I’m trying to tune my dodeca hexa X copter and after a successful separate roll and pitch autotune, the copter went into an out of control Roll/pitch dance up to 45-60 deg. I manage to land it with minor damage. When I look at the PID’s value they were unusual for my type of drone according to what I’ve been seeing here. ATC_ANG_PIT/RLL were around 2.5-3.0 ATC RAT PIT P&I was around 0.090.
So I’ve decided to do a manual tune using the PID’s from a another drone same size, type & model. And now it’s flying way more stable.
So my question is: Is it Ok to leave it like that and If you look at my bin file + param file could you tell me if there is any issue with my configuration. Bin file is a flight in stabilize after the manual tune.
Drone: dodeca hexa X copter version 4.0.3, 18 inches props , 44000mAh 6S.
MinTOW: 30lbs; MAxTOW : 55lbs
thanks in advance
your manual tune is not too bad but could make problem in wind or under pressure
the main reason of why autotune calculated wrong PIDs is noise problem
you can see some noise on your RATE values
first complete tuning process instruction
then configure harmonic notch filtering
and finally run autotune again
THANKS a lot, I will do that and try autotune again.
I’ve went through the tuning process and I forgot to mention but the flight was done under pressure, my payload was 15lbs total weight of 45Lbs.
Hi, I’ve tried autotune again with no payload today and couldn’t even finish one axes! There must be something that I’m missing??? I’ve completed the the harmonic notch filtering & the tuning process instruction again procedure with no luck. here is my files and parameters. Thanks in advance for your help.
Also, is there anyone in the Ottawa region that would be available to help me finish this project, I would like to complete the autotune on the UAV ASAP.
Could you set INS_LOG_BAT_OPT back to 0 and make another short ~1min AltHold hover flight and post that log? Your notch filter settings need a review I think.
ok I’ll try that tomorrow
I would set the notch parameters this way:
Set for post filter logging and make another flight for review in anticipation of another Autotune. Of course there isn’t much pitch/roll demand in this hover flight but what’s there doesn’t look bad (desired/actual). There does seem to be some Yaw wandering. Perhaps you should disable Compass Motor Compensation (COMPASS_MOTCT,0) while you are tuning. I never use it anyway. You can always re-enable it, the offsets will still be saved.
Hi Dave, thanks a lot for your help, I’ve done exactly what you’ve ask. Keep in mind that I’ve done the Notch filter procedure before FREQ: 128, BW: 10 and ATT:15
This is what it look like before the Notch filter:
and this is after:
So now I’m afraid to loose the filter for the 128Hz… IS it possible to apply two filter or can we just put a larger BW???
Yes Eric I know that, I saw it in the parameter file. I don’t think this will be very effective. You are missing the 1st mode, the attenuation is way too low and so is the bandwidth.
You won’t “loose” the 128Hz peak (it’s closer to 136Hz) it will be picked up as the harmonic of the 68Hz 1st mode. It’s bandwidth will be double the 1st mode or 40Hz from my suggested BW parameter.
But to answer you question about anbother filter, yes there is the static notch that can be used but I don’t think you will need it. Your 1st mode and the next 2 harmonics line-up well to be handled by the dynamic notch.
OK good, so I guess I’m ready to try another autotune?
I’m not sure what you mean by “OK good”? So you reset the filter parameters, made another hover flight with post filter logging and have reviewed it?
Ah, the log you posted above has those changes? Sorry, didn’t pick up on that. I’ll take a look.
By Ok good, I was acknowledging what you were saying. The filter parameters were set to your numbers before the flight that I’ve just posted.
Go it. Should have looked at the log 1st before commenting.
This is interesting. The filter did a good job on the 2nd mode (~130Hz) as expected and there is nothing beyond that to speak of. The gyro magnitude is very low so looking great. The 68Hz mode on the accel is still there but at greatly reduced magnitude. This is probably as good as it will get with the isolation you have on the Flight Controller. I would say you are good to go with another autotune. It’s hard to say what Aggression level you should start with. I suppose the .08 that you have configured is good.
Thanks, I will try autotune Tomorrow depending on the weather.
So yesterday I’ve tried an autotune again, there was a little bit of wind (5-10 km/h) and I was able to get to 3/4 on the roll but couldn’t complete the autotune on one axis (roll).
Here is the file:
Not much can be learned from that log and the PID values were not saved. You will have to try again.
I thought you would just save the PID’s if autotune was succesfull. I’ll try autotune again tomorrow weather permitting.