Potential of an Induction- Coil at " make." 333 



break, the maximum value indicated being approximately 

 17500 volts. 



In fig. 2 the four layers of the primary coil were all in 

 series, and the maximum potential at make is reduced to 

 14000 volts. 



In the case of fig. 3 the air-core coil (0*07 henry) was con- 

 nected in series with the primary, the layers of which were 

 still in series. The maximum potential at make is now reduced 

 to only 9600 volts. In this case the coefficient of coupling 

 has approximately the value which, when the primary capacity 

 is adjusted to its optimum value, allows the maximum ratio 

 of conversion of magnetic into electrostatic energy at break. 

 The make-potentiai of 9600 volts would scarcely suffice to 

 generate any inverse current through an X-ray tube, and it 

 may be remarked that in the case of fig. 3 the make-potential 

 would not be appreciably increased if the primary resistance 

 were much reduced for the purpose of increasing the potential 

 at break. 



As to the absolute value of the make-potential given by 

 the expression (15), if the resistances of the circuits are 

 taken at their steady-current values, the calculated value of 

 Y 2m is rather greater — about 15 per cent, in the present 

 experiments — than the value found experimentally. The 

 principal reason for this is undoubtedly that the value of Ai, 

 the damping factor of the oscillation, is greater than that 

 calculated from the steady-current resistances. The oscil- 

 lation also undergoes decay in consequence of core losses. 



The effect of one other variation of the system may be 

 briefly referred to. According to (15) the secondary capacity 

 has only an indirect influence on the spark-length at make. 

 If this capacity is increased — e.g. by connecting a leyden-jar 

 to the secondary terminals — the period of the oscillation, and 

 with it the time t 1 of rise to maximum potential, is lengthened. 

 The damping forces have therefore a longer time in which to 

 act, and a reduction of V2m is to be expected. This reduction 

 will be but slight if ky and 8 are small, that is, if Ri/Lj is 

 small. The effect might be expected to be greater if Ri/Li is 

 large. These conclusions are borne out by measurements 

 of the spark-length at make with and without a secondary 

 condenser. Thus, with the primary layers in series 

 (E = 82 volts, R 1 =16 ohms, Ri/Li = 110), connecting a 

 0*001 mfd. jar to the secondary terminals (which causes a 

 twenty-fold increase of C2) only reduced the make-spark 

 from 3*8 to 3'4 mm. When Ri was increased to 41 ohms 

 the addition of the jar reduced the spark from 3*5 to 2*3 mm. ; 

 when Ri was 82 ohms the reduction was from 3 to 1*5 mm. 



