How do you deal with quality issues?

[EBSMITH]

Here is our issue, we are running spc on some dimensions and the UCL is higher than the USL on our control chart which tells me we are not capable, we are constantly adjusting the process to bring parts into spec. now here is the kicker, we fail the gage r&r. how do I convince management to correct the real issue and quit throwing away thousands of products while trying to process in to some ghost number? Where is Tobin when you need him? lol don't answer that last question, in our current financial straits we couldn't hire him. any ideas?

[rickbatey]

Has the customer rejected parts for this dimension? Filed a complaint? If no, then someone's got to stop the Merry go round!
Can you put a $ amount on the scrap to help prove your point? Don't forget to add in the peoples time plus benefits into the equation.
Rick.

[EBSMITH]

The answer is no, and I have shown the $ amounts, but management is afraid to open the can of worms with the customer as far as centering the dimension, and the reason for that is because we don't have accurate historical data. I don't know how else to make them understand that we are basing crucial decisions on bad data, so the merry go round continues. it gets even worse, we have another job with the same customer that never makes a part out of spec, never has. we need to go out to 6 decimal places to show enough variation to plot the chart, then when the chart starts to move we go process on the press. drives me nuts.

[Suhas]

HI EBSMITH,
Here is what I would do:
1. Without worrying about dimensions, make the best parts you can - closer to the USL ofcourse but with a good stable process.
2. Mold a good quantity that your customer can test.
3. Tell the customer that this mold makes a lot of scrap and so you have to keep sorting and loosing money. Give them the 'above USL' parts and have them test it.
4. If all is good, ask them to change the print.
Not sure if this will fly with your management but this is the best and the truth is that you are already shipping parts at these dimensions.
Thanks,
Suhas

[brentb]

How old is the product?
How much historical data exists if it is an "older" product? I mean is it an inherited product?
Often, especially when new, a product will have over tight "Engineering Specs" that are not based on reality, that is fit and function, and "manufacturing tolerances" can be established that are more lenient, that is more economical

If you have a good robust stable process and UCL is greater than USL (same with LSL LCL), then talk

KOM b rent

[EBSMITH]

the product is 5 years old, it failed the capability study on the initial sample, it failed because we were .001" under the usl nothing was done at that time, so the problem landed in productions lap when it was released. the process is stable enough that if we were able to just change the print and center the dim we would never have a problem. we have been sending them parts out of spec for years without either our knowledge or the customers knowledge, parts have been performing fine in the field. Thanks for all the input. it has now climbed its way up to the #2 in command and we will see what he has to say, in the mean time I just hit the red button, maybe pushing a ship date will get it noticed.