792 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1957 



Table II — Loss or Gain of Metal from Arcing for 



Palladium or Silver 



(in units of 10-" cc per erg) 



Inactive Arcs 



Anode Tj-pe . . . 



Cathode Type. 

 Active Arcs 



Anode Type . . . 



Cathode Type. 



Cathode 



4 gain 

 1 loss 



(loss) 

 4 losst 



Anode 



4 loss 

 1 gain 



10 loss* 

 (loss) 



mound and pit 

 mound and pit 



smooth erosion 

 smooth erosion 



* This high figure refers to arcs on closure at very heavily carbonized surfaces; 

 for lightly carbonized silver surfaces the anode loss is less and most of the metal 

 is transferred to the cathode. 



t This figure refers to arcs at closure of palladium surfaces. The rate of cathode 

 loss at break of palladium surfaces is significantly less, as pointed out in Section 

 1.4(a) ; and in cathode arcs at active silver surfaces the rate of loss is still less. 



* This applies to arcs at closure. Between separating electrodes, air break- 

 down often occurs when the electrodes are too close together for air break- 

 down over the shortest path. Under such conditions, arcs between silver surfaces, 

 which are initiated by air breakdown, can become anode arcs, and the transfer 

 resulting from such arcs gives dominant anode erosion. 



breakdown, all inactive arcs at silver surfaces are of the anode type, and 

 active arcs of the anode type occur for silver only. These facts are tabu- 

 lated for reference in Table III. All of them are at once predictable from 

 the values of the critical distances for palladiimi and silver, and from 

 knowledge of the way in which breakdown distance is changed by activa- 

 tion.* 



* The difference between the transfer behavior of palladium and silver elec- 

 trodes in the active condition suggested to R. H. Gumle}- that the damaging ef- 

 fects of activation can be greatly reduced by constructing a relay with negative 

 contacts of silver and positive contacts of palladium. He tried out this idea and 

 found it to be effective. In the absence of activating vapor, a relay in which the 

 negative contacts are silver and the positive contacts palladium has no merit 

 over a relaj^ in which all contacts are palladium, but when vapor is present, the 

 erosion can, under some circumstances, be much reduced by replacing the nega- 

 tive palladium contacts by silver contacts. 



