The crew was split between working on Viton rims for the new dust dust barriers and working in and around BSC-8. At the BSC, the first activity of the day was adjusting the location of the garbing/staging cleanroom to preserve a pathway along the Y-arm/to the west bay.  The last view port was removed and pre-work FTIR (3) and wipe (3) samples were taken in the chamber. A designated cleanroom was set up for performing shed tests on C-3 covers (BSC/HAM door, support tube, etc.)that will be used in chamber. ~12 C-3 covers of various types were tested, rated, and re-packaged for later use. The outer coverings of the BSC flooring were removed in preparation for installation on Monday.

I posted a couple of pix below so everyone can appreciate the glories of this chamber surface. Note the variety of colors in the first picture and the writing on the surface in the second picture. 

1000 hrs. local 

CP5 dewar vapor pressure now 30 psig and heading toward normal.  Mechanical level gauge for CP5 in agreement with current CDS indicated level, 60-65% full. LLCV needle valve strokes normally with presence and absence of instrument air.  Oberserved needle %open in agreement with current CDS PID output %open -> Manipulated PID parameters to result in a PID output required to gain control of current CP5 level and returned control from manual to PID. Staged replacement part and tools needed to replace differential pressure transducer for CP5 but elected not to replace unit until work week.  Will monitor from home periodically until then.

GV15 annulus pump current now < 5 mA with ion pump + aux. pump cart pumping -> valved-out aux. pump cart

2055 received CP5 alarm -> Site overview from home PC indicated CP5 was 75% full, LLCV 98% and exhaust 1.7 psig -> Came onsite to investigate.  Indicated rate of change of pump level not consistent with maximum observed boil-off rate -> Probably level transducer failing. PID response to bogus low level was to overfill CP5

Also 
-> found CP5 dewar pressure relief circuit closed and over-pressure relief valve open and flowing GN2.  Regeneration line frosted indicating overfilling of CP5 (broken weld at ambient-air vaporizer allows reverse flow through this line to atmosphere, basically acting like a parallel exhaust line for pump when overfilled) -> I opened the pressure relief circuit (Praxair driver must have closed this and forgot to reopen during most recent CP5 delivery on July 27th).

I will replace transducer(s) tomorrow.  For now, I switched CP5 to manual control with LLCV @ 55% open.  



NOTE-Indicated CDS values for the pump may be intermittently bogus until transducer replacement.

Kyle, Gerardo 

(Late log entry, activity earlier in the week)

Installed O-ring type angle valves on (new/untested) 2 of 2 spool-spool annulus type O-ring joint pump ports.  Hung small turbo on valve and backed with aux. pump cart.  After ~20 hours of pumping, the pressure at the inlet of the aux. pump cart turbo was 5x10-6 torr in each case. 

New spool-to-GV15 flange annulus O-ring joint was pumped down with aux. pump cart then switched to ion pump.  Pressure increased slowly overnight and went off scale -> Valved-in aux. pump cart today and pressure now falling (7 mA at time of this entry).  Pump speed of aux. pump cart ~ = that of the ion pump implies ion pump can handle gas load unaided once pressure is below 5 mA. Once realized, new annulus joints at X-mid will be deemed okay and X-mid can be pumped down.

Crew finished painting the diode and chiller rooms today. They also cut out 4 panels into the LVEA plenum for supply air. HVAC guys did not show up today but they are due to show up on Monday. Feedthru penetrations will be cut into place on Monday.

[Thanks MichaelR for tagging in for the homestretch...]
- DennisC, DougC, EricJ, and JasonO in LVEA for IAS planning
- ICC in BSC8
- Cleaning and preparation for next week's HAM6 entry.
- H2PSL HVAC work on Diode Room by Rhodes
- BetsyB, JeffB, RobertL, FiliC by ISI and SUS test stands in LVEA for testing
- GregorioT by MY for OpLev testing
- StuartA, DanK, GregM, DanM work on new LDAS tape library and tour the LVEA with RobertS
- Farewell to PabloD, GregorioT, WalterO

R. Lane

Update to  Post 1205.

