In this thread, I began the saga of finding two inter visible NGS control marks in my neighborhood, upon which to take my new Topcon solar filter for a spin:
https://surveyorconnect.com/threads/can-i-use-two-ngs-marks-for-azimuth-control.326278/
Round One ended somewhat ignominiously, when I discovered my backsight was NOT the NGS mark, but another on a similar bridge abutment some 150 meters closer.:excruciating:
Not to be thwarted by that small setback, I returned to the scene today, and repeated the work: 16 shots (8 direct, 8 reverse), of which I had to throw out two, and came up with 175-26-8.7, or about 11.9 seconds off what Inverse3d says it should be (175-25-56.8).
I'm a little disappointed, but don't have a lot to compare it to, given my meager experience.
I did a set of observations (direct and reverse) between the errant mark and the real one, and came up with 00-09-04, which when added to my results last weekend (175-17-11.1), is 175-26-15.11, or 18.4" off what it should be.
I'm not going to mean those two results, because they're not identical measurements...that is, there's additional centering errors adding the measured angle to the previous results. But it's an additional data point.
I attribute some of the problem to the really nasty slope I had to set up on, and I did a reverse Beer Leg without thinking: the leg was precisely in line with the sun, and away from it down the slope, so I had to twist over sideways to get the shots. I couldn't stand directly in line. My neck was killing me. That definitely interrupted the flow somewhat, although my residuals weren't horrible:
It was also a little windier than I would have liked (about 8 kn).
Also, a half a mile with some heat shimmer made it really difficult to see the targets clearly. I used the big ones I had...about 10" square...They're the ones some here kidded me last year about being so big they could be seen from space.:-D I really want to try this early am, when shimmer won't be a factor.
I'm tempering my disappointment somewhat by knowing that I can definitely get better at this with a little practice.
As a bonus, when I was done, I realized the waxing gibbous moon was above the hill to my east, so did a set of 16 observations on it!
Got to figure out how to get some ephemeris data into Larry Scott's sheet for that, so I can crunch the numbers. Stay tuned.:stakeout:
rfc, post: 367483, member: 8882 wrote: In this thread, I began the saga of finding two inter visible NGS control marks in my neighborhood, upon which to take my new Topcon solar filter for a spin:
https://surveyorconnect.com/threads/can-i-use-two-ngs-marks-for-azimuth-control.326278/
Round One ended somewhat ignominiously, when I discovered my backsight was NOT the NGS mark, but another on a similar bridge abutment some 150 meters closer.:excruciating:Not to be thwarted by that small setback, I returned to the scene today, and repeated the work: 16 shots (8 direct, 8 reverse), of which I had to throw out two, and came up with 175-26-8.7, or about 11.9 seconds off what Inverse3d says it should be (175-25-56.8).
I'm a little disappointed, but don't have a lot to compare it to, given my meager experience.
I did a set of observations (direct and reverse) between the errant mark and the real one, and came up with 00-09-04, which when added to my results last weekend (175-17-11.1), is 175-26-15.11, or 18.4" off what it should be.
I'm not going to mean those two results, because they're not identical measurements...that is, there's additional centering errors adding the measured angle to the previous results. But it's an additional data point.
I attribute some of the problem to the really nasty slope I had to set up on, and I did a reverse Beer Leg without thinking: the leg was precisely in line with the sun, and away from it down the slope, so I had to twist over sideways to get the shots. I couldn't stand directly in line. My neck was killing me. That definitely interrupted the flow somewhat, although my residuals weren't horrible:
It was also a little windier than I would have liked (about 8 kn).Also, a half a mile with some heat shimmer made it really difficult to see the targets clearly. I used the big ones I had...about 10" square...They're the ones some here kidded me last year about being so big they could be seen from space.:-D I really want to try this early am, when shimmer won't be a factor.
I'm tempering my disappointment somewhat by knowing that I can definitely get better at this with a little practice.As a bonus, when I was done, I realized the waxing gibbous moon was above the hill to my east, so did a set of 16 observations on it!
