Stop Gimbal Rolling Slightly when in Follow Mode
  • When I am in Follow mode and turn the gimbal slowly around a corner or 180 degrees keeping pitch neutral, the Roll Axis begins to follow the direction, if yaw follow is right, gimbal slowly rolls to the right perhaps 5-10 degrees and then levels back off as I move forward. I tried turning up PID and power but only introduce oscillations, roll persists. I am using both YAW and PITCH follow with work great.
  • Here is a video test I made in Follow mode to demonstrate this effect. As I increase yaw rotation, roll axis rolls in direction of turn. It's not a bad effect, but one I'd like to control in some way. Is this software or a PID problem on my rig?

  • I have the same problem.
  • I have also this problem in normal mode!
    When I Pan gimbal roll about 30 degree and slowly come back !
  • I was able to correct for this in 2.2b. I kept adjusting the settings, deadband, expo, PID and it went away. I was not sure what I did to stop it. Sorry that probably does not help, but I did get it to stop.
  • When I Yaw (pan mount camera) , the Roll axis is fail! Camera roll right or left about 15 degree and remain slant!
    I calibrate gyro and accelerometer but nod leveling!
    What I do?
  • Hello everyone,

    I made my handle gimbal 3 axi few weeks ago.
    My current setup:

    - 2 Alexmos Controller V1.0 Firmware: 2.30 b5 (One for the pitch , the other for Roll & Yaw )
    - Lipo 4S / 2200mAh
    - 3 iFlight GBM 8017
    - Canon 24-105 f4 lens with 5DmkIII

    I do my tests at 24mm lens .
    First, my gimbal is perfectly balanced on all 3 axis. In any position is the gimbal held .

    My problem is in follow mode, when I throw the gimbal down or up, the pitch makes the job correctly but the Roll starts to vibrate .

    Is this a problem of torque on the Roll and Yaw ? Lack - til power ?

    My PID settings are as follows:

    Pitch: 11 / 0.2 / 12 / Power : 70/ 65 GyroTrust : No Gyro High Sensitivity / Limit Acceleration : 500 dg/sec2

    Roll : 7 / 0.25 / 13 / Power 210/ 60 Gyro Trust / No Gyro High Sensitivity / LimitAcceleration : 500 dg/sec2

    Yaw : 8 / 0,26 / 13 / Power 210/ 60 Gyro Trust / No Gyro High Sensitivity / LimitAcceleration : 500 dg/sec2

    If someone can enlighten me on a mistake I have made I will be delighted .

    Thank you in advance.

  • Hi Tristan:

    Go to the follow mode angle on the roll and set them to 90 degrees meaning Locked Roll. That is the solution. Let me know how it goes.
  • Hi kikoju!

    Thanks very much! It's working perfeclty now!
    I set the Follow start, deg to 89 and te Follow Roll mix, deg to 1.

    Thanks a lot!

  • You´re welcome Tristan. But you´ll have to go there again because there is a bug and you´ll have to do that again. Apparently the board looses some profile settings...
  • Ok! I'll see.

    The other problem is when I run, or walk fast, I've some vibrations.

    I think it's about my PID settings:



    What do you think about?
    I 've seen your video with the girl walking and runing near the sea.
    That's great!

    That's my gimbal

  • D too high and probably P too. And lower the I drastically not more than 0.2 and yaw around 0.08. Even if it looks slow it is not. It is smooth not slow. :-) It will get better with lower values and what is your gyro thrust value?
    Please uncheck skip gyro calibration on start up. It will help doing the gyro calibration on start up. Just leave the gimbal on the ground completely still and turn the power on and wait until it wakes up...
    And dom´t limit acceleration.
  • OK!
    I've tried with low D an P.
    My Gyro Thrust value is 100. I think a value between 60 and 80 is better. Isn't it?
  • I use 110. Your problem is only on the pitch axis. Work on it lowering the D or P or both and the I too. It looks to me the pitch is oscillating...
  • BTW I would lower the I on your roll not to more than .20 to .25 and pitch lower at around .1-.15
  • OK! Thank you for your advices, I'll try this afternoon with a new settings and I 'll show a feedback!
  • hi, so I reduced the PID values ​​and another problem is appeared on the Yaw, once all the settings were set on, we try out and we saw the rotation of the Pan we lose the horizon roll. I do not know what is the problem.
  • Tristan that is just a matter of getting the right PID´s. Amazingly people that get it right don´t share...Anyway don´t copy others PID. See this link and try to make use of it....

