I'm preparing a second attempt at measuring the cyclic error of my TS, and in that preparation have created some sample files using a different DC. I brought in both a Carlson RW5 (top window), and a Sokkia SDR (bottom window):
The RW5 file seems to show the slope distances rounded; the SDR does not. But more worrisome, is the fact that the Horizontal distances between the two don't even match!
Here's another version with the number of digits reduced:
I'm scratching my head. Any ideas?:-S
Are any scale factors being applied?
Showing any distance past 1mm measured from a EDM, would be just kidding your self. It just can not be measured any better. Too many significant digits in this case.
Therefore if you show to mm, the two formats ate in harmony.
The mathmeticians may disagree.
Lee Green
For example, here is what an actual SDR file with slope distances (in red) logged in meters looks like (mean computed mark-to-mark slope distances in blue):
[pre]
13TSMar-15-15 13:55
02RS 4 1.50800000 POINT
12SC 4004001111
09F1 4 113.74100000 90.00472222 259.13472222
09F1 4 218.75300000 90.14583333 78.97472222 POINT
09F2 4 218.75300000 269.85555556 258.97083333 POINT
09F2 4 113.73800000 270.00194444 79.13805556
13RSThe following MCs are derived from set(s) 1.
09MC 4 113.73950049 89.98470829 180.11125013
09MC 4 218.75299029 90.13291778 359.94763902 POINT
13RS1 DValues 0.001 0-00'07" 0-00'00"
13RS2 DValues 0.001 0-00'05" 0-00'00"
[/pre]
Here's it to three digits:
Lee is right...the numbers are the same (well, almost the same).
No scale factors. I thought Kent might be on to something regarding the ft vs. m, but I checked. I'm using Magnet field; I had the check box for "select file units" checked, and went back and checked it for both file types, and it was set to Meters, so I don't think that's the factor (it looked probable though).
I guess if the numbers are the same to 3 digits, I shouldn't worry; I get sidetracked with the minutia sometimes.
> I guess if the numbers are the same to 3 digits, I shouldn't worry; I get sidetracked with the minutia sometimes.
When you write that you created data collector files that you then imported in both RAW and SDR formats, what does that really mean? I was struck by the fact that your SDR slope distances in meters converted to distances in feet to the nearest 0.001 ft., which I'd say is pretty strong evidence that the meters were converted from slope distances measured in feet.
> > I guess if the numbers are the same to 3 digits, I shouldn't worry; I get sidetracked with the minutia sometimes.
>
> When you write that you created data collector files that you then imported in both RAW and SDR formats, what does that really mean? I was struck by the fact that your SDR slope distances in meters converted to distances in feet to the nearest 0.001 ft., which I'd say is pretty strong evidence that the meters were converted from slope distances measured in feet.
The DC captured the info. The only relevant setting I could find that I set before making the observations was in the job configuration menu, where it gives you a "distance precision". I had neglected to set that to .xxx. It was set to .x prior to the observations (I think that's just a display setting though...when I "edit raw data" on the DC, it shows the precision to one digit, but obviously it recorded it to more.
Then, after the observations I exported the job "to file" in the "exchange" menu. That's where it lets you save as a rw5, sdr, and a host of other formats. Nothing was different at that point between the saving to either file type. Could it be that Topcon Magnet's "translators" do something slightly differently to the two file types, or is there something intrinsic in the file types that they'd deal with the data differently? It is curious that the numbers convert closely as if recorded in feet.
>I was struck by the fact that your SDR slope distances in meters converted to distances in feet to the nearest 0.001 ft., which I'd say is pretty strong evidence that the meters were converted from slope distances measured in feet.
I think you're right (again...yawn);-)
While I set the units to meters from the DC, I assumed it controlled the TS. So I just booted the TS into "parameter mode" or whatever it's called, and checked the units there. Guess what? FEET.
The TS was recording in feet; the DC was converting to meters. I'll start over, trying the TS set to meters.
Not to restate the obvious, but
[pre]
5.9874912m = 19.644 ft.
