@Garug
Hi. I made this video:

http://youtu.be/-YljYJZPgJs

The gimbal is turned off, and this bouncing also occur when turned on.

In the video I'm hitting hard the roll axis, but something pretty similar happens when the gimbal is shaken roughly, like when leaving it on the table/floor.

So, is this bouncing normal? I'm guessing it happens because of the physical design and the flexibility of some parts.

I don't think that would be the cause of my roll vibrations in the roll axis in extreme situations, mostly because it would make vibrations in the pan (the yaw of the camera), not in the roll, but I need to know if this bouncing is normal for a gimbal like this, and if a perfect stabilization can be achieved anyway.

Any advice will be welcome.

Thanks a lot for the reply.

The camera is firmly attached, it doesn't move a bit from the plate.

The IMU is fixed with screws at the same part the camera is attached, and it doesn't shakes by itself or anything, I've checked.

The battery, the pitch motor and the camera seem to be perfectly balanced, I make sure everytime I put the camera.

Anyway, the “bouncing” happens even without camera, battery, an all, it seems to be a normal physical condition of the design and the materials, but I'm curious to know if it's really that normal (something similar happens to your gimbals) or if it shouldn't be that way.

Thanks again for the response. It's very appreciated, really.

Now that you mention it, when the gimbal is turned off (that is, disconnected from the battery), the pitch and the roll axes usually return to the original position, especially the roll, but also pitch if the angle is too different from normal. Is that good or bad? :S

I thought that this is how it should be since I don't know how could the axes be left in the position. They naturally return to the original angle due to the camera weight, or even the weight of the axes' materials themselves. The motors don't have any torque if there's no voltage... Anyway, the camera always return to an almost perfect horizontal angle, which is what I think it's “balanced”. Any advice will be much appreciated.

The camera is right in front of roll axis' motor (or maybe just a centimeter to the left, is that bad?), just at the same vertical level. I don't know if it's too low anyway... :S

It's also true that the camera shakes more than the gimbal parts themselves, but I guess it's because the camera is just at the less rigid part, the plate that holds it in the pitch axis. But as I said, it's weird this bouncing doesn't seem to affect the video very much (just a little bit in the horizontal when hit too hard) since the camera has a couple of stabilization systems by itself, so the roll vibrations must be originated by something else.

About the gimbal balancing. The user manual does not do a great job explaining it, this is from the 32 bit users manual.

"Mount the camera on the tray and balance the gimbal in all three axes. Stabilization quality strongly depends on balance quality. To check your balance, pick your turned off gimbal in hands. Make fast motions along all axes, try to catch resonance point and swing the gimbal. If it is hard to do - gimbal is balanced correctly."

This is one good way of testing how well the balancing was done, but does not really explain how to balance.

To balance the gimbal you should:

1. Power off.

2. Verify that angles between all axis are 90 degrees, and pitch and roll axis are in same level. verify that the gimbal fees solid and that everything is well attached to it (no loosely attached weight, that will cause problems) also verify that everything moves freely i.e that the gimbal is correctly assembled.

note: Everything that is installed during normal use needs to be on place, for example if battery is installed above yaw motor, it needs to be there when preforming the balancing.

3. Keep roll levelled, start from pitch. Set the camera to a location where the pitch orientation stays in any position it is set to. Do this by moving the camera up/down, back/front as needed.

Note: To be clear, the camera must not return to level position. It must stay in any position it is set to, lens up, lens down, level. it must stay on that position after you let go of the camera.

4. Now do the same for roll axis. it should not matter on what pitch angle the camera is as it is already balanced. There is couple of ways of balancing roll. Often camera can be moved left/right or the hole roll assembly from where it is attached to the roll motor. This depends of your gimbal.

Note: it is important that the weight is evenly distributed around roll axis. some assemblies makes possible the roll arms rotating up/down, they should be on same level and direction with roll motor axis. Also now weight above or below roll axis.

5. Balance the yaw: at this time it does not matter on what position the roll or pitch are as they are already balanced. To balance the yaw, first turn the gimbal handle 90 degrees left (or right). The yaw axis is now horizontal. Move the hole roll/pitch assembly front/back as needed so that it stays in any position it is set to. (There should be mechanism allowing moving the yaw motor position front/back related to other parts.)

