Starting Stiffener rings and o ring protectors installation on HAM4 - Apollo

(Alexa, Daniel, Keita)

After changing the RF frequency to 24.389319 MHz and adjusting the delay line phase to 257 steps, we took open loop transfer functions of the PDH loop.

PLL Board Settings:

• Input 1 enable: ON
• Input 1 pol: NEG
• Input 1 gain: 0dB
• Common Comp: ON
• Boost 1: ON
• Generic Filter: ON
• Fast Option: ON
• Fast Gain: -8db
• rest off/zero

PDH Board Settings (Nominal):

• Input 1 enable: ON
• Input 1 pol: NEG
• Input 1 gain: -1dB
• Common Comp: ON
• Boost 1: ON
• rest off/zero (unless specified)

EX_PDH_OpenLoopTF_phase/mag_Nom.txt corresponds to the data collected with the above nominal settings.

EX_PDH_OpenLoopTF_phase/mah_Boost2On.txt corresponds to the data collected wiith the above settings with the addition of Boost 2 ON

EX_PDH_OpenLoopTF_phase/mag_6dBFast.txt corresponds to the data collected with the same as the nominal settings with the addition of 6dB gain in the fast path

Heading into the LVEA to check a dust monitor (location #9) which has lost communication - Patrick

Moving, cleaning, and organizing elements for SR2 installation (HAM4 LVEA) - Jeff.B/Jodi

PSL Check:
Laser Status: 
SysStat is good
Output power is 29.3 W (should be around 30 W)
FRONTEND WATCH is Active

PMC:
It has been locked 2 day, 16 hr 11 minutes (should be days/weeks)
Reflected power is 1.1 Watts  and PowerSum = 12.16 Watts.
(Reflected Power should be <= 10% of PowerSum)

FSS:
It has been locked for 11 h and 22 min (should be days/weeks)
Threshold on transmitted photo-detector PD = 0.94V (should be 0.9V)

ISS:
The diffracted power is around 5.4% (should be 5-15%)
Last saturation event was 11 h and 28 minutes ago (should be days/weeks)

I saw a big RAM offset in REFLAIR_RF45 and didn't like it. So I changed the modulation frequency to 9099170 Hz where the RAM offset became zero. This frequency is consisent to what Alexa got (see alog 9679) in her IMC length measurement.

Before this change, the mod frequency was at 9099471 Hz and the RAM offset was 3000 counts in REFLAIR_RF45_Q with its whitening gain at 21 dB and with PRM aligned.

Since the laser power was increased to 10 W, I wanted to see if we can lock the Michelson using REFLAIR_RF45 with PRM misaligned. And yes, the signal to noise ratio was barely large enough to keep the Michelson locked.

Another big motivation of locking the Michelson was that I wanted to calibrate the BS length actuator using the Michelson fringe. I took some data and handed to Yuta who volunteered to do some fitting and so on.

Some observations in the Michelson lock:

One difficulty in the Michelson lock was that, as reported by the LLO commissioners, a bounce mode of the BS suspension at around 17.5 Hz  (see for example aLIGO wiki) limits the control bandwidth to be lower than it. A high UGF maybe able to suppress it, but then a complication is that the control signal saturates the DAC which eventually breaks the loop. This is something we should keep in mind particularly when locking PRMI.

Also because of the low signal to noise ratio at REFLAIR_RF45, the signal above 4 Hz was limited by some electronics noise which was flat. Then this high frequency noise simply saturated the DAC until I newly installed a cutoff filter at 30 Hz (which is now in FM9 of LSC-MICH).

The attached shows in-loop noise spectra calibrated in m/Hz^1/2. The Michelson optical gain was calculated to be 6.79e9 counts/m from the peak-to-peak value of the free swinging waveform which was measured to be 1149 counts. As you can see, the bounce mode was prominent at 17.9 Hz after it got excited by my swept sine measurements.

Also I am attaching the measured OLTF. The UGF was set pretty low at 2 Hz.

