You've heard it all before: New user and a lost drone

[Awrig90]

It's as the title says really and any help you can provide would be greatly appreciated.

I flew my DJI Spark up to about 40ft and then started trying to make my way to my destination. A minute or so in and I suddenly get a whole load of warnings about weak GPS so I instantly hit RTH. The RTH times out almost immediately and soon after I completely lose connection. The flight log and DAT file are below, but bear in mind that it's final location is over a hillside covered in shoulder height bracken (heading towards the hillside, possibly crashed?)

We've been out looking today to no avail, but I'm wondering if our search area was big enough. I'm also wondering what exactly I did wrong as well. Thanks in advance!

Flight log
DAT log

 

[old man mavic]

welcome to the forum,hope you find your drone

 

[Physopto]

A gut wrenching moment I would guess. Hope you find it hopefully without too much damage or non at all.

 

[THE CYBORG]

Hi Awrig90 welcome to Mavic Pilots
All I can say is I hope you get it back & working.

 

[Meta4]

We've been out looking today to no avail, but I'm wondering if our search area was big enough. I'm also wondering what exactly I did wrong as well.

This one isn't going to be easy to pin down.
You had a yaw error, probably from launching from a magnetically influenced site.
In that situation, RTH couldn't work.

See the beautiful curved flight, that's typical of yaw error flights.
But with the curve repeating, it's not a simple matter to give you a point to go to.
Perhaps @sar104 could give a better idea?

The spark had plenty of battery and could have flown a long way.
However the Spark was heading for rising ground and probably ran out of altitude (very roughly) around here .. near the red arrow.

 

[Motorcycle Rider]

Welcome to the forum...that sucks about your drone.

 

[Awrig90]

Thank you all for your welcomes and condolences! And thanks @Meta4 for that rough area. If it's there we weren't searching even close to the correct location, instead following the curving trajectory and looking to the left of those trees. It's definitely worth another search I think, although I'm not holding my breath.

 

[sar104]

This one isn't going to be easy to pin down.
You had a yaw error, probably from launching from a magnetically influenced site.
In that situation, RTH couldn't work.

See the beautiful curved flight, that's typical of yaw error flights.
But with the curve repeating, it's not a simple matter to give you a point to go to.
Perhaps @sar104 could give a better idea?

The spark had plenty of battery and could have flown a long way.
However the Spark was heading for rising ground and probably ran out of altitude (very roughly) around here .. near the red arrow.

Apologies for being a bit late with this one. Actually the possible outcomes are fairly tightly constrained because the aircraft was in ATTI mode at the end of the flight. Failsafe in ATTI mode is immediate autolanding, and so it would have descended and landed on loss of uplink, which would put it on the ground within 10 meters or so of the last location, shown by the green circle below, and most likely in the continued direction of travel. That aside, the loss of downlink was very abrupt which suggests that it may simply have hit something. The ground was rising in that direction, and the log indicates that it was down to 10 m AGL or less. In any case, it could not have made it past the 102 m contour line shown in blue.

 

[slup]

So ATTI & entering failsafe equals landing ... yeah, logical when you think about it. Good learning, thank's @sar104

 

[Prismatic]

Apologies for being a bit late with this one. Actually the possible outcomes are fairly tightly constrained because the aircraft was in ATTI mode at the end of the flight. Failsafe in ATTI mode is immediate autolanding, and so it would have descended and landed on loss of uplink, which would put it on the ground within 10 meters or so of the last location, shown by the green circle below, and most likely in the continued direction of travel. That aside, the loss of downlink was very abrupt which suggests that it may simply have hit something. The ground was rising in that direction, and the log indicates that it was down to 10 m AGL or less. In any case, it could not have made it past the 102 m contour line shown in blue.

View attachment 105286

If this was a "yaw error" event, as has been suggested, I don't understand the three "hops" in the flight path (the last of which seems to have terminated on the ground). Each hop is individually consistent with yaw error, given the smooth curve of each.

But I thought once a yaw-error event commences, the drone will fly a continuously growing spiral flight path--determined by the angle of the error (a straight-line if exactly 180°)--until the flight ends (for better or worse). This looks like three linked yaw events, with two compass "resets" along the way. But that's absurd (I think.)

What's the ding-dang-deal? ?

 

[sar104]

If this was a "yaw error" event, as has been suggested, I don't understand the three "hops" in the flight path (the last of which seems to have terminated on the ground). Each hop is individually consistent with yaw error, given the smooth curve of each.