6. perform the above mentioned test and you are done. (that test work great with big cameras and gimbals, but maybe not so great with small cameras.)

7. If you change camera lens, zoom the lens, add filter or any other accessory, the complete balancing needs to be done again. But often small unbalance is manageable, so you may be able to use the zoom like I do here.

Wow, thank you very much, Garug, it's really appreciated.

The person who sold me the gimbal, and mounted it, didn't know a thing about gimbals, I'm more positive everytime. He didn't told me any of that, and the gimbal was wrongly mounted in many ways (for example, my camera almost falls...), and the software was awfully configured. I must fix everything myself searching, trying and getting help from you, mates.

Also, it's too bad that the manual isn't as accurate or detailed as it should be in some matters... Like balancing and PID tuning.

I'll try everything you explain to see if it improves. Although I don't think that pitch and roll will keep steady in some angles...

So, isn't mandatory that the camera remains just right in front of the roll motor, it can be some centimeters right or left? Maybe I could move the center of gravity of the roll axis moving the tube as well as the camera position in the pitch to make that possible. Should the roll be balanced even without a camera, or the balance only really matters when the camera is there?

The roll arms are a little bit upwards. I found that weird, but I though it was like that because the pitch motor should be at the same level as the roll one. Should the be horizontal anyway?

Sorry for so many questions... I'll explain how it works as soon as I can. Best regards.

So I balanced the gimbal. Now the roll axis stay in level at every position, with or without camera.

But since the gimbal is a CineStar 3 copy, the camera in the pitch axis can't be moved up and down (especially up as I would need), so the pitch never stays in level, but always returns to original position... I wonder if that too-low-position could be causing the roll vibrations by inertia or something.

I'll keep testing and trying with this gimbal, but as soon as I earn some more money I'll buy a really good one. That is, if I can't achieve a perfect level of stabilization as I wish.

@Garug
Thanks a lot for the reply.

That's what I meant, not sure about the meaning of “stay level”. I'm not native English.

Anyway, the axes stay in the position I put them, now even roll, except for pitch, as I explained. Pitch works well, maybe the best-working axis for my gimbal right now, so the unbalance of it only could slightly affect roll, maybe, if anything. Maybe if I get tired soon I try to change the pitch part to be able to move up and down the camera, and improve the gimbal performance before buying a new, more expensive one. It won't be easy, though. I don't know how to even start. But I'll figure out something.

Apart from that, I'm still trying with different PID settings. I fear lowing P too much will be worse (especially for handheld), but maybe that solves the roll fast oscillations in flight. When P is high, the roll axis is much more stiff, but there are more oscillations in roll when speeding up. So maybe the question is finding just the exact amount of P and D... I guess it would be easier with fully balanced axes, but...

Do you recommend me trying Gyro LPF, or it complicates things even more, or do you think it won't help with the roll oscillations?

Thanks again.

Yeah, I know. Software-wise it's pretty accurate now I think, just need to keep trying different P and D, maybe Gyro LPF. Power and “I” values are at their best, judging by the several tests I did. Not so sure about P and D, though. I can't get them to give me concluding-enough results, possibly due to the pitch unbalancing.

That pitch balancing impossibility bothers me, because it could be the culprit of the roll occasional oscillations. I can't be sure, though.

I'll keep searching and testing and I'll comment some more.

Do you mean raising the camera position changing the pitch axis start angle?

I wonder, what if I just raise the roll arms to some higher level? I thought about that, but I don't think it would work too well, so I didn't even try it...

Anyway, I guess I won't lose anything for trying...

Thanks.

I'll definitely try raising the camera center of gravity by changing or moving the pitch part, or by making new holes to it to lift the parts, or something, whichever works best.

Thanks a million, Garug. You're of huge help, seriously. You explained useful and interesting things when nobody else seemed to know or care.

I'll be commenting when I try something more. Best regards.

@soulpatchfilms
Thank you very much! I understand perfectly, I really appreciate the assistance. Some pictures could come in handy for further reference/details, though.

I thought about doing that arm rotation fix, although I didn't think it would work too well since it wasn't so good balancing at all (the camera is still too low from the pitch motor) and I thought the arms should be horizontal anyway. But I'll definitely try it. I need to balance the pitch somehow to see if the roll oscillations disappear for good, so it's worth trying.

Best regards.