264: Prof. J. A. Fleming on the Paulsen Arc. 



Also that the time during which the electromotive force 

 can be removed without permanently extinguishing the arc, 

 is very much shorter in the first case than in the second. 

 Both these facts undoubtedly contribute to enable the pro- 

 duction of oscillations of a higher frequency to be obtained 

 with a carbon-metal arc in coal-gas, as compared with those 

 which can be reached with a carbon arc in air. 



In order that oscillations may be produced when a con- 

 denser circuit is shunted across the arc electrodes, the 

 condenser, during its charge and discharge, must cause a 

 variation of current through the arc, and this in turn must 

 cause a variation in the potential difference of the electrodes 

 in such direction as to assist the charge or discharge. Thus 

 when the condenser is robbing the arc of current and so 

 decreasing the arc current, this must automatically raise the 

 potential difference of the electrodes, and when the condenser 

 is discharging it must decrease the potential difference. The 

 greater this P.D. variation for a given current decrement or 

 increment, the greater will be the energy conveyed to the 

 condenser circuit. 



In the case of a carbon-carbon arc worked at 10 amperes 

 and say 60 volts arc P.D., the characteristic curve is rather 

 flat. Hence, if we employ a small capacity in the shunt 

 circuit we have very little variation of arc current and arc 

 P.D., and little or no energy in the condenser-circuit 

 oscillations. We can, however, obtain more energy by 

 using, as Mr. Duddell originally did, a condenser of rather 

 large capacity, say 1 to 5 microfarads. Moreover, the carbon 

 arc has a persistence which renders it irresponsive to very 

 rapid variations of current. Hence with a large current 

 carbon-carbon arc in air, we can only obtain rather slow 

 oscillations. 



If, however, we employ a carbon-metal arc, carbon nega- 

 tive and cooled copper, positive electrodes in hydrogen or 

 coal-gas, we then have an arc with a very steep characteristic 

 curve, and one which responds to exceedingly rapid variations 

 of current through it. We can, therefore, employ in the 

 shunt circuit a condenser of small capacity, and yet convey 

 to it a considerable amount of energy owing to the large P.D. 

 variation in the arc caused by a small arc-current variation. 

 Hence we can produce oscillations of high frequency. This 

 theory is confirmed by the observation that in the case of the 

 carbon-carbon arc in air for small currents, and in the case 

 of the carbon-aluminium arc, the characteristic curve is 

 nearly as steep as for large current carbon-metal arcs in 

 hydrogen ; and with these particular arcs, viz., the small 



