I wrote a script that automatically inject lines at 138 Hz in MICH, PRCL, SRCL, and measures the sensing matrix to POP and REFL RF signals.

The script is available in /opt/rtcds/userapps/release/isc/h1/scripts/lsc_sensing_matrix.py

Options are set at the beginning of the file. In particular excitations amplitudes and measurement time. The current excitations are tuned for full IFO in "high noise" state, i.e. before DC readout and before coil drivers to low noise.

The script outputs a matrix of the transfer functions from MICH2/PRCL2/SRCL2 _OUT to the POP_RF* and REFL_RF* signals. Also, it outputs a matrix of coherence of all signals with the excitation. The output is both in ASCII format and in HTML format, so that it can be easily copied and pasted into the logbook. The script saves the ASCII and HTML outputs to a file named lsc_sensing_DATE_TIME.txt and lsc_sensing_DATE_TIME.html.

A measurement performed at 2 W, before DC readout, it shown below:

The script also produces a I/Q radar plot, which is saved in lsc_sensing_DATE_TIME.png. However, I have not (yet) calibrated the LSC_*2_OUT signals in meters. Since those signals are going directly to BS, PRM and SRM suspensions, the numbers that Craig posted should be fine.

Topped off Crystal Chiller with 125mL.
Crystal Chiller filter is more discolored than the Diode Chiller's filter; will send Jeff B. an email.

~0930 - 0935 and ~0945 - 0955 hrs. local

Better late than never. All plots look to be nominally sound.

TITLE: 10/17 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    Wind: 4mph Gusts, 3mph 5min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.26 μm/s
QUICK SUMMARY:

Commissioners working on H1 after a quiet night seismically, & an ALL-GREEN CDS overview.

This morning I put the DRMI ASC engagement back in, so that we're not only doing MICH ASC but all vertex ASC loops are closed and get offloaded before we start the CARM offset reduction sequence.  This should remove the need for us to constantly be adjusting the alignment of mirrors by hand in DRMI. 

So far this morning, however, I can't sit at full RF lock with the ASC loops closed.  I didn't try before I put the DRMI ASC back in, but the change in optic position is pretty small (as one would expect) when they get offloaded, so I don't see a reason why this would cause trouble.  Both by hand and with ASC engaged, I cannot get the POP18_I_NORM above 85, when normally (before yesterday's TCS tunings) we could reliably get to 93.  So, not a big difference, but certainly noticeable. 

The oscillation we see happens after we've finished SRC_ASC, after several tens of seconds.  It's about 1Hz, so not the same as the 20Hz oscillation that they were seeing last night at 20W.  Under investigation...

I just finished adding a few lines to the ISC_LOCK guardian to lock the MICH length degree of freedom in the CHECK_MICH_FRINGES state, with the same settings we use during initial alignment.  When the BS is poorly aligned (as it was this morning, but things were good enough that I didn't feel the need to do a full initial alignment) and one needs to check the BS alignment, it can be quite slow, waiting for fringes to come and go.  So, now MICH locks and the AS Air camera exposure is increased so that we can align the BS by hand more easily.  Eventually we'll add ASC to this state if it's needed, but this should already be an improvement.

Sheila, Jenne, Gabriele, Hang, and Danny

• Today we started off with the new RH and CO2 settings suggested in alog 44576. I initially didn't consider the double pass through the compensation plate so the change in CO2 needed to be reduced by a factor of 2. Even with this new CO2/RH setting the sideband imbalance at 20 W was still off and we were losing lock on time scales that could not accommodate CO2 adjustments. 
  • TCS tuning at 10 W
    • With CO2X  at 0.1W and CO2Y at 0 W is when we saw the best upper and lower sideband balance
    • Increased CO2X and CO2Y by .3 W to restore the ITMY HWS spherical power when IFO is at 2 W
  • Powering up to 20 W
    • Found that the following IFO hot and cold settings were the best settings of the evening:
      • Cold state
        • CO2X : .5 W
        • CO2Y :  .6 W
      • Hot state
        • CO2X :  .1 W
        • CO2Y :  0 W
    • These settings are already implemented in guardian. 
    • The interferometer is still not stable at 20 W
• Side note: We only used ITMY HWS for this tuning because of the large amount of noise on the ITMX HWS. Work is being done to make this sensor more usable. 

On October 15th Danny and I left the interferometer locked with the request set to low noise when we left.  It was left this way from 3:20 UTC to around 13 UTC when Gabriele and Jenne arrived and re-aligned it. Summary page 

There were 9 locking attempts over the 10 hour period, the interferometer made it to low noise 6 times, lost lock once from the state park ALS VCO (currently this happens right after the transition to RF DARM and engaging the DHARD WFS), once while closing the beam diverters, and once while trying to transition the common arms onto the transmitted IR signals.  

