1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

GR&R gauge readings

Discussion in 'Gage R&R and MSA - Measurement Systems Analysis' started by Tiger shark, Sep 12, 2019.

  1. Tiger shark

    Tiger shark Member

    Joined:
    Sep 11, 2019
    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    1
    Hello Every one... I am trying to do a GR&R on a height gauge, The nominal as per drawing is 10±0.1mm, all my samples are reading 9.91 and few are 9.92mm, when I enter the data in Mini Tab, I got a result which is over %10 which means is Rejected GR&R. If I manipulate some samples (by deforming them) to read 10.0mm or 10.05 then I have a better result and can make it under %10 which is Good. But I don't want to do the second method, Why the result is so bad when Measuring the original samples? What I am doing wrong? any help is appreciated. ( I use 10 samples/3 Operator/3 trial).
     
    Tiger shark, Sep 12, 2019
    #1
  2. Andy Nichols

    Andy Nichols Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    5,104
    Likes Received:
    2,560
    Trophy Points:
    112
    Location:
    In the "Rust Belt"
    Welcome:
    We have some G R&R wizards here. For immediate clarification, is the height gauge readout digital?
     
    Andy Nichols, Sep 12, 2019
    #2
  3. Tiger shark

    Tiger shark Member

    Joined:
    Sep 11, 2019
    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    1
    Thanks...…. Yes it is
     
    Tiger shark, Sep 12, 2019
    #3
  4. Andy Nichols

    Andy Nichols Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    5,104
    Likes Received:
    2,560
    Trophy Points:
    112
    Location:
    In the "Rust Belt"
    OK, so we don't have a problem of resolution/discrimination. Miner? Bev?
     
    Andy Nichols, Sep 12, 2019
    #4
  5. Golfman25

    Golfman25 Well-Known Member

    Joined:
    Nov 6, 2015
    Messages:
    821
    Likes Received:
    405
    Trophy Points:
    62
    I'm no expert and nothing frustrates me more than gage r&r. But I am going to suggest you don't have enough variation to pick up. Your samples are within .01 of each other. When you increase the spread to 10.05 you're adding that variation. You're getting pinched because you have good parts. Doesn't make much sense to me.
     
    Golfman25, Sep 13, 2019
    #5
  6. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    606
    Likes Received:
    664
    Trophy Points:
    92
    Location:
    Maine
    The variation in your parts is NOT the issue. The traditional gauge R&R uses only the average SD of the repeated measurements of the parts divided by the tolerance spread. To better describe and understand what is happening can you post your data? Also check out my MSA documents and Miner’s in the resources tab...
     
    Bev D, Sep 13, 2019
    #6
    tony s and Andy Nichols like this.
  7. Miner

    Miner Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    577
    Likes Received:
    493
    Trophy Points:
    62
    Location:
    Greater Milwaukee USA
    For what purpose is the height gauge used? Is it used for inspection or for SPC? If it is used for inspection, use the % Tolerance metric, not % Study Variation. If you do use it for SPC, do you have capability data for your process? If so, substitute the capability study variation for the study variation and calculate % Process Variation. This may solve you problem if the study parts did not capture all of the expected process variation. If it does not solve the problem, remember that < 10% is the ideal situation. You can still accept studies between 10 and 30% as marginal under certain conditions. If you have a high Cpk, I would use that to justify a higher R&R.
     
    Miner, Sep 13, 2019
    #7
    tony s, Bev D and Andy Nichols like this.
  8. Tiger shark

    Tiger shark Member

    Joined:
    Sep 11, 2019
    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    1
    The Gauge is used for checking the parts. I will post the result for you soon... Thanks
     
    Tiger shark, Sep 13, 2019
    #8
  9. Tiger shark

    Tiger shark Member

    Joined:
    Sep 11, 2019
    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    1
    I will post the result soon.... Thanks for help
     
    Tiger shark, Sep 13, 2019
    #9
  10. Tiger shark

    Tiger shark Member

    Joined:
    Sep 11, 2019
    Messages:
    5
    Likes Received:
    0
    Trophy Points:
    1
    upload_2019-9-13_15-29-34.png

    Agree with Miner, if I go with %Tolerance I get good result , but % Study Var wont pass.
     
    Tiger shark, Sep 13, 2019
    #10
  11. Miner

    Miner Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    577
    Likes Received:
    493
    Trophy Points:
    62
    Location:
    Greater Milwaukee USA
    % Study Variation does not mean anything if the gauge is only used to measure product to specification. It only becomes important if you are using the gauge to measure product for SPC or capability.
     
    Miner, Sep 13, 2019
    #11
  12. Bev D

    Bev D Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    606
    Likes Received:
    664
    Trophy Points:
    92
    Location:
    Maine
    And %study variance is effected by the part variation. So if all 10 parts are very close to each other your study variance will look bad in the popular R&R method you used.
    A Youden plot will display the actual variance and you will be able to SEE waht is going on. The statistical output table of the popular R&R method is very dense and can be misleading...
     
    Bev D, Sep 14, 2019
    #12
  13. tony s

    tony s Well-Known Member

    Joined:
    Sep 10, 2015
    Messages:
    1,350
    Likes Received:
    1,055
    Trophy Points:
    112
    Location:
    Laguna Philippines
    I agree that the samples obtained for the GR&R study don't have enough variation to produce reliable results. The sample parts should represent the spread of the process variation or the tolerance spread. If you need the results for product control (i.e. conformance/nonconformance) better to use the tolerance spread. If the purpose is for process control (i.e. process stability) use the process variation instead. The number of distinct categories is also low if there's enough variation in your samples.
     
    tony s, Sep 15, 2019
    #13
  14. Gejmet

    Gejmet Member

    Joined:
    Jul 23, 2019
    Messages:
    17
    Likes Received:
    10
    Trophy Points:
    2
    Hi, im not sure if you are in the process of doing it but would it be possible to see your raw data and any charts which your software is producing please? We have the summary statistics but no indication as to the behaviour of the measurement process as yet.
    Thanks
     
    Gejmet, Sep 17, 2019
    #14