sold as an optional extra to clip onto the back of the machine. “I wasn’t part of the original design or concept; it was just given to me and I had to make it happen,” Flint said in 2011. “Sinclair had a specification which said the computer would use so much RAM and have an EPROM inside with so much memory. It needed a certain amount of logic that could be provided by off-the-shelf TTL (Transistor-Transistor Logic) chips, like counters and standard logic elements. In fact, all the hardware comprised ready-made components so there were no dedicated chips — that’s how it was made so quickly.”

Industrious industrial designers

Much of this work — essentially the process of taking Jim Westwood’s prototype and turning it into a working printed circuit onto which all the components could be mounted and which, in turn, could be fitted inside John Pemberton’s vacuum-formed plastic casing took place during the last few months of 1979. The casing now featured a raised section on top at the back of the machine to cover the large heat sink required to keep the computer’s voltage regulator cool. Gone was an indented logo area, replaced by a simple sticker to be applied to the upper casing just above the keyboard on the left side. Pemberton had originally envisaged the machine being branded the ‘Science of Cambridge ZX80 Personal Computer’, but now the computer would sport the Sinclair name in black above the ‘ZX80’ in yellow. The Sinclair logo chosen was the one that had been designed way back in the early Radionics days for the first issue of Sinclair News, and which soon became the official Sinclair Radionics logo. This was surely Pemberton and Sinclair flicking a final finger to their former NEB colleagues.

In addition to the look of the casing — and the logos it would sport, the Sinclair brand and the name of the machine — Pemberton’s work went beyond the visual: it included any part of the computer that was not a component. The machine’s heat sink, for instance, used to radiate away the heat generated by the computer’s essential voltage regulator, was designed by Pemberton because it wasn’t a standard part. Likewise the two wooden armatures across which heated millimetre-thin sheets of white plastic would be squeezed by vacuum suction to form the two parts of the computer’s case. Gaps in the case being difficult to produce using this technique, what appeared to be ventilation holes on the top of the case were actually just black lines added as sticker. The printed circuit board included space for the keyboard and would be supplied with the membrane already glued in place. The board, once populated with chips, resistors, capacitors and such, fitted into the lower part of the case; the upper part simply sat on top of the circuit board’s electronics, leaving the keyboard exposed. Six hand-pressed plastic rivets bound the three parts together. Hand pressed? Like the MK14 before it, the ZX80 was expected to be bought primarily in kit form, so it needed to be easy to assemble. The rivets were also cheaper than metals screws.

By December 1979, Pemberton was preparing to leave Science of Cambridge. He had been invited to take up a senior post in US electronics firm ITT’s newly opened Europe Design Centre in Harlow, Essex. His replacement was Rick Dickinson, a young Yorkshire-born graduate of Newcastle Polytechnic’s Design for Industry course, one of the first industrial design courses to be offered in the UK and for which he was awarded a first. During his time as a student, Dickinson had done some placement work for Pemberton while the latter was working at Radionics. After graduation, Dickinson had gone freelance, moving to Wales to be near the principality’s many electronics factories, where he might get contract work. Now, out of the blue he received a telegram from Pemberton insisting he call immediately. On the phone, Pemberton told him to come to Cambridge and meet Clive Sinclair.

To stay warm on that chill December day, the sandy haired and bearded Dickinson decided to walked the two kilometres from Cambridge station to King’s Parade. Portfolio clasped tightly against his body, he rang the bell and was ushered in. “I showed Clive some of my work and, I don’t remember exactly, I think he just said, ‘Well, when would you like to start? We’d like you as soon as possible.’ So I started in December of 1979. By then [the ZX80] was pretty advanced. They were making the tools for the mechanical parts but hadn’t made the circuit boards, so I helped with the PCB tape master. PCBs were laid out by hand on a 4:1 scale.”

