Double run leveling accuracy

Isn't the spec mm per square root kilometers? So if it's 2316 ft, your spec standard deviation would be sqrt(0.706) = 0.84 mm = 0.0028 ft and your measurement is 2.7 standard deviations off. On a single test, that isn't enough evidence to show it is out of spec.

And depending on what you used for turning points some of that error could be settling of the TP while you move the instrument, and did you balance the sight length, and calculate curvature and refraction, and how well did you plumb the rod, and is the orthometric correction significant for this run (if you used a different return path), and ...

Bill93, post: 350872, member: 87 wrote: Isn't the spec mm per square root kilometers? So if it's 2316 ft, your spec standard deviation would be sqrt(0.706) = 0.84 mm = 0.0028 ft and your measurement is 2.7 standard deviations off. On a single test, that isn't enough evidence to show it is out of spec.

And depending on what you used for turning points some of that error could be settling, and did you balance the sight length, and calculate curvature and refraction, and how well did you plumb the rod, and is the orthometric correction significant for this run (if you used a different return path), and ...

My understanding is similar to Bill93's. I am not sure what is meant by 'double-run' precision and exactly what the spec you are reading says. Typically also, with NGS specs, you would run a straight line between two "known" NGS Marks and compare your difference to the raw run of those two marks, while maintaining the appropriate sight distances and balanced turns. You can have compensating error that especially might not be seen by running out and back, for instance always having unbalanced legs with the "downhill side" always longer than the distance to the uphill turn.

Running between known marks helps check possible movement of one of the existing marks from when it was first installed. Running back is often to just have a check on your work if you are missing. Checking between more than two know marks is required for higher-order level runs.

(sorry if my post is a bit disjointed. I kind of typed as I was thinking about it)

Looking at the "Tolerances for geodetic leveling" (it always helps to look at the book), see NOAA Manual NOS NGS 3, Geodetic Leveling, 1981 page
3-7.
"One leveling of a loop of one-way length K, beginning and ending at the same point: T X (sqrt. K) mm, T= +or- 4 mm for 1st order class 1 survey."

4 X sqrt. 0.7059 = 3.36 mm (this is + or -) This is what you are allowed for 1st order work in your case. Provided you have followed all other specifications
for 1st order leveling, which you have not (but this is good work).

Please see "Standards and Specifications for Geodetic Control Networks" page 3-7 (just happens to be the same page number in both publications)
where it will tell you ( under Instrumentation) that for 1st order class 1 you need the leveling rod to be "IDS" this means Invar double scale rod.

I can't find my specifications for bar code leveling but I am sure they want an Invar Rod (?).

Hope this helps.

JOHN NOLTON

I am looking for my bar code stuff and when I find it I will post it.

0.007' off and your worried about the spec? Are you just curious about it or are you worried the gun is out of spec? Its just me thinking out loud but it seems like your are splitting hairs. There are lots of ways to lose some accuracy in running a level circuit and most likely the gun isn't it. For one, is that spec using an invar rod?

The "spec" isn't about the precision; it's about the procedure.

Did you follow the guidelines and document your procedure?

OR THIS MIGHT ANSWER YOUR QUESTION

Big Al, could not find the info. on bar code leveling BUT went to the internet and NGS web page and printed it off again ( and then found my stuff).

Also printed from the internet Topcon DL 102C Electronic Digital Level information. The specifications for your level state "accuracy (Standard deviation
for 1km double-run leveling) is 1.0mm w/Fiberglass staff". (here the 1mm is + or -)

Did you use a Umbrella?
What were your turning points like?

Its good that you check your equipment ( so many other surveyors don't). I think you equipment is working good but look at all the specifications.

JOHN NOLTON

PS your compensator setting accuracy for the DL-102C is + or - 0.5 seconds and for the DL-101C is + or - 0.3 seconds.
The specifications state for 1st order work class 1 and 2 the minimum repeatability of line of sight is 0.40 seconds.For
2 nd order its 0.50 seconds. BUT for order 1st, class 1 and 2 the rod must be Invar double scale. Second order class 1 rod must be Invar, double scale or Invar , single scale. Second order class 2 Rod must be Invar single scale. For Third order work the Rod can be Wood or Metal (I guess that also means
Fiberglass).

JOHN NOLTON, post: 350890, member: 225 wrote: but I am sure they want an Invar Rod

Affirmative.

I'm not worried about the results at all. I just don't do much levelling work, and I'm trying to learn about the process, i.e. curious.

