Vacuum tube

• 1. Introduction

  ›

  Vacuum tubes were the engine of the twentieth century's electronic revolution, and almost nobody can describe what one actually is. Picture a glass bottle the size of your thumb, with all the air pumped out. Inside, a heated wire releases a stream of electrons into the void. A second metal plate captures those electrons, and the result is a one-way gate for electric current. That simple principle, worked out by John Ambrose Fleming in 1904, opened a path to radio broadcasting, long-distance telephone calls, television, radar, and the first computers. Without the vacuum tube, none of those technologies would have arrived when they did. The questions the rest of this documentary will explore are just as striking as the invention itself: how did a device so fragile and heat-hungry come to control the modern world, and what finally knocked it from its pedestal?

• 2. The Fleming Valve

  ›

  John Ambrose Fleming was already working for two giants of early electrical technology when Guglielmo Marconi appointed him scientific advisor in 1899. Fleming had previously served as scientific advisor to Edison Telephone starting in 1879, then to Edison Electric Light from 1882, and also as technical consultant to Edison-Swan. Marconi needed a better detector for wireless signals, something more reliable than the coherer and more drivable than a magnetic detector. Fleming turned to experiments with what Edison had noticed years earlier: a current flowing in one direction between a heated filament and a nearby plate inside an evacuated bulb. That phenomenon, first formally reported by Frederick Guthrie in 1873, had been rediscovered by Thomas Edison in 1883 and become known as the Edison effect. Fleming patented his adaptation for the Marconi company in November 1904, with the patent issued in September 1905. He called it the oscillation valve because, like a valve in a water pipe, it allowed flow in only one direction. While not more sensitive than a properly tuned crystal detector, it avoided the crystal's great weakness: the delicate whisker contact that vibration or a bump could knock out of alignment, a serious problem on a ship at sea.

• 3. Amplification and the Triode

  ›

  Lee de Forest is credited with inventing the triode in 1907 while experimenting to improve his earlier diode Audion. By inserting a third electrode, a sparse grid of wire, between the heated cathode and the plate, he found that a tiny voltage on the grid controlled a much larger flow of current at the plate. A change of several volts on the control grid could swing the plate output by hundreds of volts. Because the grid drew no current when held negative relative to the cathode, the input power required was essentially zero, yet the output power change was large; this was true amplification. De Forest received a patent for the three-electrode version of his Audion in 1908. His early devices were not true hard-vacuum tubes; a small amount of residual gas produced a blue glow when the plate voltage exceeded about 60 volts. Harold D. Arnold of AT&T's engineering department, seeing a demonstration by de Forest and John Stone Stone in 1912, recognized the glow as ionized gas and recommended AT&T purchase the patent. Arnold then developed high-vacuum tubes that ran without the glow and were tested on AT&T's long-distance network in the summer of 1913. Finnish inventor Eric Tigerstedt further improved the design in 1914 while working on his sound-on-film process in Berlin. Tigerstedt made the electrodes concentric cylinders with the cathode at the center, greatly increasing the collection of electrons at the anode. General Electric, after Irving Langmuir improved an existing high-vacuum diffusion pump, began producing its own hard-vacuum triodes, branded Pliotrons, in 1915. The French type TM and the British type R followed closely, and by 1916 both designs were in widespread use by the allied military.

• 4. Tetrodes, Pentodes, and the Grid Wars

  ›

  Physicist Walter H. Schottky invented the tetrode, also called the screen grid tube, in 1919 to solve a specific problem: the triode tended to oscillate uncontrollably when used at radio frequencies because a small parasitic capacitance linked its output back to its input. Schottky added a second grid, held at a steady positive voltage, between the control grid and the plate. This screen grid decoupled the plate from the control grid, eliminating the need for elaborate neutralization circuits at broadcast frequencies. It also amplified more powerfully; while typical triodes had amplification factors ranging from below ten to around 100, tetrode amplification factors of 500 were common. Screen grid tubes reached the market by late 1927. A new problem emerged, though: electrons striking the plate with enough force knocked secondary electrons loose, and those secondary electrons could be captured by the screen grid rather than returning to the plate. Over a certain range of plate voltages, plate current could actually decrease as plate voltage rose. This was the tetrode kink, a region of negative resistance that could itself cause the circuit to oscillate. The solution arrived in 1926, when Bernard D. H. Tellegen invented the pentode. He added a third grid, called the suppressor grid, between the screen and the plate, held at cathode potential. Its negative voltage relative to the plate pushed secondary electrons back to the plate before they could reach the screen grid. The pentode became generally favored over the simple tetrode and is still manufactured in two classes: those with the suppressor grid wired internally to the cathode, such as the EL84, and those with the suppressor grid brought out to a separate pin, such as the 803.

