Please select Into the mobile phone version | Continue to access the computer ver.
Uncommanded flight deviations?
1634 21 2017-7-18
Uploading and Loding Picture ...(0/1)
o(^-^)o
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

Iv'e been noticing erratic flight behavior in my Mavic and decided to try and recalibrate everything.  The issue persists even after firmware update today, and fresh  IMU and compass calibration.. I have a 600 foot long nearly flat paved  driveway with trees on both sides. The driveway is a straight line to  the road.  I'm simply trying to fly up to the end of my driveway and  back, by flying straight up the center of the driveway at about 6 feet  off the ground ...   I am flying faster than tripod speed, but nowhere  near the max speed  in non-sport mode.  Doing this seems to go fine for  about 50 feet or so, and then the drone will entirely on its own  suddenly lose altitude to the point of scraping the  driveway unless I compensate and save it. It will also suddenly  move  laterally right or left entirely uncommanded, and on one occasion  flying a straight line it actually suddenly climbed to about 25 feet  altitude  barely missing the canopy of the trees overhead because I  caught it in time but I had to hold full down stick to get it to stop climbing.  Occasionally when flying forward in a straight line at about  half speed when I go to stop it won't just stop and hover, but will stop and then fly backwards or sideways about five or 6 feet.   Obviously the  small deviations are no big deal when you're 300 feet in the air over  an open scenic area, but when you are trying to fly  precisely in a  confined area these uncommanded deviations could result in a crash very  easily. (I also tried to do a stick calibration on the transmitter, but it wont let me.. every time I try to start the calibration it just pops up a warning to be sure to return sticks to center before hitting start, obviously they are centered, because they are self centering sticks.

2017-7-18
Use props
Scott Ferguson
lvl.4
Flight distance : 1425223 ft
United States
Offline

I have noticed that the sticks on my RC are very touchy and takes very little input at all to change path or speed but have just come to realize my thumbs doesn't move straight up,down,left or right so figure I will get input to the stick unknowingly
2017-7-18
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Has it always behaved this way or did its behavior change at some point? At least a couple of the things you described could be caused by a faulty VPS sensors, but I'm just speculating. If you upload a DAT file from a flight where you experienced the behaviors you mentioned I'll take a look at it and may be able to give you an idea of why it's happening.

The instructions for retrieving DAT files are here and once you've retrieved files for the flight(s) where you experienced these problems you'll need to upload them to DropBox or something similar and provide a link in this thread. The correct DAT files will have timestamps that correspond to when the flights occurred and will be at least tens of megabytes in size.
2017-7-18
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

Scott Ferguson Posted at 2017-7-18 15:45
I have noticed that the sticks on my RC are very touchy and takes very little input at all to change path or speed but have just come to realize my thumbs doesn't move straight up,down,left or right so figure I will get input to the stick unknowingly


I have modified my Expo and sensitivities to deal with its inherent touchiness.
2017-7-18
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-18 16:03
Has it always behaved this way or did its behavior change at some point? At least a couple of the things you described could be caused by a faulty VPS sensors, but I'm just speculating. If you upload a DAT file from a flight where you experienced the behaviors you mentioned I'll take a look at it and may be able to give you an idea of why it's happening.

The instructions for retrieving DAT files are here and once you've retrieved files for the flight(s) where you experienced these problems you'll need to upload them to DropBox or something similar and provide a link in this thread. The correct DAT files will have timestamps that correspond to when the flights occurred and will be at least tens of megabytes in size.


I will look at sending the dat files to Dropbox so you can look at them.. I wish there was a way to show the recorded footage with the stick inputs so you could see exactly what I'm talking about. I can make the videos available, if needed for comparison. Thank you! I will get to it tomorrow!
2017-7-18
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Phan-TOM Posted at 2017-7-18 17:04
I will look at sending the dat files to Dropbox so you can look at them.. I wish there was a way to show the recorded footage with the stick inputs so you could see exactly what I'm talking about. I can make the videos available, if needed for comparison. Thank you! I will get to it tomorrow!

Yes, video might be helpful so if you have one then please do include it.

As for showing the stick input, I'll be able to see the input in the DAT file, but one way you can show it in a video is to record the flight playback in the Go app with the stick option selected. Just click on the remote control icon at the bottom of the replay screen to toggle the display of the stick input and when it's selected it'll show the stick input as the flight is replayed.
2017-7-18
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-18 18:31
Yes, video might be helpful so if you have one then please do include it.

