Contribution of Statistics to the Development Program of a Transformer for L3 Carrier System

Statistical methods played a significant part in the development program of the L3 system. Experiments were designed to assist in improving the manufacture of the input and output transformers of the amplifiers. Detailed analysis of a few of these experiments is presented. I. INTRODUCTION In p r e v i o u s issues of T H E B E L L SYSTEM TECHNICAL JOURNAL t h e problems in design, development and manufacture t h a t were encountered in building the L3 coaxial carrier system are described. This system provides 1,860 one-way telephone channels or 600 one-way telephone channels plus one TV channel over each coaxial tube. T h e L3 system is capable of transmitting a television signal over a distance of approximately 1,000 miles and telephone signals, approximately 4,000 miles. From the start of the development program, statistical methods have played a significant part. Special acceptance procedures have been set up to assure t h a t the shipped product would meet certain distribution requirements. 1 Control chart techniques were generously applied both in the manufacture of component parts and for subassemblies. 2 This paper gives in part a case history of one of the difficult components. T h e viewpoint is t h a t of the experiments designed to overcome difficulties in the initiation of the manufacturing process and to explore possibilities of improvement of the component. A detailed discussion of the present manufacturing techniques of this component, the input and o u t p u t transformer of the amplifier, has already been presented by Earle.