Crash
12Next >
2906 42 2018-2-18
Uploading and Loding Picture ...(0/1)
o(^-^)o
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

P3 std had a compass calibration required, completed fine, home point recorded. Flying in mode 2 with switch set to GPS. Showed 7 sats connected, everything green for go. Took of, height about 10 ft flying forward 150 to 200 ft away aircraft moved left and flipped on its back landing upside down, broke all props gimble mount broke and all wires pulled from camera. When i got to ac motors running from low speed to full speed. Could not shut down with controller had to remove battery to stop. Aircraft new Feb 9, latest firmware and latest dji go app running Samsung s8+. Any ideas appreciated.
Regards
Mick
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

extract the flight log out from your handset which you install the APP DJI go ......... it will be inside your folder DJI and a flightrecords...the file will be in TXT format but do upload to http://www.phantomhelp.com/LogViewer/Upload/ ( follow instruction )

If its non pilot error mostly u have to send it back or to the dealer where u bought...warranty mostly will be covered fully on the Aircraft ...
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

one question ...when u found your drone laying on the ground..is all the 4 PROPS intact on all the 4 MOTORs or ??
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Siambuddhas Group Posted at 2018-2-18 08:52
one question ...when u found your drone laying on the ground..is all the 4 PROPS intact on all the 4 MOTORs or ??

Yest prop atteched but tips broke off. I had dji prop guards on but being upside down they wont help, I crashed onto grass.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Siambuddhas Group Posted at 2018-2-18 08:50
extract the flight log out from your handset which you install the APP DJI go ......... it will be inside your folder DJI and a flightrecords...the file will be in TXT format but do upload to http://www.phantomhelp.com/LogViewer/Upload/ ( follow instruction )

If its non pilot error mostly u have to send it back or to the dealer where u bought...warranty mostly will be covered fully on the Aircraft ...

Thanks I will do you have details of how i find it in the app or do i just go to the link and it will do it automatically
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

do find the txt files and upload it ... ... do remember to update your flight log in your APPs...........
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

save that txt files in your PC or so...den you upload to that URL as per normal ..
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Siambuddhas Group Posted at 2018-2-18 09:02
do find the txt files and upload it ... ... do remember to update your flight log in your APPs...........

Ok, having trouble finding but will work it out
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

sure...mostly if its not pilot errors ... mostly warranty are covered...but till den will still need your flight log to be upload...
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Siambuddhas Group Posted at 2018-2-18 09:13
sure...mostly if its not pilot errors ... mostly warranty are covered...but till den will still need your flight log to be upload...

Just uploaded file, at 1m 30 sec go from p gps to atti 8 sats 12ft 13mph 345ft gives speed error
2018-2-18
Use props
Siambuddhas Group
lvl.4
Flight distance : 116926 ft
Singapore
Offline

where is the uploaded link??
2018-2-18
Use props
RicardoGray
Second Officer
Flight distance : 4356421 ft
United States
Offline

That is a strange reaction for your phantom to make for sure. Can't imagine what would have caused it to just flip like that. Were there strong gusts of wind at that time? Yeah, the log would be helpful in figuring this out.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Siambuddhas Group Posted at 2018-2-18 09:40
where is the uploaded link??

Just uploaded again i still have it openon my phone
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

RicardoGray Posted at 2018-2-18 09:52
That is a strange reaction for your phantom to make for sure. Can't imagine what would have caused it to just flip like that. Were there strong gusts of wind at that time? Yeah, the log would be helpful in figuring this out.

Wierd, it moved left then flipped, dead calm
2018-2-18
Use props
AlanHd
First Officer
Flight distance : 797365 ft
United Kingdom
Offline

Mickus Posted at 2018-2-18 09:56
Just uploaded again i still have it openon my phone

Can you post a link to the page here so others can have a look.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

AlanHd Posted at 2018-2-18 10:50
Can you post a link to the page here so others can have a look.

http://www.phantomhelp.com/logviewer/5L3P2EBA1WWE14V1MTKT
2018-2-18
Use props
ALABAMA
First Officer
Flight distance : 10442687 ft
United States
Offline

Looks like you only had 8 satellites locked the entire flight.  Could have been a problem with the GPS.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

ALABAMA Posted at 2018-2-18 13:19
Looks like you only had 8 satellites locked the entire flight.  Could have been a problem with the GPS.

