Fri Ops DAY Shift Summary

TITLE: 05/30 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: Ibrahim
SHIFT SUMMARY:

Work for H1 continues with H1 making it to DC Readout!  

Alignment is NOT "ready for primetime" yet.  There are issues with PRX, so one needs to do a Manual Initial Alignment and be careful around the PRC.....basically, go to PRX_LOCKED and tweak PRM to maximize ASC-AS_A_DC.  (Since I had issues with one of my alignments, I made notes for aligning next time and they are at the bottom of this entry.  Ibrahim is already working on an alignment now.)
LOG:

• 1450-1510 Reset Green references (sheila remotely)
  • As Sheila drove in, she said I could run an alignment.  Had to Run a Manual Alignment because she mentioned issues with WFS for PRC.  
• 1512-1600 Manual Alignment. 
  • Alignment was mostly straightforward, but I think I missed what to do for PRC.  I skipped over (manually) PRC_ALIGN and went to PREP_PRC_ALIGN and then used PRM to maximize AS_A_DC.  For this PRM step was moved mostly in negative-yaw.
    • Moved on to SRC Alignment, but I was using INIT_ALIGN node for this alignment and during this step saw that guardian was going through PRC Align (!)---so I went DOWN and then finished off the alignment with ALIGN_IFO only.
    • Oh and another note is that the Green-Y-Arm needed to go through a Scan Alignment twice to get flashes of 1.0 to lock (flashes at 0.9 was not to lock after 1st Scan Alignment)
  • Then took H1 to PREP ASC for commissioners to touch up alignment there.  (H1 was able to get there with no issues)
• 1620 Equipment clean up from EX (mitch.randy)
  • 1640-1655 blue articulating lift begins trip from EX to Corner (mitch)
    • orange lift will be picked up by rental company next week
    • 1707-1720 small green tractor making EX-Corner trip (randy)
• 1859 GRB short
• 1927 Connection issues started to be observed (Ops Teamspeak connection lost, DARM FOM had error regarding kerberos ticket, & phones went down).  Notified Erik & Dave for recovery. 
• 2249-2315 Checking out CO2 laser in Optics Lab (matt T., julia R., caroline C.)

 

Manual Align notes with PRC Notes:

• ISC LOCK to Manual Alignment

• Take both ALSx +y nodes to Initial Align Offloaded

  • If bad, go to unlocked and touch up flashes

• Align IFO to Input Align Offloaded
• PRC Align Has issues with WFS, so we want to avoid PRC WFS steps

  • go to "PRX_LOCKED" and then use PRM to touch up AS_A_DC

• Align IFO--> MIch Bright Offloaded (can I go directly there, I think it goes to DOWN on its own)

• Align IFO --> SRC Align Offloaded
• Then DOWN and go to ISC LOCK

OPS Eve Shift Start

TITLE: 05/30 Eve Shift: 2330-0500 UTC (1630-2200 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: Corey
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 17mph Gusts, 9mph 3min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.11 μm/s
QUICK SUMMARY:

IFO is LOCKING at PREP_ASC_FOR_FULL_IFO

NLN alignment recovery is still ongoing as we attempt to get past the next few states, at which we have been losing lock.

It seems that IFO is not set up for a good PRC align so we have been skipping that in the initial alignment.

Violin damping May 30 2025

Julia, Georgia

We turned on violin damping filters during lock using the 2W Gaurdian settings.

Some problematic modes:

• ITMX mode 3; wasn't responding to damping
• ITMY modes 5 & 6
• ETMY mode 1; damping rung up the mode

For ETMY mode 1, we noticed that despite using the 2W damping setting, the gain for this filter was 0.1 (rather than 0, which is the value set for the gain in the 2W setting in the VIOLIN_DAMPING.py script). Since we were ringing up the mode with -0.1, we tried setting the gain to 0.05. This successfully damped.

We looked through the VIOLIN_DAMPING.py script and found that the 2W settings were not actually being executed for any of the modes. Georgia found a possible typo in the script where an iteration is ran through "modes.key", but the list is titled "mode" (not "modes"). Georgia then edited this line of code, changing "modes.key" to "mode.key" to fix the issue. We'll confirm this the next time we attempt violin damping when lock is aquired.

GC/CDS switch went down, control room lost offset networking

Jonathan, Erik, Dave:

At 12:28 Fri 30may2025 PDT the GC/CDS switch which provides GC networking to the MSR stopped working. Visually the switch looked functional, but it was not providing network services.

Main impact on the control room was:

 . loss of GC/CDS NAT router, no offsite access; DARM FOM stopped (unable to get references), loss of teamspeak

 . loss of all VOIP phones (they were still getting POE power, but reporting no network service).

