888 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1954 



supply excessive energy because of the large amounts of material that 

 are then burned off, and objectionable weld flash is produced. Probably 

 the major problem is to control the energy supplied by the circuit or, 

 indirectly, the control of arc duration. 



DURATION OF ARC 



In this section the variables affecting arc duration ^xi\\ be discussed. 

 During the arcing period the gun moves essentially at constant speed. 

 The arc time may then be said to be equal to the distance traveled after 

 arc initiation divided by the gun velocity. 



A . Initial Separation 



The voltage at which the arc is initiated is primarily a function of the 

 separation between the two electrodes. A series of static voltage break- 

 down tests was made in order to define the distribution of initiation 

 separation under conditions to be expected in production welding of a 

 block to a wire. The block material w^as 70-30 per cent cupro-nickel. 

 The wire was 0.040-inch diameter silicon copper. Tests were taken with 

 the wire end flat or terminated in a 60° conical point while the block 

 surface was maintained flat. Industrial contamination as may very well 

 be present in a productiln machine was simulated by the addition of a 

 thin oil film on each of the opposing surfaces. 



The results are summarized in graph form in Fig. 2. Plotted are the 

 three a limits* for the conditions indicated. Better arc initiation separa- 

 tion control is obtained with flat as compared ^^^th pointed wire ends, and 

 clean as compared ^^^th oil contaminated mre ends. The ratio of maxi- 

 mum to minimum arc initiation separations to be expected is consider- 

 ably lower for high voltages than low voltages. This fact alone makes 

 operation at voltages in excess of 1,000 volts desirable. 



In addition there exists an initiatory time lagf between the time that 

 the separation reaches the static breakdo^^^l value and the moment of 

 actual initiation. Arc duration variations due to this phenomenon are 

 reduced by an increase in applied voltage. 



B. Evaporation of Material 



The arc does not cover the w^hole surface but is concentrated on a 

 small area which is being heated, therefore, at a rate considerably'" in 



* All but three out of 1,000 welds are expected to fall within these limits. It 

 is to he noted that in view of the high relial)ility often required of this type of weld, 

 conditions even further removed than three a limits may have to be considered. 



t Field Emission of Electrons in Discharges by Llewellvn, Jones and E. T. de la 

 Perrelle, Proc. Roy. Soc. A, 216, p. 267, 1953. 



