I always turn rounds. To control and property corners. I also will do direct and reverse when tagging a Finish Floor I know a lot of people do not. Some say it’s overkill. But I also don’t have crews saying hey this control is out what shall we do. The number of rounds is dependent on project scope and requirements. But I set the bare minimum to 4. I also always request even number of rounds. 2 4 6 8 10 etc. some do 3 at our shop. But I like even numbers and is stuck in my head from doing geodetic work of 16 a lot. Of course these were all manually done.
"as traverses get more than a half dozen or so points, RTK and/or static to both connect the traverse to the real world, and to tighten things up."
Agree on the first do not on the second assuming you are talking about a closed traverse. The day RTK tightens up my 6 station traverse is the day I retire.
a six station traverse that doesnt close back on itself yes, 30 stations over 400 acres maybe
If you play around with the snaps. Also think of it this way corner to corner you can Cj that distance. You can create a point at whatever distance along that line a few ways. Coho or stake to the line go distance then you can coho points say for an added bay window from that line and points. I mean you can lay out the bedroom or kitchen if you would like. It’s all math and you seem to have a handle on that so you are going to be fine. Never be afraid to fail. Just have a back up plan until you get it down aka your pm pre comped points etc. Once you use it a while if you have not already you can add favorites program hot keys all sorts of stuff that make not having to navigate all the menus for your favorite tools. And functions. I am sure every brand has. This sorta thing. I don’t get to do that fun stuff as I am usually if in the field using someone else’s equipment and don’t want to mess there little world up. I actually got my crew chief back and so he helped me today or I helped him. He is a great young buck and works hard . I was purposely moving his hammer around today so he kept coming behind me and putting it back to where he keeps it. I finally told him look. This is your truck and your equipment. If I am not doing something right let me know. So tell me to put things back where I got them. He was like your the boss I said not today you are. He will be a great addition to the profession he is going to be an all star. I just wish I was his age again lol.
https://photos.app.goo.gl/C1AyLFfDHmmzMeyTA
Today I had a perfect example of a DXF not matching, so I looked a little closer. This one was off more than usual, often they're just slightly shifted, but still the same shape. This one some jogs were even missing. It turns out the boundary matches the points just perfectly, it's just the houses that aren't the same. I can tell by looking at it, this is what the builder sent us in their CAD file. It was likely a preliminary drawing and then the house layout and positioning obviously changed after that, but was never updated in the drawing. We only layout outside corner of wall, and wouldn't have calcs for the inside of the foundation wall, like their DXF shows. One exception to the outside corner layout, (for this builder only) we also layout a 45 degree angle on the inside of the garage, which you can see in the one picture, I have calcs and set spikes there, but they weren't in the DXF at all.
https://photos.app.goo.gl/fndkpTNp4Uxw3AX38
It also looks like they have multiple lines stacked on top of each other. Why, I don't know. But I'm sure it's part of the reason that our DXF's end up slowing down the data collector and creating a lot of lag in the interface. I will usually go through and deselect a bunch of the layers in that DXF, at least the obvious ones that I know aren't relevant. There are something like 50 of them in there usually and if I go to far, my boundary or building lines will disappear and then I have to go back in and toggle some back on, even though I didn't turn off any that sounded like boundary or buildings.
https://photos.app.goo.gl/6jZroAvaTPMWRhoe8
This is my set of hot keys which I use all the time. It drives me nuts when I use the other guys DC and I have to go through all the menus on the screen to get to something, especially the joystick or point manager.
As far as layout random points for clients on the fly, I would get a spanking if I did that. Sometimes clients send foundation plans that don't even close, and my PM catches it when he calcs and checks everything, then he sends an email back to them to fix it. I would have to be 100% confident in what I was laying out before going ahead. Other than the pre calced stuff that my PM has given me (which I am also checking in the field), I could layout extra points on line, for example a walk out basement, with large steps in the foundation, the concrete guys will often ask me for extra points on line when doing nails on footings. Or I will just do it without being asked if I feel like making their life easier.
Or if a corner wasn't excavated far enough, I can give an offset if I trust the client not to mess it up. Grades I can give out, once again only to clients we trust. With liability worries though, I can't do much more than these things without checking with the office.
