How Carefully Do You Analyze Product Test Failures?

Finding or identifying which component(s) failed is only an initial step. Knowing what the role of that failed component in the ckt., device type and specifications, knowing how the device was used in filling-in that role, and coupling all gathered information with any other noted environmental conditions that may have helped triggered the failure. These are factors that should always be considered before putting the system back in service. Identifying, addressing or correcting all these faulty conditions, incorporating them in the failure analysis, will ensure a cost effective approach in mitigating the problem. At the end of this exercise and regardless of what problem was found, economics of things will prevail which will serve as the main determinant whether to scrap or repair?

I teach courses on testing, and I always hit failure analysis hard. I tell my students that if they don't do a thorough failure analysis that they have just wasted time and money doing the test; that a cursory analysis without verifying the actual problem can be worse than none at all. I also highlight that there are times when they may not have the staff to do it but may need to go to an outside expert. In one of my courses I actually had an outside expert, and he admitted in front of the class that even someone trained in analysis cannot always tell the exact stress that caused a failure - for instance, stress from differing thermal coefficients can mimic vibration failures. That helped the rest of the class to understand the reason for doing single environment tests first as a baseline in order to help them with their failure analysis later.