After discussing with Jeff this morning, I have flipped the sign of the Coil Output filters Gain to get the damping loops going on both RM1 and RM2. These changes have been accepted in the SDF. Jeff had set it to nominal value (i.e how it was pre-vent), however this makes the suspension very angry with the WD tripping immediately. 

RM1 Coil Output Filter Gain:-
H1:SUS-RM1_M1_COILOUTF_UL_GAIN -1
H1:SUS-RM1_M1_COILOUTF_LL_GAIN 1
H1:SUS-RM1_M1_COILOUTF_UR_GAIN 1
H1:SUS-RM1_M1_COILOUTF_LR_GAIN -1

RM2 Coil Output Filter Gain:-

H1:SUS-RM2_M1_COILOUTF_UL_GAIN 1
H1:SUS-RM2_M1_COILOUTF_LL_GAIN -1
H1:SUS-RM2_M1_COILOUTF_UR_GAIN -1
H1:SUS-RM2_M1_COILOUTF_LR_GAIN 1

This sign change is done to compensate for the sign change of the osem voltages (now inmons are +ve counts, pre-vent it was -ve counts), along with wrong polarity of the magnets (which has always been there in RM2).

Both RM1 and RM2 are damping fine now.

Gerardo alerted me to the fact that PT110 is indicating an error on the H0:VAC-LX_Y0_PT110_ERROR channel. I logged in to h0vaclx and traced this to an unexpected value for the info data state readout. This should be 8 but is reading 15368. I don't know why or what 15368 means. He indicated that this appeared after they swapped in this gauge for another, so I suspect that this is the issue somehow.

J. Freed

Last week, this week, and the next week, the goal is to finalize the double mixer design, drawings and chassis. M. Pirello. has begun the drawings while I finalize the design, and we will work on the layout as well as the front plate.

Currently 83439 shows that phase noise is good except for harmonics every 4096Hz away from carrier. While any frequency more than 100Hz around the carrier should not affect SPI some of them do have an effect (namely the 8192Hz harmonic), 81593 also shows an ocsiloscope reading of the output which shows how messy the signal looks. This as such this week the focus is on reducing those harmonics.

The Q2 transistor on the Low Noise Power Board failed last week, was replaced, then it popped and replaced again. The problem was diagnosed to be the power pins on the ZFL-500HLNB+ amp. The pins stick out and touched the metal casing of another minicircuit part causing a short on the +15V supply from the low noise power board. A temporary solution of electrical tape was use to isolate the pins while a more perminate isolation for the pins will be there when it is finalized

AGBW.jpg Shows that the Agilent 4396B in spectrum analyzer mode gives less information with a smaller resolution bandwidth. The higher resolution bandwidth shows more harmonics with greater power on said harmonics. I do not know why this is, which is why I will be looking at phase noise when trying to improve harmonics.

The Agilent 4396B in network analyzer mode was used to take transfer functions of the double mixer phase delayer, IMG_1472.jpg shows the set up. The source power was 13dBm, the central frequency was 80MHz. The Span and Resolution Bandwidth were both 40kHz but changeing these values did not change much as the amplitude and phase was relativly flat around 80 MHz. A 10dB attinuator was added to the output to remove the overload message on port B that appered around 7dBm input power when I was incresing power to 13dBm.This is ok as I could calibrate the machine by replacing the phase delayer with a addaptor that let the signal go straight through. This calibration became my (0dBm 0deg) reference.  I took measuments of one port of the phase delayer at a time while terminating the other. 

This is slightly off from what was hoped for with (0dBm -90deg) difference. SPI has some 0.5dB attinuators as well as 1dB attinuators. Due to the attinuators not being their listed value at 80 MHz, Two 0.5dB attinuators on port 2 and one 1dB attinuator on port 1 seems to nearly correct the power differnece.

Continuing from here is characterizing the mixers, the summer, and RF couplers to gain inisight on reducing the harmonics.

