Hello fellow Spark (and other birds) pilots!

I'm a happy Spark owner for a couple of months now. I've spent some time flying, never had any problems until recently. I would like to share my expecience and logs and ask for help with the analysis as I have no experience in it.

Until December 31st, everything was going fine - I was flying with no problems using the RC and iPhone SE connected with an OTG cable.
On 31st, I wanted to shoot the fireworks. I took off, climbed, maneouvered a bit to get some nice angle. Then I wanted to just hover in place and wait for the fireworks to start. But all of a sudden, I got a Compas Error warning in the app and the drone started flying by itself, gaining speed. At first I was startled, not knowing what to do (as this was my first such experience). Then I tried to use the controller to command it to move the opposite direction than it was flying to - that helped and the drone stopped. Shortly after that, it got back to P-GPS and behaved just fine. I landed immediately. The flight log of this event can be found here: http://www.phantomhelp.com/LogViewer/4QG61JORZ29V098ILYXA/ Unfortunately, the log from the AC itself is no longer available (or at least DJI Assistant is not showing it in the list).

Second event similar to this one happened today. This time, I wasn't flying using the OTG (because the cable was crap and broke), but rather a WiFi between the phone and the RC. I was flying earlier that day with no issues. Then I moved to another location and took off again. No problems until I decided that it's time to get back t me and land. I switched to sport mode and headed my direction. Then I released the controls to stop the drone, and as it stopped, I got the same Compas Error, switching to ATTI error in the DJI GO app again. This time, the drone didn't start flying by itself like the last time. But as it was a bit windy, I had to make some effort with the controls to make it stay in place in ATTI mode. After couple of seconds, it got back to P-GPS again and I landed with no issues. The flight log from the app is here: http://www.phantomhelp.com/LogViewer/9G3Y59BK4RWO4YVBPOK2/
The log from the AC is here: https://drive.google.com/open?id ... OZv1y_0cmVUARrMWSea

I would like to know if I have done anything wrong and what should I do to prevent such events in the future. Every comment appreciated, thank you!

Extreme  stops etc in sport mode seem to cause this. Problem now explained, not mysterious interference.

Bright Spark Posted at 2018-1-14 09:14
Extreme  stops etc in sport mode seem to cause this. Problem now explained, not mysterious interference.

First log does not show this.

Can you let us know have you calibrated your compass lately?

Also it looks like on both flights you launched from concrete footpaths.

Bright Spark Posted at 2018-1-14 09:14
Extreme  stops etc in sport mode seem to cause this. Problem now explained, not mysterious interference.

I am more and more getting this impression too. There have been many cases posted by now where it happened in sport mode with no visible potential source of interferences.

This at last has been duplicated by several pilots. Of course it may not explain all incidences, but to reproduce a fault is worth
more than all the other analyses put together.

I've already said in some other topic I can reproduce yaw error on demand. Yes, flying fast and than suddenly changing either speed or direction and most efficiently both (like in very tight coordinated banked turn after full speed forward) yields these pesky yaw errors. My very wild guess is that sampling rate of GPS and magnetometer is not fast enough or not in sync causing poor Spark to get confused.

Hmm yeah, I kinda thought that sport mode probably has something to do with it. Although I did fly in sport mode before that happened earlier today, as well as yesterday and in those cases, everything went fine. So there has to be some other factors included I guess. Anyway, I guess I'll avoid it for now, just to be safe.

#4 hallmark007: I calibrated the compass after the first incident on 31st. Before that, I calibrated it sometime in December. Back then, I thought it's a good idea to calibrate it often because a lot of people and guides recomended it. I don't do it anymore though, after reading thru this forum. Regarding the note about taking off from a concrete paths: they are more like gravel in fact, even though I don't know what's under the gravel (and this time of a year, snow). Also, don't know if it means anything, but both times I actually launched from the black foam box the Spark comes in when you buy it.

#7 eYeSkYeYe: I missed that other topic, but thank you for the reply. It would make sense. Even though according to the logs, there were no yaw errors during my previous sport mode flights. All of them went the same: full speed forward and then release the sticks so it slows down by itself.

Speed of release is what matters.... if you release it slowly deceleration is lower than when you release it sharply..... i guess once weather permits I will test what will happen if not only I release the pitch but actually force it in counter direction. Not only will be interesting to see what will happen with AC but also with gimbal. I don't think it will handle such a forces well.

eYeSkYeYe Posted at 2018-1-14 11:30
I've already said in some other topic I can reproduce yaw error on demand. Yes, flying fast and than suddenly changing either speed or direction and most efficiently both (like in very tight coordinated banked turn after full speed forward) yields these pesky yaw errors. My very wild guess is that sampling rate of GPS and magnetometer is not fast enough or not in sync causing poor Spark to get confused.

Something like this?...

Sort of. These are attempts to do the flip I think as to me it looks like part of the routine is pulling back the throttle. I am not pulling throttle down in tight loops as I want as much lift as sparkly possible to avoid stall. Please note that forces in tight loops can be significant not to mention problems with propwash ("bad", turbulent air under the propellers) that often lead to sideslips.

