Reliability Prediction - MTBF Analysis
Click here for FMECA
AEi Systems performs parts-count and stress-based MTBF analyses. The results are used to assess the potential component failure probabilities, and when the failures may occur during the product’s lifetime. When combined with a stress analysis and a FMECA analysis, these “part-based” analyses can be very effective in pinpointing the soft spots in your design, allowing you to dramatically improve system quality.
While there are many guides that discuss reliability analyses, few discuss some of the pitfalls with calculating and interpreting the results.
An "MTBF" analysis basically sums the failure rates of the parts in the BOM. The analysis produces a single number for the systems MTBF, failure rate, and reliability. Often, an MTBF analysis is done only to compute this number, which is used mainly for marketing purposes. This is a less than optimum goal for two reasons: one, the number as discussed below is suspect, and two, the most useful aspect of the analysis goes unutilized.
The usefulness is marred by the variability of each part’s data. To get the failure rate correct requires attention to various characteristics that impact the calculation, any one of which can greatly change the result. A parts-count analysis is of limited usefulness.
Here are some of the issues that AEi Systems takes into account. If you don’t, the results for the SAME BOM can differ by several hundred percent.
• Calculation irregularities are commonly encountered
• Confidence Level (up to 2.5x difference between 60% & 90% derived FIT rates)
• Arrhenius translation between temperatures can be complicated
• MIL-Handbook-217 and Telcordia have a limited set of categories
• Is Stress Included in the MTBF computation?
• Are vendor FIT Rates used properly"
• MTBF doesn’t consider “out-of-spec” conditions
• Only considers “extreme” types of failures; results may not be useful to improving reliability
There are valuable reasons for performing this analysis such as identifying the key reliability drivers, making
MTBF comparisons with competitive products, and selecting warranty periods. There are other results from the
analyses (PI factors, ranking of failure rates by category, part type, environment, stress level, etc.), but
none is more important than the ranking of the parts in the design by their failure rate.
This is where the electrical designer can gain some insight into the design quality and choose parts that will most improve reliability.
AEi Systems supports MIL-Handbook-217 and Telcordia Technical Reference TR-332. Both standards have similar drawbacks to those listed above, though not to the same degree or specificity. MIL-Handbook-217 tends to be for more high-reliability applications (military, space), while Telcordia tends to be used more for commercial products.
Here is a link to a sample MTBF analysis.
To actually perform the analysis, you will need to provide the following Information:
Assembly & subassembly structure (if there are multiple subassemblies)
Which calculation model to use: Telcordia or MIL-HDBK-217
Temperature(s) at which you want the results computed.
Operational Duty Cycle (is the unit on all the time or does it cycle?)
Burn-in time (of parts or the system)
Expected Mission life, if known
Environment (Ground Benign, Controlled, Space, etc.)
Part Quality Levels (AEi Systems can help you figure these out)