Traverse closure?

Were you using a manual total station or a robot? Did you take single shots on each cp or were you turning sets?

A manual total station. Single sets.

I'm not completely clear on exactly what you've done, but it sounds like an angular misclosure. A sight to a target 2200' feet away in a mid-July heat wave could have some pretty significant error in it. Have you summed your angles? What is the angular closure?

In what world is 0.128' okay?

I have not summed the angles yet. I was wondering about the heat waves, but how did it get closer?

That's not awful in 7000' with 10+ setups. I'm just trying to figure out how 2 known points in the same traverse leg can be off by such a big difference.

There's any number of things that can degrade positioning accuracy. My knee jerk reaction to a linear traverse that produces a predictable or progressive error proportional to the distance would be an instrument that may appear level, but isn't. Instruments can settle in asphalt, also.

Check the tribrach. Check the level of the instrument and double center your shots. Something might become noticable. Is your prism offset correct?

My guys would be back out there with numbers like that. Like Norman said radiant shimmer and parallax can be real demons at airports. Get out there for a re-do at sun-up before the air gets soupy and shakey.

A couple of angular errors can easily give the results that you are seeing.

Fire up your CAD and do a quick test. Create points #1 and #2 approx. 2200' apart. Now draw a hypothetical traverse of similar size and shape to your traverse from today that closes perfectly on #2. Now go to your 3rd setup and select all subsequent lines including the line from #2 to #1 and rotate 30" (simulated angular bust). Now move to your 8th setup and select all subsequent lines including the line from #2 to #1 and rotate -15". Check your error from the linework (hypothetical traverse) to #2 and linework to #1. The error at each point should be different.

Well, you're at an airport..so send me a ticket and I'll land right there and will figure and fix your problem. (just kiddin)

There are a host of variables (both systematic and random) that may have produced these results and with limited insight on your procedures....sounds like technical obsolescence has been introduced.. What happens when you apply the appropriate adjustment of the 0.652' in your 12 CP'S.. Do the results fall within the tolerance of TOPO.. There was no mention of any angular misclosures.. You say you are closing 'single sets' of HA's and "apparently" they closed, within reason, on each setup. Do you, or did you, shoot each BS's HD.. What kind and version of DC is being used.. Is the Total Station's "constant" conflicting with the DC's setting.. Try computing the traverse in reverse direction and find the likely CP that may be the culprit and perhaps only need to repeat that setup.

You have a busted angle near the end of your traverse.:'(

> I ran a traverse to set control today and came up on a problem I can't quite understand.
> 2 control points on the job (#1 & #2) are about 2200' apart. I setup on #1 and got my backsight on #2, then started setting control. 1st cp set was about 1400' from #1, after that we were setting them about every 300' (about 12 total), then I recorded a check shot on my 1st BS #2, setup on #2 using my new point # and closed on my 1st setup #1. The length of the traverse was about 7100', my check shot on #1 was off .128' which is ok, but my check shot on #2 was off by .652'.
> What is going on here? How can it be out so bad on my #2?
> This is all for horizontal, we run a level loop through everything for our elevations on CPs.
> FYI I am setting this control for an airport runway overlay topo, that is why the CPs are set every 300'.

If error propagation is far from being linear then I would say that somehow there exists some blunder along the traverse. Why not perform blunder test using least squares?

> If error propagation is far from being linear then I would say that somehow there exists some blunder along the traverse. Why not perform blunder test using least squares

Just StarNet it! That fixes EVERYTHING!

(I know that isnt what you meant, but couldn't resist.

I think that your differing errors are the result of the figure created by your traverse. The linear nature of the traverse and probably your distances are far more accurate than your angles. Given the right figure, your basis of position could easily have less overall horizontal distance error than your basis of bearing.

This would be typical.

BTW, I agree with the least squares idea. Constrain to a single point, and your bearing to your original BS, and check for blunders. If you shot some distance ties from the middle of your trav to each end, it might help.

Um, that meets the standard in CT.

Sounds like an angle problem, first verify your closure distances.

Bearing distance 2 to 1 from control.

Bearing distance 1 to 2 from your first setup, you did shoot it didn't you?

Bearing distance calced last setup to 2.

Bearing distance actual last setup to 2 (new number).

Bearing distance old 2 to new 2.

Bearing distance calced new 2 to 1.

Bearing distance actual new 2 to new 1.

If you distances are within your error budget check your angles. Create perpendicular offsets from the closing error lines, new 2 to old 2 and new 1 to old 1. What angle point/points do the offsets point to? Reshoot that first and report back.

Also you should occupy old 2 backsight old 1 and turn angle and distance to your closing traverse point. Remember the caution about a very long foresight from a short backsight.

In fact any closing traverse point should be occupied twice. Once to continue the traverse forward, second to be an extension back of the traverse start. It is generally only changing the point numbers in the data collector and another set of angles. Very good insurance policy to get you on the right track when you have the problem you have.

Paul in PA

> A manual total station. Single sets.

That says a lot right there. Needs to be much tighter and check that your instrument is turning angles good.

Carl

> .. but how did it get closer?
Converging, due to the angular misclosure, I'm thinking.

> In what world is 0.128' okay?
0.128 ft in 7100 ft is 1 in 55,000. And that's okay. But I'm thinking that the real misclosure is 0.652 (1 in 11,000) - which probably still meets minimum standards.

>... Why not perform blunder test using least squares..

> ...Just StarNet it!
I agree. By 2012, LS adjustment should be minimum standard. Anybody who owns a copy of Carlson has an LS adjustment package well capable of handling this problem.

And for that matter, 2200 ft traverse legs on an airport is for the birds. This job really wants to be done with RTK.

In running your control, are you sighting on a fixed target on a tribrach, string, or a plumbing pole?

If I crunched the numbers correctly, sighting on a plumbing pole 300' away and being off just 0.07' then projecting it 2200' would account for the reduction in perceived error. If you are using a pole and not sighting the very bottom, it is pretty easy to be off several hundredths. Or if your pole or tribrach is out of adjustment, the 0.07' might show up. If you have heat waves, that 0.07' might be hard to see.

If it is equipment out of adjustment, going from the long sights to the short sights or staying with equal length sights would not show up as much. That might support a possible equipment adjustment issue.

From the information on hand, my opinion is that the figure you ran is likely to introduce error. You are going from a 300' sight to a 2200' sight. From a 300' sight, I would prefer to see no more than a 600' sight (and that would be stretching it).