178 BELL SYSTEM TECHNICAL JOURNAL 



On the basis of the nature of the electronic mechanism by means of which 

 energy is transferred to the RF field, it is now convenient to distinguish 

 three types of magnetron oscillators.^ The negative resistance magnetron 

 oscillator depends on the existence of a static negative resistance character- 

 istic between the two halves of a split anode. ^ The cyclotron frequency mag- 

 netron oscillator operates by virtue of resonance between the period of RF 

 oscillation and the period of the cycloidal motion of the electrons (rolling 

 circle or cyclotron frequency).^ The traveling wave magnetron oscillator de- 

 pends upon resonance, that is, approximate equality, between the mean 

 translational velocity of the electrons and the velocity of a traveling wave 

 component of the RF interaction field. ^ 



The magnetron oscillator with which this paper is primarily concerned 

 is of the traveling wave type. The other magnetron types are discussed 

 briefly for the sake of completeness and because an understanding of them 

 enhances one's grasp of the entire subject and places the traveling wave 

 magnetron oscillator in its proper historical perspective. 



2.2 The Negative Resistance Magnetron Oscillator — Type I: In the neg- 

 ative resistance magnetron oscillator,^ the anode is split parallel to the 

 axis into two halves, between which the RF circuit is attached. The elec- 

 trons emitted by the cathode must move under the combined action of 

 the DC radial electric and DC axial magnetic fields together with the RF 

 electric field existing between the two semicyhnders forming the anode. 

 The transit time from cathode to anode is not involved in the mechanism 

 except that it must be small relative to the period of the RF oscillation. The 

 static negative resistance characteristic arises from the fact that under cer- 

 tain conditions the allowable orbits for the majority of electrons terminate 

 on the segment of lower potential, irrespective of the segment toward which 

 they start. These electrons, being driven against the RF component of 

 the field, give energy gained from the DC field to the RF field. 



In Fig. 6 are shown the paths, plotted by Kilgore,'^ of two electrons start- 

 ing initially toward opposite segments but both striking the segment of 

 lower potential. Each path is completely traversed in a time during which 



2 The magnetron oscillator discussed by Hull, in which the magnet winding is coupled 

 to the plate circuit, is not considered as it is essentially an audio frequency device. K. 

 Okabe in his book, "Magnetron-Oscillations of Ultra Short Wavelengths" (Shokendo, 

 1937), distinguishes five types, but it is not clear just how liis types C and E are to be 

 identified. 



3 These oscillations have been called Habann, quasi-stationary, or dynatron oscilla- 

 tions, and correspond to Okabe's type D. 



■i These oscillations have been called electronic oscillations by Megaw, transit time 

 oscillations of the first order by Herriger and Hiilster, and correspond to Okabe's type A. 



^ These oscillations are the running wave type discussed by Posthumus, the transit 

 time oscillations of higher order of Herriger and Hiilster, and correspond to Okabe's 

 type B. 



« This t\pe was disclosed by Habann, Zeit f. Hochfrequenz. 24, 115 and 135 (1924). 



' G. R. Kilgore, Proc. I.R.E. 24, 1140 (1936). 



