Hi,
What can be a reason of such setpoint behavior? It’s elrs 50hz, FLTT is set to 30, lowering it to 20 does not help.
Easiest way to eliminate ELRS/Controller issue is to check RCIN channel values.
It’s logged at 10hz, nothing suspicious is there.
what radio are you using? Is the ADC filter turned on in EdgeTX? If so, the RC output after ADC filtering can be quite stepped. Maybe after Ardupilot FLTT filtering applied, it then could look like target on your graphs? Not sure about it at all, just guessing.
I would try turning off FLTT (maybe without props?) and see what the target really looks like after some RC input.
MOT_THST_HOVER,0.1 #it's at 0.3 and that's really high for this rocket.
INS_HNTCH_HMNCS, 1 #IMHO I don't think you need the extra harmonic.
Also try turning off ESC extended telemetry. Even if just for trouble shooting.
I believe the default recommendation for ELRS users and any decent flight controllers is to turn off ADC on the transmitter (which is on by default on many occasions) and to use the flight controller software for filtering if needed.
1 Like
This is the same system, in stabilized mode the target is in steps, when switching to acro it becomes smooth. On other drones with the same TX the problem is either completely absent or barely noticeable.
I wonder which RC channels are used in this particular case. Some (if not most) ELRS transmission modes have heavily discretized channels starting from 5th.
It’s usual 1-4, log file is attached in first message.
Ah yeah, I see, and discretization would manifest in a different way.
I find it moderately intriguing that PID FFT peaks are bimodal with a slight attenuation in the middle of the peak.
I also find it somewhat suspicious that INS_HNTCH_FREQ = 54, which is quite close to these 50 from above, and that the real motor frequencies frequently fall below these 54. It might be that the offending 50 is where the propeller noise starts escaping from the filters. I would quickly do a test with INS_HNTCH_FM_RAT = 0.5 just to see what happens.
I think the reason of this double peaks is that loop rate is a bit lower than 400hz, so with rc rate 50 hz it sometimes 7 and sometimes 8 periods, TC smoothes the target but after next packet it gets a bit lower, and it starts again.
I am not an expert, but that looks like d-term oscillations. How did You tune it? Did You try lowering PIDs a bit?