    Good luck
  • kikojiu thank you for posting that video from a while back. You for sure have a very good understanding of SimpleBGC and AlexMos. I was looking for this for a long time. As understanding the correlation between PID and Realtime was always hard to understand, this shows it very well at 6:30. Can I post this on my site www.jodagimbal.com?
  • Sure dbeaty do it...Just make sure people give their contribution to others after that...Most people use what we share and then vanish...
  • Thanks for the nice tutorial but what if you You increase I and the gimbal reacts slower?
  • That shouldn´t be possible in theory unless P is very low. The reaction time is not very important. Whe I setup my gimbal even running the pitch has an I of around 0.09-0.1-0.12. It is more than enough and the only way to fly the gimbal steady...what we see as reaction time going back to zero when on the stand is just a rough idea of how the motors are reacting mainly on P and D. I always set a rough value and then fly it or handheld for fine tuning...
  • tristanha what about you? Is your gimbal working ok now?
  • Hi Kikoju,
    I have not had time to make new settings these days, but tonight I'm shooting a reception at the Palais de Tokyo in Paris.

    I do not need to run or make quick movements, so I'll put the best settings I found and study the video you posted.

    I will return tomorrow!

    Thank you for sharing.
  • Hi All:

    Could you please watch this video and tell me what you think about the stabilisation? Forget the rest focus, aperture etc...it was shot to see how much unbalancing the motors would keep it stabilising.

    Thank you

  • Hi,

    About your video, I noticed that there was some vibration on the roll, and walking up the stairs is not as smooth as the other.
    There is also a slight angle in the corridor in the rapid shift rotations.

    The rest is pretty clean. I'm pretty close to the same results with my gimbal.

    Yesterday I shot, and the problem of deflection roll when turning the pan 90 degrees appeared.

    I'm 2.3b5 with 2 cards Alemox.

    I have this problem since only 2 days and strangely since I change the position of the IMU Roll.

    Before the IMU was placed under the bar Roll, like the pictures:

    Then I bought protection IMU in AerielPixels:


    I change the position of the IMU and keeping on top of the roll bar and since I have this problem deflection Roll, with or without follow mode on Pan.

    I have to test today by finding the initial position of my IMU and I'll let you know.
  • I'd have to see how unbalanced it is, someone should shoot that first. Then cut to the example. Like was the camera rolling upside down before you turned it on?
  • It's not too bad, looks like ours.
  • dbeaty it was completely off balance. rolling to the left and tilt to the front...so I am pretty happy. I repeated outdoors today with the gimbal balanced and it was pretty smooth. :-)
  • tristanha it is just PID´s...not the IMU or 2nd board.
  • I thought it was related to deadband and expo but today I did some testing. 2.2b, if I am in pitch follow and turn right 360 degrees, the faster I turn the more roll to the right. If I turn left the same speed, less roll the left. If I turn off follow pitch and turn on follow yaw, and repeat NO roll in either direction when turning 360...
  • Same problem and try so many things, but finally I replace sensor upside down and do Z to -Z, works perfectly. It might help , but not sure for you. I appreciate another way to solve this (when yaw follow, roll is change).
  • It's all about sensor not aligned and I haven't do ACC CALI. After I do proper ACC CALI according to the GUI instructions , all OK. Nice! and Happy.
  • After pulling my hair out with this same issue for weeks....what NYANA has said above worked instantly to fix the problem!!!!
    Its now rock solid when yawing......NYANA I wish I could buy you a beer!!!
    THANK YOU!!!
  • Hi there,

    I´m also experiencing this YAW - ROLL problem.