Dickinson’s first work of his own on the ZX80 was the design of a replacement for the computer’s first RAM Pack, a small box designed to take a circuit board on which owners of the computer could fit up to 3KB worth of memory chips. Like the ZX80 itself, the pack sandwiched its circuit board between two vacuum-formed plastic panels held together with plastic rivets. The complete unit then clipped onto the computer’s rear-facing expansion connector. It was a clunky offering, but it provided early buyers with a way to boost the computer’s 1KB of RAM to 4KB. Naturally, Clive Sinclair was ready to sell both the board (for £12) and the chips (£48 for the full complement of 3KB). Notionally, you could buy five of fully laden packs — £240 in all — and piggy-back them up to each other and the ZX80 to give yourself 16KB of memory to work with. Maybe one or two owners did, but with memory prices inevitably moving down, this was never going to be a popular option. Indeed, by September 1980, the 1-3KB RAM Pack had been replaced by a completely new 16KB RAM Pack, priced at just £49.95, and cased in injection-moulded ABS plastic of the kind later used to house the ZX81. This was Dickinson’s main contribution to the ZX80. The reason for the price difference? The 16KB RAM Pack used Dynamic RAM (DRAM), a cheaper type of chip than the Static RAM (SRAM) used in the 1-3KB RAM Pack (and the ZX80 itself) but one that required more complex support circuitry.

During the latter part of 1979, Hugo Davenport, who would later join Sinclair Research as its Director of Engineering but was then working on defence projects for Cambridge Consultants, was hired to write the ZX80 manual, A Course in Basic Programming, and the assembly instructions supplied with the kit. Nine Tiles’ John Grant added all of the manual’s very interesting technical details that made up its three appendices. Meanwhile, Clive Sinclair hired public relations marketing agency Primary Contact, which had been working for Radionics since the early 1970s, to prepare the adverts that would flood the computer and electronics press immediately after the new computer’s launch.

The ZX80 made its debut at the Microsystems ’80 show, held in Wembley Conference Centre between 30 January and 1 February 1980. The price: £79.95 for the assemble-it-yourself kit, or £99.95 for a pre-built model. Sinclair had indeed achieved his goal of bringing the machine out for less than £100. And unlike the Nascom 1, which had been pitched as a sub-£200 computer but wouldn’t work without a £20 ‘optional extra’ power supply board, the assembled ZX80 was ready to run as soon as you took out of the box and plugged it in. That said, the DIY crowd only really saved just over £11, since the kit ZX80 didn’t come with an AC power adaptor, for which they had to fork out £8.95.

Still, there was no doubt that Sinclair might have a MK14-esque hit on his hands. Brian Flint travelled down to London to help out on the Science of Cambridge stand at Microsystems ’80. “I’d never seen anything like it; there were so many people around the stand, it was incredible… We had an instruction book that we were selling without the computer so people could find out about it. We took a big pile of them down there and decided to sell them for a fiver each, which was a lot of money back then, and they sold out. Just a book with no product! People were hungry for information. I can’t explain it, it was like bedlam; it was absolutely incredible.”

Sinclair may have begun selling the ZX80 by exclusively preaching to the converted, but he didn’t do so for long. Almost immediately, a separate strand of advertising began to appear in more mainstream publications. This time, the ZX80 kit form was downplayed in favour of the assembled version, the better to stress the machine’s usage rather than its construction. These ads were selling a personal computer, not an electronics project. They began to talk about outselling rival machines, and features photos of earnest-looking fathers instructing their sons in the skill of Basic programming. Remember, the early days of the British home computer scene was dominated by hardware buffs who were very keen to assemble electronics devices but were less enthusiastic about using and programming them afterward. It’s not hard to imagine the typical MK14 buyer, for instance, spending time carefully assembling and testing the kit, and maybe a few hours working through the short, simple programs in the user guide, but then drifting away to other projects. They had built the kit, they had learned some Basic programming — now they wanted something else to work on. That must have occurred with the ZX80’s early buyers too, but the availability of a pre-assembled version tapped into the emerging interest in computing held be people who were neither electronics enthusiasts nor sufficiently confident to undertake that much soldering. Sinclair would later say that that was what he’d had in mind all along: “When we introduced the personal computer, there was no doubt we would sell some in the hobbies market, but we also went out with advertising promotion to the man in the street, on the grounds that there would be a completely new market there.”

