Inexplicable drop - Looking for answers before doubling down!

[raukar]

Hi Everyone!
(First off, as this is my first post, let me tell you guys how excited i am to get to start being a part of a community that seems this passionate and helpful about things. Been reading a lot here before my actual purchase of a drone).

Back to my original reason for the post: After buying my Mavic i used it a couple of times learning the different features. I've flown lots of both toy drones and RC planes in the past, so there was no steep learning curve for the actual handling. When it comes to specifics about the Mavic in particular, i felt the booklets included in the pack were a bit lackluster and in general quite uninformative about the product. So i researched as much as possible on my own before setting out to feel 100% ready.

This is why i was so surprised, at 1h36 minutes total flight time, and my 11th flight (in 5 session), about the catastrophic failure i experienced.

As i was hovering and taking shots of the family at the beach, i pull up and backwards to move away from the object, which worked perfectly until the moment at about 70m height (still in movement) when the drone just cuts out completely, disconnects from the controller and falls straight down into the water. Obviously it sank right away, without recovery possible. Complete shock of course so i can't be 100% sure but i'm pretty confident the engines all started again halfway down for a second and then cut out again to fall the rest of the way hitting the water with a splash.

Spoke to DJI support about the issue, but they only said they couldn't do much without the data from the drone itself and offered me a 20% code to buy a new one. They also promised that they'd analyse the log files from the controller (which they claim they did) but i can see that my WeTransfer containing said logs was ever downloaded at all, so they can't possibly have done the checks they claim they made.

This brings me to my plea here on the forum: In total there were 7 files: "fatal.log"; "dlm_0"; "dlm_1"; "dlm_2"; "dlm_3"; "dlm_4"; and "cp_assert.log" on my controller. And as i said they seem to never have been downloaded and checked by DJI support. I totally understand the guarantee cause is probably lost, but out of curiosity i'd like to know if anyone here can find anything else out of these than what i already know. From everything i've read, the only user misstake that could have happened is a badly attached battery but i highly doubt i missed such a thing. It also seems unlikely the battery would detach with the "downforce" pushing it in place when the drone was going up and backwards.

If i get a new one, i'd love to start back up again with an understanding of what actually went wrong. Not understanding it at all just bugs me too much.

Again, any insight is greatly appreciated! Here is the flight data from the controller: cp_assert.log and 6 more files

 

[JeepDSL]

What mobile device were you using?

The files you posted are not flight logs. There should be some .TXT files on your mobile device. If you make those .TXT log files available (check here for instructions DJI Flight Log Viewer) other forum members can help you out with some analysis.

BTW, I believe your flight logs are sync'ed to DJI servers, so they probably don't need to download your files.

 

[raukar]

Hi again!
I was using an iPhone 7. Found and exported the .TXT file and attaching it here.
Does that mean that there is no point or information to be gotten from the "black box" of the controller?

Thank you for your help.

 

[james_mav]

Hi Everyone!
(First off, as this is my first post, let me tell you guys how excited i am to get to start being a part of a community that seems this passionate and helpful about things. Been reading a lot here before my actual purchase of a drone).

Back to my original reason for the post: After buying my Mavic i used it a couple of times learning the different features. I've flown lots of both toy drones and RC planes in the past, so there was no steep learning curve for the actual handling. When it comes to specifics about the Mavic in particular, i felt the booklets included in the pack were a bit lackluster and in general quite uninformative about the product. So i researched as much as possible on my own before setting out to feel 100% ready.

This is why i was so surprised, at 1h36 minutes total flight time, and my 11th flight (in 5 session), about the catastrophic failure i experienced.

As i was hovering and taking shots of the family at the beach, i pull up and backwards to move away from the object, which worked perfectly until the moment at about 70m height (still in movement) when the drone just cuts out completely, disconnects from the controller and falls straight down into the water. Obviously it sank right away, without recovery possible. Complete shock of course so i can't be 100% sure but i'm pretty confident the engines all started again halfway down for a second and then cut out again to fall the rest of the way hitting the water with a splash.

Spoke to DJI support about the issue, but they only said they couldn't do much without the data from the drone itself and offered me a 20% code to buy a new one. They also promised that they'd analyse the log files from the controller (which they claim they did) but i can see that my WeTransfer containing said logs was ever downloaded at all, so they can't possibly have done the checks they claim they made.

