California Surveyors Forums

"I have thought about setting two control points, one on each end of the project and then tying them with GPS. So that as I traverse using conventional, I can check into the GPS point and adjust my open traverse that way."

Technically that could be considered a close traverse, since you have known coords on each end.

Both your first and second suggestions would work. You can use Star*Net and incorporate GPS, conventional and levels all into the same adjustment using either of those scenarios.

Another option is something we called a diamond traverse. It is essentially a series of skinny diamonds shaped loops daisy chained for the length of your project. At every other stationwhere you would otherwise run an open ended traverse, you set another control point a few feet away. That point is on your "return" traverse. And on those points between these double sets of points, you only set one point, using it for both your "forward" traverse and your "return" traverse, and then. ....

Oh to heck with that. If you have GPS, the diamond traverse is too much of a hassle to deal with. Do either #1 or #2 of your options listed. If elevations are critical, as they are on many road projects, run levels through your control if you set it all by GPS. If you traverse through it, measure your verticals Face Right and Face Left, Forward and Back on each leg, and take them into account in your adjustment. You will get vertical that is on par with running levels with a good optical level.

As for #3, "If it ain't broke..." The problem there is that sooner or later, it will break, and it'll be both embarrassing and expensive. If GPS is an option, why wouldn't you use it?

Although GPS has made many field procedures superfluous, it's worth mentioning that an open traverse can often be strengthened and checked for blunders by "cutting in" one or more remote objects (e.g., antennas, transmission towers, water towers, building corners). By turning angles to positively identifiable objects from as many traverse stations as is practical, you're able to position them by intersection. The accuracy can be pretty impressive when the geometry is good.

.

each end or randomly is a great idea, and then checking with the topo survey.

Basically what you are doing now is running a control traverse and then later(or maybe at the same time or very close thereto) performed a topo survey and I am positively sure you checked the backsight and forsight using topo shots at each traverse point setup.

So you essentially ran the traverse twice and probably using two different data collecting programs.

I have done what both Frame and Page have done and those are great ways to acquire a comfort level regarding whether or not I am in the right place.

Closing the traverse is always good and the best way to do it. There can be a lot of comfort in that. It sounds to me that you have been fortunate, but like I said above, I think that you are running the traverse twice in your current mode of operation.

And I wouldn't be too concerned about the angle adjustment, although it is something to consider(20 arc seconds in 5 miles is 2.5 feet and if you had twenty setups at an arc second a setup. Of course it would all have to be the same direction and it probably wouldn't) I would be more concerned with a blunder, that's the killer. Although 2.5 feet could get me in trouble real easy.

Many times I have established GPS positions at each of my project, which allowed me to check in and adjust as necessary.

I relate this to years ago when we would commence with a backsight between two known geodetic stations and end with a foresight at two other geodetic stations. If I recall correctly, we also turned a minimum of 4 angles and closed the horizon.

Agian, be careful with the vertical (unless your project has considerable room for blending any errors in between control points) using that method.

A few years (and a couple of organizations) back, I followed another crew on a project. They had done the control and we were doing the topo. Turns out they RTK'd the control, all from one base location, one set of observations. Some points had some sky visibility obstructions, but apparently not enough to block sat signal.

The checks on that topo regularly brought up discrepancies of +/- 0.1' horizontal between control points (HD from 170' to 600'), and +/- 0.3' vertical. It was just enough to be a problem for the project. We could have lived with the horizontal, but the vertical caused issues of making it appear that water should flow one direction per mapping, when it was obviously going elsewhere in the field. Better GPS field methods and a level run would have saved us a lot of headache down the line.

That crew had run the control on several projects I had taken over. They never closed a traverse loop, carried vertical only by one Face Direct observation in the forward direction, and when they used GPS, usually took one RTK observation per control point.

Sometimes they would set the base at a 2nd control monument (no direct observations between 1st and 2nd control - assumed coords provided by NGS, Caltrans, or whoever were good as gold in all cases), and take one additional RTK shot on each monument. But they would only use the 2nd set of observations as comparison to the first set. They never adjusted anything.

We had to completely redo the control on some of the projects.


Point is, make sure that you build redundancy in to your field procedures, use those observations in your adjustments, and account for the weakness in the vertical component of GPS observations.

I'm not a GPS expert, so someone jump in and correct me if this is wrong, but I think that if you were to re-occupy each point at some other time, meaning completely different satellite geometry, that would give you sufficient geometry.

Some say that if you were to just break the signal and then let it re-acquire and take a second observation right away, that works. but I've never been quite convinced of that.

