Thomas Arithmometer

It was not until the beginning of the 19^th Century that the calculator became a popular device. This happened in 1820 when Charles Xavier Thomas de Colmar (1785-1870), of France, made the Arithmometer   (French Patent N^o 1420, November 18, 1820), a machine based on Leibniz’s design which was capable of performing the four operations in a simple and reliable way. Because of its unidirectional drum, division and subtraction required setting a lever. The Arithmometer was also known as the Thomas Machine.

The Thomas Machine  was very successful and one hundred years later, during the first half of the 20^th Century, the machine was still sold. Thomas received France's Chevalier of the Legion of Honor for his machine. About 1,500 machines were made by the Compagnie d'Assurance Le Soleil, founded by Thomas, and other contractors, between 1820 and 1930.

Like every successful product, the Thomas Machine had many clones, and the term Arithmometer became synonymous with four function calculating machine. An example is the Arithmometer introduced by Arthur Burkhardt in 1878. His company, the Erste Glashütter Rechenmaschinenfabrik, started the calculating machine industry in Germany. Burkhardt’s Arithmometer was produced by the Erste company until 1920 when it merged with the Saxonia company to form the Vereinigte company.

Another machine based on the Thomas design is the MADAS (Multiplication, Automatic Division, Addition and Subtraction) introduced in 1913. The MADAS was manufactured by H.W. Egli of Switzerland. It included the automatic division mechanism patented in 1902 by Alexander Rechnitzer, of Czechoslovakia. Semiautomatic multiplication was added in 1925 in the MADAS Semis, and full automatic multiplication was included in 1927 in the MADAS Superautomat.

Another interesting implementation of the Leibniz principle is the TIM (Time is Money) introduced by Ludwig Spitz in 1907. Like other step-drum machines, it was mounted in a wooden case. In spite of being bulky and expensive, this machine achieved commercial success thanks to Ludwig Spitz's selling abilities.

In 1786, J.H. Müller, an engineer in the Hessian army conceived the idea of what later would evolve into modern computers, the Difference Engine. This was a special machine whose purpose was to evaluate and print mathematical tables by adding sequentially the difference between certain polynomial values.

Müller’s idea was forgotten until 1822, when Charles Babbage (1792-1871) and his colleague Augusta Ada Byron, Countess of Lovelace (1815-1852), of London, obtained government funds to build a programmable prototype. Due to technical limitations, budget constraints, and Babbage interest in developing a more advanced design called the Analytical Engine, the project was not completed. A partial prototype is displayed in the Science Museum of South Kensington, England.

In 1853, George Scheutz (1785-1873) and his son Edvard (1821-1881), of Sweden, built in 1853 the first working difference engine, which is now on display at the Smithsonian Institution in Washington. The Scheutz calculator was also the first calculator with printing capabilities.

Other difference engines were built in Europe and in the United States during the second half of the 19^th Century.

Due to their automatic sequential approach, difference engines are considered to be the precursors of the modern computing machines.

In 1838, Thomas Fowler (1777-1843) a creative inventor and banker from England developed a system based in tables of binary and ternary numbers to facilitate arithmetic calculations.  Later, he used a similar principle to build a ternary calculator.  His prototype, which measured 6 ft x 3ft and was made of wood, was presented to the members of the Royal Society in May 1840.  Unfortunately, Fowler died a few years after, and the machine never went into production, however this is a very interesting example of the different approaches that human ingenuity had taken towards developing a computational device.

A significant change in Leibniz’s design was patented in 1875 by Frank Stephen Baldwin (1838-1925) of the United States. In the Baldwin Principle the nine stepped drums are replaced by a pinwheel which is a variable-toothed cylinder with sets of nine radial pegs that protrude or retract by the action of setting levers through slots located in the opposite side of the cylinder. One set of pegs is used for each digit. A crank on a bi-directional handle activates the digit wheels geared in the variable-toothed cylinder. This invention was awarded the John Scott Medal for the most meritorious invention of the year.

A few years later, in 1878, Willgodt Theophil Odhner (1845-1905), a Swede working for Ludvig Nobel in Russia, patented a pinwheel device. Although Ohdner's patent is completely independent of Baldwin's invention, this device is so similar to Baldwin’s Cylinder that with the passing of time, the terms Baldwin type and Odhner type became synonymous for these kinds of calculating machines.

