470 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1952 



jorm of a hysteresis loop whose area varies approximately in proportion 

 to the square of the strain amplitude. This region is important for relays 

 for by introducing damping, long repeated vibrations — which are respon- 

 sible for considerable wear — are quickly brought down to the low wear, no 

 gross slide region with a corresponding reduction in wear. The mechan- 

 ical resistance associated with the stress strain loop is of the same type that 

 occurs in an assemblage of granular particles such as in a telephone trans- 

 mitter where the ynotion is small enough so that no gross slide occurs. 



I. INTRODUCTION 



In obtaining long life in telephone equipment such as relays, switches, 

 selectors and other mechanical devices subject to large momentary 

 forces, one of the main problems is the wear encountered in various parts. 

 This is particularly true in such small motion devices as relays where 

 even a few mil inches of wear increases the distance that the armature 

 has to travel and may eventually cause the relay to fail to make contact. 

 To obtain a design objective of one billion operations requires a very 

 careful minimizing of deleterious forces and a careful selection of the 

 best wearing materials. 



As a step toward investigating this problem several new techniques 

 have been de\'ised for measuring normal and tangential forces and for 

 producing and controlling normal and tangential motions for wear 

 studies. These methods have been applied to relays and have given 

 considerable information on the types of motion to be avoided and on 

 the best types of materials to select for various parts of the relay to ob- 

 tain long life. Specifically they have shown that normal forces cause 

 very little wear and that tangential sliding of one part over another is 

 the principal cause of wear. Fortunately, by designing the motion of 

 the armature and contacts correctly, tangential sliding can be largely 

 eliminated with a corresponding reduction in wear. 



To aid in the quantitative evaluation of wear produced by tangential 

 sliding two de\'ices have been used. One is an electromechanical vibrator 

 driven at 500 cycles per second which is capable of several mil inches 

 of motion and the other is a barium titanate longitudinal vibrator 

 coupled to a metal ''horn" which is capable of a two mil inch motion 

 at 18,000 cycles. Wires connected to these transducers are dragged over 

 materials whose wearing properties are to be tested. The normal forces 

 between the wire and material are varied as well as the length of the 

 stroke. The wear by both methods is comparable showing that the ac- 

 celerated wear testing method gives about the same wear as the slower 



