More on 4.05 Hz

More peak hunting for the 4.05 Hz bumps. I tweaked a bit the parameters for my peak finder code, adn the channel list got a bit bigger. Full list below.

Most interesting are

H1:PEM-CS_ACC_ISCT1_REFL_Y_DQ
H1:PEM-CS_ACC_LVEAFLOOR_HAM6_Z_DQ
H1:PEM-CS_TILT_LVEA_VERTEX_T_DQ

It looks like something it's shaking the ground at 4.05 Hz in the LVEA. The 4.,05 Hz peak in those channels is coherent.
 

           H1:PEM-CS_ACC_ISCT1_REFL_Y_DQ
       H1:PEM-CS_ACC_LVEAFLOOR_HAM6_Z_DQ
            H1:PEM-CS_ADC_5_26_2K_OUT_DQ
            H1:PEM-CS_ADC_5_29_2K_OUT_DQ
         H1:PEM-CS_MAG_EBAY_LSCRACK_Y_DQ
         H1:PEM-CS_MAG_EBAY_LSCRACK_Z_DQ
         H1:PEM-CS_MAG_EBAY_SUSRACK_Y_DQ
         H1:PEM-CS_MAG_EBAY_SUSRACK_Z_DQ
          H1:PEM-CS_MAG_LVEA_VERTEX_X_DQ
          H1:PEM-CS_MAG_LVEA_VERTEX_Z_DQ
    H1:PEM-CS_RADIO_EBAY_NARROWBAND_1_DQ
    H1:PEM-CS_RADIO_EBAY_NARROWBAND_2_DQ
    H1:PEM-CS_RADIO_LVEA_NARROWBAND_1_DQ
    H1:PEM-CS_RADIO_LVEA_NARROWBAND_2_DQ
         H1:PEM-CS_TILT_LVEA_VERTEX_T_DQ
               H1:PEM-EX_ADC_0_09_OUT_DQ
          H1:PEM-EX_LOWFMIC_VEA_FLOOR_DQ
         H1:PEM-EX_MAG_EBAY_SUSRACK_X_DQ
         H1:PEM-EX_MAG_EBAY_SUSRACK_Y_DQ
         H1:PEM-EX_MAG_EBAY_SUSRACK_Z_DQ
      H1:PEM-EX_TEMPERATURE_BSC9_ETMX_DQ
       H1:PEM-EX_VMON_ETMX_ESDPOWER48_DQ
               H1:PEM-EY_ADC_0_14_OUT_DQ
         H1:PEM-EY_MAG_EBAY_SEIRACK_X_DQ
         H1:PEM-EY_MAG_EBAY_SEIRACK_Y_DQ
         H1:PEM-EY_MAG_EBAY_SEIRACK_Z_DQ
         H1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ
         H1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ
            H1:PEM-EY_MAINSMON_EBAY_2_DQ
                   H1:CAL-PCALX_IRIGB_DQ
                   H1:CAL-PCALY_IRIGB_DQ

 

The 4.05 Hz line in the CS accelerometers have to be electrical coupling. They're identical in shape and amplitude ISCT1 and HAM6 despite being different DoFs and locations. They also are coherent with an unused ADC channel, PEM-CS_ADC_5_29_2K. It was used for DAQ stand testing but right now it's almost certainly disconnected. It sees a structure of odd harmonics of 1.35 Hz (4.05 is the third harmonic) which is coherent with both of the accelerometers. I think this structure is contaminating all of the electronics in the CS that Gabriele has identified. The PSDs, a zoomed PSD, and the coherence are attached.

So it's not motion, but it does mean there's electrical noise at 4.05 Hz, and it's in many PEM channels at least. The current clamp ADC_5_26 sees it, but is it actual current or just in the ADC?

In-Lock Charge Measurements

Closes FAMIS 25063

Attached are 4 images of in-lock charge measurements for 3 of the quads. ETMX was omitted as the data produced by the script was bad, alog on that here.

Measurements in the past month or so look for the most part, similar. Ryan last ran these measurements a month ago, alog here.

