Mavic Pro behaving strangely

[Jackcutrone]

Recently my Mavic Pro started behaving strangely. I was getting warnings of strong wireless interference, IMU calibration at start up most flights, compass calibration warnings and compass redundancy warnings. For example, an occasional strong wireless interference warning was coming when I was taking off from the middle of the runway at my club's RC field at a point that was over a quarter mile from the nearest power transmission wire.

Ever since a mysterious crash caused me to lose the MP about a year ago and I was unable to find it for months, I have been using a lot of items to help me find it, a Trackimo, a Tile, a strobe and florescent tape all over. I have been using all these for months with no problem. In these recent flights with the warnings, I had all that equipment in the same location of the MP as I have in months of problem free flights. However, I thought that one of these might be the issue so I removed all of them, left my iphone in my car, recalibrated the compass, and was still getting these errors frequently, though not every flight. Also had problems recalibrating the compass. Went through the procedure, got the OK from the lights on the drone, set it down and still had a compass error. Tried calibrating the compass several times and was unable to calibrate it. Weird.

I was in the process of downloading flight data from the drome using the DJI assistant and got an error message saying that there was an inconsistency in the firmware (or software) - don't remember the exact language). I had updated the firmware before these error-prone flights using the controller a few days ago. So I followed the DJI Assistant recommendation and update the firmware.

Could the firmware conflict have caused all the problems I described above? The weather has not allowed me to do any more test flights.

Also, when I tried to download the flight records from the drone, I only found the last 8 flights or so. I was trying to loo at the data from one of the flights where I had one of these errors. Unfortunately, after that flight, i did a whole lot of very short test flights without errors and so the flight with the error was not among the 8 that I found. I was able to download the more limited data from my iPad. While i can identify that flight when I playback the flight on my iPad, I cannot correlate it to the download iPad data as that does not show the date-time of the flight. I tried looking at the data with airdata but it has been so long since I used that website, I thnk I am doing something wrong. I thought the airdata site would let you replay the flight, but I don't see an option to do that on their website.

I looked on my ipad and found that the first flight that I had the problem with was 11/21/20. I uploaded that flight to airdata and it shows multiple errors. Here is the link to that flight on airdata.

 

[geelong6grant]

Recently my Mavic Pro started behaving strangely. I was getting warnings of strong wireless interference, IMU calibration at start up most flights, compass calibration warnings and compass redundancy warnings. For example, an occasional strong wireless interference warning was coming when I was taking off from the middle of the runway at my club's RC field at a point that was over a quarter mile from the nearest power transmission wire.

Ever since a mysterious crash caused me to lose the MP about a year ago and I was unable to find it for months, I have been using a lot of items to help me find it, a Trackimo, a Tile, a strobe and florescent tape all over. I have been using all these for months with no problem. In these recent flights with the warnings, I had all that equipment in the same location of the MP as I have in months of problem free flights. However, I thought that one of these might be the issue so I removed all of them, left my iphone in my car, recalibrated the compass, and was still getting these errors frequently, though not every flight. Also had problems recalibrating the compass. Went through the procedure, got the OK from the lights on the drone, set it down and still had a compass error. Tried calibrating the compass several times and was unable to calibrate it. Weird.

I was in the process of downloading flight data from the drome using the DJI assistant and got an error message saying that there was an inconsistency in the firmware (or software) - don't remember the exact language). I had updated the firmware before these error-prone flights using the controller a few days ago. So I followed the DJI Assistant recommendation and update the firmware.

Could the firmware conflict have caused all the problems I described above? The weather has not allowed me to do any more test flights.

Also, when I tried to download the flight records from the drone, I only found the last 8 flights or so. I was trying to loo at the data from one of the flights where I had one of these errors. Unfortunately, after that flight, i did a whole lot of very short test flights without errors and so the flight with the error was not among the 8 that I found. I was able to download the more limited data from my iPad. While i can identify that flight when I playback the flight on my iPad, I cannot correlate it to the download iPad data as that does not show the date-time of the flight. I tried looking at the data with airdata but it has been so long since I used that website, I thnk I am doing something wrong. I thought the airdata site would let you replay the flight, but I don't see an option to do that on their website.

