The staging building chiller has been replaced with a new unit and appears to be cooling the building very well. Pictures attached show the old chiller being removed and the new chiller ready for install.
Evan, Sheila, Ed, Nutsinee on the phone
Both TCS lasers tripped off within 20 minutes tonight, apparently because of a low temperature alarm on the chillers. We called Nutsinee and used the wiki to reset the chillers and lasers, but the lasers didn't come on although the lights on the controller were lit. The chillers tripped again within 15 minutes of being reset.
We tried setting the TCSX guardian to Down and resetting the chiller again, but the temperature dropped quickly and the chiller started beeping and had a low temperature warning message, so we shut it off using the power button.
In the attached screenshot the bottom screen is a time machine image of the laser screen from yesterday, the top screen is right now. We tried enabling the output of LZR_HD, but that didn't work.
Nutsinee wondered on the phone if this could be related to some bad settings after todays model restart. I laoded the burt that was created by the model restart, but there are no diffs so that seems OK.
We should have this problem fixed soon. There is a chasis that can be installed to fix the problem, however it also needs some changes to channel calibrations and the front end code.
I tweaked all the PUM pitch/yaw notch filters for the quad bounce and roll modes so that we can have a slightly better gain margin on the hard angular loops.
The bounce notch was tuned slightly too low given our bounce mode frequencies (9.73, 9.77, 9.81, and 9.85 Hz according to our bounce monitor filters). The roll notch seemed too wide. (Are we supposed to see Shapiro effect when driving the PUM close to the bounce and roll modes?)
I changed both notches to 3rd order elliptic filters, so the filter ripple should not extend above the 0 dB mark (previously, it went as high as 2 dB). I retuned the stopbands to give >30 dB attenuation at the bounce and roll mode frequencies (previously, the highest bounce mode frequency was only notched by 16 dB). This reworking has the added benefit of giving us 10° of phase back at 8 Hz.
These filters were installed and seem to work fine in full lock at 2 W.
Below are the past 10 day PSL trends. THe chiller woes are still apparent in these. The spare chiller has since been substituted while the former is being addressed.
It seems that the contrast defect degrades as a function of the PSL power when we power up according to the data from last night. This is not surprising.
However, this could be a cause of the recent behavior where the power recycling gain decreases as we power up. Tuning up the differential CO2 at a high PSL power can be an interesting experimental option to try out.
[An offline analysis]
I have looked at the data from last night. There were two good periods where the interferometer was locked stabily at different powers within the same lock stretch.
Because our OMC/DARM servo automatically maintains the same amount of the DC carrier light at the dark port (20 mA in DCPD SUM), increasing the PSL power from 17 to 31 W must have changed the DARM offset by sqrt(17/31) = 0.74. If there was no contrast defect, the amount of intensity noise at DCPD SUM should not change because it is only the local oscillator field who carries the intensity noise audio sidebands to the dark port and the absolute size of the local oscillator field is maintained to the same value by the OMC/DARM servo. In contrast, the intensity noise conveyed by the contrast defect simply scales with the PSL power and therefore one might expect an increase by a factor of 2 ish in the coupling coefficient for the case of the contrast-defect-induced intensity noise (assuming that the contrast defect does not change.)
The attached below shows a spectrum of the DCPD SUM for the two different power configurations.
As shown, the spectra above 100 Hz are (roughly speaking) almost always intensity noise limited. The variation of the spectral shape below 100 Hz might be due to different ASC settings which Sheila and Evan tried on the fly. The intensity noise increased by a factor of 4.5 or so rather than a factor of 2 at around 1 kHz as we powered up the PSL power. This likely means that the contrast defect became worse by a factor of 2 or so in its field strength power at the dark port. Although it is unclear how much this degradation of the contrast defect contributes to our recent low power recycling gain, it may be interesting to tune up the differential CO2 to see if we can get back to a high recycling gain.
