@kuspower Sorry that I haven’t had a chance to look at your logs.
Yes I think the ATC_ACCEL_Y_MAX could play a big part in the response you see. The default, as I have learned through my work with autotune, is very low for helicopters. This default was inherited from multicopters. I think you could raise this considerably. Probably over 60000. The only concern when doing this is that the Rate P and Rate D tuning that you have done could be too much and you wouldn’t know it because of the accel max being so low. So incrementally raise the accel parameter and just be aware that you may see it oscillate as you raise it. if so, then reduce your P or D gain until it goes away. However I think your gains are pretty low already.
I was pretty surprised in using the autotune code to see how high I could raise my P gain for yaw on my 600 size heli. What I interpreted as oscillations when I manually tuned the yaw axis were not a sign of instability but just poor tuning. Before I used autotune, I had stopped my yaw P gain tuning at 0.22. With the autotune code, it raised the gain to 0.44 with no sign of instability. It was a huge improvement on the aircraft yaw control. As far as the max yaw acceleration, it predicted my max achievable yaw acceleration to be over 100000.
I only mention this as I think raising the max accel should not cause instability problems and will definitely help with your yaw control response.
I will try to look at your data tonight. I am in the midst of cleaning up the autotune code to prepare it for merging by the new year. So I have been pretty occupied with that.