Williwaw, post: 333315, member: 7066 wrote: Late to the party but rfc, it strikes me that you're going about this all wrong for a grasshopper by depending too much on software. The best way to learn this stuff is to go old school and book everything, horizontal & vertical, direct and reverse, slope distances, ect. and reduce them using a calculator and sin/cos. You'll quickly learn by meaning angles and closing the horizon how to get a tight traverse on the first go around by spotting the errors whilst you are collecting the data. If you do a back sight check measurement and find a misclosure, that is the time to repeat your measurements rather than resorting to a detailed analysis of all your data trying to pinpoint the problem. (Deep sigh) Grasshopper must master the basics before moving on to scale factors and mapping angles. I'm afraid you have the cart on the wrong end of the horse my friend.
Williwaw:
Thanks for the comments. Not to go into all my history, but in all fairness you should know I've done everything you've suggested--last year, for the better part of a year, I had NO data collector, NO software...just my HP 11C, my field book, and Davis, Foote and Kelly. Heck, for the first 6 months all I had was an optical theodolite, and a 100' steel tape. I learned how to use both.
I learned how to do it manually, and got pretty good at it. I started learning the basics. I learned to do astronomical observations, leveling, latitudes and departures, the compass rule, and least squares, and a lot more. You might remember the multitudes of setups and hundreds of observations I did (under the tutelage of you know who), and recorded manually, entering them by hand into Starnet, all just to determine the capabilities of my new instrument. I was able to repeatedly measure directions to better than 3.5" (with a 5" instrument). Then I moved on to acquiring some tools--not to become a button pusher, but to become proficient using them in the trade, as most every surveyor here does who does it for a living (not that I ever will).
I fully admit I'm not learning surveying at the moment; I'm learning software. Big difference. Over the course of the last 5 months using a DC and software, I've shot hundreds of sets of observations on the same test course; plotted them (often manually to check the work). It's probably because I've done all this that I had a sense that something was not right with my verticals, and that it wasn't just due to sloppy work on the setups. You'd have to be off by more than a minute of arc for each of the three setups for this to be attributable to a blunder (in the field at least).
So, I would respectfully disagree with you in this case, on the premise there's something wrong with my field work. This has been a chronic problem due to something else...either instrument out of adjustment, improper set up of SurvCE, or something going on with the way TPC is reading and using the files the DC is creating. I'll find out which, but for now, I'm confident that running the same traverse another 10 times and getting the same result is not going to help find what I'm doing wrong with the software.
rfc,
I didn't intend to be harsh but remember the name of this thread is 'Back to basics'. Double check even your most basic assumptions. From my experience, it's those very simple, unassailable assumptions that will lead a guy astray. You might say it's Murphy's playground. Good luck with it.
Williwaw, post: 333366, member: 7066 wrote: rfc,
I didn't intend to be harsh but remember the name of this thread is 'Back to basics'. Double check even your most basic assumptions. From my experience, it's those very simple, unassailable assumptions that will lead a guy astray. You might say it's Murphy's playground. Good luck with it.
Certainly no apology needed, Williwaw. I've learned a ton from your posts (especially about the wildlife up there!):-O. And it's good advice to continually question the assumptions.
I'm hoping someone who runs some flavor of Carlson software can look at this .rw5 and simply see if it gives them the same answers I'm seeing.
Williwaw, post: 333366, member: 7066 wrote: rfc,
I didn't intend to be harsh but remember the name of this thread is 'Back to basics'. Double check even your most basic assumptions. From my experience, it's those very simple, unassailable assumptions that will lead a guy astray. You might say it's Murphy's playground. Good luck with it.
First thing I do is reduce it all by hand, simple enough to do, I make sure the software is giving me the same results. First principles,,,,,,,,,,,,,,,,,,,,,
rfc I applaud your efforts described above, but something sometimes evades the obvious.
Occasionally when we constantly look at survey data, calculate and recalculate the result is always the same, despite it being wrong.
It's a surveyors phenomenon,
RJS a few posts back highlighted what to me is an obvious error either in sighting or vertical angle arising from either a bent instrument or one badly out of adjustment.
My Nikon has inbuilt software that checks such things by reading both faces to the same point, vertical and horizontal.
Can yours calibrate itself?
Also, as I mentioned, what happens when you read from both ends of a line and observe readings?
Do they give good or shonky results?
It doesn't sound (look) like software when comparing those readings.
That's surely pure raw data before processing.
And what ever the software spits out is a consequence of its diet fed to it.
Like many areas of life, a poor diet (garbage in) gives a poor outcome (garbage out).
