ll<Ki:GUL.IKITlES IN LOADED TEl.r.l'HOXE CIHCIITS SKI 



Tl' at 1.000 cycles from which 5^ = 11 TU. In equation (42) .S',i is 



negative hence the No. 19-gauge will have a higher return loss than 



the No. l()-gauge circuits and the expected difference is 11—8.7 = 

 2.3 TU. 



Comparison of Calculated and Mka.-iLuld 

 Return Losses 



In order to test the methods of calculation described above a series 

 of measurements of return loss at 500, 1000 and 2000 cycles were 

 made on a group of loaded side and phantom circuits in a cable using 

 a No. 2-A unbalance measuring set. 



The representative inductance deviations were found by analyzing 

 the inductance measurements on a large group of loading coils similar 

 to those used in the cable. The representative capacity deviations, 

 not including the spacing irregularity were found by analyzing the 

 shop measurements on a number of reels of the cable. This gave 

 representative figures for reel lengths which were divided by \/l2 (in 

 accordance with the laws of probability since this cable had 12 reel 

 lengths in a loading section) to obtain the representative capacity 

 deviations due to the cable for the loading sections. The spacing 

 deviations were separately determined from the measured distances 

 between the loading points. 



The data used in the calculation were as follows: 



T.ABLE II 



Sides Phantoms 



Representative inductance deviation 0.0062* 0.0061* 



Representative capacity deviation 0.0129* 0.0138* 



Representative spacing deviation 0.0045* 0.0045* 



Combined representative deviation, H 0.0150* 0.0158* 



Cutoff frequency /c (cycles sec.) 2810 3727 



( 500 cycles 0.265 0.271 



Transmission loss ) ^^^ ^y^,^^ 0.274 0.279 



TUpermile ( ,^00 cycles 0.317 0.296 



The smooth curves of Figs. 9 to 14, inclusive, were calculated from 

 the data in Table II using the methods described above. The abscissas 

 are the percentages of a large group of circuits which may be expected 

 to have return losses greater than the values given by the ordinates. 

 This percentage is equal to 100 (1 — F). The points plotted on the 



• The figures are "fractional" deviations. Percentage deviations which are 

 sometmes used are 100 times as large. Care should be taken to avoid errors caused 

 by failure to divide percentage deviations by 100 before finding the value of Y h- 



