But it still seems like it is rolling to the wrong side during take off - If I check the log file, I can see the setpoint and estimate are going opposite ways.
I noticed on the ground some issues where the drone would vibrate quite a bit if I moved the throttle to almost takeoff power - Investigating a bit more led me to checking the motors, and it seems like one of the motors have a defect winding (it was 57mOhm where all other windings for the rest of the motors were 20mOhm).
I have replaced the motors with new ones that seems to solve this issue and it does not flip on it self now. (Now it just shoots up 10 meters when I try to take it up to 1m - after googling a bit I suspect I have to adjust MPC_THR_HOVER to a lower value since I am running a 6S1P any require little throttle to hover)
Got it flying with MPC_THR_HOVER at 20% - I think I can set it a bit higher, but it took off after giving it a little extra throttle.
I would say that it is not flying stable enough for PID tuning yet, but it is a start. I also mounted the H-Flow optical flow sensor, and I can see that sensor is working Now I just have to tune a few parameters before PID tuning I guess? My first guess is looking at the MC_[ROLL|PITCH|YAW]RATE_K