and so it begins...until this is finished I have turned off the projector and moved the striptools to the big TV (time to look for Black Friday deals to replace the projector with a 120 inch LED TV)
Attachment #1 is a spectrogram of the BS sesimometer signals for a ~24 hour period (from Wednesday night to Thursday night local time, zipped because its a large file). I've marked the nearly pure tones that show up for some time and then turn off. We need to get to the bottom of this and ideally stop it from happening at night because it is eating ~1 hour of lockable time.
We considered if we could look at the phasing between the vertex and end seismometers to localize the source of the disturbance.
The DC port of the Bias-Tee is routed to (a modified version of) the iLIGO whitening board. This has the well-known problem of the protection diodes of the LT1125 quad-op-amp lowering the (ideally infinite) input impedance of the first gain stage (+24 dB). To be sure as to how much signal we can put into this port (in anticipation of trying some variable finesse PRFPMI locking but also for general book-keeping), I tested the usable input range by driving a triangle wave at ~3 Hz and changing the amplitude of the signal until we observed saturation. We found that we could drive a 10 Vpp signal at which point there was evidence of some clipping (it was asymmetric, the top end of the signal was getting clipped at +14,000 cts while the bottom end still looked like a triangle wave at -16,000 counts). Anyway we probably don't want to exceed +/- 10,000 counts on this channel. This is consistent with Hartmut's statement of having +/- 4V of usable range (although the counts he mentions are twice what I saw yesterday).
Other discussion points between Rana, Koji and Gautam:
The clue was that the loop X arm POX loop looked to have low (<3dB)) gain margin around 600 Hz (and again at 700 Hz). Attachment #1 shows a comparison of the OLTF for this loop (measured using the IN1/IN2 method) before and after our change. We hypothesize that one of the violin filters that were turned off had non-unity DC gain, because I had to lower the loop gain by 20% after these turn-offs to have the same UGF. I updated the snap files called by the arm locking scripts.
I think I caught all the places where the FM settings are saved, but some locking scripts may still try and turn on some of these filters, so let's keep an eye on it.
We turned off many excessive violin mode bandstop filters in the LSC.
The large ground motion at 1 Hz started up again tonight at around 23:30. I walked around the lab and nearby buildings with a flashlight and couldn't find anything whumping. The noise is very sinusoidal and seems like it must be a 1 Hz motor rather than any natural disturbance or traffic, etc. Suspect that it is a pump in the nearby CES building which is waking up and running to fill up some liquid level. Will check out in the morning.
Estimate of displacement noise based on the observed MC_F channel showing a 25 MHz peak-peak excursion for the laser:
dL = 25e6 * (13 m / (c / lambda)
= 1 micron
So this is a lot. Probably our pendulum is amplifying the ground motion by 10x, so I suspect a ground noise of ~0.1 micron peak-peak.
(this is a native PDF export using qtgrace rather than XMgrace. uninstall xmgrace and symlink to qtgrace.)
Due to some feedforward work by Jenne or EQ some years ago, we have had ~10 violin notches on in the LSC between the output matrix and the outputs to the SUS.
They were eating phase, computation time, and making ~3 dB gain peaking in places where we can't afford it. I have turned them off and Gautam SDF safed it.
Offensive BS shown in brown and cooler BS shown in blue.
To rotate the DTT landscape plot to not be sideways, use this command (note that the string is 1east, not least):
pdftk in.pdf cat 1east output out.pdf
at 1 Hz' this effect is not large so that's real translation. at lower frequencies a ground tilt couples to the horizontal sensors at first order and so the apparent signal is amplified by the double integral. drawing a free body diagram u can c that
x_apparent = (g / s^2) * theta
but for vortical this not tru because it already measures the full free fall and the tilt only shows up at 2nd order
Here is some disturbance in the spacetime curvature, where the local gradient of the metric seems to have been modulated (in the "downward" as well as in the other two orthogonal Cartesian directions) at ~1 Hz - seems real as far as I can tell, all the suspensions were being shaken about and all the seismometers witnessed it, though the peak is pretty narrow. A broader, less prominent peak also shows up around 0.5 Hz. We couldn't identify any clear source (no LN2 fill-up / obvious CES activity). This event lasted for ~45 mins, and stopped around 2315 local time. Shortly (~5min) after the ~1 Hz peak died down, however, the 3-10 Hz BLRMS channel reports an increase by ~factor of 2.
