Kent said near the end of Chapter Two:
"There is a workaround for it, though. If you add four, very sharp marks at minimum distances from the instrument (in other words, the targets of the test range consist of four targets about 8 ft. away at nominally 0, 90, 180, and 270 from the instrument and five more distant ones as previously described) and
- take directions to them in the round of directions in each set, and
- measure approximate distances to all points (+/-0.5 ft. is perfect),
you can enter the directions in Star*Net to solve the standard error of a direction to the more distant targets allowing for the small differences in centering. In this procedure, if you are actually computing the centering errors for each setup, the more distant targets could be as close as 20 ft. as long as they were perfectly stable between sets. While one is at that, it would be easy enough to add extra setups with directions to the close targets to actually determine the standard error of instrument centering at the same time. Ten setups with directions to the close targets should be plenty.
The solution via Star*Net is the value of the standard error of a direction that gives a standard error of unit weight in the residuals and can be easily arrived at by trial and error."
I'm preparing this next test.
Here are my targets:
I started the test with another clunker of a TS...the display was screwed up so I had to abandon it, but I found that these targets would be super accurate at a shorter distance than 150'. It took me 45 minutes just to lay out the "pentagon" at the local elementary school. This would be way easier at 20' or even 50'.
The center mark I'm setting up on has a very tiny hole in the paper, and it's backed with black electrical tape, making a center mark about .020" in diameter, making the centering errors less than about .002' if I take my time (which I did/will). You can't see it in the photo. The dot in the optical plummet exactly covers the hole when centered.
Then, I wasn't sure that the 5 distant targets should have been shot WITHOUT re-zeroing the instrument, after sighting the 4 close targets. Does it matter, if I'm putting these into Starnet? If I zeroed in between shooting the 4 and shooting the 5, wouldn't Starnet just rotate the two sets?
Also, I still am not sure if there's any advantage to shooting with both faces, and if so, once you flip and turn the scope, do you re-zero then?
Bottom line, with this kind of digital TS, does zeroing the unit do absolutely anything at all, except internal math...i.e. nothing with the reading of the plates?
Preparing my Starnet file, ahead of gathering the final data.
Got the four close targets in, but having trouble with how to treat the five main targets.
Using B (Bearing), I get "Station Incorrectly Connected to Network: xxx"
SS gives me "missing or invalid station list".
Using Direction shots (DB, DN and DE) gives me "station incorrectly connected to Network as well.
How should these shots be entered into Starnet?
Did you give enough distances? They don't need to be precise, but must be present sufficient to locate all points.
I tend to avoid the combination records when I run something with Star*Net, and specify everything with C, D, B, and A records.
> Did you give enough distances? They don't need to be precise, but must be present sufficient to locate all points.
>
> I tend to avoid the combination records when I run something with Star*Net, and specify everything with C, D, B, and A records.
Didn't put distances in on any of the main targets; didn't think they were needed. I measured them to .1', so I entered them and that works fine. I'm all set for data now. Thanks!
Where did you think Star*Net was going to put those points? It has to find coordinates for them somehow to calculate angles with. I'm pretty sure that it works by optimizing coordinates to best fit the observations. They could have been anywhere from arm's reach to Alpha Centauri. :-O
> Then, I wasn't sure that the 5 distant targets should have been shot WITHOUT re-zeroing the instrument, after sighting the 4 close targets. Does it matter, if I'm putting these into Starnet? If I zeroed in between shooting the 4 and shooting the 5, wouldn't Starnet just rotate the two sets?
No, don't rezero during a set of directions, ever. The way that I would observe the nine targets is to take Face Left directions to the outer 5, change to Face Right (without rezeroing!) and take Face Right directions to them.
Then (without rezeroing!) return to Face Left and take one of the outer 5 targets and then the inner 4. Change to Face Right (no rezeroing, please!) and take the direction to the one of the outer 5 just observed Face Left and the inner 4.
That will give you a consistent set of directions to all nine targets that will work just fine for the purposes of the test.
> Also, I still am not sure if there's any advantage to shooting with both faces, and if so, once you flip and turn the scope, do you re-zero then?
Yes, you absolutely, positively must measure all directions both Face Left and Face Right. No slacking on that detail.
> Bottom line, with this kind of digital TS, does zeroing the unit do absolutely anything at all, except internal math...i.e. nothing with the reading of the plates?
I would set zero only once on a round of directions.
