Just a general comment
> Why would heights matter for horizontal distances?
>
Then why bother even putting in the instrument height and reflector height? Is that only necessary for measuring elevation distances?
Back to the drawing board
I must have more errors somewhere in my recording/reducing/meaning. I felt really confident that all the measurements in this traverse loop were really good, but Starnet is telling me it sucks.
Back to the drawing board! Where's my dang DC when I need it?!
[pre]
C 14 5000 5000 ! !
#B 1-2 n56-23-13w !
B 1-14 S31-02-39w
D 1-2 230.8
M 2-1-14 267-25-52 1068.22
M 1-14-23 220-53-52.5 167.745
M 14-23-25 275-36-11.5 259.122
A 25-23-14 84-23-10
M 23-25-34 182-24-16 450.015
D 25-23 259.105
M 25-34-35 229-30-05 465.700
D 34-25 450.020
M 34-35-2 223-16-02 352.950
M 35-2-1 221-08-00 230.8
[/pre]
Just a general comment
> Then why bother even putting in the instrument height and reflector height? Is that only necessary for measuring elevation distances?
Those are needed in 3-D traverse, where the elevations of each control point and point positioned from them are being computed. In 2-D work, the typical solution is just to choose an average elevation for the project and enter that in the Adjustment Options. 3-D surveys allow essentially exact scale factors to be computed at every point, whereas the average elevation approach introduces some error into the scale factor by which measured distances are reduced to grid distances for the computation of grid coordinates.
Whether the 2-D, average elevation approach is good enough or not really depends primarily upon (a) how ground elevations vary across the project (from max to min) and (b) the the target accuracies of the points positioned by the survey.
Just a general comment
As always, what factors need to be considered depends on your situation.
If you have 100 feet of elevation change over your project, expect about 5 ppm change in horizontal distances when reduced to the same elevation.
Not significant for a project of a few acres with rolling terrain. A big deal if you are working a long project up the side of a mountain.
Back to the drawing board
YOU HAVE SOME BAD ANGLES AND YOU HAVE SOME BAD DISTANCES. LOOK THROUGH YOUR LISTING FILE AND SEE WHICH ANGLES AND DISTANCES ARE FLAGGED AS BEING OUTLIERS. REMOVE THE OUTLIERS AND SEE HOW IT PROCESSES. I SUSPECT THAT YOU STILL HAVE ERRORS IN YOUR DATA ENTRY.
sorry cap locks stuck on....
Just a general comment
> As always, what factors need to be considered depends on your situation.
>
> If you have 100 feet of elevation change over your project, expect about 5 ppm change in horizontal distances when reduced to the same elevation.
>
> Not significant for a project of a few acres with rolling terrain. A big deal if you are working a long project up the side of a mountain.
It should be understood that the best approach is to do a 3-D adjustment incorporating survey data from which the ellipsoid heights of network points can be computed. The 2-D adjustment is a viable kluge for many projects in the real world, but still a kluge that isn't generally useful for ALL projects.
Back to the drawing board
> YOU HAVE SOME BAD ANGLES AND YOU HAVE SOME BAD DISTANCES. LOOK THROUGH YOUR LISTING FILE AND SEE WHICH ANGLES AND DISTANCES ARE FLAGGED AS BEING OUTLIERS. REMOVE THE OUTLIERS AND SEE HOW IT PROCESSES. I SUSPECT THAT YOU STILL HAVE ERRORS IN YOUR DATA ENTRY.
>
1. Do I understand that the high residuals for angles 2-1-14 and 35-2-1 are the place to start?
2. What if I don't have any redundant measurements (other than my doubling angles, Face Left and Face Right)?
If I eliminate them, I'll get "improperly defined coordinates" or whatever.
Back to the drawing board
Why? Can you find any reason why the residuals are soooooooo high?
Perhaps you can get a solution by using some of you angles and not all of the doubled angle data.
Remove all blunders FIRST
check to make sure that you have remove all data input errors.
remove some angles and distances..
if there are no blunders
if you have fixed all input errors
and if there is no extra measurement data that can be removed
then your only option is to re-observe
I wonder if your equipment and/or field methods are suspect?
Back to the drawing board
you may want to look at what and how you are using prism constants-
what prompted you to use 0.13 as a distance constant?
with a semi modern inst. You should be using thing around 3mm +3 ppm
What are you using for BS and FS targets?
For your angle you're probably closer to 20" unless you're turning several sets esp. given the short legs....
Back to the drawing board
> you may want to look at what and how you are using prism constants-
> what prompted you to use 0.13 as a distance constant?
> with a semi modern inst. You should be using thing around 3mm +3 ppm
> What are you using for BS and FS targets?
>
> For your angle you're probably closer to 20" unless you're turning several sets esp. given the short legs....
