652 



BELL SYSTEM TECHNICAL JOURNAL 



determined values of Cn and C22 plotted as a dashed curve. The theoretical 

 values (neglecting fringing capacitance) are shown by the solid curve. Since 

 fringing was neglected, it is not surprising that the measured capacitances 

 should exceed the calculated values by the amount shown. 



(5 10 



Q 

 < 



10 



12 



13 



01 23456789 



SPACING, S, OR S2,IN MILS 



Fig. 5 — Comparison of theoretical and experimental values of input and output ca- 

 pacitances. 



The magnitudes of yn for each grid over a range of values of Si and Si 

 are shown in Figs. 6 and 7. It will be noted that, for a given set of spacings 

 5i and S-i , yu is much greater for the parallel wire grid than for the cross- 

 lateral. This is the sort of result one would e.xpect if y^ resulted from electro- 

 magnetic coupling through the grid, since the parallel wire grid would be 

 expected to offer a better transmission path than the cross-lateral grid. 

 It was not practicable with the equipment used in these experiments to 

 measure the values of yu at low frequencies where yn would be determined 

 by the cathode-plate capacitance. Data were available, however, for the 

 low-frequency, cathode-plate capacitance of the standard, parallel-wire 



