MaVerm lvl.1 Netherlands Online

12# Hi,  I have done some test to emmulate a boat a bit...    I wanted to use the Mavic Air drone (2018) to film me on the boat. I went to the water and started the systems. Directly after take of the drone was very unstable; did decend  much even with throttle full to ascend. I was just able to get it back to the boat and catch the drone. A second trail after a restart resulted in to the same panick. Directly afterwards I tested the drone on land: everyting fine. even flying above water. I started a test: The RC is on. 1: Placed the mavic Air drone  on the ground. Powerd it on. Wait for connect and ready to fly.  take of.   result:   GOOD 2: Placed the mavic Air drono  on a smalll play toy car. not moving . Powerd it on. Wait for connect and ready to fly.  take of.   result:   GOOD 3: Placed the mavic Air drono  on a smalll play toy car. pulling it to move slow walking speed . Powerd it on.        Wait for connect and ready to fly.  Then STOP the car.  take of.   result:   GOOD 4: Placed the mavic Air drono  on a smalll play toy car. pulling it to move slow walking speed . Powerd it on.        Wait for connect and ready to fly.  Keep walking the car.  take of.   result:   UNSTABLE My conclusion:    The Mavic Air needs a sold not moving take of place.   So i can not use it on a Boat. I can be happy that I was able to recue  the drone on the water...  because the un stability is rather random and not to predict  if the drone can decend or ascend..

Additional information from the Log File:    IMU altitude at take of:    -15.4 ft
    VPS altitude at take of:    0.3 ft
During the troube some short light the VPS give a rather correct altitude. The IMU is swinging a lot.
While VPS is between 0.3  to  3.6 ft.  the Imu measures:   -22  to 21.3 ft   (totaly incorrect.)

I would expext a better senor check..

MaVerm Posted at 8-7 10:44
Additional information from the Log File:    IMU altitude at take of:    -15.4 ft
    VPS altitude at take of:    0.3 ft
During the troube some short light the VPS give a rather correct altitude. The IMU is swinging a lot.

Hi there,

Thank You for reaching out. We are truly sorry for the inconvenience you are experiencing. Kindly take note it is always adviceable to take of from ground. The aircraft flies by visual positioning. Be careful to avoid areas with specular reflection such as water. We recommend that you not fly the aircraft close to the water, as flying on the water surface will affect the normal function of visual positioning, and the aircraft may fall into the water due to the impact on positioning accuracy.  It is generally recommended to fly the drone at a distance of at least 3 m above the water surface. (Refer to the corresponding user manual for specific requirements of the flight environment.)

If the drone descends suddenly when hovering from the ground, try the steps below;
1. Please check whether the flight environment is in a high attitude area. The higher the attitude, and thinner the air. Thin air will affect the aircraft’s propulsion, resulting in propulsion overload, sudden descent when hovering, unstable flight, etc. Please ask the customer to refer to the Max Service Ceiling Above Sea Level on the official DJI website. After repeated tests, we have determined a safe flight altitude range for the aircraft.
2. Check whether the propeller is in good condition.
3. It is recommended to calibrate IMU.
4. if still reoccur kindly send us video or footage of the problem.

We are looking forward to your response for further assistance. Thank You

Hi,  Thanks for your efford to wite a complete answere and advice.  I do take care of those issues.   
In my last reply, the one before yours.  I already gave an analisis about my investigation.
It turns out that when taking off from an unstable, moving, swinging platform, the IMU sensor is out of normal measurments.
If is than take of again from a solid platform, the drone flies normal good.  So the IMU is correct. no callibration needed.
Coming back to the unstable flight:
When taking of from a moving platform, easy to do is taking the drone in/on your hand and swing it more or less. It than take off you
noitice immediatlely that the motor rotate restless..  When importing the flight record: VPS altitudeis between 0.3  to  3.6 ft.  the IMU altutude measures:   -22  to 21.3 ft   (totaly incorrect.)
So it seems that the IMU sensor is disturbed by the swinging take off.  
The Drone (software) does not notice this and let the drone fly unstable...
So What hapens in the software?  Does it use a bad inital value from the IMU at take off?  Why not using the IMU callibration values?  those are correct.
Why not blocking the take off with bad IMU sensor values?There is no message which tells the IMU (take off) status to the operator.