I have only flown about 6 times, only had a/c for a few days, but I dont recall having more than 8. Is this not good enough? Fromthe log I see everything happening when it went to atti mode and it was less than 1 second.
2018-2-18
Use props
Hellsgate
Second Officer
Flight distance : 1991555 ft
Australia
Offline

If you required a compass calibration prior to taking off that should have been a warning sign straight away.  That means you were close to or ontop of a strong magnetic field like reinforced concrete etc once the drone lifted off it moved away from the magnetic field and thats when things whent wrong because you calibrated your compass within a magnetic field you actually doomed your drone in that instant.
Once the drone leaves the magnetic field its now trying to read the earths magnetic field but it cant because its calibration is now incorrect so it doesnt know which way its facing and will behave very erratically.

Sorry for your drone but i think this is pilot error.

Should never need to calibrate the compass
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Hellsgate Posted at 2018-2-18 13:38
If you required a compass calibration prior to taking off that should have been a warning sign straight away.  That means you were close to or ontop of a strong magnetic field like reinforced concrete etc once the drone lifted off it moved away from the magnetic field and thats when things whent wrong because you calibrated your compass within a magnetic field you actually doomed your drone in that instant.
Once the drone leaves the magnetic field its now trying to read the earths magnetic field but it cant because its calibration is now incorrect so it doesnt know which way its facing and will behave very erratically.

Interesting, I was in one of my fields, maybe there is something underneath.
2018-2-18
Use props
ALABAMA
First Officer
Flight distance : 10442687 ft
United States
Offline

Labroides and Ethoderm are the best at reading flight logs.  Let's see what they have to say.
2018-2-18
Use props
AerOps
lvl.2

France
Offline

Mickus Posted at 2018-2-18 11:09
http://www.phantomhelp.com/logviewer/5L3P2EBA1WWE14V1MTKT

Don't see any tracking on the map ? GPS issue ?
I don't think that compas calibration was wrong due to magnetic interferences. You don't make the calibration on the ground, the "Sioux dance" (french expression for the procedure) is minimum 1 meter over the ground.
2018-2-18
Use props
Hellsgate
Second Officer
Flight distance : 1991555 ft
Australia
Offline

Mickus Posted at 2018-2-18 13:45
Interesting, I was in one of my fields, maybe there is something underneath.

There could be some iron in the ground were you took off from.
If ever you see the compass calibration warning come up first thing to do is shut everything down and try moving location.This may only require moving a few feet or a cple hundred meters it will totally depend on whats in the ground.
In one forum i read about a guy that had this error come up and got him interested in what was causing it so he dug up the ground and unearthed a ww2 tank.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

ALABAMA Posted at 2018-2-18 13:48
Labroides and Ethoderm are the best at reading flight logs.  Let's see what they have to say.

I do not know who they are but I will be sending unit back to dji tommorow, i presume the ac has internal log records.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

AerOps Posted at 2018-2-18 14:16
Don't see any tracking on the map ? GPS issue ?
I don't think that compas calibration was wrong due to magnetic interferences. You don't make the calibration on the ground, the "Sioux dance" (french expression for the procedure) is minimum 1 meter over the ground.

Not sure but something happened andit was extremely quick. It seems loss of gps because log shows atti mode 1/2 second before crash.
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Hellsgate Posted at 2018-2-18 14:20
There could be some iron in the ground were you took off from.
If ever you see the compass calibration warning come up first thing to do is shut everything down and try moving location.This may only require moving a few feet or a cple hundred meters it will totally depend on whats in the ground.
In one forum i read about a guy that had this error come up and got him interested in what was causing it so he dug up the ground and unearthed a ww2 tank.

Thanks for info. When dji techs determine cause i will post  here.
2018-2-18
Use props
Hellsgate
Second Officer
Flight distance : 1991555 ft
Australia
Offline

Mickus Posted at 2018-2-18 15:14
Not sure but something happened andit was extremely quick. It seems loss of gps because log shows atti mode 1/2 second before crash.

Atti mode would have initiated due to the drone being upside down and loosing contact with gps sattelites
2018-2-18
Use props
Hellsgate
Second Officer
Flight distance : 1991555 ft
Australia
Offline

Mickus Posted at 2018-2-18 15:12
I do not know who they are but I will be sending unit back to dji tommorow, i presume the ac has internal log records.

