ISO/TS 16949 - Comply SPC requirements

A

alevici

Hello i'm old date follower of the forum but never dared to post or ask...
It's my time!!
I work in a medium size metal-stamping company started the implementation of ISO/TS 16949.
We are in Automotive field since many years so we have already many requisites, but we a have a huge gap: we aren't running any SPC activity.
We currently check all SC characterist by GO-NGO gage and standing our production (900 part number in production) we can't do another way.
I read about attribute charts on AIAG manual, but standing our internal PPM rate (1200) i think i will have flat unesefull diagrams.
I read about rare event charts, but it seems they won't be anough for ISO/TS 16949 standards
On 1st of July i must start my program.....
Any suggestion?
Thank you to everybody for all the precious information i got fro the Forum!!
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
You can start with a p-chart control chart for proportion defective against your go-no go gauge. Yes, it would be better to have actual measurement data to work with, but at least with the p-chart you could tell if your percent defective changes. Yes, if the defects are really rare you can work with a rate based upon the number of good parts between failures.

Are you willing to share your data?
 
A

alevici

Sure i will... what kind of data?
My goal is to achieve ISO/TS requirements without starting variable measurement in workshop.

Some data about the company
- press from 400 to 800 ton
- about 80 workers
- metal Stamping company lecated in italy
- 900 part numner turn-over
- 35.000.000 annual part production
- Customer PPM 80
- Internal PPM 1200
- Customer requirements is to perfomr SPC on SC characteristc: no mention about the kind of chart
Be free to ask what kind of information you like to know

At the moment we analyze PPM for each part number every month: we compare actual annual value with last year value and current month wit same month of previous years. Unfortunately this is not anough for ISO/TS requirements.
I hope my english is enough good to let you understand what i mean.
 
Last edited by a moderator:

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
I guess my first question is if you have to collect data (and you do), why would you be satisfied with only attribute data?

Here is the problem - you customer requires SPC on your critical characteristics. Typically, they want capability, also. Attribute data does a lousy job of determining either. Also, they are only interested in their characteristic - not a variable that represents process control. It's sad...but true. In essence, they want report card charting - because their particular dimension may be the worst - or at least not the best - for actually controlling the process.

How long is a part typically run before torn down for another setup - minutes or hours or days?

What are your process variables? What do you adjust (X), what do you hold constant (C), and what do you have no way to control (N)? You need to look at these and determine which X dimensions are best to determine what your process is doing. Can they be attribute? Maybe - but if you sample and you need to maintain capability the attribute pass/fail gaging should be at 75% of the tolerance - not 100%.

What you miss with attribute is die health data. Is the die about to go out of spec? Is it time to repair or replace - before it does go out of spec? How quickly is it wearing? Just guessing? Feel good about that?
 
A

alevici

The only alternative to gage control will be CMM workshop control which isn't an option in this moment.
Our mean production run is 8 hours and usually we don't suffer of punch/die wearing. Our failure are usually connected to scrap marks and ruptures that can't be controlled with dimensional measures....
Just to add some information, our customer aren't asking for SPC charts, but i need to meet iSO/TS requirements that ask to control the process with statistical methods.
So just to simplify the discussion my question could be:
Is there a method that it's good for rare events that will meet ISO/TS 16929 standards?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
For the part (or a part) of interest, could you post in an excel file the number of parts made each week, the number inspected (sounds like you are doing 100% inspection?) and the number of failed parts by week say since the beginning of the year?
 
A

alevici

Hallo..i already checked the file for rare events but the question is still if the meet ISO/TS 16949 requirements.
Monday from work i will send a summary of our data.
We don't perform 100% check but e quite modest 1% that isn't much but anough to keep low the cost of eventual scrapping...
For that for same production we produce up to 40 stroke/min
 

Kales Veggie

People: The Vital Few
Hello i'm old date follower of the forum but never dared to post or ask...
It's my time!!
I work in a medium size metal-stamping company started the implementation of ISO/TS 16949.
We are in Automotive field since many years so we have already many requisites, but we a have a huge gap: we aren't running any SPC activity.
We currently check all SC characterist by GO-NGO gage and standing our production (900 part number in production) we can't do another way.
I read about attribute charts on AIAG manual, but standing our internal PPM rate (1200) i think i will have flat unesefull diagrams.
I read about rare event charts, but it seems they won't be anough for ISO/TS 16949 standards
On 1st of July i must start my program.....
Any suggestion?
Thank you to everybody for all the precious information i got fro the Forum!!

ISO/TS 16949 speaks about SPC in 8.1.1 and 8.1.2. The key phrase is "advance quality planning". If SPC is required depends on the risk as determined by the PFMEA.

So, the PFMEA is the key document here.

ISO/TS 16949 does not require the use of AIAG manuals. Your customer may require it.
 
A

alevici

For the part (or a part) of interest, could you post in an excel file the number of parts made each week, the number inspected (sounds like you are doing 100% inspection?) and the number of failed parts by week say since the beginning of the year?

Roghly this is what we do each week
Parts produced 750.000 Parts checked 3750 Fails 900
 
Top Bottom