In the Locus software...
We're doing a small survey.
2 state plane knowns as fixed points.
Good triangles - 2 vectors from the knowns going into some unknowns.
Now, if one of those vectors fail
you can exclude that vector, and adjust your survey. 1st question.. Is this recommended if only 2 vectors are going into that point?
2nd question... When you exclude that vector you can still adjust the survey and the Site Report gives you good standard errors on that unknown point.. Why?
Thanks for any insight.

Todd,

Did you adjust the processed error scaling factor ?

See prev. Thread "What's going on ?" a thread page back. Good info.

Jimbo
Modified By James Webb on 3/28/2001 at 2:43 PM

Todd,

Regarding question 1, the advantage of having 2 or more vectors to a single point is the redundancy. Think of your triangle as a closed traverse. You can perform a loop closure to check the quality of your data. By eliminating one of the 2 vectors going into your unknown point, you are depending solely on the quality of the remaining vector to compute coordinates for the point. Think of this as a side shot. If you are 100% sure that the one vector is a quality vector, then there is no problem. Of course, you cannot be 100% sure. All survey data has a probability that the results given are incorrect. Although the processing software indicates that the vector is good, it is not unheard of for the software to be wrong, i.e. assigning small uncertainties to a vector with bogus results. It is not common but it does happen. So, any point who's coordinates are based on only one vector is subject to a higher probablity of being incorrectly positioned. The more observations you have into this point, the smaller the probability of error. You have to decide what level of uncertainty you can live with.

A question for you. When you state that the vector has 'failed', is a failure in processing (QA Fail) or in adjustment (Tau Test Fail). If it is in processing, you should attempt and adjustment with the vector included. Sometimes, a vector that has failed the processing QA test is still a quality vector. The adjustment will tell you if it fits or not.

Regarding your second question, let's look at a simple example to help answer the question. Let's say you have a project containing only 1 vector between two points. One of the points has known coordinates which you hold fixed during processing. If you adjust this project, holding fixed the same point, you will get the same coordinates of the other point. The coordinates are computed based solely on the one vector. The standard error for the vector will become the standard error for the second point, i.e. an 1 cm standard error in the N, E, and U component of the vector will produce a 1 cm standard error in the N, E, and U component of the second point. So, in your example, the point with only one vector has good standard errors because the one vector into this point has good standard errors.

Hope this helps.

Bill Martin
Ashtech Precision Products