If you had to make a list of suggestions for anyone setting up a high accuracy control network, what would they be?
I'll go first... (I'm sure I'm missing something glaringly obvious)
1. Observe all observations on minimum f1/f2. Run rounds when time permits or accuracy dictates.
2. Do not use 360 degree prisms and use nodal prisms if you can.
3. Calibrate your total station often.
4. Check and adjust rod and tribrach level bubbles often.
5. Use control points with a well define center point.
6. Make sure your instruments tripod is of high quality and all fasteners are tight.
7. Adjust for temperature and pressure.
8. Always stabilize your backsight & foresight.
9. Check your level and 0 your backsight often.
10. Use common points and adjust you data with a LSA.
That's pretty well covered it, I'd add:
1. Also perform a level traverse to improve vertical.
2. Minimise distortions in the network when chosing datums by minimising constraints. If using GPS, use long duration static in positions around the perimeter of the control network, process it through OPUS, AUSPOS etc, include the covariance matrices from SINEX files in the LSA
3. We process the total station data using a script to add horizontal angles between all observed prisms (not just the backsight), it helps tighten horizontal uncertainty.
1. Determine the length of time the control will be used then multiply this number by 1.5 or 2.
2. Based on your answer to #1, determine the minimum number of control points needed then multiply this number by 1.25.
I would always "measure rounds" as Trimble calls the routine. Once you are already set up the turning of additional rounds adds almost no time to the observations. The "std dev" provided in this routine is an in field blunder check against a partially obstructed prism (pedestrians/traffic) as well as other issues.
I have gotten into the habit of what I have termed "double traversing" in my ongoing quest for the MPV (most probable value). When establishing primary control by conventional methods I will turn my rounds then rotate and re-centre the tribrachs for all prisms and the station and repeat the observation (if employing a bipod and pole I will rotate the pole 180). I have found this added redundancy improves the quality of the survey without adding that much time, as I am already there with the gear anyway.
Great points!
Don't neglect establishing stable control monuments. The tightest network in the world is useless if the monuments move...
Add curvature and refraction corrections . Also add cross ties and good geometric figure to the primary control set up even if this creates a few more points. Than needed for working control. Balance the control. Both distances and angular see geometric figure of the control network.
Aren't C&R corrections applied in almost all modern total stations?
Great point about cross ties. Cant believe I forgot that one.
So when we're not adjusting our data its very obvious to me why we cannot have a short backsight and long foresight. But if we have a control network being adjusted with a LSA, and correct me if I'm wrong here, but since the order of observation doesn't matter, does it not eliminate this issue? And if so, do we still need to balance out our legs? Is there a reduction in overall accuracy if for example we have an equidistant hexagonal traverse but one side is broken up into 3 smaller legs?
Another fantastic point about setting control in stable ground.
"Don’t neglect establishing stable control monuments. The tightest network in the world is useless if the monuments move…"
Yeah verily. If I may, add durable. Projects may be 5 years or longer between initial survey and construction completion.
I have gotten into the habit of what I have termed “double traversing” in my ongoing quest for the MPV (most probable value). When establishing primary control by conventional methods I will turn my rounds then rotate and re-centre the tribrachs for all prisms and the station and repeat the observation (if employing a bipod and pole I will rotate the pole 180). I have found this added redundancy improves the quality of the survey without adding that much time, as I am already there with the gear anyway.
+1000
It blows my mind that so few surveyors use this method. For a small additional investment in time, you can double your redundancy and check every single setup before moving up.
In your opinion do you think field checking/adjusting rod & tribrach bubbles and plummets would minimize the benefit of this method to a negligible level?
I do love the technique but it to be fair, it does add a lot of time.
And time = more observations = reduced impact from random errors.
In the time it takes me to employ this method I could run many more rounds plus probably add extra geometry via common cross ties or common point observations. Now if I had a bubble or plummet out of adjustment I'm 100% going to wish I had done this. But if I checked all my gear prior to starting, the cost of the extra time added may have had greater benefit elsewhere.
Well only if you tell it to apply them. I always applied them but many people do not. They say it doesn’t matter as they only measure short distances. I have stopped banging my head against the wall on that one. Some think C//R only affects distance some vertical so I stopped asking and just role with it.
On the short legs. To be honest this depends. If short bs and you are long fs a set up or two with no cross ties back to that set it can bite you. What I do when at times you have no choice is to re set up bs my long fs and wrap angles to the short leg close the horizon is what we use to call this. Then the means can be coed from both to see if it’s out less or more than 360. I try my best to design a network on sites I pre designed for my crews in google earth then they started getting it on there own. I like control to be usable but also a good network around a site that can’t be disturbed that allows me to recreate anything inside without causing to much additional error . I work inside out outside in. I have been to to many jobs where a hole quadrant of control is wiped out and then they have to traverse from one side to the other and wonder why things don’t match. Now this is for jobs that this is needed of course. Not all jobs require control like this because once you are done that’s it. Mostly construction sites that we know might happen and maybe starting off just boundary and Topo Well I will do the work on the front end and when design layout comes we would simply skate along as we had good tight control that we did on the front end. No re doing traverse and making adjustments etc.
That word redundancy is key. That is a word not used much now days. It goes for conventional traverse to static to rtk etc. When even doing an opus while onward is going on. The crew might be logging data all day. I make them at minimum break the set up at 4hrs re plumb the base and level at minimum. Guess what many times I get two 4hr static files. That’s meeting Gps derived height guidance. The next day they set base on a different point. What I will say Trimble made that easy. On all fronts. I am learning a new system now. Have my work cut out for me.
