GPS staking

Discussion of FieldGenius related issues and questions.

Moderators: Brian Sloman, Jason Poitras, James Johnston

GPS staking

Postby Ned Ferguson » Thu Apr 10, 2008 8:44 am

I have two coordinate points in a file with the exact same values. If I set the RTK base using one point number I get a solution that is approximately six feet different than using the other point number. WHY?
Ned Ferguson
 
Posts: 37
Joined: Wed Mar 01, 2006 6:35 pm

Postby ianw2 » Thu Apr 10, 2008 10:54 am

If you are using a base and rover on a site that is properly established and calibrated and the rod heights are all correct, this cannot happen.

If the coordinates are identical and in the same file, using the same calibration on the same site, they cannot possibly staking out six feet apart. There is something else going on.

Are the elevations the same? Are the numbers exactly the same? I can't count the times I've looked at a number and read something different.

With all due respect, when I encounter such a dilemma, 99 times out of 100, it is "operator error" on my part.

Check the other things first as well as have someone else check the coords.

Ian
Ian Wilson
Ian Wilson Land Surveying, Inc.
Temecula CA 92591
(951) 684-1044
ianw2
 
Posts: 139
Joined: Sun Nov 24, 2002 6:45 am
Location: Temecula, California

Postby Ned Ferguson » Fri Apr 11, 2008 10:37 am

Ian

If the coordinates are identical and in the same file, using the same calibration on the same site, they cannot possibly staking out six feet apart. There is something else going on.

They shouldn't be staking six feet apart, but they are. The coordinates are flat for both points, including Z. I also double checked by inversing the points. The answer returns 0.000 for HD, SD, and VD.

Yesterday we set up on the same base three times to verify what was happening.
Ned Ferguson
 
Posts: 37
Joined: Wed Mar 01, 2006 6:35 pm

Postby ianw2 » Fri Apr 11, 2008 10:54 am

OK. Something else is going on here.

Suggestion: Copy both points to new points. If one point was number 52 and the second point was 89, call the new points 1052 and 1089, respectively.

Inverse between the new points to make sure you’re on the money.

Do the copied points state out 6 feet apart?

If you have an idea about which one is “correct”, try a COGO calc on the “bad” point to create a new point 0.01 feet (or meters) away on a bearing of N 0º E. How does it stake out?

If you don’t know which one is bad, try COGOing both points 0.01 units straight North. How do they stake out?

Is one of the points part of an alignment? Do you have an offset set up for the alignment?

What GPS gear are you using?

There is something funky going on here and we’re both going to do the Homer Simpson head slap “Doh!” when we find it.

I don’t know where you are, but, if you need, call me. I’m in the office today. (Gotta lick up my son in an hour, though…)
Ian Wilson
Ian Wilson Land Surveying, Inc.
Temecula CA 92591
(951) 684-1044
ianw2
 
Posts: 139
Joined: Sun Nov 24, 2002 6:45 am
Location: Temecula, California

Postby Ned Ferguson » Fri Apr 11, 2008 8:24 pm

Ian,
Just to make sure we are not misunderstanding each other...

The two points to which I am referring are both for the RTK base. They are not points with the same values that are staking out in different locations (that truly would be weird.)

I imagine that there must be something in the raw data file causing this.

I am using the Topcon HiperLite. I have a control point that I first established conventionally, then subsequently occupied several times with the RTK so that I have about 4 or 5 different points with the same or near coordinates. Thinking it did not matter much which one I used, I picked the last point number with a description that indicated it was shot by RTK. However I am not sure why the coordinates for the two points in question are exactly the same. When setting up the base, one point works with other existing control, the other does not.

This is a file that we have been working in for about two years now, so it is difficult to keep track of exactly what happened when.

It seems that FG is very picky about the point name for control, even if the values for the point are the same. I have never understood why this must be. A point is a point. If adjustment parameters are applied to the job as a whole and one is "locked in", any point should do IMO.

If the point was obviously way off, it would not have been a problem. The fact that it was only six feet fooled us.
Ned Ferguson
 
Posts: 37
Joined: Wed Mar 01, 2006 6:35 pm

Postby ianw2 » Sat Apr 12, 2008 6:07 am

Ned:

OK…we WEREN’T talking about the same ting. I understand your situation now.

