Closes FAMIS 26040. Last checked in alog 83968.

HAM4, HAM7, HAM8, EX ST1 all have elevated peaks, mostly around 1Hz line. This is expected due to high maintenance and vent activity.

Closes FAMIS 26375. Last checked in alog 84000,

Everything is under threshold. There are different fans on than last week so difficult to compare. MR FAN3 in CS has a slightly higher noise character in last 3 days

While prepping the HAM1 ISI for testing last night, I noticed that one of the HAM2 HEPI L4Cs seemed non-responsive. It seems this has been the case for quite some time, 206 days according to the attached ndscope. Something happened on Oct 1st last year that cause this L4C to suddenly have only ~1/50th the signal the H2 L4C on the HEPI  to sees. I have poked at it a bit, checked cabling on the pier and gently hit the foot to see if I could revive the sensor. I have a couple other tests I want to do before deciding to remove the actuator to replace the sensor, not exactly eager to have to do that.

Dry air skid checks, water pump, kobelco, drying towers all nominal.  Dew point measured at HAM1, reading -43.8°C. Main turbo pumps, XBM, YBM and OMT, all nominal, and temperature is good and stable.  HAM6 turbo pump and cart are nominal.

WP12488 HAM8 ISI binary readback broken following site outage FRS33798

Fil, Erik, Jim, Dave:

Yesterday we power cycled the binary input/output and interface chassis, which did not fix the issue. Today we stopped all the models on h1cdsh8, powered down the computer, then power cycled the IO Chassis.

This also did not fix the problem. Investigation continues.

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.

FAMIS27814

Our cadence is a bit off from biweekly, but no matter since no water was added and the chillers look good (updated the T2200289 sheet). There is a small amount of blueish gunk starting to form in the wire mesh diffuser in the TCSX chiller. I didn't see any of these in our parts bin so I'll put a few on order. 

Fri Apr 25 10:08:57 2025 INFO: Fill completed in 8min 53secs

Gerardo confirmed a good fill curbside.

FranciscoL, SheilaD, JoeB

In summary:

I regenerated the mcmc fit on report 20250327T160138Z at 5, 10 and 20 Hz, to compare how the mcmc model would change given different frequency ranges for the fit. We see that the residual of the measurement to the model is better as we increase the minimum frequency threshold. We think this is due to external effects, aside SRC, confusing the model.

We want to give more validity to our observations on alog 83592 -- that the sensing model at low frequencies is a good descriptor of the sensing function. On 83592, I claimed that the observed changes in the sensing function come from changes in SRCL detuning. This is backed up by a low residual on the measured/modeled sensing function. In today's alog, we evaluate how well the model describes the sensing function by (1) having a *low* residual and (2) decreasing the optical spring frequency.

The first three figures attached ('sensing_mcmc_compare_*Hz.png') are the sensing function plots generated by pydarm for measurement 20250327T160138Z. This measurement is not a representation of a calibrated interferometer, since we tuned the SRCL detuning (83585).

From the transfer function on the left, the green dots represent the measurement, the orange trace is the mcmc model from the measurement, and the blue trace is the mcmc model from the last validated calibration measurement (20250222T193656Z). The residual from the measurement to both models are plotted on the right, matching the colors of the respective models. We are interested in the change of the orange trace when we modify the frequency range of the fit.

Each figure has a pair of vertical lines denoting the limits on the frequency range used for the fit. I regenerated the measurement report for 05, 10, and 20 Hz for the minimum frequency. The maximum frequency at 1200 Hz was not changed.

For the residual plots, we see that the magnitude is largest when fitting to 5 Hz, and that the change between the 10 and the 20 Hz fits is negligible. For illustration purposes, I plotted the different sensing models, along with the measurement, together in the last figure (20250327T160205Z_compare_models). The phase plot is not the same as the plots given by pydarm due to missing factors. We see that the three models match above 30 Hz in magnitude.

For the optical spring frequency, we see a decrease from 2.19 Hz to 0.43 Hz when fitting to 10 Hz, and 20 Hz, respectively. This decrease on the SRC spring suggests that the model does not fit the sensing function well, below 20 Hz.

So, regenerating the sensing model shows that the model is good when we see anything above 20 Hz. But we see from the measurement that the SRCL detuning does change the sensing function at low frequencies. The model, however, cannot fit the data from the measurements at these frequencies. We have seen before (80267) that angle to length cross coupling influences the sensing function in the low frequencies. A scenario of cross coupling affecting the sensing function could explain why the model cannot fit the measurement.

Repeating the analysis on cross-coupling, as done in 80267, could improve the sensing function.

TIL; pydarm edition:

The approach that enabled me to regenerate pydarm reports, albeit messy, was to

1.  Modify the pydarm_cmd_H1.yaml that is in the report to be regenerated, in this case 20250327T160138Z. In particular, the mcmc_fmin value.

2. Once the YAML file is configured, run

This command will re-run the calibration report, and, by using the configuration that lives in that directory, make a new model to the sensing function measurement.

TITLE: 04/25 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: 1mph Gusts, 0mph 3min avg
    Primary useism: 0.06 μm/s
    Secondary useism: 0.09 μm/s
QUICK SUMMARY:

IFO is in PLANNED ENGINEERING for VENT

Some planned tasks for today:

• HAM1 RM1 Install, cabling and troubleshooting
• HAM1 ISC Prep
• IOT2L - Roll up and set in location when needed
• HAM1 - ISC components install
• HEPI start software part upgrade
• PCAL work at Ends
• EY Wind Fence Work

Work safe!