The lockloss page makes it easy to look up things like this.  All of these locks were short.  

Today the violins were rung up again, either because of the large earthquake in New Caledonia last night or because of some maintence day activity.  

I've added ITMX mode1 (504.145Hz) to the guardian, it is damped using yaw with a positive gain and no phase shift.  

We are still having difficulties with the close pair of modes at 510.714 (ETMYMODE1) and 510.723Hz (called ETMY mode6).  I haven't seen any impact on 510.714 by actuating on ETMY in pit or yaw.  I am starting to wonder if that mode is really on ETMY. Mode1 is really on ETMY and can really be damped with a positive gain, although it is rung up badly right now.  

Sheila, Hang

During power-ups today we saw some ~0.5 Hz oscillations in DSOFT Pitch that kept causing locklosses. Therefore we copied our CSOFT P setup to DSOFT P to increase its bandwidth. Specifically, we blended in to original DSOFT P signal (a comb of 4 trans QPDs that decouples the TMS drift) another comb of trans QPDs that are insensitive to DHARD. The blending freq is at 0.03 Hz, same as CSOFT. After the blending, for DS P we were able to engage the HBW control filters we used for CS P and pushed the filter bank gain up to 30. We also noticed that the input matrix for DS was a factor of 2 higher than the CS value, thus our current DS P loop should have a BW twice that of CS P, sufficient to damp the 0.5 Hz oscillation.

We have put the new setup in guardian and it seems fine to engage the soft loops with this new config.

Attached plot shows 30 day trends of both BRSs DC positions and a temperature sensor in each box. Plots on the left are the DC position, plots on the right are the temps, EY is on top, EX is on bottom. It looks like both buildings have been heating up for a while, but BRSX is closer to it's limit and the temperature drift if pushing BRSX close to it's limits. The temperature in the box at EX is also about a half deg C higher than 30 days ago and that temperature is less stable now. This was going bad on it's own over the last two weeks, but the EX situation is made worse by my incursion today. I could re-balance if needed, but that's not a zero risk activity.

I have long been concerned about the IO High Power Beam Dump (IO HPBD) at the North End of the PSL table, and it's cooling hoses hanging down from the beam dump, as a source of PSL table vibration.

I talked to Jason about doing a quick test, buy putting strain relief on the IO HPBD, and then looking at the PSL accelerometers.  The hoses are taped the side of the NE corner of the PSL table, wrapped in foam, so no direct contact with the table, and the hoses are not touching each other.

The results are clear, this change reduced the accelerometers broadband noise, and eliminated peaks in the Periscope accelerometers at 413Hz.  There are some peaks and broadband noise that increased, but I'm not worried about that, since we literally used tape, and the decrease and elimination of other peaks shows that careful strain relief of cooling lines will have a positive effect on the PSL table motion.

The IO HPBD also needs a better mount, as it's a very heavy component, made from metal plates and SiC plates, and currently it is mounted with a post and fork clamp.

No a bad result, for three pieces of tape, a small bit of foam.

Plots show power spectrums for all PSL ACCs, and coherence between the Periscope X and Y ACCs and the table Accs.

In the plots, all data from yesterday, 10/15, at 19:17UTC are shown in black, and data from today, 10/16, at 19:17UTC are shown in colors other than black.

- Jason, Cheryl

During all lock stretches, the sensitivity steadily degrades over time. The plot below shows the range trend over one of the last locks.

I picked a few times during the lock, when there were likely no actions on the IFO, and computed the DARM spectrum. The evolution is shown below

Clearly, what degrades the sensitivity is very well approximated by a 1/f displacement noise.

Here's a comparison of degrading sensitivity over another long lock stretch. Similar 1/f noise

D. Barker, J. Kissel
WP 7850
ECR E1800246
IIET 11305

Some more belated documentation in support of the front-end model changes that we made last week in order to support new / updated calibration code for O3:
Below is an explicit list of front-end generated channels that need to be removed from the GDS broadcaster, and a list of the new channels that should be used instead and/or replaced. It's also attached for any user's convenience.
One can cross-reference the list of new (slow) channels in our documentation that supports the new calculation for time-dependence, T1700106.

