We saw last time how the first generation of digital computers were built around the first generation of automatic electrical switch, the electromagnetic relay. But by the time those computers were built, another digital switch was already waiting in the wings. Whereas the relay was an electromechanical device (because it used electricity to control a mechanical switch), this new class of digital switches was electronic – founded on the new science of the electron, a science born around the turn of the twentieth century. This science concretized the carrier of electrical force as not a current, wave, or field, but as a solid particle.

The device that gave birth to an electronic age, rooted in this new physics, became known (at least in the U.S.) as the vacuum tube. Conventionally, two men figure in the story of its creation: the Englishman Ambrose Fleming, and the American Lee de Forest. In fact, of course, its origins are more complex and woven from many threads, which criss-cross Europe and the Atlantic, and stretch back as far as the early Leyden jar experiments of the mid-eighteenth century.

For the purposes of our story, however, it’s convenient, and illuminating (so to speak) to begin with Thomas Edison. Edison made a curious discovery in 1880s as part of his work on a new kind of electric light, a discovery that sets the stage for our story. From there, further development of the vacuum tube was spurred by the demands of two other technological systems: a new form of wireless communication, and the ever-expanding long-distance telephone networks.

Prologue: Edison

Edison is, in the popular imagination, the inventor of the electric light bulb. This gives him both too much credit and too little. Too much credit, again, because Edison was not the only one to devise an incandescent bulb. In additional to a variety of pre-commercial predecessors, Joseph Swan and Charles Stearn in the U.K. and fellow American William Sawyer brought lamps to market around the same time as Edison. All consisted of a sealed, glass bulb containing a resistive filament. When placed in an electrical circuit, the heat generated by its resistance caused the filament to glow. The bulb was evacuated of air to prevent the filament from burning. Electric light was already commonplace in large cities in the form of electric arc lamps, used to illuminate large public spaces. All these inventors were trying to “subdivide the light,” drawing from the flaming arc a spark small enough to enter the home and replace gas lamps with a light source that was safer, cleaner, and brighter.

What Edison — or, more correctly, the industrial lab which Edison headed — did, however, was more than merely to create a light source. He – it – built an entire inter-operable electrical system for home lighting – generators, transmission wires, transformers, and so forth, of which the bulb was only the most obvious and visible component. The presence of Edison’s name in his power companies was not a mere genuflection to the great inventor, as was the case with Bell Telephone. Edison proved himself not only an inventor strictly speaking, but also an able system builder.1 To that end, his lab continued to tinker with the various components of electric lighting even after their early successes.

Early Edison lamp

As part of these researches, some time in 1883, Edison (or perhaps one of his employees) decided to seal a metal plate into the incandescent lamp, along with the filament. The accounts of why he did this do not present a clear and unambiguous picture. But it was probably an attempt to alleviate the problem of lamp blackening: the tendency of the the glass interior of the bulb to accumulate a mysterious dark substance over time. Edison (if it was him) probably hoped that the blackening particles could be drawn off onto the electrified plate. To his surprise, however, he found that when the plate was wired into a circuit with the positive end of the filament, a current flowed that was directly proportional to the intensity of the filament’s glow. When it was connected to the negative end of the filament, nothing happened.

Edison believed this effect, later dubbed the Edison effect, could be used to measure or even regulate the “electro-motive force,” or voltage, in an electrical power system. As was his habit, he filed a patent for this “electrical indicator,” then returned to other, more pressing matters.2

The Wireless

We now skip forward twenty years, to 1904. At this time, In England, a man named John Ambrose Fleming was working on behalf of the Marconi Company to develop a better receiver for radio waves.

It’s important, before we proceed further, to explain what the radio was and was not at this time, both as an instrument and as a practice. In truth, the radio wasn’t even yet radio – it was wireless. (Not until the 1910s did the former term supersede the latter in American English.) Specifically, it was the wireless telegraph – a means of conveying signals in the form of dots-and-dashes from a sender to a recipient. Its primary application was in ship-to-ship and ship-to-shore communication, and as such it was of special interest to the navies of the world.

