Can flying near radio towers interfere with Mavic Pro?

[FoxhallGH]

I must admit I'm pretty neutral in terms of who I feel is right or wrong in this debate and be honest in saying I don't have a clue about radio towers, and what effect they may potentially or not have on electronics, other than I suspected it may of had some effect, hence the post.

So far there isn't a clear answer on whether it was the tower or something in the ground from where i took off from which caused my issue but one thing I'll take away from this is I won't be flying in that field or anywhere near that site again lol. I strongly advise others to do the same, unless you want to practice flying in Atti.

I feel extremely lucky and delighted to have my Mavic back with me as I could have lost it in my first week of ownership and would have been one of the worst days of my life.

I'm grateful to everyone on this forum offering their informative and intelligent feedback in to the possible causes of what happened and interested in reading further posts about this. I think it is important for us drone enthusiasts to take extra precaution when choosing the areas to fly in.

When I get home from work this-evening I'll look to post the .DAT file if it might reveal further clues as to what had happened?

Thanks for this post adam3112, as it has raised some interesting issues ... On the one hand, I know what RF Interference can do to electronics, but I also accept what's been said about drones being used to inspect transmission towers. I still don't feel comfortable with risking flying my Mavic near a tower like that one on Reigate Hill ... But the one thing that's the real take-away from this one, is the revelation that the Mavic could in fact be twisted-up in regard to its compass and INS being compromised - without us knowing about it 'till we get up off the ground! ... And how this could be detected if you can see a difference on the controller map view where the actual drone, is pointing in a different direction from the arrow on the map.

 

[adam3112]

It's pretty clear that it had nothing at all to do with the tower.

Is it though? I have read some convincing arguments to support this idea, particularly from Congoblue. I haven't actually seen any sound explanation so far that it's to do with the ground from where I took off from either, other than it's caused by steel items or concrete being nearby, even though there was no Steel items or concrete visible within the takeoff area.

I get that it may have been under the ground, but again we don't know that with absolute certainty as Steel and Concrete are not visible.

I landed within 5ft of that same spot and took off again without any issues (other than the strong wireless interference message which again is another clue). if there was steel under the ground why did it not affect my Mavic the second time?

Imo Congoblue has given the most convincing response so far in explaining why the tower could have caused it. With the tower being the only visible hazard, it seems to carry the most weight.

 

[Congoblue]

All I was saying was that it is possible for the compass reading to be corrupted if the Mavic passed through a beam of focused RF energy, for example one of the microwave link dishes. I have never personally seen this happen on a Mavic (don't fly near towers) but I have seen it happen on other equipment.

Read up on magnetometers and how you need to keep them away from interference if you don't believe me. They can be affected by external electromagnetic fields just as much as by pieces of metal.

Anyway, personally I would not take the risk of flying close to high power radio antennas like this.

 

[Meta4]

Is it though? I have read some convincing arguments to support this idea, particularly from Congoblue. I haven't actually seen any sound explanation so far that it's to do with the ground from where I took off from either, other than it's caused by steel items or concrete being nearby, even though there was no Steel items or concrete visible within the takeoff area.

It's an open and shut case but I've completely wasted my effort trying to explain anything to you.

Imo Congoblue has given the most convincing response so far in explaining why the tower could have caused it. With the tower being the only visible hazard, it seems to carry the most weight.

Convincing???

I give up.

 

[Meta4]

Read up on magnetometers and how you need to keep them away from interference if you don't believe me. They can be affected by external electromagnetic fields just as much as by pieces of metal.

Just as much?
Really?
Just show me one case of a DJI drone with compass problems caused by external electromagnetic fields.
I've read the data from a few hundred flight incidents and not seen my first one yet.

 

[Congoblue]

Just as much?
Really?
Just show me one case of a DJI drone with compass problems caused by external electromagnetic fields.
I've read the data from a few hundred flight incidents and not seen my first one yet.

I don't think you would be able to tell from the flight logs? It would just show up as a compass discrepancy, you would not be able to tell what caused it.

Anyway, you are clearly very convinced of your own opinion and I do not want to argue further. I was just trying to point out that it was another possibility.

 

[Meta4]

I don't think you would be able to tell from the flight logs? It would just show up as a compass discrepancy, you would not be able to tell what caused it.
Anyway, you are clearly very convinced of your own opinion and I do not want to argue further. I was just trying to point out that it was another possibility.

