DoWMNG AND Haeeis — A Vibrating -Flmnc Rectifier. 173 



the flame ; and a rotary mirror was employed to observe the flame and the 

 image of the tube. On exciting the coil, the vibration maxima of the flame 

 were seen to be strictly in phase with the discharges in the tube ; and, on 

 adjusting the polarizing current so as to cut out the intermediate maxima, 

 the corresponding vacuum tube discharges likewise ceased. The milammeter 

 then registered about 20 milliamperes. 



With a view to testing the degree of rectification, a suitable oscillograph 

 not being available, the following experiments were tried : — An a.c. milam- 

 meter was connected in series with the d.c. instrument already used, and the 

 observations repeated. A rather larger current was indicated by this a.c. 

 instrument, which seemed to point to incomplete rectification. However, an 

 interesting explanation was found for this. Tested in series with d.c. currents, 

 the two instruments read the same ; but, having doubts as to the validity of 

 such a test, we tried them again in series with current rendered intermittent 

 by a tuning-fork interrupter (50 -). It was found that the a.c. instrument 

 read nearly twenty per cent, higher under these conditions. As the condi- 

 tions-are very similar in this test to those obtaining during the rectification 

 experiment, and as the observed discrepancy there is exactly accounted for by 

 this effect, we feel justified in assuming that the inverse current must certainly 

 have been less than- -^ milliampere, i.e. less than i per cent, of the rectified 

 current. . 



Our next inquiry was directed to the possibility of using such flame- 

 i"ectifiers for the purpose of charging condensers to high steady potentials 

 from an 'a.c. supply; For this purpose it is essential that there should be aji 

 exceedingly -small inverse current, or, in other words, the resistance of the 

 spark-gap -must be .very high during the intervals when the flame has dropped 

 to its minimum; were this not so, the condenser would discharge with each 

 reversal of the e;m-.f., and might even take up reverse charges. - 



To attain satisfactory conditions for a- test it is necessary in this case to 

 introduce two rectifying devices in the circuit, one on each side of the con- 

 denser, for if only one were used there would be an oscillating potential on 

 that plate of the condenser which is connected directly to the coil. For the 

 purpose of maintaining a steady, average high potential above earth, only one 

 rectifler is required, the other plate of the condenser and the corresponding 

 coil-terminal being earthed ; but under these conditions the readings of an 

 electrostatic voltmeter connected ili parallel with the condenser may be mis- 

 leading. We therefore arranged two rectifying devices, with their a.c. coils in 

 series, and having separate rheostats for controlling their polarizing currents. 

 The two were, of course, arranged to pass high-tension current in the same 

 direction, and were placed respectively in the leads from the coil-terminals to 



