18 THE BELL SYSTEM TECHNICAL JOURNAL, J.\:XUARY 1954 



(a) (b) 



INITIAL CHATTER SHOCK CHATTER 



1 DIV= 10 MICROSECONDS 1 DIV = 100 MICROSECONDS 



Fig. 12 — Views of chatter obtained by oscilloscope. 



can be used to trigger the sweep at a later time so that any portion of 

 the chatter can be observed in detail. Typical measurements are shown 

 in Fig. 12. In each of the views, the horizontal lines starting at the upper 

 left represent open contacts, the lower horizontal lines represent closed 

 contacts, and any jumping between them indicates chatter. Fig. 12(a) 

 shows the relatively high frecjuenc}^ chatter which occurs immediately 

 after contacts close. This is called initial chatter and is caused by vibra- 

 tion of contact springs in their higher modes due to the impact of mating 

 contacts. Initial chatter differs in character from the lower freciuency 

 shock chatter shown in Fig. 12(b). Shock chatter is caused by ^^^re vibra- 

 tions induced by the shock of the armature striking the core. The time 

 of each reopen of the contacts and the time between opens is much longer 

 than for initial chatter. 



Many measurements of the type just described are made in the course 

 of a relay development. They enable the relay designer to relate chatter 

 performance to design characteristics of relays so that contact chatter 

 can be reduced or eliminated. Such measurements also allow one to 

 classify chatter performance according to the circuit application. 



The Measurement of Time — Mechanical Effects 



Electrical timing measurements previously described give data on 

 over-all effective performance; to understand these results one often 

 needs to measure displacement-time characteristics both alone, and in 

 relation to current versus time and flux versus time. The string oscillo- 



