Mini 3 Atti mode crash - is this clearly user error or potentially a malfunction?

[Dale91]

Hey everyone,

My drone took off from an almost identical flight 20 off minutes earlier and wouldn't increase altitude or stop drifting forward due to GPS issues. This resulted in a crash.

The response from DJI was as follows:

1. The aircraft worked under ATTI mode after it took off.
2. Flight Time T=38.2s, Relative Height H=-1.4 m, the aircraft crashed under ATTI mode;
3. In Attitude mode, the aircraft could not hover in place. Please fly with caution.

According to the analysis, the incident was caused by the improper flight environment. Please operate the aircraft in a wide and open area and wait for the aircraft to gain enough GPS signal before launching it. We will have our quotation team to follow up on your case.

I am wondering if it is simply user error due to my actions or if there was a malfunction. Here is a link of the video of the take off so you can see there is no obstructions of clear sky and you can see the lighting conditions:
Video of flight

Thanks for you help in advance.

Nelson

DJI Flight Log Viewer - PhantomHelp.com

 

[Phantomrain.org]

So Magnetic Interference can happen when the compass gets hammered my surrounding metal , thus not being able to rely on the Compass the drone Defaults to Atti mode meaning there is no GPS guidance and you are flying it like a helicopter on your own.

So this for me is User error as you have to be aware when your take off point is a potential for magnetic interference and than shut down your drone so the Compass can reset and find another take off point or be prepared to fly in Atti mode which is no easy task for many,

This is why we wait on the Home Point to clear and gain GPS signal lock before taking off which may have also happened in your case. Taking off without GPS can cause this as well. .

Phantomrain.org
Gear to fly in the Rain. Land on the Water

 

[FLDave]

Seems to be pilot error as it was you decision to take off without GPS operation.

 

[Meta4]

So Magnetic Interference can happen when the compass gets hammered my surrounding metal , thus not being able to rely on the Compass the drone Defaults to Atti mode meaning there is no GPS guidance and you are flying it like a helicopter on your own.

There is no evidence of any compass issue in either incident.
It was simply a very impatient flyer who zoomed off before letting the drone get any GPS.

And in the kind of incident you are imagining, the compass doesn't "get hammered" and then in flight, the compass is fine.
It's having the gyro sensor initiated with (temporarily) bad compass data that causes the problem.

But again ... this incident was 100% unrelated to any magnetic interference.

 

[Meta4]

Hey everyone,

My drone took off from an almost identical flight 20 off minutes earlier and wouldn't increase altitude or stop drifting forward due to GPS issues. This resulted in a crash.

You launched with zero sats and flew off.
At 10 seconds the drone started picking up a few sats.
At 11.6 seconds you descended hard and crashed.

Next time you fly, do it in a large, open area where there's nothing to hit and you can see the drone.
wait until you have full GPS and the drone records a homepoint.
Then fly up 100 feet where there's just clear air around it and fly where you can see it.

I am wondering if it is simply user error due to my actions or if there was a malfunction.

There's not even a hint of a malfunction.

 

[Meta4]

Seems to be pilot error as it was you decision to take off without GPS operation.

Launching without GPS doesn't cause crashes.
Flying into obstacles or the ground causes crashes.
Drones are quite controllable without GPS, but they can drift with the wind and they don't have brakes.
Flying in atti mode out in the open is quite safe.
Flying in atti mode close to obstacles is asking for trouble.

 

[Yorkshire_Pud]

From the video I get impression this flight was made in low light, dusk or dawn, would that be correct? I see roughly 05:00 in the csv.
If it was low light the the vision sensors can not work and, I think of significance in this case, that would mean the downward looking sensors could not work and so neither did "VPS". That seems to be backed up by the first and fourth warning messages in the log.

"1st message HRemote controller not boundstill confirming binding status (Code: 30306).; GPS signal weak. Fly with caution. Aircraft in Altitude Zone. Max altitude set to 30m (Code: 30014).; Aircraft storage full. Back up file and clear space (Code: 10062).; UAircraft not in flight. Vision system and obstacle sensing unavailable. (Code: 180059)."

"4th message WAmbient light too low. Vision system and obstacle sensing unavailable. Fly with caution.; GPS signal weak. Fly with caution. Aircraft in Altitude Zone. Max altitude set to 30m (Code: 30014).; Aircraft storage full. Back up file and clear space (Code: 10062).; Ambient light too low. Vision system and obstacle sensing unavailable. Fly with caution (Code: 180037)."

The VPS position holding is actually much more accurate than the GPS and indoors with no GPS my Mavic Mini and Mini 2 are near cm perfect in hover when they are using ONLY VPS and is their braking etc. when the sticks are released.
It looks to me as if you took off with perhaps functional VPS, that being due to the illumination from the house, but that, as soon as the drone was flown away from the house and out of the range of the house lights, VPS was lost and the drone switched to ATTI mode.
Indoors I get the reverse of this, sat on a dark table etc. the drone thinks there is not enough light and asks if I accept responsibility for the flight, or something similar. If I accept responsibility it will take off. If I do not accept responsibility it will NOT take off. Incidentally did you get such a warning? However, as soon as it takes off, illumination from the room lights can get to the surface under the drone and the VPS can work and work well.

I suspect that, had this been a daylight flight and if your balcony? is low enough for the ground to be within the range of the VPS, the drone would have had position control via VPS and all would have been well..................*

*It might be that, with functioning VPS but absent GPS, you might have had difficulty getting the drone back to balcony height if you had descended but I am sure Meta4 will correct me if I am wrong about that.

There is one other consideration that I can think of that relates to the lighting, low levels of light probably rendered the obstacle avoidance inoperable.





