My drone is tuned very well with Cube orange running version 4.0.7. (quadcopter, 22’’ prop)
After updating to 4.3.3, I kept the same tuning, still using EKF2, and unfortunately the drone is oscillating like the PID is too high.
I also tried changing to EKF3 but still getting the same result.
Does anyone know if this upgrade requires a new PID tune? Maybe I missed something?
I really want to avoid tunning again, the platform is very stable and i want to keep it that way.
Supply a .bin log file and we can check it.
There’s been some rate and loop improvements that might have made attitude control more sensitive (or bad aspects more obvious) if it was already on the edge.
Sounds like you could enable the harmonic notch filter too.
OK. It’s also using a different IMU as primary on Cube Orange with the update. That really shouldn’t have a big, if any, effect but possibly.
But in any case it shouldn’t be much effort to retune.
I would be interested to see a “very well tuned” log. This can be very subjective…
I would say it was previously on the edge of doing these oscillations anyway.
First up set up your battery monitor, at least for measuring voltage.
Once you’ve verified it’s measuring voltages Ok, connect to MissionPlanner Initial Parameters and enter your prop size and battery cell count.
Also select “Suggested Settings”, press Calculate and accept everything it offers EXCEPT the ATC_ACCEL values (that you already have set for your copter).
Is that the Hobbywing Plantinum 60A Pro ?
It’s a bit of an odd ESC, this ESC is really set up for Planes that dont need braking, or Helis that run at fairly constant RPM and take up to 40 seconds to come up to speed.
And the specs still dont even mention the PWM range!
Ensure only one of these ESCs has it’s BEC voltage supply (red wire) connected to the servo common supply rail, or their voltage regulators will fight with each other and overheat.
I was really hoping to verify the PWM input range just to check your MOT_PWM values, but I think you will have to leave them like they are now.
I’ve got a spreadsheet that is the basis for the MissionPlanner Initial Parameters section, and I’ve added in a section that estimates the HNOTCH settings from the prop size and hover thrust. This is not going to be exact because too many unknowns affect the actual frequency, but it is near enough for a starting point.