Conclusion
This companion opened with a rover reporting a precise, repeatable, and quietly wrong position, the result of a single, unexamined assumption: that an antenna is simply a point in space.
An antenna has a physical reference point a crew can measure to, and a separate electrical phase center the receiver actually solves a position relative to. For RTK, the relationship between them is one number, added by either the crew or the software depending on which one is in use. For static and post-processed work, it is a fuller calibration the processing software applies directly.
What This Companion Actually Covered
The antenna reference point is a physical, documented mark. The phase center is the electrical point the receiver actually uses. For everyday RTK, the entire practical task is knowing whether the crew or the software adds the offset between them, and confirming that directly rather than assuming it. Common mistakes nearly all trace back to that one confirmation being skipped. Measuring height correctly, understanding radomes and multipath rejection, and understanding tilt compensation round out the physical side of setup. Static, PPP, and post-processed work asks more of that same relationship: services like OPUS and CSRS-PPP require selecting the correct antenna model and applying its full calibration, oriented correctly to the antenna's own North Orientation Mark, rather than the single average offset RTK conventionally gets by with.
None of this is exotic. It is a short, specific list of things to confirm, once per setup, and it explains the overwhelming majority of small, consistent antenna-related errors a GNSS user will ever encounter.
The Habit Worth Keeping
Confirming a setup explicitly, rather than assuming it works the way the last one did, is the same discipline this entire series has built from its first companion: understand what actually produced a number before trusting what that number means.
An antenna is not a single point in space. It has a reference point, a phase center, and a relationship between them that different software handles differently. Knowing which pattern a given setup actually uses, for RTK or for post-processed work, is what turns a small, mysterious offset into an explainable one.
The next unexpected offset this method gets applied to will very likely trace back to exactly one of the specific, confirmable causes this companion has named.