This article was initially written as part of the IEEE STARS program.
Transmission is the means by which telephone conversations get from one place to another. Transmission media have evolved over the years in ways that improved quality, removed distance limitations, and increased capacity, in line with the evolution of the predominant medium through multiple transformations. These began with iron wire and have culminated in today’s fiber optic cable. In doing so, telephone service, has become less expensive and more widely available, thus playing an increasingly important role in bringing people together: across town, across continents, and around the world. This article does not cover cellular telephony, which merits its own article.
A telephone network can be divided logically into three interlocking subsystems: 1) telephone instruments, the devices used use to make and receive calls; 2) telephone transmission, the media used to send those calls over a distance; and 3) telephone switching, the devices used to connect and route calls between subscribers. This article addresses the second topic.
To a substantial extent, the history of innovations in telephony is an American story, in part because as late as the 1950s, the U.S. had more than half of the world’s telephones, and because AT&T, operator of the Bell System and its research arm, Bell Telephone Laboratories, played a dominant role in telephone innovation. Over the 125 years following the invention of the telephone, telephone transmission improved through a long series of both incremental and major improvements, to achieve interrelated goals of better sound quality, greater distances, greater capacity, and lower costs.
The telephone lines of the late 1870s and early 1880s, such as used to connect telephones to early telephone exchanges, were single iron or steel wires with ground returns. The wires were strung over aerial poles or occasionally roof supports, and attached to cross arms on the poles via glass insulators. This technology was adapted from telegraph practice, but proved less than ideal for the higher frequencies and greater bandwidths used in telephony, because the attenuation (i.e. decrease in signal strength) was much greater at higher frequencies. While these wires proved adequate for local use, they were subject to electrical interference from atmospheric conditions, telegraph lines, and other telephone lines. Still, hundreds of local exchanges in the United States and other countries opened with single wire iron circuits.
In 1881, a Bell Telephone subsidiary opened the first, commercial, long-distance line between Boston, Massachusetts and Providence, Rhode Island, a distance of 43 miles (70 kilometers). The line worked poorly as there were major problems from both interference and attenuation.
John J. Carty solved the first problem later that year by converting the circuit from one-wire and a ground return to a two-wire or metallic circuit, an innovation patented by telephone inventor Alexander Graham Bell. Due to the added expense, metallic circuits were accepted only gradually, becoming widespread only when interference problems became a much more serious problem in the next decade with interference from wires strung for another new technology, electric power. Between 1890 and 1910, metallic circuits replaced virtually all single-wire lines, both local and intercity. There was a second major cause of interference, crosstalk, between adjacent pairs of telephone wire on a single pole cross arm. Since the distances between wires were unequal, the inductances did not cancel out, leading to crosstalk, or signal leakage. The solution to this was twisted pairs, where the two wires for a circuit were transposed, or twisted, at regular intervals, balancing and thus canceling the inductance. AT&T first tested this successfully on its new New York to Philadelphia line in 1885, and began using it widely after 1891, when Carty worked out a basic theory of line transposition.
It was widely known that copper wire had ten percent of the resistance of steel wire and thus comparably lower attenuation. It was not just the higher cost of copper that inhibited its use, but more importantly that copper wire was unsatisfactory in other ways. It lacked tensile strength, and sagged excessively when strung over miles of overhead lines. In 1877 Thomas Doolittle of Connecticut devised an improved copper wire, known as hard drawn, by drawing soft annealed copper wire through a carefully and slowly graduated series of dies. (Thomas Bolton independently developed this process in England.) In 1881, Bell Telephone tested Doolittle’s wire and in 1884 opened a successful hard-drawn copper line between Boston and New York. Going forward, transposed metallic circuits of hard drawn copper would form the backbone of the long distance plant for many decades. In 1892, using these techniques and thick copper wire, since attenuation was also inversely proportional to the diameter of the wire, AT&T opened with great fanfare a 900-mile (1,400-kilometer) line connecting the nation’s two largest cities—New York and Chicago. This was close to the limit with then available technology. In Europe, a shorter but no less significant line opened in 1894 between London and Paris, some 275 miles (or 450 km). It included a 21-mile underwater cable. By the mid-1890s, telephone transmission had evolved far from the techniques inherited from telegraphy.
