Basic Need:Initial Capability Study:No critical characteristic on print

[larry hopperrath]

Good morning!
Here is my question! We are a bracket supplier, Tier II. None of the prints we received from our customer have significant/critical characteristics on them. We do initial capability studies on an internal characteristic, even though they are not required by our customers. During an audit recently, I was asked how we proof that we meet or exceed initial capability at PPAP. I told him we do not collect SPC data. Our auditor did not like that answer and will be looking for some evidence during the next audit. What simple SPC tools can we use, not only to satisfy the auditor, but to be a benefit to our company. Currently, parts for production are approved using attribute gages on an hourly basis. Currently, our quality department is not set up good to start performing dimensional checks on an hourly basis. Any suggestions would be of great benefit to me.

 

[Steve Prevette]

What simple SPC tools can we use, not only to satisfy the auditor, but to be a benefit to our company. Currently, parts for production are approved using attribute gages on an hourly basis. Currently, our quality department is not set up good to start performing dimensional checks on an hourly basis. Any suggestions would be of great benefit to me.

It's my contention that SPC really isn't that hard. It can be done in Excel spreadsheet with a few simple formulae. I do have some reference material at https://www.hanford.gov/safety/vpp/trend.htm which people are free to use. Where are you located? There are several of us on the Cove that could be of assistance, or check your local ASQ section. Of course, there is always hiring in a consultant (or one of us, for that matter). I'll just state that I am going to Los Alamos later this month for 3 days on their nickel to help their management get an introduction to SPC.

 

[martin elliott]

Basic Need

Larry

You don't say why you do the initial study.

In a simular situation, I have used capability to prove a tool is dimensionally correct, as opposed the process, then returned to atribute methods thereafter. Unless you change/modify/insert new tool then you do not need to repeat.

If its used for proving a process then you will need to repeat, even though not called by customer, as by doing the study you are in effect declaring it a Key feature.

SPC is not difficult in itself, but the understanding of it by the operators is. simple Excel is good enough provided you understand the background

Martin

 

[larry hopperrath]

When i started with this company, no initital studies were done at all. During our initial QS audit, the auditor didnt like that! So, to make him happy, we started doing them on what our team considered ''critical'' to our customers. Our auditor we pleased with this. During our last audit, he asked if we had evidence to proof that we exceeded the capability at PPAP. This is basically what is driving this here. I just figured that if we are going to do it, lets try to actually get something beneficial out of it.

 

[lanaretriever]

Spc

Larry,

I think, if you don't have a specific tool for SPC, you could use a trend graphs, I mean, you recolect process data and in Excell you can do a trend graphs, its not necessary to elaborate always a SPC program, remember that Excell is stronger tool.

 

[Steve Prevette]

Larry,

I think, if you don't have a specific tool for SPC, you could use a trend graphs, I mean, you recolect process data and in Excell you can do a trend graphs, its not necessary to elaborate always a SPC program, remember that Excell is stronger tool.

I definitely would NOT recommend the trend line generator in Excel without the knowledge of what regression does and what its limitations are. For a direct example of misuse of regression, see my August article on "Liars figure and figures lie". You're much better off with SPC.

 

[Paul F. Jackson]

Larry, "Basically" I think that you got a lot of good basic suggestions in the other thread! If your company has been dillegently checking the stampings hourly with attribute gages then you may have a wealth of data. You can use the records of pass/fail vs. pieces checked per frequency (hour/ day) to demonstrate PPM defective for the audit. You just have to crunch the data or begin doing so. Get an SPC manual from AIAG or somewhere else and read!

Stampings can behave different statistically from other machining processes (Non-normally distributed pierced hole sizes)(variable position tolerances)(overall profile envelopes)... and it is typical that the production is checked with attribute gages. So go with it. Do some attribute control and P charts. If you must demonstrate for your audit that you are capable of monitoring and controlling something with variables data do so with feature sizes and read up on monitoring and reporting variables data in sub-groups. By doing so you don't necessarily have to understand or apply other distribution functions.

I don't think that the quality monitoring methods that have kept your company is business till now have to change completely for the sake of an audit you just have to demonstrate with data why they work and if the auditor doesn't get it right away you may have to discover it together. That or just fire him and get another!! Sorry, that's about the most caustic thing that I've ever said in this forum Mea culpa.