Hi all,

My first crash after 101 flights with 112.325km distance and 12 hours 39 minutes flight time.
As usual, done all the preflight check and made sure well calibrated.
10 GPS Satellites. All looked fine in GPS mode and no error, take off manually not touching the right stick.
Upon lift, at around 30cm from the ground, the aircraft veered to the left sharply and crashed into a vehicle parked 1m away.
Checked the Go4 App Flight Record, no input on the right stick. My only input was 25% on the left stick upwards to ascend. It all happened within 5 seconds.
There wasn't any error in the flight record too.

I checked for damage, just minor scratches on the SD Card cover, replaced all the props since they are damaged, recalibrate (compass and IMU), retry to take off manually and it went smoothly as expected.

What could have happened? I'm confused.

wow.... sorry to hear about that. I don't have any clue about this. Let's wait answers from the experts....

Before the crash did you update the firmware to 0400 ?

Cyril-Fr Posted at 2018-5-9 04:49
Before the crash did you update the firmware to 0400 ?

Yup. Its 0400.

If you upload your flightrecord.txt mayby we can find out what went wrong.  http://www.phantomhelp.com/LogViewer/Upload/
Cheers JJB

I had a similar experience with my first flight after updating firmware yesterday, but no crash.  I took off, started flying straight, but the drone started veering off to the right..  It seemed to fix itself after a few seconds and i was able to complete a 16 minute flight with no issues, but I was lucky I was in an open area.  

Any warning messages in Flight Log?  
Any warning messages shown by GO 4 app prior to takeoff or prior to crash?

Can you post a link to Flight Log / data?

I already updated to 0400. Then I went to fly for a distance of about 7000 ft . After landed there was a Firmware update require (0400), WTH!   Updated via the Go4 app and flew for another total distance of about 15000 ft. Never had such experience on my other drones.

I feel that we are more likely to have had a crash or almost a crash since this patch 0400.
I haven't flight log cause DJI Goggles for my side . Could you check my video (some post under yours) and say if you're drone do the same thing ?

"Upon lift, at around 30cm from the ground, the aircraft veered to the left sharply and crashed into a vehicle parked 1m away."

This could explain some of it. Newer take off so close to any metal. It will always bee better to stay away from a car or any other iron if possible. Follow post #5 and upload your log.

JJBspark Posted at 2018-5-9 04:53
If you upload your flightrecord.txt mayby we can find out what went wrong.  http://www.phantomhelp.com/LogViewer/Upload/
Cheers JJB

Oh this sparks another question that I feel weird.

How come the flightrecords.txt stops logging after 6th May 2018?
I've flown on 7th and 9th, but the flightrecords.txt aren't there.

Before I started this thread, I have spent hours trying to refresh and reboot my device but couldn't find any flightrecords.txt nor the MCFlight .dat files.

Is this an update for the MA.

I saw lots of pilots have issue with 04 firmware. Should reset factory, update firmware (or refresh) including batteries, recalibrate everything quad and rc then try. I also noticed my friend has the same issue when he use go app update ota,but use assistant app on pc with above steps seem a fix.

My friends MP did that a few weeks ago - the reason was that he did not put one of his props on properly and it came off a few centimeters above ground height - fortunately long grass and open filed so no damage. Might be that one prop was not installed correctly?

Hi! It might have been a gust of air

Follow the advice of JJBSpark and use the link he referenced.  Uploading the flight records give more momentary detail than you could have captured by looking at your screen.  Also, focusing on whether a firmware update was a contributor, before looking at the logs is a waste of time.   

As an added thought, check your motors, especially after replacing damaged props.  Small pieces of prop (or other debris) can get lodged in the moving parts causing rotation resistance.  I speak from experience.  The tip of the post of a Spark prop chipped off.  I had taken off the props while updating firmware.  When replacing them I notice the corner of the post was too smooth.  I looked into the motor and saw the tiny chip.  Having missed it, it would have interferred with the motor.  If it could happen on a Spark, maybe an MP/MA.

Ok. The flightrecord.txt suddenly appeared in the phone storage. No idea why it wasn't there earlier.

Probably caused by Erratic GPS signal.
hSpeed was 1.13m/s despite the aircraft has yet to take off (motor starting).

So are you going to post the flight record or not?  I see you took a look at it.

http://www.phantomhelp.com/LogViewer/9KE56DMR4V8V426HQSV9/
Here you go.

felixlim Posted at 2018-5-9 06:53
http://www.phantomhelp.com/LogViewer/9KE56DMR4V8V426HQSV9/
Here you go.

