Feature location question

[bigblue97mustang]

Hello all! In all the programs I write, I use auto features for edge points, circles, and round/square slots.

My inspection standards require that I measure surface deviation and trim length deviation, which I use S for surface deviation and T for trim length to achieve.

For the holes, it will typically ask for 2 axis values, say X and Y, and will also ask for surface deviation. For the longest I have been reporting this in S value as well, but after looking through the help files I believe that might be wrong. There are a lot of terms in the help file, that I don't fully understand. Could someone tell me which dimension option I should be using to measure surface deviation for these features? It should be noted that I use sample hits to define the constraining plane for all my circles and slots.

I had thought to report as X Y and Z deviation instead of whatever I should be using for surface deviation, but the plane most of my features set in are cross-plane.

Any help explaining this to me would be greatly appreciated.

[Matthew D. Hoedeman]

"S" should be the correct value for a LOCATION (not position) dimension of a circle (when using surface sample hits). It should also be the correct axis for EDGE points with a sample hit.

THAT BEING SAID, I stopped using "S" for edge points a LONG time ago as there is a (to me!) serious issue with that. It will work fine for a flat surface (no matter the angle, just has to be flat) but it is a hot wet mess for any edge point you measure on a curved surface (where the curve is parallel to the edge). It only know the XYZIJK at the edge, it has no idea if the surface curves away from the edge. So if you use a 3mm indent, your sample hit is 3mm away from the edge, and 3mm along the curve. If it is a HUGE curve, it might be good enough, but if it is, say, a 10mm radius, you are getting 0.46mm of error for the surface. I use a vector point 3mm from the edge, then a EDGE POINT RELATIVE TO that vector point to control the depth of hit.

[bigblue97mustang]

Thank you for the reply. Hearing you issue with S on edge points makes total sense to me. I suppose I should have noted that all my edge points are on a flat surface and flat edge.

[VinniUSMC]

I'm not sure about the "S" axis for holes, but there is an easy workaround. Construct a cast point on the hole, or a generic point with XYZIJK from the circle. Dimension the "T" of that point.

[Matthew D. Hoedeman]

I'm not sure about the "S" axis for holes, but there is an easy workaround. Construct a cast point on the hole, or a generic point with XYZIJK from the circle. Dimension the "T" of that point.

I tried it, years ago (but still in the version I'm using, V.37) and it was correct. HELP: "S prints the deviation along the surface vector" but that WAS for LOCATION, not POSITION.

[bigblue97mustang]

I'm not sure about the "S" axis for holes, but there is an easy workaround. Construct a cast point on the hole, or a generic point with XYZIJK from the circle. Dimension the "T" of that point.

Man, such a stupid easy idea. I think that is going to be the perfect solution for me. Thank you so much for the suggestion.

Is there any chance you could explain why I would use T rather than S to figure the surface deviation for the constraining plane that circle is in? From the help files, i read that T is the dimension of the 'approach vector', and I'm unsure what that even means. Thank you all again for all your input!!!

[VinniUSMC]

The approach vector for a point on a surface is the same as the approach vector for a hole (perpendicular) through that surface. Approach vector is "what direction is the probe coming from?"

"T" is the generally used deviation axis for vector deviation of a point. "S" is used when you have an edge point and you want to report the edge length (which is T), and the surface deviation (S, in this case).

[bigblue97mustang]

Thank you so much for such a detailed breakdown. I think I have a good grasp on what's going on now. Unfortunately, it looks like I have about 800 holes to go back and re-dimension X.X