Pixhawk 2.1 spontaneous severe toilet bowl crash

ArduCopter Copter 3.6
1

Hey, thanks for reading this post.

Details of the build:
Pixhawk 2.1, Copter 3.5 and tried with 3.6 beta
HERE GPS
Tarot X4 frame
T-Motor U7 420KV Motor
Tarot 1855 props
40-Amp T-Motor ESCs
6S 16AH Lumenier LiPo

So I finally got my big quad back in the air yesterday! Most recently I had to shoot it out of a tree last August after it entered a severe yaw spin during an AutoTune.

After two new props and GPS mast, the quad looked airworthy again. So I hovered in stabilize twice to make sure it holds well before entering loiter. Then I loitered a few feet off the ground for about 15 seconds and everything looked fine. Next I sent it about 20 feet or so in the air to see if it continues to behave well. It was a great success! Aside from an altitude drop during yaws, it seemed perfect.

However, when I began to descend the drone began to toilet bowl worse than I have ever seen a drone toilet bowl before. I pulled the throttle down to try to land it before it swung itself over the nearby lake. I managed to crash the quad on the grass and only break one prop and the GPS mast.

I can’t understand how it went from flying nearly perfectly to going full stupid. There are some power lines/transformer maybe 40 feet or so from where I flew… but can those really cause such a bizarre spontaneous toilet bowl like that? Or was it just a bad calibration to begin with?

If pics or anything like that would be helpful, I’ll surely provide them. Any insight any of you could provide would be massively helpful.

2
  1. You are over-propped. 420kV on 6S calls for 15 inch props, 16 inch at most.
  2. You have misaligned motors. For most of the flight, your CCW motors work, while the CW ones are close to idling speed.
  3. You mau have a loose motor mounting, as well, because at one time during the loiter the imbalance between CW and CCW motors worsenes (line 107400 and forward)
3

Asher
Asher.jpg1715×527 146 KB

4

Your CW props/motors create much more thrust than your CCW.
They are not only close to idle speed, they hit the lower limit and you loose controllability, as you set the servo limit to 1100. You configured your servo limits from 1100-1900. you should change it to 1000-2000 and you will gain more controllability at the lower end of rcout.
The big difference of up to 300ms between CW and CCW could be wrong ESC calibration or more probably the CW props geometry creating more thrust.
I had the same issue once with Tarot Propeller 15x5. After trying everything, also changing to spare tarot props, I made then the trial of turning the copter 90 degrees and changing the props accordingly and it was the same issue. So it was clear that it had nothing to do with motor alignment, esc, … only the geometry of the props.
Reducing prop size as Cornel proposed is also a good way to increase controllability as your copter is overpowered.

5

I doubt the motor mounts (twisted).
Motor 3 or Motor 4, or both.
Does the prop tips point each other , on all 4 sides?