Got to figure out how to get some ephemeris data into Larry Scott's sheet for that, so I can crunch the numbers. Stay tuned.:stakeout:
I'm a bit confused by your numbers, but a couple of things:
With your instrument 12" is about as good as you will get,,,,the d and r numbers indicate your instrument is in need of adjustment.
This is important when doing solars.
1/2 a mile isn't very far for a sight, a 1" lath is more than enough, if there is too much heat shimmer then sight something clear and turn angles in the morning or evening to the NGS point which isn't a good time for solars.
Your instrument isn't a 1" gun, don't expect 1" results
MightyMoe, post: 367493, member: 700 wrote: I'm a bit confused by your numbers, but a couple of things:
With your instrument 12" is about as good as you will get...Your instrument isn't a 1" gun, don't expect 1" results
I'm not sure of the proper way to "stack" the one sigma tolerances. Extrapolating from Bill93's assessment of the 95% confidence for the relative positions of the two marks themselves....that's about 7". My gun is a "5 second" gun, but I've tested it to be able to do 3.8" for directions and probably something less than 6" for angles. I haven't yet been able to characterize the standard errors for my solar technique, but I've read that one should be able to nail it within a second or two (The width of the reticle is about 1").
So do I add all those up and expect (at the 95% confidence level) to be within 14" (close enough to your 12", but I wasn't sure you were talking about the gun alone)?
Obviously, I'd be thrilled to be able to consistently perform astro observations as good as my hardware will allow. That's all I'm looking for.
rfc, post: 367497, member: 8882 wrote: I'm not sure of the proper way to "stack" the one sigma tolerances. Extrapolating from Bill93's assessment of the 95% confidence for the relative positions of the two marks themselves....that's about 7". My gun is a "5 second" gun, but I've tested it to be able to do 3.8" for directions and probably something less than 6" for angles. I haven't yet been able to characterize the standard errors for my solar technique, but I've read that one should be able to nail it within a second or two (The width of the reticle is about 1").
So do I add all those up and expect (at the 95% confidence level) to be within 14" (close enough to your 12", but I wasn't sure you were talking about the gun alone)?
Obviously, I'd be thrilled to be able to consistently perform astro observations as good as my hardware will allow. That's all I'm looking for.
Forget the statistical analysis of a solar observation, its not all that relevant for this exercise.
You are setting on a good point looking a known az. That is a given.
10inches, if I'm reading that correctly, is way to wide a sight in 1/2 a mile. You should plumb a pole over the point, a 1" wide rod or 5/8" pipe, something like that. If you can't see that cause of shimmer don't use it, sight a clear, random sight, turn angles from it to the NGS point in the evening or morning when the heat waves settle.
If I'm seeing your numbers correctly you have a 30" spread d and r. You need to check that out and clear that up if you want tight angles,,,,,
As far as solars, the sun moves 15" of arc every 1" of time so you aren't going to get a 1" turn with a solar. Your gun is a 5" gun with an apparent 30" d r issue,,,,,,12" (that's your number, not mine) is pretty good actually.
To test the d r, sight something with a big vertical (a natural foresight) turn angles from it to a tight backsight.
When I say 10", I don't mean to imply I'm sighting somewhere within that 10".
Here's the target I'm using. The "V" is visible at 1/2 mile; I just think the road heat shimmer was making the shots a little less primo than they might have been.
As for the spread between D and R...I'm not sure that that is the case, but there may be something wrong with that section of the spreadsheet. I'll check on that with Larry.
My backsight for the reverse shot was 179-59-55 today, and 179-59-58 last week.
I'll check the gun for that specifically tomorrow with two marks a couple of hundred feet apart, but I don't think that's a contributing factor. Thanks for the input.
Two things.
Precise leveling of instrument when steep sights are involved.
And, sighting the sun too early in the morning can be an issue with atmosphere.
rfc, post: 367502, member: 8882 wrote: When I say 10", I don't mean to imply I'm sighting somewhere within that 10".
Here's the target I'm using. The "V" is visible at 1/2 mile; I just think the road heat shimmer was making the shots a little less primo than they might have been.
