A Brief History of Calculators
Part III

Getting Ready for the 20th Century

by
James Redin

History - Part IIHistory - Part III

[ Part I ] [ Part II ] [ Part III ] [ Part IV ] [ Part V ] [ Part VI ]

The "Macaroni Box" and Dorr Felt’s Comptometer

In 1885, Dorr Eugene Felt (1862-1930), of Chicago, used a wooden macaroni box, staples, rubber bands, wire, string, and meat skewers to make a prototype of what became the first practical multiple order calculator to use a keyboard, the Comptometer. The "Macaroni Box" prototype is now on display at the Smithsonian Institution in Washington.

The design was based on a fast carry mechanism that acted while the keys returned to their original state after being depressed. A detent toothed lever controlled the wheel momentum by bringing the wheel to a full stop.

By 1886 Felt had completed his first metal prototype. In March 1887 he applied for a patent which was granted in October 11, 1887 (US Patent 371,496). Then he started the Felt & Tarrant Mfg. Co. in partnership with Robert Tarrant.

In 1889, Dorr E. Felt obtained a patent for his Comptograph which was basically a Comptometer with a printing device. This machine is now in the Smithsonian Museum. In 1890 Felt also produced a model with wide carriage able to print several columns of numbers.

In 1957 the Felt & Tarrant Mfg. Co. became the Comptometer Corporation and had manufacturing operations in the United States and England.  The English operation along with the rights to use the name Comptometer were sold in 1960 to Control Systems Limited, owner of the Bell Punch Company Ltd. and its distributor Sumlock Ltd. which became Sumlock Comptometer Ltd.  The US manufacturing operations were stopped and Comptometers were manufactured in England by using a plastic-metal mechanism of inferior quality [14].   In 1961, the Comptometer Corporation merged with Victor Adding Machine Co. to form the Victor Comptometer Corporation [8].  The new models introduced by Victor used the internal Bell-Punch mechanism.

The Comptometer used a "multiple order keyboard" also called full keyboard which consisted of a matrix with 9 rows of keys, one for each digit (1 to 9). The number was entered by pressing one digit in each column. There were no Zero keys because zero was represented by the absence of a keystroke in the corresponding column. This arrangement of keys, initially introduced by Thomas Hill in 1857, became very popular during the first half of the 20th Century.

In the Comptometer, pressing a key advanced a wheel mechanism to cause the positional value of that key to be added immediately to the displayed total. Therefore, for a given position in a number, the user could enter the non-zero digit or a combination of digits that would add-up to the digit, for example: 3 and 4 instead of 7.  Users were also able to press keys simultaneously in several columns while entering a number.  Notice that a full keyboard arrangement already contains the positional value information for a given digit, which no longer exist with modern 10-key arrangements where digits must be entered in sequential order.  All these features allowed for faster data-entry tricks used by skilled Comptometer operators. Subtraction required the addition of the nine’s complement. A lever on the right side was used to clear the display.  Comptometer operation became a formal profession and required a lot of training.  A good Comptometer operator was able not only to perform additions and subtractions at fast speed, but also multiplications and divisions by applying repeated additions and complementary subtractions, respectively.

The usage of the low value digits became so popular that some companies introduced half-keyboard versions of the Comptometer. Torpedo (1935), Plus (1935), and Contex (1950) are typical examples. These keyboards only had digits from 1 to 5.

Comptometers weighed from 17 to 25 pounds and were priced between $300 and $400.


William Seward Burroughs

William Seward Burroughs (1857-1898) worked between 1880 and 1884 in developing an adding machine with a full keyboard and printing capabilities. He applied for a patent on January 10, 1885, almost two years before Felt applied for his patent, and was granted the patent on August 21, 1888, about a year after Felt was issued his patent.

The initial implementation of the printing device printed the totals but not the individual entries. This limitation was corrected before the patent was granted.

Once the number was entered in the keyboard, a printing mechanism was activated when the user pulled a handle on the side of the machine. The adding wheels were activated when the lever returned to its original position.

In 1886, Burroughs founded the American Arithmometer Company in St. Louis, Missouri. By 1889 the company had sold about 50 machines, but they were difficult to operate. Burroughs immediately improved them by inventing the dash pot, a mechanism used to regulate the pull in the machine’s handle. He wanted to get a fully reliable and solid adding machine that would be useful for handling the financial operations in a business. In order to achieve this objective, Burroughs had to combine his invention with several functional features of Felt’s inventions. This is reflected on the patent awarded to Burroughs in May 5, 1892. Burroughs achieved his objective but he only saw the beginning of his success for he died in 1898.

