I turned off the LSC RFPD interface chassis (D1102060) for IMC length diode using the switch at the back panel, disconnected the length diode cable, connected a DB9 break out board at the front panel without connecting anything else (but the back panel was still connected to the ADC), switched on/off the chassis (again using the switch at the back) and measured how the +-18V supply for the diodes behave.

To my disappointment, every time I did this test, for about 1ms to 1.5ms there was a window where the negative power was -18V but the positive was basically non-existent (below 5V). It seems as if the negative voltage goes live immediately. See the first picture, the scale is 5V and 1ms/div.

This is the unit that we've been using since the start of IMC installation.

I did the same test for WFS interface (D1101865), and it was only marginally better than the length interface (second picture, same scale as the first one).

This is the unit that replaced the broken unit, but since this one was put into service WFS was connected/disconnected hot.

Switching off was much better for both of these units, basically both of the rails came down slowly at the same time (the last picture, note the slower 4ms/div scale).

During these tests, the only things powered by the chassis were the DC interface cards in the chassis and of course the power protection board itself. 

I don't know if the units are broken, we'll test them further though the change Rich Abbott made to diode boxes will hopefully make this behavior harmless.

These are taken with the HEPAs in the anteroom and laser room off, with the make up air fan turned to 30%. AC units are off as well.

We flew the ergo arm to the south bay for Betsy. Disassembled the E module and moved to the north side of ham 4, started disassembly of the work platform that was around bsc 1 for install at bsc 2. Need to remove (with approval granted from Hugh) the horizontal hepi accuators on the south side of bsc 3 to allow installation of the work platform section between bsc 2 & 3. These will be unbolted and rolled down will still being secured with the current hepi bolts. (This same thing was done on bsc 2 for some clearance issues.) Also assisted Hugh/Eric with hepi install. 

A month ago when we restarted the laser we noticed the frontend power monitor was reading only 6W. This was tracked down to a gain setting that was modified and not entered into the safe.snap file, so when we rebooted the computer frontends for the 2.6 code upgrade the incorrect gain value was entered. I've changed back to the correct setting and updated the safe.snap file, and we are now reading 31W.

[Kiwamu, Paul]

Just before the vent today we measured new noise spectra for the frequency and length feedback paths.

The 'whitening filters' in the IMC_X path (see Giacomo's LHO alog entry 5311) were not engaged. They were put in the model at LLO in an attempt to combat the digital noise in IMC_X above ~50Hz (see Anamaria's LLO alog entry 5821) but need to be taken account of when plotting the data.

We corrected the IMC_F frequency to length calibration to give the equivalent motion of MC2, instead of the equivalent round trip length change. This has now been edited in the filter bank for the IMC_X path; the FtoL filter has been changed to a DC gain of 5.8465e-14 (dX=df x lambda/(2FSR)). 

Relevant paramters during the measurement:

Common mode gain 20dB

Common mode filters: Compensation and 1st Boost

Fast gain -2dB

Slow path bypassed offsets and filters

MC2 M3 lock gain -1000, engaged filters: 150:4 and CLP100

MC2 M2 lock gain 0.06, engaged filters: 0.01:0.1, 0.03:1, Stab8:2, 300:1 and ELF80

Damping filter: resg and Ellip50 (the filter in IMC_X path was switched to the wresg filter to reflect this change)

The FMCS IOC needed a restart, presumably due to the failure of /ligo on Sat morning.

I shut down the HEPAs in the ante room and in the laser room, shut down the 2 ACs (though I could not see their effect, so they could be left on) and set the makeup air at 30%. Overnight the temperature on the table increased by a couple to several degrees. The IMC was still locking, though very poorly. Paul tuned up the alignment  of the IMC and the REFL_DC level returned to better than it had before the change (about 50 counts). The figure shows a comparison of IMC_F  and the accelerometer at the top of the periscope in the two modes.

[Adam M., Paul]

The IMC Guardian script can now be used to maintain the required state of the IMC.

This was achieved using the essentially the same scripts as in Livingston, but with some temporary local modifcations to account for the difference in the install state at the two sites (for example, ASC loops are not controlled by the guardian at LHO yet). The current state should be regarded as temporary, and if it's not required one can still of course lock the modecleaner in the way we have been doing up to now. (However, it may be useful for commissioning as we have already experienced the frustration of wondering where our signals have gone before realising the IMC was unlocked). The scripts can be found in /opt/rtcds/userapps/release/ioo/h1/scripts/imc

Some details:

Since we don't yet have a beam on MC2 trans QPD, we are using the MC_REFL_DC_OUTMON channel to verify the locked/unlocked state of the IMC. Currently, the threshold is set to 70 counts, because when aligned reasonably well the REFL_DC count is around 45. If the alignment state changes significantly, for example due to the change in PSL temperature or the upcoming vent, I recommend to lock manually first and realign to bring the locked REFL_DC counts back below 50 or so. If necessary, as an alternative the threshold can be changed in the imc_params.txt file. Of course if the laser power is changed significantly, the threshold should be adjusted.

