Was it High Wind That Caused This Drone Drowning?

[briano216]

Problem? Drone drowned at sea due to IMU failure

Was unit in a crash? Yes, drowning!

What have you tried so far? Claiming a warranty repair from DJI

What device are you using ( iphone , ipad, Samsung , etc)? iphone 5s

What firmware are you running ( aircraft, remote controller)? It was the latest firmware at the time of the incident on July 26

What Go app version are you using? Not sure how to find out, but I downloaded it in June

Any modification? (if so what) No

Did you change anything or install any apps? (if so what) No

Do you have a video or pictures of the problem? (if so post link) See below

Hi

I was flying my new Mavic Air over water at Edinburgh, Scotland, getting some footage near Blackness Castle, when I received a total of 11 messages stating that the IMU was malfunctioning. I also received 8 high wind warnings, however, only after the IMU started malfunctioning. Prior to that time, I had reached maximum height (120 m) and maximum distance (1000 m - my setting) and had received no high wind warnings. The drone began flying with tilts to the left and then drastic nose down attitudes, making it extremely difficult to bring home, so much so that it ran out of battery about 2 m from the shore and landed in dirty sea water. Because it landed a distance from where I was, it took about 10 minutes to get to the muddy point where it sank and about another 20 minutes to locate it floating just under the surface. It was thoroughly drenched in muddy salt water, rendering it useless.

I sent it to DJI asking for a warranty replacement because, looking at the log and the video, to me it seemed very clear that the incident was caused by the malfunctioning of the IMU. The email I received back stated that the incident was caused by flying the drone in high wind. That seemed so conflicting with what happened and what the log and vidoe showed so I send another email to them. The following is the data I sent in the email:

Attached is the historical data for the weather at Edinburgh that day, July 25. You’ll notice that at the time that I flew, around 1730, the wind speed was approximately 4 km/hr with gusts to 5 km/hr. You mentioned in point 3 that the aircraft flew backward 1.4 m/s, which equates to 5.04 km/hr. The reason the data states that the aircraft was tilted was because the IMU was playing up, as can be seen 11 times in the log data.

You’ll notice that from the time the drone took off until it reached maximum altitude at 3m 1.2s there was no notification of high wind.

At 3m 18.2s maximum distance was reached and there was still no notification of high wind.

At 7m 19.5s the first drastic attitude malfunction occurred after I had told the drone to come home. To this point no high wind notifications had been given. The high wind warnings only start to occur after the IMU began to malfunction.

The first high wind warning was given at 7m 48.3s and then 6 more were given between this time and 12m 56.2s. No more were given after that time. I propose that the IMU failure was the cause of the high wind warnings as the aircraft’s attitude was very erratic and quite drastic, as can be seen in the video at this link: DJI_0442.MP4

The IMU malfunctioning message was given 11 times in total, at the following times:
· 7m 19.5s
· 7m 28.7s
· 7m 34.9s
· 7m 38.6s
· 8m 9.5s
· 9m 38.7s
· 11m 18s
· 12m 59s
· 12m 59.2s
· 13m 0.8s
· 13m 8s

The “Abnormal compass function or GPS signal detected” message was given at 10m 22.8s

The “Switched to backup IMU” message was given at:
· 10m 31.6s
· 10m 41.3s
· 11m 10.1s
· 12m 47.1s
· 13m 5.5s
· 13m 13.2s

The “Navigation system redundancy switch” message was given at:
· 10m 43.3s
· 11m 11.2s
· 12m 48.4s
· 13m 7.3s
· 13m 14.2s

If you look at the video at the above link you will see one instance of what occurred multiple times – the drone’s attitude either tipping to the side, nose diving or climbing to the heavens. This made it extremely difficult to bring the aircraft home because I was having to turn the drone around to restore a workable attitude. But no sooner had the attitude corrected than it did it again. Hence the reason for the very erratic flight pattern (see attached file) and the difficulty I experienced in trying to bring the drone home. The segment shown in the video occurred approximately between 6m and 8m. During that time the following occurred:
· At 7m 23.4s the drone was at 303 ft travelling at 20 mph
· It then descended very rapidly, reaching a speed of 90.6 mph
· At 7m 45.2s it had descended to 243.4 ft, a descent of 59.6 ft in 21.8 secs

A similar incident occurred again:
· At 9m 36.4s the drone was at 393.4 ft travelling at 2.4 mph
· It then descended very rapidly, reaching a speed of 76.3 mph
· At 9m 55.3s it had descended to 266.1 ft, a descent of 127.3 ft in 18.9 secs

Then again:
· At 10m 11.4s the drone was at 334 ft travelling at 1.4 mph
· At 10m 21.2s it had descended to 213.9 ft, a descent of 120.1 ft in 9.8 secs

Then again:
· At 10m 45.5s the drone was at 405.2 ft traveling at 7.7 mph
· It then descended very rapidly, reaching a speed of 85.8 mph
· At 10m 57s it had descended to 298.9 ft, a descent of 106.3 ft in 11.5 secs

These were the most drastic attitude changes. However, if you move through the path of travel that occurs after the first IMU malfunction message, you will notice frequent rotations of the drone. This was because the drone was tilting to the side or facing down or up and the only way I could restore a somewhat straight and level attitude was by rotating the drone.

Clearly there was a malfunction with the recording of the flight on the receiver end as no recording occurs from about 8m into the flight.

I just received a second email back from DJI stating that their conclusion remained the same, that the incident was caused by flying in high wind. Their conclusion states:

For this case, firstly we have to stress that the data analysis should be based on DJI official data viewer, rather than other third-party viewer. Also we have reviewed the data again, and there was no abnormality or product/IMU malfunction found. The result is still the same.

The log viewer I used was the one recommended in this forum - DJI Flight Log Viewer - Phantom Help.

I have attached the log file.

Unfortunately, for some reason the video recorded by the Go 4 app did not capture the full episode but only got the first few minutes. The link to the video is: DJI_0442.MP4

 

[martinbanshee]

Interesting, my Air is about 3 months old and has now developed a faulty IMU. Luckily I noticed the error before flying and I’m now sending it back to Amazon.
It’s never had the slightest of bumps, never mind a crash and it’s down to 1 IMU with no redundancy now.
Anyway, hopefully you’ll get a discount on a replacement from DJI.

 

[sar104]

This was not a wind speed issue - winds were 10 - 15 out of the southwest - it all goes wrong at around 436 seconds when the mode is changed to "Go Home". If you compare the GPS location with the location computed by integration of the IMU recorded velocities north and east, the position values agree well up to that point. But when the aircraft attempts align yaw for RTH the IMU completely loses track of heading and the position data diverge.



After that you have an aircraft in uncontrolled flight, apart from two brief periods of ATTI mode when it successfully stabilized pitch and roll. If it had stayed in ATTI it would have been flyable but in attempting stay in P-GPS (or Sport) the FC had no chance.