Between 1873 and 1912, Baldwin made several models based on his principle. In 1900, for example, he patented the Baldwin Computing Engine, a machine that required only one stroke per digit in order to perform multiplication or division.  But the first successful commercial operation happened only when he joined efforts with Jay Randolph Monroe to adapt a full keyboard and started in New Jersey the Monroe Calculating Machine Co. in 1912. Ten years later Monroe became a pioneer in electro-mechanical calculators.

W.T. Odhner, Maschinenfabrik & Metallgiesserei started manufacturing calculators circa 1886, in St. Petersburg, Russia. In addition to the calculators, the shop also made other kind of products, including products for the Russian army during WW1. The calculators were sold under the name Original-Odhner, most of them remained in Russia and are not too common now. 30,000 calculators were made in Russia, and probably less than 20% were exported. 

In 1917, the year of the Russian revolution, his son Alexander moved the production to Göteborg, Sweden, forming the company Aktiebolaget Original-Odhner [13]. In 1918 a firm of Axel Wibel in Stockholm got the rights to manufacture this type of machine under the name Facit. In 1924, Wibel's sold his company  to AB Aetvidabergs Industri, which adopted the Facit name for all of its products.  In 1942 this company bought most of Odhner's company shares, and in 1973 Facit AB became a subsidiary of Electrolux, which discontinued the production of Facit calculators in 1977.

In 1924, Felix Dzerzinsky, founder of the Checka, started the manufacture of these pin-wheel machines as a mean to provide a source of labor for young people in Russia. Initially, Felix calculators were produced in Moscow in a factory which later became the headquarters of the KGB (Felix Dzerzinsky was also chief of the KGB).  They were manufactured until the beginning of the 80's. As is was typical in Russia, the calculators were manufactured in many factories, most of them came from a factory located in Kursk. The gears of Felix calculators were made of a low resistance Zinc alloy instead of the steel and brass used in Odhner calculators.

In 1892 Odhner sold his patent rights to Grimme, Natalis & Co. A.G. of Braunschweig for sale in Germany and some neighboring countries. This company manufactured the Odhner-type calculators under the name Brunsviga, and later, the company itself was named Brunsviga. 20,000 units were sold between 1892 and 1912. In 1910, Brunsviga introduced the Trinks Arithmotype, which was the first and probably the only writing Odhner-type calculator. In 1959 Brunsviga was merged into Olympia Werke (a typewriters manufacturer). By 1970 the company was still selling many kinds of mechanical, electromechanical and electronic calculators. Two mechanical models were marketed as Brunsvigas but identical models could be bought with the name Olympia, too.

Other machines of this type are the Dactyle, Eclair, Esacta, Minerva, Antares, Walther, Facit, Thales, Triumphator, and Alpina. The Alpina was a very small unit produced in Germany in 1961, and is considered to be one the last designs of mechanical calculators.

It was 1878 when Ramón Verea, a Spaniard living in New York invented a mechanism completely different from Leibniz and Baldwin designs (US Patent  207,918). It was an improvement to a mechanism patented by Edmund D. Barbour in 1872, and was based on a partial product multiplying mechanism able to "read" values from a notched Pythagorean table in a way similar to the Braille system. Verea was not interested in producing the machine commercially, he just wanted to "show that a Spaniard can invent as well as an American."

In 1889, León Bollée of Le Mans, France used a similar principle to develop a complex mechanism (US Patent  556,720) which later, in 1892, allowed him to build a machine able to calculate automatically the square root of an 18 digit number in about 30 seconds! Although very ingenious and innovative, Bollée's machines were not produced commercially, as Bollée focussed his attention to the construction of racing cars.

In 1892, Otto Steiger (1858-1923) also patented a multiplication machine based in Verea and Bollée's approach. This machine was manufactured between 1895 and 1935 by Hans W. Egli of Switzerland and sold with the brand name The Millionaire. Some models weighed up to 120 pounds and about 4,700 units were sold.

Another partial-product multiplying machine was the Moon-Hopkins Machine invented in 1911 by Hubert Hopkins of St. Louis. This complex machine, a combination of typewriter and calculator, was marketed by Burroughs as its Class 7 model, in the 1920’s after acquiring the rights from the Moon-Hopkins Billing Machine Co. in 1921.

Eighty years before Jack Kilby invented the first electronic pocket calculator, J.L. Troncet of France invented in 1888 an adding device called the Arithmographe. This device was based in a set of sliding digits embedded in a cardboard frame and built into a notebook. Troncet's invention became so popular that the term troncet was usually applied to refer to this type of device.

