EN RU CN
Roll Drift from zero to fully canted over time 32bit 3 axis
  • Hello, thanks for all the info provided over the last years ... I've been perusing posts hereon quite a lot!

    I have a brand new SteadyMaker Tank Plus, which I got for my Sony X70. I chose this because I got a confirmation that my camera was within the weight range and physically cleared the frame. My only complaint about it is that structurally it does flex a tiny bit when picked up by the handles. (the top connection to the rods is less substantial than I would have hoped)

    I feel that I have it adequately balanced, and it generally works after careful 6 axis calibration, but within a few power ups, and attempts/tests, the roll axis starts to drift more and more. It is exaggerated by turning myself and the rig right or left or simply yawing the rig left or right ... the horizon dips to the opposite side. (Yaw left = tilt to the right) Then it sloooowly starts to creep to the left. It just gets worse and worse.

    Performance seems pretty good for the first run .. I can tilt the frame in any direction quite a lot and don't get vibrations or shaking errors.

    I have noted zero 12C errors.

    I just performed a very careful 6 axis cal (with IMU's in gimbal .. removing them would be very hard!) and the first run and second run seemed good. Here is the second run clip (this one has camera stabilization turned on, which helps eliminate some slight vibration I have noticed, but not dealt with yet) https://www.dropbox.com/s/aqs0vujbedste5v/Clip0071.mov?dl=0

    Seems OK, right? ( I bumped the gimbal descending the stairs and again after coming back up the stairs)

    Well not just a few moments ago, I turned the rig back on, and tried again .. now, it is starting to droop. Only a few degrees, but I know from experience that it will continue to get worse, and NO settings changes will have any effect except for a full calibration.

    Note: I have not done a temp cal yet ... I need to, I know, as it is winter, but this happens inside, at the same temperature as it was calibrated.

    Info: 3 axis rig, using SimpleBGC 2.50 b3, with (I hope) the latest 3-axis AlexMos board. Here is a link to my current settings:: https://www.dropbox.com/s/wgeqvtyfmj6xlrg/Stock.profile mod before auto attempt.profile?dl=0

    I really appreciate any help.

    Thanks,

    Gene
  • There is two IMU calibrations. the accelerometers and gyro. Gyro calibration is performed at each startup unless checked off. During gyro calibration the gimbal must be still, any movement during it will cause problems.

    Your problems could be also because of joystick or joystick calibration, or something wrong on the basic setup.
  • Thanks for your response.

    The calibration I am talking about that fixes the issue is the accelerometer cal. When it starts drifting, the ACC cal is the only thing that fixes it.

    I have the checkbox for the Gyro cal OFF at startup. I am going to be using this device on a boat, and there is never "still" on a boat. I plan to take the rig to shore every chance I get in order to calibrate gyro, but it needs to last a day or three at least.

    Funny you mention joystick ... how do you calibrate joystick? I do not see that anywhere. This is a wireless joystick setup, I do not like how the joystick operates, but I do have sufficient deadband, and I do have expo curve set on it. How does a funky joystick calibration affect the horizon, if the joystick is only set to pitch and yaw?

    Thanks again!

  • Also, I wanted to mention that performing a gyro cal manually (yes, it is very still!) does no good once the roll error starts .. it just accumulates and gets worse until Acc. cal. I have the unit outside in my cold unheated shop now ... in an hour or so I will do cold temp Acc. cal, then bring in to "mud room" of house which is a little warmer, repeat for mid-temp cal, then again even later inside house.
  • Try the joystick calibration first, it is at bottom of RC tab.
  • Thanks ... already checked. Even with joystick disabled, it is a problem with temperature. I cannot go through a 20 minute Acc. calibration every time I move ((In Alaska, temperature changes a lot .. very cold, medium cold, not cold, and hot)

    Looks like temperature calibration does not work? I got the cold calibration icons all lit up, but just in now from trying a medium temperature ... it beeps when I push button, but no check mark, and doesn't make the middle thermometer icon light up/no check mark.
  • Yes, if you have high temperature changes, it is best to do the temperature calibration.

    Difficult to say why it does not work on your configuration.

    1. read the users manual and ensue you are performing it right.

    2. ensue your FW and GUI version is the same.

