First post so Hi All! Like a lot of people I got my Mavic Mini at Christmas and coming from the old Parrot AR it's a massive upgrade so really happy. I've done quite a few flights now without incident so I was beginning to get some confidence flying it until today. The weather was fine with very light winds so I went out for a couple of flights. First one was fine, went out quite a way and got some nice shots before bringing it back with no issues whatsoever. Changed batteries and sent it up again however this time it suddenly went off on it's own! At the time I wasn't putting in any control inputs but it decided to head towards a bank of trees before doing a big loop and heading back towards the river. As it came over and flew past me I managed to regain control and bring it back I have no idea why it went off?
I've uploaded the log to phantomhelp, if anyone wants to take a look and see if it's something I did wrong it would be much appreciated

https://www.phantomhelp.com/LogViewer/TR33QKBPO3WTPAVFAZK7/

Not enough sats.

You dids not wait util the aircraft set the home point before you took off, but it did record a home point later, right over the roof of the nearby cottage.

Firstly, I would suggest you never take off until your drone has locked in its home point.  This is usually when the GPS signal shows at least 4 bars and is usually at least 8 satellites.  Not waiting for a good GPS lock just raises your risk of a bad outcome significantly.

My first inclination is that you could have had a compass issue.  It looks to me like the direction the drone thinks it is facing is exactly opposite from the direction the drone in pitching/rolling in order to make the movements that it is actually doing.

Take a look at the below screenshot from PH.

The aircraft is moving from left to right at this point (red arrow), but the attitude indicator is showing that it should be moving from right to left (yellow arrow).

Did you launch your drone close to something with metal in it?  

Always a good practice to let the MM hover a minute to get the GPS signals and then do some preflight checks, check the controls and whether the aircraft respond to your controls or not...

You have a video for people to see? The flight looks crazy. I noticed your manual start has no sitck command recorded at all. How exactly did you take off?

Was takeoff from a flat roof of building?

As others have said not enough satellites,  you took over with 0 locked.  Check this post for a good list of pre flight checks - https://forum.dji.com/forum.php? ... 6orderby%3Ddateline

Yeah ... even though the flight started a bit in a hurry regarding the Navigation health you got 8 sats & a correct recorded HP before you took off for higher altitudes. My bet is a take off from a magnetic interfered area resulting in a Yaw-magYaw disagreement approx at 35 meters height were the both started to show 2 different directions & the Mini went into the famous toilet bowl curve so typical for compass errors.

At 35 meters & 72sec into the flight the Mini picked up speed which eventually reached 9m/s ... this without no real stick inputs, this is typical as the drone don't know it's heading & by that don't know which motors it should use to counter the winds that effect it ... the flight control notices that the motor inputs doesn't get it to hold & apply even more. All this in the wrong direction as the Mini doesn't have the correct compass value.

If you provide the DAT log from your phone also we there can read of the magYaw + the magnet modulation values & pinpoint this for certain ... Below were to find it (hopefully it's there) the one we need have FLY060.DAT in the end.

Mobile device DAT files (DJI GO 4 & DJI Fly)​
​
The DAT file naming convention, based on the date and time of the start of the file, is: YY-MM-DD-hr-min-sec_FLYXXX.DAT., where XXX is the flight recorder file index from the HOME_dataRecorderFileIndex field in the txt log.​
​
These are retrieved by the same method as the TXT logs. Under both iOS and Android they are in a subfolder, MCDatFlightRecords, in the folder that contains the TXT logs. In some cases, for reasons not fully explained but possibly mobile-device hardware related, and most often under Android, DAT files are not created and that folder is empty. Uninstalling and reinstalling the app sometimes fixes that. One user also found that the process required manual deletion of the app folder (apparently not deleted automatically in the uninstall process) before reinstalling the app.

I did take off before it had a GPS lock however as slup spotted I then hovered at a low level until I got a white GPS indicator before heading up. I think compass errors would fit with the behavior however I’m not sure what would cause that. The compass was calibrated on the flight before and it’s the first time anything like this has happened. Take off was from ground level on wood decking next to the house so not near any significant metal. I’m just glad I got control back before it headed off and over the river, I’ve got DJI refresh however obviously I need to be able to recover the drone for that to help!

I’ve uploaded the dat file to my Google drive https://drive.google.com/file/d/1s-tjtF8EgmbrY_MeJyNtGUJFMUYGA6yD/view?usp=sharing

Reedham Posted at 2-6 01:42
I did take off before it had a GPS lock however as slup spotted I then hovered at a low level until I got a white GPS indicator before heading up. I think compass errors would fit with the behavior however I’m not sure what would cause that. The compass was calibrated on the flight before and it’s the first time anything like this has happened. Take off was from ground level on wood decking next to the house so not near any significant metal. I’m just glad I got control back before it headed off and over the river, I’ve got DJI refresh however obviously I need to be able to recover the drone for that to help!