On Jonathan's suggestion we power cycled the MSR switch. After a very long startup boot (~8 minutes) all services were recovered.

Down time was 12:28 - 12:58 PDT

DARM FOM was restarted.

This switch is called sw-osb165-msr-0. It is in rack MSR-02 U40. It is a Ruckus ICX 7850-482P (48 port, POE, 2 power supplies).

Attached photos show front and rear of the switch. To power cycle, both rear power cords need to be pulled out, orange=UPS black=FAC (note they have locking brackets which need to be disengaged). Power cords are circled in the photo.

Morning locking on May 30 2025

Sheila has been driving us through ENGAGE_ASC_FOR_FULL_IFO, TMS_SERVO, and ENGAGE_SOFT_LOOPS.

She closed all of the WFS-based ASC loops by hand, with CHARD at reduced gain.  We are sitting in at an alignment which is not the best buildups we've seen so far, but is stable.
The CHARD P sensing matrix is now REFL A 45 I = 1, and REFL B 45 I = 0.5.

Green references have been reset to our current alignment to hopefully make ENGAGE_ASC_FOR_FULL_IFO work automatically next time we acquire, without yanking the IFO out of lock.
Screenshot of the references posted.

Now engaging ADS loops.  This cleaned up the buildups for us a bit.
Now resetting the green references again.

Now turned on the transmon servos, which made a big difference in the green references.
So we are resetting the green references again again.

Now turned on all SOFT loops, reduced gain for PRM Pitch (ADS PIT 3).
Now resetting green references again again again.

Fri CP1 Fill

Fri May 30 10:14:26 2025 INFO: Fill completed in 14min 22secs

 

PRG Estimation 2025-05-30

M. Todd, C. Cahillane, S. Dwyer

---

I ran another estimate of the PRG, following Craig's method in alog  58327, and estimate it to be around 54.2 W/W.

---

I compared the TRX power to Input Power ratio during the FIND_IR state versus the PREP_ASC_FOR_FULL_IFO state, with the PRM transmission as a normalization factor.

                       TRX_POWER_FULL_IFO  /  INPUT_POWER_FULL_IFO
PRG   =    -------------------------------------------------------------------------------------
                     TRX_POWER_FIND_IR   /  (Tp  * INPUT_POWER_FIND_IR)

I've attached the quick script I've used for the interesed reader.

PSL Status Report (FAMIS #26419)

This is for FAMIS #26419.
Laser Status:
    NPRO output power is 1.863W
    AMP1 output power is 70.07W
    AMP2 output power is 140.4W
    NPRO watchdog is GREEN
    AMP1 watchdog is GREEN
    AMP2 watchdog is GREEN
    PDWD watchdog is GREEN

PMC:
    It has been locked 11 days, 1 hr 18 minutes
    Reflected power = 23.1W
    Transmitted power = 105.5W
    PowerSum = 128.6W

FSS:
    It has been locked for 0 days 0 hr and 5 min
    TPD[V] = 0.8038V

ISS:
    The diffracted power is around 4.0%
    Last saturation event was 0 days 0 hours and 55 minutes ago

Possible Issues:
    PMC reflected power is high

Fri Ops Day Transition

TITLE: 05/30 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: Tony
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 8mph Gusts, 4mph 3min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.09 μm/s
QUICK SUMMARY:

ISC lock is in the GREEN_ARMS_MANUAL state.   H1 is listed as PLANNED ENGINEERING.

6.7 dB SQZ, 14.6 dB anti-SQZ measured on homodyne

Camilla, Georgia, Daniel, Vicky, Kevin

We looked at the squeezing on the homodyne as in 83039 and 82202.

We ran into an issue with the mounting of the homodyne PDA in the box. The power on PDA would intermittently drop to almost zero with us just standing there. After Vicky pressed hard on the PD, we could recover the same power after slightly adjusting the alignment on to the PDs again. We ultimately achieved 98% visibility after optimizing the subtraction.

The SQZ angle was recabled to use RF3 in transmission of the OMC rather than RF6 on reflection of the OPO in 83407. Today we split the phase to go to both the OMC and homodyne so that we can use either the interferometer or the homodyne without having to switch the cables as shown in the attached picture.

We then measured between 6.4 to 7 dB of SQZ, 14.6 dB of anti-SQZ, and 12.1 dB of mean squeezing. This is consistent with the measurement of 83% efficiency in 83039: based of off the anti-squeezing that would lead to about an NLG of 11.5 and 6.7 dB of squeezing. We measured an NLG of 9.35 but did see pump depletion so it's plausible that the NLG could be closer to 11.5.