This was not compass interference at startup since the initial period of flight was fine - it is clearly some kind of IMU or compass problem leading to fusion failure. I would say it is an obvious warranty replacement case.

 

[briano216]

Hi Sar104

Thanks so much for your analysis!!!!!!!!!!!! That is really helpful!!!!!!!!!! I'm really not sure what to do because DJI on both occasions have come back with the cause of the incident being high wind, but that just doesn't add up! I have the Care Refresh program, but I don't want to use up one of my two available drones when, in my view, it's a warranty issue. The first response I received back was:

For your case CAS-2068121-C3W6L0, we have finished the data analysis, and the result is as follows:
1. The GPS signal was fine and the aircraft worked under GPS mode;
2. T=07:48, h=81.9 m, d=1157.1 m, the app warned that there was strong wind;
3. T=08:24, h=94.7 m, d=1102.4 m, there was no pilot's input but the aircraft tilted 15.2 degree forward and flew backward in 1.4 m/s, which indicated that there was high wind from South-West direction affecting the aircraft;
4. The aircraft was not able to return to home at full speed as it was affected by the wind, and the battery consumed fast;
5. T=13:36, h=124.5 m, d=1167.6 m, battery 14%, Landing was triggered due to the critical low battery, and the aircraft started to descend automatically;
6. T=16:16, h=10.9 m, d=826.5 m, battery 1%, the battery drained off and the app record ended. The aircraft fell to the water finally. The last coordinate: 56.0066190 -3.5138316.
According to the analysis, the incident was caused by improper flight environment. We will have our quotation team to follow up on your case soon. Please pay attention to the flight environment and do not fly in high-wind condition.

I was so sure that it was not a high wind issue, but an aircraft instrument issue, so I wrote the email back that is posted at the top of this thread. I was very polite and very detailed. But the email I got back was on the same note:

For this case, firstly we have to stress that the data analysis should be based on DJI official data viewer, rather than other third-party viewer. Also we have reviewed the data again, and there was no abnormality or product/IMU malfunction found. The result is still the same.
We truly understand your feelings and we also feel sorry for your loss. However the higher the aircraft was, the stronger the wind it would experience. Also, the wind speed that shows in the broadcast also is the speed of near ground wind. It doesn't indicate the wind velocity that the aircraft suffer in high air.

The attitude and the speed of the aircraft strongly demonstrate the wind velocity and direction during the flight.

Please do not fly in high-wind condition or at least ensure the aircraft remains in sight in high-wind condition. Please also lower the altitude of the aircraft and RTH immediately when there is wind.

You can also refer to the Attitude Indicator in the App for the status of the aircraft during the flight.
DJI Safety Tips - How to Use the Attitude Indicator and Map in the DJI GO App:

We know this is disappointing but although we cannot find any product malfunctions or defects and it remains non warranty, a 15% off discount for a new aircraft is a courtesy to try to assist in any way we can. If you would like to use the DJI Care Refresh, please also kindly let us know.

What do you suggest I do?

 

[sar104]

Hi Sar104

Thanks so much for your analysis!!!!!!!!!!!! That is really helpful!!!!!!!!!! I'm really not sure what to do because DJI on both occasions have come back with the cause of the incident being high wind, but that just doesn't add up! I have the Care Refresh program, but I don't want to use up one of my two available drones when, in my view, it's a warranty issue. The first response I received back was:

For your case CAS-2068121-C3W6L0, we have finished the data analysis, and the result is as follows:
1. The GPS signal was fine and the aircraft worked under GPS mode;
2. T=07:48, h=81.9 m, d=1157.1 m, the app warned that there was strong wind;
3. T=08:24, h=94.7 m, d=1102.4 m, there was no pilot's input but the aircraft tilted 15.2 degree forward and flew backward in 1.4 m/s, which indicated that there was high wind from South-West direction affecting the aircraft;
4. The aircraft was not able to return to home at full speed as it was affected by the wind, and the battery consumed fast;
5. T=13:36, h=124.5 m, d=1167.6 m, battery 14%, Landing was triggered due to the critical low battery, and the aircraft started to descend automatically;
6. T=16:16, h=10.9 m, d=826.5 m, battery 1%, the battery drained off and the app record ended. The aircraft fell to the water finally. The last coordinate: 56.0066190 -3.5138316.
According to the analysis, the incident was caused by improper flight environment. We will have our quotation team to follow up on your case soon. Please pay attention to the flight environment and do not fly in high-wind condition.

I was so sure that it was not a high wind issue, but an aircraft instrument issue, so I wrote the email back that is posted at the top of this thread. I was very polite and very detailed. But the email I got back was on the same note:

For this case, firstly we have to stress that the data analysis should be based on DJI official data viewer, rather than other third-party viewer. Also we have reviewed the data again, and there was no abnormality or product/IMU malfunction found. The result is still the same.
We truly understand your feelings and we also feel sorry for your loss. However the higher the aircraft was, the stronger the wind it would experience. Also, the wind speed that shows in the broadcast also is the speed of near ground wind. It doesn't indicate the wind velocity that the aircraft suffer in high air.

The attitude and the speed of the aircraft strongly demonstrate the wind velocity and direction during the flight.

Please do not fly in high-wind condition or at least ensure the aircraft remains in sight in high-wind condition. Please also lower the altitude of the aircraft and RTH immediately when there is wind.

You can also refer to the Attitude Indicator in the App for the status of the aircraft during the flight.
DJI Safety Tips - How to Use the Attitude Indicator and Map in the DJI GO App:

We know this is disappointing but although we cannot find any product malfunctions or defects and it remains non warranty, a 15% off discount for a new aircraft is a courtesy to try to assist in any way we can. If you would like to use the DJI Care Refresh, please also kindly let us know.

What do you suggest I do?

What they have told you is pure nonsense - whoever looked at the log data either doesn't have a clue (likely) or they are deliberately trying to avoid the issue (less likely).

For example, in the two periods where the FC switches to ATTI and the pitch and roll go to zero, the resulting drift converges on the wind velocity. Because the IMU is confused you cannot rely on the x and y velocities, but you can use the GPS position data to calculate velocity, and that gives a consistent maximum value of 7 - 9 m/s (15 - 20 mph) out of the SSW. That is not enough to cause the effects that you see in the video, or cause uncontrolled flight.