To me it is quite obvious that your GPS link was too weak. Normally a minimum of 11 satellites (better more) are required for a stable GPS link. Another indication for that is the fact that there is no map shown above the flight record. Thus my assumption is that different from what is stated in the flightrecord your MA actually was in ATTI mode and was drifted by wind at take off.

I am not an expert like most of the others here, but as Wachtberger stated - you should have at least 11 satellites, so the fact that you only had 10 and then 9 would indicate that the wind got a hold of it...
See this thread for more information https://forum.dji.com/thread-138008-1-1.html

Hi,

Well IMO there is something else happening, the speed is already indicating > 0 before take-off.  Think i never saw seen that before in a flightrecord.
Meaning when lifting off the ground the GPS mode wants to compensate, that`s why it moved away!
HomeDistance increased too, guess combination of GPS latlong calculating at that time.

Or did you take-off from a moving platform?

cheers JJB

I observed many times that if there is some magnetic interference near the takeoff spot such that compass is off by 30-45 degrees, MA yaws very sharply after takeoff as soon as I release the left stick. Sometimes it happens less than a meter above the ground. Sometimes it happens at several meters if I hold throttle longer. But then it flies normally and even lands normally in the same spot.

felixlim Posted at 2018-5-9 05:31
Oh this sparks another question that I feel weird.

How come the flightrecords.txt stops logging after 6th May 2018?

Log doesn’t show anything, you say MavAir travelled a mile, so nothing recorded in your txt file, can you tell us what type of ground you launched from it looks like some concrete platforms but I can’t be sure?

Did aircraft take off in a big arc movement.

hallmark007 Posted at 2018-5-9 12:41
Log doesn’t show anything, you say MavAir travelled a mile, so nothing recorded in your txt file, can you tell us what type of ground you launched from it looks like some concrete platforms but I can’t be sure?

Did aircraft take off in a big arc movement.

It traveled 1 meter not 1 mile.

G_Sig Posted at 2018-5-9 12:47
It traveled 1 meter not 1 mile.

Sorry m stands for metre, that explains almost everything his drone was 1metre from a parked car which is one big hunk of metal , my money is firmly on magnetic interference caused by leaving drone down beside a vehicle.

felixlim Posted at 2018-5-9 06:53
http://www.phantomhelp.com/LogViewer/9KE56DMR4V8V426HQSV9/
Here you go.

Sorry about previous post, I thought 1m was 1 mile but now realize it was 1 metre, this is almost certainly a problem putting aircraft very close to what is a big hunk of metal causing compass heading to move because of magnetic interference coming from the car, you really should be taking off 10 metres from any metal objects and not on concrete which is likely to contain rebar.
Your app will not always warn you of compass problems in these situations.

小仙女的星巴克 Posted at 2018-5-9 05:11
I already updated to 0400. Then I went to fly for a distance of about 7000 ft . After landed there was a Firmware update require (0400), WTH!   Updated via the Go4 app and flew for another total distance of about 15000 ft. Never had such experience on my other drones.

Same experience here. Made two flights after updating fw and came back about 5 days later and was told my fw was only 44% installed.

Take off 1m from a vehicle - compass error for sure.

Will say that the accident is due to the pilot having takeoff before the GPS module has finished finding a good fix on the satellites. When the GPS module does not have time to initialize properly, it retains a position with a high inaccuracy in all directions, here 10 ft. I simply think the flight controller will move the to the place where it believes the drone should have been located. So the accident must be attributed to a pilot error, unfortunately.

But it's probably not entirely without DJI's fault, because I think it should not be possible to make an take off before the GPS module delivers a position with a high degree of accuracy.

The lesson of that will be, you do not have to take off before the kind lady announces that the homepoint has been set. At the same time, the top line of the APP has become green, indicating that it is now possible to fly. I would recommend people always to use the auto takeoff.

Regards Leif.

Landbo Posted at 2018-5-9 14:26
Will say that the accident is due to the pilot having takeoff before the GPS module has finished finding a good fix on the satellites. When the GPS module does not have time to initialize properly, it retains a position with a high inaccuracy in all directions, here 10 ft. I simply think the flight controller will move the to the place where it believes the drone should have been located. So the accident must be attributed to a pilot error, unfortunately.

But it's probably not entirely without DJI's fault, because I think it should not be possible to make an take off before the GPS module delivers a position with a high degree of accuracy.

Good guaranteed gps comes at 11 satellites, this aircraft had 10 before it took off even with only 10 it would not have shot off in one direction at worse it would be in opti mode with vision sensors more than able to get good fixed position on the ground, gps had nothing to do with this crash.

