Phantom 3 Professional Fly Away

Blade, do you know what a "SPEED_ERROR" is?

The only two sources for speed are going to be GPS and VPS as far as I can guess. An accelerometer could give you an idea of speed but in the wind it's going to be noisy as all get out.

How would vps know speed? It has ultra sonic and a visioning camera which does not calculate speed..

I would have to ask r&d what the speed error is triggered by.
 
I'd bet a lot of money nothing hit it and the P3 did that itself.
  • The horizon was way off beforehand. Symptom of bad compass data.
  • The heading in the animation flips about 45 degrees at 0:39. Symptom of bad compass data.
  • The gimbal keeps twitching, losing yaw position. Symptom of bad compass data.
  • The gimbal is then flipped to the other side. Symptom of bad compass data.
  • TBE ensues. Symptom of bad compass data.
  • COMPASS errors reported in data. Symptom of bad compass data.
 
  • Like
Reactions: StevenQX
Something else interesting is that in the second set of jolts at 1:37, you can see the frame in the camera. The red compass arrow on the map jolts to the left, twice, but you can see from the video that the Phantom frame actually jolts to the right. There are a variety of reasons this could be. Compass failure is definitely one of them. The quad's reaction to a mechanical failure might cause this too, though... it's really hard to tell. Have to watch it another few dozen times.
 
Here is the Video Full Screen
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
Something else interesting is that in the second set of jolts at 1:37, you can see the frame in the camera. The red compass arrow on the map jolts to the left, twice, but you can see from the video that the Phantom frame actually jolts to the right.

It does it at 0:39 as well.

Tyler, where and how did you calibrate compass? In a field? Car park? Sidewalk?
 
I'd bet a lot of money nothing hit it and the P3 did that itself.
  • The horizon was way off beforehand. Symptom of bad compass data.
  • The heading in the animation flips about 45 degrees at 0:39. Symptom of bad compass data.
  • The gimbal keeps twitching, losing yaw position. Symptom of bad compass data.
  • The gimbal is then flipped to the other side. Symptom of bad compass data.
  • TBE ensues. Symptom of bad compass data.
  • COMPASS errors reported in data. Symptom of bad compass data.

Many of these symptoms could also be caused by mechanical failure too though (heading flipping, gimbal twitching losing position because the frame is rotating too quickly, TBE caused by a weak motor / inability to maintain roll/pitch/yaw/throttle, etc.). And if the compass was awry, it could be due to mechanical damage to the compass caused by an impact either during this flight (e.g. at 0:39) or a previous flight. For example, the impact might have damaged the lead to the compass or dislodged one of its connectors. Or the impact might have broken a leg, causing the compass to dangle in mid-air. We just don't know.
 
0-39 seconds - flying normally in GPS mode
39 seconds it tilts aggressively and switches to P-Opti and P-ATTI mode back and forth
45 seconds - User switches to ATTI mode, user does not push the sticks and lets the bird float with ATTI mode (which is why bird is facing front but flying back due to being aimless.
52 seconds - RTH initiated but bird is STILL in ATTI mode, still flying backwards
1:01 seconds - Bird switches to P-OPTI mode, still not in RTH mode.
1:07 - RTH mode is enabled on bird. it switches direction to face the user and starts flying home
1:15 bird starts tilting to the left while facing forward as if there is a rotor out
1:23 Bird switches to P-OPTI mode and P-ATTI mode repeatedly until it remains in P-OPTI mode and starts flying backwards again, no user input on sticks
1:34 Bird rapidly switches between P-opti, P-ATTI and PGPS mode 5-6 times while flying backwards but no user input on sticks
1:36 Bird switches to ATTI mode and starts tilting aggressively right as user starts using the sticks
1:42 User stops using sticks and bird switches to P-ATTI mode and P OPTI mode and P GPS mode repeatedly
1:50 Bird sends abnormal compass signal and FPV and Compass and GPS signal is weak on screen
1:54 signal lost
 
0-39 seconds - flying normally in GPS mode
39 seconds it tilts aggressively and switches to P-Opti and P-ATTI mode back and forth
45 seconds - User switches to ATTI mode, user does not push the sticks and lets the bird float with ATTI mode (which is why bird is facing front but flying back due to being aimless.
52 seconds - RTH initiated but bird is STILL in ATTI mode, still flying backwards
1:01 seconds - Bird switches to P-OPTI mode, still not in RTH mode.
1:07 - RTH mode is enabled on bird. it switches direction to face the user and starts flying home
1:15 bird starts tilting to the left while facing forward as if there is a rotor out
1:23 Bird switches to P-OPTI mode and P-ATTI mode repeatedly until it remains in P-OPTI mode and starts flying backwards again, no user input on sticks
1:34 Bird rapidly switches between P-opti, P-ATTI and PGPS mode 5-6 times while flying backwards but no user input on sticks
1:36 Bird switches to ATTI mode and starts tilting aggressively right as user starts using the sticks
1:42 User stops using sticks and bird switches to P-ATTI mode and P OPTI mode and P GPS mode repeatedly
1:50 Bird sends abnormal compass signal and FPV and Compass and GPS signal is weak on screen
1:54 signal lost
Thanks!! maybe when DJI pick up the phone I can send them this.
maybe they will RMA !?
 
