(Corey, Keita)

Telescope Assembly

Now that we had a new set of machined Side Clamps (machined by Tyler and Class-A-ed over the weekend), we installed these parts on the Telescope Assy, and then installed the Primary Mirror.

Telescope Rough Alignment

With all Mirrors installed on the Telescope (the F1 [6"] Mirror is an IAS Alignment Mirror we are "iffy" on, we're waiting for some recently-ordered 6" Mirrors to see if they have a better coating), we started rough alignment of the Telescope.  This was done with autocollimators at the input and output of the Telescope and per Keita's alignment document (T1100603).  Next we were going to move to a more fine alignment with the IR Laser, but we discovered the 8" Alignment Mirror we were going to use had a frosted back surface, and this would not be useable for shooting the Autocollimator through it.  So, this is where we stopped for the afternoon.  Today (Wed), Keita found a gold-coated mirror to use for this alignment and will proceed from here.

SUS Cables Run

I grabbed two cables which run from the "roof feedthru" to the Test Stand at EY, and installed them at EX.

[Stefan, Alexa, Kiwamu]

 The IMC wasn't locking today. We found that this was because the demodulation phase changed.

The guilty RF cable:

 We took a look at the demod phase of not only the length diode but also both WFSs and found that the amount of rotation they got were identical. So we started suspecting something common, for example the cable length of the modulator. Eventually Stefan found that wiggling the N-connector of an RF cable rotates the demod phases by tens of degrees. This is the 4th output of the 24 MHz RF amplifier in the R1 rack at the field. In fact this is the one supplying the EOM with the 24 MHz modulation. This rotation happens more or less in a random fashion but the amount of rotation was relatively repeatable --- we could reproduce a couple of certain conditions with a certain demod phase by wiggling around the connector.

 We tightened the N-connector and it seems that the problem was solved. Additionally we briefly tested it by wiggling the connector again but we didn't see any more rotations.

Demod phase need to be adjusted again :

 Then an issue we got after fixing the RF cable is that all the demod phases were consistently off. Probably this may indicate that the we had been in a particular, but a different condition for a long time due to the loose connection at the RF connector. In any case we need to adjust all the demod phases associated with the IMC i.e. length diode and two WFSs.

Demod phase of the Length diode readjusted:

 We readjusted the demod phase of the length diode as a first recovery step. The new optimum setting, where the I signal is maximized, is

• 28.625 nsec = 1/8 + 1/2 + 4 + 8 + 16 nsec
• (This is set by the hardware switches at the front panel of the delay line phase adjuster)

Comparing it with the past setting of 14.4375 nsec (see alog 6282) we found that the amount of rotation we newly introduced is 123 degrees with the modulation frequency at 24078360 Hz. This resulted in a sign flip in the error signal. So we had to flip the control sign to properly lock the IMC.

Next step : readjustment of WFS demod phases :

 Apparently we have to do the same correction for all segments of the WFSs. Unlike the length diode the demod phases can be adjusted in the digital land. Since we now know that we need to rotate everyone by 123 degrees from whatever previous phase, correcting the demod phases should be relatively straightforward --- we simply punch in some numbers in the digital system. Of course we have to pay attention to the sign flip.

Just before the laser went dark, we discovered that the demod phases of both IMC length and WFS diodes had moved significantly, with all the signal ending up in the Q phase.

We found a bad connector on spigot 4 of the RF distribution panel, going to the EOM. Wriggling te cable resulted in ~30deg of phase shift.

The connector should be replaced tomorrow.

We reinstalled the primary mirror with re-machined clamp parts, and the new clamp parts worked well.

We're still waiting for the procedure to set the spring tension of the clamps right, but in the mean time we decided that 1/4 revolutions of the clamp screws after finger-tight felt about right.

Anyway, from there on we quickly set two auto collimators up, and roughly aligned the path correctly as per the procedure. 

Next we'll tune the telescope using a laser beam and mode master. That'll be about a day of setup and three days of going back and forth.

The PSL got down at around 5:50 PM in PDT due to errors in the diode and crystal chillers according to the status medm screen.

Gerardo Moreno, Scott Shankle and Mike Vargas

Attached the Face Plates. One screw on the Right Face Plate could not be attached (pic 2); hole does not line up. One screw on the Left Face Plate could not be attached (pic 1); holes do not line up.
Attached the Balance Weights and the Tongue Depressors released very nicely (pics 3 and 4).
The baffle was balanced using a Plumb Bob (pic 5).

The height needs to be checked. The Assembly Fixture and Crane will be needed if height needs to be changed.

