In an effort to help the ISC crew use the seismic platforms I looked at ITMY performance today with the various blends and levels. That is still ongoing though. Meanwhile, since Stefan is able to lock his PRMI with the ISIs off, I propose that the ISI position loops do not do enough positioning to matter (HEPI is another story.) So in an effort to make the ISI turn on more 'one button', with the ISI off I loaded the current positions in to the position loop targets. With the position loops not doing much then, the Isolation can be turned on with the one button. That isn't to say the ISI won't drift somewhere again too far, but for the moment at least it is 'one button.' You can look at the first plot of trends to see what changed. Most changes are very small but RZ is maybe 2urads. Maybe significant but again, if the alignment is OK with the position loops off, these should be just fine.
So I brought the ISI up to Hugo's configuration of 6 Dec. Not that it can't be improved or that there are no problems but at least the BLRMS looks pretty good, I think a lot better than turned off. This configuration is 250mHz blends on Stage2 with T250 blends on stage1 except X & Y with T100mHz.44NO. This is with level3 controller. The script actually engages the boost2 rather than boost3 (heard that before haven't ya) but I really didn't see a difference at higher blends when I switched it to boost3 so I don't think it matters.
Again, we certainly need to improve these ISI performance, another day.
Spent some time looking at the noise of the CPS on the BSC after rerouting the sync signals
One figure is a matlab plot of some calibrated signals using a fixed target on which there are some VERY bad BS channels
the second plot is a dtt screen of all of the BSC CPS channels in counts, (BS-V3 is looking at a fixed target, all other channels are looking at motion) with all of the ISIs OFF.
there are a number of icky channels
Phil rewired the CPS distribution box ( i think that it is now identical to what is at LLO) and the really horrible noisey channels got better
there is still some excess noise in ITMY stage 2 corner three and may itmx stage 1 H3 although that might not be real
I measured the noise floopr of the HAM-6 CPS. This is in the configuration that seemed best at LLO, a ground from the electronics rack and a local ground.
The gold colored line is twice our noise model which is what we were getting at LLO. I have been using two different targets which are at +/-4Volts and since the noise goes as the square of the
distance we would expect the noise from the channels using the different targets to be (1+0.4)^2 and (1 -0.4)^2 (there is a 1mm offest) which is more or less what we are seeing at high frequencies.
At low frequencies (were we use then in the control) it is clearly steepper then the noise model
there is also a noise spike at 0.5Hz in H2 and V2, I chased this a bit last night and found
- It doesn't look like it is coming from the CPS electronics, I turned of the power to the CPS (unplugging the calbe in the electronics room) and it was still there
- when I shorted out the inputs to the HAM SEI interface on channels #1 and 2 it popped up in channel H3 and V3
- as far as I can tell it has always been at exactly 0.5Hz
the low freqeuncy excess noise might not be real. After doing some more measurements it looks like the theremal time constant is more like a 1/2 hour (guess) then a few minutes (2" x 2" X 1/2" aluminum wrapped in foam) so I probably didn't wiat long enough for these measurments
Stefan, Kiwamu
The PRMI lock is now quite solid. After we engaged a dither alignment system for ITMY, we turned on the ring heater on ITMY with a hope it is going to improve the mode matching.
Below are Aiden's plot and remarks: I've attached some plots of thermal lens vs time for a RH running at 1W. The slope of the linear section, dS/dt, from t = 2000s to 3000s is 6.75E-9 diopters per second per Watt. The final steady-state defocus, S_steady, is about 13.75E-6 diopters. In the initial state, it takes t_S_steady1, 2037 (= 13.75E-6/6.75E-9) seconds to reach assuming an immediate and linear increase in defocus, t_S_steady1 = S_steady/dS/dt .... In other words, the 20W/2.1x = 9.4W should give the maximum mode matching after 2037s of linear increase in defocus AND in the steady-state.
I modeled the 18MHz sideband buildup in the H1 PRMI as a function of ITMY Rc to see how it compared to the observations made recently. The attached plot shows POP 18MHz I and Q signals over ITMY Rc. The leftmost side of the plot is for the measured value of ITMY Rc from the nebula page, with no ITMY substrate lens, and the measured value of ITMX Rc from the nebula page with an -80km substrate lens. MICH was locked using REFL45Q and PRCL was locked using REFL45I. The POP demod phase was tuned to minimize Q-phase signal at the starting ITMY Rc value. The input carrier power in the model is 10W, the 9MHz modulation depth is 0.1, and the 45MHz modulation depth is 0.07. No optics after the PR2 transmission were included in the model.