To illustrate in velocity space, rather than position that I showed earlier, this graph compares velocity computed from GPS location (and GPS was fine throughout the flight) with the IMU velocity from sensor fusion:



Again, as you can see, the first 430 seconds or so the velocities agree just fine. Then there are three distinct periods, 430 - 500 s, 560 - 680 s, and 760 - 800 s, where the IMU velocities are completely wrong - i.e. the FC has incorrect heading and velocity data. A closer view of the first two of these shows the problem more clearly, and note that the IMU is recording velocity excursions of up to 40 m/s (90 mph). That's not real - the actual velocities, shown by the GPS data as the solid red and green lines, never exceed 20 m/s:




In terms of DJI's example 3, at 8' 24" (504 s), this was one of the periods when the IMU was approximately correctly calculating velocity. The recorded pitch is -15° at that time, and the IMU velocity is 1.4 m/s, although the actual velocity is 3 m/s. In any case, what their analysis missed is that the aircraft had just switched from RTH to P-GPS and was decelerating. Looking forwards in time from there the pitch drops to -7° and the velocity drops to zero - consistent with a 12 - 15 mph wind speed.

Whatever they want to argue about wind speeds, the data clearly show that the IMU was giving wildly incorrect position and velocity calculations. The mobile device DAT file includes the two redundant IMU calculation sets and sensor data - it might be worth a look to try to figure out exactly what was failing.

As to what you should do, I would go back to them and point out the obvious discrepancies in the data. You are welcome to use my graphs as supporting evidence. You can also point to the AirData wind calculations:


although with the caveat that when the IMU velocity data are wrong then the wind calculations are wrong. They are good for the first and last parts of the flight, however. You can also point out that high winds do not lead to completely erratic flight of the kind logged and recorded in the video.

I would definitely not let them off the hook on this one.

 

[briano216]

Thank you sar104!!!!!!!! That is so helpful!!!!!!! This was my first drone, which I bought so I could capture the beauty of Ireland and the UK on our first ever trip well away from Australia. I only got it in June so I'm very green After DJIs second answer, I was feeling quite frustrated because, though I'm green, I was so sure what happened was not caused by high wind. You've given me such helpful information that I now know I was right and can go back to DJI and persist. Thanks so much!!!!!!

Here's a link to the last DAT file recorded by Go 4: 2018-07-25_17-27-28_FLY064.DAT

 

[sar104]

The DAT file shows a lot more information. It's actually a bit surprising that the first part of the flight was stable. It looks like there was magnetic interference at the takeoff point:



The IMU(1) yaw was initialized to 36° true (39° magnetic) from the compass. IMU(0) yaw is significantly different for some reason. On takeoff the magnetic yaw value changes to nearly north - likely as the aircraft lifts away from the source of interference, and then the FC rotates CW to attempt to resolve the immediate discrepancy between the rate gyros and the compass. You end up with a disagreement of around 45°. It flies with varying offsets until around 370 s, when a significant yaw maneuver seems to cause the FC to start registering multiple errors on both IMUs - a fairly common response. To illustrate, below are the FC log entries as the FC lost control.

It should be obvious to anyone who has ever inspected a log that these are not indicative of an aircraft in high winds, but an aircraft where sensor fusion has failed. The cause may have been purely the magnetic conditions at startup but, either way, the failure of the aircraft to abort to ATTI made it un-flyable.

