GPS Baseline

Poor Field Techniques Or Equipment, Start Over.

Backsight checks off as much as 0.13' requires you to check all of your equipment, methods and methodology.

A large survey requires more than 2 GPS points and of better quality than RTK.

Without a full explanation of what you have done, any other advice may totally miss the point.

If you are in fact setting 2 RTK points for each observation there is little in the way of checking anything. Do a minimum of 3 GPS points and observe all from at least 2 of them to check your angles and distances. That small area check allows you to better estimate the larger area to area errors. It would be much better to have at least 1 static GPS point in each area.

Paul in PA

Prism Constant??

> I wasn't sure what subject to use here. We're doing a large boundary survey using both a GPS rover and a total station for our observations. We set two working points with GPS, occupy and back sight them with our total station, and then locate found boundary corners. My back sight check horizontal errors run from 0.008 to 0.137. How are you supposed to adjust the two working points for errors?

It sounds like expecting the error of a single RTK solution to be reliably better than 0.1 foot. That would be a mistake.

If you're dead-set on using RTK rather than static, I recommend multiple (meaning *at least* two, but more is much better) RTK observations on each point under different constellations (i.e. times of day, and preferably on different days) and using the resulting vector statistics in an adjustment that includes your terrestrial measurements. Anything else is just sort of a shot in the dark, so to speak.

Are you adjusting grid to ground?

Well if I follow you, you set two GPS points then set your total station on one, and backsighted the other, where you discovered an elevation error of about 13 hundredths.

Now if your just using the GPS to bring in a coordinate system into a job site, say NAD 83 for instance, then your horizontal error is spot on.

Vertically it looks like the two GPS points aren't in agreement with each other, which is typical for RTK GPS. You could hold one elevation over the other and use your total station to re-set the vertical value of one of the points. But that assumes vertical elevations aren't super critical. If you're doing flood elevation work, you might consider tying to another benchmark.

Did you occupy the GPS points twice at different times?
Did you do a static session on one of the points (think OPUS) to compare it to the RTK values you received?

> .. My back sight check horizontal errors run from 0.008 to 0.137. How are you supposed to adjust the two working points for errors?
This is a 2-parter. First is how to improve your field practice so that you aren't "adjusting" 0.14'. Second is how you handle the data back in the office.

As Jim Frame suggested, tie your GPS points twice with some time between the ties. Half an hour or so, or more. Watch your PDOP while you are collecting- you aren't going to get good positions with high PDOP. You may be placing one or the other of your 2 point too close to the obstruction. Have your collector record the vectors and not just coordinates.

In the office, using a least squares adjustment package (I like StarNet, there are several others) simultaneously adjust the RTK vectors and the terrestrial measurements. Only the RTK base position gets held.

Poor Field Techniques Or Equipment, Start Over.

I like to resection off 3, but never for boundary. RTK + BOUNDARY = GIS (and not the good kind of GIS)

Poor Field Techniques Or Equipment, Start Over.

> I like to resection off 3, but never for boundary. RTK + BOUNDARY = GIS (and not the good kind of GIS)

Uhhh...you might need more education on how to make that RTK work. That (incorrect) conclusion is proven wrong every day by thousands of surveyors.

Poor Field Techniques Or Equipment, Start Over.

Field Dog

I run into this all the time mixing RTK and conventional. I would suggest that you not rely exclusively on RTK solutions for your control, for a variety of reasons. It's a snap shot in time and side shot as well, no real check. The temptation to keep the observation times short combined with poor satellite geometry and odd multipath often results in bad back sight checks. Here's how I deal with it. Minimum of three minutes observations time on any control point with an eye to the PDOP and potential multipath. What's an extra couple minutes. Keep back sights as long as possible. Make sure all your equipment is in good repair and adjustment. Post process your raw data from your base to your rover observations, don't rely exclusively on the RTK fix. Throw up a third receiver to cook simultaneously and provide a solid base line and process those vectors to the rover for independent solutions. Having that raw GPS data saved is a real life saver. With the addition of the third receiver you can really tighten things up using a LS adjustment. Just had one the other day that tried to bite me, side of a mountain, multipath hell. Think of all the head scratching time you'll be saving starting off with tight control.

