August 5, 1920] 



NATURM 



719 



to two valves; we can thus employ three, four, or 

 more valves in cascade, as it is called, and each one 

 multiplies or amplifies the effect of the one before. It 

 is this use of three-electrode valves in cascade that 

 has given us recently such vastly increased powers 

 of detecting wireless waves. The L^'st or final 

 amplifying valve may be made to operate a detecting 

 or rectifying valve, or perhaps a crystal detector. 



But there is an additional very valuable power pos- 

 sessed by the thermionic valve, viz. that it can 

 generate electric oscillations as well as detect them. 

 We have already seen that the fundamental property 

 of this valve is that variations •of grid potential create 

 similar variations of plate or thermionic current. 

 Supposing, then, that this latter current is passed 

 through a coil over which is wound another secondary 

 coil connecting the grid and filament (Fig. 2). 

 It is possible so to make the connections that any 

 increase in the plate current will give the grid a 

 negative charge and so immediately reduce the plate 

 current. Conversely, any reduction of plate current 

 will give the grid a positive charge which will again 

 increase the plate current. Hence the operations in 

 the plate current when once started will be main- 

 tained, the energy required being drawn from the 

 battery B (see Fig 2) in the plate circuit. The action 



=^B 



Fig. 2. — Connections for generator valve. 



resembles that in the well-known experiment called 

 the singing telephone. 



The discovery of the oscillation-producing power of 

 the valve was of great importance, because it at once 

 put it in our power to conduct wireless telephony with 

 simple, easily managed apparatus. The principles of 

 radio-telephony are briefly as follows : At the trans- 

 mitting station we have to establish in the sending 

 aerial undamped or persistent oscillations and to 

 radiate continuous waves. By means of a carbon 

 microphone we have then to rnodulate the amplitude 

 or intensity of these waves in accordance with the 

 wave-form of the sp>eaking voice. 



The arrangements for a wireless telephone trans- 

 mitter are, then, as follows : By means of a 

 thermionic valve, with its plate and grid circuit induc- 

 tively coupled, we set up, as alreadv explained, per- 

 sistent electric oscillations in the plate circuit, and 

 these are transferred by induction to an aerial wire 

 properly tuned to sympathetic vibration. High fre- 

 quency electric currents, therefore, flow up and down 

 the aerial. These produce magnetic and electric 

 effects in surrounding space which are propagated 

 outwards as an electromagnetic wave. We have in 

 the next place to vary the amplitude of these radiated 

 electromagnetic waves bv a speaking microphone, and 

 this is done by means of a control valve. This latter 

 valve has its grid circuit inductivelv connected hv a 



NO. 2649, VOL. TO5I 



transformer with a circuit containing a battery and a 

 telephone transmitter. 



Hence, when speech is made to the mouthpiece of 

 the carbon microphone, this varies the electric current 

 through it, and therefore the potential of the grid, in 

 accordance with the wave-form of the speech sound. 

 The plate circuit of this control valve is joined in 

 parallel with that of the generating or power valve, 

 and the result is that speaking to the carbon trans- 

 mitter modulates the amplitude of the aerial current, 

 and therefore the amplitude of the radiated waves, 

 in accordance with the sf>eech wave-form. 



At the receiving station these electromagnetic waves 

 impinge on the receiving aerial and create in it very 

 feeble alternating, currents, which are a copy on a 

 reduced scale of those in the transmitting aerial. 

 These are then amplified by valves in cascade, 

 rectified, and sent through a Bell receiving telephone. 

 The result is that the latter emits sounds which closely 

 imitate the speech sounds made to the distant trans- 

 mitter. We require very high E.M.F. to create a 

 thermionic current of sufficient strength for wireless 

 telephony. This is now obtained by rectifying a high- 

 voltage low-frequency alternating current by a 

 Fleming two-electrode valve. 



The whole of the appliances are usually contained 

 in a small cabinet. .A ^-k\\ . radio-telephone set as 

 made by the Marconi Co. will work over 200 miles and 

 transmit speech perfectly. More powerful arrange- 

 ments on the same principle have telephoned from 

 Chelmsford to Rome. 



For aircraft radio-telephony it is usual to provide 

 a small high-tension dynamo driven by a wind-screw 

 to give the requisite direct high plate voltage. The 

 filament-heating currents are provided from small 

 closed storage cells. The aerial wire is a long trailing 

 wire about 250 ft. in length, which is unwound when 

 required from a drum. The actual valve apparatus 

 may be placed at any convenient place in the aero- 

 plane body and yet be controlled by the pilot or 

 observer from his seat. The mere act of taking 

 hold of the microphone transmitter closes a switch 

 which lights up the valves and throws over the aerial 

 wire into connection with the transmitting valve. 

 Such aircraft radio-telephones will operate over a dis- 

 tance of fifty miles or more. So sensitive are these 

 cascaded valve detectors that it is not even necessary 

 to use a long aerial wire at all. A very few turns of 

 insulated wire wound on a wooden frame, called a 

 frame aerial, connected to the receiver suffice to col- 

 lect and detect the electric jvave signals. 



Experiments were conducted in March, 1919, by 

 the Marconi Co. to ascertain the minimum power 

 required to transmit by these valve generators 

 articulate speech across the Atlantic during daylight 

 hours. The transmitting plant consisted of two three- 

 electrode generating valves, with a third control valve 

 for speech modulation. .\ small alternator of 25 kw. 

 power supplied an alternating current which was 

 stepped up in potential to 12,000 volts and rectified by 

 a two-electrode or Fleming valve. The reception was 

 by a series of six valves in cascade, with a final 

 detector valve. The speech transmission was per- 

 fectly good and clear across the .Atlantic, and so loud 

 at Chelmsford, five hundred miles away from Bally- 

 bunnion, Co. Kerry, that it could be heard on a simple 

 frame aerial. 



Before leaving the subject of radio-telephony it may 

 be remarked that, both in connection with it and with 

 the everyday uses of radio-telegrnphv in m.T-itime inter- 

 communication, there is a great demand for an effec- 

 tive wireless call-bell. I have recently devised a form 

 of call-bell which depends upon the use of a new tvpe 

 of four-electrode valve made as follows : .\ highly 



