Home Forums Everything about everything else New type of yaw control

This topic contains 224 replies, has 20 voices, and was last updated by  mcstan 1 year, 4 months ago.

Viewing 15 posts - 31 through 45 (of 225 total)
  • Author
    Posts
  • #38173

    lauka
    Participant

    @litterbug: The not yawing in FFF has been there forever, if it’s the same thing I complained a while back. There’s a link in 0.7 Beta 1 release notes: https://rcexplorer.se/forums/topic/possibility-to-show-your-appreciation-for-triflight-development/page/3/#post-36990. I had an idea that reversing the front motors would help with that.

    The waggle is what confuses me, I get zero waggle on my baby. The waggle is so slow that it seems like I term oscillation. Log will hopefully tell what causes it. I could take a video from my copter for reference.

    Unarmed tail tune also stores the ADC calibration values from the feedback signal, but those are not available in CLI. Maybe they should be. Only way to set them is to run the speed calibration.

    Looking forward to reviewing your logs. Remember to set the debug_mode = TRI :).

    #38174

    lauka
    Participant

    I recompiled the RCEXPLORERF3 target just to be sure, but this should be the same thing. At least the binary size is same.

    #38176

    LitterBug
    Participant

    @lkaino,
    Thought I remembered the ADC values were hidden which is why I reran the speed calibration. It would be nice to see those values. Think I will do a complete disarmed tailtune check/set on both. Thinking about putting V2 on one, and leaving V3 on the other. V2 flew really good on the 2100kv frame, but never tried it on the other frame. I’ll make sure to take my laptop and a fist full of batteries to the field. Maybe the high winds were actually helping before. Was dead calm last night. May throw in a comparison flight of my best quad too.

    It would be good to see a “diff” of your settings and would love to see another video of you flying. Are you running an OSD, and if so, what UART is it connected to? I know to avoid UART2 with the MinimOSDs and have mine on UART1, RX on UART2, and logger on UART3. I may try disconnecting the OSD on one flight to see if maybe it is messing with the ADCs like it does when connected to UART2. Been struggling to get an accurate mah reading too…..

    Cheers!
    LitterBug

    #38177

    lauka
    Participant

    @litterbug:
    Was trying to do a video with the only battery I had charged after work. Was trying too hard to get cool stuff and crashed it :). The servo got a hit and the angles went off. I tried flying after that but it was doing similar waggling than in your video. It might be a good idea to double check your angles as well.

    I’ll fix the servo tomorrow and take a new video with normal flying, hoping that I didn’t bust the gears.

    I’m not running an OSD on my two babies.

    Here’s a diff I took yesterday before making the build, it shows the yaw PIDs that are now default:

    # diff
    # Triflight 0.7 Beta 2 - new control test / RCEXPLORERF3 3.1.7 Jun 15 2017 / 20:06:06 (3baeba5)
    servo 5 1315 2099 1704 40 40 100 -1
    
    feature TELEMETRY
    feature CURRENT_METER
    feature BLACKBOX
    feature AIRMODE
    serial 0 128 115200 57600 0 250000
    serial 1 32 115200 57600 0 115200
    serial 2 64 115200 57600 0 115200
    aux 0 0 2 1325 1650
    aux 1 1 3 1350 1650
    aux 2 31 3 1700 2100
    adjrange 0 0 0 975 1100 15 1
    adjrange 1 0 0 1100 1225 16 1
    adjrange 2 0 0 1225 1350 17 1
    adjrange 3 0 0 1325 1425 18 1
    adjrange 4 0 0 1450 1575 19 1
    adjrange 5 0 0 1575 1700 20 1
    adjrange 6 0 0 1725 1800 9 1
    adjrange 7 0 0 1850 1950 10 1
    adjrange 8 0 0 1975 2075 11 1
    set debug_mode = NONE
    set min_throttle = 1040
    set min_command = 900
    set motor_pwm_protocol = MULTISHOT
    set small_angle = 180
    set deadband = 0
    set yaw_deadband = 0
    set servo_lowpass_hz = 150
    set tri_tail_servo_speed = 368
    set tri_motor_acc_yaw_correction = 0
    set acc_trim_roll = -18
    set baro_cf_vel =  0.985
    set blackbox_rate_denom = 6
    profile 0
    
    set d_lowpass = 110
    set d_setpoint_weight = 149
    set p_pitch = 47
    set d_pitch = 28
    set p_yaw = 73
    set d_yaw = 26
    rateprofile 0
    rateprofile 0
    
    set rc_rate = 129
    set rc_rate_yaw = 125
    set rc_expo = 18
    set roll_srate = 76
    set pitch_srate = 76
    set yaw_srate = 60
    #38180

    lauka
    Participant

    Wtf, I wrote a longish message and it disappeared when I was trying to edit it.