This brings me to my plea here on the forum: In total there were 7 files: "fatal.log"; "dlm_0"; "dlm_1"; "dlm_2"; "dlm_3"; "dlm_4"; and "cp_assert.log" on my controller. And as i said they seem to never have been downloaded and checked by DJI support. I totally understand the guarantee cause is probably lost, but out of curiosity i'd like to know if anyone here can find anything else out of these than what i already know. From everything i've read, the only user misstake that could have happened is a badly attached battery but i highly doubt i missed such a thing. It also seems unlikely the battery would detach with the "downforce" pushing it in place when the drone was going up and backwards.

If i get a new one, i'd love to start back up again with an understanding of what actually went wrong. Not understanding it at all just bugs me too much.

Again, any insight is greatly appreciated! Here is the flight data from the controller: cp_assert.log and 6 more files

Did you do any takeoffs from the beach where sand may have gotten in between the battery & contacts?

 

[JeepDSL]

@msinger is very good with flight log analysis.

Your drone was at 72m altitude and 152m away. Data shows that the drone was moving up and away from you at the time of last entry. I wonder if you could have accidentally initiated a motor shutdown.

 

[Jacknet007]

I he was moving up and backwards the left stick would not have been close to the shutdown point although the right stick could have been. What was the battery percentage at the time? Also flight logs could show the voltages on the 3 battery cells to see if they were unusually low. Did you start the flight on a full charge?

 

[erkme73]

Posting to follow this. I'm hoping that all the resident experts can help you come to a better understanding of what happened with the data in the text file.

Using msinger's phantomhelp website, I uploaded your text file and converted it to kml to view it in Google Earth. The last leg looks rock solid. If it was a CSC shutdown, I don't know how that could have happened (both sticks down and inward) without recording that input on the flight path.

 

[JeepDSL]

CSC is down and inward or down and outward. Since the log shows the drone climbing and moving away, I guess it couldn't have been CSC shutdown assuming default stick settings.

The battery telemetry looks fine right up to the last transmission - 100% battery at start, 96% at end, all three cells performing perfectly. Could have been an instantaneous power loss due to a battery that wasn't fully seated. I don't know; this needs a better analyst.

@raukar , have you replayed the flight in the DJI GO 4 app to verify your stick movements? Did you cache video on your iphone? Does anything show up there?

 

[Budge]

So I had a similar event happen today, except I was doing the exact opposite. I was flying forward and down steadily for about 6 seconds before the props just stopped spinning and the Mavic dropped out of the sky. 100% battery at start, no warnings on the remote or on my phone.

 

[TheLightSpeed!]

So I had a similar event happen today, except I was doing the exact opposite. I was flying forward and down steadily for about 6 seconds before the props just stopped spinning and the Mavic dropped out of the sky. 100% battery at start, no warnings on the remote or on my phone.

I've read several instances here of sudden power loss as a result of battery coming unseated during flight.

I now always put the battery in, and then take one finger and push down until I here both "click" in. Otherwise you may never know.

I would say a CSC is highly unlikely based on the flight path.

 

[TheLightSpeed!]

So I had a similar event happen today, except I was doing the exact opposite. I was flying forward and down steadily for about 6 seconds before the props just stopped spinning and the Mavic dropped out of the sky. 100% battery at start, no warnings on the remote or on my phone.

Budge did your video screen go black? Like the power shut off?

 

[Budge]

Budge did your video screen go black? Like the power shut off?

I honestly don't know. The Mavic was flying over my head and going past me when I heard my friend gasp. I looked at him and saw the Mavic just fall 10 feet away. The battery popped out. I posted my txt and dat files on another thread. Wierd Crash...what happened and what to do?

 

[raukar]

Did you do any takeoffs from the beach where sand may have gotten in between the battery & contacts?

We were in a scandinavian archipelago, so the “beach” actually was just massive flat rocks The drone was just taken out of its bag before operation and put on a flat surface. As mentioned previously in the thread, i didn’t go for the “killing maneuver” right away but flew around a bit first at first to “test that everything was working” (Ended up being a false security but what can you do).

@raukar , have you replayed the flight in the DJI GO 4 app to verify your stick movements? Did you cache video on your iphone? Does anything show up there?

