The day the client called, the entire factory was on edge
Do you remember? A few years ago, a major automotive manufacturer recalled hundreds of thousands of vehicles globally due to a tiny electronic component defect. At that time, our factory was also in a state of chaos. The moment the client's call came in, sales, quality assurance, and manufacturing all became tense. Although it was eventually found not to be our problem, the pressure was so immense that I still remember it vividly. Since then, the client's demand for "Zero PPM" has been as unyielding as if carved in stone, absolutely rigid.
Where's the Problem? "Tolerance" is the Biggest Issue.
Simply put, what automotive manufacturers want is Zero PPM (Parts Per Million), with the goal of "zero defects." Sounds tough, right? But think about it, when you're driving on the road today, would you want any part in your car to have a defect? Even one in a million, that's a deadly serious matter. This is completely different from our previous mindset of "passing as long as Cpk > 1.33." Cpk 1.33 sounds great, indicating good process capability, but converted to DPMO, it's roughly 6210. With 6210 defective parts, would you dare to call it "zero defects"? Of course not.
So the key point is, we can no longer "tolerate" even the slightest defect. Previously, we might have thought, "Oh, this defect is just an isolated incident, next time will be better," but in the automotive sector, this mindset is absolutely unacceptable.
How to Do It in Practice? Prevention is Better Than Cure.
Alright, since the goal is zero defects, how do we actually achieve it? Frankly, it means shutting down all potential problem areas before they can cause trouble.
- Eliminate Risks from the Design Phase (DFMEA): When designing a part, we must consider all possible failure modes. For example, could this part deform due to temperature changes? Is it possible to assemble it in the wrong orientation? All these must be considered during the design phase, and countermeasures devised.
- Watertight Process Control (PFMEA): Once parts are designed, the manufacturing process must also be foolproof. What problems could arise at each process step? Is it a machine parameter drift? Or human operational error? All these risks must be assessed, and strict control points and inspection mechanisms established.
So the key is, you cannot wait for problems to arise before finding solutions. All problems must be resolved on paper, during the planning phase.
The Most Common Pitfall: Believing "Inspection" Solves Everything.
You might think, "Then just add more inspection stages, like Automated Optical Inspection (AOI), In-Circuit Test (ICT), Functional Test (FCT), and block all defective products, wouldn't that be enough?" I tell you, this is the most common pitfall, and I used to think that way too.
To be honest, inspection is certainly important, but it can only "filter out" defective products, not "prevent" their generation. You filter out defective products, but your process continues to produce them. What if the inspection equipment itself has blind spots, or misses a batch? Then you're done for. Moreover, every inspection adds cost; you cannot infinitely increase inspections.
I recall one time a client came for an audit, and when they saw our AOI escape rate was still 0.01%, their expression immediately changed. He said, "Do you think 0.01% is low? We ship tens of millions of parts annually; that adds up to thousands of defective parts. Is this a joke?" It was then that I truly understood: what they want is for the process itself to be stable, not to rely on backend inspection for remediation. In other words, you have to find a way to ensure your process does not generate defective products from start to finish.
One Thing You Can Do Today
Plot your process defect rate as a trend chart, and identify the highest one.