As for the spread between D and R...I'm not sure that that is the case, but there may be something wrong with that section of the spreadsheet. I'll check on that with Larry.
My backsight for the reverse shot was 179-59-55 today, and 179-59-58 last week.
I'll check the gun for that specifically tomorrow with two marks a couple of hundred feet apart, but I don't think that's a contributing factor. Thanks for the input.
Checking your D vs. R for this application needs to be done with a sight that has a large vertical, turn to something stable that is high, the edge of a roof vent or something (doesn't have to be far away, just clear, and stable), get the vertical up like you need to do for a solar, then turn to something level. This will simulate the procedure. you don't need to set anything out, just find two "natural" sights to look at, do the d and r and see if the gun has an issue.
Its something we used to check all the time, it just takes a couple of minutes.
Seeing 30" like your chart shows would give me pause. For a longer sight I do prefer a plumb rod for control work, that helps to take out some of the shimmer. That rod plumbed on the point would work well.
Take the Solar to a distant sight such as a building steeple, water or cell tower etc. then check between your control marks.
As Scott Z has stated precise level the instrument and avoid early morning or early evening sightings.
Sight uniformly and avoid parallax in sightings. Time keeping and pointing should be consistent. Time interval between direct and reverse should be consistent.
Last week's azimuth plus today's angle:
175-17-10 (az to unknown disc)
+ 0-09-04 (measured angle)
____________
175-26-14 = last week's az to Gulf Stream
175-26-09 Today's azimuth to Gulf Stream
Difference 5". That's reasonable repeatability.
Scintillation can do strange things. I'm guessing that the line of sight passes less than 1.5 m to the pavement surface.
The difference from D to R az is transiting error, which cancels outs. The plate has dual axis compensation, so only transiting err remains. The instrument has a fast level vial. 30"/2mm. So it's the trunnion axis leveling. IMO.
The 3" std dev for each D and R shows good timing and observation. And last week's az was 4" D, and 5" R std dev.
So, Bob, are these the azimuths and angle?
10-15 seconds is pretty good for solar observations in my experience. If you want better, use Polaris observations. Time is less critical due to slower apparent motion. The target is more precise.
Your difference in D and R isn't critical as long as you include a D and R backsight for each face. Using an instrument with dual axis compensation reduces errors due to leveling. So this probably isn't an issue.
Did you happen to consider deflection from normal? This accounts for the difference in slope between the geoid and ellipsoid for the observation station. It's like being out of level with the ellipsoid.
We used to only do solar obs. After 2 in the afternoon, and before 10 in the morning. Usually around 8 am, and around 4 or 4:30 in the afternoon.
This reduced the effect of deflection of the vertical.
Several other red neck tricks, were to intentionally run the observation 1" wrong, to see how much angle, one second of time made. We also were pretty careful to calibrate our time measurement devices. Based on this, we concluded that sunshots we're only good to about 5" of arc. Most of the time. When the angle is high, there seemed to be more error potential. Even due to standing and looking through the inst, at an odd angle. We'd remove the observation, that was 1" of time off, before averaging them. We used the C&G sunshot routine, that have us grid, and true.
We used to only do solar obs. After 2 in the afternoon, and before 10 in the morning. Usually around 8 am, and around 4 or 4:30 in the afternoon.
This reduced the effect of deflection of the vertical.
Several other red neck tricks, were to intentionally run the observation 1" wrong, to see how much angle, one second of time made. We also were pretty careful to calibrate our time measurement devices. Based on this, we concluded that sunshots we're only good to about 5" of arc. Most of the time. When the angle is high, there seemed to be more error potential. Even due to standing and looking through the inst, at an odd angle. We'd remove the observation, that was 1" of time off, before averaging them. We used the C&G sunshot routine, that have us grid, and true.
And a plumbing pole isn't a good backsite.
One set doesn't an azimuth make. 4 sets, 4-6 DRs each, in one session. And average.
And software isn't the cure. The computation should be the same long hand or PC.
Thank you. This is very helpful (as is the spreadsheet that made it easy in the first place). I know it's not the software that produces good results, but it sure makes one more efficient at crunching lots of numbers quickly.
