Not taking off near concrete or metal objects. Verified 11 satellites, unobstructed view of sky, Hp established, took off and SE immediately veered hard to left at 3' agl and crashed. Log shows Gps mismatch.

I wonder why Dji programming  doesn't allow gps to supercede compass information if that is the mismatch.

How can you be aware of a gps mismatch at the time, and how do you avoid this

Hi,

Best to get help/advice is to upload your flightlog for this issue

Use this link > https://www.phantomhelp.com/LogViewer/Upload/
Post the uploaded link on here.

Probably a compass - gyro conflict, this will often result in a fly away. GPS mismatch warning often bc of this fly away, it is not the cause of the fly away.

If it is a compass-gyro issue (log will show data for this) ; best way to check ALWAYS before take-off the compass heading in the FlyApp and compare to the acutal heading. If not the same, do not start a flight.

This compass heading check is the most important check before flying !!

cheers
JJB

Most of the time you see a GPS position mismatch message, it's a false alarm.
If you want an explanation of your incident, you'll need to post the recorded flight data.
It would also help if you can describe the surface you launched from.

Surface was wood deck with no metal or concrete near. Have taken off from the exact spot many dozens of times without incident.

Describing this as a false alarm is a bit of an understatement I'm afraid, as this SE had a mind of its own and veered hard to the left with absolutely zero pilot input.

Though no message requesting it before or after the crash, the compass and imu were then calibrated because that is all I knew to do. SE then took off from the same spot without incident.

For now I intend to keep a very close eye at takeoff, and will use Reagan's famous approach with Russia.  I'll trust but verify the compass before takeoff!

djiuser_UWwbJhkKO0dX Posted at 11-15 14:58
Surface was wood deck with no metal or concrete near. Have taken off from the exact spot many dozens of times without incident.

Describing this as a false alarm is a bit of an understatement I'm afraid, as this SE had a mind of its own and veered hard to the left with absolutely zero pilot input.

Surface was wood deck with no metal or concrete near. Have taken off from the exact spot many dozens of times without incident.
Wooden decks have steel screws, bolts and nails that can cause problems when they are close enough to the compass sensor.

But if you post your flight data it would confirm whether or not there was magnetic interference at the startup.
Go to: https://www.phantomhelp.com/LogViewer/Upload/
... where you'll find instructions to upload the data to that site and it will give a report
For help interpreting the data you can post a link for the report here.
Or just post the .txt file.

Describing this as a false alarm is a bit of an understatement I'm afraid, as this SE had a mind of its own and veered hard to the left with absolutely zero pilot input.
As I tried to explain to you, that message is usually a false alarm and could well have been in your incident.
You cannot automatically assume that the message had anything to do with the incident.
It's necessary to see the flight data to tell what actually caused the incident.

Though no message requesting it before or after the crash, the compass and imu were then calibrated because that is all I knew to do.
Then you don't understand anything about compass calibration.
Post your flight data if you want to learn from the incident and how to prevent similar occurrences.

Labroides Posted at 11-15 16:05
Surface was wood deck with no metal or concrete near. Have taken off from the exact spot many dozens of times without incident.
Wooden decks have steel screws, bolts and nails that can cause problems when they are close enough to the compass sensor.

Insults are not needed or helpful.
The SE raced sideways immediately upon arriving at 1 meter high, with zero inputs and HP well established. Have flown from the same spot over 50 times. Based on past experience do not believe that a few deck nails were a factor.

I am satisfied that compass and imu calibration corrected the problem and have passed on the experience so that some others might benefit if they encounter the same thing.

djiuser_UWwbJhkKO0dX Posted at 11-15 17:14
Insults are not needed or helpful.
The SE raced sideways immediately upon arriving at 1 meter high, with zero inputs and HP well established. Have flown from the same spot over 50 times. Based on past experience do not believe that a few deck nails were a factor.

Insults are not needed or helpful.
I haven't insulted you and I've tried to be help you find out what caused your incident and be able to prevent similar occurrences.
But it seems that it's not getting through to you.

The SE raced sideways immediately upon arriving at 1 meter high, with zero inputs and HP well established. Have flown from the same spot over 50 times. Based on past experience do not believe that a few deck nails were a factor.
The incident you described was a yaw error.
Based on no understanding of what a yaw error is or how it's caused, or looking at data that would show whether or not it was the issue, you've made an assumption that is almost certainly wrong.

I am satisfied that compass and imu calibration corrected the problem and have passed on the experience so that some others might benefit if they encounter the same thing.
Without knowing what caused the incident, you are satisfied that you've solved your problem?
You have no idea what you are dealing with.
It's almost certain that there was no problem with your compass or IMU and recalibrating them has not corrected the problem.

You could find out what the problem actually was and learn how to be a better flyer and prevent similar occurrences in future.
It's your choice ...