372.412 : 33058 [L-FDI][CTRL]: fault on , horiz_ctrl_fail
372.887 : 33085 [L-FDI][FDI] ns atti(0) tilt is outlier
372.887 : 33085 [L-FDI][FDI] ns atti(1) tilt is outlier
372.889 : 33086 [L-FDI]NS(0) FUSION(0): fault on , disagree
372.889 : 33086 [L-FDI]NS(0) FUSION(1): fault on , disagree
372.889 : 33086 [L-FDI]fatal error: no ns with level < 4 !!!
372.992 : 33092 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
373.658 : 33131 [L-RC]rc cmd:STD_GO_HOME
373.661 : 33131 [L-N_MIS]Init plan 0 from requester 9 for reason 8. Pre plan 5
373.670 : 33131 [L-N_MIS]Exit mis 3. New mis 12
373.670 : 33131 [L-N_MIS]Switch new mis 12 req 9 from mis 3 req 2
374.067 : 33154 [L-NS][AHRS],qg obv buffer size[218/270]
374.067 : 33154 [L-NS][AHRS],pva obv buffer size[238/270]
375.688 : 33250 [L-FDI][sensor_api_info] head(2,0,31,4).|
378.300 : 33403 [L-FDI][CTRL]: fault on , height_ctrl_fail
378.317 : 33404 [L-RC]craft ctrl failed!!!
379.238 : 33458 [L-FDI]NS(0) FUSION(1): fault on , magn_heading_err_large
380.041 : 33505 [L-RC]craft ctrl failed!!!
380.443 : 33528 [L-FDI][FDI] ns atti(0) heading is outlier
380.443 : 33528 [L-FDI][FDI] ns atti(1) heading is outlier
381.764 : 33606 [L-RC]craft ctrl failed!!!
383.488 : 33707 [L-RC]craft ctrl failed!!!
385.212 : 33808 [L-RC]craft ctrl failed!!!
385.533 : 33826 [L-FDI][FDI] ns atti(0) vgz is outlier
385.533 : 33826 [L-FDI][FDI] ns atti(1) vgz is outlier
386.936 : 33909 [L-RC]craft ctrl failed!!!
387.481 : 33941 [L-FDI]NS(0) FUSION(0): fault on , magn_heading_err_large
387.669 : 33952 [L-FDI][CTRL]: fault off, horiz_ctrl_fail
388.164 : 33981 [L-FDI]NS(0) FUSION(0): fault off, disagree
388.164 : 33981 [L-FDI]NS(0) FUSION(1): fault off, disagree
388.182 : 33982 [L-FMU/LED]action changed. Normal Flash(0)
388.182 : 33982 [L-FMU/LED]type:0, normal flash action:
388.182 : 33982 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
388.200 : 33983 [L-FDI][FDI] ns atti(0) tilt is outlier
388.200 : 33983 [L-FDI][FDI] ns atti(1) tilt is outlier
388.215 : 33984 [L-FDI]NS(0) FUSION(0): fault on , disagree
388.215 : 33984 [L-FDI]NS(0) FUSION(1): fault on , disagree
388.215 : 33984 [L-FDI]fatal error: no ns with level < 4 !!!
388.505 : 34000 [L-BATTERY]get_cell_voltage_callback_ack failed!||
388.614 : 34007 [L-FDI][FDI] ns atti(0) heading is outlier
388.614 : 34007 [L-FDI][FDI] ns atti(1) heading is outlier
388.659 : 34010 [L-RC]craft ctrl failed!!!
389.342 : 34050 [L-FDI][sensor_api_info] head(2,0,31,4).|
389.718 : 34072 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
390.383 : 34111 [L-RC]craft ctrl failed!!!
391.919 : 34200 [L-BATTERY]get_cell_voltage_callback_ack failed!||
392.107 : 34212 [L-RC]craft ctrl failed!!!
393.131 : 34272 [L-RC]mode switch changed!
393.131 : 34272 [L-N_MIS]req abort all for rc mode switch
393.134 : 34272 [L-N_MIS]Switch new mis 3 req 2 from mis 26 req 16
393.134 : 34272 [L-FLYMODE]SWITCH2HOVER!
393.830 : 34313 [L-RC]craft ctrl failed!!!
394.035 : 34325 [L-FDI]NS(0) FUSION(0): fault off, magn_heading_err_large
395.554 : 34414 [L-RC]craft ctrl failed!!!
397.278 : 34515 [L-RC]craft ctrl failed!!!
399.002 : 34616 [L-RC]craft ctrl failed!!!
400.725 : 34717 [L-RC]craft ctrl failed!!!
401.306 : 34751 [L-RC]rc cmd:STD_GO_HOME
401.309 : 34751 [L-N_MIS]Init plan 0 from requester 9 for reason 8. Pre plan 5
401.318 : 34751 [L-N_MIS]Exit mis 3. New mis 12
401.318 : 34751 [L-N_MIS]Switch new mis 12 req 9 from mis 3 req 2
402.449 : 34818 [L-RC]craft ctrl failed!!!
402.995 : 34850 [L-FDI][sensor_api_info] head(2,0,31,4).|
404.173 : 34919 [L-RC]craft ctrl failed!!!
405.436 : 34993 [L-FDI][CTRL]: fault off, height_ctrl_fail
405.897 : 35020 [L-RC]craft ctrl failed!!!
407.620 : 35121 [L-RC]craft ctrl failed!!!
408.334 : 35162 [L-FDI][FDI] ns atti(0) tilt is outlier
408.334 : 35162 [L-FDI][FDI] ns atti(1) tilt is outlier
408.974 : 35200 [L-FDI][FDI] ns atti(0) vgz is outlier
408.974 : 35200 [L-FDI][FDI] ns atti(1) vgz is outlier
409.017 : 35202 [L-FDI][FDI] ns atti(0) vgz is outlier
409.017 : 35202 [L-FDI][FDI] ns atti(1) vgz is outlier
409.310 : 35220 [L-FDI]NS(0) FUSION(1): fault off, magn_heading_err_large
409.344 : 35222 [L-RC]craft ctrl failed!!!
411.068 : 35323 [L-RC]craft ctrl failed!!!
411.580 : 35353 [L-FDI][CTRL]: fault on , height_ctrl_fail
411.679 : 35358 [L-FDI][FDI] ns atti(0) alti is outlier
411.679 : 35358 [L-FDI][FDI] ns atti(1) alti is outlier
411.765 : 35363 [L-FDI][FDI] ns atti(0) alti is outlier
411.765 : 35363 [L-FDI][FDI] ns atti(1) alti is outlier
411.811 : 35366 [L-FDI][FDI] ns atti(0) alti is outlier
411.811 : 35366 [L-FDI][FDI] ns atti(1) alti is outlier
411.935 : 35373 [L-FDI][FDI] ns atti(0) alti is outlier
411.935 : 35373 [L-FDI][FDI] ns atti(1) alti is outlier
412.106 : 35383 [L-FDI][FDI] ns atti(0) alti is outlier
412.106 : 35383 [L-FDI][FDI] ns atti(1) alti is outlier
412.277 : 35393 [L-FDI][FDI] ns atti(0) alti is outlier
412.277 : 35393 [L-FDI][FDI] ns atti(1) alti is outlier
412.447 : 35403 [L-FDI][FDI] ns atti(0) alti is outlier
412.447 : 35403 [L-FDI][FDI] ns atti(1) alti is outlier
412.464 : 35404 [L-FDI][FDI] ns atti(0) vgz is outlier
412.464 : 35404 [L-FDI][FDI] ns atti(1) vgz is outlier
412.618 : 35413 [L-FDI][FDI] ns atti(0) alti is outlier
412.618 : 35413 [L-FDI][FDI] ns atti(1) alti is outlier
412.789 : 35423 [L-FDI][FDI] ns atti(0) alti is outlier
412.789 : 35423 [L-FDI][FDI] ns atti(1) alti is outlier
412.792 : 35424 [L-RC]craft ctrl failed!!!
412.959 : 35433 [L-FDI][FDI] ns atti(0) alti is outlier
412.959 : 35433 [L-FDI][FDI] ns atti(1) alti is outlier
413.301 : 35453 [L-FDI][FDI] ns atti(0) alti is outlier
413.301 : 35453 [L-FDI][FDI] ns atti(1) alti is outlier
413.471 : 35463 [L-FDI][FDI] ns atti(0) alti is outlier
413.471 : 35463 [L-FDI][FDI] ns atti(1) alti is outlier
413.817 : 35484 [L-FDI][FDI] ns atti(0) alti is outlier
413.817 : 35484 [L-FDI][FDI] ns atti(1) alti is outlier
413.986 : 35494 [L-FDI]NS(0) COMPASS(0): fault on , interfere
414.515 : 35525 [L-RC]craft ctrl failed!!!
414.959 : 35550 [L-BATTERY]get_cell_voltage_callback_ack failed!||
415.878 : 35604 [L-FDI][FDI] ns atti(0) heading is outlier
415.878 : 35604 [L-FDI][FDI] ns atti(1) heading is outlier
416.239 : 35626 [L-RC]craft ctrl failed!!!
416.262 : 35627 [L-FDI][FDI] ns atti(0) tilt is outlier
416.262 : 35627 [L-FDI][FDI] ns atti(1) tilt is outlier
416.273 : 35628 [L-FDI]NS(0) FUSION(1): fault on , magn_heading_err_large
416.648 : 35650 [L-FDI][sensor_api_info] head(2,0,31,4).|
417.092 : 35676 [L-FDI][CTRL]: fault off, height_ctrl_fail
417.963 : 35727 [L-RC]craft ctrl failed!!!
419.686 : 35828 [L-RC]craft ctrl failed!!!
419.874 : 35839 [L-FDI]NS(0) FUSION(0): fault off, disagree
419.874 : 35839 [L-FDI]NS(0) FUSION(1): fault off, disagree
419.926 : 35842 [L-FMU/LED]action changed. Normal Flash(0)
419.926 : 35842 [L-FMU/LED]type:0, normal flash action:
419.926 : 35842 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
420.215 : 35859 [L-FDI]NS(0) COMPASS(0): fault off, interfere
421.410 : 35929 [L-RC]craft ctrl failed!!!
423.134 : 36030 [L-RC]craft ctrl failed!!!
424.858 : 36131 [L-RC]craft ctrl failed!!!
425.267 : 36154 [L-NS][AHRS],qg obv buffer size[218/270]
425.267 : 36154 [L-NS][AHRS],pva obv buffer size[238/270]
425.336 : 36159 [L-RC]rc cmd:STD_PAUSE_CUR_MIS
425.336 : 36159 [L-N_MIS]req abort all for CTRL rc cannot pause gohome but exit gohome
425.336 : 36159 [L-FLYMODE]CTRL rc cannot pause gohome but exit gohome
425.339 : 36159 [L-N_MIS]Switch new mis 3 req 2 from mis 26 req 16
425.339 : 36159 [L-FLYMODE]SWITCH2HOVER!
426.035 : 36200 [L-RC]Emergency break. rc vib!
426.070 : 36202 [L-FMU/LED]action changed. near hori or vert limit:near horiz limit(2)
427.367 : 36278 [L-FDI][CTRL]: fault on , horiz_ctrl_fail
427.930 : 36311 [L-FDI][CTRL]: fault off, horiz_ctrl_fail
427.957 : 36312 [L-FDI][FDI] ns atti(0) tilt is outlier
427.957 : 36312 [L-FDI][FDI] ns atti(1) tilt is outlier
427.964 : 36313 [L-FDI]NS(0) FUSION(0): fault on , disagree
427.964 : 36313 [L-FDI]NS(0) FUSION(1): fault on , disagree
427.964 : 36313 [L-FDI]fatal error: no ns with level < 4 !!!
428.732 : 36358 [L-FDI]NS(0) FUSION(0): fault off, disagree
428.732 : 36358 [L-FDI]NS(0) FUSION(1): fault off, disagree
428.919 : 36369 [L-FDI]NS(0) FUSION(1): fault off, magn_heading_err_large
429.130 : 36381 [L-FDI][FDI] ns atti(0) tilt is outlier
429.130 : 36381 [L-FDI][FDI] ns atti(1) tilt is outlier
429.141 : 36382 [L-FDI]NS(0) FUSION(0): fault on , disagree
429.141 : 36382 [L-FDI]NS(0) FUSION(1): fault on , disagree
429.141 : 36382 [L-FDI]fatal error: no ns with level < 4 !!!
429.483 : 36402 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
429.551 : 36406 [L-FDI]NS(0) FUSION(0): fault off, disagree
429.551 : 36406 [L-FDI]NS(0) FUSION(1): fault off, disagree
429.569 : 36407 [L-FDI][FDI] ns atti(0) tilt is outlier
429.569 : 36407 [L-FDI][FDI] ns atti(1) tilt is outlier
429.585 : 36408 [L-FDI]NS(0) FUSION(0): fault on , disagree
429.585 : 36408 [L-FDI]NS(0) FUSION(1): fault on , disagree
429.585 : 36408 [L-FDI]fatal error: no ns with level < 4 !!!​

