Took spectra of the IMC length control signal with different control configurations for HAM2-ISI and HAM3-ISI. In comparision to measurements with damping only, most improvement in the 0.5Hz-3Hz  range were seen with Level 2 isolation and 100mHz Sensor Blend on translational DOF, and 250mHz Sensor Blend on rotational DOF. 

Both HAM-HEPIS were locked, and control configurations (ISO level, and blend) are systematically the same for both ISIs. One should note however that level 2 controllers used here are specific to each chamber. I will try having the exact same filters on both ISIs to limit the differential motion injected later.

I haven't managed to calibrate these curves yet. I am working on it and it should come soon.

These measurements were performent yesterday, with the IMC locked using Matt's new MC2 M2-M3 cross-over. This recent modification of IMC controls allow keeping the IMC locked, even with the level of differencial motion injected at low frequency by the Lv2 ISO /w 100mHz X,Y,Z blend & 250mHz RX,RY,RZ blend configuration. The seismic isolation is then optimal, in comparision to what we obtained so far, in the 0.5Hz-3Hz  range.

I've noticed that this was not alog-ed yet, so here it is.

1: Oscillation

Right after WFSs were powered up originally on Jan/16, WFSA segment 1 and 4 showed some crazy thing going on, which continued for 9 hours or so (first plot, the vertical line in CH2 and CH3 indicates the time of installation). This is the head the installation crews reported to be hot, so the oscillation thing was there from the start, but this may be a different problem from the death of WFSs.

2: Death of WFSB at 13:17:47 local time propagated to WFSA, then to MC REFL almost instantly

On Jan/22 at 13:17:43 local time, when WFSB was malfunctioning and outputing a bogus offset, something happened to WFSB and the output went zero-ish. (Second plot).

Four seconds later, something drastic happened that caused huge spikes to everything.

Though you might not be able to see it, first WFSA WFSB saw a spike, then 4/16 seconds later WFSB WFSA, then 2/16 seconds later MC REFL PD, in this order. Though the timing of EPICS is sort of dubious, we're not talking about 1 sample time difference here and I'm sure this order is correct.

None of WFSA segments responded to light since then, and REFL DC developed a large offset since then.

When everything was dead, CM board power was tripped also. WFSA and WFSB are connected to the same chassis, but MC REFL is connected to a different chassis, and CM board is in yet another chassis.

Though it's not on the plot, we didn't find any suspicious thing in the PSL except obvious things (e.g. FSS disturbed by MC servo which was kicked by the incident).

3: Did WFSs die when the breaker of the WFS DC interface tripped?

As I reported earlier, at some time on Jan/22 the DC WFS interface went bad (it was the power protection board) when I was measuring WFSB supply voltage, and the breaker switch was tripped. I was the only person in front of the field rack at that time.

But it's not clear if I was in the LVEA at 13:17. Combining blurry memory of me, Filiberto and Kiwamu, plus an incomplete record of the card access system (it didn't register my swiping out when I went out of LVEA when somebody else opened the door to go out or come in, or I didn't bother to swipe out  and just went after somebody else), it's not impossible that the breaker tripped when WFSs died.

Mark B.

Preparing to do a set of Matlab TFs on IM1 to test out new provisions in the code for HAUX. The main focus at this point is for the script to run at all, so Filiberto's cabling work and other minor disturbances are not a problem.

Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot requested from 5 PM Jan. 23 to 5 PM Jan. 24. Also attached are plots of the modes to show when they were running/acquiring data. Dust monitor 10 in the H1 PSL enclosure is still indicating a calibration failure.

The data was taken from h1nds1. No data output errors were returned, but the plots were generated.
5100 seconds worth of data was unavailable on this server
1306.0 minutes of trend displayed

Many thanks to Joe and Keith at LLO for building a new l1susmc2 model. The main changes are:

l1susmc2 links to the MC_MASTER model.

In MC_MASTER, the links to the SIXOSEM_T_STAGE_MASTER and FOURSTAGE_MASTER are removed, these parts are now incorporated into MC_MASTER and modified for MC2.

The MC_MASTER has three outputs for the X_M1,2,3 stages which are filtered and summed by the IMC block.

We updated sus/common/models and sus/l1/models, then transitioned l1susmc2 to h1susmc2. 

The original H1SUSMC2.txt filter file was kept, with the three IMC filter modules appended from the LLO L1SUSMC2.txt file.

After the h1susmc2 restart, I burt restored back to 4pm this afternoon.

The DAQ was restarted as part of the install. Some mx_streamers needed restarting on other frontends.

Giacomo, Matt, Lisa

In the last few days we tried to use the Laser Head Calibrator (4mW @ 980nm, D1201258, S/N S1300150) to check the RF chain of the IMC LSC RFPD.
The idea is to use this type of LHC for all the ISC RF sensors to make sure that the signal chain from the sensor to the ADC works properly.

Before using the LHC on the LSC RFPD, we used it on an 1811 to calibrate the RF input of the laser head. We got about 13 mW/V.

We then used the LHC on the IMC LSC RFPD (24.078 MHz) installed on IOT2. We drove its RF input at 24.078473 MHz (100 Hz off from the IMC modulation frequency) to produce a 100 Hz sine in the I and Q demodulated channels. However, the amplitude of the sine was much smaller than what we expected (something like a factor of 50), given the RF input calibration factor and the diode responsivity measured with the DC response.

