592 



BELL SYSTEM TECHNICAL JOURNAL 



passing through the usual test board. At points in the circuit sensitive 

 to noise interference or crosstalk where the energy level of the trans- 

 mitted signals is very low, the circuit units are connected by means 

 of shielded pairs. This shield is connected to filament ground, as are 

 also the cases of the various transformers which are insulated from 

 the supporting metallic frame. The unavoidable noise potential 

 existing between the frame and the filament circuit cannot then pro- 

 duce any appreciable disturbance in the circuit. 



Overall Perjormance of System 

 A measurement of the transmission loss of the 2,200-mile test 

 length of B-22-N cable circuit gave results as indicated in Fig. 21. 



UJ 



_l 

 I 



O 



UJ 



a. 



UJ I 



a. z 

 < 



CYCLES PER SECOND 



Fig. 21 — Transmission-frequency characteristics of 2,200 miles of 16-ga. B-22 

 cable program transmission circuit. Curve A — Ideal characteristic. Curve B — • 

 Measured characteristic. Curve C — Line without equalizers. 



It will be observed that over the range from 35 cycles to 8,000 cycles 

 the transmission loss was practically the same at all frequencies, 

 departing only about ± 2 db. For comparison, another curve (Q 

 is given on the same drawing showing the transmission characteristic 

 which would have been obtained if distortionless amplification had 

 simply been added to the line with no attenuation equalizers. 



The delay-frequency characteristics of the 2,200-mile test length of 

 B-22 circuit are shown in Fig. 22. Two curves are given, one for 

 the circuit without delay equalizers, the other with delay equalizers. 



With respect to non-linear distortion, it was found by test that 

 when the maximum volume was held at about — 5 db, as read on a 



