DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Yannie

Well-Known Member
Joined
Feb 11, 2018
Messages
64
Reactions
41
Age
40
Hi guys,

I had a near miss. I just wanted to alert you of this so nobody loses their Mavic Pro.

I was filming a lake.
I parked my car on one side of the lake and I launched my drone from the opposite side of the lake.

I completed my filming and knew the drone's battery was pretty low at around 18%. Time to wrap things up.
I thought I would be clever and just land my drone near my car, so that I didn't need to carry it back to the car.
I landed it via the video feed and confirmed that it was actually stationary on the ground.

I unplugged my phone and turned off the Mavic controller. I completely packed everything away.

After about 5 minutes, after walking halfway back to the car, I suddenly heard my Mavic in the air!!

Confused and freaking out, I opened my bag and turned on the controller on again (without a smartphone). As soon as it connected to the drone, it immediately starts beeping, indicating that the battery was now very very low (eg, single-digit %)

Because the battery had dropped to a certain level while it sat idle on the ground beside my car, it had autonomously taken off again, climbed to the RTH altitude I previously set (100m) and heading 'home', over the middle of the lake.

It didn't have enough battery to make it home. Mid-way through it's RTH course over the lake, the battery reached the critically-low threshold and it began doing an forced-decent over the middle of the lake.

I could not override this descent - however I was fortunate on this occasion because I had a physical line-of-sight to the drone from where I was standing. I was even more fortunate that I was able to manually roll the drone sideways and steer this forced-descent onto a patch of dirt which happened to be beside this particular section of it's RTH course.
(again, my phone wasn't plugged in to the controller so I had no video feed; it was all manual)

On this occasion the drone wasn't destroyed, but this was a very, very close call.

I understand that the drone will return to home on low battery and lost signal...but I was surprised that it will do this even after the pilot has landed it somewhere else. This seems to be a loop-hole in the RTH algorithm.

If I ever want to land the drone at an alternative site, how can I land it and STAY THERE, even when the battery gets low? (assume I'm flying using just the manual controls, ie, without a smart phone)

Has this happened to anyone else?

Please be careful of this, guys.

EDIT:

Here is the flight log of the flight.
This flight was about 6 weeks ago, with many uneventful flights since. You'll have to forgive me if some of my earlier recollection of the event was inaccurate in my initial post.
I have deliberately removed all the GPS coordinates for confidentiality.
What I can tell you is: the take-off point was also the 'home point'. The landing point was not the home point.

DJIFlightRecord_2018-01-03_(11-39-13) (GPS coordinates removed).xlsx
 
Last edited:
You likely didn't keep the throttle down long enough to stop the motors after "landing" so it was still running and "in flight" even if touching the ground. So when you turned the controller off it entered signal loss RTH.

Also if you do that it would be a good idea to set home point to aircraft location once you're there...
 
can I ask for clarification on this subject, I am aware that under these circumstances this can happen but if you had shut the motors off when it landed the first time could the motors re-start under rth and lift off? Also how could you be sure that they were actually off? thanks
 
can I ask for clarification on this subject, I am aware that under these circumstances this can happen but if you had shut the motors off when it landed the first time could the motors re-start under rth and lift off? Also how could you be sure that they were actually off? thanks

I descended until it was 30cm off the ground. I gimballed down to check the ground was safe. I then pointed the gimbal to the horizon for the final landing phase (as I usually do, to reduce the chance that something on the ground scratches the front of the camera). I then further descended it completely down to the ground. Thus, I wasn't able to see whether dirt was still blowing around when I knew the drone was on the ground. I didn't pay attention to the rpm of the motors. The drone was not in the air and the altitude was 0, at the time I turned off the controller. I thought that would be the end of the story...

I have no idea why the blades would still spin when it knows it's on the ground...why would they keep spinning when it is on the ground?
 
  • Like
Reactions: Gonçalo
Becasue since you REALLY don't want them to stop spinning when they shouldn't they only stop after 3 seconds of holding the stick down and no descent being detected. That's when it knows it's on the ground. If you jsut touch and release the stick straight away the props will still be spinning.
 
I descended until it was 30cm off the ground. I gimballed down to check the ground was safe. I then pointed the gimbal to the horizon for the final landing phase (as I usually do, to reduce the chance that something on the ground scratches the front of the camera). I then further descended it completely down to the ground. Thus, I wasn't able to see whether dirt was still blowing around when I knew the drone was on the ground. I didn't pay attention to the rpm of the motors. The drone was not in the air and the altitude was 0, at the time I turned off the controller. I thought that would be the end of the story...

I have no idea why the blades would still spin when it knows it's on the ground...why would they keep spinning when it is on the ground?
It looks to me that DJI need to look at that scenario, it should not lift off the ground on its own in rth mode. Be interesting to see any further opinions from the more experienced users!
 
