Drone jumped from 60m to 70m altitude in the logs while vibration compensation was active, then appeared to drop. I want to know whether this was a real altitude oscillation or just the EKF misreading due to vibration-induced sensor errors, and the reason why this might happen.
I’m sure it was. page down to “what will happen”
Vibration Compensation
You don’t want to see that activated.
1 Like
Thanks for the link. But, to clarify the altitude change was unintentional, no altitude change command was given.
What the logs show:
BARO.Alt & GPS.Alt: Gradual climb 60m → 70m.
CTUN.Alt: Sudden jump to 70m when baro/GPS reached 70m.
Vibration Compensation: Active during this period, then deactivated after 15 seconds.
The map in the log shows the drone has descended instead of climbing.
I want to know if it is a EKF estimation error, or the Barometer and GPS were sending inaccurate data.
I am attaching CTUN.Alt, BARO.Alt & GPS.Alt as well as XKF3.IVD and XKF3.IP.
D
Motor RCOUT data while the unintentional altitude change to 70m.
That vehicle is likely not fit to fly until vibrations are reduced significantly.
Actually we took another flight after this and plenty of flight before this one, and they were all fine. No such issues
Software does not change over time. Parameters do not change over time. But hardware does change over time. It’s also known as wear-and-tear. Screws get loose, bearings get issues, stuff gets bent… etc
So if it flew well yesterday, it does not mean that it will fly well tomorrow.
It had no vibrations yesterday, it does not mean that it will have low vibrations tomorrow.
Do regular hardware maintenance to avoid such issues. And keep an eye on the vibration levels!
1 Like
That was understood. Once it’s switched to the 2-stage controller “but its altitude hold will be less accurate than normal” is an understatement.
1 Like
From a strict reading of the documentation, do “positive velocity and position innovations” mean that the accelerations are causing an estimate of the position/velocity that is more positive than it really is (in FRD, more downwards, causing a correction upwards and a flyaway), or the opposite?