In 1904 the company moved to Detroit, and in 1905 it was renamed to Burroughs Adding Machine Co. Twenty years later they had sold about one million machines and had become the largest manufacturer of adding machines in the United States. It remained as one of the leading manufacturers of mechanical office equipment until the 1950's. After WW2 it expanded its operations to include computers, and in 1953 its name was changed to Burroughs Corporation. Finally, in 1986 it merged with Sperry Corporation to form Unisys Corporation. [9]

Early Burroughs adding machines, like the Class 1 introduced in 1905, weighed more than 63 pounds and had glass side walls to display its mechanism. Because of this they were often called fish tanks [10]. They had "blind" printing mechanisms, meaning that the user was not able to see the printed numbers hidden behind the machine. A display on the frontal bottom of the machine showed the number entered or the total. Class 2, introduced in 1910, had two registers and a "transfer total" key. These models were made until 1914. Burroughs adding machines were probably the first to use direct addition of negative numbers without complementary subtraction.

William H. Pike Jr. invented the "visible" printing mechanism while working for Burroughs. This mechanism was used in the Class 3 models, commonly referred to as flatbeds. Pike left Burroughs temporarily and manufactured his own machines.

Class 3 models with mobile wide carriage were manufactured between 1911 and 1929. It is a classical example of the elegant design combined with speed, reliability and durability, typical of Burroughs’ products. The machines became very complex, some Class 3 models weighed about 50 pounds and had more than 5000 parts. Some operations required the movement of about 2000 parts at the same time! [11]

Burroughs also produced a direct adding machine, the Class 5 model, with no printing capabilities. It was so similar to the Comptometer that Burroughs was sued for it.


The beginnings of the 10-key machine

William W. Hopkins of St. Louis, invented the Standard in 1901. This machine manufactured by the Standard Adding Machine Co., had one row of 10 digit keys and marked a significant departure from the popular full keyboard.

DaltonThe following year, in 1902, James I. Dalton introduced the Dalton, an adding/printing machine designed by Hubert Hopkins, which had two rows of five digits with the sequence: [2 4 5 7 9] [1 3 0 6 8]. The Dalton was very successful and over 150 models were introduced until 1928. Earlier models had glass walls in the sides showing the internal mechanism.

Other examples of efforts made towards the development of machines with simplified key arrangements are the Adix (1903), Diera (1906) and Adix (1903)Kuli (1909) adders manufactured by the Adix Company of Pallweber and Bordt of Mannheim.  These machines had only 9 digit keys, the 10th key actually represented the number ten rather than the digit zero.  The reason was that they were used to add single rows of digits, so the zero had no usage.  The user had to keep track of the higher order digits on the results and reset the machine before re-entering those digits.  The system was not easy to use and didn't provide significant savings of time, so their production was shortly discontinued.

In Germany, an interesting case is the Astra produced by Astrawerke in 1922. It was based on the Dalton machine but had [00, 000] multi-zero keys. Another example based in the Dalton keyboard is the Madix produced in the 1950's by the VEB Feinwerktechnik Dresden, the former Ascota Werke.'

However, the modern 10-key design arranged in three rows with the sequence: [7, 8, 9] [4, 5, 6] [1, 2, 3] plus a zero key, was introduced in 1914 by Oscar J. Sundstrand of Rockford, Illinois. Later, with his brother David, he founded the Sundstrand Adding Machine Co.

The Remington-Rand is an example of a company that used the new 10-key design by 1920.

In 1927, Sundstrand sold the rights to Underwood-Elliot Fisher Co. and worked for this company until 1949. In 1950 he joined the Victor Adding Machine Co.


Christel Hamann

In 1902, Christel Hamann of Berlin, patented the Gauss, a circular calculator based on a modified Leibniz cylinder actuating on numeral gears located in a radial direction around the shaft. This was probably the basis for the design of the Curta calculators in 1943.

In 1905, Hamann designed for the Mercedes Bureaumaschinenwerke, the Euklid model. This machine, patented by Hamann in 1911, discontinued the Mercedes-Plus model, and was sold as the Mercedes-Euklid.