    Here goes again:

    @litterbug:
    I were shooting a video today after work with the only battery I had charged. Was trying too hard to get cool stuff on the video and crashed it :(. The servo got hit and the angles went off. Tried flying after that but I got similar waggling that you had on your video. It’s a good idea to redo the unarmed tail tune.

    I’ll fix the servo tomorrow and take new video with normal flying :), hoping I didn’t bust the servo gears.

    I don’t use OSD on my babies.

    Here’s a diff I took yesterday before making the build:

    # diff
    # Triflight 0.7 Beta 2 - new control test / RCEXPLORERF3 3.1.7 Jun 15 2017 / 20:06:06 (3baeba5)
    servo 5 1315 2099 1704 40 40 100 -1
    
    feature TELEMETRY
    feature CURRENT_METER
    feature BLACKBOX
    feature AIRMODE
    serial 0 128 115200 57600 0 250000
    serial 1 32 115200 57600 0 115200
    serial 2 64 115200 57600 0 115200
    aux 0 0 2 1325 1650
    aux 1 1 3 1350 1650
    aux 2 31 3 1700 2100
    adjrange 0 0 0 975 1100 15 1
    adjrange 1 0 0 1100 1225 16 1
    adjrange 2 0 0 1225 1350 17 1
    adjrange 3 0 0 1325 1425 18 1
    adjrange 4 0 0 1450 1575 19 1
    adjrange 5 0 0 1575 1700 20 1
    adjrange 6 0 0 1725 1800 9 1
    adjrange 7 0 0 1850 1950 10 1
    adjrange 8 0 0 1975 2075 11 1
    set debug_mode = NONE
    set min_throttle = 1040
    set min_command = 900
    set motor_pwm_protocol = MULTISHOT
    set small_angle = 180
    set deadband = 0
    set yaw_deadband = 0
    set servo_lowpass_hz = 150
    set tri_tail_servo_speed = 368
    set tri_motor_acc_yaw_correction = 0
    set acc_trim_roll = -18
    set baro_cf_vel =  0.985
    set blackbox_rate_denom = 6
    profile 0
    
    set d_lowpass = 110
    set d_setpoint_weight = 149
    set p_pitch = 47
    set d_pitch = 28
    set p_yaw = 73
    set d_yaw = 26
    rateprofile 0
    rateprofile 0
    
    set rc_rate = 129
    set rc_rate_yaw = 125
    set rc_expo = 18
    set roll_srate = 76
    set pitch_srate = 76
    set yaw_srate = 60
    
    #38181

    LitterBug
    Participant

    Sorry the earth jumped up and got you! 😉 I’ve been lucky and haven’t busted my servo up yet on baby. I have a spare set of gears and now having two babies makes me worry less about down time. 🙂

    I did a comparison of the default values for the most recent tests I’ve done. Looks like Yaw-P went from 84 to 73 and Yaw-D went from 50 to 26 on both my builds between trips to the field from updating to New Yaw Control V3. Yaw-I did not change. debug_mode = TRI is now also on by default. I’m looking out the window and the flags are hanging straight down. Should be able to get some more clean air testing at the field tonight.

    Looks like your servo is faster than mine @368 compared to 308, you run multishot and I’m on oneshot125, the lowpass filter has been adjusted on yours, deadbands=0, just to name a few things that could make the tunes different. Thanks for posting your diff. Gives me more things to test out at the field. I’ll try to free up my tail pivot mechanisms with some lithium lube and make sure they aren’t binding up before heading out. And hopefully I don’t mess my logging up yet again!

    Cheers!
    LitterBug

    #38182

    LitterBug
    Participant

    @lkaino,
    FYI. Wasn’t sure if you knew that d_yaw was changed in betaFlight 3.X to be yaw jump prevention rather than the typical D as it has been in the past….
    see https://github.com/betaflight/betaflight-configurator/issues/99

    LitterBug

    #38187

    LitterBug
    Participant

    Made it out the the field and got six packs in on Baby Tri 2 in almost dead calm air. Decided to focus on just one frame and play with some of the parameters that changed recently and tweak a few others. Did a complete ground tailtune from scratch setting endpoints, center, and speed before starting. Tail servo still around 310 so wasn’t really able to free it up any. Forgot to start DVR so no video of first flight. First pack felt just like the two I flew yesterday so I backed out the _yaw changes I noted previously. Changed P_YAW from 73 to 80 and D_YAW from 26 to 50 and threw another pack at it. Felt much better. Did a successful in-air tailtune which resulted in thurstfactor=54, which is the current default. Since that didn’t change, I decided to drop I_YAW from the default of 40 to 30. Felt better but after reviewing video, it was worse on the waggle. Dropped I_YAW further on the next pack to 25. Felt really odd. So went back the other way to 32 and left it there for the last pack or two. Video shows it was still not as good as when I_yaw was at 40. Punchouts rock solid, but complete throttle chop results in a wild ride of Roll/yaw flips. Just for fun, did yet another successful in-flight tailtune before last flight, and again resulted in the default 54.

