264 



NATURE 



[October 31, 1912 



thin sheet zinc or tin, the plates being immersed in 

 a metal or stoneware box of oil. 



In the case of some high-power stations, Mr. 

 Marconi employs large air condensers consisting of 

 sheets of metal hung up on insulators in a room. At 

 Nauen and at the Eifi'el Tower stations tubular or 

 plate-glass condensers are used. 



The condenser is charged by the source of electro- 

 motive force to a high potential, and then discharged 

 across a spark gap, with or without oscillations, arid 

 this discharge passes through a coil which niav be 

 one coil of a two-coil transformer, the secondary being 

 inserted in the circuit of the antenna, or else a single- 

 coil transformer, then called an auto-transformer, may 

 be made to do duty for the two separate coils in the 

 circuits of the gntenna and the storage condenser. 



.\n important element is the spark gap. In earlv 

 days when only small powers were employed, this 

 consisted simply of two stationary brass balls. When 

 large power first began to be applied, as at the Poldhu 

 Station in 1901, it was soon found that the oscillatory 

 discharge started an electric arc across the balls which 

 had to be e.xtinguished before the condenser could 

 again become charged. Also the balls became rapidly 

 worn away. To remedy these defects, various inven- 

 tions were introduced. An air blast was applied to 

 the spark gap to quench the arc. 



I devised for the Marconi Co. in 1902 a discharger 

 with revolving balls or discs driven by an electric 

 motor which overcame some of the difficulties, and 

 this type of slowly rotating disc discharger using low- 

 frequency sparks was used for some considerable time 

 at Poldhu. 



Later on Mr. Marconi invented his high-speed 

 studded disc discharger which is far more efficient, 

 and creates a quenched musical spark of the required 

 character. In this discharger a steel disc having studs 

 on it revolves at a high speed between two other 

 revolving electrodes and the passage of the studs 

 starts a condenser discharge in which any true arc 

 is instantly quenched. The kind of discharge required 

 for effective work is one in which rapidly repeated, 

 strong, highly damped discharges take place in the 

 primary condenser circuit, and these excite prolonged 

 trains of free oscillations in. the antenna. This is 

 only possible if any true arc discharge in the primary 

 circuit is entirely prevented. 



This is also achieved by the Wien or Telefunken, 

 the Peukert and Von Lepel dischargers consisting of 

 flat metal plates in close proximity. In these dis- 

 chargers the discharges succeed each other with great 

 regularity and at the rate of several hundred per 

 second. When the condenser circuit is properly tuned 

 to the antenna circuit and coupled to it not too 

 strongly (with about 20 per cent, coupling), we have 

 powerful intermittent oscillations set up in the 

 antenna, each group being very feebly damped and of 

 uniform oscillation frequency. These rapidly suc- 

 ceeding groups of oscillations are cut up into groups 

 of groups in accordance with the signals of the Morse 

 alphabet by means of a key placed in some part of the 

 circuit. Although nearly all the radio-telegrapfiy in 

 the world is now conducted by means of these inter- 

 mittent condenser oscillations, great efforts are being 

 made to perfect suitable high-frequency high-power 

 alternators, and the advent of a commercial machine 

 of this kind will no doubt make it a formidable rival 

 to the existing methods. 



Deferring for the moment the consideration of what 

 takes place in- the space between the sending and 

 receiving antenna, we may complete our description 

 of the receiving apparatus. 



In the sending antenna we have very powerful high- 

 frequency currents at the base and high potentials at 

 tlae free or upper end. Even in small stations the 

 NO. 2244, VOL. 90] 



sending antenna current maj' have a value of 5 to 10 

 amperes, whilst in large stations the antenna current 

 at the earthed end is 50 to 100 amperes, and large 

 enough to raise to incandescence quite large rods of 

 arc light carbon. 



There is, therefore, a considerable expenditure of 

 power on the antenna. A part of this is spent in 

 heating the antenna, but a large proportion is radiated. 

 Nevertheless, the over-all efficiency of the usual wire- 

 less telegraph transmitter using the ordinary un- 

 quenched condenser spark, meaning by that the ratio 

 of power radiated from the antenna to power supplied 

 by the operating dynamo or battery, is at present prob- 

 ably not more than 20 per cent, to 25 per cent, in actual 

 practice, though much higher efficiencies, even up to 

 75 per cent., have been claimed for the quenched-spark 

 system. But the evidence for these high efficiencies 

 is somewhat imperfect. 



An extremely small fraction of the whole radiated 

 energy is picked up by the receiving antenna. In 

 this latter we have currents created which are 

 measured in micro-amperes, or, at best, in fractions 

 of a milliampere. If the receiving antenna is properly 

 tuned to a closed condenser circuit inductively coupled 

 to it, the energy picked up by the receiving antenna 

 accumulates in the associated condenser circuit. 



In this last we then have feeble currents circulating 

 which imitate in mode of variation those of the distant 

 transmitting antenna. To detect them, it is now 

 most usual to employ a telephone in series with some 

 form of current rectifier, which is shunted across the 

 condenser in the closed secondary receiving circuit, 

 or else some form of current-operated detector, such 

 as Marconi's magnetic detector, which is placed in the 

 condenser circuit. 



If we merely connect a telephone across the con- 

 denser circuit, no sound will be produced in it, because 

 the frequency of the current oscillations in the receiv- 

 ing condenser circuit is too high to affect a telephone. 

 If, however, we insert some device in series with the 

 telephone which acts like a valve, it will rectify the 

 groups of oscillations into prolonged gushes of elec- 

 tricity in one direction, which, coming at the rate of 

 the much lower spark frequency, say, about 500 or 

 1000 per second, create in the telephone a shrill sound. 

 ."Xs these groups are interrupted at the sending station 

 in accordance with the Morse signals, the receiving 

 operator hears long or short musical sounds which he 

 can interpret into the letters of the alphabet. 



Amongst the rectifiers much used, my own oscilla- 

 tion valve invented in 1904, or glow-lamp detector, is 

 an interesting example. It consists of a little electric 

 glow lamp, having a metal plate or cylinder sealed 

 into the glass bulb. When the filament is incandescent 

 the space between the filament and the plate has a 

 unidirectional conductivity, and will allow negative 

 electricity to pass from the filament to the plate, but 

 not in the opposite direction. 



-Another large class of oscillation rectifier.s are the 

 crystal and contact rectifiers, the first of which, viz 

 carborundum, was discovered by Dunwoody, others 

 by Pierce and Pickard. Thus, for instance, a copper 

 point pressed against a small mass of molybdenite 

 is a good rectifier. Also the minerals chalcopyrite, 

 zincite, bornite, anatase, and hessite possess similar 

 properties, and a very sensitive rectifier is made by a 

 slight contact between two small masses of zincite 

 and bornite. Another rectifier is the galena-plumbago 

 rectifier. Also a gold point pressed very lightly against 

 an artificial surface of iron pyrites (ferric disulphide) 

 makes an excellent detector. 



In spite of much valuable work done by Prof. G. W. 

 Pierce, G. W. Pickard, and others the action of these 

 crystal rectifiers is by no means fully elucidated. It 

 appears not to be thermoelectric, since in general the 



