OK fans06b77ccc, I took another look at your RC video and I, too, am puzzled  by why the AC orientation shown on the map is not the direction the drone is pointing as evident by the camera video recording.  In my opinion this is definitely a compass miscalibration or malfunction. Other opinions?

Griffith Posted at 2017-4-24 11:38
OK fans06b77ccc, I took another look at your RC video and I, too, am puzzled  by why the AC orientation shown on the map is not the direction the drone is pointing as evident by the camera video recording.  In my opinion this is definitely a compass miscalibration or malfunction. Other opinions?

Looking at everything here now I'm also thinking this might be a compass problem, I would like to see what warning op got just before takeoff , but the way the drone took off in that circling motion has all the signs of weird compass.

fans06b77ccc, you weren't, by chance, in "Fixed Wing" mode, were you?  Because it looks like you were still moving forward without forward stick.

SkunkWerxs Posted at 2017-4-24 10:47
DroneFlying,  When I go out flying I use the tailgate of my pickup for a bench and every time I power up the RC and Mavic I get the compass error warning, as I move just the Mavic away from the truck it clears and I can fly while sitting on the tailgate even when resting the controller on it while flying I still don't get any warnings. I'm starting to wonder what would happen if I took off while getting the compass error ???
Now just to add a twist and some food for thought -- I have a large metal wielding table outside, I used this once as a bench when I power up I quickly receive compass error and strong magnetic interference warnings so I moved both the RC and AC about 20ft and then some more from it after the warnings cleared . This has happen to me when I first received my Mavic -- I have used the table as a bench in the past for other RC plane's and never had any interference, then again I never owned a system as sophisticated as the Mavic's , I do own RC equipment that is more expensive.
            Your analogy of a small change in position on the bench the op was flying at can be enough to trigger a compass error is definitely a #1 possibility it could have also effected the Controller too an caused the issue -- The Data is needed to be analyzed by DJI to know for sure what happen.

I'm starting to wonder what would happen if I took off while getting the compass error ???

I think you'd probably see behavior like what the OP recorded in his video. I had this happen to me once when I failed to notice the error before takeoff and for a few seconds the drone had a mind of its own -- but then it recovered and flew normally.

Just a Thought. They say April 23nd and 24th..   http://www.amtvmedia.com/solar-f ... -blackouts-anytime/

Midwest_Blue Posted at 2017-4-24 15:17
Just a Thought. They say April 23nd and 24th..   http://www.amtvmedia.com/solar-f ... -blackouts-anytime/

Midwest, that is true but how would you explain this op's thread he just posted ????
                  GPS Drops out

Griffith Posted at 2017-4-24 12:32
fans06b77ccc, you weren't, by chance, in "Fixed Wing" mode, were you?  Because it looks like you were still moving forward without forward stick.

No, I was in a regular GPS mode.

Oleh Artyukh Posted at 2017-4-24 23:51
No, I was in a regular GPS mode.

Is there a reason you won't tell us about warning just before flight.

OK, I've extracted logs. Attaching link to them.

Here is processed flight record file: https://app.airdata.com/main?fli ... 318d685166f11070020
Here is more information and you can see that the compass error happened during 14th second of the flight: http://www.phantomhelp.com/LogViewer/NJ35KCBRIUXVI1WI4RK9/
And log messages where we can see also some control errors: https://www.dropbox.com/s/5rt6b0 ... ogMessages.log?dl=0

A lot of info to "eat"

CelticWarrior Posted at 2017-4-25 00:37
Is there a reason you won't tell us about warning just before flight.

After starting the drone it was ok. When I took off, I tried to control the drone but couldn't, then I looked at the screen and saw compass error.

Oleh Artyukh Posted at 2017-4-25 00:50
After starting the drone it was ok. When I took off, I tried to control the drone but couldn't, then I looked at the screen and saw compass error.

