I t started quite a while ago. The earliest calculator, possibly a thousand years ago or more, was probably made of pebbles positioned in grooves in Chinese sand. By the 14th century, that calculator grew to be more stable and portable. Its Chinese inventors called it suan pan (for calculating plate), and we call it an abacus.

It was in a frame with a series of vertical rods, each carrying two movable beads above a horizontal bar, each bead having a value of five, and five below the bar, each bead having a value of one. Each rod had a position value, as do columns in today's numbering systems, with the right-most column of beads representing ones, the next to the left representing tens, and so on. Beads acquired value when moved toward the horizontal bar. The abacus provided non-volatile storage. It is possible that early Chinese accountants were called bead counters.

Even today, a visitor in Japan or China will frequently encounter a merchant who can run a calculation on an abacus faster than most of us can with more modern appliances. But even there, an American invention, the handheld calculator, has begun to transfer the merchant's abacus to a seldom-used desk drawer, as it has done to the engineer's slide rule.

But that leap took a few years and there were many calculator developments along the way-including that slide rule, which, in the 20th century, came to be known as a slip stick or guessing rod.

There was a time when the slide rule was a means of distinguishing engineers from the rest of humanity. A five-inch rule in a shirt pocket or a 10-inch rule in a scabbard dangling from a belt was an almost certain identifier of an engineer.

The slide rule traces its history to John Napier, the Scottish Laird of Merchiston, who, in 1614, invented logarithms. One could add and subtract logarithms to multiply and divide numbers. From that point, it was not a great leap to drawing logarithmically spaced lines on, say, two sticks, and adding stick lengths to perform multiplications, or subtracting lengths for division.

More sophisticated slide rules with more scales provided capabilities of finding squares, cubes, roots, reciprocals and other exponential functions. One could also solve trigonometric problems and, with special designs, a wide variety of problems tailored to specific technical or business applications.

The slide rule-even with extended length, linearly, or in aspiral or cy lindrical configuration, and even with the help of a hairline cursor-had limited resolution and depended on a user's visual acuity. Two users viewing the same slide rule might read slightly different numbers. If 463 or 465 were equally acceptable, as they were for many engineering problems, the slide rule was adequate.

But such ambiguity was not acceptable for business applications. Money matters required accuracy. And the slide rule wasn't good for addition and subtraction, mainstays of business.

Along came adding machines, side by side, in the early 17th century, with slide rules. The earliest adding machines only added. In 1642, an adding/subtracting machine was invented by Blaise Pascal. Yes, that Pascal, the 17th-century French mathematics prodigy and probability theorist, who had a programming language named for him in the 20th century. Pascal's pioneering machine, the Pascaline, used interlocking wheels and cogs. An operator would dial in 0 to 9 on each wheel in a decimal column. When the su m in a column exceeded 10, a 1 was carried to the next column to the left. Results appeared in small windows above the dials.

Later in the 17th century, in 1673, Gottfried Leibniz designed an add/subtract machine using a stepped drum. Repeating additions or subtractions, a user could perform the four popular functions-addition, subtraction, multiplication and division. Levers were used to input a number, while an indicator showed sums or differences. A revolution counter showed the number of items added or the number of times a number passed through addition or subtraction for multiplication or division.

Leibniz was the fellow who invented calculus, as did Isaac Newton, independently. He is also credited with perfecting a system that came into rather widespread use in the 20th century-binary arithmetic.

Commercial reality
The Leibniz machine was tricky to build. It didn't become a commercial reality till the middle of the 19th century.

Along the way, there came sophisticated mechanical, the n electromechanical, machines for adding and subtracting. Then, with more sophistication, more cost, more weight and more time required for a calculation, machines could multiply, divide and provide other functions.

Business offices replaced eye-shaded clerks who could add long columns of multidigit numbers and get the same sum twice in a row with costly machines. Trained operators using the latest machines could add the same columns of figures and often get the same answer. A hand crank in early models helped operators develop strong forearms.