John,

Thanks for all the help. I did not use an umbrella. We generally ran a balanced loop, keeping backsight distances equal to foresight distances. We used a SECO Leveling Rod turning plate (turtle), generally placed on pavement. Instrument was set up in earth to side of the road.

I used a fiberglass rod, not Invar.

I will review resources you have posted. Thanks again.

Al

For almost all the work I've ever done, the accuracy you achieved is good enough unless its some kind of special projects. Using a turtle usually gives a good turn.

In my opinion digital levels give much better accuracy than one could get with a optical level unless you took great care with the optical level.

Did you use a rod bubble or some waving read function

I provide the following links not seen in previous responses. The first link provides details about issues like sight lengths, setup and section distance imbalances, minimum ground clearance, etc.

I also did not see mention of the importance of performing a collimation check. This is to be performed daily and entered into the instrument for application to readings. I favor the Kukkamaki method; the Forstner method is also good.

While I imagine that the fibreglass rods are fine for construction and similar accuracy work, calibrated Invar rods should be used for precise work.

While the specifications indicate maximum sight lengths, a prudent observer does not stretch sight lengths if conditions are unfavorable.

I must also say that while you indicate this work was for a test, the determination of a new height by a closed loop from a single BM is unacceptable. Check leveling to at least two BMs at each end of a project is required for work submitted to the NGS for second-order; three at an end for first. The checks must be to monuments that are not clustered together like reference marks nor those on the same structure.

References:

FGCS specifications for digital bar-code leveling

http://www.ngs.noaa.gov/FGCS/tech_pub/Fgcsvert.v41.specs.pdf

http://www.ngs.noaa.gov/FGCS/tech_pub/1984-stds-specs-geodetic-control-networks.htm#3.5

Information about the Kukkamaki method of determining collimation error which is to be applied to the instrument at the start of each day's leveling.

ftp://ftp.ngs.noaa.gov/pub/corbin/precise-leveling-workshop/Presentation/3%20Collimation%20Check%20and%20Field%20Notes%20v2.ppt

HTH,

DMM

The condition and setting of the foot of the rod is where most leveling errors begin.
:gammon:

A Harris, post: 350956, member: 81 wrote: The condition and setting of the foot of the rod is where most leveling errors begin.
:gammon:

Could you explain further? Certainly if the foot doesn't hold the rod the same height above each mark every time, then there's a problem. But if you use a single rod and set it on every mark and TP, then it doesn't matter if the foot is worn or at some arbitrary offset. You can saw off the bottom (unknown) x inches of the rod and still get accurate results.

If you are worrying about a fraction of a millimeter, I'd be very concerned about whether the foot rests on the mark in the same manner as the rod used to originally measure that mark. The Coast & Geodetic Survey/NGS used flat-bottom rods, I think. So if you use a pointed rod you have to consider the depth of the dimple on your mark.

Bill93, post: 350963, member: 87 wrote: Could you explain further? Certainly if the foot doesn't hold the rod the same height above each mark every time, then there's a problem. But if you use a single rod and set it on every mark and TP, then it doesn't matter if the foot is worn or at some arbitrary offset. You can saw off the bottom (unknown) x inches of the rod and still get accurate results.

When a rod becomes worn and/or will not give the same height from one side to another
When a poor turtle set is used

A run of 4632ft with 0.0074ft error is a good day in my world.

Big Al, post: 350857, member: 837 wrote:
I'm trying to evaluate whether this closure meets the stated accuracy for the instrument, which is 1.0 mm standard deviation for 1km double-run levelling. How much "double-run levelling" have I performed? 2,316 feet or 4,632 feet? I'm guessing the former, but not sure.

Hello Big Al,

Depending on how you did it, and if I understand correctly what it is, you may or may not have done any double run levelling. I think double run levelling may be the BS FS FS BS -> [move on] variety, or the BS FS -> [change hi] -> FS BS -> [move on] variety of level run. If you did it like that then you may have done approximately the walking distance to the mark and back again worth of double run levelling.

Don't quote me on this though; my understanding of what double run levelling is could be completely wrong.

Double run leveling is forward-back runnings of a section. Good practice is to run them at different times of day, for example forward in the AM and back in the PM.

A run is out and back. A double run is out and back twice, or out and back with two instruments reading one rod, or with one instrument reading two rods on different points, or 1 or 2 rods on same point twice, i.e feet and meters. The last three require two level books.

Does a three wire level run meet those specs?

Paul in PA.