Isn't the reason of the motor shutdown in the video the combo stick command shutdown? Looks like the pilot is pulling back and perhaps down at the same time. I have never had the balls to try it myself (I like my Spark too much to just throw it agains the ground like this), but the manual says it should work like that. Anyway, thanks for all the replies!
But please post your findings eYeSkYeYe, I'm curious about that

I suppose that the spark's own inputs might well match these conditions for triggering errors when flown in excessively high  winds.

In that case I would expect the app to throw "High wind velocity, fly with caution" kind of warning.

eYeSkYeYe Posted at 2018-1-14 11:30
I've already said in some other topic I can reproduce yaw error on demand. Yes, flying fast and than suddenly changing either speed or direction and most efficiently both (like in very tight coordinated banked turn after full speed forward) yields these pesky yaw errors. My very wild guess is that sampling rate of GPS and magnetometer is not fast enough or not in sync causing poor Spark to get confused.

Well show us your log then.

eYeSkYeYe Posted at 2018-1-14 12:53
In that case I would expect the app to throw "High wind velocity, fly with caution" kind of warning.

Exactly, high wind was also on my list but the warning according to all my experiences so far comes quite early at windspeeds that Spark can still cope with very well. Thus there clearly must be other factors playing a key role.

hallmark007 Posted at 2018-1-14 09:22
First log does not show this.

Yes, this puzzles me as well. But many roads are leading to Rome and I start to think that more factors have to be taken into consideration. By now I have seen so many flightrecords where the flying environment makes magnetic interference as root cause at least unlikely, but then flying in sport mode was often involved. At least I believe it's good to continue these investigations and they might lead to better preparedness and understanding.

izzy26 Posted at 2018-1-14 12:36
Isn't the reason of the motor shutdown in the video the combo stick command shutdown? Looks like the pilot is pulling back and perhaps down at the same time. I have never had the balls to try it myself (I like my Spark too much to just throw it agains the ground like this), but the manual says it should work like that. Anyway, thanks for all the replies!
But please post your findings eYeSkYeYe, I'm curious about that

Not likely because I have tested it and got confirmation from two different sides at DJI that CSC will only work when Spark is in real emergency situation with serious errors occurring.

It may not be wind , though high, that flags an error. In non sport mode,normal  stick inputs may take the spark to sport mode angles of bank/tilt if it's already working hard.
I am guessing of course.
My personal evidence is that I have never had a yaw/compass error in six months, which on present info suggests it's because I don't fly in 15+ mph wnd or go too mad with sport
Mode.
Still flakey evidence, but all I have. Sport mode at the limit is, as yet, all we have to go on, IMHO.

Wachtberger Posted at 2018-1-14 13:22
Yes, this puzzles me as well. But many roads are leading to Rome and I start to think that more factors have to be taken into consideration. By now I have seen so many flightrecords where the flying environment makes magnetic interference as root cause at least unlikely, but then flying in sport mode was often involved. At least I believe it's good to continue these investigations and they might lead to better preparedness and understanding.

We know looking at second log above that errors started at 7.02 Aircraft was moving straight at 13mph when first warning appeared, there are no wind warnings , speed did increase to 29mph for a very short time , but during all the yaw errors Aircraft was at half it’s speed capacity ie less than 15mph for 80% of the time, and the Aircraft was almost always going in a straight line.

hallmark007 Posted at 2018-1-14 14:00
We know looking at second log above that errors started at 7.02 Aircraft was moving straight at 13mph when first warning appeared, there are no wind warnings , speed did increase to 29mph for a very short time , but during all the yaw errors Aircraft was at half it’s speed capacity ie less than 15mph for 80% of the time, and the Aircraft was almost always going in a straight line.

Thank you, that are very convincing points!

Is it possible that during these ,errors gps is in difficulty, and if that's how speed is calculated, is the speed accurate?

Bright Spark Posted at 2018-1-14 14:47
Is it possible that during these ,errors gps is in difficulty, and if that's how speed is calculated, is the speed accurate?

I am now thinking , initial yaw error is probably misnamed, and is due to a very poor GPS lock leading to inconsistent position and speed data .

hallmark007 Posted at 2018-1-14 13:01
Well show us your log then.

I have zero incentives of doing so... I'll just reveal my location and nothing else. Once and if DJi sorts out its bug bounty program, I might reconsider my decision. So far, yaw errors are not under scope of bug bounty program.

izzy26 Posted at 2018-1-14 12:36
Isn't the reason of the motor shutdown in the video the combo stick command shutdown? Looks like the pilot is pulling back and perhaps down at the same time. I have never had the balls to try it myself (I like my Spark too much to just throw it agains the ground like this), but the manual says it should work like that. Anyway, thanks for all the replies!
But please post your findings eYeSkYeYe, I'm curious about that

Izzy, did you crash your drone before? If not, please ensure that there is no interference, such as, high voltage line, factories, metal around your spark. If it is not caused by the environment, please provide the flight data and the black box data for further analysis and upload them to the dropbox and post the link here.

Dji Elecktra
Lol same old spiel

Congratulations of succesfull retrieval of you drone! Could be some interference comiing from here: https://goo.gl/maps/VYf1JiKaqET2 And other might be sports mode +cold weather problem. Shame there is no straight answers for compass problems...