832 THE BELL SYSTEM TECHNICAL JOURNAL, .IITLY H)57 



In Part II of this 'paper, measurements of the noise spectrum of a rippled 

 beam in the UHF region are described. These measurements reveal 

 the presence of additional forms of instability. Calculations are made to 

 account for some of these, and for aspects of rippled-beam amplijication not 

 previously understood. 



Part I — The Growing Noise Phenomenon* 



I INTRODUCTION 



When an RF probe is moved along a magnetically -focused electron 

 beam in a drift region, the noise power is at first found to vary periodi- 

 cally with distance from the electron gun. For a sufficiently long beam, 

 however, the periodic pattern is succeeded by an exponential rise, culmi- 

 nating in an irregular plateau. This so-called "growing noise" phenome- 

 non has been extensively investigated by its discoverers, L. Smullin and 

 his colleagues at the M.I.T. Research Laboratory of Electronics.^' ' 

 They have established that this noise will begin to grow at a plane nearer 

 the gun, and tend to grow at a faster rate, for electron beams (a) of 

 higher perveance, (b) with less space-charge neutralization by positive 

 ions, and (c) issuing from convergent, partly-shielded guns, rather than 

 those immersed in the magnetic field. 



The growth of microwave noise power in drifting beams has hampered 

 the development of high-power, traveling-wave tubes with acceptably 

 low noise figures,, as such devices generally have convergent, partly- 

 shielded electron guns. The problem has been evaded in the design of 

 low-noise, low-power traveling-wave tubes, by resort to confined-flow, 

 parallel beams. 



Several theories have been proposed to explain the growing-noise 

 wave: 



(1) Excitation of higher-order modes with complex propagation con- 

 stants, by electrons threading the beam transversely; 



(2) Slipping-stream amplification, due to either longitudinal or trans- 

 verse velocity gradients; 



(3) Rippled-beam amplification; ' ' and 



(4) Electron-electron interactions leading eventually to equipartition 

 of thermal energy, and thus an increase in longitudinal velocity fluctua- 

 tions. 



In Part I of this paper, measurements are presented which sho\\- that 

 the principal cause of growing noise appears to be space-charge wave 



* Presented at the I.R.E. Electron Tube Research Conference, Boulder, Colo- 

 rado, June 27-29, 1956. 





