Struggling with a root cause analysis - Customer Returns - Escape issue

[optomist1]

"This pin was reduced in size recently for safety reasons..." again dealing with a partial picture here, but given your recent responsibiltiy for this assembly/part, has the reduction in pin dimension adversely affected the "alarm" to detect a no-go condition? Just my thoughts...

 

[FloridaST]

"This pin was reduced in size recently for safety reasons..." again dealing with a partial picture here, but given your recent responsibiltiy for this assembly/part, has the reduction in pin dimension adversely affected the "alarm" to detect a no-go condition? Just my thoughts...

It hasn't adversely affected it, just made it more sensitive to the placement of the sensors that we build. I suppose from our perspective it's an adverse affect but we should have been building them properly to begin with, the criteria for us hasn't changed.

 

[Bill Levinson]

Thanks for the input everybody. I did some further research on the cove and this is a fairly common topic but your responses to my personal issue has helped get me thinking differently about it. To me, there's a condition that is allowing the operators to build these improperly and that's what I'm looking to fix. Hopefully I can implement a control poka-yoke to remove some of the burden on them.

Shigeo Shingo case studies often begin with the statement that "worker vigilance" was used to prevent errors. You could then be 100% sure errors were occurring because administrative controls are rarely 100% effective in preventing trouble regardless of how vigilant people try to be. Engineering controls (poka yoke) that make the trouble impossible will however work. From what I see of the AIAG's new FMEA manual, error proofing is the ONLY way to earn a 1 detection or occurrence rating. In fact, for process occurrence ratings, the preventive controls must not only be present, they must be "extremely effective" and the failure mode "...cannot be physically produced due to the failure cause."
From what I see of the Detection ratings, the detection control must ALWAYS detect the failure mode or failure cause (and prevent its escape) to earn a 1 rating.

I wonder if there is a way to error-proof the pin issue. I am not a mechanical engineer so I cannot give engineering advice but Shingo covers all kinds of applications of this nature. Can the parts be redesigned, and/or a jig or fixture added to the process, that forces the parts to align properly?

 

[FloridaST]

Shigeo Shingo case studies often begin with the statement that "worker vigilance" was used to prevent errors. You could then be 100% sure errors were occurring because administrative controls are rarely 100% effective in preventing trouble regardless of how vigilant people try to be. Engineering controls (poka yoke) that make the trouble impossible will however work. From what I see of the AIAG's new FMEA manual, error proofing is the ONLY way to earn a 1 detection or occurrence rating. In fact, for process occurrence ratings, the preventive controls must not only be present, they must be "extremely effective" and the failure mode "...cannot be physically produced due to the failure cause."
From what I see of the Detection ratings, the detection control must ALWAYS detect the failure mode or failure cause (and prevent its escape) to earn a 1 rating.

I wonder if there is a way to error-proof the pin issue. I am not a mechanical engineer so I cannot give engineering advice but Shingo covers all kinds of applications of this nature. Can the parts be redesigned, and/or a jig or fixture added to the process, that forces the parts to align properly?

Great post Bill, appreciate the thoughts. This pin is part of a top level assembly that our customer assembles, we merely supply a sensor that interacts with it. I pushed to have a fixture designed and we finally did design one which features a pin to keep the sensors aligned in the plastic case while they are hand soldered (they're through hole radial leaded components).

I'm still learning a lot as I came from production before entering the QA side but, as I learn more about all the various quality tools, I can see their value. Just need to lead others into the light.

 

[John Predmore]

I know the answer is because they weren't paying attention or just didn't care

maybe someone can give me another perspective.

I can offer you another perspective. 2 Types of Human Errors... | Human Error Solutions | GMP Training | By Ginette Collazo, Ph.D.

Have you ever driven past the grocery store when you were making a trip to the grocery store? The problem was not you didn't know where you were or you didn't care, the problem may have been complacency or distraction (by something going on inside the car or outside the car or inside your head). There are people who study the source of human error and people have developed effective strategies to error-proof or greatly reduce the occurrence or escape of mistakes.

Search the Cove for 3-legged 5-Whys for ways to generate additional ideas on Root Cause Identification.

 

[optomist1]

great question and thread....under the subject of "human error", believe Dr Deming had much to say about this subject....my experience with audits, run at rates etc....the Hawthrone Effect (Hawthorne effect is a term referring to the tendency of some people to work harder and perform better when they are participants in an experiment. ) is usually a factor