Downloaded from https://www.everyspec.com MIL-HDBK-816 NOTICE 1 absolute value of the parameter is the important quantity, they are nevertheless highly desirable. Figure 2 and subsequent discussions, therefore, include pre-rad measurements as part of the characterization procedure. 4.5 Considerations beyond the scope of this document. 4.5.1 Incorporating additional radiation and nonradiation information.
If information is obtained in addition to that provided by the given Sample Size Series test against the RHEPL contained in the detailed device specification, then, for particular system applications, there are several methods by which RHA part survival probabilities can be obtained which are higher than those which correspond to the Sample Size Series test alone. These methods include: The use of the next higher radiation level RHA part, radiation overtests, the use of lot rejection information, the use of variables testing for LAT, and the use of derated parameter values.
As an example, most systems manufacturing companies are already familiar with the use of derating factors for temperature and aging. Radiation derating factors may be used in much the same way and can thus be fitted into existing practices. These methods properly belong in a document on system level hardness assurance and are not further addressed here. 4.5.2 Further complex questions. The statistical questions involved in hardness assurance and lot acceptance tests range in complexity from cases for which the analysis methods are relatively straightforward to those for which the required formalism is still being developed.
In the former instances, these guidelines provide simple step by step procedures that may be used for the analysis of the characterization data. In the latter instances, which are beyond the scope of this document, these guidelines can only suggest approaches which are overly conservative or recommend that statisticians be consulted for less conservative but still valid analyses. In the latter category are cases where the within-lot variations are comparable to lot-to-lot variations (in this case the definition of sample size can become uncertain), cases of combined environments or combined parameters, and cases where not enough data is available.
The 1987 paper by Namenson and Arimura, listed in 2.2, may be consulted for an approach to multi-lot and multi-parameter data analysis. 5. DETAILED REQUIREMENTS 5.1 Characterization of piecepart performance in radiation environments (see figure 3). The first step required for developing a device specification is to characterize the radiation response of the part type in question. Such a characterization for one or more radiation environments requires measurement of its post-irradiation performance and, usually, its pre-irradiation performance as well.
The parts, generally, should be characterized through to failure so that the specifications can be written, if necessary, for the maximum capability of the part type. For the present discussion it is assumed that the purpose of the characterization measurements is to support the calculation of parameter end-point limits for qualification and LAT tests in a military standard detailed specification and that satisfactory data for the part type in question does not already exist. In this case, the characterization measurements should be such as to permit these calculations to be made for the radiation levels given in table I.
Alternatively, the characterization measurements may be required in direct support of a hardened system design or production (custom specifications). In this latter case, the parameters which are measured and the types and fluence levels of the radiations which are used should reflect the system requirements and the specific vendors who are expected to be suppliers for the system. The characterization measurements should also determine the radiation levels required to cause part failure; such information can be important for estimating the design margin when the part is used in some particular application.
To simplify the present discussion, we assume that the average post-irradiation performances of a part type may differ for different manufacturers and, for a given manufacturer, may differ from one lot to another, but that the average radiation response characteristics of a part type, for a given manufacturer, do not significantly change with time. This latter assumption may not always be warranted. It is needed to keep the initial discussion of data measurement and analysis free of the additional complexities that time variations produce. 5.1.1 Documentation of characterization information.
Experience has shown that the analysis of radiation data at a future time or by persons other than those who performed the measurements is enormously facilitated by good documentation of the test results. At the same time, a need to analyze previously acquired data or data acquired by others is common. For these reasons and because new measurements are costly, these guidelines recommend that the overall characterization measurements be summarized in a characterization report and that the conditions under which the measurements were made be documented in a test plan which is included in the characterization report.
These documents should furthermore be sufficiently complete so that independent data analyses can be performed on the results contained therein. SUPERSEDES PAGE 13 OF MIL-HDBK-816. 13