Insane out of control crash

[ml92760]

I will no longer use that guard to be on the safe side.....thanks again..

 

[Tomiw10]

Hi! I have a new Mavic Pro and I do not have much idea about it. The problem is that I can not connect both the rc and the phone at the same time. For example, if I put the aircraft in "wifi" mode, the only way to drive it is through the phone whereas if it is in "rc" mode, I can drive it throuhg the remote control but can not see anything on the phone (like the picture). I hope you undertand me.

 

[BudWalker]

Good and informative responses from the crew here. I know when I try to calibrate and I'm too close to my car, or any vehicles, I get a calibration error. I have to move at least 20 or so yards away. Surprised there wasn't a calibration error given before the your attempt to take off. Or, maybe there was but you didn't notice it when calibrating.

Exactly! A flawed calibration due to external magnetic anomalies is very unlikely. A geomagnetic field distorted enough to cause a problem will result in a rejected calibration with the "interference, please move" message.

When you calibrated on top of that car park you gave the mavic a really bad snapshot, with tons of interference and noise created by the huge amount of metal in the parking structure. Not good. The moment the mavic starts moving, that field is changing drastically which will completely confuse the poor thing and cause it to freak out.

Although commonly believed, this isn't correct. It's mathematically impossible for a calibration to detect and compensate for magnetic effects that are external to the Mavic.

All that is required for a calibration is that the geomagnetic field is straight and uniform within the confines of the compass dance. Easy to have that on a broad flat structure like the parking structure. Typically, the effects of sidewalk rebar don't extend much above 2 or 3 feet.

I've looked at the data for literally dozens of incidents like this and not one was caused by a flawed calibration.

@AJButterfield if you provide the .DAT for this flight we can know for sure what happened. The .txt would probably provide some helpful data, but the .DAT is better. Look here to see how to retrieve the .DAT. From your description it seems likely this was just an ordinary launch-from-a-geomagnetically-distorted-site incident. Nothing to do with the compass calibration.

 

[lisadoc]

Hi! I have a new Mavic Pro and I do not have much idea about it. The problem is that I can not connect both the rc and the phone at the same time. For example, if I put the aircraft in "wifi" mode, the only way to drive it is through the phone whereas if it is in "rc" mode, I can drive it throuhg the remote control but can not see anything on the phone (like the picture). I hope you undertand me.

This belongs in its own thread. No one is going to see your question here. I happened to see it but only by luck.

To answer your questions - WiFi mode is working correctly. The point of it is to skip the controller and operate the drone through your phone. Hence the RC won't connect and won't control the drone. That's the whole point of WiFi mode. No RC needed.

For your RC mode, it's probably an issue with your cable connection to your device. I would suggest dumping the cables that came with the Mavic and using your own cable and connecting it in the bottom USB port in the middle of the Mavic (like this photo)

It operates the same way as the connection on the left side of the controller, but far more reliable and efficient. Also, be sure to start the controller first, the drone 2nd, and the app last. Disconnect and reconnect the phone if you don't see an image after starting all three. I personally like to start the app and then connect the cable. Saves me from having to pull it and replug it if it doesn't "catch" the video signal.

 

[Blacksails]

Exactly! A flawed calibration due to external magnetic anomalies is very unlikely. A geomagnetic field distorted enough to cause a problem will result in a rejected calibration with the "interference, please move" message.


Although commonly believed, this isn't correct. It's mathematically impossible for a calibration to detect and compensate for magnetic effects that are external to the Mavic.

All that is required for a calibration is that the geomagnetic field is straight and uniform within the confines of the compass dance. Easy to have that on a broad flat structure like the parking structure. Typically, the effects of sidewalk rebar don't extend much above 2 or 3 feet.

I've looked at the data for literally dozens of incidents like this and not one was caused by a flawed calibration.

@AJButterfield if you provide the .DAT for this flight we can know for sure what happened. The .txt would probably provide some helpful data, but the .DAT is better. Look here to see how to retrieve the .DAT. From your description it seems likely this was just an ordinary launch-from-a-geomagnetically-distorted-site incident. Nothing to do with the compass calibration.

Interesting. This does not play well with my experience though. Many times flying from metal supported structures (most recently, flying an Inspire from a large concrete covered bridge in Iceland) the aircraft just would not fly correctly unless it was in attitude mode, ignoring the compass. It would behave erratically or begin to toilet bowl, an occurrence I have seen on the Inspire, phantoms, mavic and the old DJI Wookong-M equipped heavy lifters I used to operate.
Obviously you can't just switch to Atti with the mavic so it becomes an issue. Surely if it is enough to mess with the flight characteristics, then it is enough to mess with a calibration? Not arguing here, just trying to match my experience and maybe garner a better understanding. I am no geologist and certainly don't understand the full workings of the compass onboard DJI drones.

Again, whilst in Iceland, we flew at Godafoss. Huge magnetic interference here presumably from the volcanic rock. When I took off, the Inspire was acting strange, beginning a slow toilet bowl. Once in Atti mode and not listening to the compass, she flew perfectly. Another guy turned up with a phantom and made a compass calibration. He took off and about 2 minutes into the flight he started to experience the toilet bowl issue also, he didn't switch to Atti mode and it got progressively worse and uncontrollable (quickly) until it eventually ended up in the waterfall. Surely, calibrating in this area and then (if it hadn't of crashed) moving to a different area would have completely messed up his cal?