These machines bore prestigious names like Burroughs, Remington Rand (later Sperry Rand), Marchant, Monroe and Friden. And they bore multiple columns of 10 keys each. A 10-digit machine would sport an array of 100 keys.

What many consider the first practical calculator was produced in 1872 by Burroughs-William Seward Burroughs, a former bank clerk who probably figured that there had to be a better way. Motor-driven machines appeared in the 1920s, when Burroughs and Friden were dominant brands.

In time, electromechanical calculators, like many other electromechanical appliances, were replaced by electronic products that were simpler to use, smaller and faster, but not less expensive. Well, not at first. Since they had
no moving parts (except for versions that included printers and except for the keys), they weren't subject to the wear and tear of mechanical calculators. So they lasted longer, despite the need to replace vacuum tubes occasionally.

They lasted long enough to be replaced, starting in the early 1960s, by models using hundreds of transistors, weighing more than 30 pounds and costing upwards of a thousand dollars. According to Guy Ball, a calculator historian, collector and aficionado, there were more than 30 manufacturers, chief among them companies like Canon and Sharp. Japanese companies, thanks to the American-invented transistor, were beginning to take over the calculator business-as a springboard, it was widely assumed, for IC development.

B all's research came up with the Anita, produced by the English companies Bell Punch Co. and Sumlock-Comptometer in 1963, which laid claim to being the first fully electronic calculator. The machine weighed 33 pounds.

The following year, 1964, saw the second electronic desktop calculator, from Friden. It used a CRT display and lots of transistors and diodes. It was also the first calculator to use reverse-Polish notation (RPN), a term developed by somebody who thought it might be easier to remember than the name of its inventor, the Polish mathematician Jan Lukasiewicz.

RPN is a stack-based system in which all values are pushed on or off a stack. In operation, a user enters a number, hits Enter, enters another number and after all necessary numbers are entered, finally enters the operation to be performed.

The year also saw the first almost-all-transistor calculator, the Sony MD-5, with a novel display. It used gas-discharge tubes called Nixies, which were invented by two Haydu brothers who sold their co mpany to Burroughs. Each tube had a stack of 10 formed-wire numerals, each on an almost invisible supporting structure. A selected numeral would glow orange-red.

Burroughs had the Nixie at its research center in Paoli, Pa., by 1954. Saul Kuchinsky, an applications manager, brought it to be manufactured at the Electronic Components and Electronic Tube Division in Plainfield, N.J., where he became general manager.

Nixie gets nod
The Nixie, which required a high-voltage supply of at least 170 Vdc, was adopted as the display for most digital instruments and calculators. It cost about $10 per digit at first, then less. Some later calculators used the Panaplex, a single Burroughs tube housing a side-by-side string of six or more Nixie-like digits. Other calculators adopted the vacuum-fluorescent display developed by Ise Electronics in Japan.

Ball points out that Sony was first to drop leading zeros, first to include floating decimals, rounding off, percent calculations and reciprocals. These were m ajor contributors to the array of features that curtailed the life span of the simple four-function calculator, the "four-banger."

The earliest four-function electronic calculators provided fixed-point calculations and cost $1,000 to $2,500. Electronic calculators quickly drove mechanical and electromechanical calculators to junk piles, antique shops and the recesses of company storage rooms. At first they cost more than the mechanicals, but they provided quick calculations; they were easy to operate and easy to learn to operate; and they required little or no maintenance.

Wang Laboratories introduced a major business and scientific calculator in the mid-1960s. It sold for about $5,000 and was displaced in the marketplace by Hewlett-Packard Co.'s 9100A, introduced in 1968.

The big scientific
This programmable calculator provided a host of features including floating-point with a range of 10-98 to 1099, logs and antilogs to the natural base and to base 10, trigonometric and hyperbolic functio ns and inverses, vector addition and subtraction, polar/rectangular conversions, a magnetic-card reader/writer and an optional printer and plotter. It included a 16-layer printed-circuit board with magnetic coupling among layers for a ROM, and 16 ferrite-core registers. It used a 3.25-x-4.75-inch green CRT display.