In all fairness you only got compass warning right at the end of your flight, which is normal as you crashed .
It doesn't mean it wasn't compass problem and I guess dji will have to work that one out, where you took off from wasn't the best place and that steel tower in your video may have been a problem, I hope dji look at this kindly and sort you out.
Good luck.

CelticWarrior Posted at 2017-4-25 02:33
In all fairness you only got compass warning right at the end of your flight, which is normal as you crashed .
It doesn't mean it wasn't compass problem and I guess dji will have to work that one out, where you took off from wasn't the best place and that steel tower in your video may have been a problem, I hope dji look at this kindly and sort you out.
Good luck.

Metal tower is water tower and I flew 2 meters above it before.
According to this: http://www.phantomhelp.com/LogViewer/NJ35KCBRIUXVI1WI4RK9/ , Compass Error occured in 0.9 seconds after takeoff. And restored 2 seconds before the tree, but it was too late...

SkunkWerxs Posted at 2017-4-23 18:44
Fan's,
        I would upload the flight data to DJI let them analyze it --- I know for a fact that a compass error would not cause loss of stick control --- I know this cause I was flying on a bridge with metal guard rails on both sides and when I came below the height and in-between the rails I was getting the compass errors an it did not affect my stick inputs -- I believe the compass and GPS in the Mavic are used for stabilization and direction for the GO4 app and for autonomous flight modes .
    Anybody please feel free to correct me if I'm wrong ?

You are right. The compass is mainly used to maintain the heading. As soon compass and GPS are lost, the Flight computer relies solely on the accelerometers to try to keep the drone stable (locked on the horizon) deeming it more difficult to control but for sure fully controllable. That's why is always advisable to build confidence with a cheap-o quad that only has accelerometers forcing the person to use the sticks and constantly keep correcting it to keep it as stable as possible. But only a compass error wouldn't explain the lost of stick control.

For sure it will be interesting to see the flight logs.

AACY Posted at 2017-4-25 03:03
You are right. The compass is mainly used to maintain the heading. As soon compass and GPS are lost, the Flight computer relies solely on the accelerometers to try to keep the drone stable (locked on the horizon) deeming it more difficult to control but for sure fully controllable. That's why is always advisable to build confidence with a cheap-o quad that only has accelerometers forcing the person to use the sticks and constantly keep correcting it to keep it as stable as possible. But only a compass error wouldn't explain the lost of stick control.

For sure it will be interesting to see the flight logs.

Here you are ;)
http://www.phantomhelp.com/LogViewer/NJ35KCBRIUXVI1WI4RK9/
https://app.airdata.com/main?fli ... 318d685166f11070020
https://www.dropbox.com/s/5rt6b0 ... ogMessages.log?dl=0

Oleh Artyukh Posted at 2017-4-25 03:35
Here you are ;)
http://www.phantomhelp.com/LogViewer/NJ35KCBRIUXVI1WI4RK9/
https://app.airdata.com/main?flight=2a4fe5f651c51318d685166f11070020

Can you also provide the DAT file from the crash? The one you linked to is a log file but there will be one that ends with .DAT (as mentioned here here that will be more helpful.

DroneFlying Posted at 2017-4-25 03:40
Can you also provide the DAT file from the crash? The one you linked to is a log file but there will be one that ends with .DAT (as mentioned here here that will be more helpful.

Sure
https://www.dropbox.com/s/ddnki2ajybn62a3/FLY088.DAT?dl=0

DroneFlying Posted at 2017-4-25 03:40
Can you also provide the DAT file from the crash? The one you linked to is a log file but there will be one that ends with .DAT (as mentioned here here that will be more helpful.

I would be interested to know what you read in the .dat file, I hear some requesting them but I haven't seen anyone here arrive at a conclusion, it would be interesting to know how they are broken down, I have tried for myself but found it difficult to get exact information but I'm sure it's that I'm not reading them correctly.
Would be interesting in this case to see if more information can be extracted.

