EN CN
Some strange behaviour
  • Garug
    Introduction:

    I have 3 modes configured; Follow mode, normal mode with 'Estimate frame angles from motors' and FC mode. I also have a Joystick that is connected over SumPPM and this 'Joystick' also provides the FC_PITCH and FC_ROLL over SumPPM.

    Camera Canon 5DII currently with 24-105, well balanced at 50 mm (and all tests performed at this setting.) All motors are GB85

    PID (FW 2.4) around P: 25 +/-5, I: 0.1 +/- 0.5, D: 30 +/-10 for all axis, Power; ROLL: 150 +/- 15, PITCH: 120 +/-15, YAW: 170 +/- 20 (4S battery)

    Yaw motor gets hand warm, other colder. More power could be possible, but makes it more prone to different kind of vibrations.

    I currently use Follow Yaw and Follow Pitch, Roll: 0,0

    All RC control min and max angles are at their maximum values.

    no I2C errors. Gimbal well balanced and solid. IMU calibrated.

    The FC mode I use only as backup, when the gimbal looses control if camera did touch something etc. It is great for that. It always returns the level in couple of seconds.

    The strange behaviour:

    1. I start in Follow mode, the Yaw is always of some 5 to 10 degrees, I have tried Yaw offset, it does not help. So after start I change to normal mode, set the handle straight and go back to Follow mode. This works. (Using Joystick Yaw (or Roll) control in Follow mode causes all kind of weird behaviour, so I do not use it in Follow mode.

    2. When using the Joystick Pitch control on Follow mode, there often seems to be some connection to Yaw, i.e. Pitch control causes maybe +/- 3 degree Yaw movement. Performing Follow mode auto offset seems to help on this, but it is not consistent, next time this behaviour can be again there.

    3. If the gimbal looses control in normal mode, going to FC mode stabilises and levels the gimbal, but if going now back to normal mode it looses control again. But if going from FC mode to Follow mode it is in control and now if going in Normal mode, it remains in control ?? The same if it looses control in Follow mode, first I need to go to FC mode, then to normal mode and then back to Follow mode. Sometimes it works going from FC mode back to the previous mode where control was lost, but usually not. Cycling all the modes works almost always.

    4. Some times when going from any mode to Follow mode, the Roll tilts more than 45 degrees. Going out of the Follow mode and coming back corrects this.

    5. I had PID + Power values:

    ROLL: 30, 0.11, 30, 150
    PITCH: 30, 0.15, 30, 130
    YAW: 30, 0.1, 40, 170

    And it was working nicely, but then something changed? I do not think it was the frame, I think something in software ? and the gimbal started vibrating if camera is tilted up about 45 degrees. I think this is bot Yaw and Roll vibration. This is the only orientation where it vibrates.

    Now I am at

    ROLL: 30, 0.09, 25, 150
    PITCH: 30, 0.09, 25, 130
    YAW: 30, 0.09, 25, 170

    and I still get some occasional vibrations.

    This could be partially at least the camera installation. I think I need to improve the bottom plate and camera installation. 5DII has really clumsy camera mounting point.

    Conclusion:

    Over all I am happy when it works, but I do not feel it is reliable. No matter how much tuning there is always suddenly something strange going on. It somehow feels it remembers old situations and it depends from what mode it came to what mode etc. Often I just do complete restart, but not even this seems to be fully consistent.
  • Garug
    5. I did my best to reinstall the camera as firmly as possible and it improved the situation. I need to rethink the camera installation, now I am using a 3 mm Duraluminium plate. It just is not solid enough and the camera mounting screw is not really good enough for firmly installing the camera to the gimbal. I suppose a gage around the camera and supporting also from the flash shoe would be the best...
  • kikojiu
    I didn´t understand which version of firmware you are using 2.3? 2.4? Anyway the main power is too low for a 5D with 24-105mm. Roll should be 200, Pitch 175, and Yaw 200 too. The auto offset is not a rule. Use auto offset but Yaw has to be done by hand. Trial and error until you get it right. Still you might have to power cycle it to see if you got it totally right. Your profiles are too complicated. You should use Profile 1 to stabilize and with only follow yaw mode. P2 to follow pitch roll and i.e. 3 to be completely not following. When you have some kind of interference in P2 or P3 always put it back to P1 and it will be back to normal. Still you might have to power cycle. Until we have a 32bit board this will happen, but mine are all very reliable with my own setups.
  • Garug
    I use FW 2.4. I will give a try for the higher power settings, though my past experience is that GB85 becomes somewhat jerky especially when using joystick when power is high. (maybe I need to make the bottom plate stronger first).

    The FC mode (when used handheld) is useful only for stabilising a upset gimbal, but it is very good for that, no mater how upset the gimbal is it stabilises it. I am really waiting for the 2 IMU boards.

    I hope there would be possibility for more profiles than 3 (over SumPPM). Maybe not needed in actual use, but would be good when testing different things. I think I will not use Roll follow much on real use, it is something I have been just testing (it could have some special use for creating smooth 'flying' videos handheld.)

    I guess I have to play more with the offset values too. Slowly but certainly I am starting to understand how the FW 2.4 works and is set up, I only hope the next FW will not be too much different.

Add a Comment