The 9100 was a terrific calculator. And it was big, standing 8.25 inches high, 16 inches wide and 19 inches deep. It weighed 40pounds. And it was not cheap. It cost $4,900. And it set William Hewlett, president of HP, to thinking. Wouldn't it be neat to have such features in a shirt-pocket scientific calculator at substantially lower cost? What a dream!

While the HP-9100 was the powerhouse scientific calculator, many manufacturers and, especially, many semiconductor manufacturers saw great profit possibilities in less-powerful desktop calculators selling for much less than $4,900.

What triggered the boom in calculator production was the growing use of integrated circuits, the growing power o f ICs, the declining cost of ICs and the growing prevalence of what people called large-scale integrated circuits, a term whose meaning has escalated over the years.

With calculators in the second half of the 1960s selling for $500 or more, the market was very attractive for many manufacturers, especially semiconductor manufacturers.

Semiconductor manufacturers developed profound management theories surrounding the concept of vertical integration, which meant that they would design, produce and market both the semiconductors and the products using them. Particularly attractive were mass-market products like calculators and digital watches, which they proceeded to manufacture under their own names and private-labeled for others.

Business theoreticians developed vertical-integration theories showing how semiconductor manufacturers would vastly increase their revenue and profits.

The theories were appealing. The reality was not. The booming calculator market attracted many semiconductor manufacturers-a nd almost drowned them in red ink, leaving only two U.S. manufacturers of calculators, Texas Instruments Inc. and Hewlett-Packard, and no U.S. manufacturers of digital watches.

With electromechanical calculators selling for $1,000, Sharp made a dramatic introduction of an electronic four-function desktop calculator at the New York IEEE Show in March of 1969. The QT-8 was to be introduced at the exciting price of just $495, less than half the price of electromechanicals.

The calculator had an eight-digit LED display. It featured floating decimal and it boasted another unusual feature. If you tried to divide by zero, it signaled an error. It did not go crazy spinning wheels in a mad effort to provide a solution, as electromechanicals did.

U.S. parts, Japanese calculator
What made this product remarkable was that it was a Japanese product-a consumer and business product-based on American components, a reverse of the normal practice. In those days, when Japan was still a low-cost producer, that was most unusual. What made it all the more remarkable was that the calculator used only four "LSI" ICs and a clock chip (total cost $62.40 as part of a $30 million contract with North American Rockwell Microelectronics).

The LSI chips used 6-micron geometry, an impressive achievement then, but there was concern that North American Rockwell, noted for its work on the Minuteman missile program, didn't have a large fab in place and might not be able to deliver such dense chips in quantity. To show its confidence, Sharp reduced the price of the calculator to $395.

That pricing level in 1969 prompted the editor of a leading industry magazine to make the bold prediction that we might live to see the day when four-function calculators would sell for as little as $100-a target that was breached in three years. By the 1980s, calculator prices pierced the $10 level, but one could no longer buy a mere four-function calculator. Even the cheapest calculator was likely to offer extra features like memory, constant, s quare root, polarity inversion, rechargeable-battery power and solar power.

Manufacturers competed on the basis of packaging and styling. One machine might be offered in a notebook pocket. Another was built into a desk rule or desk-clock set. Still another was part of a woman's purse, and still another was built into a child's toy. (A child's toy calculator might sell for two or three times the price of a standard calculator.) And of course, calculators were offered in designer colors.

Sharp had been working on calculators since 1966 and had developed a calculator using as few as 145 ICs. The race was on to replace transistors with ICs and to reduce the number of ICs. The five-chip calculator of 1969 was part of a continuing race to reduce parts count. By 1972, Sharp had a single-chip calculator.

North American Rockwell Microelectronics Co., by the way, felt that its name was a bit of a mouthful, so it planned, at first, to refer to itself simply as NAR. It rejected that acronym because in German and Y iddish, "nar" means fool. It elected to use NRMEC. The Sharp relationship with Rockwell started in 1968 and continues to this day.

