Offending Y2P peak between 1 and 2Hz (pink) was removed after a new decoupling filter was put in place (green).
In the H1:SUS-ETMX_L2_DRIVEALIGN_Y2P filter, I put the same invP2P filter in FM1 as in the P2P path, then added a diagonalization filter ("Y2Pdiag") in FM2, and another filter that is a 3rd order elliptic BP in FM3. I didn't try to remove 0.4-0.5Hz Y2P peak, it's not great but probably good enough.
ITMX PUM P2Y and Y2P measurement was done, but I need to leave before generating decoupling filter.
Seems like ITMX PUM is somewhat better diagonalized than ETMX PUM, but I'll remove 1.4Hz P2Y peak anyway.
Remember, the ITM is a wire hang QUAD (the test mass is suspended from the PUM via a wire loop and prisms), and has significantly different dynamics than the ETM fiber welded QUAD. This may be an explanation for the different dynamics. I also imagine you'll have to do this process all over again with the ITM once it becomes monolithic like the ETM.
We probably need to re-do everything once it's re-hang.