From the earliest days, telephone companies put short stretches of wire inside cables to safely cross bodies of water, such as the Hudson River. Wires in cables had much higher attenuation than open wires, because while in open wires resistance was the only cause of attenuation, in cables, there were two additional causes, capacitance and induction. Cables thus were at first only used where open wires couldn’t be strung. By the mid-1880s, the sheer number of wires strung on city streets and the problems caused when they came down, as they could in a storm, led to pressure on Bell Telephone to put the wires underground in cables. This in turn led to the development of improved cables, particularly by W. R. Patterson of the Western Electric Company (Bell’s manufacturing arm). In 1888, Western Electric introduced Patterson’s cable, a new design using gas-infused paraffin as an insulator and twisted pairs to reduce interference. After 1891, paper-insulated dry-core cables supplanted Patterson’s designs, and these cables, typically containing 52 twisted pairs, but later as many as 400, became standard in urban use. Attenuation remained higher than open wire, so cable use remained limited to short, typically urban, circuits.
Around 1899, George Campbell at AT&T and Michael Pupin at Columbia University independently discovered a way to reduce attenuation—by placing inductors (soon better known as loading coils) with known values at theoretically determined intervals along a telephone line. In doing so, they reduced to practice theory developed earlier by Oliver Heaviside. Since inductors resist changes in current, these loading coils reduced attenuation by increasing induction. This increased the distance over which an electrical signal could travel and remain intelligible or, alternatively, allow the use of thinner and therefore cheaper copper wire on a circuit of a given length. Both men applied for U.S. patents and, rather than fight a patent interference suit after the patent office awarded the patent to Pupin, AT&T bought his patent rights. Pupin received over $200,000 in royalties, and AT&T saved an estimated $100,000,000 over the life of the patent by “loading” its lines and using thinner copper.
Under the right circumstances, attenuation could be reduced up to 500 percent. Loading also made it possible to use cables on longer routes. AT&T opened an aerial Philadelphia-Washington loaded cable in 1912 and the British Post Office opened an aerial London-Birmingham-Liverpool loaded cable in 1915-1916. (AT&T’s Western Electric unit had introduced loading coils in Europe in 1903.) Loading coils reduced but did not eliminate attenuation. The open-wire line AT&T planned in 1909, and opened in 1911 between New York and Denver, a distance of 2,000 miles (3,200 km), stretched the ability of a loaded line to carry intelligible conversation to its limit. In Europe, a loaded international line opened in 1914 between Berlin and Milan, a distance of 1350 km (838 miles).
What was needed to send a telephone signal further, such as across the North American continent, AT&T Chief Engineer John J. Carty knew, was a way to amplify (or in telecommunications terms, repeat) the signal. In 1908, Carty publicly committed AT&T to opening transcontinental telephone service in time for the celebrations planned in San Francisco for the opening of the Panama Canal, which was then under construction, even though he knew no means of amplification existed. In 1910 he hired physicist Dr. Harold Arnold, who began work in January 1911, to search for a means of amplification. In October 1912, Dr. Lee de Forest demonstrated his invention, the audion (the first three-element vacuum tube) to AT&T.
Vacuum tubes were the first electronic devices. The audion in particular could be used to manipulate an electric current by use of a second current passing through the grid, the central element of the three. It could produce a small amount of amplification but, beyond that, the tube gave off a blue glow, and amplification ceased. Arnold quickly resolved the problem. The blue glow was caused by the electrons interacting with the residual gas in the tube. While de Forest thought the residual ions were necessary for the proper operation of the tube, Arnold believed, correctly, that if he could increase the vacuum, the device would amplify signals more powerfully for the life of the device.
