The absolute GNSS antenna calibration with a robot is operationally executed by Geo++ since 2000. In the last years, the conducted antenna calibrations produced an extensive database of individual antennas, different antenna types and setups. The robot calibrations can provide absolute phase center and variations (PCV) of GNSS antennas for the GPS and GLONASS observables L1 and L2 as well as antenna-receiver dependent carrier-to-noise decrease pattern.

Investigations on repeatability of individual GNSS antennas and models are possible using the Geo++ GNPCVDB database. The number of individual calibrations of one antenna type gives insight into the quality of antennas series. Also long-term analysis of individual antennas have been carried out. The analysis will focus on Dorne Margoline type antennas.

The GLONASS constellation was for a long time not sufficient to perform a GLONASS PCV calibration within a reasonable time period. However, with the current constellation several calibrations for different GNSS antenna types have been executed. The GLONASS PCV calibration differs compared to GPS, because of the different frequencies of individual GLONASSS satellites. Investigations on a frequency independent modeling of GLONASS PCV are presented.

Operationally, carrier-to-noise (CN0) pattern are estimated simultaneously with the PCV during a robot calibration. The CN0 pattern depend on antenna, wiring and receiver. Comparable antenna-receiver CN0 pattern are obtained using the decrease of CN0 instead of absolute values. CN0 pattern can be effectively used for weighting of GNSS observations. The general aspects of CN0 calibration and some examples are presented.

Investigations on GNSS antenna PCV, GLONASS PCV calibration and CN0 pattern using the absolute GNSS antenna calibration with a robot the will be discussed.

The switch to the new absolute antenna phase center correction model is expected to benefit several global GPS parameters such as troposphere, orbits or the terrestrial scale. The latest model named igs05_1365.atx (1365: GPS week of the last file modification) consists of 213 different data sets containing correction values for 59 individual transmitting antennas (44 GPS and 15 GLONASS satellites) and for 154 tracking antenna types.

For only 42 of the latter, absolute robot calibration results from the company Geo++ including both zenithal and azimuthal phase center variations (PCVs) are available. However, these 42 antennas include a great many of those types dominating the IGS network. In order to complete the set of phase center corrections, relative field calibrations from the National Geodetic Survey (NGS) had to be converted to absolute corrections by adding the difference between the absolute and the relative values for the reference antenna AOAD/M_T. Unfortunately, the converted NGS PCVs are both limited to 10° elevation and to zenithal corrections. The 154 different receiver antenna types also include 48 antenna/radome combinations whose influence has been ignored within the IGS so far. Recent tests at Fortaleza where the removal of the radome caused an apparent height change of about 18 mm (J. Ray, IGSSTATION-860) have demonstrated the importance of considering the radome effect. Unfortunately, calibrations are only available for about one third of the antenna/radome combinations existing within the IGS network.

The tracking antenna information mentioned above had to be fixed in order to derive phase center corrections for the transmitting antennas on board the satellites. Due to problems with eclipse seasons and correlations with orbit parameters, long time series are essential to get PCVs and z-offsets of reasonable quality. Therefore, GFZ and TUM reprocessed 10 years of IGS data using completely independent software packages to generate a consistent set of GPS satellite antenna corrections. As the z-offsets showed a significant trend of about 2 cm/a caused by the error in the mean vertical velocity of IGb00, they were all referenced to the epoch 2000.0. Finally, CODE reprocessed about one year of data to derive consistent corrections for the GLONASS (and GPS) satellites on the basis of a combined GPS/GLONASS analysis.

At its last workshop the IGS has decided to switch to the absolute receiver antenna calibration, either obtained by robot field calibration, anechoic chamber measurements or conversion from relative calibration. After finalizing the compatible satellite antenna phase center variation models by TUM and GFZ using data from 1994 to 2005 a complete antenna model was available for testing. Since June 2005 (GPS week 1325) six analysis centers are contributing to a test with the new antenna model in parallel to the routine IGS Final products.

The parallel test has the goals (1) to test and validate the implementation of the new model in the various software packages (2) to study the effects on the IGS products and (3) to generate a new compatible IGS Reference Frame. After successfully solving initial implementation problems starting week 1341 the parallel results were in the final stage for a comprehensive validation.

The paper summarizes the results for the internal quality/consistency of the new products (satellite orbits, clocks, ERP) among the analysis centers and also the consistency to the 'classical' results. The results for the station coordinates and the updated reference frame will be presented in a separate paper. The new orbits and clocks were also tested in parallel precise point positioning applications.

Within the IGS, all the receiver antenna types being used were calibrated with respect to the Dorne Margolin (AOAD/M_T) model. There are some shortcomings with this approach (e.g.: deficient for long baselines, introduction of scale bias, satellite absolute phase center cannot be accounted for properly). It was proposed at the last workshop to switch to absolute phase center calibrations determined either from anechoic chamber results, robotic calibrations in the field, or from relative calibrations. Some calibrations include the effect of radomes. A test campaign was started last year to produce the weekly official IGS products using both relative and absolute calibrations in parallel. These weekly solutions were compared and analyzed. The effect on station coordinates was, as expected, mainly on the height (scale) component. On average, the height differences correspond to a scale change of about 2ppb, which brings the IGS “scale” closer to ITRF. Height differences at stations using the same type of antenna reach the cm level. Similar differences were also observed between different antenna types. The repeatability of the differences from week to week for each AC solutions was generally within ~2mm/~6mm(std) for the horizontal/vertical components. Some biases, sometimes exceeding 10mm, were observed between the ACs average shifts. When absolute phase centers are officially implemented, there will be a discontinuity at all stations. This may have a significant effect on the reference frame realization, the apparent geocenter and the ERP’s.

We will describe the approach developed at JPL for the calibration of the GPS transmit antennas, using the GRACE GPS antennas as reference. We will present the estimated values for the carrier phase offsets and the pseudorange offsets of the GPS transit antenna as 2-D maps and as vector offsets, and contrast them with the ground-based calibrated values. We will discuss our recent test and validation experiments in precision applications of GPS, such as orbit determination, geodesy, troposphere recovery, and the terrestrial reference frame. Finally, we will discuss the prospects for standardization in the international GPS user community.