 

[sar104]

@BudWalker - do you see anything else here? I have the same question that seems to arise not infrequently - why did IMU(0) get initialized 35° off IMU(1)? In this case the DAT file starts before GPS was acquired and yaw was initialized. And, at 2.3 s, why did the two z-axis rate gyros record completely opposite rotations?

 

[BudWalker]

The DAT file shows a lot more information. It's actually a bit surprising that the first part of the flight was stable. It looks like there was magnetic interference at the takeoff point:

View attachment 45358

The IMU(1) yaw was initialized to 36° true (39° magnetic) from the compass. IMU(0) yaw is significantly different for some reason. On takeoff the magnetic yaw value changes to nearly north - likely as the aircraft lifts away from the source of interference, and then the FC rotates CW to attempt to resolve the immediate discrepancy between the rate gyros and the compass. You end up with a disagreement of around 45°. It flies with varying offsets until around 370 s, when a significant yaw maneuver seems to cause the FC to start registering multiple errors on both IMUs - a fairly common response. To illustrate, below are the FC log entries as the FC lost control.

It should be obvious to anyone who has ever inspected a log that these are not indicative of an aircraft in high winds, but an aircraft where sensor fusion has failed. The cause may have been purely the magnetic conditions at startup but, either way, the failure of the aircraft to abort to ATTI made it un-flyable.

