More on Altitude issues

[bkushner]

So my altitude problems persist still. 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. My altimeter is showing 70 feet when I'm at 150. When the flight ended I was -170 feet. I've owned many drones and never had this issue. Also this is the SECOND M3 to do this.

DJI has addressed this a couple times with me saying it's a change in barometric pressure and it's HARMLESS. 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. Betting this effects your rth setting too. Here are some images and DJI's response.

Here is my log from THIS flight. DJI Flight Log Viewer - PhantomHelp.com

 

[ianwood]

Change in barometric pressure? That would be a significant weather event. With change in barometric pressure comes a decent amount of wind. Especially the amounts you are describing.

Does it only happen there or does it happen on all flights? How about flights over land? During the full daylight?

I am curious if there are gremlins from other sensors at play here.

 

[calmar81]

The only thing I could imagine that height is also measured with the sensors on the bottom of the drone and they cannot sense height correctly because you are over water / reflective surface.

But you will have tested it out on multiple locations and solid ground, too. So false bottom sensor readings won’t be the cause.

Can height not also be measured by GPS / GNSS?

I also have my replacement M3, acting the same as the drone before. Also being in contact with DJI.

So I hope we will all get more satisfactory answers and maybe firmware updates will fix our issues if that erratic behavior (we don‘t know from previous drones) is considered „normal“.

Good luck, buddy!

 

[bkushner]

Change in barometric pressure? That would be a significant weather event. With change in barometric pressure comes a decent amount of wind. Especially the amounts you are describing.

Does it only happen there or does it happen on all flights? How about flights over land? During the full daylight?

I am curious if there are gremlins from other sensors at play here.

I pretty much fly same location a lot. It doesn't happen all flights. Seems to be worse at dusk when lighting isn't good. I get tons of starts and stops. For instance I'm flying at 33 and go to 0 then it will only go 10 mph or so then back to 30.. I believe the sensors think they are seeing something and trying to correct. Sometimes I get the tone and the red bar on the screen from the radar when I'm high and not near anything.

 

[Meta4]

The only thing I could imagine that height is also measured with the sensors on the bottom of the drone and they cannot sense height correctly because you are over water / reflective surface.

The primary source of height data is the barometric sensor and that is the height that shows in the Fly app.
The downward sensors have a maximum range of 18 metres, but having big differences between that and the primary sensor is not a problem for the drone.

Can height not also measured by GPS / GNSS?

GPS is not used for the height as shown on the screen.

 

[Meta4]

Change in barometric pressure? That would be a significant weather event

It's not a one-off event.
Brian's getting this sort of thing on a regular basis.
His altitude sensors are just not working as they should.

 

[bkushner]

The primary source of height data is the barometric sensor and that is the height that shows in the Fly app.
The downward sensors have a maximum range of 18 metres, but having big differences between that and the primary sensor is not a problem for the drone.


GPS is not used for the height as shown on the screen.

Check out the difference between IMU and VPS.

 

[Meta4]

Check out the difference between IMU and VPS.

It's hard to tell much from that as the VPS shows height above whatever's below the drone while IMU height (if it's working properly), shows the height relative to the homepoint.

We know that your IMU height often shows very large differences from true height.
Your VPS height is likely to be quite accurate (within the abilities of the sensors)
It's quite normal for the heights to show significant differences, if flying over terrain that's higher or lower than the launch point.
And given that your IMU height is known to be showing false data, there are going to be big differences between the two numbers.

 

[ianwood]

I know that once VPS is out of range or in a confusing environment, it can get mixed up and report weird things. That's nothing new.

What's weird is it (or something else) is influencing the calculated height on your flights. There may be something off in the sensor fusion of the flight controller. The fact that this is happening over water and in low light leads me to believe the vision sensors are contributing erroneous information to the calculated height.

 

[db4476]

The only thing I could imagine that height is also measured with the sensors on the bottom of the drone and they cannot sense height correctly because you are over water / reflective surface.

But you will have tested it out on multiple locations and solid ground, too. So false bottom sensor readings won’t be the cause.

Can height not also be measured by GPS / GNSS?

I also have my replacement M3, acting the same as the drone before. Also being in contact with DJI.

So I hope we will all get more satisfactory answers and maybe firmware updates will fix our issues if that erratic behavior (we don‘t know from previous drones) is considered „normal“.

Good luck, buddy!

Altitude could be calculated with GNSS, it requires 4 satellites. 3 for position + 1 for altitude. I have not seen any evidence that DJI calculates via that method though as Barometric readings are more than sufficient for flying a drone.

Cheers!

 

[ianwood]

Altitude could be calculated with GNSS, it requires 4 satellites. 3 for position + 1 for altitude. I have not seen any evidence that DJI calculates via that method though as Barometric readings are more than sufficient for flying a drone.

Cheers!

Likely because GNSS altitude has a low degree of precision. But unlike the barometer, it does provide an absolute reference which may be helpful in certain applications.

 

[bkushner]

Just heard back from DJI Support ESCALATION and they tell me the same thing. It's all normal and rest assured IF YOU CRASH and it isn't your fault we will replace the drone under warranty. Keep in mind in addition to the barometer issue I reported the stopping and starting, drops in altitude while flying level etc...This has been a month of them telling me it's all normal don't worry. HAHA.

 

[Meta4]

Just heard back from DJI Support ESCALATION and they tell me the same thing. It's all normal and rest assured IF YOU CRASH and it isn't your fault we will replace the drone under warranty. Keep in mind in addition to the barometer issue I reported the stopping and starting, drops in altitude while flying level etc...This has been a month of them telling me it's all normal don't worry. HAHA.

Someone in there has lost the plot.

 

[Coskier]

I find it hard to trust that they will correctly attribute a crash to an error in their system for calculating altitude.

 

[bkushner]

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.

 

[dmcgrew]

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

 

[Desert Drone Works]

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 see the same behavior. Sensors off and it flies like a dream.

 

[Meta4]

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

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.

 

[canadian drone videos]

So my altitude problems persist still. 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. My altimeter is showing 70 feet when I'm at 150. When the flight ended I was -170 feet. I've owned many drones and never had this issue. Also this is the SECOND M3 to do this.

DJI has addressed this a couple times with me saying it's a change in barometric pressure and it's HARMLESS. 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. Betting this effects your rth setting too. Here are some images and DJI's response.

Here is my log from THIS flight. DJI Flight Log Viewer - PhantomHelp.com

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?

 

[canadian drone videos]

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.

The only hiccup I’ve seen on my drone is the error, obstacle sensors dirty, or something like that. Yet they are nice and clean. I too find the drone responds and handles better with the sensors off.