530 BELL SYSTEM TECHNICAL JOURNAL 



while in shorter circuits somewhat higher percentages might be en- 

 countered. In underground cable about one third as much would be 

 encountered in a year, but very little would normally occur in one day 

 since the rate of change is small. 



In order to take care of these large variations a system of pilot wire 

 regulation is used. The following discusses this system in some detail 

 in order to show what the residual variations are. This system con- 

 sists of a pilot wire extending through the cable whose circuits are to 

 be regulated, each pilot wire being perhaps 100 to 150 miles in length. 

 An automatic mechanism measures the d-c. resistance of this pilot 

 wire frequently, and makes occasional adjustments of the gain of the 

 regulating repeaters. In the case of the four-wire facilities referred to 

 above, these adjustments are made in approximately .5 db steps at 

 1,000 cycles, and other suitable adjustments are made at other fre- 

 quencies. 



This pilot wire is placed in the four-wire part of the cable (it is 

 usually obtained by compositing a four-wire circuit) and therefore 

 has very closely the same temperature variation as the four-wire pairs. 

 The position in the cable of the two directions of transmission of four- 

 wire circuits is reversed '^ at the center of each repeater section, so it 

 is possible to regulate both directions of transmission from a pilot 

 wire in either group without serious error. Since the two-wire circuits 

 are comparatively short, have generally smaller variations in decibels 

 per mile than four-wire circuits, and usually have an average position 

 in the cable, there is no serious error in regulating these from the same 

 pilot wire. 



Due to the finite steps in which these regulators operate, there is a 

 residual variation which is approximately ±.25 db per regulating 

 repeater. In addition, there may be a certain amount of lag in the 

 operation of these regulators, because of the fact that it is desirable 

 to prevent excessively frequent operation of these devices, and perhaps 

 partly because of mechanical backlash. To prevent hunting it is 

 necessary to make the adjustment in the pilot wire regulator somewhat 

 smaller than the adjustment which would be necessary to make all 

 the variation due to this cause a random matter. In other words, 

 when the temperature is changing in a given direction in many repeater 

 sections, for example early in the morning, the adjustment at each 

 of the pilot wire regulators is slightly behind what it theoretically 

 should be for the pilot wire resistance obtaining at that time. This 

 results in a directly additive effect in all regulating repeaters in a given 

 circuit during certain times of day. By careful design and routine 



'* This assumes concentric segregation which is generally used. 