FAMIS tasks 31304 and 31319, WP 12487

Procedure checklist for both stations completed.  No issues were identified at this time.

MX: Scroll pump hours: 226.8

       Turbo pump hours: 138

       Crash bearings: 100%

EX: Scroll pump hours: 7154.3

       Turbo pump hours: 1106

       Crash bearings: 100%

M. Todd, C. Compton, S. Dwyer, A. Brooks

As a continuation of the last alog, this is discussing how we can estimate thermal actuator contributions to the surface curvature to the test masses.

Summary:

The dataset used to estimate the ring-heater to surface curvature gain factor contained several ring heater settings and the corresponding HOM spacings at various "positions" in the span of time when that ring-heater power was used. For example, if the ring-heater power was changed and kept the same for 3 months, the HOM spacing data was examined at a random (long-lock time) portion of this 3-month period. This time, I tried to be a bit more systematic, such that we have a dataset with HOM spacing data at the beginning and end of each ring-heater power setting period. This can give us an idea if there are any drifts over time in the HOM spacing or coupling factors of the ring-heater power to surface curvature.

Each HOM spacing data-point is taken from a OMC spectrogram during a lock, and the errorbars associated with each are the spread of that HOM spacing peak.

Results:

The estimated ring-heater power to surface curvature gain is .7563 uD/W, which is very similar to the initial estimates done by Aiden; however, this is not the same value as the current one in TCS-SIM, indicating some adjustments may need to be made.

The blue dataset in the plot is data taken at the beginning of a ring-heater power setting period, while the yellow dataset is taken towards the end of the corresponding period. The estimated ring-heater power to surface curvature gain is the same for both datasets, however there is a relatively consistent downward drift over time by a few Hz in the HOM spacing. I'm not sure of the reason for this but it could be due difference in arm powers at each time.

Additionally, we expect the HOM spacing due to self-heating on top of the cold state to be around 5340 Hz, however TCS-SIM projects it to be around 5620 Hz.

Thus we estimate the change in HOM spacing due to self-heating alone to be around 175 Hz, where TCS-SIM projects it to be around 450 Hz.  This indicates the gain factor mapping absorbed power to surface curvature change may be to high, or our estimate of the absorbed power is too high. The latter can be estimated from the Hartmann wavefront sensors.

I took a look at the YAW angle for the HAM HEPIs associated with PR2 (incorrect by me) (HAM2), SR2 (HAM4), and SR3 (HAM5) based on this previous alog83705. I believe the IPS channel units are in nrad, so the largest yaw angle I see is on HAM2 with 1800 nrad or 1.80 urad, HAM2 and 4 are different by ~ 1.10 urad.

Wed Apr 30 10:11:57 2025 INFO: Fill completed in 11min 53secs

Gerardo confirmed a good fill curbside.

TJ was unable to run the 'guardutil archive-clone ...' command due to permission issues.  We looked at it, and the relevant error message was:

fatal: failed to copy file to 'IMC_LOCK/.git/objects/2f/826f205c25a6189503d831106e2dd97d7a39dc': Permission denied'

Tracing through paths, this was in /guardian/archive/IMC_LOCK/.git/....

The permissions + ownership of that file where 0600 (rw for the user), everything else was 0444 (r only for everyone) and owned by guardian:controls.  Changing the permissions to 0444 fixed things.

We did a search from the guardian machine for other files with that problem and found 24 files under /srv/guardian/archive, so we reset the permissions to clear up future problems.

Find command for searching: find . -type f -perm 600
Command for updating: find . -type f -perm 600 | xargs chmod 0444

I suspect this is a umask issue that someone had a umask of 077 set when they did an archive request.

Morning dry air skid checks, water pump, kobelco, drying towers all nominal.