GPS is an odd tool. Unlike every other tool in the land surveyor’s arsenal, there are no intuitive, logical checks to the measurements we make using GPS. For example, with a level, I can look down the road and see that the surface by the telephone pole way down is about my eye height. When we get there, the elevation should be about six feet higher than where I am right now. With a distance meter, we can check the shots with simple pacing. Not so with much of GPS work.

RTK is a strange beast, too. While we try to minimize the sources of discrepancies with our procedures and such, there is an inherent UNKNOWABLE error in every measurement that is different with EVERY shot! The RMS for any given series of shots can easily be calculated and predicted, but the value for a single shot is anyone’s guess.

Every once in awhile, with a total station, I’ve noticed shots that are outrageous; 36,000 feet or so. Clearly, for the attentive instrument person, this is easily caught. Not so with RTK where you’re trying to get to “zeros” on the screen. There is nothing to help you understand that you have a bad shot.

Caltrans (California Department of Transportation) RTK procedures call for RTK collection of any point at least twice using different base station locations under different satellite constellations. Values are compared and discrepant pairs are re-checked.

In my work, I NEVER accept RTK values for control points. I ALWAYS collect fast static sessions of eight to 30 minutes on EVERY control point for post processing. MY main GPS system is my Topcon HiPer Lite Plus Base and Rover. I also have a pair of Ashtech ProMark II receivers. The PMIIs are L1 only, but they act as a pair of “anchor” points for every static session I collect with the rover, giving me, effectively, three bases for each rover static shot. On top of that, in Southern California, we have a huge CGPS network that allows me to download daily raw data files from any number of CGPS stations, often within a dozen kilometers of my job site. In short, the control network I have allows me to resolve my on-site network to miniscule error ellipses.

Y9ou mentioned that if the point had been “…way off…” it wouldn’t have been a problem but that being only six feet was what fooled you. In my experience, the six foot discrepancies are uncommonly large. It is the 0.25’ to 2.0’ errors that are the ones that are most common and most damaging.

We have a surveyor around here who doesn’t fully understand the procedures that must be employed when using GPS and, particularly, RTK. As a consequence, when following behind him, I often find his points to be anywhere from 0.5’ to as much as 3.5’ out of position when I collect his points with fast static sessions and post process the data!

I often joke that these outliers are proof that neutrinos exist. As they whip through the electronics on board the receivers, it is the neutrinos that cause the huge computational errors. It’s a quantum event!

In your case, Ned, look to your procedures and tighten things up to avoid using RTK to establish control points in future.

Cheers…

Ian
Ian Wilson
Ian Wilson Land Surveying, Inc.
Temecula CA 92591
(951) 684-1044
ianw2
 
Posts: 139
Joined: Sun Nov 24, 2002 6:45 am
Location: Temecula, California

Postby Ned Ferguson » Sat Apr 12, 2008 6:04 pm

You have a point Ian, but as I said, the control point was established redundantly. All the observations, checked plus or minus 0.03' or so. So the location of the control point is not the issue. How FG is interpreting it is. Nonetheless, we should have checked into something else, but didn't since it required a muddy trip to the other side of the project. The stakes looked "about right." Another lesson learned. We were in a hurry and only staking a few curb islands. Luckily we caught the error before the curbs were set. But getting back to the original question, I still don't wee why this should ever have to happen.
Ned Ferguson
 
Posts: 37
Joined: Wed Mar 01, 2006 6:35 pm

Postby Jason Poitras » Mon Apr 14, 2008 4:45 pm

Hi Ned,

Could you send me your FieldGenius project plus some instructions on which points you were trying to setup the base on.

Make sure you give me details on what method you were using when programming the base.

I'll take a look to see if I can shed some light on why this is happening to you. My hunch is that you are trying to program the base with a point that wasn't originally shot with RTK.
Jason Poitras
General Manager
[email protected]
User avatar
Jason Poitras
Site Admin
 
Posts: 443
Joined: Tue Oct 29, 2002 5:17 pm
Location: Kelowna, BC, Canada


Return to FieldGenius

Who is online

Users browsing this forum: No registered users and 1 guest