CONQUEST OF DIST.iyXE BY WIRE TELEPHONY 387 



and again in different degrees. Moreover, there were no systematic rela- 

 tions among the effects on coils at different points in the same circuit, or on 

 coils in different circuits at the same loading points. These effects usually 

 occurred without mechanical injury to the coil windings, or to the associated 

 lightning arresters with which each coil was protected against breakdown, 

 and it was this fact that had delayed recognition of lightning as being the 

 probable cause. Confirmation of this deduction was obtained when tests 

 made on lines unexposed to lightning showed that the coil inductances were 

 close to their factory adjustment values. 



Laboratory tests showed the magnetizing effects of lightning surges to be 

 of the same order as the residual magnetizing effects of superposed direct 

 currents, ranging up to several amperes in amplitude. The high magnetic 

 retentivity of the continuous wire-type toroidal cores of the loading coils 

 was a basic factor in these phenomena. 



The necessity for accepting exposure to lightning surges as a normal 

 service experience for open-wire loading coils forced consideration of the 

 practicability of providing new designs having much greater magnetic 

 stability. Experimental work was started on (non-magnetizable) solenoidal 

 t^-pe air-core coils having finely sectionalized windings, and on toroidal wire- 

 core coils having series air-gaps in their magnetic circuit to decrease its 

 retentivity. 



Fortunately the statistical study which was made to determine the limits 

 that should be placed upon individual line irregularities in order to avoid 

 undesirable restrictions on the repeater gains showed that it would not be 

 necessary to use perfectly stable coils, i.e., the air-core coils. The concur- 

 rent work on the wire-core coils with air-gaps had indicated that by properly 

 proportioning the air-gaps the inductance changes that should be expected 

 from magnetization by lightning surges could probably be kept to tolerably 

 low values. A single air-gap would have been sufficient to provide the 

 required stability, but crosstalk considerations and other factors made it 

 desirable to have two air-gaps symmetrically located at diametral points in 

 the toroidal cores. The resultant designs were better in all important re- 

 spects than the air-core coils and were inferior only with respect to mag- 

 netic stability, which difference as above noted was tolerable. To assist in 

 the control of the inductance deviations in the lines, the new loading coils 

 were manufactured to ±1% precision inductance limits. On the older 

 standard designs, ±6% manufacturing deviations had been allowed. The 

 new coils had somewhat lower nominal inductance values than the old coils, 

 so that their average service inductance values after partial magnetization 

 by lightning surges would be about the same. 



The size advantage of the wire-core coils with air-gaps made it possible to 

 pot the three loading coils for an open-wire phantom group, connected as a 



