I did some injections on HAM1, for comparison with 42694 and 42722
I reduced the amplitude compared to 44856, so these excitations aren't showing up in DARM, but people can look for them in the REFL WFS and table top L4Cs.
The Z excitation started at 2:11:50 UTC, was stopped after a large glitch at 2:23 UTC Jan 15th.
The X excitation started at 2:27 UTC, but there were several large glitches. a relative glitch free time was from 2:37-2:44, the excitation continued until 2:44 UTC.
The Y excitation started at 2:55 UTC, and was on until 3:06. We had to turn the dither lines back on partway through that because the PRC1Y was slowly going unstable.
I know Gabriele is looking at how these compare H1 vs. L1, but just a preview of how much HAM1 motion is affecting our ASC signals...
I've taken the times that Sheila posted above, and gotten the transfer functions from the tabletop L4Cs to the different ASC degree of freedom control points (they are saved at a faster rate than the error points, although the error points are easier to calibrate). Then projected the quiet non-injection tabletop motion onto each degree of freedom.
Attached are 6 plots, looking at the DC centering loops and the WFS (CHARD and INP1), for the 3 different injection directions. The WFS themselves don't really see much with the table motion, but the DC centering loops in pitch are seeing a lot of this table motion. This is in contrast to LLO, who needs to use the tabletop L4Cs to feedforward to CHARD. None of the table motion is affecting the centering loops above about 20Hz. The L4C projections are only displayed in the plots if there is anywhere in the transfer function that has coherence during the injection above 0.7.
The ability to feedforward the L4C signals to the ASC has not yet been implemented, since without that data it wasn't clear which DoFs needed it the most. LLO hard-codes the ability to feedforward to CHARD only, which we don't want to do since that's not a degree of freedom that is bothered by the table motion. Note however that the sender parts are now in the SEI model, so if we decide to implement feedforward, it will only require an ASC restart, not any SEI restarts.
Since the injections are not strong enough to be seen in DARM, all we can do is compute upper limits, as shown below.