As for showing the stick input, I'll be able to see the input in the DAT file, but one way you can show it in a video is to record the flight playback in the Go app with the stick option selected. Just click on the remote control icon at the bottom of the replay screen to toggle the display of the stick input and when it's selected it'll show the stick input as the flight is replayed.

Can those files be reviewed by a third party? Can YOU watch the video with the stick inputs? I will try to upload as much as I can from yesterday's flights. Several of the flights were spent going up and down my driveway trying to assess the issue. Thank you again for the help!

EDIT: after revewing my devices, I have found I do not have the recorded flight videos in my "Editor" whereby I could watch the video and see the stick inputs on screen as my Cashe was full at the time. I have the onboard footage from the SD card on the Mavic itself, and the replay files with the map overlay, but am having trouble identifying which files belong to which video I have from the SD card, since the Mavic breaks the videos into segments. I will try to just upload everything I have from yesterday and hopefully you can see what I'm talking about.

While I try to retreive the DAT files, here is a sample of the onboard SD card recorded video. At 2:17 the Drone climbs uncommanded and as I turn left 90 degrees it also backs up about three feet almost into the trees. I give it a bump of forward to clear the trees. This was not the first or only deviation of this kind. several times up the driveway, I had to stop forward flight to correct for a side deviation or uncommanded descent. at one point it actually decended into the ground, but did not flip over like a Phantom would due to low CG. You asked if it had always done this, and I'm not sure, one thing I do know, is that it has never followed terrain in manual flight across the ground. For example if I fly across the ground with a slight incline it will not compensate. it will fly right towards the slope until the forward sensors stop it from flying into the slope. if it is steep enough for them to see it. Anyway. I will upload all my DAT files from yesterday for review. Thank you!
2017-7-19
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-18 16:03
Has it always behaved this way or did its behavior change at some point? At least a couple of the things you described could be caused by a faulty VPS sensors, but I'm just speculating. If you upload a DAT file from a flight where you experienced the behaviors you mentioned I'll take a look at it and may be able to give you an idea of why it's happening.

The instructions for retrieving DAT files are here and once you've retrieved files for the flight(s) where you experienced these problems you'll need to upload them to DropBox or something similar and provide a link in this thread. The correct DAT files will have timestamps that correspond to when the flights occurred and will be at least tens of megabytes in size.

Here is a link to all of my DAT files from Yesterday. If I'm not mistaken, just about every flight was spent trying to go up and down my driveway accurately. Just to be clear, I have several years of Drone flying experience with both racing drones and DJI, as well as other RC aircraft. I should have NO issues flying a straight line up my driveway.  Link -->    https://1drv.ms/f/s!An3tzeJKHTp6ukTy1GMaGLuqSfqQ
2017-7-19
Use props
DJI Thor
Administrator
Flight distance : 13602 ft
Offline

I would recommend that you sync the flight records and export the flight log, then contact the support to open a case, we'll submit a data analysis to find out the cause.
Here’s the link to contact us: http://www.dji.com/support
2017-7-19
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

Here is a link for Dat files and associated video files for a single flight today Up the Driveway.  I also took the Mavic to some train tracks this morning to have a perfectly flat Straight line referance.  I also had more room to test how far it would deviate. Before this flight, I was able to finally do a Calibration of control sticks, just in case, but it made no difference. I did not see any dramatic uncommaned descents, however lateral deviations were big sometimes 20 degrees or more off course while keeping the same heading. some deviations were abrupt and caused me to terminate forward flight to correct as the drone would suddenly drift 15 -20 feet laterally. I have video and DAT files for that flight as well if needed.n I will edit this post with a link once those train track files are uploaded.
2017-7-19
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Phan-TOM Posted at 2017-7-19 06:04
Here is a link for Dat files and associated video files for a single flight today Up the Driveway.  I also took the Mavic to some train tracks this morning to have a perfectly flat Straight line referance.  I also had more room to test how far it would deviate. Before this flight, I was able to finally do a Calibration of control sticks, just in case, but it made no difference. I did not see any dramatic uncommaned descents, however lateral deviations were big sometimes 20 degrees or more off course while keeping the same heading. some deviations were abrupt and caused me to terminate forward flight to correct as the drone would suddenly drift 15 -20 feet laterally. I have video and DAT files for that flight as well if needed.n I will edit this post with a link once those train track files are uploaded.

