Lost all control and crashed

[Just Mark]

Yes... but it can be tricky. At batteryOn the Yaw value is initialized to a value determine by the compass. After that the Yaw value is determined mostly by the IMU (gyros, and accelerometers) with the compass providing corrective inputs at low gain. If Yaw is correct but the AC is then moved to a geomagnetically distorted location the Yaw value will start a gradual change to the incorrect value dictated by the distortion. So it's possible that if launch occurs without delay after moving then Yaw will still be OK.

Your post could be summarized as asking how can it be determined if Yaw is correct before launch. The single most effective method of determining thisis correct is to look at the red triangle heading indicator on the Go App map. If that is inconsistent with the actual heading then Yaw is incorrect. It's almost guaranteed that erratic flight, possibly a fly away, will occur after launch.

This method will always work. If you just gotta launch from concrete just make sure this heading indicator is correct before launch.This method will work in those cases where your P4 is sitting in the middle of a grassy field but unbeknownst to you there is a hidden 1/4 " washer sitting directly below the compass.

Thank you!!!

 

[mark99]

The logs are what the drone did, not always what the operator did, they can and do randomly go berserk and fail to respond to any control, the blame the operator because the logs say is getting tired and lame.

 

[sar104]

The logs are what the drone did, not always what the operator did, they can and do randomly go berserk and fail to respond to any control, the blame the operator because the logs say is getting tired and lame.

Outstanding - the trusty #FAKELOGS defense.

 

[Meta4]

The logs are what the drone did, not always what the operator did, they can and do randomly go berserk and fail to respond to any control, the blame the operator because the logs say is getting tired and lame.

You don't know much about the flight logs, do you?
They are just like the black box data recorder in a commercial airliner and allow us to see exactly what was happening during a flight.
That includes what control inputs the pilot made as well as comprehensive flight data.
They help us to find lost Phantoms, identify the causes of incidents including malfunctions that end up being warranty issues and pilot error.

 

[rodgerbwg]

Hi All

I was asked by a club member to film some night flying aircraft at our club site
I set the quad up as normal, The first flight of about ten mins went well, I swapped out the battery for a new one then took the quad up again, About 15 seconds into the flight my green flashing lights turned red, The quad started drifting away from me quite erratically, i tried to drop the quad down and there was no control response know matter what I did with the sticks it was twitching all over the place and then did a final tip and ended up at the bottom of a very large pond next to my flying club, which can be seen on the map view this all happened within a one min flight, the distance travelled was about 50 to 75 meters i would guess.

One of the younger club members managed to retrieve the quad from the bottom of the pond, I removed the battery quickly, When i got home i removed the top then washed everything with warm clean water, It’s now in the drying out stage so i don’t know what the damage is to the electrics, gimbal and camera.

I’m pretty sure I have done the right thing by stripping it down and washing the dirty water off all the electronics, I used an hairdryer to dry it off, its now in the airing cupboard and been there for 3 days.

The craft was purchased in April 2017 from Clifton Cameras and is still under warranty, do i first take this up with them or DJI.

I wish i kept my phantom 3 pro now, that was solid as a rock has was my old Naza 2 on my 450.

I thought fly-aways was a thing of the past, Im totally shocked this happened with this newer tech especially with all the backup equipment it Has.
I have taken a look at the flight recording and there was a problem with the imu and then a compass error, Then the last one was a gimbal error.

17#
link to the flight record

https://www.phantomhelp.com/logviewer/53CR1WKC0E5G0L43LBTV/

Any advice welcome.

Supra59uk

Short story here about calibrating your Compass. I received a new Phantom 4 Obsidian last week. I waited for my FAA Registration stickers to arrive and be attached before putting it into the air for the first time. I had previously set all of my usual parameters, distance, height, RTH, etc. I did not calibrate the compass on that day as it was raining outside. Had a decent day arrive and fired things up inside and brought it outside. I had a Compass error warning on my tablet that cleared when I was outside which I found a bit strange but figured things were all set. I brought it up to about 25 ' and it was all over the place. It would not stay in one place, it was drifting and twisting all over the sky. I could not make it stay in any one place. It was the most erratic flight that I have ever had and quite frankly extremely nerve wracking. I was extremely close to a couple of trees and my house, I thought for sure it was the first and last flight of the new Obsidian. I was having one heck of a time trying to get it back to the driveway to bring it down. I had to climb to avoid the trees and really fought it to keep it away from the house. I have been flying for 10 or more years now and have my Part 107 Certificate. I practise flying in Attitude mode and this along with years of experience is what I feel saved the day. After a good 5 minute nerve wracking flight I made a nice soft landing. After a break I did a compass calibration. I then put it back in the air and steady as a rock. Another lesson learned and a bit more experience.