To be honest I would be surprised if it was launching whilst moving that is causing the problem but I suppose it depends on what you condsider to be unstable, perhaps you could upload a video to youtube etc. showing this.

Plenty of people launch drones from moving boats.
The thing to bear in mind is how it holds position.
There are two separate systems responsible for position holding, one is GPS which relies on GPS satellites, the other is VPS which uses visible light and the optical sensors on the bottom of the drone and MIGHT use the IR landing protection sensors.
From the behaviour I have seen in videos it seem to be that in boat launches the GPS system "rules" and the drone tends to want to hover with respect to the earth so the boat moves away from it. Hence you should generally launch from the stern.
However if VPS can lock onto something on the boat then there are various threads that suggest that VPS can, in certain circumstances, track the locked onto object.
I suppose it is possible that your drone was caught in a rare conflict between VPS and GPS position holding but I think that would be poor programming if it can happen.

If launching from a boat I would be inclined to get the drone up and away from the boat ASAP, so launch somewhere where there are no object above the drone and get it up to 10 or 20 m as quickly as possible i.e. give it full throttle at launch and hold full throttle until the drone is well above the boat I would not use an automated launch, start the drone's motor with the CSC then give it full throttle to launch and climb to safety.

With regards to the descent you noticed, does the log suggest that the drone was descending to try and maintain a constant VPS height.
In stationary, hand-launches on dry land most of my flights with mini 1 & 2 and mavic 2 drones show a little initial drift in both height ( <2ft ) and horizontal position ( <4ft ) but they quite often slowly return to at least near the launch point.

"noitice immediatlely that the motor rotate restless." that's because the drone senses movement and is trying to hold position, if your hand is still it won't fight you

Hi Thanks for your very complete reply.
I had an other manufacturer drone first,  cheaper. This one flies very well.. . But on the boat it does not take off.  It checks the gyro first, than is takes off, only
when the gyro has been in a very solid postition. could not succeed it on  boat.   So, I read the DJI performance and toke my spare mony to buy one.
Okee,  DJI is much better, more robust, more features.  Al my tests en traine on land went well.
Then I went on t water with my boat.  I let the drone toke off two times, both on divfferent places, both with very less wind. Both trails I could hardly recue the drone
from diving in the water.  The horizontal position was rather in control.  But vertical is was feighting against thedrone going in teh water.   Althou I pulled it to Ascend much to maximal it decened each time. I Was not able to get it higher then 4 or 5 meters. So I had to catch the drone both trails.   After this I tested the drone immediatly in the harbor.  There it flied perfect.  even taking of at the water edge.  even above the water decending to 0,5meter went very stable.
I found the fleight record featue.  and the flight record in the app.  Both show the stange values.  Very much swinging IMU altitude values.  vertical velocyty 0,5 to 1,5 ms up.   While the altitude did not change much. the Mode reports P-GPS (brake)  somet imes when it goed wrong.  normally it is P-GPS.
At home I started testing and researching and googling.  My results:
Place the drone on a kids car pull it,  and take off  ->  unstable
take the drone on the hand and swing more or less   take off -> unstable
Each time the IMU altitude values are far different from VPS altitude.
So my conclusion, the IMU is initialized at take off, but does noet check the stability values.   (some thing that my cheap drone does well)

to your questions:
I was not able to get the drone to 10m or higher..
I have a Mavic Air   2 hours fleigth time so far.
I will do some more tests to describe the restless motor better.

MaVerm Posted at 8-10 07:03
Hi Thanks for your very complete reply.
I had an other manufacturer drone first,  cheaper. This one flies very well.. . But on the boat it does not take off.  It checks the gyro first, than is takes off, only
when the gyro has been in a very solid postition. could not succeed it on  boat.   So, I read the DJI performance and toke my spare mony to buy one.