A completely new market? For Science of Cambridge, perhaps. But the folk at Nascom could rightfully say they had done that already, likewise Research Machines, which took computing out of the hobbies world and into schools. But what Sinclair can rightfully claim to have achieved is to have brought home computing into the mainstream. Of course, the ZX80 only set him on the path; it was its successor, the ZX81, that really brought Sinclair’s machines to the attention of ordinary consumers. However, as what was essentially the first draft of the ZX81, the ZX80 established the Sinclair way of computing: the low price of entry; the compact design that emphasised the personal rather than the shared, and the living room rather than the lab; the one-key Basic command entry; and the not-quite-powerful-enough specification that encourage the purchase of add-ons.

Press reaction

This change of emphasis was still unproven when the first buyers began to receive their ZX80s, and the press began to the review Sinclair’s new product. The response was generally good, despite the very obvious compromises Sinclair’s engineers had made to meet their boss’ component cost target. What won them over were the innovations made in software to compensate for the limitations of the hardware. David Tebbutt, the third Editor of Personal Computer World, liked the one-key Basic command entry, but was particularly taken by the was the ZX80 checked every line you entered, as you typed it in: “Syntax checking is superb — it’s impossible to go wrong,” he gushed. “Every character is checked on entry and, if the interpreter thinks that you are going to make a mistake, it signals with a reverse S (for Syntax) at the point it thinks you have gone wrong. If, later in the same line, you correct the error, then the marker disappears. What a grown up facility for such a small machine!”

Tebbutt wasn’t overly worried by the ZX80 Basic’s ability to deal only with whole numbers (integers) but not fractional values (so-called floating point values) since the machine was, to him, so clearly intended as a learning computer. “It makes Basic easy to learn, it’s small enough not for it to be intimidating and it’s cheap enough that, should you decide computing is not for you, you can give it away, sell it or whatever. Indeed it’s probably cheaper to learn Basic this way than pay for many of the courses around… It’s a fine machine on which to learn about computing.”

Tebbutt and other reviewers looked forward, however, to a promised upgrade to the ZX80’s Basic interpreter which would address many of the software’s limitations and omissions. Even as the ZX80 was going on sale, Sinclair’s engineers were already working on its successor. What the promised new code would not be able to do, however, was provide a solution to what, in use, was the ZX80’s most obvious problem: its annoying screen flicker. This was another outcome of the cost pressures placed on the computer’s hardware designers: instead of including a dedicated chip to generate the display signal, they used the ZX80’s Z80A processor. Not actually the Zilog CPU, mind, but a (cheaper, of course) version produced by NEC under licence. The system worked using the keyboard scanning software. This operated at the exact frequency of a TV signal so that if no key was pressed, the processor would read the block of memory storing the screen data, generate the picture signal and output it through an Ultra-High Frequency (UHF) modulator so that it would be accepted by a standard domestic TV. For 1980s British television operating on the PAL standard, that meant the keyboard scan routine needed to be run every 20 milliseconds to ensure 50 picture ‘fields’ were sent to the TV every second. Interrupt that sequence — by, for instance, pressing a key or running code that hogged the processor — and the TV would inevitably be deprived of a signal. The upshot: every time you pressed a key on your ZX80, the screen went black for a moment, hence the flicker.

Sinclair justified the drawback by claiming that since the machine was aimed at hobbyists rather than business users, this sort of thing wouldn’t inhibit sales. In fact, it would make explicit that the machine wasn’t suitable for more serious applications. No one would be fooled into trying to put it to improbable uses.

Of course, it didn’t take long for bright coders to work out how to hack the timing sequence imposed by the ZX80’s keyboard scanning and display sequence, and soon “guaranteed flicker free” games began to be offered by third-party software developers. Ron Bissell, co-founder of the software house Macronics, came up with one first, though a Dr Ian Logan, later an author of ZX books, wasn’t far behind. Both worked out that the ZX80’s precise timings allowed a small but usable period in which a game written in machine code could operate before the scanning routing kicked in. By ensuring a game got its work done only in that time and so was able to pause while the ZX80 did its stuff, Bissell and co. were able to prevent a flickering display.