We have seen that the arm cavity can sometimes lock quite stably (for example the 6 hours  starting UTC Jan 24 1:40) alog 9518, and that sometimes it mode hops constantly (alog 9653).  I used the scripts Keita used in alof 9653 to make a comparision of the two times.  For the stable time I am using 20 minutes of data starting at 1/24/14 1:40 UTC, for a bit less stable time I am using 1hr 20min starting at 1/24/14 6:30 UTC and a slightly less stable time of Keita's plots use 5 minutes of data. 

First histograms of the transmitted counts: the stable time, the somewhat less stable time, and Kieta's data:

On the 24 (first two plots) the SHG was off, so there is no 850-900 count offset on COMM_A_LF that you see in Keita's data from this morning.   The transmitted power this morning was only 80% of what it was durring the stable times, even when locked to the 00 mode (if you take into account the offset from the SHG).  

You can see that there was mode hopping in Keita's data. Although it seems like the "less stable" (middle) histogram has a bimodal distribution, I don't think that the small hump aroung 550 counts is due to mode hopping because it has about 80% of the power that the 00 mode has.  The mode matching should not have changed between the 24th and today, and we know we didn't have more than about 30% of our power in mode mismatch (alog 9518).   These are most likely drops in the transmitted 00 mode power due to alignment excursions. 

Here are plots of where the PIT oplevs spent most of their time, made using Ketia's script.  This morning there was about 2-3 times more pitch motion in the ETM than on the 24th (see Jeff's alog).  

These are plots made using Keita's script, of the transmitted power for different oplev positions.  (I didn't use the threshold to distinuish mode hopping since there isn't really mode hopping in the first two plots, and I used a smaller bin spacing of 0.1urad) In the first two plots, we seem to be exploring a plateu where the transmitted power doesn't change dramatically over 1urad in ETM pitch and 1/2 urad in ITM pitch.  This morning though we were misaligned, and even with less optic motion we would have been right on the edge of the mode hopping.  It seems as though we can't rely on the dither alignment when we are this misalinged, and need to tune up the alignment by hand, probably by unlocking and watching the fringes.  

The first two plots suggest we can stay locked to the zero zero mode with around 1 urad pp in the ETM, maybe more.  The plot from this morning seems to show that while the optic motion was large, our main problem was a DC misalingment. 

[Yuta, Paul]

This afternoon we set up the hardware for the IMC sideband sweep measurement, and took a test sweep of two cavity FSRs. This is in preparation for running automated sideband sweeps while stepping up the input power to get an estimate of IMC optics absorption (see e.g. LLO alog entry 9095).

First we checked the centering of the REFL beam on the in-vac LSC diodes, adjusting PRM alignment for this purpose.

Then we re-aligned the path from the ISCT1 REFL periscope to the in-air REFL diodes, after I misaligned it earlier when doing beam size measurements. Next we connected the Agilent network analyzer source directly to the 45MHz EOM input, without using the 18.8dB RF amplifier that was used last time for this measurement (see LHO alog entry 8098), and connected the 27MHz channel from REFL_AIR_B (broadband PD on ISCT1) to the network analyzer input.

Initial observations of the transfer function from N.A. drive into the EOM to a.m. measured by REFL_AIR_B around the 45.5MHz FSR showed the expected split peak structure (see LHO alog entry 8086). We then put a razor blade beam dump partially in the beam in front of REFL_AIR_B to break HOM symmetry at the PD. We misaligned the IMC to increase the HOM content injected, and then took a sweep across 2 FSRs from 36MHz to 55MHz. At some point during this process we noticed that the TEM00 peaks got much bigger, and also ceased to exhibit the split structure (see attached plot). Discussing this with Kiwamu later it appears that the IMC length shifted somehow during this time, giving a lot of a.m. in transmission with the previously used modulation frequency. It will be interesting to see if these peaks are reduced now that he has adjusted the modulation frequency to compensate.

The attached plot shows the full 2 FSR sweep from this afternoon, overlaid with the data taken in October and the output of a Finesse model. In general, in today's data we see larger contributions from the HGx0 modes than HG0x modes compared to the previous data. We suspect that is due to the razor dump mainly breaking the horizontal symmetry on the diode. Last time we used an iris, which may have been more even in blocking vertical and horizontal axes. We might wish to go back in and use a similar technique this time. Today's data also shows the much larger TEM00 peaks. Overall the SNR looks pretty good, and we'll move on to trying the automated sweeps from the control room.

