260 Prof. J. A. Fleming on the Poulsen Arc as a 



valve stops all movement of electricity in one direction, and 

 in series with the valve is placed a telephone. On listening 

 to the telephone when the arc is in operation and is quite 

 steady, a crackling sound is heard, which is not sufficiently 

 regular to be called a musical note. If the oscillations in the 

 arc circuit were continuous and absolutely persistent, they 

 would induce similar secondary currents, and my oscillation 

 valve would separate out all the unidirectional constituents, 

 and send through the telephone a perfectly steady continuous 

 current. But this is not the case. The telephone yields a 

 sound which shows that the continuous current through it is 

 interrupted irregularly, and this can only be because the 

 oscillations in the arc circuit are interrupted. If the telephone 

 is not inserted directly in circuit with the valve, but is in- 

 serted in the primary circuit of an induction-coil, or air-core 

 transformer the secondary of which is inserted in series with 

 the valve, the sound in the telephone is still more increased. 



We have therefore strong indications that the arc method 

 of exciting undamped oscillations gives rise to irregular 

 groups of oscillations which are separated by short intervals 

 of time. 



In order to be successful at all in producing high frequency 

 oscillations by the arc method, I find it essential to pay 

 attention to a number of details. One important condition 

 of success is, as Mr. Poulsen has already pointed out, that 

 the carbon electrode should be kept in slow rotation. If 

 the carbon is not rotated, a bulbous deposit of carbon soon 

 grows on the electrode and builds up a beak or pimple of 

 carbon which makes the arc irregular and soon puts it out. 

 The end of the round carbon should be cut off square to 

 begin with, but it soon gets rounded off and a little corru- 

 gated or milled, and seems to work better when it has been 

 used for a short time. Hard arc carbons are better than soft 

 for producing oscillations. Secondly, as regards the gas. 

 Pure hydrogen works with difficulty. Coal-gas is better, 

 but the coal-gas cannot be used over and over again, as 

 it is altered in composition by the arc, and this seems 

 to create a difficulty in the use of the Poulsen arc as a 

 transmitter in wireless telegraphy on board ship. The 

 deposit of soot in the chamber is sometimes excessive. The 

 coal-gas must be passed through quickly, but not too quickly, 

 and my experience is that if the sides of the containing 

 chamber are not kept cool, the deposit of soot is greatly 

 increased. The more enriched the coal-gas is the worse 

 becomes this trouble. I find it necessary in any case to 

 clean out the chamber after a few hours of continuous use. 