Ops Eve Shift Summary

TITLE: 05/30 Eve Shift: 2330-0500 UTC (1630-2200 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY: See alog84660 for a summary of the commissioning done this evening. Since then, I have been unsuccessful in relocking H1 up to PREP_ASC to do more commissioning of the ASC loops. Sometimes I can make it to that state for up to a few minutes, but eventually there's a lockloss from somewhere I can't explain, possibly the wind or something else. The lockloss tool shows IMC tags for several of the locklosses today, but I haven't been able to find a cause for why the IMC would be dropping out at these times (they certainly don't look to me like the NPRO glitches from last fall that I believe motivated the creation of the IMC tag). I'm leaving H1 in 'DOWN' for the night and commissioning can pick back up in the morning.

Engaging ASC Progress So far

S. Dwyer, J. Driggers, G. Mansell, R. Short, others

The efforts to close more of the ASC loops in ENGAGE_ASC_FOR_FULL_IFO continued this evening. Sheila was able to find some new values for ASC input matrices by stepping different DOFs, and they have been added to ISC_LOCK. With these, we were able to close every ASC loop in the main method of ENGAGE_ASC except for CHARD_Y and ADS (so, INP1_P/Y, PRC2_P/Y, SRC1_P/Y, SRC2_P/Y, and CHARD_P). Jenne relocked ALS-X as it had unlocked with Sheila's CHARD steps, and we started the ALS QPD scripts to keep green arms locked from there. Since CHARD_Y was continuing to confuse us, we decided to move on for the time being and converge ADS error signals with the hope that bringing more alignment together that CHARD_Y would be easier. Jenne tried to move PRM around to bring ADS3 closer to zero, but this caused a lockloss when moving too far. I paused in CHECK_SDF to set the ITM camera references to match the new ALS QPD offsets and accepted them in SDF (see attachment #1 for ALS-X and #2 for ALS-Y).

Before the next lock attempt, I ran an initial alignment with the new green references, doing our usual thing of only doing PRC alignment up to the PRX steps and aligning PRM by hand. Georgia also edited the ENGAGE_ASC state in the following ways with the hope of running the state now that we had several loops closed:

• [lines 3221 - 3239] in PREP_ASC_FOR_FULL_IFO CHARD P and Y intrix values (copied over from time machine)
• [lines 3489 - 3490] in ENGAGE_ASC_FOR_FULL_IFO CHARD Y input on and gain ramp
• [lines 3510 - 3511] in ENGAGE_ASC_FOR_FULL_IFO ADS DOF3 (PRM) gain turn on

Once relocked, we attempted ENGAGE_ASC_FOR_FULL_IFO, and the loops closed successfully! ISC_LOCK was then waiting for ADS to converge (which it wouldn't as ADS wasn't engaged), so Jenne tricked ISC_LOCK by having it think the DAS error signals were zero. This allowed Guardian to move into the run method of ENGAGE_ASC, which has a gain increase of PRC2 gain. This gain increase caused PRC2 to ring up and cause a lockloss.

On another attempt, we made it to PREP_ASC and tried running the loops closing steps in the Guardian shell, but we missed that PRC2_P was on its soft limiter, so it was too far off and caused a lockloss.

At this point, I've been struggling to relock up to PREP_ASC with several random locklosses along the way, even though I've been touching up DRMI alignment on almost every acquisition. Wind speeds have been increasing and just surpassed 30mph, so that may be having some impact.

Network connection to h0vacly drops briefly, apparently correlated to vacuum activity in the LVEA

During the opening of GV7 and GV5 last Wednesday I noticed that the EDC had some brief drop-outs of about 1000 channels. This Wednesday something similar was seen, and on investigation I found that HAM1 IP work was happening at this time.

This piqued my interest and on further investigation I found:

This has happened 9 times so far, the first on Monday 28apr2025.

The EDC disconnects from 1118 channels each time, the only DAQ EDC INI file with this number of chans is H0EPICS_VACLY.ini Trending channels from LY INI show flatlines around these times confirming this is the IOC in question.

The disconnection is brief, most being 8 Seconds in duration, some 9 Seconds, one 16 Seconds. The EDC disconnection number goes from 0 to 1118 in one step at the start, and usually drops back to 0 in one step at the end.

Patrick has found logs on h0vacly saying its CDS network port detected disconnections around these times.

We have no clear mechanism between opening gatevalves, working on chambers, pumping dewars and the h0vacly's network connectivity (none of these activities are anywhere physically close to the vacuum rack). Or as to why h0vacly and not h0vaclx.