AT&T bought de Forest’s patent rights, and by mid-1913 was testing the improved devices on commercial lines. The transcontinental line now became a construction project to connect Denver with the separate West Coast network. AT&T completed the line with three vacuum-tube repeaters supplementing many loading coils in July 1914, but held off commercial service until the Panama Pacific Exposition opened in San Francisco in January 1915. By 1920, all of the loading coils on this line had been replaced by an additional nine repeaters, and with this change the bandwidth doubled, and the sound quality improved.
Vacuum tubes led to many additional improvements in transmission. Used as modulators, tubes made it possible to increase capacity by employing what were called carrier circuits. While existing circuits could carry just one signal down a wire pair at voice frequencies, a carrier circuit could send multiple signals down a single circuit, using several simple waves fixed at a series of higher frequencies, and modulating these waves with the sound frequency waves carrying the conversations. Vacuum-tube electronic frequency filters, devised by AT&T’s George Campbell, allowed the several frequencies to be separated so that the individual conversations could be retrieved. This technique became known as multiplexing. AT&T opened its first multiplexed route on an open-wire line between Baltimore and Pittsburgh in 1918. It provided four additional multiplexed channels on each wire pair, and used vacuum tubes both as modulators and repeaters. By its peak of application in the early 1950s, over 1.5 million circuit miles (2.4 million km) of open-wire carrier circuits were in use in the United States alone, carrying up to 32 conversations on each set of four wires. The invention of the negative feedback amplifier circuit by Harold Black at AT&T’s Bell Telephone Laboratories in 1927 reduced the distortion caused by the non-linearity of vacuum tubes and, with this circuit, AT&T began adding carrier circuits to cables as well in the 1930s.
With great distances, and relatively low volume on most routes, the majority of long distance lines in the United States through the 1920s and 1930s were carrier circuits on open wires. Long distance cable was used only on a smaller number of high volume routes. In Europe, by contrast, in 1939, 84 percent of long distance circuits ran through cables. With a few minor exceptions, these cables did not use carrier circuits. In addition to telephone conversations, many long distance circuits throughout the world carried radio programming between radio stations. Among the notable lines built elsewhere in the world was one opened in South America in 1928 that crossed the Andes to connect Buenos Aires, Argentina and Santiago, Chile.
Radio itself was used for circuits where there was no alternative, notably to provide for transoceanic communications. Telephone service opened between the United States and Great Britain in 1927 over a long-wave radio circuit, connected to the conventional wired networks at both ends. Over the next few years, both long- and short-wave circuits were used. As a publicity stunt, in 1935 AT&T staged a round-the-world telephone call between two company executives in adjacent rooms in New York. This intercontinental service was subject to atmospheric interference and costs were very high, initially $25 per minute.
In the 1920s, experiments foreshadowing the beginning of television transmission (which would require transmission bandwidths in megahertz rather than kilohertz) and capacity concerns on a few long-distance routes led to a search for a broader bandwidth, higher capacity, transmission medium. In Great Britain C. S. Franklin patented a coaxial cable for limited use as an antenna feeder in 1928, and then Lloyd Espenschied and Herman Affel of Bell Labs (AT&T's R&D subsidiary) invented a general high-capacity coaxial cable in 1929, for which they received U.S. Patent 1,835,031 in 1931. They reduced to practice theoretical work done by several scientists, including Oliver Heaviside and Lord Raleigh, many years before that showed a system consisting of a central conducting wire centered in a conducting tube could carry an enormous signal bandwidth. The signals would be contained between the outer surface of the central core, and the inner surface of the tube.
