Vertex pump down is not trending down (just one turbo is pumping since X&Y beam manifolds are vented. Attached is plot over 10 days.

So have had a continuous exponential spectra of the WHAM6 CPS signals running on the desktop watching for the grounding or whatever-it-is noise on the Corner3 sensors.  The attached snap shows this new noise.  The reference and the thicker traces are the corner3 signals, which look fine.  The other four fine traces are the corners 1 & 2 signals.  This isn't even the worse of it but it looks very similar to the corner3 noise seen previously while messing with the cabling there the last few days.  Note that the corner3 channels are in one satellite crate and corners 1 & 3 are in the other.  I'm suspicious activity around the chamber has disturbed the ground of that crate and made it iffy.

I went back to about 1pm yesterday (1900 utc 28 March) and watched the spectra until it went noisy like this.  First time I spotted was 0130 pdt this morning. Will check with EE to see if they know of anything.

I tried to open the GV2 gate annulus valve to pump out that volume while valve is hard closed (so we don't have to hook up a pump cart later), but it won't recover with annulus valve open. It's acting like there is an o-ring leak. Valve is closed an AIP current is a little higher now.

TITLE: 03/29 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    Wind: 7mph Gusts, 5mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.13 μm/s
QUICK SUMMARY:

• I'm covering morning for Jim W. 
• Peter has been out on the floor for PSL work (FSS guardian set to "NONE" so he doesn't have to fight Guardian trying to lock the IMC).
• Jeff K working HAM6 suspensions (heard Hugh say his Corner 1 & 2 CPS are now displaying oddities...it had been the Corner 3's last week.)
• Amber staging for 9am Virtual Visit w/ South Carolina high school.
• CDS Overview:  MY is almost all INVALID/WHITE, edcu RED, TIMING RED

Sheila, Georgia, Dan, Alexei, Nutsinee, TVo

We found it odd yesterday that we needed to move PRM ~100urad in pitch and ~50urad in yaw to get PRX fringes from Feb 8th.  Sheila walked back PR2 a bit to get PRM back to where we were and this centered the beam onto the refl camera, which is good.  At this point we should have seen the beam on AS_C in HAM6 because everything seemed to be the same, however, Georgia noticed the witness sensors on SR3 M3 were much different than Feb 8 even though the alignment sliders were the same, this was due to an alignment dither offset on M1 that was still on.  Once this was turned off, we got a beam on AS_C.

Then we proceeded to HAM6 and were able to find a leakage beam on the squeezer side hitting ZM2, however, it seemed low in pitch coming out of HAM5 and not possible to coalign with the current Squeezer optics.  This was not correctable with ZM2 (railed) alone so we're in a tough position here to see *maybe* correcting the squeezer beam without jeopardizing the IFO to OMC alignment.  Sheila felt it was worth trying to nail down how much we can actually change the SR optics before clipping on the OFI, so we set out to try to find an edge of the OFI aperture by slowly moving SR3 in pitch while compensating with SR2: 

This means that SR2 has the full range of pitch before we start falling off AS_C while compensating with SR3. 

Now, If we then center on nominal with SR3, then the SR2 range is

 Based off our data, tomorrow, we'll go into chamber and see if it's possible to correct the ifo to squeezer alignment without clipping on the OFI.

J. Kissel

More details to come after I'm able to measure H1SUSOPO again, but as a result of LLO converting to the standard SUS basis (LLO aLOG 38272). and remeasuring their OPOS TFs (LLO aLOG 38308), I've made a comparison between LLO** and LHO's data again and thinks look far more consistent. I've also updated the dynamical model after the (yet to be documented) work Alvaro and I have done during the LVC meeting. In summary, we added several more elements to the stiffness matrix that explain the cross-coupling visible in a lot of the degrees of freedom.

(**Note, the confusion Stuart mentions regarding scale factors was that he didn't remove the special calibration conversion that accounted for the modified HAM-A driver, and the bad units in PRY lever arm basis change when processing the data with 
/ligo/svncommon/SusSVN/sus/trunk/OPOS/Common/MatlabTools/plotOPOS_dtttfs_M1.m. I've fixed and commited the script removing the special LLO case, and now both observatories are calibrated with the same scale factor. I've also re-processed and committed the LLO data.)

Check out the new results attached. The notable differences are in the DOFs that I've not been able to close damping loops: T and Y. T appears to somehow have a sign flip from LLO and expected. Remember we're now using exactly the same basis change matrices. The only difference in the digital system is the magnet compensation signs in the COILOUTF banks, which we believe are correctly compensating the as-built configurations. But, this is to be investigated further. For Y, it looks like the phase at high frequency is asymptoting to a different value. No idea what's going on there.

However, because the results from other DOFs are so similar, I copied over the damping loop design from LLO aLOG 38309, and tried things out. L, V, R, and P are just as successful as LLO. Nice!

Now to still figure out what's going on with T and Y...

I've started the investigation by taking (loops open) T and Y TFs with the L V P and R loops closed, hoping that it might clean things up a bit. The Qs look quite different from the 2018-03-16 measurements, but I ran out of time to investigate further. More tomorrow...

Dan:

late entry from yesterday's maintenance. Dan upgraded the firmware on the two controller cards for E18-1 (written to by h1fw1). Unlike the E18-0 upgrade, this time everything worked as advertised and both CDS writer and the LDAS readers continues to function during the staggered upgrade.

I have restarted the Beckhoff SDF system for c1plc[2,3] following recent changes to these systems. New channels were accepted and monitored. Note that the DAQ has not been restarted for the new channels.

