I’m building a custom 10 inch quadcopter. Seems to fly well and has low vibrations during a normal hover (<4 on all axis), however when I move around, the vibrations increase to around 20-30.
Any info on what the usual causes are? I mean I’d assume motor/propeller imbalance or similar wouldn’t exclusively show up during nonzero airspeed, so I’m not sure what’s causing the significant noise increase.
I think it’s got to be wind and some prop wash across the flight controller.
Also check all the wiring is not vibrating against the flight controller, and see if you can secure the motor wiring at the end of the arms.
Otherwise I could not see anything obvious wrong.
I would make some minor adjustments to the harmonic notch filter - select multi-source and lower the ATT and BW.
Thanks! Will check if wind is an issue by attaching another enclosed flight controller as well fixed cargo box below the quad.
Fair, filtering is not optimized yet, but felt it worked well enough to not be an issue so didn’t bother with optimizing all the way while still doing experimentations on the setup.
You had the harmonic notch filter very close, it could easily stay that way. The settings I advise still do the job (maybe better) but should fractionally lessen the interference with PIDs.
I attached a second flight controller in a hard mounted wind insulated container below the quadcopter, it’s the same flight controller model (h743 slimv4), but without any GNSS/RC, so it timestamps at 1-1-1980 as expected. I don’t see any significant vibration differences though.
Broadly time synced screenshot of vibes from both logs
To my untrained eye, the motor wiring, if it is as loose as in the title image, can be a real vibrational disaster. Both due to the wires hanging out right under the propwash, and the wires loose within the arms.
By looking at your noise plots (ArduPilot Filter Review), I see that propellers contribute a whole lot, especially - by matching with the reported RPM values - in the second harmonics. While fixing the notch (setting the “multi-source” bit in INS_HNTCH_OPTS) would help ignoring most of this noise in control, this will not completely remove vibrations.
The second harmonic noise may be either because you are hitting wires with propwash from every blade - as in the previous post - or suggests that you probably should spend more time balancing props+motors.
I can see that prop balancing is a common issue, but I’d then expect to see this be an issue at “idle hover” as well, not just while moving. I can see it bothering the gyro at hover, but seems like it’s filtered out well and somehow doesn’t seem to bother the accelerometer.
Will look at a good plan on how to ensure the motor wires are less movable inside the carbon tubes and perhaps get rid of the mr30 connectors to optimize things.
Likely not your case, but I was hit multiple time by severe frame resonances conveniently hidden by ESC telemetry notches.
When you hover and autotune, the machine looks very fine and smooth. When you fly for real, props start spinning faster, ESC notches stop covering frame resonances and they hit you hard through control loop feedback. With one such machine, I am currently living with four notch filters, first being ESC-based, others fixed and meticulously configured (e.g. freq 42.5, harmonic mask 80) to cover about 8 different resonance frequencies.
