Lost my Mavic Air while using Map Pilot

[larry.flemming]

WARNING - Started using Maps Made Easy (Map Pilot) with my Mavic Air yesterday to create some enhanced maps. Ran a few successful missions earlier in the day and then tried a linear mapping mission of a cross country ski trail in the evening. Waypoints were being followed perfectly (at a planned altitude of 46 m - well above any potential trees) until about 90% through the mission when things literally went south. Drone started deviating from planned flight path on its own. Flying conditions were very good, there was no indication of poor signal. RTH did nothing. What might have caused this to happen? Flight data is available at https://app.airdata.com/share/KLRLqj
Had a look in the dense brush around the last known GPS coordinate, but no dice. TIA for any insight you may have.

 

[BadWolf1]

Could you upload the orginal data the end of the file from airdata repeats the same line with different time stamps.

 

[larry.flemming]

This should be everything. Really would like to know what caused the flight plan deviation (off course and rapid acceleration) and lack of RTH functionality. Have flown this trail multiple times before with Litchi with no issue.

 

[BadWolf1]

Only the first 5340 rows of the CSV file are real data, after that it continues to log the same data.

Something failed on the drone at 17:45:29.

Open the file and in Excel or similar (freeze the top row if poss) and scroll down to row 5485, then look at time and altitude. column D and C

The drone is humming along at ~52M.
Then it drops 5M in 1 second at 17:45:30,
7M in the next second 17:45:31 ;
8.5 in the next second.17:45:32
9.6 in the next. 17:45:33
And then the data updates stop. 34m in 4 seconds

My method here is to look for an impact. We'll see a big change in speed. So lets look up and down column F which is speed. It's all 15m/sec +/- 1 so it didn't hit anything.
What else do we see if something goes wrong. Big changes in attitude, so let's look down columns I & J (Roll and pitch) going at 14 and 4 degrees respectively then suddenly, pitch goes to 44 roll goes to 30.

OK so a we dropped very rapidly at a unflyable angle.

One last thing, which I spotted column Q, battery current. We only get a new value every two seconds, and the next sample after the big lurch the current doubles. Either a motor jammed and drew a lot of current, or a prop came off and the flight controller responded by spinning the others as fast as they'd go.

Maybe one of the forum experts will shed more light.

 

[larry.flemming]

Only the first 5340 rows of the CSV file are real data, after that it continues to log the same data.

Something failed on the drone at 17:45:29.

Open the file and in Excel or similar (freeze the top row if poss) and scroll down to row 5485, then look at time and altitude. column D and C

The drone is humming along at ~52M.
Then it drops 5M in 1 second at 17:45:30,
7M in the next second 17:45:31 ;
8.5 in the next second.17:45:32
9.6 in the next. 17:45:33
And then the data updates stop. 34m in 4 seconds

My method here is to look for an impact. We'll see a big change in speed. So lets look up and down column F which is speed. It's all 15m/sec +/- 1 so it didn't hit anything.
What else do we see if something goes wrong. Big changes in attitude, so let's look down columns I & J (Roll and pitch) going at 14 and 4 degrees respectively then suddenly, pitch goes to 44 roll goes to 30.

OK so a we dropped very rapidly at a unflyable angle.

One last thing, which I spotted column Q, battery current. We only get a new value every two seconds, and the next sample after the big lurch the current doubles. Either a motor jammed and drew a lot of current, or a prop came off and the flight controller responded by spinning the others as fast as they'd go.

Maybe one of the forum experts will shed more light.

Thanks BadWolf1. Great insight into this and good observation about the high amperage draw. I had changed props a few days back but didn't think to check them again (assumed the push/twist lock was infallible...) - may well have been a thrown prop. Thanks for the prompt reply.

 

[PistonPete]

Are you able to use the find my drone feature? Don't know if that works if you were using Litchi or something else.

 

[larry.flemming]

Are you able to use the find my drone feature? Don't know if that works if you were using Litchi or something else.

The feature isn't baked into Maps Made Easy (Map Pilot), but had I had my wits about be at the time the drone went down, I should have been able to switch over to DJI app, establish connection and then use the Find My Drone feature (a beep would have made looking in the low-bush so much easier).

 

[larry.flemming]

Only the first 5340 rows of the CSV file are real data, after that it continues to log the same data.

Something failed on the drone at 17:45:29.

Open the file and in Excel or similar (freeze the top row if poss) and scroll down to row 5485, then look at time and altitude. column D and C

The drone is humming along at ~52M.
Then it drops 5M in 1 second at 17:45:30,
7M in the next second 17:45:31 ;
8.5 in the next second.17:45:32
9.6 in the next. 17:45:33
And then the data updates stop. 34m in 4 seconds

My method here is to look for an impact. We'll see a big change in speed. So lets look up and down column F which is speed. It's all 15m/sec +/- 1 so it didn't hit anything.
What else do we see if something goes wrong. Big changes in attitude, so let's look down columns I & J (Roll and pitch) going at 14 and 4 degrees respectively then suddenly, pitch goes to 44 roll goes to 30.

OK so a we dropped very rapidly at a unflyable angle.

One last thing, which I spotted column Q, battery current. We only get a new value every two seconds, and the next sample after the big lurch the current doubles. Either a motor jammed and drew a lot of current, or a prop came off and the flight controller responded by spinning the others as fast as they'd go.

Maybe one of the forum experts will shed more light.

BadWolf1, how much confidence would you have in the final GPS coordinates and altitude? I've scoured the immediate area on the ground, but to no avail. I'm a bit confused by the 17 m altitude. Would that be relative to the take-off location elevation? i.e. 17 m higher than take-off? I'm wondering if the trajectory that the drone was one would have been apt to plunk it a tree or more likely to drop it to the ground.

Below is the last screen shot available from the Maps Made Easy app. Not sure if this is even close to the view from the drones final resting place.


This is a screen shot form a GPX app that shows elevation profile. I'm confused by where it ends off. Highest trees near suspected crash zone would be 15 m.



this is the vegetation I'm combing through... urgh. a black/silver Mavic Air is a proverbial needle.



Climb/descent data elevation difference data from GPX file. Not making sense of it.


It would appear I'm 'on the mark' according to the last known GPS coordinates, but nothing to be found in a 10-15 metre radius (ground or in the tress - that I could see)

 

[BadWolf1]

BadWolf1, how much confidence would you have in the final GPS coordinates and altitude? I've scoured the immediate area on the ground, but to no avail. I'm a bit confused by the 17 m altitude. Would that be relative to the take-off location elevation? i.e. 17 m higher than take-off? I'm wondering if the trajectory that the drone was one would have been apt to plunk it a tree or more likely to drop it to the ground.

Below is the last screen shot available from the Maps Made Easy app. Not sure if this is even close to the view from the drones final resting place.

I would trust those lat/lon / alt numbers - but if you stand on that spot, the drone was some way above ground - travelling towards 110 degrees at 15m/sec, it's falling but it's momentum is keeping it going in a roughly straight line. It could have continued two maybe three seconds. That gives you a strip of 30-40m to search