The Euklid was based on a principle different from the machines made at that time. The machine had a set of ten parallel racks, one for each digit 0 to 9. Each rack moved a distance proportional to its corresponding digit while actuating over unidirectional wheels to perform addition, subtraction, multiplication or automatic division. The Mercedes-Euklid 1 had only stop-division and the racks were set by levers, but later models, like the Mercedes-Euklid 3 and the Mercedes-Kopernikus (a very rare model) had automatic division and full keyboard. A rotating crank was used for multiplication and division.

In 1925, Hamann designed the Hamann-Manus machine, which has the same exterior appearance as the Odhner machines, but, instead of being based on the Baldwin/Odhner pinwheel, uses a completely different mechanism. This machine was produced by the Deutsche Telephonwerke und Kabelindustrie Aktiengesellschaft in Berlin. There were many models based on the Hamann-Manus mechanism.


The Victor Adding Machine Co.

This company is one of the few manufacturers of mechanical calculators that survived the technological changes of the electronic revolution. Founded in 1918 by Carl Buehler of Chicago, its first product was the Victor 110, a full keyboard adding machine similar to the Comptometer with no printing capabilities designed by Oliver David Johantgen (1875-1932). Three years later this model was improved to have printing capabilities. It produced several models such as the series 300 introduced in 1923, the series 500 in 1931, the series 600 and 700 in 1939. [11]

By 1952 Victor had sold more than one million machines. As noted before, Oscar Sundstrand joined the company in 1950 and designed the Victor printing calculator introduced in 1954. Later, in 1961, Victor merged with the Comptometer division of Felt & Tarrant Mfg. Co. to become the Victor Comptometer Corp.

In 1968 Victor joined forces with Nixdorf and entered into the market of electronic desktop calculators with the Victor series 1500, and has stayed strong in this segment of the market until now.


Marchant Calculators

In 1911, Rodney and Alfred Marchant founded the Marchant Calculating Machine Co. in Oakland, California.  The company started making calculators based on the Original Odhner calculators. Their first two models were the Pony and the Standard. Later, in 1923 Marchant adapted a full keyboard to the Pony model, and in the 1940's abandoned the Odhner pin-wheel to use its own system based on a proportional wheel. One example is the Figurematic considered to be one of the fastest calculators of their time.  In 1958 the company was bought by Smith Corona, which became know as SMC. 


The Walther Company

Carl Walther, of East Germany founded the Walther Company in circa 1886. Its main products were fire arms for hunting and target practice. In 1924, as a way to recover from the problems caused by Germany's economical depression, the Walther Company started manufacturing calculators based on the Odhner pin-wheel.

In 1931 it introduced the TASMA line, which included a full keyboard and had printing capabilities. The company practically dissapeared during the WW2, but re-emerged after the war in 1948 with a new line of products. The Comtess (S32) adding machine, the Multa-32 and the Multa-33 calculators became very popular in Erurope due to its sturdy construction a high reliability.

During the 70's the company started manufacturing electronic calculators, but it couldn't survive the competition and was closed sometime at the beginning of the 80's.


Friden

Carl Friden (1891-1945), a Swedish immigrant to the United States, founded Friden, Inc. in 1934 after leaving Marchant, where he was chief designer, and introduced the Friden S, a full keyboard motor-driven calculator based on the Thomas machine. Early models included the ST and STW.

In 1952, Friden introduced the model SRW with square root function. It weighed 42 pounds and had both, a full and a 10-key keyboard. The model SRQ introduced in 1958 had also X² function.

Friden calculators did not have printing capabilities but were regarded as high quality calculators.

The company was bought by Singer in 1965 after the introduction of the Friden EC-132, one of the first electronic calculators.


The Curta, a Mechanical Wonder

It was designed by Curt Herzstark (1902-1988) of Austria, while prisoner at Buchenwald concentration camp in 1943. The Curta is a cylindrical four function calculator, similar to the Gauss calculator designed by Hamann in 1905, with a clockwise "operating handle" on top and sliding digit levers on the side. Digits are displayed in small openings on top of each lever. Its coffee grinder appearance gives one more meaning to the expression "number crunching." Paradoxically, Leibniz’s stepped drum, which had the disadvantage of being too bulky, was used, with some clever modifications, as the heart of this small piece of ingenuity.

Curt Herzstark was not new to the art; his father Samuel Herzstark (1867-1937), founded in 1905 the Rechenmaschinenwerk Austria Herzstark & Co. a manufacturer of Thomas based Arithmometers (Austria models). In 1927, Curt Herzstark also invented a mechanical memory for the Multimator, a multicolumn machine manufactured by Astrawerke in Germany.

