GPS calibration

> Ok here's the deal..

Gotcha grasshopper and you'll likely get beat up a bit by experts beyond my expertise.

Calibration, localization, local datum, or whatever you call it tosses all those "datums" out the window. You go from grid to ground and you are no longer on SPC, just fitting the local conditions. Works for me and I do it all the time. Pretty simple once you know what you're doing and how it works. Mr Robot loves it when he measures the same as my calibrated points, me too.

One point for vertical, check into at least another BM. Minimum of two points for horizontal so you have rotation and a bearing basis that can be repeated in, oh let's say 5 yrs.

I got my arse in a sling with an "inclined plane" awhile back. Talked to some guru's and found out what I did wrong. Dummy me had a line of BM's about 500 ft from my project, and it "inclined" my plane by over 4 ft, even though they all fit the numbers.

One point, or run levels around the perimeter of the project. Contractors are not stupid and they know what flows downhill, you just don't want it going your way.

Sounds like some training is in order here. Good luck.

One point calibration for me, but check other points while doing the survey.

I use 1 point localization all the time for NAD83, but the 27 datum is another animal all together. I've been told it don't play well with gps. I don't know how to advise you on this, but I don't think 1 point will get it. Proceed with caution.

depends on site characteristics, my preference is to set control outside and work in the box

IF I have the horizontal control around the site, that's what I use.

If not, a single point works great, I just shoot control as I find it to check back in the office.

I have the NEVER INCLINE PLANE option checked.

NEVER NEVER use an inclined plane unless you WANT it to be warped to "fit" your site, some folks have to.

> The scenario is Nad27 with Navd88.
The thing about NAD27 is that the best control points were established at 1:100,000 accuracy and were about 20 miles apart. For those scoring at home that's about a foot of relative error between control points - and that's as good as it gets. Any effort to tie a project into that datum is/was subject to that limitation. So if you are going to tie into control that claims to be NAD27 you had better tie and resolve all the points you can get. Just treat as you would any local control network.

Good points Norman. Kind of a good case for calibrating to local, and make your own decisions. Just check into as much as you can afford to locate and check into. At some point enough is enough.

City here is on 27 & 29 vert, all the county BM's are on 29 but NGS has really helped them get a RTN set up. Other cities around here are more gizmoed and I never know what to do, so I just calibrate my site. Whenever I need to get on 88 I just use VERTCON, and maybe run an OPUS if I'm there long enough.

If you need real NAD27 you HAVE to tie into the system (existing 1st order monuments) and then do an adjustment (sometimes called a calibration) best to do static do the "calibration" in the office, if you just want it "close" then use a file that converts from NAD83 to NAD27 and run control like you always do.

A Geoid model holding a good NAVD88 monument and then check to others, if you want really good elevations, then levels.

Don't understand what a one point calibration is. Actually would be doing nothing. :-S

> Don't understand what a one point calibration is. Actually would be doing nothing. :-S

If you go to a new site, and you don't have any "digital" data yet available to tie in to, yet you don't want to drive away with no data, you can use a "one point calibration"

You hit one point, 5000,5000 and can give it an elevation if it's a benchmark, now merely start collecting everything you want, especially all the control you can find.

Now you have something to work with back in the office. You can translate the data anyway you need.

I've done many small jobs this way, I have all (hopefully) the field work done before it's finished being calculated back in the office.

If you go to a new site, and you don't have any "digital" data yet available to tie in to, yet you don't want to drive away with no data, you can use a "one point calibration"

You hit one point, 5000,5000 and can give it an elevation if it's a benchmark, now merely start collecting everything you want, especially all the control you can find.

Yes, that's kinda what I thought, however, it really isn't a calibration, the dc is doing something and it must be creating a projection of some kind to allow the use of coordinates. Just what that is would really be useful to know. 🙂

I haven't used a calibration since about 2000. Never liked the idea from day one.

I get control points using OPUS (at one time CORS). So then I'm on a world coordinate system. Then I set up projections for my needs or use a SPC.

In essence a calibration or localization is a black box procedure for setting up a local projection system, mainly for those that don't understand how the system really works. GPS manufacturer's did it to be able to sell geodetic systems (GPS) to folks that don't have a clue.

I don't know anymore, but I suspect that the calibration function is still being sold by the venders.

Not sure why anymore things have changed so much, the big change are geoid models that are so much better. That was the real reason to calibrate back in the day, but it never worked well and as the sites got larger there were some real messy verticals where I worked.:-(

I almost never work an area with enough vertical BM's to do the 3 or more point calibration (lucky to have one good BM). I've never done one. The Geoid models seem to work good for me, I suppose if you were in an area with steep gravitational gradient the Geoid model could try to run water uphill. So it would depend on what sort of project you were doing as to how important elevation precision was.

A one point calibration would set up a projection likely centered at the GPS base (unless you tell it otherwise). I suspect that the Trimble calibration sets up a flat plane for elevation. If you do a calibration the projection settings will show up in the file. So if you understand projections you can see what is happening.

A two point calibration (horizontal) would set up a projection with a skewed meridian (unless the two calibration points were actually oriented to give a "real" North).

I've heard of more than one project where the calibration was done incorrectly and the elevations messed up and big judgments paid. One I heard about was a fish hatchery where the tanks and flows didn't work out.

Another problem with multiple point calibrations is it is going to skew the horizontal scale (maybe good or bad). If you hold a bust in a two point calibration everything else you do is going to be affected. I'd rather take the measurements as they are and then apply scaling and rotation from there as I see fit.

But what can one say. Inventing the calibration enabled the first GPS manufacturers to sell the heck out of GPS survey systems mostly to surveyors that didn't have geodetic clue (even me at the time). So they are millionaire/billionaires and I'm still just a plain old surveyor. But many of us have gone forth and figured it out and it should just plain be taught to upcoming surveyors. The black box may fail you.

pretty sure just a one point calibration puts you on the ellipsoid, would have to look at it, never did that, but some surveyors who did, when I looked at their results were really short with distances.

Calibrations are for construction sites, not boundary surveys.:stakeout: