Reducing the SR2 M1 DAMP P gain from -0.5 to -0.2 and the SRM M1 DAMP L gain from -0.5 to -0.25 did reduce the noise in SRCL and DARM between 3 and 8 Hz.
The main improvement was obtained reducing SR2 damping gain, When reducing the SRM L gain, the 1.3 Hz line appears to get larger. It's worth taking a look at the SRM damping loop and its interaction with the ASC and LSC loops to ttry to track donw the 1.3 Hz peak.
So SRCL and DARM are limited by SR2 P damping noise between 3 and 8 Hz. This change reduced DARM RMS by 20%.
I SDFed the SR2 P gain to -0.2 in observe.snap. I tried to update this in the safe.snap file by changing to the "SDF to safe" guardian state, however, it did not show an SDF diff. This is something to keep an eye out for just in case it gets reverted in safe. Tagging OpsInfo so operators are aware.
Confirmed during this relock that the SR2 M1 DAMP P gain was set to -0.2 in both safe and observe, and SRM M1 DAMP L gain was set to -0.5 in both safe and observe.
The SR2 SAFE and OBSERVE tables are the same, as seen in this table of SDF_REVERT models; alog 69140. As far as I know, no changes have been made since this table was created, except for Daniel's comment.
Tagging OpsInfo to remind people of its existence.
After the SR2_P gain reduction, it looks like the SRCL RMS between 2 and 8 Hz is now equally limited by SR2_M1_DAMP_P and SR3_M1_DAMP_L. Below 2 Hz and in particular at the 1.3 Hz peak SRCL is coherrent wiith SRM_M1_DAMP_L, however that could be recoil from the SRCL lock to M1.