Home Forums Everything about everything else Triflight 0.7 Beta 3

Viewing 15 posts - 1 through 15 (of 117 total)
  • Author
    Posts
  • #40045
    lauka
    Participant
    #40050
    Kevin_Erik
    Moderator

    Thank you Lauka!

    Downloaded and will install / fly it once the weather has improved.

    Will run Master Air Screw 6×4.5×3 propellers with F80-1900kv’s so I’ll need to retweek the tail to prevent tail-wag from all that extra torque. The upside is that the flight times are 10-12 minutes per 4S 1800 mah, far better then the 5-6 minutes i had before.

    Kevin

    #40052
    Kevin_Erik
    Moderator

    Installed !

    # version
    # Triflight 0.7 Beta 3 / RCEXPLORERF3 3.1.7 Oct 21 2017 / 13:57:41 (17fa321)

    Far as I can tell the CPU load has diminished by 2-4%. Maybe this is the effect of having a clean Firmware?

    #40054
    Kevin_Erik
    Moderator

    The new firmware is installed but yet the front left motor refuses to spin during motor test!

    All other functions ran fine, including tail-servo calibration.

    Checking to lose runs ect.

    Edit: Located a broken ESC wireโ€ฆ will reattach once the kids are in bed. I wonder if the amount of travel for the wire is enough to cover the swing of the motor arms?

    #40055
    Kevin_Erik
    Moderator

    Completed in-flight tail cal with no noticeable tail-wag using stock PID’s.
    I did however add a bit of filtering for the yaw (500) as I have found in the past that this can be a little helpful.

    Now as for the flight after the cal, OMG!

    Can we say I forgot that the firmware flash removed the RC profile settings. Previously the RC values were at 50% for each vector and once that was off, it was like flying insane mode!

    In Flight Cal and post cal crash (100% RC rate was too much for me): https://youtu.be/GaAlG2lQiO0
    Remainder of battery cut to second video:https://youtu.be/2NgAtvovnwA

    Side note: The baro was jumping all over the place in the sensor monitor. I set the LPF for it to 0.75 and it seemed to quiet down and function normally. Beyond this, the BARO Hardware setting I manually set to MS5611 and “saved” gets reverted back to AUTO at some point. Just tried it again and now everything looks ok but I am sure It will get set back to auto later. (Disconnected and reconnected to verify it took)

    DIFF Dump
    `# 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

    feature CURRENT_METER
    feature OSD
    map TAER1234
    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 align_board_pitch = 180
    set gyro_sync_denom = 2
    set servo_lowpass_hz = 300
    set servo_lowpass = ON
    set tri_tail_motor_thrustfactor = 52
    set tri_tail_servo_speed = 216
    set failsafe_off_delay = 20
    set failsafe_throttle = 1150
    set failsafe_procedure = AUTO-LAND
    set acc_lpf_hz = 4
    set baro_noise_lpf = 0.75
    set baro_hardware = AUTO (should be:MS5611)
    profile 0
    set yaw_lowpass = 500
    rateprofile 0
    rateprofile 0′

    #40063
    Kevin_Erik
    Moderator

    Baro continues to be an issue. It doesn’t show in Betaflight as being an active sensor despite the set baro_noise_lpf being set back to its default of 0.6

    RC rate now set to 50% with Yaw @ 30%
    rc_rate = 50
    Allowed range: 0 – 255

    rc_rate_yaw = 30
    Allowed range: 0 – 255

    #40064
    Terje
    Moderator

    Very nice! Any chance of basing the next build on Betaflight 3.2.1 or later? I have become addicted to camera_control on y quads and would like to have the feature on the tricopters as since the camera is pretty inaccessible.

    #40075
    Kevin_Erik
    Moderator

    @Lauka

    I have a minor bit Tail oscillation when the tricopter is idle on the ground. The oscillation becomes a full-blown wag in forward flight. It’s like it’s overshooting and recorrecting.

