Triumph-2 Antenna Model ARP

Jim Frame

Well-Known Member
(I wasn't able to reply to Kelly's thread on this subject, so I started a new one.)

First, I want to say that I saw this coming. As soon as I saw the original 2A model designations, I thought, "Uh oh." Specifying the bottom of a removable adapter as the ARP not only broke from long-standing tradition, it also just about guaranteed confusion. While it would have been better never to have gone down this path, at least an effort is underway to rectify it.

But I'm not convinced that it's been rectified yet, at least as far as OPUS is concerned. The image that Kelly posted showing the "new" JAVTRIUMPH_2A NONE isn't the one that currently appears on the NGS antenna calibration page:

http://geodesy.noaa.gov/ANTCAL/LoadImage?name=JAVTRIUMPH_2A+NONE.gif

Also, I happen to have 6 multi-hour observations on a wide-open point that collectively call into question the model, or at least the way OPUS is handling it. 4 of them were made with my JAVTRIUMPH_2A NONE between December 26 and December 31, 1 of them with aTRM33429.00+GP NONE on January 26, and 1 with a TRM22020.00+GP NONE on January 27. All occupations used SECO 2-meter fixed-height tripods, and all were reprocessed through OPUS today.

The TR2 had the 25mm adapter installed, as that's the only way I have to get the unit on the tripod's 5/8"x11 mount:

TR2+Adapter.jpg


Thus the ARP height I used for the TR2 OPUS submissions was 2.025 m. The 2 Trimble antennas go right on the tripod's mount, so I used a 2.000 m ARP height for their OPUS sessions.

The NAD83 ellipsoid heights returned by OPUS for the station using the TR2 ranged from -20.066 m to -20.087 m, for a spread of 0.021 m and a mean of -20.027 m. The Trimble 33429 session came in at -20.044 m, and the Trimble 22020 session at -20.027. The 2 Trimble antennas are very similar and well-characterized, so I consider the spread of 0.017 m and the mean of -20.036 m to be valid.

The difference between the TR2 mean and the Trimble mean is 0.040, which I regard as significant given the number and length of the observations.

If OPUS is not handling the TR2 ARP height correctly -- by using 84.3 mm as the ARP-to-phase-center, instead of 59.3 mm -- the the TR2 mean would come within 15 mm of the Trimble mean. That still seems a bit large to me, but much more reasonable than 4 cm.
 

Kelly Bellis

ME PLS 2099
5PLS
Jim,

First off, thank you for posting and starting this discussion. I purposely had started the other thread (NGS IMPENDING ANTENNA CALIBRATION CHANGES) as simply a heads up to Javad GNSS users. Vlad Prasolov made the second post in that (closed) thread as soon as NGS had thrown the switch for which I'm thankful for Vlad's quick response to NGS. At that time, NGS appears to have had things screwed up - as you've correctly noted - and Vlad's timing was perfect in documenting it. Since then, NGS has updated its illustration and is reflected in the link you gave above
http://geodesy.noaa.gov/ANTCAL/LoadImage?name=JAVTRIUMPH_2A+NONE.gif

I know that your thread is targeted to just the Triumph-2, but really some of the issues pertain to other non-conforming antenna models. NGS says that it has deleted the Triumph-LS with its ARP on the bottom of the unit and notes that the Triumph-LSA accounts for the 25 mm adapter but then their drawing is still showing the old ARP. Maybe they'll fix that soon too.

Second, this general subject of Triumph-2 related heights has been under scrutiny for quite some time and thanks to Shawn Billings for his eagle eye in noting back in April that something was looking hinky. Matt Johnson's theory that NGS might have had incorrect Phase Center Variations was eventually born out when late last year IGS completed their investigation - more on those differences later.

Unfortunately, NGS has conflated two completely different issues; i.e., 1) the little 1/4" x 20 adapters on ( e.g., TRIUMPH-2A and TRIUMPH-LSA) or off ( e.g., TRIUMPH-2 and TRIUMPH-LS) and 2) the new PCVs. The net result: making a confusing situation even more confusing.

Third, like we see in those illustrations that both you and Shawn correctly observed - this helps to highlight the reason why every surveyor needs to carefully vet their OPUS results.


non-conforming antennas 20150128.PNG


Their new drawing doesn't help me very much. And the color-coded tabular data Last modified by antenna calibration team Jan 30 2015 (link given: http://geodesy.noaa.gov/cgi-bin/red...v/sitemsgs/OPUS/non_conforming_antennas.shtml) doesn't really help me either. On top of that, their list of Non-conforming Antenna Models has changed and I'm sorry that I didn't capture the static image of their listing prior to this.
 

Matt Johnson

Well-Known Member
5PLS
Jim,

When I reprocessed some Triumph-2 data with the new calibrations (adding 0.025m for the adapter) the average height came within millimeters of the known height established from other receivers. The peak to peak spread was 4 cm. The data consisted of 24 observations of approximately 12 hours each that Shawn had recorded on a post monument he has established in his yard. With the previous NGS calibrations we were seeing considerable vertical error and this was the reason the receivers were sent to GEO++ for calibration.

I would suggest you first wait til the precise orbits are available for your recent observations and then reprocess the recent data to see if there is any change.
 

Shawn Billings

Shawn Billings
5PLS
With the previous NGS model I observed significant differences in heights with the Triumph2 processed with OPUS compared to results with other receivers processed with OPUS on the same point. I am now seeing substantial agreement with the new Geo++ calibration parameters. That NGS changed their antenna model nomenclature is going to be problematic. In my tests I have been adding the 25mm offset to the ARP height just as you did, Jim, and have seen good results.

I have heard of a potential discrepancy on the order of what you are describing in California. I hope to hear more about it soon.

I see Matt and I posted at the same time (again).
 
Top