Channels to be removed:

FAST CHANNELS:
    CAL-CS_LINE_SUM_DQ                           # Agreed, disappears (DARM_LINE no longer exists)
    CAL-DARM_CTRL_WHITEN_OUT_DBL_DQ              # Agreed, replaced by simply DARM_CTRL_DQ (which is, yes, whitened, and stored as a double)
    CAL-DARM_ERR_WHITEN_OUT_DBL_DQ               # Agreed, replaced by simply DARM_ERR_DQ (which is, yes, whitened, and stored as a double)
    CAL-INJ_ODC_CHANNEL_OUT_DQ                   # Agreed, removed because we'll never do hardware injections from the cornerstation (i.e. on the ESD)

SLOW CHANNELS:
    CAL-CS_TDEP_DARM_LINE1_REF_A_TST_IMAG        # Agreed, disappears (no more DARM_LINE)
    CAL-CS_TDEP_DARM_LINE1_REF_A_TST_REAL        #      |
    CAL-CS_TDEP_DARM_LINE1_REF_A_USUM_IMAG       #      |
    CAL-CS_TDEP_DARM_LINE1_REF_A_USUM_INV_IMAG   #      |
    CAL-CS_TDEP_DARM_LINE1_REF_A_USUM_INV_REAL   #      |
    CAL-CS_TDEP_DARM_LINE1_REF_A_USUM_REAL       #      V
    CAL-CS_TDEP_DARM_LINE1_UNCERTAINTY           #      --
    CAL-CS_TDEP_ESD_LINE1_REF_A_TST_NOLOCK_IMAG        # Agreed, disappears (renamed to SUS_LINE1)
    CAL-CS_TDEP_ESD_LINE1_REF_A_TST_NOLOCK_REAL        # Agreed, disappears (renamed to SUS_LINE1)
    CAL-CS_TDEP_ESD_LINE1_REF_C_IMAG                   # Agreed, disappears (renamed to SUS_LINE1)
    CAL-CS_TDEP_PCALY_LINE1_UNCERTAINTY                # Agreed -- made generic to P    CAL_LINE1
    CAL-CS_TDEP_PCALY_LINE2_REF_A_TST_IMAG             # Agreed -- made generic to PCAL_LINE2
    CAL-CS_TDEP_PCALY_LINE2_REF_A_TST_REAL             #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_A_USUM_IMAG            #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_A_USUM_REAL            #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_C_NOCAVPOLE_IMAG       #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_C_NOCAVPOLE_REAL       #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_D_IMAG                 #     |
    CAL-CS_TDEP_PCALY_LINE2_REF_D_REAL                 #     V
    CAL-CS_TDEP_PCALY_LINE2_UNCERTAINTY                #     --
    CAL-CS_TDEP_PCALY_LINE3_UNCERTAINTY                # Agreed -- made generic to PCAL_LINE3
    CAL-CS_TDEP_PCALY_LINE4_REF_A_TST_IMAG             # Agreed -- made generic to PCAL_LINE2
    CAL-CS_TDEP_PCALY_LINE4_REF_A_TST_REAL             # Agreed -- made generic to PCAL_LINE4
    CAL-CS_TDEP_PCALY_LINE4_REF_A_USUM_IMAG            #     |
    CAL-CS_TDEP_PCALY_LINE4_REF_A_USUM_REAL            #     |
    CAL-CS_TDEP_PCALY_LINE4_REF_C_NOCAVPOLE_IMAG       #     | 
    CAL-CS_TDEP_PCALY_LINE4_REF_C_NOCAVPOLE_REAL       #     |
    CAL-CS_TDEP_PCALY_LINE4_REF_D_IMAG                 #     V
    CAL-CS_TDEP_PCALY_LINE4_REF_D_REAL                 #     --
    CAL-CS_TDEP_REF_CLGRATIO_CTRL_IMAG                 # Agreed -- no longer needed
    CAL-CS_TDEP_REF_CLGRATIO_CTRL_REAL                 # Agreed -- no longer needed             
    CAL-CS_TDEP_REF_INVA_CLGRATIO_TST_IMAG             # Agreed -- renamed to (for a given SUS_LINE) INVA_$(ISOSTAGE)_RESPRATIO 
    CAL-CS_TDEP_REF_INVA_CLGRATIO_TST_REAL             # Agreed -- renamed to (for a given SUS_LINE) INVA_$(ISOSTAGE)_RESPRATIO 


New Channel List:

FAST CHANNELS: 
('IFO generic channel name' 'sample rate' 'data storage type')
(when sample rate not listed, rate is 16384 Hz)
   CAL-CS_DARM_ERR_DQ double
   CAL-CS_DARM_CTRL_DQ double

   CAL-CS_DELTAL_RESIDUAL_DQ double
   CAL-CS_DELTAL_CTRL_DQ 4096 double
   CAL-CS_DELTAL_CTRL_TST_DQ 4096 double
   CAL-CS_DELTAL_CTRL_PUM_DQ 4096 double
   CAL-CS_DELTAL_CTRL_UIM_DQ 4096 double
   CAL-CS_DELTAL_EXTERNAL_DQ