 

[Northern Canadian]

The logs are what the drone did, not always what the operator did, they can and do randomly go berserk and fail to respond to any control, the blame the operator because the logs say is getting tired and lame.

I get what you are saying. I had a momentary loss of control at a flight last week. It drifted away, lost about 6 feet of altitude, and was not responsive. Probably a 5 second or so incident, but it underscored that something will go wrong at some point. I am a hands on pilot. I don't use anything automatic.

 

[BudWalker]

The logs are what the drone did, not always what the operator did, they can and do randomly go berserk and fail to respond to any control, the blame the operator because the logs say is getting tired and lame.

There's an ambiguity here - might be good to disambiguate. By "not always what the operator did" do you mean that
1) the log doesn't reflect actual pilot inputs such as stick movements, or
2) the AC ignores pilot inputs. E.g. Yaw is compromised but the AC is trying to hover in a fixed position.

 

[amstar]

I have read with great interest this post, especially since unlike most the OP provided enough data which was easy to decipher. Sar and Budwalker as usual did a great job of data analysis.
But IMHO, the crux of the problem which our new members cannot reconcile is the in-explainable stupidity of how in this modern age DJI software allows a fault ridden navigational tool like a compass to over-ride a much more reliable tool like a tight GPS lock leading to a crash.
A sophisticated expensive craft like a P4P is not meant for Atti mode flying; its supposed to automate the flying so that we can concentrate on photography.
We have our RC helis if we just seek the thrill of RC flying.

 

[F3honda4me]

I have read with great interest this post, especially since unlike most the OP provided enough data which was easy to decipher. Sar and Budwalker as usual did a great job of data analysis.
But IMHO, the crux of the problem which our new members cannot reconcile is the in-explainable stupidity of how in this modern age DJI software allows a fault ridden navigational tool like a compass to over-ride a much more reliable tool like a tight GPS lock leading to a crash.
A sophisticated expensive craft like a P4P is not meant for Atti mode flying; its supposed to automate the flying so that we can concentrate on photography.
We have our RC helis if we just seek the thrill of RC flying.

Because without the compass the craft doesn't know which way it's facing. So, it can't make positional corrections to hold position via gps, because it doesn't know which motors to spin up or down in order to hold gps position.

 

[amstar]

Because without the compass the craft doesn't know which way it's facing. So, it can't make positional corrections to hold position via gps, because it doesn't know which motors to spin up or down in order to hold gps position.

All modern GPSs have an electronic compass. They can pinpoint your location and direction based on how detailed a map you load.

 

[amstar]

All modern GPSs have an electronic compass. They can pinpoint your location and direction based on how detailed a map you load.

I mean we don’t fix a compass on the dash of our car..

 

[F3honda4me]

I mean we don’t fix a compass on the dash of our car..

That's great. But we are talking about a DJI drone. Not a car.

All modern GPSs have an electronic compass. They can pinpoint your location and direction based on how detailed a map you load.

Not exactly. The DJI drone is a drone, not a car. Consumer GPS such as those in a car don't actually have a compass. But in something like a car, assuming the car has no compass, the GPS will ASSUME the direction you are facing based upon the direction you are moving, since cars are generally moving forward while driving, the GPS in your car will point the direction your car is moving.

Now if you're referring to the additional hardware that can be put in place in the form of a compass tacked onto a GPS, that's no different than the TWO compasses already in the drone. And those compasses, just like the "electronic" compasses you are referring to (which in fact, the DJI compasses ARE electronic compasses but they use magnetic readings) are prone to the same exact interference. So basically, your solution isn't a solution, it's just another compass that will have the same failings as the already existing electronic compasses inside the DJI drone.

 

[amstar]

That's great. But we are talking about a DJI drone. Not a car.

Not exactly. The DJI drone is a drone, not a car. Consumer GPS such as those in a car don't actually have a compass. But in something like a car, assuming the car has no compass, the GPS will ASSUME the direction you are facing based upon the direction you are moving, since cars are generally moving forward while driving, the GPS in your car will point the direction your car is moving.

Now if you're referring to the additional hardware that can be put in place in the form of a compass tacked onto a GPS, that's no different than the TWO compasses already in the drone. And those compasses, just like the "electronic" compasses you are referring to (which in fact, the DJI compasses ARE electronic compasses) are prone to the same exact interference. So basically, your solution isn't a solution, it's just another compass that will have the same failings as the already existing electronic compasses inside the DJI drone.

That’s not correct, all car navigational systems have, Please refer to the manuals of any or pros and cons of modern navigational equipment.
DJI NAZA is different, I will quote from an RC build:
“If you want to use the external magnetometer (built in in your GPS) and you have a FC with the same magnetometer (HMC5883L is very common), you have to disable it physically on your FC: remove chip from board or cut a trace. You can't use two identical chips/magnetometers on the same I2C bus.

