I've got an old Trimble GPS setup - a classic Zephyr Geodetic, Trimark III, 5700 base with a 5800 rover. I'm pretty much just using it for grading and building/utility location as we slowly build out our campus master plan. Lately I've been spraying grading boundaries and shape lines for a couple excavators we have working on a detention basin on campus.
I usually set up on an unknown point and calibrate the site to 2 known points (one mag nail in rock and one gin spike in pavement), using just one for vertical datum. I get what I think are good results. I measure HI (which I don't believe matters when set up on an unknown, unmarked point), I watch my rod height and each time out I finish with a check in to a tack and hub that I set several months ago. Yet, having been out there 5 times now I've literally gotten 5 different elevations at that hub, varying by up to 0.30'. Horizontal seems to be consistently accurate. I'm not using a GEOID or SPC, but just set no datum at 1.000 correction. Fortunately, for hammering a giant hole in solid rock, a few tenths here and there won't matter, but certainly will if I want to set curb form later.
Any ideas what I might be missing, how I could pinpoint the problem or possibilities of old instrument errors?
Mahalo, LSU, for all your knowledge and community service.
This Content Originally Published by a land surveyor to Land Surveyors United Network
Dear Mr. Drahn,
I see from your profile that you are in Hawai'i, which is built upon a group of volcanic oceanic islands. It has significant variations in topography and therefore one assumes, the geoid undulates significantly there.
If you use the same pole and antenna throughout your project (as you run it now) the pole height is probably insignificant. It is so because you are NOT measuring differences of elevation. You are getting 3D positions relative to the center of the Earth Fixed Earth Centered Cartesian system. This is simply plane geometry in 3D., X, Y, Z.
Elevations are not relative to this simple system. Especially in mountainous areas, the gravitational potential varies greatly from place to place. Elevations are relative to a standard gravitational potential (zero elevation such as NAVD88).Additionally, GPS does not "know" which way is up since it works globally. However, it gives good horizontal results because we are usually surrounded by satellites E-W and N-S but NOT up-down. So the local vertical data has usually 2-3 times the doubt in it compared to the horizontal.
To properly determine the elevations one needs to know where the ground's surface is, not where the antenna is. Then the geoid height must be included to get an orthometric height (or elevation).
I bleve the geoid model has recently been extended to Hawai'i but I have no experience using it. Where it covers CONUS variations of doubt may be as large as 4-8cm..
With all that said, if the area of your work is small most of that may be ignored. I do have a rule I've always followed. GPS is good enough for most vertical work but not for "hard stuff" like road surfaces, weirs, sewers, and curbs.
You can do accurate (first order)vertical measurements with GPS but it requires multiple bases and at least two periods separated by a four hour window.
Occupy the known hub and tack that you previously set from now on, apply the appropriate GEOID Model for your location, calibrate to the two monuments that you have been all along, take a check shot on another known point and go from there. There is absolutely no reason to be setting up on unknown locations when you have good relative coordinates for your project. I would expect to see up to about a tenth of vertical fluctuation throughout the course of a project but three tenths is very high and indicates you are not using correct procedures. Try the above steps and it should help you throw out allot of your existing errors. PS, always take an antenna height, the corrections you are receiving from your base station basically control the elevations that your rover will shoot. If you set up your base station with a half a foot bust in the antenna height, all of your shots you take with the rover will have the same half a foot error.
One can perform extremely accurate observations with GPS provided long occupations and repeated redundant observations. Indeed, the basis for the adjustment of the most recent realization of NAD83 is based on the CORS. Again, the results will be ellipsoid or ECEF values, not elevations.
No matter how long nor how accurate the GPS results are, in order to derive elevations or orthometric heights from them requires the application of the geoid model with its inherent doubt. Good enough for dirt work.
You sound like you are working in an old school RTK-GPS environment. So I'll give you some old and dusty RTK-GPS tips that kept me out of trouble... for the most part that is - lol
1) When you are establishing a local frame of reference, always remember YOU are the datum for the project like this. So I would never use only 2 points for several reasons. Basic plane geometry for one - 3 points fix your datum plane and with a 4th you can see some basic residual error. I never - ever - trusted a one point calibration on any but the smallest of sites when collecting data or staking out with GPS
2) J. Cavell is spot on with his comment on inherant vertical error. Even the newest GPS gear has double the vertical error vs. horizontal. A 20mm +/- vertical tolerance means that its possible to have vertical variations on the same point of up to 40mm (0.13 ft) just from the equipment. Start adding in the real world errors and your 0.30' is easy to obtain.
3) Your 5 different elevation reading can best be explained by the differences in the different SV's (space vehicles or satellites) orbits, Pdop values, number of Sv's all come into play. Because you are most likely reading 5 to 7 SV's at a time ( GPS only ) you have fewer redundant checks in your data to correct your postions.
4) Update the ephemoris data in your equipment - yup thats old school but the equipment you described is pre GPS/GLANOSS and by todays GPS standards, kind of primative. So manual data maintence is needed.
One final note - and I agree with J. Cavell - you are scraping dirt so 2 or 3 tenths here and there isn't the end of the world, but keep an eye on it because you are rubbing the upper error limits of what should be allowed.
Hope this helps
Great info. Thanks, Kevin. When you say "update the ephemeris", is that through checking that I'm using the latest firmware? Manually altering dat or job files? Or something else?
SV orbit data - some of the older GPS gear had to load the orbit data or "tracks" of the SV's. Just like if you took a star shot, you have to have the correct date and time to apply your angle reading to the star. use the wrong day and you can be a long way off. So check your manual for the 5700/5800 - you can find them online - and see if they do it automaticly or if you have to manually do it. My old GPS gear - circa 1997 - would take a few minutes when turned on to update itself. Just not sure for your equipment
I believe our 5800 is a single frequency receiver, so you are not getting the redundancy from using two bands like most of the modern GPS. This will also contribute to your error.
Whenever I used GPS, I always check two or three known points from the very beginning. Then I'd go from there.
Thanks, all. It sounds like I'll need to spend a little more setup time to (maybe) get more consistent results. Good info.
If your site is small enough, you could put monuments (beacons) in around the perimeter of the site, and some spread out inside the site as check points. Use precise leveling to determine orthometric heights for all, then use the perimeter points to calibrate for geoidal tilt and separation, check the calibration against the internal check points. The calibration will however only be good if your GPS observations are good (longer observations and good satellite geometry) and you don't have the instrument errors mentioned elsewhere in this discussion. If your GPS elevations (after adoption of calibration) agree well with the orthometric heights of all internal check points, you should have confidence in the elevations of new points within the calibration area. Always work from a known point using the saved calibration and always check on other points. Do not work outside of your calibration area. We have always used this method before adoption of geoid models, but keep in mind geoid undulation versus distance.