Other articles in the series:
- History of the relay
- History of electronic computers
- History of the transistor
- Internet history
Π we have seen how the first generation of digital computers was built around the first generation of automatic electrical switches - electromagnetic relays. But by the time these computers were built, there was another digital switch waiting to be released behind the scenes. The relay was an electromagnetic device (using electricity to operate a mechanical switch), and a new class of digital switches was electronic - based on new knowledge about the electron that appeared at the beginning of the XNUMXth century. This science indicated that the carrier of electrical force was not a current, not a wave, not a field - but a solid particle.
The device that spawned an era of electronics based on this new physics became known as the "vacuum tube". Two people are involved in the history of its creation: an Englishman and American . In fact, the origin of electronics is more complex, weaving from many threads that cross Europe and the Atlantic, and stretch back into the past, right up to the early experiments with the Leiden jars in the middle of the XNUMXth century.
But within the framework of our presentation, it will be convenient to cover (pun intended!) this story, starting with Thomas Edison. In the 1880s, Edison made an interesting discovery while working on electric lighting, a discovery that sets the stage for our story. From here came the further development of vacuum tubes, which was required for two technological systems: a new form of wireless messaging and an ever-expanding telephone network.
Prologue: Edison
Edison is generally considered the inventor of the electric light bulb. This simultaneously does him too much and too little honor. Too many, since Edison was not the only one who invented the luminous lamp. In addition to the crowd of inventors who preceded him, whose creations did not reach commercial use, one can mention Joseph Swann and Charles Stern from Britain and the American William Sawyer, who brought light bulbs to the market at the same time as Edison. [The honor of the invention also belongs to the Russian inventor . Lodygin was the first who guessed to pump air out of a glass lamp bulb, and then he suggested making an incandescent filament not from coal or charred fibers, but from refractory tungsten / approx. transl.]. All lamps consisted of a sealed glass bulb with a resistive filament inside. When the lamp was turned on in the circuit, the heat generated by the filament's resistance to current caused it to glow. Air was evacuated from the flask so that the filament would not catch fire. Electric light was already known in large cities in the form used to illuminate large public spaces. All these inventors were looking for a way to reduce the amount of light by taking a bright particle from a burning arc, small enough to be used in homes to replace gas lamps, and to make the light source safer, cleaner and brighter.
And what Edison really didβor rather, what his industrial lab createdβwas more than just creating a light source. They built a whole electrical system for lighting houses - generators, current wires, transformers, and so on. Of all this, the light bulb was only the most obvious and visible component. The presence of Edison's name in his electric power companies was more than a mere genuflection to the great inventor, as was the case with Bell Telephone. Edison showed himself not only as an inventor, but also as a systems architect. His laboratory continued to work on improving the various components of electric lighting even after their early success.
A copy of early Edison lamps
In the course of research somewhere in 1883, Edison (and possibly one of his employees) decided to enclose a metal plate inside a luminous lamp along with a thread. The reasons for this act are unclear. Perhaps this was an attempt to eliminate the darkening of the lamp - the inside of the glass of the bulb accumulated a mysterious dark substance over time. The engineer apparently hoped that these black particles would be attracted to the energized plate. To his surprise, he found that when the plate was included in the circuit along with the positive end of the filament, the amount of current flowing through the filament was directly proportional to the intensity of the glow of the filament. When the plate was connected to the negative end of the thread, nothing of the kind was observed.
Edison decided that this effect, later called the Edison effect or , can be used to measure or even control the "electromotive force", or voltage, in an electrical system. Out of habit, he applied for a patent for this "electrical indicator" and then returned to more important tasks.
Without wires
Fast forward 20 years into the future, to 1904. During this time in England, John Ambrose Fleming was working for the Marconi Company to improve a radio receiver.
It is important to understand what radio was and was not at the time, both in terms of instrument and practice. The radio was not even called βradioβ at that time, it was called βwirelessβ, wireless. The term "radio" began to prevail only in the 1910s. Specifically, he meant wireless telegraph - a system for transmitting signals in the form of dots and dashes from the sender to the recipient. Its main application was communication between ships and port services, and in this sense, maritime departments around the world were interested in it.
