Jenne and Koji

Last week Jenne has put the accelerometers on and under the PSL table immediately after the plastic sheets were removed.

So I took the same measurement as I did on 9th Aug.

Here is the comparison of the vibrational performance of the table before and after the modification.

Basically the table is now stiffer and more damped than it was before.
We don't find any eminent structure below (at least) 70Hz.
This result is obtained despite elevating of the table.

1) Attachment 1

For the horizontal comparison (top),  it is clearly seen that the large resonant peak at 20Hz was eliminated.
At least the new resonances went up to 70-90Hz region. Y is basically equivalent to X.

For the vertical comparison (bottom), it is clearly seen that the resonant peaks at around 50 & 70Hz were eliminated. 
At least no new resonance is seen.

2) Attachment 2

All-in-one plot for the measurement --- spectra, coherences, transfer functions --- after the upgrade. I put the same plot for the one before the upgrade.

 {Rana and Kiwamu}

Yesterday we disabled the sendmail daemon and the isdn daemon on allegra because we don't need these daemons always running.

-   How to disable/enable daemons:

sudo ntsysv

We changed the name of the new control room computer from kallo to rossa (since its red).

I also tried to install the nVidia graphics driver, but failed. I downloaded the one for the GeForce 310

for x86_64 from the nVidia website, but it failed to work. I installed it, but then X windows wouldn't start.

I've left it running a basic VESA driver.

Kiwamu updated the host tables to reflect the name change. We found that both rossa and allegra were

set up to look at the old 131.* DNS computers and so they were not resolving correctly. We set them up for new way.

I've taken the FSS frequency generation box out of the 1Y1 rack. It's sitting on one of the electronics benches. I'm measuring its phase noise.

In fact, many of the mounts need to be adapted to 4": the beefy steering mirrors going into the PMC, the PMC RFPD, the ISS AOM, the Faraday between the NPRO and the AOM, the NPRO itself, the ISS PDs.

Also for the FSS: the 21.5 MHz EOM, the PBS, the AOM, the refcav periscope, and the RFPD.

Its obvious, in retrospect, that we would have to do this, but it somehow didn't occur to me until actually trying to put things on the table...

The NPRO itself is already tapped with 3 (metric) M3 holes. It also has 4 (un-tapped) holes at its 4 corners which look like they are for feet. Anyone have a mount design for the Innolight NPRO already?

We also started labeling the table with the new coordinate system. In this system, the NE corner is the origin. The screw hole which is most NE is 1,1. The numbering increases in the south (+X) direction and goes negative in the west (-Y) direction.

[Rana, Jenne]

Like a new phoenix, the 40m PSL is in the process of being reborn...

We cleared many old optics and components (including Alberto's favorite periscope) off of the north end of the PSL table.  Some optics are stored on the SP table, others on the shelf inside the PSL enclosure.

The new Innolight 2W NPRO is on the table, the PMC has been moved, and the main path of the laser has been sketched out using steering mirrors. Since we still don't have a beam, we're roughly placing all of our optics, and we'll finalize the alignment after turning on the laser.

Using a leveled HeNe, I checked the height of optics we should use to match the height of optics in the chambers by shining the light at the first steering mirror in the chamber, and ensuring that the beam hit the center of that optic .  Since the new PSL table height is identical to the AP table, it's not a surprise that from now on we will be using a 4" beam height on the PSL table, rather than the old PSL 3" beam height.

On the to-do list is to make a plate with 4 through holes to raise the PMC up by 1 inch, and to make an adapter plate (or come up with another plan) for mounting the AOM that goes directly after the NPRO/Faraday, among many other things.  We also still need to make some space for the RefCav to be put in its new place on the table, and then install it with Steve's help.   

For the future SLOW controls our current plan is to keep using the VME based stuff and associated processors (Baja, Motorola, etc.). This is not

a sustainable plan since these are obsolete and eventually will die. One option is to use these boxes from Diamond Systems:

http://www.diamondsystems.com/products/octavio

 { Joe, Kiwamu }

Yes !

We now are damping all of the vertex suspensions including PRM, BS, ITMs and MCs by the new CDS. 

( Note that we are not damping SRM because we don't have it in the chamber. )

 (things to be done)

- Make the binary outputs work.

- Make DTT work

For a futher damping test, I again turned off the vertex optics watchdogs temporarily, including BS, ITMs, SRM, PRM, MCs.

 [Joe, Kiwamu]

We found that the vertex watchdogs were not correctly running.

After I powercycled c1susaux, the problem was fixed successfully.

The symptom: the watchdogs didn't disable the coil signal even when PD_VAR signals went larger than the threshold values PD_MAX_VAR.

Also we replaced the label by the correct name "C1SUSAUX" on a tag which was tied to the front end machine mounted on the new 1X5 rack.

I posted the crane safety document on the 40m wiki, vacuum page as 26 August 2010

Please add your comments and corrections.

The South End Crane will be balanced on Tuesday,  31 August 2010

This will mean that the back door of the south arm will be open on and off. Air quality will bad.

Please plan accordingly.

