Learning from failure



I was discussing failure analysis with my supervisor. I suddenly realized the subtle difference between managing defects in our manufacturing division and the automotive business. In our manufacturing division, we had everything under control, including failure analysis and immediate rectification. Failure analysis is the process of collecting and analyzing data to determine the cause of failure, with the goal of determining corrective actions to eradicate the cause of the failure. This is an established and important discipline in all branches of manufacturing. The process relies on recovering the failed components for subsequent examination of the cause of failure, using a wide array of methods, especially microscopy and spectroscopy. Non-destructive testing methods are also used to inspect and analyze the defects.

Forensic investigation into the failed process is usually the starting point of failure analysis. I remember an incident several years ago when an American customer complained about oxidized products that we shipped from the Philippines. We immediately organized a task force led by the product line manager to determine the root cause of this defect. This was a big deal at that time because the defect was being discovered at the receiving line of the customer, rather than at the end of our production line. This was puzzling for all of us. We reviewed the entire process starting from the raw materials that went into production. We checked each of the components to determine if these were susceptible to corrosion and failure. We investigated the different stages in the assembly line particularly the soldering station.

This particular step became our initial focal point because the solder points appeared to be the original source of the oxidation. We evaluated the compounds that we used as well as the manner by which we applied these compounds in the soldering stage. I remember looking at microscopic images of the solder points, trying to understand why these were eroding. We also looked at the cleaning procedures to figure out if these were sufficient. After going through the analyses, we still could not find any fault. This no fault found (NFF) situation confounded us, as the product was obviously defective.

That’s when one of our failure analysis engineers, who also happened to be a materials science engineer, raised the possibility that the packaging material used in shipping the components (specified by the customer) might be causing all these trouble. It was a long shot but we decided to take a look at that since we had nothing to go on. It was a big relief when we discovered that what appeared to be waterproof and sealed packaging was actually semi-porous. The packaging worked for domestic shipments but was affected by the longer journey from Asia. It was no wonder that the product was passing quality in all stages of production but was corroding en route.

We immediately changed the packaging material, which was a simple matter especially after going through the tedious investigation into the assembly process. Since then, we started looking at the shipping materials and even logistics handling for the possibility that out products may be affected by the environment during shipment. This gave us a valuable learning experience on packaging materials.

Compare this to fault analysis in the automotive business. Our automobiles are now equipped with a myriad of electronic sensors measuring each and every system. We automatically plug in the computer unit of every vehicle through the internet every time a car comes in for a check- up. We do this so we can simultaneously diagnose the automobile’s performance with a manufacturer’s computer located halfway around the globe. This also automatically updates the car’s computer software for any revisions.

Despite all these developments, we still need technicians to inspect and evaluate the condition of the components of the car. Our diagnostic tools would be able to isolate the component that is failing but the computer would not be able to determine why it is failing, unless there is a trend or a product fault that has been established. It is the job of the technician to understand if the fault occurred because of environmental factors, operator error or possibly a defective connection. Once we identify a component that has to be replaced or repaired, the computer records that defect into a database. The system uses this information to justify warranty claims or identify a global trending pattern.

The rectification process may involve a simple replacement or a procedure developed by the car manufacturer. The repair procedures will have to fix the combined electronic and mechanical parts of a particular component. It’s a complicated troubleshooting process that we have to accomplish with the manufacturer or its proxy computer. We try our best in all circumstances to avoid defects and failures before they even occur. In an imperfect world, though, failing is part of the journey. We treat them as invaluable opportunities to learn, so they are not repeated.

Ronald Goseco is executive vice president of the Financial Executives Institute and president of IDI-Volkswagen.


Please follow our commenting guidelines.

Comments are closed.