Mavic Pro Plat mystery drift and crash

[Hand Solo]

Hi Everyone,

I have edited the links in this post to include the now correct files - and fix my shoddy spelling!!

Following advice from @MAvic_South_Oz (cheers fella) I am posting my mobile DAT file (the SD version was not on the drone for some reason) 20-04-15-06-40-53_FLY147.DAT (Google Drive link) and DJI Flight Record (attached to this post) in regards to my Mavic Pro Platinum.

In short, DJI had told me to refresh the firmware on the drone to fix a camera jumping issue I had Camera Jumping. With the lockdown in place, in order to test, i have a narrow slither of space which is drone code friendly and so i lifted the drone off the ground and it hovered at minimum height. The drone drifted with no input until the very end where i panicked and pressed the wrong direction and ended up crashing into a garden chair. It didn't fall far, there is no visible damage to the drone (safe for a bit of light scuffing on the plastic bit that the props lock into) but my worry is that I have now invalidated my warranty as the drone has crashed and so if the camera issue persists then DJI will now not honour the repair. Secondly, have i caused any other damage and will some simple hovering allow me to ascertain if the drone is good to fly.

What I don't understand is I had done the same lift off and hover a few times previously with no drift. The only thing i did differently was put on the gold-tipped props that came with the drone - i don't know if these caused the drift or not.

Any insight would be welcome, i took a look at the DAT file in CSV view and my brain shorted - especially when it looks according to Geoplayer that i'm flying in the conservatory (the sat imagery is showing a shed that isn't there anymore which is where the drone was).

Link to original post

I think that is everything
G

 

[slup]

Hi Everyone,
Following advice from @MAvic_South_Oz (cheers fella) I am posting my Dat File (Google Drive link) and DJI Flight Record (attached to this post) in regards to my Mavic Pro Platinum.

In short, DJI had told me to refresh the firmware on the drone to fix a camera jumping issue i had Camera Jumping. With the lockdown in place, in order to test i have a narrow slither of space which is drone code freindly and so i lifted the drone off the ground and it hovered at minimum height. The drone drifted with no input until the very end where i panicked and pressed the wrong direction and ended up crashing into a garden chair. It didnt fall fall, there is no visiable damage to the drone but my worry is that I have no invlaidated my warranty as the drone has crashed and so if the camera issue does still persist then DJI will now not honour the repair.

What i dont understand is I had done the same lift off and hover a few times previusly with no drift. The only thing i did different was put on the gold tipped props that came with the drone - i dont know if these caused the drift or not.

Any insught would be welcome, i tok a look at the DAT file in CSV view and my brain shorted - especially when it l,ooks according to Geoplayer that im flying in the conservatory (the sat imagary is showing a shed that isnt there anymore which is where the drone was)

Link to original post

I think that is everything
G

That TXT log doesn't match with the supplied DAT log (it's only from a power on, no take off) ... the correct DAT log ends with FLY147.DAT for the attached TXT.

(And for your info ... it's kind of unnecessary to start up a new thread)

*EDIT*

Then looking into your first thread about this ... & the TXT file.

It seems that the MPP drifted slightly (0,8m from HP) ... can't find the hover accuracy in the MP manual, but going into mine for the MA it states for horizontal +/-0,1m with VPS sensors locked in (enough light & a pattern to lock to is required) ... only GPS +/-1,5m.

So if it's the same for the MPP & the VPS sensors had difficulties to lock in that drift you had is within spec.

 

[Hand Solo]

That TXT log doesn't match with the supplied DAT log (it's only from a power on, no take off) ... the correct DAT log ends with FLY147.DAT for the attached TXT.

(And for your info ... it's kind of unnecessary to start up a new thread)

*EDIT*

Then looking into your first thread about this ... & the TXT file.

It seems that the MPP drifted slightly (0,8m from HP) ... can't find the hover accuracy in the MP manual, but going into mine for the MA it states for horizontal +/-0,1m with VPS sensors locked in (enough light & a pattern to lock to is required) ... only GPS +/-1,5m.

