If you other station is one of the PIDs listed in the data sheet then yes, the local accuracy should be used. The (semi)proper way obtaining relative confidence between these two stations is the square root of the sum of the squares of the corresponding local accuracy values. The NGS sanctioned "proper" way involves some serious linear algebra and arrives at very similar results. Washington State's guide has a nice description of this topic:
Personally, I always use the network accuracy for all relative accuracy calculations, as well for the assigned weighting when preforming a least squares adj. (be sure to use 1-sigma!). Then there is no need to worry about simultaneous processing. What's 1.2mm anyway...really...
> If you other station is one of the PIDs listed in the data sheet then yes, the local accuracy should be used. The (semi)proper way obtaining relative confidence between these two stations is the square root of the sum of the squares of the corresponding local accuracy values. The NGS sanctioned "proper" way involves some serious linear algebra and arrives at very similar results. Washington State's guide has a nice description of this topic:
>
> WS Guide
>
> Personally, I always use the network accuracy for all relative accuracy calculations, as well for the assigned weighting when preforming a least squares adj. (be sure to use 1-sigma!). Then there is no need to worry about simultaneous processing. What's 1.2mm anyway...really...
I think Dan Martin, the local geodetic adviser agrees with you about the "serious linear algebra"....Here's his response to my question about the accuracy of the Vtrans marks:
>Bob,
>Most of there were done before the FDGC standard were incorporated by NGS (2007) and are not trivial to compute. As such, VT has never used or reported them. Generally speaking, the accuracy of the VTrans GPS points is within 2-3 cm network accuracy wise and usually sub-cm for marks that were surveyed together (local accuracy).
>The only thing you want to be careful of is the datum tag for the stations, e.g., NAD83(1992), NAD83(1996), NAD83(2007), and NAD83(2011). These tags represent different realizations (adjustments) of NAD83, and resulting coordinates can differ by 5 or more cm. The two stations you mention above are both NAD83(2011), so you are good to go.
>Dan
So the only remaining issue is Azimuth. On one end I'm fine. On the other, if I stop at the Vtrans mark (about a mile), no azimuth yet (might have to get that sun shot in after all). If I go all the way to the other NGS mark, there's another ngs mark within sight of it. Still pondering.
But I think I've answered my question: These stations are plenty accurate enough for my purposes. I think I'll be lucky to hit them within a foot (but the verdict's out on that).
Unfortunately, while the FGCS has created standards and specifications for survey work other standards exist both historically and even currently.
I like the NGS data as it complies with FGCS standards.
The "ersatz" NGS data sheet you posted in your other thread does not include FGCS standards because it has not gone through the adjustment process.
Observations (angle and distances) and their weights combined with constraints (coordinates of fixed control points included in the project) are used in the least squares process (ADJUSTMENT) to compute coordinates of the unknown points as well as the accuracies of the determination of the unknowns. Local accuracies can be determined with respect to directly connected points; network accuracies can only be determined when the project is incorporated into network.
Unlike GPS, classical geodetic surveying techniques, i.e using angles and distances observed at a number of points, accumulated error as a function of distance. Look at your EDM specifications, it includes a base error and a length-relative term (e.g. 10mm and 5ppm). Error also accumulates in a less straightforward fashion at each point observed (from and to).
You might find the NOAA Technical Memorandum on these classical surveys, Horizontal Control, to be of interest. It is found here: http://www.ngs.noaa.gov/PUBS_LIB/TRNOS88NGS19.pdf
As I understand your plan, you intend to run a traverse between two points and want to attach it to existing control points. The question of which points to use is important. For a horizontal control survey you want to start and end your survey on points with published valid positions. To perform a survey you need not only a starting point but also a starting azimuth; a position at the end and a closing azimuth are needed to determine the accuracy of your work. Working in plane coordinates and using astronomical observations requires that the astro azimuths be converted to geodetic and then plane.
Without observing a network and without adjustment software, you can check your work's accuracy by computing to the end point and computing the difference between the computed coordinates and published. Convert this to distance. The length-relative closure is the total length of the traverse divided by the distance between the published and computed coordinates of the end point.
I too would recommend a polaris observation over a solar. While I did not look very thoroughly, it appears that most of the intersection stations in your project area are gone.
Having gone on too long, I close.
DMM
> you should be able to find a decent one-second theodolite for under $500.
I second that, having recently purchased a nice T2 for about $320 delivered, and couple of weeks ago I passed on a Kern E2 at $250.
> I too would recommend a polaris observation over a solar. While I did not look very thoroughly, it appears that most of the intersection stations in your project area are gone.
>
> Having gone on too long, I close.
>
> DMM
Thanks for the link...and the additional thoughts on the project.
Now would be a good time to ask a really dumb question (buried deep within a too long thread that's probably boring most, so most won't notice):
To get azimuth via celestial objects, you need another mark. So, it's the middle of the night (unless you do it at dusk or dawn), you shoot Polaris, then what? It's too dark out to see anything. If you set a prism up as far away as you could, you can't rely on the reflection of the laser alone if you can't see the target itself.
A few ideas come to mind: 1. Leave the set up and come back in the morning and shoot the other mark. 2. Build a target with super micro LEDs embeded in it in an orthogonal pattern (or even just a single LED, since you're only interested in azimuth), so it's totally visible at night. 3. Shoot the back sight before dark then leave the setup until Polaris emerges, and take it. 4. Just do it in the day time and hope your optics, focus, and precalculation of where it is, all fall into place?
Let the flames begin for the dumb question.
Good advice in http://www.rollanet.org/~eksi/Handbook.htm
I'd look for distant lighted objects like radio towers, etc
I imagine there are some old Wild target sets on eBay. Fabricating something similar is possible.
For the first attempt, be sure to get everything set up during daylight.
Stop watches, time cubes and recording devices are handy.
The linked ephemeris was the last published. Mr Wahl who posts to this site has a page with an ephemeris for this year.
Good luck,
DMM
> I second that, having recently purchased a nice T2 for about $320 delivered, and couple of weeks ago I passed on a Kern E2 at $250.
A Kern E2 in working order for $250 ????? That's just ... mind-blowing.
> A Kern E2 in working order for $250 ????? That's just ... mind-blowing.
"Great used condition" is the way it was described, and the photos supported the description. When I turned down the $250 counteroffer (I had offered $190, as I recall), the seller relisted it, and it sold today for $305 shipped.
2014 Ephemeris
Ephemeris available here: http://www.cadastral.com/2014ephs.htm
>
> I imagine there are some old Wild target sets on eBay. Fabricating something similar is possible.
>
Wow. Beautiful set...$1200! Not in my price range. But one of them looks like a battery powered, back lit, translucent target...easy to make. Might try LEDs as well.
Thanks for the tip.
