This page gives a very brief overview of some of the major developments in the early history of electronic calculators and their underlying electronic technology.
The first electronic digital computer (the Eniac) was built using vacuum-tube technology in the mid-1940s. Although first-generation computers were available commercially in the 1950s, the first all-electronic desktop calculator (the ANITA) did not reach the market until early in 1962. The Anita was a full-keyboard machine, similar in appearance to the well-established mechanical comptometers, and was built using vacuum-tube technology.
The first solid-state electronic calculators appeared on the market in 1964. There is much debate about who was first, with competing claims from manufacturers including Sharp and Canon in Japan and Friden in America.
These early machines were built entirely with discrete components - thousands of individual transistors, diodes, and resistors. They were large, complex, and limited to basic arithmetic - add, subtract, multiply, and divide (ASMD).
As small-scale integrated circuit (SSI) technology developed in the mid-60s, so too did calculator functionality, with memory registers, square roots, and programmability following in quick succession. SSI technology allowed typical 4-function calculators to be built with only about one hundred chips, but it also allowed the construction of much larger and more capable machines. This Canon SE-600 calculator used almost four hundred SSI chips and a solid-state memory to provide computer-like programming capabilities.
The development of large-scale integration (LSI) saw a tremendous reduction in size and power requirements and a corresponding increase in capability, leading rapidly to the compact battery-powered scientific and business calculators of the early 1970s. The hundred chips were reduced to four, and then to just one, as the "calculator wars" of the mid-70s reduced the price to throw-away levels.
Subsequent advances in low-power CMOS ICs, liquid-crystal technology, and graphical displays have made things even smaller, cheaper, and more capable, but the basic development of the electronic calculator was essentially complete by about 1972 - over 30 years ago!
The examples on this page have been chosen to illustrate some of the significant early milestones, both in the calculators themselves and in the underlying electronic technology. Please check the calculator index for descriptions of other electronic machines.
Anita 1000 Series, Model 1011, S/N M2971
Functions: ASMD, constant, percent, 1 memory
Technology: Discrete-component DTL, 7 IC chips
Display: 10 digits, Nixie tubes
Dimensions: 10-1/2"W x 13-1/4"D x 7"H, weight 13 pounds
Manufactured: Bell Punch Company, England, January 1970
This 10-key Anita (Model 1011) was made in England in early 1970, although the circuit technology is more typical of the mid-60s. The machine is built almost entirely with discrete-component logic, using individual resistors, diodes, and transistors. The components are mounted on a total of 20 circuit boards, each 9" wide by 2, 3, or 4" high. The illustration shows the typical component density on a small section of one board.
Because of the very limited processing capability of these early
machines, several manufacturers (including Anita and Hewlett-Packard)
adopted the postfix (or "reverse Polish") logic system in order to
simplify the internal circuit design. This system shares the workload
by requiring the operator to enter the problem in a sequence that is
more easily digested by the calculator. Although initially unfamiliar,
the postfix system is entirely logical and rather elegant, and it
aquired many supporters amongst scientific and technical enthusiasts.
It was eventually abandoned in favour of "algebraic" logic as
calculator capabilities improved, but is still widely used in some
computer programming languages.
Canon "Canola" Model 162P, S/N 789036
Functions: ASMD, 2 memories
Programming: 32 steps, fixed program
Technology: Bipolar DTL SSI, delay line memory
Display: 16 digits, Nixie tube
Dimensions: 12"W x 15-1/2"D x 6"H, weight 14 pounds
Manufactured: Japan, December 1969 (manual dated September 1969)
Although very few (if any) of the mechanical calculator companies survived the transition to electronics, the Canon Camera Company of Japan successfully broadened their mechanical product lines to include electronic calculators, as a first step into the business equipment market. Canon's first calculator, the Model 130, was introduced in 1964. It was a 13-digit 10-key machine using discrete-transistor logic. The 16-digit 160-series followed soon after.
This Model 162P is a basic 4-function machine built using general-purpose small-scale integrated circuit (SSI) chips. The chips are from an early Texas Instruments "Semiconductor Networks" (SN) series, using diode-transistor (DTL) logic. The date code "6946" on these two chips shows that they were made in week 46 (November) of 1969. (The "later" and more common SN54/74 TTL chips were actually first made in 1964, but were restricted to the US military. They did not appear in Canon's calculators until around 1970).
The 162P has over 100 separate chips occupying four circuit boards,
each about 11" x 5", plus a considerable amount of discrete-component
logic. There are two internal memory registers, implemented using a
mechanical (ie, acoustic) delay line from NEC. The "P" suffix indicates
a programming capability, which is rather limited by today's standards.
When the manual talks of "inserting a program card into the machine", it
actually means removing the case and plugging in an 11" x 5" circuit
board, onto which a technician has soldered a diode matrix to represent
the 32 program steps that you require.
