A2 FC Firmware 2.3 - Here We ALL go Again

[kloner]

i had a superX failure that was just like that.... even the same motor....

in my case the motors were incompatible with the esc's

 

[R_Lefebvre]

Is there any description of what some of the values mean? Things like acc_x are obvious, but what about m_x? mod, svn? Not sure I'll be able to figure out what happened. The log data is pretty crude compared to what I'm used to. I would hope DJI engineers have more data than this to look at, and they're only showing some of it to the user?

Anyway, right off the bat, it's very clear than the craft started spinning way before Mot_1 started clipping. Like WAY before. I had to show this in Paint as I couldn't get the log viewer to let me plot two different Y-axis scales?

So, the yaw clearly wasn't caused by a motor hitting it's limit. What's not clear though, is if the yaw was commanded by some flight controller error, or whether there was an external factor causing it, which the A2 simply couldn't control. I can only see what the yaw heading *is*, not what the A2 is trying to do with the yaw. On our system, we log way, way more data than this. Not just what the yaw is, but what the controller wants it to be, why it wants it to be that....

 

[R_Lefebvre]

Assuming that in_A, in_E and in_R are Aileron, Elevator and Rudder input... were you sawing on the controls while the copter was out of control? Those logs show extreme movements during the out of control portion. But I'm not sure if that was you trying to control it, or if the radio control input interpretation into the A2 went nuts, and that's why it flew like that?

I have a hard time believing those were human inputs, because they're crazy. I can't believe you'd be making scrambled eggs on the sticks like that, trying to control the copter while it was doing the Harlem Shake.

I sure hope DJI has access to more logging data than this. No way you can figure out what happened from this.

 

[kloner]

when you hover or select one of those paramters, down below it has a description of what it is.... my screen grab won't show em, but there there. bottom left and bottom right

 

[BirdsEyeProd]

Is there any way for me to review this DJI log file without owning an A2? Can just anybody install the log browser on a computer. It should be trivial for a DJI engineer to show whether or not the copter lost yaw control and knew it, or whether the controller was commanding the yaw. It should always be pretty trivial to show whether or not the yaw happened before or after the motor started clipping on it's limit.

That was me on the sticks trying to get control back. Because it was over the pool area I was trying to move the sticks based off where the nose was to try and get it away from the water. It was spinning so fast that I just couldn't react quick enough but I wasn't going to give up on it.

 

[R_Lefebvre]

when you hover or select one of those paramters, down below it has a description of what it is.... my screen grab won't show em, but there there. bottom left and bottom right

Thanks, I had just noticed that. Not a lot of data.

That was me on the sticks trying to get control back. Because it was over the pool area I was trying to move the sticks based off where the nose was to try and get it away from the water. It was spinning so fast that I just couldn't react quick enough but I wasn't going to give up on it.

Ok. I've never seen somebody try that hard.

All, I can say for sure, is that the loss of control was NOT caused by a motor reaching maximum. The loss of control was clearly way, way before that.

 

[BirdsEyeProd]

Thanks, I had just noticed that. Not a lot of data.



Ok. I've never seen somebody try that hard.

All, I can say for sure, is that the loss of control was NOT caused by a motor reaching maximum. The loss of control was clearly way, way before that.

With close to $18,000 worth of gimbal and camera on the bottom I wasn't going to give up easily! (FreeFly MoVI M5 w/Canon 1DC and Rokinon Lens).

 

[R_Lefebvre]

I tend to be a little more deliberate about it when these things happen. Hands off, and just poke at it. Sometimes stirring the sticks make it worse. And if it's truly out of control, stirring doesn't help.

Again though, no idea what really happened here. Not enough data.

 

[Mojave]

FYI Quinton started a thread regarding this: How to read DJI IOSD logs (How to Instructions): http://www.multirotorforums.com/showthread.php?21095-How-to-read-DJI-IOSD-logs-(How-to-Instructions)

I will be going out this weekend to do some more test flights to confirm my motor alignment.

Is there any description of what some of the values mean? Things like acc_x are obvious, but what about m_x? mod, svn? Not sure I'll be able to figure out what happened. The log data is pretty crude compared to what I'm used to. I would hope DJI engineers have more data than this to look at, and they're only showing some of it to the user?

Anyway, right off the bat, it's very clear than the craft started spinning way before Mot_1 started clipping. Like WAY before. I had to show this in Paint as I couldn't get the log viewer to let me plot two different Y-axis scales?

So, the yaw clearly wasn't caused by a motor hitting it's limit. What's not clear though, is if the yaw was commanded by some flight controller error, or whether there was an external factor causing it, which the A2 simply couldn't control. I can only see what the yaw heading *is*, not what the A2 is trying to do with the yaw. On our system, we log way, way more data than this. Not just what the yaw is, but what the controller wants it to be, why it wants it to be that....