Hi,

if you like to get help advice here  share yiou flightlog for your issue

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

cheers
JJB

I did already uploaded it.   in another thread I copied the link.   I must check where it is ...

MaVerm Posted at 8-10 12:22
I did already uploaded it.   in another thread I copied the link.   I must check where it is ...

https://www.phantomhelp.com/logv ... GPZRZ96YUKTDWWK79EN
Perhaps

MaVerm Posted at 8-10 12:22
I did already uploaded it.   in another thread I copied the link.   I must check where it is ...

Hi,

Had a look at your flightdata.

Baro height at start log = 2.5 meter, distance 32.3 meter.
Did you power ON your drone earlier adn moved the drone on the boat ?
Normally Baro height is set to zero at powering ON, was the position on the boat 2.5 meter higher than where you power ON th drone?

With Obstacle Avoidance ON the flighmode is P-GPS (brake), OA in positon Brake or Bypass.
P-GPS in the log = OA off.
Speed shows values direct after take-off, GPS speed (calculated by LatLon data) shows speed as well, with only RC UP input.
Drone facing East and moving into direction South. If Drone is GPS stabilized oke it should stay at position.

With RC UP the Verticial speeds values shouwing climb while drone did not climb as these values suggests.
Why your drone did not ascend pushing UP, no idea.


So not much new to tell you, afaik DJI drones starting from a moving object is not quitte possible.

Luck that you were able to catch the drone and kill the motors.

Second chart : red baro height line = height not climbing (with UP input) or descending (with UP or zero UP input)

cheers
JJB    [ Charts by FRAP ; the 'best' software for data and visual flightlog analysis (www.jjbfrap.eu) ]

O wow,  thats a big analisis.  it is my uploaded fleight..  
It is the one on the boat.  it is on a quite canal, old river.  everyting was quite.  you see a little difting of my boat due to litle wind.
It was a scary fleight.  My moat is a small  4 x 1.6 meter open boat,  you sit on almost water level.
I had layed a big plate on my seats as a good take off platform.  Immediatly the drone went drifting and swinging.  Like a drunk-drone.
The data is correct:  much up throttle but no or less ascend.   
A second fleight on a place 1km further:  the same problems.
I did take off with throttles both down, propelors start rotating.  than throttle up...   that is not the take of from the button in the app.  But they seem to act the same.
Then in the harbor:  every thing good...  so Drone is not defect.

Questions:
Can you also see the IMU altitude in the data?  because this value is shown in the fleigth record in the Go4 app..  It is the main strange value which happens at instable fleigths.
(I do think that the fleigth controller is using IMU for height aswell,  an causes the decend behaviour |!)
You can imagin that Im affraid the test all.  because, what if the IMU is going to cause an uncontrolled ascend.  the drone will fly away, and gone......
Than, should I down load and install your fleigth controller program? Now i only used your online upload feature,  which shows much less information..

I'm still willing to get a goed boat fleigth.  nice to make a movie about my boat...
(DJI does not have such a nice good fleigth recorder tool?  little poor... if so)
Should DJI be able to analise this to advise and, mor important make a BUG fix in the software.  If thsi drone has been a Car, Tesla,  then it would be world news!

Do the DJI product managers also share these forums?
How can we report a bug issue to DJI?
How is customer feed back and problem reports handled?

MaVerm Posted at 8-11 03:09
O wow,  thats a big analisis.  it is my uploaded fleight..  
It is the one on the boat.  it is on a quite canal, old river.  everyting was quite.  you see a little difting of my boat due to litle wind.
It was a scary fleight.  My moat is a small  4 x 1.6 meter open boat,  you sit on almost water level.

what do you mean by "IMU altitude" ?
"Can you also see the IMU altitude in the data?"

If you mean the VPS height as measured by the downward looking sensors then below is a plot, its called OSD:sWaveHeight in the chart and in ft not metres.