Some few inventors, notably Reginal Fessenden, were by this time experimenting with the notion of a radio-telephone: point-to-point speech communication over the air via a continuous wave. Not for another fifteen years, though, would broadcasting in the modern sense emerge: the intentional transmission of news, stories, music, and other programming to a wide audience. Until then, the omni-directional nature of radio signals was mainly a problem to be overcome, not a feature to be exploited.

The radio equipment that existed at this time was well-adapted to sending Morse code, and ill-adapted to anything else. Transmitters generated “Hertzian” waves by sending a spark across a gap in a circuit. The signal thus propagated through the ether was accompanied by a dirty burp of static.

Receivers detected this emission via a coherer: a collection of metal filings in a glass tube that cohered into a contiguous mass, thus completing a circuit, when stimulated by radio waves. The glass then had to be tapped to decohere the filings and reset the receiver for the next signal – at first by hand, but it did not take long to come up with automated tapping devices.

Just coming into use in 1905 were crystal detectors, also known as “cat’s whisker” detectors. It turned out that by simply touching a wire to certain crystals such as silicon, iron pyrite, and galena (lead ore), one could pull a radio signal from the air. The resulting radio receivers were cheap, compact and very easy for anyone to try, and they stimulated a widespread amateur wireless movement, primarily among young men. The sudden surge in traffic that resulted raised a problem because of the shared nature of the radio commons. Not only did the innocent chatter of amateur “hams” accidentally interfere with naval communications, but some miscreants went so far as to send out false naval orders and distress signals. It became inevitable that the state would intervene. As Ambrose Fleming himself wrote, (no, we haven’t forgotten about him), the introduction of crystal detectors3

…was followed at once by an outburst of irresponsible radiotelegraphy at the hands of innumerable electrical amateurs and students which required the firm intervention of National and International legislation to keep it within the bounds of reason and safety.

From the peculiar electrical properties of these little crystals would emerge, in due time, yet a third generation of digital switch to follow the relay and the tube, the switch that dominates our world today. But everything in its proper place. We have surveyed the stage, now let us return our attention to the actor who has just stepped into the footlights: Ambrose Fleming, England, 1904.


In 1904 Fleming was a professor of electrical engineering at University College, London, but also a consultant for the Marconi Company. Marconi’s initial interest in recruting him was to get his expert advice on the construction of the power plant at a new shore station, but soon after he took on the problem of building a better detector.

Fleming in 1890

Everyone knew that the coherer was a poor detector in terms of sensitivity, and the magnetic detector that Marconi had devised was not much better. In order to help him find a successor, Fleming set out in the first instance to build a sensitive circuit for detecting Hertzian waves. If not a practical detector itself, such a device would be a useful for further investigations.

To build such a thing, he needed a way to continuously measure the strength of the current generated by the incoming waves, in contrast to the discontinuous coherer (it was either on – with the filings cohered – or not).4 But known devices for measuring current strength – galvanometers  required a direct (unidirectional) current to operate. The alternating current induced by radio waves reversed directions so quickly that it would produce no measurement at all.

Fleming then remembered a handful of curiosities he had sitting in a closet – Edison indicator lamps. In the 1880s, he had been a consultant for Edison Electric Light Company of London, and worked on the lamp blackening problem. During that time he received copies of Edison’s indicator, probably from William Preece, chief electrical engineer of the British Post Office, who had just come back from an electrical exhibition in Philadelphia. (Remember that it was the norm at this time, outside the United States, for post offices to control the telegraph and telephone, and therefore to be centers of electrical engineering expertise).

Later, in the 1890s, Fleming did his own studies on the Edison effect using the lamps acquired from Preece. He showed that the effect consisted in a unidirectional current flow: negative electrical potential could flow from the hot filament to the cold electrode, but not vice versa. But not until 1904, when presented with the problem of radio wave detection, had he realized that this fact could have any practical use. The Edison indicator would allow only the forward surges of alternating current to crash over the gap between filament and plate, creating a steady one-way flow on the far side.

Fleming grabbed one of the bulbs, attached it in series with a mirror galvanometer, and switched on a spark transmitter – et voila, the mirror turned and the light beam moved on the scale. It worked. He could precisely measure the incoming radio signal.

