1、 JEDEC STANDARD Assessment of Average Outgoing Quality Levels in Parts Per Million (PPM) JESD16B (Revision of JESD16A, April 1995, Reaffirmed September 2008) NOVEMBER 2017 JEDEC SOLID STATE TECHNOLOGY ASSOCIATION NOTICE JEDEC standards and publications contain material that has been prepared, review
2、ed, and approved through the JEDEC Board of Directors level and subsequently reviewed and approved by the JEDEC legal counsel. JEDEC standards and publications are designed to serve the public interest through eliminating misunderstandings between manufacturers and purchasers, facilitating interchan
3、geability and improvement of products, and assisting the purchaser in selecting and obtaining with minimum delay the proper product for use by those other than JEDEC members, whether the standard is to be used either domestically or internationally. JEDEC standards and publications are adopted witho
4、ut regard to whether or not their adoption may involve patents or articles, materials, or processes. By such action JEDEC does not assume any liability to any patent owner, nor does it assume any obligation whatever to parties adopting the JEDEC standards or publications. The information included in
5、 JEDEC standards and publications represents a sound approach to product specification and application, principally from the solid state device manufacturer viewpoint. Within the JEDEC organization there are procedures whereby a JEDEC standard or publication may be further processed and ultimately b
6、ecome an ANSI standard. No claims to be in conformance with this standard may be made unless all requirements stated in the standard are met. Inquiries, comments, and suggestions relative to the content of this JEDEC standard or publication should be addressed to JEDEC at the address below, or refer
7、 to www.jedec.org under Standards and Documents for alternative contact information. Published by JEDEC Solid State Technology Association 2017 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 This document may be downloaded free of charge; however JEDEC retains the copyright on this
8、material. By downloading this file the individual agrees not to charge for or resell the resulting material. PRICE: Contact JEDEC Printed in the U.S.A. All rights reserved PLEASE! DONT VIOLATE THE LAW! This document is copyrighted by JEDEC and may not be reproduced without permission. For informatio
9、n, contact: JEDEC Solid State Technology Association 3103 North 10th Street Suite 240 South Arlington, VA 22201-2107 or refer to www.jedec.org under Standards-Documents/Copyright Information. JEDEC Standard No. 16B -i- ASSESSMENT OF AVERAGE OUTGOING QUALITY LEVELS IN PARTS PER MILLION (PPM) Contents
10、 Introduction 1 Scope Pageii 12 Application 13 Reference publications 14 Definitions 15 Symbols 36 Assumptions 47 Procedures 57.1 Data accumulation for a single inspection 67.2 Computing AOQ and variance for a single inspection 67.3 Combining inspections within a group 77.4 Combining groups within a
11、 class 77.5 Combining classes 87.6 Data exclusion 87.7 Minimum total sample size 87.7.1 Minimum total sample size criterion (graphic method) 97.7.2 Minimum total sample size criterion (numeric method) 108 Reporting 10Annexes A Examples 12B Derivation of the minimum sample size criterion 16C Derivati
12、on of the standard deviation of AOQ 20D References 30E Differences between JESD16B and JESD16A 31JEDEC Standard No. 16B -ii- ASSESSMENT OF AVERAGE OUTGOING QUALITY LEVELS IN PARTS PER MILLION (PPM) Introduction Improvements in manufacturing technology and methodology have resulted in corresponding q
13、uality improvements for electronic devices. As a result, the traditional measure for reporting average quality levels in percent nonconforming needs to be replaced with a quantity more in line with the quality levels of today. That measure is parts-per-million or ppm. It is simply clearer to report
14、estimated average device quality as 10 ppm, rather than the more cumbersome 0.001%. The ppm terminology applies to estimating the average outgoing quality (AOQ) level of a device, from lot acceptance results. This standard was developed to provide a uniform method of measurement and calculation of a
15、verage outgoing quality levels. Minimum sample sizes and a method for aggregating data are provided. JEDEC Standard No. 16B Page 1 ASSESSMENT OF AVERAGE OUTGOING QUALITY LEVELS IN PARTS PER MILLION (PPM) (From JEDEC Board Ballot JCB-17-18 formulated under the cognizance of JC-13 Committee on Governm
16、ent Liaison.) Scope 1This standard is intended to provide a uniform method of determining fraction nonconforming in finished devices and to provide a standardized definition of the quality index referred to as Average Outgoing Quality (AOQ). The method used here is primarily directed at devices whos
17、e production or procured volume is large enough, during some predefined sampling period, to give statistically meaningful information. Application 2This standard is intended to provide a method for the derivation and reporting of fraction non-conforming. The AOQ philosophy applies to the estimation
18、of the average quality level of a product, not to lot acceptance plans. Since it is necessary to focus on accumulated lots to generate sufficient data for device quality characterization, the method in this standard should not be used to establish ppm levels for individual lots or to form the basis
19、for determining acceptability of product on a lot-by-lot or batch-by-batch basis. Reference publications 3Publications referenced by this document are listed in Annex D. Terms and definitions 4For the purpose of this standard the following terms and definitions shall be used: acceptance inspection:
20、A sampling inspection or series of sampling inspections used to determine the suitability of a lot of material for shipment. NOTE The accumulation of acceptance inspection data is used to determine average outgoing quality (AOQ). accept number (c): The maximum number of nonconforming devices in the
21、sample for which acceptance of the lot is allowed under the sampling plan. average outgoing quality (AOQ): The expected population average nonconforming, in parts per million, estimated from a series of lots. JEDEC Standard No. 16B Page 2 4 Terms and definitions (contd) class: A categorization of si
22、milar characteristics for the purpose of reporting parts-per-million (ppm) nonconforming. NOTE Examples of classes include functional (ppm1), electrical (ppm2), visual/mechanical (ppm3), hermetic (ppm4). fraction nonconforming: The unknown noncomforming proportion of the total population of devices.