Full details in FRS34265

Ops Eve Shift Start

TITLE: 05/29 Eve Shift: 2330-0500 UTC (1630-2200 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: TJ
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 13mph Gusts, 7mph 3min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.10 μm/s
QUICK SUMMARY: Work continues to close ASC loops. We've been able to close about seven of them so far with new input matric values, and we'll continue this evening.

EX windfence work complete

Work at the EX windfence has been completed. While laying out the last panel (furthest North) it was discovered that the new panel had been stitched incorrectly. An old panel that was in good shape was reused.

SPI Double Mixer Finalization Part3

J. Freed,

Spur Problem

As discussed in part 1, the Double Mixer has good phase noise around the carrier frequency to about a 100Hz out where SPI operates, however, it also has signals every 4096Hz away from the carrier.  I had assumed that it was an issue relating to phase and amplitude mismatch on the phase delayer, but after adding attinuators to correct the missmatch, it helped, excepecially the 80Mhz + 4096Hz signal saw a ~15dB improvement. DM_Att.pdf. However, the double mixer still has large signals and the largest did not change with the addition of the attinuators.

I realize now that the problem lies with the mixers themselves as by the nature of a mixer they create spurs on the output. 

DM_Mixer.pdf Shows the output of a single mixer. Along with the expected 80MHz - 4096 Hz and 80MHz + 4096 Hz there is extra peaks which i believe is caused by the Spurs of the mixer. The largest spur is at 80Mhz + 12,288Hz. This spur is actually in phase with the main signal and is thus construtivly interfered at the power combiner. 

I do not know how to get rid of this spur as it is so close to the main carrier signal. For fun, I used a combination of Marki LC filter Design Tool and LTspice to simulate adding a filter to the end of the system to reduce the 80Mhz + 12,288Hz spur, but even an 8th order elliptical filter would only reduce the spur by ~0.04dB due to how close it is to the main signal. For fun I went to 20th order elliptic and it only reduced it by ~0.1dB. In addition, with nonstandard parts for a 20th order elliptic it got reduced by ~4dB. I am unsure of any other method to reduce the spur.

---

Double Mixer is a Single Sideband Mixer

I realize now that the basic design of the double mixer actually exists and is fairly common, it is called a Single Sideband Mixer (SSB mixer). Thanks to Brian, who gave me one of these mixers to test, I have something to compair the results  

DMvsIRM.pdf Shows a plot of the double mixer vs a SSB mixer. The SSB mixer is tuned to 80MHz+4096Hz as apposed to the Double Mixers 80MHz -4096Hz but comparisons can still be made. For starters, the double mixers 80Mhz+4096Hz sideband equivalent is much better in the double mixer vs the SSB. However this comes at the cost of the 80Mhz+12,288Hz sideband equivalent which is much better in the SSB. In summary, just based on this single graph, the double mixer is more likely to be better for SPI for 2 reasons. One, we have already found out that SPI would be senstive to the 80MHz+4096Hz sideband equivalent. And two, if there is a way to attinuate these spurs by filtering, the spur further away from the carrier could be more easily attinuated.

---

High order modes filtering

Ive already discussed that adding fliters does not help the area around 80MHz. However, it does help with higher order modes

DM_Filter.pdf Shows a plot of adding a MC80-10-4AA bandpass filter from Lark engenering that has a bandwidth of 10MHz and a center of 80MHz at different positions in the Double Mixer. The best place to put one seems to be just before the amplifier, not entirly sure if adding one is nessisary to SPI but it should be good for reduceing a possible noise source.

Sideband PRG Calibration

Anamaria took some time with me today to help me calibrate the sideband PRGs the same way she calculated them here for Livingston (LLO:69872). This is mostly with the intention of providing calibrated data for the finesse modeling effort that can help us better understand IFO behavior during TCS changes/powering up.

Reminder: I ran a check of the carrier PRG calibration this week (81752), which is based on the IM4 trans and end transmission values, which confirms that at 2W, our PRG is about 55.

Method:

Measure the power in the PRC at DRMI lock and compare it to the power in the PRC at 2W lock. Knowing the gamma of the 9 and 45 MHz sideband, estimate how much of the input power is sideband power. Use the predicted sideband PRG from an IFO model with perfect mode matching, and scale the expected sideband power in the PRC by those PRGs. Compare this expected value to the actual PRC power, and then scale the sideband PRGs accordingly.