Yes the ac keeps an internal log i believe and the dji techs should be able to retrieve it
2018-2-18
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Hellsgate Posted at 2018-2-18 15:52
Yes the ac keeps an internal log i believe and the dji techs should be able to retrieve it

I just found that you can look at the A/C and remote flight info from DataViewer which is included in DJI Assistant 2, a download for windows or mac. Interesting information allowing an analysis of A/C , Compass, Sats, and stick inputs from remote. Good topic explaining capabilities here https://forum.dji.com/forum.php? ... &fromuid=274127
2018-2-19
Use props
endotherm
First Officer
Flight distance : 503241 ft

Australia
Offline

OK, lots of confusing explanations and suggestions here, I'll try to address them.

If the aircraft is reporting compass calibration required, then there is a high likelihood that there is something in your vicinity upsetting the magnetic balance, probably iron or steel in the ground, reinforcement in concrete, buildings, buried pipes, keys/coins in your pocket etc.  As suggested, move to another location and try again before taking off.  If you take off on top of what amounts to a large magnet, all your readings will make no sense when you are up in the air and the aircraft won't know which way it is facing and cause navigation errors.  It's already complaining before you start, so it will only get worse.  It might not require a full calibration, just a change of location.  In fact, it is recommended to get a good compass calibration and leave it unless you travel 100's of miles away or you have a hard crash etc.

With the Standard series of aircraft, you only have a GPS receiver, Adv and Pro also contain a GLONASS system, so it will report twice as many satellites.  The number you have reported is normal for Standards.  There is nothing wrong with your GPS.

Warranty on main components is anywhere from 6 month to 12 months, depending on what part is defective.  To get warranty, you have to show that there was a malfunction and it was DJI's responsibility.  Pilot errors and accidents don't qualify.  If you took off and flew when it was reporting errors, it is pilot error.  It is out of DJI's control, they won't take responsibility for YOUR choice to continue flying without addressing the fault.

If you crash it and the aircraft is on the ground, it does not know it has landed until you shut down the motors.  Therefore it still thinks it is in flight.  Being upside down is a serious flight condition, so it will do everything in its power to get the right way up, even ignoring any commands from the pilot's controller until it is up the right way again.  This is why you couldn't turn it off and it was spinning at a high rate.  This is normally a good thing if you hit something and tumble during flight.  Unfortunately on the ground this usually results in the aircraft grinding the prop hubs into the terrain and destroying them and the blades.  Sometimes even the ESC (motor controller) and motors can suffer and burn out as well.  All you can do is get to it and pull the battery as fast as you can to minimize damage.  So that was normal.  Ripping the gimbal off and disconnecting the wires is pretty standard because this is probably the most vulnerable part and most fragile.  As long as you didn't break the metal plates, you might be lucky and be able to reassemble the wiring and replace the rubber hangers.

I'll have a look at your flight log and see if I can see what might have gone wrong and caused the crash.  We have a pretty good success rate on the forum if you can work out how to extract the file from your controller and share it like you did.  There is also a more comprehensive flight log inside the aircraft "black box" which we can examine if necessary, but usually it isn't.  If the aircraft is lost, e.g. crashed at sea, all we have left is the minor flight record from the controller, which you've uploaded here.  DJI examine the internal record to determine warranty eligibility i.e. not a pilot error, and that something failed.  So it's best not making up stories about what happened, because EVERYTHING is recorded and can be proven.     This DAT file is larger than the TXT file flight record, and needs software to decrypt it and view it.  It is a third-party program, NOT the DJI Assistant 2 program, which works on Phantom 4 and later only.

When you crash and go inverted etc, you will get a plethora of compass errors, motor speed warnings, satellite errors etc.  This is normal.  It is a result of the crash, not the cause of the crash.  If you see a bunch of these messages coming thick and fast, you've already crashed.  If it is upside down, the GPS which is in the center of the top shell can't see the sky any more, so your aircraft will switch to ATTI mode, which means you are still in control but you lose the GPS positioning stabilization until it returns.  As I said before, if it is upside down it is more serious than just flying under an obstruction like overhanging tree branches, so it ignores your input and concentrates on just one thing -- getting the right side up.  In ATTI mode it is affected by wind and is likely to drift all over the place.  This is also normal.  You need some practice controlling the aircraft in this mode in case it ever happens to you unexpectedly.

