DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Boggling fly away - Mavic Pro

Mick M2P

Member
Joined
Mar 19, 2020
Messages
12
Reactions
11
Age
55
Location
Sweden
Hi fellow pilots.

I know there is mindblowing knowledge among you guys so I hope for some expert help :)

A friend just bought his first UAS. A used Mavic Pro in mint conditioning. All apps on his iPhone, UAV and handcontroller has the latest updates. Starting point was his house. The only place he´d been flying from, hence no GPS calibration done prior to this flight.
He flew up to 30 meters (approc. 100 feet). Ha wanted to take a photo of his daughter when suddenly the drone started drifting. No warnings or indication on the controller that something was wrong. The UAV drifted 100 meters (320 feet) before ending up in a tree.
The battery was stille in its place so it did not come off. Battery level was 4% lower compared to take off.

Does anyone have a decent explanation looking at the log below?

Log line 1: It is 20 kilometer (12 miles) from his house to the nearest airport.
Log line 3: Handcontroller did not indicate compass error (what is "Exit P-GPS Mode")
Log line 4 & 5: There is no magnetic interference around his house
Last 4 log lines is self explanatory as the UAV struck a tree

This is the log from the DAT file:
[["2021-02-11 15:45:10","","Your aircraft has entered a Warning Zone (Class C Airspace). Please fly with caution."]]
[["2021-02-11 15:45:18","","Home Point Recorded. RTH Altitude: 30m."]]
[["2021-02-11 15:46:09","","Compass Error. Exit P-GPS Mode"]]
[["2021-02-11 15:46:10","","IMU heading error. Please restart the aircraft."]]
[["2021-02-11 15:46:15","","Weak GPS signal. Positional accuracy may be compromised. Please fly with caution."]]
[["2021-02-11 15:46:19","","Motor Obstructed"]]
[["2021-02-11 15:46:23","","Motor Idling (Propeller Loose or Missing)"]]
[["2021-02-11 15:46:26","","Weak signal. Adjust antenna and avoid signal block."]]
[["2021-02-11 15:46:27","","Motor Idling (Propeller Loose or Missing)"]]
 
  • Like
Reactions: Droniac

Sometimes what happens is you take off from the same spot you have taken off from before where there is low GPS signal and because you dont gain altitude fast enough the Drone can loose the GPS go into Atti mode and drift.

There are quite a few posts like this one when flying from in-between houses / buildings

There was a really good example last week that someone posted taking off in the middle of the street surrounded by Buildings and I wish I could find it .

where the drone starts to drift lust slightly and its clear as day as to what is happening and it crashes into the building .
The Pilot had plenty of time to recognize and land but missed the drift .

Weak GPS signal can happen to anyone, the key is to look for that drift immediately or Gain Alt quickly to reestablish a signal but I am betting he was going very slowly and carefully the weak signal got him to drift as he the signal got weaker and the drone enters Atti mode.

Magnetic Interference Is also possible : Many people like to take off from there concrete driveway because it seems like the thing to do but it can result in hammering the Compass .

Phantomrain.org
Gear to fly your Mavic in the Rain / Snow and Land on Water.
 
Last edited:
Hi there fellow country man ...

In order to diagnose out from facts instead of speculating it's better that you share the flight log here. The log first needed is the mobile device .TXT log ... it's stored in the mobile screen device he flew with. Read up here (in section 3.) on how to retrieve it --> Mavic Flight Log Retrieval and Analysis Guide

*EDIT*
If you already know for certain which mobile device .DAT log that belongs to this flight ... it's actually even better if you share that ... just attach it in a new post here.

Well ... that's not quite true. Everything made of iron is magnetic & will interfere with the aircraft's compass if close enough. If the compass is near for instance a nail in a wooden deck or just above/near some hidden steel bar reinforcement that will throw the compass off. If this happens when the aircraft is powered on means that the IMU will be initializing the aircraft's heading direction into the wrong direction.

Once airborne, away from the magnetic interference the compass will correct it self ... but not the IMU. The aircraft will now think it is pointing in another direction than it does in reality. Up higher affected by winds & trying to hold position the aircraft will use different motors to fight against the wind ... but does it then with the wrong motors due to not knowing the correct heading direction. This usually creates a straight flyaway if the compass was off around 180 degrees ... a toilet bowled like flight path if it was off around 90 degrees.

Both of these messages indicate that this was the culprit ...

