Bought a mini-3, which crashed on the first flight itself.. Had decent clearance from neighboring buildings, and was only flying at a 17 feet height, when suddenly it started giving errors on "Position mismatch / NGPS signal weak" and instead of hovering or RTH descend, it flew sideways and crashed (far from the Home spot), and broke the front arm.

Wondering what the issue is, and if it is a DJI issue.

Hello there!  I'm sorry to hear about the incident involving your DJI Mini 3. We advise sending the device in for inspection and a proper diagnosis. This will also help us ensure the future flight performance of the drone, as it will undergo flight tests and quality assurance. Please get in touch with the relevant team at https://www.dji.com/support for more information regarding the request to open a case and launch an investigation. Thank you!

Is the Phantomhelp map reasonably accurate? The map suggests you took off from inside or on top of Salesforce Tower but I doubt that is correct.
The GPS was never good, if it doesn't know its location with confidence it can not RTH.
As for ''or RTH descend,'' it probably couldn't, see above.
Besides why should it? There was no loss of connection to trigger any sort of failsafe behaviour.

From PRECISELY where did you take off?
What was the lighting like?
A Google suggests you took of 1hr 18min after sunset is that correct?

Why did you take off without Home Point set? You had GPS warnings from the start. You should NEVER have taken off.  Pilot error all the way!

What were the winds like?
The following shows some pitch and roll behaviour that I do not think is explained by the elevator and aileron inputs.

RTH point was set - GPS worked or was low (which is how it allowed the manual takeoff).. Winds were maybe 10-15 miles per hour, and visibility was decent (lot of ambient light)

ShubhankG Posted at 8-17 10:30
RTH point was set - GPS worked or was low (which is how it allowed the manual takeoff).. Winds were maybe 10-15 miles per hour, and visibility was decent (lot of ambient light)

Sorry, but according to the log, the Home Point was NOT set. The RTH altitude was, but not the HP. FYI, there would have been a H location on the map view. Also, you only has a maximum of 9 satellites. PILOT ERROR!!!

ShubhankG Posted at 8-17 10:30
RTH point was set - GPS worked or was low (which is how it allowed the manual takeoff).. Winds were maybe 10-15 miles per hour, and visibility was decent (lot of ambient light)

"(which is how it allowed the manual takeoff)." I can manually launch  a DJI drone with 0, zero, satellites, if fact I often must to get it up it where it can see satellites.

Please explain what you mean by "RTH point was set "?
I see no evidence of a home-point being set or known. If you notice Phantomhelp gives no indication of distance from A home-point at any point.

Is the Phantomhelp map reasonably accurate?
From PRECISELY where did you take off and precisely where did the drone end up?

Sean-bumble-bee Posted at 8-17 10:43
"(which is how it allowed the manual takeoff)." I can manually launch  a DJI drone with 0, zero, satellites, if fact I often must to get it up it where it can see satellites.

Please explain what you mean by "RTH point was set "?

No the map is not accurate - Would be maybe 30 feet off.. I took off from the garden outside the salesforce tower (on ground), and it ended up crashing about 20 feet away from the start point into the ground. All of this is outside the SalesForce tower (not inside per the map)

I guess my question is how did it just automatically swerve sideways and crash without any trigger from the RC. Even if the home point was not set, what triggered the random move in one direction? Is that not because of a drone bug / firmware issue? How can I avoid that next time?

Hi,

Guess you never saw YT video about drones or read the manual few times.

To fly in a stabilized GPS mode you need enough satellites with good reception.
Top right GPS icon:

• Red = not enough sats to get a GPS lock
• Yellow = enough sats to het a GPS lock
• White = as yellow + HomePoint has been set.
  

In your flight the GPS reception value 2 (range 0 - 5) ; 4 or 5 is needed.
Number of sats most of the flight only 8.

Warning in the FlyApp at start :

• GPS signal weak. Hovering unstable. Fly with caution
• No GPS. Fly with caution
  

Without a good GPS lock DJI drones can do funny things.....
Hover instability and rapid moves on RC inputs, without GPS lock no braking.No GPS lock = drifting on the wind.
With good visual reference using the VPS bottom sensor it should be in a steady hover using this sensor, why it did not hover steady is the question !
Normally "low ambient light" warning are in the log, when the Visual sensors need more light to function, not in your log!Sometimes a GPS mismatch can cause a kind of fly away, drone try to correct its moved position but the position is not a steady received value.
Log shows that GPS is used and not used, that is why the GPS mismatch is not an all records.

EDIT see post #26   as well.

Exapmple of distance between 2 GPS points (0.1 second between recording)

• record 392 > 393 : 0.2 meter 0.7 feet
• record 393 > 394 : 0 meter 0 feet
• record 394 > 395 : 0.2meter 0.7 feet
• record 395 > 396 : 0 meter 0 feet
• record 396 > 397 : 10.7 meter 35.1 feet  /. speed 100 m/s   nah  incorrect GPS positions !
• record 397 > 398 : 0 meter 0 feet
  

See my chart of your data, in the polar chart the yellow line shows the movement without inputs.
As you can see the distances 'flown' are incorrect.

