Unstable Yaw and loiter fly aways APM 2.7

My main problem is trying to get my hex stable while yawing and more recently getting fly aways with no response to rc input at all.

I’ve tried 3.1.5 and now the 3.2rc2 firmware which has made no different other than very different autotune pid results. Today I ran autotune with the 3.2 firmware landed saved the results flew again and pitch/roll are fine but yawing at any speed will cause it to become unstable. The same occurred with stable 3.1.5. changing yaw rate P or yaw stable P no improvement. Everything is balanced calibrated straight and level. The multiwii pro board I was using before on this hex had no yaw issues so i’m assuming its an apm setting.

Loiter, RTL, AltHold, Auto, Land and Circle mode all work fine 90% using a Neo6m GPS. I was getting 150%+ compassmot results with the onboard compass and couldn’t get them any lower even though GPS functions seemed fine. I added an external compass and got 2% right away. logs still say compass = good no mag data ?

I ran a test Auto mission takeoff and landing in a survey grid on 3.1.5, other than unstable yawing it worked perfectly. The next attempt at an auto mission there was a Bad GPS health error on the way to the first waypoint, it then began flying high speed diagonally with no response to user input for quite a way then I got to switch to land mode briefly then RTL at which time one of the motors I recently replaced and forgot to apply loctite came loose causing a flip and crash. Why did it not compensate for the lost motor? What’s the point of having a hex if it can’t handle one motor failure?

The second fly away crash has me puzzled since the 3DR radio link with droidplanner on my phone gave no warning to a problem and tracked the gps position fine. After the autotune flight on the new 3.2rc2 firmware I loitered fine then tested yawing which made it unstable but hold altitude, it then began to casually drift with the breeze toward the only pole on the oval. My rc inputs did nothing and changing to althold still nothing. I heard the phone mention RTL about 2 seconds before the slow motion impact.

Can anyone shed some light on any of my problems?
the log attached is my final crash flight

Anyone have an idea? In the log analyser program I found it says apm believes usb is connect in flight sometimes as well

Sounds like a compass problem. No IMU data and no MAG data means that unfortunately, I can’t really help you, you’d need to change your log bitmask.