Triflight
Triflight copied to clipboard
lkaino
Strong tail oscillation
Tried the new 0.5 beta2 fw today on my 190-size tricopter (custom frame).
The yaw oscillated like crazy, changing PIDs doesn't seem to make a difference. Flashed back to 0.4 and everything works perfectly. Same PIDs and controller were used. I am using the standard BMS-210 servo. Any ideas?
Awesome work on triflight btw.
Whay PID controller did you use? I didn't tune the luxfloat at all. The gains are most likely off. If that's the case, you could try the rewrite controller.
I first tried the rewrite controller, with a fresh installation of 0.5beta2 and I didn't even touch settings as PIDs or looptime etc.
I thought stock PIDs must be optimized for bigger tricopters so I lowered them a lot (p around 2), but the tail oscillaton was still there.
The problem remained the same even after switching to luxfloat and roughly tweaking PIDs.
I did notice that the oscillation is barely there on descend, it seems to be linear to throttle input. It gets absolutely crazy on punchouts.
Thank you.
It might be related to a bug in virtual servo calculation in beta 2, will be fixed in next one.
I tried beta1 too and the problem was still present. Might be something with CF 1.12 as it worked with 0.4 (betaflight)
That is very strange as beta 1 contained very little changes to flight behavior compared to 0.4. Beta 1 is still based on same betaflight version as 0.4.
Beta 2 is based on CF 1.12.
Do you have a video or blackbox log of the scenario?
Oh you are right then, I probably never tested beta 1, let me test it later today, will report results.
I have the same issue. Crazy tail oscillations with 0.5b2 no matter the pids, but with 0.4 it flies fine
The code correction was done in commit 33dbe9c634f3980e7896a118d97df9dc5212b1c2 in the master branch. Your options are:
- add the hardware servo feedback and use 0.5-Beta2
- use 0.5-Beta1
- if you are brave, make your own unofficial hex (or use an unofficial hex from someone else)
- wait for the next official release
Just noticed beta3 is out, awesome! Will try it out and report back asap. Great job!
I noticed this behavior also on 0.7b1. The issue seems to be that the parameter tri_tail_servo_speed is used even when tri_servo_feedback=VIRTUAL. If I set tri_tail_servo_speed=0 (default 300) The issue disappears.
The servo speed is only used in virtual feedback mode. You don't have the feedback signal connected?
I didn't test the virtual feedback mode. Is the servo oscillating all the time?
My servo don't have a feedback cable and I haven't bothered to add one yet so I'm currently running with tri_servo_feedback=VIRTUAL. With tri_tail_servo_speed=300 the servo goes between what seems to be my tuned min and max as soon as the motors spin up, even before I apply enough throttle to actually lift off. This oscillation does not stop for the few seconds I've been able to keep the tricopter in the air.
After updating, I've been seeing oscillations both at idle and forward flight. Made a video of the test flight: https://www.youtube.com/watch?v=PM1rwsYw8G4&t=144s
Note: Testing yaw / roll mixing but wag can be seen clearly with mixing active and/or disabled.
# Triflight 0.7 Beta 3 / RCEXPLORERF3 3.1.7 Oct 21 2017 / 13:57:41 (17fa321)
servo 5 1144 1998 1577 40 40 100 -1
serial 0 0 115200 57600 0 115200
serial 1 64 115200 57600 0 115200
serial 2 1 115200 57600 0 115200
aux 0 0 0 1800 2100
aux 1 1 2 1700 2100
aux 2 2 2 1300 1700
aux 3 31 1 1800 2100
set min_check = 1010
set max_check = 1990
set min_throttle = 1100
set max_throttle = 1900
set align_board_pitch = 180
set gyro_sync_denom = 2
set servo_lowpass_hz = 300
set servo_lowpass = ON
set tri_tail_motor_thrustfactor = 60
set tri_tail_servo_speed = 216
set failsafe_off_delay = 20
set failsafe_throttle = 1150
set failsafe_procedure = AUTO-LAND
set acc_hardware = MPU6000
set acc_lpf_hz = 4
set acc_trim_roll = -6
set baro_hardware = MS5611
profile 0
set yaw_lowpass = 500
set p_level = 25
set i_level = 25
rateprofile 0
rateprofile 0
set rc_rate = 50
set rc_rate_yaw = 50