After the war, in 1946, Curt Herzstark was invited by the Prince of Liechtenstein to establish a manufacturing plant for the Curta in the Principality of Liechtenstein, east of Switzerland. The company Contina AG was founded for this purpose and Herzstark entered as partner and technical director. By 1949, the Curta I, one of the most impressive displays of high precision craftsmanship among mechanical calculators, was introduced. In 1950 Curt Herzstark was issued US Patent 2,525,352.

Curt Herzstark sold the patent rights to Contina and retired in 1952. Contina introduced the Curta II in 1954. Later, in 1966, Contina was sold to Hilti, a Swiss company. Hilti continued the production of the Curta until 1972 when the electronic pocket calculators took over the market.

Bruce Flamm estimates that about 80,000 of the Curta I, and 60,000 of the Curta II were made.


The First Mechanical Binary Computer

In 1936, Konrad Zuse (1910-1995) a civil engineer from Germany built up in the living room of his parent’s house in Berlin, the Z1, the first mechanical binary computer.  Zuse had made a thorough analysis of the mechanical calculators available at that time looking for an efficient way to perform the lengthy and tiresome calculations required to design building structures.  His conclusion was that the decimal approach was not the most suitable for this purpose and between 1934 and 1936 he designed a binary device able to store intermediate results in memory and perform sequences of arithmetic operations programmed in a punched paper tape (old movie film.)  In other words, he invented the binary computer.

Nowadays when we think of computer memory we associate the concept with memory chips, in the 60’s the same concept was associated with ferrite core boards, but in 1936 the concept didn’t exist.  It is hard to believe that computer memory could be made up from metal plates, and that is exactly among many other new concepts what Konrad Zuse’s genial mind conceived in those days.

Zuse’s mechanical memory device consisted of 64 words of 22 bits each, arranged in three storage blocks, one for the sign and exponent, and two for the mantissa. Each bit was represented by the position of a fixed rod inserted trough a sliding metal sheet able to move to the right or the left defining a 0 or a 1.

The Z1 was just the starting point of a series of impressive developments made by Zuse in the area of digital computers, including the invention of the first programming language, the Plankalkül.


Electromechanical Calculators

According to George Chase [1], the first motor-driven calculating machine was the Autarith. This machine was designed by Alexander Rechnitzer, of Czechoslovakia, in 1902 and manufactured by the Autarith Company, Ltd. Of Vienna. Based on the Thomas machine, this was also the first calculator to use the automatic multiplication and division mechanism patented by Rechnitzer.

In 1907, Samuel Herzstark, father of the Curta inventor, also produced motor-driven versions of his Thomas based calculators in Vienna.

In 1912, the Eclair, an Odhner type machine introduced in France, was operated by an electrical motor.

In 1913, the Mercedes Bureaumaschinenwerke introduced the Mercedes-Euklid 7, which was a motor-driven version of the Mercedes-Euklid 1.

In 1915, Marchant provided his Pony model with an electric motor drive.

In 1920, Leonardo Torres Quevedo (1852-1936), of Madrid, presented an electromechanical machine wired to a typewriter at the Paris Calculating Machine Exhibition. It performed the four operations and used the typewriter as input/output device. A number of typewriters could be supported simultaneously. The machine was never produced commercially. Torres Quevedo was a very prolific inventor, he also developed in 1895 his Algebraic Machine, an analog machine able to calculate the roots of an arbitrary trinomial equation, and an automated mechanical chess!

In 1922, Monroe introduced the electric version of its Model K non-printing calculating machine. It had a large external driving motor.

By 1925 Burroughs had most of its models, including Class 1 and 2, available with motor driven options. Motors had evolved from a continuous operation engaged through a clutch when the addition was performed (1906), to a point where, due to its high starting power, they ran only while an operation was actually performed.

In 1929, the Walther Company introduced the EMKD, an electrified version of its Odhner based machines.

In 1931, Victor introduced the Victor 511S-12, its first electric adding machine.

Friden, a manufacturer of quality electromechanical calculators is founded in 1934.

By the 1940’s electric motor driven mechanical calculators had become a common desktop fixture in business and engineering offices.

Once those tiny electrons started taking over the guts of the mechanical calculators, it took little time until they became part of its brain. In 1961, Sumlock Comptometer, of England, introduced the ANITA (A New Inspiration To Arithmetic), the first electronic calculator, and marked the beginning of the end of an era that lasted for 338 years.


Reference Sources
Internet Sources

French Translation

 
Copyright © James Redin - Revised: August 23, 2014.