CONTACT PHENOMENA IN TELEPHONE SWITCHING CIRCUITS 43 



and other characteristics of the wire, but it is (in the telephone plant) 

 rarely lower than 500,000 cycles, and on short leads it may be many 

 megacycles. Fortunately most contacts are not much affected by 

 closing a half ampere. Erosion and build-up will occur, but at rather 

 slow rates, and they are usually completely obscured by effects due to 

 the contact opening. Of course, if the contacts bounce, ^ the effect 

 will be more complex, but we are assuming for the moment that they 

 do not bounce. The structure of the relay itself, including the pair of 

 springs separated at its base by an insulating sheet, is also an oscillating 

 circuit. We have not been able to get inside of this circuit and meas- 

 ure the current surge but its oscillation frequency seems to be about 250 

 megacycles for certain telephone relays. 



Now suppose that the simple circuit of our closing contacts is compli- 

 cated by an additional wire connected to the contact spring terminal. 

 This is also charged to 50 volts before the contacts close, and being a 

 second circuit of a hundred ohms surge impedance in parallel with the 

 original wire, the current peak discharged through the contacts will 

 now be about one ampere. But now the contacts are likely to act 

 differently. About a microsecond after the current reaches its peak, 

 but before the charge in the wires has been completely dissipated, 

 the circuit is interrupted and the discharge stops. A spark, which is 

 visible in the microscope, suggests that the current carrying areas have 

 been exploded and blown apart. A few microseconds later they again 

 close and the rest of the energy is discharged, but some of the contact 

 metal must have been destroyed. 



If several "idle" wires are attached to the contact, the current surge, 

 and the number and duration of the contact reopenings, increase, but 

 not usually in direct proportion to the number of wires. If the idle 

 wires are attached to the load relay winding terminal instead of to the 

 contact, the current is smaller, as the length of single wire from relay 

 winding to contact is effectively in series with them. 



In the telephone relay circuits which we are considering, the steady 

 state current plays little part during contact closing if the contact carry- 

 ing relay is properly adjusted, as the contacts come to rest while the 

 current is still held at a small fraction of its final value by the inductance 

 of the load relay winding. 



Under some conditions which are more likely to occur in telegraph 



than in telephone circuits the contact closure phenomena are somewhat 



different from those described above. Assume, for example, that the 



potential between the open contacts may be adjusted in a range be- 



' Bounce, as distinguished from chatter, reopens the contacts after several thou- 

 sandths of a second. 