Hi there. Thank you for reaching out and we're sorry for the trouble that this has caused. Can you please confirm if you experience an error message on the DJI Fly app during this said incident? In addition, does the issue occurs in a different location? Please keep us posted. Thank you.

DJI Paladin Posted at 11-16 22:41
Hi there. Thank you for reaching out and we're sorry for the trouble that this has caused. Can you please confirm if you experience an error message on the DJI Fly app during this said incident? In addition, does the issue occurs in a different location? Please keep us posted. Thank you.

Hi Paladin,

Compass / Gyro conflict after take-off ( introduced at drone powering up) is never seen on the FlyApp screen as an error message.
Large data conflict = ATTI mode, confict below that threshold user experience a fly away....

Time for your engineers to do some thinking....

May i suggest a very simple solution : Show the compass heading in numbers in a prompt before take off!
If i read compass heading "090"  in a prompt and my drone is actual facing 185 ; i know i must not start this flight !!  

cheers
JJB

JJB* Posted at 11-17 00:41
Hi Paladin,

Compass / Gyro conflict after take-off ( introduced at drone powering up) is never seen on the FlyApp screen as an error message.

Thank you for your prompt response, JJB. I will coordinate this suggestion of yours with our team for further attention. Thank you for your continued support.

DJI Paladin Posted at 11-17 03:47
Thank you for your prompt response, JJB. I will coordinate this suggestion of yours with our team for further attention. Thank you for your continued support.

As mentioned earlier, Gps Mismatch has occurred with this SE just one time out of about 50 takeoffs from the same location. After recalibrating the compass and imu, the SE flew normally from the same location.
In addition to JBBs suggestion, would it be possible for DJI to make a small edit to the software allowing the GPS  to supercede the compass when there is a GPS mismatch?

Additionally it would be beneficial for DJI to tell us exactly what triggers the GPS Mismatch, because other posters on related  threads in the Mavic forum have said DJI has never explained the message. Not Sure if that is so.

djiuser_UWwbJhkKO0dX Posted at 11-17 06:54
As mentioned earlier, Gps Mismatch has occurred with this SE just one time out of about 50 takeoffs from the same location. After recalibrating the compass and imu, the SE flew normally from the same location.
In addition to JBBs suggestion, would it be possible for DJI to make a small edit to the software allowing the GPS  to supercede the compass when there is a GPS mismatch?

Hi,

GPS mismatch is often seen AFTER other errors !

So your problem is not a GPS mismatch but most likely a compass error.

DJI cannot fly without a working compass, so GPS use supercede over Compass is not possible.
With a large compass error drone will switch mode to ATTI.

But as you do not post your flightlog....hard to explain better without your data

cheers
JJB

JJB* Posted at 11-17 07:15
Hi,

GPS mismatch is often seen AFTER other errors !

It will be helpful if Dji themselves will explain GPS Mismatch.   While it may be true that Dji can't program the GPS to supercede the Gps, it seems that at present the compass may be currently programmed to supercede GPS.  But since we can't see their programming code, we are all making educated guesses, in good faith.

On thing DJi can do is they can program the drone to hover and allow stick control in the case of GPS Mismatch. In all the cases of GPS Mismatch that appear in the various forums, it seems this would be a vast improvement and save crashes.

djiuser_UWwbJhkKO0dX Posted at 11-17 07:39
It will be helpful if Dji themselves will explain GPS Mismatch.   While it may be true that Dji can't program the GPS to supercede the Gps, it seems that at present the compass may be currently programmed to supercede GPS.  But since we can't see their programming code, we are all making educated guesses, in good faith.

On thing DJi can do is they can program the drone to hover and allow stick control in the case of GPS Mismatch. In all the cases of GPS Mismatch that appear in the various forums, it seems this would be a vast improvement and save crashes.

ah oke,  you do not understand how a DJI drone uses compass and gps data and how the interact when one fails...

GPS mismatch will not make a drone to fly away, compass error does sometines .....
So no need for DJI to allow stick control in GPS mismatch, as stick control remains working as normal.
But i understand that you want a answer from DJI, just wait ....  ;-)
cheers
JJB

Hi JJB,

I think we agree that it would help if Dji could fill in some details on the programming behind Gps Mismatch.

In my case, I lost all stick control and anything but the usual hover after takeoff, so in cases where that occurs, surely it would help Dji drone owners, and would prevent crashes.

djiuser_UWwbJhkKO0dX Posted at 11-17 09:12
Hi JJB,

I think we agree that it would help if Dji could fill in some details on the programming behind Gps Mismatch.

GPS mismatch is seen when the calculated LatLon position (bc moving a drone from position) does not equal the actual GPS latLon position.

