Activity Feed › Discussion Forums › Strictly Surveying › Open Traverse closing method
-
No method eliminates error, all you can do is try and distribute it as close to equally as possible. I would use the compass rule, but then that’s what I usually did way back in my traverse days.
Be sure to “close” out your angles, by that I mean set up on the point you are starting from twice, at the beginning of the traverse and at the end and turn angles.
-
If you have an open traverse starting on two known points then you can only assume your angles and distances are good and calculate coordinates of each traverse point.
If you have two points at the end with only a known azimuth (say from a sun or polaris observation) then you have a blunder check on your angles and can balance your angles then calculate coordinates of each traverse point. You are still assuming your distances are good. For purposes of this discussion I am assuming your azimuth at the end is properly related to the azimuth between the two known points at the beginning. As you go east or west there is convergence of the meridians but for this discussion I will assume that is negligible.
If you have two known control points at the beginning and two known points at the end then you can balance your angles and do a Compass Adjustment. You are assuming the 4 known coordinates are without error in this case.
The other option is to “go around the block.” Start on two known control points, traverse around in a loop and close back on your first backsight. Say your control points are A and B, Set up on B bs A, fs 1. Setup on 1 bs B, fs 2. Say you go through 3, 4, 5, and 6. At 6 bs 5, fs A then setup on A bs 6, fs B to close your angles. Now you can balance your angles and do your compass adjustment.
-
Bill: just a quick comment. I prefer to use a least squares adjustment program EVEN IF there is no redundancy for two reasons…
1) I know the data is being reduced correctly (grid factors, etc)
2) It gives me an idea of the accuracy of the points. Granted, there is no check, but it uses error propagation to estimate the station confidence regions.
For example, if I have a bunch of RTK lines radiating off of a base that had a HERE position, it is very easy to input a coordinate for the base or include baselines to CORS, and then recompute all of the positions with error stats provided in the output. Each time I change control (i.e. add or delete a benchmark, etc) I know it will instantly recompute everything.
-
Posted by: @john-hamilton
prefer to use a least squares adjustment program EVEN IF there is no redundancy for two reasons…
Agreed, that is a useful practice. I use my LS program for non-redundant solutions and to do COGO.
But this sidetrack will only confuse someone who is figuring out what is needed to do any adjustment at all.
. -
Posted by: @jerrys
At the risk of exposing my own ignorance, the “angular method” adjustments I am familiar with only were designed to distribute the gross angular misclosure among the traverse stations, the assumption being that the angular misclosure was the result of systematic error in the equipment that could not be eliminated other than by adjustment once the gross angular misclosure is quantified. This error can be minimized by taken multiple reading in both face 1 and face 2 orientations of the total station.
The Bowditch or Compass Rule adjustment divides the linear misclosure proportionally among all the traverse stations, so that each leg of the traverse gets the proportional amount of its length as a percentage of the overall traverse length.
Either adjustment method presumes that there are no blunders, and that you are distributing the misclosure through the traverse. Some would first do an angular adjustment to balance the angular misclosure though all the traverse stations, as quantified by a reoccupation of the initial survey station, with a backsight taken on the last station and the angle turned to the second traverse station. If the raw error of closure before any adjustment is small, my experience is that an angular adjustment before doing the Compass Rule adjustment will result in a slightly poorer linear misclosure than prior to adjustment.
The experts on this site would encourage you to take multiple measurements to each forward station and to use least squares to do your traverse analysis and adjustment.
Right, and I never did understand why you would apply the most error (even systematic) to your longest lines when those lines are the most accurate based on the precision specs of the equipment.
-
Posted by: @jerrys
my experience is that an angular adjustment before doing the Compass Rule adjustment will result in a slightly poorer linear misclosure than prior to adjustment.
The sum of the angles is known, fixed by the laws of geometry. This is information that it seems reasonable to me should be taken into account. Even after doing face-1 and face-2 measurements (a really good idea), there is usually a residual angle error. Textbooks invariably assume the angles were adjusted prior to dealing with the lengths by compass rule.
The trick is to distribute the error in the most probable way. That might mean assigning more of the error to the angles with the shortest legs, or at the station where you had trouble keeping the instrument level on soft ground, etc. Assigning equal error to all angles is a fall-back when you lack any other information about their relative quality.
. -
Posted by: @bstrand
Right, and I never did understand why you would apply the most error (even systematic) to your longest lines when those lines are the most accurate based on the precision specs of the equipment.
How do you figure that the longest lines are the most accurate? Maybe by parts per million, but definitely not in terms of feet or cm of error.
. -
Posted by: @bstrand
Right, and I never did understand why you would apply the most error (even systematic) to your longest lines when those lines are the most accurate based on the precision specs of the equipment.