The razor dump is still partially blocking REFL_AIR_B. We disconnected the EOM 45MHz input cable from the N.A. source, and reattached it to the 45MHz port on the RF distribution panel. We reconnected REFL_AIR_B cable to the demod board.

J. Kissel, R. Mittleman

Another day, another dollar's worth of ISI problems. This morning, I received reports of more than 0.1 [urad] RMS motion of the arm cavity BSC-ISIs -- now ISI-ETMX. As I've done every day this week, I'd resolved to just install improved QUAD damping filters with higher damping gain. However, the first step in this process is to use the current performance of the ISIs as input motion. As soon as I gathered the data, I found that the performance of the ETMX was polluted by a -- you guessed it -- ~0.5 [Hz] spike feature / oscillation / resonance in the X DOF (aligned with the SUS ETMX's L DOF), which rings up the 0.44 [Hz] and 0.56 [Hz] modes of the QUAD. The worst part about it is that this particular feature had appeared over-night. 

Long story short, we did NOT find the source of the problem in the days worth of investigation, but it DID go away while we were investigating. Along the way, we've discovered several things that were poorly done on this chamber, which I'll detail below. However, now that feature is gone, I'll once again start the process of designing more beefy QUAD damping filters.

-------------
The Story, in brief (each bullet corresponds to the attachments, in order):
- Took performance spectra of X direction (ST1 T240s, ST2 GS13s, and QUAD Optical Lever Pitch), at 2014 Jan 30 15:00 UTC (before the morning meeting, local time). Found feature at 0.51 [Hz] (with a resolution 0.01 [Hz]). 
     - HEPI Isolated (Level 1 isolation filters, Position-Only Blends), 
     - ISI Isolated (Leel 1 isolation filters, ST1 XY T100mHz_N0.44Hz, all other DOFs, and ST 2 with 750mHz blends), 
     - QUAD Damped (Level 2.1 damping filters.) 
- Looked at all Cartesian degrees of freedom on ST2 and ST1, only saw spike in X.
- Suspecting individual sensor flaws again, looked at raw T240 and raw CPS with only damping loops on. Saw nothing suspicious in any DOF.
- Closed loops again, tried blending ST1 without T240s (i.e. All DOFs on ST1 and ST2). Spike disappears, but the amplitude of the residual seismic noise is higher than peak amplitude of the feature. In-conclusive test. 
- Suspecting loop instabilities, completely characterized X DOF isolation loop in both 750 mHz configuration and T100mHz_N0.44 configuration, measuring open loop gain transfer functions, closed loop gain transfer functions, plotting blend and isolation filters, etc. etc. While we found several problems, none of which would cause any instability.
- Remeasured performance of ETMX at 2014 Jan 30 21:00 UTC (during lunch / journal club local time), and feature disappeared. Good flippin' grief.
- Measured plant with isolation loops open to assess signs, appear to be self-consistent, though they don't match the conventions in T1000388. 

The problems we found (which again, won't explain the ~0.5 [Hz] peak):
(1) Both sets of blend filters installed have flaws (see the first 2 pages of the "LoopChar" attachments):
     - 750 mHz: The L4C Highpass has way too much gain at low-frequency. Now, as discussed earlier, the filter has to include the L4C response, but the roll-off of the complementary filter should start at a much higher frequency, say 200 mHz. Further, the displacement sensor low-pass should fall off (in frequency) faster than the L4C, they're currently both falling as 1/f (as indicated by the matching phase of -90 [deg]). I know these are supposed to be the "robust, OK performance" filters, but c'mon. 
     - T100mHz_N0.44Hz: The displacement sensor low-pass filter should fall off at-least faster than the L4C, if not faster than the T240. It current returns to flat in frequency, re-injecting displacement sensor noise above 100 [Hz]. 

