30th September 1925.
The Thermionic Valve - Its Origin and Development
The history of the Thermionic Valve forms one of the most astonishing chapters in the long record of scientific research and invention.
By Dr J A Fleming FRS (Sir John Ambrose Fleming, FRS).
MANY electrical inventions of great importance are extremely simple in structure or nature, so much so that they appear to be almost obvious things for which little credit need be given to any particular inventor or originator. Nevertheless, that simplicity is their chief merit, and it has in nearly all cases only been obtained after innumerable trials and failures with more complicated contrivances.
We need only mention as illustrations such inventions as the speaking telephone of Bell, the incandescent filament electric lamp of Edison and Swan, and the thermionic valve, the origin of which is the subject of the present article.
Like the speaking telephone and the incandescent lamp, the thermionic valve has rendered possible the creation of an immense industry based entirely upon it, since without the valve there would be no broadcasting, therefore no enormous manufacture of wireless telephony receiving apparatus, no construction of broadcasting stations, or large periodical literature which now instructs the amateur or aids research in the art of wireless telephony.
Nay, more, there would be no long-distance wireless telegraphy or the wonderful beam system, and no possibility of intercommunication with aeroplanes or directional wireless for ships. The thermionic valve, in short, is the chief foundation stone of modern radiotelegraphy and radiotelephony, and the amazing progress of these arts, giving employment to hundreds of thousands of persons, is the direct result of the invention and development of this marvellous appliance.
Hence it becomes interesting to look into its history, explore the stages of its evolution, and examine the nature of the scientific researches which brought it into existence.
Inventions Develop Gradually
Scientilic invention is subject to the law of evolution Things do not spring into being full grown and at once, as in classical fable Minerva is said to have sprung full grown from the head of Jove. They proceed by small stages and a careful examination shows that the pivotal ideas, observations, or inventions, were contributed at particular times by particular persons who had the insight to make them at the appropriate moment.
There is no necessity to encumber our narrative with any lengthy references to the history of wireless telegraphy. There is only one type of telegraphy which matters at the present tune, in which an interconnecting wire is unnecessary between the sending and receiving stations, and that is the electromagnetic wave or radiotelegraphy.
We cannot forget nor forbear to mention that the scientific conception of an electromagnetic wave originated in 1865 with James Clerk Maxwell, whose death in 1879 at the early age of 49 deprived the world of one of its most original scientific thinkers.
Twenty years had to elapse, even in spite of suggestions by another great philosopher, G. Fitzgerald, before H. R. Hertz, with his immense genius for experimental and theoretic research; gave us the means of creating, and in a rudimentary manner detecting, Maxwell's electric waves. When once it had been clearly recognised that beyond the actinic, luminous, and dark heat vibrations in the ether there is an almost unlimited range of longer waves to which the photographic plate, human eye, and thermopile are insensitive, immense interest began to be taken in the possible modes of generation and detection of these longer waves.
The insensitive spark gap detector of Hertz was soon replaced by the coherer of Branly and Lodge, and by it much useful work was done. Then came an era when a few select minds, such as those of Admiral Sir Henry Jackson, Sir William Crookes, Mr. Campbell Swinton, and especially Senatore Marconi, saw that this long wave radiation might be used for telegraphic purposes.
Marconi's great invention of the aerial and earth connection and his improvement of the Branly-Lodge coherer into a more certain type of wave detector translated ideas and suggestions into actual achievement, and he gave us a slow but veritable radiotelegraphy by electric waves.
On this basis, with the improved spark transmitter invented by Marconi, as described in his celebrated British patent, No. 7777 of 1900, and with the embodiment of the principles of syntony or tuning described in Lodge's 1897 British specification, spark radiotelegraphy came into existence as a practical art of immediate and immense assistance to ships and voyagers.
The invention of the arc-transmitter by Poulsen's improvement on the Duddell arc-oscillation generator, first made continuous wave radiotelegraphy possible,
Meanwhile the coherer had been replaced as a detector of the feeble oscillations induced in the receiving aerial by the magnetic detector of Marconi, and forms of self restoring coherer and elecrolytic detector tried with limited success. What was required and being looked for by many was a detector which would not require the careful adjustments of the tapped coherer and be more sensitive than existing detectors.