Like I said .. just show me one case.
This one was the same as the last 100 I've analysed that were caused by launching from steel or reinforced concrete surfaces.
But what would I know?
What a dismal waste of effort this entire thread has been.

 

[Congoblue]

What a dismal waste of effort this entire thread has been.

I am sorry you think that. I thought there had been some interesting discussion.

 

[adam3112]

It's an open and shut case but I've completely wasted my effort trying to explain anything to you.


Convincing???

I give up.

You seem to be very defensive about people giving alternative potential causes and I’m open to all ideas.

I never at any point said you were wrong, just that the only visible hazard was the tower. I was there so I should know. You keep saying this issue is because I took off from reinforced concrete or steel, all I’m saying is I took off from a muddy, grassy gravelly footpath and those things were not to be seen on the ground or anywhere in my takeoff vicinity.

On some parts of that footpath your foot could easily sink 12 inches in mud, so any concrete or steel present in that spot is going to be quite deep in the ground.

Up until this point I have appreciated your feedback and advice in this thread Meta4 but I’m finding it arrogant that you dismiss all other Hypothesies put forward by other members of the forum as though they don’t know what they are talking about when clearly some of them have worked in or experienced this quite complexed science first hand.

 

[FoxhallGH]

Just as much?
Really?
Just show me one case of a DJI drone with compass problems caused by external electromagnetic fields.
I've read the data from a few hundred flight incidents and not seen my first one yet.

Magnetometers are used in metal detectors as well as compass applications - but they are doing the same thing. It's not that they are detecting metal as such, but they are detecting the effect the metal has on the Earth's magnetic field. That effect is to concentrate the lines of magnetic force. Strong electromagnetic fields (as you'll get in microwave radiation), also concentrate magnetic lines of force. If you research Magnetometers, it seems that Micro Electronics is providing [Lorentz effect] magnetic field sensors that have immunity to electromagnetic interference (EMI) due to the use of optical and piezo components. These devices are known as MEMS and you'll find them in your smartphone (and the Mavic). The MEMS does the detection and outputs a signal onto other electronics in the Mavic and this is where it becomes a bit blurry! While the MEMS may be immune to the EMI, the amplifiers turning that signal into something that controls the motors and stability of the drone, are a different story. This is where Congoblue's experience comes in ... Sar104 pointed out the reason for the instability in regard to how the Mavic's components were working (or not working) with each other, but we don't definitely know what the 'SOMETHING' is or was that caused the misalignment in the first place. When you construct a telecom's tower, depending on the earth around it, you can have situations where a ground-plane is laid down. This is a network of wires or mesh, spanning out from the tower under a layer of topsoil. If a ground-plane is under the site, then that would be a major factor in misaligning the compass and IMU.

 

[FoxhallGH]

It's an open and shut case but I've completely wasted my effort trying to explain anything to you.


Convincing???

I give up.

Sorry if what you have contributed has been taken in the wrong way Meta4 - but - your 'explanation' of what the drone did and pulling of the flight log info has given real insight. Thank you for that. I'm just trying to go one step further and ask the question "OK - so now we know what the drone did, and why it did it - can we nail down what made it happen in the first place?" If we could get the answer to that, then it would be a useful thing to know when we check out a site as a fly zone ...

 

[BudWalker]

@adam3112 I took a look at the .txt you submitted. It certainly does appear that Yaw was compromised. This is usually caused by launching form a geomagnetically distorted site. But, not always. In this case it's hard to tell from the .txt since lateral movement is required to check against the recorded Yaw. This didn't occur until 27 secs into the flight.

Can you provide the AC .DAT? It will be FLY024.DAT. Look here for info about retrieving the .DAT. This will have the gyro, accel and magnetometer data. It will also have data from before the flight. Did you move the Mavic to the launch site after powering it up?

I have a hard time believing this was caused by the EM fields. I would have expected more erratic behavior and other systems to have been affected. But, the .DAT would probably have some indication if this was the cause.

 

[adam3112]

@adam3112 I took a look at the .txt you submitted. It certainly does appear that Yaw was compromised. This is usually caused by launching form a geomagnetically distorted site. But, not always. In this case it's hard to tell from the .txt since lateral movement is required to check against the recorded Yaw. This didn't occur until 27 secs into the flight.

Can you provide the AC .DAT? It will be FLY024.DAT. Look here for info about retrieving the .DAT. This will have the gyro, accel and magnetometer data. It will also have data from before the flight. Did you move the Mavic to the launch site after powering it up?