Oleh Artyukh Posted at 2017-4-24 02:31
Because I moved to the new place 150 km from previous flight.

I reviewed the .txt log file for this flight. It's almost certain that the cause of your incident is that you launched from a site that was geomagnetically distorted. In the video it looks like the AC was sitting on a bench with steel posts under it. I'm betting that the AC was directly above one of those posts.

The log shows the Mavic with a heading of 164 at launch but it was actually closer to 0 as shown by the video. Also, direction of flight was close to 0 but with no aileron or rudder input. These are clear signs that the compass was compromised by the geomagnetic distortion. After reaching a height of 1 meter the Mavic was clear of the geomagnetic distortion and the compass then had the correct heading. However, the IMU still has the incorrect value. It's at this point the compass error is declared.

It's almost certain this is what happened. If you want to be 100% sure you will need to retrieve the .DAT file from the Mavic. To see how to do this go to http://www.datfile.net/DatCon/retrieveV3.html
It will be large so you will need to Dropbox it and post a link.

There is nothing wrong with the compass. If the Mavic had been launched from between two of the posts there may not have been a problem.

BudWalker Posted at 2017-4-25 04:08
I reviewed the .txt log file for this flight. It's almost certain that the cause of your incident is that you launched from a site that was geomagnetically distorted. In the video it looks like the AC was sitting on a bench with steel posts under it. I'm betting that the AC was directly above one of those posts.

The log shows the Mavic with a heading of 164 at launch but it was actually closer to 0 as shown by the video. Also, direction of flight was close to 0 but with no aileron or rudder input. These are clear signs that the compass was compromised by the geomagnetic distortion. After reaching a height of 1 meter the Mavic was clear of the geomagnetic distortion and the compass then had the correct heading. However, the IMU still has the incorrect value. It's at this point the compass error is declared.

Here is DAT file: https://www.dropbox.com/s/ddnki2ajybn62a3/FLY088.DAT?dl=0

Oleh Artyukh Posted at 2017-4-25 04:12
Here is DAT file: https://www.dropbox.com/s/ddnki2ajybn62a3/FLY088.DAT?dl=0

Yes, I see you posted the .DAT while I was posting. I reviewed the .DAT and what I posted is exactly what happened. When I can figure out how to attach images to a post here I'll provide more info.

BudWalker Posted at 2017-4-25 04:17
Yes, I see you posted the .DAT while I was posting. I reviewed the .DAT and what I posted is exactly what happened. When I can figure out how to attach images to a post here I'll provide more info.

See attachment about posting pcitures in this forum.

Oleh Artyukh Posted at 2017-4-25 04:12
Here is DAT file: https://www.dropbox.com/s/ddnki2ajybn62a3/FLY088.DAT?dl=0

Look at the attached image
https://www.dropbox.com/s/wr2yuisla0rdrvm/oleg.png?dl=0
Yaw is the heading determined by the Flight Controller. Shortly after batteryOn Yaw is set to the value provided by the compass. After that Yaw is determined mostly by the IMU gyros and accelerometers). magYaw is computed by DatCon (the software used to analyze the .DAT). magYaw is computed from the magnetometers (not the IMU). Before launch magYaw and Yaw are both 164 which is incorrect. When the Mavic is launched and altitude is gained magYaw becomes correct because the geomagnetic distortion no longer affects the magnetometers. BUT, yaw remains at the old incorrect value of 164. Finally, at time 14.382 the Flight Controller declares a compass errror. The FC doesn't know if the magnetometers or the IMU is correct; only that they are different.

Oleh Artyukh Posted at 2017-4-25 04:32
See attachment about posting pcitures in this forum.

Here is the image?

CelticWarrior Posted at 2017-4-25 04:06
I would be interested to know what you read in the .dat file, I hear some requesting them but I haven't seen anyone here arrive at a conclusion, it would be interesting to know how they are broken down, I have tried for myself but found it difficult to get exact information but I'm sure it's that I'm not reading them correctly.
Would be interesting in this case to see if more information can be extracted.

