Starnet Help

>am I able to key in a known standard error (i.e the accuracy of the GPS surveyed point) yet keep the coordinates fixed

I don't think so. I've done some least squares, but only a little in Star*Net, so maybe somebody will have better info.

Your statement sounds like a contradiction, and I don't see how the other program could use standard errors without changing the coordinates a little. Fixed means you are not allowing any possible error to be adjusted, so there is nothing to do with a standard error number. Even if a program accepted standard errors, it may have just ignored them for fixed points.

Entering coordinates with the "C" code allows one to set standard errors for the coordinates

C 1 1000.00 1000.00 is free point

C 1 1000.00 1000.00 ! ! is a fixed point

C 1 1000.00 1000.00 0.02 0.05 is a point with standard error 0.02 for the Northing, and 0.05 for the Easting

If you truly want to fix the coordinates of a single point but still allow the adjustment to consider its standard errors, the only way I know of is to run the adjustment with the standard errors and then translate the network back to the fixed coordinates. The way to do this in Star*Net is via the Project Options | Other Files menu. Check the Create Ground Scale Coordinates (GND) File, then click the Settings button:

In the Ground File Settings dialog box, check the Apply Translation box and enter the point name and desired coordinates. Leave Divide by Factor and Apply Rotation boxes unchecked.

When you run the adjustment, Star*Net will create a file with the translated coordinates that you can then import into whatever application you're going to use them with.

The best way is to minimally constrain the entire network. This will give you information on tge most points.

I typically pick a single point to hold, and a bearing to another point.

The second points, and the ones I am considering to hold, are entered with the * * * after the coordinates, so they are unconstrained. This also gives a report from these provisional coordinates.

Then, you can pick points to constrain the network.

Hi guys,

Thanks for all your quick replies, will try them out.

@Bill93 - I would think it would not be in contradiction as we are living in the real world as there will be inaccuracies in our Control Points. This should be factored in our LSA to give us a truer sense right?

@Peter Lothian - I had tried that. The standard error would cause the control points to adjust by that quantum. I do not want that to happen. I would still like it to be fixed point. Tried to combine both but it will show an error.

@Jim Frame - Thanks for the suggestion, will try it and update if it is successful.

Cheers,

Jan

I like Jim Frame's answer. It agrees with me as far as I went, but he then went on to detail what could be done AFTER the LS adjustment.

The adjustment itself requires that you either allow the coordinates of point(s) to move, or allow the point(s) to be fixed which ignores any standard error they have.

@Jim - I tried what you suggested and tried the conventional way. Realised the difference I got was sub mm level difference. Good learning point for me.

New problem I have encountered. When I tried to export to dxf, it doesn't show the coordinates. Is there a way the coordinates can be listed with the point number as I am required to create a traverse sketch?

Hi Guys,

Have a new questions for Starnet and would like to consult the sages here. I have a project that is in project grid and would like to transform and rotate that to local coordinate system. Is Starnet able to do a helmert transformation for it? Read the manual and could only find a simple transformation parameters.

Ideally, I would like to do a rotation and transformation and have the residuals shown.

P.S: I have a 6 stations in both coordinate systems.

dmyhill, post: 313797, member: 1137 wrote:
I typically pick a single point to hold, and a bearing to another point.

What is the best (simplest) way to hold a bearing to another point. Does the bearing have to be calculated manually first?

[USER=7109]@squowse[/USER]

To hold a fixed bearing from Point 100 to Point 101 your inline code would look like:

B 100-101 N15-30-30E !

Where...
B = code
100-101 = From-To
N15-30-30E = Bearing you choose to hold
! = fixed

"What is the best (simplest) way to hold a bearing to another point. Does the bearing have to be calculated manually first?"

1. If you know a bearing of a line you want to hold code it as JBrinkworth shows, above.

2. If you have coordinates for both end points of a line you wish to hold then hold them fixed, the bearing between being fixed naturally follows.
For example:
C 100 5000.000 5000.00 !
C 101 5192.7183 5053.4757 !

3. If you are dealing with, and adjusting, GPS vectors you only need to hold a coordinate on the base position, and specify the state plane (or other) grid zone. The vectors will be on (very close approximation of) a grid basis of bearings, based on the ellipsoid . With the proviso that ties into CORS or actual brass monuments would result in a slight modification of that bearing basis.

Thanks guys, a common situation, for me at least, is I want to hold one station for co-ordinate and the other one for orientation only.
I need to calculate the bearing and then put this in a B record.
I always wondered if there was a shortcut.

If you hold the endpoints as in my example 2, and run the adjustment, you can then inverse the bearing /distance between the endpoints in StarNet. Take that bearing and enter it to your data as your held. Saves a step.

The manual says you can hold a bearing fixed with:
B sta1-sta2 76-54-32 !

I wish I knew how they do that. I need to study constraints in LS again so I can add it to my program.

Norman Oklahoma, post: 330804, member: 9981 wrote: If you hold the endpoints as in my example 2, and run the adjustment, you can then inverse the bearing /distance between the endpoints in StarNet. Take that bearing and enter it to your data as your held. Saves a step.

Thanks I'll try that!

You have to be careful to not hold too many entities as fixed, depending upon your settings. A fixed bearing between fixed coordinates that is off just a bit due to rounding can cause StarNet to have issues. Better to let things slide a little and look at the results and then decide what to hold as fixed, as suggested above.

Ken

Hi guys,

I have ran into another problem for starnet again. In this current project now, I am required to take multiple sets of observation for my network. Downloading the data is easy. However the project specs require me to input mean angles and their standard deviations into Starnet for adjustment. Does anyone know how to input them into starnet? The closest I could find is input of standard errors but not standard deviations.

I am using a Sokkia robotic and have the StarSDR software to convert my files. Unless there is another quicker way to input my mean and standard deviations?

Cheers and thanks in advance.

Not quite sure, but I believe that "standard deviation" and "standard error" are (almost) equivalent terms in this context.

From the StarNet manual:
"... the term ‰ÛÏstandard error‰Û is short for ‰ÛÏstandard error of the mean.‰Û It is an
estimate of the uncertainty of the mean (or average) of a set of field observations. The
equation for standard error of the mean is:
Std Error = Std Deviation / SQRT ( Number of Observations Averaged )

I don't know just how closely your results will be monitored, but it seems to me that entering the standard errors of your instrument into StarNet would satisfy the spirit of your project specification, if not the letter.

Standard Error is your best estimate of the rms error or standard deviation in the entered value.

Estimate the standard deviation of the population from which your measurements are drawn using the n-1 population std dev formula. That is a good estimate of the error in each measurement. If you are finding the mean before entering into Star*Net, then estimate the standard error of the mean as the standard deviation of the population divided by the square root of the number of measurements you averaged.

The alternative is to enter each measurement into Star*Net and let it do the averaging. That is nearly but not quite equivalent.

There is a fine point about whether the centering errors are correlated. I assume you have that set in the options for the whole project. Entering the mean and calculated standard error of the mean will let Star*Net apply the centering errors once with their full effect. Entering the individual measurements will assume the centering error is random from measurement to measurement, and hence its effect will be reduced by the averaging. If you are taking all your measurements from one setup, the centering error won't in truth be reduced, and that may be why they want you to average the measurements before entering.