INSPECTION IN MANUFACTURING PLANT 47 



An example of the operation of this chart is shown below: 



Example: 



AOQL = 1.5% 



Process Average = .56% 

 Lot Size = 4615 



On consulting the sampling table on the layout, which is information ob- 

 tained from the table for AOQL = 1.5%, Process Average Column .31 — 

 .60% (see Fig. 1), it will be found that for a lot of 4001-5000 parts the fol- 

 lowing sample sizes (SS) and allowable numbers of defectives (AN) are 

 shown : 



This means that for the first sample 55 parts should be selected and com- 

 pletely inspected for all items covered by the Sampling Layout. In order 

 that the per cent defective may be accurately determined for control pur- 

 poses, a complete inspection of the first sample must be made regardless of 

 how defective it may be. If five or more defectives are found in the sample 

 of 55 parts, the lot should be rejected or detail inspected, or otherwise dis- 

 posed of as shown on the layout. 



If no defectives are found in the 55 parts, the lot should be passed. 



If one, two, three, or four defectives are found in the 55 parts, the second 

 sample of 120 additional parts should be selected. 



In the combined sample of 175 parts, a total of four defectives is allowed, 

 if a total of five defectives is found before all of the 175 parts are inspected, 

 sampling should be discontinued and the lot disposed of as indicated on the 

 layout. 



If less than five defectives are found in the combined sample of 175 parts, 

 the lot should be passed. 



2.1 Counting the Lot Size 



The determination of the lot size may be made by weighing methods or 

 careful estimates. It has been found in practice that an estimate which is 

 within 20% of the true value of lot size is satisfactory for sampling purposes. 



2.2 Counting the Sample Size 



Since the relation between corresponding Lot Sizes and Sample Sizes 

 for a particular AOQL is not linear the same order of accuracy of count 

 does not prevail. If, in using a given sampling table, a sample smaller than 

 that prescribed is taken, the result is to increase the AOQL, and if a sample 

 larger than that prescribed is taken, the result is to decrease the AOQL and 

 increase the cost of inspection. 



Counting becomes a simple matter when articles are handled in compart- 