I'm definitely going to defer to whatever BudWalker says, as he's one of the most -- in fact, probably the most -- knowledgeable people around concerning DAT file analysis. Normally he only does this over on Thunderdome but today we got a rare treat. Thanks, Budwalker.

BudWalker Posted at 2017-4-25 04:34
Look at the attached image
https://www.dropbox.com/s/wr2yuisla0rdrvm/oleg.png?dl=0
Yaw is the heading determined by the Flight Controller. Shortly after batteryOn Yaw is set to the value provided by the compass. After that Yaw is determined mostly by the IMU gyros and accelerometers). magYaw is computed by DatCon (the software used to analyze the .DAT). magYaw is computed from the magnetometers (not the IMU). Before launch magYaw and Yaw are both 164 which is incorrect. When the Mavic is launched and altitude is gained magYaw becomes correct because the geomagnetic distortion no longer affects the magnetometers. BUT, yaw remains at the old incorrect value of 164. Finally, at time 14.382 the Flight Controller declares a compass errror. The FC doesn't know if the magnetometers or the IMU is correct; only that they are different.

This makes sense but how I as a pilot can know if there is an interference if I don't see this warning?
I know how this warning looks like because I saw it when first time turned it on in my apartment (of course without flying ) because there are Wi-Fi's etc.
The drone flew forward immediately after taking off and continued flying on its own "round" direction without moving right stick and even when I changed heading.
You can see my first analysis here:

What tool did you use for your analysis of the DAT?

BudWalker Posted at 2017-4-25 04:34
Look at the attached image
https://www.dropbox.com/s/wr2yuisla0rdrvm/oleg.png?dl=0
Yaw is the heading determined by the Flight Controller. Shortly after batteryOn Yaw is set to the value provided by the compass. After that Yaw is determined mostly by the IMU gyros and accelerometers). magYaw is computed by DatCon (the software used to analyze the .DAT). magYaw is computed from the magnetometers (not the IMU). Before launch magYaw and Yaw are both 164 which is incorrect. When the Mavic is launched and altitude is gained magYaw becomes correct because the geomagnetic distortion no longer affects the magnetometers. BUT, yaw remains at the old incorrect value of 164. Finally, at time 14.382 the Flight Controller declares a compass errror. The FC doesn't know if the magnetometers or the IMU is correct; only that they are different.

BudWalker, Thanks for your excellent analysis and explaination

BudWalker, I also want to thank you

Oleh Artyukh Posted at 2017-4-25 04:50
This makes sense but how I as a pilot can know if there is an interference if I don't see this warning?
I know how this warning looks like because I saw it when first time turned it on in my apartment (of course without flying  ) because there are Wi-Fi's etc.
The drone flew forward immediately after taking off and continued flying on its own "round" direction without moving right stick and even when I changed heading.

The geomagnetic field is distorted. A change in direction and/or strength. The Mavic can only know about a change in strength; it can't know if the direction is incorrect. Often, in these situations, the field strength hasn't changed enough for the Mavic to think there is an error.

The only way to determine the direction is incorrect is to look at the red triangle in the Go App dislplay and see if it's pointing in the right direction. In your incident the triangle would have been pointing 164 degrees but looking at the Mavic you would see that it's pointing north.

The software I used was CsvView/DatCon which can be obtained by going to http://www.datfile.net/CsvView/downloads.html

Maybe you know the answer to a question I never could figure out. How do I get this forum to notify me if someone replies to a post I made? Is it possible?

BudWalker Posted at 2017-4-25 05:26
The geomagnetic field is distorted. A change in direction and/or strength. The Mavic can only know about a change in strength; it can't know if the direction is incorrect. Often, in these situations, the field strength hasn't changed enough for the Mavic to think there is an error.