This flight should not be flown, as you already understand.

How to avoid next time : (the 2 most important checks)

• Wait for white GPS icon top left in the FlyApp.
• Before takeoff check the compass using the blue drone arrow in the map view, compare heading with the actual heading. Not the same : do not fly!
  

cheers
JJB   [ Charts by FRAP ; the 'best' software for data and visual flightlog analysis (www.jjbfrap.eu) ]

I think you flew the drone without sufficient GPS for the drone to establish ACCURATELY both its own position and the position of A home point.
With the drone not knowing its own position the drone would rely on the VPS sensors to hold position. Low down, above a suitable surface and with sufficient light, their position holding is FAR better than that of GPS.
Given the indicated heights, I suspect the drone was low enough for VPS to work, the question then becomes was the surface suitable and was the lighting sufficient?
That I can not answer.
EDIT, is this the photo you took around 41 seconds?  If so I would think that is getting to lighting levels where the grounds is insuffiently lit and the flooring pattern perhaps inappropriate in those lighting conditions.

In light of this, is it better to just do a replacement vs send to DJI for repair? Repair cost will likely be ~$25 or $0 (if DJI doesn't do it for free post the checks).. Replacement cost would be $40 (but will eat up 1 of the replacements possible as part of refresh care program)  

Sean-bumble-bee Posted at 8-17 11:02
I think you flew the drone without sufficient GPS for the drone to establish ACCURATELY both its own position and the position of the home point.
With the drone not knowing its own position the drone would rely on the VPS sensors to hold position. Low down, above a suitable surface and with sufficient light, their position holding is FAR better than that of GPS.
Given the indicated heights, I suspect the drone was low enough for VPS to work, the question then becomes was the surface suitable and was the lighting sufficient?

Yes - correct. So this lighting is not enough for the VPS sensor to work?

ShubhankG Posted at 8-17 11:26
Yes - correct. So this lighting is not enough for the VPS sensor to work?

I suspect that that is correct and or the surface beneath the drone was inappropriate for the lighting conditions.
Make no mistake, over an  appropriate, well lit surface VPS position holding is VERY good but it does need an appropriate surface and proper lighting.

Sean-bumble-bee Posted at 8-17 11:02
I think you flew the drone without sufficient GPS for the drone to establish ACCURATELY both its own position and the position of A home point.
With the drone not knowing its own position the drone would rely on the VPS sensors to hold position. Low down, above a suitable surface and with sufficient light, their position holding is FAR better than that of GPS.
Given the indicated heights, I suspect the drone was low enough for VPS to work, the question then becomes was the surface suitable and was the lighting sufficient?

Hiya,

Weird,  not to see any low light message warning in the log.

cheers
JJB

JJB* Posted at 8-17 11:37
Hiya,

Weird,  not to see any low light message warning in the log.

In light of this, is it better to just do a replacement vs send to DJI for repair? Repair cost will likely be ~$25 or $0 (if DJI does it for free post the checks).. Replacement cost would be $40 (but will eat up 1 of the replacements possible as part of refresh care program)  

I’d not recommend a beginner pilot to flight a drone relying only on VPS sensor. Additionally, I’ll say that your drone didn’t get GPS lock due to the tall buildings around your flying spot. My only advice for you  is to take some time and read the user manual and the quick start guide of your product. In there you will find valuable information that will help you to avoid committing these mistakes in the future.

I would also recommend you go to a clear open space and practice. In the city is for advanced, experienced pilots. Read the manual several times before you try again.

JJB* Posted at 8-17 11:37
Hiya,

Weird,  not to see any low light message warning in the log.

Would there be any information in the DAT?
I don't know if the mini 3's DAT is readable and I haven't looked for it is a Mini 3 Pro's DAT.
I don't remember what my Mini 1 & 2 do.

Probably crashed due to no or little GPS lock at the time of take-off, when it did finally get a lock, the drone probably thought it should have been at a different position and decided to go there, hence the crash.

You've removed the link so i can't see the log now.

One other that can make it act similarly is a compass error caused by ferrous metals near the drone at the time of turn on/power up sequence. As no one has picked up on that then its most likely the GPS issue as already mentioned.

Here in the U.S., San Francisco has the highest incidence of multipath, but obviously other large cities with tall buildings can give you (your GPS) multipath errors.  

DroneApe Posted at 8-17 20:29
Here in the U.S., San Francisco has the most incidences of multipath, but obviously other large cities with tall buildings can give you (your GPS) multipath errors.  

That wasn't his problem at all.
But hey .. why not make random guesses without seeing or understanding any relevant flight data?
I mean .. it's not like you have any credibility to lose.

