With all due respect to my learned colleagues what you are experiencing has nothing to do with heat shimmer or out of alignment tribrachs. And no, do not just use Least Squares to "fix" everything.
What is poorly understood within the profession is that Error of Closure as you describe it is not a function of your raw data set alone. Instead, Error of Closure is a function of your raw data set and where you begin your calculations.
My experience is when the "closing line" is very long relative to the other traverse lines you will find much worse closure at that location than the other traverse points. This assumes of course that your data does not contain significant blunder.
Years ago C&G came up with a report you could generate from a raw data set that would calculate the error of closure at each traverse station. Since C&G has been incorporated into Carlson, you might still be able to get that report.
Also, if you have Carlson you should be able to use SurvNet to analyze your data to see if you have a real issue or if this is one of those fake issues that looks like a problem when it really isn't.
Larry P
crap in = crap out
suck it up and go re run that sombitch with good procedure.
if you don't know what that is, ask!
Ditto all the others but will add that slop in the tripod can cause angular errors and it's often very subtle.
Check all connections in the tripod head and where the leg rods connect to fixture that attaches to the tripod head. Make sure all connections are tight....
Your closure starting occ 1/BS 2 and coming back to close on 1 seems to indicate that your loop is good within itself. Traversing through 2 just before closing on 1 and missing 2 by 0.65 indicates that you had bad starting coordinates on either 1 or 2.
How did your initial BS check compare to the inverse and how did it compare to the original field measurement from when 1 & 2 were set? Did you have any other control to check into?
Sounds to me like you need to verify your starting coordinates and if original data input and calcs check out, do some additional field checkes to determine where your previous error lies.
As for the angle error guesses, that would make sense if 2 had not been incorporated as part of the loop - that is if you had looped back to 1 without traversing through 2 and then from 1, backsighting your last trav pt in the loop, turned to 2 as an angle check.
But since you used 2 as both your initial BS and as your last station in the traverse loop, if you had a previous error in your current loop accumulating to 0.65', you would also have a very significant and propbably easily detectable blunder in your closing angle at 2 to 1 and/or in your closing distance to 1 to lose half a foot of that positional misclosure in that one leg. That seems unlikely to me.
>Error of Closure is a function of your raw data set and where you begin your calculations.
Closure of a simple non-redundant and unadjusted traverse is a vector sum, or a sum of latitudes and departures. The order of adding does not matter.
The only reason it would change is if you rotate the entire figure to assume a different basis of bearing.
How can you feel comfortable about setting control points by just turning one angle, especially on an airport project?
Traverse closure?> Larry, Whatchusay?
The error of closure can depend on where you start your calculations? That just doesn't make sense to me...
Dtp
Traverse closure?> Larry, Whatchusay?
> The error of closure can depend on where you start your calculations? That just doesn't make sense to me...
Larry's correct. Imagine a perfect traverse with perfect closures from perfect distances and perfect angles. Now, introduce a 30 second error in one of the angle observations. If you begin your closure calculations at the station where the 30 second bust occurs, you will project that 30 second error throughout the entire length of your traverse and will get a terrible closure. If you calculate the same traverse starting at the first station beyond the bust, the 30 second error will only affect the lat/dep of the last course, resulting in what could seem like a good closure and will result in perfect coordinates for all of the points despite the 30 second error.
JBS
> >Error of Closure is a function of your raw data set and where you begin your calculations.
>
> Closure of a simple non-redundant and unadjusted traverse is a vector sum, or a sum of latitudes and departures. The order of adding does not matter.
>
> The only reason it would change is if you rotate the entire figure to assume a different basis of bearing.
Not quite, Bill. The angular error is applied when determining the vector and the subsequent lat/dep. If the angles are in error, the lat/dep will differ for each successive course after the error occurs. The closure will be different depending on where you start your calculations.
JBS
> How can you feel comfortable about setting control points by just turning one angle, especially on an airport project?
[sarcasm]2 sets just takes too long! [/sarcasm]
JB and Larry are correct
:good:
Traverse closure?> Larry, Whatchusay?
:good:
Now if you ran that perfect traverse except for one 30-second bust....and you took that 30 seconds and averaged it out throughout the traverse, you will:
1. have the exact same misclosure regardless of where you start your calculations.
2. have introduced error in every angle/leg; whereas if you had isolated out the busted angle and corrected only it, you would have had a great traverse.
That is one reason to have a least-squares analysis tell you the most likely place error exists. If it's important, you could go re-turn that one angle and would have confirmed your analysis.
trojan1stdown..r u there?
Jes wunderin' if you went back out to the AP today?
Like to hear what you found.
Thanks for the help and suggestions. I'm going back out there tomorrow to reset the 2d with GPS. The office I work out of isn't totally on board with GPS so I have to borrow a unit from one of the other branches. I should have shot double angles, but I was trying to get that job setup on the way to another one so we could come back and just start shooting. I should have taken more time. I hate when the short way ends up being the long way. I am still going to mess around with my data and see if I can pinpoint the bad angle for future reference. Thanks again, this board is great.
somewhere in the traverse
use the perpendicular bisector to find the culprit
Oh yeah, I forgot to answer some questions. I did shoot the bs for a check when I zeroed and the distance checked to .03'. The coordinates for #1 & #2 are right as best I can tell. Another time we were out there with the GPS and they checked .04' and .05' from our published values. And I always try to line up the point or lowest part of the rod before shooting in a nail. It would have known to rerun te travers if the error had gotten worse, it was just the fact that it tightened up there at the end that was throwing me off.
that's a little harsh
i agree, but it is still harsh
I see it now. You have one redundant measurement if you take N angles and N distances for an N-leg traverse. With the closure point you have 2N+2 coordinate variables, minus 2 for the arbitrary coordinates of a starting point, minus 1 for the basis of bearing. That leaves one extra.
In finding an open-loop closure you ignore one of the measurements - typically the angle at the starting point.
Changing which one you ignore changes the closure. If all measurements except one are perfect, then ignoring the bad one will get perfect closure. Ignoring any one of the other measurements instead will give you a non-zero closure - all the same closure distance if the rest of the measurements are perfect. That's what Foggy and I must have been thinking of.
Larry has provided a good example in a separate [msg=152376]thread[/msg].
Not to belabor the points, but time how long it takes to set up and get a bs and fs set up vs how long it takes to turn at least two angles. You may be in a hurry, but measure twice and cut once is quite applicable as you might be seeing now.
Also, never leave a job site without adding up the angles. It's something easy to do with a calculator, and it tells a lot more than the closure. having the closure "tighten up" with introducing another angle is not very enlightening ... actually it is more deceptive than helpful in my opinion.
Get it straight, buster. I'm not here to say please. I'm here to tell you what to do. And if self preservation is an instinct you possess, you'd better farmin' do it and do it quick. I'm here to help. If my help's not appreciated, lotsa luck gentlemen.
If I’m curt with you it’s because time is a factor. I think fast, I talk fast, and I need you to guys to act fast if you wanna get out of this. So pretty please, with sugar on top, re run the farmin' traverse.
