After a transition to partial mode (not writing frames) and back it was found that h2fw0 was creating corrupt data frames. The transition was made to support LDAS QFS disk reconfiguration.

So data written by h2fw0 and served by h2nds0 in the following time period is corrupt. You should use h2nds1 to serve data in this time period

start:
4/24 22:21UTC 3:22pm local
end:
4/25 17:25UTC 10:25am local

We (CDS and LDAS) are investigating what happened. We will put in checking software to alarm if this happens again.

The new 5/16 bolts worked well with the adaptor plate, and the optic was removed from the chamber.  The optic was wrapped and secured on the Straddle Lift, and the Straddle Lift was lowered to it's lowest level, for transporting the optic to the optics lab. 

FYI: the class-B hardware for lifting the "lift table" in and out of the chamber, was installed in a way that didn't make sense to me, and others agreed, so we reworked the hardware configuration for lifting the lift table - picture will be attached.

Cyrus, Jim B., Dave B.

Beginning Monday (and finishing Tuesday), we cut over to the new CDS core switch.  From the network perspective, this is the last major central change required to support both H1 and H2 installation and is the foundation for implementing everything else remaining.  What is left is to install are the new edge switches that support the vacuum and auxiliary systems, which will likely happen in the coming weeks/months.

The initial cut over took about an hour, and we used the remaining time to clean up cabling and the configurations of the switches that will remain in the near and long term, and to perform testing.  The 'old' cdsnat0 and cdsnat1 have been decommissioned.  cdsnat0 has been replaced by rtr-msr-nat0, which is a PC running the Vyatta open source router system.  It handles the NAT functionality for the CDS control network, and the port forwarding for the limited CDS services available to the open internet.  The functionality of cdsnat1 (which never actually performed any NAT...) in routing between the internal CDS subnets is now provided directly by the Layer 3 core switch, a Cisco Catalyst 4503E/Sup7e.  The core switch also handles forwarding the udp broadcasts from the ops network to the front end network for diaggui/dtt testpoint discovery on port 5355.  Attached to the core switch are all the edge switches that support, for example, the H1 and H2 FE networks, the Ops network, and (for now) what remains of the legacy iLIGO Foundry equipment (running the vacuum and FMCS networks).

Unmounted ldas-h2-frames yesterday afternoon from all LDAS systems (gateway, lhoarchive, qfs-nfs1) and CDS h2gw0.  Explicitly enabled the 'writer' mount option on h2gw0 to avoid caching, replacing the following line in /etc/vfstab:
ldas-h2-frames                  -               /ldas-h2-frames          samfs   -       yes     -
with:
ldas-h2-frames                  -               /ldas-h2-frames          samfs   -       yes     writer

Per the man page,

# man mount_samfs
...
writer    Sets the file system to type writer.  There can be
               only one host system that has the file system
               mounted with the writer option at any one time.
               If writer is specified, files are flushed to disk
               at close and directories are always written
               through to disk.  The option atime = 1 is set for
               writer.

               Prior to the 4.0 release, the writer option was
               specified as the shared_writer option.  The older
               syntax is supported for backward compatibility.

The mount options on all LDAS systems were left unchanged:

ldas-h2-frames                  -               /dcs-h2-frames          samfs   -       yes     reader

The filesystem was then remounted and NFS-shared out, first on h2gw0, then on the LDAS QFS readers.

The LDAS systems have not logged any more 'Bad directory validation errors' relating to ldas-h2-frames since the change.  There have been warnings pertaining to cds-h2-frames:

qfs-nfs1:~ # dmesg
...
Apr 25 08:28:35 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 08:28:35 qfs-nfs1 last message repeated 3 times
Apr 25 08:55:10 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 08:55:34 qfs-nfs1 last message repeated 7 times
Apr 25 09:19:34 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 09:20:03 qfs-nfs1 last message repeated 7 times
Apr 25 09:44:05 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 09:44:33 qfs-nfs1 last message repeated 7 times
Apr 25 10:09:03 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 10:09:03 qfs-nfs1 last message repeated 3 times
Apr 25 10:35:36 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 10:36:03 qfs-nfs1 last message repeated 7 times
Apr 25 11:00:05 qfs-nfs1 samfs: [ID 756621 kern.warning] WARNING: SAM-QFS: cds-h2-frames: inode 0x2bdbf has a Bad directory validation error
Apr 25 11:00:34 qfs-nfs1 last message repeated 7 times

The 'writer' mount option was added to /etc/vfstab on h2gw1 yesterday as well, but that filesystem has not been remounted and thus the change has not yet taken effect.

• Pitch: 7.5 µrad down
• Yaw: 87 µrad CCW

These are within our spec for mechanical alignment adjustments (±160 µrad pitch/yaw), therefore IAS is now done at BSC08 and will begin to transition to BSC06 this afternoon.

