I recently did some testing to see what degree of mis levelling would cause angular error of a total station when using a compensator and when the compensator is turned off...
Very well done, thank you
I ran a similar test - not as extensively or as systematically - some years ago with a Trimble 5603 and got similar results. Up to about 3 minutes out of level the angles were consistent.
I also reached the same conclusion about centering errors. Which makes resected setups all the more attractive to me.
Another well done video.
I just have to say. Awesome job. We did similar test no distance with Wild T-3. Oh the days without single or dual axis compensators.
Just have to ask have you compared different manufacturers of the same accuracy. Wondered if different manufacturers handle the compensation differently.
Norman, I would love to sit down and do the error propagation in various scenarios to see when resections can provide better results and when setting up and back sighting provides a better quality result. Im sure it could be done in Preanalyis in STAR*NET
OleManRiver, I'm sold on compensators now.
Unfortunately I didn't have numerous instruments with me when I was at that calibration site. And we only really run Trimble right now. It would be interesting to try with other instruments. If I had to venture a guess I would imagine one would get similar results.
Which makes resected setups all the more attractive to me.
Resections can be dangerous in the wrong hands, but I love to use them.
I would suggest the resection is not dangerous at all but the lack of checking the results of the resection is where the danger comes in. You can always detect a bad resection result if you are willing to check.
We just got our Topcon GLS scanner back fresh from calibration, I levelled it, rotated it 90 degrees and it was out over 38 arcseconds, the calibration routine requires the instrument to be level, within 30 arcseconds, I levelled it again to within 30 arcseconds and ran the calibration routine, I levelled and rotated it 90 degrees again, this time it was 8 arcseconds out, so I ran the calibration routine a second time, levelled it again (zero arcseconds x and y), then rotated it through 90, 180, 270 degrees and back to zero, with a brief pause to let it settle in each position and it remained perfectly level throughout. I had expected the level compensator would have been calibrated too, it has been transported from Japan to Australia since calibration, so it could be temperature and pressure changes during the flight home, or maybe a bumpy ride. This scanner uses prisms for targets, and we include its observations in least squares adjustments with GPS total station and level traverse, so we will know next field trip exactly how well calibrated it is.
In the past we had assumed instruments returned fresh from calibration didn't need the level calibration routine performed and we had a whole heap of vertical errors in the 20 - 40 arcsecond range, but X-Y was good, which we put down to vertical refraction at the time. I think your analysis better explains the errors we were experiencing on vertical and how vertical is more affected than horizontal.
Great work.
Thank you for digging into these questions that a lot of us have.
I feel I was taught quality practices, by a phenomenal surveyor, when I was starting out, but I often find myself wondering how much of a difference some of these methodologies make.
As always I look forward to more of your content.
You said a mouth full. I love them myself. Resect is awesome when used appropriately.
Long ago (1993 to be specific) the surveyor I was working for demanded that we set up the tripod with tribrach over the point, set the legs, and level up and center before taking the instrument out of the box and setting it in the prepared tribrach. We would then fine adjust the instrument and go to work. Which is the procedure I follow to this day. This research shows that fine, final adjustment is unnecessary - in an instrument without integral optic plumb.
Im confirming your findings. I was staking out anchor bolts yesterday and an excavator set up to dig with its tracks 10 feet away from the instrument ( Geomax Zoom95). I checked the level and I've seen a 1' deviation one way and 45" the other way. I continued working until the excavator left, then I releveled the instrument and shot the backsight and I checked all the points staked while the instrument wasn't level. I got the same results ( within 1mm). I was worried that my whole setup was pushed away by the excavator, but shooting the backsight and checking another control point proved me wrong.
That’s exactly how I was taught originally. Heck I was not even allowed to touch the instrument until I could set up over a point with a plumb bob and level with fish eye aka circular then they added the prism adapter with spirit level. I then learned to fine level turn 90 adjust and spin 180 take half the error back to it until I got that down usually on a lunch break or early morning then they gave me the wild T-1A. Got that down. Then optical plummet was introduced. Early in that stage they sent me home with fixed set of legs tribrach plumb bob and I practiced on my own.
Was it the sokkia instruments total stations like the old set 2 and 3B II that had it in the manual to turn off the compensator on windy days or if equipment was vibrating it. Today’s instruments are amazing compared to us having to pay attention to a Level vile and understanding how the sun made the bubble run over time.
Fantastic work and greatly appreciated. Cheers!