>
Those targets look promising, by the way. Use a torpedo level to make sure that they are driven quite nearly plumb. You don't want the "V" (which includes the apex of the green field at the top of the "A", I assume) that is the actual target leaning left or right.
BTW, doesn't the demo version of Star*Net (which you're using, as I recall) have about a 20 data line limit? If so, you won't be able to adjust four sets with nine DM or DN lines simultaneously, but will have to break the processing down differently. Probably the simplest solution would be to post your data in Star*Net format, ready to run, let someone with the fully functioning version of Star*Net run the adjustment for you, and post the results.
> Probably the simplest solution would be to post your data in Star*Net format, ready to run, let someone with the fully functioning version of Star*Net run the adjustment for you, and post the results.
Excellent idea. I'd like to learn how to create the format manually (then I can look back fondly at all the work I did prior to buying myself a Data Collector)
The follow up explanation on how to take all the sights was great too! Only hold up now is to wait for my instrument to come back from my dealer being fully cleaned/adjusted/calibrated.
> No, don't rezero during a set of directions, ever. The way that I would observe the nine targets is to take Face Left directions to the outer 5, change to Face Right (without rezeroing!) and take Face Right directions to them.
>
> Then (without rezeroing!) return to Face Left and take one of the outer 5 targets and then the inner 4. Change to Face Right (no rezeroing, please!) and take the direction to the one of the outer 5 just observed Face Left and the inner 4.
Starting with a fully dealer calibrated instrument. Followed Kent's instructions precisely. NO RE-ZEROING.
Before repeating the setup a bunch of times (rotating the tribrach, per comment in Chapter 2), and even before putting into Starnet, I post my data, right from the field book, to make sure I haven't screwed anything up.
Also factored in was that as soon as I got all set up and took one complete round smack in the middle of the local elementary school soccer field, 50 kids showed up for a game...ARGH! Nothing's ever easy.+o(
> 
I went ahead and tried getting everything into Starnet.
I subtracted 180 from the face right measurements, and put all measurements in.
And here are my Instrument settings:
> I went ahead and tried getting everything into Starnet.
> I subtracted 180 from the face right measurements, and put all measurements in.
It looks to me as if you entered the Face Left and Face Right directions to the same targets as separate observations. They aren't separate observations. The observations are the *means* of the two directions (one F Lt and one F Rt) to each target.
> It looks to me as if you entered the Face Left and Face Right directions to the same targets as separate observations. They aren't separate observations. The observations are the *means* of the two directions (one F Lt and one F Rt) to each target.
Yes, I did. Corrected. Now Starnet is telling me "no redundant measurements"....Which is of course, correct.
Here's the updated Starnet output:
So, to recap (I re-read all of Chapter 2), where I'm going with this:
I'm going to do the above multiple times as follows:
1. Rotating the instrument 30 or 40 degrees for each set.
2. Record measurements as above (FL/FR, Mean them)
3. Enter the additional 8 observations into Starnet
I'm still unclear on two things:
1. Distance to the far targets. You say 150' is ideal, but that requirement goes away with my close in targets. Is this Correct?
I've used 20' and 50' this time...not too different. They would be absolutely stable between sets. Not sure what the close in targets are giving me in the way of data that's different than the far ones. With the field I have, I can probably get to 100' if it matters (as long as those soccer kids leave me alone:-D
2. Still unclear on whether I need to simply rotate the instrument between sets and re-zero, or physically, turn the tribrach and re-center the instrument.
I thought this endeavor was intended to provide insight to INSTRUMENT ERRORS, not "operational" setup (centering) errors.
> I'm still unclear on two things:
>
> 1. Distance to the far targets. You say 150' is ideal, but that requirement goes away with my close in targets. Is this Correct?
Yes. The near targets enable the correction for small centering errors at each setup. One thing to keep in mind is that all of the directions need to be measured in one session if you are using targets of questionable long-term stability such as the 1x2 stakes with applied targets. It's imperative, particularly at shorter ranges that the targets not move even by a whisker from setup to setup.
> I've used 20' and 50' this time...not too different. They would be absolutely stable between sets. Not sure what the close in targets are giving me in the way of data that's different than the far ones.
20 ft. is too far for the inner circle of targets. 10 ft. would be more like it with your instrument. Their purpose is to allow the computation of the instrument center for each setup in relation to the position of the instrument at the first setup. This is partly why the test has to be done at one time, not spread out over several days.