As expected, the buildup increases as PRX and PRY become better matched. In the model, the buildup increases to a maximum of around a factor of 10 from the initial value. This is a much larger increase than was seen in the experiment. This could be because the model does not include clipping at the BS, which would increase as the beam sizes increase (see lower panel of the plot). Also, it is possible that ITMY has a non-thermal substrate lens (no measurement data is available for this), putting it closer to ITMX in the no-ring-heater state. Beyond the maximum, the PRMI quickly becomes unstable. This may explain the lock loss in the experiment past the maximum POP18 buildup (though many other things could too).
We found that the pitch and yaw outputs for the y-arm initial alignment were flipped. Fixed the model and recompiled. h1asc.mdl: SVN revision 7011.
Following updates of guardian at LLO from the past few weeks, I started updating some sus guardian
It is now possible to save the aligned as well as the misaligned position of the suspensions for every optic. This can be done through the IFO_ALIGN medm screen via the Save Alignment menu (cf IFO_ALIGN.png) or with the pyhton script align_save_burt -a -m living in /opt/rtcds/userapps/release/sus/common/scripts/
For more details see llo alog 10436 and 10466
This is done through Guardian by switching between different states of the suspension. Guardian medm screens can be accessed via the ! G buttons in the IFO_ALIGN medm screen (cf IFO_ALIGN.png).
For now, only PRM PR3 and MC2 are running under the guardian and can be restored this way. For the other suspensions, it will need to be done manually (until a guardian process is created).
When the guardian medm screen is open, it is straightforward to switch between aligned misaligned damped or safe via the "REQUEST" menu (cf GUARD_PRM.png)
IFO_ALIGN.adl was updated from livingston but I had to make some modifications to get the links functioning. The new adl file was commited under the svn
1) As Kiwamu pointed out to me, the arguments under the Save Alignment link were in the wrong order (e.g. "align_save_burt MC1 -m" instead of "align_save_burt -m MC1"), so I modified it for MC1/MC2/MC3/PRM/PR2/PR3/SRM/SR2/SR3/BS/ETMX/ETMY/ITMX/ITMY/TMSX/TMSY
2) Also, the command to create a guardian medm screen (guardmedm) had the full llo path defined (/ligo/apps/ubuntu12/guardian/bin/guardmedm) which doesn't exist here so I changed it to guardmedm without the path. That should also work at llo.
J. Kissel Though we're confident that the only way to improve the X-arm Test Mass angular performance is to improve the BSC-ISI performance between 0.3 and 0.7 [Hz], I tried tweaking the Level 2.1 Damping Loop design to see if I could do any better, by increasing the Q of the boost filters on the L and P degrees of freedom. Further, I moved the frequency of the P boost down from 0.56 [Hz] to 0.51 [Hz] to account for the difference between the modelled and measured frequency of this mode. Regrettably, I could not improve the strength (at the resonances) of the boosts much more than 5 to 10 [dB] without destroying the the stability of the L & P loops -- I'd already pushed the phase and gain margins of the *lower* unity gain frequency pretty far. As such, the improvement was only in the sharp frequency regions over which I'd focused the boost, but little-to-no change in the RMS. We'll continue to work on improving the performance of the ISIs. For posterity, I post all of the design information and performance comparisons. The good news is, that my MIMO model of the QUADs can now accurately predict the optic angular motion -- especially Pitch -- at all frequencies where not limited by its sensor noise. Check out the pg 2 of the third attachment for proof! Details: --------- allquads_2014-01-31_AllGoodFibers_P.pdf -- Shows a collection of transfer functions of all monolithic suspensions. It shows that compared against the modeled first, top mass, P2P mode at 0.56 [Hz], all the measurements show that this mode is at 0.51 pm 0.01 [Hz]. This is what motivated me to focus the sharp P boost at a lower frequency. 