jeffrey.kissel@LIGO.ORG - posted 13:41, Tuesday 23 September 2025 - last comment - 14:03, Tuesday 23 September 2025(87102)
H1SUSPRM Absolute Calibration (via ISIHAM2 GS13s) Measurements Taken (Aligned vs. Misaligned are Quite Different)
J. Kissel We're beginning the process of completing all the side quests for suspensions beyond SR3 and PR3. For a summary of what these side quests are, check out the first half of G2501621. Today, after - LHO:87103 Fil upgraded satamps the lower, M2 and M3, stages of H1 SUS PRM (ECR E2400330; Side Quest 1), and - LHO:87105 Oli updated the OSEMINF filter banks to precisely compensate the new sat amp frequency response (Side Quest 3), I report on measurements taken for Side Quest 4 -- to measure the absolute, overall scale of the individual OSEMs response to suspension point motion (as predicted from the ISI GS13s projected into that basis) in the presence of ISI drive. This will eventually be processed in a way similar to LHO:86222 for PR3 top mass and LHO:84531 for SR3 top mass. However, for PRM, which is a globally controlled suspension where we drive the M2 and M3 stages (unlike PR3/SR3), the lower stages also need all of these side quests and absolute calibration. So, here I gather the data or the M2 and M3 as well since we can get them "for free," just adding the channels to the response list. Here's the position of the suspension according to the medium trustworthy "urad" while the suspension is ALIGNED -- M1 Drive request M1 OSEM M2 OSEM M3 OSEM [all "urad"] P :: -1628 -1231 -982 -1520 Y :: -60 -45 +645 -344 where I put "urad" in quotes because that's what these OSEMs are all calibrated to, nominally, but they've never been validated with anything -- until the future results from the data in this aLOG. That includes -- to the best of my aLOG searching, even the alignment sliders (which are sometimes calibrated with using the IFO beam and a distant geometrically known shape as an optical lever; see e.g. for BS and PR2 LHO:25650 or PR3 LHO:70197). Indeed, the alignment slider calibrations for PRM are still P = 1.875 [DAC ct/"urad"] and Y = 2.681 [DAC ct/"urad"] that were derived from the modeled electronics chain and actuation strength from LHO:4563. Anyways -- Here's the data with the alignment offsets ON. /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/PRM/Common/Data 2025-09-23_1810_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_X_0p05to40Hz_calibration.xml 2025-09-23_1810_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_Y_0p05to40Hz_calibration.xml 2025-09-23_1810_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_Z_0p05to40Hz_calibration.xml and I attach screenshots of these templates' results. The top mass, M1 data show similar levels of calibration adjustment needed as PR3/SR3; corrections in the range of 1/0.78 = 1.28 [urad/"mrad"] and 1/0.64 = 1.56 [urad/"urad"] are needed. Perhaps unsurprisingly, the lower stage OSEMs need similar levels of correction. PRM's M3 LL OSEM needs the most correction, needing a factor of 1/0.0915 = 10.9 [urad / "urad"]. Yikes! Note, with the suspension ALIGNED, the ADC counts for the M3 LL are still at 18660 [ct], and MISALIGNED they're 18740 [ct], both out of 2^15 = 32768 [ct], so it's not like it's at the edge of it's range, either with the flag fully or not occulting light. Nor is it like the LED emitted power dropped substantially low... huh. Welp, it'll be good to get all of these OSEMs calibrated!
Images attached to this report
Comments related to this report
Here's a zoom of data in the ALIGNED configuration on the linear scale -- above ~5 Hz, above the resonances, where we expect the calibration to be 1.0 [OSEM m / GS13 m].
Images attached to this comment
Also out of mistake and then curiosity, I also gathered data with the *misalignment* offsets ON, i.e. when the suspension is MISALIGNED. /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/PRM/Common/Data 2025-09-23_1730_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_X_0p05to40Hz_calibration.xml 2025-09-23_1730_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_Y_0p05to40Hz_calibration.xml 2025-09-23_1730_H1ISIHAM2_ST1_PRM_WhiteNoise_ISO_Z_0p05to40Hz_calibration.xml The reported necessary calibration is markedly different between ALIGNED and MISALIGNED. This is something we'll *definitely* need to consider for estimators of globally controlled suspensions operationally: (1) The calibration of the OSEMs fundamentally changes as the suspension moves across "large" alignment positions. (2) That means that we want to make sure we calibrate, and then measure estimator plant transfer functions in the suspensions' nominal low noise alignment position (at least roughly). (3) For globally controlled suspensions -- and specifically those that regularly change alignment as a result the IFO's initial alignment procedure -- we need to understand *how* much the calibration is changing even across "small" alignment position changes. (4) Also, we'd need to note -- and be able to handle -- a calibration change across how much the alignment position is shifted *during* lock acquisition as the alignment sensing and control's DC request is offloaded to the top mass drive. (5) We'll likely need to turn OFF the future estimator, if not at least expect worse performance, when the suspensions are in a "markedly different" alignment position. How "markedly" is "markedly" depends on the answer to (3). Comforting at least, is that the misalignment offsets are driving the suspensions by *hundreds,* if not a few *thousands* of "urad" from their aligned position.
Images attached to this comment