Since this measurement was supposed to be straightforward, we took this defeat personally and decided to bring the LHC back to the electronics shop and do more testing.
After using it on another 1811, a spare LSC RFPD and one of the BBPDs, we finally realized what the problem was.

One end of the R1 50 Ohm resistor which couples the RF into the laser head was not soldered .. so the level of RF which we were sending in was correct only if the cable was bent in such a way to make the connection...soldering that end helped :-)

So, for a spare LSC RFPD for 24.078 MHz (S/N S1203398) tested in the lab we get:

V_DC = 4 [mW] x Resp [A/W] x 100 [V/A] = 180 mV ==> Resp [A/W] = 0.45 (this also takes into account the overlap factor with the diode)

Vout_RF/Vin_RF @ 24.078 = 8.5 dB => with V/A = 476 and the Resp A/W = 0.45, this is consistent with the 13 mW/V calibration of the LHC RF input.


 The message 

We can now test again the laser calibrator on the IMC LSC RFPD, and hopefully we will get something that makes sense.

Mark B and Arnaud P.

We took DTT transfer functions on M0 and R0 of ITMy:

^/trunk/QUAD/ITMY/SAGR0/Data/2013-01-24_1300_H1SUSITMY_M0_*_WhiteNoise.xml
^/trunk/QUAD/ITMY/SAGR0/Data/2013-01-24_1300_H1SUSITMY_R0_*_WhiteNoise.xml

Main chain looks very good. Reaction chain is a bit messy but very similar to previous measurements from when it was H2 ITMy.

Next step is to work with IAS to set the pitch of the main chain.

Connected the fluid filled lines today, on the installed BSC9 actuators, to the valves connected to the piping.  The now fluid filled valves have not been opened to allow fluid into the piping yet.  

Kyle, Gerardo 

continued from yesterday...

SW bellows small conflat -> OK 
NW bellows small conflat -> OK 
SE bellows small conflat -> OK 

Demonstrated that the NE bellows leak was coming only from the large conflat joint -> Re-torqued (4 passes) and fixed leak for now -> Feels like an annealed gasket -> will need to recheck tomorrow

The LVEA now sits in laser hazard.  Cheryl anticipates that this status will remain through the night.

• Filiberto swapped a power supply in a H1 IO chassis this morning that brought H1SEIH23 back on line.
• Dust15 alarms:  9:52 AM, 10:53, 11:03, 11:43
• SUS and SEI alarm at BSC 1 this morning due to SEI work
• Operator CDS workshop in the control room at noon
• Laser safety makeup this afternoon

Work permits signed by ops today:  3683 (MC power budget), 3682 (4-hour CDS shutdown next Tuesday for timing network and power distrubution updates).

The crossover platform and ladders over the X-arm were removed from the LVEA, and 9 of the cable tray supports installed along the X-arm. The Apollo crew also started assembly of the  (terra universal) clean room intended for fiber welding. Disassembly was started and halted on the garbing room near the north end of the south bay. Reassembly of said garbing room is ongoing.

The Apollo crew uncrated and wiped down parts of the cleanroom donated to LHO by CIT AOS. (This cleanroom will replace the cleanroom used for Fiber Welding near the Test Stands in the West Bay.)The parts were transported to the West Bay and frame assembly started. The cleanroom will fit into the space occupied by the previous cleanroom with about a foot to spare. Tomorrow, work will continue with ceiling assembly.

Measurement of the mode cleaner power budget this afternoon will cause beams to exit the viewports on the top and the east side of HAM 2.  Access along the east side of the mode cleaner tube will be restricted.

The FMCS computer in the Control Room was down from 12:48 till 13:08 today for maint.  The system is back up and running again

Got the go head from IMC folks and SEI.  Locked up the HAM3 HEPI.
The largest displacement see from unlocked to locked condition is 300 cts or 0.0005".  The runout of the vertical shift is only 140cts so max tilting of the table is only a couple urads.  I see a similar change in the yaw position.  In conclusion, alignment should be very little affected.

late entry for wed afternoon restarts

h1peml0 was restarted to test the Id revision string substitution when the mdl file is committed to SVN.

h1ascimc model was changed to remove the WFS A,B WHITEN GAIN parts as these channels are now controlled by the Beckhoff slow controls system.

DAQ was restarted because of the ASC IMC change, it then used the modified ini files for the BSC HEPI systems. So h1hpiitmy and h1hpietmy models were restarted to load the latest code from Vincent.

At 9:22:13 UTC Jan. 24, the models on the h1seih23 computer stopped running.  All models reported ADC timeout.  Checked I/O chassis, power switch was on but found no indication of life.  The I/O chassis refuses to power up, even after disconnecting the DC power and reconnecting.  Suspect I/O chassis power supply is bad.  h1seih23 is powered off for now.

All

The repaired LSC RFPD has been re-installed on IOT2, and we could relock the IMC easily. We measured the OLTF of the fast loop: we have a 60 kHz UGF and 40 degrees of phase margin (measured with the  730 kHz notch ). Details will follow.