After looking at the model, the issue for the mysterious L2 (PUM) signals was a software error, not hardware. The ADC channel selectors in the model were updated to account for the ESD channels, and the L2 channels were not "fixed" at the same time as the other channels were. I corrected the model, committed the change to the SVN, and restarted the model. I also looked at the model for ITMX, it looked good. I did not find a final channel mapping for ETMY to see if it would include the same ESD channel mapping. The h2susbs model, does have some incorrect channel mapping, but I am not sure what they should be. The affected channels are OPLEV_SUM_IN, OPLEV_PITCH_IN, and OPLEV_YAW_IN. They will need to be fixed before they are used.

B. Bland, R. Lane, F. Clara

Further work continued on debugging the electronics, OSEMS, and cables for the UIM and PUM stages.

The mysterious signals in the PUM channels (when no OSEMs were connect to the PUM cables, signals were appearing in all of the PUM EPICs channels) was traced back to the AA Chassis as the possible culprit. No cause or resolution has been found as of yet. The temporary replacement AA Chassis appeared to be functioning, but when work continued it was found to have the same issue. One OSEM was connected, but four OSEM signals were being displayed in the four PUM EPICs channels. When the light was blocked for the one connected OSEM, no change was observed in EPICs.

It is almost impossible to seat the fan-out cables for the UIM stage without removing the UIM OSEMs completely. It also doesn't help that periodically one of the micro DB-9 connectors have mismatched fasteners. In-order to ensure a proper seat of each connector to each OSEM, all UIM OSEMs were completely removed. Open Light and Closed Light values were recorded. The Open Light offsets and gains have been loaded. The "new" flat QUAD flags are impossible to center in the BOSEM on the UIM with the existing flag mounts. The "old" round flags have been installed for the UIM BOSEMs.As well, center the flags/OSEMs for 50% counts had a lot of run-out, making it quite difficult, because rotation was so cross-coupled to the in/out. We opted not to swap the run-outs because it will be done when we swap out the Nickel-plated SMCO magnets in approximately two weeks during the cartridge install. Only one cam was used on each OSEM mounting bracket at the UIM due to the extreme difficulty in maneuvering the brackets.

ITMY has UIM BOSEMs mounted and approximately centered, the M0 and R0 chains are hanging free, and it has not been "socked".

[Responding to request during morning INS meeting...]
Dust monitor #3 is in the ISI stand cleanroom, #4 is in the SUS stand cleanroom within the LVEA.  The scale of the axes in each plot are different.

Notes:
- At no point did these dust monitors alarm this morning, indicating unacceptable levels.
- Dust peaks in ISI and SUS areas do not seem to be correlated, indicating dirty activity particular to that room, not somewhere else in the LVEA.
- H0:PEM-LVEA_DST_3_5 has been 0 since yesterday afternoon...should probably be inspected.

One of the most useful things about DTT is the ability to capture references, and plot them against current measurements. However, reuse of this feature for the same reference with several different measurements (i.e. a template) is destroyed because whenever one closes as saved DTT file that includes a reference, upon re-opening the frequency vector of the reference is shifted by some large amount (in my case, it seems to be exactly 1/BW = 100, but that may just be numerology). For example, check out my "before" vs. "after" saving plots attached.

This is a bug already in place in BugZilla, under bug 254 (and actually has been put in many many times as independent because it had yet to be resolved, and then later removed as redundant), marked with HIGH priority, but the last activity on it appears to be October 2009. Of additional note to Daniel's comment (#5), this is an example of a white-noise transfer function, where the original bug report and his comment were in reference to swept sine transfer functions alone.


The .xml file (WHOSE FILE TYPE IS STILL NOT SUPPORTED TO BE UPLOADED TO THE ALOG), can be found in the SusSVN under
${SusSVNHome}/sus/trunk/QUAD/H2/ITMY/SAGR0/Data/20110812_H2_ITMY_R0_T_TF.xml
(Though all six of the 20110812_H2_ITMY_R0*.xml's show the same bug independently).

J. Kissel, B. Shapiro, B. Bland, R. Lane

Having assumed that the QUAD was " free* " as Betsy had said in yesterday evening's aLOG, I ran a suite of transfer functions remotely before the LHO day started, hoping to see that lacing had done no harm. However, on the first average of the first transfer function (I decided Pitch is the most sensitive to anything, so I took that first), I noticed something was terribly abnormal.