   CAL-CS_CFTD_DELTAL_RESIDUAL_DQ double
   CAL-CS_CFTD_DELTAL_CTRL_DQ 4096 double
   CAL-CS_CFTD_DELTAL_CTRL_TST_DQ 4096 double
   CAL-CS_CFTD_DELTAL_CTRL_PUM_DQ 4096 double
   CAL-CS_CFTD_DELTAL_CTRL_UIM_DQ 4096 double
   CAL-CS_CFTD_DELTAL_EXTERNAL_DQ

SLOW CHANNELS:
   CAL-CS_TDEP_PCAL_LINE1_CORRECTION_REAL
   CAL-CS_TDEP_PCAL_LINE1_CORRECTION_IMAG
   CAL-CS_TDEP_PCAL_LINE2_CORRECTION_REAL
   CAL-CS_TDEP_PCAL_LINE2_CORRECTION_IMAG
   CAL-CS_TDEP_PCAL_LINE4_CORRECTION_REAL      # yes 4, not 3
   CAL-CS_TDEP_PCAL_LINE4_CORRECTION_IMAG
   CAL-CS_TDEP_SUS_LINE1_REF_INVA_UIM_RESPRATIO_REAL 
   CAL-CS_TDEP_SUS_LINE1_REF_INVA_UIM_RESPRATIO_IMAG
   CAL-CS_TDEP_SUS_LINE2_REF_INVA_PUM_RESPRATIO_REAL 
   CAL-CS_TDEP_SUS_LINE2_REF_INVA_PUM_RESPRATIO_IMAG
   CAL-CS_TDEP_SUS_LINE3_REF_INVA_TST_RESPRATIO_REAL 
   CAL-CS_TDEP_SUS_LINE3_REF_INVA_TST_RESPRATIO_IMAG
   CAL-CS_TDEP_PCAL_LINE2_REF_C_NOCAVPOLE_REAL 
   CAL-CS_TDEP_PCAL_LINE2_REF_C_NOCAVPOLE_IMAG
   CAL-CS_TDEP_PCAL_LINE4_REF_C_NOCAVPOLE_REAL 
   CAL-CS_TDEP_PCAL_LINE4_REF_C_NOCAVPOLE_IMAG
   CAL-CS_TDEP_PCAL_LINE2_REF_D_REAL 
   CAL-CS_TDEP_PCAL_LINE2_REF_D_IMAG
   CAL-CS_TDEP_PCAL_LINE4_REF_D_REAL 
   CAL-CS_TDEP_PCAL_LINE4_REF_D_IMAG
   CAL-CS_TDEP_PCAL_LINE2_REF_A_UIM_REAL   
   CAL-CS_TDEP_PCAL_LINE2_REF_A_UIM_IMAG
   CAL-CS_TDEP_PCAL_LINE2_REF_A_PUM_REAL  
   CAL-CS_TDEP_PCAL_LINE2_REF_A_PUM_IMAG
   CAL-CS_TDEP_PCAL_LINE2_REF_A_TST_REAL  
   CAL-CS_TDEP_PCAL_LINE2_REF_A_TST_IMAG
   CAL-CS_TDEP_PCAL_LINE4_REF_A_UIM_REAL  
   CAL-CS_TDEP_PCAL_LINE4_REF_A_UIM_IMAG
   CAL-CS_TDEP_PCAL_LINE4_REF_A_PUM_REAL  
   CAL-CS_TDEP_PCAL_LINE4_REF_A_PUM_IMAG
   CAL-CS_TDEP_PCAL_LINE4_REF_A_TST_REAL  
   CAL-CS_TDEP_PCAL_LINE4_REF_A_TST_IMAG
   CAL-CS_TDEP_SUS_LINE1_REF_UIM_NOLOCK_REAL  
   CAL-CS_TDEP_SUS_LINE1_REF_UIM_NOLOCK_IMAG
   CAL-CS_TDEP_SUS_LINE2_REF_PUM_NOLOCK_REAL  
   CAL-CS_TDEP_SUS_LINE2_REF_PUM_NOLOCK_IMAG
   CAL-CS_TDEP_SUS_LINE3_REF_TST_NOLOCK_REAL 
   CAL-CS_TDEP_SUS_LINE3_REF_TST_NOLOCK_IMAG


This should help alleviate some of the problems that folks have been having with the GDS Broadcaster since last week's upgrade (e.g. LHO aLOG 44454).