GPS Altitude Accuracy

[RV9APlane]

What is the accuracy of the GPS Altitude displayed on the screen(I have D1000's)? Is it normal to have an increasing discrepancy between altitude on the tape versus the GPS Altitude the higher you fly? On the ground it's pretty close but seems to widen as I go higher.

I have all parts of the Dynon system installed(OAT, etc) and had the system certified before my first flight last November. I had the Baro set to the closest airport's ATIS, which was 30.28. I wanted to fly to just under 18,000ft to see how my plane and engine performed. As I approached around 17,450, the GPS Altitude showed 17,995 so I leveled off. Not that I plan to fly that high very often, but I was thinking of the possibility that the margins of altitude spacing are reduced for aircraft flying easterly versus westerly. Thoughts? Should I make an altitude adjustment in the setup screen?

 

[KRviator]

Ignore GPS altitude. It is based on an theorietical model that bears no relation to pressure altitude except in exceptional circumstances. Do not ever fly by GPS altitude unless you have no choice.

In your example, having the baro reference set to the nearest airport and indicating 17,450 it is pertly acceptable to climb another 500'.

 

[RVDan]

At 18,000 ft the Baro reference changes from local barometric pressure to 29.92 to make the altitude reference equal to pressure altitude.  If skimming Class A airspace, might want to be aware of that. 

 

[GalinHdz]

An additional point of consideration is that over the U.S. once you cross 18,000ft all aircraft altimeters must be set to 29.92 regardless of what the local altimeter is.

The important thing is that ALL aircraft operating nearby be at the same altimeter setting so proper separation is maintained. If you are using GPS for altitude and everybody else is using Baro Pressure, you will not be at the correct separation from other traffic. This is potentially very dangerous.

As KRviator stated "Do not ever fly by GPS altitude unless you have no choice."

 

[airguy]

An additional point of consideration is that over the U.S. once you cross 18,000ft all aircraft altimeters must be set to 29.92 regardless of what the local altimeter is.

The important thing is that ALL aircraft operating nearby be at the same altimeter setting so proper separation is maintained. If you are using GPS for altitude and everybody else is using Baro Pressure, you will not be at the correct separation from other traffic. This is potentially very dangerous.

As KRviator stated "Do not ever fly by GPS altitude unless you have no choice."

And flying at 18,000' or above puts you in Class A airspace, positive control IFR only. There is an exceptional to that rule for up to 3000' AGL and above 18,000 but there is no place in the continental US where that is possible.

Up to 17,999' you are legal to be using the current altimeter setting of the nearest reporting station, but you are playing very close to the edge of an airspace violation if you are VFR.

 

[Dynon]

So GPS altitude is "very accurate", but it won't match your barometric altimeter. Here's an article that has a pretty technical description of why, but the bottom line is that all altitudes, whether it's MSL, barometric altitude, or GPS-generated all have to establish their numbers using some model. Those models are different due to what they endeavor to measure and the technology they use to do so.

http://www.esri.com/news/arcuser/0703/geoid1of3.html

 

[dynonsupport]

Do not ever fly by GPS altitude unless you have no choice.

Unless you are trying to avoid cumulogranite, which obeys only GPS altitude and doesn't care a whit about barometric altitude...

Seriously, avoiding things on the ground is more safely done with GPS altitude. This is why we have "safe altitudes" that are 1-2K higher than the actual ground, since baro can be so inaccurate. Baro altitude is useful for for man-made airspace and traffic avoidance. If we had GPS 100 years ago, we would have never invented baro based altitude at all.

That's not to say that you can't fly safely with Baro, but it's better to know what it's really all based on than just say that GPS is worthless.

 

[RV9APlane]

Great, thanks for all the replies. I had no intention of going above 18,000, that's why I decided to level off at the GPS Altitude of 17,995 in case that was the more accurate altitude at the time. This was mostly a test of my airplane's capability at high altitude than it was a practice session for future high flights. I did love it up there though!

 

[johnsteichen]

It is my understanding, and someone correct me if I am wrong,
WAAS correction of GPS is intended to give great precision of altitude AT GROUND LEVEL for purposes of precision instrument approaches. It is my experience that it is always off by sometimes hundreds of feet at higher altitude when compared to baro altitude

 

[Dynon]

Kind of. A certified GPS is designed to take the GPS model and turn it into an MSL, which should mean it more or less matches when you get close to the ground where the local observation actually reflects the air environment. As you climb, the atmosphere is almost never the "textbook" atmosphere, and so as you climb GPS's "close to perfect" altitude differs from your static-based "as accurate as the atmosphere is" altitude. Ironically, your GPS is probably more "accurate" if you took a tape measure and dropped it to the ground, but for purposes of everybody being in the same relative place, we all use BARO altitude. If EVERYBODY had a WAAS GPS, we could use that to manage all airspace. But since we don't, we use the thing that everyone does have. So as long as everybody has their BARO set the same, and they're all relatively calibrated, from day to day we'll all be at the same altitudes relative to each other. But if you compare using a measuring stick to the ground, 6000 BARO might be 5800 GPS one day and 6200 GPS the next day.

This isn't terribly intuitive, I know.

 

[swatson999]

Another thing to consider is that the accuracy of a GPS "height" (relative to either the WGS84 ellipsoid, some other ellipsoid, geoid, etc.) is going to depend on the calculations done by the receiver manufacturer *after* computing the GPS solution. The raw GPS position calculation is an [X, Y, Z, t] vector in Earth-Centered Earth-Fixed Cartesian coordinates (and hence, pretty unusable for any working purposes). And the accuracy of *that* fix depends in large part on the geometrical arrangement of the satellites used to compute the position and time.

So you have to know the raw error as well as the numerical errors introduced by the coordinate transformations.

It's all pretty interesting from a Numerical Analysis point of view