jeffrey.kissel@LIGO.ORG - posted 11:50, Tuesday 24 June 2025 - last comment - 12:15, Tuesday 24 June 2025(85281)
Table of Six-OSEM SUS M0/R0/M1 Damping Loop EPICs Gain Values
J. Kissel Following LHO:85273, I realize I've lost track of which 6-OSEM top-mass suspensions have had their damping loop gains augmented through commissioning efforts. All of the below suspensions have had their loop control filters updated with a "Level 2.0" design which, during the upgrade from Level 1.0 to Level 2.0, have their EPICs gains reset to -1.0 (and any tuning of gains is built into filter banks). However, because it's far easier to check, revert, and set without knowing details of the frequency dependence of the loop shaping, it is typical to adjust the overall loop gain by increasing or decreasing the EPICs gain in the [M0/R0/M1]_DAMP_[DOF] banks, and then check the impact on ISC loops, including the DARM noise performance and ASC or ISC loop stability. Here's a list of the current status of all top mass DAMP gains for QUADs, Triples, and Doubles that have a 6-OSEM sensor/actuator array: {'ETMX M0': [-1.0, -1.0, -1.0, -1.0, -1.0, -0.5], # Aug 2023 commissioning LHO:71927 'ETMX R0': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0], # "Gain" filters have the feedback sign in it, rather than the EPICs gain 'ETMY M0': [-1.0, -1.0, -1.0, -1.0, -1.0, -0.5], 'ETMY R0': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0], # "Gain" filters have the feedback sign in it, rather than the EPICs gain 'ITMX M0': [-1.0, -1.0, -1.0, -1.0, -1.0, -0.5], 'ITMX R0': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0], # "Gain" filters have the feedback sign in it, rather than the EPICs gain 'ITMY M0': [-1.0, -1.0, -1.0, -1.0, -1.0, -0.5], 'ITMY R0': [-10.0, -10.0, -2.0, -0.3, -0.1, -0.1]} # "Gain" filters are OFF, these are taking the place of those filters {'BS': [-1.0, -1.0, -1.0, -1.0, -3.0, -3.0], # P and Y gains have been this way since 2015 at least 'MC1': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'MC2': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'MC3': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'PRM': [-0.5, -0.5, -0.5, -0.5, -0.5, -0.5], # Aug 2023 commissioning, LHO:72130, LHO:72106 'PR2': [-0.5, -0.5, -0.5, -0.5, -0.5, -0.5], # | 'SR2': [-0.2, -0.2, -0.2, -0.2, -0.2, -0.2], # | 'SRM': [-0.5, -0.5, -0.5, -0.5, -0.5, -0.5], # V 'FC1': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'FC2': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'PR3': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'SR3': [-0.5, -0.5, -0.5, -0.5, -0.5, -0.5], # Aug 2023 commissioning 'TMSX': [-3.0, -1.0, -1.0, -1.0, -3.0, -1.0], # L and P gains have been this way since 2015 at least 'TMSY': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0], 'OMC': [-1.0, -1.0, -1.0, -1.0, -1.0, -1.0]} Even though the gains seem to be quite scattered and inconsistent at first, most have either been that way through all observing runs to date or were adjusted during a concerted Aug 2023 commissioning effort by Gabriele and Elenna.
To generate the above dictionaries of gains, I invoked the interactive guardian shell with $ guardian -i the wrote these quick for loops grab all the triples and doubles gains: gain_dict = {} for iOptic, thisOptic in enumerate(['BS','MC1','MC2','MC3','PRM','PR2','SR2','SRM','FC1','FC2','PR3','SR3','TMSX','TMSY','OMC']): gains = [] for jEUL in ['L','T','V','R','P','Y']: gains.append(ezca['SUS-'+thisOptic+'_M1_DAMP_'+jEUL+'_GAIN']) gain_dict.update({thisOptic:gains}) To grab the quad gains: quadgain_dict = {} for iOptic, thisOptic in enumerate(['ETMX','ETMY','ITMX','ITMY']): m0gains = [] r0gains = [] for jEUL in ['L','T','V','R','P','Y']: m0gains.append(ezca['SUS-'+thisOptic+'_M0_DAMP_'+jEUL+'_GAIN']) r0gains.append(ezca['SUS-'+thisOptic+'_R0_DAMP_'+jEUL+'_GAIN']) quadgain_dict.update({thisOptic+' M0':m0gains}) quadgain_dict.update({thisOptic+' R0':r0gains})To make this extra fun, some of these gains are set all the time, and some are changed as a part of the locking process. Specifically, in the LOWNOISE_ASC guardian state, all test mass M0 Y gains are reduced from -1 to -0.5, and all SR2 DoF damping gains are reduced to -0.2 (I think they start at -1 but I'm not sure).
During commissioning, Gabriele and I found that noise reinjected from the damping loops limited the RMS of other control loops (like DHARD or SRCL), which then limited the RMS of DARM at low frequency, which we have assumed can all be sourced to the noise in the BOSEMs. However, for some loops, having too low gain all the time or at certain parts of the locking process disturbs locking by causing oscillations from lack of control at suspension resonances, or instabilities because the plants used to design certain controls include the presence of the damping loop (seems especially problematic when we hand off one controller to another, like going from high bandwidth to low bandwidth ASC). As a reminder, reducing the DARM RMS has been an area of significant work throughout O4 commissioning and observation.