2 cm off on one point and 2 cm off the other way on the second RTK derived control point and walla. 0.06 - 0.07'
My guess would be that absolute accuracy, relative to Earth or the current model of its center, will be of higher value than relative precision.
In my part of the world, that PLS 100 years in the future will find that my plat coordinates are about 6 feet off.
And that might be precisely the data a California PLS needs in a hundred years!
How many more wildfires, earthquakes, landslides, or volcanic eruptions need to take place before PLSs recognize the utility in referencing their parcels to Earth? I'm not trying to pick on anyone or insinuate that anyone isn't doing a great job, I just find it somewhat shocking that most of us have the ability to show exactly where a given parcel sits on its tectonic plate but choose to omit this data because we can't envision how it could be helpful in the future.
RTK/RTN 180 seconds per point on 3 monuments that see each other, put the rover in the truck and forget about it.
Gun up, backsite, crank angle as a check.
Slide the backsite corn-nut toward the gun, forget about the check shot.
Translate and rotate record data on over and start hunting for more evidence.
Not saying it is right, but, I have done it under the direction of well respected LS's.
The purpose being to bring the survey onto some datum to meet a municipal requirement. I think the more respected way would be to do something pretty similar but drop perform a more robust adjustment.
In a similar vein if my PC has to stake on the fly… my standing instructions are to generally RTK 3 control points for 3-5’ a piece, and then resection off them on a fourth random point.
This thread boils down to “where’s the error”, and the answer is that EVERY point has error associated with it. Either set and adjust out your control to chase out error, resection to minimize it on the fly, or just accept the fact that you missed your backsight for X and move along (I wouldn’t be caught dead doing this, but YMMV).
Edit: standing instructions example assumes if no reliable/usable control nearby
SF at your location is about 50ppm, you will need to shoot a distance of 1000' to see a difference of .05'. That's at 100' elevation.
With state plane, as a rule of thumb, figure about 0.01' / 100 feet for scale distortion.
So here many set rtk control. Thats on state plane. They then will set up on one point be the other hold the total station distance and off to traversing. They assume it’s a grid to ground issue and hold one point as state plane and the bearing. Because they run a closed loop this is not an issue and such. Now the problem is which point is right or do both points have some measurement error along with a grid vs ground issue as well. Here it’s about .10 ft per 1000 ft. On average. I don’t work that way as I want any tool I choose to use at anytime be all on everything checking wise. So I use rtk and the robot gun all throughout my surveys. I use all data on same datum usually state plane. I then if job requirements call for me being on ground scale it all. Save that information. Least squares no loops no compass rule. Use the best observations from the best tools and if I can an achieve what is needed then I am happy and LS is happy. Trust me they have made me prove this and now this is the way we work. It’s very productive and great data is coming out this way. A crew can set up on points and we see very little discrepancies in total station and rtk. But you have to do it right.
Right now, we're using RTK to recover and locate as many of the 60 calculated property corners as possible. We're not only recovering and locating property corners of the parent property but also any adjacent plats including plat corners and centerline permanent control points. We've been given 10 section and fractional corners to locate as well. All missing property corners will be reset using a 5" total station. I'm not sure we're going to run a traverse due to obstructions. I'm guessing we'll set control points to be used as occupied and backsight points from which to set property corners from. As a poster pointed out, which of any 2 control points is right?
We never bother with scale factors. How does the data collector know which scale factor to use? I've never seen 1.00000000 as a SF when I create a new job.
@field-dog Are you running Trimble access? If so when you set your job up as scale 1.00000000. You are ground if your rtk job is set as say state plane. You are on that grid. That’s mixing apples and oranges. If you are running Trimble access use the same exact coordinate system for both the RTK and the robot. It scales your ground shot distances to the same datum as your RTK control. If you set your data collector to say state plane and not grid but either ground calculated scale factor or keyed in it knows based on what tool you are using RTK or robot to scale appropriately between the two. I am sure if you are using arms on or Leica all that can be done as well. I cannot see say setting control with RTK say pairs in areas you need to set a corner because it’s missing. Yet shoot all the other corners with RTK. Then set up and do the whole adjust the BS thing on distance. I would just set the corners with RTK. At least they are all relative and you can account for the uncertainty. If running Trimble access it doesn’t matter if it’s a state plane or UTM or any other known projection. It can handle all of that just keep the coordinate system manager up to date for your area. We do a lot of surveys that we never even use a robot or total station. Now correct field procedures need to be used and also understand the uncertainty in RTK positions weighted against the standards. So I am not doing that on a lot survey. As my relative precisions are much better with the robot. Now a rural boundary with some acres up and down hills through swamps rtk all day because traversing with many 300 ft average sites to get around my positional error tolerance with robot from one side to the others is not better than what can be achieved with rtk done correctly and time wise. That’s something that every project we do we say can we live with X uncertainty will it meet the state specs. Will it meet say the design survey requirements if that part of it. This dictates how and what we use to perform the survey. Which tools. Now a real world example. We did a 75 acre survey nasty nasty site. All cut over timber that’s so thick you can barely walk through it. We opted to perform the boundary all rtk base and rover. We have met our state requirements actually exceeded them. It took a few days of base and rover multiple observations and threw out some vectors and re observed. Now once they clear it we will do some further control work for what’s coming next. I have some good control outside the site we can establish some better construction control on the site from. This will be easier once they mulch all the mess. One area was checked with robot with our largest uncertainty. We still had less than .04 horizontal and vertical. Are you using NRTK or base rover.