(2) Though we measured matrix signs of the plant to be self-consistent (see 2014-01-30_1500_H1ISIETMX_Plant.pdf): at "DC," (the bottom of our band of interest)
     - CPS = Flat in magnitude, and 0 [deg] phase, i.e. in phase, and in the same direction as the actuators
     - T240 = Rising as f in magnitude +90 [deg], i.e. with the response to velocity flat above 4 [mHz]
     - L4C = Rising as one factor of f in magnitude (+90 [deg] in phase) from being a velocity sensor, as well as two more factors of f (and another +180 [deg] in phase) from the inertial sensor response not yet compensated, for a grand total of what's shown: rising as f^3, with a phase of -90 [deg]
Comparing against the new document on conventions (T1000388), the individual elements are bonkers, and some coefficients are off by ~10-20%. 

These ISIs really need some TLC.

Aidan, Thomas, Eric G.

We installed the TCS HWS HAM4 optics this afternoon per D1201098. We've finished using that area and the chamber is covered again.

Full details tomorrow.

Daniel noticed that following his inclusion of an ALS top_named part in the h1asc model this morning, the channels were being incorrectly named at H1:ASC-ALS_ instead of the expected H1:ALS-. Jim, Cyrus and myself spent some time diagnosing the problem and found that the problem is raised if the second part comes before the main part (ASC) in alphabetical order. In other words ALS was not correctly identified as a top_named part but ATT was. It appears the problem may be in the post_built python script where the "top_names" block property is parsed. We did some diagnostics with the h1asc model, so there may be some strange MEDM screens lying around. We never did a "make install" so the INI file was not changed since this morning.
The investigation continues.

I have been silently checking the signal chain of the REFLAIR and POPAIR RFPDs using the AM laser (a.k.a. PD calibrator) to make sure that they are functional expectedly.

Summary

• REFLAIR_A and POPAIR_A looked good as they showed reasonable signal gains
• On the other hand, REFLAIR_B and POPAIR_B RFPDs showed relatively large discrepancy from the expected values. I might have missed some gain factors somewhere. I will revisit these RFPDs later when I get a chance.

The RF frequency of the AM modulation was adjusted in each measurement such that the demodulated IF signal was below 50 Hz.

Calibration of the amplitude modulation depth

We recalibrated the AM laser.

The current setting of the laser was changed recently because we opened up the current driver when we thought the laser diode had been dead in the early December. Then the laser head and its current driver were sent to Rich at Caltech for his extensive testing although the laser magically fixed itself and he didn't find anything wrong. So this was the first time for us to use the AM laser which had been fixed. Because of that mysterious event, I wanted to recalibrate the laser. First of all, Yuta and I measured the power to be 2 mW with an Ophir Vega without the attenuation filter. Then we measured the modulation depth for the amplitude modulation by using a Newfocus 1611 as a reference.

The new calibration for the amplitude modulation is:

P_am =  5.13 mW x (P_dc / 1 mW) * (1 V / V_drive)

where P_dc is the laser power at DC and V_drive is the drive voltage when it is driven by a 50 Ohm source. For example, if one puts this laser to a PD which then shows a DC laser power of say 2 mW, the AM coefficient is now 5.13 mW x ( 2 mW / 1 mW) /V_drive = 10.26 mW/V_drive.

• Some details for derivation:
  • DC transimpedance of NF1611= 10k Ohm
  • AC transimpedance of NF1611 = 700 Ohm
  • Responsivity of NF1611 at 980 nm = 0.5 A/W
  • Measured DC voltage = 7.36 mV
    • => DC laser power of 7.36 mV / 10kOhm / 0.5 A/W = 1.47 uW
  • Measured AC voltage when driven by 0.5 Vpp at 10 MHz = 1.32 mVpp (measured by a 50 Ohm oscilloscope)
    • => P_am = 1.32 mV / 700 Ohm / 0.5 A/W = 3.771 uW
    • => AM coefficient is 7.54 uW / V_drive
  • Therefore the AM calibration is:
    • 7.54 uW / 1.47 uW x 1 mW = 5.13 mW x (P_dc / 1 mW) * (1 V / V_drive)

REFLAIR_A_RF9 (S1203919)

Remarks:

The signal chain is OK. The PD response is smaller by 15% for some reason.

