[Previous Part] In 1837, American scientist and teacher Joseph Henry took his first tour of Europe. During his visit to London, he made a point of visiting a man he greatly admired, the mathematician Charles Babbage. Accompanying Henry were his friend Alexander Bache, and his new acquaintance and fellow experimenter in telegraphy, Charles Wheatstone. Babbage told his visitors of his upcoming appointment to demonstrate a calculating machine to a member of Parliament, but was even more excited to show them his plans for another machine, “which will far transcend the powers of the first…” Henry recorded the outlines of Babbage’s plan in his diary:¹

[t]his machine is divided into two parts one of which Mr B calls the store house and the other the mill. The store house is filled with wheels on which numbers are drawn. These are drawn out occasionally by levers and brought into the mill where the processes required are performed. This machine will when finished tabulate any formula of an algebraic kind.

The historian cannot help but feel a chill at this kind of coincidental intersection of human lives. Here two threads in the history of computing crossed, one nearing its end, the other only beginning.

For though Babbage’s machine is often treated as a starting point for the history of the modern general-purpose computer, the connection between the two is tenuous at best. His machine (such as it was, for it was never built) was rather the culmination of the dream of mechanical computation. This dream was stimulated by the increasingly intricate clockwork devices built by craftsmen from the late medieval period onward, and first fully articulated by Leibniz. But no general-purpose computer was ever built on a purely mechanical basis – it was simply too complex a task.²

The electromagnetic relay³, on the other hand, conceived by Henry and others, could be composed with relative ease into computation circuits of previously unfathomable complexity – though this would take decades to come to fruition, and was neither foreseen nor dreamed of by Henry and his contemporaries. It was the progenitor of the myriads of transistors that make possible the digital meta-world that now overlays so much of our everyday experience. Relays filled the guts of early programable computing machines, which ruled for a brief interval before being overtaken by their purely electronic cousins.

The relay was invented several times, independently, in the 1830s. It was protean in its conception (its five inventors had at least three different purposes in mind) and, as we shall see, in its use. But it is convenient to think of it as a dual-purpose device. It could be used as a switch, to control another electrical device (including, significantly, another relay); or as an amplifier, to turn a weak signal into a strong one.

Switch

Joseph Henry combined in one person a deep knowledge of natural philosophy, mechanical aptitude, and an interest in the problem of the electric telegraph. Perhaps only Wheatstone shared this combination of qualities in the 1830s. By 1831, he had built a 1.5 mile circuit that could ring an alarm bell, using a more powerful electromagnet than anyone else yet possessed. Had he continued to pursue telegraphy with the kind of tenacity shown by Morse, his might be the name celebrated in American textbooks.

But Henry, a teacher at the Albany Academy and then the College of New Jersey (now Princeton University), built and improved electrical equipment with research, instruction, and scientific demonstration in mind. He showed no interest in turning his pedagogic instrument into a communications system. 

Around 1835, he devised a particularly clever demonstration, using two circuits. Recall that Henry had deduced that electricity had two dimensions – intensity and quantity (what we call voltage/tension, and current, respectively). He created circuits with intensity batteries and magnets to project electromagnetic force over long distances, and circuits with quantity batteries and magnets to generate large electromagnetic forces .

His new apparatus combined the two. The powerful quantity electromagnet held suspended hundreds of pounds of weight. The intensity magnet, at the end of a long circuit, was used to raise a small metal wire: a switch. Connecting the intensity circuit caused the magnet to raise the wire, opening the switch, breaking the quantity circuit. The quantity electromagnet then suddenly dropped its load, with a resounding crash.⁴

This relay – for that is what the intensity magnet and its wire had become in this configuration – was a means to make a point about the transformation of electrical into mechanical energy, and how a small force can trigger a larger one. Gently dipping a wire into acid to close a circuit caused the tiny movement of the little switch, which in turn entailed a catastrophe of falling metal, enough to crush a person foolish enough to stand beneath it. For Henry the relay was a tool for the demonstration of scientific principles. It was, in effect, an electric lever.

