Handheld Electronic Calculators

Until the 1970s, most calculating devices were constrained by either the limited number of tasks that they could perform or by their extravagant size and cost. Advances in integrated circuit research would prove to be the solution to both these problems, enabling the manufacture of miniaturised electronic calculators that were both flexible computing machines and, within a relatively short period of time, affordable to most.

The Whipple Museum is in possession of over 400 pocket electronic calculators, a collection assembled by Cambridge architect, Francis Hookham. You can download a fully illustrated catalogue of the Museum’s ‘Hookham Collection’ using this link:

Download Multiply: The Francis Hookham Collection of Hand Held Electronic Calculators (pdf)

Making microchips

In 1968, Hewlett-Packard (HP) released its HP 9100A, the first fully electronic desktop calculator: a limited yet powerful computer for its time. About the size of a typewriter and costing a whopping $4,900, it found its way into the pages of tech visionary Stewart Brand's Whole Earth Catalog, advertised next to beads and moccasins as the 'machine of the future'. Bill Hewlett congratulated his calculator design team on their achievement, but immediately set them to work on a model that was affordable and could fit in a shirt pocket.

At the same time, Texas Instruments (TI) and Sharp Electronics had also jumped into the race to make a miniaturised calculator using only four or five 'integrated circuits'. These 'microchips' (as they are now known) are small plates of semiconductor material composed of transistors and other tiny components which replaced discrete circuits made of large vacuum tubes and resistors.

A new electrical engineering company, Intel, was commissioned to make a 'microchip' for calculators manufactured by another Japanese company, Busicom. Intel bought back the rights to this chip in 1971 and began selling the Intel 4004, the world's first commercially available microprocessor, which launched a great number of developments in microelectronics that quickly swept through the computing industry.

Competing for efficiency

Numerous firms competed to exploit the microchip for miniaturised calculation. The first commercial device to truly fit the 'shirt-pocket' was the Bowmar 901B (Image 1). Bowmar was an LED company that bought its circuits and keypad from TI, and its device was a simple four-function machine, capable of addition, subtraction, multiplication, and division.

The device was bought and sold by competitors like Craig and Commodore, and it rapidly travelled overseas. However, its market lead was quickly stripped away later in the same year by Busicom's Handy LE-120A, prompting a wave of competitive engineering and marketing. By 1972, it was fairly clear how to manufacture and program four-function calculators, so competitors focused on design and pushed for miniaturisation.

More interesting was the commercial viability of scientific calculators like the HP 9100A, capable of computing transcendental functions such as logarithms and square roots. Despite initial scepticism that a market existed for pocket devices that were only affordable to laboratories and firms, the first available pocket scientific calculator, the HP-35, was a commercial hit at the price of $395 in 1972.

Few individuals could afford such a device: a problem British entrepreneur Clive Sinclair, aimed to overcome with the Sinclair Scientific. Interested in producing an inexpensive scientific calculator, Sinclair formed an agreement with TI in 1974 to produce a device with a single chip that could perform four-function calculations.

Holed up in a Texas hotel room with mathematics PhD Nigel Searle, Sinclair had to write a program for scientific functions that would use only 320 instructions. The two created a program using an efficient notation known as Reverse Polish Notation that simplified number storage and order of operations, repeated addition and subtraction for multiplication and addition, and used simplified logarithms and trigonometric functions.

Read more: RPN and scientific calculators

Cutting costs

While highly desirable, most of these early devices were simply too expensive for most users. For example, Sharp's QT-8B cost $495 in 1970 and the Bowmar 901B cost $240 in 1971. But prices soon began to fall rapidly.

Whilst the HP-35 cost $395 in 1972, within three years of its release the retail price had halved to $195. The Sinclair Scientific cost just £49.95 ($99.95) as a kit upon its release in 1974, and within two years it was possible to purchase the same model for £7. Advances in integrated circuit engineering, coupled with better programming, drove these decreases in cost - as did competition between manufacturers.

This probably contributed to a rapid change in perception: soon all but the best calculators were seen as disposable, mundane objects. Decreasing profit margins led to the eventual bankruptcy of companies such as Bowmar, or a move to other markets, as in the case of Sinclair. Only the most popular devices could enable a stable business. The TI-30 (Image 3), one of the most commonly used scientific calculators ever, cost only $25 at its introduction to the market in 1976, and the brand is so recognisable that TI has continued to update it to this day.