Activity Feed › Discussion Forums › Strictly Surveying › RTK fixed, RTK lost
-
RTK fixed, RTK lost
Posted by Harold on January 10, 2014 at 6:09 amToday, while surveying in rural Clay county, MS, I set up my GNSS rover on an iron pin between two tall pines and underneath small bare oak tree limbs. The overhead sky was fairly open. The sky to the west had some small bare limbs, but the sky and high horizon to the east had tall pines. I was running RTK fixed when I walked under the pine, and then lost it before setting up on the pin. While leveling my bipod, RTK fixed. I took a 25 count reading and stored it. While taking a second reading without moving the unit, I lost RTK fixed. I waited a minute, and then got RTK fixed again, so I finished the second reading. The two readings nearly agreed, so I inversed between this pin and another one. I was thirty feet south of where I should have been. So, I looked thirty feet north and noted to self, “this ain’t right.” My unit was still set up on the pin, and when I approached to look at the controller display, RTK lost. With a two meter pole and an antenna hanging down on my receiver, I figured that my fat water head blocked the signal. So, I lowered my head, and RTK fixed. Hmmm, sez I. I reset the ambiguities, lost RTK, and then it fixed again. I took another multiple count reading, and this time I was thirty feet north of where I was without moving the pole!
I have found out that GPS sometimes will tell you a BALD FACED LIE, and I really need to watch out for multipath problems, overhead nearby obstructions, water headed surveyors, and squirrelly RTK lost/RTK fixed problems. I have been using my system for about a year and a half, now. Thanks to this board and comments from users, I have been watching for common problems and errors using GPS equipment. So far, my system saves me a ton of time, and I learn more about its capabilities and limitations each time I get it out of the box. B-)
I am using a Sokkia GRX-1 GNSS rover with a network subscription and a Ranger controller with Survey Pro. I also have a base unit.
I saw a teacher friend Facebook post today that said, ” if you read an hour each day in a subject of a field of your choice, you will become an expert in that field in seven years.” Coupled with real world experience, the depth of knowledge would be great indeed.
I think about the guy on the movie “shooter” that told Mark Wahlberg “just when you think you have things figured out, you are wrong!”
Rant off……….B-)
wgd replied 6 years, 7 months ago 31 Members · 47 Replies -
47 Replies
-
I personally consider RTK GPS to be nature’s way of providing more interestingly messed-up surveys in a century when the technology exists to get essentially the final answer for any land boundary resurvey. The 19th century had its compass and chain, and the late 20th and early 21st has RTK. Both are just gifts that keep on giving. :>
-
Kent, that’s one of the sweetest thing I have ever heard you say. Sarcasm light is going off, and on. Every other second. That is, like a GPS reciever. Float/Fixed.
It’s sort of funny, but not funny, Fixed/Float
It’s sort of funny, but not funny, Fixed/Float
It’s sort of funny, but not funny, Fixed/Float
Yeah, like that!!
-
Anybody going to post anything about the solar flair we are experiencing right now? I read the OP and contributed the error and fix/float issue to the solar activity. Of course, I’m no GPS expert but without seeing the sky map and site that’s my .02
-
Excellent point Steven.
RTK is merely a tool.
Used improperly it WILL be disastrous.
Just like any tool. -
Also from Shooter ” I still have the shovel “
I have seen this before , So If I come into a situation like this . I will move away from the point and lay a tape on the ground and then point stake to the point . This way I can compare the distance to the tape . My thinking is it changes the receiver position enough to get a new solution through different holes in the tree. If you leave it in the same spot you could be dealing with the same multi path (Fixed) Bad solution , this way all the parameters change . -
Network RTK seems to be more susceptible to this than standard RTK. I don’t know why, but my hypothesis is in the cellular network and “dropped calls” and it trying to solve for the integers, and coming up with an answer, without all of the appropriate data.
Also, your procedure sounds like you need some tips and tactics in RTK. When you feel you have a suspect shot, the first thing you should do is lose the initialization and regain it under a different constellation at a different time. That may be 5 minutes and 100′ away. Then re-observe the point. You should use this until you find three points within congruence lest you create or have to deal with situations similar to what Kent is espousing.
-
Pablo
Just how big is that stick you tote around for stirring crap? 🙂
-
—– Warning: Rant On —–
It is my opinion that in a network or not, when the ambiguity fixing factor (if there is one) is set to the highest value (99.99%) a receiver should NEVER give you a ‘Bad Fix’. Ever.
It does not bother me if a receiver NEVER fixes. Under moderate canopy that is the correct answer. But if they fix, the fix should be dependably good. I consider a single ‘Bad Fix’ in 100 fix-dump-fix cycles to be patently unacceptable.
I am not talking about fixes that ‘fix-float-fix-float’ in two seconds with reported HRMS that are 0.2′ (which in my opinion are in indication of a bad receiver,) I am talking about fixes that hold for five seconds or more and have reported HRMS values less than 0.03′.
