Downloaded from https://www.everyspec.com A" 7MIN the actual a and 0 were shown, the source r - number of failures observed so far and explanation of the difficulties would be obvious, The 3 decisions a. each point are: Tables 3-5 and 3-6 illustrate the foilowing 1, Accept the lot If r'0) '4 rA (n) major points: 2, Reject the lot if r(n) ;P rX (n) 1, Sample size varies Invervely with ct, e, and 'y (NO flxeu). 3. Continije testing otherwise, 2. Sample size varies Inversely with KO Eqs, 3-12, and 3-13 were adapted from (y fixed). Chaptr V1ii of Ref. 23; Part il ot' Ret.
23 is an excellent discussion of sampling plane 34.2 SEQUENTIAL SAMPLING for variables and attributes, The probabilities of acceptance PF(R) have 5 special tiualitles: In order to construct a seq-tample plan, the acoopt and reject decision lines (such as PO(W) I those shown in Fig., -2(C)) are computed from: P. 1 ( 0 ) ml- a 4A Pn1 b- a RR/ (I ,/A D) ( 8) ,(A + 4A) (3.'14) (4 A U2i(/ PV(Oj) -0 4Alpe blI a, an nP 5 (Rt)1) b l oe These are nominal characteristich; the actunl bl + blo and 0 t am ,•nallor than shown. h A For seq-sampllng, the number of' itoms to kA W (ilntrcept) (3-12) be tested Is not predetermined, but Is a random variable whoa.
average Is a function of the true reliability, The average sample *n 0 T T (intercept) number ASNft (number of observations be- fore a decision Is reached) for incoming rell- Ordinarily, the a's, b's, h'W, and s are all ability levels of R a lR,, I -'s, R 10 is: positive; s is always between A. and H, A8ON, 'rhe accept line rA equation Is AO•Ij1o m (04A 4- oa,,i)I(R0b, + Bob,) VA N " k A4. + ,0 (3-13a) N (a A aj)(Rxbi) A814 1 . . (a•/~b,)(ajslbt) The reject line PR equation is ASNR 1 (PaA + n)/(Rlbi 4 Rib,) rX - hA + 8 (3-13b) ASNo , aa/bt whore where mt " the actual number of Items tested so ., I far tloI.