ECOXOMICS OF TELEPHOXK RF.LAY APPLICATIOXS 221 



may apply (a) suitable windings to satisfy the various conditions im- 

 posed by a need for sensitivity, .speed, lowest possible first cost, or some 

 other of the many specialized reciuirements to be encountered; and (b) 

 particular arrangements of contact springs which will provide for the 

 desired functions in the circuit ranging from a single "make" up to 

 complicated sequences such as 12 sets of "transfers." Every different 

 combination of the basic parts is termed a "code," meaning another kind 

 of relay built up from the parts common to the basic type in ciuestion. 

 If an increasingly large number of such codes is encouraged, each tailored 

 to some special circuit function, then the advantages of mass production 

 begin to be lost as the total demand is split into more and more sub- 

 divisions, each with comparatively low demand. Each additional kind 

 of relay is a problem in "coding economics," and can lead to excessive 

 telephone office costs unless it is properly considered. 



Satisfactory solutions to this coding problem in turn depend on a 

 detailed knowledge of all the factors governing either the first cost of 

 the relays, or the first cost of other features of the system, as controlled 

 by the relays. vVhen all such effects are properly stated, they can be 

 compared and brought into economic balance. Though such studies 

 soon branch out into many fields, the enormous dollar savings are strong 

 incentives to do whatever work is needed. 



Best economic balance in the system is formed through a series of 

 optimizing steps, generally illustrated by Fig. 1. Graph (a) shows how 

 the total cost of an office may vary with number of codes. The perform- 

 ance part of the cost diminishes with increasing codes because each code 

 will more nearly satisfy the precise circuit needs. But the manufacturing 

 costs of the relays will rise with increasing codes because the lot-sizes 

 grow ever smaller and hence more costly. The lowest point on the sum- 

 mation curve represents a desirable goal. The two base curves of (a) in 

 turn are built from detailed facts about parts costs, coil costs, power 

 plant and eciuipment costs, and many other similar data. The curve of 

 manufacturing cost, for instance, is built from graphs like (b) and (c) 

 which tell how factory costs will vary. The curve of performance cost is 

 l)uilt from graphs like (d) and (e) which tell how office costs will var}' for 

 things like power and speed. In each such case, important individual 

 economies result by following similar optimizing steps. Procedures for 

 the practical application of these ideas have been of particular help in 

 recent redesigns of the Xo. 5 crossbar system to use the newly developed 

 wire spring relay family.^ 



1 A. C. Keller, A New General Purpose Relay for Telephone Switching Sys- 

 tems, B. S. T. J., 31, pp. 1023-1067, Nov., 1952. 