Onto trying some locking now that the suspensions have settled down somewhat.
this is due to the Equivalence Principle: local accelerations are indistinguishable from spacetime curvature. On a spherical Earth, the local gradient of the metric points in the direction towards the center of the Earth, which is colloquially known as "down".
I don't understand why the z-axis motion reported by the T240 is ~10x lower at 10 mHz compared to the X and Y motions. Is this some electronics noise artefact?
Attachment #1 shows the spectra of our three available seismometers over a period of ~10ksec.
Attachment #2 shows the result of applying frequency domain Wiener filter subtraction to the POP QPD (target) with the vertex seismometer signals as witness channels.
Today I realized that pip and other python2,3 packages were installed in the conda base environment, so after running
I could run the python-GPIB scripts to interface with the Agilent.
Although, I did have to add a python2 kernel to jupyter/ipython, which I did in a separate conda environment:
conda create -n ipykernel_py2 python=2 ipykernel
source activate ipykernel_py2
python -m ipykernel install --user
I've installed pyepics on Donatella running
sudo yum install pyepics
Pip and ipython did not seem to be installed yet.
- There was a SR560+SR785 (not connected for measurement) placed near the X end which I moved; it is now behind the electronics rack by the X arm beam tube (~15m away).
- Also, for the AM measurement I moved the AG5395A from behind the PSL setup to the X end, where it now is.
- By toggling the XGREEN shutter, I noticed that the cavity was not resonant before I disconnected anything from the setup since the spot shape kept changing, but I proceeded anyway.
- Because Rana said that it was important for me to mention: the ~5 USD blue-yellow crocs (that I now use) work fine for me.
1. The cables were calibrated with the DC block in the A port (for a A/R measurement)
2. The cable to the PZT was disconnected from the pomona box and connected to the RF out of the NA, the PD output labelled 'GREEN_REFL' was also disconnected and connected to the B port via a DC block.
3. The ITMX was 'misaligned'. (This allowed the reflected green PD output as seen on the oscilloscope to stabilize.)
4. The PZT is modulated in frequency and the residual amplitude modulation (as observed in the measured reflected green light) is plotted, ref. Attachment 1. The parameters for the plotted data in the attachment were:
# AG4395A Measurement - Timestamp: Nov 07 2019 - 17:04:07
#---------- Measurement Parameters ------------
# Start Frequency (Hz): 10000.0, 10000.0
# Stop Frequency (Hz): 10000000.0, 10000000.0
# Frequency Points: 801, 801
# Measurement Format: LOGM, PHAS
# Measuremed Input: AR, AR
#---------- Analyzer Settings ----------
# Number of Averages: 8
# Auto Bandwidth: On, On
# IF Bandwidth: 300.0, 300.0
# Input Attenuators (R,A,B): 0dB 10dB 20dB
# Excitation amplitude = -10.0dBm
Update (19:13 7thNov19): When the ITMX was intentionally misaligned, Rana and I checked to see if the Oplevs were turned off and they were. But while I was casually checking the Oplevs again, they were on!
Not sure what to do about this or what caused it.
- At the X end, we set up the network analyzer to begin measurement of the AM transfer function by actuation of the laser PZT.
- The lid of the PDH optics setup was removed to make some checks and then replaced.
- From the PDH servo electronics setup the 'GREEN_REFL' and 'TO AUX-X LASER PZT' cables were removed for the measurement and then re-attached after.
- The signal today was too low to make a real measurement of the AM transfer function, but the GPIB scripts and interfacing was tested.
