EIA JEP 133C:2010

The information contained herein is intended for use with MIL-PRF-38534 for those multichip modules and hybrids that are marked as radiation-hardness-assured parts and produced under the provisions of that document or that are built to a radiation specification. Guidance is provided concerning the design, development, fabrication, acquisition and test of multichip modules and hybrid circuits that have radiation requirements. This document is not intended to provide detailed guidance about how to assure the hardness of the dice, since it is recognized that dice with a wide range of hardness will have to be used. If non-RHA dice are used, the user-developed RHA procedures found in MIL-PRF-38535 should be used. Rather, this document provides guidance as to how to assure the hardness of the entire module, given the wide range of the radiation hardness and level of hardness assurance of the individual dice to be used in the module. Specifically, four types of dice are available (in order of decreasing level of specification controls): -- Radiation hardness-assured QML controlled (or equivalent). Dice of this type can be used with no additional testing. -- Non-hardened QML (or equivalent) change-controlled dice. Such devices require radiation characterization. However once this is done, minimal lot testing would be necessary. -- Inherently radiation hard or non-hard dice that are not under a formally recognized change control system, but supplier support (e.g., change control notice, etc.) is available. -- Commercial grade or other grade die that appear to have adequate radiation tolerance, but where no supplier support is provided for the qualification or radiation-hardness assurance. This presents a worst-case situation and requires the most stringent RHA program to ensure that the radiation performance requirements of all of the modules produced are satisfied. The use of dice from any of the above noted categories, combined with the various types of MCM/hybrid suppliers, can then lead to the following categories of MCM/hybrid devices. -- Commercial module designs screened for RHA -- Commercial module designs upgraded with radiation-hardened dice -- Standard product RHA modules -- Custom product RHA modules. The relationships between the various types of dice and finished modules are shown in figure 1. The objective of this Guide is to provide guidance to allow a supplier or user to establish and complement the RHA requirements for any of these MCM/hybrid combinations. Two acquisition strategies can be inferred from figure 1. The first is where significant knowledge concerning the constituent chips is available. This approach, as exemplified by the right branch of the figure, places emphasis on component acquisition (i.e., screening and characterization) and subsequent analysis (as required) to obtain MCM/hybrid certification and qualification. It would apply to module manufacturers and to users who have access to accurate component lists, design rules, fabrication methods, etc. It is discussed in detail in subsequent sections of this Guide and should result in the most accurate and cost-effective course of action to obtain RHA qualification. The second approach is where little or nothing is known about the constituent chips. This approach, as exemplified by the left branch of the figure, represents a worst-case situation from both a technical and cost a point-of-view. If only input/output information is available, one has no choice but to try to determine the failure or response distribution of the module for each of the applicable radiation environments. This method may be acceptable if a statistically significant sample size can be tested and considerable margin exists with respect to the specified radiation levels. The key issues involved here are: an adequate sample size, the homogeneity of the sample, and the sample correlation to flight components. Thus significant effort must go into the development of the RHA test program for this technique. If some minimal knowledge of the component types is available (e.g., high speed bipolar, power MOSFET, etc.), it can be used to guide the development of the RHA test program. For example, if it is known that a power MOS circuit is used, emphasis on Single Event Effects (SEE) testing would be appropriate. Conversely, if a module is known to contain only high speed CMOS digital circuits, then neutron testing and enhanced low rate dose sensitivity (ELDRS) testing can be eliminated. Thus, despite a lack of detailed knowledge about the module, some choices can be made about the radiation testing to improve the test coverage and optimize the test effort.