The .13 distance constant is a mistake. Thanks for pointing it out. Regarding "blunders", as Starnet defines them, I've eliminated them all....doesn't mean I don't still have issues with some of my measurements. I've already identified one set of measurements around point 1 that are suspect; I'm doing them again this weekend. The comment Kent made some time ago (in this thread, lol) about not taking at absolute value some of the segments I've to this point assumed to be correct, may in fact be an issue.
I've got a plan to independently check them with additional measurements.
As for all the BS's and FS's, I'm using Topcon Prisms on tripods...all identical, and all carefully checked multiple times. I don't think that's the problem.
I'm more concerned about the angles right now than the distances (but I understand they're all certainly related.)
Thanks all for the comments.
Back to the drawing board
Yes, there is an angle problem. A simple sum of angles fails to close (match a multiple of 180 degrees) by 14 to 15 minutes. The exact amount depends on which of the values, or their mean, that you use where you have a redundant one (and the pair disagree by over 30").
The formula is 180*(N-2) for interior angles, or 180*(N+2) for your exterior angles.
While you are making more measurements, don't pass up the chance to get any easy distance values also.
Back to the drawing board
>
> The formula is 180*(N-2) for interior angles, or 180*(N+2) for your exterior angles.
>
Uh...I plead ignorance. I've not seen that in Ghitani and Wolf. That's the formula to find the mean of two angles?
For example:
14-23-25 consists of two "doubled" readings (two each FL and FR):
275-36-10
191-12-30
and
275-36-05
191-12-15
I reduced these to:
275-36-15
and
275-36-08
And then took the mean of them to make:
275-36-11.5
Subtract that from 360 (to get the interior angle)
And I get:
84-23-48.5
But as soon as I look at that, and compare to my other angle measurement of the same angle (25-23-14), I see it's way off.
That one is 84-23-10!
ARGH!
So much for picking one of my measurements at random! Let's start with the formula and I'll continue to clean up my math before I head back out.
Back to the drawing board
The sum of the interior angles of a polygon (closed traverse) is exactly 180*(N-2) by geometry. For exterior angles it is 180*(N+2).
See Wolf and Ghilani 11th ed. Section 9.7 Angle Misclosure.
If you add up your measured angles for a closed traverse, the difference (angle misclosure) between the geometrically required value and your sum should be not a lot more than your standard error for angles times sqrt(N angles). A lot more would indicate a blunder.
Back to the drawing board
Statistical Summary
Observation Count Error Factor
Angles 8 1.825
Distances 10 0.148
Az/Bearings 2 1.407
Total 20 1.241
Chi-Square Test at 5.00% Level Passed
Lower/Upper Bounds (0.522/1.480)
This was accomplished by using an error 0f 600 seconds. Previous iterations indicated that you have an error of 10 minutes in your angles. try plugging in 600 seconds. The distances seem to be okay. Now I suggest you put some thought into why the error for your angles is sooooooo large, Bad pointing, heat waves, not level, not over the point any other possible cause?
Back to the drawing board
> Now I suggest you put some thought into why the error for your angles is sooooooo large, Bad pointing, heat waves, not level, not over the point any other possible cause?
Other than one measurement that I know is suspect, I'm putting my money on bad math. It's not the measurements.
I'm going to start from scratch, doing the arithmetic again. I know this process is just teaching me a lesson, until I get a Data Collector.
Well, THAT would do it!!!!
> I'm going to start from scratch, doing the arithmetic again. I know this process is just teaching me a lesson, until I get a Data Collector.
Well, I put all the initial measurements in the small traverse segment(1-14-23-25-34-35-2-1) into my metes and bounds program and noticed that it misclosed by almost 10'! I looked at the azimuth of the closure and found that it was nearly in line with points 1-2.
I looked at what I had entered the 1-2 distance in Starnet: 230.8'.
Well, it's not. Never has been. According to an old survey it's 240'. I measured it as 239.365, and have measured it a number of times in the past very close to that.
The error factors dropped like a rock. Sheesh! I've got additional redundant measurements in process, but thought I'd pass along what I now think is the biggest thing causing my initial headache.+o(
Well, THAT would do it!!!!
>
> Well, it's not. Never has been. According to an old survey it's 240'. I measured it as 239.365, and have measured it a number of times in the past very close to that.
Why was it still, 230.8?? Just because it was an 'old" survey, does not necessarily make it a 'bad' survey, but look at what you have learned. LOL
Well, THAT would do it!!!!
> >
> Why was it still, 230.8?? Just because it was an 'old" survey, does not necessarily make it a 'bad' survey, but look at what you have learned. LOL
Who Knows. I wrote it down wrong. Regarding the survey, done in early 1982, many of the distances on it have "+/-" shown, and all the angles are only to the nearest whole minute. An old retired surveyor friend of mine looked at it and said "they did most of this in the office".
It might have been done with a precursor of Autocad (which if my memory serves me correctly, wasn't released until December 1982), but my guess is not.
At least they were more careful about writing down the numbers than I've been. They didn't have no stinkin' data collectors, either!