When using DJI NAZA gps this is not true, DJI NAZA sends compass over serial and does not use the I2C bus)
On MPU9250 board internal magnetometer is an AK8963, most GPS pucks are HMC5883L. So no need to remove hardware, only choose which one to use with cli command mag_hardware”

 

[supra59uk]

I feel your pain.

1. From my experience of dropping a P3S into Long Island Sound, nothing will bring it back to normal... even if it works for a few flights. And from my experience where a SW 1000 took off, immediately wobbled the wrong direction and crashed into a building ($5,000 worth of damage), I rarely trust GPS-mode on any flight.
2. It looks like DJI will conclude that you failed to notice the switch to ATTI mode, and then dropped it in the salt-water pond.
3. I now almost never take off in GPS-mode. I practice all flights in ATTI mode, and ensure I know which way to fly it if the winds kick up (that is, where it is pointing, what I would have to do rotate, left, right, etc.) to get it back to its home.
4. If I have to take off in GPS mode, I now always use a tether: paracord from the ground up to a little pulley under the drone, and back down to me. I step on it. If the take-off is good and control is good, I remove my foot, take the P3S up, and the end of the rope goes up, slips through the pulley and drops. At this point the P3S is un-thered, and I can fly with confidence.

But, sorry for your loss.

Thanks for the post
but let me clear something up here I have flown my phantom three and this one in atti mode, and this was nothing like flying in attimode, it felt to me the sticks was not reacting to the movements even so they was it certainly did not respond anything like flying in atti mode, so take my word for it you practising in atti mode will not help you control the quad if it does was this did, I'm sure many will disagree with my post but till you have had this situation you wouldn't know. I have flown my vortex 250 racing quad in manual mode which was no help, the clever guys who as helped and checked out the logs told me the quad had not lost connection which they are much more clued up than me so I'm sure they are correct even though it did not feel like I had any control of the situation maybe they could explain why this is ? again I do thank you all for the help you have given me

Cheers supra

 

[F3honda4me]

That’s not correct, all car navigational systems have, Please refer to the manuals of any or pros and cons of modern navigational equipment.
DJI NAZA is different, I will quote from an RC build:
“If you want to use the external magnetometer (built in in your GPS) and you have a FC with the same magnetometer (HMC5883L is very common), you have to disable it physically on your FC: remove chip from board or cut a trace. You can't use two identical chips/magnetometers on the same I2C bus.

When using DJI NAZA gps this is not true, DJI NAZA sends compass over serial and does not use the I2C bus)
On MPU9250 board internal magnetometer is an AK8963, most GPS pucks are HMC5883L. So no need to remove hardware, only choose which one to use with cli command mag_hardware”

What is an Electronic Compass? GPS Review

Also DJI Naza is not what it used in the P4P so all of that doesn't apply. Electronic compasses in and of themselves cannot provide what you're looking for because they can only tell you direction of travel in relation to north. UNLESS they are paired with a magnetic reading of some kind. Otherwise they still cannot tell you which way a drone is facing since a drone can travel in any direction regardless of the direction it is facing. To my understanding based upon what DJI has said on their forums, they are electronic compasses that use a magnetic reading, which is why they can determine the direction the drone is facing in relation to magnetic north.

At the end of the day, regardless of what technology is used in these drones, they CANNOT return to home on their own without the compass functioning. There isn't some magic compass out there that doesn't need magnetic readings in order for the drone to tell which way it's facing, which is required in order for the drone to fly itself home.

 

[Droneboondocker]

Yes... but it can be tricky. At batteryOn the Yaw value is initialized to a value determine by the compass. After that the Yaw value is determined mostly by the IMU (gyros, and accelerometers) with the compass providing corrective inputs at low gain. If Yaw is correct but the AC is then moved to a geomagnetically distorted location the Yaw value will start a gradual change to the incorrect value dictated by the distortion. So it's possible that if launch occurs without delay after moving then Yaw will still be OK.

Your post could be summarized as asking how can it be determined if Yaw is correct before launch. The single most effective method of determining this is to look at the red triangle heading indicator on the Go App map. If that is inconsistent with the actual heading then Yaw is incorrect. It's almost guaranteed that erratic flight, possibly a fly away, will occur after launch.
View attachment 90653

This method will always work. If you just gotta launch from concrete just make sure this heading indicator is correct before launch.This method will work in those cases where your P4 is sitting in the middle of a grassy field but unbeknownst to you there is a hidden 1/4 " washer sitting directly below the compass.

Thank you for this great nugget of knowledge. Would it also be helpful to check the two compass interference bars in the app settings before lauching the craft?

