232 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1954 



duced to the next lower-range unit. Such a saving can be reahzed only 

 part of the time, but it is desirable to attempt to weight its value so 

 that over the telephone system as a whole there is adec^uate encourage- 

 ment to strive for power savings. 



The method of weighting is complicated by such factors as the fraction 

 of power that may be saved in an office, by the likelihood of certain size 

 offices being representative, and by the e.xtent to which offices are de- 

 signed with safety factors for growth and overload. XcA'ertheless a 

 reasonable estimate can be made by assuming difTering amounts of the 

 office power can be saved, and by assuming that all sizes of office are 

 equall}' likely. This gives the result that the over-all incremental value 

 of power is 



the average value of power plant per kilowatt saved over the range 

 between Wo and Wo , taking into account the proportionate number of 

 plants that can be converted to the next cheaper size, where 



C = total price saving, 



W = total power saving, 



p = fraction of power saved, not to exceed about .5, (above which 



point two steps of base price saving might be realized). 

 Wo = lower limit of range, 

 W2 = upper limit of range, 

 a = incremental price of one size power plant, assumed to be about 



constant for all plants, 

 A = difference in base price to next lower plant, as rasi\ be tabulated 



for any particular case. 



By these methods, incremental values of installed power plarit may be 

 found; it is even possible in many cases to assume an average slope for 

 the family of curves in Fig. 6, without greatly changing the results in 

 practice. 



It is now desired to convert these figures of installed expense to figures 

 comparable to manufacturing cost, in the factory, of the relays which 

 will consume this power. This may be done through standard economic- 

 study practice by which first cost is converted to annual charges through 

 a knowledge of charges for interest, taxes, depreciation, etc. As these 

 factors are not necessarily^ alike for power ec}uipment and relays, the 

 factors used will differ to suit the problem. For the present study, with 

 recognition of the accounting practices then in effect, it was found that 



