After diagnosis with the tester box, as I suspected, the fully open DC voltages on the two problematic channels, LL and UR, were restored once I replaced the LM6321 ICs in those two channel paths. However, I've been puzzled by the inability to turn on the Oplev loops on ETMY. Furthermore, the DC bias voltages required to get ETMY to line up with the cavity axis seemed excessively large, particularly since we seemed to have improved the table levelling.
I suspected that the problem with the OSEMs hasn't been fully resolved, so on Thursday night, I turned off the ETMY watchdog, kicked the optic, and let it ringdown. Then I looked at the time-series (Attachment #1) and spectra (Attachment #2) of the ringdowns. Clearly, the LL channel seems to saturate at the lower end at ~440 counts. Moreover, in the time domain, it looks like the other channels see the ringdown cleanly, but I don't see the various suspension eigenmodes in any of the sensor signals. I confirmed that all the magnets are still attached to the optic, and that the EQ stops are well clear of the optic, so I'm inclined to think that this behavior is due to an electrical fault rather than a mechanical one.
For now, I'll start by repeating the ringdown with a switched out Satellite Box (SRM) and see if that fixes the problem.
While restoring OSEMs on ETMY, I noticed that the open voltages for the UR and LL OSEMs had significantly (>30%) changed from their values from ~2 years ago. The fact that it only occurred in 2 coils seemed to rule out gradual wear and tear, so I looked up the trends from Nov 25 - Nov 28 (Sundance visited on Nov 26 which is when we removed the cage). Not surprisingly, these are the exact two OSEMs that show a decrease in sensor voltage when the OSEMs were pulled out. I suspect that when I placed them in their little Al foil boats, I shorted out some contacts on the rear (this is reminiscent of the problem we had on PRM in 2016). I hope the problem is with the current buffer IC in the satellite box and not the physical diode, I'll test with the tester box and evaluate the problem further.
Short update on latest Satellite box woes.
What's more - I did some Sat box switcheroo, swapping the SRM and ETM boxes back and forth in combination with the tester box. In the process, I seem to have broken the SRM sat box - all the shadow sensors are reporting close to 0 volts, and this was confirmed to be an electronic problem as opposed to some magnet skullduggery using the tester box. Once we get to the bottom of the ETMY sat box, we will look at SRM. This is more or less the last thing to look at for this vent - once we are happy the cavity axis can be recovered reliably, we can freeze the position of the elliptical reflector and begin the F.C.ing.
The N2 ran out this weekend (again no reminder email, but I haven't found the time to setup the Python mailer yet). So all the valves Steve and I had opened, closed (rightly so, that's what the interlocks are supposed to do). Chub will post an elog about the new N2 valve setup in the Drill-press room, but we now have sufficient line pressure in the N2 line again. So Chub and I re-opened the valves to keep pumping on the RGA.
While Chub is making new cables for the EY satellite box...
While the position of the reflector could possibly be optimized further, since we are already seeing a temperature gradient on the optic, I propose pushing on with other vent activities. I'm almost certain the current positioning places the optic closer to the second focus, and we already saw shifts of the HOM resonances with the old configuration, so I'd say we run with this and revisit if needed.
If Chub gives the Sat. Box the green flag, we will work on F.C.ing the mirrors in the evening, with the aim of closing up tomorrow/Friday.
All raw images in this elog have been uploaded to the 40m google photos.
For the last week, I noticed that I was unable to turn the EY chamber illuminator on using the remote python scripts. This was turning out to be really annoying, having to turn the light on/off manually. Today, I looked into the problem and found that there is a conflict in the IP addresses of the EY Ethernet Strip (which Chas assigned a static IP but did not include detailed procedures for) and the vertex area laptop, paola. The failure of the python control of the power strip coincided exactly with when Chub and I turned on paola for working at the IY chamber - but how was I supposed to know these events are correlated? I tried shutting down paola , power cycling the Ethernet power strip, and restarting the bind9 services on chiara, but remote control of the ethernet power strip remains elusive. I suspect reconfiguring the static IP for the Ethernet switch will require some serial port enabled device...
