NATURE OF POWER SATURATION IN TRAVELING WAVE TUBES 863 



agrees qualitatively with what would be expected from the associated 

 curve of beam current. 



To determine the curves of Figs. 11 and 12 is rather stretching the 

 accuracy of the measurements as can be seen by the large discrepancy in 

 the field calculated from the two parts of the velocity curve which of 

 course should be identical. The figures do give an interesting qualitative 

 picture of traveling wave tube behavior however, and are included here 

 for that reason. 



OVERALL VELOCITY SPREAD 



Of more practical importance is the overall velocity spread in the 

 spent beam. It is often desirable to reduce the power dissipation in a 

 traveling wave tube by operating the collector at a potential below that 

 of the electron beam, and it is interesting to see how far one might go. 

 Fig. 13 shows how the velocity reduction of the slowest electron, together 

 with the output level and fourier current components of beam current 

 vary with input level. For small amplitudes, the low level theory ac- 

 curately predicts the velocity, but near overload, as we have seen, the 

 minimum \'elocity drops sharply to a value several times lower than that 

 projected from small signal theory. 



The maximum velocity spread dependence upon the space charge 

 parameter QC is shown in Fig. 14. Similar data for values of the other 

 parameters may be obtained from the velocity diagrams. 



From the foregoing data, one can deduce the amount of reduction of 

 collector potential that should be theoretically possible wdthout turning 

 back any electrons. An idealized unipotential anode could collect all the 

 current at a potential AF (in the foregoing figures) above the cathode, 

 decreasing the dissipated power by a factor of AF/Fo below the dc beam 

 power. 



STOPPING POTENTIAL MEASUREMENTS 



Information on spent beam velocity has also been obtained by a stop- 

 ping-potential measurement at the collector of a more conventional 

 4,000-mc traveling wave tube.* Two fine mesh grids were closely spaced 

 to a flat collecting plate, and collector current was measured as a func- 

 tion of the potential of second grid. The first grid was very dense, to 

 prevent reflected electrons from returning into the helix. One curve taken 

 with this arrangement is shown in Fig. 15 and for comparison we have 



* Similar measurements have been reported by Atsumi Kondo, Improvement 

 of the Efficiency of the Traveling Wave Tube, at the I.R.E. Annual Conference on 

 Electron Tube Research, Stanford University, June 18, 1953. 



