First crash

[davemorts]

Hi,

Thursday 14th December 2017 I took my mavic out for a short flight in an open space with playing fields etc..

After my home point was registered I increased my height to 30 meters and flew to a distance of 80 meters, I never flew any further on this occasion. Even though I had 18 satellites and full bars on my GPS connection the craft seemed to move on its own with no input on the controller. After 90 seconds of flight I lost GPS at a distance of 61 meters and a height of 32 meters and the drone went into ATTI mode. GPS locked on for about 10 seconds but it went back to ATTI. I tried to fly the drone back to me but it seemed unresponsive and crashed into a tree at 26 meters height and 61 meters away after 2 minutes in the air.

The drone has gone into ATTI mode on previous flights but I managed to guide it back home.

I don’t know if the crash was due to a fault in the drone.

The ribbon on the camera is broken and there is no image, the rubber mountings are also damaged.

I put a battery into it, replaced the broken rotors and it seems to fly normally.

Can anyone explain what caused this.

 

[Simmo]

Its gotta be pilot error...

 

[Franzinator]

A similar situation happened to me back in June. Good satellites, good signal strength, then suddenly unresponsive to control inputs and dreaded - “atti-mode”. For some reason people on this forum are hostile towards anyone that questions Atti-mode, as if it’s some great feature. They’ll tell you it’s your fault and to work on your flying skills. I’ve flown plenty of drones before the mavic that we’re fully manual with no gps, and I’m a really good pilot. When the mavic is in Atti-mode, it’s wild- you should never take off until gps lock. I believe both our situations were caused by some magnetic interference in the area, which cause the mavic to somehow lose its gps signal. Once that happens, RTH is impossible, as the aircraft doesn’t know where it is.

 

[davemorts]

Some more info from flight log :
I got speed error at 18.3 MPH.
Then after 42 seconds at 105 ft altitude and 205ft distance I got compass error.
At 1 min 30 secs I got Yaw error and it went straight into ATTI mode.
At 1 min 38 secs Compass error.
At 1 min 40 seconds it went into P-GPS mode for 8 seconds then back into ATTI mode.
At 1 min 49 seconds there was one mention of magnet filed interference but yaw error was continuous.
At 2 mins 4 secsonds it crashed into trees at 25 MPH 86.9ft high and 446 ft from home.

 

[davemorts]

Flight record.
DJI Flight Log Viewer - PhantomHelp.com

 

[BudWalker]

Flight record.

I took a look at the .txt. It seems that at launch there was a Yaw/magYaw separation of around 45 degrees. That's a problem but the flight appeared to proceed more or less normally until about 78 secs where there was some pronounced toilet bowling


Then at about 91 secs gpsHealth was set to 0 probably because the Flight Controller determined it couldn't estimate the GPS coords accurately enough to be used for GPS+ATTI mode.


Since these problems occurred 78 secs after launch and at a height of 32 meters it's probable they are unrelated to the initial Yaw/magYaw separation.

I'm really interested in seeing the .DAT for this incident and it may provide some additional insights as to the cause of the incident. Look here to see how to retrieve the .DAT. It will be large so you'll need to Dropbox or GoogleDrive it and provide a link.

@Franzinator I think you're right. ATTI mode in these types of scenarios is not like the P3 ATTI mode where the pilot is the only one attempting to control the AC. I've not experienced a compass error induced switch to ATTI with my Mavic. But, pilots I know to be competent report that it appears that the FC is still trying to control the AC while in ATTI mode. The pilot just has to be more assertive by providing stronger inputs than is usually required.

 

[Montypaks]

Its gotta be pilot error...

You do know it's not a must to reply right?

 

[davemorts]

Thank you Budwalker for that very comprehensive reply, I can understand a little of it.
Unfortunately I have sent the mavic to DJI for repairs so cant access the .dat file.
In your opinion does it look like a fault in the drone ?
I tried all I could to regain control of the AC while it was in ATTI mode but it didn't seem to respond to my RC inputs.

 

[BudWalker]

Thank you Budwalker for that very comprehensive reply, I can understand a little of it.
Unfortunately I have sent the mavic to DJI for repairs so cant access the .dat file.
In your opinion does it look like a fault in the drone ?
I tried all I could to regain control of the AC while it was in ATTI mode but it didn't seem to respond to my RC inputs.

Really hard to say whose fault it is without the .DAT Even with the .DAT it can often be indeterminate. And, then there are situations where a compass calibration might have prevented the incident but the Go App didn't say so. You could make the argument that the AC was uncontrollable, but then DJI could say you didn't try hard enough. This stuff hurts my head; I just analyze flights.