Can you try to provide the link again? I didn't see it in that post. I did see that you linked to a OneDrive box with 11 DAT files in it in your previous post, but without knowing which specific one to look at and which problem(s) I'm looking for it's difficult to know what to look at. If you can point me to a single DAT file -- like the one you mentioned you created today -- I'll be sure to take a look at it. And yes, please do also include the video file that you created.
2017-7-19
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-19 06:50
Can you try to provide the link again? I didn't see it in that post. I did see that you linked to a OneDrive box with 11 DAT files in it in your previous post, but without knowing which specific one to look at and which problem(s) I'm looking for it's difficult to know what to look at. If you can point me to a single DAT file -- like the one you mentioned you created today -- I'll be sure to take a look at it. And yes, please do also include the video file that you created.

Sorry, I was having trouble getting the link to work.. I made two flights today, a single battery each. One up my driveway again, and one at some train tracks. The single flight up my driveway seems to have created 5 DAT files with todays date, while the flight at the tracks generated two DAT files.  Can you describe the process or link me to a description of the process of "Syncing flight records and exporting flight logs", as mentioned in DJI Thors post above? Otherwise I only know how to link you to a folder hosting my videos and associated Dat files as before. Im sorry there were a bunch to look at, but any one of them likely contains examples of uncommanded flight deviation as it happened repeatedly on several flights throughout the day yesterday.  Todays flights may be easier, because as I said my flight at home was 5 DAT files and my flight at the tracks was 2 Dat files. Both flights were experiencing the same deviations.

The link on onedrive for todays flight at home - 3 video and 5 DAT files ->  https://1drv.ms/f/s!An3tzeJKHTp6ulIJM1bdirjyuVqq

The link for the flight at railroad tracks - 3 videos 2 DAT files -> https://1drv.ms/f/s!An3tzeJKHTp6ul2eS6OFjf9e0fCc

Is it possible that the multiple DAT files per single battery/flight is a result of switching between manual flight mode and tripod mode? Several times I switched to tripod to see if it also had the same deviations as well as for my own peace of mind that any deviations would be less pronounced.
2017-7-19
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Phan-TOM Posted at 2017-7-19 08:49
Sorry, I was having trouble getting the link to work.. I made two flights today, a single battery each. One up my driveway again, and one at some train tracks. The single flight up my driveway seems to have created 5 DAT files with todays date, while the flight at the tracks generated two DAT files.  Can you describe the process or link me to a description of the process of "Syncing flight records and exporting flight logs", as mentioned in DJI Thors post above? Otherwise I only know how to link you to a folder hosting my videos and associated Dat files as before. Im sorry there were a bunch to look at, but any one of them likely contains examples of uncommanded flight deviation as it happened repeatedly on several flights throughout the day yesterday.  Todays flights may be easier, because as I said my flight at home was 5 DAT files and my flight at the tracks was 2 Dat files. Both flights were experiencing the same deviations.

The link on onedrive for todays flight at home - 3 video and 5 DAT files ->  https://1drv.ms/f/s!An3tzeJKHTp6ulIJM1bdirjyuVqq

By any chance do you have some type of accessory attached to your Mavic such as a tracking device / locator, landing gear extension, etc? Also, can you check your props to make sure that there's no damage to them?

Can you describe the process or link me to a description of the process of "Syncing flight records and exporting flight logs", as mentioned in DJI Thors post above?

Syncing the flight records is done by going to the main page of the Go app, clicking on the three horizontal bars in the top right, then clicking on the Flight Records selection. Once there you'll see a cloud icon in the upper right corner of the screen, click that and then click on "Start Synchronization". That just synchronizes your mobile device TXT files with DJI's servers so that both contain the same list of flights; those on your device but not on their server are copied to the server and vice-versa.

With regard to "exporting the flight logs", I think he was referring to the DAT files which you've already figured out how to do and linked to, so I don't think you need to do anything else there, though I'll defer to Thor on that. Of course if we figure out what's going on before then it won't matter anyway.
2017-7-19
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-19 09:18
By any chance do you have some type of accessory attached to your Mavic such as a tracking device / locator, landing gear extension, etc? Also, can you check your props to make sure that there's no damage to them?

Can you describe the process or link me to a description of the process of "Syncing flight records and exporting flight logs", as mentioned in DJI Thors post above?

