856 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1957 



mitted to register itself directly on the chart recorder, whose motion is 

 synchronized with that of the probe. Very roughly, the detector output 

 varies as the log of input power. 



Measurements are described (a) of the UHF noise spectrum in the 

 beam, just outside the gun anode; (b) of this spectrum at the end of the 

 drift region, in a longitudinal magnetic field; (c) of the noise-power dis- 

 tribution along the axis; and (d) transverse to the axis of the rippled 

 beam in the drift region. Two calculations are then outlined, one of wave 

 propagation along the rippled beam (to explain the observed distribution 

 patterns), and the other to account for some spectacular peaks in the 

 beam spectrum (b). 



II FIELD-INDEPENDENT PEAKS 



When the noise spectrum of an electron beam is scanned by a tunable 

 receiver, it is found that an irregular array of narrow-band peaks char- 

 acterize the UHF region, below about 1000 mc. Of these peaks, some are 

 due to spurious modulation effects," and can be eliminated as follows: 



(1) Transit-time oscillations due to positive ions, secondary electrons, 

 or both. Such frequencies vary with probe (collector) position. 



(2) Resonances in the probe and receiver, excited by the pulsed- 

 voltage supply. These are unaffected by changes in collector current. 



(3) Ion oscillations in the electron gun or beam. Their frequencies vary 

 with anode voltage. 



The remaining narrow-band peaks fall into two classes, depending 

 on whether their frequencies vary with the magnetic field. 



Well-defined peaks can be detected with the RF probe stationed one 

 inch from the gun anode, with or without any focusing field. When the 

 beam is focused by a longitudinal magnetic field, these disturbances 

 propagate along the beam, and tend to increase in amplitude with dis- 

 tance, but not to change in frequency. A typical set of such frequencies, 

 within the range of the tunable receiver is as follows: 15.9, 24.3, 31.2, 

 34.0, 48.5, 63.4, 77.0, 108, 151, 166, 270.5, 372 and 481 mc. (During this 

 measurement, the anode voltage was 2,200, and the peak current about 

 40 ma.) 



No consistent relation could be found between these frequencies and 

 either the anode voltage or the cathode temperature, although unmis- 

 takable frequency changes did occur when these parameters were ma- 

 nipulated. Failure to establish such a relation may have been due to 

 uncontrolled drift in cathode activity. In any case, the measurements 

 did serve to narrow the field of possible mechanisms, by eliminating the 

 following : 



