Pitch oscilation when moving forward

ArduCopter Traditional Helicopter
1

This discussion is an evolution of what happened here: Trad heli tuning + Fport can’t send telemetry back

Since the good results at the time I made the following changes:

  • DFC head as the old one was completely wobbly and unsafe
  • upgraded esc: DS450 digital on swash, SD530 digital HS on tail
  • PWM frequency increased.

Following the upgrade the head (and swashplate) I adjusted the links and PWM settings to match the old head.

When I went to test flight, I noticed an unsolvable problem of pitch osculation in all flight modes as soon the copter start to translate more than 5m/s given or taken.

From the past experience with @Ferrosan I try to further reduce the gain for angle pitch and roll without success. I also tried to reduce further more settings “randomly” looking for some improvements but nothing.

I then decide to go “old school”. I setup a pure acro mode without leveling and assistance and begin to fly. I noticed that reducing the angle gain was counterproductive, in fact I was having more and more pitch oscillation. Considering that in acro mode most of the PID loops are not affective I concentrate myself on solving this problem.

I ended up at almost opposite of the last discussion with angle roll 4.0, pitch 9.0, yaw 7.5. In this way I can have a fully stable and controllable acro flight with head speed around 1900-2000.

Happy with this results I start to re-test stabilize, alt hold and loiter mode.

Now, regardingless what I was changing, I was obtaining either strong osculations (pitch or roll) in hovering/low speed or same as before, 1-3hz pitch oscillation when moving forward more than 5m/s.

On last flight (before Christmas) the vibration were so prolonged (following probbalu 4-5 flights of continue testing) that the tail boom become loose and I lost tail control. You will see in the log that I have to enable motor interlock in flight to “save” the copter and keep safety.

Any suggestion is welcome. As usual parameters are attached and log can be found in the following folder

trex 450 crash parameter.param (23.2 KB)

2

Only a guess here, but I’m wondering if your I gain may be too high, especially in the pitch axis (0.3).

ETA: ATC_ACCEL_X_MAX is set really high for pitch and roll too. (162000).

3

This is a major rework, I would have started with a “fresh” tuning from default on the ATC (current paramenters are messed up, as @nspector pointed out, on pitch you have I at 0.3 and FF at 0.05 ).

Also check your vibration levels with default tuning parameters, right now they are not acceptable.

1 Like