Yes, I have seen receivers fix, proffer the WRONG answer (usually obvious by an elevation that is a meter off), report HRMS values that look fantastic and HOLD said solution for minutes. (At some point the constellation changes enough that the receivers go ‘Oops’ and report a new value.)
That some receivers will routinely eagerly give you Bad Fixes is not indictment of RTK GPS or of networks, it is an indictment of bad GPS GNSS rover firmware. And bad GPS GNSS rover firmware is an indication of bad technology. And bad technology is an indication of a manufacturer that willingly sells %&*$.
I should be somewhat constrained in what I say here, as I will be taken to task for opening my mouth, but…(I can’t shut up)
Over the years I have seen receivers that work well, get new GNSS firmware and then fail miserably with repeated bad fixes. Whoops!
I have tested receivers (some of these are currently offered for sale) that are absolutely f’n horrible. In my standard test they will have 35% bad fixes. (I consider a horizontal error of 2 cm or a vertical error of 3 cm to be a bad fix.) And the receiver that is being dumped alongside them NEVER fails. Same conditions (I swap back and forth between two poles with quick releases), same corrections (network corrections fed through UHF radio so both receivers see corrections at exactly the same time), same baseline (position spoofed to network at a point 3 miles distant.) One receiver is $@%# and the other just works, perhaps taking an extra five seconds to get a Good Fix.
I have tested receivers that work okay with a matched base, but fall apart in a network or with a base from another manufacturer.
I have also run receivers, through 300 fix-dump-fix cycles, alongside failing receivers and never seen a bad fix on the ‘good’ receiver.
Here is my take on this: there are plenty of GNSS receivers that don’t have problems with my standard test. They don’t cost anymore than receivers that are routinely horrible. Why not use a dependable receiver?
Yes, yes, yes. Solar storms, horrible constellations in the afternoon, birds nesting on the base antennas, GLONASS bias uncertainty, weird stuff happening in the network provided corrections, blah, blah, blah. But the good receivers don’t fix sometimes and never (or should I say, statistically in 100 fix-dump-fix occupations don’t) get a Bad Fix.
I would check the settings on that receiver. If you can affect the statistical process and get rid of the bad fixes, then do it. (I suspect that if you are able to tighten up the fixing algorithm, the receiver will never fix.) If you can’t fix-dump-fix it on a quick release, under light canopy, 50 times without seeing a Bad Fix. Then sell the receiver on ebay to one of your competitors.
Receivers that routinely get Bad Fixes are like chains missing links, total stations with bad compensators, and drunk rod men. You don’t need them on your jobs.
I agree that I need to provide a detailed description, with pictures, of the test that I perform. I will work on it over the weekend and provide more information later.
Just my $0.02 here. This is one of my pet-peeves! And sorry to rant this early in the morning. And yes, I am a grumpy old guy.
Mark
-
For everyone: Pine trees are a major source of GPS iffy solutions. Tall pines are worse, and wet tall pines the worst of all. Don’t even think about a good solution near a pine tree. That’s just the way it is.
-
Fix Means A Mathematical Solution, Not A Good Solution
A certain times of the day satellites tend to string out in an almost straight line across two opposite quadrants. If only a few are in the other two quadrants, especially if they are low they are easily blocked out of your solution.
8 such satellites can give you a mathematical fix, but with extremely poor geometry.
If you do not regularly check you sky view you are playing blind man’s bluff with no one to say warm or cold. Satellite quantity has nothing to do with good positions, good geometry has always been a requirement in good surveying.
At those times during the day you need to bite the bullet, sit down and take a static observation. If you don’t have memory in your RTK receivers, then you are just plain planning to have bad solutions now and again.
Paul in PA
-
First, let me say that this is a good example of a professional surveyor using RTK instead of a technician. It didn’t look right and you checked it out. RTK is awesome. It can provide incredible efficiency and better results over other methods, but as in all things, this added benefits come at a cost. That cost is competent users using it competently.
The one critique I’d have to what you described is that you didn’t force a re-initialization between your first and second observations. It happened on its own (or kind of did) because the receiver reported a float solution briefly.
I’ve not experienced the same wrong answer twice between re-initializations, but I would never say never ([sarcasm]to quote my favorite musical artist, Justin Bieber[/sarcasm]). So if I’m in an challenging environment and I can force a re-initialization three times and get the same answer, I feel pretty good about it.