I have a hard time believing this was caused by the EM fields. I would have expected more erratic behavior and other systems to have been affected. But, the .DAT would probably have some indication if this was the cause.

Hi BudWalker, it is possible that I powered on the aircraft before putting it on the ground in it's takeoff position, but my memory of how and when I placed the drone down is abit foggy in all honesty. I will post the .DAT file tonight

 

[adam3112]

Here is the DAT File (attached).

 

[BudWalker]

Here is the DAT File (attached).

Don't really have an answer, just some observations.

The .DAT doesn't show any of the usual characteristics of having launched from a geomagnetically distorted site. In particular, Yaw (determined mostly from IMU data) and magYaw (determined from just the magnetometers) don't separate as the Mavic ascends. If the launch site is geomagnetically then distorted magYaw would change as the Mavic rises above the distortion but Yaw remains constant because the IMU data is saying the Mavic hasn't rotated.



There are several points in the flight where a Yaw/magYaw separation occurs. The first time resulted in the YAW_ERROR and switch to ATTI mode.

The Yaw/magYaw separations all occurred after the Mavic was rotated. The response is to then adjust the Yaw so that it aligns with the magYaw value. This from the eventLog stream.

31.900 : 9718 [L-FDI]ns req:fdi,0to0,reason:fusion.gps_yaw_err,result:fail for same index
33.104 : 9778 [L-FLYMODE][Ctrl<1>] REQ_RC_NORMAL ATTI ctrl_atti
33.164 : 9781 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
35.340 : 9890 [L-FDI]NS(0) COMPASS(1): fault off, over_large
40.492 : 10147 [L-NS][AHRS] yaw aligned 1

I've seen a few of these incidents but haven't come up with a compelling explanation. One explanation is that the Z axis gyro gain is set too low. But, then this behavior would be expected in subsequent flights. You could try a test flight. Bring the Mavic to a hover greater than 5 meters altitude. Then do several abrupt 90° rotations, each followed by 20 secs of hover (no stick input). The .DAT will show if this is a continuing problem.

As I mentioned previously it's hard to believe that any EM fields could cause this. EM radiation and a magnetic field are different physical phenomena. To actually change the magnetic field that could be measured by the magnetometers would require a very, very low frequency EM field. Although it's possible for RF energy to couple noise into the circuits that carry the data there doesn't seem to be any evidence of this.

What do you think @sar104

 

[sar104]

Don't really have an answer, just some observations.

The .DAT doesn't show any of the usual characteristics of having launched from a geomagnetically distorted site. In particular, Yaw (determined mostly from IMU data) and magYaw (determined from just the magnetometers) don't separate as the Mavic ascends. If the launch site is geomagnetically then distorted magYaw would change as the Mavic rises above the distortion but Yaw remains constant because the IMU data is saying the Mavic hasn't rotated.

View attachment 52700

There are several points in the flight where a Yaw/magYaw separation occurs. The first time resulted in the YAW_ERROR and switch to ATTI mode.

The Yaw/magYaw separations all occurred after the Mavic was rotated. The response is to then adjust the Yaw so that it aligns with the magYaw value. This from the eventLog stream.

31.900 : 9718 [L-FDI]ns req:fdi,0to0,reason:fusion.gps_yaw_err,result:fail for same index
33.104 : 9778 [L-FLYMODE][Ctrl<1>] REQ_RC_NORMAL ATTI ctrl_atti
33.164 : 9781 [L-FMU/LED]action changed. imu error:ns_abnormal(3)
35.340 : 9890 [L-FDI]NS(0) COMPASS(1): fault off, over_large
40.492 : 10147 [L-NS][AHRS] yaw aligned 1

I've seen a few of these incidents but haven't come up with a compelling explanation. One explanation is that the Z axis gyro gain is set too low. But, then this behavior would be expected in subsequent flights. You could try a test flight. Bring the Mavic to a hover greater than 5 meters altitude. Then do several abrupt 90° rotations, each followed by 20 secs of hover (no stick input). The .DAT will show if this is a continuing problem.

As I mentioned previously it's hard to believe that any EM fields could cause this. EM radiation and a magnetic field are different physical phenomena. To actually change the magnetic field that could be measured by the magnetometers would require a very, very low frequency EM field. Although it's possible for RF energy to couple noise into the circuits that carry the data there doesn't seem to be any evidence of this.

