Very strange pan balance issue. Help required!

Discussion in 'MōVI M15' started by Ben Platts, May 25, 2016.

  1. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    After upgrading to Akira we were getting really high gain settings on all 3 axis using auto tune. Around 280 on pan with an Alexa mini on the Movi. After changing absolutely nothing we went to do an auto tune on a test day and got a pan gain of around 130 so did some manual testing. It now seems that we get oscillations on the pan axis but only at a very specific point when the handle bars are situated from front to back, so horizontally across the Movi. See the attached video. It is only the position of the bars that causes the oscillations. If you undo the toad in the hole, rotate it slightly then re tighten it is still only caused by the angle of the bars and not the angle of the motor. We've checked everything for tightness, taken the toad in the hole off and on again to check for tightness. Tried it under the Alta. Reinstalled Akira. As we always use the M15 under our multi rotor I always balance every axis to a very carefully. All 3 axis are perfectly balanced. I've been balancing gimbals for 5 years from the cinestar so know the balance is good. The only time we don't get it is when you directly hold onto the toad in the hole mount with your fingers with no bars or mounting plate attached. Its a big drop in gains, literally halved with these constant, very specific oscillations and can't seem to find any explanation. Any ideas??!
     
  2. Gary Haynes

    Gary Haynes Administrator
    Staff Member

    Joined:
    Jul 6, 2012
    Messages:
    5,211
    Likes Received:
    460
    Ben, in the expert menu what are the settings you have for Pant/Tilt/Roll strength, Gyro Filter and Output Filter?
     
  3. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    Hmm. That is interesting, Ben. My mental model of how the MōVI's work is that they are a complex series of masses in motion, each configuration having one or more resonant frequencies (harmonics of a lower natural frequency).

    My first guess -- and it is only a guess -- is that you've discovered specific mechanical configurations where there are peculiar resonant frequencies depending on the exact masses in motion, moment arms, (and possibly the phase of the moon :rolleyes:) What leads me to think this is when you wrote, "The only time we don't get it is when you directly hold onto the toad in the hole mount with your fingers with no bars or mounting plate attached." That's when your hand/arm/body is effectively acting as a nice squishy "meat mass dampener."

    That said, if you could save the current configuration and post the file it would both answer Gary's questions and allow us to look at the entire configuration.

    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  4. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Hi Guys. Here's the config file.Its just very strange as we didn't change any settings apart from doing an auto tune then it started happening. We had our Alexa mini on the M15 working perfectly with Akira. Switched to a Red Dragon for a shoot with everything working perfectly, then switched back to the Alexa mini and then this issue started. Its such a specific point that the oscillations start and stop. Its like activating a kill switch by moving the bars a couple millimetres either way! And the fact that its only the position of the handle bars that causes it is very confusing!
     

    Attached Files:

  5. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    As I recall, the MōVI "remembers" where "front" is -- so when you remount the handles, you need to make sure that they handles are left/right with respect to the MōVI's front. I admit, I'm speculating furiously here -- but just curious as to whether you've been maintaining the relative orientation between handles and MōVI?

    I'll have a look at the .cfg file.
    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  6. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Just a heads up. This file was mid testing. All the settings are as they normally are but Tilt is usually on around 135, Roll on 170 and Pan on 280. This is what we were getting before with Akira and the Alexa mini. With auto tune now its the same on Roll and Tilt but Pan has gone down to about 180. We can manually tune it up to about 210 without oscillations but thats the max.
     
  7. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    I note that you're using Autotune Percentage of 85% -- going from memory, that seems significantly higher that the recommended value of 70%.
    Did you select 85% by experimentation?

    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  8. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    I haven't been maintaining any orientation but if you release the quick release on the toad in the hole and rotate it to any angle and re clamp it so the bars are in a different orientation the oscillations still only happen when the bars are in that front to back position. The motor direction seems to have no effect. The only element which is the same when the oscillations happen is the angle of the handle bars.
     
  9. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    OK. That suggests that it's a resonance-related phenomenon, I suspect.

    Andy.

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  10. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Yeah. We've been using that percentage for about 4 month without any issues. Using that percentage with Akira we've been getting fantastic results up until now. Very high gains and super smooth, locked in footage.
     
  11. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    Oy. Well, another brilliant diagnosis of mine crashes and burns. :mad:

    Unless other folks have any better ideas, I'd suggest opening a support ticket with FF (email to support@freeflysystems.com) -- save yourself some typing by referencing this thread and .cfg file.

    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  12. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Yeah thought that might be the case!
     
  13. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    Please post again when this gets resolved, Ben -- be interesting to know what the issue was.
    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  14. Katya Nelhams-Wright

    Joined:
    Jul 8, 2012
    Messages:
    216
    Likes Received:
    9
    Hi Ben
    Try using the data logging function on the MOVI as it can give you an idea of what exactly is going on on each axis and how the motors are responding, how good your GPS is etc. You just load it into excel like the ALTA log files.
    Kat
     
  15. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    Thats a good idea, Katya! It will at least tell us what the MōVI is doing, if not why.

    Andy.

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  16. Graham Futerfas

    Graham Futerfas Well-Known Member

    Joined:
    Oct 19, 2015
    Messages:
    1,208
    Likes Received:
    234
    Hi Ben, have you tried adjusting any of your Output and Gyro filters? If it's a Pan vibration, I'd look at the Pan Output Filter in Expert Mode.

    Page 20 of the Akira Beta Manual provides great information about reducing High and Low-Frequency oscillation.

    For High-Frequency buzzing or vibration, which is what it sounds like in your video, try increasing Filter values. That means, Output Filter value needs to stay at 1, but then you can increase the Output Pan Filter. If overall Output Filter is greater than 1, then it overrides the individual Pan, Tilt, and Roll values.

    This may or may not solve your problem, but it's where I'd look in the settings. Or reduce your Pan Stiffness a bit, but it sounds like you're already lower than you want.

    For convenience, here's the Akira User Manual: http://d27w8s76bjlz9v.cloudfront.net/wp-content/uploads/2016/02/08181024/Akira-Beta-Manual.pdf
     
    Wolf Schiebel likes this.
  17. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Good plan. I'll give that a go and let you know if it does anything. Speaking to freefly support at the moment too.
     
  18. Ben Platts

    Ben Platts Active Member

    Joined:
    Jan 24, 2013
    Messages:
    186
    Likes Received:
    25
    Graham has sorted it! Output filter was set to 5. Put that back to 1 then the oscillations went away. Have now got the output filters set with pan on 10, roll on 3 and tilt on 1. Gains are pan 280, roll 165 and tilt 187. Haven't flown it yet but all seems good on the ground. Cheers Graham!
     
    Graham Futerfas likes this.
  19. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,351
    Likes Received:
    1,160
    Bonus points for Mr. Futerfas! :)

    Andy

    Forensic Software & sUAV / Drone Analyst : Photographer : Videographer : Pilot (Portland, Oregon, USA): Trees=2, Ground=1, Props=11. :(
    The Ground Is The Limit™
    ---------- Forensic Drone Analyst : Forensic sUAV Analyst : Forensic Unmanned Aircraft Analyst : Forensic Drone Expert
     
  20. Derek Cooper

    Derek Cooper Active Member

    Joined:
    Apr 28, 2014
    Messages:
    527
    Likes Received:
    34
    Ben - how did it fly? Please report. We tried Akira but had dead pan motors - couldn't get it resolved. Went back to 4.02.
     

Share This Page