In preparation for the FC cleaning, I did the following:
Tomorrow, I will start with the cleaning of ETMY HR. While the FC is drying, I will position ITMY at the edge of the IY cable for cleaning (Chub will setup the mini-cleanroom at the IY table). The plan is to clean both HR surfaces and have the optics back in place by tomorrow evening. By my count, we have done everything listed in the IY and EY chambers. I'd like to minimize the time between cleaning and pumpdown, so if all goes well (Sat Box problems notwithstanding), we will check the table leveling on Friday morning, and put on the heavy doors and at least rough the main volume down to 1 torr on Friday.
The attached photo shows the two optics with FC applied.
My original plan was to attempt to close up tomorrow. However, we are still struggling with Satellite box issues. So rather than rush it, we will attempt to recover the Y arm cavity alignment on Monday, satellite box permitting. The main motivation is to reduce the deadtime between peeling off the F.C and starting the pumpdown. We will start working on recovering the cavity alignment once the Sat box issues are solved.
I had taken Satellite box S/N 102, from the SRM suspension, down to the Y-end as part of debugging. However, at some point, I stopped getting readbacks from the shadow sensor PDs, even with the Sat. Box tester hooked up (so as to rule out anything funky with the actual OSEMs). Today evening, I did a more systematic investigation. Schematic with component references is here.
The question remains as to what caused this failure mode - I can't think of why that particular IC was damaged during the Satellite box swapping process - is this indicative of some problem elsewhere in the ETMY OSEM/coil driver electronics chain?
To avoid the annoying excercise of having to manually toggle the illuminators, I solved the IP conflict. Made a wiki page for the ethernet power strips since the documentation was woeful (the way the power strips are mounted in the racks, you can't even see the manufacturer/model/make). All chamber illuminators can now be turned on/off by the MEDM scripts . Note that there is a web interface available too, which can be useful in case of some python socket issues. The main lesson is: avoid using the "reset" button on the power strips, it destroys the static IP config.
Unrelated to this work: The EY laptop, asia, won't boot up anymore, with a "Fan Error" message being the red flag. I've temporarily recommissioned the vacuum rack laptop, belladonna, to be the EY machine for this vent. Can we get 3 netbooks that actually work and don't need to be tethered to a power strip for the VEA?
I reset the remote of this git repo to the 40m version instead of Jon's personal one, to ensure consistency between what's on the vacuum machine and in the git repo. There is now a N2 checker python mailer that will email the 40m list if all the tank pressures are below 600 PSI (>12 hours left for someone to react before the main N2 line pressure drops and the interlocks kick in). For now, the script just runs as a cron job every 3 hours, but perhaps we should integrate it with the interlock process?
All the python code running on c1vac is archived to the git repo:
[chub, koji, gautam]
Attachment #1 shows the signal routing near the Satellite box. Somehow, the female 64 pin IDC connector that brings the signals from the coil driver board wasn't mating well with the mail connector on the Satellite box front panel. This is a connector specific problem - plugging the female end into one of the male connectors inside the Satellite box yielded signal continuity. The problem was resolved by re-making both connections -by driving the EPICS bias slider through its full range, we were able to see the full voltage swing at the DB connectors going to the flange
This kind of flakiness could be all around the lab, and could be responsible for many of the suspension "mysteries". To re-iterate, the problem seems to be the way the female sockets of the connector mates with the male pins - while the actual crimping points may look secure, there may not be signal continuity.
Now that this problem is resolved, tomorrow we will recover the cavity alignment and possibly start a pumpdown.
Unrelated to this work - the spare satellite box (S/N #100), which had a note on it that said "low voltages", was tested. The "low voltages" referred to the OSEM shadow sensor voltages being low when the LED was completely unobscured. The reason was that the mod to increase the drive current to 25 mA had not yet been implemented on this unit. I added the appropriate 806 ohm resistors, and verified that the voltages were correct, so now we have a working spare. It is stored in the "photodiode" cabinet along the east arm, together with the tester boxes.
Since we may want to close up tomorrow, I did the following prep work:
Rather than try and rush and close up tomorrow, I propose spending the day tomorrow cleaning the peripheral areas of the optic, suspension cage, and chamber. Then on Thursday morning, we can replace the Y-arm optics, try and recover the cavity alignment, and then aim for a Thursday afternoon pumpdown. The main motivation is to reduce the time the optics spend in air after F.C. peeling and going to vacuum.
Procedure tomorrow [comments / suggestions welcome]:
All photos have been uploaded to google photos.
Squishing cables at the ITMX satellite box seems to have fixed the wandering ITM that I observed yesterday - the sooner we are rid of these evil connectors the better.
