During the start of the mini 2 something strange happened. After starting by means of moving the 2 sticks together, I made the mini 2 rise by moving the left stick forward. After it had risen about 40 cm, an uncontrolled sideways movement began. I tried to stop it by moving the left stick down. However, the mini 2 had moved sideways so quickly that it had already landed in a bush. Fortunately only one wing was damaged. I had just received this mini 2 as a replacement for the previous damaged one. Tested inside for a while and worked stable with only the optical sensor switched on. The second flight was outside with the above result. Before I started flying I waited until there were enough satellites and the homepoint was set.

After the crash, I noticed in the flight data that the homepoint was not the same as where the mini took off. Just as I took off, the homepoint was also updated (see video) and immediately afterwards the mini 2 went sideways into the bushes.  It seems that it is important to make sure that the mini 2 and the home point match exactly during take-off. In the video you can see that I did not touch the right stick but that the mini 2 still started flying sideways.  After the crash the IMU and compass recalibrated.

The next flight paid close attention to the homepoint and had no starting problem. The horizon in the video was still a bit skewed while flying, I had already adjusted the gimbal straight. Maybe I need to calibrate the IMU again.

Update 25-1-20: Added flight log, Maybe somebody can find something strange in it.
21-01-2021 Flight-data.zip (4.25 KB, Down times: 6)

2021-1-25 Upload

Click on the file name to download attach

Thanks for sharing.

I don't have much confidence in DJI at the moment with the Mini 2 with all the stories I've heard so far.

Drifting issues, RTH completely flying in the wrong direction and your sideways issue.

The only issues I've had so far Motor error and the gimbal dipping in normal mode.

The question is who is to blame in an incident like that the operator or DJI for faulty software and hardware.

I am more inclined to think while most people will say operator I would say the manufacturer should take the blame if you was on a field and that drifted sideways into children playing I would be mortified.

Did you calibrate the IMU and the compass.
Did you do that near any metal objects, car, rebar in concrete.
Being new why didn’t you do the auto takeoff and wait till it rose and hovered and wait for you.
Doug

you are correct, Doug.  When the drone starts drifting, most of the time it is the IMU or compass that need to be calibrated. When I take off, I always hover and watch the drone's behaviour. When drifting breng it down again and callibrate compass. Try again, if not solved, calibrate the IMU.

Had drifting with my M2A before. All described action above solved the problem and could have a stable flight.

1 big rule: if the drone behaves strangly, don't fly, breng it down and solve the problem.   Flying with an unstable drone can cause damage to the drone and even more hurt people!

I agree with JohnDG & Donut Michigan to calibrate the IMU.  The IMU, Inertial measurement unit, is essential to flight telemetry and stability.

In the video you can see that I did not touch the right stick but that the mini 2 still started flying sideways.
The incident you describe sounds like a classic example of a yaw error.
What direction was the drone facing on the launch spot?
What was the surface where you launched?

After the crash, I noticed in the flight data that the homepoint was not the same as where the mini took off.  ...   It seems that it is important to make sure that the mini 2 and the home point match exactly during take-off.
Your incident had nothing to do with the recording of the home point.
If you want to avoid yaw errors (which usually end quite badly), avoid launching from reinforced concrete surfaces and check that the drone icon on the map in the app is facing the same direction as the drone.
ie .. if the drone is facing south, the drone icon should also be facing south.

JohnDG Posted at 1-23 15:05
you are correct, Doug.  When the drone starts drifting, most of the time it is the IMU or compass that need to be calibrated. When I take off, I always hover and watch the drone's behaviour. When drifting breng it down again and callibrate compass. Try again, if not solved, calibrate the IMU.

Had drifting with my M2A before. All described action above solved the problem and could have a stable flight.

When the drone starts drifting, most of the time it is the IMU or compass that need to be calibrated.
This incident had nothing to do with calibration.
Normally there is no need to recalibrate anything.

Recalibrate the IMU after a crash and the compass after modifying or rebuilding the drone.

