Tilt Radian acting stangely - ideas needed

Discussion in 'Radian' started by Richard Critchlow, Nov 2, 2013.

  1. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Hi All,
    I have a problem balancing our 3 axis photohigher AV200. The camera is as physically well balanced as I can get it. The roll and pan Radians seem to be performing as expected. The tilt Radian/servo is a different matter.

    Whenever it's initialised, it makes a sudden pan down. If the gain is set high (80% or above) and I physically hold the camera on initialisation it starts working but oscillates terribly. So I reduce the gain to get to the sweet spot which seems to be around 25-30%. However, if I switch the gimbal off and on again, it does the "usual" tilt down followed by a very dramatic tilt down to beyond the end stop causing the gearing in the servo to make horrible noises! The only solution is to switch it off in a hurry!

    A further bit of background - our system was delivered RTF as a 2 axis system in error (tilt, roll no pan). The tilt was doing this when we collected the machine but I managed to iron it out then (possibly more luck than judgement) We were then sent the additional Radian which I have fitted myself. I also realised that it had all been wired from the AR8000 rather than directly from the Spektrum Satellite. So I re-wired that which is a much better solution. All seemed to be working fine. Then the tilt problem started again and I can't get rid of it this time.

    John Cunningham is being really helpful with ideas to try but I'm wondering if anybody here has experienced the same problem or can suggest a logical way to work out the source of the issue. Why would it want to tilt down quite so dramatically - but only when the gain is low?????

    All thoughts very gratefully received
    Regards
    Richard
     
  2. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Just seen Tabb's post about info needed to solve problems. That makes sense and I'll get that done today.
     
  3. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Further to my earlier post. Here's the detail Tabb requested including video of the problem. In response to those who suggest changing to Freefly servos, would one servo v another really cause such an extreme problem. I can appreciate some may be better than others but I have two working fine and one doing very strange things - not just not great if that makes sense?

    Aircraft Type: Skyjib X4
    Gimbal Type: Photohigher AV200
    Radian Type ( 1, 2 , 3 axis) 3 axis
    Gimbal Power source: 4S 2600mah
    Gimbal Radio Type: DX8i
    Radian Communication Protocol (S.Bus, PPM, PWM, Spektrum) Spektrum

    Connections: Change these to match your setup
    Roll Sensor:
    Port
    1. Power from AR8000
    2.
    3. To Tilt sensor
    4. To Roll Servo
    Also input from Spektrum Satellite

    Tilt Sensor:
    Port
    1.From Roll Sensor
    2.
    3.To Pan Sensor
    4.To Tilt servo

    Pan Sensor:
    Port
    1.From Tilt sensor
    2.
    3.
    4.To pan servo

    Symptoms: See video for full explanation


    Steps taken to resolve:Wiring checked. Settings checked. Advice sought!

    Problem is shown in this video:
    http://play.webvideocore.net/poppla...on=#FFFFFF&colorHighlight=#fcad37&direct=true

    Radian Setting are shown in this video:
    http://play.webvideocore.net/poppla...on=#FFFFFF&colorHighlight=#fcad37&direct=true

     
  4. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,364
    Likes Received:
    1,162
    Richard:
    Do I infer correctly that you're now working directly with FF to resolve this? (The reference to email from Tabb suggests that you are.)

    Cheers
    Andy.
     
  5. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Hi Andy,
    No. When I originally wrote email, I meant post in this forum, quite understandably requesting the details I've now given above. All thoughts gratefully received!
    R
     
  6. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,364
    Likes Received:
    1,162
    Hi Richard:
    Hmmm. I recall reading your other thread that you're using Savox servos, right? My concern being that the bizarre tilt down might be the effect of Savox -- or is this happening with FF servos?

    Cheers
    Andy.
     
  7. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Hi Andy,
    Yes they are Savox servos. I can understand one servo might be better/smoother/more accurate (and that's certainly something I need to consider) than another but could one really cause such a dramatic issue result?
    R
     
  8. Brad Meier

    Brad Meier Active Member
    Staff Member

    Joined:
    Aug 14, 2012
    Messages:
    749
    Likes Received:
    66
    Normal order of radian setup is usually pan to roll to tilt. I see you have roll to tilt to pan. This may be an issue as I've never attempted wiring this way.
     
