44 BELL SYSTEM TECHNICAL JOURNAL 



tween 30 and 250 volts while the final direct current is limited by 

 circuit resistance to less than 0.5 ampere. At the low voltage, observa- 

 tions of the current and voltage transients indicate that the closing 

 contacts merely discharge the line. As the voltage is raised so that 

 the current surge peak is in the range between 0.5 ampere and 1 ampere 

 the reopenings, due presumably to overheating of the contacting areas 

 by the discharge, are observed. These become more frequent as the 

 voltage and current increase, and a new type of current surge begins to 

 appear. This is placed on the time axis ahead of the point at which the 

 initial closures have been occurring (usually 5 to 10 microseconds 

 earlier) and consists of one or more irregularly spaced heavily damped 

 pulses of current lasting only a small fraction of a microsecond and 

 evidently discharging only a minute amount of the energy stored in 

 the system. They occur perhaps once in a hundred closures at 30 volts, 

 nearly every closure at 100 volts, and several for every closure at 250 

 volts. It is believed that the transients observed indicate the forma- 

 tion of minute metallic bridges ^ between the approaching contacts due 

 to a softening of the metal by a cold point discharge and its deformation 

 by the static field, and that once formed they are exploded by the 

 discharge of current from the relay structure and adjacent wiring. 

 The high fields necessary for phenomena of this type are of course 

 due to the minute distances as the contacts approach final closure. A 

 good deal of the erosion on telegraph relay contacts operating on 

 capacitative loads or shunted by resistance-capacity "spark- killer" 

 circuits is probably due to these "preclosures," but they are not 

 thought to be of much importance at the lower battery voltages of the 

 telephone plant. 



Having now described the phenomena as contacts close in a doubt- 

 less over-simplified manner, we may consider that they have been 

 closed for a long time, the direct current and the magnetic field of the 

 load relay are established and the contacts are to be separated. The 

 action now becomes really complicated and much of it is as yet only 

 surmised. Several different things may happen, and these are influ- 

 enced by humidity, dirt, surface films, absorbed gases and many other 

 factors, including the speed of contact separation, the roughness of the 

 surfaces, and the presence or absence of a wiping motion as well as the 

 physical properties of the contact materials. 



If the steady current exceeds certain well-known values ranging be- 

 tween 0.4 ampere and 1 ampere, characteristic of the contact materials, 



2 "The Formation of Metallic Bridges between Separated Contacts," G. L. 

 Pearson, Phys. Rev., Sept. 1, 1939, Vol. 56, pp. 471-474. 



