Compass rule closure results

Ah, it's hard to comment without some more data, a 1:207,000 misclosure is less than .05' in 10,000' so you are adjusting a very tight traverse.

I used to adjust traverses probably once a week with compass rule, usually by hand. I never had a misclousre after adjustment. I don't understand that at all. Sounds like some meaningless statement from a program of some sort.

90% of the time, if you adjust the angles, and it gets worse, then you have a distance error. Now, if you data collect, then an angle issue, at a corner, will mirror a distance error.

Either way, that's a tight traverse so don't worry about it. If you want to know more, then I suggest that you adjust a few, by hand. Ted Madson had an excellent book with step by step instructions so that the person knew what was going on. I highly recommend it.

If you run your traverse starting at point 1 and end at point 1 you will get a closure (perhaps 1/200,000). But if you run it from point 2 and back around to point 2 you will get a different closure (say, perhaps 1/20,000). Each point you close on you will get a different closure prior to running an angular adjustment. (Ie add up all the interior angles it should be equal to (n-2)*180.) After adjusting all of the angles, you will get the same exact same linear error of closure on each point. It will be somewhere between the best (1/200,000) and the worst (1/20,000); say something like 1/40,000.

One way to do a better adjustment is to put more of the angular adjustment into the weakest angle.

After running the compass adjustment, you should have perfect closure by definition. I think that part of your question was slightly misworded.

When you change point 18 to 2 copy the whole line and keep 18 also.
You need several tie in points to to make the least squares thing work for you.

What software are you using? Is it Land Desktop/ Civil 3D or Carlson?
If its Civil 3D, I have never liked the adjustment routines. If I recall correctly, I think that the adjustment routine left the last course out of your measurment for total distance measured.

Say you measured 5000' total with final 250' from 17 to 18. Softdesk would leave out the 250' and make the traverse 4750' total and use the distance error at point 1/17, but the angle balance would be applied from course 17-18. You would be missing one course from the total and it would make the closure seem worse.

I think it had something to do with making 1 and 2 fixed control points and I think it had something to do with the data collector conversion to FBK and how Softdesk read the start of the file.

If its Carlson you can try changing your CL (closure) and AB (angle balance) codes in the Raw File editor pointing to 17 or 18 separately and see how it affects closure. Really, it's just an academic pursuit, but I feel experimentation can lead to a better understanding of the software.

Or are you using SurvNet, StarNet or Terramodel?

> After the adjustment point 18 cords = point 2 cords exactly.
> If 2=18 then how can the closure be 1:46,000?
I think that you are confusing yourself by focusing on the coordinates. You should not be thinking in terms of point numbers, but in terms of angles, and in latitudes and departures. Back in the day when compass rule was de rigueur the user never saw a coordinate until the very end of the process.

With a CR adjustment you typically have 3 sets of horizontal angles. The raw data, the product of the angles after angular closure adjustment, and then the angles that are computed from the coordinates that the CR adjustment produces.

> Also, when doing a Least Squares adjustment I change point 18 in the raw file to be point 2 so I cannot tell the relationship of those to points after adjustment.
Don't look a the coordinates. Look at the residuals in the measurements.

I like to run an inverse report after adjustment that shows the data in an angle right format (similar to how it was measured). That way I can compare the adjusted measurements to the original measurements (i.e. not the coordinates).

To get a feeling for how blunders are detected in a compass rule adjustment, re-enter the data with one bum distance. Compare the angular closure to the linear closure. Do the same for a busted angle.

Compass rule closure results - starting-ending points

If you run your traverse starting at point 1 and end at point 1 you will get a closure (perhaps 1/200,000). But if you run it from point 2 and back around to point 2 you will get a different closure (say, perhaps 1/20,000).
-
Please provide some proof as this defies the mathematics I have learned and used!

After a closer inspection of the results of the Compass rule adjustment I noticed that 18 does not equal 2 after the angle balance ( so that is where the after balanced closure comes from)
But... after adjustment 18 does equal 2. What happens after angle balance, during adjustment?
Below it a print out of the report.