Gautam and I were talking about some modulation and demodulation and wondered what is the power combining situation for the triple resonant EOM installed 8 years ago. And we noticed that the current setup has additional ~5dB loss associated with the 3-to-1 power combiner. (Figure a)
N-to-1 broadband power combiners have an intrinsic loss of 10 log10(N). You can think about a reciprocal process (power splitting) (Figure b). The 2W input coming to the 2-port power splitter gives us two 1W outputs. The opposite process is power combining as shown in Figure c. This case, the two identical signals are the constructively added in the combiner, but the output is not 20Vpk but 14Vpk. Considering thge linearity, when one of the port is terminated, the output is going to be a half. So we expect 27dBm output for a 30dBm input (Figure d). This fact is frequently oversight particularly when one combines the signals at multiple frequencies (Figrue e). We can avoid this kind of loss by using a frequency-dependent power combiner like a diplexer or a triplexer.
[Mirko / Kiwamu]
The resonant box has been installed together with a 3 dB attenuator.
The demodulation phase of the MC lock was readjusted and the MC is now happily locked.
We needed more modulation depth on each modulation frequency and so for the reason we installed the resonant box to amplify the signal levels.
Since the resonant box isn't impedance matched well, the box creates some amount of the RF reflections (#5339).
In order to reduce somewhat of the RF reflection we decided to put a 3 dB attenuator in between the generation box and the resonant box.
(what we did)
+ attached the resonant box directly to the EOM input with a short SMA connector.
+ put stacked black plates underneath the resonant box to support the wight of the box and to relief the strain on the cable between the EOM and the box.
+ put a 3 dB attenuator just after the RF power combiner to reduce RF reflections.
+ readjusted the demodulation phase of the MC lock.
(Adjustment of MC demodulation phase)
The demodulation phase was readjusted by adding more cable length in the local oscillator line.
After some iterations an additional cable length of about 30 cm was inserted to maximize the Q-phase signal.
So for the MC lock we are using the Q signal, which is the same as it had been before.
Before the installation of the resonant box, the amplitude of the MC PDH signal was measured in the demodulation board's monitor pins.
The amplitude was about 500 mV in peak-peak (see the attached pictures of the I-Q projection in an oscilloscope). Then after the installation the amplitude decreased to 400 mV in peak-peak.
Therefore the amplitude of the PDH signal decreased by 20 %, which is not as bad as I expected since the previous measurement indicated 40 % reduction (#2586).
Koji and I taked about cleaning up some of the flaky cable situation on the PSL table a while ago. The changes were implemented and are documented in Attachment #1. Now the Pomona box between the Thorlabs HV Driver and the NPRO head is sitting on the PSL table (sandwiched between some teflon pieces I found in cabinet S4 along the south arm), and the cables between these two devices are better strain relieved. I turned off the Thorlabs HV supply while working on the PMC table. The IMC could be locked after this work. Probably won't solve the long standing FSS mysteries but probably can't hurt.
Unrelated to this work: I also removed a Bias tee that was just hanging out on top of the FSS electronics, which was used for the modeSpec project.
Here is a comparison of the response of various DoFs in our various RFPD sensors for two different CARM offsets. Even in the case of the smaller CARM offset of ~1kHz, we are several linewidths away from the resonance. Need to do some finesse modeling to make any meaningful statement about this - why is the CARM response in REFL11 apparently smaller for the smaller CARM offset?
If you mistrust my signal processing, the GPS times for which I ran the sensing lines are:
CARM offset = ~30kHz (arm transmission <0.02) --- 1257064777+5min
CARM offset = ~1kHz (arm transmission ~5) --- 1257065566+5min
There seems to be stronger-than-expected coupling between CARM and the 3f sensors.
A coarse calibration of the CARM error point (when on ALS control) is 7.040 +/- 0.030 kHz/ct. This corresponds to approximately 0.95nm/ct. I typically lose the PRMI lock when the CARM offset is ~0.2 cts, which means I am about 1kHz away from the resonance. This is >10 CARM linewidths.