    Tail twitch can be seen in video: https://www.youtube.com/watch?v=PM1rwsYw8G4

    Do you have any suggestions as how to best tune this out?

    diff
    # 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

    feature CURRENT_METER
    feature OSD
    map TAER1234
    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

    #40076
    biggestRCEfan
    Participant

    Kevin, the oscillations I see in your video I wouldn’t characterise as tail wag but rather some oscillations in pitch and roll at the end of a maneuver. What I see is PIDs that need some tuning. Both in roll and pitch the response is not very crisp which leads me to think that roll and pitch PIDs require some work. Did you retain the tune from the previous version of triflight?

    What I also found useful in triflight based on BF is to beef up the hyper expo (is this the right term?) on roll and pitch to increase rates when doing flips and rolls. This reduces the time being upside down and less height loss. In forward flight I do a snap roll at full stick deflection without loosing much altitude.

    #40077
    Kevin_Erik
    Moderator

    My PID’s from the last Firmware were mostly stock. Beyond that, I full agree that it feels as if its constantly overshooting by a small amount. Even does this when the motors are spinning “idle” on the ground. I can clearly see a 0.5 to 1mm twitch for no discernible reason.

    Question: Can the Mini’s board / FC / BEC handle 5S or 6S batteries ?

    For now I’ll change the stock YAW PID’s to the following:
    p_yaw = 63 (was 73)
    i_yaw = 40 (stock value)
    d_yaw = 30 (was 26)

    #40087
    lauka
    Participant

    @Kevin_Erik: the oscillation is not so visible on my phone screen, but it looks like it’s overshooting a bit. You could start by lowering the P gain as you said.

    @Terje: as I mentioned in the other thread, the problem is the flash space on F3 boards. I could of course compile only for F4+, but I don’t myself have any F4 boards.

    #40103

    Duuuude this firmware is amazing! Fantastic job Lauka! Barely did any tuning on the baby and it’s pretty much spot on!

    Sucks that we cant do 3.2 ๐Ÿ™ on the F3 chip.
    Any possibility to remove some stuff not necessary for tricopters so we can cram it in there?
    No idea what those would be ๐Ÿ˜€

    Great job my friend!

    #40104
    lauka
    Participant

    @david: Good to hear :).

    Sure it would be possible to for some F3 boards, but the boards with integrated OSD will be tough ones. The 3.2 has more stuff like camera control that we would have to decide on what to keep and what not. There would be people complaining that they don’t have support for this and that. We could have a poll on what features are the most needed.

    #40106
    Terje
    Moderator

    Well, for the BabyTri you’d allways have the option to install a BetaFlight F7 all-in-one board and run Triflight with everything enabled.

    The tricky part is the integratedF3 board that the Mini and full size V4 depends upon. There are no alternatives to the integratedF3 board on those builds. Would it be possible to do a specific build for the integratedF3 board? It doesn’t support Dshot, Betaflight OSD, etc, anyway, so maybe omit features like that?

    #40107
    Kevin_Erik
    Moderator

    @David,

    Having the option of a non-integrated Flight Controller would allow for greater flexibility here. (Integrated BEC makes more sense) Granted everything being integrated makes for a lighter cleaner build. However it also makes a single component failure down the entire board/frame. Would like to see a system of similar form factor where an FC can be dropped into the center of the board. From there the power supply connections from the integrated BEC could be wired to the FC, ESC’s and servo.

    Case in point, The other day I lifted a pad…. totally my fault yes, but the entire unit is pretty much unflyable. Keep in mind that I’ve experience with multi-layer PCB repair (Naval Lab). In most all cases its either too expensive / time consuming to correct or the “fix” is forever suspect. This is why folks stopped repairing most consumer electronics and instead just replace everything.

    Note: The previously broken wire that was later reattached was the one that broken free of the pad damaging it.

    Kevin

Viewing 15 posts - 1 through 15 (of 117 total)
  • The forum ‘Everything about everything else’ is closed to new topics and replies.