A1/A2 and Feature Properties Question

[[Dy...]]

A few days ago, after a calibration ironically, we started having problems running a few of our programs. To my knowledge, we've never had any problems in the past with these programs, but now they seem to have trouble finding the nominal location of a circle path near the end of the part. I realize this may be difficult to visualize, so I will do my best to explain. I'm measuring a cylindrical part with Ø20mm and OA length of 400mm. The stylus was "skipping" along as it measures the circle on the end of the cylinder and would scrape the part as it finished the scan and moved back to the clearance plane. (Tell tale sign something isn't where Calypso thinks it is.) To remedy the nominal search alarm, I just adjusted the nominal search distance before and after nominal to a greater value to located the feature. It would now scan the part, but is showing MANY masked actual points and the scan path of this circle is off the cad approx 1-2mm on one side. This prompted me to dig a little deeper. At first, i thought this could be due to a "bend" in the part, or maybe our base alignment just wasn't getting the job done well enough. (I didn't write this program, just using it exactly how the vendor sent it to us.) I also noticed that the 3-d line used for X and Z origin, in one of the secondary alignments, had a large deviation from nominal in the A2 X/Y field. 0.528µm to be exact. (See photo of 3d line 1). When i checked the two features that are being recalled to create this 3d line 1, (circle 1 & circle 4), i also noticed a huge deviation in "circle 4" X actual. This is the secondary circle, at the end of the part, that is used to create. "3d line 1". So "circle 1 and 4" are used to create the "3d line 1" for X & Z, in a secondary alignment, but only circle 1 is used for the base alignment. The base alignment only consists of "circle 1" for X/Z origin and "point 1" for Y origin. This made me think that it's possible there was a "bend" in the part causing circle 4 to show the deviation from the base alignment X,Z nominal created by circle 1. But, I created a CAD for our "test bar", that we use at the CNC lathes, to run a similar inspection. When i created the base alignment using the same method, I saw similar, albeit, a tad bit better actuals for the secondary circle. The scan looked much better, with very few masked actual points. But, like i mentioned, there was still an unexpected deviation in the "X actual". when i created a 3-d line using these two circle paths, I also saw a noticeable deviation in the "A2 X/Y".

The more i experimented, the more i kept thinking "This has to have something to do with the base alignment." I copied the 3-d line used for the secondary alignment and made this the X,Z origin. Same results, so i switched X,Z origin back "circle 1". I then placed this copied 3d line as the "Rotation in Space". Because this 3d line included circle 4, and was part of the base alignment, to avoid an alignment reciprocity violation, I copied this circle and made "circle 5" (see photo) After i did this, all the actuals were much better! All within 1-3µm.

My questions are:

1) is the X value in the feature properties window, the deviation in X from the base alignment?

2) what exactly would A1 & A2 be for a 3d line that is down the center of a cylinder?

3) Why did making the 3d line the "Rotation in Space" make all the actuals read closer to nominal when using this same 3-d line in the base alignment for X&Z origin had no effect? Wouldn't this basically be the same thing for a cylindrical part?

4) Now that the actuals are correct for the circle, why are the actual points still 1-2mm off the CAD? (See circle 5 after rotation in space. This circle is just a copy of circle 4)

alignment before.jpgalignment after.jpg3d line after rotation in space.jpgscan path circle 4.jpgcircle 5 after rotation in space.jpg3d line 1.jpg

[[Ma...]]

I have such difficulty to read this 🤣

First of all - don't edit base alignment for seeking defferences. Use new alignment from menu.
Then - avoid using no primary and secondary - we got in trouble with that even with using pallet - after service calibration - use start alignment first to get at least granite desk. Manually touch elements, then i would call it safe.