Not sure of the typical difference between grid and ground in my area. We’re working at
28.5831367°, -081.2134467° lat, long, Florida SPC zone 901 east.
You can use: https://www.ngs.noaa.gov/NCAT/
Thanks for the real-world example. We’re using Carlson SurvPC with NRTK. In Job Settings (System) the projection is NAD83 (2011)/Florida (East). I went into raw data to find the scale factor. It’s 1.00000000. The boss wants the property corners set with the total station.
@field-dog so you are on state plane. With RTK. When you set the corners with your robot do you use same job settings for coordinate system if so I believe Carlson is doing the work anyway. So any error or difference in distance between points is in your NRTK observations. Not a grid to ground issue which down that way is probably not even as much as we normally see in the rest of conus.
I just learned that the scale factor is viewable by selecting the edit button next to projection. It is 0.999941176471. I had previously mentioned the projection was NAD83 (2011)/Florida (East), but that’s the system. The projection is Transverse Mercator. You already know all that; I’m still learning. What does false northing and false easting mean?
What does false northing and false easting mean?
At the center point of your zone there will be a place where the defining meridian and latitudinal lines meet. Normally one might assign the coordinate 0,0 to that point. But that would result in negative northings to the south of that point, and negative eastings to the west. To avoid that some other pair of numbers is assigned to that intersection point. These are the "false" northing and easting of the system. The values are assigned strategically so that resulting coordinates for different zones are significantly different in areas where two or more zones overlap, to alleviate confusion.
The CM of your zone is W81d, your project is near W81d12' and your convergence angle is 0D06' which means close to 1/2 of the longitude angle between the CM and your location (this relationship changes between the equator and the pole). The false easting 200000m should run along W81d of longitude, of course the false northing of 0m was set at the origin south of the state to keep northings positive for the zone on land. Some time ago I heard the explanation of calling those numbers false, you might look it up cause I don't remember.
@field-dog ok also norm has you going on the false northing and easting. Pay attention scale factor elevation or ellipsoid factor and combined factor. They are different to go from ground to grid or the reciprocal. Scale factor x’s the elevation factor/ellipsoid factor gives you the combined factor the combined factor takes you from ground to grid. The reciprocal of that combined factor gives you grid to ground. Look at any NGS data sheet all three are listed. NGS always published the combined factor so that is always from ground to grid. Go to NGS website and go to past webinars go to around the 2014 ish time frame look for a couple videos these are free. Dave Doyle as instructor Horizontal datum’s or state plane. Another is vertical datum’s. These are great videos you can start to watch and gain a little understanding as you learn. Also many more great training videos there. Next download the geodesy for a layman. It’s a great pdf book to really start gaining knowledge. Don’t let it scare you on what you don’t understand right away I use to read that thing back in the 90’s when we were doing static sessions for hours. I read that thing over and over again asked questions and still go to it once in a while. So your homework is define what the scale factor is. On a lambert and transverse Mercator in Florida you have both so good info to know.
Right now, we're using RTK to recover and locate as many of the 60 calculated property corners as possible. We're not only recovering and locating property corners of the parent property but also any adjacent plats including plat corners and centerline permanent control points. We've been given 10 section and fractional corners to locate as well. All missing property corners will be reset using a 5" total station. I'm not sure we're going to run a traverse due to obstructions. I'm guessing we'll set control points to be used as occupied and backsight points from which to set property corners from.
It seems that locating all the corners via RTK, then calculating other points from those RTK'ed corners, then setting random control around via RTK, then getting out a total station to set corners doesn't accomplish much - if anything. You already have RTK based error in the existing corners, the calculated corners, and the set control points. No amount of finagling with the total station is going to remove that. You just need to be able to prove that your RTK points meet any state requirement of relative accuracy. I'm not sure the example back sight check error you mentioned earlier would do that.
Once the observed points are processed and checked to comply with required tolerances, the coordinate values are the coordinate values and the back sight check shot should be somewhat in line with the relative error reported between that point pair. If that relative error meets requirements, then why do all the extra maneuvering to get a 'feel good' number to show up on the check shot screen.
Edit to add: At each control point pair, if you are just randomly selecting one control point as good and adjusting the back sight distance based on the total station measurement before setting your new corners, it seems like you would be slinging error around like a short order cook slinging out fries.
In Access, you can do a setup backsight to anything you want and can see.
You get asked for an observed point name.
The point does not have to exist -If it does the azimuth and distance will be calculated for you.
But you can type exactly any Azimuth you want to set on the instrument at its current pointing - I hope this might work for our original issue here
ProTip: if you do Angles-only you dont need to be able to measure the distance - great for distant Trig stations
Edit to add: At each control point pair, if you are just randomly selecting one control point as good and adjusting the back sight distance based on the total station measurement before setting your new corners, it seems like you would be slinging error around like a short order cook slinging out fries.
Exactly. If they really wanted to know that their corner positions were within their error budget, they'd follow up with a 5-15 minute static session on the corner after it was set and post process to get a position tighter than what they're going to get with RTK. Given the time it takes in setting up the gun and a backsight, there's no time savings or reason not to and it saves chasing around this RTK error with a gun.
Just because I'm paranoid, doesn't mean they aren't out to get me.
Here ya go. Calc'd with Corpscon v6. CF= 0.999948054
