The RF45 noise appears to be back.

Then again maybe it was just a coincidence. It saturated again and the RF45 did not appear to change.

SUS E_T_M_Y saturating (Oct 8 09:48:58 UTC)

Back to Observing at 05:49 UTC

We got to DC readout transition this time, but had similar difficulties to last night where the OMC locked on a wrong mode, switched to the dither alignment and became misaligned.  I added a check on the DCPD sum in the OMC_LOCKED state, if the sum is below 14 mA it will give a notification and not move on the the dither aligment. 

We went through a couple of rounds of turning of the OMC ASC, clearing the history, and turning it back on.  I think this failed the first time because I didn't wait for the suspension to stop swinging before turning the ASC loops on again.  The last time everything seemed fine, but we had been sitting at DC readout transition for a while.  Like last night we lost the lock when we tried to transition.  

ITMY saturated in this lockloss, caused by angular drives that were somewhat large.  It looks like many ASC loops have large drives in the 200 ms before the lockloss, but DHARD goes bad first.  

Chris Buchanan and Thomas Abbott,

Quick follow-up with omega scans. It looks like most of the power is seen in GDS-CALIB_STRAIN about eight seconds after each listed injection time, consistently for each of these three injections. Doesn't look like there are omicron triggers for these times yet, but omega scans for GDS-CALIB_STRAIN are attached.

Full omega scans generated here:
https://ldas-jobs.ligo.caltech.edu/~christopher.buchanan/Omega/Oct07_PCALX_Inj1/

https://ldas-jobs.ligo.caltech.edu/~christopher.buchanan/Omega/Oct07_PCALX_Inj2/

https://ldas-jobs.ligo.caltech.edu/~christopher.buchanan/Omega/Oct07_PCALX_Inj3/

For complete documentation of the detchar safety injections:

The injections are 12 sine-gaussians, evenly spaced from 30hz to 430hz, 3 seconds apart with a Q of 6. There are three sets with increasing SNR of 25, 50, 100 (intended). However, the SNR is limited by the PCAL acuation range at higher frequencies.

To generate the waveforms I used the script written by Peter Shawhan / Andy located here: https://daqsvn.ligo-la.caltech.edu/websvn/filedetails.php?repname=injection&path=%2Fhwinj%2FDetails%2Fdetchar%2FGenerateSGSequencePCAL.m

I tuned the injections to stay within the PCAL actuation limits referenced in Peter Fritschel's document https://dcc.ligo.org/LIGO-T1500484.

The intended time (seconds from start time of injections), freqency, snr, and amplitude (in units of strain) for all injections are pasted below:

__time__   __freq__   __SNR__    __AMP__

    0.50       30.0      25.0    5.14e-21

    3.50       38.2      25.0    4.96e-21

    6.50       48.7      25.0    2.15e-21

    9.50       62.0      25.0    2.07e-21

   12.50       79.0      25.0    1.75e-21

   15.50      100.6      25.0    1.78e-21

   18.50      128.2      25.0    1.92e-21

   21.50      163.3      25.0    2.06e-21

   24.50      208.0      25.0    2.39e-21

   27.50      265.0      10.0    1.11e-21

   30.50      337.6       5.0    8.39e-22

   33.50      430.0       5.0    8.51e-22

   36.50       30.0      50.0    1.03e-20

   39.50       38.2      50.0    9.92e-21

   42.50       48.7      50.0    4.31e-21

   45.50       62.0      50.0    4.14e-21

   48.50       79.0      50.0    3.51e-21

   51.50      100.6      50.0    3.55e-21

   54.50      128.2      50.0    3.85e-21

   57.50      163.3      50.0    4.12e-21

   60.50      208.0      50.0    4.77e-21

   63.50      265.0      20.0    2.21e-21

   66.50      337.6      10.0    1.68e-21

   69.50      430.0      10.0     1.7e-21

   72.50       30.0     100.0    2.06e-20

   75.50       38.2     100.0    1.98e-20

   78.50       48.7     100.0    8.62e-21

   81.50       62.0     100.0    8.27e-21

   84.50       79.0     100.0    7.01e-21

   87.50      100.6     100.0     7.1e-21

   90.50      128.2     100.0    7.69e-21

   93.50      163.3     100.0    8.24e-21

   96.50      208.0     100.0    9.54e-21

   99.50      265.0      40.0    4.43e-21

  102.50      337.6      20.0    3.36e-21

  105.50      430.0      20.0     3.4e-21

Here are the SNR of the CBC injections using the daily BBH matching filtering settings:

end time               SNR   chi-squared  newSNR
1128303098.986  20.35  32.86            19.86
1128303231.985  22.62  32.73            22.10
1128303398.985  23.25  21.05            23.25

Expected SNR is 18.4.