2014-01-31_H1SUSETMX_M0_DAMP_Filters.pdf -- Bode plots comparing the 2013-06-14, Level 2.1 Filters against the new 2014-01-31 Sharp Boost filters. dampingfilters_QUAD_2013-06-14_Level2p1_2014-01-30_H1ISIETMX_Seismic_SelectPlots.pdf -- Using current seismic data input, this is a few select plots from the loop design script for the 2013-06-14 Level 2.1 filters. As mentioned above, it predicts the measured P motion extremely well below ~1.5 [Hz]. Interestingly, the Y prediction is not perfect, but I suspect that below 0.5 [Hz] the signal is dominated by L and P motion of test mass, confused as Y. Especially because the two extra modes that appear at ... you guessed it 0.44 and 0.56 [Hz]. dampingfilters_QUAD_2014-01-30_Level2p1_RealSeismic_SelectPlots.pdf -- Selected performance plots with the new sharper boost. Bare in mind that I've used the *model* for the predicted transfer functions, which don't get the first P mode at the right frequency. That's why things appear like they will be unstable, and all sort of gain peaky near the lower unity gain frequency. dampingfilters_comparison_2013-06-14vs2014-01-31.pdf -- A comparison between the two designs. 2014-01-31_H1SUSETMX_Level2p1vsSharpBoost_Performance_ASDs.pdf -- A comparison of the optical lever performance between the two configurations. Of course, the spectra were taken at different times of day, so above ~0.6 [Hz] the input motion is a little different, but below 0.6 [Hz] the motion is the *same*, and one can see the expected change in shape of the hump, but no change in RMS. Oh well. Design Scripts: SusSVN/sus/trunk/QUAD/Common/FilterDesign/Scripts/ compare_quad_dampfilter_design_20140131_NormalvsSharp.m design_damping_QUAD_20130614.m design_damping_QUAD_20140123.m plotquaddampingcontroldesign.m Filters and Model saved to: SusSVN/sus/trunk/QUAD/Common/FilterDesign/MatFiles/ dampingfilters_QUAD_2013-06-14.mat dampingfilters_QUAD_2014-01-30.mat dampingfilters_QUAD_2013-06-14_Level2p1_2014-01-30_H1ISIETMX_Seismic_model.mat dampingfilters_QUAD_2013-06-14_Level2p1_RealSeismic_model.mat dampingfilters_QUAD_2014-01-30_Level2p1_RealSeismic_model.mat
Due to the non-zero chance of a second cooling unit failing, we are leaving the doors between the MSR and the hallway open for the weekend. We should keep the control room doors closed to reduce the noise.
I made a calibration of the IM4trans and MC2trans SUM channels into uW.
The conversion from W incident on the PDs to counts registered on IMC-IM4_TRANS_SUM_OUTMON is as follows:
[counts/W] = responsivity [0.16 A/W] x transimpedance gain [1000V/A] x differential to single input gain [2V/V] x whitening gain [36dB=63.1 V/V] x ADC gain [1.6384x10^3 cts/V].
This is actually the same for both MC2trans QPD and IM4trans QPDs: 2.087x10^9 [counts/W]
I used this calibration factor to calculate the power on IM4trans and MC2trans currently:
IM4trans counts ~= 57890 counts, which converts to 1.75mW. Comparing this with the expected power on IM4trans:
8.73W into IMC * 0.85 IO throughput * 2400ppm IM4 transmission * 0.1 transmission of BS between IM4 and IM4trans QPD = 1.78mW
MC2trans counts ~= 2.17x10^4 counts, which converts to 655uW. Comparing this with the expected power on MC2trans:
8.73W into IMC * (IMC gain = IMC Finesse/pi = 166 ) * 5.1x10^-6 MC2 transmission * 0.1 transmission of BS between MC2 and MC2trans QPD = 740uW.
Both these estimates for the power incident on each PD agree pretty well so I'll go ahead and add a new filter called "cts2uW" in the SUM_OUTMON paths to give the outputs in uW incident on the PDs.
The fact that the SUM output is used to normalize PITCH and YAW signals meant that adding the filter in front of just the SUM channel increased the gains on PITCH and YAW significantly, affecting the IMC ASC loops. I have now put the cts2uW filters in the input filter banks for each segment of both QPDs. It works fine now, so the MC2trans SUM and IM4trans SUM are calibrated in uW. If anyone changes the whitening gains on these PDs, please be sure to adjust the calibration factor accordingly. I may look into automatically factoring the whitening gain into the calibration to avoid this issue in future.
Just to be clear, the calibration from counts to uW for both IM4trans and MC2trans was 0.032 uW per count. The number 2.087x10^9 [counts/W] was in error: this number should have been 3.3080x10^7 [counts/W]. Although the number in the alog post was in error, the correct number was used in the filters which were applied.
Aaron, Luis, Fil, Sheila
There is now a camera installed on the IR trans path at the x end. The image is of the IR path incident on a baffle (a scratched up, old piece of aluminum painted black, to be more acurate.) There is about 30uW of green power coming out of the chamber into the IR path, you can see a square of 4 green ghost beams on the camera image, and ocassional flashes of the IR beam. Adding a dichroic to the path might make it a little more clear what is going on here.
We sent the signal to the red cable behind the monitors in the front row of the control room. I unplugged the image of the BS from the top monitor, and replaced it with this camera. IF anyone want to see the BS camera, plug the blue cable back in.
