Magnetic filed (sic) interference???

[Nemesis Looms]

Life became briefly exciting the other day whilst demonstrating the wonders of drones to a friend.

Hovering at about 20 metres over a 1.5 acre lawn, my Spark suddenly climbed to about 35 metres and shot off towards a group of trees about 50 metres away uphill, away from the established home point. I lost all GPS, telemetry and control and had to chase after it to loctate it visually. It was happily hovering inches away from the top of the tallest tree and refused to respond to the controller for about 60 seconds before re-establishing connections. Running through the flight record later on, I found that the warning 'Magnetic filed interference' (sic) appeared briefly before the fly-away.

Anyone else seen this? - My question is, which magnetic field was interfered with 20 metres above a lawn in the Highlands of Scotland and how? Was it the drone, the controller, the mobile phone or the planet?

Anyway, I got it back, so to celebrate I took it to a nearby paddock and managed to fly it into a stone wall in sport mode - but that's another story.

 

[guppy1]

Had this problem twice with my Spark at 30 meters high.(I had it only in buildup area.) Magnetic interference before a flyaway. I do think it could be some tipe of radio interference. I was lucky to find the Spark every time. I check the Wi Fi channels and select it manually. I did buy a MA and found that I get a message "radio interference", I change route or go back home.

 

[ton079]

I scrolled through te forum to see if the problem I experienced occurred before. And as in this posting: it did.

Today I lost my Spark because of this same issue. Nearby a trainstation I was before, but with some electricity pylons at around 800 mtrs away, the drone climbed up to approx. 25 mtrs, was making a video from the station but after 35 sec's it stopped with that and flew away in western (opposite from the pylons) direction. I pressed RTH quickly and although the controller was beeping, soon all connection was. On my screen I got a notification about magnetic fields.

Does anyone know what the speed is (no sport-mode) of a Spark so that I might be able to calculate with the full battery how far it could fly until power is down so that I could search for it again (you'll never know: it's and industrial area with a lot of buildings, but also a lot of roads and open fields ?)
The log only shows the location where I let the Spark go up. For two minutes.

 

[NOLADG]

With the screen shot, it looks like the drone went into Atti mode, which means it was no longer using GPS signals, but holding its altitude and floating with the wind. Pressing RTH does not work in this mode as it has no idea where home is or where IT is. So my thoughts is your drone would be flying at the speed of the current wind and at its mercy.

I would suggest to recreate a new thread with your flight logs and let the experts advise.

 

[ton079]

With the screen shot, it looks like the drone went into Atti mode, which means it was no longer using GPS signals, but holding its altitude and floating with the wind. Pressing RTH does not work in this mode as it has no idea where home is or where IT is. So my thoughts is your drone would be flying at the speed of the current wind and at its mercy.

I would suggest to recreate a new thread with your flight logs and let the experts advise.

OK ! Thanks for the advice. I will do so.

 

[Spark 317]

OK ! Thanks for the advice. I will do so.

Along with logs, the time of day and date will also help with a search pattern.

This information can be used to determine the wind speed and direction at the time of occurrence. and how long it may have been drifting in the wind until the battery is low and lands somewhere.

 

[SPark_South_Oz]

I would suggest to recreate a new thread with your flight logs and let the experts advise.

Where do you all post flight data and get analysis ?
In the Spark Help section ?

On Mavic Pilots, there is a special Crash / Flyaway section for this.
It gets a LOT of action.

and shot off towards a group of trees about 50 metres away uphill

The op's post is old, but this is the sort of feedback that often leads to a compass conflict as a cause.

Flyaways are very often a problem with the home point not being recorded correctly at take off.

When the lady says "home point recorded, please check it on the map" she means it !
When the map icon comes up, a pilot needs to check the orientation of the drone on the map, matches the way it's facing on the ground.

Not sure if this is a major problem with the Spark (I've never had this with mine or my M1P), but if the aircraft is not aligned as per the map, there there could be slight magnetic interference (not recorded and no warning) at home point, and it results in an aircraft looking like it's facing 180 degrees or other on the map.

When the aircraft gets airborne, there is conflict between compass and IMU, it usually corrects itself in a short time, but often too late, the aircraft simply flies off and loses signal before this happens.
It just keeps going until forced landing or crashing.

 

[ton079]

Where do you all post flight data and get analysis ?
In the Spark Help section ?

On Mavic Pilots, there is a special Crash / Flyaway section for this.
It gets a LOT of action.



The op's post is old, but this is the sort of feedback that often leads to a compass conflict as a cause.

Flyaways are very often a problem with the home point not being recorded correctly at take off.

When the lady says "home point recorded, please check it on the map" she means it !
When the map icon comes up, a pilot needs to check the orientation of the drone on the map, matches the way it's facing on the ground.

Not sure if this is a major problem with the Spark (I've never had this with mine or my M1P), but if the aircraft is not aligned as per the map, there there could be slight magnetic interference (not recorded and no warning) at home point, and it results in an aircraft looking like it's facing 180 degrees or other on the map.

When the aircraft gets airborne, there is conflict between compass and IMU, it usually corrects itself in a short time, but often too late, the aircraft simply flies off and loses signal before this happens.
It just keeps going until forced landing or crashing

Along with logs, the time of day and date will also help with a search pattern.

