You have your PIDs completely untuned. The default PIDs, which you use, are quite OK for a 9"-sized quad, but typically way too much for a small high-voltage quad.
Could you please provide:
the kV rates of your motors;
the overall weight of the craft (with a battery);
the pitch of your propellers?
I will try to scale the tuning of my 3" quad to your setup, so that you (hopefully) have a reasonable starting point. (After that, some adjustments, including running an autotune, will still be necessary, but the initial flight experience might be better).
ATC_RAT_YAW_P: 0.18 <— these seem to match the default ones, coincidentally
ATC_RAT_YAW_I: 0.018
Note that:
The PID values are basically my PIDs scaled by 0.4, and this 0.4 comes from my idea of how that would scale depending on kV of motors, prop pitch, craft weight and voltage. Since our frames are very similar, this might do the trick.
The ACCEL values are, again, my values scaled by the weight.
These values implicitly assume that you already set up notch filtering, in particular the one using bidirectional DShot. If not:
You may want to decrease all PIDs sharply, e.g. by a factor of 2, if you still experience high oscillations at startup.
I would strongly recommend to set up filtering anyway. This will allow for much sharper tunes and better flying overall.
It’s never going to fly with those default Rate and Gyro filter values with a 3" craft. Go to Mission Planners Mandatory Hardware>Initial Parameter Setup.
Could you maybe please read my entire post, and not just the parameters?
Since you don’t have harmonic notch set up, these particular PIDs may be too high for you. The video rather supports this. Please divide all the _P and _I values by two and repeat the test. If it gets noticeably better, but not flyable yet, repeat that.
And yet again: the default PID values, which Mission Planner seems to set up for you, are too high for your craft!
The filters, which @dkemxr was speaking about, are the following values:
which can all be safely set to 50 (and even to 100), and
INS_GYRO_FILTER
which can be set to 100 (and even to 200 together with the parameters above).
Please use these values for initial tests (50 and 100 correspondingly).
And a final remark - don’t think that Mission Planner knows what to do! It typically does not. Use your brains instead. If you cannot, for any reason, choose Betaflight.
Glad to see you have some attitude control at least.
I don’t see how MOT_SPIN_MAX would affect your flying, since in the provided logs your motor outputs never exceed 0.5. That would be really interesting if that does affect your flying. I would rather expect changes from e.g. MOT_THST_EXPO, which is usually much better to be around 0.65 than what default settings suggest.
Just as a side note, Ardu resets the battery consumption counter on powering on. So if you did some tests and then power-cycle without replacing the battery, you either want to adjust the battery capacity, or track the voltage more closely, so that you don’t drain your battery completely, as it was in your last test.
With the current state, you might want to go and run the autotune. It is best done in a nearly still weather (I know not everyone can afford that), and in your conditions, I would find some nice area with soft surface so that a crash would harm less. I would recommend doing your tune one axis at a time, given that an initial tune goes much longer than a tune starting from an already good point, and that your battery is not “infinite”.