As reported in last night entry, we tried to improve the SRCL noise coupling non stationarity by increasing the low frequency gain of the DHARD Yaw loop. In my initial elog entry, I pointed out that the SRCL coupling was fluctuating following ASC-AS_A_RF45_I_YAW_OUT. However, it turned out that this signal is basically equal to the DHARD yaw error signal. Most of the SRCL coupling fluctuations and residual motion of DHARD are at frequencies below 200 mHz, so I designed a boost to increase the low frequency gain of the loop. The improvement is visible in the first attachment (green without boost, red with boost). The loop accuracy improved significantly, as shown in the second sttachment, which is a time series of the error signal.

With this new loop, Evan injected some SRCL noise and I could do the same analysis I did yesterday to study the SRCL coupling non stationarity. Here are my main comments:

• contrarily to what stated in last night entry, the SRCL coupling is both lower and more stationary. The third attachment shows the SRCL coupling gain (average of the TF in the 100 Hz region) as a function of time. It is on average about 3.5e-11 and fluctuating by some 2e-11. Before DHARD boosting, the coupling was fluctuating between -10e-11 and +27e-11.
• Now the channel ranking gives a different answer: SRCL coupling is mostly modulated like ASC-AS_A_RF36_Q_PIT, and only partially as DHARD YAW. So it seems that we removed most of the fluctuations caused by DHARD, but now we have some other loops to improve. If I'm not mistaken, AS36 signals should be used to control some SRC degree of freedoms