Sudden Altitude Drops

[bkushner]

Anyone else experiencing sudden altitude drops? I've had it happen a good amount of the time. Here is an example.

In 10 seconds the drone drops from 103 feet to 51 feet without touching the stick.
5:23 to 5:33.

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

 

[mavic3usa]

Yes I get altitude drops but not that much and not that sudden. When flying high and fast, I gotta be careful to keep checking the altitude. When I'm at 150 knowing that everything around me is 100 or lower, no longer means I'm safe to fly however I like. Gotta remember to give it some lift every now and again; not sure why.

 

[Phantomrain.org]

You were caught in a Low Altitude zone 200 ft as well as flying closely to the Geo Zone and that alone can cause the drone to react in this manor. I made a video of how my drone reacted flying thru a Geo Zone as it started slowing and landing on its own.

Phantomrain.org
Gear to fly in the Rain. Land on the Water

 

[GadgetGuy]

Anyone else experiencing sudden altitude drops? I've had it happen a good amount of the time. Here is an example.

In 10 seconds the drone drops from 103 feet to 51 feet without touching the stick.
5:23 to 5:33.

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

This guy is reporting similar behavior during map-programmed Waypoint Missions. Programmed change in elevation is from 150 to 100 feet at the waypoint, but it descends to 50 feet instead, and clips a tree or crashes.

Waypoint mission problem

If so why not raise the programmed height accordingly? I do after the fact obviously. But it shouldn't do that in the first place. I would say the reason that the other people have not noticed this, is because they're not flying Waypoint Mission very much. Since I have had two separate mavic...

mavicpilots.com

 

[Yorkshire_Pud]

@bkushner With regards to the descent, I am wondering is it possible that the inbound flight brought the drone into conflict with an air corridor and that the drone descended to pass underneath the corridor?
I am not that well informed concerning such things as I don't fly near any known air corridors but I recollect @Suren has documented that behaviour with his Mavic 3.

 

[Yorkshire_Pud]

but it descends to 50 feet instead

Not to be too fussy but we have yet to establish whether or not shb had taken changes in ground elevation into account in their waypoint planning. Last I noticed shb seems reticent to disclose details of the flight heights, "That's classified." post #16.

 

[GadgetGuy]

With regards to the descent, I am wondering is it possible that the inbound flight brought the drone into conflict with an air corridor and that the drone descended to pass underneath the corridor?
I am not that well informed concerning such things as I don't fly near any known air corridors but I recollect @Suren has documented that behaviour with his Mavic 3.

That is certainly another possibility that can only be confirmed by knowing the exact location, and such GEO behavior and would be supported by the flight logs on his phone.

 

[GadgetGuy]

Not to be too fussy but we have yet to establish whether or not shb had taken changes in ground elevation into account in their waypoint planning. Last I noticed shb seems reticent to disclose details of the flight heights, "That's classified." post #16.

Yes, that snarky attitude is certainly not helping. I'm convinced he does understand the programmed waypoint elevations are relative to the launch point, and not AGL. However, that does not explain the 50‘ extra descent he has been experiencing, no matter what values he programmed.

 

[GadgetGuy]

Yes I get altitude drops but not that much and not that sudden. When flying high and fast, I gotta be careful to keep checking the altitude. When I'm at 150 knowing that everything around me is 100 or lower, no longer means I'm safe to fly however I like. Gotta remember to give it some lift every now and again; not sure why.

Keep in mind that elevation changes between waypoints are treated as glidepaths, so if he is flying fast, the programmed descent is happening and then exceeded, by the time the next waypoint is reached. At least that is my interpretation of everything.

 

[Yorkshire_Pud]

Yes, that snarky attitude is certainly not helping. I'm convinced he does understand the programmed waypoint elevations are relative to the launch point, and not AGL. However, that does not explain the 50‘ extra descent he has been experiencing, no matter what values he programmed.

What I am getting at is, shb may have programmed in a low point height of x ft but if the ground elevation at the low point is y ft higher than the elevation at the take off point then the AGL at the low point will be (x -y) ft.
I.e. I am not convinced that shb's drone did descend below the set height, which is why I suggested shb fly a test flight and post the flight log and, if possible, the waypoint flight plan for that test flight.
shb did say that the ground rose on the way to the drone's destination.
But really we should cease discussion of that-thread/shb in this one.

 

[BudWalker]

Anyone else experiencing sudden altitude drops? I've had it happen a good amount of the time. Here is an example.

In 10 seconds the drone drops from 103 feet to 51 feet without touching the stick.
5:23 to 5:33.

