After 25 years in the reliability field I am still a bit mystified by the humidity testing. Generally the environmental tests can be divided into two major categories: Durability tests, where some form of wear-out mechanism causes products to fail. The most common examples of durability tests are vibration and thermal cycling. The second group is the capability tests, often referred as overstress tests with the goal of determining how well the product can resist certain conditions such as high voltage, accidental drop, dust, or others.
It appears that humidity tests belong somewhere in between the two categories. On one hand there are electro-migration, corrosion, dendritic growth, and other failure mechanisms following the pattern of wear-out processes. Those failure mechanisms are indeed accelerated by the combined effect of temperature and humidity and the most commonly used acceleration models used to calculate the test durations are Peck’s and Eyring. Both models have rather limited applications and varying accuracy, but currently the best what reliability science can offer to calculate the field to test ratios. On the other hand humidity often causes the change of mechanical properties of the materials which often makes them more susceptible to failures. For example, modules of elasticity of some materials go down after absorbing moisture and some plastics become more prone to developing cracks as a result of humidity exposure. Those types of failure mechanisms can not be described by any known algebraic acceleration models. Despite that engineers often mistakenly apply them for test time calculation. The desire to use algebraic acceleration models is very strong due to their relative simplicity and ease of comprehension. The alternative to the use of acceleration models is the use of predetermined tests, like for example 164 hours of 85% relative humidity at 85 degree C. Those tests are often based on some historical data rather than on test rationale or good understanding of failure mechanisms.
Therefore, before writing a product validation plan involving any humidity testing it is important to answer the following questions:
What are the expected humidity-triggered failure mechanisms for my product?
Are any forms or electro-migration involved?
Will the humidity affect any of the material properties and will it make my product more prone to failures?
What type of humidity test is most appropriate for my products? Steady-state, cyclic, both?
Will any of the possible failure mechanisms be accelerated by higher humidity/higher temperature combination or they would remain neutral to it?
Do any of the known acceleration models apply in my case and if not, how do we determine the test duration?
What exactly will my test represents for the life of the product?
Answering those and some other questions is critical to a successful validation program involving humidity testing.
Andre, great article. This goes along with a few papers/presentations in RAMS – one by Vaishali Hegde of Respironics and one by me – on using FMEAs to help drive a reliability test plan. You first need to understand how/why humidity will affect your product before you try to specify a test using humidity, and an FMEA is a great way to determine this.
Terrific work! This is the type of information that should be shared around the web. Shame on the search engines for not positioning this post higher!