Prototype Fleming valves. The anode is set in the middle of the loop of filament (the hot cathode).

Fleming called his invention a “valve”, since it acted as a one-way gate for electricity. In more general electrical engineering terms, it was a rectifier – a means of transforming an alternating current into a direct one (rectifying it, i.e. straightening it out). Finally, it was called a diode, since it contained two electrodes: the hot cathode (the filament) which emitted electricity and the cold anode (the plate) which received it. Fleming made several refinements to the design, but in its essence it was no different from the indicator lamp built by Edison.5 Its transformation into a new kind of thing was, as we have seen before, the result of a change in mental state. A change in the world of ideas inside Fleming’s head, not in the world of stuff, outside of it.

By itself, Fleming’s valve was a useful object. It was the best field test device yet found for measuring radio signals, and a reasonable detector in its own right. But it did not shake the world. The explosive growth of electronics came only after an American, Lee de Forest, added a third electrode, making the valve into a relay.


Lee de Forest had an unusual upbringing for a Yale man. His father, the Reverend Henry de Forest, was a Civil War veteran from New York, a Congregationalist pastor, and a fervent believer in his mission as a man of God to spread the light of knowledge and justice. As such, he dutifully took up the call when invited to the presidency of Talladega College in Alabama. Talladega had been founded by the New York-based American Missionary Association after the Civil War, with a mission to educate and edify the local black population.6 There young Lee found himself caught between two stones: picked on by the local black boys for being a homely and cowardly weakling; shunned by the local white boys for being a Yankee meddler.

Nonetheless, the younger de Forest developed a firm confidence in himself. He found that he had more than a little skill as a mechanic and a tinkerer – his scale-model locomotive became a local wonder. Already as a teenager in Talladega he knew that he would make his way in the world through his inventions. Later, as a young man about town in New Haven, the religious convictions of this pastor’s son fell away — worn away by exposure to Darwinism, then shorn off in one fatal blow by the unexpected death of his father. But the core, unshakable sense of his own destiny remained – de Forest believed himself to be a man of genius, and aimed to make himself another Nikola Tesla – a wealthy, famous, and mystical magician of the electric age. His classmates at Yale, on the other hand, believed him to be a conceited windbag. He may well be the least likable human being to feature in our story thus far.7

De Forest ca. 1900

By the time he completed his Ph.D. at Yale in 1899, de Forest had set his heart on the emerging art of the wireless as his path to fame and fortune. Over the coming decades he pursued that path with great determination and resolve, but rather less scruple. It began with de Forest and a partner, Ed Smythe, working together in Chicago. Smythe kept de Forest in room and board with regular five-dollar payments, and together they developed their own radio detector, consisting of two metal leads connected by a paste that de Forest called “goo.” But De Forest was impatient for the rewards of his genius. He ditched Smythe and teamed up with a shady New York financier named Abraham White (nee, ironically, Schwartz), to form the De Forest Wireless Telegraph Company.

The actual operations of the company were incidental to both protagonists: White concentrated on using the public’s ignorance to line his pockets. He flogged the stock of the new company relentlessly, and brought in millions from wide-eyed investors afraid of missing out on the radio boom. Meanwhile De Forest, amply funded by the “suckers”8, focused on proving his genius through the development of a new American system of wireless (in contrast to the European systems developed by Marconi and others).

Unfortunately for that American system, however, de Forest’s “goo” detector didn’t actually work very well. He solved that problem in the short term by borrowing the design of Reginald Fessenden’s (patented) “liquid barretter” detector – two platinum wires immersed in a sulfuric acid bath. Fessenden soon filed suit for patent infringement – a suit he would clearly win. De Forest could not rest until he had devised a new detector that was unequivocally his own. In the autumn of 1906, he announced that he had done so. Before two separate meetings of the American Institute of Electrical Engineers, de Forest described his new wireless detector, which he dubbed the “Audion.” Its actual provenance is, however, rather dubious.