    3. if above does not help, save your profiles (good to do anyway), and save EEPROM. after erase EEPROM and perform setup again. After IMU calibrations you could load the profiles (but not EEPROM) and the gimbal should work. However I suspect there might be something wrong on basic setup also. best to perform it also from scratch and ensure it is well performed, to get inverted status, set motor poles etc.
  • Thanks for trying to help .. I feel like crying I'm so frustrated ... hours of work today and no good.

    According to manual temp compensation needs 10 degrees difference ... I had 17, acclimated for well over an hour....it would just not register. I'm not understanding.

    I can get it close to happy/good right after a 6-axis cal, but it always drifts.. The design of this gimbal is such that I cannot remove the IMU's ... so hard to know if I am getting perfect alignment for the IMUs, especially the frame. I'm using bubble level cube on surface as close to IMU's as possible, but it's very hard to do.

    Is it even possible to expect this to work reliably on a boat where I won't have the ability to calibrate every day? Or is that unrealistic? I'd like to know soon, please.
  • Update: Temperature calibration values now writing at 27 degrees difference. This is too much, so I will ahve to start over when it gets colder outside again.

    Question: Why are these not calibrated from factory, so that, say, -50 to +150 degrees is operation range? Is there any way to manually write values?
  • The IMU is inexpensive one, it does not have memory to store the calibration. Calibrating the IMU, even at factory is time consuming and expensive. The better well calibrated IMUs cost 1000 euro + :(

    "now writing at 27 degrees difference. This is too much" Why is it too much? It is better to have it calibrated to wider temperature area than normally operated.

    How is the drift? is it constant drifting or certain angle it goes to? If it is constant drifting, then it is likely the gyro calibration that is not good.

    If it is constant angle error, ensure IMU installation. It must be accurate enough (90 degrees/parallel to axis) or all kind of problems will follow.

    I suggest using thin 2 sided tape to install the IMU. There is also other good ways to instal IMU, but thin 2 sided tape works good in most cases (assuming the IMU bottom is flat)

    See also follow tab, there is offset calibration.
  • Thank you Garug,

    It's a fresh day so a fresh start. After what I found last night I'm 96% convinced it is imperfect Acc calibration and very sensitive to temperature. Understood about the IMU cal costs ... I thought the data stored in the controller .. ? Anyway.

    Ah .. 27 degrees is too much, since that 27 degrees difference was found between my cold unheated shop outside (but is only 32F, rather than 15F which is more "normal" due to climate) and inside house, where wood heat keeps us quite warm .. however, it calibrated the inside house temp to the midrange, which, unless I heat the IMU with an oven of heat gun, is error, as it never gets hotter than that ... so I have to wait until it gets really cold weather again otherwise temperature cal is not working, as the high range is unable to attain.

    The drift is random, starting from zero, which is fresh standard Acc cal, operating normal, after some time, and a few power cycles, OR if left on for a longer time (10-20 minutes) it changes, and tends to droop to the left a little, then right, and as it gets worse it is very bad, getting worse and worse until camera is all the way flopped over. Maybe I should do a time lapse!

    In the SteadyMaker Tank, the camera IMU is secured in small indent under camera stage, also secured with tape. It looks accurately installed, but is very difficult to move - I fear breaking it, and the wiring is very involved, so very difficult to remove. The frame IMU is inside the CF box enclosure that also houses the controller board, under yaw, vertically oriented, behind much wiring stuffed into box. Later, I will take apart and take photos.

    Interesting, and this "might" be a clue: The Follow offset calibration does not correct the problem once it has started, but...BUT it almost seems that when I adjust offset for YAW, it changes PITCH.... again, it does not stay, just makes camera twitch , then returns to lopsided.



  • I suspect you have some setup problem, probably basic setup. What you are experiencing is not normal.

    ACC calibration problem should not cause drifting but constant angular error and axis interaction.

    It is also possible you have a faulty IMU.

    And some times erasing the EEPROM helps, did you dry that?
  • Hmmm ... would that possibly be the outputs not responding to follow as well? But they do operate properly with regard to the joystick, I think ... pushing up makes camera tilt down, pushing down makes camera tilt up, (although sometimes it seems to reverse!) but always pushing right or left makes camera yaw right or left, appropriately.

    How would I test for bad IMU? And which one?

