[Ja...] Posted March 21, 2019 Share Posted March 21, 2019 I am scanning a face and attempting to measure a .010 profile. The profile is governing a .200 basic dimension in the X-axis. I am using 6 points and 3D alignment. The problem that I am seeing is that when I measure profile, it seems high. Looking at additional printout shows X measuring .1960. In the feature window itself, it says X measures .2002. I understand profile is taking the lowest data point, but I assume a LSQ feature would mathematically find the "average" of all data points. Exporting feature to points shows me only four or five (out of ~300) points at around .2005. The rest are all around .1960. Why is the feature window telling me X = .2002 if it is a LSQ feature and most of the points are below .200? The Z/X angle is only measuring 5'. Both the feature and the profile are reported in the same alignment. Thanks. Link to comment Share on other sites More sharing options...
[Cl...] Posted March 21, 2019 Share Posted March 21, 2019 Using FFS, or 2d curve? Profile to it's self, or to a datum? If to it's self and a FFS, what does it look like using "best fit" & create alignment? Link to comment Share on other sites More sharing options...
[Ja...] Posted March 21, 2019 Author Share Posted March 21, 2019 Clarke, I should have elaborated. The profile is called out to 3 Datums. Link to comment Share on other sites More sharing options...
[Cl...] Posted March 21, 2019 Share Posted March 21, 2019 OK, so you can't use the "best fit/create alignment" option. Link to comment Share on other sites More sharing options...
[Cl...] Posted March 21, 2019 Share Posted March 21, 2019 Is your alignment iterative (looped)? Link to comment Share on other sites More sharing options...
[An...] Posted March 21, 2019 Share Posted March 21, 2019 Please sign in to view this quote. Contribution_21_03_2019.pdf Link to comment Share on other sites More sharing options...
[Ja...] Posted March 21, 2019 Author Share Posted March 21, 2019 Please sign in to view this quote. The alignment is looped. Why can I not use 3D alignment when I have datums? Link to comment Share on other sites More sharing options...
[De...] Posted March 21, 2019 Share Posted March 21, 2019 Profile is a zone tolerance. Regardless what algorithm you choose, profile is going to report 2x the worst deviation from nominal. That said the correct algorithm to use on a feature being checked for profile would be Minimum Feature since it is specifically made for characteristics that are zone tolerances. Link to comment Share on other sites More sharing options...
[Cl...] Posted March 21, 2019 Share Posted March 21, 2019 Please sign in to view this quote. I offered this as an option before you explained that the profile was to datum's. You cannot use the "best-fit/create alignment" option under evaluation unless the profile is to its self. It creates a best-fit alignment using all the points in the feature. If you have datum and you use this, it would be cheating, and would give false/inaccurate data. Link to comment Share on other sites More sharing options...
[Ja...] Posted March 21, 2019 Author Share Posted March 21, 2019 Clarke, I don't think that I am understanding you properly. When I go into my feature evaluation menu, I don't have any choices relating to best-fit/create alignment other than the evaluation method, for which I am choosing LSQ (this may be what you are talking about?). As far as the base alignment goes, I went through [Resources] -> [Utilities] -> [3D Best Fit Alignment]. I align the part to six points, iteratively, on the fixture. I then load the part, scan a face, and try to report profile to the base alignment (actually, the base alignment offset by a rotation of 10 degrees). The only problem that I am seeing is a large difference between the feature window reporting the X-value of the feature, and the Profile additional printout reporting the X-value. They are much different. Is this because the profile finds the one point that is farthest from nominal and uses that to determine profile (by doubling etc...)? Whereas the feature window reports more of a mean? If so, my point export does not reflect this. Link to comment Share on other sites More sharing options...
[Cl...] Posted March 21, 2019 Share Posted March 21, 2019 Thought you might be using a Free Form Surface. Link to comment Share on other sites More sharing options...
[Ja...] Posted March 21, 2019 Author Share Posted March 21, 2019 Clarke, The confusion makes sense to me now. We have curve but not FFS. Sorry about that. Link to comment Share on other sites More sharing options...
[Me...] Posted March 21, 2019 Share Posted March 21, 2019 The feature location is determined by the feature origin, not a calculation like average/min/max, etc. To test this, create a point feature, make it theoretical, and input the X/Y/Z location of the face into the point nominal. The location of that point is the feature origin and should match the local feature alignment trihedron. Often the origin isn't even on the actual surface you're measuring. For example, a plane that's not square/rectangular. This is a great example of why using Cartesian distance to measure between features can be way off and inaccurate. It uses the feature origin. A good example of this is the distance between 2 lines that are supposed to be parallel to each other but aren't. Cartesian won't give you the worst case scenario. I personally use Position whenever I can for this reason. But that's a whole other conversation. Link to comment Share on other sites More sharing options...
[An...] Posted March 22, 2019 Share Posted March 22, 2019 Please sign in to view this quote. See attached!Contribution_22_03_2019.pdf Link to comment Share on other sites More sharing options...
[De...] Posted March 22, 2019 Share Posted March 22, 2019 A profile call out must include a datum reference frame, if one is not specified then Calypso checks the feature to the base alignment. The specified datum reference frame and the basic location values of the feature determines the nominal position of the profile. The measured location of the feature is only used to determine the maximum deviation of the surface from the nominal position of the feature. You could potentially use the same feature for all three datums in your profile callout, and then the evaluation method chosen in the datum features would modify your result, because those evaluation settings would be changing where Calypso determined the nominal position of the feature was. If you were looking for form (which is the only reason I could see to make profile to itself) then minimum feature would be correct algorithm to find the form of the feature to 'itself'. Link to comment Share on other sites More sharing options...
[Er...] Posted March 22, 2019 Share Posted March 22, 2019 Hi Jacob, unless I am reading your post wrong I believe the problem you're experiencing is with where the reference point for each X value is being reported from. When you open the feature window the X value is shown at that features origin point, which for a plane would be one corner, when you output the x value you can check the evaluation setting's reference point. Since you are looking for it's average you want to change this to "center of gravity". That will give you the X Value at it's average like I believe you are looking for. Additionally, I always like to switch my profiles to "two results" instead of "one result", I feel like you have more information when trying to evaluate a surfaces attitude this way. Link to comment Share on other sites More sharing options...
[An...] Posted March 23, 2019 Share Posted March 23, 2019 Just adding two basic dimensions in my PDF. Always regarding a "Non DRF Profile". See attached.Contribution_23_03_2019_.pdf Link to comment Share on other sites More sharing options...
[Me...] Posted March 25, 2019 Share Posted March 25, 2019 Please sign in to view this quote. In order to have the "X Value" compute the average location of all data points, you will need to constrain the normal vector of the feature. This assumes that you're working with a LSQ plane. Please sign in to view this quote. Keep in mind that the exported points are unfiltered, and will include any outliers that may have been identified. Additionally, they will also include the points that are automatically masked at the start of the scan (should you have that setting activated). Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in