FFT results - could the FFT DEV's please take a quick look?

ArduCopter Copter 4.3
1

I’ve been tuning a new 690-cass quad, with the gracious help of @xfacta and others. We’ve had some discussions about my FFT notch filter working as it should. It’s a light quad - with a high thrust-weight ratio.

I understand that there’s work in progress for improvements to FFT - so if there is indeed something peculiar here, maybe this example is helpful input.

If FFT is working as intended - or if maybe I have the parameters set wrong - I’d appreciate any guidance that might be available.

Thanks!

Here’s the Mission Planner FFT panels: (Please note that I zoomed in on the graphs.)

fft panels
fft panels1920×1269 145 KB

Here are the FFT parameters:
fft parameters

And here are the HNTCH parameters:
harmonic notch parameters

The log file for this flight is uploaded here:

2

@andyp1per Hi Andy, here’s a sample of FFT not getting down to the base frequency, 50Hz in this case, and instead focusing on the obvious peak at 100Hz.
Let us know if there’s a reasonable way to improve FFT to capture that base frequency, or should we wait for a future improvement.

3

If you can try 4.4 you should - it has significant improvements here

4

I look forward to giving it a try.

Is the issue in my situation typical of the problems 4.4 is correcting? Or is it an outlier?

5

Its typical of the kind of thing I fixed in 4.4

6

In the meantime the RPM referenced Notch Filter works Great!