I had changed the input pointing of the green injection from EX to mark a "good" alignment of the cavity axis, so I used the green beam to try and recover the X arm alignment. After some tweaking of the ITM and ETM angle bias voltages, I was able to get good GTRX values [Attachment #1], and also see clear evidence of (admittedly weak) IR resonances in TRX [Attachment #2]. I can't see the reflection from ITMX on the AS camera, but I suspect this is because the ITMY cage is in the way. This will likely have to be redone tomorrow after setting the input pointing for the Y arm cavity axis, but hopefully things will converge faster and we can close up sooner. Closing the PSL shutter for now...
I also rebooted the unresponsive c1susaux to facilitate the alignment work tomorrow.
[koji, chub, jon, rana, gautam]
Full story tomorrow, but we went through most of the required pre close-up checks/tasks (i.e. both arms were locked, PRC and SRC cavity flashes were observed). Tomorrow, it remains to
The ETMY suspension chain needs to be re-characterized (neither the old settings, nor a +/- 1 gain setting worked well for us tonight), but this can be done once we are back under vacuum.
[chub, bob, gautam]
[Attachment #1]: ITMY HR face after cleaning. I determined this to be sufficiently clean and re-installed the optic.
[Attachment #2]: ETMY HR face after cleaning. This is what the HR face looks like after 3 rounds of First-Contact application. After the first round, we noticed some arc-shaped lines near the center of the optic's clear aperture. We were worried this was a scratch, but we now believe it to be First-Contact residue, because we were able to remove it after drag wiping with acetone and isopropanol. However, we mistrust the quality of the solvents used - they are not any special dehydrated kind, and we are looking into acquiring some dehydrated solvents for future cleaning efforts.
[Attachment #3]: Top view of ETMY cage meant to show increased clearance between the IFO axis and the elliptical reflector.
Many more photos (including table leveling checks) on the google-photos page for this vent. The estimated time between F.C. peeling and pumpdown is ~24 hours for ITMY and ~15 hours for ETMY, but for the former, the heavy doors were put on ~1 hour after the peeling.
The first task is to fix the damping of ETMY.
[jon, koji, gautam]
I'm leaving all suspension watchdogs tripped over the weekend as part of the suspension diagonalization campaign...
I looked into this a bit today. Did a walkthrough of the lab, didn't hear any obvious hissing (makes sense, that presumably would signal a much larger leak rate).
Attachment #1: Data from the 30 ksec we had the main vol valved off on Jan 10, but from the gauges we have running right now (the CC gauges have not had their HV enabled yet so we don't have that readback).
Attachment #2: Data from ~150 ksec from Friday night till now.
Interpretation: The number quoted from Jan 10 is from the cold-cathode gauge (~20 utorr increase). In the same period, the Pirani gauge reports a increase of ~5 mtorr (=250x the number reported by the cold-cathode gauge). So which gauge do we trust in this regime more? Additionally, the rate at which the annuli pressures are increasing seem consistent between Jan 10 and now, at ~100 mtorr every 30 ksec.
I don't think this is conclusive, but at least the leak rates between Jan 10 and now don't seem that different for the annuli pressures. Moreover, for the Jan 10 pumpdown, we had the IFO at low pressure for several days over the chirstmas break, which presumably gave time for some outgassing which was cleaned up by the TPs on Jan 10, whereas for this current pumpdown, we don't have that luxury.
Do we want to do a systematic leak check before resuming the pumpdown on Monday? The main differences in vacuum I can think of are
This entry by Steve says that the "expected" outgassing rate is 3-5 mtorr per day, which doesn't match either the current observation or that from Jan 10.
As planned, we valved off the main volume and the annuli from the turbo-pumps at ~730 PM PST. At this time, the main volume pressure was 30 uTorr. It started rising at a rate of ~200 uTorr/hr, which translates to ~5 mtorr/day, which is in the ballpark of what Steve said is "normal". However, the calibration of the Hornet gauge seems to be piecewise-linear (see Attachment #1), so we will have to observe overnight to get a better handle on this number.
We decided to vent the IY and EY chamber annular volumes, and check if this made anu dramatic changes in the main volume pressure increase rate, presumably signalling a leak from the outside. However, we saw no such increase - so right now, the working hypothesis is still that the main volume pressure increase is being driven by outgassing of something from the vacuum.
