Effect of weight on tuning

ArduCopter
1

I fly a F550 with Pixhawk 2.4.8 running copter 4.0.5.

In a recent flight in circle mode, I noticed quite a bit of twitching in roll.

I wonder if this might be due to a difference in the weight of the battery I was flying from the ones I used for auto tune.

The battery I was flying for auto tune weighs about 430 grams.

The battery I flew for the flight that displayed the twitchy behavior weighs about 645 grams.

Batteries are located atop the top plate of my F550 oriented across the roll axis.

2

Weight distribution affects the flight.
Especially if added/removed weights are above the center of gravity, you may see some twitching on both roll and pitch axes.
For a detailed look could you please share your .bin logs for both flights?

3

Mustafa, thanks for responding.

Sorry for the delay. The demands of work have kept me from responding until now.

It would be difficult to identify the log files from the autotune, but attached is the log from the flight in which I observed what appeared to be overly “twitchy” behavior.

https://drive.google.com/file/d/1tGoWXZvPOMV8WMUTjvyQw4GGT-H16xBw/view?usp=sharing

1 Like
4

@Mustafa_Gokce do you have a minute to examine the log posted above?

If you have the time, the analysis of someone more experienced than I would be valuable.

5

In the logs I see a problem with vibration levels.
Check how the FC is attached to the frame, check the propeller balance and also, wiring could induce vibrations.

Then check all the initial parameters, following the tuning process instructions. Your MotBatVoltMax/Min parameters are not set and are important for PID.

Then after checking any mechanical root for the vibrations, Tune the notch filter and finally redo the tuning process.

6

Hello sir.
You have a considerably small quad considered those I am currently working on and first I notice you have a high vibration on Z-axis. And also have some vibration on the X and Y axes, too.

vibes
vibes1677Ă—889 30.8 KB

I recommend you first need to do Harmonic Notch Filtering.
https://ardupilot.org/copter/docs/common-imu-notch-filtering.html#setup
When you do this and have a quick loiter flight, we can see the vibration’s frequency and magnitude.
Please follow the tutorial for this setup.
You have also PID issues on both roll and pitch axes.
roll
roll1667Ă—905 25.9 KB

pitch
pitch1661Ă—933 27.2 KB

They are not correctly set.
You need to do AutoTune after setting filtering.
https://ardupilot.org/copter/docs/tuning-process-instructions.html#tuning-process-instructions
https://ardupilot.org/copter/docs/autotune.html
I see that ardupilot able to fly the vehicle since you have a good altitude holding on flying modes those you are using.
alt
alt1669Ă—899 18.6 KB

So you can do AutoTune after filterings are done.
https://ardupilot.org/copter/docs/common-imu-notch-filtering.html#setup
After you set the parameters for harmonic notch and logging, please post the flight log *.bin file so we can look at, and set the correct parameters for harmonic notch filtering.
After that, I think you can do AutoTune since it can fly safely even if there are some twitching.
Please note that if you have a limited battery, do AutoTune at the axes, separately. For example, roll first, then pitch, and then finally yaw.
And also note that you should wait a calm day, no wind.
Your propulsion system seems to OK, not underpowered so you are good to go.
motors
motors1669Ă—921 35.6 KB

I see that the first motor has more output than the average of others, I don’t know whether this is an issue or not.
Maybe others can comment on this.
Also please check whether or not your propellers are correctly balanced.
I hope this helps.

Mustafa

7

@Mustafa_Gokce

Thank you sir for taking the time to write such a detailed analysis. I will act on your recommendations as soon as I have the chance to fly again.

One question… My Pixhawk FC is attached to the lower plate of my F550 directly with double sided tape. Should it be on some sort of vibration isolating mounts?

1 Like
8

@andresrc thank you for taking the time to examine my flight log. I will attempt to act on your recommendations as I have time to implement and flight test them.

One question… My Pixhawk FC is attached directly to the lower plate of my F550. Should it be on a vibration isolating mount?

9

There are some foam kind of thick double sided tapes. You can use them. If your FC is pixhawk (not pixhawk cube) you may also try those mounts.

10

Yes, I would go also for the thick double sided tape, but if you can, check first the mechanical source of the vibration. Maybe propeller balance…

11

I think the F450 and F550 clones are known to have a lot of vibration and arm flex, so you will probably need quite a good dampening, or maybe dynamic notch filtering.

1 Like
12

I don’t think foam tape will cut it for these frames. As @Vabe says you need a more compliant mount. The ball plate mounts work OK. And Notch filtering should now be routine in the tuning process.

13

Thank you all for your replies. I will look into notch filter settings, which I had previously not looked at it applied, and repeat the autotune procedure after doing so.

14

This will not fix a problem with basic high vibrations. The notch filter is not even active on the accelerometers, Gyro only.

15

@Mustafa_Gokce Not sure it is relevant, but your graph of the motors appears to show only four. My vehicle is a hexcopter.

1 Like
16

Sorry, my bad. Everything is fine.