I’ve uploaded the dat file to my Google drive https://drive.google.com/file/d/1s-tjtF8EgmbrY_MeJyNtGUJFMUYGA6yD/view?usp=sharing

It doesn’t take significant amounts of metal to interfere, calibration the night before won’t help with new interference.

Some rules for checking your clear of magnetic interference. They are a guide.

Compass distances.

Natural and Artificial Magnetic Anomalies Warning
Note
The following information has not been objectively tested to determine it’s impact on a Drones compass accuracy in flight.
        1        Many things can distort the earth’s magnetic field in the area you are flying:
        •        Steel framed or reinforced concrete buildings, bridges and roadways, iron pipes and culverts, high power electric lines, heavy equipment, trucks and automobiles, steel tanks, electric motors and even computers.
        •        Flying between steel framed or reinforced high rise buildings will distort the magnetic field in addition to causing GPS multi-pathing.
        2        Safe distances for compass calibration
        •        6” (15 cm) minimum: Metal rim glasses, pen/pencil, metal watch band, pocket knife, metal zipper/buttons, belt buckle, batteries, binoculars, cell phone, keys, camera, camcorder, survey nails, metal tape measure.
        •        18” (50 cm) minimum: Clipboard, data collector, computer, GPS antenna, 2-way radio, hand gun, hatchet, cell phone case with magnetic closure.
        •        6 ft (2 m) minimum: Bicycle, fire hydrant, road signs, sewer cap or drain, steel pole, ATV, guy wire, magnets, chain-link fence, bar-wire fence, data collectors
that use a magnet to hold the stylus.
        •        15 ft (5 m) minimum: Electrical box, small car/truck, powerline, building with concrete & steel.
        •        30 ft (10 m) minimum: Large truck, metal building, heavy machinery.


How magnetic interference happens on the ground and the results.

The best way I can explain this is. If you put your Aircraft on the ground in normal circumstances, start it up, then lift it up and turn it 90 degrees to the right both your compass heading and IMU will both move together 90 degrees and no problems.
If you put your Aircraft on the ground and there is magnetic interference only the heading of your compass will change, your aircraft will still take off. But when it clears magnetic interference compass will then move to correct heading which you would think was great. But No, what happens is IMU is then conflicted and confused because of this sudden movement by compass, so you receive IMU exceptional heading warning, your aircraft cannot deal with data conflict so decides to switch to Atti mode dropping gps in favour of compass simply because aircraft can fly without gps but not compass.

While I can’t be certain you picked up interference from the ground or surrounding area it is the most common cause.

Reedham Posted at 2-6 01:42
I did take off before it had a GPS lock however as slup spotted I then hovered at a low level until I got a white GPS indicator before heading up. I think compass errors would fit with the behavior however I’m not sure what would cause that. The compass was calibrated on the flight before and it’s the first time anything like this has happened. Take off was from ground level on wood decking next to the house so not near any significant metal. I’m just glad I got control back before it headed off and over the river, I’ve got DJI refresh however obviously I need to be able to recover the drone for that to help!

I’ve uploaded the dat file to my Google drive https://drive.google.com/file/d/1s-tjtF8EgmbrY_MeJyNtGUJFMUYGA6yD/view?usp=sharing

The compass was calibrated on the flight before ....
Calibrating the compass doesn't fix anything and it doesn't prebent problems if you power up the drone in a magnetically affected area.
Take off was from ground level on wood decking next to the house so not near any significant metal.
And what holds the wooden decking together?
Are there nails in the boards?
Are there bolts in the under-frame?
All it takes is a bolt or large nail just an inch or two from your drone's compass to cause the kind of flight issue you've observed.

>>It doesn’t take significant amounts of metal to interfere, calibration the night before won’t help with new interference.

It was calibrated about 25 minutes before on my previous flight which I lifted off from the same spot. The decking is wood on a wood frame however there are metal screws holding it together, maybe that was it? The area I was in is very rural with not much in the way of structures etc that could have upset it so still seems a little odd. According to the log the aircraft didn't drop to ATTI mode (unless I've missed that?) and I don't remember seeing any warnings on the screen although to be fair I was pretty focused on trying to regain control and stop it from hitting the very tall trees next door!

Reedham Posted at 2-6 04:21
>>It doesn’t take significant amounts of metal to interfere, calibration the night before won’t help with new interference.

It was calibrated about 25 minutes before on my previous flight which I lifted off from the same spot. The decking is wood on a wood frame however there are metal screws holding it together, maybe that was it? The area I was in is very rural with not much in the way of structures etc that could have upset it so still seems a little odd. According to the log the aircraft didn't drop to ATTI mode (unless I've missed that?) and I don't remember seeing any warnings on the screen although to be fair I was pretty focused on trying to regain control and stop it from hitting the very tall trees next door!

