I think I found something interesting.
I checked all the signals with a scope and they all look OK individually:
phase1.PNG
bank 1 at sensor
IMG_2156_1.JPG
bank 2 at sensor
IMG_2157_1.JPG
bank 1 at ECU
IMG_2159_1.JPG
bank 2 at ECU
IMG_2160_1.JPG
crank.PNG
crankshaft at ECU
IMG_2161_1.JPG
And then I noticed this drawing in the manual and decided to test both channels:
wave.PNG
IMG_2162_1.JPG
IMG_2163_1.JPG
IMG_2165_1.JPG
IMG_2166_1.JPG
It looks like bank2 is advanced about 2ms relative to bank1, if they're supposed to be lined up like the manual shows. this would be about 5 degrees.
I also tested phase vs crank, but these did not vary at all with RPM. maybe IVT doesn't kick in before 3500, or the ecu disables it entirely in limp mode.
I'm not really sure the 5 degree out of phase is an error or not without knowing exactly what the gears look like, but this timing could be off.