So if it's the same for the MPP & the VPS sensors had difficulties to lock in that drift you had is within spec.

Hi Slup, thanks for the comments and help. The files I have so far uploaded here are from the time stamps i recall of the evening. I know they were a large gap between the crash and next power up as i inspected the drone and thats why I thought I had picked the correct files. Having used TXTlogToCSVtool and then run the flight log through I can see the Dat files are much younger...ie 147 as you say. I have the flight in DJIGo and I can see its a 25 sec flight - i will try and find the right flight record file from the 25 sec flight time and then that should give me the dat file.
Cheers

 

[Hand Solo]

So i went away and looked to get these files, in particular FLY147. It simply doesnt exist on the SD card - the earliest being 152 - and i havent formatted or deleted the SD card. However, I have recovered the Mobile Device Dat file. I have updated my original post with the correct files. Sorry about that, when you guys are donig this to help
G

 

[slup]

Yeah ... it was the mobile device DAT I wanted, & now we have the right one.

As the DAT is recording from power on to power off I can see that you did 3 take offs & crashed in the last of them. In the first you didn't touch anything besides the throttle to ascend & then descend. The middle flight you played with the elevator & the aileron stick also (horizontal movements) ... And in the third & last flight until you hit that garden chair, you didn't do any horizontal movements either ... so if comparing the first & last flight everything is very alike, you had a slight drift in the first flight also pretty much as large as in the last one. The motor rpm's were similar ... perhaps a tad faster in the last flight but not much, the rate gyros reported the same & the accelerometers the same.

Still believe that you didn't had a good ground pattern lock with the VPS sensors on the drone belly which made it up to the GPS accuracy to keep it in place. And hover drift with only GPS can be as large as +/-1,5m.

You can re-calibrate the IMU as that can affect how steady it hovers if this drift continues.

 

[Hand Solo]

Thanks Slup - so it sounds likle the take home message is that the ground lock wasnt great and the resultant drift was within GPS allowances - something that i need to appreciate going forward (and consider myself that the cost of this lesson was minor).

Is there any steps i need to take before flying with this drone again so i cn be sure its safe and wont put anyone at risk from a malfunctioning drone falling from the sky? - ie i have done a visual inspection and have let it hover in my conservatory - but beyond that I dont know what else i can do apart from post lockdown keep it in begnners mode and do some gentle flying and up the gears and altitude until confident its safe?

Finally, is there anything from the logs that will now stop DJI honouring the camera gimball claim i started before this happened? There is a slight scuff mark on the plastic motor housing so they could well shout 'hard landing'

Thanks again
G

 

[slup]

I would carefully check the motor shafts so nothing is bent there by observing them with motors idling without props. Then by hand feel on the motor rotation checking after any notchiness. Then new props of course. Open then up below page in GO4 with the AC powered on & check so all is green there ... if looks ok, then do a IMU calibration.

If you have synced your flights with DJI servers they surely have access to this crash flight ... but doubt that they go through everything, the scuff marks can be worse & perhaps make them ask & look into logs ... can't for sure say how they will act though.

 

[Hand Solo]

Hi Slup,
Been away from my PC for a bit but wanted to say (a little belated) thank you for all your help. Im going to retire my drone now until lockdown is done but will be flying as soon as allowed. Cant wait!!!

"I would carefully check the motor shafts so nothing is bent there by observing them with motors idling without props. Then by hand feel on the motor rotation checking after any notchiness. Then new props of course. Open then up below page in GO4 with the AC powered on & check so all is green there ... if looks ok, then do a IMU calibration."

All done, nothing bent - no notichness.

"If you have synced your flights with DJI servers they surely have access to this crash flight ... but doubt that they go through everything, the scuff marks can be worse & perhaps make them ask & look into logs ... can't for sure say how they will act though."

Spoken with DJI and they have no logs sync'd since Jan.