* FmodErr was more accurately zeroed using the RS232 connection the LNFS (thanks to Rosario). There is now a macro to do this in AlpDet "ZeroFmodErr", but some thought should be put into the automation of this process. We also adjusted the FmodErr calibration (now 0.42, was 0.65). We confirmed the effect of this tuning by intentionally mis-tuning the modulation frequency by 4Hz while in Science-Mode (see first plot). Significant improvement was seen at low frequency (10 - 40 Hz, see third plot), and the coherence with SSFS in this region is now mostly gone (see forth plot).
* Informed by the noise projections made today, the NEO filter was modified to provide less noise at 46Hz, while giving more phase margin. The intended results were achieved.
* DARM filter change related to the violin modes. The effect of this in long locks has yet to be studied.
* Tuning of the MICH-canceling Alpha was attempted, but no improvement was seen. The previous value of 0.08 seems to be nearly optimal, but there might be some improvement gained by adding a filter to the Gc (see second plot).
We made several tests of the Sensitivity Curve macro in various ITF configurations. Mostly we generated these curves with the quadrant motor auto-centering and UGF servos off. One was made with the old NEO (Gc_Algo_Number 141), the others with the new NEO (181). The best for the day was at GSP 841107790.
Overall, we are now much closer to C7 at low frequencies. At other frequencies we suffered from not having a "Ruggi Alignment" (see fith plot).