Let's leave things in this state overnight - V1 and V5 closed so that neither the main volume nor the annuli are being pumped, and get some baseline numbers for what the outgassing rate is.
I looked at the free-swinging sensor data from two nights ago, and am struggling with the interpretation.
[Attachment #1] - Fine resolution spectral densities of the 5 shadow sensor signals (y-axis assumes 1ct ~1um). The puzzling feature is that there are only 3 resonant peaks visible around the 1 Hz region, whereas we would expect 4 (PIT, YAW, POS and SIDE). afaik, Lydia looked into the ETMY suspension diagonalization last, in 2016. Compared to her plots (which are in the Euler basis while mine are in the OSEM basis), the ~0.73 Hz peak is nowhere to be seen. I also think the frequency resolution (<1 mHz) is good enough to be able to resolve two closely spaced peaks, so it looks like due to some reason (mechanical or otherwise), there are only 3 independent modes being sensed around 1 Hz.
[Attachment #2] - Koji arrived and we looked at some transfer functions to see if we could make sense of all this. During this investigation, we also think that the UL coil actuator electronics chain has some problem. This test was done by driving the individual coils and looking for the 1/f^2 pendulum transfer function shape in the Oplev error signals. The ~ 4dB difference between UR/LL and LR is due to a gain imbalance in the coil output filter bank, once we have solved the other problems, we can reset the individual coil balancing using this measurement technique.
[Attachment #3] - Downsampled time-series of the data used to make Attachment #1. The ringdown looks pretty clean, I don't see any evidence of any stuck magnets looking at these signals. The X-axis is in kilo-seconds.
We found that the POS and SIDE local damping loops do not result in instability building up. So one option is to use only Oplevs for angular control, while using shadow-sensor damping for POS and SIDE.
I guess we forgot to close V5, so we were indeed pumping on the ITMY and ETMY annuli, but the other three were isolated suggest a leak rate of ~200-300 mtorr/day, see Attachment #1 (consistent with my earlier post).
As for the main volume - according to CC1, the pressure saturates at ~250 uTorr and is stable, while the Pirani P1a reports ~100x that pressure. I guess the cold-cathode gauge is supposed to be more accurate at low pressures, but how well do we believe the calibration on either gauge? Either ways, based on last night's test (see Attachment #2), we can set an upper limit of 12 mtorr/day. This is 2-3x the number Steve said is normal, but perhaps this is down to the fact that the outgassing from the main volume is higher immediately after a vent and in-chamber work. It is also 5x lower rate of pressure increase than what was observed on Feb 2.
I am resuming the pumping down with the turbo-pumps, let's see how long we take to get down to the nominal operating pressure of 8e-6 torr, it ususally takes ~ 1 week. V1, VASV, VASE and VABS were opened at 1030am PST. Per Chub's request (see #14435), I ran RP1 and RP3 for ~30 seconds, he will check if the oil level has changed.
Pumpdown looks healthy, so I'm leaving the TPs on overnight. At some point, we should probably get the RGA going again. I don't know that we have a "reference" RGA trace that we can compare the scan to, should check with Steve. The high power (1 W) beam has not yet been sent into the vacuum, we should probably add the interlock condition that shuts off the PSL shutter before that.
[chub, steve, gautam]
Steve came by the lab today. He advised us to turn the RGA on again, now that the main volume pressure is < 20 uTorr. I did this by running the RGAset.py script on c0rga - the temperature of the unit was 22C in the morning, after ~3 hours of the filament being turned on, the temperature has already risen to 34 C. Steve says this is normal. We also opened VM1 (I had to edit the interlocks.yaml to allow VM1 to open when CC1 < 20uTorr instead of 10uTorr), so that the RGA volume is exposed to the main volume. So the nightly scans should run now, Steve suggests ignoring the first few while the pumpdown is still reaching nominal pressure. Note that we probably want to migrate all the RGA stuff to the new c1vac machine.
Other notes from Steve:
The full 1 W is again being sent into the IMC. We have left the PBS+HWP combo installed as Rana pointed out that it is good to have polarization control after the PMC but before the EOM. The G&H mirror setup used to route a pickoff of the post-EOM beam along the east edge of the PSL table to the AUX laser beat setup was deemed too flaky and has been bypassed. Centering on the steering mirror and subsequently the IMC REFL photodiode was done using an IR viewer - this technique allows one to geometrically center the beam on the steering mirror and PD, to the resolution of the eye, whereas the voltage maximization technique using the monitor port and an o'scope doesn't allow the former. Nominal IMC transmission of ~15,000 counts has been recovered, and the IMC REFL level is also around 0.12, consistent with the pre-vent levels.