The writer of this article had turned attention from its very earliest days to the subject of electromagnetic radiation. He may, in fact, describe himself as having been a personal pupil of Clerk Maxwell, since he relinquished in 1877 a teaching post at Cheltenham College to go up to Cambridge chifly with the object of working under Maxwell in the then recently erected Cavendish Laboratory.
There for two years he enjoyed Maxwell's stimulating teaching and intercourse. In the year that Maxwell died the writer was appointed scientific adviser of the Edison Telephone Company of London, and three years later to a similar position with the Edison Electric Light Company, formed to introduce incandescent electric lighting into London.
This close connection with electrotechnics .gave an opportunity for the study of certain physical appliances on a scale quite impossible in the laboratory.
The incandescent electric lamp, invented by Edison, consisted of a horse- shoe-shaped filament of carbonised bamboo sealed into an exhausted glass bulb. The filament was necessarily not quite uniform, and, when heated by a current, some parts became hotter than others. The carbon also volatilised and was deposited as a black coating on the bulb.
The writer noticed many cases of blackened lamps in which there was a white line or no deposit in the plane of the filament loop, and on the opposite side of the loop to that on which the filament had burnt through. This clearly indicated that there had been a projection of carbon particles in straight lines from that hottest spot on the carbon loop. Observations on this subject the writer communicated to the Physical Society of London in 1883.
The writer's attention was then called to an observation of Edison's when he sealed a metal plate into the lamp bulb, this plate standing between the legs of the horse-shoe filament. When the filament was rendered incandescent by a direct current it was found that a galvanometer connected between the positive terminal of the filament and the plate indicated a small direct current, but that when connected between the plate and the negative end of the filament gave no sensible current.
Edison gave no explanation of this effect, and made no application of it that had the smallest reference to rectifying electric currents.
The writer made a very extensive investigation of the effect in the years 1885 to 1890, and was thereby led to the conclusion that there was an emission of particles chiefly from the negative leg of the filament, which were charged with negative electricity.
Those days were long before Sir. Joseph Thomson had made his remarkable discovery that there are particles smaller than chemical atoms in a high-vacuum electric discharge tube, and the writer had therefore concluded that these charged particles must be atoms of carbon which were thrown off from the filament, and that each carried a negative charge.
Effect of a Plate
It was proved by the writer, experimentally, that this torrent of electrified particles was stopped by a mica or metal plate, but could pass through the apertures in a metal grid or zigzag of wire.
Also it was proved that when the two electrodes in a vacuum tube were made of carbon filaments, a very small direct voltage could create a sensible current through the rarified gas, provided that the negative electrode was made incandescent.
It was not until 1899 that Sir Joseph Thomson demonstrated that these charged particles emitted by incandescent carbon in vacuum`were electrons, having only 1/1800th the mass of a hydrogen atom, but each carrying the same electric charge as a hydrogen ion in electrolysis. Meanwhile wireless telegraphy by electric waves had been invented and put in practice by Marconi, and in April, 1899, he astonished the world by achieving wireless telegraphy across the English Channel.
In September, 1899, the writer lectured before the British Association at Dover on the centenary of the electric current, and remarkable cross-Channel demonstrations of electric wave telegraphy were made by Marconi's Wireless Telegraph Company.
In 1900 the writer became connected with that company as scientific adviser, and was entrusted with the duty of specifying and assisting to design the earliest transmitting plant prepared for trans- Atlantic wireless experiments.
Need for a Simple Detector
In this work it had become clear to the writer that some simpler and more easily managed detector must replace the coherer if speeds of signalling were to be attained comparable with those on submarine cables. The writer was well acquainted with the apparatus used in cable telegraphy, and it was not long before the idea occurred to him that if the feeble high-frequency alternating electric currents induced by the waves in the receiving aerial could be "rectified," that is, converted into direct currents, we could then use a mirror galvanometer or syphon recorder to detect them.
Accordingly, many experiments were made with devices having unilateral conductivity, but only after a time was it found by the writer that a perfect device existed in a high vacuum tube containing two electrodes, one of which was rendered incandescent by a current.
Finally, in October, 1904, the invention was made of a thermionic "valve", consisting of a carbon filament electric lamp in which the loop filament was surrounded by a metal cylinder, the latter connected to a wire sealed through the glass bulb.
When the filament is rendered incandescent by a current, the space between the Hlament and the cylinder acquires a unilateral conductivity and allows negative electricity to pass from the filament to the cylinder, but not in the opposite direction, and can therefore convert high-frequency alternating currents into direct currents.