It seems as if the transimpedance is smaller by 15% than what had been measured at Caltech (LIGO-S1203919). The cable loss from the RFPD to the rack was measured to be 0.47 dB. Be aware that the demod gain is half of the quad I/Q demodulator because this is a dual channel demod (see E1100044). The demod conversion gain is assumed to be 10.9 according to LIGO-F1100004-v4.

• Measured DC at BNC jack = 108 mV
  • DC laser power = 108 mV / 100 Ohm / 0.67 A/W = 1.61 mW
• Expected AM signal when driven by 0.1 Vpp = 5.13 mW x (1.61 mW / 1 mW) x 0.1 Vpp x 0.67 A/W x 245 Ohm = 135.6 mVpp
• Measured AM signal at the SMA jack of the RFPD enclosure = 115 mVpp
  • This is already smaller by 15%.
  • If we blame the transimpedance, it must be 208 Ohm rather than 245 Ohm.
• Measured AM signal at the rack = 109 mVpp
  • cable loss = 0.47 dB
• Expected IF signal = 135.6 mVpp x 10.9 / 2. = 622 mVpp
• Measured IF signal = 570 mVpp
  • If we take the 15% and 0.47 dB losses into account, we would expect this to be 527 mVpp.
• Expected ADC counts = 135.6 mVpp x 10.9 x 2^16 counts /40 V x 1/2 = 1209 counts pp
• Measured ADC counts = 900 counts pp
  • If we take the 15% and 0.47 dB losses into account, we would expect this to be 973 counts pp.
  • This is within 10% divation. Good enough.

REFLAIR_A_RF45 (S1203919)

Remarks:

The signal chain is healthy.

Found cable loss of about 1.5 dB. The measurements excellently agree with the loss-included expectation.

• DC laser power = 1.61 mW (the same as REFLAIR_A_RF9)
• Expected AM signal when driven by 0.1 Vpp = 5.13 mW x (1.61 mW / 1 mW) x 0.1 Vpp x 0.67 A/W x 341 Ohm = 188.7 mVpp
• Measured AM signal at the SMA jack of the RFPD enclosure = 190 mVpp
  • Very good agreement !
• Measured AM signal at the rack = 160 mVpp
  • Corresponds to loss of 1.5 dB
• Expected IF signal = 188.7 mVpp x 10.9 = 1028 mVpp
  • If the loss is taken into account, it is going to be 1028 mVpp x -1.5 dB = 865 mVpp
• Measured IF signal = 880 mVpp
  • Consistent with the loss-included expectation
• Expected ADC counts =1028 mVpp x 2^16/40 x 2 = 3369 counts pp
  • If the loss is taken into account, it should be 3368. x -1.5 dB = 2837 counts pp
• Measured ADC counts = 2784 counts pp
  • Good agreement with the loss-included model

POPAIR_A_RF9 (S1300521)

Remarks:

The signal chain is healthy.

The measurement suggests that there is loss of 1 dB somewhere. I didn't measure the cable loss this time.

• Measured DC at the BNC jack = 95.6 mV
• Estimated DC laser power = 95.6 mV / 100 Ohm / 0.67 A/W = 1.42 mW
• Expected AM signal = 5.13 mW x (1.42 mW / 1 mW) x  0.1 Vpp x 0.67 A/W x 741 Ohm = 361.7 mVpp
• Expected ADC counts = 361.7 mVpp x 10.9 x 2^16/40 counts/V = 6454 counts pp
• Measured ADC counts = 5705 counts pp
  • If we assume this discrepancy is from a loss, it corresponds to loss of 1.07 dB.

POPAIR_A_RF45 (S1300521)

Remarks:

The signal chain is OK. Though loss sounds a bit too high.

The measurement suggests a possible loss of 2.6 dB somewhere. I didn't measure the cable loss.

• Estimated DC laser power = 1.42 mW
• Expected AM signal = 5.13 mW x (1.42 mW / 1 mW) x 0.1 Vpp x 0.67 A/W x 837 Ohm = 408.5 mVpp
• Expected ADC counts = 408.5 mW x 10.9 x 2^16/40 counts/V = 7295 counts pp
• Measured ADC counts = 5391 counts pp
  • If the discrepancy is assumed to be from loss somewhere, the loss should be 2.6 dB.

