892 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1953 



JP 



^ 



> 



ilg. 



i'rccision transiormer, windings, core, assembled unit. 



transformer itself would have to be given as functions of frequency be- 

 cause of the distributed nature of the device. When, however, the high 

 impedance winding is raised above ground potential by the voltage 

 developed across the beta circuit, which is nearly the same magnitude as 

 the voltage across the coupling network, the effects of distributed para- 

 sitic capacities to ground, and the lumped capacity from the junction of 

 transformer and peaking coil, become of prime importance. The coupling 

 network, therefore, cannot be adequately represented by merely lifting 

 the circuit of Fig. 8(b) off-ground. 



In order correctly to understand and compute the amplifier gain, 

 it was necesary to develop a complex mathematical analysis of the dis- 

 tributed structure of the transformer, in conjunction with an extended 

 program of precise measurements of the transformer constants. Even 

 then, one must be content with an accuracy of a few tenths of a db and a 

 few degrees of phase, as compared with the order of magnitude better 

 accuracy which can be obtained for a two-terminal lumped-constant 

 network. The agreement between measurement and computation of 

 amplifier gain and reflection coefficient is sufficiently good, however, to 



