the High- Tension Magneto. 297 



the slower component, and to each point on the curve in that 



figure has been applied the correction appropriate to the 



slower component only. That is to say, it has been assumed 



that the wave-form is V 2 = A sin 2irnt . e~ M , so that the 



reduction in the peak voltage due to damping is given by 



* 6 2ir 



the factor sin . e ^, where tan 6— . This procedure is 



strictly accurate only when Ci is very small or very large. 



The resulting curve is shown by the dotted line in fig. 4. 

 The ratio of observed to calculated peak potential is now 0*78 

 at Ci = and increases regularly to 0'88 at C x = 4 (which is 

 not shown on the curve). Accordingly most of the dis- 

 crepancy can be accounted for by the damping, but not all. 

 It is, of course, hardly to be expected that any simple cor- 

 rection of this kind could bring theory and experiment into 

 complete accord ; for, as has been emphasized before, the 

 actual damping which produces so great a reduction of the 

 peak potential is wholly different in its nature from that 

 contemplated by the theory. 



Some Further Experiments. 



9. Similar experiments were made on the armature of the 

 same machine when it was completely removed from its 

 housing, so that the only iron in the neighbourhood of the 

 circuits was the core on which they were wound ; the un- 

 laminated pole-pieces could no longer exert any influence. 

 It was thought that the iron losses would be greatly reduced, 

 and that better agreement between theory and experiment 

 might be found. 



By exactly the same method as before the following values 

 were found for the constants of the armature: — 



1^ = 0-00045 henry; L 2 ' = 15-3 henry; T 2 = 0000174 sec; 

 c = 0'101; c' = 0"043: L 21 = 0"282 henry; /=1037. i is 

 again taken =0*5 amp. 



In fig. 6 the peak potential calculated from these values 

 is plotted against Cj in the full curve, the experimentally 

 determined values being again given by the points ©. 



The damping coefficients were not determined so completelv 

 as before^ but the following values were obtained: — 



T =0-000175 fi (obs.) =1-20 fi (calc.J =0-0124 



0-00114 1-10 0-0650 



0-00161 0-84 0-0922 



t' = 0-000232 fi' (obs.) =0-76 /(calc.) =0*296 



