Now that I have a Trimble R10 with cellular data capability, I will be using RTK more. I have of course used VRS extensively, but this is the first time i am able to do single base over longer distances. I will be doing some testing, but am interested in what distances others have been able to quickly (and correctly) initialize. Today I was able to quickly initialize at 15 km. In the past, I tested a long line from WV to my office using the WVDOT VRS, it did not initialize. I think it was about 80 or 100 km.
Any feedback or experiences on distances longer than 15 km? As i mentioned, there are two important criteria: it must init quickly (under 60 s), and it must be correct. The above 15 km lines initialized very quickly (seconds) except for one (out of six). it took about 1 minute.
I am using an R8 GNSS at the rover end. I tested a 5700 (no glonass) a few weeks ago at about 20 km, it took a few minutes but it did correctly initialize.
Whoa... Dude... Do not pass go, do not collect $200.
You're treading into a dangerous situation. Do not go past 20km. I keep it under 10km personally.
I'm on my phone, but will post something a little more robust when I get on my PC. Anyone else, feel free to post.
I made 26km yesterday - barely. I had to set the fix confidence to low and wait for several minutes. I wasn't counting on the result, just looking to see what I could achieve. The result agreed with my OPUS observation on that point at 1cm horizontal and 5cm vertical. I think that 1cm was a fluke, but that's what it was.
That was the first time I tried that. I've had great results inside 12km (where I generally work).
Base And Rover Skyview
John,
With the mountains around Pittsburgh, skyview can also be an issue.
You may have 7 or 8 sats at the base and at the rover, but only 5-6 common satellites.
You want the best base situation possible.
Paul in PA
It is refreshing to read fact based advice from an advanced user of RTK. I have been trying for years to relay this sort of information to those around me with little success. My biggest obstacle has been the propaganda of RTK adverts and 'educators' who have no real world experience quantifying results.
RTK is a valuable tool. I use it a lot. It does have limitations founded in science and math. There is no reason to fear or worship it. I am of the opinion that too many users have no understanding of what's in their hand and how the magic numbers appear.
Thanks again for the post...
Thanks Gavin. Very helpful. Struggling at 16.5 miles on Friday, I wasn't sure what "normal" is. This helps temper my expectations.
Base And Rover Skyview
This is exactly why I love VRS, the base stations are usually in places that can get many more satellites as opposed to where my project is.
It would be nice if you could do a "Network plus local base" RTK solution. Connect to your VRS, but set up a local base. You could use the local base to negate the need to model corrections, but use the network to include SVs that the local base cannot see.
Perhaps my logic is flawed, though, and there would be no benefit.
Also, I once heard that the single biggest factor to limit distance from base is the weather, because if there is differing weather at the base or rover - or conditions are inconsistent in between - it greatly affects the assumption that the tropo delay at base/rover is identical. Isn't this pretty much always the assumption when using a single base?
I've always wondered if the corrections generated by the base and the corrections needed at the rover are affected by the SV's used at each location. I usually try to place the base with good sky view all around. So the base gets all the available SV's. Then I take the rover to places where I don't get all SV's but still at least 5. So the corrections from the base is not the same SV's as the rover is using. Can this be a problem?
The base broadcasts range corrections to the SVs. It can determine these corrections because it is setting on a known point. The more SVs at the base the better, as this will increase the likelihood of similar SVs at the rover.
Having less SVs at the rover doesn't change the correction needed, so long as there are enough to produce a solution.
No one really answered my question...will it initialize with a bad solution? I am speaking of trimble here, GPS/Glonass. I would hope that it would NOT init over a long distance with a bad solution. Or if it does it would quickly catch it and drop the init.
One thing I can say about TBC/TGO/GPSurvey is that there may be an occasional false negative (i.e. data is flagged as bad but is really OK), but no false positives (flagged as good but actually bad) as long as there is a sufficient amount of data. My rules/procedures are always a minimum of 15 minutes for static, plus 1 minute per km over 15 km. I have seen a false positive when there is not enough data, say 8 minutes on a 75 km line. But that violates my above rule, so it wouldn't usually happen.
I am not going out and abusing RTK, I just want to get a feel for the maximum safe distance. I would imagine that having multiple constellations would help all of this.
It is actually easy to test it over extreme distances, plenty of free RTK single base connections out there. I have found that it will NOT initialize. The tough part is testing it over intermediate distances, by occupying known points. Now that I have the transmission medium setup (cell), I will test it this next year at all of the dam deformations we do. They all have very accurate points at the dams, at varying distances from my pedestal at the office. Problem is I cannot just use them now as they are all in restricted access areas, and I have to wait for task orders to be able to access the stations on the dams. There are a few i can use, and I will do that as I have time.
I've seen Trimble R8GNSS units initialize correctly on the north shore of Lake Ponchartrain off a base on the south shore - probably close to 42Km. However, I've also seen some pretty significant error in the vertical at that distance. An initialization at 40Km could be off as much as 0.4' in the vertical and still be within specification (more or less).
With Trimble units I wouldn't be overly concerned about a false initialization; even if it happened the software would catch it as soon as you started moving. I would be concerned about the ppm error, especially in the presence of elevated ionospheric activity.
Personally, I wouldn't be too comfortable going over 20Km but I've seen plenty of instances of people with their own reference stations doing it. In all cases that I'm familiar with they had good control to check into to verify the RTK and also understood the potential error. After Katrina several people in the greater New Orleans area put up reference stations on their buildings and used a cellular data connection for RTK. Interestingly, they were all GPS only, either NET-RS or 5700 receivers. There was (is?) a misconception that the reason R8 model 2 and 3 receivers work so much better than the original R8 is the GLONASS; in fact, an R8-2 will dramatically outperform an R8 or 5800 with or without GLONASS because it's a completely different chipset, etc.
