LEAST_SQR Vs MIN_SEP Circle

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • LEAST_SQR Vs MIN_SEP Circle

    G'day folks,

    Just wondering, in general would you use LEAST_SQR or MIN_SEP to check true position of a circle?

    Running PC-DMIS CAD++ 2019 R1

  • #2
    I typically use least squares.
    This decision is dependent upon function of the feature.
    For example: a vehicle's valve cam or other constant single-point of contact component is best measured using MIN_SEP. It will capture the high peak and low valley, and produce a mid from those extents... It would expose the worst possible extent of how that single point is affected.
    Whereas, if you used least squares for the same method on the same valve cam, it would be measuring essentially the mean of the feature, those mid-measurements would add bias to the mid of the max/min delta that the MIN_SEP exposes.

    If the cam's least squares average measured values come in spec, that just means it's overall going to work
    If you measure the MIN_SEP, you will expose any flats that might have been caused by the grinder sitting in one spot at start or end of grind.

    When MIN_SEP doesn't work well, is if you have dirty parts, debris, or rough surfaces; as it will certainly pick up and utilize the outliers as the determination for the measured value.

    Least squares is best for pretty much all other functional aspects of a circle in my opinion.

    Comment


    • #3
      Most of the time I use LEAST SQR for normal measurements, unless the feature is a datum, then I use MAX INCRIBED or MIN Circumscribed (depending on inner or outer circle).

      Comment


      • #4
        Originally posted by louisd View Post
        I typically use least squares.
        This decision is dependent upon function of the feature.
        For example: a vehicle's valve cam or other constant single-point of contact component is best measured using MIN_SEP. It will capture the high peak and low valley, and produce a mid from those extents... It would expose the worst possible extent of how that single point is affected.
        Whereas, if you used least squares for the same method on the same valve cam, it would be measuring essentially the mean of the feature, those mid-measurements would add bias to the mid of the max/min delta that the MIN_SEP exposes.

        If the cam's least squares average measured values come in spec, that just means it's overall going to work
        If you measure the MIN_SEP, you will expose any flats that might have been caused by the grinder sitting in one spot at start or end of grind.

        When MIN_SEP doesn't work well, is if you have dirty parts, debris, or rough surfaces; as it will certainly pick up and utilize the outliers as the determination for the measured value.

        Least squares is best for pretty much all other functional aspects of a circle in my opinion.
        Great explanation, thanks mate! Nice XJ btw https://www.pcdmisforum.com/core/images/icons/icon6.png

        Comment


        • #5
          This will totally depend on how many hits you are taking on a circle aslo. Unless your scanning , you are very un-likley to find the " high " and " low " spot of a feature so your Least SQ. will be alot closer to the actual size when verified with a gage pin.

          Comment


          • #6
            Originally posted by Schlag View Post
            This will totally depend on how many hits you are taking on a circle aslo. Unless your scanning , you are very un-likley to find the " high " and " low " spot of a feature so your Least SQ. will be alot closer to the actual size when verified with a gage pin.
            I totally agree with you regarding the pin, but customer wants it to be CMM'd...

            I am actually scanning, so have over 200 points. TP tolerance is 0.05mm and it can be up to 0.02mm difference between LEAST SQ and MIN SEP.

            Comment


            • Schlag
              Schlag commented
              Editing a comment
              What Anders says below.

          • #7
            The ISO standard says "circular datums *should* be calculated with MAX_INSCR (inner) or MIN_CIRC (outer)". The MIN_SEP *should* be used for circularity (but the PC-DMIS SIZE dimension does that automatically, irrespective of how the circle has been previously measured/calculated).

            The calculation method for other features should preferably be indicated on the print.
            AndersI
            SW support - Hexagon Metrology Nordic AB

            Comment


            • #8
              I do Least Square, and I show MAX and MIN. Technically the only Pin that will actually fit in the hole in my example is the MIN. But showing all 3 will kinda, notice I said kinda, give me an idea if the hole is wacky. These are threaded holes just so you know.

              Capture.JPG
              (In Memory of my Loving wife, "Ronda" who I lost March 7, 2016. I love you baby.)
              They say "Nobody's Perfect." I must be Nobody.

              Comment


              • #9
                Originally posted by AndersI View Post
                The ISO standard says "circular datums *should* be calculated with MAX_INSCR (inner) or MIN_CIRC (outer)". The MIN_SEP *should* be used for circularity (but the PC-DMIS SIZE dimension does that automatically, irrespective of how the circle has been previously measured/calculated).

                The calculation method for other features should preferably be indicated on the print.
                Would you know which ISO standard is that?

                Comment


                • AndersI
                  AndersI commented
                  Editing a comment
                  Datums: ISO 5459 - Table A.1 and the following pages.
                  Circularity, flatness etc. - some other which I don't remember right now.

              Related Topics

              Collapse

              Working...
              X