I have noticed one thing. After reading hundreds of post of something going wrong it is most of  the time not the drone.
Many are new operators. And generally an operator err.
Ice freezing to cold of environment , flying in fog or above it.
Flying low to water.
Not realizing it is to windy at RTH height.
Not making sure the battery locked and clicked when inserting.
RTH set to low  or was set to hover or to land instead of RTH.
Not checking if blades screw was loose.
Oops flying to far and not making it back.
Slamming into a tree 75 feet up.. that was me lol.
Bad wire from the controller, me buying a cheap money lol.
IMU and compass not calibrated.
Improper amount of satellites.
So when something happens think what would cause it. And it’s probably not the drone.   Sorry.
Yea there are a few drone failures but very few.

djiuser_hJmrJuQGB6Sn Posted at 1-23 14:01
Thanks for sharing.

I don't have much confidence in DJI at the moment with the Mini 2 with all the stories I've heard so far.

I don't have much confidence in DJI at the moment with the Mini 2 with all the stories I've heard so far.
DJI drones are very reliable.
Genuine faults can occur, but are extremely rare.
But confusion and mistakes caused by the drone owner are very common.

RTH completely flying in the wrong direction
I'd like to see that incident.
RTH is extremely reliable and cannot fly in the wrong direction.
The flight data would show what really happened.

The question is who is to blame in an incident like that the operator or DJI for faulty software and hardware.
I am more inclined to think while most people will say operator I would say the manufacturer should take the blame.
Since it was not caused by faulty hardware or software, it's not DJI's fault.
It was caused by the owner powering up his drone where the compass was deflected by steel or iron very close.
That caused the initialisation of the gyro sensor to have a false directional reference.
When the drone moved away from the steel and the compass was no longer affected by it, the compass data and gyro data were conflicting.
This ended up with the drone accelerating away on a curved trajectory as it tried to hold position but couldn't because of the faulty data.

Donut Michigan Posted at 1-23 18:15
I have noticed one thing. After reading hundreds of post of something going wrong it is most of  the time not the drone.
Many are new operators. And generally an operator err.
Ice freezing to cold of environment , flying in fog or above it.

I have noticed one thing. After reading hundreds of post of something going wrong it is most of  the time not the drone.
...

IMU and compass not calibrated.
There is no need to be recalibrating anything.
If the compass and IMU were working last time you flew, they will still be working properly.

Labroides Posted at 1-23 18:21
I don't have much confidence in DJI at the moment with the Mini 2 with all the stories I've heard so far.
DJI drones are very reliable.
Genuine faults can occur, but are extremely rare.

Have you looked at the forums lately? So many issues with the Mini 2 currently.

I assume you will back up DJI on this post too?
https://forum.dji.com/forum.php? ... p;extra=&page=1
So many phones on the supported list that aren't working....looks exactly like a DJI problem to me.

OverCombination Posted at 1-23 20:05
Have you looked at the forums lately? So many issues with the Mini 2 currently.

I assume you will back up DJI on this post too?

Have you looked at the forums lately?
No ... I'm only here every day.
I assume you will back up DJI on this post too?
Sorry ... I can't see where DJI made any comment on that thread, so I can't see how I could back them up.
So many phones on the supported list that aren't working....looks exactly like a DJI problem to me.
Looks that way to you?
Because you don't know what the issue is, it must be DJI?
Flawless logic.

Labroides Posted at 1-23 20:31
Have you looked at the forums lately?
No ... I'm only here every day.
I assume you will back up DJI on this post too?

DJI also haven't commented on this thread. But you are here as their fanboy saying they do no wrong.

There are people with genuine issues with their products, sometimes it is going to be user error but saying every issue is diminishes those individuals.

Go through that thread. Look how many different model phones are on there, all of which are on the supported list, but you don't think that is a DJI App issue? I can replicate the issue every single time I fly, as can others. The DJI statement is we fixed it ages ago... when they clearly have not. Their "solutions" of turn this off, turn that off, restart this, restart that with no difference shows they don't know the issue or how to solve it.

Look at the growing list of IMU issues on this forum. Silence from DJI. Are the hundreds and hundreds of posts on that user error?

