AutoTune: failing to level, manual tune may be req

I had a question about this - I would guess this set of params is for a S500 exactly as HolyBro recommends it with their kit? I have used their physical kit (frame, ESCs, motors and props) with both a Pixhawk “1”/2.4.8 clone, as well as a Matek F405 FC, and the params preset was unable to take off in both cases until I reduced the PIDs by 50% per the AP Wiki. I am still learning a lot about this, so forgive any ignorance, but the only reasoning I can think of behind this would be a)different flight controller from what was used for the “preset” b) different flight controller mounting configuration (though vibes were under limits in both cases) or c) I am flying with a 3S 5000mAh battery while HolyBro recommends a 4s battery. I would think c) to be the most likely cause, but could anyone with more experience share their thoughts on this? I could post some logs if that would be helpful at all.

Please open the log If you used 3S batteries in the 4S recommended set, it looks like the drone is probably not getting enough thrust to fly.

Have a look at your motor outputs. If they are high or oscillating that would be a clue. I agree that 3S vs 4S is probably part of the issue.

Well, it seems that was the issue! Flew the S500 with the “preset” params and a 4S 5000mAh battery today and it flew very well. Thanks for the pointers @rk2cy_u and @Allister .

It has been flying decently on 3S as well, but only with the adjusted PIDs, which makes me believe it’s just a matter of the motor speed being slower overall on 3S, making the motors’ responses to motor speed changes different, requiring the different PIDs. At least as far as my limited knowledge of PIDs goes, this makes sense to me.

I will be looking at the logs as well to check the thrust issues you both mentioned.

2 Likes

There was a test time, so I reviewed the parameter modification before proceeding, and confirmed that I did not learn about the harmonic notch configuration.

As @Leonardthall advised, is it better to reduce the pitch feed by 50% first?
Or do I have to set up a harmonic notch and tune it automatically?
I’m a little confused. :cry:

In fact, my qudeocta can be controlled stably, but the mount pulled forward to attach the gimbal shakes up and down (low amplitude).
Can you fix the problem by setting the harmonic notch?

It is written in the wiki documentation, and it has been repeated countless times in this forum, the correct way to properly tune a copter is:

Just follow the steps, without skipping a single one, and you are golden.

1 Like

I think there is a bug in new autotune for H-Copters…
I have done manual tune . And drone is flying ok. After that when i try autotune, it shows error, that autotune failed… Unable to level, please tune manually.

Perhaps it isn’t. Post a link to .bin flight log.

Please uncompress the file.
waiting for your replay. :blush:

Can i install an old version of apm … Complete the autotune… Copy the pid… Paste it to new version???

Do not go back any versions for the problems you have at the moment, just fix what is wrong now - the issue will be the same even if an older version does not warn you of the problem.

Fix these battery settings, and do not try to go too low with you battery voltage or it will be permanently damaged. If these give you trouble, then replace your battery or manually calibrate your voltage sensor - do not change these values to mask the problem.
BATT_ARM_VOLT,14.70
BATT_CRT_VOLT,14.00
BATT_LOW_VOLT,14.40
BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2
MOT_BAT_VOLT_MIN,13.2
When calibrating the voltage sensor, use a low reference voltage like 10.0 since the low voltages are far more important than the high voltages.

There’s lots of “noise” and motor outputs are oscillating badly. Set up the Harmonic Notch filter next.
Start with these:
INS_LOG_BAT_MASK,7
INS_LOG_BAT_OPT,0
INS_HNTCH_ENABLE,1 ← set this then refresh params to see the rest
INS_HNTCH_ATT,40
INS_HNTCH_BW,40
INS_HNTCH_FREQ,80
INS_HNTCH_HMNCS,3
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.15

I would also try these PIDs as a starting point
ATC_ANG_RLL_P,10.0
ATC_ANG_PIT_P,10.0
ATC_ANG_YAW_P,6.0
ATC_RAT_RLL_P,0.12
ATC_RAT_RLL_I,0.12
ATC_RAT_RLL_D,0.0075
ATC_RAT_PIT_P,0.12
ATC_RAT_PIT_I,0.12
ATC_RAT_PIT_D,0.0075
Take off and try to hover just for a very short time, then land and check motors are not hot. Listen for any strange mechanical-sounding noise. If that happens then reduce ATC_RAT_RLL_D and ATC_RAT_PIT_D to 0.005 and try again.
It might also be worth reducing these filter values until everything else is more sorted
ATC_RAT_PIT_FLTD,15.00
ATC_RAT_PIT_FLTT,15.00
ATC_RAT_RLL_FLTD,15.00
ATC_RAT_RLL_FLTT,15.00
ATC_RAT_YAW_FLTT,15.00
INS_GYRO_FILTER,30.00

After setting all that Let’s see the .bin log from a hover test, with maybe just a few gentle movements.
From that we will be able to set the Harmonic Notch filter more accurately, and hopefully move on to Autotune.

3 Likes

Thanks a lot bro… i will change the parameters and try a Hover test as u mentioned.

It worked :heart_eyes::heart_eyes::heart_eyes: . Autotune worked and now i have a better stability… Thanks a lot

I just stumbled across this thread as I experience the same error.
I have mounted the cube carrier on the 3M hard foam that came with it, could that be the cause of the problem? I think the foam is quite stiff so I wouldnt think thats the problem, but maybe it is? or is the foam that came with it fine to use?
18" 340KV 6S 4.5Kg

ok, in my case it was not the damping foam. I have now hard mounted the cube carrier using screws and I still have the problem.
I am doing it exactly like in the guide and im at a point where it seems to get better. I still get some errors though. next time I will test it with the harmonic notch filer in place.
a little question: my PID values seem to be really low now, as I decreased them a lot to remove oscillations. can somebody tell me if I can go even lower on them without entering very dangerous territory? D=0.001152 P/I=0.054

Post a link to a ,bin flight log.

@dkemxr here it is: Microsoft OneDrive - Access files anywhere. Create docs with free Office Online.

You should configure the Notch Filter with a center frequency of a 100Hz and BW at 50Hz. And set INS_LOG_BAT_MASK to 1, you don’t need to log all IMU’s it just produces very large log files.

I have already set it exactly like that, just not yet flown. will do that on monday. Other then that it should be fine right?
thanks!

this was flown with the notch in place, and autotune completed although there still were some messages. the only thing I noticed was the autoanalyzer complaining about large compass offsets… is this a thing? or are they ok?
also, I see another frequency peak at about 30Hz, would another notch filter make sense here?