Actual Altitude

[fangzz]

My altimeter and mode c agree very closely when I ask for a mode c altitude check from atc, and I have the local altimeter set. But my GPS altitude is unacceptably off: 300-1500 feet, depending on altitude. And, (and here's the kicker) I took more than one GPS, and they agreed from 0-10 feet in altitude.

I'm very inclined to believe the gps altitude, but some local feedback tells me there's nothing wrong with my mode c signal. I'm dubious, and, to me, this is too important an area to not have right-on, even tho I'm severe VFR pilot. So, while I do more checking 'round here...

One or more of you folks that know this area ( calibration, source of signal, etc) might have time for a short tutorial, which I would appreciate.

Weigh in please. Drew

[Bill Hart]

Your mode C is a pressure altitude and is corrected by ATC. (Unless you have an encoding altimeter. Not likely in a C-170.)Your altimeter is corrected for barometric differences by using the baro knob in the cloesman window. The difference between your altimeter and the mode C encoder can be no more than 125 feet different according to FAR43 appendix E. (That is if it has been checked with in the last 24 calendar months)

The read out from the GPS is a calculation of time between satellites and your unit while they may be extremely accurate what ATC sees is what you need to be flying for separations sake. So if your altimeter and encoder are with in 125 feet of each other and if you ask ATC and they have the same thing then you are good.

[ronjenx]

It has been my experience most of the error I see in handheld GPS is altitude.
I have done lots of hiking with GPS and topo maps.
GPS location is very close. Altitude constantly varies up and down quite a bit.
I have a hand held barometer/altimeter that is quite accurate for locating myself on a topo, without the aid of a GPS.

Anyway, it doesn't surprise me the GPS altitude would be off, and both of yours would agree, since they see the same data.

[Robert Eilers]

There are several types of altitude definitions. Which definition is most important depends on what you are doing at the time, i.e., absolute, density, pressure, true and indicated. When talking to ATC, what they see is indicated altitude, corrected for barometric pressure, plus or minus your encoder error. I suggest to my students and friends that it is important to have some idea of their encoder error (if any), and when negotiating Class B airspace to give themselves a margin of at least 100 ft. to avoid being busted by an encoder error. Thye may be actually flyng beneath the floor of Class B airspace and have an encoder that shows them within Class B airspace.

[markeg1964]

My old Garmin 195 would regularly report incorrect altitudes. I have read this is normal. I don’t remember how far off it was but it was almost always more than 200 feet. My Airmap 2000 is usually within 50 feet or so although I have seen it off more than 100 feet.
Here is an article I found that does a good job of explaining it:

http://docs.controlvision.com/pages/gps_altimetry.php

[Blue4]

I'm mistrustful of GPS altitude unless the unit is WAAS enabled. Its the WAAS that permits us to fly near-precision approaches GPS-only; prior to that it was non-precision only using the barometric altimeter.

In short, there's likely nothing wrong with your GPS nor with your Mode C and altimeter. Unaided GPS altitude is simply inaccurate in my experience.

[fangzz]

Thanks for all that input. It completely gibes with my recent experience. But, for the money, I would sure think my Garmin 496 would do better, especially when waas-enabled, which it is, and has been during these 'excursions.'

Thanks for the link to the article, Marc. Very helpful. I think I can now quit chasing my tail.

Thank you all. Drew

[hungstart]

Remember your GPS is absolute altitude rather than pressure altitude.

Rick C
N170MB

[fangzz]

Don't think so, Rick. Absolute altitude being one's height above ground level where ever one happens to be. As Ron Jencs pointed out earlier, gps is very useful, as I can attest, while hiking and camping for locating one's height above sea level on the side of a mountain. If it were absolute altitude, you would always read 'zero' on the ground.

One would always hope that your gps altitude would read 'true altitude', which is what pressure altitude and indicated altitude always want to be when they grow up. Alas, my expensive (for me) Garmin 496 can't seem to find its butt with both hands, height-wise. Even waas-enabled. Even after I sent it back to Garmin to fix/replace (they replaced). Not one bit of difference. So, the thought of flying gps-approaches (which I don't do since my ifr days are long over) is a scary thought, indeed. I do realize that a handheld must be very different in this way from an integrated panel mounted gps. That has to be the difference here.

[hilltop170]

I could be wrong but on a GPS WAAS approach with vertical guidance, isn't the altitude a calculated height based on what altitude you should be at based on your current position?