    I have not tried erasing EEPROM ... I am scared.

  • Erasing EEPROM did not work out so well ... I realized I should have taken notes, so I reloaded it, and it was largely different than before.

    Interesting. Back to square 1.

  • Question: One thing I have noticed during all stages of this experiment is that the YAW axis responds to roll deflection. Note: The SteadyMaker Tank is sitting on its integrated stand, so the handles are free to rotate in the air. So when I would be trying to get smooth quick recovery on roll axis PID, I would also see the handle turn when I disturbed the roll axis. Is this indicative of a problem with setup?
  • OK< I got pretty fed up chasing this ... so I have partially disassembled the rig thinking I could browbeat my way to getting the IMU's out.

    First: Upon careful inspection I see the frame IMU is canted a few degrees (not sure how much) to one side relative to the frame vertical. Is this, combined with the inherent top rods to vertical axis flexing, sufficient to cause my problems?

    Second: Steadymaker obviously thought their installation was permanent .... hot glue securing everything. I HATE hot glue in electronic installation! Very shoddy. I guess you get what you pay for. This makes it very difficult to remove the IMU's.

    If I can get these IMU's out without destroying them, can I use 12c cables from my other gimbals to calibrate the IMU's? Or does the cable length enter into the calibration?

    Thanks.
  • Further update: Don't judge me too harshly - I've been gently removing and reassembling small electronics for years. The JST connector on the camera IMU came off with only the tiniest bit of pressure as I was trying to remove the 12c connector from it (the hot melt glue was loose, having been loosened with isopropyl alcohol). I'm trying to determine whether to try to make a needle extension for my soldering iron and resolder these joints. or just order a new IMU ... I understand a screwup here could fry the board. Also, I have to wonder .. were these joints on teh JST connector to IMU already bad (why it fell apart so easy? and if so could it have lead to my issues? Or would it just have show 12C errors?

    It's just getting deeper ... what I was afraid of.
  • Ground a soldering tip to needle point, and using reading glasses and magnifying glass, I think I fixed it. I added a dab of epoxy to three sides of connector on each IMU to prevent it from happening again.

    Had to de-pin the 12c cable end to remove from frame (motor through-hole is big enough, but Steadymaker frame hole too small). I will now enlarge holes in frame so I don't have to remove cable end each time.

    Then to devise a rock-solid, repeatable method to use for calibration.
  • Couple of degrees IMU installation error would case problems.

    If you saved profiles, you can always load them back, but I suggest doing the complete setup carefully according users manual. It could be that there was something wrong on your previous setup.

    See the realtime display to ensure IMU operation. There should be small noise (+-5 units max) on the ACC and Gyro readouts. on ACC the axis that is pointing down should be about 512, other axis about 0. when IMU is stationary, Gyro should be about 0. Note, there must be the noise, and the axis must react when IMU is moved.

    It is digital information that is transferred, the cable should just be working or not, but bad power connection could cause some problems, but should show as I2C errors. if no i2c errors, cable should be ok.

    Note, for temperature calibration you should not use heat gun. The temperature must change slowly, see the users manual. There is also possibility of overheating the IMU and that could cause problems. (Also using hot clue could cause problem because overheating.)
  • OK, well there were no errors before, but we'll see after I reinstall! Hopefully my soldering is good. I think I will put the soldered IMU in the frame position just in case ... (it came from the camera IMU position)

    I am trying to do according to manual, step by step, very careful, but that is why my gimbal is in pieces ..as 6-point calibration needs to be known very accurate!

    No, I appreciate truly NOT using heat gun on IMU! At best I can use freezer, but would rather use winter shop when it gets real cold again (soon?) then mud room for middle, and inside house for high temp.

    Would be nice if we could select which range and update only that range.

    Lots of work to do ...thanks for your patience, info and help!

  • Gimbal kinda working again, thought I did a good full calibration, but I guess not.

    Used -3 in RC Roll offset to level camera, but after messing around a while I could not adjust it any more and camera lopsided.

    I think I am going insane.

    Tomorrow is another day. I worked on this ALL DAY! My wife is so happy with me ... oh yes ... and to hear there's little hope makes her even happier.

  • Did you inspect the IMU data with the real time tab? does it look ok?

