NPRO NE noise when both FSS and ISS are unlocked (Peter, Sheila, Keita)

This morning Peter wanted to measure the NPRO noise when both FSS and ISS were unlocked but he ran out of time to give IFO to commissioners, so all his measurements from the morning was FSS locked. We agreed that we need to disable FSS and measure PMC mixer instead of FSS FAST.

Later I and Sheila quickly disabled FSS and ISS and measured PSL-PWR_NPRO as well as PMC mixer. (When FSS is locked PMC mixer doesn't show anything, as expected.)

In the attached, red and blue are when FSS and IFF were off, green and brown are when FSS was off but ISS was on. (Don't pay much attention to the difference between red and green or between blue and brown, the oscillation was slowly getting smaller as we measure.)

Important message here is that the noise is there both in frequency and in amplitude without ISS nor FSS.

 

Remote noise eater operation

For what it's worth, independently confirmed that remote operation of the noise eater works.

NPRO power noise

Whilst trying to figure out whether or not the remote noise eater was operational, I measured the
NPRO power noise around the 250 Hz area to see if there was a bump there or not.  This was done
prior to measurements taken yesterday afternoon, after which the remote operation was enabled.

NPRO crystal temperature scan

Summary
There are no places with the noise eater on where there is no broad peak in the NPRO PZT spectrum.  Turning off the noise
eater gets rid of the broad peak.

Conducted a scan of the NPRO crystal temperature (aka the SLOW actuator) and made a note of where a resonance appeared with the reference cavity.
Whilst doing so, spectra of the NPRO power noise and of the FAST actuator were taken whilst the FSS was locked.


	
NPRO crystal temperature scan
	

		

			
slider setting
			
crystal temperature
			
spectral image
		
	
	

		

			
-0.706
			
-0.81
			
NE_CompP.png
		
		

			
-0.699
			
-0.805
			
NE_CompN.png
		
		

			
-0.692
			
-0.80
			
NE_CompM.png
		
		

			
-0.436
			
-0.54
			
 
		
		

			
-0.434
			
-0.54
			
NE_CompL.png
		
		

			
-0.427
			
-0.538
			
NE_CompK.png
		
		

			
-0.177
			
-0.277
			
NE_CompJ.png
		
		

			
-0.169
			
-0.28
			
NE_CompH.png
		
		

			
-0.165
			
-0.27
			
NE_CompG.png
		
		

			
0.052
			
-0.06
			
NE_CompQ.png
		
		

			
0.082
			
-0.023
			
NE_CompA.png
		
		

			
0.088
			
-0.021
			
 
		
		

			
0.332
			
0.23
			
NE_CompB.png
		
		

			
0.339
			
0.23
			
NE_CompC.png
		
		

			
0.574
			
0.47
			
NE_CompD.png
		
		

			
0.576
			
0.47
			
NE_CompE.png
		
		

			
0.578
			
0.47
			
NE_CompF.png
		
		

			
0.585
			
0.48
			
 
		
	


 The measurements were recorded with the common gain set to 20 dB and the fast gain, 3 dB.  Lowering the common gain to 16 dB did not improve things.   In NE_CompA.png, the noise eater was off for the red trace and on for the blue trace.  All other plots are relative to this plot and were taken with the noise eater on. 

ETMX effective bias measurement

To check the results from my ETMX electrostatic co-efficient measurements (alog 42572) I ran the usual optical lever V_eff measurements. The results show roughly 0 to -20V on all quadrants, which is the same ballpark as I measured earlier this week.

I'm attaching a plot of the V_eff trends since September 2016 (first attachment), a zoomed in plot of the new data (second attachment), and one of the five plots that the ESD_Night_ETMX.py script spits out, showing optical lever transfer function as a function of effective bias (third attachment). We can do more with this data than just work out Veff: from the TF value at V_bias = 0V we can work out Beta-Beta_2, and from the slope we can work out gamma - alpha.

AS_C pico, AS AIR camera

Sheila, Georiga, Craig, Gabriele

After this morning's PR2/PR3 walk, we centered on the OFI aperture using single bounce and moving SR2 with the AS centering loops on.  Georgia pico'd to center on AS_C after this was done. 

We also recentered on the AS AIR camera and added a ROC 64.4 mm lens a few inches in front of the camera because the beam was larger than the camera image.  Attached is a screenshot of what the camera now looks like in PRMI, it seems like there is a pitch mode.

Ops Day Shift Summary

TITLE: 06/20 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
INCOMING OPERATOR: None
SHIFT SUMMARY: Some commissioning work, then NPRO temp change.
LOG: See attached.

ITMY HWS Ring Heater Scan

TJ, TVo

At the end of last week, we moved SR3 to get the beam out to HAM6 without clipping on the OFI but this also required realignment on the HWS.  There was a little a bit of diffraction on the upper right corner of the return HWS beam so we tried to make a ring heater scan to see if this has any noticeable effect on the results (like extra noise), we ran the ring heater at 5 Watts input power on each segment for 3 hours (starting midnight local time).  Attached are the time trends and a picture of the contour plot.  I estimate the beam is off center on the test mass in pitch by about 1.5 cm and yaw by about 0.38 cm.  Depending on how the rest of the IFO alignment goes, such as ant SR3 moves, we might hold off on optimizing these offsets.