Start of filling in DEMOD infrastructure for PCALX Systematic Error Lines

J. Kissel

I've walked through the new front-end (LHO:68961) and MEDM (LHO:68999) infrastructure that supports demodulation of the new(ish) PCALX systematic error lines in order to get it all up and running so that we have a live, real-time estimate of the systematic error in the calibration at any given time. 

This walk through includes: 
    - Turning on the corner station local oscillators to match the parameters of the PCALX end station drive oscillators
    - Installing "BP[linefreq]" filters in each PCAL, DARM, and EXT_SIG banks which band pass the "Opposite" reference PCAL (as it's currently set up, this is PCALX), DARM_ERR, and DELTAL_EXTERNAL. For these, I've chosen, for now, I've chosen a 6th order butterworth band pass, with +/- 0.05 Hz corner frequencies symmetrically surrounding the line, i.e. for the 33.43 Hz line, the FM1 "BP33.43" design is butter("BandPass",6,33.38,33.48). The symmetry ensures that at the line frequency, the transfer function has unity magnitude and zero phase. The +/-0.05 Hz bandwidth means that the peak of the filter's impulse response is around ~15 seconds, with a 1/e time within ~20-30 seconds of the impulse.
    - Installing "LP40s" filters in each following I and Q bank, low-passing the demodulated signal to mitigate contamination of the 1f and 2f components of the demodulation process. I chose 
 and started installing SIG band passes and I and Q low passes. Here, in FM1 of the I and Q banks, I chose a 4th order butterworth, with corner frequency at 0.025 Hz, butter("LowPass",4,0.025). This filter has an impulse response of ~40 seconds, as a compromise between "don't limit my patience in getting the answer" and "don't make the impulse response of I and Q low pass filters shorter than the impulse response of the SIG bandpass."  

All of these filters, the bank inputs, and back outputs have been turned ON, and the filter bank gains have been set to 1.0. 
These settings have all been saved in the OBSERVE.snap (and therefore safe.snap) SDF file(s).
The H1CALCS.txt foton file has been committed to the userapps repo after all the above filters have been installed.

Now we "just" need to populate the EPICs records that account for the flaws in each of the DELTAL and PCAL channels, in the corresponding channels (e.g. for the 33.43 Hz line which is processed by the 5th DEMOD, filling in H1:CAL-CS_TDEP_PCAL_LINE5_DELTAL_PCAL_CORR_REAL and H1:CAL-CS_TDEP_PCAL_LINE5_DELTAL_PCAL_CORR_IMAG). Evan and I will work together to do that tomorrow morning.

error packates on sw-ey-h1fe

Interface 1/1/7 has some interface errors.  I will watch the interface over the next day or two to see if they increase.  Fil may trace the cable tomorrow and see what it goes to, there are no comments in the switch config.  I don't have good history, the switch was replaced but the network monitoring system didn't get updated so our stats are stale for that switch (I got it working in the network monitoring software today).

Some snooping around shows it to likely be connected to pwr-ey-h1fe-ac1.

The mac address table from the switch:

0019.8500.991d  1/1/7                Dynamic      99   

The arp table on the core (sh ip arp | include 0019.8500.991d):

70.3.99.10.in-addr.arpa name = pwr-ey-h1fe-ac1.cds.ligo-wa.caltech.edu.


The error counter on the switch:

1269 input errors, 0 CRC, 0 frame, 0 ignored
1269 runts, 0 giants  

Unmonitored SQZ Angle in sdf. Now no sqz observe SQZ SDF diffs.

Camilla, Naoki, Nutsinee. Unmonitored SQZ Angle (H1:SQZ-CLF_REFL_RF6_PHASE_PHASEDEG) along with related channels (H1:SQZ-CLF_REFL_RF6_PHASE_DELAY{STEP,NS}) in sdf. Ideally this would remain constant but for now we are having to change the angle during thermalization to optimize SQZ. There is now no observe SQZ SDF diffs.

End Station Microphones

WP 11121

End Station microphones at both End Stations have been upgraded.

1. Configuration of DC power supplies modified. New microphones only require 5V
2. Field power cables re-terminted changed from a 3 pin LEMO to CPC
3. BNC cables pulled/terminated from power conditioning box to microphones

Plan is to start upgrade of microphones in Corner Station next week.