Cheers! Willy

From the original post it sounds more like whoever was running the RTK and collector did not shoot the two working control points long enough, or for the short amount of time they were shot whoever was holding the rod wasn't able to plumb it perfectly. I have run into this issue before, however when the two working control points are shot for an extended period of time, I see .000 and .01 consistently. When shooting points with the RTK like all of said, it is a snapshot in time so it is important to shoot a point for a given period of time. I will probably be harassed for this, but I shoot control points for 60 seconds taking 60 positions and rarely see anything over .05 North and East and this is also county control points. Using a bipod or legs to keep the RTK good and still will help cut down on your .13 error.

I don't want to be difficult, but search this forum. There are dozens of threads about this. I say that because each one has different nuggets.

Here is my nutshell for you:

1. You must know the expected accuracy of your RTK setup.
2. Then you must use best practices (search for this info on the webs) to obtain those accuracies.
3. Then you must verify the accuracy of your actual points on the ground.
4. Then you need to analyze whether the accuracy of your control points will allow you to locate/stake your boundary within your error budget.
5.
A. Now you proceed with the boundary, using best practices (research needed).
B. Or if the accuracy of your RTK mission produces results out side your budget, try again.

Not following these, or not knowing the best practices at each stage is a recipe to prove the RTK=GIS adage as correct.

> .... to shoot a point for a given period of time. I will probably be harassed for this, but I shoot control points for 60 seconds taking 60 positions and rarely see anything over .05 North and East and this is also county control points.
The literature is sketchy on this, but I believe that length of occupation has little or nothing to do with good RTK positions. If you have marginal PDOP (>3)and/or poor multipath conditions you can get a "fix" but still get a position that is not very good, regardless of how long you sit on it. If you have good PDOP and little or no multipath perhaps 15 seconds will be plenty. Reoccupying some time later, from 20 minutes to 6 hours later, and averaging the results - preferably a weighted average in a LS adjustment - will mitigate the multipath effects.

>Using a bipod or legs to keep the RTK good and still will help cut down on your .13 error.
Agreed. That's a minimum for control.

Poor Field Techniques Or Equipment, Start Over.

Thanks for your advice! We don't use static GPS. We tried shooting all the boundary corners with GPS, but due to tree cover on most of the site we needed our total station also. I'm new to GPS, so I figured it would be smart to use three GPS points. Sort of like triangulation. However, our company only uses two points. We try to space them pretty far apart. Two were spaced over 900 feet apart. Those points were on a sidewalk because we had to shoot property corners and fences.

Poor Field Techniques Or Equipment, Start Over.

Thanks for your advice! We set our epochs to 33. Sometimes we get an "uh-oh" (bad PDOP?) from our FC-200 just prior to shooting a point. I realize that means we're about to float, but we don't have all day. Poor excuse, right? Does poor satellite geometry mean the satellites are grouped mostly in one quadrant? I have no idea what a multipath is. I once heard someone in the office mention post processing, but I've never seen it done. Maybe I should download OPUS and learn how to do it myself.

Thanks for your advice! Our prism constant is correct.

I hope that was a joke!

?!?!?!

Poor Field Techniques Or Equipment, Start Over.

Shame on your employer if he's sending you to do a survey with RTK in trees without teaching you about multi path. It's an indirect route of the satellite signal between the satellite and receiver. Imagine a tape bent through the trees. No way of knowing the actual distance from one end to the other.

There are two things a surveyor has control over with GPS: time on site and redundancy. To a lesser degree the case can be made that he has control over what the site is and the time of day he does it too. But I'd say it's mostly duration and repetition. I typically occupy control for three minutes. I also like a second observation although that's not always practical. Presence of multi path practically demands it.

Thanks for your advice! We normally have a problem getting fixed after 2:00 P.M. We adjust the antenna mast setting (default is 13) to try to alleviate that problem. I assume the vector statistics will be in the .raw file. We're using TopSURV 7 software and a TOPCON HiPer Lite rover.

Thanks for your advice! How would you adjust grid to ground? Does this mean adjusting the terrestrial measurements by using a scale factor?