Jordan, Janos, 04-24-2025, ~15:00

The wire gaskets for VBO-C and VBO-D have been leaking continuously for a very long time. Lots of vendors have been contacted, and after a lot of unsuccessful tests, it seems that we finally found the good one manufactured by Torr Scientific (UK). This gasket is extremely annealed, butter-soft, needs super careful handling, but even after some suboptimal installation steps (mea culpa), it finally sealed a new chamber successfully.

We tested it right after receiving (had a thriller with customs) with a new C+B chamber, pumped it down and tested on the same day. After the test, we let the pump running during the weekend just to see the ultimate pressure it can reach.

They are still very pricey though, 193 GBP (= 257 USD)/pc., even in the case of purchasing 100 pcs., it's 147 GBP (= 196 USD)/pc. So, certainly not an optimal, long-term solution, but a good bridge between having nothing and manufacturing our own.

Oli, Matt, Camilla 

TITLE: 04/24 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:

IFO is in PLANNED ENGINEERING for VENT

Productive day where most of the work was spent on HAM1 cabling, followed by RM1 install.

Of note:

• Some PCAL work was done
• Dave and Fil were troubleshooting HAM8 outage glitches
• CEBEX walkabout at MY
• QUAD transportation
• 100 kids enlightened with cool laser vacuum black hole facts

LOG:

Ibrahim, Dave:

RM1 exceeded its RMS motion limit, and 5 minutes later the HAM1 DACs were DACKILLed, stopping all h1isiham1 and h1hpiham1 drives.

We have bypassed RM1 and SEI-HAM1 watchdogs indefinitely for now.

Jennie Wright, Sheila, Keita, Camilla

Jennie has been working on modeling of our arm to OMC mode matching, and while she was looking for some distances we noticed a discrepancy between the as built numbers and the mode matching design.  It seems possible that there was a mix up, where design documents call for a distance of 97cm between the OMC waist and OM2, and the as built distance is 97 cm between the OMC input coupler and OM2 (so,  the OMC seems to be translated by 14 cm from where the mode matching design expected it to be).  If we have the q parameter used in the design (ROC matched to SRM, beam size there is 2.1mm), this would introduce a 1% mismatch between the OMC mode and the SRC mode (the impact could be larger for a different q parameter).   Note that the analysis done in 71145 was using the design distances, so we will want to revisit that.  I've attached an a la mode script below.

Editing to add more information:

I initially did this with the design ROC for OM2, which is 1.7meters, which is also what is in Finesse.  This results in an overlap for the design locations of 99.95% and 99% for the as built locations, with the as built waist location 12cm from the OMC waist location (there are differences in the other distances that partially cancel the 14cm difference between OMC input coupler and OMC waist location), and a waist size of 545um compared to the OMC eigenmode of 509um. In 71145 Keita says that the OM2 ROC is 2m cold and 1.75m hot. 

If we change the OM2 curvature for the as built path to 2 meters, the overlap becomes 93.8%, and the waist location becomes 14.3cm from the OMC waist location, with a waist size of 649um.  Using the hot ROC of 1.75m, the overlap becomes 98.5% and the waist is 563um 11.5cm from the OMC waist.  So, if our q parameter was the design one at SRM, the error in OM2 curvature would be compounding the error in the OMC placement to make the mode matching worse.  The single bounce measurements agree that the mode matching is worse with OM2 cold than hot, but the full IFO measurements are the opposite.

Summarizing distances:

T1000317:

• OMC waist to OM3: 26.8 cm
• OMC waist to OM2: 97.6 cm
• OMC waist to OM1: 2.37 m
• OMC waist to SRM: 6.017 m

T1200410 says we should have:

• OMC waist to OM3: 26.2cm
• OMC waist to OM2: 97 cm
• OMC waist to OM1: 2.365 m
• OMC waist to SRM: 5.936 m

.yaml file, which is based on E2100383:

• OMC input coupler to OM3: 30.4cm
• OMC input coupler to OM2: 97.1cm
• OMC input coupler to OM1: 2.416m
• OMC input coupler to SRM: 5.858 m

Camilla looking at edrawings (which agree with photos):

• OMC input coupler to OM3: 30.1cm
• OMC input coupler to OM2 95.5cm
• OMC input coupler to OM1: 2.4meters

To compare the solid works to the photos we looked at D0901822 and compared it to OM3 photo and this photo of OM2 when it was a tip tilt. 

Using the beam q from table 1 of T1200410 for just after SRM, and propagating it through the distances from T1000317, gives an overlap with the OMC waist of 99.96%.  Adjusting the SRM to OM1 distance to agree with T1200410 gives 99.95%, so that difference between the two design documents isn't significant. 

S. Koehlenbeck, J. Freed, J. Kissel, J. Oberling, R. Short

The SPI pick-off path installation on the H1 PSL table is now complete. The beam in the new SPI path has been reduced to 20mW and is currently being dumped with a razor dump between SPI-L1 and SPI-L2. Pictures attached reflect the final installation and layout, which will be be reflected in the updated as-built layout at a later date.

Associated entries: 83925, 83933, 83956, 83961, 83978, 83983 (and more to come)