Anatomy of a DJI Flyaway
- Thread starter ianwood
- Start date
Ok, back to cross fingers and knock wood to avoid flyaways,.
I hate to say it, but yeah. Here's to hoping it was a bad laminate in the NAZA PCB that heat cycled enough to separate. In sending my P2 in, I have asked that I be kept informed of any discoveries made from it. We'll see.
i m at my 4t flyaway first one caused gimbal and camera destroy
dynamics was always the same... it start with strange phantom movements like it's shaking
and then fly away oblique at full speed in one direction i have recorded 2 of them:
and this one happened yesterday
i payed it 1.000 euros... and work bad than my friend's 100 euros one...
can't belive.. .dont know how to do... i have fear to destroy it again...
i hope to find a solution... dji customer care is non existing
After every flyaway i shut down remote and phantom stabilized and start return home
so i thing is a software or hardware and software problem can't be only hardware
because when i start RTH procedure it always worked every time 4/4
(i changed compass after first crash)
maybe a naza hard reset can help??
i feel really alone
dynamics was always the same... it start with strange phantom movements like it's shaking
and then fly away oblique at full speed in one direction i have recorded 2 of them:
and this one happened yesterday
i payed it 1.000 euros... and work bad than my friend's 100 euros one...
can't belive.. .dont know how to do... i have fear to destroy it again...
i hope to find a solution... dji customer care is non existing
After every flyaway i shut down remote and phantom stabilized and start return home
so i thing is a software or hardware and software problem can't be only hardware
because when i start RTH procedure it always worked every time 4/4
(i changed compass after first crash)
maybe a naza hard reset can help??
i feel really alone
Outstanding work, Ian.
I have one question: did you check the solar activity level at the time? Has it been ruled out as a possible factor??
[
I've included a more detailed analysis of the data which is in the PDF link here: http://www.ianwood.com/docs/anatomy-of-a-dji-flyaway-v1.pdf[/QUOTE]
I have one question: did you check the solar activity level at the time? Has it been ruled out as a possible factor??
[
I've included a more detailed analysis of the data which is in the PDF link here: http://www.ianwood.com/docs/anatomy-of-a-dji-flyaway-v1.pdf[/QUOTE]
OMG... so many comments! Because we all hope it doesn't happen to us! I'll add a hardware tech's angle on this. It IS possible that you have a hardware problem. Hardware problems are ALL mechanical problems - be they a cold solder joint or an micron level defect on the chip die.
Being so, they are usually temperature sensitive...as things expand and contract with heat/cold.
The CPU in the NAZA has the most connections to the board. If one of the multitude of little IC legs has a tiny fracture, it will connect/disconnect with temperature... thus the delay in occurring until temps heat up on the board. What we do is to fire up the board..and then hit the IC with freeze spray. Keep the data logger on and sequentially freeze each IC on the board (black things with a zillion legs) See if the NAZA misfires. If so, it's not easily repairable without an SMD workstation. If it was firmware...many people would have the same problem using that exact firmware version. Freeze spray is order-able on-line.
Being so, they are usually temperature sensitive...as things expand and contract with heat/cold.
The CPU in the NAZA has the most connections to the board. If one of the multitude of little IC legs has a tiny fracture, it will connect/disconnect with temperature... thus the delay in occurring until temps heat up on the board. What we do is to fire up the board..and then hit the IC with freeze spray. Keep the data logger on and sequentially freeze each IC on the board (black things with a zillion legs) See if the NAZA misfires. If so, it's not easily repairable without an SMD workstation. If it was firmware...many people would have the same problem using that exact firmware version. Freeze spray is order-able on-line.
Neither location or solar activity could cause what happened.
The latest wrinkle is that it happened on more than one P2 which eliminates the NAZA as the most likely root cause. However, the NAZA was the principle actor in the loss of control. Something caused it to slow down to a near freeze. It was unable to sample IMU data fast enough to maintain attitudinal stability.
Here's a screenshot from the camera during a loss of attitudinal stability. The Phantom is pretty much tilted 90 degrees forward.
The latest wrinkle is that it happened on more than one P2 which eliminates the NAZA as the most likely root cause. However, the NAZA was the principle actor in the loss of control. Something caused it to slow down to a near freeze. It was unable to sample IMU data fast enough to maintain attitudinal stability.
Here's a screenshot from the camera during a loss of attitudinal stability. The Phantom is pretty much tilted 90 degrees forward.
Why is it "not possible" that outside interference could have caused this. If you were in approximately the same location especially. There are known devices that can wreak havoc on ANY electronics.
Now if these two incidents occurred more than a mile from one another I would say probably something else.
I read the entire thread...sorry if I missed the explanation.
That photo above just makes your stomach turn. That angle at that altitude. ugh.
Now if these two incidents occurred more than a mile from one another I would say probably something else.
I read the entire thread...sorry if I missed the explanation.
That photo above just makes your stomach turn. That angle at that altitude. ugh.
Last edited:
Well, let me put it a different way: how could interference cause that kind of reaction? I can think of three ways: electro magnetic disruption, hacking/spoofing of the Lightbridge signal, and finally a Lightbridge defect.
Only an EM pulse or focused high-power EM "ray" could disrupt the normal operation of the processor and make it slow down to a nearly frozen state. So unless I am being stalked by a cyber-psycho genius anarchist, the odds of being hit by a "death ray" in three different locations on three different days: 0.000000001% (give or take a few zeros).