Vern, Peter, Nutsinee
Quick Conclusion: The power readout from the front end channel (CO2_LSRPWR_MTR_OUTPUT) agrees with the analog readout at the table. So this channel is an accurate representation of what goes to the periscope.
Details: We made four measurements at various spots (see attachment). First we measured the power going to the bottom periscope mirror (point 4) to confirm that the readout was indeed agreed with the front end channel and the maximum outgoing power was really ~2W (running with CW through 18 deg RS). After that we tried to identify where we lose the power. We lased the power using PWM at 12%. The power drop between point 1 and point 2 was ~8.5% which agrees with our measurement back in early Feburary (alog25410). However, the power drop between point 2 and point 3 was 65%, in Feburary we observed only 12% power drop at this point. This suggests that the beam must have clipped on either M4A or M4 mirror. Peter was saying the beam was hugh at this point (about the size of a nickel). The power drop between point 3 and point 1 was 87%, agrees with what we measured in February. This suggests further that the clipping at M4 mirrors is a new issue.
Today I "meticulously" helium leak tested the Vertex RGA and couldn't find any external air leaks. This is unfortunate as it now confirms the worst case which is that the suspect calibration gases are the likely source of our contaminates (including air). Some history -> After the initial bake out of this RGA, The scans seemed typical until exposed to the calibration gas(es) (see https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=27100 and https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=27078) after which, they looked dirty. Hoping to still utilize these already installed leaks, I did a second bake (nominal 200C) only, this time, with both leak's isolation valves open and heated to 100C for the entire hot "soak" period before then being isolated - just prior to the temperature ramp down phase. Following this second bake I had a new problem - I couldn't get anything but a "flat line" trace on the RGA, as if the ionizer was working but the mass filter (RF) section wasn't. In addition to the standard "tapping" of the hardware, I opened the Kr cal-gas to get a burst and to confirm the symptom. I eventually discovered that a small piece of aluminum foil was shorting out some pins in the RGA's electronics socket. Once removed, the RGA worked fine with the exception that the scans looked "dirty" and had the presence of air. What would they have looked like had I fixed the problem before exposing the baked RGA to the Kr? I now feel that these NOS calibration gases installed at the end-stations and now at the Vertex ought to be removed, discarded and replaced with proven "clean" ones. To this end, I plan on temporarily installing their replacements (and the to-be-installed ones slated for the mid-station RGAs) on one of the VBOs to take a look at them before they get installed on the site RGAs.
The vacuum MEDM web screen shot software was restarted to pick up the latest vacuum MEDM screens following today's work on h0vacmr (See alog 27364).
Looked at the mode matching of the high power oscillator beam into the pre-modecleaner situated in the diagnostic breadboard (DBB). Moved ML2 closer to the pre-modecleaner and things did not appear to change much. Keita told me that an interlock prevents operation of the DBB, which I will look into tomorrow. It might be that I left something disconnected - although I don't think I did.
Sheila, Haocun We are trying to compare the amplitudes of 90MHz and 36MHz modulations at the AS port. 1. Using gpib connected to Agilent 4395A to measure the spectra of photo-current signals. The measurements are 1E-7 V/rtHz for 36MHz, 6E-8 V/rtHz for 45MHz, and 7E-8 V/rtHz for 90MHz. All three signals are at the similar level. 2. Using RF detectors to check the input signal of demodulator, which are -18dBm for 36MHz, -12dBm for 45MHz, and -28dBm for 90MHz. The 90MHz is lower, about 1/4 of 36MHz. 3. From measurements coming out from the demodulator are: 2943 counts for 36MHz, 2557cnts for 45MHz, and 17cnts for 90MHz (after being divided by gains). The factor drops to < 1/100. These results are not consistent with each other, and we will check by changing the whitening gains and the demodulator.
Updated measurement with the analyzer: -47dBm for 36MHz and -72dBm for 90MHz. There is a factor of 17 between them.
