HELP! P3 Professional went out of control and crashed.

[syotr]

I got a response from DJI on the question about compass and attitude mode. They did not explain the in flight recalibration I asked about and for some reason, they answered me in German (probably thinking Dutch and Deutsch is the same?), so let me quote it and translate. German is only my 4th language, if anyone spots any mistakes, feel free to correct me. Original is below

In pure attitude mode, the quadcopter doesnt need compass data. It doesnt stabilize its position and doesnt require absolute orientation.

If the compass is signal is disturbed, then the quad automatically goes in to attitude mode. This is reported by many pilots as "lost control". In many flight records we see that novice pilots panic and give full stick deflection in one direction, then the other. The problem with that: the quad follows the commands 1:1 and the impression is made, that its completely out of control.

In this case, the pilot must compensate the drift. It makes no difference if the quadcopter was switched in to attitude mode automatically (due to compass error) or manually via the remote control. It requires some feeling to fly in this way; you can train this with no problem if you have enough free space.

I did a test with a Phantom P1 in ATTI mode. to see how well it would hold its heading. This P1 has no compass or GPS installed. It really puts up a fight when you try to rotate it by hand. When you yaw it manually, it returns to the previous position. It will drift in the wind but holds the heading very well in spite of external forces.

 

[Vertigo]

Your video is private, but I think most of us know gyro's work just fine (at least in the short term, over longer periods, they will drift). The majority of my quads dont have a compass, and some that do hold their heading better without using the compass, as the compass is placement is not well thought out and motors influence it too much. So it will randomly yaw a bit left/right in forward flight if I enable it. Without GPS (or unless you want to use flight modes like 'head free') , there really is no point in having or using a compass.

 

[Meta4]

While on this topic, would a large amount of industrial outdoor refrigerators possibly create a error such as mine?

Refrigerators could have a magnetic influence ... if you get close enough.
But hovering above a big reinforced concrete slab is going to have a big enough magnetic effect to swamp any distant refrigeratord.
That's why your Phantom was warning you for 46 seconds before it took off.

 

[msinger]

The .DAT is from the aircraft?

Yes. If needed, you can find instructions for retrieving the DAT log here.

 

[syotr]

Your video is private, but I think most of us know gyro's work just fine (at least in the short term, over longer periods, they will drift). The majority of my quads dont have a compass, and some that do hold their heading better without using the compass, as the compass is placement is not well thought out and motors influence it too much. So it will randomly yaw a bit left/right in forward flight if I enable it. Without GPS (or unless you want to use flight modes like 'head free') , there really is no point in having or using a compass.

Try the video now. I had it set to public. Don't know why it wasn't playing. Some folks here have apparently not flown without a compass or GPS and have made statements about how they think a quad would behave in this situation. I like to actually test things rather than speculate when possible. Sometimes I am surprised and have to rethink my own beliefs. I have flown many times in ATTI mode but had never actually reached out and tried to rotate the bird by hand. I was amazed that the gyros held it so well.

 

[MikeyOnline]

I question the accuracy of compass problems taking off from concrete. I've taken off dozens of times from both my driveway and my neighbors without a single issue. I always assumed it was because the concrete slabs in our driveways didn't have rebar. Until I drove by a neighbor's house and saw a backhoe removing a slab and when they dug it up, it was full of rebar! Never even knew it but it had no effect on my compass. In addition, once you get airborne, any errors should go away after you move away from the source of metal: I could see a temporary effect but it won't "recalibrate" the compass so it's wrong the whole flight. So is there any data (video) showing influence of metal on the compass and proving it?

Mike

 

[cdronefly]

I wasn't planning on doing any distance flying, I was simply planning on hovering to burn the battery down.

Don't do this. Ever.
Don't take off with anything less than a full battery.

FWIW, I have had my P3 for over a year and take absolutely no care of my 3 batteries and they work fine. I did break them in, but that is all.
I never take off with less than a full battery - although I may land in-between and then take off with over 50 or 60%....