Dew point measurement at HAM1 , approx. -41C

TITLE: 04/30 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    SEI_ENV state: MAINTENANCE
    Wind: 4mph Gusts, 2mph 3min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.12 μm/s
QUICK SUMMARY:

The planned work today includes:

• Mostly Laser Haz with PSL_IN @200mw for initial beam path alignment work
• Beam position measurement
• HAM1 - Beam Path alignments
• Install PM1, cable up, start troubleshooting, ground checks
• HEPI upgrade software part start
• VAC - HAM1 Top flange Gauge work table side - GM?
• EY wind fence work

PCal team took PS4 to EY again today for the first every April measurement of End Y! Previous Aprils'  measurements have apparently all been EndX.

Following T1500062V-18, Nothing really out of the ordinary except I put RX here instead of bolting to the side of the breadboard like I normally do, to see if that has any effect on the load balancing of the RX optical bread board or any impact to the overall measurement during the time when WS was in the RX side.
Beam spot

python3 generate_measurement_data.py --
WS "PS4" --date "2025-03-24"
Reading in config file from python file in scripts
../../../Common/O4PSparams.yaml
PS4 rho, kappa, u_rel on 2025-03-24 corrected to ES temperature 299.4 K :
-4.701550919294612 -0.0002694340454223 4.0632996079052654e-05
Copying the scripts into tD directory...
Connected to nds.ligo-wa.caltech.edu
martel run
reading data at start_time:  1429997850
reading data at start_time:  1429998230
reading data at start_time:  1429998550
reading data at start_time:  1429998990
reading data at start_time:  1429999330
reading data at start_time:  1429999640
reading data at start_time:  1430000100
reading data at start_time:  1430000600
reading data at start_time:  1430000920
Ratios: -0.5349326615859125 -0.5432025342228874
writing nds2 data to files
finishing writing
Background Values:
bg1 =        17.394679; Background of TX when WS is at TX
bg2 =        4.438216; Background of WS when WS is at TX
bg3 =        17.479506; Background of TX when WS is at RX
bg4 =        4.610581; Background of WS when WS is at RX
bg5 =        17.487987; Background of TX
bg6 =        -0.718008; Background of RX

The uncertainty reported below are Relative Standard Deviation in percent

Intermediate Ratios
RatioWS_TX_it      = -0.534933;
RatioWS_TX_ot      = -0.543203;
RatioWS_TX_ir      = -0.527567;
RatioWS_TX_or      = -0.535369;
RatioWS_TX_it_unc  = 0.055737;
RatioWS_TX_ot_unc  = 0.055881;
RatioWS_TX_ir_unc  = 0.056510;
RatioWS_TX_or_unc  = 0.055480;
Optical Efficiency
OE_Inner_beam                      = 0.985953;
OE_Outer_beam                      = 0.985408;
Weighted_Optical_Efficiency        = 0.985681;

OE_Inner_beam_unc                  = 0.042868;
OE_Outer_beam_unc                  = 0.044311;
Weighted_Optical_Efficiency_unc    = 0.061653;

Martel Voltage fit:
Gradient      = 1637.860873;
Intercept     = 0.156301;

 Power Imbalance = 0.984776;

Endstation Power sensors to WS ratios::
Ratio_WS_TX                        = -0.927527;
Ratio_WS_RX                        = -1.383600;

Ratio_WS_TX_unc                    = 0.045873;
Ratio_WS_RX_unc                    = 0.042755;

=============================================================
============= Values for Force Coefficients =================
=============================================================

Key Pcal Values :
GS           =      -5.135100; Gold Standard Value in (V/W)             
WS           =      -4.701551; Working Standard Value             

costheta     =      0.988362; Angle of incidence
c            =      299792458.000000; Speed of Light
             
End Station Values :
TXWS         =        -0.927527; Tx to WS Rel responsivity (V/V)
sigma_TXWS   =        0.000425; Uncertainity of Tx to WS Rel responsivity (V/V)
RXWS         =        -1.383600; Rx to WS Rel responsivity (V/V)
sigma_RXWS   =        0.000592; Uncertainity of Rx to WS Rel responsivity (V/V)

e            =        0.985681; Optical Efficiency
sigma_e      =        0.000608; Uncertainity in Optical Efficiency