WP 5889 I updated the code on h0vacmr to change IP1, IP3 and IP8 to gamma controllers. Since I still had time before the end of maintenance I also took the opportunity to add IP13 (HAM1) and IP14 (HAM6). IP13 is using one controller of a dualvac. IP14 is using a gamma controller. The only hiccup was that the conversion from Amps to Torr for the gamma controllers was getting truncated to an integer. I fixed this in the library. The names changes for IP1, IP3 and IP8 are attached. I have updated the medm screens.
1. Ran field cables for the ITM ESD install. This required working on top of HAM3 to dress cables. Cables in the CER were dressed into SUS-C6.
2. Installed railing and shelves for the high voltage ESD supplies (CER Mezzanine)
3. Ken ( K Electric) ran power cable from CER Mezzanine to the SUS racks in the beer garden
4. Started installation of NGN electronics in CER Rack OAF-C1
5. Looked at power distribution for the Op Lever in LVEA. Power supply in CER was powered on, and began testing drop off points on the floor, correct voltage output.
Gerardo, Kyle, Chandra Replaced HAM 11 & 12 annulus IPs - both were spent. While at it, we re-checked the inner o-ring leak to see if it changed since many door bolts were tightened after pumping down diagonal volume. We vented the annulus space to atmosphere. Diagonal volume pressure rose an order of magnitude. Attached are pressure plots from annulus vent on 3/24 and from today. Annulus space quickly pumped down to 1e-6 Torr with hung turbo carts and now is pumping purely on annulus IPs. HAM 12 reads 2 mA (HAM 11 wiring on CDS) HAM 11 reads 5 mA (HAM 12 wiring on CDS)
Ross, Tega, Evan
Last night we were using the ESD drivers on ETMX to control a mechanical mode at 15217Hz. We used a similar method as mentioned in our previous aLOG 27153 to track the mode but now this signal can now be used to drive the ESD to damp these modes. The ESD quadrants that are used for the damping of this mode are the lower left and upper right and the damping signal is applied positively to the lower left and negatively to the upper right.
For this particular case the mode that we were trying to damp was within ~1Hz of another line around 15218Hz which may be a mechanical mode relating to another test mass. This makes it difficult for the line tracker to stay locked onto the mode if it’s amplitude is lower than the neighbouring line.
The time series plot shows the amplitude output from iwave line tracking. Before the green line the tracker settles on the pre existing line at 15217Hz. A gain of 1000 is then applied to the damping signal which exites the mechanical mode. After around two minutes a gain of -1000 is then applied to the the damping signal which damps the mechanical mode. However after the mode has been significantly damped the line tracker then changes the frequency that it is monitoring to the neighbouring line so it is no longer damping the same line. The spectrogram of OMC data underneath the time series shows how the line amplitude grows and then declines as the gain is altered.
We are hoping to repeat this test to identify more of these modes. This seems to be a good indication that this process should work if it was this mode being excited by a parametric instability.
Landed cables for ion pumps IP13 and IP14, per WP 5897.
IP13 is located on HAM1 and its controller is a dual type (Negative polarity).
IP14 is located on HAM6 and its controller is a LPC type from Gamma Vacuum (Negative polarity).
Awesome! No more red and we have more eyes on VE!
QUAD Master model change WP5899
Kiwamu, Betsy, Dave:
Kiwamu made a changed to QUAD_MASTER. All four quad models were rebuilt and restarted (h1susitm[x,y], h1susetm[x,y]
CAL model changes WP5894
Kiwamu, Dave:
All three calibration models were changed, built and restarted (h1calcs, h1calex, h1caley).
Vacuum Controls change for Mech Room IP1,3,8 WP5889
Patrick, Dave:
Beckhoff code for h0vacmr control of IP[1,3,8] changed. New EPICS database, autoBurt.req and DAQ ini files generated. This entails a channel name change, I have changed the running minute trend files. Still need to cover archived files.
Update NDS2 client code. WP5886
Jim B:
nds2 client code updated, please see earlier alog.
DMT Broadcaster channel change
Dave:
Changed the duotone channels in H0BROADCAST0.ini to reflect LLO names (use IN1 rather than OUT channels in filtermodule). This change went into effect when the DAQ was restarted.
