Why Validated Manufacturing Processes Still Fail

Lessons from two real manufacturing investigations

Most manufacturing engineers have experienced some version of the same problem: A production process successfully completes validation. The production part approval process (PPAP) is approved. Yield is stable, customer requirements are being met, and everyone assumes the process is under control. Then one day, defects begin to appear.

When that happens, the natural reaction is usually to look at the manufacturing process first. Engineers review machine settings, adjust process parameters, introduce additional inspections, and launch experiments.

Sometimes those actions appear to work. The defect disappears temporarily, everyone breathes a sigh of relief, and production returns to normal.

Then the defect comes back.

Over time, I’ve learned that one of the most expensive mistakes manufacturing engineers can make is assuming that every production problem originates in the manufacturing process itself. Sometimes the production line is performing exactly as designed. The real problem might be hidden in supplier materials, poor process discipline, environmental conditions, or manufacturing changes that occurred long before the defect was discovered.

…

Want to continue?

By logging in you agree to receive communication from Quality Digest. Privacy Policy.

Create a FREE account

Forgot My Password

Comments

Common Vs Special Cause

Hello:

The first step is to understand what the process is trying to tell you. Use a time ordered run chart. Is the measured data to be expected? Is the process stable? Is the measured value a common data point and to be expected? Or is there a special cause? Understand the variation in the process before doing anything.

Thank you, Dirk

Reply to Dirk

Thank you, Dirk. I agree that distinguishing common-cause from special-cause variation is an important step before making process adjustments.

This fits well with the broader point I was trying to make in the article: a defect appearing in production does not automatically mean that the validated manufacturing process itself is the root cause.

In the investigations described in the article, we first reviewed the available process evidence rather than immediately changing process parameters. When the process evidence did not support a process-related failure, the investigation was expanded to other potential variables.

In the first case, for example, we used a controlled DOE to isolate the material variable while holding the manufacturing conditions constant. The defect repeatedly appeared with the suspect material but not with the previously approved material. Independent chemical analysis then confirmed a material difference, eventually leading to the identification of an unapproved supplier formulation change.

Your common-cause vs. special-cause point adds an important SPC perspective: before changing a validated process, first determine whether the observed variation reflects normal process behavior or evidence of a specific change that needs to be investigated.

Thank you for adding this perspective to the discussion.

Great article. One point…

Great article. One point that really resonated with me is that process validation proves a process can work—it doesn't guarantee it will continue to perform. In battery manufacturing, I've seen validated equipment drift over time because of tooling wear, raw material variation, or subtle parameter changes. Continuous monitoring, SPC, and treating validation as an ongoing lifecycle rather than a one-time milestone are what keep a process truly under control.