iSixSigma

Gage R&R on a Spectrophotometer

Six Sigma – iSixSigma Forums Operations Manufacturing Gage R&R on a Spectrophotometer

This topic contains 3 replies, has 3 voices, and was last updated by  Dan 2 weeks, 2 days ago.

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #243034

    Kenn Armstrong
    Participant

    Hi, this is my first post.  I found a related thread, but I think it was to old for me to add a reply.

    I have the goal of developing a process using objective measurements to evaluate painted parts at our manufacturing plant.  I was thinking that I could run a gage R&R after running a Type 1 gage study in Minitab.  From there, I may be able to play with tolerances on the L, A, B, and Gloss values, to effectively turn the tool into a go / no-go gage that can also provide continuous data.

    I  am struggling with where to start.  I do have a set of “golden samples” that are approved by the customer  and were all produced from the same run.  The trick is that there is a good amount of variation within each of these samples particularly with gloss and L values.  This appears to be primarily caused by the variation in texture within each sample.  I think this type of variation is both expected and an acceptable output of the process.  The texture is from painting technique, not paint.

    My ask to this community is for advise in completing the gage study.  I am attempting to attach a short video that shows the gloss value of 33 near the center of the sample and 21 in the corner.  This is a metalic silver wet paint on plastic.

    Maybe I instruct the operators in the gage study to measure different locations on the samples.  Then maybe in production, I can write in a similar procedure.

    One other question I have is if it’s better in production to use stored L,A,B,gloss values in the spectrophotometer.  This seams like a better way to go vs. re-measuring the golden sample every time as a baseline.

    Thanks,

    Kenn

    0
    #243035

    Mike Carnell
    Participant

    @snooks98 You need to remember you have accuracy and precision. When you are doing precision I would not worry about agreeing with the clients colors i.e golden samples. You want to be the same type of paint particularly if you have a metallic finish. You will get variation that is not a gage issue from different locations on the sample. I would fixture the sample and the measurement device for the study. The point you want to make is will the device repeat and reproduce. Once you have it fixtured then do it again un-fixtured. Know you can figure how much variation is coming from the measurement device and how much due to location on the paint chip.

    Accuracy is going to be a typical study for bias but you want a known standard/golden sample.

    With the metallic finish I would fixture it and move the lighting around and change the wattage as well. There can be some very odd readings just from the way it is lit.

    Just my opinion.

    0
    #243039

    Kenn Armstrong
    Participant

    Mike – thank you – great idea with the fixture!

    0
    #243182

    Dan
    Participant

    Mike Carnell is absolutely correct in his statements above.  I just want to add one other thing.  Once you determine the accuracy / bias contributions of the device through fixturing, lighting, etc., you may find you need to fixture your samples for each subsequent study as you work to optimize the process.  I.e. each measurement you take needs to be in that controlled environment.  Cameras, color, lighting, texture, and then finally, subjective evaluation, can all lead to a tricky final result.  Good luck!

    0
Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.