Parameter analisys after autotune. 5.5 kg custom frame quadcopter

Hello,

After a lot of hit and miss in the past, i spent some time developing a new frame for my quadcopter designed with video surveillance in mind. I ended up using quality parts (recomended by the nice users here on the forum):

  • Motor KDE direct 5215XF 330KV
  • Esc APD 80F3X (with telemetry)
  • 16000 Mah tattu Battery 6S
  • 18 inch props
  • Cuav x7 Pro Flight controller with Neo 3 pro GPS
  • Herelink 1.1 for telemetry and control.
    All the parts were installed on a frame developed by me, taking into consideration vibration and a clean build overall.

Yesterday i had my maiden flight, Autotune and overall it was a good experience, except some higher than usual values after autotune in Stabilise pitch and roll P - 18.3 and Accel MA 136000. I changed those values to P - 7 and Accel MA - 90000 for both pitch and roll)

My knolwedge in reading and interpreting logs is still somewhat limited. I need an opinion regarding the current state of the quadcopter tune and if the notch filter is working as intended.
As usual, the logs are here KDE 5215xf - Google Drive

Log 1 was autotune for roll.
Log 2 was autotune for yaw, then pitch.
final flight in log 2 was in loiter mode, performing some agressive manoeuvres to see how it handles the tune.

Any feedback is highly appreciated!
Thank you in advance and i hope you have a wonderful day.

Mihai

No way to know w/o setting these.
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
And set this INS_ACCEL_FILTER,10

Make another flight that does not contain an Auto Tune session.

1 Like

Hello Dave,

I have modified the parameters specified by you. The new flight logs are on the same link.
https://drive.google.com/drive/folders/1LlXoqe6ppUbQWVIH0uuLsiEOB_ZWdC9B?usp=sharing

Log 3 Contains about a minute of loiter in place, no input from the RC. After that i continue with some aggresive maneuvers (for a quad this size).

Log 4 has some guided mode cruising followed by another short session of max stick input in loiter mode.

Thank you for taking your time to check these logs.

All the best!

Mihai

You can set the notch parameters as follows:

Notch

Disable FFT you are not using that, FFT_ENABLE,0

There remains a vibe mode ~23Hz that ESC RPM won’t pickup. It may have no significance but it might be due to something on the frame shaking. Landing gear, GPS mast, possibly what looks like a landing light.

You could make these changes and run another Auto Tune and see where you are.

1 Like

Thank you for this valuable data.

I have modified the parameters specified by you and i switched from 18 inch props to 20 inch props of a higher quality and better overall ballance.
After resetting the initial parameters with the planner plugin, i performed another autotune but sadly stabilise Pitch and roll values ( P and ACCEL MA were unusually high again (p=18,4, accel MA =122707).
I decided to lower them back to my initial values, P=7 and accel MA 50500.
i also lowered D term from 33 to 30 and flew for another 30 mins in a separate log.
Also i discovered another vibration source from the IR blasters ( as per your recomandation :slight_smile: and i took care of that.

The logs are at the same link.

TLDR: 20 inch log 1 and 2 are various maneuvers (log 2 is in heavy rain)

20 inch autotune is my attempt at autotuning pitch axis after propeller replacement.

All the best!

Mihai

It must have been the props because the Yaw imbalance is fixed and you picked up a bit of thrust so the thrust/weight looks better. Overall it’s looking pretty good! You could try another Auto Tune with aggression set to .075 but it’s in decent shape now.

Dave,

Thank you very much for your time, I would have struggled a lot without your help.
All the best!

Mihai

Hello again :slight_smile:

I did some more test flights since our last conversation, the majority of them were smooth. Yesterday i noticed an oscillation near the end of my flight, that i don’t know how to interpret and i was wondering if you can take a look on the new log to see if there is something i should modify in the pid tuning to smooth things.
Te oscillation in cause is in the 13:55:30 time mark and appears to be only in Roll axis. (* log 0 - test 7 feb)

Thank you in advance,

Mihai