For some time, much de Forest’s effort to build a new detector had centered on passing a current through a Bunsen burner flame, which he believed could act as an asymmetric conductor. As far as can be told, this idea had no merit.9 Then, at some point in 1905, he learned about Fleming’s valve. De Forest convinced himself that the valve and his Bunsen burner devices were in principle the same: simply replace the flame with a hot filament, encase it in glass bulb to contain the gas, and you had the valve. He then developed a series of patents that recapitulated the ancestry of the Fleming valve via his gaseous flame detectors. In this way he evidently thought to give himself priority of invention over Fleming’s U.S. patent, since his Bunsen burner work predated it (going all the way back to 1900).

Whether this was self-delusion or simple fraud is impossible to tell, but it all culminated in de Forest’s patent of August 1906, for: “an evacuated vessel of glass… having two separated electrodes, between which intervenes the gaseous medium which when sufficiently heated or otherwise made highly conductive forms the sensitive element…” The equipment and behavior described is Fleming’s; the explanation for its function, de Forest’s. De Forest would lose this patent suit, too, though it would take ten years.10

The impatient reader may began to wonder: why are we spending so much time on this man, whose self-declared genius seemed to consist largely in passing off the ideas of others as his own? The reason is the transformation that the Audion underwent in the last few months of 1906.

De Forest, by this point, was out of a job. White and his partners had avoided responsibility for the Fessenden suit by creating a new company, United Wireless, and leasing the assets of American De Forest to that new company for $1. De Forest was cast off with $1000 in severance pay and a few apparently useless patents, including those for his Audion. Having accustomed himself to a fairly lavish lifestyle, he now found himself in serious financial difficulty, and was desperate to turn the Audion into a big success.

To understand what happened next, it’s important to realize that De Forest believed that, in contradistinction to Fleming’s rectifier, he had invented a relay. He had set up his Audion by hooking a battery to the cold plate of the valve, and believed that the signal in the antenna circuit (connected to the hot filament) was modulating the more powerful current in that battery circuit. In fact he was quite wrong: there were not two circuits at all, the battery simply shifted the signal from the antenna, it did not amplify it.

However this false belief proved critical, because it led de Forest to start experimenting with a 3rd electrode in the bulb, to more completely separate the two circuits of his “relay”. At first he added this second cold electrode side-by-side with the first, but then, perhaps inspired by the sort of control mechanisms used by physicists to channel the rays in cathode-ray tubes, he moved it between the filament and the original plate. Concerned that this would block the flow of electricity, he then changed the third electrode’s shape from a plate to a wiggly piece of wire that resembled a gridiron – he called this the grid.

1908 Audion triode. The (broken) filament on the left is the cathode, the wiggly bit of wire is the grid, and the rounded sheet of medal the anode. Note that it still has a screws for a socket like an ordinary light bulb.

Here we have a true relay. A weak current (such as that from a radio antenna) applied to the grid could control a much more powerful current between filament and plate, by repelling the charged particles trying to pass from one to the other. This would allow it to act as a much more powerful detector than the valve, since it could not just rectify but also amplify the radio signal. And like the valve (and unlike the coherer) it could produce a continuous signal, allowing not only radio telegraphy but also radio telephony (and later the broadcasting of voice and music).

In practice, however, it didn’t actually work very well. De Forest’s Audions were finicky, prone to burn out quickly, lacking in uniformity of manufacture, and generally ineffective as amplifiers. It required bespoke electrical tuning to find the right parameters to get a given Audion to function at all.

Nonetheless, de Forest believed in his invention. To promote it, he formed a new venture, the De Forest Radio Telephone Company, but managed only a trickle of sales. His biggest prize was a sale of equipment to the Navy for intra-fleet telephony during the cruise of the Great White Fleet – but the fleet commander, unable to take the time to get de Forest’s receivers and transmitters working and to train his crews in their use, had them packed up and put in storage. Moreover, De Forest’s new company, presided over by a disciple of Abraham White, was no more scrupulous than his last; and so to add to his troubles he soon found himself under indictment for fraud.11

So, for five years the Audion went nowhere. Once again the telephone would play a critical role in the development of a digital switch, this time to rescue a promising but unproven technology from the brink of obscurity.