Some inventors of that time, in particular, , experimented with the idea of ββa radiotelephone - the transmission of voice messages over the air in the form of a continuous wave. But broadcasting in the modern sense did not emerge until 15 years later: the transmission of news, stories, music and other programs for reception by a wide audience. Prior to that, the omnidirectional nature of radio signals was seen as a problem to be solved rather than a feature to be exploited.
The radio equipment that existed at that time was well adapted to work with Morse code and badly for everything else. The transmitters created Hertzian waves by sending a spark through a gap in the circuit. Therefore, the signal was accompanied by a crackle of static.
The receivers recognized this signal through a coherer: metal filings in a glass tube, knocked down under the influence of radio waves into a continuous mass, and thus completing the circuit. Then the glass had to be tapped to disintegrate the sawdust and the receiver was ready for the next signal - at first this was done manually, but soon automatic devices appeared for this.
In 1905, only began to appear , also known as "cat's whisker". It turned out that simply by touching a certain crystal with a wire, for example, silicon, iron pyrite or , it was possible to snatch a radio signal from the air. The resulting receivers were cheap, compact and accessible to everyone. They stimulated the development of amateur radio, especially among young people. The sudden burst of air traffic that arose as a result of this led to problems due to the fact that the radio air was shared between all users. The innocent conversations of amateurs could accidentally intersect with the conversations of the Morfleet, and some hooligans even managed to give false orders and send signals for help. The state inevitably had to intervene. As Ambrose Fleming himself wrote, the advent of crystal detectors
immediately led to an upsurge of irresponsible radiotelegraphy due to the antics of countless amateur electricians and students, which required the strong intervention of national and international authorities to keep what was happening within reason and safety.
Out of the unusual electrical properties of these crystals, a third generation of digital switches will emerge in due course, following relays and lampsβswitches that dominate our world. But everything has its time. We have described the scene, now let's return all attention to the actor who has just appeared in the spotlight: Ambrose Fleming, England, 1904.
Valve
In 1904 Fleming was professor of electrical engineering at University College London, and a consultant for the Marconi Company. The company initially hired him to provide an expert opinion on the construction of a power plant, but then he took on the task of improving the receiver.
Fleming in 1890
Everyone knew that the coherer was a poor receiver in terms of sensitivity, and the magnetic detector developed at Macroni was not particularly better. To find a replacement, Fleming first decided to build a sensitive circuit to detect Hertzian waves. Such a device, even without becoming a detector in itself, will be useful in future research.
To do this, he had to come up with a way to constantly measure the strength of the current generated by the incoming waves, instead of using a discrete coherer (it only showed states on - where the sawdust stuck together, or off). But the well-known devices for measuring current strength - galvanometers - required direct, that is, unidirectional current for operation. The alternating current, excited by radio waves, changed direction so quickly that no measurement could be made.
Fleming remembered that he had several interesting things gathering dust in his closet - Edison indicator lamps. In the 1880s he was a consultant for the Edison Electric Lighting Company in London, and worked on the problem of blackening lamps. At that time, he received several copies of the indicator, possibly from William Preece, chief electrical engineer of the British Postal Service, who had just returned from an electrical exhibition in Philadelphia at the time. At that time, outside the United States, for postal services, control over the telegraph and telephone was commonplace, so they were centers of electrical expertise.
Later, in the 1890s, Fleming himself studied the Edison effect using lamps obtained from Preece. He showed that the effect was that the current flowed in one direction: a negative electrical potential could flow from a hot filament to a cold electrode, but not vice versa. But only in 1904, when he faced the task of detecting radio waves, he realized that this fact could be used in practice. The Edison indicator will only allow unidirectional AC pulses to bridge the gap between the filament and the plate, giving a constant and unidirectional flow.
Fleming took one lamp, connected it in series with a galvanometer, and turned on the spark transmitter. Voila - the mirror turned, and the beam of light moved on the scale. It worked. He could accurately measure the incoming radio signal.
Fleming valve prototypes. The anode is in the middle of the filament loop (hot cathode)
Fleming called his invention a "valve" because it only allowed electricity to flow in one direction. In more general electrical language, it was a rectifier - a way to convert alternating current to direct current. Then it was called a diode, because it had two electrodes - a hot cathode (filament) that emitted electricity, and a cold anode (plate) that received it. Fleming made several improvements to the design, but in essence the device was no different from the indicator lamp made by Edison. Its transition to a new quality occurred as a result of a change in the way of thinking - we have already seen such a phenomenon many times. The change took place in the world of ideas in Fleming's head, not in the world of things outside of it.
