Thanks Mark - I had initially followed the import instructions from Starnet's site (export fixed format JOB Xml, and use the Star*JOBXml import convertor); it looked a lot different than Starnet's "example" template, using bearings instead of measured angles.
I tried as you suggested with DC format and the Star*TSC convertor, and I've gotten it to work in 2d so far. With the import, it brought in some extra observations and redundancy, so at least the error ellipses reduced slightly. Starting with the things that I can mentally visualize - most looks like Chinese to me.
Lots to learn here, feels like one step forward and a few back, even without adding any of the extra network density I would imagine this program is designed for.
Mike J.
I am having trouble understanding you here in this post. First you show a figure( pentagon) with other lines.(very first post); then a later post
you post a figure from an earlier post. On the second post with a figure you say "I didn't use "1" as part of traverse because it's line of sight from "9".
That makes NO sense to me; please explain.
What do you want to do with the first figure(pentagon)?
JOHN NOLTON
Sorry for the confusion - the pentagon was just an example of what I'm trying to learn how to do - make a network in points instead of just a perimeter closed loop traverse.
The rectangular site (posted along w/ the starnet results, and popping up in some of my other posts) I just put up after inputting my traverse data into starnet and seeing what comes out, in case their were some obvious blunders. I haven't entered in any of the other intermediate control points that would make more of a network. Figured I should try to learn more before I proceed and confuse myself further.
Though sent you a PM, John.
Mike J. one last question WHY the point #11 and point #1, I did not understand that.
I like your pentagon so lets talk about it. First forget about point F and G. So you start at point "A" and backsight "B". Lets also say you are going
in a clockwise way around the figure.
Sitting at A you backsight B and rap up your angle to "E"
You move to "E" backsight "A" rap up your angle to "D"
Move to "D" ,backsight "E" and turn to "C"
Move to "C" turn from "D" to "B"
Move to "B" backsight "C" and turn to "A"
This now closes the traverse. One thing you will notice is that you have measured all the interior angles in your figure (in this case). This
will give you a check. The SUM of the interior angles of a polygon = (n-2) times 180. n is the number of sides in the polygon. So in this case
you have ( 5-2 ) times 180 = 540 degrees
One thing that will make your traverse stronger is if you can make a distance shot to another point. Say from "B" to "E" or "B" to "D".
Try this in StarNet and see what you get.
In your other figure if you could measure from #2 to #9 or #1 to #7; or both with GPS you will have a very strong figure.
Does this help in any way?
JOHN NOLTON
I've looked over Starnet's user / example manuals, and I see there's a lot of different means to an end; whether using "M," "T," "A," etc. functions or combinations thereof. And I can guess, given a certain site, a single person can come up multiple variations in approach both when observing data and then when inputting it.
I whipped up this drawing of my practice site since it seemed I was being less than clear I what I was trying to do. All angles and distances, I've simplified.
I have an Access question - is there a way to see the raw turned angles while working, or even an export format that would give it, as opposed to just azimuth and distance? I'm not sure if this is possible, since the TS isn't like a theodolite where I'm physically moving a plate and the turned angle *is* what's being recorded.
So far, the only way I've gotten the raw turned angle info is upon import to Starnet. I haven't found the method to get it from Access on its own.
Side point, but a seldom known feature of Star*Net...
Not to sure about modern versions, my version dates to immediately after the DOS only days... there is/was a method to do a preanalysis, it aids you designing your network. All you need is rough locations of your traverse and tie points (scaled coords work fine) and then input where the various angles and distances are measured. Based of your instrument specs it then gives you error ellipses. add or subtract distance or angle measurements and see if it strengthens the network, or if it is wasted effort.
Peter Ehlert, post: 373654, member: 60 wrote: Side point, but a seldom known feature of Star*Net...
Not to sure about modern versions, my version dates to immediately after the DOS only days... there is/was a method to do a preanalysis, it aids you designing your network. All you need is rough locations of your traverse and tie points (scaled coords work fine) and then input where the various angles and distances are measured. Based of your instrument specs it then gives you error ellipses. add or subtract distance or angle measurements and see if it strengthens the network, or if it is wasted effort.