There is another .DAT on your tablet that we could look at. It has less data than the AC .DAT but there might be something to see. Look here to see how to retrieve it. Since it has the .DAT extension it can't be attached directly to a post. But, it is small enough to be zipped and the .zip can be attached directly

 

[davemorts]

I hope this is the correct file you need. Thanks again for your help.

 

[BudWalker]

I hope this is the correct file you need. Thanks again for your help.

Yikes! That was fast but it's not the same flight as your .txt This one took place at 2017-12-15 12:27:48 GMT where as the .txt took place at 2017-12-14_[15-59-54. You can see the flight times by looking at the file name.

 

[davemorts]

Sorry, my mistake.
Try this one.

 

[BudWalker]

Sorry, my mistake.
Try this one.

That's not the right .DAT either. And, it appears there weren't enough sats to get the correct time.

They are small. Just create a zip with everything after 17-12-15-12-27-49_FLY045.DAT. Should still be small enough to attach directly.

 

[davemorts]

Apologies Budwalker, Hope I got it right this time.

 

[davemorts]

A similar situation happened to me back in June. Good satellites, good signal strength, then suddenly unresponsive to control inputs and dreaded - “atti-mode”. For some reason people on this forum are hostile towards anyone that questions Atti-mode, as if it’s some great feature. They’ll tell you it’s your fault and to work on your flying skills. I’ve flown plenty of drones before the mavic that we’re fully manual with no gps, and I’m a really good pilot. When the mavic is in Atti-mode, it’s wild- you should never take off until gps lock. I believe both our situations were caused by some magnetic interference in the area, which cause the mavic to somehow lose its gps signal. Once that happens, RTH is impossible, as the aircraft doesn’t know where it is.

 

[davemorts]

A similar situation happened to me back in June. Good satellites, good signal strength, then suddenly unresponsive to control inputs and dreaded - “atti-mode”. For some reason people on this forum are hostile towards anyone that questions Atti-mode, as if it’s some great feature. They’ll tell you it’s your fault and to work on your flying skills. I’ve flown plenty of drones before the mavic that we’re fully manual with no gps, and I’m a really good pilot. When the mavic is in Atti-mode, it’s wild- you should never take off until gps lock. I believe both our situations were caused by some magnetic interference in the area, which cause the mavic to somehow lose its gps signal. Once that happens, RTH is impossible, as the aircraft doesn’t know where it is.

Thanks Franzinator for your reply, I don't feel so bad knowing I'm not the only one to suffer this drama.

 

[Franzinator]

Thanks Franzinator for your reply, I don't feel so bad knowing I'm not the only one to suffer this drama.

During my mishap I helplessly watched the video feed as the aircraft raced sideways and a 100’ light pole passed right in front of the camera inches from view

 

[BudWalker]

Apologies Budwalker, Hope I got it right this time.

Yes, that's the right one. It seems the gyroZ and compass (magnetometers) don't agree. In the interval [63 secs - 93 sec] the gyroZ shows the AC rotates 80 degrees CW while the compass shows a 43 degrees CW rotation.


It can't be determined from this data which is correct. But, I suspect the magnetometers are incorrect.

When was the last compass calibration?
Has some type of external equipment been added such as a tracker?

When was this flight? It didn't have any of the same issues.

 

[davemorts]

That flight was the following day ( 15 Dec) I was testing to see if the drone was still operating correctly.
It was at this location ( my front garden) where I calibrated the compass on both my drones, it was done approx the end of November this year.
Yes I had a LOC8TOR attached at the rear of the drone but it doesn't switch on untill the associated hand set was pressed.
I watched a youtube video regarding the placement of this LOC8TOR and it stated that it would not interfere with the rear compass.

 

[BudWalker]

That flight was the following day ( 15 Dec) I was testing to see if the drone was still operating correctly.
It was at this location ( my front garden) where I calibrated the compass on both my drones, it was done approx the end of November this year.
Yes I had a LOC8TOR attached at the rear of the drone but it doesn't switch on untill the associated hand set was pressed.
I watched a youtube video regarding the placement of this LOC8TOR and it stated that it would not interfere with the rear compass.

So there have been flights since the calibration and this incident?
Did you calibrate with the LOC8TOR attached?
Was the LOC8TOR attached for both of these flights?
I would expect any magnetic distortion caused by the LOC8TOR would be unrelated to if it's powered on or off.