Being a fanboy and pretty much saying DJI does no wrong (or when they do its apparently super rare) is ridiculous:

Mavic Mini Incompatible Firmware
Can't Update Firmware DJI Mini 2
Compass Calibration on mid-air
DJI Mini 2 IMU Calibration Problems
DJI Fly App - Needs its screen scaling fixed as a priority.
Not Pairing
System UI isn't responding.....

The drone is targeted and sold to beginners. Beginners who are losing their money when their drone crashes, flies away or keeps throwing up errors to ruin their confidence.

The forum stinks of a product that was rushed to market in time for Christmas.

OverCombination Posted at 1-23 22:43
DJI also haven't commented on this thread. But you are here as their fanboy saying they do no wrong.

There are people with genuine issues with their products, sometimes it is going to be user error but saying every issue is diminishes those individuals.

DJI also haven't commented on this thread. But you are here as their fanboy saying they do no wrong.
Being a fanboy and pretty much saying DJI does no wrong (or when they do its apparently super rare) is ridiculous:
There are people with genuine issues with their products, sometimes it  is going to be user error but saying every issue is diminishes those  individuals.
I've never suggested that at all, making such a sweeping accusation with no evidence to support it diminishes your credibility.

Go through that thread. Look how many different model phones are on there, all of which are on the supported list, but you don't think that is a DJI App issue?
Hey Mr ignorant .. Look through that thread and count how many times I commented on anything in it.
If you want to complain to someone about your issue with configuring your phone, do it in the relevant thread but leave me out of it.

Hi,

Best to get help is to upload your flightlog using this link > https://www.phantomhelp.com/LogViewer/Upload/
(don`t bother if data is not shown in PhantomHelp, just put the upload link on here)

If you wanna keep your data private, upload your flightlog to JJBsoftware@gmail.com, i will report back to you.

It doesn`t matter if the HP is recorded not at start (better ofcourse), as long as you have GPS reception Mini will fly stabilized.
Only reason for moving uncontrolled away after takeoff is
a) compass error (wich is also seen in the log as a yaw error)
b) rarely seen a GPS flyaway as the initial GPS position is not the same as the position after takeoff; often due to bad GPS sat count and/or reception during startup.

But flightlog will explain 98% what happend.



cheers
JJB

Donut Michigan Posted at 1-23 14:14
Did you calibrate the IMU and the compass.
Did you do that near any metal objects, car, rebar in concrete.
Being new why didn’t you do the auto takeoff and wait till it rose and hovered and wait for you.

Did you calibrate the IMU and the compass.
The drone just came from DJI and they deliver fully tested and ready to fly replacements. I therefore assume that an IMU calibration is not necessary. The first time outside I did get the compass message, so I first calibrated it.

Did you do that near any metal objects, car, rebar in concrete.
Of course not, I always read the manual first. Also thought of the mobile phone (on by you earlier mentioned page 48), an explicit comment in the manual that I don't hear many people about. Controller with telephone placed on the ground so that I had a good view of the screen and drone calibrated at a height of 1.5 m.


Being new why didn’t you do the auto takeoff
What are you basing on that I'm new to flying with drones?
I just followed the procedure from the manual (page 42).
This crash is not caused by not using the autostart procedure.




and wait till it rose and hovered and wait for you.
If you read my story carefully you will see that it happened during the ascent, all this within 5 seconds. There was no time at all to wait / see if everything was functioning properly.

JJB* Posted at 1-24 01:39
Hi,

Best to get help is to upload your flightlog using this link > https://www.phantomhelp.com/LogViewer/Upload/

But flightlog will explain 98% what had.

Good tip. I'm going to look into that, see if I get to know more.

I always get the home point update when I first take off so that doesn’t seem unusual.
You’re a click away from knowing what happened, send you flight data to JJB and that will end the guessing game.