Long story short, it turns out that the suspension was *not* free, in fact, but its UIM stage had been left resting on its bottom earthquake stops. Since sense has still not yet been made of the signals coming out of the *connected* OSEMs, I made the call to do that first before starting another suite to test the do-no-harmy-ness of lacing. Onward into the fog!



Long story long, we "checked the weather by developing a model of the climate, when looking out the window would have sufficed." After seeing the first badness in the pitch transfer function with only one resonance (where there should be many), I immediately grabbed Brett, asking if such results are possible with merely cabling up the reaction chain. After a brief exchange, and Brett taking far more advantage of the quad model that I had realized was possible, he generated a model of what the TOP2TOP TFs of the QUAD would look if the UIM stage was locked. By the 4th of the 6 measurement suite, we had a full story that we believe the UIM stage was locked down some how. Of course, once we finished, Betsy went in and immediately found the the UIM stage was resting on its lower earthquake stops (see page 3 of "pics"). We had suspected the state in which the OSEMs and cables had been left (UIM OSEMs resting on the UIM stage -- pg 2 of "pics", PUM Cables secure to the stage but not connected, -- pg 1 of "pics"), but these clearly were not the primary culprit.

Some interesting points about the exercise:
- I know how to take further advantage of the QUAD model
- We now know what the TOP2TOP TFs look like when the UIM stage is locked down
- The TOP stage is clearly quite free even after cable lacing, given that the single resonance that did show up had a nicely large Q 

I attach the relevant plots and pictures. 

Positioning of the crossbeams on BSC 6 was completed yesterday with the help of Scott and Ed. They are now held close to their nominal position with the support brackets. The jib cranes for lifting the support tubes into the V-blocks were installed but it looks some modification of the cranes will need to happen. The shim under the arm needs to be shaved down as well as some mounting holes will need to be added. Spherical washers will be added to the eye-bolt as maybe moving the through hole further inboard.

We started the day in a meeting with the Apollo chamber cleaning crew to discuss the general BSC Cleaning Procedure and Hazard Analysis. Michael L. reiterated the stop work policy and answered questions about lock-out/tag-out, gate valves, nitrogen, and the possibility that the crew will need to work longer hours to help expedite the cleaning at BSC-8. We also looked at a checklist specific to BSC-8.

Since the dome and door came off yesterday, we were able to begin serious in-chamber work after lunch. Cheryl and Scott H. went into the chamber for a TCS fit check. While he was in, Scott took pix of the chamber for documentation purposes. Mark L. and Randy removed 2 viewports, an illuminator, and a camera. Zack, Chris, and Mick provided support for all these activities. 

A few noteworthy items:  
1. The OpLev laser was still energized when Cheryl and Scott entered the chamber. While not an eye hazard, it is disconcerting to find a bright red laser dot on someone's back or on your own tummy! We have added a step to the procedure to check this item prior to entering any BSC.
2. Cheryl noted that a nozzle that had housed a viewport had some very visible and easily removable debris (a streak on the pinkie finger of a glove) at the bottom of it. She and Scott H. looked at all the small nozzles for debris to see if there was a pattern of some kind. We'll investigate things further.
3. Randy and Mark noted that the upper stiffener and some other flat surfaces on the back side of the chamber exterior still had dust, etc. on them.  They encountered the issue while preparing to remove a camera and viewport. I asked the guys to work in a different location that wasn't problematic until I could talk to someone about the situation. I tried to check in with Michael but couldn't find him (Dr. appointments I think) so I asked Bubba to come out for a look. We agreed on mitigation (cover the flat surface below the port in question with foil). We will also add a step to the procedure for an actual vacuuming of the stiffeners, etc. before a BSC gets its first cleaning.
4. The space around BSC-8 is very restricted. The walkway along Y-arm in front of BSC-8 is a VERY BUSY place, especially with work going on at the test stands. With ICC cleanrooms in place, there is only a narrow passage (less than three feet) between the chamber cleanroom and the wall. We'll have to work on this.