[["2021-02-11 15:46:09","","Compass Error. Exit P-GPS Mode"]]
[["2021-02-11 15:46:10","","IMU heading error. Please restart the aircraft."]]

So ... yeah, he most probably had magnetic interference around his house.
 
Last edited:
Sometimes what happens is you take off from the same spot you have taken off from before where there is low GPS signal and because you dont gain altitude fast enough the Drone can loose the GPS go into Atti mode and drift.

There are quite a few posts like this one when flying from in-between houses / buildings

There was a really good example last week that someone posted taking off in the middle of the street surrounded by Buildings and I wish I could find it .

where the drone starts to drift lust slightly and its clear as day as to what is happening and it crashes into the building .
The Pilot had plenty of time to recognize and land but missed the drift .

Weak GPS signal can happen to anyone, the key is to look for that drift immediately or Gain Alt quickly to reestablish a signal but I am betting he was going very slowly and carefully the weak signal got him to drift as he the signal got weaker and the drone enters Atti mode.

Magnetic Interference Is also possible : Many people like to take off from there concrete driveway because it seems like the thing to do but it can result in hammering the Compass .

Phantomrain.org
Gear to fly your Mavic in the Rain / Snow and Land on Water.
Thank you for responding with possible causes. Appreciate it :)
 
Hi there fellow country man ...

In order to diagnose out from facts instead of speculating it's better that you share the flight log here. The log first needed is the mobile device .TXT log ... it's stored in the mobile screen device he flew with. Read up here (in section 3.) on how to retrieve it --> Mavic Flight Log Retrieval and Analysis Guide

*EDIT*
If you already know for certain which mobile device .DAT log that belongs to this flight ... it's actually even better if you share that ... just attach it in a new post here.


Well ... that's not quite true. Everything made of iron is magnetic & will interfere with the aircraft's compass if close enough. If the compass is near for instance a nail in a wooden deck or just above/near some hidden steel bar reinforcement that will throw the compass off. If this happens when the aircraft is powered on means that the IMU will be initializing the aircraft's heading direction into the wrong direction.

Once airborne, away from the magnetic interference the compass will correct it self ... but not the IMU. The aircraft will now think it is pointing in another direction than it does in reality. Up higher affected by winds & trying to hold position the aircraft will use different motors to fight against the wind ... but does it then with the wrong motors due to not knowing the correct heading direction. This usually creates a straight flyaway if the compass was off around 180 degrees ... a toilet bowled like flight path if it was off around 90 degrees.

Both of these messages indicate that this was the culprit ...

[["2021-02-11 15:46:09","","Compass Error. Exit P-GPS Mode"]]
[["2021-02-11 15:46:10","","IMU heading error. Please restart the aircraft."]]

So ... yeah, he most probably had magnetic interference around his house.
Tack ;)
Will keep the thread in english so others can follow. Appreciate your explanation :)
Will ask him to extract the TXT log and post it here.
 
If you already have got hold of the correct DAT log, share that together with the .TXT ... out from the DAT we can see the raw sensor data, much easier to diagnose then.
 
The DAT file will have a similar date and time stamp but one may be in the 24hr clock and the other in a 12hr clock, which can make thing's confusing.
 
Sounds very much like the typical result of IMU malfunction or magnetic interference. The log file will tell the actual reason.
 
If it did go into ATTI mode: I am not sure if you can force the MP into ATTI mode, but (in wide open field), it is good to get experience flying in ATTI. And it's fun.
 
Hi there fellow country man ...

In order to diagnose out from facts instead of speculating it's better that you share the flight log here. The log first needed is the mobile device .TXT log ... it's stored in the mobile screen device he flew with. Read up here (in section 3.) on how to retrieve it --> Mavic Flight Log Retrieval and Analysis Guide

*EDIT*
If you already know for certain which mobile device .DAT log that belongs to this flight ... it's actually even better if you share that ... just attach it in a new post here.


Well ... that's not quite true. Everything made of iron is magnetic & will interfere with the aircraft's compass if close enough. If the compass is near for instance a nail in a wooden deck or just above/near some hidden steel bar reinforcement that will throw the compass off. If this happens when the aircraft is powered on means that the IMU will be initializing the aircraft's heading direction into the wrong direction.

Once airborne, away from the magnetic interference the compass will correct it self ... but not the IMU. The aircraft will now think it is pointing in another direction than it does in reality. Up higher affected by winds & trying to hold position the aircraft will use different motors to fight against the wind ... but does it then with the wrong motors due to not knowing the correct heading direction. This usually creates a straight flyaway if the compass was off around 180 degrees ... a toilet bowled like flight path if it was off around 90 degrees.