The only way to determine the direction is incorrect is to look at the red triangle in the Go App dislplay and see if it's pointing in the right direction. In your incident the triangle would have been pointing 164 degrees but looking at the Mavic you would see that it's pointing north.

How do I get this forum to notify me if someone replies to a post I made? Is it possible?

Unfortunately, no, at least not if you mean an email notification. The only thing it provides is an icon below the bell (?) in the upper right corner of the web site pages to indicate that someone replied to your post -- and that's only if they replied directly to one of yours and not just added another message to the thread.

AACY Posted at 2017-4-25 03:03
You are right. The compass is mainly used to maintain the heading. As soon compass and GPS are lost, the Flight computer relies solely on the accelerometers to try to keep the drone stable (locked on the horizon) deeming it more difficult to control but for sure fully controllable. That's why is always advisable to build confidence with a cheap-o quad that only has accelerometers forcing the person to use the sticks and constantly keep correcting it to keep it as stable as possible. But only a compass error wouldn't explain the lost of stick control.

For sure it will be interesting to see the flight logs.

AACY, For sure it will be interesting to see the flight logs.
                  I would love to see them too!!!!
Flying a Multi Rotor drone without gyro's for stability is next to impossible you would have a better chance of getting blood out of a rock .
Flysafe-FlyFree

SkunkWerxs Posted at 2017-4-25 15:08
AACY, For sure it will be interesting to see the flight logs.
                  I would love to see them too!!!!
Flying a Multi Rotor drone without gyro's for stability is next to impossible you would have a better chance of getting blood out of a rock .

Indeed!!!

BudWalker Posted at 2017-4-25 05:26
The geomagnetic field is distorted. A change in direction and/or strength. The Mavic can only know about a change in strength; it can't know if the direction is incorrect. Often, in these situations, the field strength hasn't changed enough for the Mavic to think there is an error.

The only way to determine the direction is incorrect is to look at the red triangle in the Go App dislplay and see if it's pointing in the right direction. In your incident the triangle would have been pointing 164 degrees but looking at the Mavic you would see that it's pointing north.

Thanks for the analysis, I also agree with what you had posted regarding the compass error.

What still puzzles me is. Should not the Flight Computer immediately roll back to axis lock (using gyros) as soon the compass info was determined as being at fault?

The OP said that he was giving it stick inputs but not responding to any of them. Were you able to notice anything strange. I know that in the heat of the moment it can be very disorienting to loose yaw, but a non-response to stick inputs tends to be very obvious. OP You posted a video with the stick movements and the camera record on the side, thanks. I currently see that the Mavic was still responding to your inputs even though that the AC was drifting.  

SkunkWerxs Posted at 2017-4-25 15:08
AACY, For sure it will be interesting to see the flight logs.
                  I would love to see them too!!!!
Flying a Multi Rotor drone without gyro's for stability is next to impossible you would have a better chance of getting blood out of a rock .

Indeed!!! Flying without accelerometers (gyros) would be almost impossible.

AACY Posted at 2017-4-25 22:41
Thanks for the analysis, I also agree with what you had posted regarding the compass error.

What still puzzles me is. Should not the Flight Computer immediately roll back to axis lock (using gyros) as soon the compass info was determined as being at fault?

I don't say Mavic didn't reacted to all inputs. It did reacted to the left stick, but didn't to the right one, and you can definitely see it in the video:


BTW, I found another two videos with very similar behavior:

And here you can see the the pilot moved right stick and the drone didn't respond while still responding to the left stick:

Oleh Artyukh Posted at 2017-4-25 23:31
I don't say Mavic didn't reacted to all inputs. It did reacted to the left stick, but didn't to the right one, and you can definitely see it in the video:
https://www.youtube.com/watch?v=pYR6kD7FrKM

Great, thanks for the clarification!

Maybe @BudWalker can have an input on this.

I brought it up, because I think is probably worth having a look at it too.

AACY Posted at 2017-4-25 23:36
Great, thanks for the clarification!