Casio Model AL-2000, S/N 205664
Functions: ASMD, square root, 4 memories
Programming: 30 steps, 14 instructions, no conditionals
Technology: DTL (SSI and discrete), MOS MSI, magnetic core memory
Display: 14 digits, miniature Nixie tube
Dimensions: 13-1/2"W x 16"D x 6"H, weight 15 pounds
Manufactured: Casio, Japan, 1970. (Sold by Remington)
This Casio AL-2000 from 1970 is a very early example of medium-scale integration (MSI) technology in a user-programmable electronic calculator. Although built primarily with a mixture of discrete-component diode-transistor logic and about 90 general-purpose SSI chips from Hitachi, the machine uses two special-purpose MSI chips manufactured by (American) Philco to provide a higher level of integration for some logic functions in the arithmetic unit. The date code on this MSI surface-mount flat-pack shows that it was produced in September 1968.
The AL2000 uses a magnetic core memory module to provide a total of
512 bits of storage, sufficient for four 14-digit memory registers and
thirty 4-bit program instructions. The programs can be entered and
edited from the keyboard, and are retained in the core memory when the
calculator is switched off. Click on the link above for a more
Although the American semiconductor companies led the development of ICs through small, medium, and large-scale integration (SSI, MSI, LSI), the Japanese seemed to have the edge in calculator production and marketing. Many of the early electronic calculators sold by American companies were actually made in Japan, often using custom chipsets made by the likes of Rockwell and Texas Instruments.
In 1969 a small American company called Intel was contracted to build a chipset to take the Nippon Calculating Machine Corporation (Busicom) into the electronic age. Rather than designing a custom calculator chipset, Intel designed a general-purpose processor that could be programmed to act like a calculator, under the control of instructions stored in a read-only memory (ROM). Intel re-purchased the rights to the design, and the world's first microprocessor - the 4004 - was released to industry in 1971. Had the customer been a little more commercially astute, we might all have been using Busicom computers today!
The microprocessor architecture (processor plus program store) rapidly
replaced hard-wired logic designs, not only in calculators but in almost
all areas of electronics. The illustration shows a "clone" 4004 made by
Sharp Model EL-8, S/N 020054Y
Technology: MOS-LSI (4 chips), portable, NiCad battery
Display: 8 digits, vacuum fluorescent tubes
Dimensions: 102mm W x 164D x 70H
Weight: 0.72kg including batteries
Manufactured: Sharp, Japan, December 1970.
The Sharp company of Japan are credited with many of the "firsts" in the early development of electronic calculators. They introduced the first fully-transistorised 10-key calculator (the CS-10A) in 1964, the first MOS-LSI calculator (the QT-8D) in 1969, and the first portable battery-powered calculator (the EL-8) in 1970.
The QT-8D was a four-function desk calculator which was less than half the size (or one quarter the volume) of the Canon or Casio machines above. The hundred or more SSI chips had been condensed into a set of just four LSI chips, which were manufactured by (American) Rockwell. The QT-8 chipset was re-arranged into an even smaller package to produce the first battery-operated portable calculator, the Sharp EL-8, in November 1970. The machine illustrated was built early in 1971.
One of the charming features of early calculators is the lack of
standardisation in the user interface. The EL-8 has both Multiply and
Divide marked on the same key - which one you get depends on which
Equals key you press afterwards!
Hewlett-Packard Model HP-35, S/N 1230A80555
Functions: ASMD, trig, log, 4-register stack, 1 data memory
Technology: MOS-LSI, portable, NiCad battery
Display: 10 significant digits plus 2 exponent, 7-segment LED
Dimensions: 3-1/4"W x 5-3/4"D x 1-1/4"H, weight 9oz
Manufactured: Hewlett-Packard, USA, July 1972.
Hewlett-Packard's design brief for the Model 35 was for "a shirt-pocket-sized scientific calculator with four-hour operation from rechargeable batteries at a cost any laboratory and many individuals could easily justify" (HP Journal, June 1972). "Scientific" in this context meant a calculator with similar functions to the engineer's slide rule - powers, roots, and reciprocals; logs and exponentials; and trig functions. The HP-35 was the first "pocket" calculator to deliver these functions - along with 10-digit accuracy and 200-decade numeric range!
The calculator contains a total of eight ICs - five custom MOS-LSI chips, two display drivers, and a clock generator. There are two circuit boards, one containing the keyboard, display drivers, and the seven-segment LED display, and the other (illustrated) containing the power supply and the MOS logic circuits. The control and timing circuits and the arithmetic unit are in the two ceramic packages, with the program and data ROMs in the three 10-pin metal-can packs. Each ROM contains 256 10-bit instructions; the LSI chips in total contain around 30,000 transistors. The calculator is powered from a replaceable battery pack containing three rechargeable AA-size NiCads.
The HP-35 was introduced in mid-1972 at a price of US$395. It was
an overwhelming sucess, set a new benchmark for portable calculators,
and marked the end of the road for the engineer's slide rule.
