Autotune speed-up

Hi!

My drone without camera is about 750g. With camera it’s about 1200g.

Without camera I can do (and I already did) autotune - it worked fine.

With my camera the drone oscillates, so I would like to perform another autotune - but my battery drains before autotune algorithm would finish the tuning process.

There’s AUTOTUNE_AXES parameter to check what axes to be tunes. By decreasing the number of axes to be tuned, gives enough flight time with the battery.

Is there another way to increase autotune speed, or is this the only one?

Battery: 4s1p EVE 21700 40PL

Motors: 2306 1750kv

Props: 5x3.1x3

Yes, I think this setup is a little bit underpowered for 1200g TOW, but this is only a test-drone, to learn flying, setting things up, before building a properly sized one, with decent flight time (around 10inch props, with 6s battery)

Thank you!

You have strong not filtered vibration at 51 hz, it’s on Y axis only (roll), please address that. It will affect your autotune. Than you can setup your RPM notch something like this.

edit: y is roll axis

This vibration shall be solved via hardware, with changing the camera mount, right?

Yes, software filtering won’t be effective on such small frame size.

Thank you!

I try to change my camera mount to the frame with some softer way, not to rigidly screw to the frame.

Y̶ ̶i̶s̶ ̶r̶o̶l̶l̶ ̶a̶x̶i̶s̶,̶ ̶c̶h̶e̶c̶k̶ ̶y̶o̶u̶r̶ ̶c̶a̶m̶e̶r̶a̶ ̶m̶o̶v̶e̶m̶e̶n̶t̶ ̶l̶e̶f̶t̶-̶r̶i̶g̶h̶t̶.̶
̶I̶t̶’̶s̶ ̶c̶l̶e̶a̶r̶l̶y̶ ̶v̶i̶s̶i̶b̶l̶e̶ ̶t̶h̶a̶t̶ ̶r̶o̶l̶l̶ ̶a̶t̶t̶i̶t̶u̶d̶e̶ ̶c̶o̶n̶t̶r̶o̶l̶ ̶i̶s̶ ̶m̶u̶c̶h̶ ̶n̶o̶i̶s̶i̶e̶r̶.̶ ̶
Correction - that’s not noise, it’s oscillations.


seen here

Thank you for guiding!

I modified the camera mount (in a quick-and-superdirty way…) and hopefully tomorrow I can make a test flight.

1 Like

I was wrong about axes correspondence, the correct is:

Body axes (X, Y, Z):
+X points forward (out the nose).
+Y points to the right.
+Z points down (toward the Earth).

Roll = rotation about X (Positive roll → right side goes down)
Pitch = rotation about Y (Positive pitch → nose goes up)
Yaw = rotation about Z (Positive yaw → nose turns right (clockwise when viewed from above))

You have clear 2hz oscillations on pitch axes, not roll as I mentioned before, fixing that 51hz vibration might fix it