Reducing oscillations and Dynamic Notch filter setup for large copter

Hello guys,
I completed my maiden flight with a large quadcopter and observed some slow oscillations.

Motors: Hobbywing X6 180 Kv
Propellers: 23 x 8.8
Firmware: copter 4.3.6
Battery Li-ion: 12 Ah
AUW: 11 kg

Please help me to figure out this issue
@dkemxr @Leonardthall @amilcarlucas @xfacta

Log & Video Link:
https://drive.google.com/drive/folders/1m7Qju09vAHgw4h-6sKdRrCL7bnNIdPnm?usp=share_link

Y axis vibrations are more significant than X or Z, so maybe examine the mounting of the flight controller or if wiring is pulling on it or rubbing against it. Try to fix this as it will affect the tuning - it is already.

GPS update rate is a bit erratic, try setting
GPS_GNSS_MODE,5 or 65
whichever gives you the most reliable HDOP and a good update rate of 200ms - check in logs for GPA.Delta
and set
BRD_BOOT_DELAY,3000
to ensure the CAN GPS is ready by the time the flight controller boots up, so it is reliably detected.

Set these as a starting point

INS_ACCEL_FILTER,10
INS_HNTCH_ENABLE,1  // set this then refresh params to see the rest
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.17
INS_HNTCH_FREQ,40
INS_HNTCH_BW,20
INS_HNTCH_FM_RAT,0.7
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4

and do another test flight - it’s helpful to have AltHold and Loiter in one flight.

Have you followed this process:
https://ardupilot.org/copter/docs/initial-tuning-flight.html#initial-tuning-flight

Yes Sir
We have followed the steps.
And we have also set the initial parameters using Mission Planner.

Thankyou sir for your valuable inputs will do the recommended changes and will get back to you with updates…

It doesn’t look like you have reduced your PID parameters as this part of that page discribes:
https://ardupilot.org/copter/docs/initial-tuning-flight.html#initial-aircraft-tune

It looks like you just have the defaults set by mission planner. Am I missing something?

We had only done the Initial parameter setup in the mission planner.

We have not changed any PID values as we don’t have experience in large copters.

You have an oscillation. You need to reduce your PID’s until you get the aircraft stable.

You should not be putting aggressive inputs in until you have finished the whole tuning process.

Read through that tuning guide and follow it. I wrote up the process I take clients through when I do a commercial tune so all the fundamental steps are there.

Thank you Sir for your valuable feedback

Will follow tuning steps thoroughly

@Leonardthall @xfacta

Hello sir,
We followed the tuning documentation and reduced the pitch and roll PID values.
The oscillations have reduced and the flight is stable in loiter mode for 20 mins.

Please guide me on further steps to tune the copter.

Log:
https://drive.google.com/drive/folders/16ETeYw2N_d1wSWkNyShUTbdHaMktIk4z?usp=sharing

Video Link:

The further steps are in the documentation. Ask specific questions about specific parts of the documentation that are not clear to you.

You didnt set any of the values I specified before, and

Yes Sir

I only followed the documentation and only reduced the PID.

I am worried of changing values because this is my first big copter.

Also want to know whether I am following the right steps.

The parameters and values I gave dont directly affect PIDs or stability - but are more related to taking out noise to allow better tuning.
They should be perfectly safe to set - in fact they are standard procedure now for a new copter, setting up harmonic notch filter as soon as possible.

Thank you for replying

I will change the values that you have mentioned and post the results

Hello Sir, @xfacta

We further changed the PID values and set the following parameters

  • PSC_ACCZ_I and PSC_ACCZ_P
  • reduced the PSC_POSZ_P and PSC_VELZ_P by 40 %

After this, we enabled the notch filter with the following values.

INS_ACCEL_FILTER,10
INS_HNTCH_ENABLE,1
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.125 (= MOT_THST_HOVER)
INS_HNTCH_FREQ,40
INS_HNTCH_BW,20
INS_HNTCH_FM_RAT,0.7
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4

We flew the drone for 2 mins and here is the log.
Please help me in setting the notch filter and reduce the noise.

log file:
https://drive.google.com/file/d/1elvayHQlGm5wLSovWUyYPN2b4eXk--Ig/view?usp=sharing

1 Like

When recommending any parameter changes I have carefully considered these changes and they should suit your copter.

If you do have a Li Ion battery and not LiPo then set these:

BATT_ARM_VOLT,38.30
BATT_CRT_VOLT,36.00
BATT_LOW_VOLT,37.20
MOT_BAT_VOLT_MAX,49.20
MOT_BAT_VOLT_MIN,33.60

and definitely set these regardless of which battery you have:

BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2 or 3

Adjust these:

ATC_THR_MIX_MAN,0.5
BRD_BOOT_DELAY,3000
FENCE_ENABLE,1
GPS_GNSS_MODE,65
INS_ACCEL_FILTER,10
INS_HNTCH_FREQ,48
MOT_SPIN_MIN,0.12

that fence change will prevent you from arming in ANY mode until there is a good 3D fix and Home can be set. It seems frustrating and feels like it takes a long time at first but you get used to it. If you always just used Loiter the effect is the same, so you are not any worse off - just safer. This takes the human error out of selecting a flight mode and arming without a good home position.

There is quite some difference between clockwise and counter-clockwise motor outputs, indicating one or more are slightly twisted on their arms and not perfectly aligned with the rest. Fix that before going any further.

Shorten and tidy the battery wires if they are still the same as in that video. Long = bad.

For tuning, start by reducing these D terms

ATC_RAT_PIT_D,0.003
ATC_RAT_RLL_D,0.003

and increasing logging
LOG_BITMASK,180222

Now do another test flight with AltHold and some gentle stick movements, and maybe some Loiter too if it’s going OK. Please do all the things above before jumping straight to another test flight.

Thank You for your guidance

Will definitely make all the changes specified for the next flight and will test in ALtHold.

Will post the results in both loiter and althold

Hello @xfacta

We tested the drone with all the parameters that you suggested.

Observed oscillations in loiter mode and vibration in the x-axis jumped significantly and didn’t have the confidence to switch Althold.

Can you please help us to figure out what we are doing wrong?

Log File:
https://drive.google.com/file/d/1KYGxVdTmbxEeRcpnt21lhEUl1YGwvWQ0/view?usp=sharing

Video:

The problem is the X axis vibrations - it starts a short time after you’ve lifted off too, meaning something probably shifted. You’ll need to find what is causing this and fix it before more test flights or you could end up with a repair bill.
Have you got photo of the flight controller, mounting and wiring?
Also is the battery very secure? The wiring is still hanging down loose.

Start in AltHold mode or Stabilise, only change to Loiter once stability is confirmed.