    I would disconnect frame IMU to narrow the things that could be wrong. It will work ok without frame IMU. When it does, connect the frame IMU.

  • I did inspect Realtime data ... with motors on, an active, but no deflection, there are tiny ripples, and I am zoomed all the way in. I am trying to figure out how to show you what I see, but not make a big video ... just show activity, and settings, and problem.

    I will try to disconnect frame IMU, and also swap the actual IMU over, so the one that is frame will be now camera IMU (I tried before but controller wanted them the same as before?)

    One thing that might be very important I noticed last night after I put gimbal back together - concerning balance. Now, I hear "perfect" balance is very important, and I thought I had perfect .. in that I could put camera in any position of pitch/roll/combination and it would stay.

    However, last night I noticed that now the roll does not stay ... pitch is fine, but roll stays up to about 20 degrees each way, then after that, it rolls back to level. If I roll cage upside down, then it stays much more.

    This tells me that the roll motor is loosened up ... it staying before was just young bearings maybe. It also tells me I have a problem ... if I move camera up on the pitch axis, even though it is still balanced on pitch, it is now top heavy and does not stay, but roll stays.

    So this tells me that I need to move the arms UP in relation to the center axis of the roll motor....but the design of this gimbal does not allow that.

    I will have to re-engineer the way it is connected.

    In the meantime, can you say whether this might cause the roll problems I was having? I cannot imagine how, but maybe ...
  • I think maybe instead of trying to re-engineer this, (poor design!) I will add mass to the top of the arms. Not as sophisticated as I'd like, but it will have to do (the arms are clamped to two rods, and cannot be adjusted to bring COG of camera/camera stage in line with roll motor axis.
  • Balance is important, but should not cause the problems you have. Even more important than balance is that all axis rotate freely (within the limits they move)

    But over all it is good that you are concerned about the mechanical aspects. the gimbal needs to be mechanically good and well balanced to work good.

    When inspecting the IMU data, you can put the motors off and rotate the camera with hand. also inspect the indicators at right side of the screen, that thy move logically when you move the camera.
  • OK, I was curious about that ... I suspected it might as small errors could accrue over time, as time seems to be the one thing that is known element as change occurs. I have some weight lightly taped to the top of the roll bar, and it does help balance... obviously will have to design something a little more sexy and adjustable! But I would like to be very careful about how I proceed, so I can say YES to every detail as I go through troubleshooting.

    Motors do move freely, and all axes are free and smooth.

    The IMU data seems to make sense .. the indicators on right of screen match the movement precisely.

    The Realtime data shows me deflection of axes, and a jiggly curve as they recover, and if left alone, with motors on, or with off, is very little motion detected.... I don't know if a flat straight line is even possible? I do have scrolling speed/magnification all the way up.

    I won't have time (or permission!) to spend all day on this, but will be studying and thinking and researching more.

    Thank you!
  • I had just enough time yesterday evening to remove the IMU from the frame, and test gimbal (without doing anything) as expected, it was squirrely ... so I did a gyro and single point ACC calibration, and it was still squirrely.

    Then I took the IMU that was in the frame position and put it into the camera IMU place, leaving still only one IMU in system, and again, it was squirrely, but ... BUT .. it was different, even after doing same calibrations.

    I thought that was interesting.

    SO I spent time working with settings, making sure all correct, and then playing with PID's on PItch (got it fairly smooth, actually) and roll with Follow turned off. It holds camera level until I move around too much, then oscillations start and get really bad.

    I am fairly suspicious of my PID's, as these are radically different than before, and the sensor indicators in GUI read correctly, and very low fluctuation, but I still am unsure of whether the IMU is good, and if this one IS good, is the other one bad?

    But will continue to work with this to see if I can get fairly stable, and go from there.

  • It would be good to see a video of your gimbal and problems with it.

    You mention follow mode. I would recommend not using follow before the gimbal is well working without it.
  • OK, no video yet, as I've stripped down and started over. Do these snips mean anything to you? The Fluctuation X and Y both show a little rhythmic bump ... I reset PID's to 10, .1, 10 (yaw is off, follow is off) and I made another snip to show fluctuations still. What does it mean? (the house is still..no movement or sound right now)

    How do I post jpg?