[jrenwick]

I could be wrong but on a GPS WAAS approach with vertical guidance, isn't the altitude a calculated height based on what altitude you should be at based on your current position?

That's the definition of a glide slope. The GPS still has to know where you are in relation to the calculated slope. I think there's a lot more electronics and logic in a Garmin 430W vs. a 496, related to vertical position. The 430W won't authorize you to use the VNAV unless it's receiving a sufficient number of GPS satellites plus the WAAS satellite, and the geometry of the satellite positions is good enough to put the expected error within limits. It also knows the altitude your Mode C is reporting, and that becomes part of the solution somehow. There's a lot of black magic behind those WAAS approaches.

[GAHorn]

John hit the nail on the head....the IFR panel mounted unit reads altitude from satellites and compares it to your mode C broadcast. Remember that satellites alone...are almost NEVer directly overhead. They are off to the side, sometimes barely above the horizon. Therefore an altitude calculation/triangulation must not only be made from several satellites...they must often be calculated at extreme angles. (Imagine trying to guess an airplane's altitude as it's just slightly above the horizon 150 miles away. It may LOOK like it's only 100-feet above sea level...but it COULD be 30 thousand feet up! That's the problem of determining altitude from orbiting satellites. When you consider that...it's pretty amazing they can be within thousands of feet, much less the few hundred feet errors typical.)

[hungstart]

Don't think so, Rick. Absolute altitude being one's height above ground level where ever one happens to be. As Ron Jencs pointed out earlier, gps is very useful, as I can attest, while hiking and camping for locating one's height above sea level on the side of a mountain. If it were absolute altitude, you would always read 'zero' on the ground.

One would always hope that your gps altitude would read 'true altitude', which is what pressure altitude and indicated altitude always want to be when they grow up. Alas, my expensive (for me) Garmin 496 can't seem to find its butt with both hands, height-wise. Even waas-enabled. Even after I sent it back to Garmin to fix/replace (they replaced). Not one bit of difference. So, the thought of flying gps-approaches (which I don't do since my ifr days are long over) is a scary thought, indeed. I do realize that a handheld must be very different in this way from an integrated panel mounted gps. That has to be the difference here.

I stand corrected, what I meant was actual altitude MSL, as your altimeter is seeing pressure altitude. Based upon actual verifiable altitudes that I have been able compare my 496 to it has always had less than a 50 foot error. At times it has been right on. Most of the time I'm seeing at least 8 satellites most of the time more, with 17 foot accuracy or less.

Rick C

[GAHorn]

[...At times it has been right on. Most of the time I'm seeing at least 8 satellites most of the time more, with 17 foot accuracy or less.

Rick C

That "17 foot accuracy" is relative to a geographical point ON the earth's surface (datum) ...not distance (altitude) ABOVE the earth.

[cessna170bdriver]

The way I understand how GPS calculates position is that it measures the "time of flight" (within some known error band) of the signal from each satellite, which can mathematically can be converted to distance. Knowing the distance to one satellite puts you somewhere within a known (close) distance of the surface of a sphere centered on that satellite, which knows where in space it is.

The distance to a second satellite puts you very near the surface of a second sphere, so now it knows that you are somewhere near the intersection of two spheres, which happens to be a circle in most cases. (Think of it as being near where two bubbles touch.)

The distance to a third satellite puts you near one of the two places where the circle determined by the first two intersects the sphere of the third. The distance to a fourth satellite determines which of the two circle-sphere intersections you are near. Thus the requirement (as far as I know) for the receiver to be tracking four satellites to solve a 3-d position.

The fact that there is a slight error associated with the distance to each satellite puts you somewhere within a small 8-sided volume, the odd shape of which is determined by the random relative positions and position errors of the four satellites. Additonal satellites can cut down the size of this volume. Knowing the shape of this "error volume" the receiver could calculate the error in any direction: horizontal, vertical, or otherwise, as determined by the software in the particular unit.

In my experience, mostly with a Garmin 296, the vertical solution seems to be slightly better than the horizontal solution, at least around Tehachapi. Landing on the centerline of runway 29 leaves a track as though I had landed on the taxiway about a hundred feet to the south. I don't recall ever seeing the calculated elevation on the ground being off by more than 20 feet. Not too bad for a relatively inexpensive, "don't-you-dare-use-this-thing-in-the-clouds" unit.