The Telephone, Again

The long-distance network was the central nervous sytem of AT&T. Tying together its many local operating companies, it provided a crucial competitive advantage after the expiration of the core Bell patents. By joining the AT&T network, a new customer could, in theory, reach any of his or her fellow subscribers, hundreds or thousands of miles away – though in practice long-distance calls were rare. The network was also the material basis for AT&T’s all-encompassing ideology of “One Policy, One System, Universal Service.”

But as the second decade of the twentieth century began, that network was reaching its physical limits. As telephone wires stretched longer and longer, the signal that passed through them became weaker and noisier, until speech became entirely incomprehensible. Because of this, there were in fact two AT&T networks in the United States, divided by the continental ridge.

For the eastern network, New York City was the stake in the ground, mechanical repeaters and loading coils the leash that defined how far the human voice could roam. But these technologies could only do so much. Loading coils altered the electrical properties of the telephone circuit in order to reduce attenuation at voice frequencies – but they could only reduce it, not eliminate it. Mechanical repeaters (nothing more than a telephone speaker coupled to an amplifying microphone) added noise with each repetition. A 1911 New York to Denver line tensed this leash to its absolute limit. To span the entire continent was beyond consideration. Yet in 1909, John J. Carty, AT&T’s Chief Engineer, had publicly promised to do exactly that. And he promised to do it within five years: in time for the Panama–Pacific International Exposition set to take place in San Francisco in 1914.

The first to make such a venture conceivable, with a new electronic telephone amplifier, was not an American, but the scion of a wealthy Viennese family with a scientific bent. As a young man, Robert von Lieben bought a telephone manufacturing company with aid of his parent’s wealth, and set out to develop an amplifier for telephone conversations. By 1906 he had built a relay based on cathode-ray tubes, a common device by that time in physics experiments (and later the basis for the dominant video screen technology of the twentieth century). The weak incoming signal controlled an electromagnet that bent the cathode ray beam, modulating the stronger current in the main circuit.

By 1910 von Lieben and his colleagues, Eugen Reisz and Sigmund Strauss, learned about the de Forest Audion, and replaced the magnet with a grid inside the tube to control the flow of cathode rays – this was a much more effective design, and surpassed anything developed in the U.S. to date. The German telephone network soon adopted the von Lieben amplifier. In 1914, it enabled a nervous call from the East Prussian army commander to German staff headquarters, 1,000 kilometers away in Koblenz. This in turn led the German Chief of Staff to dispatch Generals Hindenberg and Ludendorff to the East, to enduring fame, and with weighty consequences. The same amplifiers later connected German headquarters with field armies as far south and east as Macedonia and Romania.12

Replica of the mature form of von Lieben’s cathode-ray relay. The cathode is at bottom, the anode is the coil at top, and the grid is the circular metal foil in the middle.

But the barriers of language, geography, and war prevented this design from reaching the U.S. before it was overtaken by the developments that follow.

De Forest, meanwhile, had left his failing Radio Telephone Company, in 1911, and fled to California. There he took a position at the Federal Telegraph Company in Palo Alto, founded by Stanford graduate Cyril Elwell.13 Nominally, de Forest was assigned to work on an amplifier in order to generate a louder output signal from Federal’s radio receiver (called a “tikker”). In practice, he, Herbert van Ettan (a skilled telephone engineer), and Charles Logwood (designer of the tikker) focused instead on building a telephone amplifier to secure for themselves (not Federal) a rumored $1 million prize from AT&T.

To this end, De Forest brought the Audion out of cold storage, and by the summer of 1912 he and his colleagues had a device they were ready to show to the phone company. It consisted of several Audions in sequence, to create multiple amplification stages, plus several other auxiliary components. It worked, after a fashion – it could amplify an audio signal well enough to hear the drop of a handkerchief, or the tick of a pocket watch. But only at currents and voltages too low to be at all useful for telephony. When pressed harder, a blue glow appeared inside the Audions and the signal turned to noise. The telephone men were sufficiently intrigued, however, to bring the device in to see what their engineers would make of it. It so happens that one of them, a young physicist named Harold Arnold, knew exactly how the amplifier from Federal Telegraph could be set to rights.