It's still there in the latest version, and quite useful IMO. It can sometimes be a real eye-opener with respect to how and where to strengthen your network.
The only superior evidence is that which you haven't yet found.
Starnet is THE way to go, because it is easy and totally flexible. Loops like yours can be typed in 20 minutes or so. You will learn much faster if you type the stuff in. You can do a small project like yours with the free version. Once you understand, you can then figure out how to translate data collector files. IT is easiest to learn if you put in ALL those sub loops. This procedure will work every time:
Put in a starting point at a real or arbitrary coordinate:
C 1 5000,5000 (This line is a Coordinate, and the coordinate is for point 1, and the north and east are 5000 and 5000)
Put in a real or arbitrary direction from the first point to the second point:
B 1-2 N00-00-00E (this line is a Bearing, the bearing is from point 1 to point 2, and the bearing is due north)
Put in a traverse leg using M (for measure)
M 1-2-3 14-29-12 200.18 (this line is a Measurement, backsighting 1 setting on 2 locating 3 angle right is 14-29-12 distance is 200.18 feet)
The M command can also add zenith angles and names for the point you measure to. THIS IS NOT FOR SIDESHOTS!
The same info as above, with SS instead of M, would be a sideshot. Sideshots are ignored until the traverse loops are all adjusted.
Use the M command to put in ALL remaining shots of ALL the loops. Put it all in there.
If you have a place where you only want to put in an angle, it looks like this:
A 7-8-42 12-14-25 (this line is a Angle, backsighting 7-setting on 8 and foresighting 42, and the angle is 12-14-25)
If you don't like using this angle format XX-XX-XX, you can put it in as XX.XXXX. Starnet does not care.
If there is someplace where you only need to plug in a distance:
D 18-26 412.57 (This line is a Distance line, distance is from 18 to 26, distance is 412.57 feet)
Your project should start with a coordinate and a direction, then be completed using mostly M lines (traverse angle and distance in one line) with the occasional angle only or distance only lines (A lines and D lines)
Least squares (Starnet or anything else) is the ONLY way to handle multiple loops that works. It does depend on redundancy to allow easy blunder detection, so turn your angles at least twice, and measure your distances at least twice. For example, direct and inverted angles are shown below, as well as redundant distances:
M 7-8-9 12-16-24 908.12 (An angle and distance that locate 9 from the already established 8 and 7)
A 7-8-9 12-16-27 (angle is repeated with gun upside down, second angle is input here)
(you can put in a space here to show that the following line is a new setup)
D 9-8 908.13 (after moving up, distance between 8 and 9 is measured again, but in the other direction)
You have now measured and input the 7-8-9 angle twice, and the 8-9 distance twice. This is redundancy, and least squares needs it. Without it blunders are much harder to isolate.
Starnet is very powerful and very easy to use. Once you have finished, you should have either blunders to find, or coordinates which, although arbitrary, relate to each other perfectly. You can then import those points into a survey software (I use Terramodel) and MOVE them so that one of the arbitrary STARNET coordinates (lets say point 5) is the same as it's real coordinate value, which might be a state plane coordinate. You would then ROTATE all of the starnet points around that point until one of the other points was coincident with it's "real" (state plane) coordinate value, or until some bearing between two of the points was the correct bearing. That done, you would now have the correct starting coordinate, and the correct starting bearing. You would take that information back into STARNET, and replace the starting 5000,5000 with a real coordinate, and replace the starting bearing with a real state plane bearing.
The above (hopefully) explains how you can start a traverse at point 1, and traverse thru 2 3 4 5 6 back to 1, then traverse from 3 to 7 8 9 10 back to 5, then traverse from 8 to 11 12 13 to 2, then get GPS RTK coordinates on points 5 and 11, and get it all to work. You traverse as noted above, input it all into starnet with a fake starting coordinate and a fake starting direction, process it all, import the points into some survey software, move and rotate those points so that they "fit" the "real" (GPS) coordinates. Then, take the "correct" starting coordinate and starting bearing out of the software, and plug those numbers back into the STARNET software. Reprocess the traverses with the correct starting coordinate and bearing, and now everything should fit. Now you can import the control points and the sideshots into your survey software, and do yer stuff.
Once you Starnet, you will never go back......