Downloaded from https://www.everyspec.com the results of the first sampling do not accept total sample of 60 units, 8 or fewer defectives or reject the lot. Depending on what is found are found, the lot is accepted. If 13 or more in the first sample, there are three possible defectives are found, then the lot is rejected. courseR of action. The lot may be accepted, it If more than 8 but fewer than 13 defectives are may be rejected, or the decision may be defer- found, a third sample of 30 is inspected. This red until the results of a second sampling are process is continued, until a decision is reach- obtained.
A decision is sure to be made on the ed, which in this case will be not later than the basis of a second sample, inasmuch as the re- seventh sample. jection number of a second sample 1a one more than the acceptance number. It is a common misconception that a compar- able double- or multiple-sampling plan is less An example of a double plan is as follows. A stringent than a single plan. This is not true, sample of 75 units is selected. If the sample since the acceptance and rejection numbers are contains 4 or fewer defectives, the lot is ac- selected to make them equivalent. The OC cepted.
If the sample contains 9 or more de- curves for equivalent single, double, and mul- fectives, the lot is rejected. If more than 4 tiple plans are shown in figure 5-3. defectives, but fewer than 9, are found, a second sample of 150 units is selected and inspected. 5-13. OC Curves for , Comparable Single-. If in the combined sample of 255 units fewer Doub and Multiple-Sampling Plans. It will than 9 defectives are found, the lot is accepted; be noted that the OC curves of the three plans but if 9 or more defectives are found, the lot is almost coincide, showing little difference in rejected. Double-sampling plans are sum- their severity and discriminatory power. How- marized as follows:
ever, there is a great difference among the plans in the average number of pieces per lot CumulatAve that will be inspected before a decision is sample Acceptance no. Rejection no. reached. Figure 5-4 shows the ASN (Average "Sample Number) curves for the same plans. 75 4 9 225 8 9 In the single-sampling plan the sample is al- ways completely inspected, regardless of a 5-12. Multiple-Sampli is a procedure in possible earlier decision. The ASN curve for which a final decision to accept or reject the the single-sampling plan is, therefore, a specific lot need not be riade alter one or two straight line. samples, but may require the drawing of sev- eral samples.
An example of a multiple-sam- In the double-sampling plan the first sample is pling plan is as follows: completely inspected, kt inspection of the second sample ceases as soon as the number of Cumulative defectives found in the combined samples sample Acceptance no. Rejection no. equals the rejection number. Since lots of very good quality will be accepted, and lots of 30 2 8 very bad qu dity will be rejected on the first 60 8 13 sample, 4nd inspection of the second sample is 90 12 18 curtailed, the average amount of inspection will 120 17 22 be less for a double-sampling plan than for a 150 21 27 single-sampling plan.
As will be seen, mul- I80 27 32 tiple-sampling reduces still further the amount 210 35 36 of inspection. The inspector selects a sample of 30 from a 5-14. Relationship of Sample Size to Lot Size. Lot of 4,000 itcimb. U the first sample contains The sample size is of considerably more sig- 2 or fewer defectives, the lot is accepted; if 8 nificance, with respect to severity and d's- or more defectives appear in the sample, the criminating power, than the lot size from which lot is rejected. U more than 2 but fewer than it is taken. Figure 5-5 is composed of OC 8 defectives appear in the sample, the inspector curves, all having the same sample size and draws a second sample of 30 items.
If, in the acceptance numbers, but different lot sizes. 5-6