Anatomy of a DJI Flyaway

Ian sorry to hear about both fly away.
 
Ian, in an 'on the side' discussion concerning your experiment, an interesting question popped up.

Your data logger is pulling info from the Naza at frequent intervals. Does this interaction compromise any of the FC 'thinking ability'?

Not meant to be a clever/trick question- just playing devils advocate and second guessing DJI's stance on your findings.
 
From the very little I understand about CAN bus, the only active interaction happens when a device on the bus wants to transmit. The collision detection is really neat (compared to say CSMA/CD) but, given sufficient priority that does allow a device to have a turn and make everything thing else temporarily shut up. However if all a device is doing is listening, I am pretty sure there is no interaction. It is just ear to the ground, listening to the serial data.
 
  • Like
Reactions: discv
It's a good question. CAN bus is not that different to wired Ethernet. As long as you don't introduce issues on the physical bus, it will survive most anything else you try to do with it. I suppose you can saturate it but I'm just listening for data that the NAZA is already outputting. And my first flyaway, which was identical, didn't have the logger on it.
 
Sorry to hear about your crash - Do you have any further info on the data logger?
 
Little O/T sorry... but any info on where we obtain one of these custom black boxs from :) ??
or if you made it.. any plans to sell them ??

Seems verrry interesting!!
 
The CAN logger is pretty nerdy hands on stuff. You can find most of the work done by pawelsky here: http://www.rcgroups.com/forums/showthread.php?t=2071772. He's the one who figured most of it out. Flytrex is a good option and has the data connection to go with it. I don't know if it logs this much detail. It should.
 
Ian, got a couple of questions to discard options about naza freeze,

- was your quad using fw 3.12 when flyaways happened?
- did you perform a hard reset to naza before install 3.12?

Cheers
 
3.08. There's no way to know if it is firmware specific. It could very well be an edge condition that exists in multiple firmware versions.

Not sure what you mean about the hard rest.
 
3.12 was giving compass calibration issues, and as long as I know the fix is to bridge F1 with F2 port and turn on phantom (Naza hard reset),then connect it to DJI assistant and install 3.12, power cicle disconnecting bridge before turn on, and now compass calibration is suposse to be issue less.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
Ian,
with that much data I'm surprised DJI hasn't offered you a position. You've done more research and testing than most DJI engineers have. Hats off to you again.
 
I had my first ever bonafide flyaway about 6 weeks ago. I had my theories, bought $1,000 in replacement parts, rebuilt and moved on. Then it happened again.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

This time, however, I had a data logger attached to the Phantom, a custom made miniature "black box" that records 75+ different flight parameters 5 times per second. I've spent the past day analyzing the data and I've discovered some things that I didn't suspect before.

CAN-logger-small.jpg


None of the sensors including IMU, compass, GPS, etc. were to blame. The battery, ESCs, motors did exactly as they were supposed to do. There was no loss of connection, GPS interference or any other external factors.

The NAZA locked up. It had intermittent freezes that stopped it from functioning normally. Here is a snapshot of the logged data from when the NAZA stopped working normally.

bendix-crash-table.jpg


At 12 minutes, 45 seconds, the data shows 5 anomalies:
  • The Course Over Ground (COG) freezes. It had been fluid all throughout the earlier part of the flight.
  • The speed (Total SPD) changes to 769,361.92 m/s. It had been normal throughout the earlier part of the flight.
  • The satellite count (SAT) jumps to 204 then returns to normal.
  • Erratic updates of ATTI pitch and roll values. 2Hz or slower.
  • Erratic updates of motor speed. 2Hz or slower.
The cause is likely to be one of two things:
  • Defective hardware in NAZA (e.g. cold solder joint) that took several flights of light vibration to manifest.
  • Firmware defect when flying in edge condition (corner of the envelop) that rarely manifests.
I am working on recovering more of the video (file is corrupted due to loss of power).

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
This is the kind of stuff I like and is extremely valuable instead of the usual "I don't know what happened, so...." so thank you for posting this. I wish all quads had the capability to record such detailed data. FWIW, there are services out there that can extract corrupted video recordings. I had a fly-away once and the video recording got corrupted somehow, I used an online service and they recovered the video which was very helpful in my understanding what happened in my case. I was flying a DJI P2V+ at a beach (lots of RF from various sources including an Air Force Base some about 10 miles north of the beach) and at some point the P2V+ locked on to another signal (5.8Ghz) but that signal was mechanically constant, DJI programmed the P2V+ so that if it had a mechanically constant signal mimicking the TX's signal, the video showed that the quad stopped (over the ocean) held a hover for about 30 seconds, rose to my pre-programmed altitude of 300 feet, did a 180 and RTH right where I was standing... all caught on the video clip... pretty **** cool. I've had RTH automatically invoked twice now, great feature, but I digress.
 
Ian,
with that much data I'm surprised DJI hasn't offered you a position. You've done more research and testing than most DJI engineers have. Hats off to you again.
I'm sure that DJ is well aware of the problem but for the sake of their balance sheet they would not employ him and discuss about it.
 
Just a couple of ideas... I have a background in medical testing using very sensitive recording equipment. I am also a new Phantom 2 owner with practically zero experience in RC copters. A good deal of my job involved tracking and trying to eliminate sources of interference. Seemingly simple things like electrical leakage through insulation, EMFs from any number of sources, static electricity, etc were all worthy of consideration when my recordings seemed sketchy.

Since patients were often being considered for discontinuation of life support and organ donations the medical-legal implications were enormous.

I have looked at the video several times and see a number of potential sources of interference. Frankly, I would not even consider flying in such surroundings. I am assuming that the flight took place at night although the tracking overlay is on a daytime image. That illuminated "Bendix" tower alone would have scared me off!

I would be curious what an EMF scan might show. I can only guess what types of electronic devices might be running in those buildings.

Great job trying to decipher it all and my condolences on the equipment damage.

.
 
There have been similar flyaways both in city environments and far from sources of interference, some diving into water.
DJI often promotes their products flying in urban environments so that should not be a problem.
 
If the Naza is the same from first flyaway, maybe would be a good idea to take a look inside, there's a change that something may be a little loose and when things get hot tends to expand, that could be a source for a shortcut and Naza freeze.
 
It would not be the first case of a dry joint in a Naza unit being identified as a "flyaway" cause either. The trouble is that smd solder joints on multi-layer PCBs are not trivial things to fault find on.
 
It will soon be on it's way to Schenzen for a full analysis. DJI engineers want to see if they can fin the root cause. Problem is we'll never know what they find. But at least they're interested to know themselves.
 
  • Like
Reactions: Hughie and Trumple
I hope they give you a bit more than a 30% off voucher.
 

Members online

No members online now.

Forum statistics

Threads
143,066
Messages
1,467,358
Members
104,936
Latest member
hirehackers