Espenschied and Affel designed a workable system, with appropriately sized copper wires and tubes, held in concentric symmetry by spaced dielectric washers, with signal strength retained with repeaters. Additionally, such a cable was by nature self-shielding, and resistant to interference, since the signal was on the inside surface of the concentric conductor. Bell Labs successfully tested the concept in 1929, and in 1936 installed a semi-commercial, repeatered, coaxial cable system with one tube for each direction between New York and Philadelphia, capable of carrying 240 conversations or a single television channel.
Bell Labs next designed an improved system, with four redesigned tubes in the cable, capable of carrying 480 calls or a 4 MHz television channel. Named the L-1 Coaxial cable system, AT&T made a single installation in 1940 between Minneapolis, Minnesota and Stevens Point, Wisconsin before World War II brought a halt to the system's expansion. After the war, construction of coaxial cable lines resumed. By 1948 over 5,000 route miles of L-1 coaxial cable were in service, providing telephone circuits and television networking facilities along the east coast and west to Chicago and St. Louis. Three improved and still higher capacity systems followed, culminating in the L-5 system in the 1970s, which had 22 tubes, and a total capacity of 132,000 telephone calls. By 1980, L-5 Systems provided almost 66 million voice-miles of capacity in the United States.
In the 1956, a joint U.S.-British effort led to the application of coaxial technology to the first transatlantic telephone cable (see separate STARS article on Underwater Cables). With the spread of these deep sea cables, the earlier transoceanic radio circuits were abandoned.
AT&T was not alone in pioneering coaxial cable. Two German companies joined to develop and manufacture coaxial cable for the German government. After several shorter tests, a 200-channel, long-distance, coaxial cable entered service between Berlin and Leipzig in 1938. In Britain, a 40-channel coaxial cable opened between London and Birmingham in 1936, and a submarine coaxial cable to the Netherlands the following year. In 1966, the Soviet Union opened the world’s longest coaxial cable, which ran from Moscow via Khabarovsk and Vladivostok to Nakhodka, a distance of 8500 km (5,300 miles). Three years later, this cable was connected to a new submarine cable between Nakhodka and Naoetsu, Japan, providing a direct cable route from Europe to Japan.
Microwave Radio Relay
While there had been several experiments in the 1930s on using microwave radio as a telephone transmission medium, practical development began only with the invention of the klystron, a practical microwave amplifier, by Russell and Sigurd Varian at Stanford University in 1935-1937. Intense development proceeded throughout World War II, and led to a second transmission system with the high capacity needed for increased telephone volume and television. As early as 1943, Bell Labs began planning for such a system, connecting New York and Boston with a series of seven microwave-radio relay towers between the two cities. Construction of this TDX trial system began in 1945, and the system went into successful commercial operation in 1947. It connected to a new coaxial cable between New York and Washington to provide television transmission facilities connecting stations along its route. It carried four channels at different frequencies around four GHz. Each channel could carry 480 telephone circuits or one television signal. In this, as in subsequent systems, voice or television was carried as modulation of carrier microwaves, which were transmitted between towers over line-of-site routes. The towers, at least on this first route, were placed on hills. Microwave relay had several advantages over coaxial cable: quicker construction, easier construction in difficult terrain, and no need to acquire a continuous right of way from property owners along the route. AT&T developed an improved fully commercial system, TD2, within a few years, and used it to extend its high bandwidth route west from St. Louis to the west coast.
Transcontinental television broadcasting using this combined coaxial cable and microwave system began on 4 September 1951 with the broadcast of President Harry S Truman’s address at the opening of the U.S.-Japan peace conference in San Francisco. As solid-state devices replaced vacuum tubes in the TD-2 system in the 1960s, the number of circuits per channel increased to 1500 in 1973. By the mid-1950s, microwave relay systems had been employed throughout the world, wherever terrain or required speed of construction limited the use of cable. Until the 1980s, most long distance networks in the U.S. and elsewhere were a mixture of microwave relay and coaxial cables.
