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.

MSA ANOVA _ Operator by Part interaction

Discussion in 'Gage R&R and MSA - Measurement Systems Analysis' started by YK TANG, Jul 11, 2016.

  1. YK TANG

    YK TANG New Member

    Joined:
    Nov 26, 2015
    Messages:
    4
    Likes Received:
    0
    Trophy Points:
    1
    Just a quick question, if we conduct a GRR, using ANOVA method, and found out that the p-value for Operator By Part ot be less than 0.25, therefore there is an indication of Operator By Part variation.

    And, such a case, Reproducibility % will be boken down to 2 segments, which are 'By Operator%', and also 'Operator by Part%'.

    If the results show that the "Operator by Part %' is high, which is one of the main contribution to the GRR variations, what are the possible corrective actions to be taken here? {imagine this is a micrometer, measuring the thickness of the product}

    Thanks
     
    YK TANG, Jul 11, 2016
    #1
  2. Miner

    Miner Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    645
    Likes Received:
    555
    Trophy Points:
    92
    Location:
    Greater Milwaukee USA
    One frequent cause for this interaction is when the product has significant variation in form and the operators are using different techniques with which to measure.

    For example, lets say that your product had a lack of parallelism between the two planes used to measure thickness. Let's also say that your operators used three different techniques: 1) randomly anywhere on the surface, 2) always on the edge, and 3) always x distance from edge.
    1. Operator 1 would always pick up the variation in form of the lack of parallelism
    2. Operator 2 would randomly pick up edge related anomalies such as burrs
    3. Operator 3 would not pick up as much variation in form and would not pick up edge related anomalies
    The solution is to identify the best technique, train all of the operators in the use of that techniques and document it as the standard method.
     
    Miner, Jul 11, 2016
    #2
  3. ncwalker

    ncwalker Well-Known Member

    Joined:
    Sep 21, 2015
    Messages:
    261
    Likes Received:
    170
    Trophy Points:
    42
    Location:
    North Carolina
    It may also be "clock position" as well as proximity/consistency to neighboring features. If it's a round part, imagine an imaginary mark we define as 12 o'clock. One operator may be using a visual cue on the part to always measure at the same clock position. Others may be moving around and around the part.
     
    ncwalker, Jul 11, 2016
    #3