Replies

  • Below are the results from the Autotune that I did on my TBS Style Discovery Frame.

    Before Autotune

    3692908169?profile=original

    After Autotune

    3692908180?profile=original

    The original setting did need some work but the quad was flying very well. The PID values after Autotune appear to be about half what I would have expected.

    David Ardis

  • I did an AutoTune a few days back and posted the old and new numbers a few pages back. It flys very good when I fly visual but yesterday I did an FPV flight which I haven't done since AutoTune. It has this weird little oscillation in the roll axis when flying fast forward. You don't notice it unless flying FPV and it isn't much of an oscillation but it is annoying looking through the goggles while flying. Here are my before and after numbers again and if any of the guru's can tell which to change without having to go all the way back to my original settings (It does fly better after AutoTune in most respects save the little oscillation in forward flight) I would be most appreciative. It's not that it flew bad with the original numbers it just looks smoother in slow manual flight with the new ones.

    Param             Before               After

    Stab P              4.5                  11.5

    Rate Roll

    P                     .0850              .0450

    I                       .0850              .0450

    D                     .0025              .0040

    Rate Pitch

    P                     .0850              .0400

    I                       .0850              .0400

    D                     .0025              .0040

    Thanks for any help you guys can offer.

  • I originally posted this in the 3.1 forum and then came across this one. My apologies if I posted in the wrong forum...

    I have the 3.1 on my 850 size hex and it seems to be flying pretty well. I could use some help though in getting Autotune to engage. I have been through the instructions and tips and reading all I can on it. The 2 position switch has plenty of throw and works well when I assign any other function to it. I have checked and re-checked that sticks and trims are all centered, and verified that through the radio calibration section. I assign the 3 position switch to Stabilize, Alt hold, RTL. All modes show up fine in planner. I switch to Alt hold then put the 2 position to high and nothing happens. After trying everything I could, I went to the field just to see if maybe Auto Tune doesn't show up as a mode and would work anyway, but when I switched to autotune from alt hold nothing happened. I have assigned other modes to the 2 position switch (acro, land,rtl) and they all work fine with the same switch, just can't assign autotune. When I set it and write it, it shows up as assigned, but will not switch to autotune from Alt-hold, when I do MP still shows Alt_Hold as the active mode.
    Anyone else experience this and get it to work?
    Thanks in advance..

  • Hi there,

    I have TBS discovery and I have been running with 3.01 without any problems some time already.

    Now I have installed 3.1 and I have tried several times to run autotune procedure based on your instructions. It will take only 1-2 minutes and it twitch only left and right movements, and after that it stops. So it doesn't do forward and back movements, and process is surprisingly quick. And in the end it doesn't change pids at all, so autotune doesn't seem to work.

    Any idea what I should do? Vibration levels are quite low and all calibrations are done etc...

    I have attached also here my logfile.

    Many thanks in advance!

    2013-12-26 18-07.log

  • Just for the record and with more details if requested I autotuned my heavy, slightly up-powered DJI Flamewheel Hex ("The Witch") today. It went about as seen in the above video, though the roll series took somewhat longer (perhaps because I had to intervene more often due to a tight flying area). It all went better than I expected  because I did not remove the brushless GoPro gimbal (mounted on pretty stiff rubber, not a lot of wiggle there ) nor did I remove one of the pair of 2700 mah 4S Lipos (would have been grossly out of balance). Also left small aux. battery on board and the extended landing gear. I did pull the GoPro itself (for its safety) and secured the deactivated gimbal with a zip tie. Also CG is not perfect on this machine, but it's close. So she was pretty heavy.  Anyway, the roll tuning took a while, as I said, and I was starting to think it wasn't going to "take" when it switched to tuning the pitch. It finished with that a bit faster, and that was it. No drama, everything worked as advertised even with my deviations from the recommended physical setup. After autotune it felt good and loitered on the spot, but I'll have to wait until tomorrow for a real flight test. This Hex was already flying quite decently (bus-like smooth and steady is my goal, it's a camera ship before all else) before the autotune but the process did modify some of the PIDS by quite a bit:      

    Rate_Rll_P went from 0.11 to 0.205;