In the folder "FlightRecord" where you found DJIFlightRecord_2024-08-05_[17-45-43].txt there should be a folder whose name is "MCDatFlightRecords". In that folder there should be ".DAT"  flight logs, if so and if you see one for 2024-08-05 or 24-08-05 whose name contains ''FLY057'', can you upload that file to a file hosting website, make its page public and post the URL here.
If JJB*  can read the DAT there may be a clue in there as to why the drone didn't climb.
Alternatively post the story, phantomhelp link and the link to the DAT over on
https://mavicpilots.com/forums/crash-flyaway-assistance.85/

MaVerm Posted at 8-11 03:09
O wow,  thats a big analisis.  it is my uploaded fleight..  
It is the one on the boat.  it is on a quite canal, old river.  everyting was quite.  you see a little difting of my boat due to litle wind.
It was a scary fleight.  My moat is a small  4 x 1.6 meter open boat,  you sit on almost water level.

Hi,

IMU height does not exists. Height is measured by a barometric sensor, and this height is set to zero at start. So during the flight the baro height showing in the app is always a reference to the take-off height. Values can be plus or minus ofcourse during the flight.
It is 'normal' to see a small negative number if you land at the same spot used for take-off.

In my chart you see both height values ; ALT = baro height, IR ALT = ir sensor.
The bottom Infra-red sensor is measuring the actual height drone to surface below, only ofcourse if this distance is within its measuring range AND the signal back is received in good order (depens on the structure of the surface)

The RC with screen has a screen recording feature, if you use the one without a screen, find a screen recording app for your mobile device.


cheers
JJB

on the phantom help site,  the link above.  I see a column  IMU altitude (4th column) .  I think that this is a altitude calculated from the IMU.
As im my test the IMU gives strange value variuoations after a bad take off. this sensor makes it suspicious.
I will upload the requested DAT file for you...

Hi JJB,  thenks for helping us aswell.
in this picture the start fleight valus are shown...  You See  IMU  does not start at zero....  strange...
Notice that for a good and stable take off, you are completely right.  Then I see those altitude values being the same.
Any how,  still the problem is:  NO ascend while throttle is UP  upt max!   see the first graph in this threath...
Which proces is blocking this?

the DAT  file at:
https://we.tl/t-ABOs22fHcA

MaVerm Posted at 8-12 04:12
on the phantom help site,  the link above.  I see a column  IMU altitude (4th column) .  I think that this is a altitude calculated from the IMU.
As im my test the IMU gives strange value variuoations after a bad take off. this sensor makes it suspicious.
I will upload the requested DAT file for you...[view_image]

Hi,


The Field names are created by the originator of PhanthomHelp.
All signals are processed by the the IMU, but as said height is measured by a barometric sensor.
So there are imo 2 heights in data;  Baro height and IR height.

cheers
JJB

MaVerm Posted at 8-12 04:19
Hi JJB,  thenks for helping us aswell.
in this picture the start fleight valus are shown...  You See  IMU  does not start at zero....  strange...
Notice that for a good and stable take off, you are completely right.  Then I see those altitude values being the same.

Hi,

At starting the motors (from that moment the flightlog is written) the baro height is set or reset to zero.
Same after landing, the moment the motors are stopped the baro height is set to zero again.

If baro height is not zero at start that there something fails in the power ON initialisation sequence and /or at start sequence, mayby more signals in error ....

Do you have data from a normal flight, powering ON at starting motors while drone at the same position at ground ?




cheers
JJB

MaVerm Posted at 8-12 05:02
the DAT  file at:
https://we.tl/t-ABOs22fHcA

Hi,

oops  many many errors in the DAT.

do you have a DAT file from a good flight ?