What do you think @sar104

I just finished the same analysis and came to the same conclusion. Looking at the overall yaw performance for both flights, which are covered by that DAT file, shows reasonable agreement:



Looking at the first 50 seconds of the first flight, the trigger for the ATTI switch was a separation of compass and IMU yaw just after 30 seconds that was, as you pointed out, unrelated to height, and so was not caused by magnetic interference at launch. I also agree that it was most unlikely to have been due to EM emissions from the tower.



The behavior of the aircraft was normal for the 10 seconds of ATTI mode before the FC realigned the yaw and switched back. It looks like the OP was simply unfamiliar with ATTI flight characteristics. The aircraft had been accelerated with full forward elevator for 6 seconds to nearly 8 m/s when it went to ATTI. Full elevator was maintained for another 2 seconds after the switch, and then released, at which point the aircraft responded just to the wind, which was around 15 mph out of the south.

 

[adam3112]

Thanks guys this is Interesting stuff, although I must admit this leaves me with more questions than answers and it's quite complicated with the limited knowledge I have on how the IMU / Compass mechanics work.

So my understanding of this (which I hope is correct) is that the Yaw/magYaw were getting separated whilst rotating the Mavic throughout the flight, but the first such separation caused the atti / lost gps? Is this a common phenomena and am I expected to get this on subsequent flights?

What are the roles of the Yaw / MagYaw and Why did the Mavic only lose gps once and not every time they separated?

It is a little bit disturbing that there is no root cause or explanation for this and I fear it will trigger again on a future flight where I could lose or crash the Mavic.

When I get the opportunity (probably on the weekend) I'll take the Mavic out and do the test of hovering and turning 90 degrees with 20 sec intervals like Bud suggested, although is it likely to trigger what has happened already, again?

PS. Sorry if I have got the wrong end of the stick with that was explained from the flight data as I'm new to the world of drones.

 

[BudWalker]

Thanks guys this is Interesting stuff, although I must admit this leaves me with more questions than answers and it's quite complicated with the limited knowledge I have on how the IMU / Compass mechanics work.

So my understanding of this (which I hope is correct) is that the Yaw/magYaw were getting separated whilst rotating the Mavic throughout the flight, but the first such separation caused the atti / lost gps? Is this a common phenomena and am I expected to get this on subsequent flights?

What are the roles of the Yaw / MagYaw and Why did the Mavic only lose gps once and not every time they separated?

It is a little bit disturbing that there is no root cause or explanation for this and I fear it will trigger again on a future flight where I could lose or crash the Mavic.

When I get the opportunity (probably on the weekend) I'll take the Mavic out and do the test of hovering and turning 90 degrees with 20 sec intervals like Bud suggested, although is it likely to trigger what has happened already, again?

PS. Sorry if I have got the wrong end of the stick with that was explained from the flight data as I'm new to the world of drones.

It didn't actually lose GPS. GpsHealth went to 0 but the numSats actually increased from 13 to 18. GpsHealth is the confidence that the FC has in the coords that it computes from GPS, IMU data, and. to a lesser extent magnetometer data. Since there was a problem with either the IMU or magnetometer data the FC didn't have confidence in the coords that it's computing. Confusing? That's why gpsHealth has been renamed navHealth in the newest CsvView/DatCon version.

Don't know why this happened on just the first rotation. There were Yaw/magYaw separations on other rotations that didn't cause the FC to lower the navHealth.

The purpose of the test is to see if this behavior is triggered.

 

[adam3112]

I will conduct the test on my next outing.

If the test fails and I get this problem again, I guess I'll just have to fly the Mavic straight backwards and forwards without turning it from now on lol.

On a serious note I'm losing a lot of confidence in my Mavic.

Edit: Is is worth re-calibrating the compass and IMU?

 

[FoxhallGH]

Keep that piece of advice that was given earlier by ac0j in mind adam3112 ... If you look at the Google map picture (your log files) and compare that to your take-off video, it appears that the arrow depicting the Mavic on the Google Earth map, is pointing about 20 degrees left of the direction your video shows the Mavic to be looking in. That may confirm some earlier analysis of the initial data. As part of your pre-takeoff check, once the Mavic tells you it has established GPS - expand the map view and check that the arrow on the map, is pointing the same way that the actual Mavic is on the ground in front of you.
I think you had some bad luck first-flight. My first flight ever went way wrong (not a DJI drone) and ended up 20 metres up a tree when I tested RTH - I still don't know where ti thought home was ... Took me a week to get it down! Fly safe & have fun.