How many distance measurements do you take?

all TS software multiple angle routines that I am familiar with required the distance and angle for input.
I guess somewhere there maybe a setting not apply distances but why would one check that selection.

Those ratios (i.e. 1:50,000) are somewhat unhelpful. Particularly on those short leg traverses like you mention (250' legs).

Consider 2mm+2ppm applied to 76.2m (250') is 2mm+0.15mm or 2.15mm. The ratio of error would be 2.15mm/76.2m or 1:35k which is a bust already, and we haven't even considered setup errors.

The problem is the constant part of the error (2mm). Until the lengths become long enough to render those constants to be insignificant, they bite you in the butt when looking at a straight error ratio (1:x).

> Those ratios (i.e. 1:50,000) are somewhat unhelpful. Particularly on those short leg traverses like you mention (250' legs).
>
> Consider 2mm+2ppm applied to 76.2m (250') is 2mm+0.15mm or 2.15mm. The ratio of error would be 2.15mm/76.2m or 1:35k which is a bust already, and we haven't even considered setup errors.
>
> The problem is the constant part of the error (2mm). Until the lengths become long enough to render those constants to be insignificant, they bite you in the butt when looking at a straight error ratio (1:x).

Thanks, Shawn. Perhaps I'm asking the question the wrong way. Or perhaps the assumption that distances measured being much better than angles, is erroneous. Here's the specs on my instrument:

What's the right question to ask if one is striving for providing Starnet a "balanced" number of measurements (both angles and distances), to provide the best results. Feel free to ask the question I'm trying to ask a different way.:-/

reading multiple sets of distances does two things, one is to develop a good mean, and the other is to help prevent busts. Always read at least one more distance, and try to look at it with fresh eyes forgetting the first read. Also, when I worked as a instrument man, we always took one extra shot in meters as a method of double=checking for a bust.

But I know that's not your question. Granted that the readings don't vary much from reading to reading, I would still take a number of readings. It just doesn't take much more time to read four readings than one. your main time is travelling to the point and setting up. Second most time is multiple sets of angles. Having the edm read a few more distances is just not worth worrying about. Definitely shoot both ways. (and bring an extra battery if that's your worry). Redundancy while you're already set up is much better and more time=saving than coming back to the site to retake a measurement.

I know you want a mathematical/statistical answer. I'll leave that to others to address.

Here's an algebraic way to consider your question:

At what distance is a 5" error equal to 2mm+2ppm?

2mm+2x10^-6x(D)=sin(5")xD

Solving for D, I get 90.9m. So a five second error at 90.9m is 0.0022m or 2.2mm and the distance error at 2mm+2ppm at 90.9m is 0.0022m or 2.2mm. Less than 90.9m, 5 arc seconds will be less than 2mm+2ppm, beyond 90.9m 5 arc seconds will be greater than 2mm+2ppm. This DOES NOT include pointing errors and setup errors. Just a straight comparison of the instrument's distance precision compared to its angular precision.

A Measurement Is A Set Of Observtions

For traverse work my data collector is set to take a distance every time it takes an angle. However it is also necessary to take measurements from the ahead station to the traverse backsight point. Having done that, multiple observations in each direction, you have accomplished a measurement.

Paul in PA

A Measurement Is A Set Of Observtions

OK. Question answered. More is better than less. I'll proceed to set the DC to observe SD with every pointing...4 per Measurement x 3 (or 4...haven't decided yet) measurements per station pair (BS and FS).

I think my TS also has a built in setting that makes as many observations as you want EACH time you ask it to observe the distance and averages them, so if I'm taking a total of 12 or 16 observations of distance routinely, I can probably shut that off (turn it down to ONE, each go).

Thanks again for the input.

I shoot important stuff at least twice.

The question really is: "At what point do repeated EDM range measurements not contribute anything significant to the resulting estimate of the distances between the ground marks between which the measurement was attempted to be made."

Suppose that you have a total station with an EDM claimed to measure ranges with a standard error of +/-2mm (= 0.0066 ft.), and that the uncertainties of centering of instrument and prism are:

+/- 0.002 ft. (instrument)

and

+/- 0.005 ft. (target)

Note: The above are examples only, Your Mileage May Vary.

Note also: The above uncertainties of instrument and target centering are the standard errors of the components of centering that are parallel with the line between the stations occupied by instrument and target.

So that means that the contribution of instrument and target centering to the net uncertainty of the range measurement between the stations will be:

SQRT[(0.002)^2 + (0.005)^2] = +/- 0.0054 ft. (s.e.)

In other words, if the basic range measurement from instrument to reflecting prism were essentially perfect, the distance between the ground marks over which the instrument and prism were set up would still have an uncertainty of +/-0.0054 ft. (s.e.)

Naturally, the basic range measurement won't be perfect. If one range measurement has an uncertainty of +/-0.0066 ft. (+ +/-2mm), then the resulting uncertainty in the distance between the stations will be:

SQRT[(0.0054)^2 + (0.0066)^2] = +/-0.0085 ft.

How much the uncertainty of the distance between the stations actually decreases by taking the average of repeated measurements depends mostly upon the nature of the errors in the EDM ranges in the first place.

That can really only be determined by test.

All traverse points and corners get angles wrapped and three distances measured and averaged on the forward and reverse. I'm not walking back to the backsight to get another distance, except with my robot because it forces me to.