964 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1956 



fore service reactions occur. Such measurements will also be of aid in de- 

 termining the effectiveness of efforts along the first and third lines. 



As discussed earlier, the presence of the operator on every call was of 

 material assistance in the detection of unsatisfactory trunks. On operator 

 or direct distance dailed calls, there will be little or no operator conversa- 

 tion over the intertoll trunk connection. As a substitute, the maintenance 

 forces may need to make more frequent checks of the transmission per- 

 formance of the trunks unless the stability of individual systems and 

 components of systems can be improved. Manual methods have been 

 used by the maintenance forces in the past to measure trunk losses. Semi- 

 automatic measuring methods have been developed to reduce the time 

 and effort required. In many cases the necessary number of measurements 

 will be economical only when made by automatic devices. One form of 

 such gear is described in a companion paper. ^ 



The ability to measure over-all trunk losses simply and frequently is 

 of direct aid in detecting when loss deviations exceed maximum toler- 

 ances. Such measurements in themselves, however, are insufficient to 

 detect incipient troubles or to indicate the component part responsible 

 for unsatisfactory transmission. An attempt has been made to achieve 

 these objectives by using statistical analysis of the measured data as 

 an aid to diagnosis. The following sections discuss the application of such 

 analysis. 



Use of Transmission Loss Data 



It has been shown that considerable variation can be expected in trunk 

 losses even in the absence of trouble conditions. For any given group of 

 trunks selected for analysis, the performance is described by the distri- 

 bution grade and the bias. If a group of trunks is found to have bias, it 

 is usually an indication of some assignable cause. One such cause might 

 be a change in gain of an amplifier common to the group. Another cause 

 might be improper gain adjustment for channel units of a carrier termi- 

 nal associated with the group. 



If a group of trunks is found to have a greater distribution grade than 

 the distribution grade for all the trunks in the office, this may indicate 

 excessive instability in a component part common to the trunks in the 

 group. If analysis of all the trunks terminating in an office shows a higher 

 distribution grade than is usually fomid in similar offices, the fault may 

 be due to maintenance routines being inadequately or improperly ap- 

 plied. 



* H. H. Felder, A. J. Pascarella and H. F. Shoffstall, Automatic Testing of 

 Transmission and Operational Functions of Intertoll Trunks, page 927 of this 

 issue. 