372.412 : 33058 [L-FDI][CTRL]: fault on , horiz_ctrl_fail
372.887 : 33085 [L-FDI][FDI] ns atti(0) tilt is outlier
372.887 : 33085 [L-FDI][FDI] ns atti(1) tilt is outlier
372.889 : 33086 [L-FDI]NS(0) FUSION(0): fault on , disagree
372.889 : 33086 [L-FDI]NS(0) FUSION(1): fault on , disagree
372.889 : 33086 [L-FDI]fatal error: no ns with level < 4 !!!
372.992 : 33092 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
373.658 : 33131 [L-RC]rc cmd:STD_GO_HOME
373.661 : 33131 [L-N_MIS]Init plan 0 from requester 9 for reason 8. Pre plan 5
373.670 : 33131 [L-N_MIS]Exit mis 3. New mis 12
373.670 : 33131 [L-N_MIS]Switch new mis 12 req 9 from mis 3 req 2
374.067 : 33154 [L-NS][AHRS],qg obv buffer size[218/270]
374.067 : 33154 [L-NS][AHRS],pva obv buffer size[238/270]
375.688 : 33250 [L-FDI][sensor_api_info] head(2,0,31,4).|
378.300 : 33403 [L-FDI][CTRL]: fault on , height_ctrl_fail
378.317 : 33404 [L-RC]craft ctrl failed!!!
379.238 : 33458 [L-FDI]NS(0) FUSION(1): fault on , magn_heading_err_large
380.041 : 33505 [L-RC]craft ctrl failed!!!
380.443 : 33528 [L-FDI][FDI] ns atti(0) heading is outlier
380.443 : 33528 [L-FDI][FDI] ns atti(1) heading is outlier
381.764 : 33606 [L-RC]craft ctrl failed!!!
383.488 : 33707 [L-RC]craft ctrl failed!!!
385.212 : 33808 [L-RC]craft ctrl failed!!!
385.533 : 33826 [L-FDI][FDI] ns atti(0) vgz is outlier
385.533 : 33826 [L-FDI][FDI] ns atti(1) vgz is outlier
386.936 : 33909 [L-RC]craft ctrl failed!!!
387.481 : 33941 [L-FDI]NS(0) FUSION(0): fault on , magn_heading_err_large
387.669 : 33952 [L-FDI][CTRL]: fault off, horiz_ctrl_fail
388.164 : 33981 [L-FDI]NS(0) FUSION(0): fault off, disagree
388.164 : 33981 [L-FDI]NS(0) FUSION(1): fault off, disagree
388.182 : 33982 [L-FMU/LED]action changed. Normal Flash(0)
388.182 : 33982 [L-FMU/LED]type:0, normal flash action:
388.182 : 33982 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
388.200 : 33983 [L-FDI][FDI] ns atti(0) tilt is outlier
388.200 : 33983 [L-FDI][FDI] ns atti(1) tilt is outlier
388.215 : 33984 [L-FDI]NS(0) FUSION(0): fault on , disagree
388.215 : 33984 [L-FDI]NS(0) FUSION(1): fault on , disagree
388.215 : 33984 [L-FDI]fatal error: no ns with level < 4 !!!
388.505 : 34000 [L-BATTERY]get_cell_voltage_callback_ack failed!||
388.614 : 34007 [L-FDI][FDI] ns atti(0) heading is outlier
388.614 : 34007 [L-FDI][FDI] ns atti(1) heading is outlier
388.659 : 34010 [L-RC]craft ctrl failed!!!
389.342 : 34050 [L-FDI][sensor_api_info] head(2,0,31,4).|
389.718 : 34072 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
390.383 : 34111 [L-RC]craft ctrl failed!!!
391.919 : 34200 [L-BATTERY]get_cell_voltage_callback_ack failed!||
392.107 : 34212 [L-RC]craft ctrl failed!!!
393.131 : 34272 [L-RC]mode switch changed!
393.131 : 34272 [L-N_MIS]req abort all for rc mode switch
393.134 : 34272 [L-N_MIS]Switch new mis 3 req 2 from mis 26 req 16
393.134 : 34272 [L-FLYMODE]SWITCH2HOVER!
393.830 : 34313 [L-RC]craft ctrl failed!!!
394.035 : 34325 [L-FDI]NS(0) FUSION(0): fault off, magn_heading_err_large
395.554 : 34414 [L-RC]craft ctrl failed!!!
397.278 : 34515 [L-RC]craft ctrl failed!!!
399.002 : 34616 [L-RC]craft ctrl failed!!!
400.725 : 34717 [L-RC]craft ctrl failed!!!
401.306 : 34751 [L-RC]rc cmd:STD_GO_HOME
401.309 : 34751 [L-N_MIS]Init plan 0 from requester 9 for reason 8. Pre plan 5
401.318 : 34751 [L-N_MIS]Exit mis 3. New mis 12
401.318 : 34751 [L-N_MIS]Switch new mis 12 req 9 from mis 3 req 2
402.449 : 34818 [L-RC]craft ctrl failed!!!
402.995 : 34850 [L-FDI][sensor_api_info] head(2,0,31,4).|
404.173 : 34919 [L-RC]craft ctrl failed!!!
405.436 : 34993 [L-FDI][CTRL]: fault off, height_ctrl_fail
405.897 : 35020 [L-RC]craft ctrl failed!!!
407.620 : 35121 [L-RC]craft ctrl failed!!!
408.334 : 35162 [L-FDI][FDI] ns atti(0) tilt is outlier
408.334 : 35162 [L-FDI][FDI] ns atti(1) tilt is outlier
408.974 : 35200 [L-FDI][FDI] ns atti(0) vgz is outlier
408.974 : 35200 [L-FDI][FDI] ns atti(1) vgz is outlier
409.017 : 35202 [L-FDI][FDI] ns atti(0) vgz is outlier
409.017 : 35202 [L-FDI][FDI] ns atti(1) vgz is outlier
409.310 : 35220 [L-FDI]NS(0) FUSION(1): fault off, magn_heading_err_large
409.344 : 35222 [L-RC]craft ctrl failed!!!
411.068 : 35323 [L-RC]craft ctrl failed!!!
411.580 : 35353 [L-FDI][CTRL]: fault on , height_ctrl_fail
411.679 : 35358 [L-FDI][FDI] ns atti(0) alti is outlier
411.679 : 35358 [L-FDI][FDI] ns atti(1) alti is outlier
411.765 : 35363 [L-FDI][FDI] ns atti(0) alti is outlier
411.765 : 35363 [L-FDI][FDI] ns atti(1) alti is outlier
411.811 : 35366 [L-FDI][FDI] ns atti(0) alti is outlier
411.811 : 35366 [L-FDI][FDI] ns atti(1) alti is outlier
411.935 : 35373 [L-FDI][FDI] ns atti(0) alti is outlier
411.935 : 35373 [L-FDI][FDI] ns atti(1) alti is outlier
412.106 : 35383 [L-FDI][FDI] ns atti(0) alti is outlier
412.106 : 35383 [L-FDI][FDI] ns atti(1) alti is outlier
412.277 : 35393 [L-FDI][FDI] ns atti(0) alti is outlier
412.277 : 35393 [L-FDI][FDI] ns atti(1) alti is outlier
412.447 : 35403 [L-FDI][FDI] ns atti(0) alti is outlier
412.447 : 35403 [L-FDI][FDI] ns atti(1) alti is outlier
412.464 : 35404 [L-FDI][FDI] ns atti(0) vgz is outlier
412.464 : 35404 [L-FDI][FDI] ns atti(1) vgz is outlier
412.618 : 35413 [L-FDI][FDI] ns atti(0) alti is outlier
412.618 : 35413 [L-FDI][FDI] ns atti(1) alti is outlier
412.789 : 35423 [L-FDI][FDI] ns atti(0) alti is outlier
412.789 : 35423 [L-FDI][FDI] ns atti(1) alti is outlier
412.792 : 35424 [L-RC]craft ctrl failed!!!
412.959 : 35433 [L-FDI][FDI] ns atti(0) alti is outlier
412.959 : 35433 [L-FDI][FDI] ns atti(1) alti is outlier
413.301 : 35453 [L-FDI][FDI] ns atti(0) alti is outlier
413.301 : 35453 [L-FDI][FDI] ns atti(1) alti is outlier
413.471 : 35463 [L-FDI][FDI] ns atti(0) alti is outlier
413.471 : 35463 [L-FDI][FDI] ns atti(1) alti is outlier
413.817 : 35484 [L-FDI][FDI] ns atti(0) alti is outlier
413.817 : 35484 [L-FDI][FDI] ns atti(1) alti is outlier
413.986 : 35494 [L-FDI]NS(0) COMPASS(0): fault on , interfere
414.515 : 35525 [L-RC]craft ctrl failed!!!
414.959 : 35550 [L-BATTERY]get_cell_voltage_callback_ack failed!||
415.878 : 35604 [L-FDI][FDI] ns atti(0) heading is outlier
415.878 : 35604 [L-FDI][FDI] ns atti(1) heading is outlier
416.239 : 35626 [L-RC]craft ctrl failed!!!
416.262 : 35627 [L-FDI][FDI] ns atti(0) tilt is outlier
416.262 : 35627 [L-FDI][FDI] ns atti(1) tilt is outlier
416.273 : 35628 [L-FDI]NS(0) FUSION(1): fault on , magn_heading_err_large
416.648 : 35650 [L-FDI][sensor_api_info] head(2,0,31,4).|
417.092 : 35676 [L-FDI][CTRL]: fault off, height_ctrl_fail
417.963 : 35727 [L-RC]craft ctrl failed!!!
419.686 : 35828 [L-RC]craft ctrl failed!!!
419.874 : 35839 [L-FDI]NS(0) FUSION(0): fault off, disagree
419.874 : 35839 [L-FDI]NS(0) FUSION(1): fault off, disagree
419.926 : 35842 [L-FMU/LED]action changed. Normal Flash(0)
419.926 : 35842 [L-FMU/LED]type:0, normal flash action:
419.926 : 35842 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
420.215 : 35859 [L-FDI]NS(0) COMPASS(0): fault off, interfere
421.410 : 35929 [L-RC]craft ctrl failed!!!
423.134 : 36030 [L-RC]craft ctrl failed!!!
424.858 : 36131 [L-RC]craft ctrl failed!!!
425.267 : 36154 [L-NS][AHRS],qg obv buffer size[218/270]
425.267 : 36154 [L-NS][AHRS],pva obv buffer size[238/270]
425.336 : 36159 [L-RC]rc cmd:STD_PAUSE_CUR_MIS
425.336 : 36159 [L-N_MIS]req abort all for CTRL rc cannot pause gohome but exit gohome
425.336 : 36159 [L-FLYMODE]CTRL rc cannot pause gohome but exit gohome
425.339 : 36159 [L-N_MIS]Switch new mis 3 req 2 from mis 26 req 16
425.339 : 36159 [L-FLYMODE]SWITCH2HOVER!
426.035 : 36200 [L-RC]Emergency break. rc vib!
426.070 : 36202 [L-FMU/LED]action changed. near hori or vert limit:near horiz limit(2)
427.367 : 36278 [L-FDI][CTRL]: fault on , horiz_ctrl_fail
427.930 : 36311 [L-FDI][CTRL]: fault off, horiz_ctrl_fail
427.957 : 36312 [L-FDI][FDI] ns atti(0) tilt is outlier
427.957 : 36312 [L-FDI][FDI] ns atti(1) tilt is outlier
427.964 : 36313 [L-FDI]NS(0) FUSION(0): fault on , disagree
427.964 : 36313 [L-FDI]NS(0) FUSION(1): fault on , disagree
427.964 : 36313 [L-FDI]fatal error: no ns with level < 4 !!!
428.732 : 36358 [L-FDI]NS(0) FUSION(0): fault off, disagree
428.732 : 36358 [L-FDI]NS(0) FUSION(1): fault off, disagree
428.919 : 36369 [L-FDI]NS(0) FUSION(1): fault off, magn_heading_err_large
429.130 : 36381 [L-FDI][FDI] ns atti(0) tilt is outlier
429.130 : 36381 [L-FDI][FDI] ns atti(1) tilt is outlier
429.141 : 36382 [L-FDI]NS(0) FUSION(0): fault on , disagree
429.141 : 36382 [L-FDI]NS(0) FUSION(1): fault on , disagree
429.141 : 36382 [L-FDI]fatal error: no ns with level < 4 !!!
429.483 : 36402 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
429.551 : 36406 [L-FDI]NS(0) FUSION(0): fault off, disagree
429.551 : 36406 [L-FDI]NS(0) FUSION(1): fault off, disagree
429.569 : 36407 [L-FDI][FDI] ns atti(0) tilt is outlier
429.569 : 36407 [L-FDI][FDI] ns atti(1) tilt is outlier
429.585 : 36408 [L-FDI]NS(0) FUSION(0): fault on , disagree
429.585 : 36408 [L-FDI]NS(0) FUSION(1): fault on , disagree
429.585 : 36408 [L-FDI]fatal error: no ns with level < 4 !!!​

