Please or Register to create posts and topics.

Completely Out of Control

#11 · September 6, 2024, 12:31 pm

Quote from Kris Morgan on September 6, 2024, 12:31 pm
Dave, there is a lot of moving parts in your question, and personally, I think you're overthinking this. I'll try to summarize.

My thinking is that you will find MUCH more utility if you put your survey on the grid and stay on the grid. The connection to the grid via CORS or OPUS is too easy and allows you to move wherever you want and have a repeatable experience.

First, you want Ground coordinates. Not really, what you want is a ground distance. If you want to run on the ground, then don't really mess with an SPC. When you make a "here position" and then you "calibrate" it to "a value", you essentially are making a mapping projection, complete with scale factors and angles of convergence. The further you move from your base point, the more it gets off. If you had a statistics class in college, think of the words "relevant range". Now you can calibrate to points "around" your site through a host of procedures (horizontally or vertically but I'll get to the Z component) and it gets even worse then since you have now a polygon of control and when you step outside that theoretical polygon, scale rears it's ugly head, and the worse your control was, the faster you fall off.

Vertically, you want one of two methods to constrain to. The first is ONE SINGLE POINT. This keeps your plane level (for the most part) and is a vertical transformation. The next is SEVEN OR MORE points surrounding your parcel. Anything less and you "tilt the plane" and the verticals get nasty quick.

Consider this as an alternative, and I think you'll find MUCH better bliss through it. Make a "here position" in an SPC environment. Post Process that point to OPUS. Yes, yes yes. I hear you. You want ground distances. Just hang on. Keep it on the grid. Then when you do a tract a mile away, DO THE SAME THING. Now, the two data bases literally tie together and you're WELL on your way to making a solid data base.

What's that? Oh yeah. Ground distances. Okay, draw the whole thing up in AutoCAD, like everything. Wait till the VERY last thing, and scale all of your linework up by the reciprocal of the combined scale factor. Annotate your lines. You'll have grid coordinates that aren't fussy, and surface distances. Also, at some point, you'll hate that you chose this and eventually come around to keep it on the grid.

I've typed up the culmination of the past 25 years of surveying with GPS and how best to relay it to the clients and for longevity. I've only EVER worked on one project that the deliverable was so large, for the scale factor to have made a change in the acreage. I reported both.

There is nothing wrong with the LDP or the crowd that promulgate them. If you find you "move around" a lot, I don't think you'll like it and having to play in the right sand box. They're not hard to make but at the end of the day, it's one more thing to keep up with. If your scale factor is greater than 0.9999, then it doesn't matter on a 40 acre survey (mostly). I would save you the nearly decade long "beat my head against the wall" session to come around to this thinking that all kinds of things. I had some great teachers (they didn't know they were though) regarding the methodology of this. Not to step on any one's toes, but Shawn Billings is one of the best at this and if you want to chase this rabbit, call him. He's one of the nicest and smartest guys I know of the subject. He and I do it differently, but we've followed each enough that we read the field notes and make each others data sets relative to each others before we ever leave for the field.

Good luck.

Dave, there is a lot of moving parts in your question, and personally, I think you're overthinking this. I'll try to summarize.

My thinking is that you will find MUCH more utility if you put your survey on the grid and stay on the grid. The connection to the grid via CORS or OPUS is too easy and allows you to move wherever you want and have a repeatable experience.

First, you want Ground coordinates. Not really, what you want is a ground distance. If you want to run on the ground, then don't really mess with an SPC. When you make a "here position" and then you "calibrate" it to "a value", you essentially are making a mapping projection, complete with scale factors and angles of convergence. The further you move from your base point, the more it gets off. If you had a statistics class in college, think of the words "relevant range". Now you can calibrate to points "around" your site through a host of procedures (horizontally or vertically but I'll get to the Z component) and it gets even worse then since you have now a polygon of control and when you step outside that theoretical polygon, scale rears it's ugly head, and the worse your control was, the faster you fall off.

Vertically, you want one of two methods to constrain to. The first is ONE SINGLE POINT. This keeps your plane level (for the most part) and is a vertical transformation. The next is SEVEN OR MORE points surrounding your parcel. Anything less and you "tilt the plane" and the verticals get nasty quick.

Consider this as an alternative, and I think you'll find MUCH better bliss through it. Make a "here position" in an SPC environment. Post Process that point to OPUS. Yes, yes yes. I hear you. You want ground distances. Just hang on. Keep it on the grid. Then when you do a tract a mile away, DO THE SAME THING. Now, the two data bases literally tie together and you're WELL on your way to making a solid data base.

What's that? Oh yeah. Ground distances. Okay, draw the whole thing up in AutoCAD, like everything. Wait till the VERY last thing, and scale all of your linework up by the reciprocal of the combined scale factor. Annotate your lines. You'll have grid coordinates that aren't fussy, and surface distances. Also, at some point, you'll hate that you chose this and eventually come around to keep it on the grid.

I've typed up the culmination of the past 25 years of surveying with GPS and how best to relay it to the clients and for longevity. I've only EVER worked on one project that the deliverable was so large, for the scale factor to have made a change in the acreage. I reported both.

There is nothing wrong with the LDP or the crowd that promulgate them. If you find you "move around" a lot, I don't think you'll like it and having to play in the right sand box. They're not hard to make but at the end of the day, it's one more thing to keep up with. If your scale factor is greater than 0.9999, then it doesn't matter on a 40 acre survey (mostly). I would save you the nearly decade long "beat my head against the wall" session to come around to this thinking that all kinds of things. I had some great teachers (they didn't know they were though) regarding the methodology of this. Not to step on any one's toes, but Shawn Billings is one of the best at this and if you want to chase this rabbit, call him. He's one of the nicest and smartest guys I know of the subject. He and I do it differently, but we've followed each enough that we read the field notes and make each others data sets relative to each others before we ever leave for the field.

Good luck.

#12 · September 6, 2024, 5:23 pm

To start a JOB file for site calibration, choose “ no project, no datum” since the site calibration will create that projection.

Just like a resection, you may find NOT using every control point observed giving you lower residuals since pieced together total station data over five years will have points that don’t match into other points as well as others do.

The purpose of the calibration is to use GNSS on a local coordinate system. I rarely use a calibration, but I would think if you switch the coordinate system setting back to SF 1, you would not be able to export GNSS observed points. And TA has had improved scalding function for export, but I’m not sure if 2017 has correct scaling working on export. You could fairly easily scale values in excel.