HELP! P3 Professional went out of control and crashed.

[Meta4]

Is this a fact? Most of my drones dont even have a compass, they are perfectly controllable using gyro's and accelerometers alone. Obviously the drone cant navigate without compass (well, technically it could, the pixhawk will manage without working compass, guestimating its bearing from GPS position data, but its tricky and the result not pretty), but basic control where the pilot controls the yaw should certainly not need one?

Flying without a compass is one thing, but you can't turn the compass off in the Phantom.
If it's supplying bad data to the flight controller, you are going to have to deal with a screwy Phantom and the flight record for this one shows you how tricky that can be.

 

[Vertigo]

But the phantom reports that the compass data is garbage, its not difficult to compare compass and gyro data to figure this out (and the phantom almost certainly does this to figure out the data is no good); why would it use it when the quad is under pilot control and there is no need for compass data ? I cant believe this is a problem in attitude mode.

 

[Meta4]

If you can't believe it just look at the flight record above in phantomhelp.com. much of the flight was in atti mode but it didn't make it any more controllable. The moral of the story is don't screw up your compass and you can fly safely without the worries

 

[Vertigo]

Thats not conclusive at all. The pilot stated he never switched out of P GPS mode; the phantom did this a few times automatically, for relatively brief periods, but I cant tell from the logs if it was controllable then, or even if that is the same as manually selecting attitude.

Note that the software will instruct you to switch modes when this happens "Compass Error, Exit P-GPS Mode". I really hope that shouldnt not read 'Compass Error, Exit P-GPS Mode but you are still f*cked'.

I think we all agree its best to ensure good working compass, but if its true a phantom can not be controlled reliably in attitude mode with a failing compass, I would consider that a huge problem. Of all sensors, the compass is the one that is the least reliable and most easily influenced (and arguably, least necessary to maintain manual control).

 

[Meta4]

Thats not conclusive at all. The pilot stated he never switched out of P GPS mode; the phantom did this a few times automatically, for relatively brief periods, but I cant tell from the logs if it was controllable then, or even if that is the same as manually selecting attitude.

It was most of the flight with very brief times in P-GPS between them.
When the Phantom is in atti mode, it is in atti mode regardless of whether you or the Phantom switches.
For all of the red trace below the Phantom appears to have been out of control.

Of all sensors, the compass is the one that is the least reliable and most easily influenced (and arguably, least necessary to maintain manual control).

As long as you don't give your compass a bad calibration, the compass is very reliable.
A compass error isn't a failure of the compass, it's a properly compass telling you that there is a problem with the magnetic environment.

 

[Vertigo]

As long as you don't give your compass a bad calibration, the compass is very reliable.
A compass error isn't a failure of the compass, it's a properly compass telling you that there is a problem with the magnetic environment.

Im just saying a compass reading is (far) less reliable than readings from say gyro's or accelerometers as its kinda hard to influence the gravitational environment. What you say is true, but it doesnt really matter if its the environment or the instrument that makes the reading unreliable and a common failure that any drone should be able to cope with, even if with pilot assistance.

If the effect of not having a working compass in attitude mode is just that heading may drift a little, like with most non GPS quadcopters, then thats fine, we have a stick to compensate. But it should be easily controllable and Im yet to be convinced it isnt with the phantom.

Are you saying the opposite based on this log, or is this common knowledge?

 

[MikeyOnline]

I have to agree with Vertigo on the compass in Atti mode. Atti is supposed to be manual/pilot control. I could see it still using the compass to try to display telemetry, but only for display purposes. I don't see any circumstance where it should try to control or alter the control of the AC with the compass or GPS while flying in Atti mode. On what basis would it make alterations to flight characteristics based on compass anyway... when you are supposed to be under manual control? I don't see the logic. Unless there is some sort of override where Atti mode gets overridden when there is critical battery?

Mike

 

[Vertigo]

On what basis would it make alterations to flight characteristics based on compass anyway... when you are supposed to be under manual control? I don't see the logic.

Even without GPS, on naze32/pixhawk/multiwii/ etc flight controllers, the compass, if available, is used to help maintain a constant heading. Gyro's primarily control yaw, but gyro's drift over time. On some of my quads, I can turn the compass on or off with a switch, and truthfully, it doesnt make a real difference. You only notice while hovering: without compass, the quad will slowly yaw in some direction. Flying hands on, its completely unneeded.

