[bhuism]

@lkaino, Yes, this was all careful LOS flying (am on vacation, don’t want to break stuff I can replace now), and in general, I don’t dare to use the goggles yet, but I will.

What do you mean by back and forth, I think the rccommand[roll] is all me 🙂 you mean this as attached? gyro oscillations seems to be within acceptable limits.

Attachments:

1. 

   Screenshot-from-2017-06-11-21-06-41.png
   

[bhuism]

@lkaino, mine flies great again, no yaw anymore after rolls, or other strange yaw behavior, almost all defaults, except P and D yaw. For some reason the servo min+middlemax was 300+ points off, maybe from a crash the servo shifted. After an unarmed tail tune everything was fine again. I think, this also explains the constant high yaw I.

https://drive.google.com/open?id=0ByiRgpuS3FxsdkQwbEc0RVB5NTA

# diff
# Triflight 0.7 Beta 2 / BETAFLIGHTF3 3.1.7 Jun 4 2017 / 13:24:55 (1698ba5)
name baby
resource MOTOR 1 B07
resource MOTOR 4 NONE
resource MOTOR 5 NONE
resource MOTOR 6 NONE
resource MOTOR 7 NONE
resource MOTOR 8 NONE
resource MOTOR 9 NONE
resource SERVO 1 B09
resource PPM 1 NONE
servo 5 1091 1941 1515 40 40 100 -1

map TAER1234
serial 0 32 115200 57600 0 115200
serial 2 8192 115200 57600 0 115200
aux 0 0 0 1300 2100
aux 1 28 0 1750 2100
aux 2 26 0 1300 2100
aux 3 31 1 1300 2100
rxfail 5 s 1000
set rssi_channel = 7
set debug_mode = GYRO
set digital_idle_percent = 1.120
set motor_pwm_protocol = DSHOT150
set frsky_vfas_precision = 1
set bat_capacity = 1500
set ibat_scale = 247
set gyro_sync_denom = 2
set deadband = 2
set yaw_deadband = 2
set tri_tail_servo_speed = 318
set tri_servo_feedback = EXT1
set acc_trim_roll = 8
set pid_process_denom = 1
set blackbox_rate_denom = 2
profile 0

set p_yaw = 65
set d_yaw = 20
rateprofile 0
rateprofile 0

set rc_expo = 40
set rc_yaw_expo = 30
set thr_expo = 70
set roll_srate = 75
set pitch_srate = 75
set yaw_srate = 65

[bhuism]

@phasor, yes it does, like (almost?) all F3 boards 🙂

What current scale factor are you using for your BFF3 board?

[bhuism]

@lkaino, thanks again! I’ll try to explain the pitching (which gets corrected with a full deflected servo at the end of the roll) during a roll.

And try to keep yaw 0 during a roll 🙂

Another question, the I term of yaw is pretty consistently high through the flight (only not during the roll), is that normal behavior? In the past I did not see this.

ps: I did tail servo endpoint tuning with Davids servo tuning tool and unarmed tail speed tune, not armed tail tune.

[bhuism]

@phasor, what d’you use for bff3?

[bhuism]

The current state of affairs.

Attachments:

1. 

   IMG_20170604_204735-2.jpg
   

[bhuism]

@lkaino: yeey! 🙂 directly put it on. I’m sure the release is fine but I get hard uncommanded yaw’s after a roll. Took some time to get good on camera, but I got it. Sync with +2.6 sec.

log / video. If you want to take a look, I am very interested in whats wrong.

The roll ends on frame #57894 (14.891 sec) then a hard uncompounded yaw to the right.

I had acc_yaw_correction on 0 this flight.

relevant non standard settings:

set servo_lowpass_hz = 40
set servo_lowpass = ON
set yaw_lowpass = 50
set d_yaw = 80

complete diff:

# diff
# Triflight 0.7 Beta 2 / BETAFLIGHTF3 3.1.7 Jun 4 2017 / 13:24:55 (1698ba5)
name baby
resource MOTOR 1 B07
resource MOTOR 4 NONE
resource MOTOR 5 NONE
resource MOTOR 6 NONE
resource MOTOR 7 NONE
resource MOTOR 8 NONE
resource MOTOR 9 NONE
resource SERVO 1 B09
resource PPM 1 NONE
servo 5 1102 1993 1554 40 40 100 -1

feature AIRMODE
map TAER1234
serial 0 32 115200 57600 0 115200
serial 2 8192 115200 57600 0 115200
aux 0 0 0 1300 2100
aux 1 31 1 1300 2100
rxfail 5 s 1000
set rssi_channel = 7
set debug_mode = GYRO
set digital_idle_percent = 1.120
set motor_pwm_protocol = DSHOT150
set frsky_vfas_precision = 1
set bat_capacity = 1500
set gyro_sync_denom = 2
set deadband = 0
set yaw_deadband = 0
set servo_lowpass_hz = 40
set servo_lowpass = ON
set tri_tail_servo_speed = 305
set tri_servo_feedback = EXT1
set tri_motor_acc_yaw_correction = 0
set pid_process_denom = 1
set blackbox_rate_denom = 2
profile 0

set yaw_lowpass = 50
set d_yaw = 80
rateprofile 0
rateprofile 0

set thr_expo = 50
set roll_srate = 75
set pitch_srate = 75

[bhuism]

Is this whats happening?

Baby tricopter yaws after punch outs

Now I need to increase acc_yaw_correction again I guess.

[bhuism]

@bhuism: Did you get the twitches now after changing the value of the parameters? They seem to be caused by either of the front motors, but I can’t figure why. The twitch is always happening to same direction, either the left motor is loosing sync or right motor is accelerating for some reason. The FW is doing nothing that should cause them.

@lkaino, you were right about the twitches, I replaced the front right motor, (and went to dshot150) the twitches are now gone.

[bhuism]

quick tip, don’t blindly put the results of a diff command from BF 3.1.6 (TriFlight beta 1) in BF 3.1.7 (TriFlight beta 2) The numbers identifying modes moved a couple of places. Now my transmitter switch was inducing fail saves, in stead of air mode, which was worrying and -in hindsight- hysterical.

I also saw the quick pitch change when airmode disabled, not to mention slow roll’s.

Detail: blackboxlog viewer still identifies the flightModeFlags as AIRMODE, which really was failsafe.

[bhuism]

I really like this beta2 as well, I already lowered P and D of yaw.

The weird twitches are gone , don’t ask me why. I changed to many things to know what caused them. I reset the parameters and reconfigured from scratch (maybe I did set gyro + pid to 8khz, with which it’ll go to 100% cpu) I also took the servo apart and put it back together again, wanted to check for flat gears. And reversed the direction of the front motors.

This is the first flight with motors softmounting, yey!

1 think I don’t understand, I think the all the gyro’s are pretty clean. (when hovering inside, mind you) How can the yaw D be so much more noisy?

quick hove inside:

https://drive.google.com/open?id=0ByiRgpuS3FxsZ2J1VmctR0RKbDA

outside, again the dreaded oscilations I can’t get rid off:

https://drive.google.com/open?id=0ByiRgpuS3FxsSWVTSVpQSnFNVjQ

Both where with tri_motor_acc_yaw_correction = 0, outside flight with relative normal PID settings.

If there is anything I can test lemme know, got more time tomorrow.

[bhuism]

@lkaino, see log + video (log sync: -6.251), I am afraid I could not reproduce the drift in this flight, mainly cause I can’t fly fpv (yet), need more open space. Accidentally I allready tuned the d-term filters more as you suggested. After this flight I reset all PID settings (not filters) to defaults, and that flew better. I do have weird twitches all of a sudden, they can be heard good, and cleary visible in the gyros at: 00:25.333.

[bhuism]

@grahamwho here you go. with docker it’s easy to build these days.

https://github.com/betaflight/docker-betaflight-build

Attachments:

1. triflight_3.1.6_SPRACINGF3.hex_.zip

[bhuism]

If only in angle mode, then I suspect the autolevel pid settings? or a combination, did you lower the auto level term?

[bhuism]

@db1, drift on roll? my first guess would be vibrations in the frame? Can you post a BB log? so we can look at the gyro’s.

here is mine, baby tri, flies good except some drift on yaw in turns.

https://drive.google.com/open?id=0ByiRgpuS3FxsZlZBaEdHMlFKTkk

@lkaino, what do you think of these logs, better he?

[Author]

Posts