==========================================================================================

OK, I've examined your flight record.  One of the simplest I think I've ever seen

I see nothing out of place.  Some gentle climb and forward commands at low speed.  Nothing to indicate pilot error and nothing to indicate a fault occurring.

Looking at the data, all I can say is that it shows there was no wind (if there were, we would see the aircraft pitch and roll angles change to counter the wind).   Your last command was right stick forward, then centered.  I then saw the aircraft react to that with active braking, tilting in the opposite direction to come to a standstill.  The angle reported was 16° nose-up, which seems pretty severe, almost half of the maximum angle it is capable of.  Perhaps you have your control settings unnecessarily set to very, very responsive, which has the downside of making changes in direction quite abrupt and violent, and could take a toll on the aircraft over time (cracks, loose components etc).  If the aircraft has been flown "hard" up until now, maybe something was loose or damaged?   Finally the spreadsheet shows the fall and crash without any external stimulus.
spreadsheet.png

At this point we can't identify what caused it.  We generally trot out a couple of likely suspects: birdstrike, or the prop wasn't attached properly and flew off (you reported all props were intact after the crash, so that is eliminated), or one of the props snapped off in flight, due to faulty manufacture (unlikely) or previously damaged by the user from a hard landing or crash.  The latter may be confirmed if they check your previous flight records that have been synced (or present in the flight recorder) showing a crash/hard landing.  Theoretically it is possible to examine the motor speeds and loads to determine if the prop broke in flight, but the math is pretty involved and hard for me to determine if that is the case.   DJI probably have a better idea.  Sometimes viewing the video can offer further clues that aren't evident in the log.    It could show collision with a power line or a tree or other obstacle that doesn't show up on the (old) satellite map.  If you don't have a deliberate video recording, often the lower resolution live feed can be extracted from the smart device and shared so we can take a look at it.  

As you've had the aircraft less than a year and it's still under warranty, I'd definitely give the warranty claim a go as it appears you have done nothing wrong.  I don't know what the outcome would be, but good luck.

Hope you followed the crash investigation steps as I've explained them and learned something in the process.

2018-2-19
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

endotherm Posted at 2018-2-19 12:46
OK, lots of confusing explanations and suggestions here, I'll try to address them.

If the aircraft is reporting compass calibration required, then there is a high likelihood that there is something in your vicinity upsetting the magnetic balance, probably iron or steel in the ground, reinforcement in concrete, buildings, buried pipes, keys/coins in your pocket etc.  As suggested, move to another location and try again before taking off.  If you take off on top of what amounts to a large magnet, all your readings will make no sense when you are up in the air and the aircraft won't know which way it is facing and cause navigation errors.  It's already complaining before you start, so it will only get worse.  It might not require a full calibration, just a change of location.  In fact, it is recommended to get a good compass calibration and leave it unless you travel 100's of miles away or you have a hard crash etc.

Endotherm I want to thank you for the post, I really appreciate the information, in fact I am impressed with how you are able to interpret the log.
I have learned a great deal from this, maybe an expensive learning experience, but useful nonetheless.I am fully prepared for dji to come back with pilot error and would not argue against it for the simple reason that I have only flown about6 times and frankly am not sure if I did anything to cause it. I am going to go over your post andthe log more in depth tommorow on my PC.
Until then once again thanks, its people like you that make these types of forums invaluable
Regards
Mick Steele
2018-2-19
Use props
endotherm
First Officer
Flight distance : 503241 ft

Australia
Offline

Mickus Posted at 2018-2-19 15:18
Endotherm I want to thank you for the post, I really appreciate the information, in fact I am impressed with how you are able to interpret the log.
I have learned a great deal from this, maybe an expensive learning experience, but useful nonetheless.I am fully prepared for dji to come back with pilot error and would not argue against it for the simple reason that I have only flown about6 times and frankly am not sure if I did anything to cause it. I am going to go over your post andthe log more in depth tommorow on my PC.
Until then once again thanks, its people like you that make these types of forums invaluable

No worries, you're welcome.

Once again, I don't think you did anything to cause it, but that doesn't mean you won't end up with the liability for "bad luck" events (e.g. weather/birdstrike).  DJI may take responsibility if it can identify a component failure etc or other reason that the system wasn't operating properly, and made them responsible.  They are generally pretty good at taking blame if they are responsible.

