Using StarNet - What Does a 1" Direction Error Mean - Trimble S9 Total Station

Direction in the Star*Net context refers to a bearing or azimuth, i.e. an angle with respect to a known orientation. Regardless of the manner in which the direction was determined (e.g., astronomic measurement, GNSS, gyroscope, compass), your needed to tell the software what the standard error of that direction is.

Jim Frame, post: 444768, member: 10 wrote: Direction in the Star*Net context refers to a bearing or azimuth, i.e. an angle with respect to a known orientation. Regardless of the manner in which the direction was determined (e.g., astronomic measurement, GNSS, gyroscope, compass), your needed to tell the software what the standard error of that direction is.

I don't think that is what the standard error of a direction is intended to represent in Star*Net, though. On the menu that appears in the OP, the standard error of a direction means the standard error of a direction angle as something quite distinct from the standard error of an azimuth or bearing.

The standard errors of directions are applied when direction sets in DN or DM format are entered as input. Otherwise, the standard errors of angles apply in A, M, and T lines.

Kent McMillan, post: 444774, member: 3 wrote: I don't think that is what the standard error of a direction is intended to represent in Star*Net, though. On the menu that appears in the OP, the standard error of a direction means the standard error of a direction angle as something quite distinct from the standard error of an azimuth or bearing.

The standard errors of directions are applied when direction sets in DN or DM format are entered as input. Otherwise, the standard errors of angles apply in A, M, and T lines.

Kent, Jim

I found the following website http://helpdesk.microsurvey.com/index.php?/Knowledgebase/Article/View/1232/35/starnet-instrument-input-settings

As you can see it states that the direction value is taken from the instrutment itself, I have attached the spec sheet for my intrutment where in this document woulf you find the direction value?

The Trimble spec sheet gives standard deviations based upon DIN 18723, which is the standard error of a direction.

It isn't clear to me how Trimble distinguishes the instruments that measure directions with standard errors of 0.5" from those that have standard errors of 1" on the spec sheet.

Maybe I should have waited until I had the software in front of me before posting. Does Start*Net use bearing as the name for a known direction?

(Just boarded a plane for London; no Star*Net on my phone!)

Star*Net does use the letter "B" for a known bearing and I've used it in the past.

Jim Frame, post: 444781, member: 10 wrote: Maybe I should have waited until I had the software in front of me before posting. Does Start*Net use bearing as the name for a known direction?

(Just boarded a plane for London; no Star*Net on my phone!)

Well, when you get to the UK, be sure to ask if they have any "stone messuages" available on offer.

Steward Souten, post: 444767, member: 12714 wrote: Hello All

Stu with another basic question.

I am currently using StarNet to teach myself about survey adjustments.

StarNet requires instrument settings to determine the error/uncertainly in my observations, specifically Direction as shown below:-

I was wondering if anyone could explain what a Direction is (I know that the value entered for Direction affects data types such as DB DE etc) I just don't know what it means to have a 1"sec Direction error - can anyone explain?

I am currently using a Trimble S9 Total station and I have been told that I just take the direction value from the specification of my instrument. However I have looked through the spec sheet for my instrument but cannot find any values relating to direction. I have attached the spec sheet for reference. Can anyone help locate where on my spec sheet the direction value is?

Hello Steward,

You're not going to find it on the spec sheet, but unless Trimble are trying to mislead us all, just take it that a direction has an uncertainty of 1" for a 1" S8 under ideal conditions. I would expect that since Trimble mention 1" angular accuracy then no matter the DIN or ISO spec then they reasonably expect that whoever buys the instrument is wanting to see a fair amount of sub-second residuals in an adjustment done by a package like star*net. So even if you used 0.7" as an input SD for a direction if you took their words as literally true (1" angular accuracy) then you shouldn't get into trouble. Once you are using a 1" and 1mm instrument like yours, your adjustments in star*net won't change much if you use 1.5", 1" or 0.7" as your input direction deviation. You'll be playing with fractions of a mm.

I hope the above is of some help, I don't want to muddy the issue for you.

Regarding the micro survey documentation you linked to, they state:

Angle Accuracy - The Angular Accuracy = 2 x (DIN 18723) / Square Root of the Number of Pointings.

I would understand this to not be true. It seems based on a probably faulty assumption that the angular or directional accuracy of an instrument is due to random, normally distributed errors. It seems to me that the predominant error sources that give rise to angular specifications are systematic ones.

Conrad, post: 444809, member: 6642 wrote: I would understand this to not be true. It seems based on a probably faulty assumption that the angular or directional accuracy of an instrument is due to random, normally distributed errors. It seems to me that the predominant error sources that give rise to angular specifications are systematic ones.

Yes, I find it both puzzling and incredibly annoying that the level of technical documentation reflected by that sheet of specifications is at the used car lot level. It's not as if the user shouldn't care what the uncertainties in measurements made with some specific instrument are and what strategies tend to reduce them below the specified levels of uncertainty.

Kent McMillan, post: 444812, member: 3 wrote: Yes, I find it both puzzling and incredibly annoying that the level of technical documentation reflected by that sheet of specifications is at the used car lot level. It's not as if the user shouldn't care what the uncertainties in measurements made with some specific instrument are and what strategies tend to reduce them below the specified levels of uncertainty.

I got to see the sqrt(no. of readings) assumption by a software vendor recently when using their field software with a GNSS unit. I was observing a point with a code solution and it started off with about 1m accuracy at the first epoch, then roughly 0.7m, 0.6m, 0.5m, 0.4m after 5 seconds. The assumption with their quality indicator being that somehow each epoch was degraded by instantaneous and perfectly uncorrelated effects. All anyone has to have done is observe the slowly meandering trace of an autonomous or differential GNSS position to know that sqrt(n) isn't describing any part of the situation. I can assure anyone reading that the leica quality estimates do not exhibit this kind of behaviour.