OK I will sync the logs. To answer your question on the condition of my mavic, it is mint. Ive only had it for maybe a month or so and have only flown it maybe 20 times. Its never been crashed or even so much as brushed a tree branch. there are NO accessories or modifications to the factory aircraft at all except my stupid registration sticker on the rear wall of the battery bay.  - I was watching my flight log on my Phone of my flight at the train tracks this morning, where I made several back and forth attemts to follow the dead straight tracks, I made several passes both forward flight and backward flight at different speeds. You can clearly see the deviations on the overhead map, and watching the sticks you can see there is no input other than forward on the right stick unless intentional corrections are made.  I did not run into any alarming altitude changes at the tracks, but on the log screen the "height" almost always reads 0.0 ft.  On my driveway flight this morning it did once again lose altitude and drift making contact with the ground.
2017-7-19
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

Spent the day trying to figure this out. Several sessions with DJI support chat, four different support agents, multiple times refreshing firmware, calibrating IMU, compass, sticks on the controller and recalibrating the forward and downward facing sensors, and still no change. It flies straight for 10-20 feet, then dives for the ground, slides left or right several feet or climbs until I stop it. No telling which way it will go either. It just doesn't want to fly in a straight line. Numerous attempts to recalibrate everything and test fly have not helped.. and really, neither has an entire day working with DJI support.
2017-7-19
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Phan-TOM Posted at 2017-7-19 17:22
Spent the day trying to figure this out. Several sessions with DJI support chat, four different support agents, multiple times refreshing firmware, calibrating IMU, compass, sticks on the controller and recalibrating the forward and downward facing sensors, and still no change. It flies straight for 10-20 feet, then dives for the ground, slides left or right several feet or climbs until I stop it. No telling which way it will go either. It just doesn't want to fly in a straight line. Numerous attempts to recalibrate everything and test fly have not helped.. and really, neither has an entire day working with DJI support.

I did look at your files a little a couple of days ago and was hoping to come up with something more conclusive, but at this point all I can say is that your concerns about your Mavic are justified, because it doesn't seem to be functioning normally. I was going to review the videos again this morning, but it looks like you took them down and to be honest I'm not sure I'd gain more insight anyway.

The drifting appears to have a couple of different causes. First, both the videos and DAT files indicate that there's a bit of yaw drift occurring even when your Mavic is stationary. I've seen a few other people mention this happening to them lately but unfortunately I'm not aware of there being a definitive cause or solution. I know you already mentioned that you've done IMU and compass calibrations, and the DAT file shows no stick input so I'd guess that calibrating the remote also won't help, though I suppose it couldn't hurt either.

The more dramatic lurches you're seeing in your driveway may be caused by the reduced GPS health that occurs due to the thick canopy of trees around your driveway. In other words, this may be the Mavic trying to adjust its position after it regains a GPS location that it can trust. The DAT file does show a number of brief drops in the GPS health and also a handful of GPS_POSITION_NONMATCH errors, though the correlation between them is far from perfect so I'm not at all certain this is really what's causing it. I did notice that at least in the early minutes of the railroad track videos -- where you were flying in a more open area -- that you appeared to only encounter the less dramatic yaw drifting and not the more serious sliding experienced in your driveway.

The VPS data, on the other hand, is a bit more clear-cut: unless there's something physically interfereing -- such as an obstructed sensor or some attachment that's interfering with the readings -- it definitely appears that the VPS readings are sporadically incorrect, and that's why you're seeing the sudden drops in altitude. In other words, the Mavic suddenly thinks that its altitude is higher than it really is and drops down to compensate.

So while the data related to not being able to fly straight is inconclusive, the bottom line is that there does appear to be a problem with your Mavic and it's probably one that should be covered under warranty. At least that's my personal opinion and hopefully DJI agrees. I'm sorry that I don't have a more definitive diagnosis for you, but I am pretty confident in saying that this is something you should continue to pursue with DJI, and I personally would be pretty insistent on getting a warranty replacement.

Good luck and let us know how things turn out.
2017-7-22
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-22 05:14
I did look at your files a little a couple of days ago and was hoping to come up with something more conclusive, but at this point all I can say is that your concerns about your Mavic are justified, because it doesn't seem to be functioning normally. I was going to review the videos again this morning, but it looks like you took them down and to be honest I'm not sure I'd gain more insight anyway.

The drifting appears to have a couple of different causes. First, both the videos and DAT files indicate that there's a bit of yaw drift occurring even when your Mavic is stationary. I've seen a few other people mention this happening to them lately but unfortunately I'm not aware of there being a definitive cause or solution. I know you already mentioned that you've done IMU and compass calibrations, and the DAT file shows no stick input so I'd guess that calibrating the remote also won't help, though I suppose it couldn't hurt either.


