SAMPLING INSPECTION TABLES 9 



use a complete set of such solutions for lot tolerance values from 0.5% to 

 10%. Each table is based on a particular value of lot tolerance per cent 

 defective, and each solution, comprising a sample size, n, and allowable 

 defect number, c, covers a range of lot sizes and a range of process average 

 values.* The value of n given in the tables is based on the largest lot size 

 for each lot size range, and the value of c corresponds to the mean lot size 

 in each lot size range and to the mean value of process average in each 



.PROCESS 

 N AVERAGE 

 ^ p = 0.45°/< 



ACCEPTANCE 



LINE 



LOT TOLERANCE 

 P4.= 3 % 



•SAMPLES OF I7( 



0.45% DEFECT 



PRODUCT 



CONS. RISK 

 R.=O.IO 



/ 





M 



/ 



T" 



-T 



PRODUCERS RISK 

 Pp=0.042 



> FROM 

 VE 



k 



■^ 



, — SAMPLES OF 170 

 FROM A 3 % DEFECTIVE 

 LOT OF 1000 PIECES 



\ 



\ 



\ 



\ 



^^ 



3 4 5 6 7 8 9 



NUMBER OF DEFECTS IN SAMPLE, d 

 I I I J 



2 3 4 5 



PER CENT DEFECTIVE IN SAMPLE 



Fig. 2 — Relation between consumer's risk and producer's risk 



process average range, as indicated in Fig. 3. This procedure is followed 

 for all of the sampling tables presented with this paper. 



For the lot quality protection tables for both single and double sampling 

 (SL and DL Tables), these choices are made to insure that, for the lot 

 size range covered, the risk will not exceed the specified value (0.10) and 

 to give on the average^ for the process average range covered, the most 

 economical plan. For reasons found advantageous in practice, sample 

 sizes for samples of over 50 pieces are given to the nearest 5 units. For 



* The extremely small process average range in the first column of each table has been 

 specifically provided for those cases, increasingly common with long continued use of 

 these inspection procedures, where the process average per cent defective is for all prac- 

 tical purposes zero. 