Added flight log, Maybe somebody can find something strange in it. (see first post for logfile)

DronePerspectiv Posted at 1-25 13:40
Added flight log, Maybe somebody can find something strange in it. (see first post for logfile)

Hi

Your flightlog is a CSV file....what i need to analyse is your flightlog.txt. like "DJIFlightRecord_2021-01-21_[09-09-09].txt"

One question, do you remember wich heading your drone was heading at the ground/at this auto-takeoff?  Roughly within 10 degrees?

cheers
JJB

JJB* Posted at 1-26 00:45
Hi

Your flightlog is a CSV file....what i need to analyse is your flightlog.txt. like "DJIFlightRecord_2021-01-21_[09-09-09].txt"

Try this:
https://1drv.ms/x/s!Ak13vG6CU8yVgnYPA20bh9UUy4dd?e=FktHwa

DronePerspectiv Posted at 1-25 13:40
Added flight log, Maybe somebody can find something strange in it. (see first post for logfile)

Your flight data shows nothing unusual about the homepoint.
You had good GPS right from the start and the drone recorded it's homepoint.
At 4.9 seconds as you started to climb, the homepoint was updated.
This is completely normal.
The difference between the 1st and 2nd homepoints is less than a metre and had nothing at all to do with the incident.

At around 7.8 seconds the drone started to move without any corresponding joystick input.

Everything about this incident says Yaw Error.
I asked some questions back here:
https://forum.dji.com/forum.php? ... st&ptid=234302&pid=
Can you answer them?

FYI .. I explained how a yaw error happens here:
https://forum.dji.com/forum.php? ... 302&pid=2402227

I used up two battery charges to try and duplicate the yaw error, could not to the degree the drone was not controllable.
I would turn the drone on when it was sitting on a steel plate. Plenty of errors and the attitude indicator was way off but as soon as I was 1 to 2 feet above the steel everything corrected.
It did not hold the faults.
Yes you have to fly the drone until it corrects but nothing that caused me to sweat.

I don’t dispute your warnings I just can’t get it to be an uncontrollable aircraft, seems to me it would have to be the perfect storm to cause an uncontrollable flyaway.
It does wander and look confused but if the operator stays with it he or she can recover.

I was prepared for the drone to fly strange so I would think that helped so one would need to pay attention during takeoff and not dismiss it as a simple event.
Be ready for the unexpected when taking off and you should be able to fly another day.

If the condition is caused by faulty hardware then It’s probably game over and not much you can do.

Labroides Posted at 1-26 01:54
Try this:
https://1drv.ms/x/s!Ak13vG6CU8yVgnYPA20bh9UUy4dd?e=FktHwa

Thanks, but i need the original flightlog.txt.
I know how to get or read this CSV file.

cheers
JJB

Lisa3 AKA Karen Posted at 1-26 03:05
I used up two battery charges to try and duplicate the yaw error, could not to the degree the drone was not controllable.
I would turn the drone on when it was sitting on a steel plate. Plenty of errors and the attitude indicator was way off but as soon as I was 1 to 2 feet above the steel everything corrected.
It did not hold the faults.

I used up two battery charges to try and duplicate the yaw error, could not to the degree the drone was not controllable.
I would turn the drone on when it was sitting on a steel plate.
All that proves is that you didn't induce a yaw error.
But lots of drones have been lost through accidentally induced yaw errors.

To get a yaw error you would need to have some steel close enough to the compass to deflect it which would cause an incorrect initialisation of the gyro sensor.
But the steel can't be so big/close that it sets off a compass error.
The effect on the compass has to be a similar strength to the earth's normal magnetic field, but in a different direction.

Be ready for the unexpected when taking off and you should be able to fly another day.
If you are ever unlucky enough to have a yaw error (of more than just a few degrees), there's nothing you can do.
The drone would zip off at very high speed before you could think of doing anything, and probably wouldn't respond to your controlls anyway.

Labroides Posted at 1-26 04:10
I used up two battery charges to try and duplicate the yaw error, could not to the degree the drone was not controllable.
I would turn the drone on when it was sitting on a steel plate.
All that proves is that you didn't induce a yaw error.

If you are ever unlucky enough to have a yaw error (of more than just a few degrees), there's nothing you can do.
The drone would zip off at very high speed before you could think of  doing anything, and probably wouldn't respond to your controlls anyway.