Data:

DRMI lock values (measured after DRMI ASC converges):

Channel	Value
POP A LF INMON	12.5 ct
POPAIR RF18	151 ct
POPAIR RF90	27.25 ct
IM4 TRANS NSUM	1.98 W

2W full lock values (measured after full IFO ASC converges, including ADS):

Channel	Value
POP A LF INMON	283.8 ct
POPAIR RF18	154.5 ct
POPAIR RF90	28.5 ct
IM4 TRANS NSUM	1.98 W

Other useful numbers:

Parameter	Value	Notes
Gamma 9	0.213	22.7 mW 9 MHz per W of input power, 62883
Gamma 45	0.275	37.8 mW 45 MHz per W of input power, 62883
PRM Transmission	0.031	https://galaxy.ligo.caltech.edu/optics/
PR2 Transmission	229 ppm	https://galaxy.ligo.caltech.edu/optics/
M12 Transmission	0.054	Splitter on POP path, 63625
2W lock carrier PRG	55	Original measurement, checked again recently
PRG9 from model	119	Anamaria's Optickle model
PRG45 from model	13	Anamaria's Optickle model

The work:

At 2 W lock, POP LF INMON measures 283.8 counts (this accounts for the dark offset). With an input power of 1.98 W and a PRG of 55, this should be 109 W. So we set a scale that 283.8 ct POP diode == 109 W PRC power.

At DRMI lock, we measure 12.5 counts in the POP diode (this accounts for the dark offset), so this gives us 4.8 W PRC power in DRMI lock.

At 1.98 W input power, 44.9 mW is 9 MHz power and 74.84 mW is 45 MHz power. We multiply by the perfect mode matching sideband PRGs and get that 5.3 W should be 9 MHz and 0.972 W should be 45 MHz in DRMI lock, so a total of 6.3 W expected PRC power in DRMI.

4.8 W / 6.3 W = 0.76, so 0.76 * 119 = 90.6 and 0.76 * 13 = 9.9, so we can say that our 2W PRG9 = 91 and PRG45 = 10.

Notice that this method ties the sideband PRGs to the measured carrier PRG. The method is perhaps not so good for the 45 MHz sideband, since we know there is so much less 45 MHz power than 9 MHz power. But these values are sensible.

Why did we do it this way?

You might notice we used the POP LF inmon value instead of the POP LF out value. There is a calibration in the POP A LF filter bank labeled "to_uW", which I assume means "microwatts of power on the POP diode". We started by using this calibrated value, and putting it into Watts of PRC power using the known transmission of the M12 splitter and PR2 (LLO calibrates POP A LF into Watts in the PRC). This resulted in a lower sideband PRG, (76.7 and 8.4 respectively). However, it also predicted that the carrier PRG should be 46. The carrier PRG calibration Craig and I performed does not rely on the POP diode at all, which is good because there is enough sideband power in the PRC to confuse the measurement. So, Anamaria and I decided to trust the independent carrier PRG measurement based on the arm transmission. Furthermore, we looked at how the H1:LSC-PR_GAIN_OUT16 channel is calibrated, and noticed that it ratios the H1:LSC-POP_A_LF_INMON value with IM4 trans. This motivated us to ignore the uW calibration filter (which then motivated us to correct for the dark offset of 3.8 ct).

Comparisons to mod depth method

If you have a good enough alog memory you might know that Craig and Jennie W have done a sideband PRG calibration in the past, using the modulation depth step test such as here: 70865 and 76358. Upon further inspection of the results, I am not sure they are so good. Mainly, the 45 MHz sideband PRG that results from this test is around 27-30, which is huge. I looked at these measurements again to try to understand why the values are so different. I noticed that when Craig steps down the 45 MHz sideband, the POP A LF value actually increases. Anamaria and I think this is because with lower 45 MHz modulation, the carrier input power increases, and the high carrier PRG (compared to the 45 MHz PRG) amplifies this excess power further. I can't find anywhere in Craig's code that accounts for this. For now, I am going to trust the method described in this alog over the mod depth step method.

Channel calibration factor

These values will calibrate the POPAIR B RF18 I ERR and RF90 I ERR channels into sideband PRGs accordingly.

calibration factor = (2W sideband PRG value) / ( 2W POPAIR count value / IM4 trans)

POPAIR18 > PRG9 calibration factor = 91 / ( 154.5 / 1.98) = 1.17, and normalize by input power

POPAIR45 > PRG45 calibration factor = 10 / (28.5 / 1.98) = 0.7, and normalize by input power

Note that if you use the POPAIR B RF18/90 I NORM MON channels instead of the ERR channels, these are already normalized by input power.

You are right that when the 45 MHz goes down, the carrier goes up.  
That's why POP LF and the arm transmissions all rise when we lower our mod depths.
The mod depth code does account for this, otherwise it would not work at all, see Daniel's original work.
It gives an estimate that the arm powers are 100% carrier, for instance, which comes from the fact the arm transmissions rise.