  • Here are screenshots of deflection. I have done basic re-cal again, and set PID's to lowest, and still there!

    https://flic.kr/s/aHskpjyZ6q
  • It is max zoom, how many units are those pums, if less than 5, probably nothing to worry. I will check later on how the gyro signal looks at max magnification. Is the gimbal totally still during the shots?
  • Yes, those pics were when gimbal still.

    How does "less than 5" equate? What is this unit? Is it what I see on left side of realtime display (changing numbers) for instance here are numbers I currently see: (gimbal is still, motors on, will include PID's in this post)

    This is with one IMU. Still trying to figure out how to determine if an IMU is bad. This IMU is the one from frame, put into camera position.

    Is possible board is bad?

    Have cleared EEprom, started over (again)

    Acc X -16 to -20
    Acc Y 1 to 5
    Acc Z 510 to 514
    Gyro X -5 to 2
    Gyro Y -5 to 7
    Gyro Z -1 to 2
    Err Roll -2 to 0
    Err Pitch -2 to 2
    Err Yaw -17 to 15

    Pwr Roll 145
    Pwr Pitch 106
    Pwr Yaw 127

    PID Settings https://flic.kr/s/aHsks1Xo4L

    INteresting that I've had to change RC roll offset twice to keep camera level.


  • Regarding calibration, I find that WHEN THE MOTORS ARE OFF, all gyro values should stay within the range of -3 to 3 when nothing is moving. If they don't stay in that range, calibrate them again. You don't have to be off very much on gyros to really thrash your performance. Accelerometer calibration seems to be a bit more forgiving.
  • You mean calibrate gyro again? I just checked with motors off, and I see some max deviation -4 to 3 ... mostly in the -1 to +1 range though. I don't know how to get much more still than this cor calibration ... no wind, music is soft, cats aren't jumping around, ...
  • Also I find this strange -- was working "kinda decent" with pitch and roll set up like I have it, but when I activate YAW, it is doing one thing it always has been doing, regardless of settings, calibration ... the YAW axis pulses slowly and moves.

    Enough so if I hold it, it moves my arm.... just pulses and pulses. Funny error ... if I let it go, it crashes into the remote and bounces back and forth.

    It has always done this even before I started messing with settings (basic setup from factory.)

    Trying to figure out how to make a concise video showing you how this thing acts. Would like to get it to act the same way for a while first so we have a baseline!

    I do appreciate help! I am hoping to learn enough in this process so I can make a video explaining what my problem was, and how to fix it.
  • You have too many things going on. It is best to disable follow and RC, set everything not used. Set all RC offsets etc. to 0. The gimbal must work ok without them.

    Also the gimbal works on basic operation ok, with 1 IMU, Z axis ACC calibration performed correctly and gyro calibration performed still. If it does not there is something wrong.

    It is best to start from simplest possible setup and add the features (follow, joystick etc) when the simple setup is first working.

    However, if it is 3 axis gimbal, it is better keep all axis activated from start.

  • I thought I posted this last night .. I guess not. While it has changed now, due to progress, I will leave this text here so future solution seeking persons might find it and glean from it useful information:

    12+ hours old (see next post)

    Thought it was almost working enough to make a video, so I powered down and removed USB and reconnected frame IMU. Apparently it still thought frame IMU was cam IMU, and my calibration was messed up, so I did a new cal on both and still gimal no work (floppy wild) so I disconnected all, and removed frame IMU again, repowered, recalibrated, and now it holds if I don't touch it after another calibration, but it doesn't work anywhere near as well as it was a few minutes ago.

    Problem here is that I don't know what changed. How can I do this work if there is no constant?

    Everything changes all the time, not same from power up to power up, and same settings, identical care in calibration, but different results?

    Very tricky situation.

    I don't know how to give you enough information in a way you can use it to help me, so I am trying to understand this enough so that I can help me to help you help me... if that makes sense.

    But so far, it seems very random each thing I try .. so each calibration of IMU means different PID's, and very careful calibration each time seems different, using same levels, etc.

    Probably time for bed.
  • Garug - thank you!

    I have learned more tuning stuff, and here is my findings, and this is with Follow=OFF, and RC=OFF, and RC Offsets=OFF, except for after I get tuning, I change Roll offset= -2.4 or so to level camera. This seems constant at this point.