Warnings about compass, GPS, etc. should be heeded...very strongly.
There is no warranty against crashes and losses - but that is not my department.

 

[BudWalker]

I question the accuracy of compass problems taking off from concrete. I've taken off dozens of times from both my driveway and my neighbors without a single issue. I always assumed it was because the concrete slabs in our driveways didn't have rebar. Until I drove by a neighbor's house and saw a backhoe removing a slab and when they dug it up, it was full of rebar! Never even knew it but it had no effect on my compass. In addition, once you get airborne, any errors should go away after you move away from the source of metal: I could see a temporary effect but it won't "recalibrate" the compass so it's wrong the whole flight. So is there any data (video) showing influence of metal on the compass and proving it?

Mike

The P3 uses the IMU for heading info. The compass info is only used indirectly for this purpose. Soon after powerOn the compass heading is used to initialize the IMU's heading value. After that the IMU updates it's heading value mostly from the gyros, accelerometers, and to a much lesser extent the compass.

If launch occurs in a geomagnetically compromised environment the compass can be pointing the wrong way and the IMU will be initialized with the wrong value. After launch the compass changes to read correctly but the IMU won't because there haven't been any turns. It can be 2 minutes before the IMU converges on the, now, correct compass value.

The sudden change in compass value causes a COMPASS_ERROR_LARGE error and the switch to ATTI is made. But, this error will "go away" and be replaced with a YAW_ERROR_LARGE and the AC can end up in GPS_ATTI with the IMU heading having the wrong value. @Krum04 had flight where this happened. It was made worse because RTH occurred since the connection to the RC dropped.

It seems that even if the AC is in GPS_ATTI an incorrect IMU heading can make the AC hard to control. If this happens DON'T engage RTH.

Sorry I can't post more detailed info. I'm on my stupid iPad for another couple of weeks. I'll post some more info then.

 

[sethtinsley]

Refrigerators could have a magnetic influence ... if you get close enough.
But hovering above a big reinforced concrete slab is going to have a big enough magnetic effect to swamp any distant refrigeratord.
That's why your Phantom was warning you for 46 seconds before it took off.

Fortunately I wasn't hovering over a "big reinforced concrete slab" so that is not the case.

 

[sethtinsley]

Yesterday I received a response from DJI... Here it is:

"Dear Seth Tinsley
After having our data analysis team take a look at the flight log from the flight in question, we were unable to determine the cause of this unfortunate event.

There does not seem to be any user error, but without the craft, to analyze the flight records there is no way to be 100% sure.

With that being said we have decided to cover your unit by warranty, and offer the craft only at no cost to you.
We are in the process of finalizing a new alternative for our customers in who find themselves in situations similar to yours."

Hats off to DJI for the analysis and overall way the situation was handled, I can say this is not what I expected and can't be happier to read it. Thank you to all of you for the advice, support and knowledge.

 

[MikeyOnline]

The P3 uses the IMU for heading info. The compass info is only used indirectly for this purpose. Soon after powerOn the compass heading is used to initialize the IMU's heading value. After that the IMU updates it's heading value mostly from the gyros, accelerometers, and to a much lesser extent the compass.

If launch occurs in a geomagnetically compromised environment the compass can be pointing the wrong way and the IMU will be initialized with the wrong value. After launch the compass changes to read correctly but the IMU won't because there haven't been any turns. It can be 2 minutes before the IMU converges on the, now, correct compass value.

The sudden change in compass value causes a COMPASS_ERROR_LARGE error and the switch to ATTI is made. But, this error will "go away" and be replaced with a YAW_ERROR_LARGE and the AC can end up in GPS_ATTI with the IMU heading having the wrong value. @Krum04 had flight where this happened. It was made worse because RTH occurred since the connection to the RC dropped.

It seems that even if the AC is in GPS_ATTI an incorrect IMU heading can make the AC hard to control. If this happens DON'T engage RTH.

