834 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1953 



improved. While on the surface it might appear more practical to defer 

 this work until the data are more adequate it has been found that these 

 speculative studies are the only sure guide to ever getting the right data. 

 Faced vsith such a complicated problem it is difficult to select from the 

 vast amounts of things that might be important those relatively few 

 items on which success or failure depends. In the L3 system there have 

 been a number of critical problems which required intensive effort to 

 find acceptable solutions, for example, the design of stable regulators to 

 permit operation of 700 regulators in tandem, the choice of shapes for 

 manual and dynamic equalizers and the selection of equalizer adjust- 

 ment methods. 



The design of equalization for the L3 system is not yet complete 

 since the dynamic equalizer shapes are still subject to considerable un- 

 certainty and the details of some of the final television mop-up equaliza- 

 tion are yet to be settled. However no major difficulties are anticipated 

 in equalizing the telephone system to be installed from Philadelphia to 

 Chicago during 1953. Although amplifiers having the final gain charac- 

 teristic were operated in the trial line between New York and Philadel- 

 phia for the first time in November, 1952, it was immediately possible to 

 transmit quite satisfactory television pictures over the 200-mile loop 

 using existing equalizers. It therefore appears that equalization will not 

 limit the rate of field installation of the L3 system. 



THE PROBLEM 



The 4,000-mile coaxial cable has a gain distortion of nearly 40,000 db 

 between 0.3 and 8.5 mc. Although the amplifiers reduce the distortion to 

 perhaps 200 db, (and 100 microseconds), they leave a residue charac- 

 teristic that is considerably more difficult to equahze. Further it is 

 necessary to deUver service to intermediate offices spaced on the average 

 about 120 miles apart. This requires equalization of high precision at 

 numerous intermediate points. A further problem is the variability of 

 the transmission characteristic due to manufacturing deviations plus 

 time and temperature changes. Also, gain distortion cannot be per- 

 mitted to exceed about 5 db at any point in the line or the signal misalign- 

 ment will result in degraded signal-to-noise ratios.^ 



The overall transmission objectives^ are of the order of 0.25 db and 0.1 

 microsecond which, if allocated among the over 1,000 amplifiers, lead 

 to rather unrealistic amplifier requirements. In fact, individual am- 

 plifiers do not always meet the 4,000-mile overall requirements. Thus 

 it is the problem of the equalization designer to provide a mop-up sys- 



