Garmin 1000 altitude is not geoid corrected?

[peter-h]

One G1000 owner reports that the reading is 150ft off in the southern UK, where 150ft is indeed the geoid variation.

Page 4 here suggests that is as intended.

I find this amazing because every reasonably modern GPS has a means (usually a lookup table) of correcting for this. For example - here and here.

Why would have Garmin disabled that correction on the G1000?

Is it a user configuration option, which may have got mis-set by accident?

This was a known bug in the SIRF-II chips which were used in most GPSs until about 7 years ago. The SIRF-III chip did the correction OK.

 

[TMetzinger]

I've never even looked to see where the GPS-derived altitude in the G1000 is. What page was he looking at?

 

[peter-h]

He refers to the "GPS page".

Garmin 1000 Perspective.

 

[MachFly]

I've never even looked to see where the GPS-derived altitude in the G1000 is. What page was he looking at?

It's on the 3rd auxiliary page, two lines bellow the coordinates.

 

[somorris]

Not a GPS expert, Peter, but my understanding is that it may require more time on station to accurately establish the surface (or msl) elevation than you may be able to get from a mobile station. The geoid elevation is much easier to establish. Maybe that is part of the explanation?

 

[Piloto]

On the G1000 the altitude on the ADI dislpay is always baro altitude, subject to the QNH setting. On the other two GPS you indicate it shows GPS altitude which is independent of QNH setting and is always in agreement with the actual elevation.

José

 

[MachFly]

Page 4 here suggests that is as intended.

Are you referring to this?

WARNING: The altitude calculated by G1000 GPS receivers is geometric height above Mean Sea Level and could
vary significantly from the altitude displayed by pressure altimeters, such as the GDC 74A Air Data Computer,
or other altimeters in aircraft. GPS altitude should never be used for vertical navigation. Always use pressure
altitude displayed by the G1000 PFD or other pressure altimeters in aircraft.

 

[MachFly]

On the G1000 the altitude on the ADI dislpay is always baro altitude, subject to the QNH setting. On the other two GPS you indicate it shows GPS altitude which is independent of QNH setting and is always in agreement with the actual elevation.

José

G1000 is actually capable of showing both, you just can't get your GPS altitude to show on the front page.

 

[flyingron]

Not a GPS expert, Peter, but my understanding is that it may require more time on station to accurately establish the surface (or msl) elevation than you may be able to get from a mobile station. The geoid elevation is much easier to establish. Maybe that is part of the explanation?

Position is position, satellite visibility determiners that vertical position will always be about three times worse than horizontal. However, once that is determined, mapping it to an ellipsoid is just math. Mapping it to a geoid takes math and a geoid database. That used to be a significant thing but now that memory is cheap (it's not really more than a megabyte) it's not a big thing.

 

[peter-h]

I am aware that vertical accuracy on a GPS is not quite as good as horizontal accuracy, etc, but there seems to be a really basic issue here.

Here we have somebody with a G1000 who says his GPS altitude is 150ft off the real altitude, on the ground.

This is simply way way out of line with any even remotely modern GPS performance, never mind a G1000.

But it does correspond to a SIRF-2 GPS chip which, due to a bug, didn't do the geoid correction, so it didn't display real altitude above MSL. Where I am, southern UK, the error is indeed about 150ft.

I have no access to a G1000 so I cannot check it, but I find it hard to believe that Garmin would be selling a GPS with such a defect - regardless of the old warnings about not using GPS altitude etc etc etc. Their handhelds do this correctly, after all (within 10ft on my G496).

As regards how the geoid correction might be done - there are 2 ways I know of. One is a lookup table (which I believe is how every modern GPS does it) and the other is a polynomial which provides a reasonably good "best fit" over the entire earth's surface (which is easy enough to find on google but I have no idea if any manufacturer actually uses it; unless one was exceptionally cunning with fixed point maths, it would require the microcontroller in the GPS module to do 32-bit floating point).

If somebody can confirm their G1000 shows the correct altitude above MSL (within say 20ft) and it does so at a place where the geoid error is known to be way more than that, then it looks like my contact has a defective or misconfigured G1000.

 

[seand]

Of what consquence is the GPS altitude in a G1000 other than curiosity? Is it used for any of the position calculations in vertical navigation or terrain imaging?