Macronics, Quicksilva, Timedata and others all offered flicker-free games for the ZX80. They were just three of the many software developers who, inspired by the Sinclair machine, began offering programs commercially. This wasn’t a new phenomenon; various companies had been founded to offer software for previous home computers, the Nascom in particular. But if the sheer number of adverts appearing in the computing and electronics press at this time is anything to go by, the ZX80 software development community grew much more quickly and grew larger than any other such group up to that point. And not only software: very soon authors were writing books to help users make the most of the machine. Ian ‘flicker free’ Logan, Bob Maunder and Terry Trotter together produced The ZX80 Companion for Maunder’s publishing operation, Linsac. Publisher Melbourne House, later to become more famous for its ground-breaking graphical adventure game, The Hobbit, started out with a book, 30 Programs for the ZX80 1K. Phipps Associates had The ZX80 Pocket Book. And prolific writer Tim Hartnell quickly penned Making the Most of Your ZX80. Hartnell, an ebullient, afro-haired Australian in his late twenties, had come to London as a jobbing journalist. Back then he wasn’t a computer specialist, though he had a gift for maths. However, he was exactly the kind of buyer Sinclair had in mind for the ZX80: not an electronics tinkerer but an intelligent professional interested in the possibilities presented by microcomputing. Hartnell bought a ZX80 early on in April 1980 and was quickly hooked. As Robert Young, later Hartnell’s business partner, recalls, it was the young journalist’s own frustration trying to get to grips with Basic that made him realise other user were probably having just as hard a time and would appreciate some help. Hartnell founded the ZX80 User’s Group with a brief, three-line classified ad in Personal Computer World magazine. Within weeks he had a list of more than 3,000 members. In 1982, Hartnell became the launch Editor of Argus Specialist Publications’ ZX Computing magazine, though he also wrote books and articles on almost all of the home micros of the period, for nearly every computing publication in the UK before returning to Australia in 1984 to set up a local technology publishing and software firm, Interface Publications, and where in 1991, at the age of 40, he died of cancer.

Customer support

While Hartnell was writing and recruiting, Science of Cambridge was getting to grips with growing demand for its kit. The production of the casing and the circuit board was initially contracted out to local firm Tek Electronics, which had been a Radionics partner in the early days. But as demand grew, Sinclair established a production partnership with Timex, which maintained a couple of contract electronics plants in Dundee and would take the strain when the ZX80’s successor came out.

“I was getting a load of kits coming back at one point, usually because someone had tried to solder the parts together and made an awful job of it,” recalled Brian Flint in 2011. “I’d fix them and send them back out again. Most of it I could do by visual inspection; in a certain light I could see where they had created a solder bridge. I simply wiped over it with the iron and it invariably worked.”

Still, the ZX80’s reliability was better than had generally been achieved under the Radionics regime; the Radionics instruments products being a notable exception. Science of Cambridge’s Production Controller, Dave Chatten, claimed in late 1980 that the return rate on the computer was 2.4 per cent — 24 machines in every 1000 sent out would come back for repair. By contrast, Radionics had suffered a return rate of 20-30 per cent on the Microvision pocket TV and the Black Watch.

By April 1980, Science of Cambridge was selling around 1000 ZX80s a month, and of course there were the inevitable delays in getting kits and complete units out to their buyers. That in turn prompted the inescapable grumbles and, when the magazines began asking questions, the obligatory Science of Cambridge bromide. For example, Personal Computer World readers complained that ZX80 deliveries were taking “much longer than the 28 days promised in the early advertisements”, to which Sinclair responded: “Production has been exactly according to schedule from the start of the programme, but orders were far greater than expected. As a result, delivery is taking eight weeks instead of the four weeks we planned for. Production is rising rapidly, however, so we expect this time to reduce soon. There is not much difference in the delivery of kits and built, but kits are a bit easier at the moment.”

It’s undoubtedly no coincidence that around this time, Science of Cambridge dropped any claim that it would deliver in that 28-day period from its ads.