Regarding leveling the instrument, I didn't just rely on the bubble, which is a 30"/div. vial. 
I went to the compensation screen where it tells you how far X and Y are off, then aligned the scope such that two tribrach screws are in line with X and the third at right angles, and work until the compensators are near zero.
It's really hard though; they're very touchy. If you breath on the tripod, it moves a few seconds.
And let's have a big round of applause for Bob.
He repeated an Astro azimuth on two successive days with only 5" of difference (or there abouts) observing one set each time, using a plumbing pole, with adverse atmospheric conditions.
And that level vial is the weakest link.
It's unnecessary to precise level from compensator. The compensator compensates for the values that you see.
Shawn Billings, post: 367826, member: 6521 wrote: It's unnecessary to precise level from compensator. The compensator compensates for the values that you see.
Shawn:
OK then, but what does the last sentence (before the bold section) mean? Is the "vertical indexing adjustment" they're talking about the "Adjustment of Compensation Systematic Error of Instrument"? The "VCo, HCo, and HAx" thing? I asked about this eons ago and if I'm not mistaken it was you that helped me figure out how to do it. I did go through this procedure on Friday, only hours prior to the latest round of solars.
I have noted that when the compensators are near zero, the bubble isn't perfectly centered. Is that what they're talking about when they refer to the "true level condition of the instrument"? So is what you're saying "Don't worry about precise leveling; the compensators do that....BUT make sure the compensators are adjusted properly"? Is that it?
I'd really like to fully understand these matters, and be able to perform adjustments available to the user, but if I'm screwing it up, I might just have to bite the bullet and send it off to a pro. My recollection is that a year or so ago when I did that, they didn't even touch these adjustments.
I've probably shot over a thousand astro azimuths and a couple of hundred solar azimuths. Your set up and a precise way to measure time are key elements. Best practice for solar obs were also (if I remember correctly) to be made 2 hours apart. Astro shots were usually 4 or two a night so waiting was less preferable to using a different T3 to repeat the measurement. If your monument is not suitable for a super stable set up it is preferable to move to a temp az station and turn in your intended line.
The adjustment of the bubble is important.
However, if the bubble is centered in D, and off a bit in R, then 'level' is halfway in between.
So, the bubble may not be centered when it is truly level. The key is that the bubble does not move - at all - when changing instrument face. And that bubble is not sensitive enough.
Most instruments (my experience) the bubble's behavior is slightly temperature sensitive. I setup, get the bubble level. Meaning if it's slightly to one side D, and in the exact same location in R, it's level. An hour later the bubble may be in a slightly different location, but, does still does not move as the instrument is rotated. It's level. With a 30" bubble it's very hard to spot very small movement. And above 20+ degrees vertical angle, the horizontal angle measurement will not be true in either face, and the average D/R hor angle will be true, as long as the plate is level (via compensator), but the trunnion axis is not. Which I believe is your situation. And by observation procedure, D/R average, the horizontal angle should be true.
I looked at your notes and I see that the delta elevation D and R are grossly different: -2.97 and -1.77. I assume that's feet at 2300 feet. Again, average vertical difference -2.37, that's a big diff. And also indicates that there is an adjustment issue, around a minute. In manual instruments, even a T2, a minute error in vertical index is not uncommon, and DR average is true. But fully compensated auto vertical index, I'd expect better. It would be wiser to record the vertical angle instead if vertical difference. So, setup and record H and V angles, D & R, to a high target. 30å¡ elevation or higher. And read a backsite that is near the horizon. Then evaluate differences in the angles.
Pmoran:
Having that much solar obs experience, how good are solar azimuths?
Bob is using NTP time source via smart phone app. I use the same and compared to WWV, it's well below 0.05". And the UTC to UT offset (DUT) is applied. The ephemeris used is reckoned in UT. The Laplace correction of -3.9 seconds, is also applied.
(His 2 observations were on different days, and the 2 azimuths are only 5" different. However a dozen seconds different than inverse. The line of sight is over 2000 feet of asphalt and he reports high scintillation.)