 

[L.Adamson]

Of what consquence is the GPS altitude in a G1000 other than curiosity? Is it used for any of the position calculations in vertical navigation or terrain imaging?

If I had to rely on the GPS or barometric pressure, to avoid terrain, I'll take the GPS every time. At least here in the mountain west of the USA & with WAAS. If I need to avoid other aircraft, then it's pressure that's important.
I use a Garmin 696. If I get a current altimeter setting, the 696 panel page and the aircraft's altimeter will exactly follow each other for a few minutes. Then they'll drift apart. It will be the altimeter that's off, due to fluctuations in pressure, and not the GPS. With WAAS, it really is that good. I've been doing this comparison for years now, with the former 496 & current 696 portables.

L.Adamson

 

[gismo]

Of what consquence is the GPS altitude in a G1000 other than curiosity? Is it used for any of the position calculations in vertical navigation or terrain imaging?

GPS altitude is used for any GPS approach with vertical guidance and AFaIK all terrain warning systems available to GA (light) airplanes. A 150 ft error with WAAS correction is very unlikely but without WAAS that's well within the GPS vertical performance standard. If this was in England (where there is no WAAS) it could be perfectly normal.

 

[peter-h]

My experience of flying with GPSs around Europe is that 150ft is definitely not normal, and anywhere near it.

10-20ft is normal.

 

[Henning]

If I had to rely on the GPS or barometric pressure, to avoid terrain, I'll take the GPS every time. At least here in the mountain west of the USA & with WAAS. If I need to avoid other aircraft, then it's pressure that's important.
I use a Garmin 696. If I get a current altimeter setting, the 696 panel page and the aircraft's altimeter will exactly follow each other for a few minutes. Then they'll drift apart. It will be the altimeter that's off, due to fluctuations in pressure, and not the GPS. With WAAS, it really is that good. I've been doing this comparison for years now, with the former 496 & current 696 portables.

L.Adamson

No WAAS in Europe though...

 

[peter-h]

We have EGNOS which is the same thing.

A 430W (e.g.) will use EGNOS in the same way it uses WAAS in the USA.

The result is a small improvement in the vertical accuracy. I have a Garmin 496 which will use EGNOS and it improves it maybe 3x.

 

[olasek]

Why would have Garmin disabled that correction on the G1000?

G1000 is IFR certified, their handhelds are not, so it is perfectly easy to imagine (from the legal standpoint) they did not want you to have access to 'real' GPS altitude to tempt you to do dumb things. They are only responsible for providing you the correct vertical guidance when you shoot WAAS (or EGNOS) approaches, this is where the use of the geometric-GPS altitude is made but it is all behind scenes, you have no access to it.

 

[L.Adamson]

G1000 is IFR certified, their handhelds are not, so it is perfectly easy to imagine (from the legal standpoint) they did not want you to have access to 'real' GPS altitude to tempt you to do dumb things. They are only responsible for providing you the correct vertical guidance when you shoot WAAS (or EGNOS) approaches, this is where the use of the geometric-GPS altitude is made but it is all behind scenes, you have no access to it.

I prefer facts, instead of speculation.

 

[olasek]

I prefer facts, instead of speculation.

You got it above, in post #7.

 

[L.Adamson]

You got it above, in post #7.

Yes, and I responded to post #7. As I said, I'd take the GPS altitude with WAAS, over pressure altitude..........if it meant avoiding terrain; unless I had a last minute altimeter setting from the exact area in question. The GPS based altitude..........really IS that good! It's not good for aircraft seperation, when everyone needs to be on the same page, altitude wise. And that's really, what I see in post #7.
IMO.........Garmin wouldn't be making a distinction between certified panel mounts and portables, as far as the terrain databases are concerned. That's where is see some "guessing" on your part. And I can't say for sure, myself. That's why I'd like to see facts.

L.Adamson

 

[Henning]

I'm more suspicious of GPS altitude. I started using GPS back in the 80s with the first civilian installation on Westward (the owner had the company that made the displays) and have been using it since. Altitude has always been the least reliable feature and I have seen it read several hundred feet of altitude while sitting on the water even in more recent units.

