Getting to the root cause of product failures

When product failures such as the Takata airbag incident occur, it is tempting to simply come up with a quick fix to solve the problem. But more important is to determine the actual cause of the failure, in order that it does not happen again. The recent spate of incidents involving Takata airbags in some Japanese cars like Toyota, Mazda and Honda have prompted recalls for certain affected models. At present, investigators are focusing on the role of moisture in prematurely activating the airbags.

We see safety and reliability issues in many aspects of our lives, not just in the automotive industry, but also in aviation, in electronics, even in our cookware. Ideally, we should see these problems before the product is released to market. Unfortunately, sometimes with disastrous results, we only see these failures when the product is already being used by millions of people.

Most people take it on faith that brands are there to be trusted, even with their lives. If a company fails to live up to that expectation, it could lead to the end of that company. Does anyone still remember a tire company named Firestone whose manufacturing process problems led to deadly tire blowouts in the mid-1990s?

Detecting the root cause of failures for critical components, especially if the failures are intermittent and do not happen with regularity, can be challenging for failure engineers who often try to subject these components to various tests. The dangerous part is when a certain set of conditions produce a failure that is not part of the test procedure for these components.

Failure engineers often use various techniques like statistics, to determine whether a failure is design related (meaning it was there from day one), manufacturing related or materials related (meaning the reliability could be questionable), or the component itself could be good but used in conjunction with others in a system, it could be the wrong choice. Or sometimes, as in the case of the space shuttle Challenger explosion, knowing the unsafe conditions when not to use the product. In the case of the Challenger, investigators concluded that the very cold weather that morning produced ice in the O-rings that caused fuel to leak and explode.

Design engineers try to design some redundancy and put in a safety margin when they select a component. Hence, if a certain amount of stress will be placed on a component, they could choose to double the amount of stress that a component could handle.

Test engineers try to stress the component and see when it breaks. By getting the statistics of failure, they can compute when it needs to be replaced. Engineering managers try to qualify their suppliers and the materials they use, as well as the manufacturing process, to ensure that the product they are making is the safest and best product they can consistently come up with. Test techniques such as Fault Tree Analysis (FTA) and Failure Mode Effect Analysis (FMEA), and the use of statistics such as Weibull Analysis, often help design and manufacturing engineers make sense of the failures they see to develop better and safer products.

To a test and reliability engineer, a productive workday normally means he/she has found a new failure in a component that had not been found before and has not shown up in previous tests and starts working on a root cause and eventually a fix. Take note that getting to the root cause of why a problem occurs is just as important as fixing it. Because understanding why a problem happens adds to the design engineers’ list of things to consider.

The best companies and brands are defined by the amount of engineering and testing they do to ensure that a product that is used by millions is safe and thoroughly tested. If incidents do happen despite this, praiseworthy companies take it upon themselves to determine the cause and prevent it from happening again. Take for example the intentional tampering and sabotage of Johnson & Johnson Tylenol painkiller which led to deaths in the early 1980s and how it was used by management to redesign its products to be tamper resistant. J&J had shown the public that it would never place profit over the public’s safety, and this helped them bounce back from this tragedy.

Contrast this with the way Merck handled the Vioxx (Rofecoxib) test results for their revolutionary painkiller even if there was statistically significant data to show that there was an increased risk of heart attack and stroke for certain individuals. Yet they still chose to release the drug into the market.

It is easy for companies to be tempted to take shortcuts to go after a quick profit, after all many times the problem does not show up. However, when it does show up, the results are often tragic and no amount of finger pointing and recrimination can make up for what should have been done in the first place.

Trust is then lost, and all that is decided by how a company acts.

© JapanToday