You can calculate speed with optical flow. I am not saying it is. But it is possible.

Not at this altitude and dji does not use the system for speed measurements.
 
0-39 seconds - flying normally in GPS mode
39 seconds it tilts aggressively and switches to P-Opti and P-ATTI mode back and forth
45 seconds - User switches to ATTI mode, user does not push the sticks and lets the bird float with ATTI mode (which is why bird is facing front but flying back due to being aimless.
52 seconds - RTH initiated but bird is STILL in ATTI mode, still flying backwards
1:01 seconds - Bird switches to P-OPTI mode, still not in RTH mode.
1:07 - RTH mode is enabled on bird. it switches direction to face the user and starts flying home
1:15 bird starts tilting to the left while facing forward as if there is a rotor out
1:23 Bird switches to P-OPTI mode and P-ATTI mode repeatedly until it remains in P-OPTI mode and starts flying backwards again, no user input on sticks
1:34 Bird rapidly switches between P-opti, P-ATTI and PGPS mode 5-6 times while flying backwards but no user input on sticks
1:36 Bird switches to ATTI mode and starts tilting aggressively right as user starts using the sticks
1:42 User stops using sticks and bird switches to P-ATTI mode and P OPTI mode and P GPS mode repeatedly
1:50 Bird sends abnormal compass signal and FPV and Compass and GPS signal is weak on screen
1:54 signal lost
It is _NOT_ flying normally 0-39 seconds .. look at the horizon. This was not a collision this was a failure, whether caused by a previous incident or not we can't say with the evidence we have. But I would stake my paycheck on this not being a bird hit.
 
Mechanical failures don't keep you in the air that long. When a motor or ESC goes bad, the only possible direction is down. And if one motor should slow for whatever reason, it will zip off in one direction. Possibly something damaged the compass but that would be a huge impact. It's pretty solidly on there and the wires are tucked in nicely within the arch of the leg. And it's rare that a bird will make a direct hit without some warning passes first. They almost always swoop around and warn you first. Bird strike also doesn't explain the bad horizon beforehand.

People, if your horizon is way off, land! You have a compass problem.

calibrated in the park prior to this flight as well, park, no keys, phone metal etc.

On grass? No concrete or storm drains within 20ft?

Not at this altitude and dji does not use the system for speed measurements.

That would be my assumption too. But if it was on and trying to determine speed from it, it would conflict with GPS and result in a "speed error" of some sort. Just a wild guess.
 
when it responds yes

calibrated in the park prior to this flight as well, park, no keys, phone metal etc.

But Tyler you didn't try controling the craft it atti? You gave no stick inputs.

To stop the craft an atti mode you would have to cancel out any movement.

Something caused bad compass data, that's about the only hard facts we have..
 
It is _NOT_ flying normally 0-39 seconds .. look at the horizon. This was not a collision this was a failure, whether caused by a previous incident or not we can't say with the evidence we have. But I would stake my paycheck on this not being a bird hit.
The horizon is often off on the P3. Mine is perfectly fine but the horizon is usually 5-10 degrees off during normal flight. Don't read too much into that.
 
one of my helis has a flybarless controller with trainer mode. It keeps the bird upright, but it won't account for drift, which requires pilot input. I'm thinking the later part of the video was like this. But wouldn't it keep altitude even if he wasn't controlling it? It started pitching on it's own.
 
one of my helis has a flybarless controller with trainer mode. It keeps the bird upright, but it won't account for drift, which requires pilot input. I'm thinking the later part of the video was like this. But wouldn't it keep altitude even if he wasn't controlling it? It started pitching on it's own.

Correct. That's what ATTI does. The bird should stay level but will drift with the wind. It will actually pitch and roll less without stick input as it isn't fighting the wind.
The horizon is often off on the P3. Mine is perfectly fine but the horizon is usually 5-10 degrees off during normal flight. Don't read too much into that.

Disagree. Bad horizon is bad IMU and/or bad compass data.
 
Blade strike, if you have contact with DJI R&D, this would be a great opportunity to suggest to them that they start sending raw IMU data (gyro, accelerometer), raw compass data, and motor outputs, in the telemetry stream. If we had that data available in the .TXT files, we would have a much better idea of what happened here. The .DAT files have that and they're a great addition, but if you crash in water, they are lost...

I guess we have to hope that we see more crashes like this, and eventually we get one where we can get our hands on the .DAT file.
 
Correct. That's what ATTI does. The bird should stay level but will drift with the wind. It will actually pitch and roll less without stick input as it isn't fighting the wind.


Disagree. Bad horizon is bad IMU and/or bad compass data.

Incorrect. Slightly bad horizon is normal.
 

Members online

Forum statistics

Threads
143,066
Messages
1,467,358
Members
104,935
Latest member
Pauos31