This information can be used to determine the wind speed and direction at the time of occurrence. and how long it may have been drifting in the wind until the battery is low and lands somewhere.

I used the Phantomhelp page to get my log readable. Is that what's being meant? Or does one need the other files, like KML, CSV or TXT?
Please see what I mean:

DJI Flight Log Viewer | Phantom Help

Upload and view DJI Phantom, Mavic, and Spark flight logs online

www.phantomhelp.com

 

[SPark_South_Oz]

Thanks ton, yes I've used that a few times with the simple TXT files, I find PhantomHelp log reader quite ok to get the basics.
It's also a bit of fun revisiting a flight in such a way.

Some logs need more in depth analysis with DAT file, csv / graphing etc, way beyond my interpretation level.

I just meant what segment of the forum, I'm guessing Spark - Help, but didn't see too many assistance threads for crashes / flyaways on the first page.

 

[ton079]

Thanks ton, yes I've used that a few times with the simple TXT files, I find PhantomHelp log reader quite ok to get the basics.
It's also a bit of fun revisiting a flight in such a way.

Some logs need more in depth analysis with DAT file, csv / graphing etc, way beyond my interpretation level.

I just meant what segment of the forum, I'm guessing Spark - Help, but didn't see too many assistance threads for crashes / flyaways on the first page.

Agreed. Anyway, I also dropped an email at DJI. They will answer as soon as they can, but I understand that may cost some time seen the circumstances we're all living in these days.
Thanks for the replies and stay safe !

 

[RotorWash]

The op's post is old, but this is the sort of feedback that often leads to a compass conflict as a cause.

Flyaways are very often a problem with the home point not being recorded correctly at take off.

When the lady says "home point recorded, please check it on the map" she means it !
When the map icon comes up, a pilot needs to check the orientation of the drone on the map, matches the way it's facing on the ground.

Not sure if this is a major problem with the Spark (I've never had this with mine or my M1P), but if the aircraft is not aligned as per the map, there there could be slight magnetic interference (not recorded and no warning) at home point, and it results in an aircraft looking like it's facing 180 degrees or other on the map.

When the aircraft gets airborne, there is conflict between compass and IMU, it usually corrects itself in a short time, but often too late, the aircraft simply flies off and loses signal before this happens.
It just keeps going until forced landing or crashing.

Good points, especially about the AC symbol being properly aligned on the map. That alone would prevent a lot of issues. Maybe also worth noting that if the bird flips into ATTI mode, it's still very flyable, actually more flyable in some ways, just won't hold position. However without proper pilot input on the sticks the situation can easily get out of hand and end up as you mentioned.

 

[ol' Sparky]

I've had that happen a couple of times. Both times at about 60 feet up and 600 out. It lost connection and satellites, went into ATTI mode. First time it took off away from me, but regained signal after 70 or eighty feet and responded to the controller. I think it may have had to do with the phone I was using being almost dead. Controller was charging it, but was still slowly dying. It might have tried to shut down apps to save it's battery and lost connection. Second time same situation near the same spot. This time it flew toward me and regaind signal/control about a hundred feet out. Later checked log and found It showed massive IMU and compass errors. Recalibrated both and later found that one of the prop tabs had broken off and am thinking there may have been excessive vibration that caused it. Lucky it didn't fail and drop out of the sky. Now I ALWAYS check my props have both tabs and are secure before a session. I think the point is...there's usually an explanation for a fly-away other than "it just took off on me", though I have no doubt that HAS happened. Now I don't fly unless all batteries are fully charged and a proper pre-flight. Haven't had a problem since. Knock on wood.

 

[ton079]

Thanks ton, yes I've used that a few times with the simple TXT files, I find PhantomHelp log reader quite ok to get the basics.
It's also a bit of fun revisiting a flight in such a way.

Some logs need more in depth analysis with DAT file, csv / graphing etc, way beyond my interpretation level.

I just meant what segment of the forum, I'm guessing Spark - Help, but didn't see too many assistance threads for crashes / flyaways on the first page.

You reminded me of syncing the data into the cloud via the app, after I've contacted DJI TechSupport. They advised me to do that, opened a ticket and promised me to provide me their investigation results within a few days. They did. After 3 days they informed me that the problem was accepted as warranty-claim and that I would receive a new Spark (identical of what I had) within a few days.

Great service, nice help.

So, SPark_South_Oz, thanks for bringing the fact of the data to my attention. It helped a lot ! Happy person again ;-)

 

[Northwood Mediaworks]

I had something like this once, and it was related to hidden metal in the ground near my take off point. Also, metalic items on your person can affect things negatively.

 

[I B Spectre]

Would pressing the PAUSE button have helped in any way when it went into ATTI mode?

 

[ton079]

Would pressing the PAUSE button have helped in any way when it went into ATTI mode?

I will never know, but what I remembered iks that I pressed, what I thought was useful, but nothing helped.

Anyway, after having investigated my flightlogs, DJI accepted responsibility and send me a brandnew Spark within 3 days ! Old spark was 13 months old.