GPS Altitude

[cbretana]

Is the GPS altitude displayed by the D3 with respect to MSL? or above a statandarized earth ellipsoid (specifically WGS84) like other non-aviation commercial GPS receivers, (cell phones etc.)

And more specifically, how accurate is it? Does it use WAAS (if it's available?)

 

[Rhino]

Presumably AGL, but I'd ask Dynon to confirm that.

I've seen nothing indication the D3 uses WAAS, and I doubt it does, but Dynon could confirm that too. No GPS derived altitude can be 100% accurate 100% of the time, due to several factors. Rely on your static system, and only use GPS altitude as a last resort.

 

[FORANE]

No, it's not agl. It's MSL or above standard elipsoid. Comparing various pics in flight I see it's typically 100-400 higher than the altitude displayed on the HDX.

 

[cbretana]

"No GPS derived altitude can be 100% accurate 100% of the time, due to several factors..." Of course. Nothing can be 100% accurate 100% of the time, even if it's from the most accurate instrument created by mankind in all of history. Is that really what you thought I was asking for?

No, I was asking about what the accuracy actually is. Is it plus or minus 5 feet?, 100 feet?, 500 feet? or 10,000 feet? If you don't know, then just say so. As to your advice about "Rely on your static system, and only use GPS altitude as a last resort.", it's clear you don't understand why I am asking this. I would have thought that the topic being about the D3 would have made that obvious. Let me explain.

So you understand. I do not have a backup altimeter. I am installing the D3 as a backup for the unlikely scenario that I lose all electrical power in IMC conditions. In that case my main Dynon HDX screen (which is where my static altitude would be displayed), might no longer be available. In that (admittedly unlikely event) it might be useful to know how the D3's GPS altitude works and how accurate it is.

And I know it's not AGL, I didn't ask that. And I can compare HDX static altitude with the D3's GPS altitude myself. In the unlikely event that I experience complete electrical failure in IMC, I would like to know what the maximum possible error that the D3 (functioning properly) can be. I don't want to rely on anecdotal recollections.

 

[Rhino]

I wasn't being quite that literal. Nobody knows for sure what the answer is. Because of the different factors affecting GPS accuracy, there is no way to predict maximum error. It changes with location, installed equipment, atmospherics, clock errors, rounding errors, etc. As a backup during a failure or emergency? Sure, I would take it, especially VFR.

 

[Marc_J._Zeitlin]

... I would like to know what the maximum possible error that the D3 (functioning properly) can be. I don't want to rely on anecdotal recollections.

So this will depend on what you're defining "error" to be. If you want to know how far off a non-WAAS (since the D3 docs say nothing about WAAS) GPS can be vertically from the ACTUAL GPS altitude above whatever ellipsoid is being used as the surface reference, this:

GPS Performances - Navipedia

seems to indicate that it could be <156 meters with 95% certainty. And that kind of matches up with the maximum errors I've seen on non-WAAS GPSs of about 200 - 400 ft.

However, if what you define the error to be the difference between the D3's displayed altitude and a barometric altitude, well, there's no way to know that, because the altimeter setting will have a large effect on what is displayed, and the WGS-84 ellipsoid and MSL are NOT the same thing. However, I don't think I've ever seen a difference between a non-WAAS GPS altitude and a barometric altitude of more than 500 ft., and with a WAAS GPS, no more than 200 ft. or so.

So if _I_ got stuck in IMC with only the non-WAAS GPS to keep me from hitting something, I'd add 500 ft. to whatever altitude I was supposed to be at (which already theoretically has margin built into it) and fly there (IOW, if I was assigned 8,000 ft., I'd climb to the D3's 8,500 ft.).

But if I actually only had one large EFIS screen and one ADAHRS to feed it and was going to fly IFR, I wouldn't use a D3 as the backup - I'd get a G5, GI-275 or an AV-30, so that I had a FULL EFIS with air data, albeit without a moving map.

 

[cbretana]

Rhino,

Yes, you are correct; all the points you mention are true. They are, obviously, true for every altitude sensor, regardless of the technology. The FAA nevertheless manages to publish reliable maximum allowable errors for all altimeter technologies, including GPS, (to determine minimum Decision Heights or MDAs for published approaches when using GPS measured altitudes), etc. etc.

So people do know (to a predictable level of statistical certainty), how accurate they are. Even static altimeters have maximum allowable errors (we are supposed to check them before we fly and have them calibrated and inspected every few years to ensure they are still accurate to within those tolerances (plus or minus 75 feet when the right altimeter setting is set).