Martel Voltage fit :
Martel_gradient         =        1637.860873; Martel to output channel (C/V)
Martel_intercept   =        0.156301; Intercept of fit of     Martel to output (C/V)

Power Loss Apportion :
beta          =        0.998844; Ratio between input and output (Beta)  
E_T          =        0.992240; TX Optical efficiency
sigma_E_T          =        0.000306; Uncertainity in TX Optical efficiency
E_R          =        0.993389; RX Optical Efficiency
sigma_E_R          =        0.000306; Uncertainity in RX Optical efficiency

Force Coefficients :
FC_TxPD          =        9.160033e-13; TxPD Force Coefficient
FC_RxPD          =        6.229843e-13; RxPD Force Coefficient
sigma_FC_TxPD          =        5.093831e-16; TxPD Force Coefficient
sigma_FC_RxPD          =        3.305931e-16; RxPD Force Coefficient
data written to ../../measurements/LHO_EndY/tD20250429/

This adventure was brought to you by, Francisco L & Tony S.
 

Calibration is currently regenerating the O4b uncertainty budgets.

Due to a missed change to the ETMX UIM suspension filters, which was found in LHO alog 82804, and then fixed only on Feb 27, 2025 (see LHO alog 83088), we need to add a correction TF to be included in the uncertainty budgets.

I attach a graph of that correction TF (corrected model / original model) and the text file needed for that correction.

We will be using correction TF in all our uncertainty budgets for O4b, and up to Feb 27, 2025 around 20:00 UTC, or GPS 1424721618 for O4c.

Summary: This work aims to understand the shape of the DARM sensing function as affected by mode mismatches and input power. It seems that the mismatches affect the detuning of the SRCL DOF, that is it affects the locking point, but not the shape of the DARM sensing funciton. However, the input power, which causes radiation pressure, has a bigger effect on it. If I do not include QUADsuspension option, then I get a flat response function regardless of the mismatch or the input power. 

For the past week I have been studying the effect of SCRL DOF detuning (or sometimes I refer to it as offset) on the DARM readout using FINESSE.
In FINESSE, that is done by changing SRCL.DC, and looking at the transfer function from DARM.AC.i to AS.DC.o

The reason being to understand the measurement of the sensing function taken and show here. 
My system include up to maxtem=4, QUADsuspensions,     InitialLock > DARM_RF_to_DC.

Before we start doing that, let’s understand the error signal of SRCL DOF. This is read at POP45_I channel. The ideal tuning of this loop is SRCL.DC=90, that is the case where the carrier beam is anti-resonant in the SRC, while the sidebands are.  At that detuning, the error signal is 0. However, since we do not have a perfect model, there is some mode mismatch between SRC and the arm cavities. This will inject some higher order modes carrier and sidebands into the SRC and that changes the locking point. Using the default cold state LHO model, the detuning at which the error signal is 0 is SRCL.DC=-90.4643
This is shown in the plot below (Error_signals_vs_detuning.pdf)

Let’s look at this offset as the mode mismatch changes. I adjust ITMlenses focal lengths and measure the mismatch between SRX and XARM cavities, and SRY and YARM cavities. I also have an amplitude detector to measure the power in the HG20/02 modes at the AS port, which might be partially resonant in the SRC due to its low finesse. (Detuning_vs_SRC_ARMS_mismatch.pdf) 

On the x-axis is the SRY to YARM mismatch percentage, and on the y-axis is the SRX to XARM mismatch percentage. I’m not sure why there is such a big difference of mismatches in the default cold state model, but the qualitative behavior is what we need here. 
The colors represent SRCL.DC offset while the numbers on each cell represent the power in the 2nd high order mode in W. 
We notice that as the SRY to YARM mismatch gets closer in magnitude to SRX to XARM mismatch the offset decreases and goes towards 90, and the power in the 2nd HOM also gets smaller, which suggests that the presence of these modes affects the offset, as we would expect(see also This old technical note)
I believe the reason that the HOM gets smaller as the mismatches get closer together can be imagined similar to what the Michelson interferometer does to the fundamental mode for HG20/02 modes that are at the same phase, they cancel each other out at the AS port. 

