Altitude, just to clarify:

[cheddarman]

P3A drone (not that it matters).

There is a small hill about ½ mile to the north of our village with a height of about 770 feet above sea level.
Most of the village is at 100ft above sea level.

If I take the drone up onto the hill, and take off from there, and go up to the legal height of 400 feet, I will be at 1170 feet above sea level.
If I then fly south towards the village, as the drone clears the hill, will it drop down and follow the contours at 400feet above the ground or will it stay at 1170 feet as per the barometric pressure, or does it depend on which mode one is flying in?

 

[msinger]

will it drop down and follow the contours at 400feet above the ground or will it stay at 1170 feet as per the barometric pressure

The drone will follow your command. If you continue to fly straight forward, it will not descend (or ascend).

 

[cheddarman]

That's what I though Mike, thanks.

So before I take off from the hill, will the App register 0 or the 770 above sea level due to the barometer?

 

[MapMaker53]

Elevation is measured in feet (or meters) relative to sea level. Your drone shows altitude, which is measured in feet (or meters) relative to land surface. The drone's barometer is set to zero when the motors start so the reading (altitude) is actually relative to your initial launch surface -- which is the important part. You can ascend 400 ft straight up from the top of that hill and be legal. But if you venture away from the hilltop horizontally in a straight line, even though you will still be only 400 ft above your takeoff point and your dashboard will show you at 400 ft altitude (seems legal), you would now be flying increasingly higher than 400 ft altitude above the ground as the land surface slopes away -- and thus would be illegal.

Similarly, were you to start your motors at the base of the hill, the barometer would register that as zero altitude. If you were to then fly up the slope of that hill and reach the summit at 670 ft, your dashboard would show you at 670 ft altitude (seems illegal), but you would be legal because you would not be greater than 400 ft above the top of the hill's land surface.

 

[cheddarman]

Many thanks, great description! I hadn't taken into account that the barometer always resets to zero on motor start!±

Cheers

 

[Meta4]

So before I take off from the hill, will the App register 0 or the 770 above sea level due to the barometer?

Your app never has any idea of what or where sea level is.
And the drone has no way to measure its height above sea level or above the ground below it (except the very short range VPS sensors)

 

[pdqdl]

So long as we are on the topic of altitude...

Using the DJI app, if I were to launch from the same hill described above to film something along the seashore, with my battery subsequently telling me it is to come home, what happens to altitude using RTH?

I have my drone set to fly back on autopilot at 200' altutude, but I have no idea how it handles rises & falls in terrain. I have always presumed that it uses 3 dimensional GPS to keep track of altitude, but I don't know how well it's internal maps know the geography.

 

[cheddarman]

Good question. I also have my RTH altitude set at 200'. Would it fly into the hill or climb up following the contour??

 

[pdqdl]

I'm not too worried about the hill, since I presume object avoidance would take care of that. 100' tall trees (30 meters) become a much bigger problem however, since object avoidance doesn't work too good against trees.

I've already climbed one tree to get the P4P down, and I don't plan on repeating that little exercise.

 

[pdqdl]

My presumption is that it records the altitude of the launch point, and then rises to that altitude prior to navigating. That might work very poorly in almost anything but relatively level ground. Either the machine might end up flying far above max altitude, or it might "rise" into the ground if located at a higher altitude when the command was given.

Now I have always presumed the system was smarter than this, but I sure don't know how much smarter that might prove to be.

 

[Meta4]

I have my drone set to fly back on autopilot at 200' altutude, but I have no idea how it handles rises & falls in terrain. I have always presumed that it uses 3 dimensional GPS to keep track of altitude, but I don't know how well it's internal maps know the geography.

GPS is not used at all for altitude sensing.
Your drone has no sensor that can detect the terrain below the drone or know how far away it is. (except for the very short range VPS sensors).
The internal maps don't know the geography at all.

The altitude data comes from the barometric sensor in the IMU.
It measures changes in air pressure, which just gives the altitude relative to the launch point.

 

[Meta4]

Now I have always presumed the system was smarter than this, but I sure don't know how much smarter that might prove to be.

Just like real planes, the drone relies on the pilot having some smartness.

 

[pdqdl]

Relying on this pilot's smartness ended up parked in a tree once.

I guess my reliance on RTH is going to be a bit more cautious. Fortunately, we don't have too many big elevation changes in my part of the world.

 

[CeeBeePh4]

I agree that Your drone measure from the starting point wherever It is. So if You start at the top of the hill and let the drone fly to the village You are not flying legal when You are over the village if You haven't reduced the height to minus 270 ft.
I have only a little remark. The figure You will see on Your screen Is the height. Among pilots we use "altitude" as height over sea level. When flying an airplane You use three ways to tell how high You fly. If You set Your altimeter on the actual barometer pressure (given from the tower) You read how high over the see level (MSL) Your airplane is. If You put Your altimeter on zero You know how high over the airfield You fly (AGL- above ground level) and finally if You put Your altimeter at 1015 You will see Your height over a standard level.
Your drones altimeter always is put on zero when starting and measures AGL.
Sorry for a. long explanation why using height is better than altitude.