It seems like it must have been something like that because this was exactly what happened to me.
Have been thinking what the direction must have been during the take-off and estimated it was about 95/100 degrees.
The flight log says 116 degrees. This would have been a fairly significant deviation of 15/20 degrees.
The starting place is ordinary paving stone. I have already made many starts from this place without a single problem.
It seems that in addition to enough satellites and right homepoint, I have to pay more attention to whether the compass is also giving the right direction.
Hardware and software must be good because the flights after the crash did not cause any problems.

OK, today was the final day of testing.
I did everything to Sunday to cause an issue.
Start up near ferrous metal at different distances and locations around the drone - warnings but zero loss of control or flying issues.

Gave up on metal and brought out the magnet and this is when it got interesting.
Again startup at different distances and positions - it had to be flown but fully controllable.
Until I hovered over the magnet. It flew hard right with a CW yaw, fast but I was able to control. It was a fight for a time but it was recovered.
I don’t know if the OP could have controlled his occurrence because there were no stick inputs that it was attempted, other than a down blip.
Other times it would start to yaw quickly but stay in place.
One time I actually landed on the magnet, on purpose, the magnet helped the final distance and the Spark stuck.

There can be a perfect storm to cause what the OP experienced but I’m convinced it would have to be a true external magnetic interference not just ferrous metal (maybe the metal disrupting the earths field but I could not duplicate)

I will not dismiss the drone itself as the cause, on start up it could have had an issue in the IMU or compass but not out of range to give a warning and when the drone was flown it got lost.
The OP did note that he had used this same area before, maybe since, with no issues so who knows.
I equate it to a sensor not throwing a fault. Its reading could be incorrect, compared to actual, but it is not out of an acceptable range.

Hopefully we’ll find out definitively what happened.

Lisa3 AKA Karen Posted at 1-26 15:10
OK, today was the final day of testing.
I did everything to Sunday to cause an issue.
Start up near ferrous metal at different distances and locations around the drone - warnings but zero loss of control or flying issues.

The OP did note that he had used this same area before, maybe since, with no issues so who knows.
This is often the case because the issue is caused by something small and close to the compass.
A small difference in positioning the drone is all it takes.

Labroides Posted at 1-26 16:57
The OP did note that he had used this same area before, maybe since, with no issues so who knows.
This is often the case because the issue is caused by something small and close to the compass.
A small difference in positioning the drone is all it takes.

That’s scary, I don’t have a metal detector but if I did I can’t imagine the quantity of metal bits I would find.
If this is the case, when your drones number is up it’s up.

Thank you for your input and information.

It happened something similar to me once, in December.

Just when I was climbing after the take off  the home position changed midflight to around 30 meters away from the start position and the drone started to go in that  direction by itself. I realized just in time before it hit a building.

In the log you could see it had some GPS  interference just after the take off. I took off with all the sats synced, as allways, 10 secs after GPS interference, home position changes to another position and the mini 2 starts flying by itself to the side to reach that new home position while it kept climbing.

Never happened again though, and I've done 182 flights.

Huginn Kenningar Posted at 1-26 18:19
It happened something similar to me once, like two or three months ago.

Just when I was climbing after the take off  the home position changed midflight to around 30 meters away from the start position and the drone started to go in that  direction by itself. I realized just in time before it hit a building.

In the log you could see it had some GPS  interference just after the take off. I took off with all the sats synced, as allways, 10 secs after GPS interference, home position changes to another position and the mini 2 starts flying by itself to the side to reach that new home position.

That's confusing.
What actually happened, must have been something else.
Your drone doesn't make any effort to reach your home position unless it's returning to home.
If you post the flight data, it might show the cause of your incident.

Lisa3 AKA Karen Posted at 1-26 17:16
That’s scary, I don’t have a metal detector but if I did I can’t imagine the quantity of metal bits I would find.
If this is the case, when your drones number is up it’s up.

If you are launching where it might be a concern, you can check that the drone icon in the map window is pointing the same direction as the real drone.
ie.  If the drone is facing south but the icon is facing east, that's showing a problem.

Labroides Posted at 1-26 18:35
In the log you could see it had some GPS  interference just after the take off. I took off with all the sats synced, as allways, 10 secs after GPS interference, home position changes to another position and the mini 2 starts flying by itself to the side to reach that new home position.

