Yes, you can do that, but I've always been wary of leaving out data. That's where statistics sometimes turns into lies.
If I were doing this, I would adopt the surveyor's values for these 5 points and use these formulas or something else to adjust points the surveyor didn't occupy or resurvey those points. Perhaps do the 5 math adjustments and resurvey a sample to see how it works.
It's a really good learning example and thank you RFC for sharing it.
I can see from my spreadsheet that 3102/1:2 has the largest difference of squares, but how do you know that the problem is with the Surveyor's coordinates and not with RFC's?
but how do you know that the problem is with the Surveyor's coordinates and not with RFC's?
In reading the data posted earlier, the surveyor refers to the existing corners as "EIR/.5RODUP12, EIP/.75UP16, EIR/.5REBARUP9, and EIR/.5REBARUP4". That would prompt me to wonder if the two parties are measuring to the same point. I am assuming that "UP#" refers to how far the monument is sticking up from the ground. Some surveyors just measure center top of the exposed pin while others measure where the pin enters the ground, and yet others spin the pin and try to get the center of that motion.
So we know there's a difference, but how do we know that there's a problem with Surveyor's coordinates and not RFC's coordinates or no problem other than a plumbing problem?
So we know there's a difference, but how do we know that there's a problem with Surveyor's coordinates and not RFC's coordinates or no problem other than a plumbing problem?
Exactly. At this point we only know there is a difference. That difference is considered a problem, but we don't yet know the nature of the problem.
The suggestion it was potentially a bad RTK measurement has potential, but we don't know that as fact at this point.
I can see from my spreadsheet that 3102/1:2 has the largest difference of squares, but how do you know that the problem is with the Surveyor's coordinates and not with RFC's?
I might be able to add a partial answer:
As far as I know that’s the only point that is showing a significant difference with my points once I’ve translated but not rotated my survey.
I observed point 2 from 3102, 3102 from point 2,with a total station; each observation consisting of multiple measurements both forward and reverse with the scope. probably two dozen or so individual measurements.
My total station had been recently calibrated by a ccompany qualified to do so. I would think that was such redundancy with the total station, I might be out .01’or .02’ after StarNet is done with things, but certainly not .28 feet.
I am still absorbing landbutcher’s advice to translate, but not rotate my survey. math teachers advice is way above my pay scale lol, but it is on my list of things to absorb and try to understand.
if I can settle the issue by going back out and remeasuring the 3102– 2 – 500 triangle a gazillion more times, I absolutely will.
if I can settle the issue by going back out and remeasuring the 3102– 2 – 500 triangle a gazillion more times, I absolutely will.
IMO, that would not settle the issue with the already introduced level of redundancy you state. You would just confirm you do not not have a mistake in your measurements - which you have probably confirmed by already having multiple observations and checking your observations through StarNet.
Since he used RTK methods, I expect your surveyor noted relative positional accuracy on the drawing or whatever means he communicated the results to you. That should mean he also had redundancy in measurements and reducing that data to be able to support his results. I start from the position of giving him the benefit of the doubt and accepting the premise that he conducted the survey in a manner to achieve whatever the required standards are.
Since there was only a few hundredths difference in the other three lines, the 0.28' difference is certainly something to figure out.
But, if both of you have performed the work correctly and indicate you have confidence in your answers based on proper analysis, then you both should continue to get the same answer (within small amounts) that each of you came to with any additional redundant measurements.
Potentially:
1) As has been suggested earlier, the RTK measurement might be bad. If so, it should have been caught by redundant measurements as well as then processing the data to be able to report the results of the survey. Again, my hope is that the surveyor had redundancy and reduced his results in some manner to support his work. So, I hope this option is not the case.
2) You have both done a good job of measuring accurately. If so, then that leaves the likelihood that you are not measuring to the same point.
Looking at the sketch Landbutcher put together, the three other sides are so close to matching each other. They are within what I would expect between a traverse and a RTK survey.
What are the conditions of the monuments at each end of the 239' line? Are they sticking up and plumb or are they bent and leaning?
0.05' split is a bit much. Every piece of your equipment can have a little error and shooting from both ends is a good procedure to reduce errors. I would just average those distances for that 239 foot leg. If you have unadjusted traverse angles at the 4 corners (calc'd from traverse or directly measured) you could adjust your 4 sided traverse using a "Lats & Deps" adjustment. OR (if you love a math headache) you could put all the field data into one big traverse and do a "Lats & Deps" adjustment.
Here is an excel file of a "Lats & Deps" adjustment I did for a student on another chat. You need a sketch showing all your angles (left or right) and horizontal distances. Then pick a line for a Basis Of Bearings and calc all the bearings around your traverse (it does not have to close, the file will close the angles and adjust the traverse). Then enter those bearings and distances into the file.
The cells with formulas are locked but you can view them by clicking on them. You can also unlock the file but I recommend you do a save-as with a new name first (like change to 01, 02 etc). If you need more than 10 legs I can make it bigger or if you are good with excel feel free to experiment.