Though a recovered SNR of 20 (about 10% percent difference from 18.4) is comparable to some of the SNR measurements when doing injections with CALCS in aLog 21890. Note this is the same waveform injected here except in aLog 21890 it starts from 30Hz. In both cases the matched filtering starts at 30Hz. The last two have a bit higher SNR though.

I edited Peter S.'s matlab script to check the sign of these PCAL CBC injections.

Looks like the have the correct sign. See attached plots.

To run code on LHO cluster:
eval '/ligotools/bin/use_ligotools'
matlab -nosplash -nodisplay -r "checksign; exit"

Also in hindsight I should have done a couple CALCS CBC injections just to compare the SNR at the time with the PCAL injections.

A time-domain check of the recovered strain waveforms is here: https://wiki.ligo.org/Main/HWInjO1CheckSGs.  I found that the sign is correct, the amplitude matches within a few percent at most frequencies, and the phases are generally consistent with having a frequency-independent time delay of 3 or 4 samples (about 0.2 ms).  Details are on that wiki page.

Thomas Abbot, Chris Buchanan, Chris Biwer

I've taken Thomas/Chris' table of recovered omicron triggers for the PCAL detchar injection and calculated the ratio of expected/recovered SNR and added some comments:

Recovered time      time since                 frequency recovered expected  recovered/expected        comments
                               1128303531 (s)          (Hz)           SNR        SNR           SNR
1128303531.5156	0.515599966	         42.56	34.07	25	            1.3628
1128303534.5078	3.5078001022	        61.90	39.41	25	            1.5764
1128303537.5039	6.5039000511	        64.60	28.29	25	            1.1316
1128303540.5039	9.5039000511	        79.79	23.89	25	            0.9556
1128303543.5039	12.5039000511	1978.42	21.38	25           	0.8552                                  suspicious, the frequency is very high
1128303546.502	15.5020000935	 144.05	26.24	25	           1.0496
1128303549.502	18.5020000935	 185.68	26.38	25	           1.0552
1128303552.502	21.5020000935	 229.34	26.29	25	           1.0516
1128303555.501	24.5009999275	 918.23	27.34	25	           1.0936
1128303558.501	27.5009999275	 315.97	11.05	10	           1.105
1128303564.5005	33.5004999638	 451.89	6.76	          5     	1.352
1128303567.5156	36.515599966	        50.12	68.53	50	          1.3706
1128303570.5078	39.5078001022	 61.90	78.23	50	          1.5646
1128303573.5039	42.5039000511	 76.45	52.04	50	          1.0408
1128303576.5039	45.5039000511	 91.09	48.42	50	          0.9684
1128303579.5039	48.5039000511	 116.63	47.73	50	         0.9546
1128303582.502	51.5020000935	 144.05	52.59	50	         1.0518
1128303585.502	54.5020000935	 177.91	52.3	        50	         1.046
1128303588.502	57.5020000935	 261.81	54.8	       50	          1.096
1128303591.501	60.5009999275	 323.36	55.64	50	          1.1128
1128303594.501	63.5009999275	 414.01	19.67	20	          0.9835
1128303597.501	66.5009999275	 390.25	9.55	       10	        0.955
1128303600.5005	69.5004999638	 481.99	9.34	        10	          0.934
1128303603.5156	72.515599966	         48.35	136.81	100	          1.3681
1128303606.5078	75.5078001022	 71.56	156.91	100	         1.5691
1128303609.5039	78.5039000511	 76.45	102.72	100	         1.0272
1128303612.5039	81.5039000511	 138.03	102.85	100	          1.0285
1128303615.5039	84.5039000511	 134.83	95.52	100	         0.9552
1128303618.502	87.5020000935	 1283.14	104.17	100	         1.0417                 frequency seems a bit high
1128303621.502	90.5020000935	 211.97	107.18	100	         1.0718
1128303624.502	93.5020000935	 261.81	104.53	100	         1.0453
1128303627.501	96.5009999275	 323.36	109.66	100	         1.0966
1128303630.501	99.5009999275	 414.01	42.15	40	        1.05375
1128303633.5005	102.5004999638	 959.39	19.11	20	        0.9555                  this last injection had some kind of glitch on it

