This was in Majestic mode from the hand unit. I did a auto cal at 75% hanging from the machine. Roll and tilt values where to high. So I brought the values back down. I did find bringing the strength values up from 7 to 20 helped the performance. Much higher near 30 and I got an oscillation. I held the machine in the air and slow small movements gave pitch performance as abysmal as the previous video. If weight is the solution and to add more glass or lead weights I need someone to say that. thoughts...? -Cody
Cody, I believe Gary did suggest adding weight in Posting #15, however, what camera are you using? Can you also verify that you have the camera well-balanced with the MōVI powered off? Will it stay in any position that you put it in, please? Thanks Andy.
Cody let us know the camera and lens setup. Photo of the rig wouldn't hurt. My next suggestion would be to do a reset to factory defaults and then an autotune at the factory default setting. Basically start from a baseline. Based on the camera/lens and their weights you may need to tweak the gyro/output but start with the stock settings.
Alpha 6000 with stock lense. It's a starting camera. Looking at an alpha 7 or other. I'll add weight tomorrow.
The Alpha 6000 is waayyy light, Cody. You really need to bulk it up to the weight of say, a GH4. Just for example: The Alpha 6000 weighs 344 grams/12.13 oz (body only).The GH4 weighs 560 grams / 19.75 oz (body only). The M5 needs a certain mass to "push back" against the motors and the Alpha 600 is just not there. So see if you can securely add on some mass -- just as an experiment use PVC tape to strap on some weight under the camera. Be sure to rebalance the camera/gimbal so, with the M5 powered off, the camera will stay in any position in which you put it. Also, be sure that you're not using a lens hood -- the downwash from the props can shake a camera like crazy. Andy.
K. Thanks. I'll add weight for an experiment. I feel like I just built a nice computer in a sense and bottle necked it with only 4gig of ram.
Cody, that's not really what's going on. I realize it can be frustrating to tune one of these devices, so, in the hope that it helps create a good enough working mental model of the problem for you (and perhaps other folks), what you actually have is a mechanical system subject to the laws of physics and complex interactions controlled by proportional, integral, derivate (PID) software controllers using nanoscale sensors and driving stepper motors to move the camera to remove rotational motion about the tilt, roll and yaw axes. (As an aside it cannot remove translational motion along these axes, only rotational motion -- so, for example, as you walk it cannot remove the up/down motion of each step.) Bottom line, PID controlled stepper motors need to sense a certain amount of inertia from their "payload" otherwise they oscillate. The parameters you are setting in the app. (e.g. stiffness) are the parameters for the PID controllers. See https://en.wikipedia.org/wiki/PID_controller A real life example for a PID is you. If you decide to accelerate your car up to 55 mph on a freeway, you become a PID controller, increasing the throttle as you move from a standstill, then as your car approaches the required speed you back off the throttle a bit to avoid over speeding, and then you add a little more to avoid slowing down to much....that's a PID in action. The parameters you specify control how aggressively you get to 55 mph (gentle or burning rubber), and how rapidly you back off the throttle when you approach 55mph and how quickly you react to and correct over speeding and under speeding. If you are too aggressive in correcting, then you would travel down the freeway going 58 mph, then backing off to 52 mph, then going back up to 58 mph and so on. And that's the oscillation you feel in the MōVI when the parameters are set wrongly. Now imagine you're driving an 18-wheel truck, not a car....you would have to adjust the way you drive because it weighs 95,000 lbs and cannot accelerate or decelerate anywhere as fast as a car weighing 3,000 lbs. The mass in motion alters the parameters you need for the PID. So you have to adjust the apps' parameters based on the mass in motion -- and for a light camera that requires a whole different set of settings. It's hard to see because the movement is small and rapid, but the Autotune is basically waggling the MōVI back and forth in each axis, tilt, roll, and pan, and sensing how the system reacts, then setting the PID parameters to the best values that the firmware in the gimbal control unit thinks will work. You can then tweak the parameters as you see fit. Hope this helps... Andy.