REFLAIR_B_RF27 (S1200234)

Remarks:

The signal gain is bigger than the expectation by a factor of 2.3.

• Measured DC at the BNC jack = 836 mV
  • This corresponds to a laser power of 836 mV / 2000 Ohm / 0.4 A/W = 1.045 mW
• Expected AM signal = 5.13 mW x (1.045 mW / 1 mW) x 0.05 V_drivepp x 0.4 x 1000 Ohm = 107.2 mVpp
• Expected ADC counts = 18 dB x 107.2 mVpp x 10.9 x 2^16/40 counts/V = 15206 counts pp
  • I assumed that the gain factor of the diplexer (LIGO-D1300989) to be 18 dB at 27 MHz.
• Measured ADC counts = 35431 counts pp
  • This is bigger than the expected by a factor of 2.3

REFLAIR_B_RF135 (S1200234)

Remarks:

The signal gain is bigger than the expectation by a factor of 1.5

• Measured DC laser power = 1.045 mW
• Expected AM signal = 5.13 mW x (1.045 mW / 1 mW) x 0.0014 Vdrivepp x 0.4 A/W x 1000 Ohm = 3 mVpp
• Expected ADC counts = 43 dB x 3 mVpp x 10.9 x 2^16/40 counts/V = 7573 counts pp
• Measured ADC counts = 11689 counts pp
  • This is bigger than the expected by a factor of 1.54

POPAIR_B_RF18 (S1200236)

Remarks:

The signal gain is bigger than the expectation by a factor of 2.3

• Measured DC at the BNC jack = 744 mV
• DC laser power = 744 mW / 2000 Ohm / 0.4 A/W = 0.93 mW
• Estimated AM signal = 5.13 mW x (0.93 mW / 1 mW) x 0.1 Vdrivepp x 0.4 A/W x 1000 Ohm = 191 mVpp
• Estimated ADC counts = 191 mVpp x 10.9 x 2^16/40 counts/V = 3404 counts pp
• Measured ADC counts = 7803 counts pp
  • This is bigger than the expected by a factor of 2.3

POPAIR_B_RF90 (S1200236)

Remarks:

The signal gain matches with the expected value, but I don't believe this.

• DC laser power = 0.93 mW
• Estimated AM signal = 5.13 mW x (0.93 mW / 1 mW) x 0.1 Vppdrive x 0.4 A/W x 1000 Ohm = 191 mVpp
• Estimated ADC counts = 191 mVpp x 10.9 x 2^16/40 counts/V = 3408 counts pp
• Measured ADC counts = 3549 counts pp
  • This suspiciousely shows an excellent agreement with the expected one.

(Correction Jan/31: Sheila's made a similar plot for a period (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=9691) where the cavity was locked all the time, and it seems like the plot presented here was sampling only a small edge of the 00 lock range.)

As of now, it seems like both ETM and ITM should be within 0.3 urad pk-pk from the optimal alignment to stay locked on 00 mode.

Plot 1: I took OL and green transmission data for 5 minutes when the cavity was going back and forth between 00 and 01 mostly without losing lock.

In the top panel, the transmission time series is actually the sum of the SHG and the transmission, but it's on 00 mode when it was 1400 counts, 01 when it was between 1000 and 1200 counts.

Plot 2: 2D color map of the above data.

X axis is ITM OL PIT, Y is ETM. I arbitrary set the threshold of 1300 counts for green transmission and subtracted that from the transmission data, and set the color map such that orange-red color is above 0 (i.e. locked to 00 mode), yellow-green is negative  (01 mode).

Each cell is normalized by the number of data points in that cell. Blue means there's no data point in that cell.

Sorry that X and Y axis scaling is different. To aid your eyes, I put white and pink diagonal lines showing the direction of common (soft) and differential (hard) misalignment.

Plot 3: 2D plot of where the OLs were staying.

You cannot tell from plot 2 where the OL was staying the most, so here is the number of data points in each cell divided by the total number of data points. Adding everything on this map together you'll get 1.

