Fleming originated the oscillation valve illustrated in Fig. 118. It consists of an ordinary incandescent lamp with a carbon filament, having a metal cylinder, C, placed around the filament, but attached to an independently insulated platinum wire sealed in the glass. When the lamp is lighted by passing a current through the filament, the incandescent carbon liberates negative ions. If oscillations are then set up in a circuit which includes a pair of sensitive telephone receivers and is formed by connecting the negative terminal of the filament with the platinum cylinder, negative electricity will be enabled to pass from the filament to the cylinder but not in the opposite direction, and so sounds will be produced in the telephone receivers. High frequency oscillations themselves could not be made to pass through the telephone receivers because of the choking action of the iron cores of the electromagnets.
Fig. 119. Flame Audion.
The simple but sensitive form of detector illustrated in Fig. 119 is not of practical value for commercial work, but is very interesting as the progenitor of the audion, and provides a good field for amateur investigation. Its only drawback is that the gas flame is very difficult to keep steady and every flicker registers as a sound in the telephone receivers.
A Bunsen burner using coal gas furnishes the flame, and a salt of an alkaline metal heated in the flame, the ions. The hydroxides of caesium, potassium and sodium give the best results in the order named.
The salt is contained in a piece of trough-shaped platinum foil, about 3/8 inch long and 1/16 inch wide. This trough is made the cathode or negative of the telephone circuit and placed in the outer oxidizing flame just above its juncture with the interior reducing flame and must be kept incandescent. The upper electrode or anode is a piece of platinum wire about 1/16 inch above the trough.
Fig. 120. Circuit of Flame Audion.
The arrangement and construction of the detector is clearly indicated by the drawing so that it is unnecessary to go into details. The block, E, which fits on the tube of the Bunsen burner, is made of fiber. Two double binding posts, D, are fastened to E to support the rods, R, which are fitted at the tops with binding posts, B, into which the electrodes may be clamped.
Twelve dry cells are connected with a multiple point switch so that an electromotive force of 6-18 volts, varying in steps of one cell at a time, may be secured. The flame is best provided with a mica chimney to protect it from drafts. By keeping plenty of salt in the trough and carefully adjusting the voltage, this detector may be made marvelously sensitive.