Henry was probably the first to connect two circuits in this way – to use electromagnetism from one circuit to control another. The second, so far as we know, were William Cooke and Charles Wheatstone, though with an entirely different purpose in mind.

In March 1836, soon after seeing a telegraphic demonstration in Heidelberg using a galvanic needle for signaling, Cooke was inspired by a music box. Cooke believed that using needles for signaling in a real telegraph would require several needles, and therefore several circuits, to identify a particular letter. Cooke wanted instead to use an electromagnet to activate a mechanism, which could then be as complex as needed to indicate the desired letter.

The machine he had in mind would resemble a music box, with a barrel surrounded by a number of protruding pins. On one side of the barrel was a dial inscribed with the letters of the alphabet. One such device would sit at each end of the telegraph. A wound spring could make each barrel turn, but most of the time a detent would hold them in place. When the telegraph key was depressed, it closed the circuit, activating electromagnets to release both detents, allowing both machines to turn. Once the desired letter was showing on each dial the key was released and the detents dropped, stopping the motion of the barrels. Without knowing it, Cooke had recreated the chronometric model behind Ronalds’ telegraph of two decades before, and the early Chappe telegraph experiments (though those used sound rather than electricity to synchronize the dials).

Cooke realized that a similar mechanism could be used to solve another longstanding problem in telegraphy – how to notify the receiver that a message was incoming. A second circuit with another electromagnet and detent could be used to activate a mechanical alarm bell: close the circuit, pull up the detent, and the alarm rings.

In March of 1837, Cooke began collaborating with Wheatstone on telegraphy, and around that time one or both of the partners conceived of the secondary circuit. Rather than having an independent circuit for the alarm (and the miles of wire that entailed), why not simply use the primary telegraphic circuit to also control the alarm circuit?

By this time Cooke and Wheatstone were back to needle-based designs, and it was straightforward to attach a small piece of wire to the back of the needle so that, when its tip was pulled up by electromagnetism, its tail would close a second circuit. This circuit would set off the alarm. After a decent interval (allowing the receiver to rouse himself from his nap, disconnect the alarm, and prepare pencil and paper), the needle could then be used to signal a message in the usual fashion.⁵

Twice on two continents in the span of two years for two different reasons, someone had realized that an electromagnet could be used as a switch to control another circuit. But there was another way of thinking about the relationship between the two circuits.

Amplifier

By the fall of 1837 Samuel Morse had confidence that his idea for an electric telegraph could be made to work. Using Henry’s intensity battery and magnet, by way of Morse’s New York University colleague, Leonard Gale, he had sent messages as far as a third of a mile. However, to prove to Congress that his telegraph could span the continent, he needed to do a great deal more than that. It was clear that, however powerful the battery, at some point the circuit would become too long to provide a legible signal at the far end. But Morse realized that even with its potency greatly diminished by distance, an electromagnet could open and close another circuit powered by a separate battery, which could then send the signal on. This process could be repeated as many times as necessary to span arbitrary distances. Hence the name relay for these intermediary magnets – like a relay horse, they take the electrical message from their fatigued partner and carry it forward with renewed vigor.⁶

Whether this idea was influenced by Henry’s work is impossible to say, but what was certainly new with Morse was the purpose to which he put the relay. He saw the relay not as a switch but as an amplifier, which could turn a weak signal into a strong one.

Across the Atlantic and around the same time, Edward Davy, a London pharmacist, had the same notion. Davy probably began tinkering with telegraphy sometime in 1835. By early 1837, he was doing regular experiments with a one-mile circuit in Regent’s Park, in the northwest of London.