But there was still a calculator world to be conquered. And that world, the handheld calculator, was conquered by the man who, in 1958, designed the first integrated circuit-TI's Jack Kilby-along with TI colleagues Jerry Merryman and James Van Tassel. They were awarded a basic patent in 1974 for a miniature electronic calculator.

Texas Instruments was always dedicated to expanding the use of semiconductors. So in 1965, it began work on a handheld calculator, the Cal-Tech, just as, a decade earlier, it had designed and helped a small manufacturer produce and market the first transistorized pocket radio-an event that created a massive market and added the word "transistor" to the popular vocabulary. A transistor was a pocket radio.

In contrast to the 1955 radio, which was to promote the use of transistors, the 1965 calculator program was to promote integrated circuits and, mor e specifically, TI's single-chip calculator. The design was completed in March 1967. It was a four-function, battery-powered calculator without a traditional display. Instead, the machine included a thermal printer and provided printed output on a strip of paper tape that emerged from a slot on the left side of the calculator near the top edge. Printing was serial, one character at a time, with results of computations up to 12 digits long.

As was customary at the time, TI worked with a Japanese calculator manufacturer, so the Cal-Tech prototype pocket calculator reached the market, in the fall of 1970, as the Canon Pocketronic. It sold for under $400, weighed 1.8 pounds and was promoted as suitable for carrying in a pocket. A large pocket.

TI entered the market under its own name in 1972 with a couple of desktops and the handheld TI-2500 DataMath, a four-function, floating-point machine that could operate from its rechargeable battery or from the ac line. It used an eight-digit LED display and had a $120 retail price.

Triggering a revolution
Another relationship of a Japanese calculator manufacturer with an American semiconductor manufacturer was to have an effect far beyond anybody's wildest imagination.

In mid-1969, Busicom, a Japanese calculator manufacturer, asked Intel Corp. to design chips for a family of programmable desktop calculators. Busicom chose Intel because, unlike many other semiconductor companies, it did not have ties with other Japanese companies. In those days, the major Japanese calculator makers were partnered with U.S. IC manufacturers: Canon and Ricoh with American Microsystems, Sharp with Rockwell and Seiko with Signetics.

Busicom's design called for five or more chips, possibly as many as 12. Designing those chips would have been a painful drain on Intel's small staff, which was fully occupied with the 1101, a 256-bit DRAM chip, forerunner of the 1103, the 1-kbit DRAM that knelled the death of ferrite-core memories.

Marcian (Ted) Hoff, Intel's 12th employee, took ti me off from work on the 1101, then the 1103, to take on the Busicom project. Hoff saw that it would be difficult to meet the cost target with the proposed design, so he urged some simplification. Busicom's engineers politely told him to get to work on their design.

Hoff had been using Digital Equipment Corp.'s PDP-8 minicomputer and saw that it had a very simple instruction set. This triggered the thought that he could eliminate a lot of the complex logic in the Busicom calculator, do more in ROM and make the instruction set simpler.

In October 1969, Hoff presented his idea to Busicom's managers, and showed that Intel's architecture would result in a more general-purpose solution. They accepted. His team came up with a four-chip design that included a ROM chip (the 4001), a RAM for moving and storing data (the 4002), a shift register for I/O (the 4003) and a CPU chip.

A CPU chip! That chip, later called the 4004, was the first commercial microprocessor. Other companies-notably TI, Rockwell, GI and GMe-c laimed that, in developing calculator chips, they had, in fact, developed microprocessors, perhaps somewhat earlier. But the 4004 was the first device to be offered to the marketplace as a computer in itself. By later standards it was a rather simpleminded computer, but it had the processing power of the first electronic computer, the Eniac of the late 1940s. It was a lot smaller.

Hoff's team had developed a general-purpose computer that could be programmed to be a calculator and, it turned out, could be programmed for a vast array of other applications. But in early 1970, the chip set was designed for Busicom, which had paid Intel $60,000 to develop it.

