ELECTRON TUBES FOR A COAXIAL SYSTEM 1121 



these parts is difficult when production numbers are considered. The micas 

 in particular presented a serious problem. Mica sheet is comf)osed of a large 

 number of laminations many of which are of the order of 0.0001" in thick- 

 ness. When the mica discs are punched out, these laminations leave, not 

 smooth edge holes as do metal stampings, but rather a large number of 

 minute jagged edges. The method used to check these was an optical one 

 in which the mica was projected at about 40 times size onto a glass screen 

 on which engraved lines acted as go-no-go gages. This reduced tool and 

 human error considerably. The cooperation of several industrial concerns 

 which supply some of the critical parts and the measuring instruments 

 was very helpful in obtaining the desired tolerances. 



It was evident from the start of the development of the L3 tubes that 

 the performance requirements for high gain conventional structure tubes 

 would be pushing to the limit the available process controls and measuring 

 techniques. A statistical quality control program was put into effect on the 

 tubes after the final laboratory design had been crystallized. The statistical 

 study covered the tube dimensions and the data collected from those tubes 

 after they had been processed. The net result of the study was to indicate 

 that better measuring methods and process controls are needed. 



With the amount of d-c. feedback employed in the working circuits, the 

 space current does not vary too rapidly with tube geometry. In the case 

 of the most critical spacing, that between the grid and the cathode, a 10% 

 change in the spacing would be expected to cause only about 2.5% change 

 in space current. However, the transconductance is more sensitive to the 

 grid-cathode spacing, with a 14% change in transconductance to be ex- 

 pected for a 10% change in the spacing. This comes about because the trans- 

 conductance is a function of the spacing, even at a fixed space current. 



Since a 10% change in spacing is only 0.00025 inch, the importance of 

 close tolerances on the parts dimensions controlling it is evident. The test 

 specification limits on transconductance permit a variation of about ±25%, 

 so that the 0.00025 inch change in spacing would use up over half of the 

 allowed deviation. Preproduction runs at the Laboratories have shown 

 that the tubes are practical and that their performance in the amplifier 

 circuits has justified their design. 



3.2 Electrical Characteristics 



The nominal electrical characteristics are shown in Table II. The corre- 

 sponding characteristics for the earlier types 386A, 6AK5 and 404A are 

 also given for comparison. The last row in the table shows figure of merit 

 values which are a measure of the circuit performance. The tabulated values 

 for the figure of merit were calculated, taking into account the effect of 



