Crash - Pilot error or drone issue?

[theburninator]

Hey guys,

I had a fly-away crash last month and DJI is so far failing to take any responsibility at all. I am a new pilot, so maybe there's something I don't understand, but there are 2 major issues I haven't been able to reconcile nor have they been able to answer. Here is the scenario:

I flew the drone straight up, winds were roughly 5mph according to local weather. Upon reaching altitude, I got a wind warning and noticed that the drone was drifting at roughly 1.2 m/s. Not being very experienced yet, I hit the RTH button as I felt I was in over my head and didn't want to risk anything. Upon hitting the button, the drone accelerated to 4.2 to 5.2 m/s in the opposite direction of the home point (which was set correctly before take-off). After the drone continued to move away from me, I hit the control stick to descend thinking that the wind was causing an issue. That brought the drone down to the RTH height, but then the drone stopped and began to descend on its own and nowhere near the home point. I pressed up on the control stick to stop the descent and it would not respond. The drone ended up crashing about 80 ft. high in the top of a tree and was not retrievable.

I opened a ticket with DJI and their response is that the wind caused the flyaway and there was no fault with the drone. I have asked multiple times about why the drone accelerated in the wrong direction upon RTH activation and why it landed without input when there was still plenty of battery left (70%). I have yet to receive an answer for either of these questions and continue to get the response that the wind is responsible for everything.

I would just like some input on whether the drone did malfunction or if there is an explanation I am just not understanding. I understand that I am a new pilot and that this could easily be 100% my fault or just a lack of understanding on my part. I honestly just don't know for sure and would love to have a more definitive answer. I uploaded the flight log, hopefully it has the correct information. Please let me know if anything more is needed and try to be easy on me, I just want to learn from this either way

Thanks in advance for you help!

 

[Prismatic]

View media item 3206Check the Flight Log Retrieval and Analysis Guide by @sar104 to make sure you've collected all the available data for analysis. Best of luck!

 

[RodPad]

@theburninator,

Welcome to the Forum!

DJI Flight Log Viewer | Phantom Help

Upload and view flight logs from your DJI drones

www.phantomhelp.com

Rod

 

[steve495]

I'm new at this, but uploaded your text file to Airdata. In my "new guy" opinion, DJI was correct. The flight was in winds beyond the published limit of 8 m/s, which is about 18 MPH. As pilots learn - quickly with a small drone like the Mini - the winds aloft can be significantly different than at the surface.

It looks like your home point recorded at the beginning of the flight was wrong due to GPS signal strength. Its recorded point is way over on the east side of N Leavitt Rd NW behind Shafer Industrial Services. There was no "Home Point Updated" in the log, so I'm not even sure the Mini new where it was, let alone where its home point was. Any GPS device needs some time to figure out where the heck it is. It's not instantaneous. (Actually, not sure if the Home Point Updated is supposed to display in the log I'm looking at, but there is still a big concern due to the star point shown far to your southwest.)

Although the winds at the surface seemed to be about 5 MPH out of the north, the average in-flight wind speed seems to be between 20 and 27 MPH. This is calculated by determining how much the drone needs to tilt to stay in place. These numbers in the screen shot are probably about 30% too high as compared to what it really was. Still, it was beyond the limit of the drone.



You pretty much took off, ascended to 115 feet in less than 40 seconds, and continued up to 390 feet or so within another 60 seconds. The "not enough force" warnings - related to the winds aloft - started to display 22 seconds into the flight at about 50 feet. You pretty much went straight up into more and more wind. Those "not enough force" warnings continued through your flight.

You hit the RTH button at about 2:55 into the flight. That was the wrong move and a rookie mistake. With the high wind, your first priority is to descend as much as possible and hope for lighter winds. Then, switch into Sport mode to drive home as it will allow for a greater tilt angle and more speed to work against the wind.

You canceled RTH at 3:52 into the flight, but stayed in P mode and finally did start descending on its own a bit as it was struggling trying to push into the wind. At 4 minutes, you jammed the throttle up for more altitude, then back down to descend.

Assuming your remote was in the default Mode 2, you had occasional inputs with the left control stick increasing and decreasing power, but you rarely touched the right controller to get forward drive to come home. Almost never that I can see. She just started drifting into the wind, 2 minutes into the flight at almost 400 feet.

Sorry, but this clearly looks to be pilot error due to not recognizing winds aloft.

Edited to add. I do find it interesting that when you hit RTH, the drone did turn towards home, but the speed did increase from about 3 to 9 MPH with the drift in the wind. It does not look like the wind increased, but it was still well beyond the published 8 m/s limit.

