I was maidening a mini H-quad with a Pixhawk running AC3.1.  On my first hover in stabilize mode, it worked beautifully. It was hovering perfectly still without any trimming needed. About 4 minutes into the hover, with the quad at about 2 meters high, all the motors suddenly cut off, and the quad just fell to the ground. I have attached the Pixhawk log file and hope some knowledgeable kind soul can help me decipher what went wrong.

Thanks for any help rendered!

Staq

PS: The quad suffered no damage, only 2 broken props. Really wish to know what really happened up there. Dare not fly it at the moment.

Views: 2846

Attachments:

Reply to This

Replies to This Discussion

Sorry, I as about to recheck the log, and I must have been looking at wrong log, or failed to see the cause:

Plot ATT.Roll vs ATT.DesRoll   - , you'll see a huge error near line 16000

same for ATT.Pitch vs APP.DesPitch  ,

now you can plot RCOUT.Chan1...4 , you'll see that all are reducing thrust, except for motor 4 - that is asked for 100% at line 16212.   - that did not help the attitude.

Conclusion is simple; you had an ESC or mechanical failure on rear-right motor.

Basically Tim was right,  I was commenting on one of the other 40 logs I had in download folder :)

no, if all ESC's was out of power/underperforming for any reason, you'd see logs indicate a huge increase in motor PWM and ThrOut.

If motor 4 was actually delivering 100% when commanded, and the other close to nothing, you'd see it tumble vey fast.

What you see is attitude get out of control, and the three other motors dropping throttle in order not to flip the quad,  as motor 4 lost it's power.

Once you rear-right arm was lower than the rest, it had to proportonally reduce the throttle to the other motors, if it did not,  but increased/maintained throttle, it would get lateral speed, and then tumble uncontrollably in the air

Please have a look at the attached graph. I have plotted RCOut together with Voltage, Current and Vcc. Notice that all voltages and current and Vcc already stopped recording just *before* the moment RCOut Channel 4 spiked. So, that could mean no power is going into ESC 4 even though Pixhawk send in a 100% signal to it. In other words, I cannot conclude an ESC problem when there is no power going to all the ESC. But why is Pixhawk still able to log its final dying moments? Could be the backup power from its servo rail as I have a single ESC +5V connected to the servo rail. Is this a symptom of a faulty power supply module (original 3DR component)?

Any thoughts?

Attachments:

Andre, please see my attached graph in reply to Tim's post. Why does all current and voltage readings cut out just before the spike in signal to ESC 4?

ok, I guess you still won't believe my conclusion, like before - but what you *actually* can see, it Chan3 drop proportionally to 4, because motor 3 is the opposite motor of 4.

Like I told you, you lost lift on 4, so the opposite motor *must* reduce power not to flip.

motor 1,2 ?  - they rotate in same way, they only adjust yaw at the time, and are forced to reduction, unless Pixhawk planned to spin it wildly around...

Servo outputs are behind a polyfuse, so pixhawk would not reboot  . but ESC's usually have their own working voltage.

Stop overthinking it, you had a ESC, sync, motor,  propeller or  propeller-mount problem.  - thatt's it.

Pixhawk only logged until crash made it lose power or reboot. - the final few lines may have not been written or not in an cluster that got updated in FAT.

oh, and the voltage drop is about ~35mV,   insignificant.

you whole flight varied between 5.250 and 5.170  - well within limits. ending record of 5.173 is not a problem.

Most of Pixhawk works at 3,3V anyway.  and it would be out-of spec around ~4.8 or less.

Ok thanks. I'll replace ESC 4 and try again.

Hi,

Any solution on this topic? What was your final solution Staq?

Just asking because testing my first Y6B, and same issue with stopping motors.

you need to post dataflashlog - until then, I doubt it's related :)

Can help me somebody to find out what the problem could be?

Analyzing logs too big task for me, as a Pixhawk beginner..

http://www.filedropper.com/10_8

the attached log contains no flight,

the airframe is mostly level within few degrees, your input asks for more, on final "crash detect" the airframe was level within 1degree, (on ground) , your input requested ~27 roll to the left.

then another Error 12.1  Crash detected, which stops your motors.

It seems you need to do a decisive takeoff,  I guess you are so close  to ground, the airframe cannot achieve some of your requested angles, and detects a crash.. (being unable to reach desired attitude)

fly in the air, not on ground :) 

Hi Andre,

You're right. It wasn't a real flight, but a test run on my desk.
Is that normal, that on a test run, without props, the motors cutoff sometime?

Is it safe to make a test flight in this state of readyness of my drone?

Or chance that will motor cutoff in real flight too?

G.

Reply to Discussion

RSS

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service