Video Bounce P2 H3-3D; GoPro 3+ Black

Apr 7, 2015
Reaction score
Occasionally, my video from the gopro 3+ black has a bounce (up and down pitch motion) in it when flying.
The video moves up and down about 1-2 degrees (Guessing). The frequency of this bounce
changes over time. It is slow, maybe 1 to 5 cycles per second.
Gimbal is the H3-3D.
I have decreased the Tilt sensitivity for smoother tilt control. I though maybe
this had an effect on the bounce correction feedback control (if there is any). First tried the tilt sensitivity at 30-shot some video, then changed sensitivity to 2-shot some video. Doesn't seem to make a lot of difference. Very still day wind wise.
All other parameters are set to default.
This problem comes and goes, sometimes stable/sometimes not??
By the way, I have an ND filter on the GoPro with counter-balancing weights on the back side. With this, the gimbal is more balanced in the static condition.
Lee Snook
Last edited:
Might not be gimbal..
What controller are you using? New with the tilt dial or the old with the lever?
Have you modified the tilt channel in anyway? Plug into assist and see if there are values showing. I know the rear pot for the lever is very sensitive and maybe sending out ghost signals.
I am using the old controller with the lever.
As in my original post, I changed the tilt gain, didn't make a difference.
The rear pot, on the lever arm, is very rigid and tight. Anyway, I don't
suspect it because it's reaction time is too slow to produce the pitching frequency.
I did add series resistors on both sides of the Yaw pot to slow it down.
You know the yaw-mod affects the gimbal behaviour. The controllers pots are not built perfect - there is always a bias. The calibration in the RC12 and Phantom assist "hides" this. But when you do the yaw mod, the issue manifests itself.

Do this. Turn on your phantom and controller. Without touching the sticks, watch the camera when you enable and disable the yaw switch. Notice it looks to one side?

Try flying with the yaw mod disabled. The gimbal will behave as it should. Now enable it and see if the issue you have is occurring.

I did the yaw mod and there was a bias everytime I enabled the yaw-slow. The bias was to the left and it made the camera jerk/wobble as I descended/ascended. The h33d is compensating the yaw but because its so minute, it cannot compensate correctly.

You need install a 10k pot on the resistor that is reading higher. Then goto into the phantom assist, enable the yaw switch and then dial down the pot so the azimuth (triangle) is in the middle = 0..

Test again, turn on and off the yaw switch. If the camera does not move then you are done.
Last edited:
Thanks for the great suggestion. I sorted my resistors by using an ohmmeter. The effect that you described, camera yaw when switching the yaw switch is slightly noticeable, is minimal on my unit.
I am not sure how the yaw behavior would affect the pitch bouncing movement.
Getting resistors that are equal does nothing as the bias is in the pots inside the sticks.
Anyways, check your X1 channel in the Phantom assist and make sure its not moving. I don't know how you find dead centre using the rear level as it doesn't have return to centre.
I have the same problem as well if you mean this
I have the old controller but this has nothing to do with it i guess. Its just a bad copy of the gimbal i would say, will send it in for repair. It never crashed or landed hard, the power filter board which should prevent this! did not changed anything. The Gimbal has also massive Jello effect
Here is an example of the pitch/bounce problem that I see occasionally.

Here is my Yaw circuit modification.
I measured the pot balance it was 2.78k ohms on both sides with the arm centered.
Then, I opened the switch and measured 7.44k ohms on both sides.
Seems balanced to me.


  • Yaw Circuit.png
    Yaw Circuit.png
    20 KB · Views: 522
Here is my Yaw circuit modification.
I measured the pot balance it was 2.78k ohms on both sides with the arm centered.
Then, I opened the switch and measured 7.44k ohms on both sides.
Seems balanced to me.

Saw the video and never seen anything like that..
Can you try disabling the yaw mod and doing the video again? This would determine if your yaw mod is causing it.
Can you also default your gains to stock.

Nice diagram btw.

I did the same but introduced a 10k pot to the dominant side. You need to install the 10k pot on the side that is higher so you can lower the resistance.

It's not the resistors. Your resistors can be perfectly balanced and it still will not work.
The inherent problem is in the actual controller itself. The left controller stick has a bias when it was manufactured. Start the RC12 and click calibrate. Notice the azimuth isn't centred? The calibration then fixes this and hides the bias. Now when you enable the yaw mod, the bias comes back but at a lower value (you will see in the phantom assist).
Thanks, good advice.
I did disable the yaw mod and did not seem to make any difference. Although, the bouncing problem was not bad for this test. My gains were set to stock.
I understand what you are saying about the unbalance of the controller pot. But, my measurements do not show this to be true for the controller that I have. If you look at the schematic, you will see that the resistance on both sides of the pot are equal at the center position and when introducing the external resistors, the resistive measurements are equal.
I will do the RC12 test tomorrow.
By the way, I am and R.F. Engineer/ 45 years, so I appreciate what you are saying.
Check out some of my projects at;
All of this is fine and good but, still does not address the pitch/bounce problem.
Thanks again,
Lee Snook

Members online

Forum statistics

Latest member