@Meta4, with regards a possible compass issue caused by magnetic metals in the table etc., if that had been the case am I correct in thinking that the OSD:yaw [360] data stream would have probably shown a significant change in value as soon as the drone was out of the range of the magnetic interference?

 

[Meta4]

*It might be that, with functioning VPS but absent GPS, you might have had difficulty getting the drone back to balcony height if you had descended but I am sure Meta4 will correct me if I am wrong about that.

Yes ... but it's not a case of if he descended.
He crashed the drone by descending at speed into the ground.

@Meta4, with regards a possible compass issue caused by magnetic metals in the table etc., if that had been the case am I correct in thinking that the OSD:yaw [360] data stream would have probably shown a significant change in value as soon as the drone was out of the range of the magnetic interference?

That's correct ... and the data didn't show anything like that.

 

[Yorkshire_Pud]

Yes ... but it's not a case of if he descended.
He crashed the drone by descending at speed into the ground.

That's correct ... and the data didn't show anything like that.

I was not questioning how the flight ended.
I was suggesting that if this flight had been flown in daylight, the VPS would have been able to work (providing that the ground was within range) and the drone would have been 'safe'.
Then, assuming he took off from a balcony, if he had then descended and flown around a bit and then tried to climb back to balcony height he might, because of the lack of GPS, have been unable to climb to balcony height ..... In just the same way as someone who flies a drone from the lip of a canyon down into that canyon and loses GPS whilst doen there may not be able to regain the lip of the canyon because of the 'no-GPS' ceiling that comes into effect.

That's correct

Thanks.

 

[Gringorio]

In just the same way as someone who flies a drone from the lip of a canyon down into that canyon and loses GPS whilst doen there may not be able to regain the lip of the canyon because of the 'no-GPS' ceiling that comes into effect

What is this "no GPS ceiling" you're referring to? I've never heard of such a thing. I've flown my P4P down into canyons and back out in atti mode with no issues.

 

[Yorkshire_Pud]

What is this "no GPS ceiling" you're referring to? I've never heard of such a thing. I've flown my P4P down into canyons and back out in atti mode with no issues.

post #2 thread Dumb Question About Altitudes - Flying Into Canyons?
Meta4 Jul 21, 2021
"In other words, if you have poor or weak GPS, your drone will not be able to climb more than 16 feet.
That's not a problem going down, but if your drone's downward facing sensors see something below them, due to poor programming, the zero height is reset and you might not be able to bring it back up again."

I seem to remember at least two other threads where this occurred but the one that sticks in my mind is some guy in the US or Canada flying from the roadside down into a canyon, losing GPS and the drone then being unable to climb back up to the lip. He saved the drone by flying it along the canyon until the lip and road dipped and he was able to remotely land the drone by the roadside

 

[Meta4]

I've flown my P4P down into canyons and back out in atti mode with no issues.

The big issue is that Mavics using the Fly app can't switch off the VPS sensors.
In atti mode, they can only fly up 5 metres.
I think they have no problem descending any distance in atti.
But if they get within VPS range of the bottom or a ledge etc, DJI's stupid programming limits them to 5 metres above what the VPS has seen.

The Phantom and other drones using DJI Go 4 can disable VPS and fly to 30 metres without sat reception.

 

[Dale91]

Yes ... but it's not a case of if he descended.
He crashed the drone by descending at speed into the ground.

That's correct ... and the data didn't show anything like that.

Thanks for your help everyone!
Just to clarify it didn't crash into the ground, it crashed into the top of the trees which were close to take off height.

By the time i realized it couldn't ascend any higher and then realized it would not stop drifting forwards, I think my reaction was to drop the height to minimize the fall distance after hitting the trees.

So to prevent a similar crash, my main issue was the limited sats? Would this in combination with the low light explain why i couldn't ascend any higher?

Cheers again everyone

 

[Meta4]

With full GPS and a wide open area, your drone is so easy to fly that a 7-year old could do it.
But when you step outside the simple basics and do things like you did, there are lots of things that can go wrong.
You need to learn the basics and then learn about what could go wrong and how to make sure they don't.

The manual is a good place to start and you'll find it here:

https://dl.djicdn.com/downloads/DJI_Mini_3_Pro/UM/20220519/DJI_Mini_3_Pro_User_Manual_v1.0_en.pdf

So to prevent a similar crash, my main issue was the limited sats? Would this in combination with the low light explain why i couldn't ascend any higher?

The section of the manual that explains this is halfway down page 60.

 

[Dale91]

With full GPS and a wide open area, your drone is so easy to fly that a 7-year old could do it.
But when you step outside the simple basics and do things like you did, there are lots of things that can go wrong.
You need to learn the basics and then learn about what could go wrong and how to make sure they don't.

The manual is a good place to start and you'll find it here:

https://dl.djicdn.com/downloads/DJI_Mini_3_Pro/UM/20220519/DJI_Mini_3_Pro_User_Manual_v1.0_en.pdf

The section of the manual that explains this is halfway down page 60.

Cheers, I'll make sure I brush up on all that stuff.

Just had a bunch of flights where nothing went wrong so hadn't needed to deal with any issues. Need to be better prepared.

Thanks

 

[Yorkshire_Pud]

So to prevent a similar crash, my main issue was the limited sats?

I believe ....No it was not, the reason was the COMBINATION of low sats and low light.

Low light disables two systems VPS and Obstacle Avoidance.

If you had had GPS OR Obstacle avoidance the drone would not have been in ATTI mode and would not have drifted.

it crashed into the top of the trees

LOW LIGHT will disable Obstacle Avoidance, so even if you nominally had obstacle avoidance switched on it could not protect the drone