The calibration was done by sweeping the CARM offset (no PRM) and identifying the arm cavity FSRs by looking for peaks in TRX / TRY. Attachment #1 shows the scan, while Attachment #2 shows a linear fit to the FSRs. In Attachment #2, the frequency axis is taken from the phase tracker servo, which was calibrated by injecting a "known" frequency with the Marconi, and there is good agreement to the expected FSR with 37.79 m long arm cavities. There is much more info in the scan (e.g. modulation depths, mode matching to the arm cavities etc) which I will extract later, but if anyone wants the data (pre-downsampled by me to have a managable filesize), it's attached as a .zip file in Attachment #3.
Full analysis tomorrow, but I collected sensing matrix measurements with lines driven in PRCL,MICH and CARM at a couple of CARM offsets. I also wanted to calibrate the CARM offset to physical units so I ran some scans of the CARM offset and collected the data so I can use the arm cavity FSR to calibrate CARM. Koji suggested using REFL165_I for PRCL and REFL165_Q for MICH control - this would allow us to see if the problem was with the 1f sideband only. While the lock could be established, we still couldn't push the arm powers above 10 without breaking the PRMI lock. While changing the CARM offset, we saw a significant shift in the DC offset level of the out-of-loop REFL33_I signal. Need to think about what this means...
I removed the Trillium T240 DAQ interface unit from 1X4 for investigation.
It was returned to the electronics rack and all the connections were re-made. Some details:
Update 445pm: Seems to have done something good - the old feedforward filters reduce the YAW RMS motion by a factor of a few. Pitch performance is not so good, maybe the filter needs re-training, but I see coherence, see Attachment #2 for the frequency domain WF.
I am still unable to achieve arm powers greater than TRX/TRY ~10 while keeping the PRMI locked. A couple of times, I was able to get TRY ~50, but TRX stayed at ~10, or even dropped a little, suggestive of a DARM offset? On the positive side, the ALS system seems to work pretty reliably, and I can keep the arms controlled by ALS for several tens of minutes.
I created a spreadsheet (Attached) by taking Koji's c1psl sheet from slow_channel_list and filtering out the channels that do not need an Acromag. I added in the QPD channels that are relevant to the PSL from the c1iool0 sheet.
I began mapping the PSL related Eurocrates connectors to their respective VME channels starting with the PMC electronics.
I am confused about the TTFSS interface (D040423): While it is a Eurocrate card, in the schematics it seems to have 50 pin connectors.
I found old wiring schematics that might help with identifying the channels once the connector issue is clarified.
I did some re-alignment of the POP beam on the IX in air table. Here are the details:
Tangentially related to this work - I took the nuclear option and did a hard reboot of the c1susaux Acromag crate on Sunday to fix the EPICS issue - it seems to be gone for now, see Attachment #5.
There are many versions of the POP22 signal path I found on the elog, e.g. this thread. But what I saw at the LSC rack was not quite in agreement with any of those. So here is the latest greatest version.
Since the 2f signals are mainly indicators of power buildups and are used for triggering various PDH loops, I don't know how critical some of these things are, but here are some remarks:
Thanks. Please update this wiki page too.
I reproduced Gautam's sketch of the 1x1 and 1x2 Eurocrates into a pdf image that contains links to the appropriate DCCs in the legend (see attachement).
PMC got unlocked at ~4am. I re-locked it. Also tweaked the input pointing into the cavity. The misalignment was mostly in pitch.
There was also a loud buzzing in the control room due to the audio cable being improperly seated in the mixer. I re-seated it.