New SUS PI models
Tega, Ross, Dave
New h1susitmpi, h1susetmxpi and h1susetmypi models were installed.
ADC card installed in h1oaf0 IO Chassis for NGN model
Dave, Jim:
We installed a 6th ADC card into h1oaf0's IO Chassis. This is a long-term-temporary card (will be removed post O2) which will provide the h1ngn model with 30 sensor channels. Powering down h1oaf0 meant that the following models were restarted: h1ngn, h1pemcs, h1oaf, h1tcscs, h1odcmaster, h1calcs.
The h1iopoaf0 model was edited to add the 6th ADC.
The h1ngn model was modified to use the new ADC.
In the IO Chassis we decided to locate the ADC between the DAC and DC power supply rather than group it with the other 5 ADC cards, This minimized the models which needed to be modifed to just 2.
Check 18bit DAC card in h1susb123 which failed its autocal
Dave, Jim:
Following the reboot of h1susb123, h1iopsusb123 reported that the first 18bit DAC card had failed its autocal. Today we restarted h1iopsusb123 three times, and each time the first DAC is reported to have failed autocal.
h1asc filter file loaded
Dave:
h1asc was showing a modified filter file. Investigation shows another repeated-filters problem (DHARD this time). I performed a full load.
Additional Ion Pumps in MechRoom
Patrick, Dave:
on Tue afternoon Patrick added two new Ion Pumps (13 and 14) to h0vacmr. I rebuilt H0EDCU_VAC.ini and restarted the DAQ. The autoBurt.req file was also updated.
WP 5894, 5899
related alogs: LLO log 26126, LHO log 27336
In order to update the online calibration model and some related infrastructure, I have recompiled, rebuilt and restarted the following front end models this morning.
h1susitmxh1susitmyh1susetmxh1susetmyh1calcsh1calexh1caleyAll of them are burt-restored to 7:10 AM this morning. The changes I made on h1cal* are already summarized in alog 27336. As for the change on h1sus*, it is only quad suspensions' master library, QUAD_MASTER.mdl, that was modified; I have replaced a calibration oscillaltor on the L3 stage with a new fixed phase oscillator.
Two updates related to h1calcs:
With the elevated intensity noise currently being injected into the interferometer, we can passively make an estimate of the CARM pole frequency.
The attachment shows the transfer functions that take interferometer input power to transmitted arm power (as measured by the four end-station QPD sums) with the interferometer locked at 22 W. There isn't good coherence around the CARM pole frequency itself. However, if we normalize each signal to RIN, this fixes the dc value of the transfer function at 1. Hence, the magnitude of the slope is sufficient to extract the CARM pole.
At 10 Hz, the transfer function is 0.0631, which implies a CARM pole of 0.63 Hz. There is about 1.5 % uncertainty from the spread in the values from the four transfer functions, and about 2 % uncertainty from the drifts in dc intensity for the four QPD sums during the measurement period. Together that makes 2.5 % uncertainty. There may also be additional uncertainty from the whitening and antiwhitening of the QPD signals.
If we assume an ETM transmissivity of 4 ppm, an ITM transmissivity of 1.45 %, a PRM transmissivity of 3.0 %, and 50 ppm of loss on each test mass, we expect a CARM pole of 0.64 Hz (see T1500325).
Obviously a superior method is to take a driven transfer function that has good coherence at and around the CARM pole frequency, and to use the phase information to make a true fit.
Kiwamu, Evan
Similarly, the transfer function from input intensity to POP dc can give the recycling gain of the 9 MHz sidebands. This transfer function is attached, in RIN/RIN. The ac magnitude of this TF is about 0.04.
Assuming the 45 MHz sideband power is negligible in the PRC, assuming the 9 MHz modulation depth is 0.22 rad, and assuming the carrier power recycling gain is 35 W/W, I believe this implies a 9 MHz recycling gain of 60 W/W or so.