I did some tests of the electronics chain today.
Hypothesising a bad connection between the sat box output J1 and the flange connection cable. Indeed, measuring the OSEM inductance from the DSUB end at the coil-driver board, the UL coil pins showed no inductance reading on the LCR meter, whereas the other 4 coils showed numbers between 3.2-3.3 mH. Suspecting the satellite box, I swapped it out for the spare (S/N 100). This seemed to do the trick, all 5 coil channels read out ~3.3 mH on the LCR meter when measured from the Coil driver board end. What's more, the damping behavior seemed more predictable - in fact, Rana found that all the loops were heavily overdamped. For our suspensions, I guess we want the damping to be critically damped - overdamping imparts excess displacement noise to the optic, while underdamping doesn't work either - in past elogs, I've seen a directive to aim for Q~5 for the pendulum resonances, so when someone does a systematic investigation of the suspensions, this will be something to look out for.. These flaky connectors are proving pretty troublesome, let's start testing out some prototype new Sat Boxes with a better connector solution - I think it's equally important to have a properly thought out monitoring connector scheme, so that we don't have to frequently plug-unplug connectors in the main electronics chain, which may lead to wear and tear.
The input and output matrices were reset to their "naive" values - unfortunately, two eigenmodes still seem to be degenerate to within 1 mHz, as can be seen from the below spectra (Attachment #1). Next step is to identify which modes these peaks actually correspond to, but if I can lock the arm cavities in a stable way and run the dither alignment, I may prioritize measurement of the loss. At least all the coils show the expected 1/f**2 response at the Oplev error point now. The coil output filter gains varied by ~ factor of 2 among the 4 coils, but after balancing the gains, they show identical responses in the Oplev - Attachment #2.
They have been stored on the 3rd shelf from top in the clean optics cabinet at the south end. EX
5 PR3/SR3 optics from FiveNine Optics were delivered. The data sheets were scanned and uploaded to the following wiki page. https://wiki-40m.ligo.caltech.edu/Aux_Optics
As it turns out, now ITMY has a tendency to get stuck. I found it MUCH more difficult to release the optic using the bias jiggling technique, it took me ~ 2 hours. Best to avoid c1susaux reboots, and if it has to be done, take precautions that were listed for ITMX - better yet, let's swap out the new Acromag chassis ASAP. I will do the arm locking tests tomorrow.
[Attachment #1]: Computed spectral power transmissivity (according to my model) for the coating design at a few angles of incidence. Behavior lines up well with what FNO measured, although I get a transmission that is slightly lower than measured at 45 degrees. I suspect this is because of slight changes in the dispersion relation assumed and what was used for the coating in reality.
[Attachment #2]: Similar information as Attachment #1, but with the angle of incidence as the independent parameter in a continuous sweep.
Conclusion: The coating behaves in a way that is in reasonable agreement with our model. At 41.1 degrees, which is what the PR3 angle of incidence will be, T<50 ppm, which was what we specified. The larger range of angles was included because originally, we thought of using this optic as a substitute for SR3 as well. But I claim that for the shorter SRC (signal recycling as opposed to RSE), we will not end up using the new optic, but rather go for the G&H mirror. In any case, as Koji pointed out, ~50 ppm extra loss in the RC will not severely limit the recycling gain. Such large variation was not seen in the MC analysis because we only varied the angle of incidence by +/- 0.5 degrees about the nominal design value of 41.1 degrees.
Several housekeeping tasks were carried out today in preparation for the Y-arm loss measurement.
Rich came by the 40m to photocopy some pages from Hobbs, and saw me working on the 60 Hz hunting. As I suspected, the problem was being generated in the D040060. This board receives the photodiode signal single-ended, but has a different power ground than the photodiode (even though the PD is plugged into a power strip that claims to come from 1Y4). The mechanism is not entirely clear - the presence of these 60 Hz features seemed to be dependent on the light level on the TRY photodiode (i.e. they were absent when the PSL shutter is closed, and were more prominent when TRY was 0.9 rather than 0.5) but the PD certainly wasn't saturated - the DC signal was only ~100 mV when viewed on a scope. In any case, Rich suggested the simplest test would be to ground the BNC shield bringing TRY to the rack, to the local ground on the board, which I did using a crocodile clip. This did the trick, the TRY signal RMS is now dominated by the ~1 Hz seismic-driven variation.
