GRID CURRENT MODULATION . 115 



coefficients to b's, and the Zk of the grid circuit to the Zk of the plate 

 circuit, we obtain the expressions 



T _ b, / ^v y y (12) 



-^^ (1 +^Zo)Vi +6iZj 'J 



Now it is apparent that each of these two terms contributes something 

 to the sideband frequency — the first by amplification of the grid 

 sideband potential, and the second by modulation of the two funda- 

 mental components in the plate circuit. The net sideband current in 

 the plate circuit may accordingly be expressed as 



/. = 



M^2 bia^Zj 



(1 + 6iZ^)(l +b,Z,) 1 -faiZ, 



nEpE 



y 



(13) 



+ &iZ.)(l +aiZp)(l -faiZJ 



Under normal conditions both grid current and plate current charac- 

 teristic curves are concave upward, so that bi, b^, and a^ are all positive. 

 The two terms of eq. 13 are then in phase opposition, a condition 

 which is responsible for failure to work certain modulators to fullest 

 advantage. For efficient plate modulators the grid modulation term 

 should be suppressed, which may be accomplished by making the 

 external grid impedance to the modulated product of interest equal to 

 zero, or by keeping the grid potential negative at all times. For 

 efficient grid modulators the plate modulation term should be reduced 

 to a minimum by suppressing the fundamental currents in the plate 

 circuit, in which case Zp and Zq are made large. Of course the 

 possibility exists of phasing the two sideband components to add 

 rather than to subtract (arithmetically) — and this, it will be readily 

 seen, is obtained by having the phase angle of the entire plate circuit 

 approach 90° at each fundamental frequency when the grid circuit 

 sideband impedance is large. This condition cannot be met without 

 lowering the amount of plate current modulation, so that the first 

 mentioned plate circuit condition is the more practical one. 



Other possibilities of more favorable phasing exist by working 

 within appropriate regions of tube operating characteristics where 

 either 62 or a^ becomes negative. Generally speaking, operating 

 points of this nature are not stable with variations in tube potentials, 

 nor are they adaptable to large power outputs approaching the 

 maximum load capacity of the tube. Finally, for straight amplifi- 

 cation purposes the two terms of eq. 13 should be made equal and 

 opposite in sign. 