I went back into BSC8 this morning to fix the ISI's weight problem. Ended up bolting 7 of SUS's dog clamps at various points around the optical table, mostly around the top of the quad. I tried to be careful to not to pinch any cables or interfere with any moving parts. We may not be exactly back to where we were ~2 months ago, but everything is back into spec, or at least as close as we can get. IAS is in now taking final measurements.

Measurements on the M0 & R0 masses are being run on H2 SUS ITMY.  The estimated completion time is ~4:00 AM PST.

Today:
With Jason on the IAS EQ, I finished the fine alignment of the ITMy chains in pitch and yaw. This required knocking the reaction chain top stage blades to swing it into yaw tolerance.  Jason will likely post numbers tomorrow.

With Garcia, I centered up all 20 OSEMs. 

Closed the ring heater.

Finished adding the test mass EQ stops and brackets.

Handed off the SEI to unlock the ISI.

Garcia may run a TF overnight to make sure we're not rubbing anywhere.  Tomorrow, after SEI finishes fussing about their minor weight issue (ehem), Jason will take a final look at the fully floating SEI/SUS BSC8 assembly for any last minute alignment tweeks.

PSL installation team


pulled back the pump light fibers from LAE to Diode room and disassembled them from the diode boxes. They are coiled now and ready to be put into the new rail between the H1 LDR and the H1 LAE.

Moved cables and electronics from the H2 racks next to the H2 LAE to the H1 racks

Installed the PSL water manifold underneath the (new) optical table. The manifold for the "table water" (powermeters, etc.) is not installed yet, because the CPC connectors are not self-sealing (so they are leaking, if nothing is connected to them). Also, the pressure valve for this circuit needs to be mounted to the new filter cartridges.

Since laser diode rack is not going to be moved in one piece, we started to get the shelves into the new laser diode rack (to be continued).

HughR, HugoP, JimW, some sus people...

We attempted unlocking and level Stage 2 of the ISI today. The unlock went well, but we were unable to balance the stage. Access to the walls of Stage2, where the balance mass would go, would normally be difficult at best, but in this case it was completely blocked by the Arm Cavity Baffle. Our CPS's indicate that the payload is light, with all sensors reading a few thousand counts high. This is especially true in corner 2, which is currently reading ~14000 counts. That translates to around .003-.004" high. We currently plan to bolt a couple of SUS's large dog clamps to the table as an improvised ballast. Should be a quick fix first thing in the morning. We do have masses designed specifically to mount to the optical table, but we have none of the smaller sizes on site. Even then, the smallest of those masses are 2.5 kg (~5 lbs) which is likely too big to fine tune the balance from the optical table.

[Jeff B, Deepak K, Andres R and Gerardo M2]

Monday & Tuesday 23rd/24th April 2012

 - The wires for the lower mass was assembled and assembled to the intermediate mass.

 - The Lower mass was hung on the loop after properly placing the wires in the grooves.

 - The intermediate mass is then hung on the lower blades.

 - The Lower blades are then hung freely from the upper blades.

 - The masses were initially aligned and now waiting to be properly aligned.

 - The suspension is fully assembled and all the masses are freely suspended.

May have affected dust counts in OSB labs when stopped and when restarted(?)  

Found that the pitch actuator ram operates normally when decoupled from the fan variable pitch/vane assemblies (toggling power with 4-20ma signal removed resets position counter and causes ram to fully retract.  Re-landing 4-20ma signal wires, once reset, allows remote control of ram position).  

The exhaust variable pitch/vane assembly is easy to manually stroke through its range when mechanically decoupled from the actuation system but the intake variable pitch/vane assembly feels dry and has excessive friction.  It is quite difficult to stroke through its range even whilst decoupled from the actuator/crank/exhaust vane.  I had to use a heavy sledge hammer and "tap" with authority to get it to stroke and to change pitch -> For now, I manually left the pitch at reduced flow from the "as found" state then re-coupled everything but I did not re-energize the actuator ram.  As such, no remote pitch adjustment can be made.  The mechanical force needed to stroke the front pitch/vane assembly is too much for the 500lb max. force of the actuator ram.  Also, SF06 is "viby" when compared to SF05.

Frank Seifert, Jonathan Berliner

We deinstalled the eLIGO EX Pcal system this afternoon. Everything, except for the aluminum support structure, was brought to the LSB Optics Lab. The aluminum support structure was left outside of the End Station by the roll-up door, along with the other scrap. The viewport gate was left closed and the viewport window was not inspected. Though the viewport gate key isn't locked or tagged out (as shown in photo), there are assorted warning stickers on the viewport gate enclosure. Some beam dumps weren't clamped down...we should be more mindful of this in the future.

[Jenne, Jan]
We have started placing accelerometers and cables in the end station, so please be careful when walking near BSC10 and the old H2 racks and the dome.  All accelerometers have cones over them, so step around the cones.  All cables that are in the walking path currently are very short stretches, and are taped to the concrete with cleanroom tape (per Jodi's instructions), and are right near several cones.  

Anyhow, please respect the cones, and step carefully.  