Monthly Dust Monitor Vacuum Pump Check

    All three vacuum pumps (CS, EX, EY) are operating within normal temperatures and vacuum pressure. Made minor tweaks to the End-X and the CS vacuum pressure. End-Y was OK. Closing FAMIS task #7526. 

End station Inficon gauges

FRS 9721

Inficon gauges at EX were relabeled and now match the D1700275 drawing.

PT-526 gauge is lower gauge on chamber 5
PT-525 gauge is upper gauge on chamber 5

ITMY HWS streaming code

History: The new code to stream the HWS camera images that we received May(?) of last year works well on h1hwsmsr (ITMX), but we haven't been able to get it run on h1hwsmsr1 (ITMY). It would bring up a single image with an error: "KeyError: ".

Today, I poked around the code to see if I could figure out where the error was coming from and found that the FuncAnimate function from matplotlib had set the refresh interval from a default 200ms to 0ms, and turned on blit to help with image processing. I changed the interval time back to its default and then turned off blit and now we can stream the image and the GUI works.

Now we are just left with the error that the old streaming script also had:

"Unable to configure ring buffer
rc = -1
pdv_multibuf(0x0x21ab010, 4): Invalid argument"

It is odd that this error is only on h1hwsmsr1 and not h1hwsmsr. TThe error comes somewhere from the script /opt/EDT/pdv/take. My C knowledge is minimal, so I am a bit stuck for now.

At least we have a working streaming script!

More PRMI alignment

Following up on last night's work 42590, more PRMI alignment. At the beginning I realigned MICH at dark fringe, and then PRX. Power were initially at similar levels as yesterday (POPAIR_B_RF18 ~ 20, POPAIR_B_RF90 ~ 27)

Most of this morning's improvement was due to moving PR3 and PR2 together, both in pitch and yaw. Powers improved by a further 15-25% with respect to yesterday's night: POPAIR_B_RF18 ~ 23, POPAIR_B_RF90 ~ 34)

I also moved the input beam by acting in IM3 and IM4. I saw some effect, but not very large.

Attached plots with trends and offsets.

Remote noise eater operation

The noise eater switch is not toggling.  This was checked by:
 - measuring the NPRO power noise out to 1 MHz
 - measuring the voltage across the connector at the front panel of the control box, which is effectively the output
     of the KL2612 terminal

    TwinCAT recognises a change of state when one is requested.

Update
    The cable from the NPRO power supply was modified.  Remote operation of the noise eater is now possible.  On the
FSS MEDM screen, on the left hand side there are two buttons ON (green) and OFF (red) and an indicator next to the
"NPRO RRO:NOISE EATER" indicator.  These two buttons operate the noise eater.  If it's on the indicator to the right
of the two buttons is green.  Off, red.




  Fil / Peter

Slightly better PRMI buildup

[Sheila, Craig, Georgia]

Summary

We had some clipping, probably in the PRC. We adjusted IM4 and now have 25% improvement in PRC buildup as measured on POPAIR_B_RF18, and 10% as measured on POPAIR_B_RF90 (compared to last night). This is still lower than we expect, but it is a step in the right direction.

Details

We checked the PRMI build-up again, locking PRX, MICH, then PRMI, and found it similar to yesterday (alog 42570).

We checked the beam on the POPAIR_B diode and found that the beam is overfilling the diode, we should install a shorter focal length lens in front of this but it is not urgent.

We then tried to check for clipping: we set the IFO for single bounce off ITMY, and made a feedback loop between AS_C and ITMY. We moved IM4 in yaw and found that we lost power straight away. We checked the AS_A and AS_B centering loops and found that we were clipping on the OFI, which we then fixed with SR2. We then tried to check for clipping again but this time using AS_A, we moved IM4 in pitch and found that the sum output of AS_A increased, which suggested clipping. To maximise the AS_A sum we had to move IM4 roughly -2000 counts in pitch (from original alignment of -400 to -2400).

Meanwhile Craig checked the spot positions on the PRC mirrors and found we were off in pitch, he'll attach his results in a comment.

We went back to aligning PRX, MICH and PRMI walking IM4, PRM, and PR2, screenshot of alignment sliders attached. We found a 25% improvement in build-up according to POPAIR_B_RF18. We re-ran Craig's spot position measurement and found we were closer to the centres of PR2 and PR3 in yaw. This is a good start. Stay tuned!

 

PRC mirror spot estimates after PRC A2L minimization using Hang's decoupling script and Kiwamu's HSTS and HLTS alpha to mm:

All numbers are in mm from the LL coil.


 PRM P      PR2 P      PR3 P      PRM Y     PR2 Y       PR3 Y
-2.4890     0.3589    -3.1097    -1.6346    17.1775    -15.8104
-2.5006    -0.2489    -2.8373    -1.5536    16.5192    -13.5315


Jenne's coil signals
Koji's alpha definition
More Kiwamu

I calculated the corrections to Kiwamu's alpha to mm values, according to Koji's 2017 correction to elog 2863

 
Corrected HSTS and HLTS alpha to mm factors:
============================================
HSTS alpha to mm: 39.28
HLTS vertical alpha to mm: 50.93
HLTS horizontal alpha to mm: 87.43


Also attached alogs 42600 and 42601 to Kiwamu's original alogs to avoid future confusion.