The basic principle used in Troncet's adding device was not new, it was already used by the Additionneur of Claude Perrault (1613-1688) in 1685, by C. Caze in 1720, and improved by Kummer in 1847. According to Robert Otnes [3], in the United States, S.S. Young patented a slide adder in 1849, G.B. Fowler patented one in 1863, which was improved in 1890 and sold under the name The Universal Adding Machine. Then, C.E. Locke patented one in 1901.

These devices were inspired by the Abacus. Beads and strings were replaced by slide bars with notches or holes moving inside a flat case. Notches were accessible through a slot. The slot had digits marked on its side. Digits were also marked on one part of the slide and hidden by the body of the case, except for one which was displayed through a small window in the display. One slide was used for each relative position in a number.

Early implementations had the slides arranged horizontally. Troncet arranged them vertically to facilitate its use as a hand-held device. Addition was performed by pushing down the slide, with the help of a stylus, a number of notches equal to the digit to be added. At the top, the slot bent to the left and then one unit down, allowing the stylus to be moved to the left and then down while displacing the left slide by one unit when the result exceeded 9. This rudimentary but practical carry mechanism was the improvement made by Kummer in 1847.

Troncets were made in Germany, USA and several Asian countries with a wide range of features. They were cheap and became so popular that some were still used when electronic pocket calculators took their place in the 70’s.

In Germany,  Addiator was a manufacturer of high quality slide adders. Addiators became so popular that they became synonymous of slide adders of Troncets. Typical Addiator's models were: Addiator Standard (1920), Maximator (1925), Duplex (1950) and Arithma (1960).

As opposed to slide rules which were analog devices, slide adders were discrete devices. However, because of its sliding nature, slide adders were more related to slide rules than to mechanical calculators. Faber-Castell, a well known slide rule manufacturer, bought Addiator and incorporated its device in the back of their slide rules.

Some examples of slide adders are: Addiator, Addifix, Addimax, Correntator, Exactus, Produx, German, Kingson, Groesbeck, Locke, Fowler, Universal, Bair, Rapid Computer, Arithstyle, Dilworth, Calculator, Ve-Po-Ad, Omega, Omega Rechner, Addimult (Tower,) Kalkometer, Magic-Brain, Midget, MBC Pocket, Tasco, Valiant, Vanguard, Windsor and Wolverine.

A special variation is the Slide Band Adders which were basically an American invention. In 1886, Charles Henry Webb of New York invented the Ribbon Adder. In this device, slide bars and notches were replaced by continuous bands and holes. Around 1909, J.H. Bassett & Co., of Chicago, introduced another unit which was sold as a novelty until 1938. These devices are also called Slide Chain Adders, and eventually the ones made in Germany and Japan dominated the US market.

Some examples of slide chain adders are: Bassett, Clark, Comptator (Rapid Computer Adding Machine, 1920), Perplex, Stima, Gem (1904), Hasbro, and Baby.

Less successful were the circular implementations of the slide adders patented in the United States by Mendenhall (1867), Loomis (1868), Taylor (1874), Hart (1878) and Briggs (1879) [4]. This was due to the limitation imposed by the number of digits in the numbers to be added, since slide bars are replaced by concentric discs.

Some examples of circular adders are: Addograph, Stephenson Adder, Quick Reckoner, Sebastian, Webbs, and Eureka.

Between 1850 and 1887, many attempts were made to develop a calculating machine that would use keys as means to enter data. In Europe, key-driven machines were made by V. Schilt (1851), F. Arzberger (1866), A. Stettner (1882), Bagge (1882), d’Azevedo (1884), Petetin (1885) and Maq Meyer (1886). In the United States: D.D. Parmelee (1850), T. Hill (1857), G.W. Chapin (1870), W. Robjohn (1872), D. Carrol (1876), Borland & Hoffman (1878), M. Bouchet (1883), C.G. Spalding (1884), A. Stark (1884), L.W. Swem (1885), P.T. Lindholm (1886) and B.F. Smith (1887) [7].

These attempts illustrate the difficulty in adapting the usage of keys to the wheeled mechanisms. Most of them were single-order machines with no carry mechanism which added only one digit in each keystroke. The few with multiple order capability had slow carry mechanisms, and lacked an efficient way to control the momentum applied to the wheels by the key action.