the High-Tension Magneto. . 385 



points lie much more closely to a smooth curve, and they 

 could be repeated on successive trials very nearly within the 

 limits of observational error. The nature of the result was 

 thus entirely changed and observation brought much more 

 closely into concordance with theory. 



It was, of course, suspected at once that the great difference 

 was due to the fact that the external choking coil was free 

 from hysteresis and eddy-currents, and not merely to the 

 presence of the external coil ; but it was thought desirable 

 to establish the fact by substituting for the air-core coil a 

 coil of the same self-inductance wound on an iron core. 

 Fig. 12 B shows the results which were then obtained ; the 

 sharp minimum has not altogether disappeared, but it is 

 much reduced, while the whole curve is depressed some 

 12 per cent. Now the iron losses in the external coil must 

 have been very much less than they would be in a coil of 

 similar self-inductance constructed like the armature of a 

 magneto, for the core on which the coil was wound con- 

 sisted of a straight laminated bar and not of a nearly closed 

 magnetic circuit. Accordingly, the difference in the curve 

 introduced by the substitution of such a coil for one strictly 

 free from iron losses is sufficient to indicate that the main 

 reason for the complete absence of sharp minima in the 

 •curve taken without the external coil is to be found, as has 

 always been suspected, in the iron losses. On the other 

 hand, it is by no means clear immediately why the mere 

 addition of an external coil, even if it is free from iron 

 losses, should change so completely the character of the 

 results ; some further investigation into the matter seemed 

 desirable. 



The first step was to discover whether the damping of the 

 oscillations had been greatly diminished by the air-core 

 choking coil. The logarithmic decrement was found to be 

 reduced, but only by about 10 per cent., an amount much 

 too small to account for the great difference in the peak 

 potential. But at the same time a remarkable feature 

 appeared. The primary self-inductance without the choking 

 coil in series was '00623 henry, and it was accordingly 

 suspected that with the choking coil in series it would be 

 about *0072 ; it was actually found that the self-inductance 

 was only * 005 78, less than that of the armature without any 

 external coil. It is clear that the presence of the external 

 coil modifies profoundly the oscillations in the rest of the 

 primary circuit. 



Phil. Mag. S. 6. Vol. 37. No. 220. April 1919. 2 E 



