58 BELL SYSTEM TECHNICAL JOURNAL 



device was very large, for tlie velocity of the outer beam was far from that 

 at which interaction with the helices occurs. 



A signal from the probe at the end of the central portion of the tube 

 23 db greater than that from the probe at the beginning of this section was 

 observed in a comparison of the second type. This can probably be taken 

 as a measure of the increase in signal in this portion of the tube due to the 

 double-stream interaction alone, although the probe arrangement may also 

 be subject to some remaining complicating effects. Overall gain for the 

 device in this measurement was 32 db. Further interaction of the same 

 kind occurs in the portions of the tube outside of the space between the 

 probes. 



Measurements of the gain of an amplifier with helix output as a function 

 of velocity separation between the streams have been made. For fixed 

 mean voltage and current, theory predicts an increase in gain from zero db 

 at zero separation to a maximum and then a decrease to zero as the velocity 

 separation is further increased. A maximum gain was observed experi- 

 mentally as velocity separation was varied, and in the neighborhood of the 

 predicted optimum value of velocity separation for the current used. 



In the amplifier tube with gap output it was possible to evaluate the a-c. 

 component of current in the electron stream produced by the amplified 

 signal since the impedance across the gap was known. The power output 

 from this tube at saturation was 0.1 mw, a little less than the maximum 

 shown in Fig. 3. For 75 ohm output impedance this power corresponds 

 to 1.15 milliampere r.m.s., or about one third of the total d-c. current to the 

 collector in both streams. The output power, although relatively low, is 

 thus of the right order of magnitude for the currents used. 



Acknowledgment 



The writer wishes to acknowledge his indebtedness to J. R. Pierce for 

 valuable suggestions and discussion, and for supplying unpublished calcu- 

 lations concerning the relation between hollow and solid beams, the efifect of 

 the proximity of the conducting wall, and the effect of the separation of 

 the beams in space. 



Thanks are also due to A. R. Strnad for assistance in mechanical design 

 and to R. E. Azud for construction of the amplifier tubes. 



