MEASUREMENT OF POTENTIAL DIFFERENCE 263 



before the passage of the 

 spark that the dielectric 

 strength of an air gap is 

 very dependent on the hu- 

 midity. This is illustrated 

 by Fig. 148. 



The density of the air af- 

 fects the results obtained 

 with any form of spark gap. ~. 130 



The irregularities of the I 120 

 needle-point gap have 

 proved so troublesome that 

 a substitute for this form of 

 gap has been sought. For 

 voltages above 50 kv. the 

 use of two spherical elec- 

 trodes of equal diameters is 

 now recommended by the 

 American Institute of Elec- 

 trical Engineers. The ad- 

 vantage is that if the dis- 

 tance between the electrodes 



5 10 15 20 25 30 35 40 45 50 55 60 

 Spacing in cm 



FIG. 148. Showing effect of humidity 

 is less than three times the on the action of a needle-point spark 



radius of the spheres, the g 



corona does not form before the gap breaks down. The erratic 



0.35-t w 



.0.35.1 



T 



i 



\, 

 ?( 



-i.t 



( -Q.28A 



4.07^ >| 



Note; 



A variation of 1 cm. in thickness and 

 width of woodtu parts is permissible 



l''n;. 149. Spark gap with spherical 

 electrodes. 



effects due to broken down 

 air around the electrodes are 

 thus avoided. Humidity has 

 no influence on the results. 



An additional advantage is 

 that the length of the gap is 

 much reduced; with spheres 

 25 cm. in diameter, the spark- 

 ing distance for 200 kv. is 

 about 13 cm. against about 

 52 cm. for the needle-point 

 gap. 



The sphere gap is especi- 

 ally adapted for measuring 