23、 NOTE Estimates of fraction nonconforming are derived from samples. group: A subdivision of a class based on inspection conditions or criteria, e.g., device type, product family, test temperature, sample size. NOTE Lots that are members of the same group receive the same set of inspections in that g
24、roup. inspection: The assessment of a characteristic and its comparison to a standard. NOTE Examples of inspections include low-temperature electrical test, room-temperature test, and visual inspection. lot: An aggregate of devices from which the sample is selected. lot acceptance rate (LAR): The ra
25、tio of lots inspected over a sample period. nonconformance: A device characteristic that does not conform to an individual specified criterion. nonconformity: A single device that has one or more nonconformances. NOTE These nonconformities may be placed into classes for reporting purposes. parts per
26、 million (ppm): The unit of measurement used to describe Average Outgoing Quality giving the number of nonconformities for each million units. sample period: The period of time selected by the manufacturer to accumulate data for the calculation and reporting of average outgoing quality (AOQ) or lot
27、acceptance rate (LAR). JEDEC Standard No. 16B Page 3 Symbols 5The following symbols are used in this standard: Symbol AOQ Formula Definition True value of the average outgoing quality. AOQ Estimate of AOQ. c Sample plan acceptance number. L Total number of lots inspected. LR Total number of lots rej
28、ected. LAR 1(LR/L) Lot acceptance rate. m S/L Average sample size per lot N Total number of devices in lots inspected. D Total number of nonconforming devices found from all samples.S mxL Total number of devices sampled. p D/S Fraction of nonconforming devices found in all samples. UB1Upper bound of
29、 a one-sided confidence interval for AOQ. LB2Lower bound of a two-sided confidence interval for AOQ. UB2Upper bound of a two-sided confidence interval for AOQ. 2Variance of AOQ. 2Standard deviation of AOQ. JEDEC Standard No. 16B Page 4 Assumptions 6The following assumptions are used by this standard
30、. Users should confirm that these assumptions are met for the device whose quality is to be reported. 6.1 Attribute sampling inspection is being conducted on devices that have completed manufacturing processes affecting the criteria being reported. 6.2 Equipment must be calibrated according to ISO 1
31、0012-1, ANSI/NCSL Z540.1, ANSI/NCSL Z540.3, or ANSI/ASQC-M1. Accuracy and precision must be in accordance with manufacturers internal documents or applicable procurement requirements. Statistical controls for the test process shall be in accordance with JESD557. For the purpose of estimating AOQ, te
32、st equipment can be used interchangeably provided the requirements of this paragraph are satisfied. 6.3 Lots of devices that fail acceptance inspection are reprocessed 100%, and all nonconforming devices are removed from the lot or the lot is removed from consideration for shipment and discarded. Th
33、e nonconforming devices that were removed may or may not be replaced by conforming devices. In the case of c0 sampling schemes, nonconforming devices found in the sample of accepted lots are removed from the sample and may or may not be replaced by conforming devices before the sample is returned to
34、 the lot for shipment. 6.4 All confirmed nonconforming devices observed during the first submission to acceptance inspection are included in the calculation of the total number of nonconforming devices found in samples (D) unless excluded according to 7.6. Data from resubmitted lots are not used in
35、the calculation of AOQ. 6.5 Single, multiple or skip lot sampling may be used if appropriate (e.g., multiple temperature or hermeticity inspecting). Double counting of nonconforming devices is not allowed. 6.6 The total number of devices sampled (S) shall include only devices actually inspected. 6.7
36、 AOQ values shall be reported for inspected characteristics only. Characteristics actually inspected and accompanying inspection results comprising the data for AOQ calculations shall be included in the manufacturing internal documentation and shall be available for review upon request. 6.8 Nonconfo