    Here is what I sent to SteadyMaker as progress report ... after I type this I will continue work, to try to stabilize the jitters that happen when I go too far.

    Here's a great tuning video


    I have removed the IMU from the frame, and taken that IMU (I marked them) and put it in the camera IMU position.

    With one IMU, and specifically, this one, I can get the gimbal to operate somewhat reliably, although this is totally different from the stock settings as shipped, (partially due to different IMU and only 1 IMU for sure)

    However, attempts to replace that IMU with the original IMU from camera position wreak havoc.

    "So far" anyway ... I have not gotten to the point where I can confirm this scientifically, but the tuning is "fair" but not as good as it needs to be ... not by a long shot, especially confusing since it is not yet as good as it was before I started yanking IMUs....howeve,r one notable change is that using only one IMU and this one in particular, it is operating more reliably ... in that when I turn it on, it returns to the same attributes as before, even after a few hours, whereas before I changed IMU's around, it would be different each time, even after 10-20 minutes.

    I do not know yet whether I would ask for a replacement IMU. Further testing is required. While I feel it is certainly possible, verging on probable, at this point.

    I will update as further info gleaned.

    Perhaps by tonight I will publish new PID's and other settings, and have test video to share.
  • OK< I spoke too soon. I was really hoping I had a good start on this. But I saw gimbal was still, realtime looked great, and then I picked it up, and it freaked out.

    It freaked out in a way that it was NOT doing before ... I had it "fairly" tuned, then went out for a couple hours. Came back, powered on, connected GUI. Then it just sat there after leveling camera properly, and stayed put.

    UNTIL I moved it, then freaked out and started oscillating on what seemed like all three, but mainly YAW, which really started to spin around.

    So I messed with it a bit, could not find anything changed or wrong, and then reloaded that profile that I had saved earlier today after getting it pretty good ... and it came back to pretty good.

    Then I was able to pick it up and play with it a little, and saw some small oscillations, and made note of PID's, and changed slightly, and then tested new adjustments, and was OK.

    I turned it off to change battery.

    Battery new, now gimbal seems OK, but then when I move it camera freaks out, suddenly very little power to Roll, and I can put camera in any position, all motors on all settings back to the same, still same result.

    Power off/on, same.

    Reload known "OK" profile seems OK again.

    Power off, then back on, seems OK. (sitting still, level)

    Power off, then back on, pick up, move around a little, seems OK, but those oscillations still there, so I want to tune PID's more.

    So I connect to GUI, make a small change, write ... now ROll is off by a few degrees (easy to see as I have level glued to camera)

    Disconnect, power off, power on, seems OK<

    (sigh)


  • Back to fairly stable ... rebuilt PID's,

    Yaw is not as tight as I'd like ... +/- 3 degrees, generally +/- 1 degree, but sometimes more. Adjusting tighter makes vibration.

    (just back from taking a break, power was off, and now yaw (with follow) oscillates whereas just a moment ago it was working nicely ... and also caused pitch/roll vibration quickly)

    Now the yaw handle is slowly rotating. Before, it was not.

    SUDDENLY as I typed last sentence, pitch and YAW freaked out and became dangerous.

    Noticed batt level at 60% (yes, I calibrate voltage set). Changed battery. Now gimbal is on, camera is ROLL back the other way.

    Cannot pick up gimbal or it freaks out.

    I just don't understand why it's changing unless it is just reacting to changing voltage. Is this possible? I had considered this, and set the compensate voltage drop but things change either way.

    But with the same settings as I had not too long ago I cannot pick up the gimbal and move it around at all or it freaks out.

    One good thing is that roll is not veeeery slowly oscillating like it used to ... it has not done that since I removed the frame IMU.


  • Well, I got it working again, at least a little bit, and you cannot say I'm not doing my homework. I might not be good at it, but I am really trying very hard.

    Here is album with my current settings (several GUI pages to see) https://flic.kr/s/aHskrT1VLz

    And here is how it looks ... a few clips rendered and stuck together, nothing fancy, but showing with all current settings Follow OFF, and same settings Follow ON.

    You can see it has basic function, but twitchy. Then, what you don't see, is that it randomly freaks out and doesn't work, even just sitting here as I write it did this again. Battery voltage in this case is 12 volts

    https://youtu.be/NnoXF3dlLeI

    I will await your advice before doing anything else ... I am running around in circles, it seems!