I looked on my ipad and found that the first flight that I had the problem with was 11/21/20. I uploaded that flight to airdata and it shows multiple errors. Here is the link to that flight on airdata.

Good day from Melbourne Australia. You post is exactly what’s happening with my mavic pro 1.
ive just had the bird fixed after an accidental meeting with a tree ( not the aircrafts fault / my stupidity flying in too much wind )
But the guy who fixed it said he updated the firmware and he gave it a test flight and said all was ok , but I nearly had a fly off yesterday Compass error etc. all the same things you stated Took the rth button to get it back bc I’d lost all control over it.
i purchased the bird as a refurbished unit in sept and have done 140’ trouble free flights not one issue with compass etc

I rang the My repairer who suggested the compass possibly might have been damaged in the crash , but he asked me to test a set up with no props etc and I found imu error and severe compass error came up
took 3 attempts to get the compass flight ready Never had this issue before.

I believe the firmware has caused a major bug / glitch
Cheers. Grant

 

[slup]

Recently my Mavic Pro started behaving strangely. I was getting warnings of strong wireless interference, IMU calibration at start up most flights, compass calibration warnings and compass redundancy warnings. For example, an occasional strong wireless interference warning was coming when I was taking off from the middle of the runway at my club's RC field at a point that was over a quarter mile from the nearest power transmission wire.

Ever since a mysterious crash caused me to lose the MP about a year ago and I was unable to find it for months, I have been using a lot of items to help me find it, a Trackimo, a Tile, a strobe and florescent tape all over. I have been using all these for months with no problem. In these recent flights with the warnings, I had all that equipment in the same location of the MP as I have in months of problem free flights. However, I thought that one of these might be the issue so I removed all of them, left my iphone in my car, recalibrated the compass, and was still getting these errors frequently, though not every flight. Also had problems recalibrating the compass. Went through the procedure, got the OK from the lights on the drone, set it down and still had a compass error. Tried calibrating the compass several times and was unable to calibrate it. Weird.

I was in the process of downloading flight data from the drome using the DJI assistant and got an error message saying that there was an inconsistency in the firmware (or software) - don't remember the exact language). I had updated the firmware before these error-prone flights using the controller a few days ago. So I followed the DJI Assistant recommendation and update the firmware.

Could the firmware conflict have caused all the problems I described above? The weather has not allowed me to do any more test flights.

Also, when I tried to download the flight records from the drone, I only found the last 8 flights or so. I was trying to loo at the data from one of the flights where I had one of these errors. Unfortunately, after that flight, i did a whole lot of very short test flights without errors and so the flight with the error was not among the 8 that I found. I was able to download the more limited data from my iPad. While i can identify that flight when I playback the flight on my iPad, I cannot correlate it to the download iPad data as that does not show the date-time of the flight. I tried looking at the data with airdata but it has been so long since I used that website, I thnk I am doing something wrong. I thought the airdata site would let you replay the flight, but I don't see an option to do that on their website.

I looked on my ipad and found that the first flight that I had the problem with was 11/21/20. I uploaded that flight to airdata and it shows multiple errors. Here is the link to that flight on airdata.

That indeed looks like a yaw error due to a magnetic disturbed compass at AC power on ... in order to say for sure the .DAT log is needed.

If the AC DAT is overwritten ... you instead most probably still have the mobile device .DAT still in the device you flew with. Go here and read up on how to retrieve --> Mavic Flight Log Retrieval and Analysis Guide (read under section 3.) the correct one ends with FLY028.DAT. Attach it then in a new post here.

Then do this to play the .TXT log in Airdata.com ...

 

[slup]

Here is the flight record of the near fly away As I said I had no control whatsoever

If you really want to know what happened there my advice is to create your own thread to avoid confusion ... & post the logs to avoid speculation.

Logs is KING ... they usually don't speculate

 

[Jackcutrone]

Here is the dat file from the iPad for that flight.

 

[slup]

Here is the dat file from the iPad for that flight.

Mmm ... this isn't straight forward unfortunately.

Below is the flight from the mobile device DAT log ...

(Click on the pics & charts below to make them larger)




In above chart at point 1. you do a compass calibration, and below from the DAT log event stream we can see that it was a success ...