Both of these messages indicate that this was the culprit ...

[["2021-02-11 15:46:09","","Compass Error. Exit P-GPS Mode"]]
[["2021-02-11 15:46:10","","IMU heading error. Please restart the aircraft."]]

So ... yeah, he most probably had magnetic interference around his house.
Hey friend. Got hold of the TXT file and attaches it in this post/thread. Is it rocket science to extract the data and interpret it? I have tried open it with several different program that was recomenden on the web. No success though.
 

Attachments

  • DJIFlightRecord_2021-02-11_[15-45-10].txt
    105.1 KB · Views: 11
Thank's for the .TXT log ... in order to definitely tell the reason the mobile device .DAT log is needed so we can access the raw sensor data. The correct .DAT log ends with FLY037.DAT ... try to retrieve it & attach it in a new post.

Out from the .TXT we can see that the start was in a bit of a hurry, your friend didn't wait for a HP to be recorded before he commanded take-off through the app. The HP was then recorded when the navhealth became 4 (of 5) & that happened 7 sec into the flight.

He then flew out & everything appeared OK until 54sec into the flight ... there he releases the Elevator (right stick forward) & a bit earlier the Aileron stick (right stick left) ... but the AC continued it's heading speed up there on 11m.

Have placed the markers in the chart where the incident starts ... The sat. pic shows where the AC is by then & the yellow bar the IMU indicated heading direction.

The meaning of the graphs can be read out under the chart ... click on it to make it larger. (Take the black graph=heading speed times 10 for the correct m/s speed)

1613327779175.png
1613327794939.png

From 54sec the AC starts to fly away in a pretty straight path ... but at 59sec the flight controller drops the navhealth to zero (...even though it have 18 satellites locked), it have totally lost the position confidence ... just after the AC goes into ATTI mode ... meaning only height hold but no horizontal hold & the AC starts to drift with the wind which comes from north.

During the drift towards the crash site several stick inputs is made that are completely the opposite relative the indicated heading direction & what would have been the logical to stop the flyaway.

So far it's clear that the flight controller lost the position confidence & just before the FC gives up & go for ATTI the AC performs uncommanded movements. Furthermore stick commands given during the flyaway points to a IMU heading difference towards reality of approx 180 degrees.

All this points to 2 reasons ... either a magnetic disturbed compass during power on ... or a IMU failure.

This is also confirmed when looking at the velocity differences between the GPS & the IMU (the differences should be nearly 0 usually)

Already at 41 sec the velocities start to deviate ...

1613328945148.png

This can also be shown by the logged flight paths ... the red the GPS & the green the IMU, they don't agree.

1613329068192.png
 
  • Like
Reactions: SpitFire
Thank's for the .TXT log ... in order to definitely tell the reason the mobile device .DAT log is needed so we can access the raw sensor data. The correct .DAT log ends with FLY037.DAT ... try to retrieve it & attach it in a new post.

Out from the .TXT we can see that the start was in a bit of a hurry, your friend didn't wait for a HP to be recorded before he commanded take-off through the app. The HP was then recorded when the navhealth became 4 (of 5) & that happened 7 sec into the flight.

He then flew out & everything appeared OK until 54sec into the flight ... there he releases the Elevator (right stick forward) & a bit earlier the Aileron stick (right stick left) ... but the AC continued it's heading speed up there on 11m.

Have placed the markers in the chart where the incident starts ... The sat. pic shows where the AC is by then & the yellow bar the IMU indicated heading direction.

The meaning of the graphs can be read out under the chart ... click on it to make it larger. (Take the black graph=heading speed times 10 for the correct m/s speed)

View attachment 123846
View attachment 123847

From 54sec the AC starts to fly away in a pretty straight path ... but at 59sec the flight controller drops the navhealth to zero (...even though it have 18 satellites locked), it have totally lost the position confidence ... just after the AC goes into ATTI mode ... meaning only height hold but no horizontal hold & the AC starts to drift with the wind which comes from north.

During the drift towards the crash site several stick inputs is made that are completely the opposite relative the indicated heading direction & what would have been the logical to stop the flyaway.

So far it's clear that the flight controller lost the position confidence & just before the FC gives up & go for ATTI the AC performs uncommanded movements. Furthermore stick commands given during the flyaway points to a IMU heading difference towards reality of approx 180 degrees.

All this points to 2 reasons ... either a magnetic disturbed compass during power on ... or a IMU failure.

