Flyaway after rpm filter.

[rancid vine]

Hi, I flashed my Speedybee f405 from the stock Esc blheli_s to blue Jay(both 24 and 48Khz. I enabled the bidirectional, dshot300 and the rpm filter. Tried a test flight, on 24 was quite noise, after the take off it suddenly opened the gas a bit and didn't get back down only until I disarm! Then tried the same thing with the 48! Amazed how smooth and controllable the drone took off until the same thing happened, suddenly it gave some throttle and even with throttle full close, drone didn't came down, disarmed again. Then I tried to disable the gyro low pass 1 and moved the sliders a bit ok the right side. Tried the drone quite the same thing, the first time it accelerated quite immediately, the second time it was hovering for some seconds before this happened again. All the times that this happened was also noise! Disabled the rpm and the drone works ok(a bit noisy)The drone is a 7 inch with 1300kv 2807 motors. Any recommendations?

[rancid vine]

I tried with UAV tech preset , it end up even worse, it starts the throttle and the noise the second it takes off and goes quite fast also!

[wide tide]

Motor screws loose or touching the windings? Loose wire touching the FC?

[hot gate]

Can you send a blackbox log?

[rancid vine]

Here are the logs from yesterday! I don't know which is which but if it's too messy dont bother. I think it's the 117 is the one with the UAV tech filter presets

[rancid vine]

@hot gate
And this is a file from a flight before changes to bidirectional and rpm of it helps
https://quickshare.samsungcloud.com/aFUkHzAUA8gs

[rancid vine]

Checked it now, I can't say they were super loose, but they took 1-2 turns! I'll try it again! Sounds interesting, I hope this is the issue!

[hot gate]

I can see why its flying away: you have a lot of filtering and theres some oscillations that are feeding back into the system and the pid controller is acting on them causing a runaway situation

roll in particular, these oscillations are 100hz and your filter setup as it is wont touch them. Unchecked it goes into motor command and will vibrate the drone and the gyro will pick back up on them and the process repeats

these oscillations are around 100hz, thats the lower end of your dynamic notch range allowed

its 7" and I'm used to seeing resonances lower than a micro or 5" but this seems really low. In anycase you should use an rpm filtering preset since you flashed Bluejay

change the min hz to lower than 100. I wouldnt set it lower than 75-80

You may try this preset if the above change gets the drone under control. If they don't help then this preset might make things fly away worse though. Just fair warning:

[rancid vine]

Thanks for your time and your help! 🤗🤗🤗 I'll try them!!!

[rancid vine]

@hot gate right now I have flashed it back to bsheli and disabled rpm and bidirectional. You think it's a good idea to try to lower the filter first before flash it again to see if it will help with the noise?

[hot gate]

I think it's a good idea to use bluejay and rpm filtering. Rpm filtering was working and it will be less filter delay vs 2 gyro low pass and 3+ dynamic notches (needed for traditional non rpm filtering)

[rancid vine]

Maybe I can try to put in at 90 just for a safe start to see if it's better or...

I see

But In any case if I get it right problem should be the 100hz area that's not filtered?

[rancid vine]

Question, the dterm value that shows here, is the actual filter or a "recomended filter" I mean was it 75 for the d term, or dterm is another story?

[hot gate]

what you see in that spectral plot is your filters as they were when you logged that flight

that is you had stock dterm lowpass and stock gyro lowpass.

[rancid vine]

I see, but the dynamic notch is the one that will do the job for me, right?

[hot gate]

yes. set dynamic notch min lower and see if it helps. keep in mind that something on the drone is certainly resonating at that 100hz. If its the frame then the notch will help, but what would help more is fixing the mechanical problem instead.

[rancid vine]

Great!

There was some not so tight motor bolts

And also I have some play in two motors

One is slight, the other a bit more

I'll try to see with all screws checked

And we talk again🤗

[hot gate]

is anything pressing directly against the AIO/FC? like something sandwiched under it between the PCB and the frame?

[rancid vine]

@hot gate Nah, it's just the DJI camera cable under the stack and in between there is the beeper cable

[rancid vine]

@hot gate looks fucking good or im wrong???

can it be from the change to 90 from 100 or just from the screws that i checked?

[outer hare]

The principle of operation of the RPM filter is not vibration analysis, but clear knowledge through the number of Revolutions Per Minute divided by 60. Therefore, there is no point in guessing where the lower frequency of the motors is. Set the minimum available, it works for all frame sizes and all motors.
set rpm_filter_min_hz = 30
save

According to the log, the frequency of the first harmonic is 105 Hz. The filter works with 100 Hz plus 50 fade. It is not enough.
They use FFT_FREQ mode for logging.

[hot gate]

this entry duration is so long that my computer wont even display any spectral plots for me in BBE

I wonder if it will work if I set an in and out point. It WORKS

what drone is this? log in 1/2 or even 1/1 rate to see upper harmonics clearly

and why are we not using bidirectional dshot on this? what esc is it?

[rancid vine]

It's the one I had issues with the rpm and brought it back to the non biderctional mode🤓 7 inch, 1300kv, Speedybee f405

I just wanna make sure everything is fine with the noise before trying rpm again!

[outer hare]

The motors create vibrations regardless of the selected protocol ))
Biderctional dshot allows you to get rid of them. I wrote the command that needs to be executed to make this happen.

[rancid vine]

I see, but in my first try I had flyaways, so I wanted to make my drone flyable.

[rancid vine]

So! After some flights with the old settings, all was working great! Until... Today, I went to fly it and then the same thing, noise in the motors and flyaway. I disarmed and it felt from 3 meters or so. Tried to arm again, the beeping was wrong, you know that sound that it makes when you are plugging the battery, it's not complete now, no osd on googles, receiver beeping continuously. I came home to take the blackbox logs and FC doesn't even been recognized from the PC although some LEDs are on. I took the logs anyway, can someone understand what the hell is going on?! 😭

The first is when it happened and the second when I try to arm it again

[rancid vine]

I tried to plug another battery this time and it works...

Can it be the battery? (Was first try on a diy actually)

It connects also to the computer!

I'll be a schizophrenic with this drone

[rancid vine]

In the logs it has minimum voltage value of 13.22v!

[rancid vine]

Maximum ampere 50!! That's so wrong. It should need around 10! What's your thoughts? Motors? Battery? Something else?

[rancid vine]

I have removed the electrical tape from the battery and indeed connection between two cells was broken. I don't know if it was from the beginning for example when I tighten the strap or because of the drop

[rancid vine]

Update: it's not the battery, tried it again with another battery and same thing

I'm starting to think something is wrong with the motors

[rancid vine]

Hmmm motors are fried I guess

Should it be this?

Or motors are fried because of the issue?!

[rancid vine]

Update: flashed bluejay 48pwm

Changed also dshot from 300 to 600

And it is flyable now

And I have double the flight time holy cow

Was barely having 7-8 minutes and now 16!

Plus I have a feeling that now I can control my drone easier, I landed just next to me!