QST, April, 1920, pages 5-7 (opening sections of article):

The Vacuum Tube as a Detector and Amplifier
By L. M. Clement
 Presented at Meeting of the Radio Club of America, Columbia University, January 16, 1920
 
A large part of the practical wartime development work of the Western Electric Co. in vacuum tubes is creditable to Mr. Clement. This article therefore may be accepted as an authoritative presentation of the subject in question, and fills a need we have long felt in QST.--Editor.

    The audion or vacuum tube, as it is sometimes called, due in large measure to the rapid development work during the great war and partly to the pre-war development of several of the large companies, is a necessity to all fully-equipped radio telephone and telegraph stations.
    The audion, as we all know, consists of an evacuated glass tube or bulb which contains the filament, the plate, which generally surrounds it, and the control electrode or grid which is generally mounted between the filament and the plate. When this device is associated with the proper apparatus and operated under the proper conditions, it can be made to function as a detector, an amplifier, an oscillator, or a modulator.
    In order to be able to design apparatus which will operate under proper conditions, we must first know something about what goes on in the tube under different conditions.
    Edison, during some of his experimental work with the incandescent lamp, found that current could be made to flow from a plate inside the bulb to the filament but not in the reverse direction. This phenomenon is called "Edison Effect" and it remained unaccounted for a number of years but was finally explained by the Electron Theory.

Filament

    The filament of the audion, generally in the form of a wire, is heated by the electric current to such a temperature that a large number of electrons attain sufficient velocity to leave the surface of the wire. The filament then is deficient in negative electricity, or, in other words, has a positive charge which attracts the electrons which have left its surface. If no outside forces are active, a state of equilibrium is said to exist when the number of electrons leaving per unit of time is equal to the number falling back on the filament in that time.
    Because of its very high melting point, allowing of very high operating temperature, tungsten is a very suitable substance for vacuum tube filaments. Platinum wire filaments which have been coated with certain metallic oxides have been found to emit a copious supply of electrons at comparatively low temperatures. This accounts for the long life of the so-called "coated" type of filament.

Plate
Fig 1. Study circuit
    If the filament is placed in an evacuated bulb, it is obvious that the electrons will penetrate the space surrounding the filament to a greater distance than they will in air, because of the removal of the large gas molecules. Suppose now an electrode in the form of a plate is introduced in the bulb and a potential positive with respect to the filament is applied to it. The negative electrons will flow to the plate. Instead of thinking of a flow of electrons from the filament to the plate, we generally think of a flow of current from the plate to the filament; that is, from the positive to the negatively charged body. If a negative potential were applied to the plate, no electrons would be attracted to the plate, and hence no current would flow because of the lack of these carriers of negative electricity. If an alternating potential were applied to the plate it is obvious that current would only flow through that part of the cycle when the plate was positive with respect to the filament.
    The two-element tube, due to its unilateral conductivity, has found some application as a rectifier. The General Electric Company has built some commercial types of rectifiers and some tubes have been built for potentials of 180,000 volts.
    Fleming recognized that this device could be used to rectify or detect radio frequency signals. He called the rectifier for this purpose a receiving valve. The device was not very generally used as it was not far superior to the ordinary crystal detectors which required no external battery.
 

Ionization

    In these old valves for receiving work, due to the low voltages employed, a high degree of vacuum was not necessary. In rectifiers similar to the Kenotrons of the General Electric Company, which are designed to operate at high voltage, it is necessary to remove the gas molecules to prevent ionization by collision. Ionization is generally accompanied by a pink or blue glow which fills the tube. The presence of gas in the tube causes irregularity of action at low voltages and excessive ionization or even arcing at high voltages.
    In order to overcome these difficulties it is necessary to so evacuate the tube that excessive ionization does not occur at the operating voltage. This is accomplished by special vacuum pumps; and the heating of the elements, and even tubes themselves, to drive off occluded gases.

Temperature Saturation
Fig 2. Temp saturation
    In order to study the action of these valves under different conditions, a circuit shown in Figure 1 is used. The data in Table I shows the effect of filament current (which is a measure of the filament temperature) on the plate current for different plate potentials. These results are shown in the curves of Figure 2. For a constant plate potential, we observe that the plate current increases with filament current until saturation point is reached, after which no rise in the plate current takes place.
    On the section of the 60 volt curve, Figure 2, AB, the plate voltage is drawing to the plate all of the electrons emitted, but beyond the saturation point the filament is supplying more electrons than the plate potential can draw to it. This is due to the resultant charge of the cloud of electrons between the filament and the plate which causes the excess of electrons to be returned to the filament. This is sometimes spoken of as the space charge effect.
    As is to be expected the saturation points for the 30 and 15 volt currents occur at lower filament temperatures. In general, tubes should be operated beyond the temperature saturation points as then a small change in filament current produces practically no change in plate current.
 

Voltage Saturation
voltage saturation
    The voltage-current characteristics of the valve were taken at three different filament temperatures. This data is contained in Table II and plotted on curves in Figure 3. As the plate voltage is increased the plate current rises until the saturation point is reached, beyond which an increase of voltage does not produce an increase of current.
    Below the saturation point the filament is emitting more electrons than can be drawn to the plate at the plate voltage applied, due to the space charge effect. Beyond the saturation point all of the electrons emitted by the filament are drawn to the plate. The saturation occurs at lower plate currents when the filament current is less because fewer electrons are available.
 

The Audion or Vacuum Tube

    Dr. DeForest found that the plate current or flow of electrons between the filament and the plate could be controlled by the application of potentials to a third grid-like structure placed between the filament and the plate.
    The tubes (or audions as they were called by Dr. DeForest) made several years ago were not pumped very well and consequently contained a large number of gas molecules which made their action more or less unsteady. The high vacuum tubes manufactured by the British, French, General Electric Company and Western Electric Company are pumped to a very high degree of vacuum and consequently little trouble from ionization is experienced.