    Rate_Rll_I went from 0.055 to  0.205;

    Rate_Rll_D went from 0.006 to 0.014;

    Rate_Pit_P went from 0.11 to 0.24;

    Rate_Pit_I went from 0.055 to 0.24; 

    Rate_Pit_D went from 0.006 to 0.015;

    Stb_Rll_P went from 3.8 to 8.475;

    Stb_Pit_P went from 3.8 to 8.475;

    That's all for now -

  • I've run autotune on my DJI 450 frame

    Everything seems to be OK apart from rate_d has not been changed

    here is a log for first two stages of first axis, for some reason it stuck on 0.004000, detailed autotune log is also attached (only ATDE, ATUN and EV messages)

    RateD_UP

    ATUN, 0, 0, 8963.720, 9991.153, 0.090000, 0.004000, 4.500000
    ATUN, 0, 0, 8356.156, 9292.127, 0.085000, 0.004000, 4.500000
    ATUN, 0, 0, 8653.599, 10107.86, 0.080000, 0.004000, 4.500000
    ATUN, 0, 0, 6542.139, 8493.349, 0.075000, 0.004000, 4.500000
    ATUN, 0, 0, 9603.799, 10780.62, 0.075000, 0.004000, 4.500000
    ATUN, 0, 0, 6035.020, 7513.669, 0.069999, 0.004000, 4.500000
    ATUN, 0, 0, 9185.005, 10302.42, 0.069999, 0.004000, 4.500000
    ATUN, 0, 0, 6834.020, 6834.020, 0.064999, 0.004000, 4.500000
    ATUN, 0, 0, 6751.679, 7769.967, 0.074999, 0.004000, 4.500000
    ATUN, 0, 0, 10378.24, 11586.91, 0.074999, 0.004000, 4.500000
    ATUN, 0, 0, 6352.092, 7564.765, 0.069999, 0.004000, 4.500000

    RateD_DOWN
    ATUN, 0, 1, 8362.400, 9456.907, 0.069999, 0.004000, 4.500000
    ATUN, 0, 1, 9454.492, 10499.21, 0.064999, 0.004000, 4.500000
    ATUN, 0, 1, 6632.829, 6632.829, 0.059999, 0.004000, 4.500000
    ATUN, 0, 1, 6273.583, 7642.478, 0.064999, 0.004000, 4.500000

    Changed PIDs

    RATE_RLL_P, 0.090000 > 0.074999
    RATE_RLL_I, 0.000000 > 0.074999

    RATE_PIT_P, 0.090000 > 0.075000
    RATE_PIT_I, 0.000000 > 0.075000

    STB_RLL_P, 4.500000 > 7.875000
    STB_PIT_P, 4.500000 > 7.875000

    autotune_detailed.txt

  • Tried Auto Tune Today. 

    Frame DJI 550, Multistar 2216 800kv, 4S 6000mah, 10x47 Carbon Mixed Slowfly.

    After tuning, my hexa feels loose and mushy, probably because of the low P roll value. My previous PID was using APM Default it flew better. 

    3692903994?profile=original

  • Is it possible for auto tune to not start if APM is not mounted at the cog? I have a dead cat like frame that is having trouble carrying out auto tuning. When I activate auto tune, nothing happens, besides the copter drifting away. 

  • Developer

    Someone asked for extra details on the ATUN and ATDE messages in the dataflash log so I've slightly updated the dataflash wiki page to include the fields and a brief description of what it means.

  • Hi all and specifically Marco Robustini,

    Auto Tune on a HEAVY Hex:

    So I am trying to summerize what I need to do to get a good Autotune given that my first one on my newly built Hex Frame did not turn out very well tuned:

    Result of the Autotune starting from standard params was that it became very solid but the whole thing vibrates rapidly. It also does not respond to Pitch and Roll well, it does it at a very shallow angle (with sticks at max) and vibrates even more at the same time. The Good news is that it is rock solid in its position while in loiter (but still vibrates).

    Configuration:

    - APM 2.5, AC3.1 mounted in the center top frame which connects directly to the arms. Mounted on Kyosho Zeal tape.

