LIGHTNIN( 



-PEEK 



189 



the voltage evenly between the units as well as to direct the arc 

 away from the string. 



These tests arc of practical imi)ortance, since lightning voltages 

 higher than these rarely occur on operating transmission lines. The 

 lightning spark-over voltage was found to be twice the GO-cycle 

 spark-over voltage. The impulse ratio was increased by the shield. 

 The wet and dry lightning flashes clear the shielded string while 

 on the nonshielded string the flash cascades and is likely to rip off 

 the skirts. 



It will be noted that the three typical gap arrangements have dif- 

 ferent characteristics. Both the needle gap and the insulators 

 require a lightning voltage about double the 60-cycle voltage to 



220 

 200 

 180 

 160 



ii/ao 



</zo 



I 30 

 ^ 60 



ao 



20 



o V\ I I I I I 1 I I 

 o 2 a- 6 e ^o /z /-^ /6 /8 so ^^ ?* as sa jo jz 



S/oacing '■ Cms. 



Fig. 26. — Needle-gap spark-over curves for GO cycles and for impul.se 

 waves Nos. 1 and 2. Wave No. 1, single Lalf sine wave impulse; 

 wave No. 2, impulse with steep front, but wilh a long tail 



cause spark over, while the sphere spark-over voltage is the same 

 for 60-cycle and a wide range of impulse voltages. This character- 

 istic is of great practical importance, since it is desirable to have 

 an arrester gap that discharges at a low lightning voltage and to 

 design insulators that have a high lightning spark-over voltage. 

 The reason for this is time lag, and it seems w^orth while to discuss 

 it briefly. 



TIME LAG 



A fixed minimum voltage is required to spark over a given gap 

 when the time of application is not limited. Energy is necessar}'^ to 

 rupture gaseous, liquid, and solid insulation ; this introduces a time 

 element. 



