How to react when all check points fail?

[Jim Wynne]

Hello everyone
First check point, receiving raw material from supplier, but didnt check the paper´s thickness

Did the supplier send you the wrong material? What is the documentation provided to the incoming inspector? Is there a specification sheet? How is acceptance documented?

Second check point, Before starting production, have to check thickness, and was too confident, he thought was the same product as always , no need to check it.

How does the person responsible for checking the thickness know what to do in verifying the material? Why is this done if the material was accepted as conforming at incoming inspection?

Third check-point Quality coordinator, even if procedure, states to perform the final checking against customer requirements before the release, checked only some aspects not all what is required, he supposed everything was ok.

If the material is checked twice before the Quality Coordinator does the final check, why would thickness of the paper be checked again?

IT seems overconfidence , and root cause???? not easy to find it nor eliminate it.

The root cause is that the supplier of the paper sent the wrong material. You shouldn't have to verify it three times in order to have confidence that the material is as specified.

I may apply resources to find something but nothing really applicable under a QMS.

What do you do in these cases?, What have you done in the past, or how do you analyze a case like this?

To start off, I´m writing up a Corrective action in the final process where was detected.

So you're writing a CAR for the operation where someone did the right thing? What about the supplier?

Also, this is not a case of shared responsibility as has been claimed in previous replies. Each person had individual responsibility in this case. It would be shared responsibility if two or more people conferred and decided jointly what to do.

 

[Michael_M]

A defective product was detected one step before the shipping (a brochure was made of paper thicker then required)

On a positive note, the product was detected at the 4th check point (one step before shipping).

Although I agree with Jim that each step is responsible for their step, I also agree with Steve that this was a shared responsibility in everyone thought everything previous to them was done right. I would focus on the attached forms and find out why they are ineffective at catching the problem. I am making a HUGE assumption here, but I think if you delve into the forms you will find some of the data to be 'just filled in to save time' (again, huge assumption based on my past experance).

 

[Jim Wynne]

Although I agree with Jim that each step is responsible for their step, I also agree with Steve that this was a shared responsibility in everyone thought everything previous to them was done right.

Again, that's not shared responsibility. Also, if we follow the standard PFMEA process (perhaps not a factor in this case) we should always assume that material coming into a given process step is conforming, and concentrate failure mitigation efforts on the current step. Part of the problem here is that having multiple layers of verification of a given product feature assumes that upstream checks were not done, or were done incorrectly. That sort of thing can become a self-fulfilling prophecy.

Also, it appears that the checklist(s) were completed as though the required verification had been done, which is a case of falsification of the record. That should be considered a very serious problem.

 

[John Broomfield]

Thanks John
Could you give ideas to focus on to improve the process?

You’ll have completed a study to determine the process variation that is attributable to:

A. Common causes
and
B. Special causes

Eliminate the special causes from the inputs to the process and the process itself.

You are then left mainly with common cause variation or variation caused by the system. These usually are largely predictable so you may be able to live with them.

If you’re still paying the price of nonconformity then determine the 4% of common causes that result in 50% of the nonconformities and eliminate these causes from the system. This phase of process improvement my take a while.

Best not to rely on inspection by humans which at best is 85% effective.

 

[Mike S.]

I know nothing about the OP's process. There is some good advice and options above.

I will say, I know some processes, at their best, cannot produce 100% no-defects product and thus require inspection. To pretend otherwise is folly.

 

[John Broomfield]

I know nothing about the OP's process. There is some good advice and options above.

I will say, I know some processes, at their best, cannot produce 100% no-defects product and thus require inspection. To pretend otherwise is folly.

Then install automated inspection.

 

[Jim Wynne]

If you’re still paying the price of nonconformity then determine the 4% of common causes that result in 50% of the nonconformities and eliminate these causes from the system. This phase of process improvement my take a while.

Do you have a definitive source for this 4%/50% claim?

Best not to rely on inspection by humans which at best is 85% effective.

This idea has been debunked repeatedly in these forums.

I know nothing about the OP's process. There is some good advice and options above.

I will say, I know some processes, at their best, cannot produce 100% no-defects product and thus require inspection. To pretend otherwise is folly.

Then install automated inspection.

What if the cost of automated inspection (which isn't 100% foolproof) outweighs the risk and cost of nonconformity?

 

[Mike S.]

Then install automated inspection.

Automated inspection is also not 100% accurate, with type 1 and 2 failures, and often relies on humans to look at the flagged parts and make a determination.

 

[Mike S.]

I think the 4%/50% mentioned is a derivative of the Pareto Principle applied to itself as a power law, which really gives 4%/64%.

 

[Bev D]

I think the 4%/50% mentioned is a derivative of the Pareto Principle applied to itself as a power law, which really gives 4%/64%.

The so called Pareto principle where any percentages are stated are MYTHS. Old wives tales. Fake. I don’t know how many times we have to say it or point to the literature debunking it - not the least of which are Pareto’s own writings.