RF Intensity Noise Monitor for Squeezer Laser

Nutsinee Daniel

We added an 1811 photodetector to monitor the RF intensity noise of the squeezer laser. The new photodetector is mounted next to the 1611 which is used for the laser locking. There was a 50:50 beam splitter in the path of the laser to the 1611 that dumped about 3.6mW of light onto a beam dump. We added a 03 ND filter and a 150mm lens to this path and ended up with about 400uA of DC current on the 1811. The RF and DC readouts use cables ISC_SQ_540_1 and ISC_SQ_541_1, respectively. They are routed thru the PEM panel that had 2 TNC feedthru adapters available. The RF is then routed by cable ISC_SQ_540_2 to the same RF patch panel used for homodyne and OMC RF at position 3. We diconnected the RF of the homodyne from the demodulator and hooked up the new 1811 RF channel. The DC is routed by cable ISC_SQ_540_2 to the D-sub breakout panel just beneath it on position 4. The corresponding filter module is H1:SQZ-EXTRA_AI_4_OUT.

Ops Eve Shift Start

TITLE: 04/25 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Commissioning
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 13mph Gusts, 8mph 5min avg
    Primary useism: 0.06 μm/s
    Secondary useism: 0.10 μm/s 
QUICK SUMMARY:

- IFO has been locked for ~1 hr and PCAL/CARM measurements are currently ongoing

- CDS/DMs ok

Tuesday Afternoon OPS Shift End Update

TITLE: 04/25 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
SHIFT SUMMARY:
LOG:
IFO Is currently Locked in NLN_CAL MEAS

Today's Lockloss List :
First lockloss happened right after I requested SEI ENV to go to Maintenance mode.
https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/index.cgi?event=1366470047

Maintenace:
-(7:45am-8am) In-lock Charge Measurements if IFO is locked.

-8am - FAC - More Vinyl Work-                                                                      

-APS door security vendor at FCES/VEA door and high bay LVEA door

-8am - EY Charge (Rahul)

-8am - Corner, then EX, then EY - Jim /CDS - Restart BSC ISI and SEIPROC models.
Continuing ECR updates from HAM work last week, seiproc is getting some dolphin
channels for related to eq mode. Will require daq restarts, and time to copy filters in.

-CDS - DAQ restart to take in 2 new Guardian nodes and remove one node .

- Update to CAL-CS and PCAL Models (Kissel)

-CDS - Marc - Replace two identified overheating power supplies in the CER.
Mezanine Rack C1 U29-U31 LeftHandSide (LHS) +18V 5A -- (SUS C6 ITMs BS) --
Mezzanine Rack C5 U13-15 RightHandSide (RHS) -18V 5A -                                            

-EY - CDS -changed the connections for the daq data and regular FE ip traffic between end-y and mid-y. We changed from standard SFPs to bi-directional sfps.

-EY - VAC - perform functionality test on turbo/hepta pumps.

Post Maintenance Locking:
Tried to use H1 Manager Gaurdian for initial alignment.
It struggled to get green arms just right, but was able to get through initial alignment just fine.
After that it re-requested initial alignment agian, TJ and I then took over and started the normal locking process with out H1 Manager.

Once I started locking it was smooth sailing all the way up to NOMINAL_LOW_NOISE
Comissioning button set so Elena can do some tests.

NUC 30 (DARM screen) needs to have GDS removed from the channels to run correctly for more than an hour or 2.

22:05 IFO taken to NLN_Cal_Meas

22:18 Erik Von Reese turned GDS back on the DARM screen. we are trying to determine how long it takes to freeze the DARM display.
                                                                                                                                                                                                                                                                                                                        