On a more pessimistic note - it looks like the elliptical reflector moving did not work, and the clipping in the Y arm persists . I am able to recover TRY~1 with the yaw offset on the ETM (which is still lower than the 1.06-1.07 Koji reported in Aug 2018, but I can believe that being down to the MC transmission being a few % lower at 15000cts rather than 15500), while the maximum I see without it is ~0.9. This is puzzling, because when the chamber was open, we saw that there was ~1.5" clearance between the edge of the reflector and the beam on an IR card. I suppose the input pointing could have been off by a small amount. So one of the primary vent objectives wasn't acheieved... But I will push ahead with the loss measurement.
Since we changed the HeNe, I updated the calibration factors, and accepted the changes in the SDF.
To measure the Y-arm loss, I decided to start with the classic reflectivity method. To prepare for this measurement, I did the following:
I'm running a measurement tonight, starting now (~1130PM), should be done in ~1hour, may need to do more data-quality improvements to get a realistic loss number, but I figured I'd give this a whirl.
I'm rather pleased with an initial look at the first align/misalign cycle, at least there is discernable contrast between the two states - Attachment #2. The data is normalized by MC transmission, and then sig.decimated by x512 (8**3).
Another arm loss measurement started at 6pm.
There are still several data quality issues that can be improved. I think there is little point in reading too much into this until some of the problems outlined below are fixed and we get a better measurement.
As an interim fix, I'm going to try and use the Oplevs as a DC reference, and run the dither alignment from zero each time, as this prevents the runaway problem at least. Data run started at 11:20 pm.
Attachment #1 shows estimated systematic uncertainty contributions due to
In all the measurements so far, the ratio seems to be < 1, so this would seem to set a lower bound on the loss of ~35 ppm. The dominant source of systematic uncertainty is the 5% assumed fudge in the mode-matching
Bottom line: I think we need to have other measurements and simultaenously analyse the data to get a more precise estimate of the loss.
One of the XT1111 units (XT1111a) in the new vacuum system has malfunctioned. So all valves are closed, PSL shutter is also closed, until this is resolved.
Pressure of the main volume seems to have stabilized - see Attachment #3, so it should be fine to leave the IFO in this state overnight.
The whole point of the upgrade was to move to a more reliable system - but seems quite flaky already.
From the measurements I have, the Y arm loss is estimated to be 58 +/- 12 ppm. The quoted values are the median (50th percentile) and the distance to the 25th and 75th quantiles. This is significantly worse than the ~25 ppm number Johannes had determined. The data quality is questionable, so I would want to get some better data and run it through this machinery and see what number that yields. I'll try and systematically fix the ASS tomorrow and give it another shot.
Model and analysis framework:
Johannes and I have cleaned up the equations used for this calculation - while we may make more edits, the v1 of the document lives here. The crux of it is that we would like to measure the quantity , where is the power reflected from the resonant cavity (just the ITM). This quantity can then be used to back out the round-trip loss in the resonant cavity, with further model parameters which are:
If we ignore the 3rd for a start, we can calculate the "expected" value of as a function of the round-trip loss, for some assumed uncertainties on the above-mentioned model parameters. This is shown in the top plot in Attachment #1, and while this was generated using emcee, is consistent with the first order uncertainty propagation based result I posted in my previous elog on this subject. The actual samples of the model parameters used to generate these curves are shown in the bottom. What this is telling us is that even if we have no measurement uncertainty on , the systematic uncertainties are of the order of 5 ppm, for the assumed variation in model parameters.
The same machinery can be run backwards - assuming we have multiple measurements of , we then also have a sample variance, . The uncertainty on the sample variance estimator is also known, and serves to quantify the prior distribution on the parameter for our Monte-Carlo sampling. The parameter itself is required to quantify the likelihood of a given set of model parameters, given our measurement. For the measurements I did this week, my best estimate of . Plugging this in, and assuming uncorrelated gaussian uncertainties on the model parameters, I can back out the posterior distributions.