    Reset ibat_scale to default 360 since it appeared that I have been chasing it in the wrong direction. Current used was close, but a little low and dropped ibat_scale to 300. Pretty close to dead on for my build. Max current seems low on the post flight status screen, but I’m more concerned with how much capacity has been used than peaks.

    I’ll add Video shortly….
    Flight Log #1
    Flight Log #2
    Flight Log #3

    Cheers!
    LitterBug

    #38189

    LitterBug
    Participant

    WTH… my post disappeared after adding links to the log files…. Fortunately I backed it up….

    Made it out the the field and got six packs in on Baby Tri 2 in almost dead calm air. Decided to focus on just one frame and play with some of the parameters that changed recently and tweak a few others. Did a complete ground tailtune from scratch setting endpoints, center, and speed before starting. Tail servo still around 310 so wasn’t really able to free it up any. Forgot to start DVR so no video of first flight. First pack felt just like the two I flew yesterday so I backed out the _yaw changes I noted previously. Changed P_YAW from 73 to 80 and D_YAW from 26 to 50 and threw another pack at it. Felt much better. Did a successful in-air tailtune which resulted in thurstfactor=54, which is the current default. Since that didn’t change, I decided to drop I_YAW from the default of 40 to 30. Felt better but after reviewing video, it was worse on the waggle. Dropped I_YAW further on the next pack to 25. Felt really odd. So went back the other way to 32 and left it there for the last pack or two. Video shows it was still not as good as when I_yaw was at 40. Punchouts rock solid, but complete throttle chop results in a wild ride of Roll/yaw flips. Just for fun, did yet another successful in-flight tailtune before last flight, and again resulted in the default 54.

    Reset ibat_scale to default 360 since it appeared that I have been chasing it in the wrong direction. Current used was close, but a little low and dropped ibat_scale to 300. Pretty close to dead on for my build. Max current seems low on the post flight status screen, but I’m more concerned with how much capacity has been used than peaks.

    Flight Log Pack#2 in the next post should match up to this video

    Cheers!
    LitterBug

    #38190

    LitterBug
    Participant
    #38218

    LitterBug
    Participant

    @lkaino,
    I had someone ping me on my youtube video who also saw better results with the previous P_YAW=80 and D_YAW=50.

    Cheers!
    LitterBug

    #38219

    phasor
    Participant

    I was the one that pinged you and those settings eliminated a lot of my tail wag. This is using
    triflight_new_control_3_RCEXPLORERF3-1.zip. It’s too windy here to get a good FPV flight with the baby tricoper but when I do I will post back here with the results.

    Jerry

    #38222

    lauka
    Participant

    Weekend was busier than I expected. I managed to replace the servo gears today and did a quick angle setup. Had time to fly two packs, here’s a raw video of the second flight: https://youtu.be/h7ZANBCwgKE (still uploading). The yaw felt a bit sloppier, maybe didn’t get the angles quite right. I upped the P and D gains, used in flight adjustments so not sure of the exact values. P maybe up by 25 and D up by 15. I lowered I by 10.

    I tried checking LitterBug’s logs but they are somehow broken. Could you try increasing the blackbox denom?

    Will read through your messages on better time next week.

    #38225

    bhuism
    Participant

    Nice flying! for some reason my motor 1 goes to full wheb I enable airmode using the new_control branche. Blackbox also doesn’t play nice when this happens (no log, or incomplete log file), I will investigate further tomorrow.

    #38229

    LitterBug
    Participant

    Nice Flying @lkaino! HD video is SO much better to watch than my current FPV feed DVR. What camera are you using and how are you mounting it? I need to set up a runcam mount on my baby or look into one of the new Runcam Split cameras when they come out. Probably ought to try a Runcam mount, as I would really like to do a rear view camera.

    Cheers!
    LitterBug

Viewing 15 posts - 31 through 45 (of 225 total)

You must be logged in to reply to this topic.