In most cases looks like the ratio is within 0.1 of 1. On a quick glance I see 10 injections that were not within this range.

Finished hardware injections. More details to come.

Spectra attached. The performances of the X and Y QPDs are now comparable.

Glad it was found. Attached are before/after spectra from good science times at 9UTC yesterday and today. The first two show the differences in the strain channel. The biggest feature removed is a nasty wide bump around 78Hz. The last three show the differences in the QPD signals. Very different. For detchar folks that might be wondering about what the beam diverter does, see the intro in T1100252. 

Finally, the last three plots shows another difference I spotted. There is a nasty feature at 640Hz that is gone today. I'm not sure if this is associated with the diverter and will keep an eye on it. 

Stan pointed out that a line around 1278Hz also disappeared. See spectrum below. The spectrogram of how this line varies is quite similar to the 640Hz line, and lo and behold 640*2 = 1280.

I still don't know if the 640 and 1278 are for sure from the diverter. If the diverter was switched off during lock, and I knew the exact time, I could tell. But I haven't been able to find the time or the diverter switch channel to read in my MEDM hunting, please advise. 

Josh -

The beam diverter was closed in between lock stretches.  Because it is mounted on the suspended TMS, moving the diverter moves the TMS, which requires realignment of the interferometer. 

It was open through the lock stretch that ended ~14:00 UTC on 6 Oct 2015, and was closed before the beginning of the lock stretch that started ~00:30 UTC on 7 Oct 2015.

The beam diverter readback channels are:

 H1:SYS-MOTION_X_BDIV_A_POSITION (similar for Y) 1 is closed.  

Thanks Sheila and Jenne, the 78, 640, and 1278Hz stuff was all there at the end of the last lock when the X BDIV was zero (open). They were all gone on the first lock when X BDIV was 1 (closed). 

Just to note, I got a chance to make a few excitations on ETMX ISI last night (WP5528 again)  and it seems that there is still a bilinear coupling from ETMX ISI motion to DARM which is much larger than the linear coupling from ETMY ISI motion; that this coupling at 75 Hz hasn't changed much with the beam divereter closed; and that noise from this is not neglibigle in the DARM noise budget over a broad range of frequencies from about 50-90 Hz.

Procedure to perfrom Hardware injection through Pcal:

1. Turn of the pcal excitation.

2. There are two filters in H1:CAL-PCALX_SWEPT_SINE filter bank. Turn on both F1 and F2 if your injection is in units of Strain. Turn on F1 only if it is in units of meter.

3. Perform an injection.

4. Turn off the filters after you are done

5. Turn back the pcal lines on and make sure the SDF monitor for CALEX is all clear.

Looking at the checked in analyze_invactpcal.m, I noticed that you correct not only for the anti-aliasing analog and digital filters in the readback path (which you do need to get to get meters as read by the PCAL), but you also correct the plot for the anti-imaging analog and digital filters.  This implies that the actual induced motion of the mass is off by the anti-imaging analog and digital filters as you had to apply this correction in post-processing.

So the plot you attached tells us how good of a match we have if we pre-warp the injections by the known anti-imaging filters or take them into account in post-processing when looking for them in search pipelines.

In the case where that is not done, then I believe the attached RxPD_over_EXC_no_IOP_compensation.jpg is closer to how the magnitude and phase will be.  All I have done to the analyze_invactpcal.m script is removed the IOP upsampling filter , "par.A.antiimaging.digital.response.ssd" in the calculation of RX_over_EXC_corrected variable.  The plot is still correcting for the analog anti-imaging (as its mostly delay like) and the zero order hold (again delay), which generally just creates a shift in the start time of the injection.