The following is an extract from a GNSS survey guidance document that recommends avoiding receiver mixing. Of particular interest to me is the last paragraph concerning the use of RINEX data-
"There are a number of reasons why it is not recommended to mix equipment and antenna types. Firstly, it is vital to only use antennas which have been designed specifically for the GNSS receivers proposed for the survey. The electrical characteristics and preamplifiers used
in antennas are all different, especially between different manufacturers. Use of an antenna not designed for a particular receiver may result in poor, irregular or incorrect tracking of the GNSS signals and at worst, damage to both antenna a receiver.
The second reason is that mixing receiver types within a survey project can result in systematic height errors if the antenna phase centres are not modelled correctly. The processing software must be capable of correcting for the different antenna types. The model of antenna used on site (including the antennas at any COGRs used) must be known and the correct antenna phase model used in baseline processing. Antenna phase models must be either all relative or all absolute, and should be from the same source (e.g. NGS).
In addition, phase centre variations (paragraph 5.3.3) may not be included in the software models. In real time surveys, the manufacturers often use a bespoke data format for the transmission between base and rover. Thus, it is extremely unlikely that a rover unit from one manufacturer will work with the base station from another. Even if such an arrangement is claimed to work by the manufacturer, it is not recommended as best practice.
Apart from the above two reasons it may simply be found on completion of the fieldwork that data from one manufacturerÛªs receiver will not load correctly into processing software from another manufacturer. Data may subsequently be available in RINEX format, but manufacturer specific information can be lost in conversion to this format. Advanced processing techniques may be needed for RINEX datasets, as specialist knowledge can be needed to reduce the possibilities for error."
For static surveys I use CORS data on a regular basis and the data is supplied in RINEX format. The antennas used at the CORS sites are Leica but I may use anything from Trimble to Ashtech on the survey points. To me the last paragraph seems to suggest that RINEX may not be suitable for static surveys. Also what advanced processing techniques for the RINEX datasets are they referring too. Do they mean stripping out the GLONASS data?
Totalsurv, post: 376856, member: 8202 wrote: To me the last paragraph seems to suggest that RINEX may not be suitable for static surveys.
RINEX was created and subsequently developed with static surveys in mind. A properly-designed baseline processor will handle RINEX data correctly.
I have no experience with multi-constellation RINEX data, as my campaign equipment is GPS-only, as are most of the continuous trackers I use. But I have confidence that the folks behind RINEX know how to record non-GPS constellation data, so it falls to the processing software publishers to implement the RINEX standard correctly.
I remember reading and hearing a lot the danger of mixing equipment. With the development of antenna phase models, especially the absolute models, large errors in height determinations have been eliminated.
Working with mixed antennas on a project spanning multiple years where only one antenna type was used previously we found decimeter-level height errors until a model for our new antenna was provided. With the new model, the problem went away.
As Mr Frame points out, the RINEX format was developed for the exact purpose of allowing the mixture of receivers and antennas. By all reports, users can successfully process mixed data using RINEX data and the antenna phase models. Good a posteriori orbits help too.
Assuming the authors of the quoted document are in earnest, there are issues with mixing receivers that are not widely known by surveyors. As an example, take a look at this IGSMAIL message discussing the issue of mixing cross-correlating and non-cross correlating receivers.
https://igscb.jpl.nasa.gov/pipermail/igsmail/2002/003811.html
HTH,
DMM
The only reason that makes any sense to have all the same type of equipment in use would be that all the settings, Height, antenna and other small variables would be constant for all setups leaving less to checklist and cause an unexpected result.
I have never had a problem mixing drinks or antennas.
I had a whopper III on the roof hooked up to GSR 2300, 4 marine antennas hooked up to other GSR 2300's, 2 topcon hiperXTs, 2 CHC X91+s and 4 nearby CORS using various Leica units. I have uploaded to OPUS and I have processed through TopCon Tools after a few Rinex conversions. All roving units are on fixed height tripod measured in the vertical.
My favorite sessions are to establish the whopper - long data sets on 4 NGS points all about equidistant (1.3-1.5 miles) from the Whopper. Fish Pier w/ NOAA Tide gauge, Coast Guard Station, Airport and Dunkin Donuts with an NGS BM set in 1996. Fairly long sessions and all mix and mathc with great results.
That document sounds very dated to me. As long as you have a good antenna model there's no problem with mixing data from different manufacturers receivers, or if there is I've never experienced it. Even doing RTK, the position that's broadcast is to the base phase center, so as long as the base and rover are speaking the same "language" there's no problem.
I think it enhances static surveys to mix up receivers and antennas, maybe it will slightly degrade the accuracy, but it greatly improves the redundancy which is far more important.
That being said I would be very careful about mixing different brands for RTK surveying.
MightyMoe, post: 376929, member: 700 wrote: I think it enhances static surveys to mix up receivers and antennas, maybe it will slightly degrade the accuracy, but it greatly improves the redundancy which is far more important.
That being said I would be very careful about mixing different brands for RTK surveying.
Isn't that what an RTN is doing with mixed CORS serving to all brands of rovers?
I am having an issue right now with post processing RINEX data from an RTN with the receivers being Topcon. Trying to feed the RINEX data into Leica LGO and post process to my Leica receivers, so far no go.
I would tend to agree with original post until proven otherwise and I think it has been discussed here a bunch that running your own base/rover provides more consistent results than using an RTN and I suspect this is somewhat related to all data isn't data and all software works differently until proven otherwise. I think a large percentage of the time RINEX/mixed receivers/mixed antennas/mixed networks work fine, BUT sometimes it doesn't either.
SHG
The RI in RINEX means Receiver Independent. I've never observed an issue with mixing RXs as long as the correct antenna model is used and assuming the RXs are like quality.
I think the 'first reason' doesn't refer to mixing Rx's in the sense of BASE + ROVER, it refers to mixing Rx and antenna at one end. Like using a Trimble Rx connected to a Leica antenna.
Neil
