[ in ] 



A VIBKATING-FLAME EECTIFIER FOR HIGH-TENSI 



CUEEENTS. 



Bv JOHN J. DOWLING, M.A., F. Inst. P. ; 



AND 



J. T. HAEEIS, B.Sc. 



[Read December 21, 1920. Published February 21, 1921.] 



Incomplete rectification of an alternating high-tension current can be effected 

 by placing a small flame near one electrode of a spark-gap in the circuit. The 

 unilateral conductivity of this arrangement is due to the large supply of 

 negative carriers of high mobility derived from the flame. It occurred to one 

 of the authors that, by causing the flame to vibrate in synchronism with the 

 pulsations of electromotive force, a supply of ions would then be made avail- 

 able when the e.m.f. was directed in one sense through the circuit, while 

 there would be practically none available to carry the reverse component of 

 current. Almost perfect rectification would be obtainable in this way. After 

 a few trials, a satisfactory arrangement was devised, the simplicity and novelty 

 of which seem to warrant this description. 



From the foregoing outline of the principle it will be clear that the flame- 

 vibrations must be strictly in phase with the high-tension electromotive force, 

 and must so occur that the flame rises and falls once, while the electromotive 

 force makes one complete alternation. These conditions are fulfilled in the 

 following way : — The flame-vibrations are produced by a Koenig's manometric 

 capsule, the membrane of which is either a thin iron plate or an india-rubber 

 diaphragm, to the centre of which a small piece of iron is attached. Mounted 

 close to this is an electromagnet supplied with two independent windings. 

 Through one of these windings the full primary current of the high-tension 

 transformer passes. The other winding carries a direct current from a 

 storage battery, a suitable rheostat being introduced for regulating the 

 current, as described below. 



The alternating current alone would cause a pulsation of the manometric 

 flame to occur twice in each complete alternation, since the manometric disc 

 would experience an attraction in whichever way the magnet happened to be 

 magnetized. If the direct current in the other magnet winding be now regulated 

 until its magnetizing effect is almost equal to the maximum value of that due 

 to the a.c. winding, it is clear that the magnetization will no longer undergo 

 periodic reversals, but will, on the contrary, vary between two points on the 



gOIENT, PBGC, R.D.S., \0h. XVI, NO. XVI- 2 C 



