Robotic sighting

It really depends on your tolerances...I normally use the 360 prism, anything with a tighter tolerance I end up using a regular -30mm prism.

With our Leica, it's pretty dialed in. We have used the 360 for control and never had issues when we set a round prism up over it later. Just make sure you have your constants right.

One of my best guys and friends used to tell me that no robot could sight better than him, he used to say he had an eagle eye! It was a good back and forth between us until he used the auto-aim on a semi-long traverse (12ish turns) and it closed at 1:500,000 unadjusted. He hasn't tried to eye a target since...

Tom

I have had the opposite occur using a Trimble 5603.
For an experiment in traversing, we ran a four point loop behind the office using the robot. We used tribrachs and RMT 360 prisms on back and foresights. Each leg was around 300' +/-. We had the robot turn two sets at each station. The instrument had recently been calibrated. Resulting closure was under 1:18,000 before balancing the angles. It came up after balancing but not much. We were quite disappointed.

Later that same day I ran through the same points by myself. Using a Nikon DTM-520, the same tripods and tribrachs, but regular -30 prisms, I closed 1:250,000.

I reviewed the DC file for the 5603, and noticed that on the reverse face, it was reporting angles ranging from 180-00-15 to 180-00-50, yet the final mean turned angle was listed as closing within tolerance. I think we had it set for 10 seconds. It felt like the robot was cheating.

Anyway, we never did use the robots for traversing, only for topo and location work. I never did figure out the real problem. Now that you guys mention it, results more than likely would have been better using the Trimble traversing prisms, but management would never have paid for a set.

Instruments all appear to "take" the best signal they get back, so with a 360 degree prism where there will be more than one reference visible, the resultant pointing can vary slightly, a few seconds either way, depending upon how the prism is orientated relative to the instrument. For really accurate traversing the 360 prism is probably not the best tool.

Going back to the early days of the GDM4400, I carried out a number of experiments with regard to horizontal deflection of the sightline. Any object close (say within an inch or so) to the sightline can result in a horizontal deflection. I first noticed it with when a stand of waving grasses just clear of the sightline made the coordinate change as the wind blew - hence the investigation. Subsequently it showed up when observing alignments of machinery when the line of sight passed very close to equipment. It can also happen (often) with wire mesh fencing on line.

This is caused by the "Best Signal" syndrome. It isn't the same as the horizontal refraction that can occur against tunnel walls (or objects which have been soaking up heat (ie. parked vehicles).

I would sugest that it it good policy to ensure that traverse sightlines are at least a foot off any object (and double that vertically above ground surface).

> The instrument had recently been calibrated.

The first thing I do when I get an instrument back from calibration is perform the onboard collimation routines. When I get my instrument back, all of these values are set back to zero resulting in the wrong corrections being applied.

Also with the trimble, enter the temp and pressure on the faceplate only, not in the DC. In other words, enter it in one place only.

I had some very wierd results in my 5601 once. Turns out on the temp & pressure screen on the DC, the PPM was set at zero, not NULL. This will really screw with your measurments.

For what it's worth, I always noticed more vertical error in F1/F2 measurements than I expected. If I ran a traverse in F1 only, I had great closures. Theoretically, if the gun is calibrated and collimated, the F1 measurements should be as good as turning sets.

Now I run an S8 and I'm just amazed the accuracy and precision of it.

Ck it on something you have already closed. Or do two sets of measurements, one robotic and one manually. It takes a little time before you get comfortable with it. They work. Jp

The 360 is fine for general stake out. For things that have to be right, I recommend using a good quality mini prism and a short layout pole.

As a very good layout surveyor once told me "you have to get down on the ground"

Thanks for all the feedback.
I have just got the instrument and hadn't had time to run any tests.
This morning I set up the Trimble 360 prism on a tribrach with a friction puck. After the initial shot I rotated with prism with the gun on Autolock, the instrument followed the prism 1' off of the original pointing. Over the 22m(12ft)line of the sight that makes for 6mm(1/4") of error in the angle.
Questioned the Trimble sales rep about this and he said not to worry that the instrument and prism are ideally suited for anchor bolts.
I think I will will stick with my tried and true "peanut" for any structural layout and "get down on the ground" as Tom suggested.

Do the fancy Leica style (big chunk of glass) prisms have the same offset as they are turned?

It seems there are a number of unnecessary potential errors built into the systems that are just accepted in the hope they are small enough. Seems everyone would want to be using the nodal prisms for stakeout.

I use an A7 360 prism and have had no trouble traversing with it..What I have discovered however, is when rotating the rod, the pointing will deviate slightly from center..so..I use the following technique to ensure best results...find and point the prism class in "near perfect" alignment to the instrument and the center pointing becomes quit perfect.

I first discovered this "non-perfect-pointing" of 360 prisms some time ago when we used older 360 "circular" prisms (not of the inverted triangular type) in conventional 'tracking' mode and while observing the rodman run the rod, I could see that the prism was not pointed directly at me resulting in 0.04'+/- (diameter of the rod) and when rotating the rod, pointing alignment became centered. Having experienced this..I observed the similar using the A7 360.

Another test one can perform is to stake to a known point with the rod in various rotational positions and observe the "left-of / Right-of"...

> I use an A7 360 prism and have had no trouble traversing with it..What I have discovered however, is when rotating the rod, the pointing will deviate slightly from center..so..I use the following technique to ensure best results...find and point the prism class in "near perfect" alignment to the instrument and the center pointing becomes quite perfect.

Ditto.

I just took that as common sense. I traverse with my Topcon 360° and have no discernible issues. I do the same thing when shooting property corners too. I keep my rod in good plumb and check my instrument on a known distance baseline at lease every 2 weeks too.

Agreed.....A point I was attempting here is..looking through the gun, watching the rod being rotated, the crosshairs would move off center of the rod to its' edge (0.04'+/-), thereby recording the angle at that time of the shot..resulting in the point being inaccurate (for the traverse point).. therefore..for best results, one must find the correct position (rotational direction) of the glass pointing to the gun to minimize the inherent inaccuracies of 360 prisms...The A7 is much less, however.

If I follow correctly the prism tracking is not concurrent with optical sighting, meaning that the gun can have a solid lock on the prism but not look right through the crosshairs.

There is this odd learning curve with robotics where you end up having to unlearn some fundamentals of conventional total station surveying.

When I was trained it was conveyed to me to not leave the gun unattended, however once getting into robotics you leave the most expensive gun unattended and jacked up higher than a human HI in order to get more out of your set ups.

> If I follow correctly the prism tracking is not concurrent with optical sighting, meaning that the gun can have a solid lock on the prism but not look right through the crosshairs.

You are correct, my Topcon is that way, but it can be adjusted by the user. If I am doing a lot of traversing where I might stake backsights, I adjust the tracking to match the optical. Most of the time I just use a prism with the robot tracking on it when I zero the backsight, then the slight difference between robot and human are not part of the equation.