Your best GPS solution would be to do as Evan says, occupy a point in the morning and then re-ocupy the same point later in the day or later in the day on a different day. You can break and re-start your GPS occupation right after the first, but your best practices solution is to have a different satellite configuration.

Yeah, would be nice to keep the road within the ROW

"...Some say that if you were to just break the signal and then let it re-acquire and take a second observation right away, that works..."

If you are measuring your control with RTK I think it's a great idea to break (or dump initialization, as you may have heard it called) and then re-initialize. What you are doing is protecting yourself against taking measurements with a [potentially] bad initialization. If you take only one shot you have no redundancy. If you take consecutive shots with the same initialization, the two may agree, but could be significantly in error if you managed to get a bad initialization (or "fix"). Of course if you do not compare those two shots while in the field, you will miss the opportunity to easily take more should they happen to disagree.

This is a separate issue from satellite geometry, which certainly does change throughout the day and results in varying positions. I believe the others whom have suggested redundant observations at varying times in the day offer good advice.

Assuming your GPS receiver performs to its marketed standards, more satellites (redundancy) and lower DOP values (better satellite geometry) should increase the quality and repeatability of your positioning.

So good procedures would include both breaking initialization and taking another shot right after the first and reobserving at another time with different geometry.

Thanks for the explanation of the reason to break signal and reinitialize, Nick.

Definitely the more shots you have, the more redundancy you have. With RTK, you are generally taking a single tie shot to that point. The more shots you have to that point, the more checks you have.

When using RTK for boundary work I like to have several shots from one base, at different times of day, with re-intialization (is that a word?) and then use a second base and also take multiple shots at different times of day. The Caltrans Survey manual has excellent standards for GPS surveys. It can be found on their website.

"So how would i do any kind of traverse correction"

Does the Leica network provide solution quality data (e.g. a variance-covariance matrix)? With single-base RTK you can record the matrix (the QC2 record in Trimblese contains the matrix), which you can subsequently use in an adjustment.

However, limiting yourself to only a single visit still leaves you vulnerable to error sources that can't be quantified from the collected data. Multiple reinits and longer shots can't hurt, but returning on a different day at a time that will yield a significantly different constellation is preferable.

.

I would urge you to take a look at Caltrans for methodology. They have standards that are based upon what type of accuracy and precision is needed for a particular type of survey work. Some of the older surveyors will remember these standards as being refered to as First, Second and Third order work. There are new standards,guidelines if you will,NSSDA and these days folks are interested in positional tolerance expressed in terms of local or network accuracies.

I probably would not use RTK for anything except topo shots, that is just me. By the time you do everything you need to do for the RTK shot,so that it will be usable for control, you will have spent almost as much time visiting the point as you would if you were collecting data for a faststatic session. So why not collect the data in 2 static session?

The reasons for visiting the same point more than one time and at different times of day is to make sure that you have collected data that is subjected to different atmospheric conditions and different satellite geometery. The data collected on the first visit could be affected by multipath. How will you know? Atomospheric conditions do not seem to be much of an issue, given the generally moderate climate we experience in Ca. and for short base lines. Do stay away from the lightining storms!

I would not be a fan of dumping the initialization, because nothing much has changed if the few minutes that it takes to re-initialize and what is to say that the new initalization isn't bad? You will not get new satellite geometery,nor change in atmosphere, so it would not count as a statistically independant observation, especially if you are also using the same base station on the re-initialization. Something that is not dicussed often is space weather. This is cyclical and we have been in a relatively quite period, but more solar activity is expected in the coming years.

The fact is, that the better the quality of the control, the more time it takes.
There is no real way to shortcut this.

Going way back to your original statement regarding methods, I would like to add a thought regarding No. 1. If you have one control point at each end, you can not really adjust your traverse. You need a initial backsight and a closing angle. Thus, two control points at each end would be preferred. Now if you are only going 1000 feet, I would not bother.

"Atomospheric conditions do not seem to be much of an issue, given the generally moderate climate we experience in Ca. and for short base lines. Do stay away from the lightining storms! "

Tropospheric (low altitude) conditions may not be much of an issue in California, but ionospheric (high-altitude) conditions can introduce significant signal delay no matter where you're located. Propagation speed through the ionosphere and SV constellation are the two reasons to reobserve on a second day.
.

Jim, thanks for the futher clarification. Both components are critical and I did not mean to over look the ionospheric issues in my comments, but that is exactly what I did. The standard of visiting the point on two different days at differents times of the day is not meant to be overlooked just because we have mild sunny weather.

I am glad to see that RobCT chimed in with very useful links just before my last post....