Observations - the Face Plates are very flimsy and rattle a lot. They are tied to the baffle on the outer circumference only.

SLC crew - EX working on Cryopump Baffle

CDS crew - Rebuild of data concentrator - did not work, re-stored to previous working version

Filaberto- Running PEM Cables

Pcal - Working in the H2 PSL Enclosure

Robert - Disconnect dust monitors at ISCT1

Prax Air - Refill CP2

Hugo and Jim W - Taking overnight transfer functions starting at 9 PM PT

Jonathan Hanks - Reconfigure router on GC

Filaberto and crew - Run HAM 5/6 cabling to feed through(s)

Apollo - Install feed through(s) on HAM5/6 and install 16" conflat with new gasket on HAM6

Launched a delayed overnight measurement on HAM3-ISI, with the help of JimW. 

Measurement should start around 9.50pm PCT. Horizontal isolation loops will be turned OFF then. Only vertical isolation loops will be left ON for the duration of the measurement.

Measurement will be finished by 8.30am PCT. Isolation loops will be turned back ON by then.

Installed new framecpp-1.19.24-p1 package, obtained from Livingston.

Built new daqd from advLigoRTS/trunk, installed on x1dc0, x1nds1, x1fw1.

Restarted daqd on the three systems, ran tests to verify that the dataValid flag works properly.  dataValid is definitely non-zero for bad channels, and 0 for good channels.  Note that the value of dataValid is the same as the DAQ status, which can be various values depending on the cause of the bad data.

Plot shows 80 days of pumping at the vertex.

I found and fixed two issues:

• I had applied $(OPTIC) -> $(OPTIC)$(num) the other day when I shouldn't have - this screen was already expecting OPTIC = IM (not OPTIC = IM1 or the like).
• There were some left over references to signals with R0 in the name (apparently left over from adapting SUS_CUST_QUAD_DACKILL.adl).

I committed the fix.

WP4045, DAQ partial upgrade to fix incorrect dataValid flag in frames.

We tried the upgrade this morning during maintenance and could not get it to work. We tried the simple one-line change made at LLO in Jan on both the frame writer and the data concentrator, but this caused half of the front ends to be marked as bad. It was only resolved by backing out the change. We then built the DC against branch2.6, which did not fix the problem. Finally we ran LLO's DC binary, with the same results. At this point we ran out of time and reverted the original binaries back.

Today we went into the PSL because the reference cavity was no longer locking. The main problem was that the autolocker range was not large enough to find the resonance. In addition to fixing this, we tweaked the alignment into the reference cavity and into the PMC. We also tweaked the ISS, and Rick wanted to note that it was surprising that the REFSIGNAL went from about -2.0 to about -2.5. Additionally we experimented and found that the makeup air needs to be shut off before it can be reduced below 100%. And finally, we put the cover on the PMC to reduce index of refraction noise. The PSL is now running with the internal fans off, the AC off, and the makeup air at 20%.

Rick, Sheila, Robert

(Robert, Stefan)

Good news: The Rev cavity now locks quickly.
Bad news: the MC reliably kicks it out of lock again when its gain is increased.

Today's Team - Gerardo Moreno, Mike Vargas and Apollo's Randy and Scott

We bent the Springs (pics 1 and 2) and attached them to the Class B Top ring.
Lifted the baffle and attached it to the Springs at the Flexure Rods (pic 3).
It was very difficult to attach the baffle to the blades because of the tight spacing (pic 4).

Homework for the night - dream about how we can accomplish this task in situ.
The spacing between the Tube Wall and the Baffle is 1.25"; the Rings are 0.5" thick.
There is very little room (0.25") to insert and attach two 1/4-20 nuts with washers.

By adding the new PEM channels to H1PEML0 we exceeded the DAQ hard coded limit of 250,000 channels. This limit applies to the sum of Fast Channels, Slow Channels and Test Points.

For a short while this afternoon I removed h1peml0 from the DAQ to allow it to run. I then wrote a script to calculate how many channels we have in the DAQ sans-h1peml0, it is 249,461. Initially I  hand edited H1PEML0.ini, put back in the H1:PEM-CS_ACC channels (there are 37 of these) and added h1peml0 back into the DAQ. The DAQ channel total was now precariously close to the limit at 249,990 (only 10 spare).

We then removed the SUS BS test stand front end from the DAQ and put PEM back in its entirety. The channel total is now 246,254 giving a 4k buffer.

We will work tomorrow on extending this limit as part of the DAQ upgrade to fix the channel status word in the frame.