Landbo Posted at 2018-5-9 14:26
Will say that the accident is due to the pilot having takeoff before the GPS module has finished finding a good fix on the satellites. When the GPS module does not have time to initialize properly, it retains a position with a high inaccuracy in all directions, here 10 ft. I simply think the flight controller will move the to the place where it believes the drone should have been located. So the accident must be attributed to a pilot error, unfortunately.

But it's probably not entirely without DJI's fault, because I think it should not be possible to make an take off before the GPS module delivers a position with a high degree of accuracy.

You can't make explicit conclusions regarding GPS signal quality judging from the number of satellites alone. There is the "signal strength" indicator that shows signal quality and status ribbon. Pilot is safe to take off if the ribbon is green. If the signal strength indicator shows 4 out of 5 bars this means that signal quality is very good and it can be this way with as few as 6 satellites. The OP needs to replay his flight log in DJI GO 4 to see what signal quality they had. Phantomhelp doesn't report it.

Often deep in a river gorge, it can take me 10 minutes to acquire a decent lock due to the high, vertical sides of the river, so I'll often pop it up in the air with no lock, makes a big difference to acquiring just putting it up 10-20ft to grab sats. I've never had it skew off however even with only a few satellites, daylight helps with the VPS holding. I'm thinking this is interference with metal either in the ground or nearby vehicles.

hallmark007 Posted at 2018-5-9 14:37
Good guaranteed gps comes at 11 satellites, this aircraft had 10 before it took off even with only 10 it would not have shot off in one direction at worse it would be in opti mode with vision sensors more than able to get good fixed position on the ground, gps had nothing to do with this crash.

And then Mr. Haldmark did not think it's strange that the vessel's groundspeed is 2.4 mph from the moment the engines start up ?

The number of satellites does not determine how accurate a GPS shows one's position. It makes the satellite's location in the sky and if they can be seen by the antenna as well as whether the radio signal is reflected on its way down to earth.

It is hopeful neither a natural law that we all have to agree with Mr Halmark ?

Regards Leif.

asaw Posted at 2018-5-9 14:54
You can't make explicit conclusions regarding GPS signal quality judging from the number of satellites alone. There is the "signal strength" indicator that shows signal quality and status ribbon. Pilot is safe to take off if the ribbon is green. If the signal strength indicator shows 4 out of 5 bars this means that signal quality is very good and it can be this way with as few as 6 satellites. The OP needs to replay his flight log in DJI GO 4 to see what signal quality they had. Phantomhelp doesn't report it.

The OP needs to replay his flight log in DJI GO 4 to see what signal quality they had. Phantomhelp doesn't report it.

That's right, we can not see what DJIGO4 shows in Phantomhelp. But do you have an explanation of why the groundspeed is 2.4 mph from the moment the engines is turning  on ?

Regards Leif.

Landbo Posted at 2018-5-9 15:23
And then Mr. Haldmark did not think it's strange that the vessel's groundspeed is 2.4 mph from the moment the engines start up ?

The number of satellites does not determine how accurate a GPS shows one's position. It makes the satellite's location in the sky and if they can be seen by the antenna as well as whether the radio signal is reflected on its way down to earth.

I think you might have to agree with mr asaw as well as his comment equally makes sense, you also ignored that this aircraft has VPS opti mode etc and the fact that it was within 1 metre of potential magnetic interference, that it already had 10 satellites, needing only 6 for good gps lock so at worse it had 5 gps 5 glonass (worse case scenario)
I note you also ignored ground speed.


And I have to say sarcasm is the lowest form of wit.

Thanks for all your comments.

There was someone that asked did I take off on a moving platform? No. It was on a tarmac in a car park that was not very populated with vehicles.

Yes the VPS should have picked it up. I've flown from indoor to outdoor where GPS was from 0 to 12 suddenly and never had issues. It switched from Opti to GPS smoothly.

My guess is it could also be some fallen leaves on the tarmac because my friend told me the leaves were blown away when the rotors started. It could be possible that the aircraft has captured the leaves being blown away and tried to compensate for the positioning.

As for magnetic interference, I think it is highly unlikely as the aircraft rolled left, not yawed left. It is clearly visible in the log.

The GPS signal was no lower than 4 bars and the aircraft stayed in GPS mode all the while.

felixlim Posted at 2018-5-9 16:05
Thanks for all your comments.

There was someone that asked did I take off on a moving platform? No. It was on a tarmac in a car park that was not very populated with vehicles.

You can clearly see yaw movement in your log.

Yaw happened upon impact with the vehicle.

It’s recommended by DJI to calibrate the IMU after upgrade firmware. That maybe why issues occur on MA after upgrade because IMU not calibrate.