I suspect that IMU 1 was being used and that's why IMU 0 Yaw wasn't initialized.

This seems to be a textbook example of launching at a geomagnetically distorted location. Using the ultrasonic height sensor it's clear that the Yaw/magYaw separation occurs as the MA leaves the ground.

 

[sar104]

I suspect that IMU 1 was being used and that's why IMU 0 Yaw wasn't initialized.

This seems to be a textbook example of launching at a geomagnetically distorted location. Using the ultrasonic height sensor it's clear that the Yaw/magYaw separation occurs as the MA leaves the ground.
View attachment 45361

Yes - the separation was clearly with takeoff. It looks as if somewhat stable flight is resulting from a 45° separation. We know that 90° produces the toilet bowl effect, for obvious reasons, and 180° leads to the linear runaway - again as expected. It was definitely using IMU(1) - I'm just not sure why it would not have initialized IMU(0); it's presumably not much use as a backup IMU if its yaw is not initialized.

 

[briano216]

Hi sar104 and BudWalker

Thanks so much for your research!!!!! You've left the DJI analyst I had for dead!!!!!! I wanted to know what had caused the incident, as I was so sure it wasn't high wind, and you've confirmed that so clearly for me. I really appreciate your help.

So it seems that magnetic interference at takeoff was the cause of the incident (I thought I took off from bitumen, but I guess there's still no guarantee it doesn't have metal under it). If that's so, then that would seem to be my responsibility, not DJI's. I'm in the DJI Care Refresh program so they've said it will only cost $129 AUD to get the aircraft fixed (I'd rather a replacement) so sounds like I need to accept that option. Am I on the right track there?

Again, thanks for your generous help!!!!!

 

[sar104]

Hi sar104 and BudWalker

Thanks so much for your research!!!!! You've left the DJI analyst I had for dead!!!!!! I wanted to know what had caused the incident, as I was so sure it wasn't high wind, and you've confirmed that so clearly for me. I really appreciate your help.

So it seems that magnetic interference at takeoff was the cause of the incident (I thought I took off from bitumen, but I guess there's still no guarantee it doesn't have metal under it). If that's so, then that would seem to be my responsibility, not DJI's. I'm in the DJI Care Refresh program so they've said it will only cost $129 AUD to get the aircraft fixed (I'd rather a replacement) so sounds like I need to accept that option. Am I on the right track there?

Again, thanks for your generous help!!!!!

DJI can certainly argue that it is your responsibility to ensure that the takeoff location is free from magnetic interference, although they still do not provide any guidance on how to do that (the best method is to ensure that the aircraft heading arrow in the GO app is pointing in the direction that the aircraft is actually facing). Additionally, if the FC behaved more like it does in earlier DJI models, and switched to ATTI mode when this happens rather than flailing around trying to resolve the problem (which it rarely seems to succeed in doing), then the aircraft would still be flyable. And they often seem to replace aircraft under warranty in these situations, without much explanation, although they are more reluctant without the crashed aircraft to examine.

I would contact them again, point out that their previous analysis is clearly incorrect, that the log data are completely unambiguous that sensor fusion failure occurred (no need to go into why that happened) and argue that the failure of the aircraft to switch to ATTI mode prevented you from flying it back and thus directly caused the loss.