Of you're able to upload your DAT file, I think the experts may be able to tell you what the tilt was and what the motors were doing.

 

[steve495]

Again, sorry about the loss. It sucks. As a courtesy, here is the Airdata information you can scroll through. Note that I will end up deleting this at some point in the near future. The winds aloft calculation is not included in the free plan, but I have it...

Go to General > Notifications to see more details on the flight.

 

[Meta4]

I had a fly-away crash last month and DJI is so far failing to take any responsibility at all. I am a new pilot, so maybe there's something I don't understand, but there are 2 major issues I haven't been able to reconcile nor have they been able to answer.

You have managed to give a clear and accurate description of the incident, which is uncommon and makes it easier to analyse the data.

You climbed and the drone started to drift when you passed 244 ft with the drift increasing as the drone went higher.
At 2:14, hovering at 390 ft the drone was being blown off course at up to 9 mph by a strong wind from the south west.
At this stage the drone is 300 ft from home, but that's increasing while the drone is left up in the wind.
After lots more hovering up there, the drone is now 600 ft away at 2:55.1 when you initiated RTH.
It turns to face home and starts the RTH but because the wind is faster than the drone's RTH speed, the drone continues to be blown backwards at 9-10 mph.
At 3:52.8 (now 1370 ft away) you cancel RTH and bring the drone down to about 270 ft where the wind is less but still blowing the drone further away.

A separate issue becomes noticeable around 3:54 when the drone started to sink without you mushing the left stick down.
This is a problem that is being seen in a number of flight incident logs. (uncommanded descent)

At 4:15.7 (alt 233 ft) you push the left stick full forward to counter the uncommanded descent but it makes no difference and the drone continues descending at about 7 ft/sec.

All this time you've done nothing to bring the drone back and it's now >1800 ft away when the data ends at 4:35.6 with the drone still sinking despite the left stick being pushed full up.

It looks like signal was lost shortly before the crash in a tree which does not show in the data.

To answer your questions, pilot error or drone issue?
The answer is both.

Pilot Error
You pushed the drone up where the wind was too strong for the drone to fight against and left it there being blown further away.
Wind at height is always stronger than on the ground.
Go higher and the winds get even stronger.
In a strong wind situation like this, it is important to get the drone down out of the wind that is causing the trouble.
Leaving it up is only letting the problem get worse.
If you had brought the drone down to 100 ft or so, it would have been easy to bring it home.
Not noticing the drift when it sbecame visible at 244 ft is another mistake.

DJI Issue
The uncommanded descent is a big problem for the Mini that DJI don't seem to be doing anything about.
I'm seeing it a few times each week and when it starts to happen, it doesn't seem that anything fixes it.
I would make a point of mentioning this to DJI because they don't seem to even be aware of it.

 

[MAvic_South_Oz]

Mavic landing battery 70%, is it assumed then that it 'landed' at the NW end of the yellow flight line on the Airdata map ?
You saw it in the tree ?

If there is a possibility it's landed differently for some reason, flown further after loss of signal, maybe @sar104 can give an opinion on location ?
sar has a great track record for helping drone pilots find their aircraft !

 

[steve495]

Could the uncommanded descent - at least in this instance - be related to the motors loosing thrust due to the constant "not enough force" reports throughout the flight? Is it just getting tired, or does that not matter?

Also, did you note what I saw in the logs? Looks like there was no inputs to the right stick to fly the drone into the wind (back home).

 

[Meta4]

If there is a possibility it's landed differently for some reason, flown further after loss of signal, maybe @sar104 can give an opinion on location ?
sar has a great track record for helping drone pilots find their aircraft !

No need for high tech.
At end of data the drone was only 83 ft above launch height and falling fast.
The ground there was about 10 ft lower than the launch point so it can't have gone far.

Could the uncommanded descent - at least in this instance - be related to the motors loosing thrust due to the constant "not enough force" reports throughout the flight? Is it just getting tired, or does that not matter?

I believe that those not enough force messages (that only show in Airdata) are just advising that the drone is sinking and can't maintain height.
It has nothing to do with getting "tired", the drone should be able to maintain height until critical low battery initiates autolanding.
In all of these uncommanded descents, the drone keeps sinking with a good battery and full left stick up.

Also, did you note what I saw in the logs? Looks like there was no inputs to the right stick to fly the drone into the wind (back home).

Yes, I saw that the drone was left hovering (being blown away) for much of the flight.

 

[theburninator]

Thank you for all the great information. It's obviously not what I was hoping for, but if it was my screw-up, then so be it! I clearly didn't have enough understanding of higher altitude winds as well as the best course of action to take when things start to go south. I'm still learning and this was a tough lesson!