The Fleming valve itself was helpful. It was the best field device for measuring radio signals, and a good detector in itself. But he did not shock the world. The explosive growth of electronics only began after Lee de Forest added a third electrode and turned the valve into a relay.
Listening
Lee de Forest had an unusual upbringing for a Yale student. His father, the Reverend Henry de Forest, was a New York Civil War veteran pastor , and firmly believed that, as a preacher, he should spread the divine light of knowledge and justice. Obeying the call of duty, he accepted an invitation to become president of Talladega College in Alabama. The college was founded after the Civil War by the American Missionary Association based in New York. It was intended to educate and instruct the local blacks. There, Lee felt himself between a rock and a hard place - local blacks humiliated him for his naivety and cowardice, and local whites for being .
And yet, as a young man, de Forest developed a strong self-confidence. He discovered a penchant for mechanics and invention - his scale model of a locomotive became a local marvel. As a teenager, while studying at Talladega, he decided to devote his life to inventions. Then, as a young man and living in the city of New Haven, the pastor's son threw off his last religious convictions. They gradually left because of their acquaintance with Darwinism, and then they were blown away by the wind after the untimely death of his father. But the feeling that he had a destiny did not leave de Forest - he considered himself a genius and strove to become the second Nikola Tesla, a rich, famous and mysterious magician of the era of electricity. His classmates at Yale University considered him a self-satisfied windbag. He may well be called the least popular person we have ever met in our history.
de Forest, c.1900
After graduating from Yale in 1899, de Forest chose mastering the burgeoning art of wireless signaling as his path to fortune and fame. In the decades that followed, he stormed this path with great determination and confidence, and without any hesitation. It all started with the joint work of de Forest and his partner Ed Smythe in Chicago. Smythe kept their venture afloat with regular payments, and together they developed their own radio wave detector, consisting of two metal plates held together with an adhesive that de Forest called "paste" [goo]. But de Forest could not wait long for awards for his genius. He got rid of Smythe and teamed up with a dubious New York financier named Abraham White.ironically changed his name from that given to him at birth, Schwartz, to hide his dark deeds. White / White - (English) white, Schwartz / Schwarz - (German) black / approx. transl.], opening the De Forest Wireless Telegraph Company.
The very activity of the company was secondary for both of our heroes. White used people's ignorance to line his pockets. He was swindling millions from investors struggling to keep up with the expected radio boom. And de Forest, thanks to the plentiful flow of funds from these "suckers", concentrated on proving his genius through the development of a new American wireless information transmission system (in contrast to the European one developed by Marconi and others).
Unfortunately for the American system, the de Forest detector did not perform particularly well. For a time, he solved this problem by borrowing Reginald Fessenden's patented design for a detector called a "liquid barretter"βtwo platinum wires immersed in a bath of sulfuric acid. Fessenden sued for patent infringement, a lawsuit that he obviously would have won. De Forest could not rest until he came up with a new detector that belonged only to him. In the autumn of 1906, he announced the creation of such a detector. At two separate meetings at the American Institute of Electrical Engineering, de Forest described his new wireless detector, which he called the "Audion." But its real origin is in doubt.
For a while, de Forest's attempts to build a new detector revolved around passing current through a flame. , which, in his opinion, could be an asymmetric conductor. The idea, apparently, was not crowned with success. At some point in 1905, he learned about the Fleming valve. De Forest got it into his head that this valve and its torch-based device were basically the sameβif you replace the hot filament with a flame, and cover it with a glass bulb to restrict the gas, you get the same valve. He developed a series of patents that retraced the history of pre-Fleming valve inventions using gas flame detectors. He obviously wanted to take precedence over the invention by circumventing Fleming's patent, since the work on the Bunsen burner preceded Fleming's work (they had been going on since 1900).
It is impossible to say whether this was self-deception or fraud, but the result was de Forest's patent of August 1906 for "an empty glass vessel containing two separate electrodes, between which there is a gaseous medium, which, when sufficiently heated, becomes a conductor and forms a sensitive element." The equipment and operation of the device are due to Fleming, and the explanation of its operation is due to de Forest. De Forest ended up losing the patent dispute, although it took ten years.
