c.iik'kiik ri:i.rriii>xy ox men jiw/.m,/ iim^ i.v 



the iin[K'<laiui' of t>pir.il transformer l)anks was im-asiiri-d. In 

 Fig. 4 is shown the ini(K'(lanre versus fre(|iK'nr\' characteristic of a 

 thrif phase, llO.OtM) titKM) \\. 12.(MK) K.V.A. transformer hank con- 

 nccte<l "star" on the lii^h side wiiii ilie neutral ^numded and "delta" 

 on the low side. As shown l)>' the diajjrani, these measiirentents 

 were made l)etween phases on the high side with the low side optyr 

 circuited and short circuited. The coincidence of these curves for 

 frequencies aliove 50 K.C". indicates that at these fre(|uencies the 

 dominant characteristic is the distributed capacity of the \\\g\\ windinjj 

 and the iniptnlance is probably unaffected by changes on the low 

 p<itential side of the transformer. Below .')() K.C, however, the 

 impedance changes rapidly both with lre(|ueiuv- and with the low 

 potential termination. 



A study of Figs. 3 and 4 and other data shows that the dtsirabk- 

 frequency range in which to operate a power line carrier telephone 

 circuit is that from ."lO K.C. to 1.10 K.C. In this range the attenua- 

 tion is not excessive, it is very little affected by the associated power 

 apparatus, and it is independent of the conditions on the low potential 

 power circuits. The curve shown in Fig. 3 indicates that, contrary to 

 the common l)elief, the attenuation in this range is a relatively smooth 

 function of freciucncy. This conclusion is supported by the fact that 

 in the various installations of power line carrier telephone equip- 

 ment which have liecn made since the attenuation measurements on 

 Fig. 3 were obtained, no [X)vver Imes have been encountered where the 

 attenuation was a critical function of frequency. Another important 

 argument for the selection of this frcquencv' range lies in the fact tnat 

 it is well above the range emplovcd for midtiplex telephony on com- 

 mercial telephone svstems and therefore precludes any interference 

 with such systems. 



Coii'i.iNc; Bi;r\vi;iiN C.arrikr Hoi'i'^'iiNr and Powkk Line 



Probably the most ililficult problem to solve was that of providing 

 a satisfactory method for connecting the carrier eciuipment to the 

 power line. The u.se of power transformers has not been found prac- 

 ticable for if frequencies low enough to be efficiently transformed were 

 employed, the attenuation of the circuit would be a function of the 

 conditions in the distributing network and a change in the number or 

 arrangement of transformers would result in an appreciable change 

 in the attenuation. Such a method of coupling to the power line 

 would also have the objection that communication would not be 

 possible when the p<jwer transformers were disconnected from the line. 