 

[Helirazor]

Just a thought, we had a pot go bad on our dx9. Luckily we picked it up before any damage was done. We couldn't arm the motors and looking on the assistant it showed the rudder was way off centre.
Further testing showed on the servo monitor if a small input was introduced it would go full extreme, intermittently!

 

[R_Lefebvre]

Just a thought, we had a pot go bad on our dx9. Luckily we picked it up before any damage was done. We couldn't arm the motors and looking on the assistant it showed the rudder was way off centre.
Further testing showed on the servo monitor if a small input was introduced it would go full extreme, intermittently!

No, I don't think that was it. Movement was on 3 channels, and look like frantic stick movements, not a bad pot.

 

[gtranquilla]

I had a similar issue only yesterday.... MR was jittery, 3 motors were extra hot after landing..... so back to the test bench.

Frantic but small stick movements appearing in DJI Assistant could be traced to emi or rf noise getting in to the ESC servo wires etc. Move the main power leads etc. away from servo leads to see if the noise can be reduced or eliminated.
Same situation if there are other rf transmitters on the MR......

 

[R_Lefebvre]

Again, that is not the situation here. These were not small movements. Also, no datalogging system will record what the signal getting into the ESC is. The only thing the flight controller sees is the signal coming in from the Rx, and the signal it is sending out to the ESC. The ESC signals usually look pretty erratic in any case. The Rx inputs sometimes have a small jitter, about 1-2 ms. If EMI is so bad to make it worse than that, you'll usually lose the signal completely.

 

[Quinton]

Here is a video of my A2 test when I first got it..

Heres a link if people can not see the embedded video..
https://vimeo.com/88513428

It was doing it completely by itself, I had copied settings in my tx from a Wookong setup and changed them accordingly.
They only went away when I started completely from scratch with setting up the A2, but as you can see there was wild erratic inputs from the FC without the sticks being touched, not sure if it helps any.

 

[kloner]

I've seen this here before, not good.....

 

[rilot]

I see the A2 was brought up in a recent meeting with the CAA: https://docs.google.com/file/d/0BwXwWQiGyCE_Q09MTkFtMmxKdmM/preview

AB opened the discussion by making the meeting aware of his understanding that EuroUSC are currently conducting an investigation into the DJI A2 flight controller. He went on to state that he has canvassed several members of the SUAS community and has found that there is a 50 percent failure rate with the A2 controller. He viewed the results of the survey as significant and stated that, in his opinion, the A2 controllers should not be used for commercial operations. Post a discussion the

following actions were placed:
• AB to provide the CAA with evidence of the fragility of the A2 controller; and,
• The CAA to consider, having reviewed the evidence, issuing a Safety Notice or taking further action.

 

[Quinton]

I see the A2 was brought up in a recent meeting with the CAA: https://docs.google.com/file/d/0BwXwWQiGyCE_Q09MTkFtMmxKdmM/preview

Thanks for sharing.
Point 10 was actually more interesting..

10. As a result, NATS has subsequently decreed that its units are no longer permitted to grant such permissions. A meeting was held between the CAA and NATS recently in order to address the issue. NATS’ main concern centred around the use of the term ‘permission’ and the liability issues that fell out of this (as it was felt that this term removed any responsibility from the small UAS pilot). A number of possible solutions were discussed and hopefully, a way ahead will be agreed fairly shortly. The outline plan is to issue a General Exemption, with the wording of ‘permission’ likely to be replaced by ‘approval’,
plus a removal of any need to seek ‘permission/approval’ for +7kg operations within controlled airspace at 200ft or less and beyond 2.5nm from the runway (ie. outside the ATZ). More details will follow as soon as they are firmed up.

 

[R_Lefebvre]

Where do you find a 99.99+% reliable controller, at any price?

 

[Scott Drysdale]

You only need the FC system to be a little bit more reliable than its reliable 5 volt power supply IMHO. But I don't quite see solid evidence of that with the WKM and the A2.

 

[Jimi Handmix]

The A2 system has taken out my copter. DJI is saying I had an ESC fail and go full throttle (M1) which caused the copter to start spinning... only problem was it started spinning way before the esc went full throttle according to the iOSD Logs. Here are screenshots of the logs from the motors along with the video showing it going down. I am also including my iOSD file so you can take a look for yourself and see if you can find anything. I honestly don't believe that an ESC caused this. The copter started spinning clockwise then stopped and went counter clockwise. I will be inspecting all my motors and escs today and then sending them back to KDE for them to inspect and tell me if they find anything wrong.

Video: http://youtu.be/bgHHrnVx9HM
iOSD File: http://www.filedropper.com/fly011


Hey birdseyeprod,
your video link wouldnt work right. can you please re-post the link again?! would be nice.