PART II — THE UHF NOISE SPECTRUM 857 



(1) Transverse positive-ion oscillations, for which the freciuencies 

 vary as the square root of anode voltage. 



(2) Transverse electron plasma oscillations (near or beyond the 

 anode), for which the frequencies would be too high. 



(3) Longitudinal electron plasma oscillations at the potential mini- 

 mum, for the same reason (should be near 2,500 mc). 



(4) Longitudinal diode oscillations.'* When the electron transit angle 

 through the diode is approximately (n -\- j) periods, where n is an in- 

 teger, the real part of the diode conductance becomes negative, permit- 

 ting oscillations to occur. Again the frequencies of such oscillations would 

 be too high, (2,200 mc and higher) for the gun used, to conform to the 

 observed values. 



There is, however, one published theory for which an order-of -magni- 

 tude correspondence does exist between the measured and calculated 

 frequencies. Klemperer^' ^ has shown that a strip beam tends to break 

 up into clusters of "pencils" at the cathode. He ascribes these to standing 

 waves resulting from transverse oscillations in the space-charge cloud, 

 and offers an expression for the wave velocity in this medium. Applica- 

 tion of his formula to the cathode used in the present experiments results 

 in a least frequency of 3L3 mc. Other observers, such as Smyth' and 

 \^eith,* have also reported evidence of interaction between electrons in 

 a retarding-field region and RF fields, which may underlie these oscilla- 

 tions, 



III FIELD-DEPENDENT PEAKS 



With the RF probe stationed ten or more inches from the gun anode, 

 narrow-band peaks can be found in the noise spectrum of the beam. 

 The amplitudes of these peaks increase and their frequencies decrease 

 with decreases in the magnetic field. For each probe position, the process 

 of finding the peak of greatest amplitude involves repeated adjustments 

 of the focusing field, the magnetic field at the cathode, and the receiver 

 frequency. 



When the fields have been so optimized, it is found that the probe is 

 located at or near the first beam-diameter minimum, following that at 

 the entrance to the drift space. When the field is doubled, and the 

 "tuning" process repeated, the greatest peak is found to have about twice 

 the frequency of the fir.st, and the probe is found to be located at or near 

 the second beam waist. It is convenient, therefore, to think of these 

 peaks as "proper" frequencies of the A'' = 1, etc., modes of the rippled 

 beam, where A^ is the number of ripple wavelengths between gun and 

 probe. 



