Monthly Archives: February 2010

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.

Design News, January 28, 2010:  “Toyota’s sticking gas pedal was an almost-unforeseeable problem, experts say, and the best course of action now is for engineers to ensure that drivers can handle the failure if it happens again.”

Q. I would like to hear how the Reliability Engineering community feels about this.  Was it unforeseeable?  How can we “ensure that drivers can handle the failure” if the failure is unforeseeable? 

Other manufacturers using throttle-by-wire added software that uses a signal from the brake pedal to override the signal from the gas pedal.  Apparently someone foresaw the problem.  

 Q. As more and more of the functions of the driver are gradually taken over by computerized equipment, with the long term goal of eliminating the driver’s participation entirely,  how do we assure the “passengers” that all is well?  How do we perform Reliability testing holistically?

And finally, though it is not technically related, is the public ready to surrender complete control?

http://www.designnews.com/article/print/446480-Toyota_s_Problem_Was_Unforseeable.php

http://www.nytimes.com/2010/02/08/08webtoyota.html

http://www.washingtonpost.com/wp-dyn/content/article/2010/02/02/AR201002020316

San Jose Mercury News, Sunday, February 14, 2010 Page A1, “Auto electronic controls drawing scrutiny”

We are providing a FREE 1/2 Day Design for Reliability seminar on Thursday, February 11th at Wyle Labs in El Segundo, California. If you are interested in attending, please let us know and we can provide you more info.

Last week was the annual Reliability, Availability, and Maintainability Symposium (RAMS) conference. We presented 5 papers/tutorials. If you were not able to attend, let us know and we can get you copies of our presentations.

Next week is the Medical Design and Manufacturing (MD&M) show in Anaheim California. We will be presenting a paper on Monday entitled “Reliability Challenges in the Medical Industry”. Hopefully you can join us. If not, let us know and we can get you a copy of that paper.

Also next week is the Components for Military and Space Electronics CMSE Conference in Los Angeles. We will be presenting a paper on Wednesday entitled “New Fast Methods for Determining MTBF using HALT Results”. Hopefully you can join us. If not, let us know and we can get you a copy of that paper.