Slow yaw at start of flight without control inputs

[BorisTheSpider]

Good mavic, probably 10 hours of flight on it. FW .400. No major issues so far.

Today, at the start of a flight, after putting it in a hover at about 10M to get precision landing data, then leaving the sticks alone, I observed it slowly yaw right about 10 or 20 degrees over the course of maybe 15 seconds.

Never seen that before, not sure whether I stopped it with a stick input or it stopped itself eventually, but once it stopped I then had an uneventful flight for a few minutes.

I'll pull the DAT and look at it tomorrow, but just wondered if anyone had seen that before - it could have been (and I'm entirely speculating since I haven't pulled the DAT yet) a magyaw vs yaw deviation. Like I said though, no further undesirable behaviour and it stayed in GPS mode throughout the flight.

The only factor I can identify that seems a likely cause was power lines, quite a few of them near the takeoff area, maybe 4M above ground, but they weren't high-tension, just ordinary distribution lines carrying probably 230V.

 

[Kilrah]

Had that several times in one location, taking off from something that contained metal. Stabilized itself after a few seconds an no further issues.

 

[Xcel]

I'm on FW .400 as well. I've had that happen to me three times in 47 hours of flight time. It starts off exactly like you said slowly yawing to the right but then it gets worse though, after a couple of minutes it switches to ATTI mode for me even though I have 20+ satellites. This happens all of a sudden and it can switch back and forth during a flight. It was only last time it happened that I learned to recognize the symptoms, it always starts off with the slow yaw to the right then it goes into ATTI at some point.

No idea what causes it, I never calibrate my compass and don't take off near any metal. I've had it happen at home twice and I always take off from the same location over 100 times so I don't know what's up.
The next time I see it though I'm going to land asap and restart the quad.

 

[BorisTheSpider]

OK so I reviewed the DAT.

Yaw and MagYaw start out well matched differing by the usual 5 degrees or so, then at T=0 at takeoff, begin diverging and reach a maximum of 16 degrees of difference at T=10s, there's then no control inputs from T=10 (when the throttle input to climb initially after takeoff ends) and T=30s when I made a yaw input.

During that interval with no control inputs, MagYaw gradually returns from -107 to -116 whilst Yaw stays at -123 throughout. That 20 second interval must have been when I observed the uncommanded yaw to the right which the mavic made of around 10 degrees.

Throughout the event GPSHealth was at 4 and flight mode was GPS_ATTI

The DAT file is at https://www.sendspace.com/file/n4mv1z

I'd be extremely grateful for any input from BudWalker or anyone else with experience reviewing DAT files, to see if I'm just looking at a user error here in taking off too close to power lines, and whether there's more to review about the data that I haven't considered. In particular I can't recall how to interpret the MagZ values to see what they tell us about what happened.

 

[Xcel]

I should have saved my latest incident from a week ago but alas I didn't. I might have a DAT from an earlier incident saved though but I'm not sure. I'll check tomorrow, I'm only on my laptop right now.

 

[BudWalker]

OK so I reviewed the DAT.

Yaw and MagYaw start out well matched differing by the usual 5 degrees or so, then at T=0 at takeoff, begin diverging and reach a maximum of 16 degrees of difference at T=10s, there's then no control inputs from T=10 (when the throttle input to climb initially after takeoff ends) and T=30s when I made a yaw input.

During that interval with no control inputs, MagYaw gradually returns from -107 to -116 whilst Yaw stays at -123 throughout. That 20 second interval must have been when I observed the uncommanded yaw to the right which the mavic made of around 10 degrees.

Throughout the event GPSHealth was at 4 and flight mode was GPS_ATTI

The DAT file is at FLY152.DAT (251.16MB) - SendSpace.com
I'd be extremely grateful for any input from BudWalker or anyone else with experience reviewing DAT files, to see if I'm just looking at a user error here in taking off too close to power lines, and whether there's more to review about the data that I haven't considered. In particular I can't recall how to interpret the MagZ values to see what they tell us about what happened.

I looked at FLY152. In summary, the Mavic was launched from a geomagnetically distorted site. But, it wasn't too bad and the Mavic was able to recover.

Before launch the magYaw/Yaw difference was about 5 degrees (the geoMagnetic declination is -5.12 degrees). When the Mavic was launched and reached 5 meters the magYaw/Yaw difference had increased to 15 degrees.

No control inputs were provided for 30 secs but the Mavic rotated itself so that the magYaw/Yaw separation was decreased back to the original 5 degrees. The Flight Controller holds the Yaw value constant while rotating the Mavic until magYaw (i.e. the compass/magnetometers) agree with the Yaw value.

The fact that the Mavic is being rotated while Yaw is constant has been verified in other flights where a video was provided. A video wasn't provided here but we can look at totalGyroZ to confirm the Mavic was being rotated. TotalGyroZ is computed by DatCon by integrating and then summing the gyroZ value. Here we can see that totalGyroZ shows the Mavic rotating.


The magnetometers are shown here.


It appears the launch site was possibly on concrete? It's hard to tell exactly.

 

[BorisTheSpider]

Thanks Budwalker - very kind of you to take the time to look at this - the launch site may have had concrete around, I'm usually cautious of concrete because of the possibility of rebar so it wouldn't have been bare concrete but it may have had shale or grit over it and I didn't check carefully enough - there's certainly concrete around as you can see the large concrete water container next to the launch site so the foundations of that could have been underneath the gravel.

The map view also doesn't seem to show the power lines which were nearby, so perhaps the imagery is old or they just don't show clearly.

Either way, it seems it's user error and I was lucky the impact wasn't worse. I'll be more careful around power lines and possibly hidden concrete in future.

 

[BudWalker]

Thanks Budwalker - very kind of you to take the time to look at this - the launch site may have had concrete around, I'm usually cautious of concrete because of the possibility of rebar so it wouldn't have been bare concrete but it may have had shale or grit over it and I didn't check carefully enough - there's certainly concrete around as you can see the large concrete water container next to the launch site so the foundations of that could have been underneath the gravel.

The map view also doesn't seem to show the power lines which were nearby, so perhaps the imagery is old or they just don't show clearly.

Either way, it seems it's user error and I was lucky the impact wasn't worse. I'll be more careful around power lines and possibly hidden concrete in future.

The geoMagnetic distortion exists only up to 5 meters. If the power lines are the source of the geoMagnetic distortion then it would get stronger with altitude. What is the height of that concrete water container?

 

[BorisTheSpider]

The geoMagnetic distortion exists only up to 5 meters. If the power lines are the source of the geoMagnetic distortion then it would get stronger with altitude. What is the height of that concrete water container?

The concrete water container was around 3M high.

 

[WEOTN]

Weren't some solar flair activities expected this week? We can see what it does for "northern lights", could it affect a drones nav system?

From space weather.com: "Earth is exiting a stream of solar wind flowing from a large hole in the sun's atmosphere. As a result, the chance of polar geomagnetic storms is subsiding from 35% today to no more than 10% by the weekend. Solar wind speeds will soon return to background levels."

 

[BorisTheSpider]

Weren't some solar flair activities expected this week? We can see what it does for "northern lights", could it affect a drones nav system?

From space weather.com: "Earth is exiting a stream of solar wind flowing from a large hole in the sun's atmosphere. As a result, the chance of polar geomagnetic storms is subsiding from 35% today to no more than 10% by the weekend. Solar wind speeds will soon return to background levels."

Thanks for the suggestion but it seems the solar activity was fairly normal at the time of the flight - I think rebar in the concrete is the most likely cause.