All that was necessary. therefore, was to include this filament and cylinder rectifying space in series with a galvanometer to detect such currents. When using spark oscillations, which come in groups or trains, a telephone could be used instead, and each train therefore gave a single sound in the telephone, and the groups of trains constituting the signals gave audible dash and dot signals.
This was produced and patented in Great Britain on November 16th 1904 - the now well- known "Fleming valve."
It began to be used in practical wireless reception by Marconi's Wireless Telegraph Company very soon after, and provided an easily managed detector not liable to be put out of adjustment by atmospherics.
It was, in fact, used for experimental reception across the Atlantic before long. This valve was absolutely the first technical application of the emission of electrons from an incandescent conductor in vacuo, and the importance of the invention is not to be measured merely by the actual device then first patented by the writer, but by the degree to which it opened up an entirely new field of research and invention. The present year 1925 may therefore be called the year of the "Coming of Age" of the Thermionic Valve.
Importance of the Invention The importance of the invention is also shown by the determined attempt made by American wireless men to claim the invention for themselves and deprive the present writer of credit for it and remove his name from connection with it by re-christening identically the same invention by other strange names, such as Audion, Kenotron, Tungar or Diode.
The name "valve" first given to it by the writer, is a simple, appropriate, easily pronounced English word of five letters, and has now become so firmly fixed in British wireless language that it will never be displayed by other out-landish words.
The endeavour on the part of one United States patentee to claim the invention for himself gave rise to much prolonged litigation, but the decisions of a Court of First Instance, given by His Honour Judge Mayer, later on confirmed by the unanimous decisions of three judges in the United States Court of Appeals, settled the question beyond dispute that the writer was the first and true inventor of the rectifying thermionic valve, by whatever name it may be called.
In Great Britain Mr. Justice Sargant, in 1918, declared the invention to be one of "unusual utility," and that, in his opinion, the three-electrode valve would never have come into being but for the previous invention of the 1904 Fleming valve.
The early Fleming valves were made with carbon filaments, but in 1908 the writer found that tungsten wire possessed advantages in that it could be heated to a higher temperature.
Also, in a British Patent specification first mentioning this, he described a method of using the valve for reception which consisted in applying to the anode circuit such a voltage that it brought the point of working on the characteristic curve just to a place where that curve has a change in curvature. This method is capable of giving great sensitivity in the case of certain soft valves.
The great use of this rectifying valve now is not to rectify the feeble high-frequency oscillations in receiving sets, but to rectify a low-frequency but high voltage electromotive force required (e.m.f) to keep the anodes of three-electrode valves at a high direct potential.
This leads us to the next stage in the invention, namely, the means for controlling the electron flow from the filament to the anode cylinder. In experiments described to the Royal Society of London, in 1905, the writer had shown that a source of e.m.f., now called an anode battery, might be inserted in series with the valve, and that the electric conductivity of the vacuous space between hot filament and cold cylinder or anode depended upon this voltage.
In the paper above mentioned he gave the first delineation of the curve now called a characteristic curve of the valve expressing the relation between impressed voltage and anode current.
Introduction of the Grid
The writer was well aware that this current could be reduced by holding near the valve a permanent magnet, but unfortunately it did not occur to him in sufficient time that this anode current could be controlled by inserting a grid or spiral wire or metal mesh cylinder between the anode cylinder and the filament, and giving to this grid small positive or negative potentials. This however, was done in a rudimentary form in 1907 by Lee de Forest in the United States, who had been following carefully the work of the writer.
De Forest interposed a zigzag of wire carried on a separate terminal between the anode plate and filament, and thus made the first form of three-electrode valve, or, as it is sometimes called, triode. The introduction of the grid or controlling electrode was of great importance; it enabled the valve to become not only a detector but an amplifier for electric oscillations.
The next step in invention involved the improvements in making the vacuum. The writer had specified in his patent specifications that the vacuum should be as high as possible, but the attainment of that aim was limited by the power of the vacuum pumps then in existence.
Prior to 1904 only modifications of the Sprengel mercury pump were available, as used in incandescent lamp manufacture. But even an exhaustion to a millionth of an atmosphere leaves still some ten billion molecules of air per centimetre cube in the bulb. In 1904 Sir James Dewar gave us his beautiful process for making high vacua by charcoal cooled with liquid air.