37、rmities that are not device-related shall not be included in the calculation of AOQ. 6.9 A single device with nonconformities in more than one class should be included only in the class that would have the most severe impact. 6.10 The estimator of AOQ given in equation (1) of 7.2 is appropriate if,
38、for each lot, the sample size for the lot does not exceed 10% of the lot size. 6.11 The estimates of p and LAR shall be determined using the same samples. 6.12 All devices selected for acceptance inspection are assumed to be randomly sampled from the population being analyzed. (see Hahn and Meeker (
39、1991) P10ff). JEDEC Standard No. 16B Page 5 Procedures 7Figure 1 Procedure Flow Chart YES (all groups) Minimum Sample Size Criterion See 7.7.2 NO (any group) Collect N,D,S,L,LR (each inspection) see 7.1Compute LAR,p,m (each inspection) see 7.2Compute AOQinsp,2insp(each inspection) see 7.2 Collect Dg
40、rp,Sgrp,Ngrp,Lgrp, AOQinsp(each group) see 7.3Compute pgrp,AOQinsp,LB2grp, UB2grp,2grp(each group) see 7.3Is (0.7)( UB2grp)0 is: UB1= (AOQ) nullexp null1.282AOQnullnull (7) When D = 0 the upper 90% confidence bound (UB1) given by Louis (1981) is: UB1= (1 (null/null) 106 (8) where = 0.10. 7.7.1 Minim
41、um total sample size criterion (graphic method) Figure B.1 in Annex B can be used to determine if the minimum total sample size requirement in 7.7 is achieved. The x-axis is AOQ in units of ppm. The y-axis is the total number of lots sampled (L). If the average single lot sample size (m) is between
42、100 and 200, then use the m = 100 minimum sample size line. If the average single lot sample size is greater than 200, then use the m = 200 minimum sample size line. If the average single lot sample size is less than 100, then use the numeric method in 7.7.2. If the intersection of AOQ and L falls a
43、bove the appropriate minimum sample size line on the graph, then the minimum sample size has been achieved. Example Suppose the single lot sample size is m = 125, that the AOQ for these samples is 300 ppm, and that L = 2000 lots were inspected. Since m = 125 is between 100 and 200 use the m = 100 li
44、ne. The intersection of the 300 ppm vertical line and the 2000 horizontal line falls above the m = 100 minimum sample size graph line on Figure B.1. Therefore, the minimum sample size is achieved. JEDEC Standard No. 16B Page 10 7.7.2 Minimum total sample size criterion (numeric method) A numeric min
45、imum sample size test may be performed when desired. This procedure is used when results from the graph method are ambiguous (e.g., when the intersection of AOQ and L is very close to the appropriate minimum sample size line on Figure B.1) or when the average single lot sample size is less than 100.
46、 The numeric method requires computing a two-sided confidence interval for AOQ using equation (9), then checking to see if the AOQ estimate is within 30% of each confidence bound. This numeric method is illustrated by example in annex B. An approximate two-sided 90% confidence interval given by Hahn
47、 and Meeker (1991) when D0 is: nullLB2, UB2null = nullAOQexpnull1.645AOQnull,(AOQ) exp null1.645AOQnull null (9) NOTE This confidence interval is not symmetrical around AOQ. The minimum sample size requirement is satisfied if (0.7)(UB2) AOQ (1.3)(LB2). (10) Reporting 88.1 To provide consistent repor
48、ting, this standard shall be specifically referenced if it is stated or implied that the AOQ has been derived in compliance with this standard. 8.2 Alternative estimators of AOQ may be used instead of the JESD16 estimator if comparable accuracy and precision can be demonstrated. Rules would need to
49、be derived for the alternative estimator that govern the minimum sample size and confidence bounds. Conditions under which to apply the minimum sample size and confidence bound rules would apply as with the JESD16 estimator. 8.3 If nonconforming units are not found in the samples, it is recommended that an upper confidence bound be reported according to equation (11). 8.3.1 Report AOQ with an upper 90% confidence bound. EXAMPLE AOQ = 0, S = 50,000 upper boundppm= null1 - (0.10)150,000null 106 = 46 ppm (11) 8.3.2 Format AOQ = 0 with an upper 90% confi
