"Trig Points" in the US

[Plumbob2]

We're kinda short on "trig points" here in the U.S. Most of our benchmarks are pretty crudely placed having been placed eons ago, and it's hard to know which of the ones whose notes indicate that their coordinates were later corrected by "GPS" were corrected with anything accurate enough to help.

 

FYI - here in the US, and almost world wide, there are many thousands of Survey Grade GPS points. Most are better then a cenimeter of accuracy - some even have a published bearing and velocity to track movement. If you're trying to compare hand held GPS units, may I suggest you call your local Land Surveyor, City or County Surveyor, NGS, or NOAA for the location of a survey monument with a published Lat & Long.

 

Bob in Seattle

[+Indotguy]

Testing the accuracy of consumer grade GPS devices against the coordinates of a professionally set bench mark is pretty futile. Consumer GPS devices have an error range of approx. 3 meters so even if you are spot on the bench marks coords. it is more luck than accuracy. If you wish to check your GPS's accuracy try marking a waypoint with your GPS at a fixed location (yes, a BM will work) and continue doing this over a period of days. Have the device turned on with sats acquired for a half an hour before performing this and remain at the mark for the same amount of time (a few minutes) before you mark. A plot of these waypoints in relation to the fixed location will tell you more about your GPS's accuracy than comparing it to a professional survey device.

Edited March 28, 2010 by Indotguy

[+Walts Hunting]

Or to find one the easy way you go to geocaching.com, click "find a Benchmark" and put in the zip code for a list and tha last day found.

 

As to checking accuracy at one that only gives the accuracy given current conditions, if you went to several you would get different readings particularly when the constellation of satellites changes.

[+ecanderson]

FYI - here in the US, and almost world wide, there are many thousands of Survey Grade GPS points. Most are better then a cenimeter of accuracy - some even have a published bearing and velocity to track movement. If you're trying to compare hand held GPS units, may I suggest you call your local Land Surveyor, City or County Surveyor, NGS, or NOAA for the location of a survey monument with a published Lat & Long.

 

Bob in Seattle

 

Are the survey grade points actually published as a separate list anywhere? Are we talking about established "benchmarks" in the NGS list? If so, how does one identify in the descriptive information that a particular point is of such accuracy? I've seen the listings that note that the coordinate data was updated via GPS, but I don't know how to distinguish between a garden variety update that is likely no more accurate than a typical handheld and one that has been done with the kind of precision you mention.

[+ecanderson]

Testing the accuracy of consumer grade GPS devices against the coordinates of a professionally set bench mark is pretty futile.

That would depend upon the nature of my test. As it happens, it isn't futile at all.

Consumer GPS devices have an error range of approx. 3 meters so even if you are spot on the bench marks coords. it is more luck than accuracy.

No surprise there.

If you wish to check your GPS's accuracy try marking a waypoint with your GPS at a fixed location (yes, a BM will work) and continue doing this over a period of days.

As was the case in a previous discussion, repeatability has been confused with accuracy, and both are to be tested under a variety of satellite (constellation) conditions, both intentionally good, and intentionally poor. Also being tested are best and worst user practice with the GPS itself (e.g., antenna orientation, time to settle, etc.).

 

But to answer your concern, the reported vs. actual (the issue that started this being "actual") position will be one of the things to be measured numerous times under several conditions. We have in hand a unit a friend was prepared to discard that developed what appeared to be an offset in one axis that is of interest.

 

In short, I will be interested in any notable differences in not just the statistical radius of each unit, but whether the centroids of the recorded positions for any of the units are somehow unusual vs. a specific "known" point. That will help to determine if, for some reason we haven't been able to imagine as yet, one or more of these units has some sort of built-in bias.

 

Already, the repeatability portion (which does not require a known point of reference) has been pretty fascinating. For that, any point can be chosen that provides the desired view of the sky that is appropriate for that portion of the test.

 

Another thing that is turning up some really interesting data points is the "settling time" differences between units. I'm defining that as the time from "on" (understanding that prior almanac knowledge is one of the variables) to establishment of readings that fit, statistically, within a decreasing radius to a more or less stable number. Watching the tracks that these things build from a stationary point has been illuminating! It's also one reason that I'm a bit annoyed that we still haven't seen the WAAS lock fix for the Cartesio chip units yet ... when no lock can be established within any reasonable period after power on, you can quickly see the results in the data. Yup, WAAS really does help (why does that still surprise some people?) to reduce the statistical radius of the measurements, and sometimes even moves the centroid a bit. Anyway, I'm miffed because in taking the data, the luck of the draw with WAAS lock on Oregon and Dakota units screws up the measurement set if I'm trying to work with apples to apples, and when the objective of a particular part of the test is to get "best case" data, I can't just turn WAAS off on all units to make them all apples. I'm getting tired of restarting side-by-side tests with other units that include Oregon and Dakota just to see if a Cartesio chip might be kind enough to pull up some data for 48 or 51 on the first pass through the list.