Intel delivered the chips and Busicom sold 100,000 calculators using them. Busicom, however, was in poor financial condition (it went into bankruptcy in 1974), so it was ready when Intel offered to lower its chip prices in return for the rights to the microprocessor design and the rights to market the microprocessor for non-calculator applications.

The Busicom calculator design led to the world's first commercial microprocessor. It would now be possible to design a new calculator without replacing all the old chips. More important, the microprocessor opened a new world.

Toward smaller pockets
A new race was on-now for calculators aimed at smaller pockets. Manufacturers boasted that their calculators were no thicker than a pack of cigarettes. At one point, the English manufacturer Sinclair Radionics offered the world's thinnest calculator, the Sinclair Executive. This machine with an eight-digit LED display was as thin as a cigarette. It was not a wonderful calculator, but it could easily fit in a shirt pocket and it cost $400.

By 1972, calculators had reached the point where their cost had crashed through the $100 barrier; they were fully portable; and they could provide a handful of functions to supplement the

basic four. But the handheld calculator couldn't provide the functions needed by an engineer or scientist. An expensive and heavy de sktop was needed for those functions. Until the HP-35. Bill Hewlett's dream.

A pocket scientific
The HP-35, introduced in January 1972, was the first pocket calculator with scientific functions. It came to market at $395, was immediately back-ordered for months and it killed the slide rule. (The first slide-rule maker to fail was Dietzgen, in 1973.)

Some professors objected to it because engineers would forget how to use logarithms or trig tables, just as earlier professors had objected to the basic four-function calculator because children would forget how to do long division. The professors failed to stop the exploding growth of scientific calculators, fathered by the 35, which was named for the number of keys on its keypad.

HP followed the 35 with other scientific, then business-oriented pocket calculators. These met with successes that were far beyond the most ebullient expectations. The ultimate miniaturization came in 1977, when the company introduced the HP-01, a wristwatch calculator t hat started at $600 in a steel case. The 01 came with a small probe for pressing the tiny keys.

That calculator was impressive. It was also a commercial failure. It weighed 14 ounces, someone recalled, and prompted one engineer to comment: "I'll wait till they bring out the portable model."

In time, really portable calculator wristwatches did appear on the market-for substantially less than $600. One fellow recently purchased a calculator wristwatch on New York's Canal Street for $6. A block later, his companion comforted him when he saw the same calculator for $5.

The trend toward miniaturization reached the point where some calculators weren't much thicker than credit cards, then reversed as many people felt their fingers weren't small enough.

For a while, calculators were used as premiums. Publishers used them to induce new subscribers. Credit-card companies used them as incentives to switch to their cards. Banks offered them to new depositors. But calculators quickly became too cheap to be an indu cement for anything.

In time, the calculator, especially the pocket calculator, generated a following of enthusiasts and collectors and even an international association. Two of the leading enthusiasts and collectors, Guy Ball and Bruce Flamm, recently published Collector's Guide to Pocket Calculators , a 200-page paperback full of pictures, specs and history of a tremendous array of calculators. (The book is available for $23.95 plus $4 for first-class mailing from Wilson/Barnett Publishing, P.O. Box 345, Tustin, Calif. 92781.)

Over the years, calculators acquired value through functionality, packaging, merchandising, promotion and, more recently, through rarity and condition. A Ball-run Web site features historical articles, vintage ads, photos and buy/sell offerings of rare calculators ( www.oldcalcs.com ). One recent offering featured the Commodore SR4148R, a rechargeable handheld with wallet and user manual, for $40. An accompanying notation said: "Excellent condition, not working."

Back to

Back to

	eeProductCenter Launches SpecSearch®, New Parametric Parts Search Engine In our continuing effort to enhance our site, eeProductCenter introduces SpecSearch® powered by GlobalSpec. Click here.
Free Subscription to EE Times First Name Last Name Company Name Title Email address   Click here for your Free Subscription to EETimes Europe