Hi everyone,


The other day I was flying my new DJI Mini Pro 3 (updated to last firmware). All was going well but at a certain moment I was suddenly unable to control the drone for a few seconds. The drone started to fall down slowly (by that I mean it didn't fall like a stone):



When it happened I could still see the image on the remote control, had good connection/signal and 59% left of battery. The drone didn't RTH (it was set properly), but about 3-4 seconds later I gained controls back and was able to fly it to safety manually. The weather was hot (32-35°C) but during the whole session I didn't have any overheat notifications. Later I was able to fly normally, however I'm wondering if the drone or battery has an issue. Did I miss something or do something wrong?

Here are more details:
https://www.phantomhelp.com/LogViewer/6K5MJ9RRISDVIZBH3O09/
(The issue happens at 9 min 28)



I noticed several things while looking at the logs:
- At first, a bit before the incident, the cell deviation dropped to 0.11V during 0.8 sec (not sure if that's abnormal)
- Data shows that it's like if I had pull the joysticks down (which I didn't)








Thank in advance!


Sakat


Edit: Added flight log.

Hi Sakat,

best to get help/advice is to share your flightlog fot this issue.

use > https://www.phantomhelp.com/LogViewer/Upload/
Post the uploaded link on here



cheers
JJB

JJB* Posted at 7-1 00:26
Hi Sakat,

best to get help/advice is to share your flightlog fot this issue.

Hi JJB,

I didn't realize that I could simply share the URL of the log:
https://www.phantomhelp.com/LogViewer/6K5MJ9RRISDVIZBH3O09/

Thanks for the tip!

Sakat

Sakat Posted at 7-1 00:28
Hi JJB,

I didn't realize that I could simply share the URL of the log:

Hi,

Had a look at your log.
No issues in the log.

Battery cell values and deviation all oke, even high cell deviation does not make a drone loose height.

It looks that you, while not aware of this, momentarily pushed your drone down.
FRAP algortihm does check for uncontrolled going down or up, not in your log.

See my charts of your data.

cheers
JJB   [ Charts by FRAP ; the 'best' software for flightlog analysis ]

JJB* Posted at 7-1 00:55
Hi,

Had a look at your log.

Thanks for the details. Altough I'm not sure to understand all the details FRAP looks great!

I can't guarantee for sure that I didn't push down at the moment it happened. However I'm certain that I didn't hold down fully during two seconds (I was flying about 6-8m above water). When I saw it losing height by itself I directly pushed to go up, without response. Then confused I released all control (while imagining how I would go in the water to recover it for a quick moment) and rised the remote control higher than my head while holding up and got control of the drone again.

Sakat Posted at 7-1 01:15
Thanks for the details. Altough I'm not sure to understand all the details FRAP looks great!

I can't guarantee for sure that I didn't push down at the moment it happened. However I'm certain that I didn't hold down fully during two seconds (I was flying about 6-8m above water). When I saw it losing height by itself I directly pushed to go up, without response. Then confused I released all control (while imagining how I would go in the water to recover it for a quick moment) and rised the remote control higher than my head while holding up and got control of the drone again.

Hi,

Start using FRAP to learn more  ;-)

Tip  fly always with a screen recorder active, and analyse eacht flight using logs and screen recording.

Data of the FlightLog RC stick inputs are for 99.9%  correct...
it always take some time to stop the down motion (3 m/s down) to  a up motion.
In your log about 1 second to climb again.  see the V-z column in my chart.

I fly a lot low above water, but always with my eyes watching the drone and nothing else! (what you did)   and always ready to pull up!  

cheers
JJB

JJB* Posted at 7-1 01:37
Hi,

Start using FRAP to learn more  ;-)

But if it it was just the drone taking some time to stop down, why it wasn't responding when I was pressing to go up during the incident? Look at the video of the RC I posted, it took at least 4 sec until I regained control (about the moment the gimbal tilt straight down, which I also don't recall doing by myself)

Sakat Posted at 7-1 02:11
But if it it was just the drone taking some time to stop down, why it wasn't responding when I was pressing to go up during the incident? Look at the video of the RC I posted, it took at least 4 sec until I regained control (about the moment the gimbal tilt straight down, which I also don't recall doing by myself)

nah, check the height in your video, the momeny you pulled UP height from 1 meter to 14.
no 4 seconds delay (happily not , otherwise your drone was tranferred into a boat  ;-) )

You must have touched the cam wheel as well.

cheers
JJB

JJB* Posted at 7-1 03:20
nah, check the height in your video, the momeny you pulled UP height from 1 meter to 14.
no 4 seconds delay (happily not , otherwise your drone was tranferred into a boat  ;-) )

I can't say for the gimbal but I really lost control of the drone for a couple of seconds, during which I couldn't go up. I added timecode to the sequence in Premiere; it lasted a bit more than 3 sec.