Sigh... hard loch
While I was trying to lock the PRMI this evening, I noticed that I couldn't move the REFL beamspot on the CCD field of view by adjusting the slow bias voltages to the PRM. Other suspensions controlled by c1susaux seem to respond okay so at first glance it isn't a problem with the Acromag. Looking at the OSEM sensor input levels, I noticed that UL is much lower than the others - see Attachment #1, seems to have happened ~100 days ago. I plugged the tester box in to check if the problem is with the electronics or if this is an actual shorting of some pins on the physical OSEM as we had in the past. So PRM watchdog is shutdown for now and there is no control of the optic available as the cables are detached. I will replace the connections later in the evening.
Since I couldn't find anything wrong, I plugged the suspension back in - and voila, the suspect UL PD voltage level came back to a level consistent with the others! See Attachment #2.
Anyway, I had some hours of data with the tester box plugged in - see Attachment #3 for a comparison of the shadow sensor readout with the tester box (all black traces) vs with the suspension plugged in, local damping loops active (coloured traces). The sensing noise re-injection will depend on the specifics of the local damping loop shapes but I suspect it will limit feedforward subtraction possibilities at low frequencies.
However, I continue to have problems aligning the optic using the slow bias sliders (but the fast ones work just fine) - problem seems to be EPICS related. In Attachment #4, I show that even though I change the soft PITCH bias voltage adjust channel for the PRM, the linked channels which control the actual voltages to the coils take several seconds to show any response, and do so asynchronously. I tried restarting the modbus process on c1susaux, but the problem persists. Perhaps it needs a reboot of the computer and/or the acromag chassis? I note that the same problem exists for the BS and PRM suspensions, but not for ITMX or ITMY (didn't check the IMC optics). Perhaps a particular Acromag DAC unit is faulty / has issues with the internal subnet?
The POP beam coming out of the vacuum chamber is split by a 50/50 BS and half is diverted to the POP22/POP110/POPDC photodiode (Thorlabs PDA10CF) and the other half goes to the POP QPD. This optical layout is still pretty accurate. I looked at the data of the POPDC and POP QPD SUM channels while the dither alignment was running, to see if I could figure out what's up with the weird correlated dip in REFLDC and POPDC. While the POPDC channel shows some degradation as the REFLDC level goes down (=alignment gets better), the QPD sum channel shows the expected light level increase. So it could yet be some weird clipping somewhere in the beampath - perhaps at the 50/50 BS? I will lock the PRMI (no arms) and check...
I changed the shape of the low pass filter to reduce high frequency sensor noise injection into the MICH control signal. The loop stability isn't adversely affected (lost ~5 degrees of phase margin but still have ~50 degrees), while the control signal RMS is reduced by ~x10. This test was done with the PRMI locked on the carrier, need to confirm that this works with the arms controlled on ALS and PRMI lcoked on sideband.
Tried a bunch of things tonight.
One possibility is that the arm buildup is exerting some torque on the ITMs, which can also change the PRC cavity axis - as the buildup increases, the dominant component of the circulating field in the PRC comes from the leakage from the overcoupled arm cavity. We used to DC couple the ITM Oplev servos when locking the PRMI. The TRX level of 1 corresponds to ~5W of circulating power in the arm cavity, and the static radiation pressure force due to this circulating power is ~30 nN, rising up to 300nN as the TRX level hits 10. So for 1mm offset of the spot position on the ITM, we'd still only exert 300 pN m of torque. I don't see any transient in the Oplev error signals when locking the arm cavity as usual with POX/POY, but on timescales of several seconds, the Oplev error point shows ~3-5 urad of variation.
I set up a photodiode (PDA10CF) in the IFO REFL beampath and the Agilent NA is sitting on the east side of the PSL enclosure. This was meant to be just a first look, maybe the PDA10CF isn't suitable for this measurement. The measurement condition was - PRM aligned so we have a REFL beam (DC level = 8.4V measured with High-Z). Both ITMs and ETMs were macroscopically misaligned so that there isn't any cavity effects to consider. I collected noise around 11 and 55 MHz, and also a dark measurement, plots to follow. The optics were re-aligned to the nominal config but I left the NA on the east side of the PSL enclosure for now, in anticipation of us maybe wanting to tune something while minimizing a peak.
