Pre- filtering it looked like my first peak was at 33hz, but there looked to be 6 peaks, maybe even 8. So i set 7 harmonics.
Post filtering It looks like everythings flattened out and all good, but is setting 7 harmonics bad practice? am i just filtering out 90% of the useable gyro data?
the vehicle also had oscillations it didnt have before, but that doesnt concern me as we havent tuned yet.
Params: (Logs is below 4mb but isnt being uploaded for some reason)
When you load the FFT screen select Magnitude (bottom left corner) then load your log file.
Everything will make much more sense.
There probably wont be 7 harmonics.
What size props have you got?
28.2 inch props, 100kv motor, 12S.
Ctun thrust out was average .17
I’ll have to check the magnitude checkbox later, away from the log file now. What does the magnitude checkbox do?
33Hz will probably be exactly what you need for the INS_HNTCH_FREQ
but INS_HNTCH_BW should probably be about 17 or as narrow as you can determine from the updated FFT graphs. You dont want the notch filtering to spill over into the very low frequency movements and attitude data that is essential.
For example, think of the required control frequencies as say 20Hz and below, and the lower extent of your HNOTCH settings as reaching INS_HNTCH_FREQ - (0.5 x INS_HNTCH_BW). So with your settings you would have hnotch affecting frequencies down to 35-20=15Hz
Set INS_HNTCH_FM_RAT,0.7 to allow the notch filtering to start working at lower than hover throttle.
The magnitude option will give a more sensible graph like in this example from the doco
That graph is actually with the “magnitude” checkbox unchecked. I did get a look at the FFT graph with the magnitude button checked, and even before notch filtering it seems like theres very little noise in the system post-filtering. Does everything look right here? Is it just the gyro filters filtering (what i think are) the little amounts of noise out?
Even though you seem to have everything set correctly and HNOTCH is working well, I’m a bit concerned about the crazy number of harmonics in that screenshot - GYR0
I’d be looking for what’s causing that and trying it eliminate it - maybe some parts are too flexible??
hey man, i just want to say i appreciate all the help you’ve lent throughout the thread.
Everything is pretty stiff. There are some long landing gears and a stick attached to the vehicle with around 30-50g hanging off the end of it. also a winch with the hook, but that should be tight on the vehicle. Those might be acting as tuning forks, but idk if they’re heavy enough to register vibrations hard enough to be picked up like that. Still need to read a vibrations textbook.
I’ve never seen that many harmonics either, at most its two spikes. All the rotors are spinning at the same RPMs as well. There was some slight looseness in one of the arms we fixed earlier, i’ll check at some point if that was causing it.
Otherwise, I tried to move the gyro filter up to 60hz, the post-filter seemed a little shaky
so i moved it back down to 30hz and called it there. Notch filter at 35, bandwith at 10hz (maybe the gyro filter should be moved down to 25hz?), attenuation still at 40. Took care of the PID tune here as well.
Hope everything looks decent here. Thanks again for all the help.