Edited March 28, 2010 by ecanderson

[+ecanderson]

Or to find one the easy way you go to geocaching.com, click "find a Benchmark" and put in the zip code for a list and tha last day found.

 

As to checking accuracy at one that only gives the accuracy given current conditions, if you went to several you would get different readings particularly when the constellation of satellites changes.

 

Understood, but perhaps missing what I'm trying to test. "Find a Benchmark" isn't any better than using the same spot in my backyard time after time, and the location requirement for the repeatability portion of the test doesn't require anything other than an understanding of whether the sky view is being obstructed in some way.

 

As for the constellation, that is also understood, and is already intentionally incorporated into the testing. Unfortunately, I can't get the US Gov to work on my schedule, so being sure I'm hitting some true peaks and valleys of constellation configuration at my location has had me testing at some pretty peculiar hours! There are a couple of good calculators that help to determine best/worst moments at a given latitude and longitude. They take all of the satellites and project the quality of the view of each to a given point across whatever period of time you specify. The results are a somewhat arbitrary graph of the results, but clearly show the peaks and valleys of the number and quality of signals that can be expected at a given point and time.

[xyzee]

I lost most of my NGS url's in a previous XP crash, but still had the one for the DATA SHEET POINT RADIUS FORM.

www.ngs.noaa.gov/cgi-bin/ds_radius.prl

IE can't open it, so I did a Google for the main NGS site, and it won't open either. The main site had a lot of pages that helped me understand what the "daffynitions" were.

I wonder, changed url, down for maintence, my pc, or?

 

How long does the Cartesio chip take to lock on to the WAAS Sats after start up?

[+ecanderson]

I lost most of my NGS url's in a previous XP crash, but still had the one for the DATA SHEET POINT RADIUS FORM.

www.ngs.noaa.gov/cgi-bin/ds_radius.prl

IE can't open it, so I did a Google for the main NGS site, and it won't open either. The main site had a lot of pages that helped me understand what the "daffynitions" were.

I wonder, changed url, down for maintence, my pc, or?

Bummer. Well, I really do appreciate the attempt. I was checking another site this morning (http://svinetfc4.fs.fed.us/rastergateway/scp/page_40105.html) and discovered that they'd moved that one, too. Musical chairs at the .gov and .fed.us sites seems to be a job security requirement

 

How long does the Cartesio chip take to lock on to the WAAS Sats after start up?

Seems to be a total crap shoot. The Cartesio based units I've worked with have all had issues with WAAS. It seems that the current GPS firmware has difficulty pulling down the WAAS almanac data, even with a booming full bar signal indication for 48 and/or 51. After several minutes, it times out, gives up and moves on to the next possible satellite, including all of the EGNOS satellites - which is pointless here in Colorado. If the unit already knows from the uncorrected data if it's in Europe or the U.S., looking for satellites that can't possibly be seen makes zero sense. That's something that they just plain need to correct, even if they clean up the almanac time-out problem. Seems like it can take forever to cycle through all of the possibles (half an hour?) before it gets around to trying the "local" ones again.

 

So the answer to your question ... it's going to lock either during the first couple of minutes like you'd expect from any other unit (and the odds of this happening in our location might be about 20%), or it's half an hour later when the unit gets around to trying sensible satellites again. Speaking from the U.S. side (reverse to EGNOS if you're in Europe - I assume the same issues apply over there, but don't know that for a fact) if you don't get WAAS lock on an Oregon or Dakota on the first pass, it's statistically MUCH faster to just power the unit off and on again and let it take another crack at 48 and 51 without the delay of looking for every other satellite numbered 33 and above! The problem is that it can take several of these attempts to get a lock. Sometimes you get lucky - sometimes you don't.

 

A secondary but equally annoying problem is that the Cartesio seems to lose WAAS lock in a way that doesn't occur with the other chips. It is my feeling from my experience that the following occurs: if the unit is shaded from a locked WAAS satellite such that the continuous signal is lost for even a moment, the WAAS almanac data that you were lucky enough to receive to get the WAAS lock in the first place was not cached, and the entire process (including the high odds of missing a lock again) starts over from the beginning! That bites.

[+EScout]

I find it useful to go to super-accurate benchmarks. It gives confidence in your GPSr. It is amazing how close you usually get with your GPSr. I mean very close.

 

If you are in the USA, you want the NGS Datasheets.

http://www.ngs.noaa.gov/cgi-bin/ds_county.prl

Go to the state, county and choose "GPS sites only". You can choose some near you then narrow to those benchmarks with public access.