The impatient reader may already be wondering why we spend so much time on this man, whose self-proclaimed genius was to pass off other people's ideas as his own? The reason lies in the transformation that the Audion underwent during the last few months of 1906.
By then, de Forest had no job. White and partners avoided liability in connection with the Fessenden lawsuit by creating a new company, United Wireless, and lending it the assets of American De Forest for the sum of $1. De Forest was fired with a $1000 settlement and several worthless patents, including the Audion patent. Accustomed to a lavish lifestyle, he faced serious financial difficulties and was desperate to turn the Audion into a big success.
To understand what happened next, it is important to know that de Forest believed he had invented the relay, in contrast to Fleming's rectifier. He made his Audion by connecting a battery to a cold valve plate, and believed that the signal in the antenna circuit (connected to the hot filament) was modulating a more powerful current in the battery circuit. He was wrong: these were not two circuits, the battery simply shifted the signal from the antenna, and did not amplify it.
But this error became critical, as it led de Forest to experiment with a third electrode in the flask, which was supposed to further disconnect the two circuits of this "relay". He first added a second cold electrode next to the first, but then, perhaps influenced by the control mechanisms used by physicists to redirect beams in cathode beam devices, he moved the electrode into place between the filament and the primary plate. He decided that this position could interrupt the flow of electricity, and changed the shape of the third electrode from a plate to a wavy wire that resembled a griddle - and called it a "grid".
1908 Audion triode. The thread (broken) on the left is the cathode, the wavy wire is the grid, the rounded metal plate is the anode. It still has a thread, like a regular light bulb.
And it really was a relay. A weak current (such as that produced by a radio antenna) applied to the grid could control a much stronger current between the filament and the plate, repelling charged particles that tried to pass between them. This detector worked much better than the valve, because it not only straightened, but also amplified the radio signal. And, like a valve (and unlike a coherer), it could give out a constant signal, which made it possible to create not only a radiotelegraph, but also a radiotelephone (and later - the transmission of voice and music).
In practice, it didn't work particularly well. De Forest's audions were finicky, burned out quickly, lacked consistency in manufacturing, and were ineffective as amplifiers. In order for a particular Audion to work correctly, it was necessary to adjust the electrical parameters of the circuit for it.
Nevertheless, de Forest believed in his invention. He organized a new company, the De Forest Radio Telephone Company, to advertise it, but sales were meager. The biggest success was the sale of equipment to the fleet for intra-naval telephony during a round-the-world voyage "". However, the fleet commander, not having time to get de Forest's transmitters and receivers working and to train the crew in their use, ordered them to be packed and left in storage. Moreover, de Forest's new company, led by a follower of Abraham White, was no more respectable than the previous one. In addition to his failures, he soon came under charges of fraud.
For five years Audion achieved nothing. Once again, the telephone will play a key role in the development of the digital relay, this time rescuing a promising but untested technology that was on the brink of oblivion.
And again the phone
The long distance communications network was the central nervous system of AT&T. It tied together many local companies and provided a key competitive advantage when Bell's patents expired. By joining the AT&T network, a new customer could, in theory, reach all the other subscribers thousands of miles away - although in reality, long-distance calls were rarely made. Also, the network was the material basis for the company's overarching ideology of "One policy, one system, one-stop service."
But with the beginning of the second decade of the twentieth century, this network reached its physical maximum. The further the telephone wires stretched, the weaker and noisier the signal passing through them became, and as a result, speech became almost indistinguishable. Because of this, the US actually had two AT&T networks separated by a continental ridge.
For the eastern network, New York was the peg, and mechanical repeaters and - a leash that determined how far a human voice could reach. But these technologies were not omnipotent. The coils changed the electrical properties of the telephone circuit, reducing the attenuation of the voice frequencies - but they could only reduce it, not eliminate it. Mechanical repeaters (just a phone speaker connected to an amplifying microphone) added noise with each repeat. The 1911 line from New York to Denver brought this tether to its maximum length. There was no question of extending the network to the entire continent. However, in 1909 John Carthy, AT&T's chief engineer, publicly promised to do just that. He promised to do it in five years - by the time the in San Francisco in 1915.