 

[COFlyBoy]

The interface specifcation for GPS users states that the satellite ephemeris is referenced to the WGS 84 ellipsoid model of the Earth. So the resulting altitude calculations are likely actually height above the ellipsoid, not height above MSL.

 

[L.Adamson]

I'm more suspicious of GPS altitude. I started using GPS back in the 80s with the first civilian installation on Westward (the owner had the company that made the displays) and have been using it since. Altitude has always been the least reliable feature and I have seen it read several hundred feet of altitude while sitting on the water even in more recent units.

Once you get moving, I'm seeing very reliable GPS readouts. I've been making comparisons in mountainous areas for quite a long time now. This is been with a Garmin 296, 496 & 696. As previously mentioned, I can get a current altimeter setting, or one from XM weather..........and watch the Garmin's "six pac" panel page altimeter exactly duplicate the airplanes altimeter, for several minutes, before they drift apart. Yet, when the barometric pressure is again current, the aircraft's altimeter once again matches the GPS. This scenario happens over and over. On occasion, they will become several hundred feet apart. I'll attribute that to pressure settings.

And naturally, being a GPS "fanatic" that I am, I'll always be measuring known airport altitudes to that of the GPS. Once again, WAAS is included with my "tests".

edit: My tests never get within 2000' of sea level......

L.Adamson

 

[peter-h]

Apparently the G1000 generates a log which can be downloaded.

This has 3 columns:

1 - baro altitude
2 - geoid-corrected GPS altitude
3 - uncorrected GPS altitude

It has been suggested (I am still hoping somebody can verify this) that the GPS altitude displayed by the G1000 is 3 and not 2, which (if true) would be different from perhaps every other modern GPS, handheld or panel mounted.

 

[N2124v]

Errors in altitude can be magnified by the location of the satellites the gps is receiving signals from. If they are low to the horizon that can create issues.

 

[L.Adamson]

Errors in altitude can be magnified by the location of the satellites the gps is receiving signals from. If they are low to the horizon that can create issues.

Original topic is for the U.K. area, which "may not" directly apply to what I'm saying. With WAAS, and eight to eleven satellites to choose from at any one time, ...............this altitude error won't be existing. If this was the case, GPS approaches in IMC, would still be just a thought for the "future".

 

[John Collins]

Peter,

From everything I have read and playing with the simulator, the G1000 displays GPS - MSL altitude on the Satellite page. It is displayed to the nearest foot. The same value is also displayed to the nearest 20 feet on the Terrain page when you have the aviation data activated. On the satellite page it is just labeled Altitude, but on the Terrain page it is labeled GPS MSL Altitude. It can also be viewed as a white line on the terrain legend when on the Map page. You might recognize the location where I did the simulation.



Note the GPS altitude marker on the terrain legend, I circled it in yellow

 

[Henning]

Original topic is for the U.K. area, which "may not" directly apply to what I'm saying. With WAAS, and eight to eleven satellites to choose from at any one time, ...............this altitude error won't be existing. If this was the case, GPS approaches in IMC, would still be just a thought for the "future".

You speak with much faith in the box filled with magic smoke...

 

[peter-h]

John,

Many thanks. Yes I recognise the location

However EGKA is about 10ft above MSL, and every GPS I have played with (except the known-defective ones) show it as that (plus or minus say 20ft) so the G1000 should not be showing 120+ ft as shown in your second image.

That's if I understand you correctly.

 

[John Collins]

Peter,

It is reading 13000 feet, not 120 feet, compare the MFD with the PFD, the three graphics of the simulator are at the same place and altitude . If I adjust the trainer to the database reported field elevation of 7 feet MSL, the altitude on the GPS page reads 0 feet, the terrain page shows 0 feet, the white line on the legend for the MFD Map page is on the brown - blue boundary, and the baro altimeter reads -0 feet.

 

[peter-h]

I wonder how accurate the simulator is... would they really simulate the NMEA data coming from the GPS, and that data coming from a piece of software simulating the GPS receiver?

 

[John Collins]

I wonder how accurate the simulator is... would they really simulate the NMEA data coming from the GPS, and that data coming from a piece of software simulating the GPS receiver?

It is pretty representative, I guess that it uses the same control code as a version of the real box and the GPS position information is simulated. I don't understand the reference to NMEA data, isn't this a serial output format used by some GPS units to interface with external devices?

