514 BELL SYSTEM TECHNICAL JOURNAL 



A power tube similar to the 1553 might therefore be larger in electrode 

 area, might have a coarser grid and wider input spacing, and perhaps 

 would differ somewhat in output configuration, particularly if the plate 

 voltage were raised. Any cathode development permitting a higher cur- 

 rent density would improve the power output more than the gain, and 

 might well lead to a drastic anode redesign to permit larger plate dissipa-' 

 tion. 



Similarly, a design to optimize noise figure would lead to still a third 

 version of the tube, in which one might consider such things as critical 

 relationships between input and output spacings. 



For the 1553 at 4000 megacycles the following quantitative data may 

 be quoted in order to check the gain-band product estimates.^ 



I F21 I = 39.10~^ mhos 



Gn = 73 • 10~ mhos 



Note that the transadmittance is less than the dc value of 45 • 10~' 

 mhos by only about 15%, while the input conductance, instead of being 

 equal to the transadmittance as at low frequencies, is almost twice as 

 large, on account of loading of the input gap by electrons returning to the 

 cathode. Using the active capacitance C22 of .477 /x/x/, the intrinsic gain 

 band product is: 



VB = F21 V47rGnC22 - 3480 megacycles. 



With the somewhat optimistic capacitance degradation factor of .81 com- 

 puted in Appendix 2, the gain band product would be reduced to 2820 

 megacycles. 



The experimental average value is about 1100 megacycles. The differ- 

 ence is probably due in part to resistive loss in the passive input circuit, 

 which may be calculated as follows: Neglecting feedback, the input' cir- 

 cuit may be represented as containing a resistance i?,, in series with the 

 short-circuit input admittance gn + jbn . Robertson gives the following 

 values for these elements: 



^11 = 73 • 10~^ mhos 



611 = 26-10-3 mhos 



Rg ^ 7.6 ohms 



Accordingly, the input degradation factor Ru/{Rii + Rs) should be 

 11.2/(11.2 -f 7.6) = .60, giving a computed overall gain-band product of 

 1690 megacycles. The best tubes sometimes exceed this figure. Tubes 



« S. D. Robertson's measuremcnls at 4000 megacycles, B. S. T. J., 28, 619-655 (Oc- 

 tober 1949). 



' A. E. Bowen and W. W. Miiniford "Microwave Triode as Modulator and Amplifier," 

 this issue of B. S. T. J. 



