AC3.3rc7 almost hurt me bad, flipped on AUTO into ground - hope someone helps out on this one - Discussions

Hi,

I have had about 20 flights or so on AC3.3rc7 on Pix....I usually fly an AUTO mission and I manually takeoff (even thou I have a TAKEOFF entry of 10m) and then switched to AUTO at a low altitude and then it takes it from there, been doing this for months.....never a problem.

Today I did same thing, armed, took off manually and climbed to a few meters then flipped the flight mode switch to AUTO....copter abruptly did a 180 turn heading straight into the ground and at me....I quickly jumped out of the way and fell over some bushes...incredibly neither me or the copter were hurt/damaged....the copter crashed into very soft landscape and was intact.

Obviously very concerning because the copter headed straight at me at nearly full throttle upside down, me or someone else could have gotten hurt bad.

No errors that I can find on the rather short log file attached...I did add a second GPS last week but have had several flights since then....I have now reverted to 3.2.1 and all is fine apparently....flew an AUTO mission with no problems

onboard vide link of the crash....

https://youtu.be/mxRshWNRNmM

2015-07-18 18-33-32.bin

2015-07-18 18-33-32.log

IvanR,

Thanks for the report. It's uncovered a couple of issues although I can't directly tie them to the flip.

What's clear is that the althold controller (used in AUTO) is pulling the throttle very low after the vehicle is switched to Auto. This is because the Throttle Accel gains are a bit high:

ACCEL_Z_P = 0.898 (default is 0.5)

ACCEL_Z_I = 1.796 (default is 1.0)

and also because the vehicle is trying to slow rapidly from a 3m/s climb rate when entering auto to a 1.8m/s climb rate.

I think there's at least two possibilities:

1. the low throttle reduced the roll, pitch, yaw gains so much that the vehicle could not maintain it's attitude

2. some of the motors stopped when the throttle was pulled low. I see the MOT_SPIN_ARMED is set to zero on this vehicle so it would be difficult to see if the THR_MIN value is high enough to keep the motors running. I'm wondering if you could try setting MOT_SPIN_ARMED to 130, then arm in stabilize with throttle at zero and check that all motors are spinning.

During the investigation I found THR_MIN is being scaled incorrect in two places (angle boost calculation and stability patch). I'm unsure if these could cause your crash but I'm fixing and investigating further.

Thanks again for he report.

Replies

Ivan R July 28, 2015 at 6:36am

hey Randy, just wantd to provide update on this was a motor going bad....thanks for your help
- Developer
  
  Randy > Ivan R July 28, 2015 at 7:04am
  
  Ivan,
  
  Ok great, thanks for the update. The bug we uncovered related to THR_MIN is fixed now in -rc8.
Developer

Randy July 20, 2015 at 1:09am

IvanR,

Thanks for the report. It's uncovered a couple of issues although I can't directly tie them to the flip.

What's clear is that the althold controller (used in AUTO) is pulling the throttle very low after the vehicle is switched to Auto. This is because the Throttle Accel gains are a bit high:

ACCEL_Z_P = 0.898 (default is 0.5)

ACCEL_Z_I = 1.796 (default is 1.0)

and also because the vehicle is trying to slow rapidly from a 3m/s climb rate when entering auto to a 1.8m/s climb rate.

I think there's at least two possibilities:

1. the low throttle reduced the roll, pitch, yaw gains so much that the vehicle could not maintain it's attitude

2. some of the motors stopped when the throttle was pulled low. I see the MOT_SPIN_ARMED is set to zero on this vehicle so it would be difficult to see if the THR_MIN value is high enough to keep the motors running. I'm wondering if you could try setting MOT_SPIN_ARMED to 130, then arm in stabilize with throttle at zero and check that all motors are spinning.

During the investigation I found THR_MIN is being scaled incorrect in two places (angle boost calculation and stability patch). I'm unsure if these could cause your crash but I'm fixing and investigating further.

Thanks again for he report.
- Ivan R > Randy July 20, 2015 at 4:24am
  
  thanks so much Randy...I will adjust the settings as you recommended thou I have had this settings for quite sometime w/o issues..... should I try AC3.3rc7 again? do you think any of this is related to it? or should I stay on 3.2.1? appreciate your help...
  - Developer
    
    Randy > Ivan R July 20, 2015 at 4:56am
    
    Ivan,
    
    I've pushed a couple fixes into master and these will go out with -rc8 so it would be a great help if you could retry with -rc8 once it's available. I think it'll go out before this coming weekend.
    - Ivan R > Randy July 20, 2015 at 8:19am
      
      ok thanks, will do an report back

This reply was deleted.