After the calibration you have around 4-5 degrees difference between the compass (magYaw) & imuYaw which is really good.

Then at point 2. above in the chart, very large pitch & roll movements are noted together with several Gyro errors in the DAT log event stream ...

You walked over to this place from the street ... did you do an IMU calibration there or what?



Then you walked back to the street & took off ... the disagreement between the compass (magYaw) & imuYaw remained then around +25 to -17 degrees until landing ... not enough to create a yaw error & within the normal.

But in the middle of the short flight the FC initiated ATTI mode at 16m height as seen in the chart at point 3. ... that's strange as you had around 15 satellites locked ... This is what's shown in the DAT log event stream there.


The ATTI mode period only lasts for 3sec then it's back to normal & at 53sec you initiate autolanding from the app ... but then again a lot of compass errors is shown in the DAT log event stream...


Just plain out from the sensor data, the only that shows a suspicious behavior is the gyro ... it doesn't follow the IMU or the compass as it should in some spots ...



Just before the ATTI mode the difference in GPS & IMU velocities also indicate that something isn't as it should ... can't really say why, but suspects IMU hickups though ...



@BudWalker ... can you pinpoint anything?

 

[BudWalker]

Mmm ... this isn't straight forward unfortunately.

Below is the flight from the mobile device DAT log ...

(Click on the pics & charts below to make them larger)


View attachment 119121

In above chart at point 1. you do a compass calibration, and below from the DAT log event stream we can see that it was a success ...

View attachment 119128
After the calibration you have around 4-5 degrees difference between the compass (magYaw) & imuYaw which is really good.

Then at point 2. above in the chart, very large pitch & roll movements are noted together with several Gyro errors in the DAT log event stream ...

You walked over to this place from the street ... did you do an IMU calibration there or what?

View attachment 119130

Then you walked back to the street & took off ... the disagreement between the compass (magYaw) & imuYaw remained then around +25 to -17 degrees until landing ... not enough to create a yaw error & within the normal.

But in the middle of the short flight the FC initiated ATTI mode at 16m height as seen in the chart at point 3. ... that's strange as you had around 15 satellites locked ... This is what's shown in the DAT log event stream there.

View attachment 119131
The ATTI mode period only lasts for 3sec then it's back to normal & at 53sec you initiate autolanding from the app ... but then again a lot of compass errors is shown in the DAT log event stream...

View attachment 119132
Just plain out from the sensor data, the only that shows a suspicious behavior is the gyro ... it doesn't follow the IMU or the compass as it should in some spots ...

View attachment 119133

Just before the ATTI mode the difference in GPS & IMU velocities also indicate that something isn't as it should ... can't really say why, but suspects IMU hickups though ...

View attachment 119134

@BudWalker ... can you pinpoint anything?

Yes, quite the puzzler. I've tried a couple of theories that didn't graduate to the level of being compelling.

@Jackcutrone what happened in the [-120.0, -9.0] secs interval? This is where the MP1 had been set down and then was picked up, carried to that box, and then returned to the launch point. I gotta ask was some external object added or removed from the MP1 in this interval. The magnetic properties of the MP1 may have changed, although it's hard to tell because the MP1 was returned to a slightly different location and orientation.

 

[Starz]

IMU calibration at start up most flights,

Any particular reason for this?

 

[Jackcutrone]

Reply to SLUP: That flight scared the hell out of me, and I think you can see why. I don't remember any wild stick movements. I do see some stick movements replaying the flight on Airdata toward the beginning of the flight but they don't seem to be that radical to me but then what do I know. Nothing stands out about the fight to me except for the MP going nuts.

 

[Jackcutrone]

Any particular reason for this?

Reply to STARZ: I haven't a clue. I did not initiate the IMU calibration. The aircraft did that on its own. As a rule, I never do an IMU calibration just compass calibration when flying any significant distance from the last flight.

 

[Starz]

I did not initiate the IMU calibration. The aircraft did that on its own.

I’m thinking that you are seeing a message that references an
“imu initializing”. DJI aircraft do not calibrate imu’s on their own.
A possibility?