The GPS home point information is really confusing for me though. When I look at the log in the DJI Fly app, it shows the home point is set at 0:16 with 11 satellites visible and it is in the correct location. The map you posted shows it in an entirely different area. Why would the DJI app show the HP in the correct location (where I actually was) if the underlying data shows something else? I'm attaching a screenshot of that time from my log if that helps.

I'm new at this, but uploaded your text file to Airdata. In my "new guy" opinion, DJI was correct. The flight was in winds beyond the published limit of 8 m/s, which is about 18 MPH. As pilots learn - quickly with a small drone like the Mini - the winds aloft can be significantly different than at the surface.

It looks like your home point recorded at the beginning of the flight was wrong due to GPS signal strength. Its recorded point is way over on the east side of N Leavitt Rd NW behind Shafer Industrial Services. There was no "Home Point Updated" in the log, so I'm not even sure the Mini new where it was, let alone where its home point was. Any GPS device needs some time to figure out where the heck it is. It's not instantaneous. (Actually, not sure if the Home Point Updated is supposed to display in the log I'm looking at, but there is still a big concern due to the star point shown far to your southwest.)

Although the winds at the surface seemed to be about 5 MPH out of the north, the average in-flight wind speed seems to be between 20 and 27 MPH. This is calculated by determining how much the drone needs to tilt to stay in place. These numbers in the screen shot are probably about 30% too high as compared to what it really was. Still, it was beyond the limit of the drone.

View attachment 92652

You pretty much took off, ascended to 115 feet in less than 40 seconds, and continued up to 390 feet or so within another 60 seconds. The "not enough force" warnings - related to the winds aloft - started to display 22 seconds into the flight at about 50 feet. You pretty much went straight up into more and more wind. Those "not enough force" warnings continued through your flight.

You hit the RTH button at about 2:55 into the flight. That was the wrong move and a rookie mistake. With the high wind, your first priority is to descend as much as possible and hope for lighter winds. Then, switch into Sport mode to drive home as it will allow for a greater tilt angle and more speed to work against the wind.

You canceled RTH at 3:52 into the flight, but stayed in P mode and finally did start descending on its own a bit as it was struggling trying to push into the wind. At 4 minutes, you jammed the throttle up for more altitude, then back down to descend.

Assuming your remote was in the default Mode 2, you had occasional inputs with the left control stick increasing and decreasing power, but you rarely touched the right controller to get forward drive to come home. Almost never that I can see. She just started drifting into the wind, 2 minutes into the flight at almost 400 feet.

Sorry, but this clearly looks to be pilot error due to not recognizing winds aloft.

Edited to add. I do find it interesting that when you hit RTH, the drone did turn towards home, but the speed did increase from about 3 to 9 MPH with the drift in the wind. It does not look like the wind increased, but it was still well beyond the published 8 m/s limit.

Of you're able to upload your DAT file, I think the experts may be able to tell you what the tilt was and what the motors were doing.

 

[theburninator]

Thanks for this information as well! I see now that I made a number of critical errors. I did not push the drone back home because I was unsure of the orientation of the drone (still my fault obviously). At the end, I was trying to get it to a reasonable height and then worry about flying it home, but it had that uncommanded descent. Rookie mistakes all over the place, but I am trying to learn from this and you guys are helping out a lot!

I realized from the beginning that I have blame in this, but I really thought the descent at the end could be a drone issue and so far DJI hasn't even tried to explain what happened there. I guess that's the only avenue I have left to pursue (although it doesn't look like I'm going to get anywhere). I don't even expect them to send me a new drone, It would just be nice for them to at least meet partway or something...probably just wishful thinking though

You have managed to give a clear and accurate description of the incident, which is uncommon and makes it easier to analyse the data.

You climbed and the drone started to drift when you passed 244 ft with the drift increasing as the drone went higher.
At 2:14, hovering at 390 ft the drone was being blown off course at up to 9 mph by a strong wind from the south west.
At this stage the drone is 300 ft from home, but that's increasing while the drone is left up in the wind.
After lots more hovering up there, the drone is now 600 ft away at 2:55.1 when you initiated RTH.
It turns to face home and starts the RTH but because the wind is faster than the drone's RTH speed, the drone continues to be blown backwards at 9-10 mph.
At 3:52.8 (now 1370 ft away) you cancel RTH and bring the drone down to about 270 ft where the wind is less but still blowing the drone further away.

