Blaming it on the mechanics of the gimbal sounds like a cop out to me. I guess every gimbal in the world was designed wrong. What a waste of time and money.
Vibration when tilting has nothing to do with being "designed wrong". Any gimbal design has some compromises according to their specific use case. Standard pitch-roll-yaw gimbals are compromised in that when pitching down, the lens mass becomes perpendicular - rather than axial - to the roll motor, and that causes gains to be incorrect for this situation. Yaw can be affected too, but less so (because the weight of the camera is a smaller part of the overall yaw axis mass - the other being the pitch and roll assemblies).
There are ways to go around it with software and we're doing that, but the source of the problem is inherent (and in fact has nothing to do with the pitch axis which people think is the problem source, incorrectly).
How bad this behavior is depends on three factors:
1. The weight distribution of the camera
2. How aggressively the controller is tuned in the straight ahead position
3. The rigidity of the gimbal structure
There are other gimbal designs like the Letus Helix which avoid this by placing the pitch as an external axis, but they have other limitations (like gimbal locks when yawing).
Had this problem been in the CP software, you'd see it on any gimbal, but it's far from reality (it does happen on many gimbals, but certainly not all or the majority of gimbals). We don't see it often with square payloads like a GoPro.
The reason this problem is more prominent in the CP comes from the tuning precision the autotune achieves (which is why manually reducing the gains solves this problem many times).
Most consumer cameras aren't very long or don't have very heavy lenses, so you might not be seeing it with less precise controllers. It doesn't mean it's not there and we have a customer flying a 10kg camera with heavy lenses that had the same exact problems on Alexmos before.
Obviously when the forces the controller needs to deal with change, the tuning becomes invalid. It's physics, as always. Mass that was axial and now placed far away, perpendicularly, is going to behave completely differently (gain changes up to 100% according to our tests!).
The rigidity of the gimbal structure is another matter - some gimbals deforms when the camera is pointed down, due to the weight of the camera being supported by a mounting single bolt that's experiencing shear forces (instead of resting on the gimbal pitch shelf, as it would in the straightahead position).
. I think if we had a video on how to work the different features that would help . If your gimbal is doing this when you run the analyzer then you can try doing this, etc,
