Thank you Thomas.

I ran the Spirit BEC tester dozens of times on this heli, without any signs of weakness. Servos performed exactly as expected every single time. Rx and Spirit voltage remained very consistent every single time. Lowest voltage I logged was around 7.65v for an instant. As you know, the BEC tester moves the swash up and down, and also the tail serve left and right progressively faster until its conclusion. No cyclic movements though. Always +- collective only, so the swash should stay flat throughout the test. If the swash tilts, that would be an indication of a bad servo. This never happened in all my tests. Always 100% flat swash. I even used my hands and a towel on the main blades to give it resistance during the testing. Perfect every single time. Sure seems that if I had a problem servo or servo's even, it would have reared its head in these tests.

Also, all 3 of the incidents exhibited full + collective and then full cyclic. Not just a full cyclic in one direction either. It was a rotating cyclic; rotating back elevator to left aileron into the ground in this case. Like stirring the cyclic in piro maneuvers. Except I was not telling it to do that. This seems to be a coordinated effort, and would rule out a servo misbehaving randomly.

In these pics, you can also see how I used the 3M tape to secure and as a vibration barrier (covered up with black electrical tape) between Spirit and the wire as it is routed along the side. In the first two incidents, I was using the stock Spirit cable that is about 6” long and smaller gauge wire. The latest setup was using high quality wire from a Kontronic BEC lead custom made to be as short as possible (minimize losses and interference possibilities). Also, all servo leads were secured with shoe-goo before flight.

control path:
transmitter -> receiver -(EX Bus)> Spirit -> cyclic servos -> main rotor

If the problem is not:
Transmitter - works with all other models
receiver - REX7 #2 in this last event replacing first one used in the first 2 events.
EX Bus cable - stock for 1st 2 events, new larger gauge and shorter custom cable for last event.
cyclic servos - tested OK during dozens of BEC tester sessions with load added by my hands.

That leaves the REX7 firmware(v1.0), the EX Bus protocol and the Spirit unit as the only possibilities for coordinated un-commanded control of the cyclic and pitch like in all 3 events.

What perplexes me with these 3 possibilities is that all 3 incidents began during high G maneuvers. I got 7 flights in without incident where I wasn't pushing the heli as hard. First full power smack and it happens again. Maybe the EX Bus signal is getting mixed up when I am stick banging fast? No errors in Spirit log to support, but you told me before that these types of messages were not logged (protocol errors, checksum errors, etc) like VBAR.

On a positive note, I was able to get the tail and cyclic tuned perfectly prior to the last crash & burn flight.

Really liked how it felt in flight compared to its VBAR twin.
Ended up with:
Flight style = 8
cyclic ring = 180
tail gain at 60 with 1.40x rudder common gain
cyclic rotation speed =12
rudder rotation speed = 12
CFF = 7
revomix=2
piro consistency=170
rudder delay=0

Raising the flight style and CFF mad a big difference in getting the agility I like.