 

[peter-h]

Yes; NMEA data is what come out of the GPS receiver module. I would think the G1000 uses a separate module, as do most panel mount GPSs.

 

[olasek]

I wonder how accurate the simulator is... would they really simulate the NMEA data coming from the GPS, and that data coming from a piece of software simulating the GPS receiver?

Very good question but I think John is mistaken in thinking there is any GPS simulation going on in the trainer.
I also have the trainer and I checked three vastly different locations in the world and they all gave me absolutely identical image of satellite positions/strengths, etc. In my opinion the trainer does absolutely nothing to simulate the GPS signal/satellites, errors, etc.

 

[peter-h]

Interesting... I also would not expect the trainer to simulate the GPS receiver performance/accuracy.

I have a feeling that perhaps I have not explained the issue very well. It is nothing to do with the fact that VNAV (altitude) GPS accuracy is worse than LNAV accuracy (nobody disputes that).

It is a very specific question as to whether (and if so, why) the G1000 displays (in the GPS page(s)) what appears to be the initial WGS84 ellipsoid altitude figure, rather than the figure corrected for the geoid variation.

The former is basically useless, because (even with WAAS/EGNOS) it will be anything up to a few hundred feet out.

To establish the answer, somebody will have to test a real G1000, at a location where the geoid variation is known to be substantial (say at least 100ft), and see what altitude it reads.

It will read either the correct altitude (within say 10-20ft) or it will read the faulty one and that will be quite obvious because the error will be equal to the geoid variation.

 

[TMetzinger]

I still don't understand, Peter, why you care what the GPS altitude displayed is? It's not a figure a pilot would use for any in-flight decisions, is it?

 

[John Collins]

Very good question but I think John is mistaken in thinking there is any GPS simulation going on in the trainer.
I also have the trainer and I checked three vastly different locations in the world and they all gave me absolutely identical image of satellite positions/strengths, etc. In my opinion the trainer does absolutely nothing to simulate the GPS signal/satellites, errors, etc.

I seriously doubt it is simulating the GPS constellation, but that it is simulating the output of the GPS reciever function with a 3D position and other related stuff that normally comes from the GPS receiver board. All I demonstrated to myself was the relationship between the altitude field on the GPS page and the terrain page were the same value, which is identified as GPS MSL Altitude, or the raw GPS value corrected to the Geiod. As Peter indicated, there is no use for displaying the WGS84 value.

 

[John Collins]

I still don't understand, Peter, why you care what the GPS altitude displayed is? It's not a figure a pilot would use for any in-flight decisions, is it?

It is a backup to the pitot system failure and useful for terrain due to its independence from non standard pressure lapse rate and temperature effects.

 

[John Collins]

Interesting... I also would not expect the trainer to simulate the GPS receiver performance/accuracy.

I have a feeling that perhaps I have not explained the issue very well. It is nothing to do with the fact that VNAV (altitude) GPS accuracy is worse than LNAV accuracy (nobody disputes that).

It is a very specific question as to whether (and if so, why) the G1000 displays (in the GPS page(s)) what appears to be the initial WGS84 ellipsoid altitude figure, rather than the figure corrected for the geoid variation.

The former is basically useless, because (even with WAAS/EGNOS) it will be anything up to a few hundred feet out.

To establish the answer, somebody will have to test a real G1000, at a location where the geoid variation is known to be substantial (say at least 100ft), and see what altitude it reads.

It will read either the correct altitude (within say 10-20ft) or it will read the faulty one and that will be quite obvious because the error will be equal to the geoid variation.

IMHO there is no way it is the WGS84 uncorrected value to the Geiod. There is always a possibility that there is an error in the correction database.

 

[L.Adamson]

I still don't understand, Peter, why you care what the GPS altitude displayed is? It's not a figure a pilot would use for any in-flight decisions, is it?

I suppose the important part of knowing GPS altitude, is in making comparisons to the altimeter altitude. Since GPS altitude is the one, that will be incorporated with synthetic vision, then it's of value to know that it's usually the more correct & actual altitude (depending on how current your altimeter setting is). Other than that, it's GPS altitude that determines the terrain & obstacle warnings.

L. Adamson