Maximizing Prototype Potential by Demonstrating Reliability with Variable Runtimes

SUMMARY & CONCLUSIONSZero Failure Testing, also known as Success Run Testing, is a widely used method for the demonstration of minimum product reliability. Typically, either a Binomial distribution approach or a Chi Squared distribution approach is used to evaluate such data. The Binomial distribution is generally used when the runtimes are identical across all tested units, whether for the target field lifetime or for a higher or lower test duration.In practical application, however, runtimes often vary between different tested units. This especially holds true for prototype tests close to the end of the product development phase. To ensure that all tested data is included in the calculation, the Binomial distribution approach is extended to allow for individual runtimes on each considered test unit. It is shown that in this case, the results are equivalent to those from the Chi Squared distribution approach.Furthermore, both approaches are compared when failures are observed in the test. Data from a real-life application is used as an example for calculating the reliability, both with and without failures. Different results for the reliability emerge from each method, with the reasons for these differences being analyzed and their respective advantages and disadvantages highlighted. The Binomial method is using more detailed data, while the Chi Squared method provides a good initial estimate and is simpler to calculate. Depending on the context, one calculation method may be more suitable than the other. Overall, by incorporating individual runtimes and failures in the test evaluation, a more accurate and comprehensive evaluation of prototype test results is possible.