In the 1970s, communications satellites began to be used as an additional transmission medium, especially as a supplement to undersea cables. However, satellite telephone circuits were inferior, because of the half-second delay resulting from the distance that the radio waves had to travel. Satellites have proven far more satisfactory for the one-way transmission of television signals. The use of satellites for telephone transmission rapidly declined in the 1990s after the spread of fiber optics.
In 1938, Alec Reeves at the ITT (International Telephone and Telegraph Company) research laboratory in Paris developed the theory for telephone signals to be converted into digital signals for efficient transmission and then back. In this system, known as Pulse Code Modulation, the amplitude of an analog signal is periodically sampled, and the sample translated into a digital binary code. But he could not reduce the idea to a practical device with available technology. With the availability of solid-state electronics, PCM was finally used in 1962 for AT&T’s T-1 digital transmission system, over twisted pairs for interoffice trunking (i.e. connecting nearby exchanges.) T-1 spread quickly; by 1967 the Bell system had over 16,000 systems, and 330,000 miles of T-1 lines. A second system, T-2, designed for moderately long coaxial cable lines, up to a few hundred kilometers began being installed in the AT&T network in 1972. It had four times the capacity of the T-1 system. AT&T devised and added digital-under-voice systems, providing purely digital transmission for computer data communications to its microwave relay systems in the mid-1970s. Digital transmission seemed to represent the future, even as the local loop remained analog.
Researchers throughout the industry continued to look for a still higher frequency and hence higher capacity transmission medium. For example, Bell Labs spent many years developing a system of higher frequency millimeter radio waveguides traveling inside highly-polished metal tubes. It was never implemented because fiber-optics proved a far superior approach.
In 1966, Charles Kao, at ITT’s Standard Telecommunications Laboratory in England, demonstrated that there was no theoretical reason preventing a sufficiently pure glass fiber from having a low enough attenuation to allow it to be used as a medium for light waves bearing information. Light waves have a much higher frequency than microwaves. In 1970, a team led by Robert Maurer at Corning Glass developed the first suitable glass, which Corning then continued to improve. That same year, a team at Bell Labs developed the first room-temperature semiconductor laser, providing a practical pulsing light source suitable for a digital optical system.
Test systems in several countries were quickly followed by field trials with customers. GTE installed a test fiber optic cable system in Long Beach, California, in 1977. AT&T quickly followed with one in Chicago, and the British Post Office with a system at Martelsham Heath. Finally, in 1976 J. Jim Hsieh at MIT Lincoln Laboratory developed a laser that emitted light at the same wavelength, 1.3 micrometers, that a fiber developed by Masaru Horiguchi at NTT could optimally transmit, providing for a higher capacity, lower loss, and more efficient system. Other advances followed over the next few years. In 1983, U.S. long distance company MCI, working with Corning, opened a commercial, 1.3 micrometer, fiber-optic cable system between New York and Washington, which AT&T soon followed with a competitive line. Since fiber optic transmission was digital it was well suited for the ever increasing quantity of digital computer data being sent over the world’s telephone lines.
From the mid-1980s, fiber optic installations expanded rapidly all over the globe, and generations of improved systems followed quickly one after the other. Fiber had enormously higher capacity, which increased even further with each generation, and much cheaper operating costs. For example, the last copper transatlantic cable, TAT-7, opened in 1978 with a capacity of 4,000 calls; the first fiber cable, TAT-8, opened in 1988, with a capacity ten times greater. That was just the beginning of a massive increase in capacity; by the late 1990s, new generations of fiber optic systems could carry millions of calls, though in practice by this time most of what was transmitted was data, and not conversation. Or to put it in data terms, coaxial copper cable carried millions of bits, or megabits, per second; early 1980s fiber optic cable, hundreds of megabits; 1990s fiber, gigabits; and 2000s fiber, terabits. Fiber optics rendered all previous telephone network transmission media obsolete. By 2000, copper wire for the most part persisted only in local loops, and microwave systems had been largely decommissioned. The cost of transmitting a phone call to any place on Earth within reach of a fiber-optic cable rapidly approached zero, thus knitting the planet more closely into a single instant communications web, greatly facilitating global commerce. The widespread adoption of fiber optics, among other things, made the global internet possible.