New miscentering in PRC after alpha to mm factor corrections (plus additional alignment happening right now)

All numbers are in mm from the optic center using LL as the fiducial coil. 


  PRM P     PR2 P     PR3 P     PRM Y     PR2 Y     PR3 Y
-2.7622    1.1651   -0.7635   -1.2032   15.3530  -12.3856

unclipping on PR2 baffle

Alexa, Kiwamu,

We spent most of the daytime today for

• Measuring the spot position on various optics
• Moving the spot on PR2 by 12 mm to minimize the risk of clipping on the PR2 scraper baffle

There were a few noticable effects after the PR2 spot was moved:

• the right side of the BS elliptic baffle as seen by the analog camera became brighter.
• POPAIR_RF18 increased to 260 uW from 220 uW in the PRMI sb lock.
• However, the optical gain did not change either in PRCL of MICH, indicating that the recycling gain did not change.

 

Summary of the spot measurements before and after the unclipping.

		before [mm]	after [mm]
PRM	Pitch	-2.82	NA
	Yaw	0.402 toward West	NA
PR2	Pitch	-6.04	-5.44
	Yaw	2.26 toward West	14.1 toward West
PR3	Pitch	-24.5	NA
	Yaw	0	NA

 

(Some more notes)

We aimed for a displacement of 16 mm on PR2 in order to bring the beam to the ideal point according to Keita's measurement (see alog 14640).  However, later, I realized that I had picked a wrong number from his entry -- I must have chosen eitehr 14 or 10 mm. Anyway, the resultant dither measurement suggested that we moved it by 11.5 mm to the right direction on PR2. So it happened to fall into a OK displacement. I think that we anyhow improved the clipping situation on the PR2 baffle.

The POP extraction path was then corrected by using the pico-motorized mirror in HAM3. This was not a difficult task as the beam was hitting the swiss cheese baffle and therore it was visible. We then went out and did fine tunings. The pico motor was adjusted to recover the highest green TRX. We steered a steering mirror for each POP diode.

In order to compute the off-centering on HLTS, I used the mirror sizes described in D1101381.

• R = 132.5 mm
• L = 101.4 mm
• D = 206 mm for vertical 
• D = 120 mm for horizontal
• conversion factor for vertical = 52.5 mm per alpha
• conversion factor for horizontal = 94.9 mm per alpha

For computing the HSTS spot position, I used 42.2 mm per alpha (see alog 13765).

In 2017, Koji corrected his coil balancing calculations from 40m elog 2863

Now,  instead of , where  is the ratio of moment of inertia and optic mass,  is the coil force imbalance factor, and   is the distance between coils.

Corrected alpha to mm for HLTS optics: 
Vertical: 50.93 mm / alpha
Horizontal: 87.43 mm / alpha

For HSTS optics, see alog 42601

PRM spot position: good

I measured the spot position on PRM by using the usual angular dither technique (see for example LLO alog 5010) with PRY locked. It turned out to be pretty good.

• measured vertical off-centering = too low by 2.0 mm
• measured horizontal off-centering = off toward PR3 by 1.0 mm

 

(some settings):

• ITMX and SRM are intentionally misaligned.
• PRY is locked on the carrier light by REFL_A_RF45_I with a demod phase of 6.9 deg
• No whitening gain, but whitening stage 1 and 2 are engaged to obtain better signal-to-noise ratio.
• Feedback on PRM and PR2 as usual
• servo gain in LSC = -50 which should give a UGF of 30-60 Hz.
• Excitation in SUS-PRM_LOCK_P(Y) with an amplitude of 1000 counts at 10.1 Hz
• I used dtt to take a transfer function of (REFL_A_RF45_I) / (PRM_LOCK_P(Y)) 
• Before the measurement, PRM was tweaked to maximize the intracavity power

(results)

See the plot below.

(some notes)

• At the beginning of the measurement, the beam on REFL_A was at the edge of REFL_A. This resulted in bogus data because it introduced some kind of wave-front-sensing effect as the PRM angle was modulated.
  • This was fixed by moving the tip-tilt mirrors in HAM1, namely RM1 and RM2.
• I used a conversion factor of 42.2 mm for converting the imbalance alpha (see 40m elog 2863 for the definition) into the position of the rotational nodes.
  • According to D1101377-v2,  I used the following numbers to derive the conversion factor:
    • radius of the mirror: R = 75 mm
    • mirror thickness: L = 75 mm
    • magnet separation D = 95.45 mm

In 2017, Koji corrected his coil balancing calculations from 40m elog 2863

Now,  instead of , where  is the ratio of moment of inertia and optic mass,  is the coil force imbalance factor, and   is the distance between coils.

Corrected alpha to mm for HSTS optics: 39.28 mm / alpha

For HLTS optics, see alog 42600.