More on Altitude issues

[Meta4]

My friend with the Air2s had a similar issue a few weeks ago, except his started going up with no joystick input, likely caused by false barometer readings?

That's unlikely, obstacle avoidance is the most likely cause.
But the cause should be easily identified with a look at the recorded flight data.
Barometric sensor malfunction hasn't been a problem with DJI drones before the M3.

 

[sar104]

That's unlikely, obstacle avoidance is the most likely cause.
But the cause should be easily identified with a look at the recorded flight data.
Barometric sensor malfunction hasn't been a problem with DJI drones before the M3.

I really doubt this is a problem with the barometer. It does appear that the M3 incorporates visual sensor data more dominantly into a number of parameters, including altitude. While I applaud their cutting edge development, it does seem that there are still some issues to be ironed out.

 

[dmcgrew]

No-one flew into a bridge.
The OP was obviously looking where he was flying, all the more so because he was aware of the altitude problem that his drone was exhibiting.
The bridge is a handy way of illustrating the issue, since its height is known.

I know that.. but the OP said " Well I don't think it's harmless if I fly into a structure because I think I'm at 80 and I'm at 150."

 

[Speed]

So if its barometric pressure... wouldn't that effect the altimeter of other drones in that location?
If you have the ability I would put another drone in the air. If its altimeter readings are correct this would completely discount DJI's barometric pressure claim. It may also point directly to a M3 platform issue.

As others have mentioned....I'm having a hard time with DJI's barometric pressure claim.

 

[Meta4]

I know that.. but the OP said " Well I don't think it's harmless if I fly into a structure because I think I'm at 80 and I'm at 150."

You misunderstood what he was saying.

 

[Meta4]

So if its barometric pressure... wouldn't that effect the altimeter of other drones in that location?
If you have the ability I would put another drone in the air. If its altimeter readings are correct this would completely discount DJI's barometric pressure claim. It may also point directly to a M3 platform issue.

There's no need to do that.
DJI drones have been flying with barometric sensors for many years and the sensors have been very reliable.
A simple look at Brian's data is all that's needed to see that his issue has nothing to do with air pressure changing during the flight.
That might account for a 10 ft difference perhaps, but not >100 ft.

 

[Phantomrain.org]

After much playing around I've found that the drone performs 90% better when sensors are TURNED OFF. I've said all along this is a sensor issue. When flying at dusk the sensors are switching on and off constantly as light changes and it's causing the stops and starts and declines in altitude. Tonight I few 1/2 hour after sunrise so getting pretty dark and did not have a single anomaly.

I just realized that all my flights with the Mavic 3 have been at night and thus why I was not having any issues. Now I see says the blind man.

Phantomrain.org
Gear to fly the Mavic 3 In the Rain . and Land on the Water.

 

[Sky Angels]

There's no need to do that.
DJI drones have been flying with barometric sensors for many years and the sensors have been very reliable.
A simple look at Brian's data is all that's needed to see that his issue has nothing to do with air pressure changing during the flight.
That might account for a 10 ft difference perhaps, but not >100 ft.

With over 100 DJI Drones I have never see a correct altitude number on any DJI craft. Except enterprise solutions. As soon as you fly below take off altitude it’s screwed up. Come back up and you must rely on the camera and VPS system (dual sonar sensors).
It will NEVER be perfect unless you rely on VPS and LOS visual checks. Or a RTK unit.

 

[sar104]

With over 100 DJI Drones I have never see a correct altitude number on any DJI craft.

It depends on what you mean by "correct". Deviations of a few meters are perfectly normal for fused barometric/inertial solutions.

Except enterprise solutions.

Maybe - if the FC is using RTK for altitude correction.

As soon as you fly below take off altitude it’s screwed up. Come back up and you must rely on the camera and VPS system (dual sonar sensors).

That's incorrect. The barometric sensor doesn't care whether it is above or below takeoff height - it simply converts pressure to relative altitude assuming a standard atmosphere. The sources of errors in either direction are the same.

It will NEVER be perfect unless you rely on VPS and LOS visual checks. Or a RTK unit.

You probably should define "perfect".

 

[db4476]

It depends on what you mean by "correct". Deviations of a few meters are perfectly normal for fused barometric/inertial solutions.

Maybe - if the FC is using RTK for altitude correction.

That's incorrect. The barometric sensor doesn't care whether it is above or below takeoff height - it simply converts pressure to relative altitude assuming a standard atmosphere. The sources of errors in either direction are the same.

You probably should define "perfect".

A modern digital Barometer has a resolution of +/- 1 meter and position above or below a point makes no difference to the Barometer's accuracy. What is not known is how DJI incorporates the data into the IMU such as complementing GNSS altitude or vice versa.

