From a statistical (LSA) point of view, which will produce higher certainty of azimuth:
Many Polaris shots from a single station in the network, or,
Fewer Polaris shots from multiple stations in the network. Assume proper interpolation of Convergence, LaPlace, etc. at each station.
As here:
If it is truly random error, you are better off with more measurements.
If some systematic error dominates, then doing it under different conditions, even if fewer times, will provide more benefit.
Was taught to take the time and make a minimum of six readings for any solar and route angles.
😉
I would say multiple points. If you have a least squares program with simulation capabilities it should be easy to see which is more beneficial. You would need to estimate the accuracy of each possibility. It all depends on your instrument, methods, and somewhat the skill of the observer, although with polaris that is less critical than with a faster moving object.
If you use an instrument without dual axis compensation, instrument mislevelment can be one of the largest sources of error and it is NOT eliminated by D+R.
> Was taught to take the time and make a minimum of six readings for any solar and route angles.
> 😉
Yes. Understood. I intend to do at least 6 on one station. I posed the question as a proportion (many on one vs fewer on more stations).
> I would say multiple points. If you have a least squares program with simulation capabilities it should be easy to see which is more beneficial.
I'm beginning to use TraversePC, which has LSA; not sure I can do simulation with it. With this few points though, the demo of Starnet might do simulation (not sure; I'll look).
> You would need to estimate the accuracy of each possibility. It all depends on your instrument, methods, and somewhat the skill of the observer, although with polaris that is less critical than with a faster moving object.
I believe I've characterized most of my errors fairly well. I'll tweak upon crunching.
>
> If you use an instrument without dual axis compensation, instrument mislevelment can be one of the largest sources of error and it is NOT eliminated by D+R.
Have dual compensation. One problem I've yet to figure out is that even after running the V angle 0 datum routine multiple times, I'm still getting 5"-7" of difference between D & R.
Thanks for the input.
The difference in D+R does not matter as long as you observe on both faces and use both in your comps.
Since it is a bit harder to observe on an astronomic object than the backsight, I usually take 1 backsight, then 3 foresights to the star/planet/sun/moon, then flop, 3 more foresights to the star/planet/sun/moon, and then backsight. That constitutes one set. Ploaris is easier than the others since its apparent movement is so slow, but you still have a steep vertical and may be fumbling with the stopwatch, etc.
Your Backsight Should Be Horizontally Near Polaris
I would go with John Hamilton's method with one exception.
BSD, 3*FSD, BSD, BSR, 3*FSR, BSR plus you record time on all shots, BS and FS.
The two extra Backsights add very little observation time to the set.
Paul in PA
Your Backsight Should Be Horizontally Near Polaris
> I would go with John Hamilton's method with one exception.
>
> BSD, 3*FSD, BSD, BSR, 3*FSR, BSR plus you record time on all shots, BS and FS.
>
> The two extra Backsights add very little observation time to the set.
>
> Paul in PA
John and Paul:
Thank you. That is great information.
Now that I have a data collector, as long as I check that my time in the DC is accurate, do I need to fumble with a stopwatch at all? Can't I just sit there and fire away, letting it record the time of the shots?
Your Backsight Should Be Horizontally Near Polaris
Several issues with using the time in the DC. It depends on what data collector hardware and software you are using. For example, I use Survey controller on a TSC2. If I export a .dc file, the time is not in there. However, if I convert the .job file to a .jxl file, the time is in the file for each shot. I have been meaning to do a test to see if there is any delay, but have not done it yet. You could to a "calibration", but it is difficult to set the time exactly without an NTP client program.
Polaris at elongation does not need super accurate time. Polaris at upper or lower culmination does need good time. In between it varies. For other astronomic objects, time is much more critical, that is one of the advantages of polaris.
Buy an inexpensive stopwatch. I used to buy them at Radio Shack but I don't know if they are still around. The start the watch at the minute beep (see below). Do not use the displayed time on cell phones, as I pointed out last year that different phones have delays. I do have an app called Navy time that does display time sync'd to an NTP server, but there still might be slight delays.
An inexpensive shortwave radio can pick up WWV time signals, which is the best way to get time. Alternatively you can call the US Naval Observatory at 202-762-1401 or WWV in Colorado at 719-567-6742.
Be aware of DUT1 on these signals, although as I mentioned above it is not so important for polaris, especially if you avoid upper or lower culmination.
Your Backsight Should Be Horizontally Near Polaris
> Several issues with using the time in the DC.
I output raw rw5 files. I'll do a test to see how close the time is. I'm pretty sure it'd be within a second or two.
I thought for Polaris that's good enough
> Polaris at elongation does not need super accurate time. Polaris at upper or lower culmination does need good time.
How good? My recollection last fall was that it was on the order of a tenth of an arc second per second, or an arc second per 10 seconds or so.
I only have a 5 second instrument, with a proven accuracy of less than 4", so I thought that if I get the time to within a second or two in time, I'd be way OK.
> Buy an inexpensive stopwatch. I used to buy them at Radio Shack but I don't know if they are still around. The start the watch at the minute beep (see below). Do not use the displayed time on cell phones, as I pointed out last year that different phones have delays. I do have an app called Navy time that does display time sync'd to an NTP server, but there still might be slight delays.
>
I've compared my iPhone app "Skytime" with several computers sync'd to an NTP server, and it's been dead on after months and months (I believe the app sync's as well).
But be it a stopwatch, iPhone or something else, it's just one more thing to mess with. Granted, the last time I did this there was a foot of snow on the ground (into which I dropped my laptop), and 10 degrees, so it might be a heck of a lot easier to deal with, sitting in a lawn chair sipping cold ones the way Scott Zelenak does it; I just thought that if my DC can do it for me, well, why not?:-D
Your Backsight Should Be Horizontally Near Polaris
yes, it would be great to use the dc. I initially thought it would be impossible until I realized that trimble has an ascii file converter that creates the .jxl file, and the timetag data is in there. As soon as the weather is nice (like you said), I am going to do some testing. Should be easy to do. I recently purchased a Leica T3000, and I am anxious to try it out.