APPLYING CARRIER TELEPHONE SYSTEMS 585 



through a city area. In other cases certain main stations are not 

 manned 24 hours per day and control and alarm circuits have not been 

 terminated at such points. In one case a main repeater station has 

 terminated in it the control leads from eight auxiliary stations. Four 

 of these are connected through the partially attended main stations on 

 either side. 



Completion Tests and Overall System Adjustments 



The usual completion tests were made on each unit of equipment 

 after it was installed and on each cable pair between repeater stations 

 after it was unloaded, in order to insure readiness of each item to be 

 connected to form the overall carrier system. The gains of the re- 

 peaters were given a final adjustment by connecting each repeater 

 input to the cable pair with which it was installed to work, then 

 sending a predetermined amount of power at 28 kc into the cable pair 

 at the adjacent office and adjusting the gain of the repeater until it 

 delivered the desired output level. The flat gain master regulator was 

 adjusted with respect to its pilot wire so that it would adjust the gain 

 of the amplifier to maintain the desired output level at 28 kc at all 

 cable temperatures. 



The repeater sections beginning at one end of each twist regulating 

 section were measured progressively at the output of each repeater. 

 In each case transmission was checked at ten frequencies throughout 

 the range from 12 to 60 kc. In this way a check was obtained to deter- 

 mine whether proper equalization was being provided at each line 

 amplifier. It was necessary in some cases to change the type of 

 equalizer provided because the transmission characteristics of specific 

 cable pairs differed from the average which had been assumed in 

 providing equalizers. Measurements were also made on the overall 

 twist regulating section and transmission checked at ten frequencies 

 throughout the 12 to 60 kc range, since the output of a perfectly cor- 

 rected twist section would be the same at all frequencies within this 

 band. 



Overall measurements similar to these were made on the high-fre- 

 quency line between terminal points and Fig. 20 shows the results for 

 typical New York-Charlotte and New York- Washington systems. 

 The most desirable characteristic would be a straight line and it will 

 be noted that this curve differs materially from such an ideal. This 

 difference is due to inadequate compensation by means of equalizers 

 for small deviations from linearity in the individual line amplifiers. 

 This lack of linearity in the overall high-frequency line has not ma- 

 terially affected the systems being operated at present but might be- 