This is also confirmed when looking at the velocity differences between the GPS & the IMU (the differences should be nearly 0 usually)

Already at 41 sec the velocities start to deviate ...

View attachment 123850

This can also be shown by the logged flight paths ... the red the GPS & the green the IMU, they don't agree.

View attachment 123851
Thanx a million buddy :)
 
For nearly a minute of the flight, the GPS and IMU data were in close agreement. Then suddenly, they were not. That seems like the fingerprint of an IMU failure.

It's very possible that the "Compass Error. Exiting P-GPS" message you asked about meant that flight control was switching to ATTI mode on account of inconsistent data from the GPS and IMU. If so, the aircraft would tend to drift unbraked with whatever breeze was present.

The OP says this was the pilot's first UAS, and they'd only flown a few flights at home. Well, to an unschooled and inexperienced pilot, a sudden switch into ATTI mode can feel like you have lost the ability to control the aircraft. In an important sense that's true.

Probably unaware they were in ATTI mode--not to mention what that even means--the pilot steered the craft somewhat wildly, as @slup notes. In that situation, most people would be naturally and understandably a bit panicked and disoriented. While the pilot technically had control of the bird, that had no practical significance during the few remaining moments left of the flight.

That said, an IMU failure isn't pilot error, and a beginner can be forgiven if they don't find ATTI flight exactly intuitive.

Every flight is a bit like playing a slot machine. Yes, the slot machine might pay a king's ransom on this pull, but it probably won't. Likewise, there might be a severe mechanical failure on this flight, but there probably won't.

But both gambles pay off sooner or later--"sooner", I think, in the unfortunate case of the OP's friend.
 
For nearly a minute of the flight, the GPS and IMU data were in close agreement. Then suddenly, they were not. That seems like the fingerprint of an IMU failure.

It's very possible that the "Compass Error. Exiting P-GPS" message you asked about meant that flight control was switching to ATTI mode on account of inconsistent data from the GPS and IMU. If so, the aircraft would tend to drift unbraked with whatever breeze was present.

The OP says this was the pilot's first UAS, and they'd only flown a few flights at home. Well, to an unschooled and inexperienced pilot, a sudden switch into ATTI mode can feel like you have lost the ability to control the aircraft. In an important sense that's true.

Probably unaware they were in ATTI mode--not to mention what that even means--the pilot steered the craft somewhat wildly, as @slup notes. In that situation, most people would be naturally and understandably a bit panicked and disoriented. While the pilot technically had control of the bird, that had no practical significance during the few remaining moments left of the flight.

That said, an IMU failure isn't pilot error, and a beginner can be forgiven if they don't find ATTI flight exactly intuitive.

Every flight is a bit like playing a slot machine. Yes, the slot machine might pay a king's ransom on this pull, but it probably won't. Likewise, there might be a severe mechanical failure on this flight, but there probably won't.

But both gambles pay off sooner or later--"sooner", I think, in the unfortunate case of the OP's friend.
Thanks for your valued input :)
 
For nearly a minute of the flight, the GPS and IMU data were in close agreement. Then suddenly, they were not. That seems like the fingerprint of an IMU failure...
Not necessarily ... the velocity differences usually increase when the uncontrolled flight starts, if no major need of fighting something to be able to hold position, the velocity differences are usually rather low.

It was somewhat difficult to find a DAT from an older case where it was confirmed that it was a a wrongly initialized IMU that was the cause. Could find below though ... even though the Yaw/magYaw difference (the red) isn't constant (the compass wasn't healthy) it was confirmed that the IMUYaw was off by 127 degrees.

As seen here below the velocity differences starts when the uncontrolled flight kicks in & the heading speed increases ...

1613464780236.png
 
Not necessarily ... the velocity differences usually increase when the uncontrolled flight starts, if no major need of fighting something to be able to hold position, the velocity differences are usually rather low.

It was somewhat difficult to find a DAT from an older case where it was confirmed that it was a a wrongly initialized IMU that was the cause. Could find below though ... even though the Yaw/magYaw difference (the red) isn't constant (the compass wasn't healthy) it was confirmed that the IMUYaw was off by 127 degrees.

As seen here below the velocity differences starts when the uncontrolled flight kicks in & the heading speed increases ...

View attachment 123913
Wow, IMU Yaw off by 127 degrees is huge and makes a heck of a difference.
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,583
Messages
1,554,081
Members
159,586
Latest member
DoubleBarS