It's difficult as you can appreciate to fully comprehend, understand, what is happening for us on the other end of all of this.
Hang in there, don't overlook the obvious, and perhaps try not to blame software at the expense of other issue(s).
So RFC, you have acceptable data when you look at it from a field book aspect (meaning direct/reversed vertical angles and comparing forward elevation difference vs back elevation difference), but are getting different errors when you run it through your software?
MightyMoe, post: 333376, member: 700 wrote: First thing I do is reduce it all by hand, simple enough to do, I make sure the software is giving me the same results. First principles,,,,,,,,,,,,,,,,,,,,,
Did that. See post #11 above.
Couple of quick thoughts. When you level your instrument, rotate the barrell and watch the horizontal angle. If it moves more than 10", then you need to colminate the instrument, but that ONLY affects horizontal noticabley.
For the vertical, EVERY time we do anything where elevations are critical, then the rod is verified as reading correctly in the morning, noon, and at the end of the day. Takes about 1 minute and catches ALL kinds of problems and helps ISOLATE problems when you find one.
Also, depending on your scale, since it's scaling distances, then over time, it will affect the elevations.
However, unless your scale is 0.999, that shouldn't amount to a 0.8' in half a mile so I'd say it's in the rod and HI. For this reason, the rod is set at a certain rod height, say 5.5 feet, and the markups are double checked.
I have an easier time when things are simple..in a field book format.
Let's assume I ran through the horiz./vert. columniation routines prior to starting my work and verified the instrument is in adjustment.
I go out and begin my traverse. I start by occupying #1 and measuring the vertical difference to #2.
The split between my direct and reversed vertical angles looks acceptable (shy of 360å¡ by only 10 seconds).
If it wasn't acceptable, I would check my setup and measure the angles again. If I can't get the angles to check I do not proceed.
Assuming they are acceptable, I move ahead to #2 and measure back to #1.
This time my vert. angle split is unacceptable. I want to discover this in the field so I can correct the problem without making a return trip.
If I run with the data, I have introduced approx. .03' of elevation error into my traverse.
The second red flag is how well the vertical differences compare. .07' error might not be bad for a long shot with lots of relief, but for a short leg with little elevation difference, the elevation differences should compare nicely.
Assuming no red flags, I would average the vertical differences and apply the mean to the elevation of #1 to obtain the elevation of #2.
How well I check into #1 after I close my traverse is dependent on the total length of the traverse, the length of the traverse legs, and the amount of relief.
If I was constantly getting bad splits, I would take a look at the tripods to see if the bolts are snug where the legs attach to the head.
If I had good data when I look at it from the perspective below, but was getting bad closures when the data was entered into my software, I would seriously question what the software was doing.
If I wanted to utilize an additional check (to prevent having to return to the field), I may measure my instrument height, target height, and distances in meters AND feet.
----------------------------------
..............Direct 88-30-01
........Reversed 271-28-15 (split = -00-01-44) <--- First Red Flag!!
SD=337.860 @ 88-30-53 (Adjusted)
INST. = #2 HI=5.28
BS = #1 ..HT= 5.12
VD = +8.917 <--- Second Red Flag!!
-------------------------------------
................Direct 91-34-55
.........Reversed 268-24-55 (split = -00-00-10)
..SD=337.879 @ 91-35-00 (Adjusted)
FS = #2 ........HT=5.15
INST. = #1 .. HI=5.64
VD = -8.846
imaudigger, post: 333395, member: 7286 wrote: I have an easier time when things are simple..in a field book format.
Let's assume I ran through the horiz./vert. columniation routines prior to starting my work and verified the instrument is in adjustment.
I go out and begin my traverse. I start by occupying #1 and measuring the vertical difference to #2.
The split between my direct and reversed vertical angles looks acceptable (shy of 360å¡ by only 10 seconds).
If it wasn't acceptable, I would check my setup and measure the angles again. If I can't get the angles to check I do not proceed.
Assuming they are acceptable, I move ahead to #2 and measure back to #1.
This time my vert. angle split is unacceptable. I want to discover this in the field so I can correct the problem without making a return trip.
If I run with the data, I have introduced approx. .03' of elevation error into my traverse.The second red flag is how well the vertical differences compare. .07' error might not be bad for a long shot with lots of relief, but for a short leg with little elevation difference, the elevation differences should compare nicely.
Assuming no red flags, I would average the vertical differences and apply the mean to the elevation of #1 to obtain the elevation of #2.
How well I check into #1 after I close my traverse is dependent on the total length of the traverse, the length of the traverse legs, and the amount of relief.If I was constantly getting bad splits, I would take a look at the tripods to see if the bolts are snug where the legs attach to the head.