Start Time	System	Name	Location	Lazer_Haz	Task	Time End
14:54	FAC	Karen	EY	N	Technical cleaning	16:33
14:55	FAC	Kim	EX	N	Technical Cleaning	16:16
15:07	SUS	Ryan C	CTRL Rm	N	EY Charge measurements	16:03
15:16	FAC	Vinyl	LVEA	N	Flooring work	19:16
15:19	CDS	Dave Barker		N	CDS Model & DAC Restarts	19:17
15:32	EE	Fil	CS,EX&Y	N	PEM Power Supply & Microphone work	19:17
15:38	VAC	Gerardo	CS , EX &Y	N	Vacuum System work	19:18
15:40	FAMIS	Ryan S	CS, EX & Y	N	HEPI & dust Pump FAMIS Tasks	16:31
15:41	FAC	MItchel & Randy	EY	N	Wind Fence repair work	18:40
16:08	FAC	Chris	LVEA	N	Filling up forklift batterys	16:45
16:15	FAC	Apollo	FCES	N	Installing locks and security	21:14
16:23	SEI	Jim	CTRL Rm	N	Restart BSC ISI AND SEIPROC models	17:03
16:30	VAC	Jordan, Travis	EX & EY	N	Vacuum work	19:15
16:41	EE	Marc & Fernando	CER	N	Replacing 2 Power supplies, SUS taken to Safe!	17:38
16:44	OPS	Ryan S	EY	N	Wind Fence tractor driving	18:24
16:46	FAC	Chris	EY & EX	N	Filling fork lift Batteries	18:03
16:50	Tour	Gabriele + 2	LVEA	N	LVEA tour	17:35
16:51	PSL	Jason	LVEA WBay	N	Checking out Test stands	18:51
17:00	EE	Betsy	CER	N	Working with Marc	17:38
17:16	FAC	Cindy,Karen,Kim	LVEA	N	Technical cleaning	18:32
18:00	CDS	Jonathan & Fil	EY	N	Taking PEM Y offline for a few moments	19:44
18:03	SEI	Jim	CER	N	Resetting signals to watchdogs via CER racks.	18:04
18:04	FAC	Chris	Wind - Y	N	Helping with the Wind fence at End Y	19:17
18:51	ASC	Elena	LVEA	N	Pluging in SR75 into rack	19:16
19:09	PEM	Adrian	EX	N	Checking sensors	19:35
19:20	Vac	Nitrogen	EX	N	Nitrogen delivery since 11:30	20:10
19:52	SEI	Jim	Ctrl RM	N	SEI Transition checks and testing	20:06
19:56	LVEA	LAZSER HAZARD	LVEA	YES	THE LVEA IS LASER HAZARD	07:56
20:05	SQZr	Nutsinee, Danelle	LVEA	SQZr Bay	Checking SQZ Table	22:35
21:28	VAC	Travis	EX	N	In mechanical room working on pump for past hour.	21:28
21:35	SQZr	Naoki	SQZ table 0	YES	Helping Nutsinee	22:11
21:50	ALS	Elena	Ctrl Rm	N	Open Loop Measurment of CARM	22:10
22:11	EE	Fil	LVEA SQZT0	YES	Running Cables to SQZr Rack.	22:22
22:42	Commish	Elena	CTRL RM	N	CARM & PCAL Meas. & PCAL -> Darm TF	22:57

New CAL_AWG_LINES Guardian now managed by ISC_LOCK

R. Short, J. Kissel

I've added into ISC_LOCK management of Jeff's new CAL_AWG_LINES guardian. In LOWNOISE_LENGTH_CONTROL, CAL_AWG_LINES will be requested to the LINES_ON state and will remain on. The lines will turn off in NLN_CAL_MEAS and will turn back on when moving back to NLN.

This change is committed to the svn.

Filter Cavity End Station (FCES) Access Control

Advanced Protection Services installed all of the hardware for the access system at the FCES.  The panel still still needs to be programed and a perminant power source installed by us.

They may be back tomorrow to program the panel.

Foton complains of Filter errors for PCALXY Comparison DEMOD filters

J. Kissel

As I'm trying to make changes to other filters in the h1calcs model -- whose foton file is 
    /opt/rtcds/userapps/release/cal/h1/filterfiles/
        H1CALCS.txt 
-- I keep seeing foton's "File Errors" red box appear next to the filter file listed in the GUI interface -- see attached.

The error message is referring to the low pass filters in the I and Q banks of the new(ish) PCAL XY comparison, e.g. 
Warning: filter 1 module CS_TDEP_PCAL_X_COMPARE_ERR_DEMOD_I had at least one pole with a small
  frequency. Poles with frequency less than 10^(-6) as fraction of
  sample frequency will suffer from low precision.
  Consider running the filter at a lower sample frequency.