But I thought once a yaw-error event commences, the drone will fly a continuously growing spiral flight path--determined by the angle of the error (a straight-line if exactly 180°)--until the flight ends (for better or worse). This looks like three linked yaw events, with two compass "resets" along the way. But that's absurd (I think.)

What's the ding-dang-deal? ?

It looks like there were stick inputs that modified the course.

 

[Awrig90]

Thanks for the help everyone. I went back with a group to search for it yesterday and covered both @sar104 and @Meta4 suspected landing sites but came out empty handed, sadly.
Looks like this one is a goner, which is annoying as I still don't actually know what / if I did anything wrong or whether it was just a technical malfunction. Oh well, you live and you learn!

 

[slup]

Thanks for the help everyone. I went back with a group to search for it yesterday and covered both @sar104 and @Meta4 suspected landing sites but came out empty handed, sadly.
Looks like this one is a goner, which is annoying as I still don't actually know what / if I did anything wrong or whether it was just a technical malfunction. Oh well, you live and you learn!

The most possible cause for this are shown in @Meta4 post --> #5

If you attach the mobile DAT log ending with FLY150.DAT directly without uploading it to Airdata it should easily be seen there.

 

[Awrig90]

Thanks, I've attached that here

 

[slup]

@sar104 ... please validate my conclusion below when you have time, may very well be wrong here ? .

From what I see this isn't as easy as a take off from a magnetic disturbed spot, something else is going on ...

Looking at how the imuYaw (Black) & the magYaw (Green) follow each other it's not immediately apparent that the magYaw diverge from imuYaw at height. But something is going on from 21sec ... the agreement between Green & Black isn't good, instead choppy & are changing with a fluctuating magMod (magnetic field). It looks as if the compass (magYaw, Green) is adjusting the imuYaw (Black) step wise ... all following the magMod.



If we then check out how well the gyro (Green) follows & register the imuYaw movement (Black, & shown unwrapped) & taking in the rudder commands it's showing a pretty good agreement until again 21sec but then gets differences.



If we after this take a look at the difference between IMU velocity & GPS velocity (Black & Green) which in severe Yaw errors "due to magnetic disturbed take off spots" usually show big deviations ... Here we see deviations also but not as large, starting again about from 21sec ... and the deviations are following the magMod (Red).



And a last check ... will the magMod (Red) change when the AC pitches (Blue)?

Yep ... it does, pretty well from about 28sec.



So all this together leads me to think that something on the Spark or attached to it is magnetized & by the pitch the magMod changes ... & by that effect the compass (magYaw) ... which in turn adjusts the imuYaw creating that nice bow shaped flight path.

Eventually the flight controller reacts & switch to ATTI mode at 55sec ... with all that means regarding no braking & no position hold.

 

[sar104]

@sar104 ... please validate my conclusion below when you have time, may very well be wrong here ? .

From what I see this isn't as easy as a take off from a magnetic disturbed spot, something else is going on ...

Looking at how the imuYaw (Black) & the magYaw (Green) follow each other it's not immediately apparent that the magYaw diverge from imuYaw at height. But something is going on from 21sec ... the agreement between Green & Black isn't good, instead choppy & are changing with a fluctuating magMod (magnetic field). It looks as if the compass (magYaw, Green) is adjusting the imuYaw (Black) step wise ... all following the magMod.

View attachment 105533

If we then check out how well the gyro (Green) follows & register the imuYaw movement (Black, & shown unwrapped) & taking in the rudder commands it's showing a pretty good agreement until again 21sec but then gets differences.

View attachment 105535

If we after this take a look at the difference between IMU velocity & GPS velocity (Black & Green) which in severe Yaw errors "due to magnetic disturbed take off spots" usually show big deviations ... Here we see deviations also but not as large, starting again about from 21sec ... and the deviations are following the magMod (Red).

View attachment 105538

And a last check ... will the magMod (Red) change when the AC pitches (Blue)?

Yep ... it does, pretty well from about 28sec.

View attachment 105539

So all this together leads me to think that something on the Spark or attached to it is magnetized & by the pitch the magMod changes ... & by that effect the compass (magYaw) ... which in turn adjusts the imuYaw creating that nice bow shaped flight path.

Eventually the flight controller reacts & switch to ATTI mode at 55sec ... with all that means regarding no braking & no position hold.

View attachment 105541

Yes - I think the problem is bad compass data. If you do the pure inertial yaw calculation, just using pitch, roll and the rate gyros, and compare that to the IMU and magnetic yaw you get the following:



I think that the inertial solution (red) is probably correct, but the FC is being influenced by the magnetic yaw (blue) and so the fusion solution (green) ends up off by 60 - 70°, which would account for the flight path.