Limit Wind fighting in Auto mode

From lot’s of anecdotal evidence (I don’t have Cubes),and hard facts posted as logs before and after, “Hard mounting” with the supplied foam tape only works well on very rigid frames. The internal isolation is insufficient for compliant frames. But there is a steady drumbeat of “must be hard mounted” with these FC’s so take it as you see it.

With Vibes and clipping events off the charts you are lucky Vibration Compensation FS hasn’t been activated. Then things can go sideways quickly.

This is not what you should be worrying about. Also, running this on default parameters is likely not wise. With 17" props on 6S at least set these basic initial tuning parameters, round up on the filters, after fixing the Vibe issue.


Then read thru the process for setting the dynamic notch filter. Complete that then do an Auto Tune.

thank you for the reply, i did another test today with the small battery and reduced the speed to 10m/s vibration stayed well below 25 all the time (this is ok right?)

yeah, it is almost a religious question if the cube should be mounted on foam or not :wink: what kind of foam / vibration dampening do you recommend?

i also investigated the mentioned GPS issue and solved it, was a faulty cable. nothing to worry i am flying with 2 GPSes anyway.

i activated the dynamic notch filter and it might fly better, difficult to say. do i need to redo the auto tune now? (already done before) also do i have to do the process with the values you have mentioned if the auto tune is already done and the copter hovers very well and stable?

btw, the frame is pretty strong, aluminum frame. will post a photo of it tomorrow.

You can’t just activate the notch filter. It requires a minimum of 2 flights with multiple parameters to set during the process. Each craft is unique.
Tune sequence:
Set initial parameters
Set dynamic notch parameters
Perform Autotune

The ball mount plates work well in most cases.

Please do follow the instructions posted by Dave. I will probably solve your issues.

Hey guys
thank you for your advice. sorry for letting you wait for a while. i had a issue with the GPS, was solved with replacing the carrier Board of the cube.

i then adjusted the values as proposed by Dave and made a test flight. flown pretty well, similar like before but a little less aggressive, especially the yaw.

so i performed the auto tune in Loiter, after it was done i tested the flight behavior and noticed that it was worse than before. especially in loiter, holding the position was way less accurate than with the old PIDs. so i landed with the switch Autotune Off, charged the batteries and performed the auto tune in AltHold mode. The tuning for the Yaw took a little longer than the Pitch and Roll. anyway it succeded, flight behavior was better than the (loiter) autotune before but still not better than the old PIDs. so i landed again in Autotune Off to return to the old PIDs.

i must have missed something, the first time i made the autotune with the notch filter off, it was a huge difference after…

i am pretty happy with how it behaves now (at least with the small battery, didn’t test the big), but i am wondering why the auto tuning was no satisfying success.

The Log Files are available under the same download link like mentioned above.

Thank you for your help.

Here some Pictures of the Quad:

We never saw the logs for the multiple flights required to set the Notch filter parameters. Simply enabling them with default parameters is pointless. If you want to do it right then set INS_LOG_BAT_MASK to 1 then make another Hover flight in AltHold for ~1min. The post that log. Then the notch parameters will be determined and set and you will make another flight to check to see if it’s effective. Then you would run Auto Tune.

Thank you dave for your help. i learned a lot about the dynamic notch now…

unfortunately the manual does not go much in detail if using BLHeli telemetry data…

so i did the hover like you requested (finally the weather allowed it) once with
INS_LOG_BAT_MASK = 1 (named Hover1.BIN) i then adjusted some values like i believe it is right and made another flight with INS_LOG_BAT_MASK = 2 (named Hover2.bin)

(the logs are available under the same download link like the others)

Thank you for helping me out.

Not bad. I would change these parameters as follows:

You want the filter active over the frequency range seen and there is small peak ~46Hz (frame mode probably). The harmonic of this is the big peak and it will be attenuated with a BW 2X the 1st harmonic setting or 46Hz in this case.

I would make these changes and then run Autotune and you should be set.

cool, thanks. i’ll try tomorrow.


Edit: heavy snowfall today, no flight permission :wink:

Looking at the pictures I would suggest to fix some of the cables in place. It looks like they can flap around in the wind and the GPS cables even hitting the FC.

40 is out of range, i’ll try with 30?

Ignore the out of range message

ok guys i did everything like advised by you and would like to say thankyou first :wink:

it completed the first longrange flight today (15km at 10m/s) first with the small, then with the big battery
the vibration increased only a little when using the big battery and stayed well below 30 now (mean was 21)

i will do another test tomorrow at higher speed to compare with the failed flight before.

We did some Speedtest Flights today. i made a mission, Do Set speed, then 1km flight, turnaround and come back. then increase the speed and do it again… we did 8, 10, 11, 12 ,13 and 14 meters per second. (Check the log called Speedtest8t14ms.bin)

The Z-axis is getting high when it comes to speed like 13…14 m/s… lower than with the old tuning but i am still not very comfortable…

i guess all i can do now is mount the FC on vibration dampers? we would like to reach about 18m/s cruise speed

Would there be anyway to increase the thickness of those aluminum plates? Usually high vibrations is a byproduct of a frame that is not stiff enough. I have an 18" propeller quadcopter flying at 20 m/s and vibrations in the Z-axis average are well below the limit.

I agree with Manav. Aluminum is not the greatest material for multirotor frames. But short of that put the FC on a ball mount plate. That has worked well for many craft with flexi frames. Dump out the “it must be hard mounted” cool-aid :grinning:

yes, i can either stack 2 of those plates and bolt them together or go to a metal shop and get thicker plates… but the frame feels pretty stiff, you cannot bend anything by hand…

i will see if i can enforce the frame, if not i will mount it on ball mounts…

btw. i have been reading that the ball mounts should lead to the center of the IMU like this thing here does:
hovewer i do not feel comfortable to put the FC on a fragile 3D printed part. so do you think i can mount it on a regular ball plate?

Stacking two of the plates is a good start to see if it will make any difference. Not being able to bend something by hand usually means the airframe is strong - but for vibrations you are looking at stiffness. The two are slightly different metrics - although they are related to one another.

By stacking two plates you will also be increasing the mass of the center module - which should help with vibration damping as well.

As per putting it on a regular ball plate - I can’t say for sure what your results are going to be. If you already I have one in-house I would def give it a shot in just normal hover to see what it looks like.

The 3D printed mounts are strong enough for the job. The FC shouldn’t experience force that is strong enough to break 3D printed parts. Just print in higher infill if you are worried.

thank you for the tip, i have another idea how to stiffen out the frame, i will talk to my engineer coleague at work tomorrow…

i am not sure if i have one in house, but was thinking about CNC milling one out of alluminium…
i can also give the 3d printed a try and compare then…