Attribute Control Chart Question - I plan to use to np chart

[Link Xue]

Hello all
Right now our workshop is assembling a plastic part, we purchase all the parts from the suppliers and just assemble it.
The inspect way is see the surface if ok, no marks and burrs, that's all, really a simple thing.
From my point of view, cause here hasn't attributes data, so X-R chart isn't available, I plan to use to np chart, also we only need to know the scrap number. Am I right?
Any comments or suggestions will be most appreciated.


Link

 

[Sam]

If you are QS9000 certified your "scrap" data needs to be devewloped based on the requireemnts of 4.10.1.1. This is zero defects for attribute sampling plans.

 

[Link Xue]

Thanks Sam, but I still confused


Yes, we use the C=0 sampling plan for incoming goods, but does this has the direct relation with the control chart?
Our client need the spc chart in process, my question is which spc chart is the most appropriate and helpful to us in our situation?

Link

 

[D.Scott]

Just finished lunch

The type of attribute chart you use should be determined by your process and what you are looking to represent with the data.

If you process the parts, inspecting 100%, your best chart would probably be, as you say, the NP chart.

The criteria for failure has to depend on the needs of the customer. If ALL nonconformances are "scrap", then that is all you need to record.

It would seem to me that some of the rejects would result from different components and/or even the assembly itself. If this were the case, I would want a Pareto analysis of the defects found. Maybe the assembly defects could be re-worked and are not really "scrap". Also I would want to know which supplier I was having the problems with.

Hope this helps confirm what you it sounds like you have already done.

Dave

 

[Sam]

Link,
The np-chart is used for recording a defective item,i.e., is the part good or bad.
The C-chart is used for recording defects per part, i.e., how many defect types per part. Each chart requires a constant sample size.

As Dave said, whatever fits your process is what you should use.

IMO a c-chart and a pareto chart would be the best combination.

 

[Link Xue]

Many thanks, Dave and Sam

I really appreciate your instructive help.

 

[db]

China, a fantastic place, 1.3 billion people warmly welcome you.

Can you imagine the size of that group hug?