The other possibility would be someone hacking and spoofing the Lightbirdge uplink. Not only would they need to demux, sync and remux the time division perfectly, they would then have to reverse engineer the proprietary Lightbridge protocol and then the unique binding hash. Then they would need to know how to confuse the NAZA via the Lightbridge enough to get it to go bezerk. Odds of that happening: 0.00000001% but only if by the same cyber-psycho genius anarchist.
The last possibility is a systemic flaw in the Lightbridge system that allows a spurious signal over the same frequency to disrupt it's normal operation and then somehow confuse the NAZA so much as to make it unrecoverable. If that were the case, I wouldn't be the only one reporting this issue. I haven't seen or heard of any other similar reports to this.
What is most perplexing is the dozens of flights on the same hardware in the same conditions that had no issues.
Only an EM pulse or focused high-power EM "ray" could disrupt the normal operation of the processor and make it slow down to a nearly frozen state. So unless I am being stalked by a cyber-psycho genius anarchist, the odds of being hit by a "death ray" in three different locations on three different days: 0.000000001% (give or take a few zeros).
The other possibility would be someone hacking and spoofing the Lightbirdge uplink. Not only would they need to demux, sync and remux the time division perfectly, they would then have to reverse engineer the proprietary Lightbridge protocol and then the unique binding hash. Then they would need to know how to confuse the NAZA via the Lightbridge enough to get it to go bezerk. Odds of that happening: 0.00000001% but only if by the same cyber-psycho genius anarchist.
The last possibility is a systemic flaw in the Lightbridge system that allows a spurious signal over the same frequency to disrupt it's normal operation and then somehow confuse the NAZA so much as to make it unrecoverable. If that were the case, I wouldn't be the only one reporting this issue. I haven't seen or heard of any other similar reports to this.
What is most perplexing is the dozens of flights on the same hardware in the same conditions that had no issues.
I think that interference could most certainly cause flyaways, especially if they all occurred in the same general area. I'm an engineer, and the likelihood of two different birds having the same hardware issue is very low, unless there is a systemic defect. I doubt this is the case, or we would see many more similar issues.
You didn't read the thread. They didn't happen in the same general area. And not on the same day. And there were several dozen normal flights in between. The Lightbridge air unit was common between all three incidents and is most likely to be the cause.
As for interference having anything to do with this, no, there is nothing in the data suggesting any interference be it magnetic, GPS L band, ISM band or otherwise.
As far as I am concerned, only magnetic interference can cause a Phantom to flyaway and that is strictly user error. All other claims of interference causing flyaways is simply blind guessing.
As it looks, we call that a computer glitch. And yes the onboard computer of the Phantom can glitch and even has been know to re-boot. When it does that it's really bad news as all onboard data is corrupted and some find it as a fly a way for no known reason. What causes this anomaly.. We don't know yet. We see the same thing some time with PC's and laptops rebooting for no known reason. Can be a electrical loss of power for a split second or even an erroneous command interpreted as a reboot procedure. All it needs is for ones and zeros to get out of place.
I now own a used Phantom 1. I have lots of RC experience (decades). Here are two immediate observations:
1- When the USB cable is connected to a PC, it causes the phantom CPU to enter a completely "non-flight" mode. And the phantom bare usb connector just rests on plastic in the battery compartment.
- Could a static spark to the USB connector cause the CPU to lock out of flight mode, and into "I'm talking via the USB to a computer mode" ????? USB works on 5 volt signals, or maybe the 11.4V power needs to trigger this mode - either one is way less than the 4kv-20kv we can generate in static electricity as humans moving around.
2- The USB connector to the LED PCB is easily dislodged, especially by a larger battery pack. I just did this, so I know it can happen. Can a crooked/loose usb cable to PCB connection cause the phantom to lock up the flight mode?
1- When the USB cable is connected to a PC, it causes the phantom CPU to enter a completely "non-flight" mode. And the phantom bare usb connector just rests on plastic in the battery compartment.
- Could a static spark to the USB connector cause the CPU to lock out of flight mode, and into "I'm talking via the USB to a computer mode" ????? USB works on 5 volt signals, or maybe the 11.4V power needs to trigger this mode - either one is way less than the 4kv-20kv we can generate in static electricity as humans moving around.
2- The USB connector to the LED PCB is easily dislodged, especially by a larger battery pack. I just did this, so I know it can happen. Can a crooked/loose usb cable to PCB connection cause the phantom to lock up the flight mode?
Subscribed.
Had a flyaway yesterday, P2 H3-3D.
Had GPS lock before take off, and was flying around 500m away and it took off flying away from me. Hit return on the controller but nothing happened. Switched controller off an it did the return to home .
I'm blaming some nearby low volt power lines but on a subsequent flight around 200m from the power lines I got solid red LEDs on all four legs and odd flight movements, but was able to land.
Had a flyaway yesterday, P2 H3-3D.
Had GPS lock before take off, and was flying around 500m away and it took off flying away from me. Hit return on the controller but nothing happened. Switched controller off an it did the return to home .
I'm blaming some nearby low volt power lines but on a subsequent flight around 200m from the power lines I got solid red LEDs on all four legs and odd flight movements, but was able to land.
Similar threads
