ARCIXG OF ELECTRICAL CONTACTS IN TELEPHONE SWITCHING CIRCUITS 555 



Iniiiaiion VoUcujc of Glow Discharge 



The cant lie vor bar setup previously used for similar measurements 

 of arc initiation voltage as function of separation" was used iiere. By 

 varying the separation the corivsponding glow initiation voltage was 

 measured. For each separation a measurement was also made of the 

 arc initiation \'oltage. The results are given in Table III. The results 

 indicate that both arc and glow are initiated at the same voltage for 

 the same separation. One may, therefore, conclude that at least the first 

 few steps invoked in the process of the breakdown are the same whether 

 theij lead to a glow discharge or to an arc. In many cases, it was observed 

 that the arc was preceded by a period of glow discharge. This was not 

 found, however, to be the general case as discussed in the following sec- 

 tion. 



Table III — Glow and Arc Initiation Voltages as Functions of 



Contact Separation for Pd Contacts in Dry 



Atmospheric Air at 25 °C. 



Glow-arc Transition 



The experimental setup used is shown in Fig. 13. By systematic varia- 

 tion of the circuit parameters Vo, L and C, a variety of transients 

 was obtained and recorded. Samples of typical cases are shown in Fig. 14. 

 For transient stability and reproducibility, it was found necessary to 

 exercise extreme care in securing good contact surface cleanliness and in 

 maintaining it during the experiment. The presence of organic vapors, 

 humidity, films of grease or oil, fingerprints, etc., usually led to erratic 

 results. The general effect was an inclination towards more arcing and less 

 glow discharge. Only by proper cleaning of the contact surfaces and allow- 

 ing the contact to arc heavily for about 20 minutes was it possible to 

 obtain fairly reproducible results. Table IV shows a summary of results 

 obtained from one of several sets of experiments performed. 



Before stabilization of the transient, it was generally observed that the 

 glow period was first short then gradually increased until it reached a 

 limiting value which it did not exceed. These limiting values are given 

 in column 5 as fractions of the half period iriLCf'. They range from zero, 

 actually glow was not detected with a time resolution of 1 per cent of the 



