Hi,
Anyone know what the actual phase center is from the bottom of the Antenna, the sketch that I have from
Trimble says 0.1491m the IGS file I've seen states 0.1284m which is the correct height?
On the Trimble sketch it also says Nominal Phase center is the reference surface for Trimble Phase center correction tables,
where could I find these correction tables?
Lee D, post: 337889, member: 7971 wrote: Antenna Calibrations
Yeah I had a look at the NGS website it has the height of 0.128m in there Antex file for up, its just when I complete my processing I manually enter 0.1491 I obtain results closer to what I was expecting
I thought there was a little more info on the NGS page. When the R10 came out I wasn't sure what they were using as the reference plane but at least they cleared that up.
They have a picture showing the reference point they are using which is the same as the Trimble sketch, so I'm just trying to figure out where the 2.1cm in difference is, I'm assuming its the difference between the Nominal Phase center which Trimble states and the actual phase center, but I'd just like to see a document confirming my suspicions 🙂
Two centimeters is a significant difference and I would certainly query both Trimble and NGS about this.
In the not to far distant past, there were differing methods of calibrating antennas that caused quite a bit of confusion. Fortunately NGS is now using "absolute" calibration methods that are superior to their former "relative" methods that calibrated phase center variation relative to a Dorne Margolin antenna. Regardless, the vertical difference from the ARP to the nominal L1 or L2 phase center needs to be closer than 2 cm.
The correction tables referred to are probably the phase center variations tables that are test results put out by several agencies. The "actual" phase centers are different for different frequencies and both move around a bit. "Back in the day" the movement was much greater than it is now. The difference in phase center positions from the GNSS satellite broadcasting antenna to the GNSS receiver phase center is where the phase measurement is determined. The user can't do anything about the satellite phase center; however it is modeled to a great extent within the higher accuracy orbital ephemerides. So the L1 and L2 phase observables are slightly different, and the difference down on the ground can be modeled. The phase processing software hopefully models this difference and reduces the output to one measurement rather than two. Usually the L1 phase center is what is asked for in antenna height measurements, and, I suppose, what is used by the software for the actual distance measurement. The positions of the moving phase centers are functions of azimuth and elevation angle of each satellite. The NGS (as well as other agencies) publishes tables of their test results of measured variations for individual antennas. I expect these are the "correction tables" you were looking for.
(A similar kind of phenomena occurs with multipath as it too biases the range and is likewise a function of azimuth and elevation angle of the satellite. It would be nice to have a multipath model for specific sites and software that could respond to it, but that hasn't happened yet as far as I know).
All the above, with more detail, can be found on the NGS website under their "Antenna Calibration" page. Additional info is also available from several agencies in Germany and, I think, the IGS. Essential knowledge for anyone doing highest accuracy positioning with GNSS.
I agree that it's a little unsettling that there's such a big difference in, apparently, the "nominal" phase center and the "absolute" phase center. I can say, though, that we reference all of our heights to either the bottom of the quick release or the bottom of the R10 as shown in the NGS drawing, and the elevations that we've been getting with the R10s are really good.
Normally one expects to see ~18.2 mm of difference between an Absolute and Relative antenna calibration. (See Absolute Ant for some more details.) 18.2 mm is nearly 2 cm, right? And I expect to see the Relative Model L1 offset BIGGER than the Absolute.
