8Kg Copter crash on take off

ArduCopter Copter 4.0
,
1

I have a very strange crash today.

  1. Frame size: 890mm
    2.Motor & ESC: DJI E2000 PRO
    3.Battery: 12S ( 2 x 6S 22000mah)
    4.FC CUAV X7
    The thing is, I have a successful take off and run some hover test in Loiter and PosHold beause the DJI ESC’s problem I always take off on Stabilize mode. and I’ve test this copter for many days, This is the first crash. I really trust this copter so I push the throttle really fast to avoid some take off motor oscillations, the copter didn’t take off just roll over, although It’s a soft ground but still 1 and 3 motor hit the ground very hard. log’s here https://mega.nz/file/t40D0CzS#Kmx4Ph3BlnvlD2BykSSgx12YweJYAWbcawG-quMUy64
    image
    image2880×1635 357 KB
2

There is a NavEKF2: lane switch 1, and EKF variance on the take off moment

image
image638×548 54.7 KB

4

I think the vibration is caused by the copter hit the ground, It a bit delay, and I think It’s some kind of sensor data failure make the EKF very unhappy. as you can see 17:01:11 I have a Roll/Pitch inconsistent by 14 deg. I never the Roll/Pitch inconsistent before.

6

Yes, I always notice the vibration level. but I think we should look at RCO instead of RCIN,

image
image1284×621 41.6 KB
as we can see the vibration is quiet acceptable at this time.

7

And This is today’s frist take off loghttps://mega.nz/file/Z9th2CiQ#a4DQ9CmSMR56wfaYYc0mS5TVnawXoZpFi7rvI9R9Zog
nothing very different from yesterday’s

8

the ekf message is 5 seconds early the high vibration and crash, so my guess is something wrong about the ekf…

9

Yes, you had 2 sensors inconsistent. Probably the reason it changed lanes.

10

Yes i was checked only first gyro data the problem was second gyro that reported different data

Screenshot_4
Screenshot_41698×520 10.5 KB
Screenshot_6
Screenshot_61701×523 9.55 KB
Screenshot_7
Screenshot_71696×525 9.03 KB

i think we can ignore other sensors like baro and mag because of stabilize flight mode
@XinChengGe sorry for wrong analysis , so i remove them to don’t mislead others

11

No worrys @hosein_gh, @dkemxr thank you for your guys help me analysis. The problem is the ekf should work out with this situation, right? @rmackay9

12

just a little update. I replace the ICM20689 with ADIS16470, and set EKF2_IMU_MASK to 7, but the log shows the second IMU sensor’s failure, I tried again, The second IMU’s data is matching the first and third IMU’s now, So It’s like some kind of sensor gltich??? should I disable the second IMU(BMI088) ?