I’ve tested many receivers. I’ve fooled them all. As Mark said, there are some that are obvious. He doesn’t count them because there are indicators there beyond “Fix” or “Float” that should immediately flag the user to problems (high RMS and a moving solution while the receiver is static, and quickly alternating fix-float status being examples). Then there are some that will show low RMS and still be float. I had one two weeks ago on a topo survey. I didn’t know it until the contours around a ditch just didn’t look right. I had low RMS and the solution kept its fix for several minutes in fairly open sky. After a return trip to the field, I found that I had about 5-10 bogus shots. This was in an open field with trees to the North (Southern, Eastern and Western skies were completely unobstructed as well as directly overhead). This is the monster in the closet that only comes out at dark. But I had enough checks in our system to trap it and fix it. Even with the return trip, it was still far faster than any other method we could have used to finish the job. For this job I had a series of shots diagonally that were not part of the TIN that I compared to the final contours. I placed them in difficult locations, such as just below the top of a break line or between a ditch flowline and top bank. This also tested our density to make sure the interpolation was good.
As to fooling receivers, I’ve certainly not handled them all, but I have seen every receiver I’ve test report a bogus fix. I’m holding out hope in a new receiver that reportedly has tremendous redundancy built into the solution, but until I test it, I remain skeptical. The best advice I can give is to build redundancy into your system, particularly on critical shots. Test your receiver in practice scenarios and learn to read other indicators beyond fix and float. Look at your points (as you did) and see if they make sense. RTK is fast, but it comes at the expense of a more sophisticated user. Not all of our ranks are up to that challenge. Better they stick to total stations.
-
amen. our experience too.
I’ve asked people that should know why this is. I suggested that needle length might somehow be attenuating the GPS signal, but they assured me this could not be. Their answer was moisture content in the needles. Could be! Still hard to understand how pine trees, specifically, seem to be the worst trees to work around for GPS. Oh, and guess what? We work in a region called the Piney Woods! 🙂
-
Harold,
You fail to mention what Network RTK software the local RTK Network provider is using.
If it’s a Trimble VRS Now installation, this could very well be the culprit. Recently, Trimble has been upgrading the older VRS software with the newer, Trimble VRS Now software.The TN VRS Now software now requires the RTK Rover send it’s NMEA $GPGGA message to the VRS Now server at least every 30-seconds. In some instance of the VRS Now Networks around the globe, the local VRS Now Administrator can set the requirement for the NTRIP RTK Rover send in it’s NMEA $GPGGA message every 10-seconds.
On the NTRIP RTK Rover side of things, the RTK Rover will often revert from RTK Fix >> RTK Float >> RTK Fix >> RTK Float all day long, definitely a very frustrating thing when your trying to get some work done.
IMHO, just another way TN screws with everybody. Me thinks their trying to minimize the investment in Server hardware, limit how many RTK Rovers can be connected to the VRS Now servers at any given time. If the NTRIP RTK Rover doesn’t send it’s $GPGGA message to the VRS Now Server often enough, it doesn’t think your working any more, automatically disconnects the RTK Rover. Most RTK Smart Antenna Rover’s internal GSM modems will automatically re-connect to the NTRIP Caster, but the RTK Rover user gets stuck in this vortex.
How does a RTK guy combat this problem?
Upgrade your Data Collector software to the currently released version. The newer data collection software will have a new feature to allow the RTK Rover user control how often the GSM, CDMA modem or Data Collector Internal device sends the NMEA $GPGGA message to the VRS Now Network server. The user can choose to send the $GPGGA message at any interval from 1-300 seconds. Try sending the $GPGGA at 30-seconds, if you still observe the RTK Fix >> RTK Float behavior, increase the $GPGGA send rate to every 10-seconds.-BbB B-)
-
> Just my $0.02 here. This is one of my pet-peeves! And sorry to rant this early in the morning. And yes, I am a grumpy old guy.
>
> MarkHell, I like grumpy old guys. 🙂
-
It Is The Needle Length And The Needle Array
Moisture is required to make an evergreen needle an effective antenna, but is the shape and overall configuration that is the killer. Besides the moisture in the needles is the other moisture in the tree that sends the weak GPS signal to ground.
The needles surround a stem in a 360 degree array and stems point in every direction. Take a close look at the needle array on a stem and you will see multiple yagi antennas.
It is very easy to understand why your receiver cannot see the satellites.
Paul in PA
-
Due to screwy results , last year we pretty much said the heck with VRS .Horizontal was decent most of the time , but I kept seeing a couple tenths between points every once in a while . Vertical was worse.
We now bring in OPUS then set up our base and RTK .
Also we had a lot of trouble getting VRS to work in some areas and it hardly ever worked in the afternoon, I traded in our network rovers for Hiper II ,base and rover and have not looked back .
-
It Is The Needle Length And The Needle Array
Yes, it is the H20 in the pine needles that cause the problems with RTK .
Everyone in the Deep south has witnessed this effect with RTK. Even Glonass does not help either even thought some dealers like to pitch that it does. -
It Is The Needle Length And The Needle Array
I wouldn’t say Glonass doesn’t help. Just probably not as significant as some would like to believe (or convince others to believe, as the case may be).
Log in to reply.