Satellite signal lost, erratic flying, almost lost Mavic!

[daithi23]

Finally got a response from DJI:
"As to conclusion from analysis team, they gave a conclusion that it's an environment issue. They also ask you to change to different locations to do some more test, as no abnormal evidence founded in DAT files. Please kindly note"

So, fly it a bit more and hope that I can recover it again if there's an attempted flyaway?

 

[grizzard]

If it does fly away then it appears it would be covered since it was at their instruction.

 

[daithi23]

OK, not sure if this is related or should I start another thread but I had seen a bit of drifting before this and had put it down to wind or a calibration needed. Today I made sure I was up to date on all the firmware, calibrated compass and IMU and recreated same thing as yesterday...

I flew up, turned around approximately 180 degrees and fly higher. On the YouTube video you can see it yawing to the right but I wasn't touching the yaw at this point...

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Here's a plot of the Yaw and magYaw along with the relativeHeight so it can be compared with the YouTube video. Tried to add Control:Rudder but it said "empty not plotting".


I saw this almost exact behaviour yesterday but lots of similar drifting the past few weeks, any idea if it could be related to previous fly away? @Mavicsailor ?
Uploaded the DAT here: DJI_ASSISTANT_EXPORT_FILE_2017-03-15_18-07-28.DAT

 

[BudWalker]

OK, not sure if this is related or should I start another thread but I had seen a bit of drifting before this and had put it down to wind or a calibration needed. Today I made sure I was up to date on all the firmware, calibrated compass and IMU and recreated same thing as yesterday...

I flew up, turned around approximately 180 degrees and fly higher. On the YouTube video you can see it yawing to the right but I wasn't touching the yaw at this point...

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Here's a plot of the Yaw and magYaw along with the relativeHeight so it can be compared with the YouTube video. Tried to add Control:Rudder but it said "empty not plotting".
View attachment 8555

I saw this almost exact behaviour yesterday but lots of similar drifting the past few weeks, any idea if it could be related to previous fly away? @Mavicsailor ?
Uploaded the DAT here: DJI_ASSISTANT_EXPORT_FILE_2017-03-15_18-07-28.DAT

This is wonderful. Thank you.
This behavior is "normal". At 22.5115 secs the AC had turned through about 180 degrees. For whatever reason the FC's Yaw value accumulated about 25 degrees error in this turn. The "drifting" is the FC slowly correcting by turning the AC so that the actual heading becomes what the FC thinks the heading is.

I've seen this in other incidents.

With the Phantom 3 the response to a Yaw/magYaw separation is to declare a compass error, switch to ATTI and hope things don't get nasty as the pilot is then being relied upon. The Mavic instead attempts to fix this kind of error.

What's great about this post is that there is a video that verifies that magYaw (computed by DatCon from magnetometer data) is giving the correct heading while Yaw (coming from the FC) is not.
@ScrappyMavic @Cyberpower678 @Robbyg

 

[Andrew F]

So return it for repair as it appears you have one bad sensor. That's all it takes for anomalous behavior like this. But I do have to say I've seen this with mine and then had it resolve on its own. But I've always noted it shortly after take off and I land immediately. If I first let the Mavic "settle" and it is rock solid before flight, it has always stayed true. Knock on wood.

 

[BudWalker]

I can't resist one more post on this. At issue is the question "Is it Yaw or magYaw that has the correct value". In the interval 22 - 68 secs Yaw is constant and magYaw is changing.

There is also the value totalGyroZ that verifies that magYaw is correct. TotalGyroZ is computed by DatCon not the Flight Controller. For each gyroZ data point an integration is done over the interval since the last data point. These integrated values are then summed to obtain totalGyroZ. totalGyroZ yields heading info independent of the Flight Controller since it is dependent only on gyroZ.

totalGyroZ was implemented to measure gyroZ error or the "drift rate" over time. When the AC is sitting stationary prior to launch any change in totalGyroZ will be the "drift rate". In this case "drift rate" is -0.076 degrees/second. In the above the totalGyroZ change was +0.26 degrees/second so it can be safely assumed that the gyroZ error was not a factor.
@ScrappyMavic @Cyberpower678 @Robbyg

 

[Robbyg]

Interesting Find. I have to take a better look at this at home and see the amount of rotation in the video versus what is happening in the graph.