For convenience, I separate the parameters into two groups - (i) All the model parameters excluding the RT loss, and (ii) the RT loss. Attachment #2 and Attachment #3 show the priors (orange) and posteriors (black) of these quantities.
So that the experts on MC analysis can correct me wheere I'm wrong.
The pressure is still 2e-4 torr according to CC1 so I thought I'd give ASS debugging a go tonight. But the arm transmission signal isn't coming through to the LSC model from the end PDs - so a resurfacing of this problem. Rebooting the sender model, c1scy, did not fix the problem. Moreover, c1susaux is dead. The last time I rebooted it, ITMY got stuck so I'm not going to attempt a revival tonight.
I restarted c1scy, c1rfm (so both sender and receiver models were cycled) and power-cycled the c1iscey and c1sus machines. The TRY PD is certainly seeing light - it is just not getting piped over to c1rfm. dmesg doesn't give any clues. I'm out of ideas.
P.S. The new reality seems to be that getting ITMY stuck in the event of a c1susaux reboot is inevitable. As is the practise for ITMX, I tried slowly ramping the PIT and YAW biases to 0 slowly - but in the process of ramping YAW to 0, the optic got stuck. I am ramping in steps of 0.1 (in units of the PIT/YAW sliders, waiting ~3 seconds between steps), I guess I can try ramping even more slowly.
Update: I power cycled the physical RFM switch. This necessitated reboot of all vertex FEs. But seems like things are back to normal now...
Note: to unstick ITMY, seems like the best approach is:
A more permanent fix than a crocodile clip was implemented. Should probably look to do this for the X end unit as well.
sudo mkfs -t ext4 /dev/sdb
sudo dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync
controls@c1vac:~$ sudo dd if=/dev/sda of=/dev/sdb bs=64K conv=noerror,sync
[sudo] password for controls:
^C283422+0 records in
283422+0 records out
18574344192 bytes (19 GB) copied, 719.699 s, 25.8 MB/s
To complete the story before moving on to ALS, I decided to measure the X arm loss. It is estimated to be 20 +/- 5 ppm. This is surprising to say the least, so I'm skeptical - the camera image of the ETMX spot when locked almost certainly looks brighter than in Oct 2016, but I don't have numerical proof. But I don't see any obvious red flags in the data quality/analysis yet. If true, this suggests that the "cleaning" of the Yarm optics actually did more harm than good, and if that's true, we should attempt to identify where in the procedure the problem lies - was it in my usage of non-optical grade solvents?
In an earlier elog, I had claimed that the suspected clipping of the cavity axis in the Y arm was not solved even after shifting the heater. I now think that it is extremely unlikely that there is still clipping due to the heater. Nevertheless, the ASS system is not working well. Some notes:
We have to systematically re-commission the ASS system to get to the bottom of this.
I've suspected that the TTs are drifting significantly over the course of the last couple of days, because despite repeated alignment efforts, the AS beam spot has drifted off the center of the camera view. I tried looking at IPPOS, but found that there was no data. Looking at the table, the QPD was turned backwards, and the DAQ cable wasn't connected (neither at the PD end, nor at 1Y2, where instead, a cable labelled "Spare QPD" was plugged in). Fortunately, the beam was making it out of the vacuum. So as to have a quantitative diagnostic, I reconnected the QPD, turned it the right way round, and adjusted the steering onto it such that with the AS spot on the center of the CCD monitor, the beam is also centered on the QPD. The calibration is uncertain, but at least we will be able to see how much the spot drifts on the QPD over some days. Also, we only have 16 Hz readback of this stuff.
I leave it to Chub to take the high-res photo and update the wiki, which was last done in 2012.
Already, in the last ~1 hour, there has been considerable drift - see Attachment #2. The spot, which started at the center of the CCD monitor, has now nearly drifted off the top end. The ITMX and BS Oplev spots have been pretty constant over the same timescale, so it has to be the TTs?
Last year, I worked on the ALS delay line electronics, thinking that we were in danger of saturation. The analysis was incorrect. I find that for RF signal levels between -10 dBm and +15 dBm, assuming 3dB insertion loss due to components and 5 dB conversion loss in the mixer, there is no danger of saturation in the I/F part of the circuit.
