PART II — THE UHF NOISE SPECTRUM 859 



At each of these optimum field settings, several weaker "satellite" 

 peaks can also be detected, most readily those at the cyclotron frequency 

 itself, and at 0.707 times the latter; i.e., the "plasma" frequency, as 

 shown in Fig. 1. In addition, smaller peaks have been repeatedly ob- 

 served at harmonics (up to the sixth) of the proper frequency, and one 

 at slightly less than half of that frequency. (When a proper frequency 

 was simulated by means of a signal generator, only its first harmonic 

 could be detected in the receiver output.) 



At the fields corresponding to AT" = 4 in Fig. 1, the cyclotron frequency 

 (312 mc) was found, but not the proper frequency. The highest proper 

 frequency observed was 240.5 mc, in the A^ = 3 mode. The proper- 

 frequency peaks decrease with increasing focusing field, whereas the 

 field-independent peaks excited in the electron gun tend to increase, at 

 the far end of the drift region. 



IV SPATIAL DISTRIBUTION OF UHF NOISE CURRENTS 



In Figs. 2 to 5 are shown synchronized chart records of collector cur- 

 rent, one or more UHF narrow-band peaks, and microwave noise power 

 near 4,000 mc — all as functions of distance from the electron gun, for 

 the A'' = 1 to 4 modes, respectively. In all runs, the beam was pulsed 

 with a 1,000-cycle square wave, and the collector aperture set at a 0.100 

 inch diameter. The magnetic fields at the cathode and in the drift space 

 were adjusted before each set of readings, with the probe at a common 

 reference position, for greatest amplitude of some UHF peak. In Figs. 

 2 and 3, these were proper frequencies, whereas in Figs. 4 and 5 they were 

 field-independent frequencies. 



The content of these distribution curves can be summarized as follows: 



(1) At the low fields employed (none quite equal to the nominal 

 Brillouin value), the beam ripples are quite large, both in amplitude and 

 wavelength. 



(2) The proper-frequency traces have two or three maxima near each 

 beam waist, and their amplitudes grow more rapidly with distance from 

 the gun than any of the satellite frequencies. 



(3) The patterns of the cyclotron and "plasma" frequencies do not 

 differ significantly from those of the field-independent frequencies, and 

 usually display two peaks near each beam waist. 



(4) The collector-current maxima decrease with distance from the gun, 

 although their minima change little. (The first maximum is sometimes 

 flat-topped due to beam interception before it enters the drift space.) 

 The rate of decrease of these maxima, and the rate of increase of proper- 

 frequency amplitude, are greater, the longer the ripple wavelength. 