  • Before I put away tools and do something else for a while, I turned motors back on just to see. With Follow off, it is working again. (Follow was off before when sitting here freaking out). So it is OK, and I know that if I leave it sitting it will go crazy again.


  • More info (if it is helpful to you when looking at my settings) I contacted Steadymaker to confirm motor poles and such: yes, 14 poles, motor is custom made by the supplier, but the pitch and roll is 6008, the yaw is 7008

    Is this helpful? I was a little worried my roll motor not strong enough, but I think it's OK.
  • If you need to apply RC roll offset to get the camera leveled, something is wrong.

    1:03 on the video, the gimbal will not work correctly if you hold it from anywhere else than from the handle, there is no point for that kind of test.

    If the two IMUs work differently, likely one of them is broken.

    What FW are you using. I would suggest, save all your settings (save EEPROM and profiles) and Erase EEPROM load the latest beta FW. Make sure you always use same version of FW and GUI.

    Would be good to inspect the motors and measure with ohm meter that they are ok. Each phase should have about same resistance, and no phase should be connected to motor metal parts.

    With camera IMU only, the gimbal will work ok +-30 degree handle positions. do not expect more, for wider handle movements frame IMU is needed (or estimate angles from motors, but no reason to go there, when it is working ok +-30 handle position, instal the frame IMU)

    The gimbal operation will change when battery voltage changes. what voltage are you using now? what voltage does the manufacturer recommend. There is battery voltage compensation, you should set up that correctly. When done so, it will eliminate the effects of battery voltage changes.

    It could be that your battery voltage is too low to provide enough power. If you try higher voltage, mind though not overheating the motors.

    Pictures of the gimbal and especially camera and IMU installation would help, I suspect the camera is not steadily enough installed and there might be problems with IMU installation too.
  • OK ... I get that ... I thought maybe something was wrong, but what? For a while I thought Acc calibration was off, leading to -2.xx degrees RC offset to correct, but if my calibrations were off, it would be different after each one ... not the same! Wierd, that.

    Understood about holding gimbal by stem not useful ... I wanted to show how yaw handle starts rotating strangely (sometimes it does not)

    I think one IMU is broken ... I am not sure which to suspect. Probably the first one that was in camera position ... the reason the gimbal was not operating right in the first place maybe? The Chinese manufacturers suggested I broke the 12c cable or connector by taking it apart too many times ... but I think that is incorrect, and I would see 12c errors if that was the case, no?

    Firmware 2.56 b6 (beta) This is reported by GUI to be latest, is it not?
    GUI=2.50 b3 ... I looked, but could not find newer. I will look again.

    I will also inspect motors as you suggest.

    I will save and reload EEPROM after GUI check.

    Thank you very much for your time and help!





  • Oh ... there is no angle sensor in the controller is there? Mine is canted slightly from vertical, I noticed.
  • Some editing on above post regarding battery etc.

    When you update the FW, see the text window, it gives the link to latest GUI. It must be correct or problems will follow.
  • Some controllers have frame IMU installed on them. On that case you should not use a separate Frame IMU and the Frame IMU will not work as Camera IMU because of the IMU addressing.

    If there is no IMU on controller board you can use Frame or Camera IMU as Camera IMU.
  • Thanks, will check above. Looks like my GUI is latest and FW too ...

    Using manufacturer supplied 12 volt 3s 500mAh batteries.

    I do have the voltage calibration set as per reasing and calibration, as well as battery voltage compensation (although I have switched it on/off several times, it does not help with sudden surprising changes)

    I'll get those pix of camera, gimbal, IMU install, etc.
  • I don't think there is IMU on this board ... I cannot select frame IMU as I have the one supplied removed. Here are the photos ... in the wiring photo, the frame IMU is removed .. you can see the grey patch of sticky ape underneath the controller board ... the IMU was there.

    The camera is very firmly attached.

    https://flic.kr/s/aHskppyNUx
  • Just checked resistance of motor phases and for continuity to chassis. All is perfect. No continuity between phases and case/chassis. Resistance is identical between phases of pitch and roll motors (individually), and yaw is different, but all same (different motor)