If I had good data when I look at it from the perspective below, but was getting bad closures when the data was entered into my software, I would seriously question what the software was doing.If I wanted to utilize an additional check (to prevent having to return to the field), I may measure my instrument height, target height, and distances in meters AND feet.
----------------------------------
..............Direct 88-30-01
........Reversed 271-28-15 (split = -00-01-44) <--- First Red Flag!!
SD=337.860 @ 88-30-53 (Adjusted)INST. = #2 HI=5.28
BS = #1 ..HT= 5.12
VD = +8.917 <--- Second Red Flag!!
-------------------------------------
................Direct 91-34-55
.........Reversed 268-24-55 (split = -00-00-10)
..SD=337.879 @ 91-35-00 (Adjusted)FS = #2 ........HT=5.15
INST. = #1 .. HI=5.64
VD = -8.846
You're absolutely right about putting things in a different format. Here's some of my calculations:
And here is the Carlson .rw5 file:
JB,NM100-700-500-100,DT08-22-2015,TM13:17:44
MO,AD0,UN2,SF1.00000000,EC0,EO0.0,AU0
--SurvCE Version 2.07
--CRD: Alphanumeric
--VT NAD83
--Equipment: Topcon GTS/GPT Series
--TS Scale: 1.00000000
--EDM Mode: Fine
SP,PN500,N 428158.3170,E 1618336.8190,EL1037.6300,--
SP,PN100,N 428378.1570,E 1618499.8080,EL1002.8800,--=1
--Set Collection with Obs Order 123...321...
OC,OP100,N 428378.15700,E 1618499.80800,EL1002.880,--=1
LS,HI4.7800,HR4.8500
BK,OP100,BP500,BS216.3311,BC0.0000
LS,HI4.7800,HR4.8500
BD,OP100,FP500,AR0.0000,ZE83.0820,SD275.079000,--
FD,OP100,FP700,AR31.1316,ZE83.5502,SD388.050000,--
FR,OP100,FP700,AR211.1326,ZE276.0510,SD388.048000,--
BR,OP100,FP500,AR180.0001,ZE276.5207,SD275.080000,--
--SS,OP100,FP500,AR0.0000,ZE83.0807,SD275.079500,--
--SS,OP100,FP700,AR31.1321,ZE83.5456,SD388.049000,--
--Set Collection with Obs Order 123...321...
OC,OP700,N 428232.20968,E 1618142.61106,EL1043.941,--
LS,HI4.8500,HR4.7800
BK,OP700,BP100,BS67.4632,BC0.0000
LS,HI4.8500,HR4.7800
BD,OP700,FP100,AR0.0000,ZE95.5824,SD388.040000,--=1
LS,HI4.8500,HR4.8500
FD,OP700,FP500,AR42.5406,ZE92.0907,SD208.104000,--
FR,OP700,FP500,AR222.5429,ZE267.5047,SD208.103000,--
LS,HI4.8500,HR4.7800
BR,OP700,FP100,AR180.0045,ZE264.0048,SD388.041000,--=1
--SS,OP700,FP100,AR0.0000,ZE95.5848,SD388.040500,--=1
--LS,HI4.850000,HR4.850000
--SS,OP700,FP500,AR42.5355,ZE92.0910,SD208.103500,--
--LS,HI4.850000,HR4.780000
--Set Collection with Obs Order 123...321...
OC,OP500,N 428158.31700,E 1618336.81900,EL1037.630,--
LS,HI4.8500,HR4.8500
BK,OP500,BP700,BS290.4951,BC0.0000
LS,HI4.8500,HR4.8500
BD,OP500,FP700,AR0.0000,ZE87.4307,SD208.104000,--
LS,HI4.8500,HR4.7800
FD,OP500,FP100X,AR105.5200,ZE96.4741,SD275.072000,--
FR,OP500,FP100X,AR285.5222,ZE263.1242,SD275.070000,--
LS,HI4.8500,HR4.8500
BR,OP500,FP700,AR180.0014,ZE272.1818,SD208.103000,--
--SS,OP500,FP700,AR0.0000,ZE87.4224,SD208.103500,--
--LS,HI4.850000,HR4.780000
--SS,OP500,FP100X,AR105.5204,ZE96.4729,SD275.071000,--
--LS,HI4.850000,HR4.850000
To sum up as best I can:
The bust on the vertical is .80286, that I'm sure. But it doesn't match what TPC says it is (.78) That bothers me.
I think I have two possibly related, but separate, problems:
1. Something is up with the instrument calibration.
2. Something is up with exactly how TraversePC is reading the .rw5
I've got an email into them, but in the meantime, if someone ran the file through Carlson Survey, it'd rule #2. out (or not).