A separate issue becomes noticeable around 3:54 when the drone started to sink without you mushing the left stick down.
This is a problem that is being seen in a number of flight incident logs. (uncommanded descent)

At 4:15.7 (alt 233 ft) you push the left stick full forward to counter the uncommanded descent but it makes no difference and the drone continues descending at about 7 ft/sec.

All this time you've done nothing to bring the drone back and it's now >1800 ft away when the data ends at 4:35.6 with the drone still sinking despite the left stick being pushed full up.

It looks like signal was lost shortly before the crash in a tree which does not show in the data.

To answer your questions, pilot error or drone issue?
The answer is both.

Pilot Error
You pushed the drone up where the wind was too strong for the drone to fight against and left it there being blown further away.
Wind at height is always stronger than on the ground.
Go higher and the winds get even stronger.
In a strong wind situation like this, it is important to get the drone down out of the wind that is causing the trouble.
Leaving it up is only letting the problem get worse.
If you had brought the drone down to 100 ft or so, it would have been easy to bring it home.
Not noticing the drift when it sbecame visible at 244 ft is another mistake.

DJI Issue
The uncommanded descent is a big problem for the Mini that DJI don't seem to be doing anything about.
I'm seeing it a few times each week and when it starts to happen, it doesn't seem that anything fixes it.
I would make a point of mentioning this to DJI because they don't seem to even be aware of it.

 

[hashfly]

400 feet up and almost a mile out with 30-mph gusts, if you remove these pilot errors from the equation then you might have been able to save it from the un-commanded descent. The Minis operating limits are well documented and unfortunately you pushed the bird well beyond it's limits and your own

the landing happened well out of VLOS which again is a mistake.

Tough lesson, are you sure you can't pay someone to climb the tree?

 

[theburninator]

I was able to locate the drone, just unable to retrieve it. It was something like 80 feet up in a tree that was not really thick enough to climb and in the middle group of trees to further restrict access. It did eventually fall down and I actually just received it today, but the gimbal mount is broken and one of the ribbon cables in torn in half. I'm not sure if it's worth trying to repair or if I'm better off just buying a bare drone and having the other one for spare parts/etc.

Mavic landing battery 70%, is it assumed then that it 'landed' at the NW end of the yellow flight line on the Airdata map ?
You saw it in the tree ?

If there is a possibility it's landed differently for some reason, flown further after loss of signal, maybe @sar104 can give an opinion on location ?
sar has a great track record for helping drone pilots find their aircraft !

 

[theburninator]

I'm still not sure what's going on with the whole GPS thing. I clearly messed up with the winds. It may not be a big difference, but the drone landed about 2,000 ft away from where it took off in reality and according to the DJI app. I'm still trying to figure out how the GPS log data differs so much from what it shows on the app when I view the flight there :-/

400 feet up and almost a mile out with 30-mph gusts, if you remove these pilot errors from the equation then you might have been able to save it from the un-commanded descent. The Minis operating limits are well documented and unfortunately you pushed the bird well beyond it's limits and your own

the landing happened well out of VLOS which again is a mistake.

Tough lesson, are you sure you can't pay someone to climb the tree?

 

[Meta4]

I'm still trying to figure out how the GPS log data differs so much from what it shows on the app when I view the flight there :-/

The GPS issue is not a great mystery
You must have launched from a spot where the drone did not have a clear sky view, perhaps houses and trees were blocking a chunk of its horizon?
That meant that it had trouble getting enough sats at first.
It acquired enough sats for a basic fix at 10.7 seconds while the drone was still on the ground.
However the flight controller recognised that this fix was not reliable and gave it GPS Health = 0
This is the first GPS point on the map below.
As you can see it is quite inaccurate - the controller was right to give it a score of zero.

You launched vertically and at 14.1 secs and 8 feet higher, the Mini was able to see more sats and get a reliable fix and this is where it recorded the home point.
The first point is in the data but was not recorded as a home point.

I would still push DJI about the uncommanded descent.
It was a big factor in the incident (even if not the major factor).
Don't let them bully you with your mistake on this one.
They should be grateful to have a serious issue pointed out to them with good data to support it.
It has caused the loss of more than a few Minis and they need to be made aware of it and do something to fix it.
I've never seen anything similar in 6 yrs of flying other DJI craft.

Copy and paste any of my assessment that is useful - just leave out all the mention of wind at high level and blowing away.
Looking at the data, it looks like the sinking started earlier, around 3:0.7 but became worse later at the time I pointed out earlier.

 

[sar104]

As @Meta4 points out, the unusual track is just a result of taking off without a position lock. There was no position solution until an inaccurate one at 10.7 seconds - hence the incorrectly recorded home point. At 11 seconds an accurate solution was achieved, but that didn't reset the original incorrect home point.