Rob

 

[SteveM]

I had the same experience. I had noticed that the drone had not been flying straight the day before. So I did a full firmware, IMU, compass and controller calibration to try and resolve. On the following flight the "non-straight flying" continued but everything else seemed the same. All looked good, on locked on to 18 satellites. But then when I had flown out 250m and turned to come back I suddenly got "Weak GPS" & "IMU Heading exception Warning" and went into ATTI mode. After a few seconds, it seemed to regain GPS. But then when coming into land 1 min later. The same thing happened. It was over a clear field in line of sight at 20-40m high. Looking at flight playback, looks like full throttle for most of flight and full throttle and full turn for a second or so just before 1st loss of signal.
I have flown a Phantom for about a year, but still really a learner. I have just sent it straight back to DJI as I don't think it should do this or is safe. It feels exactly like many of the other recorded events on the web. I will let you know what DJI say when they have analysed / fixed my drone.

 

[Andrew F]

I had the same experience. I had noticed that the drone had not been flying straight the day before. So I did a full firmware, IMU, compass and controller calibration to try and resolve. On the following flight the "non-straight flying" continued but everything else seemed the same. All looked good, on locked on to 18 satellites. But then when I had flown out 250m and turned to come back I suddenly got "Weak GPS" & "IMU Heading exception Warning" and went into ATTI mode. After a few seconds, it seemed to regain GPS. But then when coming into land 1 min later. The same thing happened. It was over a clear field in line of sight at 20-40m high. Looking at flight playback, looks like full throttle for most of flight and full throttle and full turn for a second or so just before 1st loss of signal.
I have flown a Phantom for about a year, but still really a learner. I have just sent it straight back to DJI as I don't think it should do this or is safe. It feels exactly like many of the other recorded events on the web. I will let you know what DJI say when they have analysed / fixed my drone.

Sorry for your experience. Fortunately, it ended well. The problem with these Mavics: the more features the more potential problems. There are so many "break points." I'm gonna keep my fingers crossed that I don't get one when I'm "out" because most of my flying is over the Pacific. Hard to recover out there. Haha. Let us know how it turns out. My guess is DJI will replace with a refurbished unit and won't tell you the cause. Upload your flight file and folks on here can analyze it amazingly well.

 

[SteveM]

Sorry for your experience. Fortunately, it ended well. The problem with these Mavics: the more features the more potential problems. There are so many "break points." I'm gonna keep my fingers crossed that I don't get one when I'm "out" because most of my flying is over the Pacific. Hard to recover out there. Haha. Let us know how it turns out. My guess is DJI will replace with a refurbished unit and won't tell you the cause. Upload your flight file and folks on here can analyze it amazingly well.

Hi. Thanks. Newbie question : I have sent drone back to DJI, so I don't have it at the moment. But I still have DJI Go app flight record and SD card from drone. Is the flight file on the drone or phone or SD card?

 

[msinger]

Is the flight file on the drone or phone or SD card?

The TXT flight log is on your phone. See this link for instructions on retrieving it.

 

[SteveM]

Thanks.
Here is the file : (.txt)

 

[SteveM]

Error message on 1st drop

 

[SteveM]

2nd occasion : IMU Heading error and weak GPS

 

[Andrew F]

Thanks.
Here is the file : (.txt)

Give the experts some time to analyze. All I can see is a very scary event! Haha. Have you flown much in just ATTI mode? I think without practice and experience doing that, there can be a lot of extra stick input that creates an appearance of things gone crazy when it's really just stick confusion. For example, when the Mavic is coming toward you, the normal left/right controls are mirrored. If it's windy, things can get really dicey really quickly.

 

[SteveM]

Thanks. Never flown in ATTI before. But it has inspired me to want to have a practice in ATTI on my Phantom 4 when I get a low wind day. Could help me - as long as it responds to controls. Sounds like for some people it doesn't.

 

[Andrew F]

Thanks. Never flown in ATTI before. But it has inspired me to want to have a practice in ATTI on my Phantom 4 when I get a low wind day. Could help me - as long as it responds to controls. Sounds like for some people it doesn't.

Yeah. It's really had to say without analyzing the stick vs. Drone movement. Hopefully one of the analyais experts weighs in.

 

[andreww78]

Hey Mate, had any luck with the lost gps signal? still been flying ok? Mine did it yesterday not sure how to proceed.

 

[SteveM]

I sent mine back for repair. They replaced with another drone. So hopefully ok.