The ZX80 production delays were eventually eased, and by September 1980, Science of Cambridge was able to say it had sold more than 20,000 units — double the number that Nascom achieved in its first year and well in excess of year-one MK14 sales. On the back of the money the machine was bringing in, the company expanded: there were now 12 staff in King’s Parade, up from the five who Rick Dickinson joined in December 1979, and six engineers at The Mill who were dedicated to the flat television tube project. There was one other, in a different location: Nigel Searle, the mathematician Sinclair had brought in to Radionics to help develop the company’s first programmable calculator, but who had gone to the States and later parted ways with Radionics. In January 1980, after a visit to the Consumer Electronics Show in Las Vegas, Sinclair travelled to Boston to meet Searle. He showed his formed colleague the new computer, and soon Searle was establishing Science of Cambridge’s North American sales operation. The ZX80 was formally launched in the US in August, but it wasn’t until the Autumn that the machine became available to local computer buffs. It was theirs for $199.95, but only in a fully assembled form, not as a kit.

In part the delay in shipping the US units was as much the product of the ZX80’s on-the-cheap video system. The same technique — get the processor to drive the display — had to be applied, but because the US NTSC television system operated at a different frequency than the British PAL standard, the electronics needed to run at a different speed and so did the software routine that scanned the keyboard and controlled the display. That required some tweaks to be made. So did US Federal Communications Commission regulations on electromagnetic emissions. The ZX80 was not permitted to be sold in the US unless it was given the green light by the FCC, and bursts of radio noise from the machine stopped it from receiving that approval. So foil inlays — one above the circuit board, one below, both attached to the case and joined at the edges — were placed inside every ZX80 destined for North America. With this DIY Faraday Cage in place, US sales could begin.

Those were the technical reasons, and the broader production issues helped delay the US release too, but there was also a third reason: the ZX80’s first clone.

MicroAce or micro imitation?

The press were alerted to the story when Science of Cambridge went to court to obtain an order blocking the sale of a machine called the MicroAce. It’s not clear how Sinclair got wind of the MicroAce, but somehow he discovered that New Barnet, Hertfordshire-based computer dealership Comp Shop was planning to produce a Z80A-based kit computer with 1KB of RAM, membrane keyboard attached to the front of the circuit board, and a small, vacuum-formed plastic case, albeit black rather than ZX80 white. The circuit board layout was different. The ZX80 used standard, off-the-shelf parts, so it was hard for Sinclair to grumble about the hardware similarities, but he could take action if the software had been used or imitated. “This proposed rip-off of my ZX80… if it is based on my software, I will act with the full force of the law,” he thundered. His anger was not unjustified: the MicroAce claimed a “unique one-touch key word entry”, and the MicroAce’s keyboard layout, complete with keywords, was identical to the ZX80’s. Neither Comp Shop, nor another company named in Sinclair’s affidavit, Product Launch, were represented in court, so Sinclair walked away with an ex parte injunction banning the promotion and marketing of the clone.

Comp Shop proprietor Chris Cary, when questioned shortly afterwards by Personal Computer World’s Guy Kewney, denied that he was behind the MicroAce, but admitted he was supplying the chips it would contain. “Somebody has approached me for prices on a list of components which would fit exactly into an imitation ZX80, and there aren’t many other things you can build with those chips, in that volume,” he brazenly told the journalist. Kewney didn’t believe a word of it but he was careful not to say so in print, instead writing of Cary: “He was behind the well-known ‘nearly copy’ of the Ohio Superboard, the Compukit UK101, and anyway he is the only known operator with the chutzpah to try to undercut Sinclair on the ZX80 price.”

Within a month Science of Cambridge and MicroAce had come to an arrangement, reached out of court and therefore beyond the public gaze. The MicroAce, they agreed, would not be sold in the UK, but it could be made available overseas. An office was established in Santa Ana, south of Los Angeles, and the $149 kit was soon being promoted in the US computer press, albeit not widely — Sinclair won MicroAce’s agreement not to promote its knock-off in the key titles, which Sinclair reserved for himself. The deal wasn’t limited to the US. In Australia, the Dick Smith electronics retail chain took the MicroAce onto its shelves, offering the computer for 199 Australian Dollars.

How had Sinclair failed to block the MicroAce entirely? He had the judge at the court hearing to thank for that. The contents of the MicroAce ROM, since it was not exposed and therefore couldn’t readily be viewed, could not be assessed to be the same as the equally opaque ZX80 ROM, the judge said. Science of Cambridge could not there and then show the contents of the two chips in a way that would immediately satisfy the court as to their similarity. Presumably Sinclair’s legal team forgot to bring the kit required to read each ROM and disassemble its contents. The MicroAce keyboard, on other hand, was clearly copied from the ZX80, said the judge, and on that basis granted Sinclair the requested injunction.

