Elements of Radio Telephony, William C. Ballard, Jr., 1922, pages 14-23:
It has been known in scientific circles for several hundred years that heated bodies exhibit peculiar electrical effects. For example, early investigators found that by bringing a charged electroscope near a red-hot iron ball the electroscope would lose its charge. As a part of his researches in connection with the development of the incandescent lamp, Edison noted that the filament had the power of giving off electricity when heated.
Modern discoveries in physics indicate that all material bodies are composed of minute divisions of matter termed molecules and that these molecules are in constant motion inside the substance. The rapidity with which these particles move and the temperature of the body are thought to be closely related, the higher the temperature the greater the molecular velocity. On the other hand as the temperature is lowered the molecular activity becomes less and less until the molecules come to a standstill at the absolute zero of temperature which corresponds to 273 degrees below zero in the Centigrade scale.
If the molecular structure of a material is such that electricity can pass through it, we call it an electrical conductor. Scientists think of electricity as being composed of large numbers of infinite divisions which have been termed "electrons." An electron is quite a small value when compared with the amounts of electricity we deal with in everyday life; for example, in order to light an average-sized incandescent lamp of fifty watts capacity it requires a stream of about 3,000,000,000,000,000,000 electrons per second.
When the molecular structure of any material prevents the passage of electrons more or less completely we rate this material as an electrical insulator.
When water is exposed to the air a certain amount of it vaporizes, and it evaporates as we say. The rapidity with which it evaporates is dependent upon several conditions, one of which is the temperature of the water. This is explained according to the "molecular motion" theory of heat by saying that some of the water molecules have attained sufficient speed due to the temperature to break through the surface and be shot off into the air. Water does not have to be in liquid form to evaporate; it is a well-known fact that ice will evaporate at temperatures considerably below the freezing point. This fact can be checked by hanging a wet cloth out on a very cold day; if conditions are right it will dry almost as rapidly as in the summer.
Other solid substances act in the same way as ice but usually require rather high temperatures before much evaporation takes place. One example is the gradual blackening of incandescent lamps which have been used for some time, the high temperature at which the filament is operated causes a certain amount of evaporation which condenses upon the cool walls of the glass bulb.
Tungsten is particularly suitable for lamp filaments because it can be heated very hot and give off large quantities of light without appreciable evaporation.
If high temperature produces extreme velocities of the molecules composing the material it is reasonable to suppose that it may also be true with regard to whatever electrons happen to be in the material. As the temperature is increased more and more some electrons will attain a sufficient velocity to break through the surface of the material and be shot off into space. This idea of "evaporation" of electricity from heated bodies is further substantiated by the fact that the numerical relation between temperature and amount of evaporation are very much the same in the case of the actual evaporation of the material and "evaporation" of the electrons or thermionic emission, as it is usually called.
It has been found that the electricity given off from heated filaments, such as are used in electron tubes, is negative as compared with the commonly accepted idea of positive and negative electricity. Years ago the assumption was made that an electric current consisted of a motion of electricity from positive to negative. This was an arbitrary assumption made on account of a lack of information one way or the other, and it seems now that it was incorrect for we believe that the electron passes from negative to positive and has the same effect as a passage of so-called "positive" electricity in the opposite direction. Thus, according to the original assumption an electron is negative electricity and passes from negative to positive.
Fleming Valve.--The earliest application of the thermionic emission principle to radio telegraphy was in the Fleming valve used as a rectifier in radio receiving circuits. The apparatus consisted of an incandescent lamp filament mounted together with another electrode in the form of a plate inside a glass bulb and exhausted in the usual manner. When the filament is brought up to incandescence large numbers of electrons are shot off. As each electron carries a negative charge, the removal of electrons or negative electricity imparts a corresponding positive charge to the filament, the removal of negative electricity having the same effect as the addition of positive electricity. Unlike polarities attract and like polarities repel, so the positive filament will attract back all the electrons unless some other electrical force is acting upon them. However, if the plate is charged positively by connecting it to a battery, a certain number of the electrons will be attracted over to the plate.
This action takes place only when the plate is positively charged with respect to the filament; if the connections to the battery be reversed then the plate will be negative with respect to the filament and repel the electrons back into the filament even more vigorously. Thus current will flow when the plate is positive but not when the plate is negative. If an alternating voltage is applied between plate and filament, electrons will pass and a current flow when the plate is positive but no action takes place when the plate is negative. The tube acts essentially like an electrical check valve, allowing current to pass in one direction but shutting it out absolutely in the other. Just how this rectifier action is utilized in the receiving circuit will be discussed in a later chapter.
