THE L3 SYSTEM — DESIGN 807 



on quality control to the specific needs of the L3 system. The most im- 

 portant point to appreciate in this connection is that the control of the 

 process of manufacture (as well as the end electrical requirements) of 

 individual elements is being used as a basic factor in the design of the 

 system. 



2.233 Repeater Accuracy 



In developing the equalization plan it is a logical and straight forward 

 operation to provide shapes and ranges in the equalizers that will com- 

 pensate for the random variations of known elements. Likewise, real 

 but indeterminate parasitic elements can be taken into account by 

 specifying the final characteristics of the line amplifier feedback network 

 and the equalizer fixed shapes (design deviation equalizers) on the basis 

 of measurements on a rigidly controlled group of amplifiers that are 

 deemed to be representative of the final product. However, having once 

 specified the equalization on this basis the design elements and indeter- 

 minate parasitic elements must be held to the values and ranges upon 

 which equalizer location, shapes and ranges are specified. This point of 

 view has led to rigid mechanical control and the omission of component 

 adjustments in the line amplifier which represent a departure from other 

 transmission systems. These features are discussed in detail in the com- 

 panion amplifier paper.^ 



2.3 NEW YORK-PHILADELPHIA TRIAL 



The first installation of L3 has been made between New York and 

 Philadelphia. Since the middle of 1952, this installation has been used 

 to test components, to verify values of important system parameters 

 used in system analyses, and to gather data for the further design and 

 development of equalizers. 



Random noise measurements have confirmed theoretical values (Fig. 

 5) to an accuracy of better than 2 db. In general, the measurements have 

 indicated that the theoretical values have been conservative. 



Measurements of system modulation performance, made with single 

 frequency tones, also confirm the theoretical values used in analyses. 

 Third order modulation measurements are in almost complete agree- 

 ment with theory while second order measurements have been generally 

 two to three db more favorable than the analytic values used. 



Transmission measurements have confirmed that equalizer networks 

 designed so far are satisfactory for systems to be installed in the near 

 future. Further measurements are required to determine automatic 