Doublechecked on the stick movement and as erkme73 said the sticks really didn’t move from their up and back positions att all, and then complete cutout.


What really worries me for potential future flights after an experience like this is safety. Now i wouldn’t have flown over large groups of people or done anything crazy anyway, but not understanding what went wrong puts the hardware in question for me. I honestly would prefer to know i DID mess up somehow, so that i know i can avoid it in the future.
It seems from other accounts that the battery not being properly attached is a likely cause, but it seems so unlikely as well as i took all precautions having read up here beforehand (And hadn't really used it enough to get too cocky either).

Attaching the “Final Moments” video from the iPhone cash. As you can see, it just cuts out without interference or anything. Just ends.

 

[msinger]

I wonder if you could have accidentally initiated a motor shutdown.

That was not done per the flight log.

 

[Jacknet007]

Another person flying a Mavic or other drone in the area?

 

[BuzzyOne]

Well, there obviously was a hard cut off to the power of the MP. That is a very rare occurrence that could be caused by only a few things. The most probable would be a battery connection issue. I was just playing with my MP battery and it certainly can be latched incorrectly and not noticed. In fact, the clips act a a sort of spring and if improperly seated they will force the battery to pop out a bit if the battery is backed out slowly. It doesn't take much for the MP to lose power, the battery doesn't have to fall all the way out of the MP to kill it.

It is also easy to unseat the clips just by picking up and handling the MP. The little buttons on the side are right where one would grab the thing. If one tab is unseated, it is possible to pull the battery out. Just play around with it and you will see that it doesn't take much and an improperly seated battery is unnoticeable just looking at the MP.

I would vote battery disconnect.

 

[JeepDSL]

I would vote battery disconnect.

Based on the comments and evidence, I agree.

 

[erkme73]

If that turns out to be the case, why doesn't this qualify as a design flaw? I mean, you would think that installing a battery should be binary... Either it's in, or it's not. I don't recall anyone complaining about this for the first 6 months of the product release. I've certainly not been able to provide a sometimes-working battery install with my 1yo MP. Did something change that now not only makes it possible, but even likely that someone can inadvertently provide shaky power? If so, how does this not qualify as a safety issues which triggers a safety recall like the Karma?

Seems to me there may be some DJI blame-shifting go on here... And not just to avoid individual warranty liability. If they cop to enough of these, one could make a reasonable argument that they have a systemic problem.

 

[Jacknet007]

Battery disconnect sound plausible except force on the drone with the direction he was moving would not have likely to unseat a loose battery unless he let go of the throttle just at that moment.

 

[raukar]

Battery disconnect sound plausible except force on the drone with the direction he was moving would not have likely to unseat a loose battery unless he let go of the throttle just at that moment.

Yes the direction and angle the MP was going in makes it unlikely a "jerk" reaction could have displaced the battery (In that case it actually should have happened when flying around in different directions earlier in the flight). I can't recall doing anything strange myself or the drone reacting in any way (slowing/stopping) before dropping. As someone mentioned it was at about 150m away so hard to tell for sure but it seemed like the whole unit dropped as one without separating mid-air.

Either way, it seems some kind of power failure must have been the cause for the crash.

I am 100% sure the battery wasn't visibly attached the wrong way though, which i guess means there is a margin of error of even 1mm or less that can cause a complete cutout? (i.e. not visible to the eye on inspection?).

This makes me super nervous as it would mean you can't know for sure if you truly are good to go or not, even if you add pressing on the battery as a pre-flight check. The sound of the MP hitting the water from 70m height was pretty scary. Imagine hitting a roof, or anything (anyone) that would potentially have passed by. Messing up or acting like an idiot is one thing, but having an issue that seems to happen this easily (and from looking at the forum often) makes me ill at ease. After all we don't fly in the vacuum of space, so you want to be confident you've done everything right to avoid accidents.

Does anyone have any examples of image/video of how you could notice if the battery isn't correctly installed? Not counting the case of it not being properly clamped in of course

(As a side note: This whole thing is really surprising to me. When i upgraded to the Mavic from the Bebop 2 i was impressed in how solid and well made both controller and unit felt, like upgrading from a Kia to a BMW. Even though plastic-y, i've never felt unsafe with the Bebop 2 though, and i flew that thing A LOT. Too bad the image quality just isn't there.)