ShubhankG Posted at 8-17 11:39
In light of this, is it better to just do a replacement vs send to DJI for repair? Repair cost will likely be ~$25 or $0 (if DJI does it for free post the checks).. Replacement cost would be $40 (but will eat up 1 of the replacements possible as part of refresh care program)

Hi,

Had another look at your log, and checked few things in my log database. (similar cases)

You started this flight with  8 satellites and a 2 for 'GPS quality'.
So the top left GPS icon FlyApp was yellow =  GPS lock.
Seen also in the log by GPS used value (true) and the flightmode as P-GPS **
Without a GPS lock the flightmode had been ATTI mode.
If in ATTI mode a warning and audio is there.

** can you post the video of this flight, the video playing back to flight in the FlyApp?Love to see wich FlightMode is shown on the video.
Log shows that the GPS is used and not used, but not used means ATTI mode!  good question for DJI to answer.

Because flying between buildings the GPS position was not a steady position, it changed by meters. (see my post #12)
This fluctuating of GPS positions was seen by the FlyApp ; few times a "GPS Position Mismatch" in the log.
This is seen when the next GPS position is not possible as the drone did not move actual to that position.

IMO FlyApp software should not only recognize this but should also bring the drone in ATTI mode, if happens not just in 1 or 2 records.
This will prevent that the drone starts to move on its own, chasing the GPS positions.

Fair chance imo to claim this error as a warranty.

About hover precision ; with a GPS lock GPS is the main info for keeping it steady at position in a hover. The visual system will aid this stability but is not in the lead!
Only in ATTI mode and Visual sensors active (withing sensor range and good lightning) the Visual system provides good hover precision.
In your flight no warnings about low light conditions, so the visual system was active, but as said, GPS will overrule this info.

Seen this in your log in the same record " No GPS. Fly with caution | GPS signal weak. Hovering unstable. Fly with caution"
Uh? No GPS is not the same as GPS signal weak !!
No GPS = ATTI
GPS signal weak = P-GPS but not a steady hover....
So what is it DJI ??

cheers
JJB

Did you fly inside the building?
I ask because your flight was not very high, if yes then this would also cause GPS issues.
If not, then just the height of the buildings will also block a lot of the satellites.
If the drone thinks its not where its supposed to be when it gets a lock it will become erratic in nature

ShubhankG Posted at 8-17 11:22
In light of this, is it better to just do a replacement vs send to DJI for repair? Repair cost will likely be ~$25 or $0 (if DJI doesn't do it for free post the checks).. Replacement cost would be $40 (but will eat up 1 of the replacements possible as part of refresh care program)

Hi,

I have a question, afaik you calibrated your compass before this flight.

If i look at the picture, taken at approx 41 second in flight, i see bulidings behind you.

But if i position your drone on the map, see picture, than the heading of your drone is not towards the building. Drone heading is 220 degrees, so heading is the opposite direction.When did your drone started to fly away ?  if at the end of this log than.....a compass / yaw error.

cheers
JJB

JJB* Posted at 8-18 10:27
Hi,

I have a question, afaik you calibrated your compass before this flight.
[Image]

Started flying away towards the very end.. you can see post 1 min the altitude dropping and right speed at 4m/s and left at 0

ShubhankG Posted at 8-18 11:49
Started flying away towards the very end.. you can see post 1 min the altitude dropping and right speed at 4m/s and left at 0

Hi,

oke, you did not say anything about the take off position and my remark about the heading and compass calibration.

It looks like that a compass/yaw error causes the fly away.
Because the compass was offset by approx 180 degrees. Drone heading taken picture not the same heading in the flightlog.

To prevent this in all your next flights, check before takeoff the blue arrow in the map view of the bottom left view flyapp and compare this with the actual heading.
This is THE most important check before any flight!
Blue arrow heading and actual headinh not the same : do not fly!
Switch off drone, find another takeoff spot and power ON drone. Check again

Once drone comapss is good calibrated, no need to do this more times in next flights!

When in flight the drone position is changed (not by RC input) then old and changed position is known.To correct back to the old position drone compass data is needed.
Knowing heading of the drone = knowing wich motors gets more or less revs to move back to the old position.
If drone compass is not the actual heading than ofcourse the wrong motors are changed in revs.
So drone moves the wrong direction = more offset from old position = more incorrect correction = etc etc = fly away.


cheers
JJB     [ Charts by FRAP ; the 'best' software for data and visual flightlog analysis (www.jjbfrap.eu) ]

JJB* Posted at 8-19 00:43
Hi,

oke, you did not say anything about the take off position and my remark about the heading and compass calibration.

Hey there,

Thanks for sharing your opinion!  I hope you found the video in my previous post informative.

I found a couple of other multipath links for you.   I sure hope the DJI engineer gets back to you. Good luck!

https://forum.dji.com/forum.php?mod=viewthread&tid=279355

https://forum.dji.com/forum.php?mod=viewthread&tid=295599