Profile inspection suggestions?

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

  • Profile inspection suggestions?

    Hello Friends,

    Got another question for y'all. I'm trying to perform a profile inspection on a part and I'm having a whole lot of issues. I need to inspect the shown feature as a profile and I initially tried to do it with a linear open scan. That resulted in point spacing that didn't really make sense without taking about 60 points and data that made even less sense. Instead of re-iterating the last 4 days of troubleshooting I've been going through I'm just going to ask, how would you all go about measuring this wonky profile? Ideally I'd like to be able to measure it with only a TTP, but if the best way to do it is with continuous scanning then I can limit this to only run on our scan capable machines.

    Thanks!

    Tortoise

    profile qeustion.jpg
    Attached Files

  • #2
    single points, at whatever spacing you want. If they then want the 'useless' PROFILE tolerance, construct a feature set out of the points and dimension that. However, you could also dimension (location) each individual point using +/- 1/2 of the profile tolerance (assuming it is not un-equal profile tolerance)
    sigpic
    Originally posted by AndersI
    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

    Comment


    • #3
      Yeah I think you are having paralysis by analysis on this one. There is nothing wrong with using vector points instead of scanning. Yes scanning is more data collected but more is not always better. Just go to vector points, create a set from the points, and profile the set.

      Good luck!

      Comment


      • #4
        Scanning is best for profile. It can tell you alot about what s going on when that shape is made. It's not useless, only if your reporting with the new reporting. I use it everyday to trouble shoot many different manufacturing issues.
        sigpicTAU ALPHA PI INDIANA DELTA CHAPTER
        "Due to the highly confidential nature of my job, I am not allowed to know what I am doing" - author unknown

        Comment


        • #5
          Originally posted by Benedictj1 View Post
          Scanning is best for profile. It can tell you alot about what s going on when that shape is made. It's not useless, only if your reporting with the new reporting. I use it everyday to trouble shoot many different manufacturing issues.
          For this type of feature would you use a linear open scan or is there something different that is preferred? I'm pretty new to programming and haven't yet taken the scanning course so I'm just not sure which rabbit hole I should dive down for this particular part. We have a bunch of other parts like this one too so it seems worth the time to figure out how to do it right.

          Comment


          • #6
            I would do this:

            Taking of the points...
            1) Align to datum scheme. Alignment shall recall STARTUP and shall LEVEL, ROTATE, XYZ ORIGIN in that order.
            2) Take a bunch of single autovector points on the geometry. Cover your surface thoroughly & evenly but don't overkill it.

            Dimensional output...
            1) Construct the points into a featureset. Output the SURFACE PROFILE of the featureset.
            2) Output the T VALUE for EACH individual point. Use the tolerance of your surface profile. If bilateral, ± 1/2 your tolerance.
            SF7107(PCD), SF454(PCD), 152614(Quindos), 9159(Quindos), 7107(Quindos), B&S Manual, M&M Gear Checker

            Comment


            • #7
              A linear open scan would be more accurate. if your radius features aren't exactly positioned to nominal, your vector points aren't correct. And the additional sensors in the scanning head will sense the vectors and surface deviations better. Plus you suffer more error triggering a bunch of times rather that one touch and scan full surface.

              Comment


              • #8
                Originally posted by BKulpa View Post
                A linear open scan would be more accurate. if your radius features aren't exactly positioned to nominal, your vector points aren't correct. And the additional sensors in the scanning head will sense the vectors and surface deviations better. Plus you suffer more error triggering a bunch of times rather that one touch and scan full surface.
                Yes, what he said. Linear open scan, try 10 pnts/mm to start, increase or decrease depending on actual size of that feature. Make sure execution tab is set correctly. RELEARN, FINDNOMS (RADIUS OF PROBE), CleraPlane (if necessary), CAD COMP ON.Setting tab, set scan speed to accommodate smallest feature in scan. Scan speed depends on probe diameter and offset force. Look in manual for your probe head. Same with offset force, depends on probe tip/shank diameter and specific probe head.
                sigpicTAU ALPHA PI INDIANA DELTA CHAPTER
                "Due to the highly confidential nature of my job, I am not allowed to know what I am doing" - author unknown

                Comment


                • #9
                  You don't have to do the whole surface in one scan feature. If you want minimum points on a straight surface, create a linear scan there and lower the points per mm, then create a second scan on the radius and bump up the points to catch as much radial surface as you want. Create a feature set of all the scan, or just select all of them in the Xact dimension menu.

                  Comment


                  • #10
                    Originally posted by Sora5 View Post
                    You don't have to do the whole surface in one scan feature. If you want minimum points on a straight surface, create a linear scan there and lower the points per mm, then create a second scan on the radius and bump up the points to catch as much radial surface as you want. Create a feature set of all the scan, or just select all of them in the Xact dimension menu.
                    Good point. You can use a number of different features for different areas of the profile as needed and report them together as one profile. That opens you up to the options of patch scans, perimeter scans, etc. I guess it all depends on how thorough you need to be.

                    I suggest you look at all the requirements for the part you are measuring. Some customers require a certain point density for profile measurements. On the other hand, I have dealt with some customers that just want a few vector points on each surface with snapshots of the model with DimInfo Boxes showing profiles for each point - just a little data to give them warm fuzzies about the part.

                    See if you can get more information about the requirements. Ask the designer or customer. Also look hard for any requirements buried in the documentation. There could be a simple note on the print or PO or any other such document that calls out a specification you should be following.

                    Comment


                    • #11
                      You can still do it in one scan, just change the min and max steps based on angle change(unless that has been eliminated, I'm on a pretty old version).

                      Comment

                      Related Topics

                      Collapse

                      Working...
                      X