Figures ▾

Telephone Switching History · Volume 1

Operators and Switchboards

Figure 1 — A working step-by-step "band" on display: a line finder, a first selector, a second selector and a connector side by side. Together these four switches can carry one call from a lifted handset to a…
Figure 1 — A working step-by-step "band" on display: a line finder, a first selector, a second selector and a connector side by side. Together these four switches can carry one call from a lifted handset to a ringing bell in a 10,000-line office, and every volume of this dive comes back to them. Telephone Museum, Waltham, Massachusetts. — ArnoldReinhold, via Wikimedia Commons (CC BY-SA 4.0)

1.1 About This Dive

A telephone by itself is a pair of microphones and earpieces joined by wire. It becomes a network only when something can connect any one telephone to any other on demand, then take the connection apart when the call is over. That something is the switch, and the story of the telephone after 1878 is largely the story of how switching was done: first by people at switchboards, then by machines that obeyed the caller’s dial, then by machines that thought before they acted, and finally by computers that turned speech into numbers.

This dive follows that story from the first commercial exchange in 1878 to the point, around 1976–1982, where digital switching took over. It concentrates on the electromechanical system that carried more calls than any other for most of the twentieth century: the step-by-step system, which its inventor’s company called the Strowger system. Most of the equipment I collect is Strowger step-by-step gear. The middle volumes are written so that the same pages serve a newcomer who has never seen a two-motion switch and someone with a shelf of them on the bench who needs to make a small system work.

The eight volumes:

  1. Operators and Switchboards (this volume). The first exchanges, the multiple board, common battery, and the first private branch exchanges.
  2. Strowger and the Automatic Exchange. The undertaker’s patent, La Porte, the dial, Automatic Electric, and how the Bell System came round.
  3. The Two-Motion Switch. The mechanism itself, part by part.
  4. The Switch Train: Following a Call. Line finders, selectors and connectors, and how a dialled number becomes a path.
  5. Inside the Circuits. Line circuits, the selector’s relays, battery feed, ringing, tones and supervision.
  6. Private Exchanges and Small Systems. PBX, PAX and PABX, and how a 100-line system is built.
  7. After Step-by-Step. Panel, crossbar, electronic switching, and the arrival of digital.
  8. Setting Up and Troubleshooting a Small Strowger System. Power, adjustment, test equipment and fault-finding.

Every figure drawn for the dive comes from one script kept with the project, so it can be corrected and redrawn. Reference documents (patents, period textbooks, Bell System and Automatic Electric publications) are in the site’s Reference Library, linked from the text where they are used.

Figure 2 — Telephone switching from Bell's patent to the first digital local office. Red points above the line are milestones in automatic switching and the Bell System; blue points below are milestones in th…
Figure 2 — Telephone switching from Bell's patent to the first digital local office. Red points above the line are milestones in automatic switching and the Bell System; blue points below are milestones in the manual era, in Strowger's own development and abroad. The bands at the foot show when each kind of switching entered service. All of them ran long past the right-hand edge: step-by-step offices were still working in the late 1990s. — Original diagram drawn for this dive

1.2 The First Exchange: New Haven, 1878

Alexander Graham Bell’s telephone patent, US 174,465, issued on 7 March 1876 (Google Patents). The first telephones were sold or leased in pairs, wired directly to each other: a doctor’s office to his home, a factory office to its warehouse. A telephone that could reach only one other telephone was useful, but limited, and within two years someone had built the obvious improvement.

On 28 January 1878, George W. Coy, operating as the District Telephone Company of New Haven, opened an “Exchange” with 21 customers “who could be interconnected as they desired by means of a central switchboard.” That is the Bell Laboratories history’s own wording, and it is the first commercial telephone exchange (A History of Engineering and Science in the Bell System: The Early Years, ch. 1; IEEE ETHW). The office was a storefront in the Boardman Building at Chapel and State Streets, and subscribers paid $1.50 a month (Connecticut History). Coy’s switchboard is usually described as home-made, assembled from carriage bolts, handles from teapot lids and bustle wire. That detail is widely repeated but rests on secondary accounts. A one-page directory with 50 names followed on 21 February 1878.

Coy was also the first customer for a signalling device: Thomas Watson’s call buzzer (US 199,007, issued 8 January 1878), installed first at New Haven and nicknamed “Coy’s chicken” (BTL Early Years, ch. 3; patent). Before a telephone could be switched, someone had to know it wanted to be, and that problem of signalling runs through the whole history of switching alongside the problem of connecting.

