The brucoVirgo code has been modified (thanks to Gabriele) to work under CL7.  The CL7 compatible version is v1r5.

You can check it out of svn: svn co https://svn.ego-gw.it/svn/advsw/brucoVirgo/tags/v1r5

This latest version of the brucoVirgo package is now also available in  /virgoDev/brucoVirgo/v1r5/src/bruco-master

Since about May 13th Bruco stopped working, as also noted by mwas. The problem was not an easy one, but Gabriele and Valerio managed to spot it, and Bruco is again operational. The problem was a memory overflow error associated to the processing of channels with 16384Hz sampling. After adding channels *AdcMaxVal* e *AdcMinVal* to the exclusion list BruCo runs again without problems.

The change has been checkedout in /VirgoDev/brucoVirgo/v1r2. So to run bruco now use (example):

olserver115> nice +30 /virgoDev/brucoVirgo/v1r2/src/bruco-master/bruco.py --ifo=V1 --channel=LSC_DARM --gpsb=XXXXXXXXXX --length=300 --nproc=5 --minfs=100 --outfs=10000 --naver=100 --dir=/data/procdata/web/bruco/2017-MM-GG/LSC_DARM_XXXXXXXXXX/ --top=100 --webtop=20 --xlim=10:5000 --ylim=1e-11:1e-4 --excluded /virgoDev/brucoVirgo/v1r2/src/bruco-master/share/virgo_excluded_channels_LSC_DARM.txt

The detchar doc page has been updated accordingly.

T help debugging this memory problem Franco added this python library: http://pynash.org/2013/03/06/timing-and-profiling/

In [1]: import line_profiler
In [2]: import memory_profiler
In [3]: line_profiler
Out[3]: <module 'line_profiler' from
'/virgo/InstallArea/Linux-x86_64-SL6/lib/python2.7/site-packages/line_profiler.pyc'>
In [4]: memory_profiler
Out[4]: <module 'memory_profiler' from
'/virgo/InstallArea/Linux-x86_64-SL6/lib/python2.7/site-packages/memory_profiler.pyc'>

The BruCo software has been updated by Gabriele to solve the following issue :

# 2017-03-08 - Virgo sampling rate is not always a power of 2, and this caused a crash
#              for some auxiliary channels with sampling rate lower than the desired
#              output. Implemented an upsamping of the aux channel to solve the issue

This latest version has been tested and installed at Cascina, see documentation page.

A minor drawback is that not physical coherence is generated for the upsampled channels at the upsampled frequencies, these shoud be ignored.

This morning we re-locked the north arm to verify the alignment of NI was good, and then aligned the PR to overlap with the beam from NI.

In this state, with the NI misaligned, we checked the alignment of the B2 beam on SIB2.  The beam was re-centered on both PDs and the camera.  When we compare the DC signal of the B2 PDs in this configuration to the IMC transmitted power, we find a drop in coherence between 1-10Hz, see Figure 1.  Somehow the prompt reflection from the PR is not agreeing well with the transmission of the IMC.  We can see the spectrum of the B2 PDs breathing in this region, and while we were on the table we observed some dust particles flickering in the B2 beampath (which is 6W when it enters the bench).  If the disagreement between IMC_TRA and B2_DC is due to dust, we can expect the coherence to improve when we bring the bench under vacuum.  The coherence had not improved ~30min after the work on the bench had finished, so either the dust is a continuous problem or there is some additional clipping on the beam path.

In the afternoon, we locked the PRNI cavity, using first the 6MHz signal from B2 and then the 8MHz signal from the same photodiode.  The noise was largely the same, except for a large 3023Hz line in the 8MHz signal, see Figure 2.  During the lock we apparently excited the 219Hz violin mode of the PR.  Previously we had excited the harmonic of this mode at 438Hz, so I added a 219Hz notch to the PRCL locking filters and the noise was reduced, see Figure 3.

With PRNI locked, we aligned the beam on the SPRB and SDB2 benches.  On SPRB this was easy, the beam was well-centered on the optics and we improved the alignment usig the final steering mirrors before the PD and the camera.

On SDB2 there were many issues, some of which we fixed, some of which should be fixed next week:

• The B1p beam was about 1cm too low when it exited the mini-link onto the bench.  Michal, on the phone from Annecy, used the SDB1 M4 picomotor to increase the height.  We aligned the B1p beam on the photodiode.
• the B5 beam is offset by almost a centimeter in the horizontal direction, compared to the lens and mirrors on teh beam path.  We did not fix this, but these optics should be moved to accomodate the beam position (it is awkward to move the beam from inside the SDB1 chamber).  We centered the B5 beam on the PD, but the beam is clearly clipping somewhere.  This should be OK for locking but needs to be fixed.
• We aligned the B1 beam on the DC PD and the camera, although this beam will need to be realigned once the 2nd OMC alignment is resolved.
• The B1s1 and B1s2 beams are missing some photodiodes, and some of these bright beams are not dumped and exiting the table.  This is not ideal from a laser safety perspective, and probably scatters a lot of light.
• One extra beam is dumped on the bench using a razor dump, which is certainly not clean and not vacuum compatible.  This dump should be replaced with something that doesn't risk making the in-vacuum components (on the detection bench!) dirty.
• The B1s1 and B1s2 beams are very large on their photodiodes, at least a few mm across.  Since these beams appear to be much larger than the PDs we did not bother to center them.  Are the lenses in these beam paths not the final versions?  They are not focusing the beams as we expected.

So, the various PDs in the corner building are in a better-aligned state with respect to the arms, and we are ready to begin CITF locking.   The beam was cenetered on the B2 and B4 cameras, these can be used as a reference for whether the beam is centered on the PDs.  We didn't center the B5 camera since this path will be realigned.