1002 THE BELL SYSTEM TECHNICAL JOURNAL, JULY 1953 



temperature or a combination of both. Homogeneous lots of grids, are 

 next run from this spool of wire and are processed up to the sizing opera- 

 tion using the same standard conditions. This test lot of 25 grids is run 

 whenever a new spool of wire is introduced. 



The first few grids from each lot are sized on the standard sizing blades 

 and checked against a reset-run chart, ^ a form of sequential analysis. 

 The set-up is satisfactory if the distribution of the minor axis dimension 

 is found to be within 1-sigma (O.OOOS'O of nominal. If grids pass the 

 reset-run chart the balance of the lot is sized on the standard blades. 

 If the sample does not conform to the reset-run chart, a final correction 

 is made by substituting a larger or smaller set of sizing blades as required. 



The distribution of the minor axis after sizing is plotted on control 

 charts. A sample of 5 grids is taken from each tray of 36 grids and X 

 and R plotted. If in control the lot is passed. If out of control, the lot 

 is 100% inspected and packaged into three cells. The grids are then 

 shipped to the mount assembly line in the usual 1-3-1 or 0-5-0 distribu- 

 tion. By application of the above controls of processing, variation in the 

 process average from lot to lot has been reduced to ±0.0003''. 



In order to keep an accurate record of the grid lots, relating the pro- 

 cesses used to the resulting grid distribution, a special lot ticket has been 

 developed. This ticket contains information such as specified tension 

 and temperature along with the reset-run chart and XR chart. A ticket 

 of this type accompanies each lot of grids through processing. 



3.635 4S7A Plate 



Since the 435A and 436A are tetrodes, the dimensions of the plate in 

 these tubes are not nearly so critical as for the 437A tube, a triode. In 

 the 437 A, the plate assembly is made up of two sections welded together 

 on a mandrel. An air press sizing operation is employed to control the 

 inside diameter of the assembly. The drawing requirement for the plate 

 assembly minor axis inside dimension was set at 0.0750 db 0.0005". 

 Initial production was measured using a tool maker's microscope sighted 

 on the edge of the plate assembly. Misalignment of the part under the 

 microscope and burrs on the edge of the material contributed to variation 

 in the measured values. In addition, the measurement was made at the 

 edge of the assembly rather than the center which was desired. Results 

 of measurements made with the tool maker's microscope are shown on 

 the first portion of Fig. 12. The plate assemblies appeared to be oversize, 

 although this was at variance with electrical test results on the tubes 

 made from these parts. The standard deviation of 0.00101" and 3-sigma 

 limits of 0.00303" were entirely too wide to meet the design intent. 



