I was flying down to pixhawk2.1. Can you take a look at the flight log?

I made a 1100 class HEXA drones. I installed a 21.5 inch prop and I used a motor of 185kv. I used APM: copter V3.5.4 HEXA. and Pixhawk 2.

During the first flight, we performed roll, pitch autotune, and on the second flight we proceeded to All Autotune. I have successfully completed two autotunes, but the problem appeared on the third flight. I started in althold mode when I took off, and when I switched to loiter mode, there was a problem with the drones.

I think this problem is caused by the abnormal temperature of the IMU sensor.

Log.zip (675.5 KB)

These do not look like successfully autotuned numbers.Roll looks high (but my 960 started off in the .66 region before dropping to .44.But Pitch is set exactly the same as Yaw ( Nope - I is set at 2 and not 0.2).Somehow those numbers have got corrupted.I suggest investigating around there first.Those numbers could explain why three of the motors were bottoming out in flight.And why it flew like a turkey.

PARM, 61260221, ATC_RAT_RLL_P, 0.7362748
PARM, 61260262, ATC_RAT_RLL_I, 0.7362748
PARM, 61260305, ATC_RAT_RLL_D, 0.03575641
PARM, 61260349, ATC_RAT_RLL_IMAX, 0.5
PARM, 61260393, ATC_RAT_RLL_FILT, 20
PARM, 61260437, ATC_RAT_RLL_FF, 0
PARM, 61260824, ATC_RAT_PIT_P, 2
PARM, 61260873, ATC_RAT_PIT_I, 2
PARM, 61261052, ATC_RAT_PIT_D, 0.2
PARM, 61261102, ATC_RAT_PIT_IMAX, 0.5
PARM, 61300831, ATC_RAT_PIT_FILT, 20
PARM, 61300882, ATC_RAT_PIT_FF, 0
PARM, 61300941, ATC_RAT_YAW_P, 2
PARM, 61300995, ATC_RAT_YAW_I, 0.2
PARM, 61301059, ATC_RAT_YAW_D, 0
PARM, 61301120, ATC_RAT_YAW_IMAX, 0.5
PARM, 61301612, ATC_RAT_YAW_FILT, 5
PARM, 61301674, ATC_RAT_YAW_FF, 0

Your VCC varies a bit wildly.You could really do with some current monitoring in a beast like that.

The abnormal temperature of the IMU ? They are heated ?

Flying at temperatures below zero. As soon as I switched the power on, I took off.
Do IMU sensors require preheating?

It would help, apparently.Getting info about the Pixhawk 2.1 is like extrcting teeth but they do claim that you can fly below the ice point with it.Once fully heated up I guess that means.Operating temperature is 60C.

Lipos don’t perform very well at below zero.DJI provides battery heaters because of that.In the open source wortd you have to make your own,or insulate,or pre-heat.Heated bags are available as well.

Ignoring that for the moment - it’s your PIDs that look to be the problem to me.The cold may have been a factor but that copter won’t fly very nice with those pitch PIDs.

Try setting them the same as the roll axis.That should get you a bit closer to steady flight.
PARM, 61260824, ATC_RAT_PIT_P, 0.7362748
PARM, 61260873, ATC_RAT_PIT_I, 0.7362748
PARM, 61261052, ATC_RAT_PIT_D, 0.03575641

And I’d advise to get a power module fitted and calibrated as soon as you can.On a beast like that you really need to know when the battery is failing.It also shows certain problems up in the logs.

Also set the ATC_RAT_RLL_FILT, and ATC_RAT_PIT_FILT, to 10.Halving the value.This is recommended for large copters with big props.You’ll get a better tune and feel from doing that.If you change the pitch values to match roll as well it should at least stay in the air.

The PIDs are the problem here.If it still performs badly after balancing those then it’s time to dig in the logs again.