Downloaded from https://www.everyspec.com I I I i 1 I I !i I SAMPLE SIZE ACCEPT NO. REJECT NO. - SINGLE SAMPLING N x 75 I 2 ---- DOUBLE SAMPLING Nis 50 0 3 1.00 N 2 , t00 (ISO) 2 3 MULTIPLE SAMPLING NI a 20 0 2 .90 N 2 , 20(40) 2 w N 3 s 20(60) 0 2 N 5 , 20 (2oo i 31 I. 7 0 - N 6 20(120) I 3 70 N7' 20(140) 2 3 .60 - - # ACCEPTANCE NOT PERMITTED UNTIL THREE SAMPLES HAVE SEEN INSPECTED. j.50 - PLANS BASED ON MIL'STD105A TABLE VIJ -AOL' 0. 65 * I 4 L9 .20 - - " i .10* 0 1 2 3 4 5 6 7 a 9 -0 1 p PERCENT DEFECTIVE Figure5-3.
Oterat g Characteristic (O0 curves for sigl.-, double-, and UPmltit4e -BA mplt lng tas It is readily observed that the OC curvesfor lot Although perceut sampUng appears logicalfrom sizes of 1,000 and over aro virtually equivalent a cost or time aspect, it is decidedly not so with to that for a lot size of infinity. Thus it is often respect to the deoree of protectionoffered. The said that the lot size has almost no effect on absolute size of the sample, it has been noted, . Ithe sampling plan. This statement holds, pro- affects the characterisUcs of a plan more sig- vided the sample size is a small fraction of the nificantly than the lot size.
A proportional lot size (normally less than 10 percent). Fig- change in both the sample and lot size, there- ures 5-6 and 5-7 demonstrate the effect of fore, will result in different degrees of pro- varying sample size for finite and infinite lot tection. Figure 5-8 Is an Illustration of this s i ze, respectively. phenomenon. It. 16