Took several steps backwards today, although it did start off promising.  I re-aligned the output of the front end
laser and got ~90 W output power with a nicer beam.  Spent time optimising the alignment into the pre-modecleaner.
The mode matching was still pretty poor.

    Then the bad news, the diode currents on the PSL Computer were not those on the datasheet.  After setting the
diode currents and temperatures to the datasheet values, the output of the amplifier was ~84 W.  However the
output beam did not appear as good.  The mode matching into the pre-modecleaner was less than 50%.  The beam
propagation was re-measured.



Ed / Peter

All fours fibers have been welded and annealed.  We ended up with a PUM-to-ETM differential pitch of ~3mRad.  We'll attempt to reduce this with a further round of annealing tomorrow.  

M&M finished wrapping the entire bake enclosure with Al foil and taping the seams. Bottom air temp has reached a new high of 95C (20C less than top air temp). This morning they swapped CP4's GN2 regen heater with CP3's - if we can resurrect that system we may have a chance to bring the entire enclosure up to uniform temperature. I raised the set point of bake enclosure to 130C (from 115C). 

Turbo pump back end measures 55C with 35C cooling water.

I have cleaned up the ISC guardian code (ISC_LOCK, ISC_DRMI, OMC_LOCK, ALIGN_IFO, ALS_XARM, ALS_YARM, ALS_COMM, ALS_DIFF, ISC_GEN_STATES, ISC_library and lscparams). 

My goals were to remove depricated code (it's been growing since ~2014, and many functions clearly haven't been used since about then), merge nearly-identical functions (we don't need 15 different functions to determine if a cavity is locked), ensure that functions from sub-files (eg. ISC_library) are clearly called out (by removing all "from _______ import *", which Sheila had already done a lot of work toward), and make the code generally easier to read. 

I was going to do a bit more syntax checking, but since we have IR light into the vacuum earlier than had been anticipated and several people have been working on locking the IMC, I deployed the new code this afternoon.  All guardians were checked in before I did anything, and then I copied over my files from a side folder into the main userapps guardian folders. 

So far, the IMC guardian is able to lock the IMC (after a few boost changes from Sheila, TVo, DanielBrown and Alexei...they'll write a separate alog).  The ALIGN_IFO guardian is able to change the SUS configurations, although it hasn't been tested farther than that, since PeterK is working on the PSL. 

This cleanup has removed ~4,000 lines of unnecessary guardian user code.  I think it'll help make things easier as we go forward with a new commissioning phase.

Stephen A., Chandra R.

This morning Stephen helped/instructed me on assembly of the first LHO chevron baffle. Baffle SN 001 was installed in nipple housing SN 001. We iterated several times on bending the louvers to securely set them in place and ended with an N2 "top gun" blow off before bolting the two baffle halves together and fastening to the nipple housing. The eight bolts that secure the baffle to the housing still need to be torqued. We placed the two 16.5" Cu gaskets on knife edges and wrapped with Al foil for temporary storage in VPW until we can install on IP11 at EY later this week. Stephen will post pictures later.

Refer to LLO aLOG for more detail on steps of installation. https://alog.ligo-la.caltech.edu/aLOG/index.php?callRep=38375

Evan G., Jeff K.

We moved a compensation zero out of the ESD output filter bank for each quadrant that had generated an uncompensated 5.2 kHz pole in the DARM loop (see LHO aLOG 33927). This resulted in an effective delay of ~20 usec. There is adequate roll off already from a low pass filter in the ESD driver, so moving this out of the DARM loop will be minimally impactful. Instead, the compensation zero will be a compensation pole in the CAL-CS path and doesn't result in an additional 5.2 kHz zero being added to the Foton filter.

We leave the filter in place and turned on, but the design string is now simply zpk([], [], 1, "n")

The new filter is installed in CS_DARM_ANALOG_ETMY_L3, FM5: zpk([], [3245.33075], 1, "n") (this comes from the mean of the summing node poles, see 27619)

Removing this from the DARM path means that we need to include these as an uncompensated poles in the DARM model and reduce the unknown actuation delay by 20 usec. This will bring the unknown actuation delay down to ~40 usec.

SudarshanK, RickS

To study the downconversion of Pcal excitation we used a DB9 breakout at the back of the Pcal chassis and routed the excitation channel through H1:CAL-PCALX_WS_PD for now. This will be returned to its original configuration once the investigation is complete.

A new line at 86 Hz has appeared in the DARM channels after the vent.

After running BruCo, I've seen that that frequency was coherent with some PCAL channels: https://ldas-jobs.ligo-wa.caltech.edu/~pep.covas/bruco/1181560000/

I attach two different plots showing the differences between before/after the vent for the DARM channels and for  three different PCALX channels . The "before" data is from 07/05/2017 09:00 UTC, and the "after" data is from 15/06/2017 09:00 UTC.

Evan G., Joe B., Jeff K.

Joe B. pointed out that there is an uncompensated ~5.215 kHz pole in the ESD output filter bank because our filter design zpk([3229.1150], [], -1, "n") results in Foton providing a pole to compensate the zero at high frequency. So the actual installed filter is zpk([3229.115],[5215.189175235227],0.9999999999999999,"n").

The ~3.229 kHz zero compensates for a pole in the analog electronics. Unfortunately, the ~5.215 kHz pole is uncompensated. Attached is a Matlab figure showing the effects from 1 Hz to 10 kHz. The band of most interest to us is <200 Hz since above 200 Hz the actuation is sufficiently rolled off that it does not contribute to the response of the interferometer.

Below 200 Hz, there is very small residual magnitude error. Phase, however, is incorrect by ~1.1 degrees at 100 Hz. The equivalent delay is ~30 usec. This could explain part of the unknown actuation delay, which--for LHO--is 61 usec.