610 BELL SYSTEM TECHNICAL JOURNAL 



condenser. The maximum voltage to which C2 is charged in this interval 

 is made much greater than that of the d-c. power source {E^. The ratio of 

 these two voltages depends upon the ratio of the inductance charging time in 

 the preceding interval to the oscillation period. Both factors can be varied 

 over wide limits, and step-up ratios of roughly ten to twenty are generally 

 used. 



The third interval starts with magnetization of the non-linear core near 

 point d on the loop, where the inductance again drops. This situation is 

 precisely the same as that previously described for the low-power pulser. 

 As a result the condenser discharges through the load resistance at the time 

 indicated in Fig. 3c, driving the core far into saturation with a field of the 

 order of a hundred oersteds. This field extends too far to the left of point d 

 to be shown in Fig. 3(d). Here the differential permeability approaches 

 unity, and the correspondingly low inductance permits a rapid build-up of 

 pulse current. Evidently but one pulse is produced each time the tetrode 

 conducts, and the number of pulses produced per second is changed simply 

 by varying the input frequency without requiring any circuit change, power 

 dissipations permitting. 



Energy storage in the linear coil depends upon its inductance, upon the 

 bias current, and upon the peak current reached during the tetrode conduc- 

 tion interval. A plot of the current in L\ against the product of time and of 

 voltage across the coil permits this energy to be represented as an area 

 (Fig. 3e). Evidently a given area can be made up by var}'ing the relative 

 sizes of its component triangle and rectangle, only the latter varying with 

 bias current. If for example the bias is reduced to zero, the rectangle would 

 vanish and the peak current would have to be increased to attain the original 

 amount of stored energy. The higher maximum current requires more 

 cathode emission of the tetrode and leads to greater plate power dissipation. 

 Thus in addition to determining the energy stored, the amount of bias is 

 one of the factors determining power dissipation capacity and emission which 

 must be provided in the driving tube or tubes. Additional factors enter to 

 make a bias corresponding to d (Fig. 3d) the most favorable from an efficiency 

 standpoint. 



The operating principles developed above in terms of a simplified circuit 

 have been applied to a number of practical circuit forms which are described 

 in the sections following. 



Load Circuit 



In radar applications the useful load is a magnetron which takes the 

 place of the linear resistance previously considered. Since the magnetron 

 viewed at its input terminals acts essentially Hke a negatively biased rectifier, 

 additional means must provide for the flow of condenser charging current in a 