The time has come to discuss how it is that the valve and Audion actually worked. The crucial knowledge for explaining their function came from the Cavendish Lab in Cambridge – the intellectual center of the new electron physics. There, in 1899, J.J. Thomson had shown convincingly via experiments on cathode ray tubes that a particle with mass, later known as the electron, carried the current from cathode to anode. Over the next few years, Owen Richardson, a colleague of Thomson’s, developed this basic premise into a mathematical theory of thermionic emission.14

Ambrose Fleming, an academic engineer who worked just a short train ride away from Cambridge, was familiar with this body of work. Therefore it was clear to him that his valve functioned by thermionic emission of electrons from the heated filament, which then crossed the vacuum gap to the cold anode. But the vacuum in the indicator lamp was far from complete: an ordinary light bulb did not require such a thing; it was sufficient to remove enough oxygen to prevent the combustion of the filament. Fleming therefore realized to make the valve work as well as possible, it should be evacuated beyond normal levels, in order to prevent residual gas from interfering with the passage of electrons.

De Forest, on the other hand, did not realize this. Because he came to the valve and Audion by way of his Bunsen burner experiments, he believed exactly the opposite – that hot, ionized gas was the working fluid of the device, and that too-perfect evacuation would destroy its function. This was the reason for the Audion’s inconsistent and disappointing performance as a radio receiver, and for the fatal blue glow.15

AT&T’s Arnold was perfectly placed to correct de Forest’s error. A physicist who had studied under Robert Millikan at the University of Chicago, he was recruited specifically to apply his knowledge of the new electron physics to the coast-to-coast telephony problem. He knew that the Audion tube would function best at a near-perfect vacuum, knew that the newest pump designs could achieve that, knew that a new kind of oxide-coated filament along with a larger plate and grid would also help increase the electron flow. In short, he transformed the Audion into the vacuum tube, wonder-worker of the electronic age.16

AT&T now had the powerful amplifier it needed to build their transcontinental line, it lacked only the legal rights to use it. Their representatives remained diffident in the conversations with de Forest, but opened separate negotiations through a third-party lawyer, who managed to acquire the rights to the Audion as a telephone amplifier for $50,000 (roughly $1.25 million in 2017 dollars). The New York-San Francisco line opened right on time17, though as a triumph of technical virtuosity and corporate publicity rather than of human communication; rates were so exorbitant that hardly anyone would use it.

An Electronic Age

The true vacuum tube formed the root for a whole new tree of electronic components. As with the relay, so too did the vacuum tube diversify and diversify again, as engineers found ways to tweak the design just so to suit the needs of a particular problem. The growth of -odes did not end with diodes and triodes. It continued with the tetrode, which added an additional grid to sustain amplification as the number of elements in the circuit grew. Pentodes, heptodes, even octodes, followed. There were thyratrons filled with mercury vapor, which glowed an eerie blue; miniaturized tubes as small as a pinky finger, or even (eventually) an acorn; indirectly heated tubes to prevent the hum of an alternating current power source from disturbing the signal. The Saga of the Vacuum Tube, a book describing the growth of the tube industry to 1930, references on the order of 1,000 different models by name in its index, though many were bootleg knock-offs from fly-by-night independent brands: Alltron, Perfectron, Supertron, Voltron, etc.18

Tubes of all shapes and sizes: diodes, triodes, pentodes, etc. Contrast the now-standard pins with the connections on the early Audion.

Even more important than diversity of forms was the diversity of applications enabled by the vacuum tube. Regenerative circuits transformed the triode into a transmitter – a transmitter that generated smooth, continuous sine waves, with no noisy spark, and thus able to transmit sound perfectly. With coherer and spark in 1901, Marconi was barely able to heave the smallest fragment of Morse code across the narrowest part of the Atlantic. In 1915, with the vacuum tube as transmitter and receiver, AT&T was able to project the human voice from Arlington, Virginia to Honolulu, over twice the distance. By the early 1920s they were combining long-distance telephony with high-quality sound broadcasting to create the first radio networks. By such means would entire nations soon bend their ear to a single voice, be it that of a Roosevelt, or a Hitler.