 

[Meta4]

I could see it still using the compass to try to display telemetry, but only for display purposes. I don't see any circumstance where it should try to control or alter the control of the AC with the compass or GPS while flying in Atti mode. On what basis would it make alterations to flight characteristics based on compass anyway... when you are supposed to be under manual control? I don't see the logic. Unless there is some sort of override where Atti mode gets overridden when there is critical battery?

Compass ... telemetry?? That's not what a compass does.
And despite wishing that a bad compass calibration won't make your Phantom difficult to control, that's what it does.
It's possible to have your compass calibration slightly out and flying big curves rather than straight or your compass can be further out and impossible to control.
Your Phantom needs it's compass to be properly calibrated to fly straight.
And if you don't screw up your compass calibration or hover over a large area of steel for 45 seconds ignoring the compass error on your screen, you won't have that sort of problem.

 

[MikeyOnline]

Compass ... telemetry?? That's not what a compass does.
And despite wishing that a bad compass calibration won't make your Phantom difficult to control, that's what it does.
It's possible to have your compass calibration slightly out and flying big curves rather than straight or your compass can be further out and impossible to control.
Your Phantom needs it's compass to be properly calibrated to fly straight.
And if you don't screw up your compass calibration or hover over a large area of steel for 45 seconds ignoring the compass error on your screen, you won't have that sort of problem.

Telemetry as far as being able to look at your device and see which direction the AC is pointed, yes. And still relying on the compass after it throws multiple compass errors in the log? Really? Seems odd. I've seen numerous threads here where people get a compass error, the AC starts to fly off, and then the pilot switches to Atti mode and brings it right back in. Seems like the compass was ignored after the switch to Atti (which it should be after a compass error).

Mike

 

[Meta4]

I've seen numerous threads here where people get a compass error, the AC starts to fly off, and then the pilot switches to Atti mode and brings it right back in. Seems like the compass was ignored after the switch to Atti (which it should be after a compass error).

I see lots of people suggest it but not many having any success - because atti mode makes the flight controller ignore GPS data but doesn't do anything to compass data.
If it did, the Phantom that started this thread would be at home in its box.
Being in atti made no difference for the OP.

 

[MikeyOnline]

I see lots of people suggest it but not many having any success - because atti mode makes the flight controller ignore GPS data but doesn't do anything to compass data.
If it did, the Phantom that started this thread would be at home in its box.
Being in atti made no difference for the OP.

Yikes! Well, thanks. Seems like the longer I'm here, the less I feel like I know. I can accept that, but it still seems beyond odd to me as I don't see the logic. I've seen more than one log where a "compass error" is thrown and the AC switches to Atti mode automatically. I just don't see the logic in switching to Atti mode without pilot input on a compass error, and then still using the errant compass information plus shutting off the only (most times) reliable positioning system that is left (GPS). Seems like something else DJI needs to address. We all know we can fly just about any bird without a compass. I have several cheaper quads that don't have a compass and gyros don't drift fast enough that you can't keep a heading while flying on gyros only. At least if you have LOS.

Mike

 

[flyNfrank]

As soon as he launched in the air the a/c was being told to RTH (not seen in the data) as well as the fact he uses the App Auto Takeoff feature. Basically a string of commands all at once corrupted the flight. App Auto Takeoff, Smart PowerGoHome, Smart PowerLanding, The result to start with was the Large Compass Error followed by Speed ErrorLarge, Yaw ErrorLarge. There were error messages in this flight that I have not seen before.

 

[flyNfrank]

Yikes! Well, thanks. Seems like the longer I'm here, the less I feel like I know. I can accept that, but it still seems beyond odd to me as I don't see the logic. I've seen more than one log where a "compass error" is thrown and the AC switches to Atti mode automatically. I just don't see the logic in switching to Atti mode without pilot input on a compass error, and then still using the errant compass information plus shutting off the only (most times) reliable positioning system that is left (GPS). Seems like something else DJI needs to address. We all know we can fly just about any bird without a compass. I have several cheaper quads that don't have a compass and gyros don't drift fast enough that you can't keep a heading while flying on gyros only. At least if you have LOS.

Mike

This flight would be a good example for you. It was left in P-mode on the R/C but the app had the A/C in one of two Atti mode types throughout the flight. I'm think it's common though due to not having enough satellite signals.