When you look at these events under a microscope, the graphical "flight replay" on the app isn't really detailed enough to examine.  The usual way is to upload the logs to PhantomHelp and have the site decipher it.  Even the summary displayed on their web page isn't always enough, the best tools are to click on the links under the map and download the csv file (comma separated value data) and open it in a spreadsheet like Excel, and the kml data (opened and viewed in 3D overlayed in Google Earth).  I add some colour and "pretty it up" with bar graphs to make the data more visual and easier to interpret. I hide a lot of irrelevant columns, but the data contains dozens of parameters, recorded every tenth of a second.  For a completely forensic analysis, the DAT file in the aircraft records many more parameters in finer increments.   It isn't worth viewing them most of the time because it can be recording many esoteric parameters in way too much detail, to become confusing.  This flight was so short and simple that it is perfect for learning how to read the data.
2018-2-20
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

endotherm Posted at 2018-2-20 02:41
No worries, you're welcome.

Once again, I don't think you did anything to cause it, but that doesn't mean you won't end up with the liability for "bad luck" events (e.g. weather/birdstrike).  DJI may take responsibility if it can identify a component failure etc or other reason that the system wasn't operating properly, and made them responsible.  They are generally pretty good at taking blame if they are responsible.

Thanks Endotherm, I believe I have a full understanding of your analysis. I do however have a couple of questions. But first I should explain the flight path as shown on google earth. From the start point looking forward along the path. The sat picture is from spring last year. To the left is my house, the start point where I am standing is concrete pavers, this is where I calibrated compass, and to the right is a small wooden greenhouse going forward where the a/c does a small zig zag (that I do not recall happening) is a 6 car garage that is metal, I fly between 2 trees and over a wood horse fence, the next building on left of flight path is the barn with metal roof then my workshop which is all metal.
Regarding compass, I can understand that if the compass has an issue the a/c would not know which direction it is facing but I do not understand how it would be an issue if you are flying visually. My understanding is it would require the pilot to input yaw  to correct the deviation, unless we are considering the compass swinging extremely wildly. Now if you were flying auto pilot ie waypoints etc I could see an issue.
I downloaded the CSV file. Would I be correct that roll is the side to side movement of the a/c? and if so a negative figure equates to movement to the left? It would seem that there is a continuous bias left.
I found looking at the google earth track from ground level view interesting, it shows movement left slight gain in altitude with movement backwards then down, going forward and backward until hitting the ground.
You mention stick sensitivity, I presume this is adjustable in the app?
Now even though from the data it appears that at time of incident I was not doing any inputs, and though it happened very quickly I find it difficult to believe that I would have had no reaction, even if that reaction would have been incorrect. Because of this lack of control data I am left wondering if I lost connection between a/c and remote? I do know from the time of take off to after I had removed the a/c battery I did not look at the app or remote at all. Hopefully DJI tech will shed some light on it.
Once again thanks for your time and effort in this, I appreciate the knowledge you have passed on to me.
Regards

Mick Steele
2018-2-20
Use props
endotherm
First Officer
Flight distance : 503241 ft

Australia
Offline

Mickus Posted at 2018-2-20 06:02
Thanks Endotherm, I believe I have a full understanding of your analysis. I do however have a couple of questions. But first I should explain the flight path as shown on google earth. From the start point looking forward along the path. The sat picture is from spring last year. To the left is my house, the start point where I am standing is concrete pavers, this is where I calibrated compass, and to the right is a small wooden greenhouse going forward where the a/c does a small zig zag (that I do not recall happening) is a 6 car garage that is metal, I fly between 2 trees and over a wood horse fence, the next building on left of flight path is the barn with metal roof then my workshop which is all metal.
Regarding compass, I can understand that if the compass has an issue the a/c would not know which direction it is facing but I do not understand how it would be an issue if you are flying visually. My understanding is it would require the pilot to input yaw  to correct the deviation, unless we are considering the compass swinging extremely wildly. Now if you were flying auto pilot ie waypoints etc I could see an issue.
I downloaded the CSV file. Would I be correct that roll is the side to side movement of the a/c? and if so a negative figure equates to movement to the left? It would seem that there is a continuous bias left.

