Why I do believe there's a HUGE problem with the Mavic Mini - Flew AWAY the 2nd day

[Dave Maine]

I used to fly frequently in a large park in Cambridge, MA, in a large park ringed by apartments and condos. WiFi problems were a constant problem with the Phantom 3 4K, Mavic Air, and Phantom 4 adv. Frequently the effects were present starting at 600 ft out. Sometimes you could clear them by climbing to 400 ft, mostly not. Shifting flight direction sometimes helped, sometimes not.

The problem was easily seen using a laptop with a software program which measured the WiFi signal environment. These measurements consistently detected more than 50 active wireless sources. Their signal strength equaled or exceeded the drone and controller frequently. Only the RF pairing allowed any operation to work at all.

The outgoing control signals would generally be more robust, and work further out than the video returns. Drone WiFi operation will always be problematic in urban environments due to the physics of radio frequency transmission. Ocusync, Lightbridge, and other DJI inventions may make minor improvements, but they won’t change the underlying physics of interference. Blaming DJI for this problem is totally irrational.

In any case, data is needed to evaluate what happened scientifically, rather than emotionally.

 

[Mikey's]

Sorry for your lost. I am not so experienced pilot and can't analyse flight data so I have no clue why it did not return to home but I am a WLAN specialist. Please, never ever fly your AC relying on WLAN control in urban area out of your sight. 300 m is tooooo much. Even in this case the first message you got from your RC was: "Curent WiFi environment’s experiencing high interference. Fly with caution." In urban area you can lose the WLAN signal very suddenly and unexpectedly even after couple of meters. Be careful especially in hotels neiborhood.

 

[Deleted member 94047]

If the cable between the RC & device came loose the RC would still be in control but the video stream should have stopped in the device as shown in below pic. In this case the RC LED is still green & no auto RTH will be initiated.

I think you make a very good point. A cable disconnect can indeed explain the abrupt ending of the log file and why there was no subsequent RTH by the AC. However, wouldn't the AC still return home upon reaching the low battery threshold? From the MM manual p. 13



Assuming OP stayed long enough at the home point, they would have been able to recover the AC if this was the case.

 

[slup]

... However, wouldn't the AC still return home upon reaching the low battery threshold?...

Assuming OP stayed long enough at the home point, they would have been able to recover the AC if this was the case.

Yeah ... but did he stayed long enough considering the battery were at 76% in the end of the log?
And what happened during that period of battery depletion until low battery RTH kicked in ..? The area were the log ended looked like this & the red cross marks approx end of log.
The Mini could easily have drifted at 33 m altitude into something during that long period?

 

[sar104]

Pilot error
You flew into 22 mph (9.8m/s) winds.
Wind direction was towards you.
Check along the flight path slightly to the north.
Please use UAV forecast to tell you of upper level winds.
It shows 2.7 mph (1.7m/s) winds on the ground which can be deceiving.
See the following log of yours with a subscription account.
Weather, in flight winds and wind map

Unfortunately this event demonstrates that the AirData wind estimate is not correct for the Mini - perhaps because it is using an aerodynamic model for the Spark. If you look at aircraft speed and aircraft tilt, and focus on the periods of the flight when the aircraft was stationary, you will notice that the tilt was no more than 1°. That indicates that there was very little wind.

 

[sar104]

Here it is 19-12-26-04-33-04_FLY026.DAT

Also... I have message #45 under moderation approval, I am sorry not to be able to promptly respond to one analysis of the flight data.

There are a number of errors, but nothing definitive that I can see in the DAT file. In particular there are cell voltage time out reports, and some GEO position error reports, but those don't look likely to account for the sudden disappearance of the link and failure to RTH. Have you sent both the txt and DAT logs to DJI?

 

[Dakarenduro]

Yeah ... but did he stayed long enough considering the battery were at 76% in the end of the log?
And what happened during that period of battery depletion until low battery RTH kicked in ..? The area were the log ended looked like this & the red cross marks approx end of log.
The Mini could easily have drifted at 33 m altitude into something during that long period?
View attachment 89512

Did someone check the roofs??

 

[Capt KO]

DJI, we have a HUGE problem with the Mavic Mini, NEW Pilots! Your phones must be ringing off the hook. Glad I learned on Phantoms, makes my Mini a fun little camera, but with more limitations. All good. Have fun, fly safe.
Capt?

 

[Johnny Diaz]

You’re not the only one to lose your Mavic mini in a fly away, too bad you didn’t read about all the others who have experienced the same problem before buying. There has been quite a few. Sorry to hear about your loss, sometime the cheapest model isn’t the best, obviously.

 

[Sticksandgrips]

again out of line of site and multiple warnings for bad connection. this guy is asking for trouble and wont man up and say i made a mistake. nevermind he was flying it above streets with people and cars. there is so much wrong with this post. I am new to drones and my mini but seeing his video of the flight is proof enough for me this was 100% user error.


And the fact he is combative about it makes it even worse.

 

[FoxhallGH]

[Post removed at the request of @m80116 }

 

[Capt KO]

Not the answers that the OP was hoping for. But for the 1000 guests watching this thread, interesting reading. ? New Mini pilots! Pay Attention! Don’t panic and you will have a blast with your new Mavic mini.

My Mini is a sparrow compared to my eagle-like Phantoms and treat each accordingly. No comparison, different uses.