Agree on the first do not on the second assuming you are talking about a closed traverse. The day RTK tightens up my 6 station traverse is the day I retire.
a six station traverse that doesnt close back on itself yes, 30 stations over 400 acres maybe
I mean, I can count on one hand the number of traverses that close on themselves, or have no cross-ties, that I have seen in the past 5 years.
Just for kicks I looked at a couple of jobs I've processed in the past week or two. A six-station link traverse with four 2-3 minute RTK observations scattered throughout improves the average horizontal error ellipse (95% confidence) by about 0.015' and had no appreciable improvement in the vertical. Not much better, but certainly doesn't make it worse, and ties it to the NSRS, which is an additional benefit for not much effort.
Now, I also have a ~3500' traverse with about a dozen points. When I include about eight to ten 2-3 minute RTK observations throughout those points, the average horizontal ellipse goes from 0.19' to 0.07', and the vertical goes from 0.06' to 0.04'. A pretty nice improvement for only another hour's work with a rover - brings those relative positional precision values down to within statutorily mandated tolerances for boundary work.
Using the RTK data and holding the base station fixed, as opposed to hard-holding the beginning and ending point pairs, brings my reference factor from 1.32 down to 0.96, and now passes the chi-squared test at 95%.
I should note that this is only network RTK, not base-rover, observations out on the edge of the subnet.
Mixing and matching observations is rarely a bad thing when properly weighted.
Agree on the first do not on the second assuming you are talking about a closed traverse. The day RTK tightens up my 6 station traverse is the day I retire.
The RTK doesn't tighten up the traverse. Just the opposite. I hit the points with RTK to connect to the datum, then traverse to tighten things up - the elevations, principally - then simultaneously adjust.
When I check collimating I might change a second or two, a 1 second instrument. On new control points and property corners on parcel take projects I make sure 0 is good on the backsight, observe both faces and 2 RTK observations using integrated surveying while turning my rod 180 between each observation. I’m an engineering road/utility surveyor, so little need to run traverses
Yes yes yes. The difference between accuracy and precision is also in this scenario. Accuracy to the datum. If you set two points in the same scenario with rtk network vrs. And close on yourself very well. A loop. That doesn’t mean your outer points from that initial start point pair are tied to bars as accurately as your beginning. They could actually be off more than the uncertainty in your closed loop precision. Why because of a rotation. I am seeing even better results like you described using base rover . I did a test this past week as I will be teaching someone least squares. So way over kill but I did vrs. To a small site 4 points. Then set base up shot all 4 points again twice as same with vrs. Then traversed through all points and had two extra points within that traverse. I also did some cross ties and also some closed the horizon on those as well so bs my fs etc. anyway I purposely on rtk shots only I know which ones leaned the rod out or moved the Tribrach off the point. Now I have and know the good set ups because I need to introduce error to teach how to trouble shoot. Don’t worry I had two days to do a 1 day project for the bean counters. Yes I actually gave some wrong HI and target heights. Again on purpose. The good data shows exactly what you are seeing. Reducing the error eclipses through redundancy. Base on two different points multiple observations different times of day and all . All the training I see from manufacturers are so canned that it doesn’t help someone solve realistic problems. So I invented my own issues. Hopefully this will help me teach better to someone so they don’t just learn what buttons to push or click. But to analyze and remove blunders and perform a better quality survey adjustment. Also how to spot sloppy work or a tribrach or rod out of adjustment as well. What I am amazed at is that the level of precision relative and the accuracy by incorporating multiple measurements and tools and just how good rtk has actually become if good field practice or procedures are done. I honestly think if done correctly from some of my experiments are better than manufacturers specs. I do need a few more scenarios for good level data running through all control but I think I could be confident in .03 ft vertical differences in height with rtk alone on sites under a certain size from my results so far. Now I do not have enough samples yet but am working on it when the budget allows. But it takes discipline and a good field person to pay attention and think when doing observations. I have about 8 so far with very good digital levels and rtk alone base rover holding one leveled vertical point appropriately placed under a maximum .04. Ft was the worst outlier. Most were better but it takes a lot of samples to know for sure. Of course in different areas. Then I would have the issue of size area this could be achieved to see. Now same data sets with some robotic work gets even better when incorporated into a final adjustment. I think you are spot on. For sure