Soon after Cooke and Wheastone’s March 1837 meeting, Davy caught wind of the competition, and began to think more seriously about building a practical system. He had noticed that the strength of the deflection of a galvanometer’s needle diminished notably with the length of the wire. As he wrote many years later:⁷

It then occurred to me that the smallest motion (to a hair’s breadth) of the needle would suffice to bring in contact two metallic surfaces so as to establish a new circuit, dependent on a local battery; and so on ad infinitum.

Davy called his idea for making a weak signal strong again the “Electrical Renewer.” However he would never bring this nor any of his telegraphic ideas to fruition. He was granted his own telegraph patent in 1838, independent of Cooke and Wheatstone’s. But he sailed to Australia in 1839 to escape a failed marriage, leaving the field in Britain clear for his rivals. Their telegraph company bought up his idle patent several years later.⁸

The Relay in the World

We tend to pay a great deal of attention to systems in the history of technology, while rather neglecting their components. We chronicle the history of the telegraph, the telephone, the electric light, and bathe their creators (or those who we deem retrospectively as such) in the warm rays of our approbation. Yet these systems are made possible only by the combination, recombination, or slight modification of existing elements which have quietly grown in the shade.⁹

The relay is just such an element. From its ancestral forms, it quickly evolved and diversified as telegraph networks began to grow in earnest in the 1840s and 50s. It then found its way into electrical systems of all kinds over the following century. The earliest change was the use of a stiff metal armature, like that in the telegraph sounder, to close the circuit. A spring pulled the armature away from the circuit when the electromagnet was off. This was a much more reliable and durable mechanism than the bits of wire or pins used in Henry, Cooke, and Wheatstone’s experiments. Default-closed models were also devised, as complements to the original default-open design.

Relays were only occasionally used as amplifiers or “renewers” in the early decades of the telegraph, since a single circuit could span 100 miles or more without them. They were very useful, though, for joining a low-current long-distance line to a high-current local circuit that could be used to power other machinery, like the Morse register.¹⁰

Dozens of U.S. patents from the latter half of the nineteenth century describe new forms of relay or new applications for them. The differential relay, which split the coil so that it the electromagnetic effect was canceled in one direction but reinforced in the other, enabled a form of duplex telegraphy: two signals passing in opposite directions over a single wire. Thomas Edison used the polarized (or polar) relay to make his quadruplex, which could send 4 simultaneous signals over a single wire: two in each direction.¹¹ In the polarized relay, the armature itself was a permanent magnet such that it responded to the direction or polarity of the current, rather than its strength. Permanent magnets also allowed for the creation of latches, relays which would stay open or closed, whichever way they were last set.

In addition to their role in new telegraphic equipment, relays also became essential components of railway signaling systems. When electrical power networks began to appear at the end of the century, they found uses in those circuits, too – especially as fault-protection devices.

Yet even these networks, vast and complex as they may seem, did not demand of the relay more than it could give. The telegraph and railroad touched every town, but not every building. They had tens of thousands of endpoints, but not millions. Electrical power systems, for their part, did not care where in particular its current ended up – it bathed a neighborhood-wide circuit in electricity, and each home or business could pick up what it needed from the wire.

The telephone was another story. Because the telephone had to make a point-to-point connection from any arbitrary home or office to any other, it would need control circuits on a scale never before seen. Moreover, the imprint of the human voice wavering across the telephone wire was a rich signal (far richer than Morse code), but a feeble one. So long-distance telephony would also demand new, better amplifiers; and it would turn out that these new amplifiers could be switches, too. It was the telephone networks, more than any other system, that would drive the evolution of the switch. [Next part]

Further Reading

Historical work on the electromechanical relay is scanty. I have not yet found any source wholly dedicated to the topic.

A brief technical account can be found in James B. Calvert, “The Electromagnetic Telegraph“, which is also an excellent source on the technical history of the telegraph in general.

Franklin Leonard Pope’s Modern Practice of the Electric Telegraph (1891) is a cornucopia of information on late-nineteenth century telegraphic equipment in the United States.