If the collective advice here is that that great an error is solely the result of rookie error, even with careful, full tripod, "zero centering error" setups, then I'll go back and do it again.
Not familiar with the Carlson format. I always handle the control data myself and only allow Carlson to reduce radial data.
That gives me a warm and fuzzy feeling. Usually there is not that many points that it really saves any time anyway.
A quick check on your data shows this...
Direct Zenith = 95-58-48
Reversed Zenith = 264-00-48
SUM=359-59-36
Adjusted Zenith Angle=95-59-00
SD=388.04'
INST = #700 HI = 4.85
BS= #100 HT = 4.78
VD= [hl]-40.379'[/hl]
-------------------------------------
Direct Zenith = 83-55-02
Reversed Zenith = 276-05-10
SUM = 360-00-12
Adjusted Zenith Angle = 83-54-56
SD = 388.05'
INST = #100 HI=4.78
FS = #700 HT=4.85
VD = [hl]+41.061'[/hl]
I see that your vertical angles don't look horrible.
However the vertical difference looking forward vs backwards don't agree between 100 and 700, nor do they agree for 700 and 500.
Are you sighting the correct crosshair in the scope? (I'm sure you are, but I have seen problems in the past).
The only other place to look would be the instrument height or the distance. It appears that when you swap target and gun you have the same measure up?
I suspect the distance. I would at the min. tape a baseline and shoot the distance in traverse mode and see what it records. Almost seems like a scale factor issue.
Personally I like to survey on the ground and convert my data to grid later to avoid issues with scale factors.
Some funky data you have going on there.
Am I misreading that data?
imaudigger, post: 333425, member: 7286 wrote: Am I misreading that data?
I'm the wrong person to ask that question of, lol!
Here's some more:
Sums of both faces, from/to each point:
From 100 to 500: 360-00-27
From 100 to 700: 360-00-13
From 700 to 100: 259-59-02
From 700 to 500: 259-59-54
From 500 to 700: 360-01-25
From 500 to 100X: 360-00-23
imaudigger, post: 333421, member: 7286 wrote: Not familiar with the Carlson format. I always handle the control data myself and only allow Carlson to reduce radial data.
That gives me a warm and fuzzy feeling. Usually there is not that many points that it really saves any time anyway.
A quick check on your data shows this...Direct Zenith = 95-58-48
Reversed Zenith = 264-00-48
SUM=359-59-36
Adjusted Zenith Angle=95-59-00
SD=388.04'
INST = #700 HI = 4.85
BS= #100 HT = 4.78
VD= [hl]-40.379'[/hl]
-------------------------------------
Direct Zenith = 83-55-02
Reversed Zenith = 276-05-10
SUM = 360-00-12
Adjusted Zenith Angle = 83-54-56
SD = 388.05'
INST = #100 HI=4.78
FS = #700 HT=4.85
VD = [hl]+41.061'[/hl]I see that your vertical angles don't look horrible.
However the vertical difference looking forward vs backwards don't agree between 100 and 700, nor do they agree for 700 and 500.Are you sighting the correct crosshair in the scope? (I'm sure you are, but I have seen problems in the past).
The only other place to look would be the instrument height or the distance. It appears that when you swap target and gun you have the same measure up?I suspect the distance. I would at the min. tape a baseline and shoot the distance in traverse mode and see what it records. Almost seems like a scale factor issue.
Personally I like to survey on the ground and convert my data to grid later to avoid issues with scale factors.
Some funky data you have going on there.
I thought everyone on this forum used Carlson.
I just realized, though, from your calcs, that I did not (in my manual calculations above), account for HI and HR. But I assume SurvCE does.
They are as follows: 100: 4.78; 700: 4.85, and 500: 4.85. I could have screwed up the LS records. There was a very confusing screen, when looking at the back sight at each station for the HR. I thought it defaulted to what it already knew, but maybe not. Look at the very last line of the file:
--LS,HI4.850000,HR4.850000
If that's telling me that it thinks the foresight (100X) is at 4.85, then I've screwed up. I can't tell whether the LS records precede the shot or come after.
There are a few others that, if I'm reading them correctly now, then they're wrong. I don't think they could all add up to .8' but I'll do that math manually now.
Thanks for the help.:-)
Addendum: I just checked the SurvCE Data Structure manual. It clearly says that the LS record should precede any traverse and sideshot records.
