366 BELL SYSTEM TECHNICAL JOURNAL 



on average and shorter loop conditions where transmitter battery 

 supply is greater, with consequent higher sidetone, and to favor trans- 

 mitting and receiving efficiency on longer loops where battery supply 

 is low. Since loop losses are greater for transmitting than for receiving 

 because of transmitter battery supply loss, the ratio of the transformer 

 is such as to favor the transmitting efficiency of the set somewhat in 

 comparison to the receiving efficiency. This has the advantage of 

 raising the transmitted speech level further above line noise. The 

 same idea, of course, was followed in the design of the sidetone set. 



The resultant anti-sidetone circuit adopted and here discussed, as 

 compared with the sidetone circuit previously in general use, when 

 equipped with the same transmitter and receiver and on the same loop 

 and trunk, reduces sidetone on the average by about 10 db. Under 

 the most unfavorable conditions of use, the reduction is unlikely to 

 be less than about 7 db compared with the corresponding sidetone 

 connection. Under the best conditions of balance encountered the 

 reduction may be as much as 12 db. On the effective basis of trans- 

 mission the average net improvement in transmission which results is 

 about 6 db. 



From the electrical circuit standpoint alone, the efficiency of the 

 anti-sidetone arrangement is below that of the sidetone set in the 

 order of about one or two db in transmitting and in receiving, which 

 is necessitated by the limitations of practical design and circuit con- 

 ditions discussed above. 



Figures 4a and 4b show for transmitting and receiving, respectively, 

 the difference in efficiency, with respect to frequency, of the anti- 

 sidetone set from the sidetone set, each with the same instruments. 

 Two subscriber loop and trunk conditions are shown : an average loop 

 and trunk, and a long cable connection. 



Figure 4c shows the variation in sidetone reduction with frequency, 

 of the new set as compared with the corresponding standard sidetone 

 set, for the same two circuit conditions as above. The curves are 

 indicative of the effect of variation of circuit impedance on sidetone 

 balance, in changing not only the magnitude, but the frequency range 

 in which the best balance occurs. 



Data of this sort alone do not, of course, indicate the relative 

 transmission performance of the two sets. The beneficial effect on 

 the telephone user of the large reduction in sidetone must be evaluated 

 on the same yardstick as the losses in transmitting and receiving 

 efficiencies which, in the practical case, accompany this reduction in 

 sidetone. McKown and Emling have shown the effect of changes of 

 this sort on the results obtained by the ordinary telephone user, in 