Edit: but the data shows this had plenty of gps signals. The flight was corrupt.

 

[MikeyOnline]

Yeah, sorry to get off topic a bit since this particular flight had a lot of issues. I'm more worried now about what I should do if I have a compass error during flight. I always make sure I have a good calibration and don't take off from metal surfaces, etc. But if I have a compass error during flight, I always thought I could switch it to Atti mode and fly it home like I fly my cheap quads that don't have a compass: by LOS and the gyros. Seems like my assumption was wrong. So what's the best course of action if you see a compass error and the AC starts flying in the wrong direction? If I can't switch to Atti and have true manual control... what to do? I just like to be prepared for such situations and I thought I knew what to do. Apparently not!

Mike

 

[Vertigo]

I can accept that,

I cant accept it. It would be literally unacceptable if the phantom relied on compass sensor data that
1) it knows is unreliable or unusable at that moment, and even reports this
2) is not needed at all in attitude mode, as can be seen by flying any quadcopter that doesnt have one.

I mean no disrespect to Meta4, but I think he's wrong here and it will take more convincing evidence than the OP's log to make me change my mind on this.

Its believable that a slightly miscalibrated compass affects flight, compass and gyro data may be different and cause some unwanted heading adjustments, but the moment the phantom itself says the data is not usuable and you should take manual control, I dont buy it. Why would it even bother to detect and report a compass problem ?

 

[flyNfrank]

Yeah, sorry to get off topic a bit since this particular flight had a lot of issues. I'm more worried now about what I should do if I have a compass error during flight. I always make sure I have a good calibration and don't take off from metal surfaces, etc. But if I have a compass error during flight, I always thought I could switch it to Atti mode and fly it home like I fly my cheap quads that don't have a compass: by LOS and the gyros. Seems like my assumption was wrong. So what's the best course of action if you see a compass error and the AC starts flying in the wrong direction? If I can't switch to Atti and have true manual control... what to do? I just like to be prepared for such situations and I thought I knew what to do. Apparently not!

Mike

I just posted in another thread how I have now flown over 350 flights with my P3 and I have never had any type of compass error before or during a flight. I properly calibrate the compass before each flight. I also have the data for each of those flights.

 

[sethtinsley]

As soon as he launched in the air the a/c was being told to RTH (not seen in the data) as well as the fact he uses the App Auto Takeoff feature. Basically a string of commands all at once corrupted the flight. App Auto Takeoff, Smart PowerGoHome, Smart PowerLanding, The result to start with was the Large Compass Error followed by Speed ErrorLarge, Yaw ErrorLarge. There were error messages in this flight that I have not seen before.

With that being said does that lead you to believe this could have been some sort of abnormal error, maybe due to a software malfunction or something of that nature? This aircraft would still be here today if I was able to gain control, but the toilet bowl flush and the complete unheard of tone of the motors leads me to think something more went wrong. Again, I'm clearly not as knowledgable as the majority of you regarding the series of events here, but leading of the the chaos I believe I did everything as expected.

I'm really trying to get down to what DJI will have to say. I filled out the form and synced the logs today and have messaged with a person with the problem ticket.

 

[Imabiggles]

It is, I was unaware of the danger of metal when flying. To clear it up, does this only pose an issue when taking off from an area with large amounts of metal near it or is it at anytime during flight avoid metal?

IMHO - the 'dangers' of metal are way over spoken about here. If you calibrate the compass in your car - yeah not a great idea. If you calibrate on concrete - its not an issue (do it all the time). Calibrate next to a 'gasp' metal building - not an issue. calibrate next to huge rare earth magnets or high voltage equipment running lots of amps? - not a great idea.

 

[Wibble]

Yeah, sorry to get off topic a bit since this particular flight had a lot of issues. I'm more worried now about what I should do if I have a compass error during flight. I always make sure I have a good calibration and don't take off from metal surfaces, etc. But if I have a compass error during flight, I always thought I could switch it to Atti mode and fly it home like I fly my cheap quads that don't have a compass: by LOS and the gyros. Seems like my assumption was wrong. So what's the best course of action if you see a compass error and the AC starts flying in the wrong direction? If I can't switch to Atti and have true manual control... what to do? I just like to be prepared for such situations and I thought I knew what to do. Apparently not!

Mike

That is also what I thought - manual over ride basically!
I would never take off with such a low battery either but I can't see it would cause that erratic a flight.