 

[amstar]

What is an Electronic Compass? GPS Review

Also DJI Naza is not what it used in the P4P so all of that doesn't apply. Electronic compasses in and of themselves cannot provide what you're looking for because they can only tell you direction of travel in relation to north. UNLESS they are paired with a magnetic reading of some kind. Otherwise they still cannot tell you which way a drone is facing since a drone can travel in any direction regardless of the direction it is facing. To my understanding based upon what DJI has said on their forums, they are electronic compasses that use a magnetic reading, which is why they can determine the direction the drone is facing in relation to magnetic north.

At the end of the day, regardless of what technology is used in these drones, they CANNOT return to home on their own without the compass functioning. There isn't some magic compass out there that doesn't need magnetic readings in order for the drone to tell which way it's facing, which is required in order for the drone to fly itself home.

No one is questioning the value or role of a compass in navigation.
What is not clear is why a compass error would trigger a flight mode switch from a GPS hold to Atti, which would definitely make matters worse in terms of controlling the AC for an inexperienced OP.
The P4 is a great flying machine with many redundancies built in. This particular software algorithm has never been satisfactorily explained on any forums.

 

[Meta4]

This particular software algorithm has never been satisfactorily explained on any forums.

The Phantom can be flown without GPS but it can't be controlled without the compass.
When the flight controller is getting conflicting data from the compass and GPS, it can't make sense of the data conflict.
Since it can't work without the compass, it drops the GPS data.

 

[sar104]

No one is questioning the value or role of a compass in navigation.
What is not clear is why a compass error would trigger a flight mode switch from a GPS hold to Atti, which would definitely make matters worse in terms of controlling the AC for an inexperienced OP.
The P4 is a great flying machine with many redundancies built in. This particular software algorithm has never been satisfactorily explained on any forums.

It's certainly a good question and has caused much confusion, but I think the answer is actually fairly simple even though the underlying control processes are not. It hinges on the exact method that the FC uses to determine attitude (pitch and roll), orientation (yaw), velocity and position. Inferring from the existing literature on 9-sensor IMU-controlled flight, it turns out that the GPS data are not primary in any of those for flight control, largely because the resolution is too low both spatially and temporally to be part of any fast control feedback loop.

The 9-sensor IMU is almost certainly doing all the fast stuff using some variant of one of the standard sensor fusion method. In these methods the 3-axis rate gyros and accelerometers provide the primary time-resolved data on all the variables listed above but, since they are measuring first and second derivatives with respect to time of position and orientation respectively, they are subject to accumulating errors (effectively drift) in the integrated (orientation and velocity) and double-integrated (position) data that build over time. Those are corrected for at a lower rate using the sensor's 3-axis magnetometers (for orientation), the GPS carrier wave Doppler data (for velocity) and the GPS position data (for position).

This combined approach allows the FC to detect changes in orientation orientation fast enough (200 Hz) to apply the motor adjustements necessary to correct for them. It also allows it to know the absolute values for position and orientation well enough that in P-GPS mode it can apply the appropriate motor adjustments to correct for horizontal positional drift - i.e. it knows, relative to itself, exactly which direction in which to move. Vertical position (altitude) is easier because the up and down directions are uniquely and absolutely defined by gravity, whereas compass heading is only known relative to the earth's magnetic field.

So now it becomes clearer what happens if the magnetometer data are compromised. Yaw changes (rotation) are still detected by the rate gyros and so yaw can be stabilized, but the FC no longer knows in which direction it is facing. That means that when positional drift is detected from the slower GPS data it is unable to know the appropriate motor adjustments to correct for that. Simply using trial and error - i.e. applying a random correction and seeing how that changes the long-term position drift is not a viable fast control strategy. Due to wind if nothing else, the FC does not even know the relationship between heading and track, and so at best it would end up flying around in circles hunting for a location. For that reason it switches to ATTI mode in which it does not attempt to figure out heading (it assumes that it cannot correct for rate gyro yaw drift because the compass data are bad) and therefore cannot hold horizontal position either.

Those consequences are undesirable but inevitable, and it would require a completely different trial-and-error-based navigational algorithm for the FC to attempt to fly in a specific direction or to a specific location. I'm not aware of any current work on that problem - does anyone else have any insights in that area?

 

[BudWalker]

Thank you for this great nugget of knowledge. Would it also be helpful to check the two compass interference bars in the app settings before lauching the craft?

I'm not sure I know the answer to this since I'm not really sure what the interference bars represent. They almost have to be based on field strength; similar to checking the MagMod value to see if it's in the proper range. A field strength outside it's normal range isn't a problem per se, it's just an indication of a different geomagnetic field which could be a problem. But, like I say, I don't really know what the interference bars represent so I can't say if it helps to check them prior to launch.

But, it's clear that Yaw can be incorrect yet there will be no indication from the interference bars. The only way to know this is by looking at the App heading indicator.