Sakat Posted at 7-1 03:39
I can't say for the gimbal but I really lost control of the drone for a couple of seconds, during which I couldn't go up. I added timecode to the sequence in Premiere; it lasted a bit more than 3 sec.

https://youtu.be/6aLEvzqAKPg

Hi,

With respect to your opinion, the log does not show this.
Drone reacted to your RC input, see my chart. Check time, ALT and RC Updn column.

Multiple inputs on 4 directions, drone did exactly what is seen  in the video.
cheers
JJB

JJB* Posted at 7-1 03:54
Hi,

With respect to your opinion, the log does not show this.

No problem I don't take it bad at all and really appreciate your help to understand what happened!

There really was a moment during when I couldn't control the drone. The 33% RC UpDn input could be mine, that I don't recall. But you can imagine that when I saw my brand new drone going down like that while above water my first reaction was to rise the altitude and for sure not to hold it full speed down (RC UpDN 100%) during two seconds (09m 29.8s - 09 min 31.8s).

Could it be an issue with the RC sending "ghost" input? The screen of the remote was quite hot.

Sakat Posted at 7-1 04:15
No problem I don't take it bad at all and really appreciate your help to understand what happened!

There really was a moment during when I couldn't control the drone. The 33% RC UpDn input could be mine, that I don't recall. But you can imagine that when I saw my brand new drone going down like that while above water my first reaction was to rise the altitude and for sure not to hold it full speed down (RC UpDN 100%) during two seconds (09m 29.8s - 09 min 31.8s).

Down 33% plus 100% after that for while...
Did you move from postion yourself, so by accident you touched the controls?

Never say never, but don`t think 'ghost' input ever happend.

About remote screen, uh  i fly alot using the RC and cannot remember its was getting hot.
But guess all depends on the air temperature if wich you are flying.

so imo nothing wrong with your drone, happy many landings!
cheers
JJB

JJB* Posted at 7-1 04:39
Down 33% plus 100% after that for while...
Did you move from postion yourself, so by accident you touched the controls?

I was standing in the same spot, especially when I saw it going down

But I'll give it more hours of flight over solid ground from now on to see if the issue happens again or not!

Sakat Posted at 7-1 04:58
I was standing in the same spot, especially when I saw it going down

But I'll give it more hours of flight over solid ground from now on to see if the issue happens again or not!

But I'll give it more hours of flight over solid ground from now on to see if the issue happens again or not!
The "issue" will only happen if you hold the left stick down again.

Here's an extract of the relevant raw data to show what happened in your incident.
The Throttle column shows the joystick input to move the drone up/down.
The red numbers show 364, indicating that the left stick was pulled fully down for a full two seconds.

Over on the left, the grey highlighted numbers show the height as the drone descended.
The descent reached 3 metres/sec, which is the max descent speed in Normal Mode.

When you pushed the left stick up the Throttle numbers go to 1684 (max up), the descent speed slows, but the drone takes a second to overcome inertia and start climbing and another second before the climb speed reaches max.

There's nothing at all strange about the data and the way the drone responded to your joystick input.

Labroides Posted at 7-1 05:19
But I'll give it more hours of flight over solid ground from now on to see if the issue happens again or not!
The "issue" will only happen if you hold the left stick down again.

Thanks for the answer.

I understand that what is shown in the data corresponds to what is seen in the video. Same for your great explanation about the inertia decreasing before going up. What I'm saying is that I didn't hold the left joystick down at full speed during 2 seconds. The 809 value is possible though, this one I can't say for sure.

Sakat Posted at 7-1 07:06
Thanks for the answer.

I understand that what is shown in the data corresponds to what is seen in the video. Same for your great explanation about the inertia decreasing before going up. What I'm saying is that I didn't hold the left joystick down at full speed during 2 seconds. The 809 value is possible though, this one I can't say for sure.

What I'm saying is that I didn't hold the left joystick down at full speed during 2 seconds. The 809 value is possible though, this one I can't say for sure.
The data is telling a different story.