Soon after Gaede invented the molecular pump, and at a later stage Irving Langmuir the mercury condensation pump. Still more recently Holweck has given us an improved molecular pump, and there have been several modifications of the mercury vapour injector pump. The result has been to render possible far higher vacua, and the process of exhaustion is now aided by chemical means by introducing metallic magnesium into the bulb which absorbs the residual oxygen and nitrogen.
The result is a hard or high vacuum valve in which the whole of the anode current is conveyed by electrons which come out of the filament, and no ionisation of gas molecules takes place.
Generation of Oscillations
Another great stage of improvement was when it was discovered that the valve could act as a generator of oscillations both in its two- and three- electrode form. In the latter type the grid and cylinder circuits are coupled together inductively in such fashion that the changes of current or potential in each circuit tend to augment each other. When once this had been discovered it rendered it possible to employ the valve in an improved manner by means of so-called reactive coupling as a more sensitive detector, as done first by C. S. Franklin and E. H. Armstrong.
Also, by gradual stages larger and larger bulbs were employed in generating valves of gradually increasing power.
This is the place to say a word or, two about improvements in the emitting filament. After carbon, with its rather low volatilising temperature, hard drawn tungsten wire, came to be employed, since it has a very high melting point (near 3,500 C), and the electron emission increases rapidly with temperature.
The Western Electric Company, of America, made a departure in the use of an oxide-coated metallic filament. It was discovered by Wehnelt that the oxides of earthy metals had the power of greatly increasing the electron emission of hot metals. The Western Electric filament is, therefore, made of a thin strip of' platinum-iridium or other metal coated with oxides of barium and strontium.
Another important discovery was that thorium had a similar power when placed on tungsten, and that a thoriated filament could give the required emission for detector valves at a dull red heat. Hence came the so-called "dull emitter" valves, which can be worked with one or two dry cells for reception.
Other great improvements were made in the bulb. In large transmitting valves the electron emission from the filament bombards the anode cylinder and makes it red-hot. This heat has to be dispersed by radiation, and with very large glass bulbs there is great risk of fracture.
One great advance was in the use of silica for the bulb, because it can be heated and cooled unequally without risk. The Mullard Valve Company have very successfully grappled with the problem of making silica-bulb high-power transmitting valves.
The greatest advance was made when it was discovered that the anode cylinder could form part of the bulb, and thus, being external, could be water-cooled. This was rendered possible by the discovery that a copper or copper-plated nickel tube with sharp edge could be sealed to a glass tube.
The cooling of the anode cylinder enables generating valves of very large power to be made, even up to many hundred kilowatts output; by this invention it is now possible to equip high-power stations entirely with valve generators. In fact, it is doubtful if either the Poulsen arc or the high-frequency alternator will be able to compete with the valve as the radio generator of the future.
The high power generating valve may involve in the future only an incandescent rod of tungsten and a water-cooled enclosing anode, and, in fact, be a simple impproved type of Fleming valve.
When a filament is heated by a current the magnetic field round it tends to prevent the escape of electrons. Hence, if that current is an alternating current, the electrons escape to the anode or gushes at the instant when the heating current passes through its zero values. If, then, that intermittent anode current is passed through a transformer, the secondary current will be a pure alternating current of double the frequency of the filament-heating current. In this manner it may be possible to step-up frequency and create powerful high-frequency oscillations.
It is now nearly, 21 years since the writer invented and gave to the radiotelegraphic art the simple two-electrode rectifying valve, but that invention has been proved to possess a wonderful fertility -and stimulated the inventive powers of many able minds, so that in the short space of two decades we have become possessed of an appliance having the most astonishing range of properties for the generation and detection of Maxwell’s electromagnetic waves, and of which we have not yet exhausted all the possibilities.
In its power of creating short electric waves it has no rival, and it has been the starting point for the great inventions resulting in the marvellous beam system of radiotelegraphy of Senatore Marconi, perhaps the most remarkable of his many great inventions, in the development of which he has been so ably assisted by Mr. C. S. Franklin.
The history of the thermionic valve forms one of the most astonishing chapters in the long record of scientific research and invention. At one end of the chain we have the initial observations of the writer on some curious effects in connection with blackened carbon filament incandescent lamps, and at the other end a perfected appliance which enables a single human voice or musical instrument to make itself heard almost over the whole habitable surface of the globe.
Go to Early Radio Contents Page