Sorry I can't post more detailed info. I'm on my stupid iPad for another couple of weeks. I'll post some more info then.

Great info, thanks! Guess I've just been lucky, but I still wonder how much effect embedded non-magnetized rebar in concrete really has on the compass. Has anyone ever tested that?

Mike

 

[Tenly]

Yesterday I received a response from DJI... Here it is:

"Dear Seth Tinsley
After having our data analysis team take a look at the flight log from the flight in question, we were unable to determine the cause of this unfortunate event.

There does not seem to be any user error, but without the craft, to analyze the flight records there is no way to be 100% sure.

With that being said we have decided to cover your unit by warranty, and offer the craft only at no cost to you.
We are in the process of finalizing a new alternative for our customers in who find themselves in situations similar to yours."

Hats off to DJI for the analysis and overall way the situation was handled, I can say this is not what I expected and can't be happier to read it. Thank you to all of you for the advice, support and knowledge.

Hmmm....who was it on the forum here who suggested this should be the outcome....trying hard to remember....


Whilst so many were looking for any possible reason to belittle this pilot and claim it was his error, a very few of us actually listened to what happened and came to the same conclusion as DJI. No proof of pilot error whatsoever!

This is awesome news! Congratulations! It is no less than what you deserve and I'm glad to see that DJI has acknowledged that!

This makes the 3rd aircraft in the past month (that I'm aware of) that has been replaced at a 100% discount even though the crashed bird could not be returned - so those who claim that this very rarely happens may need to update that tale!

I also hope that these 3 stories will serve as an example to those judgemental and superior types that love to chime in on these threads and assign blame to the pilot with little or no evidence to back them up. DJI makes a great UAV - but it's not infallible A little more "benefit of the doubt" to our fellow pilots would make these forums a much nicer place to hang out!

This has been a great "I told ya so" for me! I'm so glad that this worked out for you.

From the time I received that email - it took 3 days and a phone call before I actually received the coupon code (which has to be used within 30 days and can only be used towards the same unit that was lost). After you submit your order, expect 2-day free shipping from China and arrival in a plain brown box containing the aircraft, a battery and 1 set of propellers!

Congratulations again!!!

 

[syotr]

I question the accuracy of compass problems taking off from concrete. I've taken off dozens of times from both my driveway and my neighbors without a single issue. I always assumed it was because the concrete slabs in our driveways didn't have rebar. Until I drove by a neighbor's house and saw a backhoe removing a slab and when they dug it up, it was full of rebar! Never even knew it but it had no effect on my compass. In addition, once you get airborne, any errors should go away after you move away from the source of metal: I could see a temporary effect but it won't "recalibrate" the compass so it's wrong the whole flight. So is there any data (video) showing influence of metal on the compass and proving it?

Mike

Here's an idea. Just get a simple little toy compass and hold it near the concrete. If the compass changes orientation strongly, then maybe there is something there to be concerned about.A magnetometer app on your phone may work as well.
I have never found an issue launching from the concrete road or sidewalk. Not saying it can't happen, but I have not seen it.

 

[skipilot1]

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?

I have to agree. If Nothing else, provided you have stick control, you should be able to fly FPV. Even without the compass, left is left and right is right. Even if the controls are reversed there would have to be a way to point it toward your home point.

 

[skipilot1]

Hmmm....who was it on the forum here who suggested this should be the outcome....trying hard to remember....


Whilst so many were looking for any possible reason to belittle this pilot and claim it was his error, a very few of us actually listened to what happened and came to the same conclusion as DJI. No proof of pilot error whatsoever!

This is awesome news! Congratulations! It is no less than what you deserve and I'm glad to see that DJI has acknowledged that!

This makes the 3rd aircraft in the past month (that I'm aware of) that has been replaced at a 100% discount even though the crashed bird could not be returned - so those who claim that this very rarely happens may need to update that tale!