    Makes no difference if the Copter yaws in flight, or I simply yaw the gimbal with the RC. If I activate or deactivate the follow mode, or set follow start and mix to 90 and 0 ...
    ROLL is always following YAW

    Drives me nuts!
    Any ideas?
  • Are you using estimate angles from motors. if not, try it.

    Also carefully perform follow offset auto.

    Make sure also that joystick is calibrated and enough dead band.

    It is very important that each axis is 90 degrees to each other and IMU installed correctly. If not, it causes the problem you are having. Is there a picture of the IMU installation?
    I dont know how to properly explain this but Here is what i found for now, while checking the real time data for the roll i noticed that the Gyro_X remains leveled in the center line where as the ACC_X goes out of the center when i YAW. the ACC_X however comes back to the center line eventually aligning the roll of the gimbal but after a long delay. When i YAW clockwise, the blue line (ACC_X) goes "above" the Red line (GYRO_X) & When i YAW anti clockwise, the blue line (ACC_X) goes "below" the Red line (GYRO_X). Seems like the controller is noticing this improper data but not doing anything to rectify it.
    Please check the below pictures of the Real time data.
  • You can tune the interaction between ACC and Gyro using gyro trust.

    It is important that when the gyro is calibrated, it is totally stationary (at floor, not moving at all). IMU position does not mater, but it must be stationary. I skip the gyro calibration at startup and do it when needed.

    I also advice starting in non follow mode profile and changing to follow mode profile after startup. Not so important with 32 bit boards and 2 IMUs, but remember that with 8 bit board I always did it like that to avoid problems. (The gimbal gets easily confused if it is forced in follow mode.)
  • @Garug
    Thanks for showing interest in helping me out.

    Are you using estimate angles from motors. if not, try it.
    Ans) At he the moment it is deactivated but i did try that not long ago which was of no help.

    Also carefully perform follow offset auto.
    Ans) Will try this now

    Make sure also that joystick is calibrated and enough dead band.
    Ans) Honestly i don't know how to calibrate the Joystick

    It is very important that each axis is 90 degrees to each other and IMU installed correctly. If not, it causes the problem you are having. Is there a picture of the IMU installation?
    Ans) I have tried my best to get all the Axises 90 degrees to each other & Surely i don't think the balancing is the problem as i have precisely balanced all the three axis to the best.

    Here are the pics of IMU Mount, P.S. I did try using Nylon nuts & bolts to mount the IMU but made no difference in performance.

  • IMU installation looks ok. Joystick calibration RS subtrim auto and set deadband to 30, you can lower it when everything works.
  • I just did RC subtrim auto without doing any changes to the deadband setting (it remains 15), well i think it seems to have worked, the terrible & unnecessary roll movement has stopped while doing YAW movement. What do you advise Garug, should i play around with the deadband setting?

    However I would like to deeply appreciate your help. You have really helped me out with this precious piece of information. I just cant believe that this was all due to "RC subtrim auto", this had got me head scratching for weeks now.
  • Deadband basically just is how much you need to move the joystick before it moves the gimbal. If the joystick is working ok and not moving the gimbal when it is not touched, then the Deadband is ok.

  • Note: on RC tab you might want still to verify that functionalities that are not assigned for anything (like probably RC Roll) do not have any unconnected input assigned for them.
  • The joystick seems to be working fine so i think i wont be messing with the deadband settings unless necessary. Your point noted, In the RC tab i have made sure that nothing is assigned for ROLL motion, whereas the YAW & the pitch are the only ones that are activated. Thanks again GARUG, you are a Great person or else very less people who have gotten it right are willing to share their knowledge these days.