The flight path you describe is exactly how the kml data records the flight.  I also examined the historical imagery to see if there were any clues from previous structures which may affect the present.  I notice you took off from what looks like the site of a swimming pool or tank that has been removed.  That could indicate underground pipes and cabling.  Otherwise you were well clear of any metal or structures.
pool.png

I should point out that you did not appear to have a compass problem.  There were no error messages recorded at all in the log file.  The discussion regarding compass errors was to answer a lot of speculation that was being posted earlier on.  It was just to clarify what they were, and what you might see as a result of bad calibration or interference.  This wasn't applicable to your flight, I'm sorry if the explanation wasn't clear in that regard and worried you unnecessarily.

Your little zig-zag occurred around 54 seconds to 62 seconds of the flight, and was as a direct result of slightly rocking the right stick right and left.
zigzag.png

Many people panic when they see compass error warnings.  You are right, it really just indicates that the aircraft is confused about its orientation, or the data doesn't correspond with data from the other sensors.  It is the app's way of telling you the data is currently unreliable (though it may still be correct!) and the auto-navigation could be off.  It doesn't indicate it is about to fall from the sky.

The "roll" is the bank or tilt of the aircraft, traditionally the dipping of the wings.  The roll angle shows the tip of an imaginary left wing with respect to the fuselage, so positive angles mean the left wing goes up, negative values mean it goes down.  The sides with respect to the front of the aircraft is always fixed, despite the X-profile craft moving backwards/sideways or some arbitrary angle.  The accelerometers on the main board are fixed in the X/Y/Z directions with respect to the body.

In still air, a negative figure equates to the left side dipping down.  That should result in a "crabbing" movement laterally left (the props are pushing air to the right).  However, the aircraft may roll to oppose incoming wind, and maintain its position or course.  So just because you see the values change doesn't necessarily mean the aircraft is moving sideways.

The same explanation applies to pitch (the nose going up and down, positive angles means up, negative angles are down).

The yaw value shows the compass heading angle with respect to North.  Negative angles represent counter-clockwise moves through to South (-180°).  Positive angles are clockwise from North through South (0-180°).  So a yaw of -90° would mean the aircraft is facing West.  You could keep moving clockwise, increasing from 0 to 180,  Continuing clockwise would then see the angle decreasing with a negative number, e.g. 178, 179, 180, -179, -178...   This number is completely dependent on what you are using to decrypt the flight data.  This description is true for Phantom Help.  Using AirData to decrypt the same information will return positive compass headings of 0 through 360°.

What you are seeing as a left bias (-1° or -2° or so) is probably in response to a light breeze coming in from the N/E (aircraft left side), so it banks left a bit to keep flying in a straight line, without being blown to the right.  This is in P (GPS) mode.  In A (ATTI) mode, the breeze would indeed be moving the aircraft right and blown off your straight line course.  You would need to manually push the stick left to cancel out that force.  It's not an easy thing to do, as the wind keeps changing.  In GPS mode, the aircraft is making LOTS of corrections for you that you don't even notice, making the aircraft incredibly easy to fly and maintain position.

During the crash and tumble, the aircraft is changing position and orientation very quickly, quicker than the normal point-to-point distance of the normal flight track.  As a result, the track might appear like a lightning bolt (zigzag) and going backwards etc.  The positioning might even disappear (e.g. the 1:29.5 data point is missing)  when the aircraft is upside down and the GPS can't see the sky.  The mapping software will then just join the dots between two known recorded positions.  It doesn't necessarily mean you flew in straight line between them, it's just that there is no further data to show where you really did go.  Your track is probably indicative of flipping backwards and tumbling.  Nice and "curvy", but the kml track looks "pointy".  The track shows where the GPS module was, not the landing gear which could be forward or backwards or side to side or upside down etc... Your aircraft centre-of-gravity probably fell straight down in a straight line (draw an imaginary trend-line through the points of your "lightning bolt").