Compucorp "Micro Scientist" Model 324G, S/N 4201133
Functions: ASMD, trig, log, special functions, 10 memories
Programming: 2 x 80 steps, all keystrokes, no conditionals
Technology: MOS-LSI, portable, battery/AC
Display: 13 significant digits plus 2 exponent, 7-segment neon
Dimensions: 5-1/2"W x 9"D x 3"H, weight 3 pounds (ex batteries)
Manufactured: Compucorp USA, 1972-73.
The Computer Design Corporation of Los Angeles built several related models of portable battery-powered calculators in the early 1970s. The "320 Series" were all based on a set of seven 40-pin LSI chips from Texas Instruments. The different models (Scientist, Statistician, etc) were produced in the factory simply by changing the labels on the keyboard and fitting the appropriate ROMs to the plug-in memory board. (The extension of this idea to provide user-installable ROM modules with replaceable keyboard overlays first appeared in the HP 9810 desk calculator, and (much later) in pocket calculators such as the Texas Instruments SR-59 and the HP-41 series. The need for these supplementary modules soon disappeared as ROM chip capacities increased).
The illustration shows the memory board from the Model 324G. There is a 40-pin memory controller from the Texas Instruments chipset, three 28-pin ROMS from various makers, and an empty space for a fourth ROM. Four 2102 static RAM chips from Intel provide a total of 4096 bits of memory for user programs and data storage. The chips on this board are dated 1973, although the rest of the machine is dated mid-1972.
The Model 324G "Micro Scientist" was close to the state of the art in
electronic calculators at that time. It was portable and battery
powered, had 13-digit precision, used scientific (exponent) notation for
200-decade numeric range, calculated trig functions in degrees, radians,
or grads, converted rectangular to polar coordinates, converted degrees,
minutes, and seconds to decimals, calculated powers and logs, did the
inverse of all the above, had ten storage registers, did
direct-to-register arithmetic, and could hold two independent user
programs of up to 40 keystrokes each. (Click on the link above for
a more detailed description).
Sharp Model EL-501, S/N 72032099
Functions: Scientific, pocket/battery
Technology: MOS-LSI (single chip)
Display: 8 digits, 2 exponent, vacuum fluorescent display
Dimensions: 3-1/8"W x 5-1/2"D x 1-1/4"H, weight 8oz (225g) incl. batteries
Manufactured: Sharp, Japan, September 1977.
The rapid development of LSI technology during the 1970s sparked a period of intense competition in the calculator market. Manufacturers and chip builders leap-frogged each other with models that were smaller, cheaper, and more capable, leading eventually to single-chip pocket calculators with scientific functions at throw-away prices.
This Sharp EL-501 calculator from 1977 was cheap enough to be widely used by senior school children. It is smaller, lighter, and has more fancy functions than the HP-35, although it is noticeably less accurate.
The illustration shows the logic board and display of the EL-501, with the connection to the membrane keypad at the bottom. All of the calculator logic and the display driving circuitry is contained within the single 36-pin LSI chip.
The calculator is powered from two disposable AA-size batteries via the discrete-component inverter in the lower section of the board. Five separate voltages are needed to run the MOS IC and the display. The total battery drain is about 60mA at 3V, or 200milliwatts.
The MOS LSI and vacuum fluorescent displays were soon replaced by CMOS LSI chips and liquid crystal displays, reducing the power requirements by a thousand times (to around 200 microwatts) and providing a true "single-chip" calculator with no additional components.
Hewlett-Packard Model 9100B, S/N 0938A03791
Functions: ASMD, trig, log, special functions, 32 memory registers
Programming: All keystrokes, subroutines, 392 steps, magnetic card storage
Technology: Discrete diode-resistor logic, magnetic core memory
Display: 10 significant digits, 2 exponent, 3-register CRT display
Dimensions: 16"W x 19"D x 8-1/2"H, weight 40 pounds
Manufactured: USA, October 1969
The Hewlett-Packard 9100 calculator combined a full range of scientific functions with an extensive programming capability, including conditional jumps, subroutines, and an in-built magnetic card reader for storage of programs and data. Optional accessories included a dot-matrix printer and an 11" x 15" flatbed plotter.
Although, functionally, the HP 9100 looks like the culmination of early calculator development, this machine was in fact Hewlett-Packard's first calculator, and is older than all of those illustrated above. (The Model 9100 was introduced in mid-1968; the illustration is the "B" version, with increased memory capacity).
The HP 9100 is built entirely with discrete transistors, diode-resistor logic, and magnetic core memory - there are no integrated circuits. The illustration shows a few of more than one thousand diodes and resistors on the main circuit board.
This remarkable calculator was at least five years ahead of its time, and served as the benchmark against which all others were judged. Its price of just under US$5000 was four times the price of a top-line mechanical calculator, twice the price of most electronic machines, and more than a year's wages for many people at that time.