Sorry, sir, that is just wrong. Exactly the opposite is true.
It is true that a single long measurement between two points may indeed yield more accurate results than the sum of a series of shorter lines. That is because you are eliminating instances of centering errors inherent in multiple setups.
-
I’m sorry but the poster seems to be be asking if an optical traverse between two known control points should be adjusted for azimuth/distance. I’ll state a Compass adjustment will suffice (least squares with error estimates will outperform, but not by much), but you’ll be hanging your hat on the accuracy of the control points. Personally, I’ve always conducted closed loop traverses in the 1-2 mile range so I know my internal accuracy outperforms the (supposed) accuracy of the control points, which are subject to egregious provenance problems quite often.
5-12 miles between alleged NGS/State/County/etc. control monuments, yep, I’ll hold them as true and adjust to them without a closed loop and it’ll tighten my survey. Of course, GPS surveying makes optical surveying at such distances moot; you’ll be more accurate and faster using GPS.
The thing that bothers me is the GPS button pusher who translocates using OPUS, doesn’t search for monuments or occupation evidence and declares the whole local cadastre/subdivision off by 15 feet and slams stakes in somebody’s garden. The result is a lot of aggravation among landowners who’ve been in repose for decades.
Open, closed, distant surveying no matter how accurate should honor local conditions to protect private interests.
-
Posted by: @mike-marks
I’m sorry but the poster seems to be be asking if an optical traverse between two known control points should be adjusted for azimuth/distance.
No, he posted this at the bottom of page 1 of this thread:
Sir it's a open Traverse which starts on the known point and end on unknown point
after 4 kilometer in length from starting known point. It's parallel to the road.Note that he is in Pakistan, hasn’t mentioned GNSS, and may not have many official control points to choose from.
. -
Yes, ppm, which is why you want long backsights. And compass rule doesn’t discriminate between random error and systematic error so you end up smearing unwarranted error into your most accurate lines. Anyway, it was done this way long before I came along so I’m sure the old timers do what they do for a reason.
-
Posted by: @bstrand
Yes, ppm, which is why you want long backsights.
While you do want long backsights for your angles in order to minimize the angle error due to centering, I don’t see how ppm, which is usually used to talk about distances, would affect that.
And compass rule doesn’t discriminate between random error and systematic error so you end up smearing unwarranted error into your most accurate lines. Anyway, it was done this way long before I came along so I’m sure the old timers do what they do for a reason.
You usually don’t know which are the most accurate, but if you do you can take that into account when distributing the latitude/departure errors instead of holding strict proportion by compass rule. The books have always recommended some form of adjustment, so I figure that experience has usually supported it.
. -
Assuming that you have zero errors in your traverse, no centering, no angular or distance measuring error, in other words, the perfect traverse, you still have the problem of not knowing how precise the starting known points were to begin with. For every 1″ of error in your beginning backsight azimuth you would be left with 0.7 meters of error over the total 4 kilometers distance traversed.
Willy -
@john-hamilt
My bad, didn’t convert to dd. I stand by that OP likely started out w/ significant error. W
Willy -
Agreed. But I have worked in third world countries, sometimes they are lacking any modern equipment and therefore have no way to establish a GPS coordinate at the other end. The best thing he could have done without any GPS or existing control would be to observe an astro azimuth at either end. That can be done with minimal equipment, If using polaris time is not critical (if done at the right time), if using the sun, etc it would be necessary to have fairly accurate time. There are apps for smartphones that will try to sync with an ntp server, not sure how well they work overseas, especially if the time server is located far away
-
That is what I have been familiar with and what was the general operating procedures decades ago when running cross country transmission lines plus taking solar shots from every setup and computing everything to account for the curvature of the earth’s surface to obtain better actual global positions.
In construction, the actual measured distances would be used for design.
For TxDot work we would take 6 repetitions of angular measurement at each setup.
We did one TxDot project where for each setup we had to fill out a two sheet form for angular measurements that required over 20 observed readings on the instrument to obtain a balanced mean angle that was recorded to the hundredth of a second on paper for the final report. It did not matter if we used a 1second or a 5second instrument, the results always came out the same.
0.02
-
Posted by: @dave-karoly
If you have two known control points at the beginning and two known points at the end then you can balance your angles and do a Compass Adjustment.
I have recently discussed this with management. We are starting a traverse on a pair of our GPSed points and ending on a pair of our GPSed points. Management says we cannot close on the ending points because they aren’t published points. We have to run the traverse back to the beginning. If you argue that by beginning on a pair of GPSed points establishes a beginning azimuth, then why doesn’t it apply to closing the traverse on a pair of GPSed points?
MH
Log in to reply.