Finally, let us look at the DARM sensing function, which is our goal of this investigation. 
We would expect the sensing function to be flat starting around ~7Hz and keep like that until it starts rolling off at ~ 200 Hz. 
Now, here is the surprise, I plotted the response function in the default case where there is mismatch between SRC and the arm cavities, and then I changed ITMXlens focal lengths to get almost perfect mode matching, and it turns out that they are not that different, as seen below (Comparison_with_different_MM.pdf)

What seems to be affecting the response function is not the mode matching (even though it affects the locking point), but the input power value. If I reduce the input power to 1W instead of 60W that is used in the comparison plot, I get the following response function (yes, I checked the error signal zero point is still at 90.464 (1W_DARM_readout_vs_detuning.pdf)

And so, from this investigation, it seems that the input power affects the response function more than the mode matching, while the mode matching affects the locking point. 

Next, I reset the model and I started changing the reflectivity of the SRM, narrowing the linewidth and increasing the finesse of the cavity. As the finesse increases, the detuning goes towards 90 degrees. This could be from reducing the effect of the HOM on the fundamental mode in SRCL.  (SRM_reflectivity.pdf)

Sheila, Keita, Elenna, Camilla, Jason, TJ  WP# 12491.

With MC2 misaligned, TJ transitioned to laser hazard and opened the PSL IR light pipe.

Keita checked that the MC REFL path on IOT2L was aligned-ish. We had a beam on the LSC REFL DC and on WFS A. Not much on WFS B. Sheila moved the input PSL PZT to improve pointing onto WFS A.

We were getting flashes on MC2 trans so went to the control room where Sheila zero’ed H1:IMC-WFS gain (was 0.04) and continued to move PSL PZT to maximize MC2 trans flashes (could also see flashes on IM4 trans).

Once these flashes were maximized, Sheila tried to lock with the guardian (which automatically sets gains and thresholds based on input power). Locked but didn’t look great. Sheila and Keita edited the IMC_power_adjust_fuc() of ISC_libary.py so nothing is done when the PSL input power < 1W. IMC then locked great.

During the IMC locking, we had some troubleshooting here as we kept getting “PMC unlocked” notification taking IMC_LOCK to FAULT. Jason explained this could get it getting to the end of its range and while it swaps to the other side of its lock might trigger the notification, even though we could not see it on ndscope. We also had the FSS unlock and the autolocker not able to relook. Jason took PSL FSS common and fast gains to zero, let it settle then brought common and then Fast gains slowly back. This happened a once more where we also had to turn off /on the auto-locker to make it settled.

On IOT2L, everything  (camera and diode) in the IMC trans path was fine, left as is. In the IMC REFL path, the beam was not centered on the high power beam dump (maybe wasn't originally) so we moved IO_MCR_M3. We then checked we were on the plateau of the the LSC diode with IO_MCR_BS1 and checked we weren’t near the edge of the TRIG diode.  We moved the IO_MCR_BD6 beam dump so it was centered on the beam. We moved IO_MCR_BS2 and BS3 to center the beams on WFS A and B.  Sheila moved IO_MCR_M12 up to get MC TRANS on the GiGe camera. Everything on IOT2L is now aligned. 

Checked in-vac and in-air powers are correct-ish by comparing them to 2W IMC locked/ unlocked. Sheila and Elenna locked IMC with WFS centering servos back on and offloaded the WFS. 

As Ibrahim checked yesterday 84153, the IMs are a little different but Elenna checked the beam centering on IM4 trans is close to when we were locked: P was 0.24 now 0.2, Yaw was -0.08 now -0.13. 

Keita restored SR2, SR3, BS, ITMX to a time 31 days ago when IFO was locked and could see the beam on the AS AIR camera. 

We have left the IMC offline and PSL IR light pipe closed. 