Ok sorry I thought for some reason you said night before doh ;+) but even 25 minutes before same applies, it may well be the case that craft didn’t go to atti mode but this is supposed to be the failsafe when IMU and compass are conflicted .

Yes nails screws are all problem and although you had no problem first flight and probably many other flights then that is more about good or bad luck. I think looking at your log it wasn't severe so this is probably why it remained in P mode and you had some control .

A tip when you are taking off again, just check small red triangle on map in lower left corner if the heading on this is same as heading of craft then you should be ok, if it has moved you need to shut down an move and restart.

A simple exercise below will help and particularly to show if their are any problems with craft before you fly off.




1/ Never calibrate Compass unless prompted to.

2/ start AC leave until you receive enough gps lock ,

3/ if you get Compass interference turn off AC and move to another location,
without interference.

4/ In bottom left hand corner on your map you will see small red triangle, check to make sure that this triangle is pointing in the same direction (heading) as your AC, this will show good compass on the ground.

Raise AC to height of 8ft
Hover for 20 seconds
Fly forward 2ft
Backward 2ft
Left 2ft
Right 2ft
Up 2ft
Down 2ft
Yaw left
Yaw right
Each time returning to hover position
If you have a good horizontal each time, you will then know you have good GPS good IMU and good Compass, and your ready to fly.

This exercise takes less than 2 minutes and is worthwhile doing before every flight,

hallmark007 Posted at 2-6 04:59
Ok sorry I thought for some reason you said night before doh ;+) but even 25 minutes before same applies, it may well be the case that craft didn’t go to atti mode but this is supposed to be the failsafe when IMU and compass are conflicted .

Yes nails screws are all problem and although you had no problem first flight and probably many other flights then that is more about good or bad luck. I think looking at your log it wasn't severe so this is probably why it remained in P mode and you had some control .

I would be interested in how many cases there are compass errors. You could then check the production date for these.
I've been traveling by car and plane with a Spark for about a year (Germany, Greece, Spain) and with the MP for 2 years now (Germany, Croatia, Spain, Italy, Mauritius). During this time, I had no compass or IMU errors, and I never thought of calibrating anything. Except on my desk, there is a steel beam underneath, there is for all a compass error. Move it 20 cm and everything is fine. 1.5 m behind me is a "continent (Australia, approx. 2 sqm)" made of steel with countless magnets. On the right is a pin board 2 m away, also full of magnets. My MM and MP fly here in a confined space without errors (ATTI). Not to forget the WLAN networks, computers and bluetooth devices that are in operation all day. For screwing on the MP I use magnetic tools for the small screws.
Just as it is common with mobile phones that the same components are not always installed in the same model, it could also be here ... ??

A short compass check after taking off is also mandatory for me. Doesn't take 10 seconds ...

I'm not sure if reading this forum has saved my Mini from flying away, but it's a distinct possibility. After reading about some fly aways and digging deeper into compass and yaw issues, I started paying attention to my drone's orientation after each and every takeoff. And sure enough, a few days ago, as I raised my Mini to eye level and was waiting for it to get more satellites, I checked the map and the drone's orientation was about 30 degrees off from the actual orientation. I immediately landed and performed a compass calibration, which fixed the issue. I wonder if I would have gotten a flyaway had I not noticed that and proceeded with my flight. Checking drone heading is now a must for me before each flight.

DaMa Posted at 2-6 06:44
I would be interested in how many cases there are compass errors. You could then check the production date for these.
I've been traveling by car and plane with a Spark for about a year (Germany, Greece, Spain) and with the MP for 2 years now (Germany, Croatia, Spain, Italy, Mauritius). During this time, I had no compass or IMU errors, and I never thought of calibrating anything. Except on my desk, there is a steel beam underneath, there is for all a compass error. Move it 20 cm and everything is fine. 1.5 m behind me is a "continent (Australia, approx. 2 sqm)" made of steel with countless magnets. On the right is a pin board 2 m away, also full of magnets. My MM and MP fly here in a confined space without errors (ATTI). Not to forget the WLAN networks, computers and bluetooth devices that are in operation all day. For screwing on the MP I use magnetic tools for the small screws.
Just as it is common with mobile phones that the same components are not always installed in the same model, it could also be here ... ??

There was something similar done over on MP1 forum, and results were a little surprising, it seemed early batch compass was not so well sealed, some members were able to nail it down to a certain month when users received their drone, so yes possible.