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

Kinda interesting. It would seem the FC was thinking that the M3P's downward velocity (OSD:zSpeed) was negative. Or, the upward velocity was positive; i.e. the M3P was ascending. To compensate the FC caused a descent.

The green plot shows the negative downward velocity. The red is the FC's altitude data.

From the pilot's observation it would seem that OSD:zSpeed is incorrect. This value is computed by the FC and is dependent on gyro, accelerometer, GPS and magnetometer data. Unfortunately, this data is only available in the .DAT log which is encrypted on the Mavic 3.

 

[GadgetGuy]

What I am getting at is, shb may have programmed in a low point height of x ft but if the ground elevation at the low point is y ft higher than the elevation at the take off point then the AGL at the low point will be (x -y) ft.
I.e. I am not convinced that shb's drone did descend below the set height, which is why I suggested shb fly a test flight and post the flight log and, if possible, the waypoint flight plan for that test flight.
shb did say that the ground rose on the way to the drone's destination.
But really we should cease discussion of that-thread/shb in this one.

You are overlooking that shb's drones are, in fact, descending 25-50' below the waypoint set value, similar to the OP's 50 foot drop here with no downward stick input. shb inputs a waypoint value of 150' and the next waypoint at 100' on the map, both relative to his launch point. Instead of arriving at the second waypoint at 100’ above the launch point, it's arriving 25-50' below the input value, or 50-75 feet above his launch point. The terrain below is only relevant when the excessive descent causes the drone to hit a tree or run into the ground. It is still an unprogrammed descent of up to 50' during forward flight, which should not be happening, just like here.

 

[bkushner]

You were caught in a Low Altitude zone 200 ft as well as flying closely to the Geo Zone and that alone can cause the drone to react in this manor. I made a video of how my drone reacted flying thru a Geo Zone as it started slowing and landing on its own.

Phantomrain.org
Gear to fly in the Rain. Land on the Water

I fly this area 7 days a week for the past 8 years. I know the zones like the back of my hand. The lowest area is 194 feet. Drone was under a 100 feet when so that wouldn't have an effect. Also I've owned every drone made since the Phantom 2 and never had this issue. I currently fly an Air2S and Mavic mini 3 Pro no issues..

 

[bkushner]

@bkushner With regards to the descent, I am wondering is it possible that the inbound flight brought the drone into conflict with an air corridor and that the drone descended to pass underneath the corridor?
I am not that well informed concerning such things as I don't fly near any known air corridors but I recollect @Suren has documented that behaviour with his Mavic 3.

Not the issue.

 

[GadgetGuy]

I fly this area 7 days a week for the past 8 years. I know the zones like the back of my hand. The lowest area is 194 feet. Drone was under a 100 feet when so that wouldn't have an effect. Also I've owned every drone made since the Phantom 2 and never had this issue. I currently fly an Air2S and Mavic mini 3 Pro no issues..

I think you are onto something. There is something fishy in this autonomous change in elevation, which is "dependent on gyro, accelerometer, GPS and magnetometer data." It is very similar to what the other guy shb is also reporting. The drone is descending on its own, beyond stick inputs, and beyond preset waypoint elevation values.

 

[BudWalker]

I think you are onto something. There is something fishy in this autonomous change in elevation, which is "dependent on gyro, accelerometer, GPS and magnetometer data." It is very similar to what the other guy shb is also reporting. The drone is descending on its own, beyond stick inputs, and beyond preset waypoint elevation values.

There is also this incident.
Mavic mini 1 stuck at full throttle
Although it's different in that 1) it was a Mini 3 and 2) it was an uncommanded ascent. Since the .DAT was available for that incident it could be determined the cause was a faulty Z axis accelerometer. My question is; do the Mini 3 and Mavic 3 have the same accelerometer HW?

 

[BudWalker]

Anyone else experiencing sudden altitude drops? I've had it happen a good amount of the time. Here is an example.

In 10 seconds the drone drops from 103 feet to 51 feet without touching the stick.
5:23 to 5:33.

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

If you are up for some experimentation could you try this? Just using the throttle go full positive and wait until max ascent rate is achieved. Then go full negative. Then try the opposite - full negative until max descent rate and then full positive.

You might want to start with plenty of altitude

The purpose of the experiment is to see if max Z axis acceleration can cause the error.

In the flight you submitted the incident occurred after full elevator was applied. That causes some negative Z axis acceleration since the M3 will be tilted in the direction of travel.

 

[Wreckhunter]

Not the issue.

Well your doing something wrong