Pinpointing Causes for Rare Faults

Contrary to popular belief it is nearly impossible to identify and track each and every process variation that can or will occur over the life of the process (or product). It is also not cost effective to do so. Most process are set up with a pass/fail check at points along the process because it is simply too expensive to catch every fault.

Perhaps one of the most effective programs ever developed is the Failure Mode and Effects Analysis (FMEA) because it makes a good starting point for where to look. By that, I mean to emphasis that even a good FMEA will not cover every combination of screw ups to which most processes are subject. Consequently there will be surprises along the way. What a good FMEA does do is that it usually identifies the parts of the process that have the most influence. The failures come from not knowing what is significant and what is not. That being said, the issue here is tolerance and limits which are often chosen arbitrarily without any reasonable analysis. This is always costly whether the limits and tolerance band is too big or too little. Rarely is any analysis done to determine the impact of variation beyond a certain limit.

Most processes are monitored in a monthly report. This is a lot like driving a car using only the rear view mirror to see where you have been. You certainly know that you are moving, but you only know for sure where you have been.