 

[Jackcutrone]

Yes, quite the puzzler. I've tried a couple of theories that didn't graduate to the level of being compelling.

@Jackcutrone what happened in the [-120.0, -9.0] secs interval? This is where the MP1 had been set down and then was picked up, carried to that box, and then returned to the launch point. I gotta ask was some external object added or removed from the MP1 in this interval. The magnetic properties of the MP1 may have changed, although it's hard to tell because the MP1 was returned to a slightly different location and orientation.

Reply to Budwalker: If I remember correctly (which is frequently not the case) I started up the MP in one location but then moved it since I wanted to start from level ground, but then realized that the electric box shown in the picture, might cause interference issues, so I moved it to another place before I took off. I did not remove any of the devices I use for tracking, the Trackimo, the TIle or the strobe at any point from startup to end of flight.

To all of you kind people who are trying to help, the reason I uploaded the dat file from this flight rather than the other ones I was talking about, is that this is the first flight where the MP started behaving weirdly.

I am also very puzzled about something else. From the TXT files on the iPad, I can identify two flight from a few days ago that I would like to analyze. One shows the compass interference in the middle of the flying field - at least a quarter mile form any electric transmission lines and with no Trackimo, Tile or strobe attached. The other shows multiple yaw and compass errors. The problem is that I cannot locate the DAT files from the flight either on the aircraft or on the iPad.

As far as the aircraft, after the problem flights, I did multiple very short test flights not realizing that the aircraft will only store so many DAT files before it starts overwriting them. The problem flights were all overwritten by the later short test flights. And the alter short flights were fine. I did not experience any problems at all on those.

I am puzzled about the iPad. It has the TXT files from all the flights that day, but only has some of the DAT files. The dat files that it has are not consecutive, nor are the the last flights (as would be the case if earlier ones were just overwritten.

The DAT files are from flight 038 at 12:43; 037 at 12:57; 038 at 12:58; and 039 at 1:15. The flights where the errors occurred are flight 032 at 12:27 and 033 at 12:28. Why don't I have DAT files for these flights. Lacking the DAT files I am posting the TXT files form those two flights, although I realize they do not contain as much detail but hopefully one of you very smart people can take a look and see if you find any reason for the errors.

It so frustrating when you have intermittent errors and the records that would help you solve them are gone.

One other piece of information that might affect this. One Dec. 10, 2020, at about the time I was flying, we were having strong solar flare activity and coronal mass ejection, the "shock wave" having hit the night before. I initially discounted that as the cause since I had good satellite coverage but the more I think about it, a geomagnetic storm might be the culprit for December 10th though that would not help explain what happened on the November 21st flight, so maybe the solar flares have nothing to do with my problems


Anyway, thanks for all the time and effort.

Mystified, Jack Cutrone

 

[Jackcutrone]

I’m thinking that you are seeing a message that references an
“imu initializing”. DJI aircraft do not calibrate imu’s on their own.
A possibility?

Yes, I did see a message that said IMU initiallizing on several start ups and did not do them myself.

 

[Jackcutrone]

Thanks for responding to my messages guys.

SLUP explained why I thought I was not seeing DAT files and those for the error-laden flights. Now I know that they are there in the first DAT file of the day. I am attaching that file in hopes that someone can examine it and find the source of the the wireless interference error message and the yaw and compass errors. Here is that DAT file.

As was suggested to me, I will perform the IMU calibration before the next flight. What would cause the aircraft to need a recalibration of the IMU? Did I do something to it?

 

[Jackcutrone]

On further review, one of the error flights might be in this DAT file

 

[BudWalker]

Thanks for responding to my messages guys.

SLUP explained why I thought I was not seeing DAT files and those for the error-laden flights. Now I know that they are there in the first DAT file of the day. I am attaching that file in hopes that someone can examine it and find the source of the the wireless interference error message and the yaw and compass errors. Here is that DAT file.

As was suggested to me, I will perform the IMU calibration before the next flight. What would cause the aircraft to need a recalibration of the IMU? Did I do something to it?