Attachment #1: Results of a coarse sweep from 5 MHz to 100 MHz. The broadband RIN level is not resolvable above the dark noise of the photodiode, but the peaks at the modulation frequencies (11 MHz, 55 MHz and 29.5 MHz) are clearly visible. Not sure what is the peak at ~44 MHz or 66 MHz. Come to think of it, why is the 29.5 MHz peak so prominent? The IMC cavity pole is ~4kHz so shouldn't the 29.5 MHz be attenuated by 80dB in transmission through the cavity?
Attachment #2: Zoomed in spectra with finer IF bandwidth around the RF modualtion frequencies. From this first measurement, it seems like the RIN/rad level is ~10^5, which I vaguely remember from discussions being the level which is best achieved in practise in the 40m in the past.
Check the RAM due to the EOM? Perhaps the pointing / polarization control into the EOM got degraded.
I looked at some signals for a 10 second period when the PRMI was locked with at some CARM offset, just before the PRMI lost lock, to see if there are any clues. I don't see any obvious signatures in this set of signals - if anything, the PRM is picking up some pitch offset, this is seen both at the Oplev error point and also in the POP QPD spot position. But why should this be happening as I reduce the CARM offset? The arm transmission is only ~5, so it's hard to imagine that the radiation pressure is somehow torquing the PRM. There are no angular feedback loops actuating on the PRM in this state except the local damping and Oplev loops.
The 1f signals are also changing their mean DC offset values, which may be a signature of a changing offset in the 3f MICH and PRCL error points? The MICH error signal is pretty noisy (maybe I can turn on some LPF to clean this up a bit), but I don't see any DC drift in the PRCL control signal.
I ran some sensing measurements at various CARM offsets to check if the PRCL-->REFL33 and MICH-->REFL165 signals were being rotated out of the sensing quadrature as I lowered the CARM offset - there was no evidence of this happening. See Attachment #2. Other possibilities:
The IMC went into some crazy state so I'm calling it for the night, need to think about what could be happening and take a closer look at more signals during the CARM offset reduction period for some clues...
In calculating the above numbers, I assumed a DC transimpedance of 10 khhms and an RF Transimpedance of ~800 V/A.
[Elog14480]: per these calculations, with the NewFocus 1611 PDs, we cannot achieve shot noise limited sensing for any power below the rated maximum for linear operation (i.e. 1mW). Moreover, the noise figure of the RF amplifier we use to amplify the sensed beat note before driving the delay-line frequency discriminator is unlikely to be the limiting noise source in the current configuration. Rana suggested that we get two Gain Blocks. These can handle input powers up to ~10dBm while still giving us plenty of power to drive the delay line. This way, we can (i) not compromise on the sacred optical gain, (ii) be well below the 1dB compression point (i.e. avoid nonlinear noise effects) and (iii) achieve a better frequency discriminant.
Temporary fix: While the gain blocks arrive, I inserted a 10dB (3dB) attenuator between the PSL+EX (PSL+EY) photodiode RF output and the ZHL-3A amplifiers. This way, we are well below the 1dB compression point of said RF amplifiers, and also below the 1dB compression point of the on-board Teledyne AP1053 amplifiers on the demodulator boards we use.
Nest steps: Rana is getting in touch with Rich Abbott to find out if there is any data available on the noise performance of the post-mixer IF amplifier stage in the 0.1 -30 Hz range, where the voltage and current noise of the AD829 OpAmps could be limiting the DFD performance. But in the meantime, the ALS noise seems good again, and there is no evidence of the sort of CARM/DARM coupling that motivated this investigation in the first place. Managed to execute several IR-->ALS transitions tonight in the PRFPMI locking efforts (next elog).
No new Teledyne AP1053s were harmed in this process - I'll send the 5 units back to Rich tomorrow.
back on new Rossa from Xi computing
Update: Sun Nov 3 18:08:48 2019
Update: Fri Nov 15 00:00:26 2019:
We think we got to the bottom of this issue today. The RF signal level going into the demod board is too high. This electronics chain needs some careful gain reallocation.
