[Ushiba, Akutsu(R), Nakano(R)]

The ALS DARM servo has been modified, and the residual error has been reduced to 6 Hz in the RMS.
This would help the PRFPMI lock acquisition.

Summary

Roll-off modification.

The ALS DARM was modified for better suppression. The high-frequency component in the control signal was reduced by the notch filters, which help the feedback signal reduction at high frequencies. Thanks to these modifications, the roll-off ELP filter could be relaxed. The error signal(top) and the feedback(bottom) signal are shown in the first attached figure. Red is the one with new filters and green is the one with old filters. Also, the OLG was measured and shown in the second attached figure. The top two of the left are the gain and the phase of the OLG. Red is the new OLG, and blue is the old one. We have an additional phase margin of 20 degrees while the RMS of the feedback signal is the same. The gain margin was too small and we see the gain peaking at 20Hz, but with new filters, that gain peaking has been gone.

New boost

Then, we designed the boost. Since we gain 15 degrees of phase, we can design a more aggressive boost from 8Hz. The residual loop error was reduced by a factor of 2 in the RMS as shown in the third attached plot. The UGF is the same as before 20Hz and the phase margin is 33 degrees. The measured OLG will be reported soon.

Action items

- Stable engagement process of the second boost.
- Check the cross-over frequency of the ETMX-IM and TM.
- Check the L2A coupling of the ETMX

Activity log

1332 Measured OLG of DARM. The phase margin seems to be too small, and the ELP105 roll-off filter eats too much phase. We reduced the order of the ELP from 4 to 2. 
1343 Investigate the ALS error signal and found unknown harmonic peaks with a distance of ~10 Hz from several tens of Hz. We need the help of PEM!!
1353 Keep investigating the ALS DARM signal. Several peaks dominate the RMS of the error signal. The UGF of the ALS DARM is 20Hz so we notched the main peaks above 20Hz. The notched peaks are 58.75Hz, 60Hz( and harmonics), 80Hz, 104Hz, 149Hz, 153Hz, 156Hz, 179Hz, 179Hz, and 300Hz.
1407 Locked the ALS DARM with the notch filter. The RMS was reduced by a factor of ~1.5.
1420 Added the broadband notch at ~370Hz to reduce the broad bump around this frequency. This reduced the RMS by another factor of 2.
1432 Try to increase the DARM gain by 1.5. The control signal starts to saturate.
1502 Changed the ELP frequency from 105Hz to 200Hz. 
1517 Modify guardian.(L174 was commented out)
1618 Measured OLG. While the RMS of the control signal above 1Hz is the same, we got additional 15 degrees at the UGF (20Hz) from the DARM loop modification. The measured OLG, the error signal, and the feedback signal spectra are shown in the attached plot.

1628 Start new boost implementation. A new boost was put at DARM1 FM8 (two poles at 0Hz and complex zero at 8Hz with Q of 1). 
1645 The Boost was not able to be engaged, we doubt the order of the DARM filter (DARM2 FM6), so move it to DARM1 FM6. 
1704 Stil does not work, move new boost to DARM2 FM6 which is equivalent to the old BST, but still, it didn't work.
1723 Changed whole boost configuration. Put integrator from 1Hz at FM8, the boost from 0.1Hz to 1Hz at FM6, and 2nd order boost from 0.1 Hz to 8Hz at FM7. All of the filters are engaged with a ramp time of 3 sec. The order of engagement is FM8->FM6->FM7. Now, all of the new boosts can be engaged.