The first person to make such an undertaking possible with the help of a new telephone amplifier was not an American, but the scion of a wealthy Viennese family with an interest in science. Being young With the help of his parents, he bought a telephone manufacturing company and set out to make an amplifier for telephone conversations. By 1906, he had made relays based on cathode ray tubes, which by that time were widely used in physical experiments (and later became the basis for the video screen technology that dominated the XNUMXth century). The weak incoming signal controlled an electromagnet that bent a beam that modulated a stronger current in the main circuit.
By 1910, von Lieben and his colleagues, Eugene Reise and Sigmund Strauss, had become aware of de Forest's Audion and had replaced the magnet in the tube with a grid that controlled cathode rays, a design that was the most efficient and superior to anything made in the United States at the time. The German telephone network soon adopted the von Lieben amplifier. In 1914, thanks to her, a nervous telephone call was made by the commander of the East Prussian army to the German headquarters, located 1000 kilometers away, in Koblenz. This forced the Chief of Staff to send Generals Hindenberg and Ludendorff to the east, to eternal glory and with dire consequences. The same amplifiers later connected the German headquarters with the field armies in the south and east as far as Macedonia and Romania.
Copy of von Lieben's improved cathode ray relay. The cathode is at the bottom, the anode is the coil at the top, and the grid is the round metal foil in the middle.
However, language and geographical barriers, as well as the war, meant that such a design did not reach the United States, and other events soon overtook it.
Meanwhile, de Forest left the faltering Radio Telephone Company in 1911 and fled to California. There he took a job at the Federal Telegraph Company in Palo Alto, founded by a Stanford graduate . Nominally, de Forest was to work on an amplifier that would increase the output of the federal radio. In fact, he, Herbert van Ettan (an experienced telephone engineer) and Charles Loguewood (a receiver designer) were building a telephone amplifier in order to win a rumored $1 million prize from AT&T.
To do this, de Forest took out an Audion from the mezzanine, and by 1912 he and his colleagues already had a device ready for demonstration at the telephone company. It consisted of several Audions connected in series, which created amplification in several stages, and several additional auxiliary components. The device, in principle, worked - it could amplify the signal enough that you could hear a handkerchief falling or a pocket watch ticking. But only at currents and voltages too small to be useful in telephony. When the current was increased, the Audions began to emit a blue glow, and the signal turned into noise. But the telephonists were interested enough to give the device to their engineers and see what they could do with it. It so happened that one of them, a young physicist Harold Arnold, knew exactly how to fix the amplifier from the Federal Telegraph.
It's time to discuss how the valve and the Audion worked. The key understanding needed to explain their work came from the Cavendish Laboratory in Cambridge, the intellectual center of new electron physics. In 1899 there, J. J. Thomson showed in experiments with cathode ray tubes that a particle with mass, and later known as an electron, carries current from the cathode to the anode. Over the next few years, Owen Richardson, a colleague of Thomson's, developed this assumption into a mathematical theory of thermionic emission.
Ambrose Fleming, an engineer who worked a short train ride from Cambridge, was familiar with this work. It was clear to him that his valve worked due to thermionic emission of electrons from the heated filament, crossing the vacuum gap to the cold anode. But the vacuum in the indicator lamp was not deep - for an ordinary light bulb this was not necessary. It was enough to pump out so much oxygen that the thread did not catch fire. Fleming realized that for the best performance of the valve, it must be emptied as thoroughly as possible so that the remaining gas does not interfere with the flow of electrons.
De Forest didn't get it. Since he came to the valve and the Audion through experiments with the Bunsen burner, his belief was the opposite - that hot ionized gas was the working fluid of the device, and that its complete removal would lead to the cessation of work. That is why the Audion worked so unstable and unsatisfactorily as a radio receiver, and why it emitted blue light.
Arnold at AT&T was in the perfect position to correct de Forest's mistake. He was a physicist who had studied under Robert Millikan at the University of Chicago and was hired specifically to apply his knowledge of new electronic physics to the problem of building a coast-to-coast telephone network. He knew that the Audion tube would work best in near-perfect vacuum, he knew that the latest pumps could achieve such a vacuum, he knew that a new type of oxide-coated filament, together with a larger plate and grid, could also increase the flow of electrons. In short, he turned the Audion into a vacuum tube, the miracle worker of the electronic age.
