DESIGN FACTORS OF THE 1553 TRIODB 505 



whereas without such conditions not only was the yield low but it was 

 difficult to ascertain just what factors were operating to inhibit emission 

 and to cause cathode-grid shorts. 



A summary of the pertinent low-frequency characteristics of the 1553 

 triode is given in Table I. It should be noticed that, at plate currents of 

 25 milliamperes, the transconductance per milliampere is about 2000, that 

 is, about one-fifth of the theoretical upper limit. At lower currents this 

 ratio is higher: at 10 milliamperes, for example, it is 3000 micromhos per 

 milliampere. Diodes with the same spacings have about twice these values 

 of transconductance per milliampere, showing that the grid is fine enough 

 to obtain fifty per cent of the performance of an ideal grid. 



Triode Design Requirements 



Analysis of the figure of merit can well begin by devoting attention to 

 the band-limiting capacitance Cout of the output circuit. First, some ques- 

 tion may be raised as to the applicability of the concept of a simple 

 L-C shunt resonant circuit at high frequencies, where the circuit parame- 

 ters are actually distributed, not lumped. Suppose the actual circuit 

 admittance is Yx = Gx + jB^. In order to represent it as a simple shunt 

 resonant circuit of admittance Vp = Gp -\- jo^Cp + 1/joiLp, we need only 

 require that the two be equal and have equal derivatives with respect to 

 frequency at the center frequency /o = wo/27r. Accordingly the "effective 

 values" of the actual admittance are given by the following equations: 



Gp = Gx (cco) 



Cp = \{Bx + 5x/coo) (3) 



1 



u 



\{(ji(? Ex — OJoBx) 



From this development one sees that the representation neglects Gx, 

 the first derivative of the conductance, but otherwise is correct to first 

 order as a function of frequency. 



There are important cases where this representation as a simple circuit 

 does not hold. For example, double-tuned circuits having two local reso- 

 nances have a fundamentally different band shape. However, such compli- 

 cation of the circuits has been excluded from the figure of merit on the 

 ground that it is purely a circuit "broad-banding" problem: having de- 

 termined the performance of the tube for simple circuits, any broad- 

 banding (double-tuning, staggering, etc.) will give a calculable improve- 

 ment which does not depend upon the tube. Accordingly, to compare 

 tubes it is sufficient to consider standard simple circuit terminations, 

 tuned to the same frequency. 