> 2. Still unclear on whether I need to simply rotate the instrument between sets and re-zero, or physically, turn the tribrach and re-center the instrument.
You need to rotate the circle, and with your instrument that means physically rotating the tribrach that the instrument is attached to, and then recentering the instrument after rotation as well as you can. The inner circle of targets will allow the correction for any remaining centering errors.
> I thought this endeavor was intended to provide insight to INSTRUMENT ERRORS, not "operational" setup (centering) errors.
The instrument errors are partially dependent upon which part of the circle the directions are taken on. So, to eliminate the errors from that, the circle must be physically rotated so that the measurements are taken over quite different parts of it. This is easy to do in geodetic-grade instruments which have circles that can be rotated without disturbing the centering of the instrument.
>
> The instrument errors are partially dependent upon which part of the circle the directions are taken on. So, to eliminate the errors from that, the circle must be physically rotated so that the measurements are taken over quite different parts of it. This is easy to do in geodetic-grade instruments which have circles that can be rotated without disturbing the centering of the instrument.
Perfect. Thanks. I'm on it.
> >
> > The instrument errors are partially dependent upon which part of the circle the directions are taken on. So, to eliminate the errors from that, the circle must be physically rotated so that the measurements are taken over quite different parts of it. This is easy to do in geodetic-grade instruments which have circles that can be rotated without disturbing the centering of the instrument.
>
> Perfect. Thanks. I'm on it.
An important thing that I forgot to mention is that you should use the "DM" and "DN" data types for the input file, not the "A" and "D" types. The reason is that you aren't measuring separate Angles, but are observing rounds of Directions which are related to each other differently than individual angles would be. The DM lines would be for the very first set of directions, with approximate distances to the targets +/-0.10 ft. Then all later sets of directions would be entered as DN lines, pure directions, no distances.
That is, the program you should be following is just taking directions to a series of targets Face Left, then changing to Face Right and taking directions to the same series, not measuring separate angles to each target from one reference target. This detail should not be overlooked.
>
> That is, the program you should be following is just taking directions to a series of targets Face Left, then changing to Face Right and taking directions to the same series, not measuring separate angles to each target from one reference target. This detail should not be overlooked.
What's the difference between "taking directions" and "measuring angles"?
I point to target A1; zero the instrument; turn to A2; record; continue to A3; record, etc. I'm not measuring A1 to A2; A2 to A3, etc., if that's what you mean.
I understand not recording distances after round 1 though.
> What's the difference between "taking directions" and "measuring angles"?
> I point to target A1; zero the instrument; turn to A2; record; continue to A3; record, etc. I'm not measuring A1 to A2; A2 to A3, etc., if that's what you mean.
> I understand not recording distances after round 1 though.
You are recording directions to the targets and those directions should be treated as the DM or DN data type, not as individual, independent angles. Each direction is more closely related to every other direction (has a smaller relative standard error) in the round than if separate angles were measured to each target from the reference target.
The other extremely important thing to note is that you will have to assign a different name to each setup (O1, O2, O3 ... or whatever) so that Star*Net will recognize that the different setups may have slightly different coordinates. Otherwise, Star*Net will treat them as being the identical point with the identical coordinates, which defeats the purpose of the inner ring of targets.
> > What's the difference between "taking directions" and "measuring angles"?
> > I point to target A1; zero the instrument; turn to A2; record; continue to A3; record, etc. I'm not measuring A1 to A2; A2 to A3, etc., if that's what you mean.
> > I understand not recording distances after round 1 though.
>
> You are recording directions to the targets and those directions should be treated as the DM or DN data type, not as individual, independent angles.
OK. I get it now. You're describing how I enter the data in Starnet, not how I observe the targets. One step at a time, lol.
I'm going to concentrate right now on just capturing all the data properly; I'll deal with Starnet later. If I can't do it with the Demo version, I'll figure out how to create a Starnet ready data file for someone here to help with.
Off to the field:-)
> I'm going to concentrate right now on just capturing all the data properly; I'll deal with Starnet later. If I can't do it with the Demo version, I'll figure out how to create a Starnet ready data file for someone here to help with.
It may be that the limit on the demo version of Star*Net is the number of points, not observations. I don't know. If so, you should be able to adjust all of your observations in demo mode, but the offer stands to run the adjustment if you post a Star*Net file substantially ready to run.