 

[pdqdl]

That's all well and good, and I understand AGL perhaps a bit better. What remains unaddressed is how DJI handles altitude differences upon receiving an RTH command.

For example, up until today I didn't even know the P4P had a barometer. I thought that it would be far more accurate if measured with GPS, although that is admittedly a problem if you lose signal.

My fear is that using RTH will just dial in a crash or illegal altitude if there are significant ground level altitude differences along the flight path.

 

[Meta4]

What remains unaddressed is how DJI handles altitude differences upon receiving an RTH command

There's no mystery to the process and it's very well understood by most pilots.
Your drone does not handle variations in terrain heights when in RTH.
It cannot because it has no way to sense the terrain under it.

In RTH your drone goes to the height you have set and returns home at that height.
It's up to you to understand how this works and any limitations there may be in your local environment.
Ten minutes of test flying should give you a good understanding of how it works.

.

For example, up until today I didn't even know the P4P had a barometer. I thought that it would be far more accurate if measured with GPS, although that is admittedly a problem if you lose signal.

1. GPS would be less accurate, which is why your drone uses a barometric sensor, just like real planes.
2. GPS would only give a height above a theoretical sea level, not a height above the terrain.
3. Losing signal wouldn't make any difference

My fear is that using RTH will just dial in a crash or illegal altitude if there are significant ground level altitude differences along the flight path.

Then you need to do what everyone else does and set an appropriate RTH height to clear any obstacles that might be between you and any obstacles where you fly.
It's not complicated at all.

 

[pdqdl]

Thanks. That clarifies quite a bit.

That being said, "testing" isn't quite so straight forward. Since we were talking about the response to large terrain changes under RTH flight, I couldn't possibly test that. I don't have any large terrain changes. About the biggest hill we have in this area is 150' tall (I checked using actual maps with contour lines), so I can't possibly evaluate how it might respond to a 700' increase in terrain altitude. At least not while remaining within sight distance, so as to terminate the flight before a crash occurs.

Furthermore, designing a test that calls for the drone flying itself into the ground or into illegal airspace doesn't exactly sound like a good plan. It seems to me like a better idea would be to discuss the problem with folks that might know the answer.

I must differ with you, however, as to whether GPS can be more more accurate than a barometer. I'm not suggesting that our little drones necessarily are that accurate, but the surveyors that mark out our properties seem to think that the GPS signals are quite accurate, both with respect to longitude and latitude, as well as altitude. I went on a flight with a guy in a little Cessna 20 years ago, and he was quite proud of his GPS device that revealed his exact altitude and position, and projected that upon a map of known terrain altitude and registered objects that were much taller. So it is possible...

 

[Meta4]

so I can't possibly evaluate how it might respond to a 700' increase in terrain altitude.

Testing would show that it's just as I said.
Your drone does not do anything to respond to any change in altitude.
It just comes straight home in level flight.

I must differ with you, however, as to whether GPS can be more more accurate than a barometer. I'm not suggesting that our little drones necessarily are that accurate, but the surveyors that mark out our properties seem to think that the GPS signals are quite accurate, both with respect to longitude and latitude, as well as altitude.

Your surveyors use survey grade differential GPS which cost them thousands of dollars and requires special training to use.
Your drone has civilian grade GPS which is nowhere near as accurate.
Your barometric sensor is sensitive to altitude changes of a couple of inches, your drone's GPS is not
And your drone has no way to compare height above a theoretical sea level with terrain height to let you know how far your drone is above or below the ground.

 

[pdqdl]

I see that you deleted my closing sentence, thereby dodging the entire point of my post.

"So it is possible..."

You also completely dodged the point of my mental exercise concerning testing. Pushing the RTH button and then watching what happens isn't a test. A test is putting the drone in adverse circumstances and then observing what happens, hopefully under conditions that lead to knowledge acquisition and catastrophe avoidance. Repeating the mantra of "do a test" doesn't address any of my previous points.

That's ok. I get the point. It will crash if conditions include running into the ground or other obstructions at the preset altitude if not otherwise detected.

 

[Meta4]

Repeating the mantra of "do a test" doesn't address any of my previous points.

Everything in your previous posts was addressed back in post #6 where I said:

the drone has no way to measure its height above sea level or above the ground below it.​

But that message didn't seem to get through.
Your drone doesn't respond to terrain height differences in normal flight so why you would expect it to in RTH.