    - Hex Frame F800-EVO (DJI S800 EVO (stiff Arms) Clone) Similar to Marco Robustini's https://www.youtube.com/watch?v=-igl79Iw_sI But the EVO version with retractable gear.

    - Retractable Gear (not retracted while in Autotune)

    - Nex5t Camera and gimbal.

    - 8000mah 6s.

    - Hover at 48% in and out.

    - Total Weight: 6.1kg

    - Balanced CG.

    Wind condition: Moderate to low about 5-7mph.

    Possible solutions:

    1) There are 6 vibration ( anti-vibration kit by Secraft ) dampers between the landing gear frame and rails (where the battery is attached) part and the body part where the arms are attached (APM mounted on the body part) . Possibly remove them and attach directly to the landing gear and rails part?

    2) Gear up or down during Autotune?

    3) Remove the gimbal altogether during AT? (Although I had a matching weight replacing the camera attached to it during the AT and I secured the movement of the Gimbal servos (Using Duct tape, but I still had the gimbal dampers unsecured) I could see the gimbal swaying back and forth during the AT. 

    4) Leave the Gimbal on but secure it so that its dampers and servos are not moving at all. This will maintain the weight distribution of the system during the AT?

    5) A combination of any of the above?

    What else would you guys suggest? 

    My results are attached after the Autotune starting from standard.

    Thanks guys for all your efforts.

    Screen Shot 2013-12-19 at 3.56.41 PM.png

    2013-12-19 19-15.log

This reply was deleted.

Activity

DIY Robocars via Twitter
RT @breadcentric: See how the April AWS DeepRacer races have ended and a couple bits of news: https://blog.deepracing.io/2021/05/09/aws-deepracer-league-2021-update-11-end-of-april-special/ #AWSDeepRacer #Machin…
21 hours ago
DIY Robocars via Twitter
RT @sunilmallya: Representation Learning +Instance Transfer to learn new reward functions along with advantage based filtering of new exper…
yesterday
DIY Robocars via Twitter
Apr 27
DIY Robocars via Twitter
Apr 27
DIY Robocars via Twitter
RT @f1tenth: Sliding (autonomously) into the weekend like ... 🤖😎 #f1tenth #robots #AutonomousVehicles @OpenRoboticsOrg @NVIDIAEmbedded @Aut…
Apr 25
DIY Robocars via Twitter
RT @chr1sa: One of the problems with autonomous car racing is that watching software drive is not a very exciting spectator sport. To help…
Apr 25
DIY Robocars via Twitter
RT @SmallpixelCar: Replaced AGX Xavier with @NVIDIAEmbedded Jetson Xavier NX. Now both cars look fast and clean https://t.co/jlcoY2EjZf
Apr 25
DIY Robocars via Twitter
RT @chr1sa: Yesterday we had a record number of competitors (36) in our monthly @DIYRobocars virtual autonomous car race. We kept it going…
Apr 25
DIY Robocars via Twitter
Apr 16
DIY Drones via Twitter
RT @chr1sa: After more than a year of only virtual races, @DIYRobocars returns to the newly renovated @circuitlaunch on May 22 for the resu…
Apr 11
DIY Robocars via Twitter
RT @DAVGtech: And now available with LiDAR 🔥 https://twitter.com/Heavy02011/status/1381137016381964293
Apr 11
DIY Robocars via Twitter
RT @Heavy02011: #VirtualRaceLeague: @DIYRobocars Race #9 - #ParkingLotNerds #thunderhillracetrack, CA Join us for the next race April 24th,…
Apr 11
DIY Robocars via Twitter
RT @DWalmroth: Weather's finally cooperating, looking forward to racing 1:10 scale autonomous cars outdoors again! @diyrobocars, @NVIDIAEm…
Apr 7
DIY Robocars via Twitter
RT @AIDRI_: I finally succeeded in optimizing the trajectory and speed of a car on a #racetrack. Next step: implement a 2d controller and…
Apr 7
DIY Robocars via Twitter
@jetdillo @circuitlaunch Actually the second *in person* event in a year. We do virtual races every month
Apr 2
DIY Robocars via Twitter
Update: we're moving it back one day to Sunday (the 4th) at 11:00am instead
Apr 2
More…