I only stated this because it was the first error thrown and the other week you said you didn't know yourself why these errors were being thrown from the sonar even at high altitudes, have DJI come up with a reason for this now.
How about the video of the flight, you dont need the craft for that just go into director in the pilot app and pick moments and all your flights will be there to view.
Sorry about your P3 by the way, just watched your flight record. Not good, i don't like this function the P3 has were if the compass gets any strange data it knocks it self out of GPS and into Atti. My old phantom never went out of GPS and that was without glonass lock as well. Many people are reporting the P3 going into Atti mode with compass error now, and nearly all crashes that i have known have gone into Atti around the time of the crash. This compass error and going into Atti is far to common for my liking now.
I only stated this because it was the first error thrown and the other week you said you didn't know yourself why these errors were being thrown from the sonar even at high altitudes, have DJI come up with a reason for this now.
Thanks for uploading the .TXT file! Here's a selection of the columns from the "Comprehensive CSV" exported using http://www.DJILogs.com. The flight controller is definitely complaining about the compass error being large (COMPASS_ERROR_LARGE) and that's one of the reasons it switches from GPS mode to Atti mode, but there are also a lot of SPEED_ERROR_LARGE messages, and I'm not convinced the compass was actually wrong.
Also - at around 6/2/2015 2:01:05 AM (row 405) there is a sudden unexpected roll and yaw shock. It goes from 3.5 degrees roll to 20 degrees roll and the yaw goes from 340 degrees to 285 degrees in a split second (100ms). It also pitches back, from -21 degrees to -9 degrees. At the same time, the craft veers to the right as shown by the course over the water, even though the right stick is directly forward. My take on this is that you either hit something, most likely a bird, or one of the props/motors/ESCs started to go bad. After that point, the craft behaves very oddly, before finally beginning a rapid uncontrolled descent at row 1061.
My bet is that a bird strike damaged one of the props.
Also - at around 6/2/2015 2:01:05 AM (row 405) there is a sudden unexpected roll and yaw shock. It goes from 3.5 degrees roll to 20 degrees roll and the yaw goes from 340 degrees to 285 degrees in a split second (100ms). It also pitches back, from -21 degrees to -9 degrees. At the same time, the craft veers to the right as shown by the course over the water, even though the right stick is directly forward. My take on this is that you either hit something, most likely a bird, or one of the props/motors/ESCs started to go bad. After that point, the craft behaves very oddly, before finally beginning a rapid uncontrolled descent at row 1061.
My bet is that a bird strike damaged one of the props.
In the video of the flight record, from :43 to 1:26, he never uses his sticks while the craft is all over the place including a SHARP right turn during a RTH of all things! I guess you could say a bird hit at that hard right turn, but what would explain :43 - 1:26? Everything about this video defies logic. I've seen bird strikes and they're sudden and sharp... and the quad either recovers within a second or two or is flipped and comes STRAIGHT down. This does absolutely nothing like that.
In the video of the flight record, from :43 to 1:26, he never uses his sticks while the craft is all over the place including a SHARP right turn during a RTH of all things! I guess you could say a bird hit at that hard right turn, but what would explain :43 - 1:26? Everything about this video defies logic. I've seen bird strikes and they're sudden and sharp... and the quad either recovers within a second or two or is flipped and comes STRAIGHT down. This does absolutely nothing like that.
I agree the instability started there but we have no idea why. Birds don't cause compass and speed errors unless they're made of iron! And those errors started at the same time.
P.S. What is a "speed" error? Never heard of this before.
I don't think a bird strike is going to end in a gradual crash. Physical damage to props will end quickly. And again birds don't cause compass/speed errors. Something else was going on.
I know I know... I'm guessing that if it was a bird strike, the failure was gradual. Perhaps part of the propellor came off and the quad was able to stabilize for a while, but the motor/prop which got hit remained weak and gradually got worse - maybe more bits of propellor came off over the next minute, or the motor overheated because the shaft was out of alignment, or something.
The thing is that IMU or compass failure wouldn't really explain the sharp jolt. It is hard to know though, without the raw IMU data from the .DAT file.
OpDon't think you really read my response. Oh well.
Receiving bad data from rapid temp change. This would makes sense since he took off very fast and then over water. This would be a prime example of rapid temp change.
Is it possible that one of the props was fractured from the prior crash and failed causing the sudden jolt? There was another post on here where somebody didn't notice that they had a failed prop.I don't know what the "speed" error is, but I'm guessing it happens when the quad's internal estimate of its speed differs substantially from what the GPS is reporting. That could happen easily if a prop or motor failed. The compass errors could be caused by that, too. It's even possible an impact damaged the leg with the compass on it.
I find it hard to imagine that IMU or compass failure would cause such a sharp jolt (30 degrees in 100ms), and in any case, if the IMU or compass did fail, then the roll/pitch/yaw values wouldn't agree with the course-over-ground values (lat/lon, HEADING.compass etc.). What I see is that they do agree. The craft flew to the right, and it knew it had rolled to the right. If the IMU or compass had failed, it would fly off but that wouldn't be reflected in the roll/pitch/yaw values, which would show the craft being stable in the air.
Could also have been a cascading failure. Perhaps an initial impact confused the flight controller, which then got gradually more and more confused even though everything was mechanically sound. Or perhaps the impact caused an initial mechanical failure which got progressively worse over the following minute or so, until eventually the props were unable to provide enough lift to keep the quad in the air.