I do think that we seen similar with spark many yaw error leading to drone going to atti mode etc, these drones are pretty low to the ground, I noticed before Christmas I was using a small landing pad for my mini and steel sprung ring could affect compass heading if craft was not in the center of pad, I just done away with it and just always check before I head off, Mavic mini lends itself to proper checking before flying, it has such good battery that one shouldn’t be in any hurry to launch before checks .

During the last couple of outings (MM CE Version) I have problems with the IMU and the compass not calibrating. If I power the MM down and re start the MM then sometimes after the 3rd or 4th time the IMU and Compass will calibrate and I can fly with no problems. I am starting to think the MM is very sensitive to its surrounding magnetic and radio interference. It may be due to the radio and MM signal being limited in the CE country regions.  I also now make sure the MM is within 30 meters from me when the battery reaches 50% and make smaller movements on the controls to limit the power being used by the motors and if I see any strange movements from the MM I land it immediately and replace the battery.

Droneflier Posted at 2-6 07:15
During the last couple of outings (MM CE Version) I have problems with the IMU and the compass not calibrating. If I power the MM down and re start the MM then sometimes after the 3rd or 4th time the IMU and Compass will calibrate and I can fly with no problems. I am starting to think the MM is very sensitive to its surrounding magnetic and radio interference. It may be due to the radio and MM signal being limited in the CE country regions.  I also now make sure the MM is within 30 meters from me when the battery reaches 50% and make smaller movements on the controls to limit the power being used by the motors and if I see any strange movements from the MM I land it immediately and replace the battery.

This is very commendable and currently maybe sensible. The firmware shall be actually responsible for this, not the pilot. It should not happen that flight maneuvers result in overload!

Reedham Posted at 2-6 01:42
I did take off before it had a GPS lock however as slup spotted I then hovered at a low level until I got a white GPS indicator before heading up. I think compass errors would fit with the behavior however I’m not sure what would cause that. The compass was calibrated on the flight before and it’s the first time anything like this has happened. Take off was from ground level on wood decking next to the house so not near any significant metal. I’m just glad I got control back before it headed off and over the river, I’ve got DJI refresh however obviously I need to be able to recover the drone for that to help!

I’ve uploaded the dat file to my Google drive https://drive.google.com/file/d/1s-tjtF8EgmbrY_MeJyNtGUJFMUYGA6yD/view?usp=sharing

So ok ... from the DAT file the readout clearly tells us that this event was due to a compass error coming from a magnetic disturbed take off area.

It's a lot of graphs, I know ... but will explain what we see.

The dark blue one is register the magnetic field & as seen it's start to differ in a major way already at 70sec. & then peaks around 85sec.

If we then look at the purple (that's AC yaw angle according to the gyro)
and the yellow (which is the magnetic yaw angle according to the compass)

Usually these 2 angles follows each other well with a slight constant difference of  perhaps max 30 degrees when everything is healthy & good to go ...

At 70sec. & especially at 90sec. these to angles start to differ in a major way ... at 90sec. the difference is 55 degrees and that makes the flight control confused about the correct directions. (Actually the angle difference is even worse when you luckily are going in for landing at 208sec. the difference is then nearly 100 degrees.

This then leads to that the AC tries to correct the position hold when wind starts to affect it to the wrong direction ... which leads to worse position hold, and more counter action from AC in wrong direction ... and the speed picks up.

All this could have been avoided ... not by calibrating the compass, but with looking at the map view in the Fly app & the drone icon there. Most probably the drone icon didn't point in the right direction there compered to reality. If you had seen that your correct action should have been to move to another take off spot well away from the initial.

Hi, I am sorry to know about this unfortunate incident and thank you for sharing your informative insights on what happened. I would recommend for you contact our DJI support team to start up a ticket at ( https://www.dji.com/support?site=brandsite&from=nav ). We have the professional team who would do their best to  assists you with regards to this matter. Thank you.

Thanks for the advice, especially slup for the useful analysis of the logs. I'll always check the compass orientation when taking off from now! I do think that DJI need to work out a more elegant solution to dealing with this kind of data mismatch, I was very lucky in this instance that my flight level meant I cleared the trees etc however it could have been a very different story.

My Mavic also asks for a compass calibration on pretty much every other flight, is that normal?

Reedham Posted at 2-7 03:48
Thanks for the advice, especially slup for the useful analysis of the logs. I'll always check the compass orientation when taking off from now! I do think that DJI need to work out a more elegant solution to dealing with this kind of data mismatch, I was very lucky in this instance that my flight level meant I cleared the trees etc however it could have been a very different story.

My Mavic also asks for a compass calibration on pretty much every other flight, is that normal?

It depends I find similar with my own moving location always triggers calibration.

If you continue to check small compass arrow you should be ok and similar problem with compass IMU conflict shouldn’t occur .
It's great to see we now have somebody who can post charts and give correct explanations, charts are no good if they cannot be explained correctly good job .