// ** copy paste from DAT.000 :    8164 [L-FMU/MOTOR]Start. reason:MR_REQ_RC_NORMAL

-141.891 :    1051 [L-FDI]NS(0) FUSION(0): fault off, disagree
-141.891 :    1051 [L-FDI]NS(0) FUSION(1): fault off, disagree
-140.894 :    1100 [L-BATTERY]get_cell_voltage_callback_ack failed!||
-139.896 :    1150 [L-BATTERY]get_cell_voltage_callback_ack failed!||
-139.696 :    1161 [L-FDI]NS(0) FUSION(1): fault on , alti_err_large

//
more....
//
.000 :    8164 [L-PWM]set actuator mode:CTRL
.000 :    8164 [L-FDI]tilt when motor on = 0.125250, static flag = 0
.000 :    8164 [L-FLYMODE]MR_REQ_R_RC_ASST_TAKEOFF
.002 :    8164 [L-FMU/FSM]state changed. cur: on ground
.004 :    8164 [L-N_MIS]Init plan 3 from requester 2 for reason 9. Pre plan 5
.004 :    8164 [L-TAKEOFF]enter assit takeoff plan
.004 :    8164 [L-N_MIS]Exit mis 3. New mis 4
.004 :    8164 [L-N_MIS]Switch new mis 4 req 2 from mis 3 req 2
.020 :    8165 [L-IOC]Set CL.  87.4 Degree
1.510 :    8239 [L-N_MIS]Exit mis 4. New mis 5
1.510 :    8239 [L-N_MIS]Switch new mis 5 req 2 from mis 4 req 2
2.380 :    8283 [L-FMU/FSM]state changed. cur: in air
2.388 :    8283 [L-N_MIS]req abort all for assit takeoff finished
2.393 :    8283 [L-TAKEOFF]exit assit takeoff plan
2.394 :    8284 [L-HOME]Update. lati:51.7122014 longti: 4.9112889 alti: 11.60
2.398 :    8284 [L-N_MIS]Switch new mis 3 req 2 from mis 26 req 16
2.398 :    8284 [L-FLYMODE]SWITCH2HOVER!
3.032 :    8316 [L-RC]Emergency break. rc vib!
3.032 :    8316 [L-FMU/LED]action changed. set home0)
3.757 :    8352 [L-FDI][FDI] ns atti(0) alti is outlier
3.757 :    8352 [L-FDI][FDI] ns atti(1) alti is outlier
3.770 :    8353 [L-FDI]NS(0) FUSION(0): fault on , disagree
3.770 :    8353 [L-FDI]NS(0) FUSION(1): fault on , disagree
3.770 :    8353 [L-FDI]fatal error: no ns with level < 4 !!!
3.810 :    8355 [L-FDI]NS(0) FUSION(0): fault off, disagree
3.810 :    8355 [L-FDI]NS(0) FUSION(1): fault off, disagree
4.289 :    8379 [L-FMU/MOTOR]safe_near_grd:false
4.290 :    8379 [L-FMU/FSM]not near ground
4.555 :    8392 [L-FDI][FDI] ns atti(0) alti is outlier
4.555 :    8392 [L-FDI][FDI] ns atti(1) alti is outlier
4.568 :    8393 [L-FDI]NS(0) FUSION(0): fault on , disagree
4.568 :    8393 [L-FDI]NS(0) FUSION(1): fault on , disagree
4.568 :    8393 [L-FDI]fatal error: no ns with level < 4 !!!
4.608 :    8395 [L-FDI]NS(0) FUSION(0): fault off, disagree
4.608 :    8395 [L-FDI]NS(0) FUSION(1): fault off, disagree
4.954 :    8412 [L-FDI][FDI] ns atti(0) alti is outlier
4.954 :    8412 [L-FDI][FDI] ns atti(1) alti is outlier
4.967 :    8413 [L-FDI]NS(0) FUSION(0): fault on , disagree
4.967 :    8413 [L-FDI]NS(0) FUSION(1): fault on , disagree
4.967 :    8413 [L-FDI]fatal error: no ns with level < 4 !!!
5.007 :    8415 [L-FDI]NS(0) FUSION(0): fault off, disagree
5.007 :    8415 [L-FDI]NS(0) FUSION(1): fault off, disagree
5.127 :    8421 [L-FMU/LED]action changed. Normal Flash(0)
5.127 :    8421 [L-FMU/LED]type:0, normal flash action:
5.127 :    8421 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
5.370 :    8433 [L-NS][AHRS] reset baro just after takeoff
6.006 :    8465 [L-FDI][FDI] ns atti(0) alti is outlier
6.006 :    8465 [L-FDI][FDI] ns atti(1) alti is outlier
6.024 :    8466 [L-FDI]NS(0) FUSION(0): fault on , disagree
6.024 :    8466 [L-FDI]NS(0) FUSION(1): fault on , disagree
6.024 :    8466 [L-FDI]fatal error: no ns with level < 4 !!!
6.064 :    8468 [L-FDI]NS(0) FUSION(0): fault off, disagree
6.064 :    8468 [L-FDI]NS(0) FUSION(1): fault off, disagree
6.405 :    8485 [L-FDI][FDI] ns atti(0) alti is outlier
6.405 :    8485 [L-FDI][FDI] ns atti(1) alti is outlier
6.423 :    8486 [L-FDI]NS(0) FUSION(0): fault on , disagree
6.423 :    8486 [L-FDI]NS(0) FUSION(1): fault on , disagree


