Completely Out of Control
I'm not sure if I can describe this correctly but basically, I've pieced control together over a 40 acre site for the last 5 years using mostly standard and robotic total station and hopefully growing levels of understanding. I recently got a couple R8-3's and am playing with them using a TSC3 and Access 2017.
Yesterday I set a job up initially as 1.0 scale factor only and COGO settings set to "ground", knowing that a site calibration would probably create a transformation. I set the base up over one of the control points (local coordinate system), did the autonomous HERE and did my first HV calibration on our site vertical benchmark. That put my CAD in the ballpark on screen and I went and shot corners and CPs that I had located before. I moved through seven CPs taking 'observed control points' on them, then started playing with calibration pairs. I don't remember when in the process I was prompted for a projection but I do remember accepting Transverse Mercator and setting the geoid as well.
The goal was to check in to as many previously shot corners and control points as I could in an afternoon and see if I could get my gnss control to match (or provide an adjustment to) my piecemeal control.
In the end, the best I can get playing with calibration pairs is that I'm off of all CPs by 0.06-0.18 with a 17.8" rotation and a SF of 0.99970924 (local CSF is 0.999985... but I don't think that's relevant because it's a calibration calc).
Question is, and maybe I'm answering it by writing this: If I want to export ground coordinates should I eliminate all the calibration points except for the initial "known" benchmark shot? This would presumably bring the SF back to 1.0
Next question is: How best to go about shooting ground distances with GNSS if you're not shooting on SPC and adjusting in post? Do I calibrate to a single known point, then shoot a corner or two and do a 'manual' rotation to get best fit before proceeding (assuming it's reasonable)?
I think in general I'd like to either spend the time figuring out my own LDP or shoot everything on SPC and translate as needed for mapping.
Unfortunately I don't own TBC or StarNet yet so can't even do that very efficiently and, probably like most of you, certain things like this I need to know that I know and when I know that I don't I live bugged.
Mahalo.
There's a lot to unpack here. Any tight control, in other words control that has good geometric internal consistency, can be merged from terrestrial to GPS using an LDP. Or it can be rotated, translated against State Plane. That's horizontally.
Vertically a calibration might work, however, we're discussing Hawaii and I'm guessing the Geoid contours are about as extreme as anywhere. What that means is there won't be a plane tilt that will match no matter how it's done if the figure is too large. How large is too large? I have a small airport that no amount of calibration would "fit" level runs. So 1-1/2 miles N-S, by 1-1/4 miles E-W defeated any attempt at calibration. On the islands that size limit will probably shrink considerably.
It's best to apply a Geoid model, I don't know about Hawaii but for that small airport we now can use GPS with Geoid18 and hit control vertically less than a cm consistently and that's RTK. Using calibration for the same site there would be more than 5cm in places.
I would first collect everything in State Plane. Then inverse points and chart how the angles and distances compare to my terrestrial survey. In Access there is an option button when you inverse to see ground distances or grid distances. Your terrestrial distances should match closely to ground when you inverse. Once you calibrate you lose that clean comparison.
If all that data looks good you have a choice, design a local LDP that matches your terrestrial survey, scale up and probably rotate SPC, or calibrate.
Without TBC and only Access on a controller this will be difficult to accomplish.
Calibration isn't my preference ever. I quit doing it almost thirty years ago and haven't regretted that decision. Without TBC I'm not sure a stand alone LDP can be designed for your site using the controller.
Dunno where the idea came from that TBC or StarNET is necessary for developing an LDP.
Check out appendix E of this document:
https://www.ngs.noaa.gov/PUBS_LIB/NGSRealTimeUserGuidelines.v2.1.pdf
The equations for computing grid scale factors are a bit tedious but easily automated. Especially if you're working with just TM or LCC projections, not Oblique Mercator...
In any case, there are ways to develop LDPs without purchasing specific software packages.
I'm with Mighty Moe on this one. However, i have no idea how you would even begin to do a lot of what you want without TBC. It's pretty much a requirement when using Trimble GPS. Otherwise you're going to be fighting data.
I would tie the data on state plane, probably just an opus based solution. Move all your existing control coords to state plane, set your scale factor in TBC and you're done as far as setting up your system. How your existing control 'fits' the GPS is another thing entirely.
I'll second that on computations. It's mostly 8th grade stuff.<div>
Work in SPC.
Multiply the scale reduction by average 1/csf and you have the ldp scale.
Use sin of lat X delta long = covergence as a basis for the new central meridian by applying your rotation to the convergence and working backwards.
Apply the north and east shift from your base coordinate to desired coordiante to the false northing and easting.
Test, test, test...done!
</div>
Rover, I don't think it's a question of only being able to do it with TBC. From what the OP stated is all he has is Access and a DC, even if the projection parameters are developed by any formula or program, how then is it imputed to Access using only his DC.
I can't answer that cause I tried unsuccessfully once, got frustrated in the field, and simply collected data and fixed it later.
If anyone knows how that can be done, I hope they chime in.
Rover, I don’t think it’s a question of only being able to do it with TBC. From what the OP stated is all he has is Access and a DC, even if the projection parameters are developed by any formula or program, how then is it imputed to Access using only his DC.
The ability has been there for a long time - certainly long enough that v2017 can do it.
Create job, go to properties, then coordinate system, select "Key in parameters", then "Projection", and enter in the type and parameters.
Then on to Datum transformation, which will almost always be a "three parameter" null transformation.
Ignore the Horizontal adjustment and Vertical adjustment sections, since this is not a site calibration.
Tap Accept and voila! The job is now in a custom projection.
Rover, I'm doing a terrible job of explaining what I mean. In Access it's possible to set up a job with a custom projection. What I think the OP wants is to set up a projection that incorporates his control point values (XYZ). That I can do in TBC, however, in Access the options are more limited (at least from my experience, I have always controlled projections from the office software).
The easy way would be to locate the points with GPS, place your origin point on a control point, holding the GPS L, L, H and the record XYZ of the control point.
Use a scale factor that places the LDP at surface. Inverse the points and apply a rotation angle to then mimic the GPS projection to the XYZ values. If that's an option in Access I missed it in the field, but in TBC it's quite available.
The more difficult way is what Doctor Herb discussed in a seminar which is to place the origin point so the Lat, Long, XYZ values pass through the points.
Frankly, it's best to set up a good regional LDP, locate project control points and rotate the old data to the new values. Sometimes, that isn't possible if the control is for a finished design and needs to be held, but everything I work on these days is GPS based anyway.
Thanks to all for your time in this discussion. Very helpful in wiping a couple more cobwebs from those lonely parts of my brain. Mahalo.
I think in general I’d like to either spend the time figuring out my own LDP
You will never regret the effort it takes to do that. You are headed down the right path. When you begin a control network using total station observations you pick a point pair as a basis of bearing. I like to use that same basis for rotation when designing an LDP to fit an existing coordinate system. Also like to use the first setup point in the network as the origin point of the LDP. Designing an LDP gives you so much more control over what is held in calculating the parameters than a calibration. If I've said it once I've said it a thousand times. Friends don't let friends field calibrate. Field calibration assumes the measurements made to the held control points are either perfect or equally imperfect. That is never true.
