Oh Great!
It took a long time to convince myself to try the experiment. It has been on the list for a couple of years and I always doubted it would be worth the effort. Turns out it works fine, and no doubt I will get better with the plumb bob with practice.
Oh Great!
> It took a long time to convince myself to try the experiment. It has been on the list for a couple of years and I always doubted it would be worth the effort. Turns out it works fine, and no doubt I will get better with the plumb bob with practice.
I'm having some difficulty calling a target centering error of +/-0.02 ft. s.e. "working fine". One obvious test is whether the relative uncertainties of points positioned by that method meet the ALTA specifications at 95% confidence. The other thing I'd point out is that if you work to better than the minimum standards, your work survives longer. Otherwise, five years from now when it no longer meets the new minimum standards, you're left to tell a client that somehow it all turned into a pumpkin at midnight.
I Doubt Those Are D&R Values
I have shot a lot of prism on bob and 0.02' in any direction is just a start of the probable error.
I have never shot prism on bob on my own projects. Shots to stakeout prisms and mini rods in abundance.
Paul in PA
Oh Great!
Who is "we" ? Are you the King now, or just hearing voices?
You are mixing errors (plural) with error (singular)
It gets into hairsplitting whether you want to call it horizontal angle error or centering error, and how you would test for each and what combination passes the chi squared test. I would have to see a better explanation of why you think something is amiss.
I suspect that I have systematic error in how I am eyeballing the tip of the bob "over" the point.
Oh Great!
Been testing all my fieldwork to the ALTA spec since 2005 with star*net. Nothing new there.
You can pat yourself on the back for mentioning it.
I greatly doubt that any minimum standards anywhere will become more stringent than the ALTA spec any time soon. Nice attempt to instil fear of future regret though.
A well adjusted rod will work as well as a tribrach and tripod/bipod. I like the tripod for stability and for the reasons that Kent stated.
We always use tripods and tribrachs when possible because we feel it is faster.
The old guys all think a plumb bob gets you closer, but I have not seen that hold true in practice. My sighting errors are less than the wiggle of the string.
There are specific moments when the plumb has its place, but not for this.
Oh Great!
> Who is "we" ? Are you the King now, or just hearing voices?
"We" means all of us who have a clue as to what the term "standard error" means. It's a well established usage in the surveying literature.
> I suspect that I have systematic error in how I am eyeballing the tip of the bob "over" the point.
This ain't that difficult. The way you test it is to point MANUALLY at some point that can be clearly seen in the telescope of the instrument and then have the robot point on your plumb bob/prism combo eyeballed over that same point. Remove the bob/prism combo and repeat for ten or fifteen times. Repoint MANUALLY at the same point that your eyeball was aiming for.
That gives you a way to actually measure the errors in the angles taken to the prism/bob and to characterize the uncertainties by both computing the standard error of the angles measured to them and to check for systematic errors that the mean may show.
Oh Great!
> Been testing all my fieldwork to the ALTA spec since 2005 with star*net. Nothing new there.
> You can pat yourself on the back for mentioning it.
Well, the obvious reason for mentioning it is that if you are surveying so loosely as to have standard errors of +/-0.02 ft. in all of the angles and distances, it will be difficult to actually meet the ALTA spec in many cases where the separation distances are as short as they are on many residential and urban-sized parcels.
From what you've posted, I don't get the idea that you've actually bothered to test the standard error of centering of the prism/bob combo. That is something far short of what modern surveying expects.
Oh Great!
You are stretching the truth to inflate the horizontal target centering error into an error in all the angles and distances.
With sufficient redundancy, a trivial and momentary effort with a robot to take an extra set or two, the 0.02' will shrink to a meaningless quantity.
Sigma over square root of 'N'.
Oh Great!
"Modern surveying" is a vague abstraction ... 0.02' and now I have offended modern surveying!
The King hears voices! I had better get my shovel !
Oh Great!
> You are stretching the truth to inflate the horizontal target centering error into an error in all the angles and distances.
No, actually not. If the process of centering a prism over a ground mark has a standard error of +/-0.02 ft. that means if the angles are being measured to that same prism as a target, that the uncertainty of +/-0.02 ft. is also the uncertainty of target centering. That's the uncertainty in every direction angle measured to every target.
The nature of eyeballing a plumb bob over a ground mark is such that if you have those sorts of errors in the component perpendicular to the line to the instrument, the errors in distance are unlikely to be much smaller. That's the standard error of every distance measured by the total station to a prism on a plumb bob eyeballed over some ground mark.
So it carries through the work, and in ugly ways.
Oh Great!
