Visual Inspection - Capability Study on Attribute Data

[Barahir]

Is it possible to do a capability study and get a Cpk or Ppk value on an attribute dimension? Specifically I'm looking at an end finish where the inspection method is visual - there is no numeric value for which I can measure or test - it is a pass/fail attribute. Boss wants to know if I can get a process capability study on something like that. Do I do visual inspections and put a 1 for pass or 0 for fail? I'm not sure that something like this is at all possible.

 

[ScottK]

Re: Capability study on attribute data?

If you can't measure it you can't get a Cpk.

That's not to say you can't do a capability study on an attribute.
The goals are just different.
You can define defects per million opportunities and use that as a measure, comparing to what your acceptable level would be. Or Z. That would work too.

For example - your end finish is the only attribute you're concerned with.
So you have two opportunities per part for this defect because there are two ends. (I'm making an assumption here - it may be more or less because I don't know what you make)
You inspect 100 randomly chosen parts and find one defect.

Therefore you have:
.010 defects per unit
.005 defects per opportunity
5000 defects per million opportunities

This would correlate to a Z value of 2.58 I believe.

 

[Barahir]

Re: Capability Study on Attribute Data? Visual Inspection

Thanks, that confirms what I was thinking as well. I told him that I didn't think it would be possible to get a Cpk or Ppk on an attribute dimension, but that we could base capability off defects/rejects. The end finish is only on 1 end of the product, and if we've produced roughly 20,000 parts with zero defects that would (to me) prove the process as either:
1. capable, or
2. really lucky.

 

[bhartigambhir]

Re: Capability Study on Attribute Data? Visual Inspection

Hi

For Attribute Data Like Visual Inspection A Study Called Kappa In Msa Is Availabile.... I Have A Soft Copy To Conduct This Attribute Study......

In This Study They Give Us An

- False Alarm
- Miss Rate

 

[Bev D]

Re: Capability Study on Attribute Data? Visual Inspection

I beleive the last poster is confusing MSA with process capability of categorical data (aka attributes).

Of course you can calculate a Cpk (technically a Ppk value) from categorical data. Many customers actually insist on the Ppk value instead of the yield number. go figure
You simply go "backwards" from your defect rate to a Z score to the Ppk/Cpk value. Ppk = Z/3

Thre are 2 reasons we wantto know a Ppk score:

1. what is projected yield
2. what is spread of process (for stackup issues etc.)

only 1 really applies for categorical data.

The true complicating issue here is when you have a highly capable process and you have no defects for large volumes. I typically calculate the upper confidence interval for the exact binomial and use that value as my "yield" number. (yes it's a worst case estimate, but it provides context between a sampel size of 20 and 20,000) there are other accepted approaches and I can post them later if someone is interested...

 

[massfrompak]

Re: Capability Study on Attribute Data? Visual Inspection

can someone let me know how attributes can be deal
regards
ahsan

 

[himanshugosain]

Re: Capability Study on Attribute Data? Visual Inspection

Hiya,

u can use Binomial Process Capability Chart if u know the number of defective parts / %of defectives.

If u know the number of defects then use Possion Capability Study Chart.

Cheers, Himananshu

 

[rjfavro]

Re: Capability Study on Attribute Data? Visual Inspection

I'm in the same boat. My boss wants defect tracking on visual defects of automotive glass. There are 30 visual reject types.

any help or examples
thanks

 

[Sean Kelley]

Re: Capability Study on Attribute Data? Visual Inspection

We have done a visual capability study on defects in steel. We used 3 inspectors and had 50 samples with 25 good and 25 defective. I was quite surprised how well they did with less than 10% error in identifying the good vs defective. I used Mintab and it has a very user friendly format for these if you have never done one before.

 

[Bev D]

Re: Capability Study on Attribute Data? Visual Inspection

Again there seems to be two interpretations here:

1. a capability study on the process: what is the stable rate of defects that is produced by this process (capability study, Cpk)
2. what is the capability of the inspectors for these visual defects (MSA, kappa score, McNemer's test)

it would be helpful if rjfavro would elaborate on what he is really looking for. saying that 'his boss is looking defect tracking' makes me think he is simply looking a control chart...