MIL-STD-690-SUP Failure Rate Sampling Plans and ProceduresThis standard provides procedures for failure rate FR qualification, sampling plans for establishing and maintaining FR levels at selected confidence levels, and lot conformance inspection procedures associated with FR testing for the purpose of direct reference in appropriate military electronic parts established reliability ER specifications. Figures and tables throughout this standard are based on exponential distribution. Weibull distribution will be acceptable in certain components such as capacitors. Use of Weibull distribution for any component must be approved by the qualifying activity. This standard also provides guidance to specification writers in the use of this standard see appendix A and references material for users of ER parts.

Pages 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 - 14 - 15 - 16 - 17 - 18 - 19 - 20 - 21 - 22 - 23 - 24 - 25 - 26 - 27 - 28 - 29 - 30 - 31 - 32 - 33 - 34 - 35 - 36 - 37 - 38 - 39 - 40 - 41 - 42 - 43

MIL-STD-690D

6.  NOTES

(This section contains information of a general or explanatory nature that may be helpful, but is not mandatory.)

6.1  FR level conversion.  FR levels may be converted to mean time to failure (MTTF) as follows: MTTF = 100,000 + %/1,000 hours.

Example:  Where FR level P = .1 percent per 1,000 hours.

MTTF = 100,000 + .1 = 1 failure per 106  hours.

6.2  Computation of the unit hour requirement.  Table VIII is a complete Poisson distribution table that is suitable for computing unit hours.  To calculate unit hours with a "C" number of 0 (for table I, table II, or table IV), determine the probability of acceptance (Pa) by subtracting the FRSP value (.60, .90, or .10) from 1.

Example:  Pa = 1 - .60 = .40

From the Poisson table, find .40 under the C = 0 column.  This value of "m" is found, by interpolation, to be .916 for level M qualification and "m" in the table is the total of failure rate A (lambda) multiplied by the time (test hours).

m = A × t

unit hours = m + A (1%/1,000 hours).

unit hours = m + A = .916 + .00001 = .0916 million hours.

Values for P, R, and S levels are found by multiplying the previous level by 10.

6.3  Computation of the true failure rate.  Table VIII is used again to compute the true failure rates for qualifying 19 of 20 times and for rejecting 9 of 10 times.  To calculate the true failure rate to accept 19 out of 20 times (=.95), look up .95 in the C = 0 column (interpolation is needed).  The value of "m" is found to be .051.  Referring to 6.2, A (true failure rate) is found by dividing "m" by the time for FRSP-60.

Example:  A = m + t = .051 + .0916 million hours = .06 percent per 1,000 hours. Repeat the process to determine the true failure rate for FRSP-90.

To calculate the true failure rate to reject 9 out of 10 times (Pa = 1 - .90) look up .10 in the C = 0 column (interpolation is needed).  The value of "m" is found to be 2.30.  Referring to 6.2, A (true failure rate) is found by dividing "m" by time for FRSP-60.

Example:  A = m + t = 2.30 + .0916 million hours = 2.51 percent per 1,000 hours. Repeat the process to calculate the values in table V.

6.4  Subject term (key word) listing.

Confidence levels                                                   Failure rate sampling plan (FRSP) Cumulative unit hours                                             Maintenance of qualification Established Reliability                                            Qualification

Extension of qualification                                        Qualifying activity

Failure rate (FR)                                                     Truncation of data

6.5  Changes from previous issue.  Marginal notations are not used in this revision to identify changes with respect to the previous issue due to the extensiveness of the changes.

21