Try measuring between corners of the polygon on either side of the trace, and subtract trace width?
Try setting 0.01mm grid, zoom in and measure it yourself?
Looks the same to me, but all I've got is a blurry video to look at.
On principle, I don't like functions that compact a huge amount of data into a single result. This is such a function. Some problems that immediately come to mind:
- In what location did it measure 0.159mm between?
- How many locations/regions have the same measurement?
- The shapes have complex geometry; how can I verify that it's picked the correct distances?
- What is the exact decision tree that led to this number being given to me?
I don't know if this particular function is reporting correctly or not -- I rarely use it, personally. I have no particular reason to doubt that it's reporting correctly.
I also have no particular reason to doubt that the polygon is being poured correctly. My experience with polygons has been quite good, but goofs do appear from time to time, and polygons in general tend to be one of the more glitch-prone aspects of most EDA tools.
In any case, you seem to be overemphasizing the criticality of this dimension -- CPW is more sensitive to trace width, for those proportions. I find it very unlikely that an "ANT" trace needs precision impedance, and you aren't going to get much better than 10% on an impedance controlled fab anyway, or 20% or worse for generic proto. I'm making a number of assumptions here, so do fill in anything I'm missing.
Speaking of fabs, generic proto fabs may enhance the shape-to-polygon clearance for you. I've seen 7 to 10 mil (up to 0.25mm) clearance set by the fab.
Tim