Space Charge.--When applying comparatively low positive voltages on the plate we find experimentally that the current obtained represents only a fraction of the current corresponding to the number of electrons emitted from the filament. Thus some of the electrons sent off from the filament must have been attracted back into the filament even though the plate is charged positively. This action is attributed to what is generally known as "space charge." Each electron carries a negative charge, similar charges repel each other, hence there is always a mutual repulsion between electrons. After an electron has left the filament there are two forces acting upon it, one the force of attraction of the positively charged plate and the other a force of repulsion due to the electrons which have left the filament ahead of it and which on account of the repulsive action between similarly charged bodies are pushing it back into the filament. Whether the electron leaves the surface of the filament or not depends upon whether the force exerted by the plate or that due to the mutual repulsion action is the stronger. If the first predominates it will pass over to the plate, if not it will be repelled back into the filament. Each electron carries the same charge, hence two electrons a given distance away will repel another electron with twice the force that a single electron would. Thus the repulsive effect on the electron just leaving the filament depends upon how many electrons are clustered around the filament in their motion toward the plate, in other words the electron density in the space just around the filament. The greater the density of electrons moving away from the filament the greater the current flowing to the plate, thus for any given positive plate voltage there will be a certain electron density around the filament which will just neutralize the pull of the plate. This prevents any increase in the rate at which electrons are pulled away for if more are pulled away the density around the filament will be increased and the electrons would be pushed back into the filament with more force than the plate exerts to pull them away. If the voltage on the plate is increased the current will increase to the point where the increased electron density around the filament is again sufficiently great to nearly balance the increased pull of the higher voltage on the plate. There is a limit to the number of electrons which can be drawn over to the plate, however, and this is determined by the total number of electrons emitted by the filament which in turn is controlled by the filament temperature. The current corresponding to a complete utilization of all the electrons sent off by the filament is generally spoken of as the "saturation current."
Fig. 6 shows how the current flowing to the plate varies as the filament current is maintained at a constant value and the plate voltage varied. The saturation current is indicated by the sudden flattening out of the curve. This indicates that all the electrons sent off from the filament are being utilized and hence the number cannot be increased by increasing the plate voltage.
Grid Electrode.--The addition of the third electrode to the rectifying valve is due to Dr. Lee DeForest, who first investigated the action of the grid electrode. DeForest's three-element tube was of the same construction as the Fleming valve except that it contained a third electrode operating cold and located between the filament and plate. The third electrode enormously enhances the value of the vacuum valve and increases its functions from that of simple rectifier to amplifier, amplifying rectifier, oscillator and modulator.
In considering the action of the third electrode, the grid acts merely in the capacity of a second plate in so far as the electric field around the filament is concerned. The space charge limitation of current is no longer determined solely by the potential of the plate but is the result of the combination of the potentials of both plate and grid. When the grid is located between the filament and plate it usually has a much greater relative effect on the limitation of the electron current by space charge than does the plate. In other words the grid acts like a second plate located in a very much more advantageous position. In some large-sized tubes used for power purposes this ratio of effectiveness runs as high as 400, although usually much less. This means that if a given decrease in plate current required the reduction of the plate voltage from 800 volts to 400, the same effect could have been produced by lowering the potential of the grid by one volt. This ratio of effectiveness of grid and plate is usually spoken of as the "amplification constant." Grids are usually made in the form of a wire network; sometimes wound in the form of a helical spring with the filament in the center and with plates of cylindrical form outside, sometimes built in the form of a flat gridiron with fine wires closely spaced for use with flat plates. The amplification constant may be increased by placing the grid closer to the filament or by decreasing the opening between grid wires, but a tube with very fine grid and high amplification constant gives a relatively smaller plate current for a given plate voltage since the shielding action of the grid makes the plate voltage relatively less effective in drawing the electrons across.
Tungsten and Coated Filaments.--Most elements require an extremely high temperature before they give appreciable electron emission, hence only a few with very high melting points can be used practically. Tungsten with its extremely high melting point is very satisfactory where large electron currents are required and is used extensively for both transmitting and receiving tubes.
It has been found that numerous chemical compounds exhibit the same thermionic effects as pure metals and in certain cases at very much lower temperatures. The oxides of barium, strontium and calcium are particularly valuable in this respect and can be operated at a fairly low red heat as compared to the brilliant white heat required for tungsten. Since the power input to the filament is used very largely to produce heat in the filament and the heat produced in the filament is necessarily a total loss, emission at low temperatures represents quite a saving of filament power. The usual procedure is to coat a platinum wire or strip with certain compounds of the three metals which change into the oxide when heated. The author has built a number of tubes using other materials in place of platinum with very good success. The saving in power for a given filament emission is around 75 per cent as compared with the tungsten filament and makes it almost imperative to use coated filaments in tubes designed to be operated from dry cells.
The tubes shown in Fig. 7 have oxide-coated platinum filaments and those illustrated in Fig. 8 have uncoated tungsten filaments.