P3S Gimbal Geometry - Something NOT Right

[JB63]

I've had my P3S for barely 10 days, and, as already stated in a different post, am fascinated with all the technology and 'smarts' inside the unit. Yet, today I found something rather 'odd'. If we want a gimbal to be very 'efficient', all three axes must intersect at the same point (as in gyroscope gimbals). We also want that intersection point to be in the center of the camera's sensor. If you look closely at the P3 gimbal, the three axes appear to intersect at the same point ... but the camera's center/sensor is offset about 10mm to the right. Why is that? I'm thinking one of the four possibilities:

1. Gimbal/camera were not designed/optimized for one another
2. Oversight from the design team (highly doubtful)
3. Some performance/design conflict that the engineers could not overcome
4. Someone wanted to avoid 'gimbal lock'.

 

[NM_Quad]

I think you are over analyzing things. There is no requirement for all 3 axes to intersect at the same point, nor for the axes to intersect at the camera's sensors. What you want is the camera to sit on a "stable platform" (proper gimbal terminology) isolated from the aircraft movements. That's exactly what the DJI gimbal provides - that stable platform. I also believe the pitch axis is slightly ahead of the camera to achieve proper camera balance for minimal servoing. Fly your Phantom up 100 feet, let it hover and shoot about 30 seconds of video looking at something and see if you can detect any noticeable movement. I can't. Rock solid. Shows the stable platform is working perfectly, well designed, and optimized for the aircraft and camera. Trust me, if your gimbal isn't working quite right, you will know it immediately as it dances around or very unstable video. Put trust in your Phantom and enjoy your flights.

 

[solentlife]

As per post above .... over thinking it.

The gimbal in fact is not a true 3 axis item, but a 1+1 affair.

The horizontal pan axis is not there for us to use, but is purely to counter any jerks or offsets caused by the model when we rotate it or by wind etc in flight.

We only have control of vertical tilt and that basically means centre of action is less critical.

 

[JB63]

I'm neither over-thinking nor over-analyzing. To some of us in this hobby, it's purely about the machine and what it can do. To others it's how much $$$ they can throw at the hobby, and for a third group, it's about how the machine is built so it delivers the expected performance. Now, back to my original point: One can 'easily' write the math equations that relate the effectiveness of the gimbal suspension upon vibration/rotation of the video frame and for the 'best' performance, the 3 axes need to intersect exactly at the center of the camera sensor. However, and like for any other design, there are other constraints the designers had to wrestle with. One of those could be minimum effort from the motors to get the camera balanced, and there are many more I'm sure. The other 'truth' is that, if the object being photographed is very far (as is the case with these birds), then the gimbal offset has minimal effect. One fact remains: the gimbal is 'not' centered and at least we can agree on this. Whether that has serious implications or not, we're all entitled to our own conclusions ... Happy Flying.

 

[JB63]

The gimbal in fact is not a true 3 axis item, but a 1+1 affair.

Can you elaborate a bit more on this?

I see three axes of rotation in my gimbal: Pitch/Roll/Yaw
The end-user (you/me) has manual control over the pitch, to frame the camera subject. However, all three axes are dynamically stabilized against sudden motions of the drone.

 

[solentlife]

Reason I say 1+1 is the control aspect we have of it. You yourself admit that tilt is the only control we have or pitch if you prefer.

The other two - roll and yaw are not available to you and are purely there to allow the model to keep camera stable.

Yes the gimbal if it was installed to give you the pilot full pitch / roll / yaw control would be a full 3 axis unit. But it is not. It is installed in a 'crippled' format where we only have control of vertical pitch. Therefore in practical terms 1+1.

I agree that there are different folks with different strokes and I find that I often encounter this in my quest for real info. I applaud persons wish to gain more knowledge - I fall into that category. But I also strike a balance between a hobby and technical desires.

Nigel