Also, looking at the data file above, all the LS records themselves are correct. But the ones at the end of each Set Collection, that are commented out ("--") seem wrong. And in the second and third sets, right after each SS record, is another LS record. WTF are those? They're the ones that seem messed up.
rfc, post: 333447, member: 8882 wrote: I thought everyone on this forum used Carlson.
I just realized, though, from your calcs, that I did not (in my manual calculations above), account for HI and HR. But I assume SurvCE does.They are as follows: 100: 4.78; 700: 4.85, and 500: 4.85. I could have screwed up the LS records. There was a very confusing screen, when looking at the back sight at each station for the HR. I thought it defaulted to what it already knew, but maybe not. Look at the very last line of the file:
--LS,HI4.850000,HR4.850000
If that's telling me that it thinks the foresight (100X) is at 4.85, then I've screwed up. I can't tell whether the LS records precede the shot or come after.
There are a few others that, if I'm reading them correctly now, then they're wrong. I don't think they could all add up to .8' but I'll do that math manually now.
Thanks for the help.:-)
I'm pretty sure you have a bad measure up, try using feet and meters on the hi's oh's next time.
there should be a way to isolate it, if it's only one, if you can't then there are more than just the one.
forget all the math of angles and distances, just reduce it all to the vertical difference, run it backwards and forwards like a checkbook and the error should jump out.
And no we don't all use Carlson, never even looked at it.
rfc, post: 333442, member: 8882 wrote: I'm the wrong person to ask that question of, lol!
Here's some more:
Sums of both faces, from/to each point:
From 100 to 500: 360-00-27
From 100 to 700: 360-00-13
From 700 to 100: 259-59-02
From 700 to 500: 259-59-54
From 500 to 700: 360-01-25
From 500 to 100X: 360-00-23
My suggestion....
Check the adjustment of your gun by going through the collumination routines.
Take it to a known baseline and verify you do not have a scale/prism offset problem.
Establish a neat form of keeping notes. Set your notes up before traversing and fill in the blanks.
Let the DC also collect data.
Verify your data in the field while you are collecting it and preserve the checks on your field notes.
Know what the answer should be so you can check what the software is spitting out.
Then go back and try and make head or tails of your DC file.
I'm sure if you posted up a proposed field book form, you could get some good suggestions based upon lots of experience.
In my experience the control work is minimal compared to hand recording thousands of side shots.
I use Carlson for reducing radial surveys and hand record and calc. control in a field book that gets archived for later use.
Here is what my Carlson raw file looks like for a radial survey.
just reduce it all to the vertical difference, run it backwards and forwards like a checkbook and the error should jump out.
:good:
This is what I do when carrying elevations on a control traverse. Note the vertical delta direct and reverse on the foresight, mean the two assuming they agree within +/-0.03', plus up the HI, add or subtract the vertical delta, minus out the HR to get my new elevation, on the fly and into the book. Do a vertical back sight check. If they don't agree, rinse and repeat. It's much easier to repeat a setup or get a new measure up than it is to wade through reams of raw data files looking for the proverbial needle in the haystack or having to run the traverse over again. If the vertical deltas direct and reverse don't agree fairly close, that would point to a collimation issue with the instrument. A bad measure up will jump up and b*tch slap you. 😉
just reduce it all to the vertical difference, run it backwards and forwards like a checkbook and the error should jump out.
:good:
This is what I do when carrying elevations on a control traverse. Note the vertical delta direct and reverse on the foresight, mean the two assuming they agree within +/-0.03', plus up the HI, add or subtract the vertical delta, minus out the HR to get my new elevation, on the fly and into the book. Do a vertical back sight check. If they don't agree, rinse and repeat. It's much easier to repeat a setup or get a new measure up than it is to wade through reams of raw data files looking for the proverbial needle in the haystack or having to run the traverse over again. If the vertical deltas direct and reverse don't agree fairly close, that would point to a collimation issue with the instrument. A bad measure up will jump up and b*tch slap you. 😉
just reduce it all to the vertical difference, run it backwards and forwards like a checkbook and the error should jump out.
:good:
This is what I do when carrying elevations on a control traverse. Note the vertical delta direct and reverse on the foresight, mean the two assuming they agree within +/-0.03', plus up the HI, add or subtract the vertical delta, minus out the HR to get the new elevation, on the fly and into the book. As a general rule, don't go beyond 500'. Do a vertical back sight check. If they don't agree, rinse and repeat. It's much easier to repeat a setup or get a new measure up than it is to wade through reams of raw data files looking for the proverbial needle in the haystack or having to run the traverse over again. If the vertical deltas direct and reverse don't agree fairly close, that would point to a collimation issue with the instrument. A bad measure up will jump up and b*tch slap you. 😀
