New build of an unorthodox X8. Autotune came up with some pretty wild gains but it flies fine-ish with either the Autotune gains or the defaults. Yaw control is good, not great.
I have about 30 minutes of flight time on this and today while simply hovering it suddenly went into a full right yaw and spun until I crash-landed. Props are tight, RPM’s seem normal, only clue I’ve found so far is that ATT.Yaw and ATT.DesYaw seem fishy. I know my frame is weird and the tune needs work, but it flies OK other than this one crazy incident. Any theories would be greatly appreciated.
All is sound mechanically. What seems unusual about the motor outputs? Odd motors decreased while even motors increased, properly corresponding to the right yaw.
I haven’t notch filtered it as I’m just in the early stages of tuning, but vibration doesn’t seem bad and the Kakute H7 is mounted on standard gummies. If there were some vibration-induced aliasing of the gyro signal, would I be able to spot it in the log? Like the mag rotating one way while the gyro is reading the opposite?
The desired yaw lagging the actual yaw is a sure sign of a physical issue.
I just would have expected more separation in outputs between the CW and CCW motors, but some could be up or down to maintain stability. With a flat arrangement of motors it’s much more obvious and simple.
The other possibility is you have some of the spin directions wrong. If the order was wrong we could expect to see uncontrolled instability.
The frame is flat and square, it’s just very top-heavy and high drag, but that shouldn’t be a direct contributor to the yaw issue. And note that it flies fine and has decent yaw response, near-zero drift, and is fully controllable. It just did this crazy spin randomly in the middle of an otherwise benign flight where it was hovering almost perfectly still and then suddenly started spinning while remaining fairly level in pitch/roll (despite my illogical stick reaction). There were no significant stick or wind inputs and no datalink. I thought for sure it lost a motor or a prop came loose but the RPM telemetry suggests everything was running as commanded.
I agree that the desired yaw lagging is very strange, as is the sharp kink when the desired yaw begins rising. But even that mystery may only explain the continued pirouetting - why did it begin yawing in the first place? Is it possible that some integrator is backwards somewhere? Or some mode change occurred - I see the mag reset command but that doesn’t seem to be the culprit.
