866 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1956 



3. The observed effect of small beam radius upon efficiency is not as 

 pronounced as was found in the computations. The reason is not kno^^^l 

 but may be due to modulation of the beam diameter at large signal levels. 

 This effect would be neghgible with the larger Tro's, due to the focusing 

 fields being relatively much larger. 



4. The computations, and also those of Nordsieck, Poulter and Rowe^ 

 indicate a much higher efficiency than has been observed at elevated 

 beam voltages and small C and QC. The reason for this may be that 

 the limited number of "electrons" used in the computational models 

 fail to adequately account for the very sharp space charge cusp that 

 forms under low QC conditions, or that interpolation between their 

 points should not be linear, as assumed in making the comparison. 

 On the other hand it would be difficult to be sure that nonuniformities 

 in electron emission were not influencing the measurements in the case 

 of the large beams by giving a larger QC than calculated. 



5. The increase in efficiency to be had by elevation of beam voltage is 

 much smaller than is indicated by the computations. This may be a real 

 difference, or it may be that at elevated voltages, the measurements are 

 beginning to feel the influence of overloading in the attenuator. The 

 margin of safety on attenuator overloading is not as great as one would 

 like at the higher frequencies. 



6. The velocity curves, Fig. 8, compare the computed and measured 

 data on three runs. For small QC, Fig. 8(a), the agreement is remarkably 

 good considering the fact that in the computation only 24 "electrons" 

 were used to describe a rather complicated function. The effect of the 

 lumping of space charge in the artificial 'disc' electrons causes a scatter- 

 of points which is different from that in an actual tube as is especially 

 apparent in Figure 8c. In spite of this the computational results indicate 

 a velocity spread and current distribution not greatly different from that 

 observed. 



CONCLUSIONS 



The large scale model traveling wave tube is a means for the deter- 

 mination of non-hnear behavior, and has been valuable in determining 

 relationships and limitations important to efficient operation of such 

 tubes. It has shown that there is a broad optimum in tube parameters 

 around C = 0.14 QC = 0.2 and 7ro = 0.5 for which values it is possible 

 to obtain efficiencies well above 30 per cent. The measured ac beam 

 velocity and current near overload show that it is unlikely that signifi- 

 cant increase in efficiency can be obtained by any simple expedients such 