Figure 3 — "First exchange switchboard with twenty circuits equipped, in 1878": the plate as printed in Casson's 1910 History of the Telephone. Casson does not name the exchange, and it should not be taken as…
Figure 3 — "First exchange switchboard with twenty circuits equipped, in 1878": the plate as printed in Casson's 1910 History of the Telephone. Casson does not name the exchange, and it should not be taken as Coy's New Haven board without better evidence, but it shows the scale of the first boards: twenty circuits on one wall-mounted frame. — Herbert N. Casson, The History of the Telephone (1910), via Internet Archive

1.3 Boys, Then Women

The first operators were boys, recruited from the telegraph offices where messenger boys were already on hand. The experiment was short. By about 1882 “women were beginning to be used and more decorum was introduced into the switchroom,” as the Bell Laboratories history puts it. A Cincinnati manager told the National Telephone Exchange Association in 1881 that women’s service was “very much superior to that of boys and men. They are steadier, do not drink beer and are always on hand” (BTL Early Years, ch. 6). One of the original boy operators in Boston, John J. Carty, went on to become AT&T’s chief engineer.

The first woman operator is generally given as Emma Nutt, who started at the Edwin Holmes Telephone Despatch Company in Boston on 1 September 1878. Her sister Stella started a few hours later (Smithsonian). The details of her later career vary between sources and can’t be settled here.

Figure 4 — The central telephone station in Paris, as engraved for Guillemin's El mundo físico (1882). Operators stand at tall boards; each vertical strip of the board is a block of subscribers' lines, and th…
Figure 4 — The central telephone station in Paris, as engraved for Guillemin's El mundo físico (1882). Operators stand at tall boards; each vertical strip of the board is a block of subscribers' lines, and the cords that join them hang below. — Fondo Antiguo de la Biblioteca de la Universidad de Sevilla, via Wikimedia Commons (CC BY 2.0)

1.4 How an Operator Connected a Call

The operator’s tool was the cord pair: two plugs on flexible cords, joined through a small circuit at her position. Every subscriber’s line ended at the board on a jack and a signal: first a mechanical “drop” (a little shutter that fell when the subscriber cranked a magneto), later a lamp. The sequence was the same everywhere:

  1. The caller lifts the receiver (or cranks the magneto). The line’s signal operates.
  2. The operator puts the answering plug into that line’s jack, throws her listening key, and asks for the number.
  3. She tests the wanted line (is it already in use?), puts the calling plug into its jack and presses the ringing key to send ringing current down it.
  4. The called party answers. The two talk through the cord circuit.
  5. When both hang up, supervisory signals on the cord tell her the call is over, and she pulls down the cords.

Each step became a job for the machine. Detecting that a subscriber wants service, finding a free path, testing the called line for busy, ringing it, noticing the answer, and knowing when to take the connection down: every one of these has a counterpart in the step-by-step equipment described in Volumes 4 and 5. A selector or connector does what an operator’s hands and eyes did, one fixed piece of logic at a time.

Figure 5 — How a common-battery operator connected a call. Each line ends on a jack with a lamp; the cord pair, with its listening key, ringing key and supervisory lamps, joins any two lines. Every numbered s…
Figure 5 — How a common-battery operator connected a call. Each line ends on a jack with a lamp; the cord pair, with its listening key, ringing key and supervisory lamps, joins any two lines. Every numbered step here reappears in the automatic switch train as a relay or a switch motion. — Original diagram drawn for this dive

1.5 The Multiple Switchboard

One board serving a few dozen lines could be worked by one or two operators. A board serving thousands could not, because no single operator could reach every jack. The answer came from Chicago. At the American District Telegraph Company of Chicago, General Manager Leroy Firman worked out in late 1878 and early 1879 a board in which every subscriber’s line appeared at several operating positions, so that each operator could reach every line within arm’s length of her own seat. That repeated appearance is the multiple, and the word recurs in automatic switching, where a line’s terminals “multiple” through the banks of many switches. Firman’s patent, US 252,576, was granted on 17 January 1882 and sold to Western Electric (BTL Early Years, ch. 6; patent). The practical multiple board was developed at Western Electric under Charles E. Scribner, who held 441 patents in his career.

A line that appears at several positions can be plugged into at more than one of them, so the multiple needed a way for an operator to tell that a line was already in use elsewhere: the busy test. In later boards the operator touched the tip of her calling plug to the sleeve of the wanted jack and listened for a click. The automatic connector does the same test electrically on its sleeve wiper, as Volume 5 explains.