I was demonstrating to Koji a strange feature I had noticed in the ALS control, whereby when applying a CARM offset to detune the arms, the two arms seemed to respond differently (based on the transmission levels). This kind of CARM-->DARM coupling seemed strange to me. Anyway, I also noticed that the EPICS indicators on the ALS MEDM screen suggested ADC saturations were going on. I had never really looked at the fast time series of the inputs to the phase tracker servos, but these showed saturating behavior on ndscope traces. I went to the LSC rack and measured these on a scope, indeed, they were ~20V pp.
The output of the BeatMouth PDs are going to a ZHL-3A amplifier - we should consider replacing these with lower gain amplifiers, e.g. the Teledyne AP1053. This is relegated to a daytime task.
Other findings tonight:
While working on the PSL table, I somehow put the IMC FSS into a bad state, reminiscent of this behavior. Seems like this is linked to some flaky connection on the PSL table. One candidate is the unstable attachment of the Pomona box between the NPRO PZT and the FSS output - we should install a short BNC cable between these to avoid the lever arm situation we have right now.
The Trillium T240 seismometer needs mass re-centering. Has anyone done this before, and do we have any hardware to do this?
I went to the Trillium interface box in 1X5. In this elog, Koji says it is D1000749-v2. But looking at the connector footprint on the back panel, it is more consistent with the v1 layout. Anyway I didn't open it to check. Main point is that none of the backplane data I/O ports are used. We are digitizing (using the fast CDS system) the front panel BNC outputs for the three axes. So of the various connectors available on the interface box, we are only using the front panel DB25, the front panel BNCs, and the rear panel power.
The cable connecting this interface box to the actual seismometer is a custom one I believe. It has a 19 pin military circular type hermetic connector on one end, and a DB25 on the other. Power is supplied to the seismometer from the interface box via this cable, so in order to run the test, I had to use a DB25 breakout board to act as a feedthrough and peek at the signals while the seismometer and interface boards were connected. I used Jenne's mapping of the DB25--> 19 pin connector (which also seems consistent with the schematic). Findings:
I am holding off on attempting any re-centering, for more experienced people to comment.
I'd like to revive the PRC angular feedforward system. However, it looks like the coherence between the vertex seismometer channels and the PRC angular motion witness sensor (= POP QPD) is much lower than was found in the past, and hence, the stabilization potential by implementing feedforward seems limited, especially for the Pitch DoF.
I found that the old filters don't work at all - turning on the FF just increases the angular motion, I can see both the POP and REFL spots moving around a lot more on the CRT monitors.
I first thought I'd look at the frequency-domain weiner filter subtraction to get a lower bound on how much subtraction is possible. I collected ~25 minutes of data with the PRC locked with the carrier resonant (but no arm cavities). Attachment #1 shows the result of the frequency domain subtraction (the dashed lines in the top subplot are RMS). Signal processing details:
The coherence between target signal (=POP QPD) and the witness channels (=seismometer channels) are much lower now than was found in the past. What could be going on here?
I wanted to restart the c1oaf model. As usual, the first time the model was restarted, it came back online with a 0x2bad error. This isn't even listed in the diagnostics manual as one of the recognized error states (unless there is a typo and they mean 0x2bad when they say 0xbad). The fix that has worked for me is to stop and start the model again, but of course, there is some chance of taking all the vertex FEs down in the process. No permutation of mxstream and daqd process restarts have cleared this error. We need some CDS/RCG support to look into this issue and fix it, it is not reasonable to go through reboots of all the vertex FEs every time we want to make a model change.
I looked into the seismometer situation a bit more today. Here is the story so far - I think more investigation is required:
Attachment #2 has some spectrograms (they are rather large files). They suggest that the increase in noise in the 0.1-0.3 Hz band in the BS seismometer X channel is real - but there isn't a corresponding increase in the other two seismometers, so the problem could still be electronics related.