On Monday Scott (Apollo) installed the Protection Shrouds on the HAM2 BNC FeedThrus--None of the other FeedThrus are very vulnerable so we didn't install the Shrouds there yet.
There is a "No Step" placard on the Shrouds--do please comply.

- De-install of BSC 1,2,3 
- Accelerometers being laid down at end station
- Unifirst came
- H2 Front End PSL shutdown
- Mouse and Key: Door Locks
- Mid Columbia:  Working on forklifts 
- Shutdown of H2 PSL front end

The deinstall progressed to the point where the lazy susan is in place, the lift table, teflon highway, and adaptor plate are in place.  However, the shoulder bolts for the tall dog clamps have gone missing and were not found.  Some alternatives were considered and rejected, so Apollo crew and Kyle have come up with an alternative that replaces the missing bolts.  The replacements are being cleaned and then baked overnight, so should be ready for use tomorrow, allowing us to complete the deinstall.

- Cheryl, Jodi, Zack, Caleb, Randy, and Mark D.

I am now running new dust monitor code on h0epics2, which is a Ubuntu workstation. It is running under screen.

There has been a bug in the previous code, such that data does not appear to have been recorded since March 15, 2012 for the following:

Dust monitor in clean room over BSC8
H0:PEM-LVEA_DST15_3     any data for counts > 200
H0:PEM-LVEA_DST15_5     any data for counts > 200

Dust monitor in clean room over test stand in LVEA
H0:PEM-LVEA_DST4_3      any data for counts > 200
H0:PEM-LVEA_DST4_5      any data for counts > 200

Dust monitor in OSB optics lab
H0:PEM-LAB_DST1_3       any data for counts > 20,000
H0:PEM-LAB_DST1_5       any data for counts > 8,000

Dust monitor in OSB bake oven room
H0:PEM-LAB_DST3_3       any data for counts > 20,000
H0:PEM-LAB_DST3_5       any data for counts > 8,000

This was because the audible alarms were set for these dust monitors at these levels. When these alarms went off a signal was sent which was read by the CDS code as an error. The CDS code then set the epics channel to INVALID, and the counts were not updated. This should be fixed in the new code.

 (Doug, Margot, Jeff)
We setup the Newport laser collimator and retro-reflected the beam from the First Contact thin film coated HR surface and measured the power both in front of the optic and measured the transmitted power through the optics (2 optics and 2 separate measurements) with and without the First contact. Measurements were taken at ~3M and ~6M distance from laser collimator to HR side of sample optics. Sample optics were coated for 1064nm with both HR and AR coatings and First Contact sprayed on the HR surfaces.
I didn't see a need to set up a BS and try to measure a true reflect from the HR surface. The small difference would still be int the absorption and diffused beam if present and is within usable range.
Parameters:
Newport laser collimator beam dia  ~3.0 cm dia.
Ophir power meter aperture 1.2 cm ²

optic s/n 11: @ 3.0 meters
power at 0º incidence in front of optic 44.4µW
power transmitted through First Contact 39.4 µW = 85.6% T
power transmitted through without First Contact 38.0 µW = 88.7 % T (higher reflectivity)

optic s/n 38: @ 3.0 meters and 6.0meters
power at 0º incidence in front of optic 42.4µW
power transmitted through First Contact 36.5 µW (3.0 meters) = 85.6 % T
power transmitted through First Contact 36.4 µW (6.0 meters) = 85.6 % T
power transmitted without First Contact 34.2 µW (6.0 meters) = 80.7 % T (4.9 % higher reflectivity)

No retro-reflections interference noticed.

"This may work"
Doug

Apollo (Mark & Chris) pulled the FeedThru (F2-2) to rotate it 180° to put the 'D' right side up in the usual way.  I didn't know the cables were attached and they were unable to pull them far enough out to allow the cable disconnect enabling them to put in a new copper gasket.  I went inside the chamber to disconnect the cables from the FeedThru.  Refer to D1003085 for flange locations and connection descriptions.
Few things here:
1) The In-Vacuum cables on F2-1 have screws for securing to the FeedThru but they were not secured.  I secured them; these would be the ISC-Picomotors & the Transmon 1064nm 25pin cables.

2) The In-Vacuum cables on F2-2 do NOT have screws in the cable and so they aren't secured.  As we flipped the FeedThru and flipped the cables, I felt it prudent to record the cable numbers:
2a) F2-2C1 (ISC Transmon 532 nm) Cable D1000924, S1104106
2b) F2-2C2 (ISC Transmon Beam Diverter) Cable D1000223, S1104077
I did not trace the cables back to the Transmon so don't know from whence they come.

3) While the FeedThru was install upside down from our usual 'D' faces down orientation, my concern about the in-vac cable possibly being installed incorrectly and damaging the FeedThru or Cable was misplaced.  The 25pin cable cannot be installed upside down.  That misplaced concern comes from the 3pin FeedThru where the Vacuum Side can be installed reversed wrt the outside--see the attached photo if you like.