Ac4.0.7 Notch filter based in throttle issue

Activating notch filter based on frequency reference based in throttle conduct to behavior as throttle increased during flight due to battery discharge … reality is that mean mot rpm is the same during the whole flight…

Istead of using fft ref based on throttle it will be more relyable to use reference based on watts…

What is the normal throttle beahavior?
Increase during flight due to battery voltage drop?
Or normaly throttle isnt affected by battery discharge when mot trust expo is correctly set?

I think what you are saying is:
As motor thrust and commanded RPM is affected by battery voltage drop, MOT_THST_EXPO keeps the expected thrust and MOT_THST_HOVER “the same”. So that MOT_THST_HOVER isn’t constantly changing due to varying battery voltage.

Are you proposing HNOTCH be based on battery power usage? I think that would essentially be the same as throttle except maybe less direct and less accurate. Throttle-based makes sense since throttle will equate to RPM - and RPM (motors and props) is the source of vibrations and noise.
Don’t let me stop you from following up on your thoughts and questions. @andyp1per is the developer for the Harmonic Notch filter and he can definitely explain much better than me.

Are you setting up Harmonic Notch filtering and you have more questions or need help?

The Harmonic Notch filter seems very reliable and is in widespread use, with the two most common modes (that I am aware of) are:
INS_HNTCH_MODE,1 throttle based
INS_HNTCH_MODE,3 ESC telemetry based

“Throttle Out” which is what the harmonic notch uses is not the same as “throttle”, throttle out is adjusted for battery voltage sag already so its pretty much as accurate as it can be for an approximation. ESC telemetry and FFT are more accurate and so recommended for that reason - there is no need to create another approximation.

I’m not sure I understand the point being made. Expo is a linearization function not directly correlated to battery voltage as I understand it.

Yeah… MOT_THST_EXPO is about relating RPM to actual thrust output, I think.
And the RPM will drop as battery voltage drops, for a given load on the motor. So they are all interacting.

RPM and thrust are always the same ie If the prop is spinning at x rpm it will always produce y thrust (assuming no changes in atmospherics). Thus MOT_THST_EXPO is not related to this. What it actually relates to is that the ESC output verses the ESC’s input. So if you have a motor that can produce 100lbs you might expect that a throttle set at 50% would produce 50lbs of thrust but this is not the case more often than not. It may actually be 60lbs of thrust at 50% throttle. This will vary on the ESCs used. What MOT_THRST_EXPO does is try to better match the ESC’s actual thrust to throttle settings thus creating better control of the thrust output. Well, that’s my understanding of it anyway.

For the harmonic notch what this means is that if the prop max speed is 1000rpm then without the EXPO function it might expect the motors to be spinning at 500rpm when throttle is set at 50%. But as we now know the motors may be actually spinning at lets say 600rpm. The expo function helps the FC realise that in this example the 50% throttle translates to 600rpm not 500 and thus the vibration filter is moved to match this better computed rpm.

Ok thank you so much for this explanations.
So as I understand now, ThrO is the reference for hnotch filter and ThO is the same during the whole flight.

ThO dont vary due to battery discharge as there is compensation for it, so it’s good reference for hnotch filter.

Motor RPM (ESC telemetry or Bi-directional Dshot) or FFT (in Master) is a better reference for the Notch Filter.

Thanks a lot Angus, that was the explanation we were look8ng for.