According to the FC there were 2 fly aways in this .DAT. From the eventLog stream
251.845 : 14850 [L-FDI][CTRL]: fault on , fly_away
279.845 : 16250 [L-FDI][CTRL]: fault off, fly_away

837.845 : 44150 [L-FDI][CTRL]: fault on , fly_away
863.845 : 45450 [L-FDI][CTRL]: fault off, fly_away

During the second fly away there was some major league toilet bowling


There was also some suspicious behavior with the magnetometer data.

What we really need at this point is one test flight where 1) no external equipment is attached 2) takes place at a location with no possibility of geomagnetic distortion, and 3) a compass calibration is done before flight. So far what we have are multiple .DATs each recorded under indeterminate conditions.

 

[slup]

Agree with @BudWalker ... a proper benchmark with known parameters is needed.

Then ... I don't like the look of the gyro here in the DAT with 3 flights ... the yaw-magyaw is reasonably low but the gyro do weird stuff & doesn't seem to agree. First circle AC just sitting still on ground, the second is in gohome state.

 

[BudWalker]

Agree with @BudWalker ... a proper benchmark with known parameters is needed.

Then ... I don't like the look of the gyro here in the DAT with 3 flights ... the yaw-magyaw is reasonably low but the gyro do weird stuff & doesn't seem to agree. First circle AC just sitting still on ground, the second is in gohome state.

View attachment 119332

The first interval is just showing the Z axis gyro error that integrates to a drift rate of 0.28°/sec. Seems a little high but 0.2°/sec is not that uncommon.

TotalGyro:Z is 100% consistent with yaw only if pitch and roll are 0°. Usually, pitch and roll are small enough that totalGyro:Z will be close to yaw and comparisons between these are useful. In the second interval the MP is oscillating in both pitch and roll ±25° which causes totalGyro:Z to be incorrect.

As an example if the MP is pitched -25° then Z axis isn't vertical and will detect any roll that takes place.

 

[Jackcutrone]

According to the FC there were 2 fly aways in this .DAT. From the eventLog stream
251.845 : 14850 [L-FDI][CTRL]: fault on , fly_away
279.845 : 16250 [L-FDI][CTRL]: fault off, fly_away

837.845 : 44150 [L-FDI][CTRL]: fault on , fly_away
863.845 : 45450 [L-FDI][CTRL]: fault off, fly_away

During the second fly away there was some major league toilet bowling
View attachment 119330

There was also some suspicious behavior with the magnetometer data.

What we really need at this point is one test flight where 1) no external equipment is attached 2) takes place at a location with no possibility of geomagnetic distortion, and 3) a compass calibration is done before flight. So far what we have are multiple .DATs each recorded under indeterminate conditions.

Actually all the flights I posted were done without external equipment, at a location with no possibility of geomagnetic distortion and just after I had calibrated the compass prior to the first flight. The first flight that is from take-off to landing took 5 minutes, 39 seconds and is the first flight shown in the flight 35 dat file I posted earlier.

At this point I am going to take your advice and calibrate the IMU. Then I will go back to the same location which is in the middle of a field and will do a compass calibration. I will then do several tests flights but will shut everything down in between so that the DAT files will all be a discrete flight. Then I will post one or more of those DAT files so you guys can take a look.

Checking the weather for tomorrow, although it will be cold, there will be light winds so I will be able to get in some flights and post those.

Again, I am so appreciative for the help and the time and efforts you all are putting in to this.

Jack

 

[Jackcutrone]

BTW - a couple of questions about IMU recalibration. I watched the DJI video and understand, as you guys said, that I must do it on a level surface (via spirit level measurements) and that colder, up to a point, is better. Also I read a number of posts on calibration, some of which said to ensure it is not near metal surfaces when I do it.

I plan on doing it tomorrow and then do some test flights. I was going to do it in my garage and level a piece of plywood on some plastic sawhorses. The outside temperature will be about 31F, 0C, and it will be a little warmer in the garage.

The only metal there would be the track for garage door opener and some metal shelving, both of which will be at least 6 feet away from the drone. I did not notice anything in the DJI video tutorial about metal and was curious as to whether it would really affect the IMU. Given that the IMU has accelerometers, a gyro, a barometer and a thermometer, it didn't seem like any of those would much be affected by metal nearby. But I wondered if that is correct, and a six foot distance form any metal is okay?

Thoughts?