• 5. Computing on Glass and Wire

  ›

  Tommy Flowers, who later designed the Colossus codebreaking machine at Bletchley Park, had noticed something that contradicted the received wisdom of his day. The prevailing belief was that vacuum tubes, like light bulbs, were too unreliable for large installations because too many would fail in too short a time. Flowers had introduced valves into telephone equipment in large numbers before the war and discovered that if they were never moved and never switched on and off, they would operate reliably for very long periods. In 1934 he built an experimental installation using over 3,000 tubes in small independent modules; when one tube failed, a single module could be switched off while the rest continued, limiting the cascade of further failures. That installation was accepted by the Post Office, which operated telephone exchanges. Colossus I, built for Bletchley Park during World War II, used about 1,600 valves; its successor Colossus II used about 2,400. The earlier Heath Robinson machine, which Colossus replaced, used only around a hundred valves. On Colossus's first day at Bletchley Park, a problem with a known answer was run. After four hours, with each run taking half an hour, the answer was the same every time. That consistency astonished the codebreakers, because the Robinson had not always given the same answer. The 1946 ENIAC carried the approach even further, running over 17,000 tubes and suffering a tube failure on average every two days; locating and replacing the failed tube typically took fifteen minutes. Advances in special-quality tubes specifically designed for computing followed: Sylvania's 7AK7 pentode of 1948 was the first such computer tube, developed to address cathode poisoning that cut off conduction after long periods at cutoff. By the late 1950s it had become routine for special-quality small-signal tubes to last for hundreds of thousands of hours when operated conservatively.

• 6. Power, Heat, and Endurance

  ›

  A vacuum tube generates heat from two sources simultaneously: the filament or heater, and the stream of electrons constantly striking the plate. Removing that heat from inside a sealed glass envelope is a genuine engineering challenge. Convective cooling is impossible inside a vacuum; the plate must radiate heat as infrared light through the glass. For small and medium power tubes this is sufficient, but for high-power applications the anode is sometimes made part of the metal tube envelope itself, in thermal contact with a heat sink cooled by forced air or water. The water-cooled tube designated the 8974 weighs 80 kilograms, produces 1.25 megawatts of dissipation, and remains among the largest commercial tubes available. The highest-power tube currently available is an Eimac forced-water-cooled power tetrode capable of dissipating 2.5 megawatts; by comparison, the largest power transistor can dissipate only about 1 kilowatt. At the other extreme, the longest recorded valve life belongs to a Mazda AC/P pentode, serial number 4418, in service at the BBC's main Northern Ireland transmitter at Lisnagarvey from 1935 until 1961, accumulating a documented 232,592 hours of operation. The BBC maintained meticulous records of valve lives with periodic returns to its central valve stores. Water used in tube cooling systems must be deionized, because ordinary water conducts electricity and would cause high-voltage leakage through the cooling circuit to the radiator. Such systems typically include a built-in conductance monitor that shuts down the high-tension supply if the water's conductance rises too high.

• 7. Decline, Survival, and Niche Persistence

  ›

  In the 1940s, the invention of semiconductor devices opened a path to solid-state electronics: smaller, safer, cooler, and more economical than thermionic tubes. Beginning in the mid-1960s, the transistor began displacing thermionic tubes from most applications. The cathode-ray tube, functionally an electron tube though not usually classified as one, survived longer, remaining in use for television receivers, computer monitors, and oscilloscopes until the early years of the twenty-first century. Some thermionic applications have proven difficult or impractical to replace. The magnetron at the heart of microwave ovens remains a thermionic tube. High-frequency amplifiers in some applications still use tubes. Audio enthusiasts continue to prefer tube amplifiers for what they describe as a warmer sound, and electric guitar players deliberately overdrive tube amplifiers to achieve distortion that defines the instrument's characteristic tone. At the extreme end of power, no solid-state device comes close to matching what a water-cooled tube can dissipate. Special-purpose tubes persist in radar transmitters, nuclear medicine imaging equipment, X-ray systems, and in industrial welding controls. A class of tubes that contain small amounts of radioactive material, including isotopes such as tritium, krypton-85, and cesium-137, uses ionization of an internal fill gas to ensure fast and consistent switching. The Western Electric 346B, for example, contains radium-226. That so many categories of the vacuum tube endure more than sixty years after the transistor's rise says something concrete about the limits of what solid-state devices can yet replicate.