Closure Results (Before Angle Balance)
Starting Point 2: N 888708.991 E 1711353.016 Z 798.640
Closing Reference Point 2: N 888708.991 E 1711353.016 Z 798.640
Ending Point 18: N 888709.002 E 1711353.026 Z 798.516
Azimuth Of Error: 41°10'19"
North Error : 0.01109
East Error : 0.00970
Vertical Error : -0.12404
Hz Dist Error : 0.01473
Sl Dist Error : 0.12491
Traverse Lines : 16
SideShots : 79
Store Points : 1
Horiz Dist Traversed: 4008.689
Slope Dist Traversed: 4025.058

Closure Precision: 1 in 272080

Starting Point 2: N 888708.991 E 1711353.016 Z 798.640
BackSight Azimuth: 292°01'40"

18 AR112.2759 88.4331 192.110 4.810 4.830 888709.002 1711353.026 798.516

Angle Balance
Angular Error: 0°00'32.00" for 16 traverse sides
Adjusting Each Angle: 0°00'02.00"

Closure Results (After Angle Balance)
Starting Coordinates : N 888708.991 E 1711353.016 Z 798.640
Closing Reference Point 2: N 888708.991 E 1711353.016 Z 798.640
Ending Coordinates : N 888708.932 E 1711353.088 Z 798.516
Azimuth Of Error: 129°15'54"
North Error : -0.05854
East Error : 0.07161
Vertical Error : -0.12404
Hz Dist Error : 0.09249
Sl Dist Error : 0.15472
Traverse Lines : 16
SideShots : 79
Total Hz Dist Traversed: 4008.68923
Total Sl Dist Traversed: 4025.05800

Closure Precision: 1 in 43343

18 AR112.2757 88.4331 192.110 4.810 4.830 888708.932 1711353.088 798.516

Compass Closure
Adjusted Point Comparison
Original Adjusted
Point# Northing Easting Northing Easting Dist Bearing
18 888708.932 1711353.088 888708.991 1711353.016 0.092 N 50°44'06" W

Compass rule closure results - starting-ending points

Mr. Schumann, I will provide you a traverse with variable misclosures if you tell me that you will look at it to evaluate my premis. (I don’t want to do the work if you don’t want to look at the results).

Consider first, that to run a closure, I need to run every angle and distance Point 1, to 2, to 3…to “n” and a final angle and distance back to point “1”. The only missing angle not used in the calculation is the angle at point 1. If you later run the closure from point 2 and around, the missing “redundant” angle will be the one turned at point “2”.

If that is not sufficient-enough explanation, let me know and I will provide you a traverse to evaluate.

I am pretty sure you get this, because I know you are more experienced and knowledgeable than I am. But maybe I didn’t word it so well.

Note that the second closure is after angle balance and prior to compass-rule adjustment. That is what I was trying to address above.

A better angle balance would be to adjust a different value at every angle point based on the strength of that angle. The weakest angle getting the most adjustment.

The way I would come up with the weakest angle would be to run a closure starting at every point which would leave out the angle turned at the starting point. When that angle is eliminated from the closure, and you get a higher closure level, the angle at that point would be the weakest.

> Before adjustment the closure happened to be 1:207,000.

It's Miller time.

Once your traverse is closed with a result of 1 in 46,000 you make an angle adjustment.

The result improves the closure to 1 in 207,000.

That is fantastic any day of the week.

At this point I would do a Crandal adjustment because it is so near perfect considering it is 16 legs with a total length of 4,008 feet.

Any closure routine should have a result of perfect match with the beginning point and the next point will be very close to the original coordinate. The difference will be that of the adjusted length of that leg.

Perhaps you are basing your results upon closing upon your second point instead of your beginning point.

The closing upon your second point should still be out by approximately the original amount.

The only difference would be the bearing between 1 and 2 would be the same as from 17 to 18.

BTW, the use of Compass or Crandal is a highly debated topic and we do what we do for our own benefits in my opinion. After all, either deals with very small amounts to obtain a perfect closure.

It is just a personal choice for me.

😉

ML

You've never noticed that if you take the same raw numbers, and use different beginning and ending points of a traverse the raw closure is different? C&G has a routine that looks for errors by evaluating this PROOF.

BTW, the use of Compass or Crandal is a highly debated topic and we do what we do for our own benefits in my opinion. After all, either deals with very small amounts to obtain a perfect closure.

It is just a personal choice for me.

Shhh!! Be very quiet about it, but I've had some very tight traverses, and I didn't even adjust them, I'm so ashamed:whistle:

I agree and do the same unless the size of the job requires it.

With today's technology and quality of instruments, with proper use, there is no real reason to apply a closure routine to comply with minimum state requirements.

B-)

:good: :good: :good:

And remember...if you record and average both the back and fore bearings of your compass...there is no reason to adjust your "angles"...

DDSM:beer:

Compass rule closure results - starting-ending points

Mark Chain is absolutely correct on this. Experiment with your own data, you'll find it to be true.