720 



NATURE 



[August 5, 1920 



exhausted glass bulb contains a straight filament of 

 tungsten, which is rendered incandescent by a 6-volt 

 battery. Around the filament are arranged four 

 narrow curved metal plates having their curved sides 

 facing the filament and very near to it. Each of these 

 plates is carried on a wire sealed through the glass 

 bulb. The plates are arranged round the filament, as 

 shown in Fig. 3. 



Two of these plates on opposite sides of the fila- 

 ment, viz. 3 and 4 (see Fig. 3), are called the potential 

 plates, and the other two the collecting plates. The 

 collecting plates are joined together outside the bulb 

 and connected to the positive terminal of the filament- 

 heating battery, and a galvanometer G or telegraphic 

 relay is inserted in that circuit. The electronic ernis- 

 sion from the filament then creates a current which 

 fiows through the galvanometer or the relay, as in the 

 Edison experiment. If the two other plates have a 

 small potential difference made between them, either 

 of constant direction or else a high-frequency alter- 

 nating difference, this suddenly reduces the thermionic 

 current. The potential difference of the potential 

 plates introduces a new electric force into the field 

 which deflects awav the electrons proceeding from the 

 filament and prevents them from reaching the collecting 



Fig. 3.— 



• 3-— Fleming four-anode valve, i and 2 are the 

 collecting plates. 3 and 4 are the potential or 

 deflecting plates. B is the filament-heating battery, 

 and the central dot is the end-on view of the 

 straight filament. G is a relay or galvanometer. 



plate. If, then, we connect the potential plates to the 

 ends of a resistance of about 15,000 or 20,000 ohms, 

 and include this resistance in the plate circuit of an 

 ordinary three-electrode valve, the thermionic current 

 of the latter flowing through the resistance will create 

 a terminal potential difference which arrests the 

 thermionic current of my new valve. Hence the relay 

 does not operate. If, however, we give an extremely 

 small negative potential to the grid of the three- 

 electrode valve, then this reduces the thermionic cur- 

 rent of the latter and increases that of the other valve, 

 which again in turn causes the relay to close contact, 

 and it may be caused thereby to ring a bell. The 

 negative grid potential can be derived from the oscil- 

 lations in an aerial wire as above described. In this 

 manner I have constructed an arrangement by which 

 the ordinary feeble antenna oscillations can be em- 

 ployed to ring a call-bell. The operator can then 

 switch over the aerial to an ordinary valve receiving 

 set and listen to the telephone. 



It remains to say a few words on the methods 

 by which the thermionic valve is employed in the 

 reception of signals made by undamped or continuous 

 waves. By far the best method of receiving signals 



NO. 2b49, ^f L. 105] 



by these waves is by the so-called beat-reception. If 

 two sets of waves of slightly different wave-length 

 are superimposed, no matter what sort of waves they 

 may be, the result is to produce a compound wave 

 with periodically increasing and decreasing amplitude. 

 These augmentations are called the beats. 



If a continuous electric wave falls on an aerial it 

 creates on it continuous oscillations. Suppose, then, 

 that we generate also by some local means in the 

 aerial wire undamped oscillations differing in fre- 

 quency, say by 1000, from the incident waves. The 

 result will be to produce in the aerial electrical beats 

 having a frequency of 1000. These act to a receiver 

 just as do damped trains of waves with a train 

 frequency of 1000. They can be rectified and detected 

 by a valve and telephone, as already explained. It is 

 now quite easy to produce high-frequency oscillations 

 of any required periodicity by coupling a three- 

 electrode valve to the aerial and then coupling the 

 grid and plate circuits of the valve. Sometimes a 

 separate three-electrode valve is used to rectify and 

 detect the beats. Capt. H. J. Round has, however, 

 invented ingenious methods by which one and the 

 same thermionic valve can be used simultaneously to 

 generate and to detect the beats. 



We must, in the last place, glance at the uses of 

 the thermionic valve in connection with ordinary tele- 

 phony with wires. When the rapidly fluctuating elec- 

 tric currents which are propagated when a speaker at 

 one end of a long line converses by telephone with 

 an auditor at the other flow along a copper telephone 

 line, two effects take place which militate against 

 clear and audible speech transmission. First, the 

 current generally is enfeebled as it flows, and this is 

 called the attenuation. Secondly, the different har- 

 monic constituent currents which go to make up the 

 complex wave-form which corresponds to each articu- 

 late sound are differently enfeebled. 



The vibrations of high pitch are more enfeebled 

 than those of lower pitch. The first effect reduces the 

 loudness of the speech received, and the second its 

 articulate clearness or quality. The cause of the 

 general enfeeblement is the resistance of the line, 

 which fritters away the energy of the speech electric 

 currents. Until lately the only known method of 

 overcoming it was by putting sufficient copper into 

 the line, but this, of course, means cost. 



The thermionic valve is, however, able to make a 

 very large economy in copper. It has already been 

 explained that the three-electrode valve can act as 

 an amplifier. Suppose, then, that we cut a long 

 telephone line in the middle and insert on one side a 

 transformer, the secondary terminals of which are 

 connected to the grid and filament of a valve, whilst 

 the plate circuit also contains a battery and a trans- 

 former of which the secondary circuit is in connection 

 with the continuation of the' line. Feeble telephonic 

 currents arriving at the valve would vary the potential 

 of the grid, and this, as just explained, would fluctuate 

 in lil<e manner, but with increased energv, the plate 

 current. The transformer in the plate circuit would 

 then re-transmit the speech current, but with exalted 

 amplitude. The valve can thus be used to counteract 

 the effect of resistance on the Ijne. In practice, how- 

 ever, the arrangements are a little more complicated, 

 because a telephone line has to be used in both 

 directions. 



If our trunk telephone line system in Great Britain' 

 had to be laid over again, it is perfectly certain that 

 a very great economy in copper could be made by a 

 widespread use of the thermionic valve as a repeater 

 and relay. It repeats so perfectly that we may cer- 

 tainly say it has completely outclassed all previously 

 invented forms of microphonic relay. 