 

[slup]

Looking at the map in post #54 and doing a bit of basic trig' - shows that the Mavic had an extremely narrow 'window' that it would have been in, to be LoS with the controller. The flight took place down the Via Roma in San Goivanni Valdarno, toward the square and a building called the Pallazo D'Arnolfo (that houses the 'Museo delle Terre Nouve'). I've not been able to find a stat' that gives the exact height of the tower on that building, but if you look at Google earth and other photo's, and use people at the base of it as a reference, the tower of the building must be very close to 100ft tall.

View attachment 89553

Why that's significant, is that at the furthest point [1,096 ft distant] in the flight, the Mav' Mini appears to have been at an altitude of 108.3 ft. The angle between the horizontal view down the Via Roma and the drone at that point, was around 5.7 degrees. The photo I have attached, is the [Google Earth] view down the Via Roma, from the park where the flight originated - and unless the operator moved a lot, would have been very close to being his view of the flight operations area.
The significant thing is that at the last GPS reported point of the flight (identified as point J on the GPS view) the Mavic was only slightly higher than the Pallazo D'Arnolfo tower that you can see at the end of the street, and about 150 to 200 ft this side of it. The course shown on the map shows the Mavic drifting right on its way out (at the second pedestrian crossing), which would put it behind the large 6-floor building you can see on the right, about 1/3 of the way down the street - and effectively shadowing the Mavic from the controller.

I may be wrong, but that street looks residential. If so, you could expect a lot of WiFi usage for internet access in the apartments - and therefore a lot of cross-channel interference.
In addition to that, to maintain 100% clear line of sight between the controller and Mavic, you have a very small window (less than 10 degrees wide) to fly the Mavic within, while it is at the 108 ft height and 1,000 ft distance.

Since the RTH was not triggered, I'd propose that the Mavic still thought it had signal - due to the interference from other local WiFi sources, but the pilot's controller signal would have been severely attenuated due to signal block and bounce off the buildings down the Via Roma.

I believe that under Italian regulations, Drones must be identified with an Operator ID [ Drone Laws in Italy | UAV Coach (2019) ], and therefore, if this Mavic has dropped into somebody's garden, the Operator ID may get it identified & returned to the OP.

Will not argue about your thesis ... pure visual it would have made me back off from flying there at those distances.

The disturbing with this is that non of that is seen in neither the .txt or .DAT log ... until everything just abrupt shuts off.

The log have only recorded short cutoffs in the beginning & just past half flight ... all just glitches. If WiFi pollution were to blame alone it should have been more of these vertical pink bands in the chart, especially near to the end.

 

[m80116]

Pilot error
You flew into 22 mph (9.8m/s) winds.
Wind direction was towards you.

There was almost no wind at ground level and I believe very mild wind at flight altitude. I have footage and data showing the drone didn't have to fight to stay put in the air.

Those wind historic forecast models are highly unreliable for micro-environments and topographical level heights. I might upload a video frame from the video showing the chimneys above the roofs emitting a perfectly straight plume of smoke. Still picture extract is here

I know it's not pleasant when data don't match with reality.

 

[sar104]

There was almost no wind at ground level and I believe very mild wind at flight altitude. I have footage and data showing the drone didn't have to fight to stay put in the air.

Those wind historic forecast models are highly unreliable for micro-environments and topographical level heights. I might upload a video frame from the video showing the chimneys above the roofs emitting a perfectly straight plume of smoke.

I know it's not pleasant when data don't match with reality.

The wind speed that @pmshop was quoting was calculated by the AirData website based on aircraft tilt and ground speed - they are not from forecast models. However, as I posted above, they were incorrect for this flight. The wind was calm, as shown by the < 1° tilt of the aircraft when holding position.

 

[m80116]

This is highly speculative ... but can't find reasons that make it unlikely.

Thanks for your time and interesting analysis.

I just want clarify the buldge in the path is entirely due to my control input stick movement (yaw) which is also shown in the video of the flight path, the DJY Fly app makes it easy to notice with control sticks graphics overlay. I can give you the link to that vid.

Also I had a notice of Aircraft disconnected, and the remote was flashing all of its LEDs.

 

[m80116]

In theory also the same could have happend as happend to me few days ago. [...]

Do you have an idea when the last firmware was released ? My fly-away case happened the 26th of Dec. and it was up-to-date to that moment.

 

[m80116]

WiFi problems were a constant problem with the Phantom 3 4K, Mavic Air, and Phantom 4 adv. Frequently the effects were present starting at 600 ft out. Sometimes you could clear them by climbing to 400 ft, mostly not. Shifting flight direction sometimes helped, sometimes not.

The problem was easily seen using a laptop with a software program which measured the WiFi signal environment.

One of the judgement errors I've made is that I've scanned both 2.4 GHz and 5 GHz bands with my mobile phone (I have a phone dedicated to data hotspot, Wi-Fi scans and other tasks) assuming the 5 GHz band would be the same as the Wi-Fi standard because I've heard it said several times in the videos so I just assumed... and I assumed it wrong.

I have no idea what there was on the 5.8 GHz spectrum. This I admit is a novice error.

If say the 5.8 GHz was full, considering the 2.4 GHz might have been completely saturated when sampling all the lengths and heights this still doesn't explain why I haven't got notices of interference, weak connection, adjusting antennas or else...

 

[m80116]

In urban area you can lose the WLAN signal very suddenly and unexpectedly even after couple of meters.

I start to believe probably my mistake was to trust the signal strength indicator which stayed full and no warning notices for several minutes (just two and very short mid-flight for the whole record).

 

[sar104]

I start to believe probably my mistake was to trust the signal strength indicator which stayed full and no warning notices for several minutes (just two and very short mid-flight for the whole record).

Perhaps, but that doesn't explain why it didn't RTH.