Downloaded from https://www.everyspec.com AMC706 7"19 Example No. 28 Same data as Example No. 27 but ause Ref. 17 to solve the problem, Procedure Examp e 1. Pick several plans that have 1. In version B, Rev. July 1969, there are 2 S- 10%. plans which might be applicable: V and VI. 2. Convert the time axis from kT 2. Multiply times in Plan V by V 3 3, and in to XtT, because this example Is Plan VI by v - 5. given in terms of Xj. 3. Find XIT. (truncation). 3. Plan V. XJTX,, 10.35 Plan VI, XT - 6.25. 4. Choose a plan. 4. Because of the "Tmx - 8000 he" (XiT - 6) constraint, Plan VI in preferred over Plan V. It has m - 5, and actual a, P of about 13%, (see p. 60, Ref. 17). 5.
Check the OC curve and an-ex- 5. (See p. 88, Ref. 17). To get a rejection pected test time curve, probability of 4% requires X/hi w 10, and of 2% requires X/.I ow 15. The o-expeoted test time for very good equipment is about "_/ht - 3000 hr." So the test plan that fits the problem constraints requires, in essence, that the equip- ment be 10 times as good as required, just to keep from having an unreasonable chance of failing the test when the equipment is much better than needed. This example illustrates the paradox brought on by the exponential distribution and expensive tests. 3-37