I also hope that these 3 stories will serve as an example to those judgemental and superior types that love to chime in on these threads and assign blame to the pilot with little or no evidence to back them up. DJI makes a great UAV - but it's not infallible A little more "benefit of the doubt" to our fellow pilots would make these forums a much nicer place to hang out!

This has been a great "I told ya so" for me! I'm so glad that this worked out for you.

From the time I received that email - it took 3 days and a phone call before I actually received the coupon code (which has to be used within 30 days and can only be used towards the same unit that was lost). After you submit your order, expect 2-day free shipping from China and arrival in a plain brown box containing the aircraft, a battery and 1 set of propellers!

Congratulations again!!!

View attachment 58152

Congratulations. Makes me feel pretty about the company I decided to go with.

 

[Susruth]

Yesterday I received a response from DJI... Here it is:

"Dear Seth Tinsley
After having our data analysis team take a look at the flight log from the flight in question, we were unable to determine the cause of this unfortunate event.

There does not seem to be any user error, but without the craft, to analyze the flight records there is no way to be 100% sure.

With that being said we have decided to cover your unit by warranty, and offer the craft only at no cost to you.
We are in the process of finalizing a new alternative for our customers in who find themselves in situations similar to yours."

Hats off to DJI for the analysis and overall way the situation was handled, I can say this is not what I expected and can't be happier to read it. Thank you to all of you for the advice, support and knowledge.

Looks like DJI has a standard copy paste response for anyone they decide to refund ir is it? I saw this exact message elsewhere

 

[Susruth]

Hmmm....who was it on the forum here who suggested this should be the outcome....trying hard to remember....


Whilst so many were looking for any possible reason to belittle this pilot and claim it was his error, a very few of us actually listened to what happened and came to the same conclusion as DJI. No proof of pilot error whatsoever!

This is awesome news! Congratulations! It is no less than what you deserve and I'm glad to see that DJI has acknowledged that!

This makes the 3rd aircraft in the past month (that I'm aware of) that has been replaced at a 100% discount even though the crashed bird could not be returned - so those who claim that this very rarely happens may need to update that tale!

I also hope that these 3 stories will serve as an example to those judgemental and superior types that love to chime in on these threads and assign blame to the pilot with little or no evidence to back them up. DJI makes a great UAV - but it's not infallible A little more "benefit of the doubt" to our fellow pilots would make these forums a much nicer place to hang out!

This has been a great "I told ya so" for me! I'm so glad that this worked out for you.

From the time I received that email - it took 3 days and a phone call before I actually received the coupon code (which has to be used within 30 days and can only be used towards the same unit that was lost). After you submit your order, expect 2-day free shipping from China and arrival in a plain brown box containing the aircraft, a battery and 1 set of propellers!

Congratulations again!!!

View attachment 58152

Yeah ...saw the same msg before

 

[Susruth]

Looks like DJI has a standard copy paste response for anyone they decide to refund ir is it? I saw this exact message elsewhere

Drone just fell out of the sky

 

[MikeyOnline]

Looks like DJI has a standard copy paste response for anyone they decide to refund ir is it? I saw this exact message elsewhere

Makes sense. I do that in my own business for certain types of messages. What happens is that you reply to a customer and tell them that you are going to cover it under warranty and you get a couple replies to that asking "Did you find any pilot error or things I should avoid?" So you modify your response to include the part where they say they were not able to find any pilot error. Then you notice a couple people asking "Did you analyze the logs and did you find the cause?" Then you have to reply saying you did that but the root cause was not found. And so on.

What you end up with is a boilerplate message that, when the conditions warrant (like analysis complete, no pilot error found, root cause unknown, etc.), you send that "honed" message that you have determined from past experience clearly and concisely explains what was done, what you will be getting, and even a note about future plans to help people in the same situation. So I for one am not bothered by it being a "standard" response.

Mike

 

[Tenly]

Looks like DJI has a standard copy paste response for anyone they decide to refund ir is it? I saw this exact message elsewhere

Yep. Exact same message they sent me!