Cheers!

 

[sar104]

A modern digital Barometer has a resolution of +/- 1 meter and position above or below a point makes no difference to the Barometer's accuracy. What is not known is how DJI incorporates the data into the IMU such as complementing GNSS altitude or vice versa.

Cheers!

The weighting isn't known, but the basic scheme is well established. Displacement is tracked primarily by the inertial system - accelerometers and rate gyros. The barometer data are incorporated at lower gain via a Kalman filter or similar. GNSS altitude has not been used, at least for past versions.

 

[Sky Angels]

It depends on what you mean by "correct". Deviations of a few meters are perfectly normal for fused barometric/inertial solutions.

Maybe - if the FC is using RTK for altitude correction.

That's incorrect. The barometric sensor doesn't care whether it is above or below takeoff height - it simply converts pressure to relative altitude assuming a standard atmosphere. The sources of errors in either direction are the same.

You probably should define "perfect".

Well then perfectly normal is not good at all!

Very accurate. Maybe not quite perfect.

Yes the altimeter system is flawed it will read wrongly every time you fly below take off altitude. Way off! The VPS can be used to judge very accurately.

 

[Meta4]

Well then perfectly normal is not good at all!

Very accurate. Maybe not quite perfect.

Yes the altimeter system is flawed it will read wrongly every time you fly below take off altitude. Way off! The VPS can be used to judge very accurately.

What you've described is not what almost every DJI drone owner experiences.
The indicated height in the app is reasonably accurate.
During the time of a flight you can see variations of +/- 10 feet or so due to air pressure changes, but that has no practical effect on drone users.
And whether the height has a + or - sign in front makes no difference at all.

 

[sar104]

Yes the altimeter system is flawed it will read wrongly every time you fly below take off altitude. Way off! The VPS can be used to judge very accurately.

No - that is incorrect unless you have a faulty unit. All the (many) DJI and other aircraft that I've used handle altitude just fine below the takeoff altitude.

 

[db4476]

The weighting isn't known, but the basic scheme is well established. Displacement is tracked primarily by the inertial system - accelerometers and rate gyros. The barometer data are incorporated at lower gain via a Kalman filter or similar. GNSS altitude has not been used, at least for past versions.

No - that is incorrect unless you have a faulty unit. All the (many) DJI and other aircraft that I've used handle altitude just fine below the takeoff altitude.

Well then perfectly normal is not good at all!

Very accurate. Maybe not quite perfect.

Yes the altimeter system is flawed it will read wrongly every time you fly below take off altitude. Way off! The VPS can be used to judge very accurately.

Flying below your takeoff point does not change accuracy at all. Barometric sensors are more than accurate enough for the purpose of flying a drone. If you need cm level of accuracy such as for mapping/surveying then an RTK equipped system would be the way to go.

Cheers!

 

[Mantrain]

How would you ever hit the bridge or anything else assuming you are flying in N or C mode?

I understand it could be frustrating not to have the accurate information but I do not understand how it would be catastrophic. anyways, you should be entitled to an exchange model. Have you requested an exchange?

 

[dmcgrew]

You misunderstood what he was saying.

Not sure how to interpret that sentence any other way.

 

[Meta4]

Not sure how to interpret that sentence any other way.

As you haven't explained what sentence, I'd have to go back and work it out.
BK originally said
>>Today I was heading for a bridge in a section I know to be 150 feet. So I usually drop to 70-80 feet to go under it. Well, as I approached I could clearly see I was dead on to hit it.

To that you replied:
>>Why would you fly into a bridge if the height is incorrect? Do you not look at where you are going?

BK didn't fly into a bridge, he flew towards the bridge.
And because he was looking where he was going (why would you ask?), he was able to stop and hover in front of the bridge to record those screenshots to illustrate the problem he's having with his indicated height in the app.

 

[bkushner]

The debate is getting deep. Obviously I didn't hit bridge and saw I was too high. To sum it up, I'd like to have accurate altitude reading and not be off -170 feet and there are many reasons for that.

 

[Sky Angels]

Flying below your takeoff point does not change accuracy at all. Barometric sensors are more than accurate enough for the purpose of flying a drone. If you need cm level of accuracy such as for mapping/surveying then an RTK equipped system would be the way to go.

Cheers!

Please man,
I’m saying the Mavic cannot read correct altitude ever once you fly below take off point!
Period. Do you understand what I am saying?
I don’t care how it gets to the OSD. It’s WAY OFF AND TOTALLY UNRELIABLE!

“None”of the DJI Drones below Enterprise Level have ACCURATE ALTITUDE READINGS IF YOU FLY BELOW THEN BACK ABOVE TAKE OFF POINT!

101 DJI drones owned.

None of them ever read correctly