As said, often seen after other errors. In a fly away you will see also very high speed values in the log, wich are often incorrect, this if you know also the actual GPS speed.

Compass check heading before take-off is the most important check to perform.
If you do not share your log, check yourself  ; heading craft and if you remember the actual heading before take-off. 99% these 2 values differs a lot.....

cheers
JJB

Assuming that's correct, there may be an easy way for Dji to help prevent Gps Mismatch crashes.  They can revise their software to initiate immediate hover in the event of Gps Mismatch, and to allow stick control. I can't see why they wouldn't, but they may have higher priorities.

djiuser_UWwbJhkKO0dX Posted at 11-17 11:17
Assuming that's correct, there may be an easy way for Dji to help prevent Gps Mismatch crashes.  They can revise their software to initiate immediate hover in the event of Gps Mismatch, and to allow stick control. I can't see why they wouldn't, but they may have higher priorities.

, there may be an easy way for Dji to help prevent Gps Mismatch crashes.
You've assumed, without any evidence, that the GPS mismatch message caused a problem,
It's highly unlikely that it was a problem at all.
If you wanted to find the actual cause of your crash, that would be a simple matter to investigate the flight data which will probably point it out clearly.
But despite having this suggested multiple times by two of the people who could help you, you have no interest in getting to the bottom of your incident.

I can't see why you continue to discuss the incident when you have no interest in finding the facts.
But I see that you've done the same with your last incident too.
https://forum.dji.com/forum.php?mod=viewthread&tid=277255

That one also made it clear that you:
1.  Know very little about how your drone works
2.  Have no interest in finding out what caused your incident.

They can revise their software to initiate immediate hover in the  event of Gps Mismatch, and to allow stick control. I can't see why they  wouldn't, but they may have higher priorities.
If you don't want help, why keep yammering on about things you know nothing about?

djiuser_UWwbJhkKO0dX Posted at 11-17 11:17
Assuming that's correct, there may be an easy way for Dji to help prevent Gps Mismatch crashes.  They can revise their software to initiate immediate hover in the event of Gps Mismatch, and to allow stick control. I can't see why they wouldn't, but they may have higher priorities.

uh, there are no GPS mismatch crashes !!  only fly away (often result in a crash) from a data conflict between compass and gyro.

Powering up drone with a incorrect compass heading (due to metal closeby ect).
Initial gyro value gets incorrect compass heading (so compass = gyro)
After take-off the compass gets the actual heading ; away from the interference....
But gyro is not updated !
So conflict between compass and gyro data.

For explame ; compass is 90 degrees changed after take off.

If drone move backwards due to wind, than the correction must be forward ofcourse.
But correction is made to the wrong motors, bc of the 90 degrees data conflict.
Result in moving right or left (depends plus or minus 90 degrees change).
Result is drone not closer to its posistion but further away in different direction = more speed to the motors = more off set = etc etc.
Result is drone will fly to its max speed to try to fly back to the position....
As the calculated LatLon postion does not equal the actual position : warning GPS mismatch.

cheers
JJB

JJB* Posted at 11-17 12:37
uh, there are no GPS mismatch crashes !!  only fly away (often result in a crash) from a data conflict between compass and gyro.

Powering up drone with a incorrect compass heading (due to metal closeby ect).

Not withstanding the really good comments above, it seems that jjb, in a nice way, has roughly equated the gps mismatch to a fly away. It sure looked like one.

If I had attempted to fly from a cliff or canyon edge that day, the SE would have been likely lost if it had operated in the same way.

It's  reasonable to believe that if Dji would lock a drone in a hover, upon encountering the Gps Mismatch, they could prevent most or all crashes associated with Gps Mismatch.  It's absolutely something they can and should do, imo.

They have a failsafe for rc disconnect, and critically low battery, and imo  they should have a failsafe for Gps Mismatch.

djiuser_UWwbJhkKO0dX Posted at 11-17 14:58
Not withstanding the really good comments above, it seems that jjb, in a nice way, has roughly equated the gps mismatch to a fly away. It sure looked like one.

If I had attempted to fly from a cliff or canyon edge that day, the SE would have been likely lost if it had operated in the same way.

Why do you keep posting nonsense here?
You've completely ignored any advice from people who could help you.

You have no idea what you are talking about.
You have no idea how your drone works.
You have no idea what actually happened in your incident and no interest in finding out.
Without any facts or understanding, everything you post here is nonsense.

djiuser_UWwbJhkKO0dX Posted at 11-17 14:58
Not withstanding the really good comments above, it seems that jjb, in a nice way, has roughly equated the gps mismatch to a fly away. It sure looked like one.

If I had attempted to fly from a cliff or canyon edge that day, the SE would have been likely lost if it had operated in the same way.

oke, i will not continue to explain it again to you.
You not even understand it all for 1%

Happy many safe landings,
cheers
JJB

DJI management and engineering has spoken. They determine, as suspected, this is a manufacturing defect.