Yesterday, Koji and I noticed (from the wall StripTool traces) that the vertex seismometer RMS between 0.1-0.3 Hz in the X-direction increased abruptly around 6pm PDT. This morning, when I came in, I noticed that the level had settled back to the normal level. Trending the BLRMS channels over the last 24 hours, I see that the 0.3-1 Hz band in the Z direction shows some anomalous behaviour almost in the exact same time-band. Hard to believe that any physical noise was so well aligned to the seismometer axes, I'm inclined to think this is indicative of some electronics issues with the Trillium interface unit, which has been known to be flaky in the past.
Attachment #1 - comparison of the POP QPD PIT and YAW output signal spectra with and without them being normalized by the SUM channel. I guess the shape is different between 30-100 Hz because we have subtracted out the correlated singal due to RIN?
This did not have the effect I desired - I was hoping that by normalizing the signals, I wouldn't need to change the gain of the ASC servo as the buildup in the PRC changed, but I found that the settings that worked well for PRMI locked with the carrier resonant (no arm cavities, see Attachment #2, buildup RIN reduced by a factor of ~4) did not work for the PRMI locked with the sideband resonant. Moreover, Koji raised the point that there will be some point in the transition from arms off resonance to on resonance where the dominant field in the PRC will change from being the circulating PRC carrier to the leaking arm carrier. So the response of the actuator (PRM) to correct for the misalignment may change sign.
In conclusion, we decided that the best approach to improve the angular stability of the PRC will be to revive the PRC angualr feedforward, which in turn requires the characterization and repair of the apparently faulty vertex seismometer.
The closest thing I can think of is here.
Is there a loop model of green PDH that agrees with the measurement? I'm wondering if something can be done with a compensation network to up the bandwidth or if the phase lag is more like a non-invertible kind.
During our EX AM/PM setups, I don't think we bumped the PDH gain knob (and I hope that the knob was locked). Possible drift in the PZT response? Good thing Shruti is on the case.
I made a change to the c1ass model to normalize the PIT and YAW POP QPD outputs by the SUM channel. A saturation block is used to prevent divide-by-zero errors, I set the saturation limits to [1,1e5], since the SUM channel is being recorded as counts right now. Model change is shown in the attached screenshots. I compiled and installed the model. Ran the reboot script to reboot all the vertex FEs to avoid the issue of crashing c1lsc.
I tried implementing a basic PRMI ASC using the POP QPD as a sensor. The POP22 buildup RMS is reduced by a factor of a few. This is just a first attempt, I think the loop shape can be made much better, but the stability of the lock is already pretty impressive. For some past work, see here.
The EX PDH setup had what I thought was insufficient phase and gain margins. So I lowered the gain a little - the price paid was that the suppression of laser frequency noise of the end laser was reduced. I actually think an intermediate gain setting (G=7) can give us ~35 degrees of phase margin, ~10dB gain margin, and lower residual unsuppressed AUX laser noise - to be confirmed by measurement later. See here for the last activity I did - how did the gain get increased? I can't find anything in the elog.
Looking at the sensor and oplev trends over the weekend, there was only one event where the optic seems to have been macroscopically misaligned, at ~11:05:00 UTC on Oct 19 (early Saturday morning PDT). I attach a plot of the 2kHz time series data that has the mean value subtracted and a 0.6-1.2 Hz notch filter applied to remove the pendulum motion for better visualization. The y-axis calibration for the top plot assumes 1 ct ~= 1 um. This "glitch" seems to have a timescale of a few seconds, which is consistent with what we see on the CCD monitors when the cavity is locked - the alignment drifts away over a few seconds.
As usual, this tells us nothing conclusive. Anyways, I am re-enabling the watchdog and pushing on with locking activity and hope the suspension cooperates.
The satellite box was reconnected and the suspension was left with watchdog off but OSEM roughly centered. We will watch for glitches over the weekend.