 

[NAVMAV]

At the risk of raising something may be totally useless, we wonder if those "Falling" speeds of 85-95 MPH are correct? Since this occurred in Great Britain, speed is usually calculated in KmPHs, not MPHs, right?

We assume that's an Operator Setting, BUT you might double check the setting to ensure you were actually using Miles, rather than Kms? No point accidentally giving DJI an "Incompetent Operator" lever they might try to employ in further negotiations.

Rgds, NavMav

 

[TuT]

When my Phantom 3 fell from the sky after a firmware upgrade DJI provided me with 3 conflicting reasons for my crash. After I proved each explanation incorrect they would make up a new one.

I have Zero trust in DJI warranty. Folks in the USA seem to have a better experience.

 

[sar104]

At the risk of raising something may be totally useless, we wonder if those "Falling" speeds of 85-95 MPH are correct? Since this occurred in Great Britain, speed is usually calculated in KmPHs, not MPHs, right?

We assume that's an Operator Setting, BUT you might double check the setting to ensure you were actually using Miles, rather than Kms? No point accidentally giving DJI an "Incompetent Operator" lever they might try to employ in further negotiations.

Rgds, NavMav

The logs use metric units - there is no ambiguity.

 

[briano216]

When my Phantom 3 fell from the sky after a firmware upgrade DJI provided me with 3 conflicting reasons for my crash. After I proved each explanation incorrect they would make up a new one.

I have Zero trust in DJI warranty. Folks in the USA seem to have a better experience.

Yes, I've now received 3 emails, each stating that it was caused by high wind and that there's no evidence of an IMU malfunction and yet the log shows that fact 11 times.

 

[briano216]

DJI can certainly argue that it is your responsibility to ensure that the takeoff location is free from magnetic interference, although they still do not provide any guidance on how to do that (the best method is to ensure that the aircraft heading arrow in the GO app is pointing in the direction that the aircraft is actually facing). Additionally, if the FC behaved more like it does in earlier DJI models, and switched to ATTI mode when this happens rather than flailing around trying to resolve the problem (which it rarely seems to succeed in doing), then the aircraft would still be flyable. And they often seem to replace aircraft under warranty in these situations, without much explanation, although they are more reluctant without the crashed aircraft to examine.

I would contact them again, point out that their previous analysis is clearly incorrect, that the log data are completely unambiguous that sensor fusion failure occurred (no need to go into why that happened) and argue that the failure of the aircraft to switch to ATTI mode prevented you from flying it back and thus directly caused the loss.

The puzzling thing on the magnetic interference at take off is that on that flight, I did exactly what you said. I actually do it for every flight. I check that the compass in Go 4 is pointing in the same direction as the drone and, if not, I do a compass re-callibration. I hold it above my head and rotate as specified. I found rotating it in my hand was too bumpy so I do it above my head. If others are around I feel a bit silly, but it works much better! I actually had to do it before that flight. So when I put the drone on the ground, the Go 4 compass and the drone were pointing in the same direction.

I'll do as you've suggested and persist with them. In Australia we have bodies we can work with if we're not happy with a companies outcome so I'm going to start working with them I think.

 

[sar104]

The puzzling thing on the magnetic interference at take off is that on that flight, I did exactly what you said. I actually do it for every flight. I check that the compass in Go 4 is pointing in the same direction as the drone and, if not, I do a compass re-callibration. I hold it above my head and rotate as specified. I found rotating it in my hand was too bumpy so I do it above my head. If others are around I feel a bit silly, but it works much better! I actually had to do it before that flight. So when I put the drone on the ground, the Go 4 compass and the drone were pointing in the same direction.

I'll do as you've suggested and persist with them. In Australia we have bodies we can work with if we're not happy with a companies outcome so I'm going to start working with them I think.

Firstly - a compass calibration is the last thing that you should do if you experience interference - that is not going to help at all. Secondly, you can see from the data that as the aircraft takes off and rises, the magnetic yaw changes even though the aircraft does not rotate - that is strong evidence of magnetic interference. It's not your main argument in this case though. Did you point out to DJI support the log entries showing sensor fusion failure?

 

[TuT]

The puzzling thing on the magnetic interference at take off is that on that flight, I did exactly what you said. I actually do it for every flight. I check that the compass in Go 4 is pointing in the same direction as the drone and, if not, I do a compass re-callibration. I hold it above my head and rotate as specified. I found rotating it in my hand was too bumpy so I do it above my head. If others are around I feel a bit silly, but it works much better! I actually had to do it before that flight. So when I put the drone on the ground, the Go 4 compass and the drone were pointing in the same direction.

I'll do as you've suggested and persist with them. In Australia we have bodies we can work with if we're not happy with a companies outcome so I'm going to start working with them I think.

I see you are also in Australia. I hate to say it but I believe DJI Australia are liars, cheats or just plain incompetent. Closer to the first two unfortunately.
I tried Australian Consumer Law but got no traction.
In the end I just rejected their $799 repair quote and fixed it myself for $99.

The worst part of the whole experience was, as an electronic engineer, I could disprove their excuses with their own logs. 3 times they made up NEW reasons for my Phantom's failure, their final being that I hit a 5m telegraph pole that was over 35m below the drone in a valley.

I can't say Rise Above Drones Australia was any help. They just towed the DJI line and even threatened to report me to CASA for flying near sunset. Avoid them as well.

I wish you the best of luck.

 

[briano216]

Firstly - a compass calibration is the last thing that you should do if you experience interference - that is not going to help at all. Secondly, you can see from the data that as the aircraft takes off and rises, the magnetic yaw changes even though the aircraft does not rotate - that is strong evidence of magnetic interference. It's not your main argument in this case though. Did you point out to DJI support the log entries showing sensor fusion failure?

Since each of the three emails I've received from DJI state that they take no notice of third-party data, I sent the following email to them:

Thanks for your email.


Given that DJI does not take notice of evidence from third-party software, could you please email me the following graphs from your DJI official software to support transparency in this process:

• The graph that compares the GPS location to the IMU-computed location, as shown in attachment 1.
• The graph that compares the GPS and IMU velocities, as shown in attachment 2.

Also, if I am not satisfied with the findings from DJI, I understand that the next step is to contact ACCC (Australian Competition and Consumer Commission). Is that correct?

Have a great week.

I thought I would attempt to get the same graphs from DJI and then pursue the sensor fusion failure avenue.

Being very new to drones, I'm keen to learn. You said the way I did the compass re-callibration wouldn't have helped. How do you tell if the compass just needs re-callibration because it's been in the car too long or otherwise and when it actually doesn't need re-callibration because the discrepancy between the way the drone is pointing and the direction the Go 4 app says the drone is facing is due to magnetic interference? I'm guessing you would walk to a nearby area that is most likely free of metal, such as a grassed area and see if the two agree there.