  9. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,364
    Likes Received:
    1,162
    The conventional wisdom seems to be that you have to use FF servos. I agree with you though -- I don't see why Savox would trigger such extreme differences. I don't know whether there are other forum members using Savox with Radians, but, if there are, I'm sure they'll chip in the comments....

    Cheers
    Andy.
     
    Richard Critchlow likes this.
  10. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    That's something I'll look at tomorrow, but the theory suggests that the order should make no difference. The order you describe is the logical order on a cinestar gimbal. Mine was already wired so I just adapted the wiring for the third Radian. They all know their role and seem to be acting as expected - apart from tilt.
     
  11. Richard Critchlow

    Richard Critchlow New Member

    Joined:
    Apr 30, 2013
    Messages:
    21
    Likes Received:
    0
    Hi All,
    Following a lot of hard work, frustration and invaluable assistance over the weekend the problem is sorted (or at least it will be once I've got another potentiometer) I hope the solution will help others to better understand how these things work. The issue was the resistors that had been fitted to our gimbal where incorrect. I now understand that the 'Servo Out' indicator in the Radian Software needs to be at zero. Ours were both at +55% (Tilt and Roll) but pan was at 0. I had no idea either what that meant or that this would cause a problem. Apparently, that figure shows how much work the servo is having to do to hold the axis steady. If it's too far off centre, it effectively runs out of travel and can't work properly. I think that we may have got away with it with a lighter camera. That's my lay person's understanding of the problem.

    The solution was to borrow a potentiometer out of a servo, attach this to the servo and tune the 'servo out' to get it to zero i.e. the servo is effectively in it's centre position when the system is level. Having done that, the tilt now initialises without a single twitch and is so much smoother than it has been! We just need a second pot to do the same to the roll and I think we'll be there.

    I can't find any reference to this stuff in any manual or am I wrong? Maybe the FF servos and resistors come pre-matched but given that they are sold as seperate units, it might be helpful to make clear what the resistors are actually doing and what you need to try to achieve with them.

    All credit for this solution has to go to John Cunningham who has been kind enough to respond to all my emails this weekend and spent an hour on Skype with me yesterday to diagnose the problem and suggest this solution but I thoughlt it hegely important the solution is shared on here. He's been a real star

    Regards
    Richard
     
  12. Gary Haynes

    Gary Haynes Administrator
    Staff Member

    Joined:
    Jul 6, 2012
    Messages:
    5,211
    Likes Received:
    460
    Richard the resistors come with each Radian. And they work flawlessly with Freefly servos. Sounds like the Savox servos need a different resistance and you were able to determine that with the old pots.
     
  13. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,364
    Likes Received:
    1,162
    The FF resistors can be measured with a Ohmmeter -- they are two 2.6K Ohm resistors with a center tap. I suspect they do precisely what Richard ended up doing using a potentiometer (yeah, the original "POTI" of RC fame) -- namely centering the servo.

    For more information, I found several relevant hits doing a Google search for:
    rc servo resistors​

    This one seems to confirm my hypothesis: http://www.seattlerobotics.org/guide/servohack.html
    which states, in part:

    The theory behind this hack is to make the servo think that the output shaft is always at the 90 degree mark. This is done by removing the feedback sensor, and replacing it with an equivalent circuit that creates the same readings as the sensor being at 90 degrees. Thus, giving it the signal for 0 degrees will cause the motor to turn on full speed in one direction. The signal for 180 degrees will cause the motor to go the other direction. Since the feedback from the output shaft is disconnected, the servo will continue in the appropriate direction as long as the signal remains.​

    It uses 2.2K ohm resistors, but what's 0.4K ohms between friends, eh? :)

    Please correct my understand if this is wrong. I have a ghastly reputation for often being wrong but rarely sounding uncertain. :rolleyes:

    Andy.
     

Share This Page