Suggest a Poka Yoke (Mistake Proof) Method for my Product

[Michael_M]

It sounds like exact count on the inserts is what will be used however, this brings up another couple of issues that need to be addressed.

1. What are the chances that there will be a miss-count on the inserts before starting?

2. What are the chances an insert will be dropped or lost during the insertion process and not known?

 

[Wilderness Woody]

Weigh?...

We would really need to know more about the part... size, form, weight, etc. to describe realistic solutions. Weigh count of the prepped inserts is one thing, but what happens if insert is "lost" during assembly? Root cause of missing inserts may reveal weaknesses and point to a simple solution.

 

[normzone]

If this is anything like the sheet metal chassis I deal with, [DRAMMAN] has it right.

" I have worked on many issues like this. kitting parts and the like helps but never solves the problem. What happens when the kitters make mistakes and provide less parts? What has worked best for me is creating a well organized process with motivated and engaged employees. "

Redesign is not an option, small lots make dedicated tooling impractical.

 

[Ronen E]

Not sure what you meant by

what happens if insert is "lost" during assembly?

but I was referring to weighing the assembled object, not the bunch of inserts. This would detect any single-error (at present / not-present sensitivity), and good chances that even some (or most) of the common multiple-mistakes - depends on the details. It's not 100% proof, but most acceptance programs allow some non-zero defect rate. And most important - it's simple, practical and not expensive (assuming that it's not a huge assembly).

Keep it simple, guys. Cameras and sensors??? What's next?... Install a robot to do it?

Cheers,
Ronen.

 

[DRAMMAN]

Weighing the final part should work. Scales these days are incredibly accurate. A single missing insert could easily be detected.

 

[Wakefield QM]

Good stuff Thanks for input !
Looks like we will improve our inspection :mg:

 

[jeffr]

I would weigh several assembled parts for variance in the event the variance is as much as an insert.

 

[DRAMMAN]

This is a must do. That could easily prevent using weight verification.

 

[Ronen E]

The applicability of weighing as a control method can't be discussed without knowing the details such as weights - nominal and tolerance; available weight accuracy and precision; variance within lots and lot-to-lot, etc. A careful, quantitative analysis must be taken.

It could be anything from "totally useless" to "WOW".

Cheers,
Ronen.

 

[silentrunning]

This is a common problem in sheet metal shops so after posting my last idea I did some more investigation and think I have come up with a ?better? idea. My last suggestion was to check the parts on a CMM, but that adds an extra operation and time to the process. I am now working with our maintenance man to build a counter that can be programmed to count the number of strokes on our PEM Setter. It will have a red LED and a green LED. The setup man can program the number of strokes the PEM Setter is supposed to cycle based on the number of inserts in the part. The red light will remain on until the correct number of strokes is counted and then the red will go out and the green will come on. At this point the operator will know that the correct number of PEMs have been inserted. The operator will then grab a new part and hit the reset button turning out the green light and relighting the red light to start the cycle again. This will preclude having to handle the parts another time and thus cut down on the cost. Theoretically this should work.