In alog 42711, I found that ETMX ISI St2 is moving a lot more than the other BSCs. I suspect this is probably some mechanical rubbing somewhere, so I've been looking for ways to mitigate the loss of performance. Turning on the ST2 RX/RY isolation (high gain, dc coupled, with 750mhz cps/gs13 blends) loops helps a lot. We typically leave these off because they inject GS13 noise into the X/Y dofs, but for this chamber it seems to be an improvement over a large frequency band in both the rotational and translation dofs. Attached image is for RY & X, but RX & Y is simlar. The dashed lines are with the loops rotational loops off, solid lines are with them on, green is the X gs13 signal, brown is the RY. Most interesting is the big lump around 50mhz that goes away when the loops are engaged, I guess the "rubbing" is causing extra RY, X sees that *g/w^2 and closing the cps loop suppresses that extra motion? The motion, especially ~10-30hz is still than the other chambers, but turning these loops on makes the motion below ~5hz down to 30mhz mostly better.

I noted that the foreline pressures were all < 5 x 10-2 torr.  This just confirms that all four Vacuum Sentry valves are working -> they isolate the scroll pumps at the loss of power to the scroll pumps. 

   Posted are the weekly CPS noise spectra plots. 

Continuing on from yesterday's work.

    SCRN0001.jpg and SCRN0002.jpg show the power noise from the 70W amplifier, with the ISS
on and off.  In both cases there are four relatively narrow peaks at ~19 kHz, ~38 kHz, ~54 kHz
and ~73 kHz.  They are present whether or not the ISS is on or off, and when both the ISS and
the AOM driver is off.  In the case that the AOM driver is off, there are a sharp accompanying
higher frequency peaks.

    SCRN0003.jpg shows the output of the front end.  SCRN0004.jpg shows the output of the front
end with the AOM driver on and off.  There is a small difference between the two but oddly
enough with the AOM driver off, there is a broad peak between ~40 kHz and 50 kHz.  The broad
peak around 20-25 kHz is from the NPRO, as shown in SCRN0005.jpg.  In SCRN0005.jpg features
are present irrespective of the condition of the noise eater.

    The power noise from the 70W amplifier seems to be quite sensitive to vibrations.  For
example dropping a screw on the table surface, or tapping the the table surface results in a
number of peaks appear in the spectrum.  It just might be possible that the cooling lines
are vibrating causes the four narrow peaks.  At present the cooling lines are merely drawn
over a metal plate which is fastened at one end.  The more rigid mount for holding the cooling
lines might be more appropriate.  However the peaks do not appear to diminish when the cover
of the amplifier is pushed down upon.  Failing that it would be worth looking at the output
spectrum of the pump diodes.

After reaching LOCKING_ALS, Sheila and I tried to lock both arms to red in the CHECK_IR state.  We lost lock every time we tried to reach CHECK_IR, and were unable to achieve ALS_DIFF locking robustly.

We tracked our issues down to a regular ~1 Hz in-loop glitch in ALS_DIFF, which with higher gain settings caused saturations in our L1 and L3 ETMY stages.

Sheila turned off the length control to ETMY and high passed the DIFF PLL control signal.  The glitches were still there. 

We also noticed that the glitches were appearing in COMM as well.

This indicates something is probably an issue with the sensing path, could be the amplifier removal that was done today to remove RF saturation on the ALS DIFF PD.  We may need to handle COMM RF saturations by removing its amplifier as well.

Sheila, Craig, Hang

After Georgia and Sheila locked ALS COMM this morning, we proceeded to lock ALS DIFF.
Our first attempts to lock failed as the ETMY ESD linearization was off.  When we switch it on ALS DIFF locking worked.  We were also able to get the ISC_LOCK guardian to reach LOCKING_ALS.  

The plot shows the spectrum of H1:ALS-C_DIFF_PLL_CTRL_OUT_DQ, which is already calibrated into micrometers just like ALS COMM.

ALS DIFF RMS = 6.8 Hz = 48 pm

Attached are the old and new coefficients used to calculate the percentage of diffracted power.  Based on 65.6 W incident
on the pre-modecleaner.

    The old polynomial was 107.476x**2-83.072x+16.063.  The new polynomial is 129.252x**2-100.701x+19.7562.

    The change was accepted into SDF.

[Sheila Craig Georgia]

This morning Sheila aligned the ALS common beat note onto the PD on ISC-T1, the beat note strength is 0 dBm, slightly lower than it used to be (+6dBm), investigation ongoing.

The PLL, which locks the ALS_COMM VCO to this beat note locked without a problem.

We had to fix a problem with the ALS_COMM guardian to finish locking ALS_COMM, where the IMC length control is transitioned from the usual IMC reflection PDH (IMC-MCL) to the common mode VCO (LSC-MCL). The guardian edges needed to be updated: in 2016 there was no PREP_FOR_HANDOFF state, and commands in this state were included in the HANDOFF_PART_1 state. The next svn commit was with Jenne's guardian revamp, somewhere between these two commits the PREP_FOR_HANDOFF state was created but the edges were not defined. We have now fixed this, and locked ALS_COMM with the X-arm. The UGF of this loop was roughly 700 Hz (note:we rerouted Craig's IMC OLTF SR785 excitation from the MC servo board to the IFO-common mode servo board (LSC-REFL_SERVO) to take this transfer function).

Meanwhile Team Hartmann are working on the ITMX SLED alignment. We've tweaked up the y-arm alignment, beamsplitter alignment (locking MICH), and SR3 alignment (locking SRY) to hopefully aid their quest.

These will be de-energized at the next opportunity. 

I went down to X end and repeated the ALS measurements done for Y end.

ALS X green visibility is 30%.  Free swinging PDH peaks give ~750 mVpp demodulated reflection response.  
With a cavity pole of 275 Hz, we roughly calibrated the PDH error signal with 2*cavityPole/PDHvolts = 733 Hz/V.

Plot 1
PLL UGF is 27 kHz
PDH UGH is 8 kHz

Plot 2
PLL error spectrum

Plot 3
PDH error spectrum

Plot 4
Estimated transmission frequency noise from PDH error spectrum with arm pole applied.
In loop ALS X Transmission Frequency Noise RMS is 1 Hz.

Now that ALS COMM is locking, we will get an out-of-loop measurement of ALS frequency noise soon.

I have installed advLigoRTS-3.2.3 and created a corresponding build area for this release ( /opt/rtcds/lho/h1/rtbuild-3.2.3 ) I have made the necessary release symbolic link changes.

I compiled the SUS QUAD models using RCG-3.2.3 in preparation for their upgrade to permit additional violin mode damping filters to be added.