Accuracy of tribrach clamping

There are other sources of errors in the procedure. I am not sure that results expressed accuracy of tribrach clamping.

I've done a lot of highway work in metric, I've never heard or used any surveying instrument that is that accurate, looks more like numbers for machine work. Your stated numbers basically have no error.

Unless I'm reading them incorrectly do you actually mean 0.05mm? Or do you really mean 5mm?

> From discussions in the latest TS accuracy thread, I have tried to quantify the accuracy/consistency of the instrument position when re-clamped back in a tribrach after being removed and reclamped. I tested both keeping lug positions the same and also changing to the other 2 lug positions. The test was done indoors on a concrete slab floor. The instrument is a 5" leica TCA1105, GDF22 design Leica tribrach and leica budget wooden tripod. I had targets placed about 2.8 m distant at 0-90-180. Results are from an angle-only resection.
>
> The first station was assigned coordinates of 0,0 mm. The first three are remove and re-clamp in same lug position. The next two are the other lug positions. The station results are in mm.
>
> COORDINATE..... 11..... 0.0000..... 0.0000
> COORDINATE..... 12..... 0.0072..... 0.0204
> COORDINATE..... 13..... 0.0072..... 0.0204
> COORDINATE..... 21..... 0.0207..... 0.0070
> COORDINATE..... 31..... 0.0346..... 0.0207
>
> Note the 2nd and 3rd stations are identical coordinates. This is because the 5" instrument turned identical angles. SD for coordinates is reported as about 0.05mm E and 0.08mm N with an input standard deviation of 5" for directions. The angles came out better than that though.

It's interesting that those values fall somewhere in between the value of 0.006mm reported by P. Berton Jones in "Survey Review" and the value greater than 0.1mm that Jim Frame mentioned. I'd suspect that the accuracy of replacement probably also varies with the age and condition of the tribrach.

What was the condition for which 31 was derived?

> I've done a lot of highway work in metric, I've never heard or used any surveying instrument that is that accurate, looks more like numbers for machine work. Your stated numbers basically have no error.

What Conrad is testing is the centering accuracy of simply removing and replacing a fixture (in this case a total station) in a Leica tribrach, both with the lugs in the same holes on the tribrach and in different holes.

For your test the angular accuracy at that distance is negligible. You can quantify it mathematically as a function of 5" over a distance of 2.8m and you end up with .0000677m, but the instrument you are using has an EDM accuracy spec of around 1mm.

A quick google search showed the TCA1105 as being accurate to 2mm, but even if you go with the 1mm value it isn't accurate enough to perform your test.

You state that you are using an angles-only resection, but you had to measure a distance somewhere at some point when deriving the values for your targets. Also, your tripod seems to be the weaker component than your tribrach.

> A quick google search showed the TCA1105 as being accurate to 2mm, but even if you go with the 1mm value it isn't accurate enough to perform your test.

Note, though, that the test is to check the *relative* positions of subsequent replacements of the instrument in the fixed tribrach. That solution isn't highly sensitive to errors of that magnitude in the distances, at all.

> Results are from an angle-only resection.

That wouldn't be my choice of methods to test this; it's too indirect. I'd set up 2 tripods (pillars would be better!), one with the tribrach and one with a decent theodolite. I'd mount a small scale with appropriate graduations on the tribrach insert and directly observe the graduations from the theodolite at each insertion of the insert.

> > Results are from an angle-only resection.
>
> That wouldn't be my choice of methods to test this; it's too indirect. I'd set up 2 tripods (pillars would be better!), one with the tribrach and one with a decent theodolite. I'd mount a small scale with appropriate graduations on the tribrach insert and directly observe the graduations from the theodolite at each insertion of the insert.

For measuring displacements below 0.1mm, though, the problem will be finding a scale with graduations that fine or you'll be back to measuring angles to interpolate the scale interval, which would be no better than the resection solution.

Where the resection solution could be improved is by redundancy, as in targets at 0-90-180-270. The geometry of that arrangement means that solution of relative coordinates of successive replacements is quite insensitive to small errors in the distances to the targets.

Actually, you are correct.

But I would still contend that the tripod is introducing enough error as to render error introduced from the tribrach negligible, given that it isn't a bargain basement model.

> But I would still contend that the tripod is introducing enough error as to negate error introduced from the tribrach, given that it isn't a bargain basement model.

That's easy enough to test as well. If the joints are properly tightened as Leica recommends and the legs anchored for the indoors setup, I wouldn't expect that much play.

Agreed, but to what magnitude would you quantify "that much"? The OP is kicking around values of 0.06mm?!?!?

Yes I understood what he is doing

No offense, but ....

this is somewhat a ridiculous discussion. The test should have concluded with a statement to the effect that the differences are negligible - end of story.

His worst case (if I have done my math correctly) difference is 0.0015" which is beyond what the human eye can discern, beyond what that instrument can measure, beyond what the average survey crew set-up capability is, and way beyond anything that matters to a surveyor.

In fact, this error is beyond the leveling capability of the level bubble on the average survey instrument - even assuming it is in adjustment.

Now it might be important in setting up the nuclear collidor in Switzerland, but beyond that it doesn't matter.

Try it on a Monument like this

A monument like this would remove all the tripod error from your calculations.

Here is a picture of the monument looking to the West. The monument is set 4 feet deep in concrete with a 14' heliz below it. I think this is a stable monument.

The GPS has been running for 9 days. We changed the battery yesterday and installed a trickle charger. We are going to run it for a month and compare the OPUS solutions.

Here is a view looking south. The 1 meter extension gets the antenna well above the fence.

Here is a Google Earth Image of the site. The red line shows the path of the extension cord that we are using to power the charger. All of the fill has been in this area for over 40 years, and the Helix dontinuse down into natural ground.

:beer: :beer: :beer: :beer: :beer: :beer: :beer: :beer: :beer: :beer: :beer: :beer:

yeah- if you strap GPS on a 16G, it'll cut subgrade to 0.05mm.........o.O

2.8 m =~ 10'

20" =~ 0.01'/100'
20" =~ 0.001'/10'

Get a bigger room with targets set 28 m away.

I don't think you will get meaningful results testing that way, the measurement component throws a wrench in the data by adding errors on its own, how can you differentiate one from the other? I can't. "That's the system I use, that's the expected error that comes with it".

You'd have to find a way to measure how the instrument is sitting time after time when clamping into the tribrach. I am thinking some kind of micrometers, or some devices designed to measure the infinitively small movements. Shooting targets I don't think will achieve that.

Back in my youth, I worked in a plant building portable propane gas plants, one of the buildings had a machine shop, which fabricated parts. Those guys had tools to measure 0.02mm, but surveying equipment...........oh well

Thanks for the good number of responses gents; I appreciate the replies.

I neglected to do what a good scientist should do and post all the numbers and methodology. I'll give you my raw numbers and assumptions when I get more time. It'll give you a chance to play with the numbers and point out where the solution could mislead. I wouldn't necessarily rely on intuition alone as it doesn't always give good results.

I'll be happy to have any misunderstandings I have about the method corrected. If I learn something then I'll be happy.

The EDM does determine the initial coordinates of the targets but this geometry should be insensitive to this error. You can push the coordinates away or toward the instrument several mm but you get about the same relative answers.