546 BELL SYSTEM TECHNICAL JOURNAL 



"static compression" exists, but rather with how much gain can be real- 

 ized without exceeding the dissipation ratings of the tubes. 



With this in mind data were taken on 25 of the experimental tubes. In 

 each case they were matched to the input and output waveguides and the 

 cathode current was stabilized at 30 ma. After driving the tube to a 

 high level of output power, the circuits were rematched and the resulting 

 "compression" curves revealed the capabilities tabulated. 



Table II 

 Summary of Data on 25 Experimental Close-Spaced Triodes 



Low level gain 



Gain (500 mw output) . . . . 

 Power Output (3 db gain) . 



Highest 



12.3 db 



7.0 db 



950 mw 



Lowest 



3.8 db 

 -8.0 db 

 50 niw 



Average 



7.8 db 



1.82 db 



455 mw 



It can be seen from the table that we might expect to obtain a gain of 

 20 or 25 db with three or four stages with a power output of about 500 

 mw and a flat band of over 20 mc. 



Three St.\ge Amplifier 



A three-stage amplifier with 24 db gain has been assembled using an 

 earlier type of circuit and loop tested at low levels on the equipment of 

 Messrs. A. C. Beck, N. J. Pierce and D. H. Ring.^ This amplifier had a 

 bandwidth of about 30 mc to the 1 db points and while it does not 

 represent the best that can be done with the 416A tube, the results of the 

 loop test are interesting. 



The recirculating pulse test, or loop test, is performed on a repeater 

 component to determine its ability to reproduce a pulse faithfully after 

 repeated transmissions. The output of the amplifier is connected to its 

 input through a long delay line and an adjustable attenuator. The overall 

 gain of the loop thus formed is adjusted to unity or zero db so that an 

 injected pulse will recirculate through the loop without attenuation but 

 accumulating distortion with each round trip. After allowing the pulse to 

 recirculate long enough the amplifier is blanked out or quenched and the 

 recirculating i)ulse amplitude dies out, thus preparing the loop for the 

 next injected pulse, when the process is repeated. With a pulse length of 

 one microsecond and an overall delay of two microseconds, one hundred 

 round trips occur in 0.2 milliseconds, thus allowing the process to be 

 repeated at the rate of two or three thousand times per second. A cathode 

 ray oscilloscope is used to examine the pulse shapes, and its sweep is 

 synchronized to the injected pulse so that successive corresponding pulses 

 are superposed, enabling the operator to examine the pulse after any 