I'm never sure if foton counts from 0 or 1 in its messaging, so I'm not sure if "Filter 1 module" is FM1 or FM2. So, I list all three filters that are in these banks below:

FM1 "LP300s"
    Command: butter("LowPass",4,0.0033)
    Alternate:     zpk([],[0.003048802457154851+i*0.001262977001249471;0.003048802457154851-i*0.001262977001249471;
        0.00126285532697085+i*0.003048788768825331;0.00126285532697085-i*0.003048788768825331],
        0.9999794439618445,"n")

FM2 "LP30m"
    Command: butter("LowPass",4,0.000555)
    Alternate: zpk([],[0.000512753140477464+i*0.0002128241445590586;0.000512753140477464-i*0.0002128241445590586;
        0.0002123893051084369+i*0.0005125483540322335;0.0002123893051084369-i*0.0005125483540322335],
        1.000081799778788,"n")

FM3 "LP60s"
    Command: butter("LowPass",4,0.017)
    Alternate: zpk([],[0.01570595205275844+i*0.006505637221348701;0.01570595205275844-i*0.006505637221348701;
        0.006505618350371988+i*0.01570594367905692;0.006505618350371988-i*0.01570594367905692],
        1.000000060520675,"n")

My guess is that it's the 30 minute low pass in FM2 "LP30m" (so counting starts at 0 for this error messaging system), since that would be the low pass filter lowest frequency poles. This filter has poles at 0.55 mHz, which means they're 0.55e-3 Hz / 16384 Hz = 3.39e-8, which is definitely lower than 10^-6.

It looks like the team has chosen to use FM1 "300s", browsing through the DEMOD settings as they stand; see e.g. the attached screenshot of PCAL_Y_COMPARE_DEMOD overview.

So -- we should just delete these filters that foton complains about.

But, rather than just deleting it, I aLOG it 
    (a) for the team to understand what the issue is, 
    (b) for them to conscious for these warnings next time, and 
    (c) to document it for the CDS team, since I wonder what impact this has*** and if we can instead push this message to the user as soon as the design string is written such that the user knows that this is against the rules.

*** Impacts?
    - What happens if I switch to *using* such a filter that has this error? Does it round the frequency up and still low pass?
    - What happens if I save the file after making an entirely independent filter?

After consulting with Tony and Erik, I've deleted all of the offending "LP30m" filters and moved on further editing the H1CALCS.txt file in other banks.

CDS Maintenance Summary: Tuesday 25th April 2023

WP11151 Guardian new CAL_AW_LINES and THERMALIZATION nodes

Jeff, Dave:

A new H1EPICS_GRD.ini file was generated with the new nodes added. DAQ+EDC restart was needed.

WP11138 latest CAL and PCAL code

Jeff, Erik, Dave:

We built the latest h1calcs, h1calex and h1caley models. It was found that the end station models were removing two slow channels from the DAQ frame, which would require an ECR. Jeff undid that particular change so these models are only adding channels to the DAQ.

model	num slow added	num slow removed	num fast added	num fast removed
h1calcs	1,220	0	0	0
h1calex	3	0	0	0
h1caley	3	0	0	0

WP11128 Reduce EY-MY fiber usage, add ADMIN-LAN access to MY

Jonathan, Erik, Fil:

EY to MY fiber usage was reduced by replacing fiber pairs to a single fibre using bi-directional SFPs. The network connections for H1FE and H1DAQ were upgraded to bi-dirs, while the timing was kept as a fiber pair. This freed up two fibers, one of which was used to send the H1ADMIN network to MY using ethernet-fiber media converters at each end.

We now have IPMI control of h1pemmy-mgmt. In a future TM the H1ADMIN lan connection to MX will also be done.

WP11121 Pulizzi PEM Coil power controls

Fil, Erik, Dave:

Fil installed three more Pulizzi ethernet AC-power controllers in the corner station: CER, SQZ and VERTEX. I reconfigued the sw-lvea-aux Cisco switch ports 2, 43, 46 to VLAN105 to connect to these units. I created three new EPICS IOCs on cdsioc0 to run these boxes. The CER unit as intermittent errors, and does not report outlet status except when a SCAN command is explicitly given.

