Attribute MSA Visual Inspection

S

Sergii

Hi all,

I was given a task to perform an MSA study for our new project. I don't have mach of experience with MSA. I read the MSA Reference manual 4th edition a several times and that about all my experience that I have. So I really appreciate any help.

In our manufacturing process we don't do much of actual measurements, but instead we have a lot of visual inspection. We have a few hand soldering processes that were identified as a critical in the control plan. So I decided to start the study form them. As I understood the sample part selection is very important for the study. It is not clear for me how should I select parts for the cross-tab method. Specially if our line produced mostly a good parts. Should I deliberately create some defects in order to have a bad parts?

Thanks,
Sergii
 
I

Ingo1966

Re: Attribute MSA visual inspection - Need Help!

For your Attribute MSA Experiment you need good parts and also bad parts.

Attached Cross Tab Excel Calcualtion acc. AIAG 4th.

can help you to understand the Cross Tab Kappa Method
 

Attachments

  • Cross Tab_Vers3.xls
    495.5 KB · Views: 806
Last edited by a moderator:
M

milind dhakad

Re: Attribute MSA visual inspection - Need Help!

Yes, I aggreed with sir, as without bad parts you cannot evaluate the operator competancy. How your inspector indentify the bad parts in good parts ?? and for same purpose we do MSA (attribute) so we need the bad parts also
 
I

Ingo1966

Re: Attribute MSA visual inspection - Need Help!

I did a minor correction on Cross Tab Excel File the Description in the % Probability Sheet was not really correct.

With Version 4 it is o.k.

Sorry for that
 

Attachments

  • Cross Tab_Vers4.xls
    495 KB · Views: 628
B

blody

Re: Attribute MSA Visual Inspection - Need Help!

Well, this is an old post and probably the situation wel handled but I wish to share my experience on visual control MSA study I have done and yet to do. In our facility we have two different sorts of visual controls (and many measurements).

First type is rather easy, it is more or less to check if the detail is the right detail (we produce more than 40k different products) so this literally is about comparing the technical drawing and the part and saying that they are similar.
or to check with the system if the details are marked with correct paper.

Second type of visual control is rather more difficult and rather subjective, everyone can have different opinions. This is the one I am yet to do since opinions can vary so much, I need to find the perfect parts (that shall contain both acceptable and non acceptable parts) that will eliminate most of the subjectivity.

For the first one (even though it is super easy) I needed to proove to the auditors that our operators can actually look in to a drawing and compare it with the real product. So what I did was two very similar products and one of those products drawing under the tittle "to be delivered " and said to the operators that this product will be sent to this customer and please sort these 50 pieces out. The result as suspected turned out to be 99,8%=super good.

Sometimes for the sake of prooving we do silly business IMO.:notme:

Ps: :Dcould not help but giggle on the fact that picard, janeway and sisko managed the do an MSA together.:D
 
Top Bottom