Maybe @BudWalker can have an input on this.

He already looked at it end explained what could happen. See above in this thread.

Oleh Artyukh Posted at 2017-4-25 23:42
He already looked at it end explained what could happen. See above in this thread.

You refer to this ?

"The log shows the Mavic with a heading of 164 at launch but it was actually closer to 0 as shown by the video. Also, direction of flight was close to 0 but with no aileron or rudder input. These are clear signs that the compass was compromised by the geomagnetic distortion. After reaching a height of 1 meter the Mavic was clear of the geomagnetic distortion and the compass then had the correct heading. However, the IMU still has the incorrect value. It's at this point the compass error is declared."

But this doesn't explained why after the compass error was declared, the Flight Computer didn't go into axis (yaw) lock. You did give some rudder input, but it was still going bananas on the yaw axis.

On the other hand, had DJI answered back to you on this?

AACY Posted at 2017-4-25 23:52
You refer to this ?

"The log shows the Mavic with a heading of 164 at launch but it was actually closer to 0 as shown by the video. Also, direction of flight was close to 0 but with no aileron or rudder input. These are clear signs that the compass was compromised by the geomagnetic distortion. After reaching a height of 1 meter the Mavic was clear of the geomagnetic distortion and the compass then had the correct heading. However, the IMU still has the incorrect value. It's at this point the compass error is declared."

I created request with all explanations and logs 11 hours ago, they have up to 48 hours to respond . Waiting.

Oleh Artyukh Posted at 2017-4-25 23:56
I created request with all explanations and logs 11 hours ago, they have up to 48 hours to respond . Waiting.

My best wishes and fingers are crossed for you mate!

SkunkWerxs Posted at 2017-4-25 15:08
AACY, For sure it will be interesting to see the flight logs.
                  I would love to see them too!!!!
Flying a Multi Rotor drone without gyro's for stability is next to impossible you would have a better chance of getting blood out of a rock .

Hi SkunkWerxs,

I read your previous discussions on this thread about the magnetic compass and thought I could offer a bit of a clarification on the matter.

Compass correction consists of two elements: magnetic variation (also called declination) and magnetic deviation. Variation is caused by the irregular position of the magnetic poles and the this changes over time in space. Deviation is caused by external magnetic influence in the vicinity of the compass - so steel for example. Now - when you calibrate the compass you are actually giving the unit the info about the deviation of the compass and that's why the manual states you should carry this out in an area free of magnetic influence (including on your own body - so phone, keys, coins etc). The deviation changes all around the 360 degrees plane of calibration. Due to the fact that the drone will not remain always parallel to the ground this is why you need to rotate it in that position.

Now comes something I do not know - which is how the aircraft determines its course of action. From my point of view - the declination is totally unimportant to the craft because you are not telling it to fly a course 045deg true - you only tell it to fly "forward" or "right" or whatever. For this the deviation needs to be known, the variation can be anything.

As I understand it from the log analysis - the craft recorded the compass heading in the IMU and right after take off - when the metal legs of the bench were not there anymore - the one recorded in memory was much different than the one the compass was reporting. I'm guessing that both compasses were so influenced and that's why the error popped up.

So at this moment you have a craft which doesn't really know in which direction it is headed. That's bad. But this doesn't explain this behavior. If it doesn't know the heading and you try to give it "forward" motion it should - in my opinion - beep and alarm and say "can't do, no proper compass". It should not start flying off to the side. Especially since the gyros must be telling it "you are motionless" and the GPS is telling it more or less the same. This is the moment where the pilot should receive a warning about the compass and have the option of controlling the craft only with the left stick to bring it down. He can't control the right stick because the craft doesn't know which motors to rev up and which motors to rev down.

Just my few cents to this discussion. I'm keeping my fingers crossed for Oleh to get this sorted out properly with DJI - wishing a lot of luck with this!

Cheers,

Jakub