After centering a stick movement, you will see a short burst of power and change of angle in the opposite direction.  This represents the GPS mode trying to stop movement in the last direction and remain stationary.  Think "retro-rockets" on the lunar lander, or SpaceX boosters on the Falcon (depending on your age ).  This is also known as active braking.  Depending on how fast and powerful it does it will result in how quickly it comes to a stop.    A light, longer burst will result in a slower but smoother brake.  A quick heavy brake will be rapid and it will stop on a dime.  However this is analogous to stomping on the brakes in your car -- it can do it, but braking like that all the time is likely to take its toll eventually with rapid wear or something breaking.  This can be modified by adjusting the sensitivity of your sticks.  You'll want to look for "Gain& Expo Tuning" in the "MC Settings" in the Go app.  This is about the Mavic, but it is explained pretty well in this video or here    I couldn't find a modern one specifically for the Phantom.  Just from looking at your flight data numbers, it looked a bit "severe" to me, but it might be the factory setting and quite normal in reality.  Up to you if you change it, but most of us do once you get more experienced, usually to smooth out rotations/pans while you are recording video.

During your crash, you actually applied a little bit of forward stick.  This could have been unintentional or subconscious.  You can see that in the spreadsheet screen grab in my post #30, at the bottom of the image -- column AJ:798 to AJ:810.  These four columns AH to AK are added in by me to the spreadsheet downloaded from PhantomHelp.  This is so I can view bargraphs representing stick inputs, recorded in hidden columns of numbers in AC to AG.  If you see no grey bar graph, that doesn't mean there is no data.  It is in fact recording a value of 1024, which represents a centered stick.  So you still have control data and a connection to your aircraft.  If the connection is lost, you will be missing entire lines of data, best identified by looking at the time column.  You want to see 0.1 second intervals, 10 lines = 1 second of flight.  The lines of data will resume once the connection is re-established.  You can then see the missing time and determine if the connection was broken and for how long.

Hopefully that explains your questions.
2018-2-22
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

endotherm Posted at 2018-2-22 03:29
The flight path you describe is exactly how the kml data records the flight.  I also examined the historical imagery to see if there were any clues from previous structures which may affect the present.  I notice you took off from what looks like the site of a swimming pool or tank that has been removed.  That could indicate underground pipes and cabling.  Otherwise you were well clear of any metal or structures.
[view_image]

Endotherm, I really like you. Your explanations are concise, make sense, and are based upon data rather than conjecture. This does not mean I think I am not at fault I still find it hard to believe that I would not have reacted when the A/C veered left, and also I do not see any data to indicate any attempt to shut down the motors, i.e. left stick down. This leads me to believe some loss of communication possibly between a/c and remote but definitely with the app. I am presuming there will be more data available to the DJI techs.
The google earth image that you have shows a 16ft swimming pool that my wife makes me put up every year for the grand children. My confusion about the compass was in not understanding the relationship with the other sensors, but even then I would expect that unless the compass was spinning wildly you would still be able to gain control by correcting compass error with the sticks.
I watched the video regarding stick sensitivity  a couple of days ago and thought it was a very good explanation of not only how to, but also the benefits for adjusting, my sticks are currently set however  they come from factory and are very sensitive.
When I get further information from DJI regarding the incident I will post here.
This has been a good learning process for me and I appreciate your input. All I need now is an A/C to fly!
Regards

Mick Steele
2018-2-22
Use props
endotherm
First Officer
Flight distance : 503241 ft

Australia
Offline

Mickus Posted at 2018-2-22 05:19
Endotherm, I really like you. Your explanations are concise, make sense, and are based upon data rather than conjecture. This does not mean I think I am not at fault I still find it hard to believe that I would not have reacted when the A/C veered left, and also I do not see any data to indicate any attempt to shut down the motors, i.e. left stick down. This leads me to believe some loss of communication possibly between a/c and remote but definitely with the app. I am presuming there will be more data available to the DJI techs.
The google earth image that you have shows a 16ft swimming pool that my wife makes me put up every year for the grand children. My confusion about the compass was in not understanding the relationship with the other sensors, but even then I would expect that unless the compass was spinning wildly you would still be able to gain control by correcting compass error with the sticks.
I watched the video regarding stick sensitivity  a couple of days ago and thought it was a very good explanation of not only how to, but also the benefits for adjusting, my sticks are currently set however  they come from factory and are very sensitive.