So all I am asking for is how does the D3 do this, so I have a rough idea as to how much I can trust it. If I'm IMC, and I know approximately where I am, it would help to know how low I can descend trying to break out below an overcast without running into the ground.

Nevertheless, in spite of perhaps giving the impression that I do not appreciate your contribution to this thread, I do appreciate (and enjoy participating) and if I have given any other impression I regret it.

 

[cbretana]

Marc,

Yes, this is exactly what I was interested in. The answer (< 156 m) is larger than I would like, but it answers my question.

On the other note, .. according to my reference, GPS systems use the WGS84 ellipsoid. But this must be implemented in the algorithm that determines altitude above some reference surface after the algorithm that calculates position in 3D space from the time delays. Both these algorithms must be implemented in each individual GPS receiver.

But I have not found any reference in my research that specifies/requires that any standard is required for all GPS receivers, although I would suspect it is.

 

[Rhino]

So people do know (to a predictable level of statistical certainty), how accurate they are. Even static altimeters have maximum allowable errors (we are supposed to check them before we fly and have them calibrated and inspected every few years to ensure they are still accurate to within those tolerances (plus or minus 75 feet when the right altimeter setting is set).

Theoretically, yes, but those measurements assume everything works perfectly, and conditions never change. They don't take into account the factors I mentioned. What you really want to know is how much variability might be there from any set standard. I can't answer that.

As for system approvals, the FAA implements several different standards, under different TSOs, for different systems and usage, such as VFR or IFR, etc. That's discussed to a certain extent here. But if you read the intro, they also discuss variables like the ones I mentioned.

 

[cbretana]

Rhino,

I think we can agree on this fine point. We have drifted here into the realm of Philosophical Epistemology. In the real world, no measurement is perfect, and things always change. The world is not black and white. This is always the case, and it will always be the case, we assume it and it need not be stated. So we ignore that, and state things as though they are known to a degree of precision beyond what is justifiable. I would do the same here, but your statement that "those measurements assume everything works perfectly, and conditions never change" impels me to state the obvious. Every measurable quantity is on a scale, where the absolute exact value is unknown. Measurements do NOT assume everything works perfectly. All we can say with perfect certainty is that the true value is within some range of values, with a probability between 0 and 100%. We strive to make that range as small as possible, so that the measurement will be useful to some degree. There's an entire field of study (Measurement and Error Analysis), that analyzes this.

But you are incorrect in "What you really want to know is how much variability might be there from any set standard. ". I just want to know how well it works, to just that sufficient level of precision that allows me to make intelligent, useful decisions, without either assuming it works perfectly or that it doesn't work at all.

 

[cbretana]

But if I actually only had one large EFIS screen and one ADAHRS to feed it and was going to fly IFR, I wouldn't use a D3 as the backup - I'd get a G5, GI-275 or an AV-30, so that I had a FULL EFIS with air data, albeit without a moving map.

In principle, I agree. I would also look at the Dynon D30, which is similar to and more of a competitor to the Garmin G5.

However, when I originally planned my new panel, I compared all the options, and based on price and end goals, decided on the D3.
My reasoning:
I have decided, as a policy, to fly in IMC only when I absolutely have to. I intend to only fly when the forecast weather for the majority of the flight is VMC and the forecast ceilings along the route of flight are VMC/VFR. I often fly west from AZ to the coast to visit family, and the weather there is often low clouds from the marine layer, but VMC above 3-4000 MSL. If I get stuck above an undercast and the electricity and the EFIS screen go away I want to be able to let down to where I can break out under the weather without having to look for a sucker hole. I carry multiple wide screen EFB devices (with iFly EFB) that I can use for VMC navigation. The Dynon D3 gives me that capability at the lowest price point. Aircraft Spruce recently reduced the D3 to $800. Compared to the prices of the more expensive D30 and the Garmin G5, the D3 provides that capability and is a bargain. If I win the lottery, I would likely spring for one of the more capable units and put it in the same space in my panel.

 

[Rhino]

Rhino,

I think we can agree on this fine point. We have drifted here into the realm of Philosophical Epistemology....

Yeah. You could say that.

But you are incorrect in "What you really want to know is how much variability might be there from any set standard. ". I just want to know how well it works, to just that sufficient level of precision that allows me to make intelligent, useful decisions, without either assuming it works perfectly or that it doesn't work at all.

Kinda the same thing. But yeah. That's what I was getting at.