Recently i have made an new quad and is in the process of going to do the tuning, but as a part of my first test i have to make sure that the quad is flyable atleast in altitude hold so that auto tuning could be done. But normally my practice is that i fly in alt hold if all goes well then fly in loiter( if flyable) to i do a quick tune first then followed by auto tune in althold.
however i have found some strange behavior where the drone will suddenly tilit backward and lost altitude. checking the logs i could not figure out the issue because i checked the pwm out and it looks normal, motors are in pwm out ch13,14,15,16. could it be a fault in the hardware? or perhaps on my configuration or some settings (like perhaps i should us the RPM instead of fft to drive the filter or some others )?
my setup is as follows:
ESC - blujay (for bi-directional dshot)
Motors - f90 7 inch prop
FCC - pixhawk 6c
i appchiated any and all help.
please find the attach logs:
Edit: i have observed the motors get really hot
i have done a second flight with FFT disabled (filter using esc telem . rpm via bidirectional dshot)
same behavior is observed
Sure they will with the output oscillation you have plus it’s a bit underpowered/overweight. A 7" will not fly on default PID’s and vertical acceleration controller parameters. This craft is in no shape to run either Quick Tune or Auto Tune.
And yes, always use ESC telemetry as the Notch reference if it’s available.
Couple other tidbits:
Motor Ranges are at default.
You have ESC errors maybe because the Dshot loop rate is too low.
You do not need this configured SERVO_BLH_MASK
This should not have left the ground in this state.
Please choose only pre and post logging raw logging enabled.
Review your filtering, for multi-source filters the width and attenuation should be 1/4 of frequency.
Thank you for pointing me the right way. Much appchiated , i totally made some mistakes and have followed AMC with no issue. now there is a large difference between what how it fly’s but the tendency is still the same (loss in altitude and hot motors) i suspect is the hot motors that causes lost of thrust and fall (since no trust loss warnings are present). i have a question what can be causing that phenomena?
shown bellow is the new PWM graph after following AMC
so the oscillation has been drastically decrease. i also used RATE.*out to check if any pitch/roll/yaw is averaging above 0.15 and it does not look like it. i checked using despitch vs pitch, desroll vs roll, desyaw vs yaw and the PID is tracking okay (should be enough to atleast maintain hover with pilot feed back)
i have also changed the filtering using FFT values to have a better value. (band with and frequency is from FFT)
regarding this i found that without servo_blh_mask enable bidirectional Dshot does not work, only if sett to the corresponding motor ch does rpm show on ESC telemetry. following your recommendation i have used throttle to drive the filtering instead
regarding this yes i noticed but i couldn’t find the source, what i suspect is the vibration from the prop guard does it pose a significant problem? this is the new filtering graph:
You have disabled RPM filtering and PIDs logging, it was working before, I recommend to configure it as it was before.
To have bdshot working usually you just need
MOT_PWM_TYPE
SERVO_DSHOT_ESC
SERVO_BLH_BDMASK