Fil looked at the end station VEA units which had been installed some time ago. We found that they have incorrect IP addresses and are on the wrong VLAN (in H1VID, not H1SLOW). Also, exrpc does not respond to pings, even though its ethernet link lights are on.

WP11142 ISI BSC upgrades, long-range_dolphin end-station to seiproc

Jim, Dave:

We installed Jims latest ISI BSC model changes, including adding an End-Station to h1seiproc long-range Dolphin channel from each end station.

Models upgraded: h1isiitmx, h1isibs, h1isiitmy, h1isietmx, h1isietmy, h1isiham8, h1seiproc.

DAQ changes are summarized below

model	slow channels added	slow channels removed	fast channels added	fast channels removed
h1isiitmx	2,904	876	6	0
h1isibs	3,012	876	6	0
h1isiitmy	2,904	876	6	0
h1isietmx	2,905 (+1 RFM0 chan)	876	6	0
h1isietmy	2,905 (+1 RFM1 chan)	876	6	0
h1isiham8	32	0	0	0
h1seiproc	8	0	0	0

add/remove channel names are available in attached tar file.

A new cdsrfm long-range dolphin MEDM was generated to show the additional EX->CS and EY->CS IPC channels.

GDS Broadcaster Additional Channels

Jonathan:

9 new slow channels were added to GDS0 and GDS1 broadcast channel lists. The additional channels are listed below. I've "shown my work" in how to get this list using the new channel-lists archiving system, finding the diffs between the 2am and 2pm files.

david.barker@opslogin0: cd /ligo/www/www/exports/daq/chan_archive/2023/04/25
david.barker@opslogin0: diff h1_gds0_channels_2023_04_25_02_00.txt h1_gds0_channels_2023_04_25_14_00.txt|grep ">"
> H1:CAL-CS_LINE_1_OSC_CLKGAIN
> H1:CAL-CS_LINE_2_OSC_CLKGAIN
> H1:CAL-CS_LINE_3_OSC_CLKGAIN
> H1:SUS-ETMX_L1_CAL_LINE_CLKGAIN
> H1:SUS-ETMX_L2_CAL_LINE_CLKGAIN
> H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN
> H1:SUS-ETMY_L1_CAL_LINE_CLKGAIN
> H1:SUS-ETMY_L2_CAL_LINE_CLKGAIN
> H1:SUS-ETMY_L3_CAL_LINE_CLKGAIN
 

DAQ Restart

Dave, Jonathan:

The DAQ and EDC were restarted to support the model changes, the EDC changes and the GDS broadcast list changes.

GDS0 needed a second restart to resync its channel lists, but not GDS1.

FW0 spontaneously crashed after running for about 10 minutes. FW1 was not fully operational at this time and we may have lost a full frame.

Reboot/Restart Report

Tue25Apr2023
LOC TIME HOSTNAME     MODEL/REBOOT
09:22:42 h1oaf0       h1calcs     
09:23:22 h1iscex      h1calex     
09:23:52 h1iscey      h1caley     
09:32:05 h1seiex      h1isietmx   
09:32:34 h1seiey      h1isietmy   
09:33:14 h1seib1      h1isiitmy   
09:33:40 h1seib2      h1isibs     
09:34:09 h1seib3      h1isiitmx   
09:34:39 h1cdsh8      h1isiham8   
09:35:13 h1oaf0       h1seiproc   
09:42:01 h1daqdc0     [DAQ]
09:42:11 h1daqfw0     [DAQ]
09:42:12 h1daqnds0    [DAQ]
09:42:12 h1daqtw0     [DAQ]
09:42:19 h1daqgds0    [DAQ]
09:42:38 h1susauxb123 h1edc[DAQ]
09:43:15 h1daqgds0    [DAQ]  <<< 2nd GDS0 restart
09:45:11 h1daqdc1     [DAQ]
09:45:21 h1daqfw1     [DAQ]
09:45:21 h1daqtw1     [DAQ]
09:45:24 h1daqnds1    [DAQ]
09:45:31 h1daqgds1    [DAQ]
09:47:21 h1daqfw0     [DAQ]  <<< FW0 crash
12:28:54 h1pemmy      ***REBOOT***
12:29:51 h1pemmy      h1ioppemmy  
12:30:04 h1pemmy      h1pemmy     
 

