The GNSS antenna phase center is the electrical point at which the antenna measures the satellite signal, and it is not fixed. This millimetre-level shift can turn into centimetre-level error in RTK. This page explains what the phase center is, why it shifts, how it is calibrated and how it affects antenna selection.
In satellite positioning, the computed point is not the outer housing of the antenna but the phase center, where the signal is received electrically. To tie a position to a point with centimetre accuracy, four concepts must be kept apart.
The physical, measurable reference point of the antenna; usually the center of the bottom surface where the mounting thread is located. Antenna height is measured to this point, and all calibration values are given relative to it.
The electrical point at which the antenna effectively receives the satellite signal. It cannot be seen and is not a physical part; its position changes with frequency and the direction of the incoming signal.
The three-dimensional position of the mean phase center relative to the ARP. The vertical component is usually a few centimetres, and each frequency (L1, L2, L5) has its own value.
The deviation of the phase center from its mean position depending on the elevation angle and azimuth of the signal. It is in the millimetre range; the elevation-dependent component is usually larger than the azimuth-dependent one.
A GNSS antenna cannot behave identically in every direction electrically. A satellite signal arriving at a low angle near the horizon and one arriving from directly overhead are "collected" at different points of the antenna element. As a result, the phase center:
Phase center stability describes how small and repeatable these deviations are. In survey-grade antennas, PCV is kept below a few millimetres; in simple patch antennas it can be larger and harder to predict.

The elevation-dependent component of PCV feeds directly into the height computation. An uncorrected phase center shows up mostly as error in the vertical coordinate.
When the same antenna is used at base and rover, PCV effects largely cancel out on short baselines. With different antennas this cancellation does not occur; without correction, the error can reach centimetre level.
Because of azimuth-dependent PCV, antennas should face the same direction. This is why survey-grade antennas carry a north mark.
Height must be measured to the ARP and entered correctly in the software. A height entered against the wrong reference adds a constant vertical error equal to the PCO.
The mounting surface and nearby metal structures distort the antenna's radiation pattern. Even a calibrated antenna departs from its calibration values on a poor mount.
As the distance between base and rover grows, satellite angles at the two points diverge; PCV effects no longer cancel and calibration becomes more important.
Phase center errors cannot be eliminated, but they can be measured and corrected in software. For this, the antenna type is calibrated and the results are published in a standard file format.
In projects aiming for centimetre-level positioning, it is not enough for the antenna to support the GNSS bands; phase center stability and a known calibration value are also required.
The datasheet should state PCO/PCV or a phase center repeatability value.
L1, L2 and L5 support, with the phase center of each band defined separately.
A suitable ground plane, circular polarization (RHCP) and a low axial ratio.
Sufficient gain and a low noise figure to compensate for cable loss.
For example, the Harxon HX-CVX600A full-band survey antenna specifies a phase center repeatability of ≤2 mm. For a multi-frequency alternative, see the VLG full-band GNSS antenna page. Antenna calibration data and ANTEX files are published by the IGS.
How RTK works, NTRIP, public transport and V2X applications, and Poynting combination antennas.
No. Antenna height is measured physically from the ground to the ARP. The phase center is usually a few centimetres above the ARP, and this difference (PCO) is added in software from the calibration file.
In survey-grade antennas it is below a few millimetres; datasheets typically show values such as ≤2 mm. For antennas without this value, a field test is recommended when centimetre accuracy is the goal.
Yes, in public transport, autonomous vehicle and V2X projects that require lane-level positioning. In applications where metre-level accuracy is sufficient, such as fleet tracking, its effect is negligible.
The software cannot apply a phase center correction. Using the same antenna at base and rover, aligning the antennas in the same direction and keeping the baseline short reduces the error. If the manufacturer provides its own calibration file, that can be used.
Yes. That is why ANTEX records include the radome type along with the antenna type; use with and without a radome can have different calibration values.
Web alanımızı en uygun hale getirmek için programlar kullanıyoruz.