Edgard, Oli.

Follow up to the work summarized in 84012 and 84041.

TL;DR: Oli tested the estimator on Friday and found the ISI state affects the stability of the scheme, plus a gain error in my fits from 84041. The two issues were corrected and the intended estimator drives look normal (promising, even) now. The official test will happen later, depending on HAM1 suspension work.

____

Oli tested the OSEM estimator damping on SR3 on Friday and immediately found two issues to debug:

1) [See first attachment] The ISI state for the first test that Oli ran was DAMPED. Since the estimator was created with the ISI in ISOLATED (and it is intended to be used in that state), the system went unstable. This issue is exacerbated by point 2) below. This means that we need to properly manage the interaction of the estimator with guardian and any watchdogs to ensure the estimator is never engaged if the ISI trips.

2) [See second attachment] There was a miscalibration of the fits I originally imported to the front-end. This resulted in large drives when using the estimator path. In the second figure, there are three conditions for the yaw damping of SR3:
           ( t < -6 min )          OSEM damping with gain of -0.1.
    ( -6 min<  t  < -2 min)   OSEM damping with a gain of -0.5, split between the usual damping path and the estimator path.
     ( -2 min < t < 0 min)     OSEM + Estimator damping.

The top left corner plot shows the observed motion from every path. It can be seen that M1_YAW_DAMP_EST_IN1 (the input to the estimator damping filters) is orders of magnitude larger than M1_DAMP_IN1 (the imput to the regular OSEM damping filters).

The issue was that I fit and exported the transfer functions in SI units, [m/m] for the suspoint to M1, and [m/N] for M1 to M1. I didn't export the calibration factors to convert to [um/nm] and [um/drive_cts], respectively.

____

I fixed this issue on Friday. Updated the files in /sus/trunk/HLTS/Common/FilterDesign/Estimator/  to add a calibration filter module to the two estimator paths (a factor of 0.001 for suspoint to M1, and 1.5404 for M1 to M1). The changes are current as of revision 12288 of the sus svn.

The third attachment shows the intended drives from the estimator and OSEM-only paths. They look similar enough that we believe the miscalibration issue has been resolved. For now we stand by until there is a chance to test the scheme again.

I keep forgetting to reset the dust monitor alarm levels after IOC restarts so I wrote a little script to check all of the dust monitor alarms levels and reset them if needed based on E1600132. It does not reset the LVEA ones, it just notifies, as these ones alarm levels are sometimes modified.

The scripts is called check_dm_alarms_lvls.py and lives at /ligo/cds/userscripts/. I've also updated the wiki with this info.

Some changes to the RM1/RM2/PM1:

• Zeroed the alignment offsets of RM1 abd RM2.
• Changed the signs of the coil outputs for RM1 and RM2. This should account for the sign change of the OSEM voltages.
  For reasons unknown the coils between RM1 and RM2 have opposite signs.
• Went through the PM1 model and made sure all the settings and filters are identical to RM1/2.
• The ISC/LSC input on these models is a constant. This constant was 1 and it was changed to 0. (Requires a model boot to become active.)

All other things equal the local damping loops should have the correct sign now. Thic change will also require a sign change in the ASC centering servos.

Late entry

4-16 (Wednesday) activities:
- The previously received, poorly packaged Inficon gauge has been tested, and found functional. These tests continue, and also the HAM1 vacuum interlock gauge will be installed soon
- The BSC8 Annulus Ion Pump - after some aux cart pumping - is now able to hold the annulus pressure, at the mid- E-6 Torr region
- The rough pumpdown of the corner has started at 16:20, at the OMC turbo station, with a pair of ISP-1000 mobile pumping carts. In 4 hours, ~185 Torr was achieved, which means 1930 l/s effective pumping speed, which well corresponds with the 2000 l/s theoretical speed of the pumps
- The rough pumping still needs to be terminated in the end of the days, and so it was at 20:20
- All the turbo stations have been prepared for the HV pumping at the corner