I also browsed the DIN article linked to in the micro survey article and I must say I'm none the wiser.

Isn't the DIN Spec per pointing, so a foresight and back-sight constitute 2 error sources? So that would make the 2 x DIN / Sqrt. of pointings make sense to me.

G

gary_g, post: 444837, member: 1026 wrote: Isn't the DIN Spec per pointing, so a foresight and back-sight constitute 2 error sources? So that would make the 2 x DIN / Sqrt. of pointings make sense to me.

G

Hello Gary,

it makes no sense to me, as this formula pretty much ignores systematic biases in instrument specifications, and assumes domination of temporally and spatially random effects that can be brute-forced away with enough pointings. If you are able to spec an instrument to 0.5" seconds using procedures specified in DIN 18723 then the random effects present in the testing procedures must be pretty small. How then to produce these perfectly random variations in instrument readings? We would be left, perhaps, with mischievous little sprites dancing within the instrument to produce moment to moment variations in sensor voltages and component positions! Or something else

Steward Souten, post: 444767, member: 12714 wrote:
I was wondering if anyone could explain what a Direction is (I know that the value entered for Direction affects data types such as DB DE etc) I just don't know what it means to have a 1"sec Direction error - can anyone explain?

Hey Steward:
If you've got time for some reading, check out this last installment of an epic 6 part endeavor to determine precisely what the REAL number is for your instrument, regardless of manufacturer's data sheets:

https://surveyorconnect.com/community/threads/ts-direction-accuracy-chapter-6.284927/#post-284969

Under the tutelage of Kent, I set up an array of identical targets in a field, and got my answer. I think you can too.:)

Steward Souten, post: 444767, member: 12714 wrote: Hello All

Stu with another basic question.

I am currently using StarNet to teach myself about survey adjustments.

StarNet requires instrument settings to determine the error/uncertainly in my observations, specifically Direction as shown below:-

I was wondering if anyone could explain what a Direction is (I know that the value entered for Direction affects data types such as DB DE etc) I just don't know what it means to have a 1"sec Direction error - can anyone explain?

I am currently using a Trimble S9 Total station and I have been told that I just take the direction value from the specification of my instrument. However I have looked through the spec sheet for my instrument but cannot find any values relating to direction. I have attached the spec sheet for reference. Can anyone help locate where on my spec sheet the direction value is?

What I would do is -

If you have a 1" instrument then start with 2", 1" and 2" for angle, direction and zenith errors.
azimuth/bearing errors would not apply to the adjustments I do.

the centring error affects these errors as well, maybe start with 1mm for each of these to start with.

when you run an adjustment you will get the average standardised errors for each of the observation types. you can then tweak the instrument error values to try to achieve 1 for all of these values.

I believe the above is fairly standard practice. If you are using multiple instruments or other input it does start to get a bit more complex.

Steward Souten, post: 444767, member: 12714 wrote: I was wondering if anyone could explain what a Direction is (I know that the value entered for Direction affects data types such as DB DE etc) I just don't know what it means to have a 1"sec Direction error - can anyone explain?

Am I correct in assuming you don't know the meaning of a "direction"?
Some instruments cannot, or need not to be set to zero to measure an angle.
Point at the backsight and read whatever is on the plate/display, turn to the foresight and read whatever is now on the plate/display. You will get the angle by subtraction.
(You can tell I'm old by the use of "plate", the horizontal circle with the angle divisions etched or grooved)
Using directions is quite a timesaver. Point and read, point and read, point and read...

Dave Lindell, post: 444920, member: 55 wrote: Am I correct in assuming you don't know the meaning of a "direction"?
Some instruments cannot, or need not to be set to zero to measure an angle.
Point at the backsight and read whatever is on the plate/display, turn to the foresight and read whatever is now on the plate/display. You will get the angle by subtraction.
(You can tell I'm old by the use of "plate", the horizontal circle with the angle divisions etched or grooved)
Using directions is quite a timesaver. Point and read, point and read, point and read...

[SARCASM]You forgot the obligatory "GET OFF MY LAWN"[/SARCASM]

Dave Lindell, post: 444920, member: 55 wrote: Am I correct in assuming you don't know the meaning of a "direction"?
Some instruments cannot, or need not to be set to zero to measure an angle.
Point at the backsight and read whatever is on the plate/display, turn to the foresight and read whatever is now on the plate/display. You will get the angle by subtraction.
(You can tell I'm old by the use of "plate", the horizontal circle with the angle divisions etched or grooved)
Using directions is quite a timesaver. Point and read, point and read, point and read...

Hello Dave,

How is an angle determined with any method other than subtraction of one direction from another made from a static circle? To my mind it is besides the point how the circle was settled in its position before reading. I want to make sure I'm not missing something as I didn't think any instrument needed to 'set to zero' to measure an angle. Every instrument could be used as a directional instrument no?

The assignment of a part of a circle to be 'zero' during construction, in my mind, seems to be an arbitrary assignment with no significance other than for the maths.

SEE:

https://surveyorconnect.com/community/threads/understanding-din-specs.29064/

I think the only time I've used DN is with resections- It's the way Star*TDS reduces a resection in a survey pro .Raw file.
In those cases, I've usually ended up with a direction value the same as my angle value.

Do any of the Star*Net gurus have other instances where they use it?