270 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1957 



70 

 65 

 60 

 55 

 50 

 45 

 40 

 35 



V) 30 



_l 



LU 



tn 



O 25 



LU 

 Q 



2 20 

 15 

 10 



< 



5 





 -5 

 -10 

 -15 

 -:?0 

 -25 



100 200 300 400 500 



FREQUENCY IN KILOCYCLES PER SECOND 



Fig. 13 — Repeater gains and losses. 



600 



was monitored on recorders in both directions. Owing to the insertion 

 and withdrawal of repeaters in the power circuit from time to time, the 

 repeaters were subjected to several power-switching operations and 

 temperature cycles. 



Stability of electrical characteristics, particularly gain, between the 

 pre-housing tests and the completion of the 'confidence trial' some four 

 months later was regarded as an important criterion of the reliability 

 of a repeater. Unfortunately test conditions and differences ])etween, and 



i 