The starting common mode gain is 0dB, which ramps up to 16dB upon lock acquisition. If there are problems to do with locking to bad modes, it may be necessary to reduce the starting gain. This can be edited on line 44 of the mcdown script. The locked gain can be edited on line 20 of the mcup script. Currently it takes two steps of 8dB from 0dB. If the starting gain is edited, this should also therefore be edited, or it could be rewritten to just make one step to the required gain.

Commencing Matlab TFs on PRM.

At midnight this morning the main control room NFS server suffered a scsi error and the /ligo file system went into read-only mode. This means you cannot modify or create files on the /ligo file system, which includes user home directories and the cds wiki pages. This is a known problem with this type of integrated computer and file server, and the only resource is to power cycle the machine, which cannot be performed remotely.

The immediate problem this causes is the hourly autoburt backups are not recording. I can reroute those to the h1boot file system temporarily until I can get to the site to reboot the server.

The problem occured at 4 mins past midnight, Sat 9 Feb.

Vincent was keen to test IPC between his new BSC2 ISI model and the BS SUS model, but the latter hadn't been created so I waded in and created it. I adapted the existing BSTST model and then compared it at the end with the L1 version - they were gratifyingly similar, except for the IPC stuff, which only exists in the H1 version so far. I got it running and updated the sitemap to point to it. I installed projection matrices, filters, and default gains and offsets, and created a safe.snap. 

Along the way I discovered that the SUS_CUST_BSFM_ADC_MONITOR.adl screen had been hardwired to L1, so I fixed that.

I noticed a few anomalies that need further debugging:

* The output fields on the SUS_CUST_BSFM_OVERVIEW.adl screen between the USER DACKILL and the IOP DACKILL (H1:FEC-31_DAC_OUTPUT_2_4 etc) are white. There's also whiteness in the H1CDSOVERVIEW.adl screen.

* The M2LR OSEM is showing a non-zero, fluctuating value (≈±2500 counts) despite having no OSEM connected.

* There's red in the Dolphin IPC bit on the H1CDSOVERVIEW.adl screen.

Apart from all of that(!), very possibly when the physical suspension is connected up it will damp.

[Paul, Kiwamu, Keita, Dick, Rodica]

Today we measured the mode cleaner cavity length by modulating the EOM at 45 MHz and doing a sweep around this frequency. We measured the spectrum using an HP 43968 Spectrum Analyzer. The maximum source power was 19 dBm from a Mini-Circuits RF amplifyier with gain 19 dB. The bandwidth for the highest zoom on the peak was 300 Hz.

The frequency of the dip in the middle should correspond to the 5th FSR away from the carrier resonance. The calculated length is 9.09918 MHz +/-  100 Hz, only 291 Hz away from the design value of 9,099,471 Hz. This is equivalent to 1.05 mm difference from the design length.

We tried a full sweep over one FSR but we realized that the modulation away from resonance would be too small to give us any useful results, so we focused on the 45 MHz resonant peak. Attached is a plot of the data over a full FSR, that also shows a peak corresponding to the non-resonant polarization, and few other peaks possibly due to higer order modes.

Similar measurements at LLO are described in alog 4702.

Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot requested from 5 PM Feb. 7 to 5 PM Feb 8. Also attached are plots of the modes to show when they were running/acquiring data.

Data was taken from h1nds1.
1440.0 minutes of trend displayed

[Hugo, Rodica, Paul]

Today we were finding it impossible to lock the IMC again. When we looked at the SUS-MC2-M3_LOCK_L_OUT_DQ channel, the signal appeared to be always railing (on both sides). We have seen this before at times, due to very noisy seismic states (see e.g. entry 5145). However, even with what we expected to be a quiet siesmic and ISI state, we still found this signal railing at ±40,000 counts, making it impossible to lock using the length path. In the end, Hugo realised that when the model was recently updated, the limit on the MC2 M3 lock filters had reverted to its initial old value of 40k counts. Apparently at some point this was increased to 400k counts, but not saved in a safe.snap file. Hugo increased the limit back to this value and the mode cleaner began locking again.

We should update the safe.snap file so that any further reboots revert to the 400k counts limit.