The initial difficulty flying the aircraft has already been explained - the winds were close to 30 mph out of the SW - far too high for the Mini to handle.

You actually got lucky that the second issue arose, however. Without the uncommanded descent the aircraft would have flown much further away. The uncommanded descent looks just like a number of other Mavic Mini incidents, and there is clearly a firmware or hardware problem with this model. Looking at the battery and flight status:



The aircraft is being blown away in RTH mode while applying only 10° of pitch to try to fight the wind (should be 20°), until RTH is cancelled at 233 seconds. That results in a large pitch excursion followed by pitch and roll instability and the uncommanded descent:



That resulted in it landing in the tree behind 308 Aspen Drive NW, where you found it. This was pilot error to launch in windy conditions.

This is, however, just the latest in a string of such uncommanded descents by Mavic Minis, all with the same tell-tale signature of a pitch excursion followed by attitude instability and loss of altitude. It's hard to avoid the conclusion that there is a significant firmware bug or hardware design issue, although I don't expect DJI to acknowledge that at this stage.

I'm in the process of compiling those events into a single post to illustrate the issue.

 

[Thomas B]

One last recommendation @theburninator .... consider DJIRefresh on the next drone.., another or the replacement DJI will send if they “repair”. It is painless, much less cost than the drone or repair, and doesn’t matter why a crash occurred to get a new or refurb drone.
Just something to consider.

 

[Bekanne]

I'm new at this, but uploaded your text file to Airdata. In my "new guy" opinion, DJI was correct. The flight was in winds beyond the published limit of 8 m/s, which is about 18 MPH. As pilots learn - quickly with a small drone like the Mini - the winds aloft can be significantly different than at the surface.

It looks like your home point recorded at the beginning of the flight was wrong due to GPS signal strength. Its recorded point is way over on the east side of N Leavitt Rd NW behind Shafer Industrial Services. There was no "Home Point Updated" in the log, so I'm not even sure the Mini new where it was, let alone where its home point was. Any GPS device needs some time to figure out where the heck it is. It's not instantaneous. (Actually, not sure if the Home Point Updated is supposed to display in the log I'm looking at, but there is still a big concern due to the star point shown far to your southwest.)

Although the winds at the surface seemed to be about 5 MPH out of the north, the average in-flight wind speed seems to be between 20 and 27 MPH. This is calculated by determining how much the drone needs to tilt to stay in place. These numbers in the screen shot are probably about 30% too high as compared to what it really was. Still, it was beyond the limit of the drone.

View attachment 92652

You pretty much took off, ascended to 115 feet in less than 40 seconds, and continued up to 390 feet or so within another 60 seconds. The "not enough force" warnings - related to the winds aloft - started to display 22 seconds into the flight at about 50 feet. You pretty much went straight up into more and more wind. Those "not enough force" warnings continued through your flight.

You hit the RTH button at about 2:55 into the flight. That was the wrong move and a rookie mistake. With the high wind, your first priority is to descend as much as possible and hope for lighter winds. Then, switch into Sport mode to drive home as it will allow for a greater tilt angle and more speed to work against the wind.

You canceled RTH at 3:52 into the flight, but stayed in P mode and finally did start descending on its own a bit as it was struggling trying to push into the wind. At 4 minutes, you jammed the throttle up for more altitude, then back down to descend.

Assuming your remote was in the default Mode 2, you had occasional inputs with the left control stick increasing and decreasing power, but you rarely touched the right controller to get forward drive to come home. Almost never that I can see. She just started drifting into the wind, 2 minutes into the flight at almost 400 feet.

Sorry, but this clearly looks to be pilot error due to not recognizing winds aloft.

Edited to add. I do find it interesting that when you hit RTH, the drone did turn towards home, but the speed did increase from about 3 to 9 MPH with the drift in the wind. It does not look like the wind increased, but it was still well beyond the published 8 m/s limit.

Of you're able to upload your DAT file, I think the experts may be able to tell you what the tilt was and what the motors were doing.

Steve 495 you did a wonderful analysis and explanation.

 

[RayOZ]

How many of these uncommanded descend is preceded by "Not Enough Force/ESC Error"? Would the motors be overheating? What's the current draw on the motors?

 

[Meta4]

How many of these uncommanded descend is preceded by "Not Enough Force/ESC Error"?

Probably all of them, but the message only shows in Airdata ... not in the app or flight data.

Would the motors be overheating?

No .. The problem is that the drone doesn't tilt enough.
It ignores throttle inputs to climb.