Figure 6 — "Section of typical New York City telephone exchange": operators seated at a common-battery multiple board, as printed in Casson (1910). The ranks of jacks above their heads are the multiple, every…
Figure 6 — "Section of typical New York City telephone exchange": operators seated at a common-battery multiple board, as printed in Casson (1910). The ranks of jacks above their heads are the multiple, every subscriber's line repeated within reach of every position. — Herbert N. Casson, The History of the Telephone (1910), via Internet Archive

1.6 Magneto, Then Common Battery

The earliest exchanges were magneto (or local-battery) systems. Each telephone carried its own dry cells to power its transmitter and its own hand-cranked magneto generator to signal the exchange. Batteries had to be replaced at every subscriber’s premises, and every call began with a crank.

The cure was the common battery: one large battery at the central office feeding talking current to every line over the line wires themselves. The first significant installations came in the 1890s. The Bell Laboratories history tabulates a common-battery non-multiple board with drops at Lexington, Massachusetts, in December 1893, using Hayes’s repeating-coil battery feed; a multiple with lamps but local-battery talk at Worcester in 1896; and full common battery with lamp signals at Louisville in 1897 (BTL Early Years, ch. 6). Scribner patented the rival “bridged-impedance” feed in 1896. Lamps replaced drops as line signals from about 1894, and by 1900 “complex operating procedures … had been reduced to a few simple manual operations controlled by lamp signals.”

Common battery mattered for automatic switching in two ways. First, a telephone that draws current from the office signals by drawing current: lifting the receiver closes the loop, and the office can detect that with a relay. From about 1908, when the Strowger system became a straight two-wire, common-battery design, every automatic system started a call that way. Second, the office battery became the standard source of power for everything in the building, talking and switching alike. In North America it settled at a nominal 48 volts, positive side earthed, the figure that runs through Volumes 5 and 8.

Figure 7 — A two-position magneto switchboard from Kempster Miller's American Telephone Practice (1905). Each line ends on a drop and jack; the plugs and keys of the cord pairs sit on the shelf in front of th…
Figure 7 — A two-position magneto switchboard from Kempster Miller's American Telephone Practice (1905). Each line ends on a drop and jack; the plugs and keys of the cord pairs sit on the shelf in front of the operators. — Kempster B. Miller, American Telephone Practice, 4th ed. (1905), via Internet Archive

1.7 The First Private Branch Exchanges

Businesses wanted switching of their own almost as soon as towns did. The Bell Laboratories history credits the Dayton, Ohio, Telephone Company with the first PBX “in accord with present concepts,” installed in 1879 at the local Soldiers’ Home: seven extension stations and at least one trunk to the town exchange (BTL Early Years, §6.1). In July 1902 AT&T announced its first standard PBX, later called the No. 1 PBX: a cord switchboard in 30- and 80-line sizes with ten trunks and ten cords. In October 1903 followed the No. 2, cordless and key-operated, for offices too small to employ a full-time attendant.

Kempster Miller’s 1905 textbook gives a definition that still holds:

By “private branch exchange” is meant an exchange, complete in itself … It is a “private” exchange because it is individual to the particular institution which it serves. It is a “branch” exchange because it operates in conjunction with and forms a part of a larger exchange, hence the term “private branch exchange.” (Miller, American Telephone Practice, 1905, ch. XXIII)

Manual PBXs stayed in service for most of the twentieth century. The Western Electric 555, a cord board with plug-in circuits for up to 120 lines and 15 cords, was introduced in 1948, more than twenty years after the Bell System’s first step-by-step PBX, the 740A of 1926 (BTL Switching Technology, ch. 13). Volume 6 follows the PBX into the automatic era, where the small step-by-step PABX became the direct ancestor of the system I am building.

Figure 8 — A Western Electric 555 cord PBX at the Telephone Museum, Waltham, Massachusetts. Introduced in 1948, it is a direct descendant of the 1902 No. 1 PBX: line jacks and lamps above, cord pairs and keys…
Figure 8 — A Western Electric 555 cord PBX at the Telephone Museum, Waltham, Massachusetts. Introduced in 1948, it is a direct descendant of the 1902 No. 1 PBX: line jacks and lamps above, cord pairs and keys on the keyshelf, worked by one attendant. — Daderot, via Wikimedia Commons (CC0)

1.8 Why the Operator Had to Go

By 1900 the manual exchange was a mature, well-engineered system, and it had three problems that more engineering couldn’t remove.

  • Labour. Every call needed an operator’s attention at least twice: to set it up and to take it down. Traffic grew faster than the number of operators the companies could hire and train, and wages rose with every year of growth.
  • Scale. A multiple board had to repeat every line at every few positions. As offices grew, the multiple grew roughly with the square of their size, because there were more lines to repeat and more positions to repeat them on. Large boards became enormous, and expensive in copper and floor space.
  • Privacy and error. An operator could listen, and could connect the wrong number. One undertaker in Kansas City became convinced that an operator was doing worse than that.

Volume 2 takes up his story.

1.9 In the Reference Library

Sources