Assessing the Precision of Your Kit Using a Known Reference

[Quantum]

Here are a few notes for those who would like to test their own kit for precise accuracy on the globe. You guys are trying to help me, so I'm going to help you. That's the way it works in hackerland.

I found the Seattle Calibration Baseline: http://www.ngs.noaa.gov/cgi-bin/env_cgi?name=AI3599
... with the help of a local GPS wonk, who really knows his stuff. He'd measured the 150meter disc's location down to sub-centimeter with his $10k survey-quality receiver,
then post-processed in software to base stations, for WGS84:
47°40'30.97724"N 122°15'07.73659"W -15.303m (ellips.)

Now, calibration baselines are a layout of several brass discs up to 1km apart, so surveyors can check the accuracy of their equipment. The location of a given disc is officially given in a format called NAD83 (surveyer). This must be converted to WGS84 (GPS), for our purposes. To do this, use this (impossible to find) handy calculator:
http://www.ngs.noaa.gov/cgi-bin/HTDP/htdp.prl?f1=4&f2=1

In the interest of scientific process, be careful trusting info about NAD83 vs WGS84. When GPS began they were considered equivalent so some converters might not really change anything. But at that time most consumer GPS didn't get near a meter anyway (which is the approximate difference although it varies across the US). If you care about the last centimeters, WGS84 sort of keeps changing because it improves with each "realization" (when someone important decides to recompute everything up to date). The US Coast Guard used to use NAD83 for its WAAS corrections, but WAAS reference stations now is more likely to be using WGS84 (not NAD83). Caveat: I don't know if Coast Guard has changed from NAD83. Regardless, this would only affect that last meter.

So anyway, I put my pod over the + on the disc (knowing where its patch antenna is), and took readings at intervals. I wrote them by hand, every 30 seconds for twenty minutes, but I just realized that you may be able to use the logging function of VisualGPSce, and import the result into a spreadsheet.

My readings varied slightly around
47°40.51601"N 122°15.12871"W
... and from 6-9 meters altitude. Notice how the reference coordinates are in minutes' & seconds", and VisualGPSce's are in minutes and decimal fractions of a minute? I converted the reference to my format by dividing 30.97724 by 60 for .51629 as the fraction of minutes, and similar /60 for longitude. The converted reference coordinates are:
47°40.51629"N 122°15.12894"W

Subtracting this from my mean average, gives my error of:
-0.00028N -0.00023W

... so too far east and south.

Now, a second of latitude here in the Pacific Northwest equals around 31 meters, and a second of longitude is about 21 meters. So let's convert my errors (minute fractions) to meters:
(31m*60seconds)*-0.00028error fraction of a minute, lat
(21m*60seconds)*-0.00023error fraction of a minute, long

... which means my error is:
-0.5208mN == 1.71' south
-0.2898mW == 0.95' east

HEY, wake up! This is the answer.

I'll go into altitude if I feel like it. It's not what you'd think.

My thanks to Doug Hogarth, for his kind, generous, and capable tutorage in this subject:
http://niceties.com/gps.html

Hope this helps.

[Quantum]

Some more corrections. Lutz, can you please make these corrections to my first post and delete these second two?
-----------------------------------------------------
NAD83 for its WAAS corrections, but WAAS reference stations now is more likely
... s/b
NAD83 for its WAAS corrections, but WAAS reference stations now are more likely
-----------------------------------------------------
NAD83). Caveat: I don't know if Coast Guard has changed from NAD83.
... s/b
NAD83). Caveat: I don't know whether the Coast Guard has changed from NAD83.
-----------------------------------------------------
and import the result into a spreadsheet.
... s/b
and import the result into a spreadsheet. I put my readings into a page each for modes SBAS+DOP, SBAS-only, DOP-only, HDOP-only, and NoCorrection. (I'd changed the mode of my pod with the excellent SiRFTech app, by Dennis Groning:)
http://w5.nuinternet.com/s660100031/SirfTech.htm

Then I used the first page to do averages, standard deviations, and to make charts. I made a scatter-chart for coordinates in each of the modes, and a tracking line-chart for the altitudes over time. I would share this spreadsheet, but I do not use Microsoft products, and thus it is in OpenOffice Calc OpenDoc format.
-----------------------------------------------------

[Quantum]

Hm, I notice that my requested corrections have not been made. I would do it, if I could...

[philpugh]

All very interesting but what are you going to use it for?

The amount of variation will be affected by:

the satellite configuration (which is constantly varying)
the local weather conditions - including temperature.
the local vegitation / building skyline conditions
let alone anything the US military may decide to do in the future.
I would expect most (all?) general purpose GPS units to give very similar results given the same conditions.

Impressive though 0.5m sounds - it's still the difference between being at the top of a tall cliff and at the bottom!

Try repeating the exercise at intervals of a week or two. I would expect the variation to change.

Here in the UK there are no shortage of trig points that are at a defined lat/long (or OSGB grid ref) and altitude - should anyone want to use them - just check any OS map.

We see you’re using an ad-blocker. We’re fine with that and won’t stop you visiting the site.
Have you considered making a donation towards website running costs?.
Or you could disable your ad-blocker for this site. We think you’ll find our adverts are not overbearing!