8:53-9:43 Checking SUS cables on HAM 4 (LVEA) - Filiberto 9:07-10:10 Moving, cleaning, and organizing elements for SR2 installation (HAM4 LVEA) Jeff.B/Jodi 9:40-10:40 Going into LVEA to check a dust monitor which has lost communication Patrick 9:47-11:20 Heading to LVEA West Bay (cleaning/organizing ACB components) – Mitchell 10:15-12:15 Stiffener rings and o ring protectors installation on HAM4 – Apollo 10:33-11:53 Going to HAM 4 to install SR2 Pre-installation Plate(Cookie cutter) – Jeff B. 12:40-12:46 DAQ restart and new h1asc model – Dave 12:54-13:17 Going to BSC2 to work on HEPI – Hugh 13:30-15:11 Heading to LVEA West Bay (cleaning/organizing ACB components) – Mitchell 13:44-15:12 Going to HAM 4 to install SR2 Pre-installation Plate(Cookie cutter) – Jeff B. 14:22-15:05 Heading into the LVEA to search for cables - Corey
J. Kissel, R. Mittleman, R. McCarthy Looking for functional STS-2s to use for BSC-ISI sensor correction at ETMX, one suggestion that came up was the Vault STS-2. However, Rich immediately reminded us that it's over 10 years old and hasn't received any TLC in a while, so it's most likely functioning poorly at low-frequency and will need to be shipped back to Quanterra for maintenance. To confirm, I've measured its low-frequency coherence with the functional corner station STS-2s. Ideally, we'd go grab the VAULT STS2, and huddle test it against one of the LVEA STSs, but ... it's cold, and for the frequencies we're concerned about it shouldn't matter. In summary -- yes: the VAULT STS-2 is poorly coherent with any of the corner. If functional, we'd expect to see coherence down to 50-60 mHz in the Z directions. Richard suggest we'll grab in the Spring. Details: -------- Serial Number Layout: HAM2 -- SN 89922 HAM5 -- SN 100145F ITMY -- SN 89941 VAULT -- (Known, but not yet by me) (ETMY -- SN 89938) Calibration Details: HAM2, HAM5, and ITMY -- 1e-9 [(m/s) / (nm/s)]. These aLIGO GND STS2s, with channels H1:ISI-GND_STS_${CHAMBER}_${DOF}_DQ, are already calibrated in the front end to (nm/s), so all I have to do was convert to (m/s). VAULT -- 3.052e-7 [(m/s) / ct]. Although pem.ligo.org suggests that the calibration should be 0.0076e-6 [(m/s) / ct], it did not match any of the corner station STS-2s at the microseism, where the motion should be identical. So, I just scaled the VAULT to match. Further, the channels are different from pem.ligo.org, they're now H1:PEM-VAULT_SEIS_1030X195Y_STS2_${DOF}_DQ
At COB yesterday we had roughed in the longitudinal, lateral, pitch and yaw positions of the ETMy test mass within the suspension structure. We had to push the structure ~3mm to achieve this. Today, while bringing the height into alignment, we threw the yaw out of tolerance (because the height mechanics are grossely coupled to yaw). After finally getting the height into tolerance (iterate: go up, whoops too far, go down, no up, now down!) we re-tuned yaw to within ~300uRad. We still need to work on yaw more but our backs were literally breaking getting to these adjustments (see pic) so we opted to stop there and start fresh on Monday. We now have the long, lat, and height within spec with the pitch and yaw almost in spec for the main chain. We also adjusted some bad roll on the reaction ERM which for some reason induced pitch. We'll attack that chain Monday. IAS plans to bring the gap measuring aparatus out on Monday which we'll need to use soon.
It is now using the file /ligo/lho/data/conlog/h1/pvlist_1391196748 (attached). There are 99,264 channels in this list. 5,755 are currently unmonitored.
There are 62,966,023 rows in the database.
Suspension has been stopped and Pre-installation plate installed. Work completed!
We've observed some burn marks on the shielding of the ETMy ring heater cables which occured sometime during the 3 prior weld sessions (May 2012, Dec 2013, Jan 2014). The burns are on the lowest ring heater cable that laces around the test mass and makes a connection between the test mass and the PUM. There is a burn on the right segment and the left segment. I dug up a picture that shows that the burn on the "right" segment (as viewed from the back of the suspension) was there just after the May 2012 weld session, so that one is not new. However the "left" cable burn is new from the Dec or Jan welding. We did not see when this actually occured.
Filiberto tested the ring heater cable and found that all pins are operating as per spec, although he thinks pin 1 is shorting. We are tracking down what this means (did we test it correctly, when was it last tested, is it possible that the burn is contributing to the short, even though it does not look like it is, etc.).
This is the picture of the RH "right" cable taken in May 2012.
And here is the "left" RH segment burn which happened in the Dec or Jan weld.
For the ring heater cable, the following pins were tested. Pins 2,3,4, and 5 are tied together. Pins 14,15,16, and 17 are tied together. The resistance between these two sets is 47.5 ohms. Pins 8,9,10, and 11 are tied together. Pins 20, 21, 22, and 23 are tied together. The resistance between these two sets is 45.9 ohms. Pin 1 is tied to shield and shorted to ground. Filiberto Clara