The ruling prevented the MicroAce from going on sale in the UK, but there was no limit to its sale beyond the court’s jurisdiction. However, Sinclair might sue in every territory in which the MicroAce was launched and succeed on the back of the similarities between the keyboards even if he failed to convince anyone that the ROM had been copied. MicroAce could not change its keyboard; too many had been manufactured and its layout was too closely tied to the ROM. The result: impasse. Science of Cambridge had strength enough to limit MicroAce’s actions, but not enough to block it altogether.
“It is a straight copy,” said Melbourne House boss Alfred Milgram after MicroAce had set up shop in the States. “The only thing that MicroAce has done is to add another 1KB on board. Apart from that it is exactly the same machine [as the ZX80]. It has the same routines which all run in the same places, the same operating system, the same number of chips, it is basically the same machine. I can’t see how Sinclair can design anything that would not be compatible with the MicroAce.”

Sinclair had one advantage over MicroAce: the work on the ZX80’s successor has already begun and with it a new, improved version of Basic that addressed the limitations of the ZX80’s implementation of the language. Not only would this quickly supersede the current ROM, but the new machine’s design could be altered to make duplicates like the MicroAce impossible. The law may not have been entirely on Sinclair’s side, but time certainly was. He would make sure no one could easily rip off his products ever again.

Science of Cambridge announced the new ZX80 ROM — at 8KB, double the capacity of the current version — on 11 September 1980, alongside the 16KB RAM Pack. The new chip would provide those missing floating-point maths functions and some other besides, and come with a new keyboard overlay to stick down on top of the current one, all for £19.95. Many a ZX80 owner rubbed their hands with glee… only to be disappointed a month later when the updated ROM was formally delayed until early 1981. By way of compensation, Sinclair said the new chip would contain not only extra features but also support for a new printer. “We wanted to bring out the 8KB as soon as possible, but on the other hand we did not want to be severely criticised at a later stage when we would have to bring out another version,” he said.

Very shortly, Science of Cambridge would itself reappear in a new guise. In October, it became Sinclair Computers, a sign that its founder had decided upon the focus of the business, at least for the medium term. Or maybe not. In December, the company’s name changed again, this time to Sinclair Research, the name by which it would be known through the coming home computer boom. Wait a minute, didn’t Sinclair Research already exist? Indeed it did. In May 1979, as Sinclair’s final efforts to persuade the NEB to restructure Radionics were falling on deaf ears, a new company, called Ravenfinch, was incorporated. The following December, Sinclair renamed it Sinclair Research, the role it had undoubtedly been formed to play. Sinclair’s suggestion to the NEB had included the formation of Sinclair Research to devise the products that a slimmed down Radionics would license and take to market. Now renamed, it became a sister company to Science of Cambridge. Just over a year later, toward the end of 1980, as Sinclair Computers became Sinclair Research, the existing Sinclair Research became Sinclair Research UK, subtle shift that shows just how far Sinclair had come since the Summer of 1979. Then he had clutched at a new cheap micro as at a straw to remake his company. Now,18 months later, it had not only done so but had established Science of Cambridge/Sinclair Computers/Sinclair Research as a microcomputer manufacturer with an international reach.

Clive Sinclair and his new computer business was on a roll. With Timex’s Dundee plant now punching out around 10,000 ZX80s a month, 70 per cent of them going for export, Sinclair decided he deserved a new car. His Rolls-Royce had been sold in the Summer of 1979 to help fund the development and launch of the ZX80, but the success of machine now allowed him to lay down £20,000 in March 1981 on a limited edition Porsche Carrera GT, the turbocharged version of the 924. The price, one wag estimated, was the profit on a mere 800 ZX80s. On that basis, the ZX80’s profit margin was an impressive 25 per cent. The best part of 30,000 of the compact computers had been sold to eager punters. “I don’t know why he got it — it’s quicker for him to walk to the office,” said one unnamed Sinclair Research employee. Maybe, but the flash new set of wheels was Sinclair’s way of saying he was back in business.

Previous Instalments