Attribute G&R for all Visual Controls in Control Plan


Involved In Discussions
Need your help.. during a TS audit, the auditor indicated that for all visual controls we should have an attribute gage R&R completed. It make sense.. for variable data we have no problem... we are covered.. but for attribute, we normally have those for the ones indicated by the customer as special characteristics. However, not all the visual control had one. Any ideas how could we handle the attribute GR&R established in control plans? This is what we have defined so far:

Identify the type of visual control => Based on special characteristics / others, using a go-no-go gage, using pictures or visual aids.

Create families of those visual controls

Conduct the R&R.

At this time, we still have many questions about it.. like the frequency, if the families are OK...does any one has a best practice that could share with us? or any suggestion how can we control/establish a rule for these visual controls?

thank you in advance!!

Bev D

Heretical Statistician
Super Moderator
I've rarely families of visual defects, mainly because most of the visual defects I typically deal with are so unique. One could have families for such things as scratches, finish, color match, dirt, and the like where the defect would be common across multiple types of parts.

Although a R&R study is typically only done once to qualify the system, visual defects often have substantial operator component and inspectors often drift and shift in their repeatability. SO, I treat visual R&Rs as an annual (or semiannual for critical and difficult defects) event for all inspectors. More like calibration than R&R...

this is tough to create families for. unless the dimensions and the gage are very similar your results will be artificially biased towards the 'great' side.
Frequency for this is once. you will have to recalibrate any rebuilt, repaired or replaced gage, but an R&R would not be called for unless the gage is designed differently than the one you replaced
Top Bottom