What did the AC status say? If motors are completely shutdown, shouldn’t it say “Ready to Takeoff”? I guess it’s worth testing out in a controlled scenario. Set low battery and critical battery to a higher level, and test with the AC >50m away?
 
Becasue since you REALLY don't want them to stop spinning when they shouldn't they only stop after 3 seconds of holding the stick down and no descent being detected. That's when it knows it's on the ground. If you jsut touch and release the stick straight away the props will still be spinning.
(Seeing as you bring this up: Who in their right mind thought it would be a good idea to make the emergency engine stop accessible just by triggering a certain position simultaneously with the sticks? There have been a number of times that I've been trying to orbit around something in the ocean by moving back/right while simultanously both descending and turning left (ie to maintain the camera view on the object as I move to the right). I'm very worried that I'm going to accidentally activate an emergency engine stop, as the left stick is back-left and the right stick is back-right. I wish it could be configured to require additional buttons to be pressed on the controller to force an engine shutdown, not just certain stick movements)
 
I wish it could be configured to require additional buttons to be pressed
It is configurable (and configured that way by default) to only activate in case of an FC malfunction being detected.

But if you have both sticks really all the way in the corners your footage must be really ugly to look at.
 
  • Like
Reactions: Lake_Flyer
(Seeing as you bring this up: Who in their right mind thought it would be a good idea to make the emergency engine stop accessible just by triggering a certain position simultaneously with the sticks? There have been a number of times that I've been trying to orbit around something in the ocean by moving back/right while simultanously both descending and turning left (ie to maintain the camera view on the object as I move to the right). I'm very worried that I'm going to accidentally activate an emergency engine stop, as the left stick is back-left and the right stick is back-right. I wish it could be configured to require additional buttons to be pressed on the controller to force an engine shutdown, not just certain stick movements)
Try that CSC command in the simulator and see for yourself how a max pitch sideways, max descending, max spinning corkscrew would NOT (as in could never) help you get any usable shot.
 
can I ask for clarification on this subject, I am aware that under these circumstances this can happen but if you had shut the motors off when it landed the first time could the motors re-start under rth and lift off? Also how could you be sure that they were actually off? thanks
You must always see the drone with your bare eyes during landing and taking off. Actually, you MUST ALWAYS have the drone in sight (or at least be able to see it). Period. Rule number one!

How you can be sure that everything is OK? By looking at it. By actually being there.

And no, when the motors shut down they will never turn on again without your input.
 
  • Like
Reactions: bobj151
Thanks for all info on this, although I like to think I am fairly experienced and knowledgeable I am still learning from different scenarios you guys experiences - thanks.
 
It is configurable (and configured that way by default) to only activate in case of an FC malfunction being detected.

But if you have both sticks really all the way in the corners your footage must be really ugly to look at.
I have decreased the sensitivity of the sticks to make movements slower and more cinematic, but I appreciate your opinion of my footage nevertheless.
 
Becasue since you REALLY don't want them to stop spinning when they shouldn't they only stop after 3 seconds of holding the stick down and no descent being detected. That's when it knows it's on the ground. If you jsut touch and release the stick straight away the props will still be spinning.
I wonder if in this scenario, the OP even could have sent a 3 second 0-throttle command at all. The RC probably lost LOS anyhow with the MP at the ground at a distance. It probably lost signal a second after touchdown. After 'hovering' a few seconds on the ground, waiting for a control signal, it went into RTH.
 
  • Like
Reactions: helmut
You seem to not be aware that you can adjust the sensitivity of the sticks.
Of ONE stick (yaw), on the others you can only adjust expo which has no effect at full deflection.

I wonder if in this scenario, the OP even could have sent a 3 second 0-throttle command at all. The RC probably lost LOS anyhow with the MP at the ground at a distance. It probably lost signal a second after touchdown. After 'hovering' a few seconds on the ground, waiting for a control signal, it went into RTH.
That is possible and already happened to someone else. Here he said he still had a picture though.
 
I wonder if in this scenario, the OP even could have sent a 3 second 0-throttle command at all. The RC probably lost LOS anyhow with the MP at the ground at a distance. It probably lost signal a second after touchdown. After 'hovering' a few seconds on the ground, waiting for a control signal, it went into RTH.
The delay between landing and taking off again was about 5 minutes - significantly longer than the standard delay between first losing connection and starting the RTH sequence (from my experience this is something like 5-8 seconds)
 
You seem to not be aware that you can adjust the sensitivity of the sticks.
You mean a slightly slower sideways descending corkscrew would serve you?
Show me those shots. I'm curious how they would look.
You can do this already by just having both sticks a fraction towards the CSC command. We used to call that piloting:rolleyes:. I don't see a reason to adjust your settings because you want to have both sticks FULL down/in.
 
Lycus Tech Mavic Air 3 Case

DJI Drone Deals

New Threads

Forum statistics

Threads
130,584
Messages
1,554,099
Members
159,588
Latest member
gfusato