Andy, My degree is in instrumentation partially. I had worked on larger and much lower speed systems in loops like oven and heat control. I understand the gain and the PID. Where things got gray quickly is that purchasing the system I had not heard of a minimum mass/ weight spec. I completely understand your description above and it does make sense. The systems need feedback and this system this Movi gimble needs a force to apply to. Its not 'windowed' correctly. Thanks -Cody
Hi guys, Because I have a very similar problem I thought about just posting it here as well I have a X8 which I have usually flown with a custom built alexmos gimbal and it worked really well like that with a BM 4K camera but recently I wanted to test the Movi m5 on it and first used the GH4 and the results were horrible.. Everything was set to auto on the camera on that flight but for the next flight I balanced all the props which are CF T-motor props. And on the second flight I had exposure set to 1/50 but the results were still far from being good. And on the third flight I flew a BM pocket with the same lens as before and with IS set to off. But still horrible results... Can you guys please help me to determine what is causing these vibrations? I have not yet balanced the motors because it is hard to get all the plugs out on this copter and I also think it would not make too much of a difference. The motors are T-motor U7. Why did I not see any vibrations with the BM 4k and the alexmos gimbal? Might it be because the BM 4k has global shutter but the others not? Might this be because the whole gimbal setup is much lighter than before and the vibration dampeners are then less effective? Or is it some tuning which has to be done with the Movi? Which btw has worked really well in handheld mode. Here are videos from the flights: Flight one: Flight two (notice the roll movement which you can see for example at 2:50): And the third flight with BM pocket: I have tried to get some info from the DJI OSD mark II but the software is just far from being user friendly and I have not been able to read any vibration info from it... I would really appreciate if someone could help me with this
How is the MōVI mounted to your frame? Do you have any vibration isolation? It would be helpful if you could post: Some photos of how the MōVI is mounted on the copter A screenshot of your Configurator settings for the MōVI. How did you tune it? Autotune? And what firmware is it running?
In which case, profound apologies for trying to explain things that you already knew! Always a risk in the life of a forum moderator -- so thanks for being so gracious! Andy.
Thanks for the reply Here is a picture of how it is mounted. You can also see the vibration dampeners which are 6 in total which I think might be too much.. I just tuned it manually with 50% in autotune percentage. The camera is also perfectly balanced. I don't know what firmware it is but I think it should be the newest one. I will find out by tomorrow.
I'm no mechanical engineer, but I think your problem is those long bars that are holding the quick release rails. It all looks like it's going to amplify any movement, and there's nothing to dampen side-to-side movement. The Cinestar-style vibration isolators are designed to mount directly (or very close) to the MōVI itself and to be configured in a square or circle.. I think you just have too much hardware on there. I'd look into using the star plate on the copter attached to 4 isolators, then use a Toad in the Hole if you want a quick release. They work great.
Andy, I don't know anymore than anyone else. I take in every words to process it. Its nice for reminders as well. I know little and am always learning more. I appreciate any help from anyone. These forums are helpful for people who fight what seems a loan battle. If in the manual it would have said something to the effect the systems is designed for use of cameras with X weight, and that if lighter then weight must be added I would have applied the reasoning. -Cody
How long are those axial posts? I have extended mine to 5cm does not make any difference the thread is here http://forum.freeflysystems.com/index.php?threads/different-stand-offs-for-vibration-dampening.6288/
Thanks for the reply! I think I will shorten mine to about 4-5cm. I made a little test yesterday where I did put a small brushless motor to the frame itself and then stuck a screw with a tape to the side of the motor to create vibrations and it did make a lot of vibrations but still it did not have any affect to the camera. So I guess that the vibration dampeners are doing its job. Take a look: So for the next test flight I will only use 4 of the vibration dampeners instead of 6. I will also lower the stiffness for all axis which might have been a little too high before.. And I might also play a little with the gyro settings. Do you guys have any other suggestions to what I should do for the next flight? The only problem now is that the weather is horrible at the moment and will be for the next few days...