Thank you so much for your time, looking into this for me, I really appreciate it!  I have to admit I'm surprised you didn't see more definitive data on the deviations I had at the railroad tracks,as the deviations were more pronounced, and I had more room to let them go uncorrected, to make them more evident. On the driveway I could only let it deviate so far before I would run into tree branches.. I did start a repair request with DJI, and got the shipping label, and will be sending it back for them to analyze. In the meantime, for the sake of comparison, (and because frankly I wanted one anyway), I picked up a P4P+ and attempted to experiment with the same flights scenarios.  With the P4P, I have not yet had a chance to go to the railroad tracks, but I did run it up and down the driveway a couple times. The phantom loses altitude over the driveway the same way the mavic does, and some browsing of the forums leads me to believe that the VSP and altitude holding issues are a result of the firmware. It seems to be a fairly common problem.  I also saw some lateral deviation with the P4P, but it was not as pronounced as it was with the Mavic. When it stops raining here I will get the P4P out to the tracks for a test flight. I don't think these problems will be a major crisis when flying at higher altitude's over open areas, but when I'm trying to do close-in Dolly shots, it can be extremely annoying.
2017-7-24
Use props
rydfree41
Second Officer
Flight distance : 153268 ft
United States
Offline

Just a thought , and maybe I missed it skimming the previous post , but have you tried turning off all sensors , Obstacle avoidance , precision landing settings etc and flown it to see if it acts the same ?  If it behaves normal then turn on one at a time to see if you can narrow it down to one setting that causes the issue .
2017-7-24
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

I will give that a try, but honestly I'm starting to think it might be some kind of magnetic interference with the compass or other GPS interference. Why it would not maintain altitude in forward flight is still beyond me, but there seem to be more people on other forums that have had the same issue with theirs. To further support my theory, I took my new P4P out and flew the exact same flights in the exact same locations including the train tracks, and saw it behave almost identically. Even the P4P drifted slightly in yaw, and went off track in forward flight, and even in almost the EXACT same places along the flight path. It also was losing altitude in slower forward manual flight. I really do think it may be a fault in the firmware. Either that or its a magnetic interference as I stated.. Not sure, but I'm going to hold off sending out my mavic for the moment.
2017-7-25
Use props
DroneFlying
Captain
Flight distance : 10774613 ft
United States
Offline

Phan-TOM Posted at 2017-7-25 10:07
I will give that a try, but honestly I'm starting to think it might be some kind of magnetic interference with the compass or other GPS interference. Why it would not maintain altitude in forward flight is still beyond me, but there seem to be more people on other forums that have had the same issue with theirs. To further support my theory, I took my new P4P out and flew the exact same flights in the exact same locations including the train tracks, and saw it behave almost identically. Even the P4P drifted slightly in yaw, and went off track in forward flight, and even in almost the EXACT same places along the flight path. It also was losing altitude in slower forward manual flight. I really do think it may be a fault in the firmware. Either that or its a magnetic interference as I stated.. Not sure, but I'm going to hold off sending out my mavic for the moment.

One thing I did notice is that when you flew over the railroad tracks that you apparently took off from your truck's tailgate. The steel can interfere with the Mavic's compass, so I'd suggest making sure that there's no metal nearby when you take off. On the other hand, your driveway appears to be asphalt and there's normally no ferrous (e.g., rebar) material there, but even so I'd suggest that you take off from a grassy area when you do more testing.
2017-7-25
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

DroneFlying Posted at 2017-7-25 10:17
One thing I did notice is that when you flew over the railroad tracks that you apparently took off from your truck's tailgate. The steel can interfere with the Mavic's compass, so I'd suggest making sure that there's no metal nearby when you take off. On the other hand, your driveway appears to be asphalt and there's normally no ferrous (e.g., rebar) material there, but even so I'd suggest that you take off from a grassy area when you do more testing.

My truck is Aluminum, but the train tracks are steel lol! anyway I also noticed it with my mavic over a cornfield one day, it took alot of side stick to track a straight flight path. I'm going to try the P4P at the corn field also, so We will see.
2017-7-25
Use props
Phan-TOM
lvl.3
Flight distance : 61184 ft
  • >>>
United States
Offline

Confirmed at cornfield location as well. this is where I first noticed the Mavic not flying a strait line, because its very pronounced there. Mavic was flying a straight line then would veer off to the right requiring a good deal (50% or more) left stick to counter the drift.. the P4P did literally the exact same thing in the exact same spot. multiple times. Who knows what is causing it, but I no longer think its the drone that is the problem. Its either firmware, or local magnetic interferance. only thing I can think of. I'm going to go out there with a magnetic compass, and see if it happens while walking.
2017-7-27
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules