Spider or Radar graph measurement question

[[Ga...]]

Our customer is requesting that a spider or radar graph be performed on the outside diameter of a tube. To perform a radar plot or spider graph we measure the tube radii at 10 degree intervals on 3 planes equally spaced along the tube. The way that we are performing this is cutting a section at the 3 planes on the tube diameter, creating a 3d curve at each location and measuring 30 points around the tube. Once we measure the 3D curve we are selecting the 3D curve feature, checking the best-fit box, selecting the strategy, and then right-clicking on the segment and performing a list of the points. The numbers are the deviation from the nominal point on the tube to the measured point on the tube. We take this data and put in into our excel spider graph chart. Our issue is the first and/or last point is spiking or giving a false number. Example shown in the PDF is point 1 being .144mm deviated from tube nominal. To verify that it was not the tube we measured a gage pin and got similar results. The results are not the same for each tube or measurement. Is it possible that the probe is deflecting too much when taking the first and or last reading? The PDF attached shows the green points (the results of a diameter with the 1st point removed) and the blue points (the point in question). I am not sure if this is the best way to do this plot if anyone has any suggestions it would be greatly appreciated.

Thanks
Gary Soulliere

Spider Plots.pdf

[[Br...]]

Please sign in to view this quote.

It is. Especially when you are scanning. Its typically advised to measure a circle more than 360 degrees, (e.g. 380) and eliminate the beginning and end of the scan. You can do this with curve too if that's what you desire. You can copy/paste the beginning and and end points so that they overlap, and mask out the overlapped points. Alternatively you can set your curve scan to make contact prior to the actual points, but sometimes this gets a little funny on smaller diameters.

As far as the plot goes, you can do a profile of a line tolerance and plot it out. You also could do circularity on a circle feature instead of a curve to get that "spider web" look, but that wont take your size into consideration.

[[Ga...]]

Thanks for the help, it was driving us crazy why the numbers were so off.