On average EX=-28.12, IX=-6.28.

Conclusion:

1. Thin band on 00 mode area, elongated along the white (soft or common) line,  shows that the problem is along the pink line (differential or hard) due to cavity geometry discussed before by people.
2. In order to stay locked on 00, both ITM and ETM should be within 0.3 urad pk-pk or somethibng, otherwise you'll drop out of the locking range along the pink line.
3. The DC alignment was not optimal despite dither alignment. [ITMX, ETMX] could have been centered around [-6.4, -27.2 urad] (instead of [-6.28, -28.12]). Not sure what to make out of this, as the dither servo was going on and off each one second.
4. A factor of 2 or 3 improvement will make ITM great, but ETM needs more (but note that the data was when it was bad).

Note that this is dependent on RF sideband frequency.

In preparation for acceptance review of the ETMX quadruple suspension, I took a set of ASD of the osems in the local and euler basis. ETMX is in chamber, under vacuum and the ISI was in its level 1 isolated state (ST1 = T100mHz blend on X and Y, T750mHz for the other DOF, ST2 = 750mHz blend). Measurement was taken successively with suspension damping on and off.

The main resonnances of main and reaction chains are well damped for all degrees of freedom.

The minor things to notice are :

Damping for the vertical DOF of the reaction chain is not having much effect on the first resonnance (@~ 0.56Hz) (see p9 of 2nd pdf).

The upper left osem of L2 level has its noise floor above the usual sensor noise we see on other osems (see p34 of 3rd pdf).

Otherwise everything else looks good.

The results attached are :

1) Comparison damping off/on for the top mass sensors of the main chain

2) Same for the reaction chain

3) Comparison between different quads (H1 ETMX H1 ETMY L1 ITMY L1 ITMX) of all osem sensors in local and euler basis.

A safe snapshot will need to be taken when possible since I found the normV/R/P filters disengaged, and a gain of 10 in the L2L L1 sensalign matrix. I also corrected the channel list of the plotquadspectra.m since it was requesting two times the L1_WIT_L channels. The modified script was commited to the svn

Written by Yuta, posted by Koji, while he is waiting for renewal of his ligo.org account:

In the entry alog #9381, Sheila explained how the PDH signal is distorted by the broadened resonance
of the higher-order modes due to low finesse of the cavity.

Here in this entry I explain how the shape of the PDH signal can be modified by changing the sideband frequency.

[Motivation]
ETMX transission for green was larger than designed(designed:5% -> measured:36%) and cavity length lock does not stay long.
The PDH signal looks strange(see alog #9381). To explain the situation and see how we can improve the PDH signal,
we calculated PDH signal including HOMs.

[Method]
1. Calculate PDH signal including the effect of carrier HOMs and sideband HOMs.
2. Change sideband frequency to see how PDH signal changes.

The parameters I used are the same as the ones listed in alog #9381.

Calculated TMS is 5.076kHz and the sideband frequency before we have changed last night was 24.407079MHz.
HOM content and vertical scale are arbitary in the following plots.

[Result]
1. HOMPDH_sb24_4MHz.png and HOMPDHIQ_sb24_4MHz.png show the calculated transmission, PDH signals,
and XY plot of I-phase and Q-phase PDH signals, when the sideband frequency was set to the original value. This
IQ plot is very similar to what we have seen(see video in alog #9381) and agrees well with Sheila's calculation.

2. HOMPDH_sbonres+0_5TMS.png and HOMPDHIQ_sbonres+0_5TMS.png are the plots when one of the sideband
is at the middle of the TEM00 and TEM01/10 resonances. PDH signal gives zero crossing at TEM00 resonances,
but it also gives zero crossing at other HOMs. So, mode hopping rate is expected to be high.

3. HOMPDH_sbonres+TMS.png and HOMPDHIQ_sbonres+TMS.png are the plots when one of the sideband is
at the TEM10/01 resonance. PDH signal does not give zero crossing at TEM00 resonances, but if the correct offset
is given, mode hopping rate should be low. The IQ plot will be somewhat simple "8" shaped plot in this case.

See alog #9379 for what we have done using these results.