5.9877960 = 19.645
5.7393940 = 18.830
5.3556408 = 17.571
[/pre]
... and so on.
OK. Instrument set to record in meters. Saved the files again (in meters).
They're still different.
Not by much, but if I'm about to embark on an attempt at super precise cyclic error determination, I should be eliminating these vagaries, shouldn't I?
> They're still different.
> Not by much, but if I'm about to embark on an attempt at super precise cyclic error determination, I should be eliminating these vagaries, shouldn't I?
>
>
>
So, what are you doing? Are you logging the measurements to the same device and then exporting them in the two different formats or what?
Are you exporting them to two different projects in some Topcon data transfer software, one with the fields formatted for different precision of expression?
I'd want to know what accounts for the difference, myself. If one data collector format is doing a better job of logging measurements, then that's the one I'd use. Basically, what you want to know is which DC format is logging the actual input from the instrument and which is getting creative.
For the purposes of the cyclic error test, the limiting value is probably the uncertainty in your tape measurements of the distances to the holes on the rail to which the distances will be measured. I'm going to guess that over 5m that uncertainty is probably on the order of about 0.5mm or more, mostly depending upon how well defined the centers of the holes are.
> So, what are you doing? Are you logging the measurements to the same device and then exporting them in the two different formats or what?
Yes. Exactly.
>
> Are you exporting them to two different projects in some Topcon data transfer software, one with a the fields formatted for different numbers of decimal places?
No. Same "project" in Topcon Link. I just imported both data files and showed them on the screen one below the other.
>
> I'd want to know what accounts for the difference, myself. If one data collector format is doing a better job of logging measurements, then that's the one I'd use. Basically, what you want to know is which DC format is logging the actual input from the instrument and which is getting creative.
Duh. Never thought of just looking at the screen on the instrument as soon as it reads the distance. I've torn down the setup now, but tomorrow, prior to the cyclic test, I'll do that.
>
> For the purposes of the cyclic error test, the limiting value is probably the uncertainty in your tape measurements of the distances to the holes on the rail to which the distances will be measured. I'm going to guess that over 5m that uncertainty is probably on the order of about 0.5mm or more, mostly depending upon how well defined the centers of the holes are.
I've got an arrangement that I think will get me better than .5mm. Stay tuned.
> > I'd want to know what accounts for the difference, myself. If one data collector format is doing a better job of logging measurements, then that's the one I'd use. Basically, what you want to know is which DC format is logging the actual input from the instrument and which is getting creative.
>
> Duh. Never thought of just looking at the screen on the instrument as soon as it reads the distance. I've torn down the setup now, but tomorrow, prior to the cyclic test, I'll do that.
One thing you can look at now is the actual DC file which ought to be readable as ascii text in a text editor. Both files should give values for the slope distances.
Adjusting several different measurements of the distances to the holes on the rail in Star*Net, one set of measurements from each direction and one set more from each direction using a different part of the tape ought to give a reasonably good estimate of the uncertainty of the distances as long as the tape is corrected for temperature and properly tensioned.
> One thing you can look at now is the actual DC file which ought to be readable as ascii text in a text editor. Both files should give values for the slope distances.
Here are both files, opened in a text editor. I can't make heads or tails at all out of the SDR file. It doesn't even look like the SD's are there; I could be reading it wrong.
As for the RW5, it looks like the SD's are only recorded to 3 decimals. Go figure.
I'm not sure this explains anything.
Im going to have to decide which to use later today, so, failing any dramatic insights to the contrary, I'd go with RW5, as my other DC (the old Jett running SurvCE), defaults to RW5. One less thing to deal with, and I can download/upload the same files to either.
The SDR file lists these ranges (which I assume are in meters):
[pre]
Pt S.Dist
26 6.0016
27 6.0090
28 6.0086
29 6.0088
30 6.0088
31 6.0014
32 6.0016
[/pre]
I'd say it's logging distances to the nearest 0.1mm. I'd use the SDR format.
