356 BELL SYSTEM TECHNICAL JOURNAL 



open-wire phantom group of 104 mil conductors, between Newtown Square 

 and Brush ton (test stations near Philadelphia and Pittsburgh, respectively) 

 installed during August 1910. 



By design, the side circuit transmission characteristics were substantially 

 identical with those of the loaded non-phantomed circuits then in extensive 

 use. A slight impairment resulted from the increase in circuit resistance 

 caused by the inserted phantom loading coils. These coils were installed 

 at the same points as the side circuit loading coils, at systematic intervals 

 of about eight miles, a distance set by the need to coordinate the coil spacing 

 with the line transposition systems. The phantom loading coil inductance 

 (0.163 henry) was chosen to provide a theoretical cut-off frequency close 

 to that of the side circuit (approx. 2400 cycles). This resulted in a phan- 

 tom circuit impedance approximately 60 per cent of that of the side circuits, 

 and an attenuation nearly 20 per cent lower than that of the side circuits. 

 This advantage resulted in the phantom being preferred for long-haul 

 service. The increase in transmission efficiency obtained by loading the 

 phantom (about 2.5 to 1) was practically as large as that obtained in the 

 side circuits. 



Since this was a pioneering project, it is understandable that the cross- 

 talk results were not all that could be desired. There was some real satis- 

 faction, however, in the fact that the crosstalk was not too close to the bor- 

 derline of being intolerable. The crosstalk was due to unbalances in the 

 line and in the loading coils. In commercial service, unbalances in the 

 phantom-deriving repeating coils and in the composite telegraph sets were 

 also factors in the crosstalk performance. In the course of time, in conse- 

 quence of improvements in the phantom transposition systems and expe- 

 rience in the manufacture of the line and terminal apparatus, substantial 

 improvements in the service crosstalk characteristics were secured. 



The loaded phantom circuit was much more susceptible to noise induc- 

 tion than the side circuits, and increased the need for good line mainten- 

 ance. 



At this point, a few remarks regarding the conservative policy followed in 

 this phantom loading development are appropriate. So far as the loading 

 apparatus development work itself was concerned, a trial installation could 

 have been made much earlier than the summer of 1910. The early labora- 

 tory work on the proposed initial loading coil designs showed several minor 

 changes to be desirable from the crosstalk standpoint. After these were 

 made, the designs appeared to be free from inherent dissymmetry that might 

 cause crosstalk. The question as to whether the coils would have satis- 

 factory balance when manufactured on a quantity production basis, how- 

 ever, could only be determined by undertaking manufacture of a sizable lot. 

 When the question of making a trial installation was first considered in- 