"Our engineer has finished the data analysis and the result is as follows:

1.     Unit was in GPS mode and was responsive to RC input. The unit too off with good GPS signal.
2.     At t=00:04s relative height=1.1m, distance from home=1.2m, unit had mismatch between GPS speed and position.
3.     At=00:21s, relative height=1.2m, distance from home=2.2m, unit recorded external impact.
4.     Unit was flying with mismatched GPS speed and position, which made it drift and crash.

Conclusion: The incident was caused by manufacturing factors. Therefore, it is concluded to DJI's responsibility. We will provide you with warranty service."

Further, They will consider a software edit to force 'hover" upon a gps mismatch, per DJI management.

Thanks to all who contributed!

djiuser_UWwbJhkKO0dX Posted at 12-5 06:14
DJI management and engineering has spoken. They determine, as suspected, this is a manufacturing defect.

"Our engineer has finished the data analysis and the result is as follows:

Good news for you!

As you did not share your flightlog it is difficult to understand DJI conclusion.
Very rarely a GPS mismacth is the cause of a fly-away, more often a error message after the fly away has started due to a compass errror.

But who am i do discuss the DJI diagnosis ....   ;-)

If they can change their software such that it will force a hover in case of GPS mismatch fly away, would be great! Next to change is force ATTI mode in case of a compass error fly away.

cheers
JJB

JJB* Posted at 11-17 00:41
May i suggest a very simple solution : Show the compass heading in numbers in a prompt before take off!  If i read compass heading "090"  in a prompt and my drone is actual facing 185 ; i know i must not start this flight !!  

May i suggest a very simple solution : Show the compass heading in numbers in a prompt before take off!
If i read compass heading "090"  in a prompt and my drone is actual facing 185 ; i know i must not start this flight !!  

Great idea.

djiuser_UWwbJhkKO0dX Posted at 11-17 07:39
It will be helpful if Dji themselves will explain GPS Mismatch.   While it may be true that Dji can't program the GPS to supercede the Gps, it seems that at present the compass may be currently programmed to supercede GPS.  But since we can't see their programming code, we are all making educated guesses, in good faith.

On thing DJi can do is they can program the drone to hover and allow stick control in the case of GPS Mismatch. In all the cases of GPS Mismatch that appear in the various forums, it seems this would be a vast improvement and save crashes.

If I understand you correctly.......
For the moment let's just say DJI did change the programming to give GPS precedence over the compass and let's imagine the compass goes screwy so GPS 'takes over'.
GPS gives the drone it's position 'on a map' but ........................ in what direction is it pointing and where is home etc. etc. etc.?

Sean-bumble-bee Posted at 12-5 10:14
If I understand you correctly.......
For the moment let's just say DJI did change the programming to give GPS precedence over the compass and let's imagine the compass goes screwy so GPS 'takes over'.
GPS gives the drone it's position 'on a map' but ........................ in what direction is it pointing and where is home etc. etc. etc.?

The most common FlyAway is when during powering the drone the compass and thus the gyro is set with a wrong heading value. (close to rebar, metal car roof ect)

The moment in flight, compass now oke heading bc away from the 'metal', if correction is needed to stay in stabilized postion the wrong motors gets corrected ; fly away in progress. Larger mismatch in heading = a more angry rapid fly-away.
The correction is based on gyro heading and not on compass heading....
It must be easy to check and compare both values, if fly away starts and the next corrections are larger than before...so this can be a trigger to kill the fly-away by stopping corrections and let the drone hover or get drone into ATTI mode.
Pilot can than try to fly drone back home manually, wich is obvious better that a uncontrolled fly away with max correction = max speed !
Change of manul recovery from this fly away = imo 1%
As the compass indication is oke, manual back home using the radar compass display is possible.

Seen also few flight where out of the sudden the compass heading goes 'funny', sometimes 180 degrees switch.
If this happens after many minutes in flight and compass data before is looking normal, it is just a hard compass error.
But not always does the drone switches to ATTI mode!.

The other FlyAway is a GPS mismatch one...(saw 2 flights over 5 years time...), same as above. make the drone just stop by killing the corrections.


BTW as we never saw the flightlog...hope that DJI is correct in their analysis.

cheers
JJB

JJB* Posted at 12-5 12:17
The most common FlyAway is when during powering the drone the compass and thus the gyro is set with a wrong heading value. (close to rebar, metal car roof ect)

The moment in flight, compass now oke heading bc away from the 'metal', if correction is needed to stay in stabilized postion the wrong motors gets corrected ; fly away in progress. Larger mismatch in heading = a more angry rapid fly-away.

ok & thanks