MIL-HDBK-217 Reliability Prediction
The most widely known and used reliability prediction handbook is Mil-217. It is used by both commercial companies and the defense industry, and is accepted and known world-wide. The most recent revision is "Military Handbook, Reliability Prediction of Electronic Equipment", MIL-HDBK-217, Revision F, Notice 2, which was released in February of 1995. It contains failure rate models for numerous electronic components such as integrated circuits, transistors, diodes, resistors, capacitors, relays, switches, and connectors, to name a few. MIL-217 requires a greater amount data entered into the model. It also is a little harsher in the calculation of failure rate data than the Bellcore standard. Typically, but not always, MIL-217 calculated results will show a higher failure rate than Bellcore standard for the same system. This difference in the standards obviously stems from the original intended use of the MIL-217 standard for aerospace and military, or mission critical applications.
Maintaining reliability and providing reliability engineering is an essential need with modern electronic systems. Reliability engineering for electronic equipment requires a means for a quantitative baseline, or a reliability prediction analysis. The MIL-217 standard was developed for military and aerospace applications; however, it has become widely used for industrial and commercial electronic equipment applications throughout the world. Using the Mil-217 standard for reliability prediction produces calculated Failure Rate and Mean Time Between Failures (MTBF) numbers for the individual components, equipment and the overall system. The final calculated prediction results are based on the roll-up, or summation, of all the individual component failure rates.
"Military Handbook; Reliability Prediction of Electronic Equipment", MIL-HDBK-217F Notice 2 dated December 2, 1991 developed by Rome Laboratories, and the United States Department of Defense. The purpose for developing this handbook was to establish and maintain consistent and uniform methods for estimating the inherent reliability of military electronic equipment and systems. The handbook is intended as a guideline, not a specific requirement, to increase the reliability of equipment being designed.
The handbook contains two methods of reliability prediction, Part Stress Analysis and Parts Count Analysis. The two methods vary in the degree of information required to be provided. The Part Stress Analysis Method requires a greater amount of detailed information and is usually more applicable to the later design phase. The Parts Count Method requires less information such as part quantities, quality level and application environment. It is most applicable during early design or proposal phases of a project. The Parts Count Method will usually result in a higher failure rate or lower system reliability, a more conservative result than the Parts Stress Method would produce.
Part Stress Analysis
The Part Stress Analysis method is used the majority of time and is applicable when the design is near completion and a detailed parts list, or BOM, plus component stresses are available. By component stresses, the standard is referring to the actual operating conditions such as environment, temperature, voltage, current and power levels applied, for example. The MIL-217 standard groups components or parts by major categories and then has subgroups within the categories. An example is a "fixed electrolytic (dry) aluminum capacitor" is a subcategory of the "capacitor" group. Each component or part category and it's subgroups have a unique formula or model applied to it for calculating the failure rate for that component or part.
Failure Rate and pi Factors
The failure rate formulas referred to above include a base failure rate, pib, for the category and subgroup selected. ("pi" is used here to represent the parameter variable used in the euations). The base failure rates apply to components and parts operating under normal environmental conditions, with power applied, performing the intended function(s), using base component quality levels and operating at the design stress levels. The standard then applies many pi factors, or multiplying factors, to the base failure rates in order to factor in the actual operating conditions, environment and stress levels referred to above. Base failure rates are adjusted by applying the pi factors, which range from 0 to 1.0, to the underlying equation or model provided for each component category.
