Elevation affect on staking with GPS

Are you staking to geodetic values (LLH) or to grid values (NEZ)?

If you're just changing the elevation while staking to the same grid NE values, you're not going to see a difference....because you're still staking the same northing and easting.

(Edit to add that depending on where you are working, the shift due to elevation - really ellipsoid height - may be miniscule...or noticeable.)

I don't know why anyone would do building layout in SPC. It's a job for a TS with 1000, 1000, 100 origin.

I don't know why anyone would do building layout in SPC

Maybe because the Architect designed it over a boundary plan done in SPC?

If you are on any projections in most of today’s software.  You cant use the old i only care about North and East. We must have either an ellipsoid height or elevation and the geoid. This doesn’t matter if you are using gps or a robot or total station it will burn you. Especially if you use zero or 0 as an elevation. 0 or the famous-99999999 that we all used for years is bad juju. In Trimble use no number.  If your stake out coords are defaulted from a cad package as 0 or -99999999. Just open the csv file and get a close elevation or average.  It will probably be close enough.  When i am cogo in Trimble from a plat that i know is on spc and i am working in state plane. I look at my site or even take a ground shot so i can have a elevation to correctly scale my ground keyed i. Distance to grid as i search ama stKe it out for trying to find.  The video by surveying with robert gary posted. Is an example. Now i love robert but some technical terms are not precise but he shows what can happen. I stopped placing 0 and -99999999 as a default elevation when survey controller was used i had a macro that ran through and changed it to blank because the robot would look at the center of the earth lol. 

@gary_g  i had to send this video to my crews when he first posted it. I try and make them aware of issues before they have them for sure. I have caught a few in house but luckily for us it didn’t have that 0 elevation or that -999999999 one. That cad does.

When you created the job you gave it an average elevation which the data collector then used as the "plane" to place your northings and eastings on. The only relevance of your stake point having an elevation is the cut/fill that will be computed. The differing distances between your 2 points at various elevations that are shown in your sketch are accounted for when you specify the average height for the job. Specify the average height of 10,000 feet and stake your points. Then specify an average height of 100 feet and stake your points. You will now see the difference you were expecting.

I didn't use the terms elevation and height correctly. Your job will ask for an ellipsoidal height when you create it. If you say 10,000 feet then your points will be placed on the ground with the distance between them that you would expect for a plane at 10,000 feet. Like wise if you give your job a working ellipsoidal height of 100 feet, those same points when staked on the ground will have less distance between them than when staked using an ellipsoidal height of 10,000 feet.

@jimcox That is the question of the day. I files that state nad 83 state plane all the time I even called the PE and the surveyor on one as i had found some issues on the control. Surveyor had control good. Gives to engineer somehow it gets moved. Engineer takes original control on state plane. Does some type of scale in c3d i don’t know he said to make it work on state plane. I get all and have to stake it. Finally figured out it had been scaled to ellipsoid from grid. As they used scale factor not cf. and the few points left on site were moved accidentally.  Luckily on this site the ellipsoid and ground distances were close enough. But that is something that with the new datum coming is will we account for correct elevation or ellipsoid heights on a grid system and not the famous 0 or -9999999. That will mess terrestrial systems up . Why I love TBC. I don’t have to flatten a boundary line to elevation 0 to get the correct distance along the line when labeling the bearing and distance. I also can label it either grid or ground or both. Especially when at higher elevations. Ground and grid will be closer but grid will still be projected not assumed.  I think a lot of people treat the grid coordinates as numbers and a coordinate and not on a projection unfortunately on the design side which messes us up.

I didn't notice the sketch yesterday before I replied...that sketch is not representative of how projections and elevations work. The plane does not move up and down with the point elevations - it stays put.

Projections are inherently 2-dimensional. (You mentioned that you did a site calibration, which creates a custom projection for your job site.) Even if you enter in a geodetic ECEF or LLH value, the northings and eastings are computed solely from the latitude and longitude. Height is computed separately, either from the ellipsoid surface or with a geoid added in.

