ECA TEP140-1974 Proposed Acceptance Sampling Procedures for Small Lots《提议的小批量验收样品程序》.pdf

1、I c- - e EIA TEP140 74 = 3234600 0008387 2 APRIL 1974 PROPOSED ACCEPTANCE SAMPLING PROCEDURES FOR SMALL LOTS FORMULATED BY JEDEC ELECTRON TUBE COUNCIL JEDEC PUBLICATION NO. 40-A EIA TEP140 74 W 3234b00 INTRODUCTION O0 0838 m These procedures were formulated by JEDEC Committee JT-16 on Government Lia

2、son and approved by the JEDEC Electron Tube Council. It is a revision of the 1963 publication which first expressed the application of these statistical principles. Definitions and symbology follow those of MIL-STD-105D. Published by ELECTRONIC INDUSTRIES ASSOCIATION Engineering Department ZU01 Eye

3、Street, N.W., Washington, D. C. 20006 e EIA TEP140 74 W 3234600 0008389 b W The problem of Verifying the quality of small lots has loaw larger because of the -1ications posed by the publication of P%R-1, sometimes referred to as the Darnell report. develaped by an electron tuk Carranittee, the plan

4、is perfectly general in application and can be applie wherever the small lot problem exists. -le the approach outlined in this rep i.e., in table 2 (CH) at the rm (n) colurrm (N) intersection for n=10 and N=100 again noting PA(.95)=0.5% and PA(.10)=20%. Gohg down the colmm we note that PA(.lO) decre

5、ases as the sample size increases. Lets say PA(.10)=4.0% muld be an acceptable value for an LTPD. We note that our sample size has quahpled, n nm equals 40, but worse, we note that the PA(.95) value has dmppd from 0.5% to 0.1%. has increased substantially. uts go to table 2 (c=i) and again start at

6、the um (n) column (N) intersection for n=lO, N=lOO and note that PA(.95)=3.9% and PA( .19)=33%. These are not acceptable so we go dm die column again looking for a suitable CanQdse between sample size, PA( .95), and PA( -10). th we get to a PA(.lO) that falls belm 4.0% were at n=80 and PA(.95)=1.1%.

7、 Here a PA(.95)=1.1% seems acceptable as a ccmpr&se but this is only achieved at scmie expense (i.e. , 8 times the ariginal sqle size). If this value of PA(.10)=20%. is not Lets see how these tables might help us “his is not a gcd ccmprdse as our risk of rejecting good lots By the It should be appar

8、ent that in small lot prcduction certain degrees of statistical assurance are prohibited if sampling is to be used at all. of very high egrees of quality or reliability in the cases of small lot manu- facture must. be obtained by other thds. 1. The performance of a nondestructive test, which is ham

9、to correlate with the desired use condition, on all prcduct as a 100 percent screening. The nonquantitative approach calling for the utmst in care, control, and attention at all pints in the manufacturing process. Therefore, assurance Included aniong these might be: 2. 3. Perfmce of a 100% screening

10、 test. IV. EIA TEP140 74 3234b00 0008391 4 3 Sampling Procedures It& suggested that the acoeptance sarqling operating procedures for sml lots utilize all applicable procedures contained in ML-SD105D except that Reduced 3i-1pctior-1 n& be allawed. %en Tightened inspection is required, go to the next

11、largest sample size in the table being used. For example, if n=lO c.4 is in use and the switching rules require Tighten& rilspection, go to n=16 o=O (see table i) v. canp ling Plan Specification Provision for dl lot sanpling should be made in the lot definition for tfae particar prcduct you are work

12、ing with. That is, a statenmt in the lot definition should be made which allm use of the small lot sqling tables. For exanple: One definition which has been used is: “If the production rate is 25ounitc or less for a maximum period of seven consecutive calendar days, the provisions given in “Hypergec

13、mtric -ling Plans for ml iots“ may be used. “ -ling plans for smal lots are given in tables 1, 2 and 3 which oorrecpond to plans with acceptane nmbers of zero, one, and tm, respectively. The tables are entered by lot size and the desired AQL. any sqling plan (n,c may be used for which the value of P

14、A(. 95) is equal to or less than the desired AQL. The values of PA(,lO) given in the tables correspond to LTPD (iot,Tolerance Percent Defective). The PA(.lO) is given so that the risks associated with small lot carpling are knam. in the lot size column, EIA TEP140 74 8 O v) CJ O O OJ O W d 5: rl O C

15、u rl O O rl O W O W % 4 O c3 O hl O rl z O rl VJ o) O O ?I VJ o) a O rl u3 o) b O rl O v) o) O O rl . m o) . O 4 O u3 o) . O rl . v) o) 2 O v) o) O FI O 3 O O rl v) o) O rl u3 CR l . e . c I 1 I PROBlEfl4 HARD COPY es EIA TEP140 74 3234600 0008393 8 8 O v) cu O O F1 O 3 8 rl O cu rl O O rl O a3 O (o O m s O m O CJ O rl z O & y$ - PROBLEM HARD COPY EIA TEP140 74 = 3234b00 0008394 T 8 O CR cu O O Cu O (o ri I I I %S% I