Also any additions or corrections or improvements are welcome.
Here is an excel file of a "Lats & Deps" adjustment I did for a student on another chat. You need a sketch showing all your angles (left or right) and horizontal distances. Then pick a line for a Basis Of Bearings and calc all the bearings around your traverse (it does not have to close, the file will close the angles and adjust the traverse). Then enter those bearings and distances into the file.
Also any additions or corrections or improvements are welcome.
-- attachment is not available --
This is very cool. Thank you Sir. Read about Lats and Deps but never done one. I'll take it for a spin.
But the unanswered question is: If further analysis confirms that the RTK position of 3102 is in fact off by .28, should I mention it to the surveyor? Should I accept a survey that is known to be "inaccurate"? I mean, if he stamps it, and we record it no one on the earth (except maybe some future surveyor looking to create a pin-cushion) will ever know about it let alone give a rip.
Not considering mentioning it to him but thought I'd ask the sage professionals here the question.
I've compared my surveyors elevation of the base station to mine and adjusted all four corners (in addition to the base point) accordingly.
I'm left with deltas for the four corners: .46', .48', -.30' and -.64'
How accurate is GPS/GNSS equipment (assuming no post processing)? Curiously, if I average the four numbers above, I come out with 0', but that could just be total coincidence.
My plan is to adjust both position and elevation of my entire survey pre-Starnet adjusted points according to the surveyor's base point, then put them back through Starnet.
+0.46' to -0.64' is huge and I am going to assume that that error is between the surveyor elevations and your traverse elevations.
1.) Did you double (face1-face2) your vertical angles and correct before calc'ing the delta H?
2.) Did you close the delta H's back to where you started?
3.) Just like a normal level circuit, the sum of the +H's must equal the sum of the -H's. In your case 0.2' to 0.3' would be fine. Then you take (the error divided by the total distance of your traverse) and correct each leg's delta H based on the distance of that leg. Make sure you correct the H's so that the new sums will come out equal or you did it backwards.
ALSO: I found an error in my Lats & Deps excel file so here it is corrected. Delete the first one. There are several items that need manual entry, they are highlighted in yellow. You do not need to use all 13 legs but you do need to clear all cells that you are not going to use.
+0.46' to -0.64' is huge and I am going to assume that that error is between the surveyor elevations and your traverse elevations.
1.) Did you double (face1-face2) your vertical angles and correct before calc'ing the delta H?
2.) Did you close the delta H's back to where you started?
-- attachment is not available --
Yes, the difference between the surveyor's elevations and mine.
Yes, from control point to control point, I did multiple (usually three) pairs of face1 and face2, and did the same for backsights from the next points. But the answer to the second question is harder to answer. The 24 acres is mostly wooded; I did most of the traverses in the summer times when the visibility through the woods was horrible. I created almost 50 points, each visible by at least two others, to get the 2000' between the furthest points on the property. At the time, I wasn't about to chainsaw avenues through the woods.
In the very earliest days, I did not appreciate the importance of elevation and probably screwed up several HI's or Target Heights. I tried weeding most of the errant observations out in Starnet, but looking back on it now, I don't trust my work from those early days. I'm fine with doing it all over again, but now, with 6" of snow on the ground, It'd be a pain to even find the points (marked with 12" nails) but I don't have a Shonstedt lol.
I could however re-do some traverses now, and get from the farthest points apart in perhaps a dozen or so setups; I'd do traverses up trails I've built since the early days. It's very steep in some places, but with a chain saw and some patience, I think it'd be a fun project.
That said, when the surveyor recorded the corners, the canopy was still very much in place; that's why I asked about GPS accuracy for elevation.
Or I could inherit some GPS equipment and get into the 21st century of surveying technology and start my learning all over!
This is an elevation only question:How do you calculate the difference in elevation between ground and grid for any specific location in your survey?
I ask as I'm now looking at the discrepancies between my elevations and those determined by my GPS surveyor. I may have errantly mixed ground and grid along the way for some observations. Could this be a factor?
@rfc "That said, when the surveyor recorded the corners, the canopy was still very much in place; that's why I asked about GPS accuracy for elevation."
That is where a complete raw file might help. The number of satellites and the PDOP should be in the raw file. Those numbers tell you the potential accuracy of each shot. They will give you a way to weight each shot. More sats and a lower PDOP indicate a potentially (not always true) more accurate shot. Vertical accuracy suffers first and the most with PDOP less than 2.5 and at 4.0 (the lowest I use in trees) can be off 0.5'+ vertically but only 0.2' off horizontally when checked with my robot.
@rfc "I ask as I'm now looking at the discrepancies between my elevations and those determined by my GPS surveyor. I may have errantly mixed ground and grid along the way for some observations. Could this be a factor?"
If you had your EDM scale factor set on 1.0 then everything you measured was ground distance and elevation.