This is why you can go to the NGS NCAT and enter in only a latitude and longitude and get a SPCS northing and easting. Ellipsoid height is not in any of the calculations - take a look at manual NGS NOS 5 to see that elevations are not in any of the equations.

Refer to the attached image, attempting to depict ECEF coordinates being projected onto a plane at surface elevation.

Your attempt to depict ECEF Should look more like this. Every plane coordinate with the same N and E but different elev. will result in very different xyz in the ECEF system. When you are surveying in a plane system there is lots of transformation going on behind the curtain. Pay no attention to the man behind the curtain.

Thanks for the responses all!  The information you gave was very helpful; especially Rob's video.  I did more testing this morning with changing the job height in the properties and also mimicked Rob's procedure with GPS points and total station EDM reading between computed points vs GPS measured points.  Over my short distance of 117' between points I saw a variation of about 0.08' between computed points at zero elevation and points at 10000' elevation.  That seems to match Rob's 250' between points and 100000' difference in elevation.

For those who are concerned that this is for building layout: it isn't.  These are rough points with tolerances of +/- 0.1' and no elevations.  I was curious if giving my computed points a zero elevation as opposed to a near-true elevation would affect the horizontal position.  That answer appears to be "no" as long as my job height is approximately accurate.  However, the site has local control that I will be calibrating to anyway, so this was just a thought-experiment.

As I understand it, my job height sets the location of the plane for my N,E,Z coordinates to be computed from, but the Z values will always be perpendicular to that plane.  The N,E values are derived from lat/long and are unaffected by Z, but are affected by initial job height.  Their locations relative to each other will remain correct (250' between points will be 250').

remain correct (250' between points will be 250').

Not really, 250' between points will be 250' +- 250(the PPM correction for the projection at your location). In other words if the combined scale factor of the projection at your location is 0.9997 or 300PPM your distance will be reduced by 3 parts in 10000, or 3 ft in 10,000', .3' in 1000', basically 1 inch in 250'. 

This will happen if you assign 0' to the elevations or 5000'. 

For some surveyors that projection deviation from the real world is "good enough".

I recently sat through "training" that "explained" why not being concerned about that shrinkage is important. 

At least I got some CPE's.

One thing to understand is that you're always surveying on a plane projection. Even when you set up your instrument you're creating a plane. Imagine a 40 acre tract sliding up a big hill. The southwest corner and northeast corner are intervisible but the northeast corner on the hill is 400' (steep tract) above the southwest in the valley. Place the instrument on the southwest corner and you will create a plane level with that instrument, you locate the northeast corner then go occupy it. At that point you create a second plane 400' above the first. This introduces two planes in the survey and they are 400' in elevation apart or 20ppm. Distances measured from the high one will be 20ppm longer than from the lower one. For most purposes that's fine. Cutting across a 40 acre tract is a bit shorter than 1900' so the error is about 1/2 inch. But we always have this effect in our surveys and it's important to understand it.

GPS usually picks a plane and puts everything on it. Then it doesn't matter what the elevations on the ground are because they are projected to the plane. To match ground and plane closely that plane needs to be configured close to the ground. 

@bill93 Not necessarily true, I work in a few counties where our plans are required to be on the NJSPC system and that is started from the beginning of the project through final as-builts.  We are required to provide complete as-built drawings in SP so that they can drop them into their GIS systems.

@chris-bouffard help me out. I assume you are using civil 3d for drafting as most people do. When bringing in the points file from whatever means gps total station etc that are all on state plane and you begin the drafting for as builts. How do you draw on grid projection in civil 3d correctly. I am learning civil 3d so i am a ignorant users. And i don’t get much time in it as i am mostly doing comps and all the other stuff. In TBC where I perform most of my comps I am in whatever I want i just have to pay attention doing cogo on grid vs grnd when reducing that to grid from a plat etc.