Difficult to maintain altitude

Cube Autopilot
1

Hello everyone!

We have transformed a Joyance drone to an ArduPilot platform. We are using an Orange Cube, plenty of power, etc. The usual setup.

We are using the same ESCs, motors, propellers, power battery, and wiring that the drone originally had.

We were successful, and the drone flies pretty well.

However, we have an issue: the drone can’t maintain a constant altitude. If you leave the drone with no input, it keeps going up and down—sometimes a few centimeters, sometimes at least 1 meter.

For example, I can see the drone going up while the barometric altitude is going down.

Here is a picture of the flight controller installation. The FC is inside the frame, with no wind inside.

image
image563×423 86 KB

Here is a fast-forward video where you can see what is happening, with no RC input.

I was checking the logs, and I can see that one of the barometers is a bit noisy compared to the other.

image
image1942×1067 201 KB

The altitude is reached easily, but it is still quite noisy

image
image1740×1083 87 KB

I have been working on the vertical acceleration tuning, but I can’t get the drone to hold its altitude.

Is it possible that the barometers are not working properly?

I hope you can help me. Thanks.

Here are two logs.

Thanks!

2

Is that power connector resting on top of the flight controller in flight? That won’t help.

3

Thanks, noted, we have change the Cube for a new Orange+, and improve the wiring. Any other idea to work on? Thanks

image
image496×679 122 KB

4

How tightly sealed is the fuselage? If too much, it will take the pressure sensor a while to sense the change.

5

….at least has one screw hole open…. but anyway, is not sealed…

6

….ok, so, we have change the Orange cube for a brand new OrangeCube+, did a quiktune, and the problem still there….

In this test, we have added a radar rngfnd sensor in the bottom, to compare the results, and I think the sensor works great, but as this is a drone to clean buildings, anyway we need the baros ok…

This are the baros now, I think there is improvement, but I still saw how the drone went up and down with no imput.

image
image1842×1159 189 KB

Any ideas?

here is a log, where I enable and disable the RNGFND with RC7 in the flight to compare the results.