Flyaway Flyer Update – Emergency Retrograde
Naza-M V2 / 3.12 Advisory
This post may update over several days...
In my last crash, 3 blogs back, until today I have been reluctantly blaming a GPS disc mount fail for that loss of control and resulting accident.
I have never been more than 20% happy with that explanation. Yes I can think of reasons why it may have come loose, but overriding everything is the fact that in the on-board footage, I can see no sign of anything moving independently of the craft itself. At several points, as the flyer tumbles I am fairly sure I should be able to see some sign that the GPS disc had come off, and even looking frame-by-frame (and I have the entire crash at 60 FPS !) I do not once see any evidence that is it is anywhere other than its post.
But, knowing the Naza’s flight characteristics quite well, as I thought I did, I was convinced nothing else could cause such an unpredictable and unusual type of crash.
If it had dropped like a stone, I might have been expecting that – any number of easily explainable things can cause that type of crash. My decision to fly here was based largely on the worst case scenario being just that, and if that had happened, no people or buildings would have been in danger.
But to have a craft that has been perfectly under your control for 2 minutes suddenly unstoppably accelerate in a random direction, only stopping when it collides with something or power is cut to its motors – that requires a definite answer, and a definite solution before I can fly anywhere near people or buildings again. Any tests after this will have to be in open fields, miles from anywhere, and I will have to be prepared, in the event of problems, to first go into Manual mode, which removes the possibility of GPS and gyros making it all go wrong, but if that doesn’t work, to cut motors and manually crash at the first sign of a fly-away or any loss of control.
And then I discovered this. Thanks Chris B.
A helpful list of people who have had pretty much the same experience as me since upgrading the firmware to 3.12.
Some of them are also having the yaw issues I have been experiencing…
Help us, DJI !
I’ve contacted DJI and given them all the relevant information, and the link above. I eagerly await their response…
But what can we do ?
Tell the CAA for a start – it may be that recent multi-rotor accidents are being blamed on GPS IMUs becoming dislodged from their posts mid flight, and that is not the case.
Their advisory recommends manually checking the security of the GPS unit before each take-off, and this still remains sound advice.
Update: That advisory refers to an incident involving the Wookong-M, not the Naza-M
But I need to go flying, and in more places than 3 tedious fields with nothing in them.
So – what I’m going to do is downgrade the Naza-M to the last stable version of V1, which, assuming that goes OK, should eliminate the problem at one stroke because V1 software has never flown away in a whole year of flights, or in any report that I’ve ever seen. I am fairly sure this a new problem, specific to V2 firmware update, and the 3.12 revision. I can live without auto-take-off.
Ultra-safe test conditions will remain until I am convinced that my machine will never leave my control and fly away again.
In all the time I’ve flown, I’ve always had excellent performance from the Naza, and I’ve come to install its updates without question, and not expect problems, which is a testament to DJI’s thoroughness in design and testing of its flight controllers.
And I was very pleased to be able to go up to V2 firmware without having to upgrade the hardware, so if it wasn’t for the fact that it occasionally flies off, I’d still be singing their praises, and will be again if / when they fix it…