The above procedure calculates the predicted failure rate at the actual operating conditions for each component in the project. The procedure to determine the overall system level or equipment failure rate is to sum, or roll up, the individually calculated failure rates for each component. Most manufactures of electronic equipment assemble a majority of the components to various types of printed circuit boards (PCBs) or as part of a hybrid construction. A failure rate is determined for the PCB or hybrid device by the summation of the failure rates for the numerous components, solder joint connections and other types of construction involved. The failure rate for each connection made to the PCB by electrical connectors is also included. The failure rate for wire between electrical connections is assumed to be zero. MIL-217 does not utilize a basic circuit board model. It sums the failure rates for each individual connection type times the quantity and adds that overall PCB (Block) connection rate to the sum of the attached component failure rates. However, MIL-217 does have a models for PCBs with plated through holes (PTH), surface mounted technology (SMT) and a model for hybrid
The design quality or "as purchased" quality of the component utilized has a direct effect on the part failure rate and appears in the models as a pi factor, piQ. Many of the components covered by the MIL-217 specification are available in several quality levels and each has an associated pi factor, piQ. It is especially important to take notice to microcircuits and integrated circuits (ICs) quality specifications and the resultant pi factors. Parts purchased under older specifications are referred to as "Non-established Reliability" (Non-ER) or they can be broken down into two additional quality levels labeled, "MIL-SPEC" or "Lower". Non-ER parts purchased in complete accordance with a particular MIL specification should be entered for the applicable MIL specification. If some of the quality requirements are waved for the purchased component or if it is a commercial component, the "Commercial", "Lower" or Non-ER rating should be used. Each quality designation has an associated pi factor, piQ.
Environmental stress is of major concern in establishing the failure rate for components and parts included in a system per the MIL-217 model. Environmental stresses can be quite different from one application environment to another and can subject the equipment to a controlled environment with constant temperature and humidity, or an environment with rapid temperature changes, high humidity, high vibration and high acceleration, for example. The environmental designations included within MIL-217 are included in the formulas as piE,
Ambient and operating temperatures have a major impact on the failure rate prediction results of electronic equipment, especially equipment involving semiconductors and integrated circuits. The MIL-217 standard requires an input of ambient temperatures and more definitive data required for the calculation of junction temperatures in semiconductors and microcircuits.
A thermal analysis should be a part of the design and reliability analysis process for electronic equipment. Ambient temperatures for overall equipment should be the ambient temperature close to or between the equipment involved. Individual component or part ambient temperatures should use the operating ambient temperature inside the equipment where they reside. The ambient temperature for components or parts located within the area of hot spots should be adjusted for the higher ambient temperature in the area.
Typical MIL-217 Failure Rate Model
A sample MIL-217 failure rate model for a simple semiconductor component is shown below. Many components, especially microcircuits, have significantly different and more complex models.
Failure rate = pib * piT * piA * piR * piS * piC * piQ * piE Failures/million Hours
The above listed pi factors are based on a simple component and are shown for example. There are also pi factors for items such as learning factor, die complexity factor, manufacturing process factor, device complexity factor, programming cycles factor, package type factor, etc. Each component or part group and it's associated subgroup has a base failure rate plus numerous pi factor tables, unique to that component or part, that list factors that are used in the model to adjust the base failure rate.
A solid tantalum fixed electrolytic capacitor, for example, has a MIL-217 model as follows:
Failure rate = pib * piCV * piSR * piQ * piE Failures/million Hours
MIL-217 Parts Count Analysis
The MIL-217 Parts Count Reliability Prediction it normally used when accurate design data and component specifications are not determined. Typically, this will happen during the proposal and bid process or early in the design process. However, this stage in the design process is where design decisions and project specifications, allocations, etc. can be determined with help from preliminary reliability prediction data.
Minimal information is required for a Parts Count Reliability Prediction. The formula for a parts count analysis is simply the summation of the base failure rate of all components in the system. (Refer to the MIL-217 standard for the specific equation)
Equipment operating in multiple environments will have the calculations applied to a portion of the equipment in each environment.
The MIL-217 standard provides tables for the component groups (same groups as the Parts Stress analysis) listing generic failure rates and quality factors for the different MIL-217 environments.
The predicted failure rate results will normally be more harsh using the Parts Count method than using the Part Stress analysis. The Parts Count analysis does not factor in the numerous variables and uses worst case generic or base failure rates and pi factors.