 

[BudWalker]

Interesting. This does not play well with my experience though. Many times flying from metal supported structures (most recently, flying an Inspire from a large concrete covered bridge in Iceland) the aircraft just would not fly correctly unless it was in attitude mode, ignoring the compass. It would behave erratically or begin to toilet bowl, an occurrence I have seen on the Inspire, phantoms, mavic and the old DJI Wookong-M equipped heavy lifters I used to operate.
Obviously you can't just switch to Atti with the mavic so it becomes an issue. Surely if it is enough to mess with the flight characteristics, then it is enough to mess with a calibration? Not arguing here, just trying to match my experience and maybe garner a better understanding. I am no geologist and certainly don't understand the full workings of the compass onboard DJI drones.

Again, whilst in Iceland, we flew at Godafoss. Huge magnetic interference here presumably from the volcanic rock. When I took off, the Inspire was acting strange, beginning a slow toilet bowl. Once in Atti mode and not listening to the compass, she flew perfectly. Another guy turned up with a phantom and made a compass calibration. He took off and about 2 minutes into the flight he started to experience the toilet bowl issue also, he didn't switch to Atti mode and it got progressively worse and uncontrollable (quickly) until it eventually ended up in the waterfall. Surely, calibrating in this area and then (if it hadn't of crashed) moving to a different area would have completely messed up his cal?

Well, I'd have to see the .DAT for one of these flights to know for sure. But, from your description it sounds like these incidents are probably caused by launching from a geomagnetically distorted site. I.e. it's not related to the calibration, it's the launch site itself. Basically, what happens is that Yaw gets initialized to a compass heading that's incorrect because of something like rebar in the concrete, manhole cover, etc. After the AC rises to 2 or 3 feet the compass becomes correct because the AC rises above the extent of the magnetic distortion. But Yaw remains at the initial, incorrect value. Yaw is computed mostly from gyros, and accelerometers; magnetometer data is used only at a very low gain. All this happens with a non-flawed calibration.

If you're interested here are a couple of incidents where this happened.

Compass error

And, this one seems to be popular
Looking for Trouble ??

 

[tracer bullet]

I've launched from concrete w/ rebar numerous times and never noticed an issue. Choosing that word specifically as opposed to saying I factually did not have any issues. I know not to calibrate there, but didn't' think much about launching from it. Right now I'm feeling really lucky.

Hypothetically speaking if you needed (or admittedly just really wanted) to launch from a parking garage upper deck, is there anything to check other than the heading indicator in the app and comparing it to real life (as mentioned in the first link above)? If you weren't excited about a hand launch, would it be OK to get say 4 feet off the ground, hover and wait and look for issues, then get up a bit higher, wait and check again, and then start flying away?

 

[BudWalker]

I've launched from concrete w/ rebar numerous times and never noticed an issue. Choosing that word specifically as opposed to saying I factually did not have any issues. I know not to calibrate there, but didn't' think much about launching from it. Right now I'm feeling really lucky.

Hypothetically speaking if you needed (or admittedly just really wanted) to launch from a parking garage upper deck, is there anything to check other than the heading indicator in the app and comparing it to real life (as mentioned in the first link above)? If you weren't excited about a hand launch, would it be OK to get say 4 feet off the ground, hover and wait and look for issues, then get up a bit higher, wait and check again, and then start flying away?

Checking the heading indicator in the app is the single most effective means to prevent launch time fly aways. If this is correct then Yaw is correct. The incidents that I referred to above all had an app heading indicator that was incorrect.

Launching and looking for issues could be a problem. While hovering the AC is probably going to want to make some small corrections. To do that it needs to know which way to adjust. If Yaw is compromised then it will move the wrong way.

Another check that can be made prior to launch is to insure that gpsHealth is 4 or 5. Number of satellites isn't as effective a measure. gpsHealth can be seen by looking at the gps bar graph on the RC.

 

[FrankIII]

Happy to help.

Definitely not possible. On the mavic you have a selection between sport and GPS mode, both of which use the compass.

On drones like the phantom/Inspire/etc you have a 3 position switch with attitude mode in the middle.

There is a way (through hacking) of swapping the sports mode switch to go between gps mode and Atti mode, but I wouldn't recommend that. I tried it and although Atti became switchable, it still seemed to be attempting use compass data, as it behaved erratically indoors, exactly as it does in gps mode with interference.

How about putting a piece of aluminum foil on top of the MP where the GPS antennas are located? That would block / shield the GPS signals causing the AC to stay in the Atti mode wouldn't it???

 

[Blacksails]

How about putting a piece of aluminum foil on top of the MP where the GPS antennas are located? That would block / shield the GPS signals causing the AC to stay in the Atti mode wouldn't it???

It would. I believe quite a few people have done this, however... it does mean your beautiful Mavic ends up looking like a flying packed lunch.

 

[royster]

I do think that my Phantom in ATT mode is much more predictable to fly that when the Mavic goes into ATT mode, is this just me?
Or maybe the design/flying characteristics of the Mavic (in ATT) are just that much different than a Phantom?
Whatever the reason, i;m much more condfident& enjoy flying the P3 in ATT mode,