Moreover, the ability to create transmitters tuned to a precise, stable pitch also allowed telecommunications engineers to finally realize the dream of frequency multiplexing, which had lured Alexander Graham Bell, Edison, and others forty years before. By 1923, AT&T had a ten-channel voice line from New York to Pittsburgh. The ability to carry many voices on a single copper wire would greatly reduce the cost of long-distance calling, which had always been too expensive to use for all but the wealthiest of individuals and businesses.19 Once they saw what vacuum tubes were capable of, AT&T sent their lawyers back to buy more rights from de Forest, and then still more – to secure the rights to the application of the Audion in all imaginable fields, they paid him a total of $390,000, roughly $7.5 million today.20

Given their evident versatility, why did vacuum tubes not dominate the first generation of computers in the same way they dominated radio and other telecommunications equipment? It was obvious that a triode could act as a digital switch in the same fashion as a relay[^22]; so obvious that de Forest was convinced that he had built a relay before he actually managed to do so. And the triode was far more responsive than the traditional electromechanical relay, because there was no need to physically move an armature. Whereas a relay typically took several milliseconds to switch on or off, the effect on the flow from cathode to anode from a change in the electrical potential on the grid was nearly instantaneous.

However tubes had a clear disadvantage vis-a-vis relays: their tendency, like their ancestor the incandescent bulb, to burn out. The lifetime of de Forest’s original Audions was so poor – a mere 100 hours or so – that he had a back-up filament placed in the bulb to be wired up after the first inevitably failed. This was exceptionally bad, but even later, high-quality tubes could not be expected to last for more than a few thousand hours of normal use. In a computer with thousands of tubes, whose computations might take hours to complete, this was a serious problem.

Relays, by contrast, were, in the words of George Stibitz, “awesomely reliable.” So much so, that he claimed that21

if a set of U-type relays had started in the year 1 A.D. to turn a contact on and off once per second, they still would be clicking away reliably. Their first contact failure, or misfire, would not be due until more than a thousand years from now, around the year 3000.

Moreover, no experience existed with large electronic circuits comparable to that of telephone engineers with large electromechanical circuits. Radio receivers and other electronic equipment might contain five or ten tubes, but not hundreds or thousands. No one knew whether a, computer with, say, 5,000 tubes, could be made to work. By going with relays over tubes, computer designers made the safe, conservative choice.

In our next installment we will see how, and why, these doubts were overcome.


Hugh G.J. Aitken, The Continuous Wave (1985)

J. A. Fleming, The Thermionic Valve (1919)

Anton A. Huurdeman, The Worldwide History of Telecommunications (2003)

Paul Israel, Edison: A Life of Invention (1998)

Tom Lewis, The Empire of the Air (1991)

Gerald F. J. Tyne, Saga of the Vacuum Tube (1977)