BSC model update today

Much like the HAM-ISI update from last week, the BSC-ISI got their O4 update. The majority of this was for the blend fader infrastructure, meaning now all of the different installed blend filters are being continuously computed, and we just "fade" from the current blend to the next blend. This means that we no longer have to do the old dance of waiting for 2 different filter turn on transients, and blend filter changes will now take 7 seconds (kind of arbitrary right now, it can be changed by changing an guardian variable, which controls an epics value), instead of previous 1-2 minutes.

There hasn't been much testing of the 7 second blend changes yet, but while TJ and Tony were doing initial alignment we changed the ETMX St1 node to 3 or 4 different configurations and back and didn't see any change in the X arm ALS controls. If that holds up, the whole earthquake transition will now take 30seconds total, and be limited by the ramp time of the sensor correction mode change.

There are a couple new screens, more or less copied from the HAMs, updated from what I posted last week. The overview also has a different name temporarily, but looks the same while I work with Huyen to get her ready to update LLO models.

First attached screen shot is the new blend screen for Stage1. The blends can be changed manually here by clicking engage, or changing the "Next Blend" epics variable. I don't like having the engage buttons here, it's possible to turn on an empty blend path and trip the ISI, they might go away. The engaged blend filter is highlighted with a dark green box, a dark blue box will appear around a blend that is being ramped to. The little read and yellow boxes on either side of the blend name banks indicate the inputs and outputs for each bank, and there is a little green bar that lights up if the gain of that filter bank is =1. So a blend that uses the CPS, T240 and L4C for stage 1 there should be 6 yellow boxes and 3 little green bars. I also want to add some status lights to indicate that filter modules are actually engaged.

The second screen is the filter banks for one blend. No a lot new here, but you can see that this blend uses all 3 sensors, inputs, filter modules, gain and outputs are all engaged, so this blend is ready to be used. All of the unpopulated blends have been left off.

There are also 2 sensor versions of these screen for the Stage 2. We don't normally change the Stage 2 blends though. 

SR 785 plugged in for CARM measurements

Daniel, Naoki and I pluggeg in the SR785 at the PSL racks to run measurements of CARM during lock and determine if we can increase the CARM gain (see alog 68967). This means we have unplugged the cable that allows us to run frequency noise injections, We will try to get carm measurements as we thermalize, and then plug in the frequency injection cable so we can return to those measurements later in the lock.

CO2X steps

Elenna, Dan

We wanted to test the original asymmetric CO2 powers we had back at 60W to see if we could get back to lower frequency noise.

Elenna had been doing some steps and measurements and I took over when she left. Elenna had set the nominal requested CO2 in lscparams to 2.7W (which is actually 4W of actual annular CO2) and I was going to come in a few hours later and take some measurements. However the TCS guardians didn't seem to take us there, probably not reloaded. I changed CO2X by hand up to 4W with the aim of coming back a few hours later to do the measurement.

About 1.5 hours later there was a lockloss, PRG became very noisy. Looking at POP_A_RF9_{I,Q}_INMON they both doubled in counts. I didn't measure PRCL gain but maybe this CO2 change in combination with the new fixed thermalisation gain scaler increased it too much.

I didn't get a chance to do the measurements we wanted (contast, spring, frequency coupling) so I've set the nominal CO2X to 2.7W requested and reloaded the guardians. I've set the THERMALISATION guardian to IDLE to see if this will hold a longer lock. I'll come back in a few hours and see how it's doing and try to take the measurements if it survives.

Looking at the lines before lockloss, it seems going back to an asymmetric CO2 setting 4W and 1.7W on X and Y reduces the 200 and 4500 Hz frequency noise coupling a little but increases the 25.2Hz coupling. Arm power is also slightly down by a few kW. This also looks like it will eventually reduce both the optical gain and cavity pole judging by their trends.