Testing – Triflight rebased on 3.2.2 Betaflight

[Kevin_Erik]

Ubuntu is installing… will look for.

check triGetMotorCorrection() in mixer_tricopter.c
if (isAirmodeActive() && tailServo.feedbackHealthy)
{
correction *= tailMotor.pitchCorrectionGain;
}
isAirModeActive() might not work

And report back.

[Kevin_Erik]

Got the program installed and running

However… will take took long to get into this today as I’ll have to learn how Ubuntu is setup.

[Kevin_Erik]

@Lauka

Did two flights today (Acro mode), one with 4S and one with 6S, using 1-shot settings on the ESC and YAW / ROLL P&I settings that are half of stock. Tried performing an inflight Tailtune but as far as I can tell it failed as I didn’t hear the pretty beeps just the simple double beep. Beyond this, the tri felt sloppy in the turns as if the I-term was too low. Oh and as always the wag is ever present.

Tricopter Baby, 4S LiPo 4-14-2018
Note: Flight starts off with Tailtune so there is about a minute of hovering about.
LOG: https://drive.google.com/file/d/1kwu5Cp9SJhX7gTkR1HelNW-Ck5aUKb3a/view?usp=sharing
Video: https://www.youtube.com/watch?v=EiOgq_PM_gw&feature=youtu.be

Tricopter Baby, 6S LiPo 4-14-2018
Note: 6S flight felt much more stable and locked in than the previous flight, however i forgot to tailtune it for 6S. Also notice the very stupid person walking thru my flight path with her two dogs. She could care less if she’s hit apparently.
LOG: https://drive.google.com/file/d/1IrvttKoifKMfhPkA6z52RplxDdzY4cuq/view?usp=sharing
Video https://www.youtube.com/watch?v=9DshHA3_ovs&feature=youtu.be

# Triflight on Betaflight 3.2.2 v2 / MATEKF405 (MKF4) 3.2.2 Dec 31 2017 / 15:24:22 (30f68c8) MSP API: 1.36

# resources
resource MOTOR 1 C08
resource MOTOR 2 C09
resource MOTOR 3 C07
resource MOTOR 4 NONE
resource MOTOR 5 NONE
resource MOTOR 6 NONE
resource SERVO 1 C06
resource SERIAL_TX 11 A15

# servo
servo 5 1154 1830 1516 100 -1

# feature
feature AIRMODE

# beeper
beeper -ON_USB

# map
map TAER1234

# serial
serial 30 32 115200 57600 0 115200

# aux
aux 0 0 0 1700 2100
aux 1 1 2 900 1300
aux 2 2 2 1300 1700
aux 3 37 1 1700 2100

# master
set dshot_idle_value = 200
set failsafe_off_delay = 30
set failsafe_throttle = 1100
set failsafe_procedure = AUTO-LAND
set align_board_yaw = 270
set bat_capacity = 1250
set ibata_scale = 165
set tri_tail_servo_speed = 358
set tri_servo_feedback = RSSI
set p_roll = 30
set i_roll = 20
set p_yaw = 30
set i_yaw = 20

# rateprofile
rateprofile 0
set rc_expo = 25
set rc_expo_yaw = 25

[Kevin_Erik]

Today i succeeded with inverting the Tri-Baby’s motors and servo.

I did this after the the realization of how the change will impact the copters Yaw / Roll relationship. The standard copter Rolls away from the direction of Yaw, forcing the need for counter corrections. With everything inverted, the copter should Roll towards the direction of Yaw, much like a regular aircraft banking to turn. Granted the copter isn’t really designed to fly this way as it would require drastically changing how the copter is built. However I thought it would be interesting to try while waiting to see Laukas response / advice to my prior posts. Even did a little test flight that i may post a video later on.

Note: Still wags though (sad-face)

https://farm1.staticflickr.com/815/39684570820_fb884ec88b_z.jpg

[Kevin_Erik]

[Kevin_Erik]

OMG!

Was just out for a flight and it was fucking fantastic!

It was so stable and fast with only minor amounts of wag. Will post a video and logs shortly.

[Kevin_Erik]

Video

Note: As of the time I posted this it was still loading onto youtube.

[fomoco_44]

@Kevin_Erik,

Nice one!!…I’ve got a Baby Tri on the way so I might give that a try when it arrives, just trying to figure out how to do it on my Mini Tri now!

[Kevin_Erik]

I don’t think it would be possible on the tri-mini without some serious modifications to the airframe.

[Terje]

Any chance the github page will be updated with the current release of triflight? Maybe with a meaningful version numbering and changelog? 🙂

[Kevin_Erik]

Now that I have a 2nd busted Matek FC, i am thinking of trying out the BARDWELL F4 FLIGHT CONTROLLER which uses the CLRACINGF4.hex

Anyone have experience with either this FC or the CLRACINGF4.hex ?

Or is there something else that can be recommended?

[fotoamg]

Hi!
I do not see CC3D build, how to get one?

Also I’ve downloaded latest triflight for cc3d and flashed I have that problem:
Triflight 0.7b3: zero and low throttle tail twitching, flight is not bad only if you reduce throttle. tri_yaw_boost = 240 default setting makes this twitching, but if you reduce it’s value it will just spin and yaw around and no yaw stabilization so uncontrolled yaw

Is there any way to fix it?
This PWM pulsing at zero throttle is not caused by the servo as it seen on output channel in betaflight gui even when servo is not powered but throttele is low, copter is at the ground.
It wants to compensate.

Check output 6 pwm oscillation on GUI:

[MadPoet]

Hi Lauka,

Sorry to bother you, hope you’re having a great spring. Would it be possible to get a HEX file for OMNIBUSF4 (OBF4)?

I want to play around with a HGLRC f428 flight stack and triflight.

[fotoamg]

Hi!
I received now an emax servo and added feedback wire as VIRTUAL seems not working fine in Triflight at low throttle or not spinnnning motors 🙁

still cannot figure out where to connect servo feedback on a cc3d?

I would need that rebased branch if possible as I need to build different targets, whan I am missing now are:
– cc3d
– omnibus F4 SD

Thanks if anyone can help!
p.s.: now I went down in weight of trifecta as much possible, with small 550mAh 2S battery I could reach 331grams without FPV stuff etc.

Attachments:

1. 

   trifecta_light1.jpg
   

[Kevin_Erik]

@Fotoamg

-Default your filtering and other PID’s until your tail wag gets sorted out. It will make it easier for others to help you if your Firmware is running defaults.
-Increase you Yaw P gains considerably, otherwise the Servo maybe slow to respond. (I’m using 170 currently)
-Add Yaw D gain to help prevent overshoot.
-The Servo feedback wire either needs RSSI or EXT to work. The exact FC pad is different for each controller.
-In Virtual mode, the Bench-Tune Yaw range will complete but the speed test will not without a feedback wire properly installed.

[Author]

Posts