The author thanks members of the STARS Editorial Board and others for review and constructive criticism of this article, with special thanks to Emerson Pugh, Alex Magoun, and Bill Caughlin. The author further thanks Bill Caughlin of the AT&T Archives and History Center for the photographs used in this article.
- 1878, First telephone exchange opens in New Haven, Connecticut.
- 1881, First Commercial Long Distance Line opens, Boston-Providence.
- 1881, Alexander Graham Bell patents the metallic, or two-wire, circuit.
- 1884, Hard-drawn copper wire begins to replace iron wires.
- 1891, Twisted pairs incorporated into telephone lines by John J. Carty.
- 1899, Loading Coil theory developed independently by Michael Pupin and George Campbell.
- 1915, First U.S. transcontinental telephone line, using vacuum tube amplifiers developed by Harold Arnold.
- 1918, First installation of carrier circuits, based on work by George Campbell.
- 1927, Transatlantic telephone service opens via radiotelephony.
- 1929, Broadband coaxial cable invented by Lloyd Espenschied and Herman Affel.
- 1941, First U.S. commercial coaxial cable installation, Minneapolis, Minnesota to Stevens Point, Wisconsin.
- 1947, First microwave relay system in the telephone network, New York to Boston.
- 1956, First transatlantic telephone cable opens, Newfoundland-Scotland.
- 1962, T-1, first digital transmission system installed.
- 1965, Charles Kao conceives of using light sent over glass fibers as a transmission medium.
- 1982, First long-distance, fiber-optic transmission system, New York to Washington, D.C.
- 1988, First transatlantic fiber-optic cable opens, New Jersey to England and France.
References of Historical Significance
B. Gherardi and F. B. Jewett. 1919. Telephone Repeaters. Proceedings of the AIEE 38, 1287-1345.
L. Espenschied and H. Affel. 1931. Concentric Conducting System. U.S. Patent 1,835,031, 8 December 1931, filed 23 May 1929.
L. Espenschied and M. E. Strieby. 1934. Systems for Wide-Band Transmission Over Coaxial Lines. Bell System Technical Journal 13, 654-679.
George A. Campbell. 1937. Collected Papers. New York: American Telephone and Telegraph Company, 1937
C. Kao and G. Hockham. 1966. Circuit Waveguides for Optical Frequencies. Proc. IEE 113, 1151-1158.
References for Further Reading
F.G.C. Baldwin. 1925. The History of the Telephone in the United Kingdom. London: Chapman & Hall
E. F. O’Neill, ed. 1985. A History of Engineering and Science in the Bell System: Transmission Technology. Murray Hill: NJ, AT&T Bell Laboratories
M.D. Fagan, ed. 1975. A History of Engineering and Science in the Bell System: The Early Years 1875-1925. Murray Hill, NJ: Bell Telephone Laboratories
Anton A. Huurdeman. 2003. The Worldwide History of Telecommunications. Hoboken, NJ: Wiley-Interscience
Jeff Hecht. 1999. City of Light: The Story of Fiber Optics. Oxford: Oxford University Press
About the Author
Sheldon Hochheiser is archivist and institutional historian at the IEEE History Center in Hoboken, New Jersey. Prior to joining IEEE, he spent sixteen years as corporate historian for AT&T, acting as both subject matter expert on AT&T history and manager of the corporate archives. While at AT&T, Dr. Hochheiser curated historical exhibits, completed oral histories with company executives, and studied every aspect of the history of the telephone in the United States. He earned a Ph.D. in the History of Science at the University of Wisconsin, and a B.A. in Chemistry-History at Reed College.