That's confusing.

https://www.phantomhelp.com/logviewer/NRVMM9HQVVMO76454ZCQ/

The real take off point was the same as the landing point. The take off point that appears in the log was set midflight, so the log doesn't represent the reality. You can see that i was just holding the right stick forward to go up (I use the controller in mode 3) but it also went sideways due to the GPS interference problem.

In the log you can see the mini is sliding to the south, but that's wrong, I took off from the landing spot and it slided to the north, to where  the new home point was set XD

Huginn Kenningar Posted at 1-26 19:18
https://www.phantomhelp.com/logviewer/NRVMM9HQVVMO76454ZCQ/

The real take off point was the same as the landing point. The take off point that appears in the log was set midflight, so the log doesn't represent the reality. You can see that i was just holding the right stick forward to go up (I use the controller in mode 3) but it also went sideways due to the GPS interference problem.

There's definitely something odd about the flight and it's difficult to work out exactly what happened.

The real take off point was the same as the landing point.

According to the data, the landing point is 49 metres south of where the drone was at the start of data.
So you launched to the west, outside the curved building rather than east of the curved building, between it and the pool?


The take off point that appears in the log was set midflight, so the log doesn't represent the reality.
Midflight ??
The data starts with the motors starting and the altitude at zero.
The weird thing I see about the location shown at the start of the data is that it suggests that the drone is 7.5 metres away from where it recorded a homepoint.
Unless you had moved the drone, that's an indication of poor GPS location data


You can see that i was just holding the right stick forward to go up (I use the controller in mode 3) but it also went sideways due to the GPS interference problem.
In the log you can see the mini is sliding to the south, but that's wrong, I took off from the landing spot and it slided to the north, to where  the new home point was set XD
I can't work out what was going on.
Did you see the drone make these movements or are you basing your description on the track shown in the data?

As the drone lifted off at 2.3 seconds the data shows that the homepoint was updated (which is not unusual).
The new homepoint was about 9 metres south of the original one.
But there's no data to show the drone moving in any direction in response to this.

The data shows a lot of GPS Position NoMatch warnings for the first 21 seconds of the flight but none after you cleared that curved building.
When there are just a few, it's usually a false alarm.
But the number in this data suggest a real problem with the GPS data close to that building.

Labroides Posted at 1-26 18:38
If you are launching where it might be a concern, you can check that the drone icon in the map window is pointing the same direction as the real drone.
ie.  If the drone is facing south but the icon is facing east, that's showing a problem.

Yes, thanks, I did notice the orientation was off when I was playing.
It reset and was accurate once I was clear of the disruption.

I will pay closer attention when flying and not trying to crash.

DronePerspectiv Posted at 1-26 11:51
If you are ever unlucky enough to have a yaw error (of more than just a few degrees), there's nothing you can do.
The drone would zip off at very high speed before you could think of  doing anything, and probably wouldn't respond to your controlls anyway.

Hi,

Best is to always check before flying if the actual heading is the same as the heading in the map view flyapp. if not the same don`t fly.

The power button is at an akward position on the mini!  Powering up while holding mini upside down and moving during the powering up cycle can have effect on the compass and gyros.

Best is (well my opinion) to get use to power up holding craft top up horizontal, finger under de drone to 'master' the on button.

cheers
JJB

Labroides Posted at 1-26 21:28
There's definitely something odd about the flight and it's difficult to work out exactly what happened.

The real take off point was the same as the landing point.

Yep, both homepoints are false XD

The real flight was more like this.



I did another flight in the same location that day, just before the incident and it went ok, I was in control all the time. Here's the log of the other flight: https://www.phantomhelp.com/logviewer/3D1XZ69IYPUNYWT9JLI1/

Then I swapped batteries and that happened. On both flights I took off/landed from the same spot, just like you can see in the flight that had no issues. I never take off before all sats are synced and everything is ok (no warnings/calibrations/updates).

I've just flown those two times in that location and the drone has never done that thing again since then, and it's been through some firmware and app updates. Since then I pay more attention to every take off and don't start to look at the screen till it's way up, just in case XD