AT&T had the powerful booster needed to build a transcontinental line - only it didn't have the rights to use it. Company representatives were incredulous in their negotiations with de Forest, but started a separate conversation through an outside lawyer who managed to acquire the rights to use the Audion as a telephone amplifier for $50 (about $000 million in 1,25 dollars). The New York-San Francisco line opened just in time, but more as a triumph of technical virtuosity and corporate advertising than as a means of communication. The cost of calls was so cosmic that almost no one could use it.
electronic era
The real vacuum tube has become the root of an entirely new tree of electronic components. Like the relay, the vacuum tube has continually expanded its uses as engineers have found new ways to tailor its design to meet specific needs. The growth of the "-ode" tribe did not end with diodes and triodes. He continued with , which added an additional mesh that maintained gain with the growth of elements in the circuit. Then appeared , , and even . Thyratrons appeared, filled with mercury vapor, glowing with an ominous blue light. Miniature lamps the size of a little finger on a leg or even an acorn. Lamps with an indirect filament cathode, in which the buzzing of the alternating current source did not disturb the signal. The Saga of the Vacuum Tube, a book describing the growth of the vacuum tube industry before 1930, lists over 1000 different models by their index - although many of them were illegal copies from untrustworthy brands: Ultron, Perfectron, Supertron, Voltron , and so on.
More important than the variety of forms was the variety of uses for the vacuum tube. Regenerative circuits turned the triode into a transmitter - producing smooth and constant sine waves, without noisy sparks, capable of perfectly transmitting sound. With a coherer and sparks in 1901, Marconi could barely transmit a small piece of Morse code across the narrow Atlantic. In 1915, using a vacuum tube as a transmitter and receiver, AT&T could transmit the human voice from Arlington, Virginia to Honoluluβtwice the distance. By the 1920s they had combined telephony over long distances with high quality sound broadcasting and created the first radio networks. Thus, soon the whole nation could hear the same voice on the radio, whether it was Roosevelt or Hitler.
Moreover, the ability to create transmitters tuned to an accurate and stable frequency allowed telecommunications engineers to realize the long-held dream of a frequency multiplex that attracted Alexander Bell, Edison, and others forty years ago. By 1923, AT&T had a ten-line voice line from New York to Pittsburgh. The ability to transmit multiple voices over a single copper wire radically reduced the cost of long-distance calls, which, because of the high cost, has always been affordable only for the richest people and businesses. Seeing what vacuum tubes were capable of, AT&T sent their lawyers to buy out additional rights from de Forest in order to secure the rights to use the Audion in all available applications. In total, they paid him $390, which is equivalent to about $000 million in today's money.
Why, with such versatility, didn't vacuum tubes dominate the first generation of computers in the same way that they dominated radio and other telecommunications equipment? Obviously, the triode could be a digital switch just like a relay. So obvious that de Forest even believed he had created the relay before he actually created it. And the triode was much more responsive than a traditional electromechanical relay because it didn't have to physically move the armature. A typical relay took a few milliseconds to switch, and the change in flux from cathode to anode due to a change in electrical potential across the grid was almost instantaneous.
But lamps had a clear disadvantage over relays: their tendency, like their predecessors, light bulbs, to burn out. The life of the original Audion de Forest was so short - about 100 hours - that it contained a spare filament in the lamp, which had to be connected after the first burned out. It was very bad, but even after that, even the best quality lamps could not be expected to run more than a few thousand hours. For computers with thousands of lamps and hours of calculations, this was a serious problem.
Relays, on the other hand, were, according to George Stibitz, "fantastically reliable." So much so that he claimed that
If a set of U-shaped relays started operating in the first year of our era and switched a contact once a second, it would still work today. The first failure in contact could be expected not earlier than in a thousand years, somewhere in the year 3000.
Moreover, there was no experience with large electronic circuits comparable to the electromechanical circuits of telephone engineers. Radios and other equipment may contain 5-10 lamps, but not hundreds of thousands. Nobody knew if it would be possible to make a computer with 5000 lamps work. By choosing relays over lamps, computer designers made a safe and conservative choice.
In the next part, we will see how and why these doubts were overcome.
Source: habr.com