@dkemxr is right, you need to fix the yaw bias. You can try the filter’s I’ve suggested above for now, but you’ll need to dial them in more once you fix the bias.
2 Likes
SIR, @Allister WHAT DO YOU THINK ABOUT THIS FILTER SETTING? , IS THIS TRUE? " * You must not use a dynamic, throttle-based notch filter on frequencies this low. This frequency is inside the “control bandwidth” of the PID controllers. Filtering it will introduce severe phase lag (delay), which will “blind” the control loop and lead to severe instability, oscillations, or a loss of control."?
The vibration characteristics of the craft will change after you have fixed the Yaw Bias so your priority is out of order.
Looking at the photos of that craft with sq aluminum arms I don’t suppose you can adjust the motor mounts. So, get 2 adjustable wrench’s and twist the arms that are causing the problem. Unless you can do something at the arm frame mounts.
3 Likes
Yes that statement is true.
1 Like
Yes, they’re low. Don’t use them if you’re worried. Fix the yaw bias first, then evaluate the filters. They will change.
1 Like
Sir, I have searched fourm and followed instructions, according to that I did the mapping on floor of propeller tips, and found the deviation to be ± 1mm. I suppose that will not cause yaw bias. Though arm lengths are very long (quadcopter size is 850mm). Does compass misalignment cause yaw bias? Or any other reason other than frame?
No, yaw bias is caused only by twisted motor mounts or damaged props.
Looks like you are aligning in the horizontal plane istead of the vertical plane. No wonder you do not find it.
1 Like
Align the props tip to tip and you will see which ones are out of alignment.
A damaged prop as @amilcarlucas said or a damaged motor can also cause it.
2 Likes
1 - Google Drive, sir it seems the propellers were causing yaw bias or it is hard to tell from osscilations from RCOU if yaw bias is completed eliminated. can you please check @dkemxr ?
It looks like it is corrected. Now set the Motor Ranges as has ben stated before.
And re-set the PSC_ACCZ_P&I parameters based on the Hover thrust which looks to be ~.34
2 Likes
done, can u check if which is best propeller based on these logs(please check only vibe and notch filter fft analysis)? also i noiced the max amplitude after post filter(estimated) is -27 db, so what should be the max amplitude after configurating the notch filter? if the amplitude is still higher than threshould amplitude after configuration, does it mean do i have to do mechanical changes to reduce vibrations at that frequency? can u please help me with harmonic notch filter so that i can do further steps. @amilcarlucas @dkemxr @Allister .
Use the 1247 props, much less vibration and a bit more thrust. There seems to be a frame mode that can be addressed with a fixed notch along with a dynamic throttle notch. I would try this:
Disable Raw logging and run Auto Tune.
2 Likes
Sir, what does frame mode means? Also the quad was hopping up and down when in Hover is this the phase lag?(I was on notch filter setup in amc). Also what should be the max amplitude after applying notch filter so that I don’t have to do mechanical changes?
It’s a mode of vibration that is fixed not changing dynamically with motor RPM. With what you have there is could be anything. After screwing around with that bucket for months I would lower your expectations and just tune it the best it will get and start flying it. Unless you want to make a career out of posting on the forum (which some seem to do).
Well, <50dB but perhaps you can’t get there. As I said do the best you can and start flying. Apply the Notch Filter settings and run Auto Tune. Or, use the “Estimated Post-Filter” feature and determine you own settings and then run Auto Tune.
3 Likes
To back add to what @dkemxr has said, that spike at 32 hz is there and more or less the same with both sets of props. This is a good example of a problem with the frame/build. You’re not going to get this one below -50dB without over filtering. On a good build, I will try to set the least amount of filtering to get to just -50dB. There’s no advantage to going significantly past, it’s wasting resources and potentially inducing lag. With this drone, I’d say that the filter’s that have been suggested are the best you’re going to get and if you want to make further improvements then the it’s back to the drawing board.
Take what you have and go fly it.
3 Likes
For a visual understanding of the issue (and for the benefit of future readers) here is how you can identify a frame resonance.
Using the IMU Spectrogram section of the filter review tool, you can see the darkest band clearly related to the motors’ rpm at 75ish Hz. If you look closely, you’ll see a straight orange line under 50Hz, and note how the line does not change much as motor rpm changes.
3 Likes
Sir, does it mean I have to use static notch filter on it? Also my quad is moving left during alt hold mode and stabalize, can I use rc channel trimming to fix that? If yes, how can I do this from AMC or MP parameters list? ( I want to use autotune but the drone keeps going left so I have to constantly give inputs to keep it one place which lengthens the autotune process. @Joshua_Kordani @dkemxr
No, all RC transmitter trims must be centered an you must never change them.
It moves left or does it drift from the wind?
These are very different things.
1 Like
No sir the Drone is not drifting because of the wind it is tilted on the left side when in air(I did the accelerometer calibration) , also my transmitter does not have function for trim, so how to do it from AMC or MP? Or is there any parameter that we can change?
Calibrate the accelerometer again.
Do not use the trims. There is nothing in MP or AMC to change the trims and is good that it is so.
No parameters… just accelerometer
2 Likes