This morning, I laced up the 156" and 121" D1002522 extension cables which switch back through the R0 top mass, UIM, and PUM.  This involved rebuilding the 5 cable harness brackets in these stages in order to add the fluorel shim stacks.  I dropped 8 nuts in the process and had to remove completely 2 of the brackets for further adjustment of binding.  Cross your fingers we do not need to take these apart when the suspension is built with glass.  Once at the top, I strain relieved these 2 cables loosely to the Upper structure.  There are appropriate brackets which I will attach and lace tomorrow.  After this work, I verified that the chain alignment was not too far out via the 50% OLVs of the Top Mass.  So far so good.

After lunch, I attached the 12" UIM D1002523 and 17" UIM D1002524 quadra-puss cables to the extension cables setting them such that they do not short the isolation of any suspension stages (like I did with the extension cables).  However, when I went to plug the microD-9pin connectors into the BOSEMs on the UIM (a BEAR due to lack of room to maneuver for attachment) and the AOSEMs on the PUM, we ran into immediate troubles understanding the signals.  One set had bean reading some random signals prior to plugging anything in - pointing to rack problems much further away.  The other set showed 1 OSEM connection, but was stale when I blocked the light on that OSEM.  Long story short, looks like we could use some EE help first thing tomorrow if possible.  Robert noted a few other fishy things in the rack.

I left the suspension hanging free* but not covered, in the event Kissel wants to run TFs in the morning to at least tell us how the added cables and bracket affect the overall chain.  I also left the 2 top UIM BOSEMs perched near their mounting brackets, plugged in, and not touching anything critical (other stages or wires, etc).  I did this because remounting them is very difficult and they need stay off until we sort out the signal issues.

*as far as I could tell without TFs.

Pictures to be posted soon.

Priming done today. Tomorrow they will paint.

Gerardo -> 
removed viewports from the 1, 3 and 7 O-clock positions of the iLIGO adapter spool at X-mid and installed blanks in the 1 and 3 O-clock positions.  He installed a 10"-2.75" zero-length reducer at the 7 O-clock position.  The 2.75" CF Tee with the pirani and CC gauges which had been removed from the X-mid iLIGO BSC was installed on the zero-length reducer. 

Kyle -> 
removed viewport from the 9 O-clock position of the iLIGO adapter spool and installed a blank in its place.

Similar to what Corey reported in LHO aLOG 1160, we have two more Triax feed thru failures. These occurred in Clean and Bake processes but no one knows how they broke. I think this just testifies to the fragility of these connectors. As Corey reported, the connector was just found to be broken. I attach a number of photos here trying to focus in close but with only marginal success at good visibility. The real life observation is that there appears to be a band of metal that intrudes into the ceramic space-presumably for strength. The ceramic appears to fail very near this band as the band looks to remain, at least partly in the broken free connector. The amount of ceramic looks decidedly less that the BNC style for the HAM CPS. While we have had some failures of these HAM BNC connectors, that has only occurred during shipping to my knowledge.
I don't know the solution to this problem. It looks to me that the connector would be stronger pushing in rather than pushing out, much the way the vacuum would stress the feed-thru--so that is good. In that light, I've directed C&B to always set them down such that when they must/should rest on the feed-thru, they push the connector in rather than out.

Continuation of mudding the drywall. Painting will start tomorrow.

This is a follow-up entry of this one in ilog. The only difference is I'm injecting 20W instead of 2W into MC.

While students were adjusting squeezing table, I quickly swapped the VCO of H1 from the old one to the aLIGO test unit and measured the MC feed back signal going to the VCO.

New VCO has a smaller range, and I needed to increase VCO gain by 8dB.

Green and blue traces in the attached are the VCO feedback signal (H1:IOO-MC_F) with the old VCO from two different lock stretches, and the red is the same thing with the new VCO but scaled down by -8dB. With this setting, OLG of the MC servo with new VCO was only about 0.4dB smaller than the old one.

The new VCO is not doing any harm. In kHz range it is clearly better than the old one.

Considering the fact that the previous test with 2W into MC showed a smaller improvement, it's also clear that the measurement is limited by the sensing noise of MC.

For a better test you need to wait for the full H1 IFO to come back, but this is already good enough in that nothing seems to be wrong.