> "Modern surveying" is a vague abstraction
Only to those who are clueless, I'm afraid. It's painfully obvious to the rest of us.
Oh Great!
Purely ignorant generalization on your part.
If you are already accounting for the centering error in the network model, and if everything is passing the chi square test, the centering error does not magically increase the angle error and the distance error.
It is true that you could decide to have an artificially small centering error and throw all the error into the angles or the distances, but then those would not pass the chi squared test.
There is a place where everything will pass the chi squared test at close to 1 and that is when you know you have realistic assumptions (if you have enough redundancy). It maybe tell you something you don't like about your overall measurement system but it will tell you something.
So Kent, if we already know the extent of the centering error for a data set with a large enough number of samples, and we know that everything is passing the chi square test, where is all that extra error coming from? Are the voices telling the King there is "extra unpredicted error" ?
Oh Great!
I only see one of you. Please introduce me to your invisible friends?
Oh Great!
> If you are already accounting for the centering error in the network model, and if everything is passing the chi square test, the centering error does not magically increase the angle error and the distance error.
Well, I'm afraid that is simply contrary to fact. Centering errors, whether they are the centering errors of targets or instruments, ALWAYS act to increase the uncertainties in the basic angles and distances measured by a total station. That proposition should be so obvious as not to even require mention.
For example, in a least squares survey adjustment program like Star*net, the standard error of angles is the basic instrumental standard error, in effect the best possible result for a measurement. The centering error of that instrument over the point from which the angles were nominally measured introduces an error, as do the centering errors of the targets at the FS and BS stations.
There is no magic to this. This is a purely predictable result of centering uncertainty.
Oh Great!
> I only see one of you.
Well, if you want to actually familiarize yourself with the surveying literature, I'd suggest buying any standard surveying textbook such as the 5th edition of "Surveying Theory and Practice" by Davis, Foote, Anderson & Mikhail. It shouldn't be particularly difficult to find.
Oh Great!
That one is out in the shed. I will go ask them if they have been whispering in your ear.
Oh Great!
> That one is out in the shed.
Well, put it on your shelf and your chances of eventually getting licensed may improve. Trust me on this. The surveying world is full of technicians who don't really understand what they're doing. If you can speak knowledgeably about the standard errors of survey measurements, you'll be stepping out of that pack and into the light, if briefly.
Oh Great!
I'm looking at page 105 in Ghilani and Wolf, Adjustment Computations: Spatial Data Analysis, 4th Edition.
I am trying to figure out what you are missing or what we are miscommunicating, beyond your usual presumptive tone. Mrs, Bubble is out of town so I have to do my own sarcasm, or I probably would not have sparred so long past my bedtime.
What I was getting at above is that star*net allows you to isolate the centering error from the angle error and the distance error by using the chi square test.
Once you know what the centering error is, you know how it will affect the angles/distances.
I did do a similar test where I shot a point 8D&R with a glass on a rod in a bipod. I also spun the rod a quarter turn and re-leveled every 4 shots. (The rod has not one but two 8' bubbles, and I check them on a jig whenever they disagree in the slightest.) I then shot 8 D&R to my new plumb bob rig -- same glass. In this way I got an average "height" of the tied-off bob string on the glass (4.66', +/- 0.04' one sigma, and also got an average of the horizontal, which on that day was within 0.015 of the rod average and had a major ellipse of 0.06', one sigma. So, no, probably didn't test it nearly enough yet.
The thing I am noticing, looking at page 105 of Ghilani et al, is that they are assuming you are shooting a fixed target for all pointings in your set. Whereas the bob-on-a-glass is not fixed -- it is in a random "best intent" spot for each shot, which hopefully averages to something that is very close to the point.
With a fixed target that has the same centering error for each pointing, the centering error cannot be reduced in size by taking multiple pointings. (edit, took this out of quotes because it's a paraphrase, not a direct quote)
But with the handheld plumb bob (or the handheld rod) we are making what was a fixed error into a random error, one that differs in a random direction with each pointing, and can be reduced by taking multiple pointings.
Oh Great!
While closure and repeatability of measurements are not a direct measurement of accuracy & precision, they offer better confidence in the quality of work than abstract analysis.
People like to bloviate about how accurate and wonderful they are on the internet. Just the nature of the beast. But Kent, you take the cake! The real surveyors are out surveying leaving the rest of us to bloviate about how it should be done. LOL.
I just can't figure out how you can possibly make a profit in your business if you over-analyze and over-complicate everything. Yes Kent, I do carpentry work quite frequently but my measuring tape is in for calibration so I'm taking a week off.