 

I have not gone to the site in some time,and it does not currently work. I have 7 of these benchmarks in my local area that I have gone to. A few years ago, they had big white Xs painted over them, probably for aerial photos (the ones in unpaved area had a plastic X fastened to the ground.)

[+ecanderson]

I'm familiar with the benchmark lists that the NGS publishes. I'm also aware that some benchmarks have been updated with GPS coordinate information. But are all of those annotated as "Location Adjusted" to be trusted for the kind of accuracy that can be achieved with today's equipment? Some of these GPS adjustments were done quite a few years back (e.g., 1995 - since NGS is toast today, see the copy of the sheet at gc.com at http://www.geocaching.com/mark/datasheet.aspx?PID=AA6418), and I don't know how to gauge the true quality of these older "Location Adjusted" horizontal coordinates. How accurate are they really compared to what can be done today?

 

And yeah -- the whole NOAA NGS server is down today. Forgot to pay the electric bill again, I guess.

[xyzee]

Short break from outside work, fantastic day her in mid Utah, January Thaw is finally here!

 

Thanks for the NGS web status. I'm on the end of the phone line from the station, so everything is suspect.

 

The MTK in my 60Cx, when connected to the ex ant on the roof, acts the same as your description of the Cartesio. I was on the phone with a Garmin rep, when after 1 hr and 49 min, it locked on 48. I figured it was the chip, but the rep said no, to reset, which I had several times already. He did say that they are not allowed to discuss "chips". Sometimes it will connect to 51, like the rest of thr 60'ies, sometimes like you have to do with the off/on thing to make it find 51. Otherwise, it starts with 33, just like it will normally do when it loads up the WAAS Sats for the first time. Same with both ant's on the roof. But, when in a south window, or when outside, it grabs 51 like the rest of the units do.

I can't figure the roof thing, in the clear, no trees, has 51 all night as I capture the tracks on the map page, then change batteries, and it wants to start on 33. Need to test with the short 3 ft ant/ground plane/pole setup, and see what happens.

I put faith in the GPS corrected sites, even if a few yeare old.

[Plumbob2]

We're kinda short on "trig points" here in the U.S. Most of our benchmarks are pretty crudely placed having been placed eons ago, and it's hard to know which of the ones whose notes indicate that their coordinates were later corrected by "GPS" were corrected with anything accurate enough to help.

 

FYI - here in the US, and almost world wide, there are many thousands of Survey Grade GPS points. Most are better then a cenimeter of accuracy - some even have a published bearing and velocity to track movement. If you're trying to compare hand held GPS units, may I suggest you call your local Land Surveyor, City or County Surveyor, NGS, or NOAA for the location of a survey monument with a published Lat & Long.

 

Bob in Seattle

[Plumbob2]

We're kinda short on "trig points" here in the U.S. Most of our benchmarks are pretty crudely placed having been placed eons ago, and it's hard to know which of the ones whose notes indicate that their coordinates were later corrected by "GPS" were corrected with anything accurate enough to help.

 

FYI - here in the US, and almost world wide, there are many thousands of Survey Grade GPS points. Most are better then a cenimeter of accuracy - some even have a published bearing and velocity to track movement. If you're trying to compare hand held GPS units, may I suggest you call your local Land Surveyor, City or County Surveyor, NGS, or NOAA for the location of a survey monument with a published Lat & Long.

 

Bob in Seattle

 

Also, I see some conversations about using GPS for elevations too. Usually, under good conditions, the vertical has an error of 2.5 times that of the horizontal error. So a 10' coordinate error could introduce a 25' error in the vertical altitude - notice I did not say elevation. The "GPS Elevation" is not based on a true sealevel datum. If you think you might be in a 100 year flood zone, don't use you handheld unit for elevation.

 

Bob in Seattle

[+ecanderson]

I'm not going to get wrapped around the problems of elevation. That's (fortunately) not of concern, and is a much more difficult problem to tie down, if for no other reason than the ball we live on ain't exactly round, so fudge factors must be employed just in the description of "elevation" when it come to GPS.

 

Anyway -- the above begs the question. I find it difficult to believe that an entry for a benchmark showing that it was location "Adjusted" by GPS back in 1995 is an indication of sub-meter accuracy. Or maybe it is. I have no idea what the folks at NGS were using back then, only what is available now.

 

It would be helpful if someone could tell me, in rough terms, that by the year 19XX or 20XX, the only hardware for GPS locates being used was capable of producing location with an accuracy of 1 meter or less. Or who knows -- there may have been a considerable transition period from old to new hardware, and no "year" or short range of years will serve as an accurate indicator of the probable error of an NGS locate by GPS. I just don't know. And for that reason, I don't know whether I should be skeptical of benchmark sheet entries that indicate a GPS adjust or new GPS locate. Was it really as good in 1995 as it is now? Centimeter??? I'd settle for a foot.

Edited March 29, 2010 by ecanderson