There's no doubt ... whoever had their hands on the controller pulled that joystick down and held it there for two seconds.
When you are only a few seconds away from an obstacle (that includes the ground or water), you have to be very attentive.

Labroides Posted at 7-1 07:44
What I'm saying is that I didn't hold the left joystick down at full speed during 2 seconds. The 809 value is possible though, this one I can't say for sure.
The data is telling a different story.

I also see the data and what they tell. There is no doubt that it shows a 2 sec hold down at max speed. That said, I'm not crazy, I didn't do that input and saw what I saw: the drone stopping to fly forward and losing height.

Totally agree about being attentive while flying near objects. As said, I'll give it more flying time above solid and easily accessible ground to see if that happens again. Considering your feedback that won't be an issue since all went normal according to my inputs and there was no malfunction at all.

Sakat Posted at 7-1 16:13
I also see the data and what they tell. There is no doubt that it shows a 2 sec hold down at max speed. That said, I'm not crazy, I didn't do that input and saw what I saw: the drone stopping to fly forward and losing height.

Totally agree about being attentive while flying near objects. As said, I'll give it more flying time above solid and easily accessible ground to see if that happens again. Considering your feedback that won't be an issue since all went normal according to my inputs and there was no malfunction at all.

I also see the data and what they tell. There is no doubt that it shows a 2 sec hold down at max speed. That said, I'm not crazy, I didn't do that input and saw what I saw: the drone stopping to fly forward and losing height.
After analysing flight data from hundreds of flights since 2015, I'd believe what the data shows rather than what a flyer thinks happened.
If you didn't pull the joystick down for two seconds, who did?

Labroides Posted at 7-1 16:38
I also see the data and what they tell. There is no doubt that it shows a 2 sec hold down at max speed. That said, I'm not crazy, I didn't do that input and saw what I saw: the drone stopping to fly forward and losing height.
After analysing flight data from hundreds of flights since 2015, I'd believe what the data shows rather than what a flyer thinks happened.
If you didn't pull the joystick down for two seconds, who did?

Put differently, if bugs and malfunction never happens why even bothering with a warranty?

Tell me if I'm wrong but since that input is recorded in the log of the RC. At least it wouldn't be caused by the drone itself. The RC is based on Android. I've been using that system on different devices for 10+ years daily and there were times something didn't react properly, forced shut an app, needed to reboot, .... Could it be that for some reason the RC froze for a couple of second and sent the wrong command? It was about 32°C, sunny morning and the remote was on for about 15-20 min, with screen recording on the whole time so it got hot. A bit later I actually stopped flying for a while because it became really uncomfortable to touch the screen.

Sakat Posted at 7-1 17:13
Put differently, if bugs and malfunction never happens why even bothering with a warranty?

Tell me if I'm wrong but since that input is recorded in the log of the RC. At least it wouldn't be caused by the drone itself. The RC is based on Android. I've been using that system on different devices for 10+ years daily and there were times something didn't react properly, forced shut an app, needed to reboot, .... Could it be that for some reason the RC froze for a couple of second and sent the wrong command? It was about 32°C, sunny morning and the remote was on for about 15-20 min, with screen recording on the whole time so it got hot. A bit later I actually stopped flying for a while because it became really uncomfortable to touch the screen.

Could it be that for some reason the RC froze for a couple of second and sent the wrong command?
And somehow unfroze and perfectly recorded you pushing the left stick up after two seconds?
It's a lot easier to believe that the data perfectly shows the joystick input.
Could it be that you hooked the joystick on a strap or clothing etc and inadvertently pulled the stick down?

In all the flight logs I've looked into, I've only ever seen one case of a faulty controller causing spurious control input and it was obvious in the data.
I don't see any obvious fault in the data from your incident.

Labroides Posted at 7-1 17:24
Could it be that for some reason the RC froze for a couple of second and sent the wrong command?
And somehow unfroze and perfectly recorded you pushing the left stick up after two seconds?
It's a lot easier to believe that the data perfectly shows the joystick input.

Yes, like lagging for a while and then get the UP input that I was holding. A bit like when the keyboard freeze for a while when you are typing and then all the chars entered come at once.

Not really, for the clothes, I was wearing a quite tight tshirt. RC is also new and neat clean so nothing could have been stucked in the joystick. At that moment the right joystick was also not responding.

Except that incident, all was fine (three batteries). Didn't see it happening despite going up and down multiple times.