The time it took between when it started to flip and the end of the record is 1 second, so I didn't expect to see much reaction in that time.  It would have crashed before you knew what was going on.  We did see a slight right stick forward movement while it was crashing, which wouldn't have affected it at all.  After it crashed, the record in your controller stopped, which is pretty typical.  That's usually because the antennas in the legs are pointing away from you from the crash and not easily visible.  The transmission is obscured by the body and your controller/app considers the flight is over, so stops.  Sometimes we see the data record for ages after crashing, but usually it stops.  The aircraft keeps recording parameters internally until such time the battery is turned off or removed.  DJI can access and view that, but it isn't going to add anything to the analysis -- the crash has already happened by that stage.  When you got to the crash site, were the motors still spinning or had it shut down?  Possibly the aircraft saw it was stationary and the props were obstructed, so it considered itself landed and turned itself off?  There may not have been anything further recorded in the app/controller if you put it down on the ground after the crash and went to the crash site to turn the aircraft off.

It doesn't really matter too much how you reacted just before the crash in response to the flip.  What we need to determine is why it randomly went out of control without you doing anything to initiate the flip.  That is the key incident to be determined in this case.  It may have been an ESC chip that just decided to fail, sometimes new electronic components just die suddenly.   DJI will see that when they power it up and test it, and that will be a warranty condition.
2018-2-22
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

endotherm Posted at 2018-2-22 11:46
The time it took between when it started to flip and the end of the record is 1 second, so I didn't expect to see much reaction in that time.  It would have crashed before you knew what was going on.  We did see a slight right stick forward movement while it was crashing, which wouldn't have affected it at all.  After it crashed, the record in your controller stopped, which is pretty typical.  That's usually because the antennas in the legs are pointing away from you from the crash and not easily visible.  The transmission is obscured by the body and your controller/app considers the flight is over, so stops.  Sometimes we see the data record for ages after crashing, but usually it stops.  The aircraft keeps recording parameters internally until such time the battery is turned off or removed.  DJI can access and view that, but it isn't going to add anything to the analysis -- the crash has already happened by that stage.  When you got to the crash site, were the motors still spinning or had it shut down?  Possibly the aircraft saw it was stationary and the props were obstructed, so it considered itself landed and turned itself off?  There may not have been anything further recorded in the app/controller if you put it down on the ground after the crash and went to the crash site to turn the aircraft off.

It doesn't really matter too much how you reacted just before the crash in response to the flip.  What we need to determine is why it randomly went out of control without you doing anything to initiate the flip.  That is the key incident to be determined in this case.  It may have been an ESC chip that just decided to fail, sometimes new electronic components just die suddenly.   DJI will see that when they power it up and test it, and that will be a warranty condition.

Thanks Endotherm. I did put the remote down after attempting to stop the rotor  without success. The A/C was upside down in the grass with rotors spinning, I had installed the DJI prop guards which I feel helped prevent further damage in this incident. I attempted to shut the A/c down using the battery power button (3 times) without success. I then picked the A/C up and while trying to remove the battery, which took some time, I noticed the motors speeds were hunting in unison i.e. high revs to low revs.
I will be interested to hear what DJI finds, probably sometime next week. Hopefully they will share the data they collect but I doubt it. Whatever they inform me of I will post when I get it. It has been a good learning experience for me.
Regards
Mick Steele
2018-2-23
Use props
DJI Susan
Administrator
Offline

Hey Mick, could you tell us your case number? I'd like to check the exact status and see whether I can help.
2018-2-25
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

DJI Susan Posted at 2018-2-25 23:00
Hey Mick, could you tell us your case number? I'd like to check the exact status and see whether I can help.

Hi Susan, the repair shop got the a/c last Friday, #CAS-1480398-Y5L8S7. I expect they will start an initial investigation today and contact me when they have some results. The repar shop is in California, Thanks for your interest,
Regards
Mick.
2018-2-26
Use props
Mickus
lvl.3
Flight distance : 601677 ft
United States
Offline

Mickus Posted at 2018-2-26 01:24
Hi Susan, the repair shop got the a/c last Friday, #CAS-1480398-Y5L8S7. I expect they will start an initial investigation today and contact me when they have some results. The repar shop is in California, Thanks for your interest,
Regards
Mick.

As a conclusion to this thread, I received an email from DJI support, "warranty per data analysis". They do not share the data but from the repair parts is included "Wi-Fi main board and Wi-Fi left antenna". So I am assuming a loss of control connection is what they think occurred.
I would like to thank everyone who responded, especially Endotherm who took the time to interpret the data and give me insights into data analysis which has given me increased knowledge and understanding.
Regards
Mick Steele
2018-2-27
Use props
12Next >
Advanced
You need to log in before you can reply Login | Register now

Credit Rules