The key is that the MOSFET mixer used in the demodulation circuit drives an I/F current and not voltage. The I-to-V conversion is done by a transimpedance amplifier and not a voltage amplifier. The confusion arose from interpreting the gain of the first stage of the I/F amplifier as 1 kohm/10 ohm = 100. The real figures of merit we have to look at are the current through, and voltage across, the transimpedance resistor. So I think we should revert to the old setup. This analysis is consistent with an actual test I did on the board, details of which may be found here.
We may still benefit from some whitening of the signal before digitization between 10-100 Hz, need to check what is an appropriate place in the signal chain to put in some whitening, there are some constraints to the circuit topology because of the MOSFET mixer.
One part of the circuit topology I'm still confused by is the choice of impedance-matching transformer at the RF-input of this demod board - why is a 75 ohm part used instead of a 50 ohm part? Isn't this going to actually result in an impedance mismatch given our RG405 cabling?
Update: Having pulled out the board, it looks like the input transformer is an ADT-1-1, and NOT an ADT1-1WT as labelled on the schematic. The former is indeed a 50ohm part. So it makes sense to me now.
Since we have the NF1611 fiber coupled PDs, I'm going to try reviving the X arm ALS to check out what the noise is after bypassing the suspect Menlo PDs we were using thus far. My re-analysis can be found in the attached zip of my ipynb (in PDF form).
To debug the issue of the suspected drifting TTs further, I temporarily hijacked CH0-CH8 of ADC1 in the c1lsc expansion chassis, and connected the "MON" outputs of the coil drivers (D010001) to them via some DB9 breakouts. The idea is to see if the problem is electrical. We should see some slow drift in the voltage to the TTs correlated with the spot walking off the IPPOS QPD. From the wiring diagram, it doesn't look like there is any monitoring (slow or fast) of the control voltages to the TT coils, this should be factored into the Acromag upgrade of c1iscaux/c1iscaux2. EPICS monitoring should be sufficient for this purpose so I didn't setup any new DQ channels, I'll just look at the EPICS from the IOP model.
In my effort to understand what's going on with the suspensions, I've kicked all the suspensions and shutdown the watchdogs at 1235366912. PSL shutter is closed to avoid trying to lock to the swinging cavity. The primary aims are
All the tests I have done so far (looking at free swinging data, resonant frequencies in the Oplev error signals etc) seem to suggest that the problem is mechanical rather than electrical. I'll do a quick check of the OSEM PD whitening unit in 1Y4 to be sure.But the fact that the same three peaks appear in the OSEM and Oplev spectra suggests to me that the problem is not electrical.
Watchdogs restored at 10 AM PST
PSL NPRO PZT voltage showed large low frequency (hour timescale) excursions on the control room StripTool trace, leading me to suspect the slow servo wasn't working as expected. Yesterday evening, I keyed the unresponsive c1psl crate at ~9 PM PST, and had to run the burtrestore to get the PMC locking working. I must have pressed the wrong button on burtgooey or something, because all the FSS_SLOW channels were reset to 0. What's more, their values were not being saved by the hourly burt-snap script, so I don't have any lookback on what these values were. There isn't any detailed record on the elog about what the optimal values for these are, and the most recent reference I could find was Ki=0.1, Kp=Kd=0, which is what I've set it now to. The servo isn't running away, so I'm leaving things in this state, PID tuning can be done later.
I also added the FSS Slow servo channels to the burt snapshot requirement file at /cvs/cds/caltech/target/c1psl/autoBurt.req, and confirmed that the snapshots are getting the channels from now onwards.
While looking at the req file, I saw a bunch of *_MOPA* channels and also several other currently unused channels. Probably would benefit from going through these and commenting out all the legacy channels, to minimize disk space wastage (though we compress the snapshot files every few years anyways I guess).
Reminder that this (unrelated) issue still needs to be looked into... Note also that the new vacuum system does not have burt snapshot set up (i.e. it is still trying to get the old channels from the c1vac1 and c1vac2 databases, which while has significant overlap with the new system, should probably be setup correctly).
IMC was not locked for the past several hours. Turned out MC autolocker was stuck, and I could not ssh into megatron because it was in some unresponsive state. I had to hard-reboot megatron, and once it came back up, I restarted the MCautolocker, FSS slow servo and nds2 processes. IMC re-locked immediately.
I was pulling long stretches of OSEM data from the NDS2 server (megatron) last night, I wonder if this flakiness is connected. Megatron is still running Ubuntu12.