// **
cheers
JJB

Hi,  Ithing I toke the correct files.  For a good test in the harbor.
https://we.tl/t-z1JPyX3OXk
if correct this is a fleigth from land to water.

Hi JBB,  I now see your LOG dump in the thread.  Okee seems not well all those messages...   
You want a correct DAT to verify the tool if it has read the file correctly....   I can understand.
By the ways,  als this analisis is done with that Phantom tool..   I gone install it .  And will surely redo tests to get a reliable compare between good and bad...
I appreciate your feedback very much

MaVerm Posted at 8-12 04:12
on the phantom help site,  the link above.  I see a column  IMU altitude (4th column) .  I think that this is a altitude calculated from the IMU.
As im my test the IMU gives strange value variuoations after a bad take off. this sensor makes it suspicious.
I will upload the requested DAT file for you...[view_image]

"I see a column  IMU altitude (4th column) "
Whoops, I have never noticed that before or not paid any attention to it lol.

Hi,  Do you now mean that the column "IMU altitude"  has a wrong name?   And if so,  what should it be?

Hi Sean and JJB,
The nice graph and data that you showed here.  Which app should I install?
Is it the Flightreader?   On that site I do not see it in the feature list...

I just up loaded this one:
https://www.phantomhelp.com/logviewer/ELXEAOJF9XR44TTZUUEID7OTGR2XMMIO
I must say that many of my take off disturbance test resulted in a good take off..
This is the only wrong one. What I did:  just toke the drone on my hand and swinged it rathe much.  (when to much the mobile app complains about unstable signal)
What the drone did this wrong time, not even speed up the propellors.  to is would fall from my hand instead of ascending..
The 4th column is very wrong altitude!
This was my todays first fleight, just to show everything is working fine.
https://www.phantomhelp.com/logviewer/64NGHZZ01FH6B4AKNC4KKI4E99SZTXY9

MaVerm Posted at 8-12 11:48
Hi Sean and JJB,
The nice graph and data that you showed here.  Which app should I install?
Is it the Flightreader?   On that site I do not see it in the feature list...

JJB*'s is from his own "Frap" see the thread
https://forum.dji.com/forum.php?mod=viewthread&tid=203659
The one I used is CsvView
https://datfile.net/CsvView/downloads.html

MaVerm Posted at 8-12 12:02
I just up loaded this one:
https://www.phantomhelp.com/logviewer/ELXEAOJF9XR44TTZUUEID7OTGR2XMMIO
I must say that many of my take off disturbance test resulted in a good take off..

Yikes 70ft at motor start .... interesting, I think you have a sick drone.

MaVerm Posted at 8-12 11:42
Hi,  Do you now mean that the column "IMU altitude"  has a wrong name?   And if so,  what should it be?

I don't know it is incorrect but I have always just looked for the OSD:height column in the phantomhelp CSV or something similar in the CsvView menus.

