Help with PIDs. Blackbox video

[paulshort]

Hi guys. I used to consider myself pretty good at tuning stuff. This one has me confused. Its a Rcexplorer tricopter v3 and t motor 2216 930kv motors running 4s and 8040 props. Here are the PIDs at the time of this video

set p_pitch = 44
set i_pitch = 26
set d_pitch = 25
set p_roll = 47
set i_roll = 30
set d_roll = 25
set p_yaw = 200
set i_yaw = 20
set d_yaw = 33
set roll_rate = 55
set pitch_rate = 48
set yaw_rate = 60
set tri_motor_acc_yaw_correction = 0
set gyro_soft_lpf = 90.000
set tri_dynamic_yaw_minthrottle = 200
set tri_dynamic_yaw_maxthrottle = 70


Here is a link to a video

https://youtu.be/7L36ir8hLV0

Basically it has a slight wobble that I cant tune out. No matter what I do it still shakes slightly. I have tried p gains as high as 7.0 and as low as 3.0

My ESCs are HK20a flashed with SimonK.


Any help would be much appreciated.


Paul

 

[Ocean]

Do you have the raw log that I could take a look at?

I'm not sure which trace relates to which variable in the video.

PS: Nice Caterham

 

[paulshort]

Log file

Here is a link to the log file. Any help would be much appreciated.

(you can just right click and save as)

http://godrone.co.nz/wp-content/uploads/2016/10/blackbox_log_2016-10-24_112711.txt

Paul

 

[PsyBorg]

I am not a tri copter person so I will assume PID's work the same in a tri they just handle control of the yaw axis differently. I am probably way off base because of this but anything I have flown so far usually as long as it is balanced properly likes sitting around the 4-6-9 method for P gains.

Currently your p gains are are 4.7-4.4-20 With pitch so low and yaw so high I see you have D terms cranked way up. I will take a guess you get shorter flights and everything comes down warmer then normal and possibly into the "Hot" range. I think the wobbles you have are D related because they are so high and they are compensating for the lower pitch P gains as you can see in the video the copter is bobbing a lot in a slow steady manor. If it were intermittent oscillations I would say it was high p gains but since it is a constant wobble it is more D related (or something is seriously un balanced to be so strong and steady in the black box like that)

If no one is able to give you a solid answer I would say retune using Skitzo's method after doing a motor calibration and a tail auto tune and start with roll 4.0 20 10 Pitch at 6 20 10 and yaw 9 20 0

 

[paulshort]

Thank you. I will give that a shot. Those are the stock PIDs recommended for a tricopter using 2217 1300 kv motors on 8 inch props. I brought the motors used so they may be a bit shaky. I will have a shot a balancing them maybe.


Paul

 

[Ocean]

Seems like theres some vibration getting to the gyro. Maybe try soft mounting or decreasing the gyro lowpass.

I think this is then effecting the PIDs traces a lot. The D term looks like its working really hard and so is the P term. I say keep these the same and first sort out the vibration issue. Then see how it flies.

 

[PsyBorg]

Easy test for that would be remove the props and run each motor up in software. The gyros will show any imbalance immediately. If it is a prop(s) causing that much vibration it would be seen on more then one axis like it seems to be. You can see the quad tilting up and down in the video. If it were vibrations I would think it would manifest more as jello not a steady pitching like that.

 

[Ocean]

I've had a deeper look into the log more.

It's definitely vibration that's the issue. See the image below, the traces are really thick on both the roll and pitch axis. The yaw seems ok at this stage.

Your Tricopter log:

What an Ideal log would look like (ignore the huge spikes that's roll/flips, look at how thin the lines are in comparison)

As a result the P term and D term are going crazy.

The first thing as PsyBorg said is to find the source of vibration and try to stop it. If this is impossible then set a low pass filter on betaflight. Theres a good section on the wiki about the different types of filters.

What version of clean/beta/triflight are you using?

 

[paulshort]

Thank you so much. It is nice when someone halfway around the world will take the time to help with your problem. I will spin the motors up one at a time and then balance them. I have mounted my FC on foam already. I spent some time this morning tuning it and it flies much better now but there is still an issue of a super slow oscillation. It just ruins any video I try to shoot. Ill post a vid when the weather clears. Dropping the D gains to 10 really helped.


Paul