But in the case of the R10, from the NGS website we find the the relative and absolute antenna calibrations are exactly the same:
[Relative Model]
TRMR10 NONE L1/L2/L5/G1/G2/G3/E1/E2/E5ab/E6/B1/B2/B3 NGS ( 5) 13/03/14
-0.9 -0.6 [hl]128.4[/hl]
0.0 -0.2 -0.6 -1.0 -1.0 -0.7 -0.2 0.2 0.2 -0.3
-0.8 -1.1 -0.8 0.0 0.7 0.8 0.2 0.0 0.0
[Absolute Model]
TRMR10 NONE L1/L2/L5/G1/G2/G3/E1/E2/E5ab/E6/B1/B2/B3 IGS ( 5) 13/03/14
-0.9 -0.6 [hl]128.4[/hl]
0.0 -0.2 -0.6 -1.0 -1.0 -0.7 -0.2 0.2 0.2 -0.3
-0.8 -1.1 -0.8 0.0 0.7 0.8 0.2 -0.6 0.1
There is more information in the new atx Calibration file:
[INDENT=1]TRMR10 NONE TYPE / SERIAL NO
ROBOT Geo++ GmbH 5 14-MAR-13 METH / BY / # / DATE
5.0 DAZI
0.0 90.0 5.0 ZEN1 / ZEN2 / DZEN
4 # OF FREQUENCIES
IGS08_1854 SINEX CODE
# Number of Calibrated Antennas GPS: 005 COMMENT
# Number of Individual Calibrations GPS: 010 COMMENT
# Number of Calibrated Antennas GLO: 005 COMMENT
# Number of Individual Calibrations GLO: 010 COMMENT
# GLONASS PCV COMMENT
# derived from Delta PCV per 25.0 MHz COMMENT
# for frequency channel number k=0 COMMENT
G01 START OF FREQUENCY
-0.91 -0.65 [hl]+128.36[/hl] NORTH / EAST / UP [/INDENT]
Since five R10's have been calibrated at Geo++ (in Germany) one could look up the Absolute Calibration directly in Geo++'s database [ here ] if you have paid them $125 Euro's for access. (I don't have an account, but perhaps someone else [Gavin?] here does?)
I bet that the R10 was initially calibrated by Geo++ as an absolute antenna model and that a relative calibration was never performed by NGS. Which means that the R10 would not normally appear in the relative model listing at NGS.
Perhaps you can just add 18.2 mm to the absolute model to get a valid relative offset?
That said, is there a legitimate reason to use a relative offset for any purpose now? (That might be another thread.)
M
RCliffWilkie, post: 337969, member: 10285 wrote: Two centimeters is a significant difference and I would certainly query both Trimble and NGS about this.
In the not to far distant past, there were differing methods of calibrating antennas that caused quite a bit of confusion. Fortunately NGS is now using "absolute" calibration methods that are superior to their former "relative" methods that calibrated phase center variation relative to a Dorne Margolin antenna. Regardless, the vertical difference from the ARP to the nominal L1 or L2 phase center needs to be closer than 2 cm.
The correction tables referred to are probably the phase center variations tables that are test results put out by several agencies. The "actual" phase centers are different for different frequencies and both move around a bit. "Back in the day" the movement was much greater than it is now. The difference in phase center positions from the GNSS satellite broadcasting antenna to the GNSS receiver phase center is where the phase measurement is determined. The user can't do anything about the satellite phase center; however it is modeled to a great extent within the higher accuracy orbital ephemerides. So the L1 and L2 phase observables are slightly different, and the difference down on the ground can be modeled. The phase processing software hopefully models this difference and reduces the output to one measurement rather than two. Usually the L1 phase center is what is asked for in antenna height measurements, and, I suppose, what is used by the software for the actual distance measurement. The positions of the moving phase centers are functions of azimuth and elevation angle of each satellite. The NGS (as well as other agencies) publishes tables of their test results of measured variations for individual antennas. I expect these are the "correction tables" you were looking for.
(A similar kind of phenomena occurs with multipath as it too biases the range and is likewise a function of azimuth and elevation angle of the satellite. It would be nice to have a multipath model for specific sites and software that could respond to it, but that hasn't happened yet as far as I know).
All the above, with more detail, can be found on the NGS website under their "Antenna Calibration" page. Additional info is also available from several agencies in Germany and, I think, the IGS. Essential knowledge for anyone doing highest accuracy positioning with GNSS.
Yes, I agree with you so I got in contact with my Trimble dealer, still waiting for him to get back to me!
Surely if there was a greater height difference in the phase centres then Trimble would give various heights in their sketch rather than just one.
I just have an inkling that the although it states the same reference point as Trimble on the NGS website they've actually used a different reference point. Anyhow I'll just have to wait and see.
Anyway, thought I'd keep you informed,
I received a reply from my Trimble dealer and he states:
"there is a 20.7 mm difference between the ÛÏNominal Phase CentreÛ and the ÛÏmean phase centreÛ for L1".
Which gives then gives me the same height as per the antenna calibrations, of 0.1284m.
Bit unclear now why Trimble in their sketch have the height to the Nominal Phase Center and not to the L1 and L2
phase centers, and why when converted to Rinex when logging to phase center they use the Nominal Phase center height?
Perhaps 0.1284 is the L1 phase center and 0.1491 is multi-frequency, multi-constellation?