Thanks again.  Im spending more time on the forum than Im flying with the drone...   :-(   
good response is helpfulll
And i earn a lot of points....

The following comes from analyzing FLY057 that took place 2024-8-5 15:43:30 GMT. In summary, the MA thought it had an upward velocity > 0.0 when, in fact, it was 0.0; This caused the Flight Controller to apply a downward bias to compensate. The OP stated that the MA would descend even with full positive throttle. This isn't quite correct. The MA did ascend with 100% throttle. But, 100% throttle was mixed with intervals where the throttle was [0 - 80]%. The downward bias would cause the MA to descend in the intervals where the throttle was [0 - 80]%.

Looking at the interval before motorStart (time = 0.0). The IMU temperature gets to the 65° C at time -58 secs nad the barometer (presure) value stabilizes at about -31.5. From this I assume the MA height, altitude, etc is stable. During this interval the GPS veldD data is 0.0 indicating the MA is not ascending or descending. However, the veldD values computed by the Flight Controller has the MA ascending at about 0.6 m/s.


Looking at the vertical velocity response to the throttle it can be seen that 100% throttle does cause the MA to ascend. VelD is the downward velocity; negative values mean the MA is ascending. During 0% throttle intervals the MA begins to descend.



I'm at a loss to explain why the FC thinks the MA is gaining altitude when it's not. An accelerometer with a non-zero bias would cause the vertical speed to be changing; i.e., accelerating. A stucj barometer would result in a vertical velocity = 0.0.

The other .DAT submitted (FLY059) doesn't show this phenomena.

I gone sit to understand your analisis.
but is seems that I'm able to get the controller/algoritms to a strange behaviour.   Maybe I should go for a tester Job at DJI..  I'm very good in creating unexpected situations...
For back ground.  I'm an electronics designer with a broad experience.  I I had the time I could be able to develope my own...  All those sensors and opto equipment is rather known stuff for me.
So te get them into stress is not difficult, unless the application is clever enough to prevent errors.  I really like the good flight behavior of the DJI drones
for these pictures you used  https://datfile.net/CsvView/downloads.html.
Thanx again!

For this error full log file.   I had the drone all the time in my hand.  it would not ascend at all.  the rotor speed did not boost up to ascend speed...  
i don't know if the log contains rotor speed?  

MaVerm Posted at 8-13 07:09
I gone sit to understand your analisis.
but is seems that I'm able to get the controller/algoritms to a strange behaviour.   Maybe I should go for a tester Job at DJI..  I'm very good in creating unexpected situations...
For back ground.  I'm an electronics designer with a broad experience.  I I had the time I could be able to develope my own...  All those sensors and opto equipment is rather known stuff for me.

"I had the drone all the time in my hand."


There must be some mistake. This flight (FLY057) attained an altitude of 4.19 meters over launch point. The altitude at motorStart (time = 0.0) was -22.39. At time 12 secs the altitude was -18.2 meters.


The accelerometer data also shows the MA rising and moving about.

Yes, motorSpeed data is available but it won't help much with this.

Hi,   a new uncontrolled flight:
https://we.tl/t-sk9fzlrYPK

I was some more days away, had less time to proceed the DJI analisys...

This flight went wonder full uncontrolled.  
I started all up.  placed the MA on the ground.  and toke off rather fast.  
I got some message on the screen but did not read them very well..
The MA went into the air but after a few second I noticed it was not reacting as it should: unstable..
after 15 or 20 seconds I was able to land it into the gras, there it kept rotating te propelors.
I could take it out of the gras after 20 seconds (I had to get the keys of the gate I had to pass)
When taking out of the grass, the propelors rotated still, i turned the droen and rotation stopped.

In the DJI app I can open the fleigth record. and I notice that it had no GPS...  at half the recod it has GPS and my location jumped to there..
Normay the MA has its GPS very fast in 5 seconds..  it sems to rememer them..  now the MA was about 5 days unpowered.
typical systems need more than a minute to get all GPS sattelites