Copter is losing orientation on aggressive turns

ArduCopter Copter 4.5
1

Hi all. Still fighting with AP tuning.

Problem is when I’m frlying in ACRO copter is losing oruentation on fast(agressive) turn (this is NOT PROP wash). I’ve increased ATC_THR_MIX_MAN = 3.5 - does not help! :frowning:

I’m not sure about filters configuration. I do have 3.5" 4s quad in about 310g weight with battery.

It’s flying OK in Betaflight. I’ve lost in about 3 weeks trying to find the problem. And now I’m thinking to go back to BF :frowning:

Can somebody please advise?

Logs: 5 01.01.1970 1-00-01.bin - Google Drive

And thanks in advance!

2

I’ve done lots of series tuning for acro in particular - presumably you have looked at them? https://www.youtube.com/playlist?list=PL_O9QDs-WAVzX7e3-P9Ofv4ZPQTWzR8vs

1 Like
3

Hi man!
I was looking your series about chimera (thanks for that!) - I have it too :slight_smile:
I was not looking on this specific series of your videos. But was trying to do in same way as you did with chimera.

Problem for me I did not get how to define right filtering. So I do have (in my opinion) good chart - but I’m not sure.

4

The noise looks ok. I think you could go narrower on the notch and the reduced phase lag will let you get a tighter tune. Given that you are using ESC telemetry you should be able to use FREQ 40/BW 10 and it all work as intended.

Your problem simply sounds like tuning. Fast aggressive turns generate a lot of noise which can affect the attitude controller. The control doesn’t look bad. Try the narrower notch and retune. What AGGR setting did you use when you did autotune?

Also you have Yaw D set but no FLTD for yaw which means you will get noise coming through yaw - set yaw FLTD to GYRO_FILTER/4 and retune

5

Hi Andy

I’m using AGGR = 0.1. Then tried with manual tune to increase P - was no lucky :slight_smile:
Thanks for that. I’l try to update filter. And let you know

What do you think about GYRO_FILTER - is that OK for that quad?

6

Try AGGR 0.075 when you have made the other changes - you may have also created instabilities through gains that are too high

7

Thanks - I will try. Instability was before manual changes :frowning:

8

Nothing changed :frowning:
is still trying to tip the drone in the opposite direction when making a sharp turn.

The situation is slightly improved by ATC_THR_MIX_MAN = 4, but this is the maximum value that the AP allows to use (resets to 4 on reboot)

Logs from the next flight right after autotune

9

You have a fairly pronounced spike at 182Hz:

image
image1318×510 65 KB

Most likely this is frame resonance since you are blasting about and your RPM is well away from this. This will prevent you increasing your filters any further so I would use the second harmonic notch to take this out. When you have done that you should be able to raise your filters so that you can get a tighter tune.

2 Likes
10

Hi Andy, thanks for that.
I’ve tried to change props on smaller ones from 3.5 to 3. Looks like you are right.
I’ve got something like this:

image
image1603×1679 319 KB

I hope that’ will help :slight_smile:

11

Multisource on the fixed HNTC2 is just strange.
And setting only the 6th and 7th harmonics at 6 * 39 = 234Hz and 7 * 39 = 273Hz respectively is also strange, but it does seam to work…

1 Like
12

What motors do you have?
I’m working on the harmonic notch filter for you.

You will want to set:
ANGLE_MAX,6000
and maybe dial back some logging, and add some normal things in
LOG_BITMASK,407291

1 Like
13

Hi
@xfacta : 1404 IFlight XING 3000KV 4S.
Thanks

Looks like main problem was gone after static filter. I increased ATT to 40DB

image
image1920×1026 77.5 KB

The only thing I’m thinking was it over filtering?

image
image1619×1897 350 KB

14

That motor is 12 Poles, you will need to set:
SERVO_BLH_POLES,12
and I would try this filter configuration with a flight to see what the log is like then:

INS_HNTCH_BW,20
INS_HNTCH_FM_RAT,1
INS_HNTCH_FREQ,80
INS_HNTCH_HMNCS,1
INS_HNTCH_MODE,3
INS_HNTCH_OPTS,2
INS_HNTCH_REF,1

It is nearly the same as you had in the last log but just standardised/minimised.
The number of poles explains why even in steady flight the notch was not matching the noise peak:

image
image1000×380 23.3 KB

but it should be good now.

2 Likes
15

Thank. I was double checked nomber of pols - in some kind of reason I was missed N and P :person_facepalming:
Next few days I can’t test changes - because hawy rains here. I’ll let you know when I’ll get results.

Thanks all

16

So. Looks like problems was in number of poles. After setting correct value - it’s flying well!
I removed static filtering. Changed notch to 80/40. And no spikes after filter (even with 40/10)

@xfacta , @andyp1per thanks for your help!

1 Like
17 
INS_HNTCH_BW,20

Use 80/20 at most if using per-motor notches
INS_HNTCH_OPTS,2

1 Like