  1.  Paul Israel, Edison: A Life of Invention (1998), 167-190. 
  2. Some accounts report that it was in fact another member of Edison’s lab who found “his” effect, which is far from improbable, given the scope of the operations for which Edison was by then responsible. Other accounts confuse the Edison effect with lamp blackening itself, and the absence of blackening that sometimes appeared in the plane of the filament. This line was indeed related to the same underlying phenomenon of thermionic emission as the Edison effect. But it required no special perspicacity to notice it, it was obvious from just looking at burnt out lamps. Explaining it was another matter. 
  3.  J. A. Fleming, The Thermionic Valve (1919), 4. 
  4. In this sense the coherer was in fact a digital switch, but a very poor one. 
  5.  Fleming was nonetheless able to secure an American patent easily; Edison’s patent had since expired, and in any case did not foresee applications to radio. 
  6. A number of other black colleges of the South, such as Howard, Fisk, and Hampton Universities, were also founded by the AMA. For the most part, southern whites had no interest in educating southern blacks. 
  7. Tom Lewis, The Empire of the Air (1991), 29. For instance, at the outbreak of the Spanish-American war, he was all patriotic bluster, but only enlisted after waiting long enough to be certain the war would be short and easy for the United States – “I will get the benefits–and the glory (!) of the campaign without the danger. I do not risk my destiny,” he wrote. Lewis, 33. 
  8.  Hugh G.J. Aitken, The Continuous Wave (1985), 186. 
  9. The idea originated an experiment done in 1900 when working with Smythe, in which a sparking transmitter seemed to affect the brightness of a burner flame. It turned out the effect was due to the sound of the sparkbut the idea of some kind of connection between hot gases and wireless reception stuck with De Forest. Hugh G.J. Aitken, The Continuous Wave (1985), 196-199. 
  10. De Forest pulled the same trick with the regenerative circuit, which allowed a triode to generate oscillating waves, and become a radio transmitter. Exactly as with the Fleming valve, de Forest convinced himself after hearing of the invention that he had already invented it. Unfortunately, in this case the Supreme Court eventually found in his favor, to the shock of most of the engineering community. Lewis, Empire of the Air, Chapter 8. 
  11. De Forest escaped with a stern warning from the judge, but three of his co-defendants were found guilty. 
  12. Gerald F. J. Tyne, Saga of the Vacuum Tube (1977), 239-240; On the fateful telephone call see Winston S. Churchill, The World Crisisvol. 5 (2015 [1931]), 122-128. 
  13. Thus what follows is (to my knowledge) the first event significant to the history of computing to take place in the future Silicon Valley, or indeed in the Western U.S. generally. 
  14. Richardson would later receive the Nobel Prize for this work. 
  15. Hugh Aitken has made the case that De Forest was not wrong, and in fact his Audion worked in a different mode from later “hard vacuum” tubes, a mode based on the conductivity of ionized gas: Hugh G.J. Aitken, The Continuous Wave (1985), 221. I am neither a physicist nor an electrical engineer, but I find this hard to credit for two reasons: 1) There’s no reason I know of that ionized gas would have one-way conductivity – which means there could be no rectification or amplification through such a medium. 2) It seems clear from the 1912 telephone amplifier work at Federal that the Audion began to break down as an amplifier exactly when the gas began to ionize (the blue glow), i.e the ionized gas mode was dysfunctional. 
  16.  At around the same time, in 1913, a General Electric research physicist named Irving Langmuir also came into contact with the Audion, and reached the same conclusions about how to improve it as Arnold. Fritz Lowenstein, a former co-worker of de Forest’s, and former assistant to Tesla, had been doing his own experiments with building a telephone amplifier from the Audion. His prototypes came to Langmuir’s attention via another GE engineer, Ernst Alexanderson. GE thus developed its own high quality vacuum tube technologies in the lab, but was constrained from employing them by lack of patent rights. 
  17. The Panama–Pacific International Exposition, on the other hand, did not – it was delayed to 1915. 
  18.  Gerald F. J. Tyne, Saga of the Vacuum Tube (1977), 331-366, 471-480. 
  19.  John Bray, Innovation and the Communications Revolution: From the Victorian Pioneers to Broadband Internet (2002), 58. 
  20. By the end of World War I, the importance of radio as a military and commercial technology was evident, and virtually all of the important patents in the field of radio fell under the control of such corporate giants, closing the era of independent invention. In the U.S. that meant primarily AT&T, General Electric, and the American Marconi Company. The last was transformed into the Radio Corporation of America (RCA) shortly after World War I, under pressure from the U.S. Navy. American Marconi was subsidiary to its British parent company, and the Navy wanted to ensure that American radio technology, which had become central to naval operations, was under wholly American control. For all the gory details, see Aitken, The Continuous Wave
  21. Quoted in David Ritchie, The Computer Pioneers, 44. 
  22. It’s actually possible to build limited forms of logical circuits strictly from diodes (i.e. Fleming valves), see https://en.wikipedia.org/wiki/Diode_logic on the limitations.

2 thoughts on “The Electronic Age

  1. Thanks for writing this blog – I read it cover-to-cover now! I am impressed by the level of detail and how you weave together the development of technologies, human stories, and over-arching trends in politics and society. Your blog a high-quality book!

    As a physicist and engineer I am very interested in the history of science and technology. What I find most fascinating is how difficult it sometimes is to really understand dated publications on things that ‘we know’ today, because you need to translate the language and concepts used in earlier times to their modern counterparts first before you can try to understand the actual science.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s