I think op540m has finally bitten the dust.  I noticed that both of its screens were black, so I assumed that it had crashed due to known graphics card issues or something.  But upon closer inspection, it is way more dead than that.  I checked that it does have power (at least the power cable is securely plugged in at both ends, and the power strip its on is successfully powering several other computers), but I can't make any lights or anything come on by pressing the power button on the front of the computer tower.

Immediate consequences of op540 not being operational are the lack of DMT, and the lack of Alarms. 

Joe is doing an autopsy right now to see if its really dead, or only 'mostly dead'.

EDIT: Joe says maybe it's the power supply for the computer.  But he can't turn it on either.

 was down.  I restarted the version in the 2.7.5 folder.  It went down again almost immediately but stayed up after the second restart.

[Joe, Kiwamu] 

Woooo Yeaaaah ! 

With the new CDS we succeeded in damping of PRM and BS !!

We are in the process of doing a damping test with the real time code and have turned off the vertex optics watchdogs temporarily, including BS, ITMs, SRM, PRM, MCs.

The tile work was done yesterday after noon.

This morning Mike Gerfen and me lowered the enclosure frame to normal height.

Keven- janitor and I removed plastic covers from chambers, racks, SP, MC2 and clean tool boxes.

The afternoon Jenne, Kiwamu, Joe and Aiden cleaned the enclosure inside out. The particle count measured zero inside the enclosure with HEPAs on when the covers were

removed. The MOPA and all other components were happy to see us in excellent condition.

This table height is very user friendly!

Safety grounds were reconnected.

Atm1, new tiles around the concrete slab

Atm2, frame lowered with low cross bars reinstalled

Atm3, the enclosure frame's north west foot is connected to ground

Atm4, PSL optical table is connected to ground at the north east corner through 1 Mohm

Atm5, PSL optical table level at a stimulating, back-friendly height

Turns out the elog version 2.8.0 uses a different encryption method than 2.7.5.  This mean the encrypted passwords stored in the elogd.cfg don't work with the new code.  elogd includes functionality to generate encrypted passwords, but unfortunately I don't know the administration passwords for some of the logbooks.  So I'm going to leave 2.7.5 running until I can get those added properly to the 2.8.0 cfg file.

Thoughts on where to take the pickoff for the SHG for the PSL-green?  We discussed today at the meeting the possibility of putting a 90/10 beam splitter right after the PMC, so that the green team would get somewhere between 100-200mW. 

In a manner similar to the now classic 'Mode Matching from PMC to IMC' entry, I have calculated the lenses and positions needed to match the 2W NPRO beam into the PMC.

The added complication is that we also want to have a reasonable beam size to get into the Faraday and the AOM. It seems that this should be possible using one lens.

After the beam comes out of the AOM, there's another lens to match to the PMC. Its possible to do this with more lenses, but this is just an effort to minimize the number

of surfaces in the beam.

So apparently the themes/configurations didn't work so nicely on some of the logbooks with 2.8.0, so I'm reverting to 2.7.5 until I can figure out (assuming I can) how to get them to display properly.

So unfortunately, I made the start-elog-nodus script smart enough to kill the debugging run I had (although thats probably good since there might have been issues with continuing to run - just poor timing on part of the crash).

In related news, I have gotten the latest version of the elog code to actually compile on Nodus.  I had to hack the cryptic.c file (elog/src/cryptic.c) to get it to work though.

The following was copied from the #ifdef _MSC_VER section of the code into the #else directly following that section. 

#define MAX(x,y) ((x)>(y)?(x):(y))
#define MIN(x,y) ((x)<(y)?(x):(y))
#define __alignof__(x) sizeof(x)
#define alloca(x) malloc(x)
#define mempcpy(d, s, n) ((char *)memcpy(d,s,n)+n)
#define ERANGE 34

I also removed #include <stdint.h> as the functionality it provides is covered by inttypes.h on Solaris machines, which is automatically included.

This new code was released August 5th 2010, while the old elog code we were running was 2.7.5 and was released sometime in 2008.  There are several crash fixes mentioned in the version notes so I'm hoping this may improve stability. I'm in the process of making a copy of the elog logbooks into the elog-2.8.0 install (so as to have a backup with the original 2.7.5).  I'm also copying over all the configuration files.   In a few minutes I'm going to try switching over to the new elog.  If it doesn't work, or is worse, its easy enough to just start up the current version.

All files are located in /cvs/cds/caltech/elog/elog-2.8.0 (the old directory is elog-2.7.5).  I've made  a new startup script called start-elog-nodus-2.8.0.  To start the new one, just run that script.  To start the old one, just go to the elog-2.7.5 directory and run the old start-elog-nodus script.

took two again

 Restarted the elog using the script. I had to do it twice for it to work. This is not the first time this has happened---does anyone know why this might be?

I'm going to take the elog down for one minute and restart it under gdb (using a copy of gdb stolen from fb40m since I couldn't figure out how to install an old enough version on nodus from source).  The terminal with information is running on Rosalba under the "Phase Noise" panel, so please don't close it.  Ideally, the next time the elog crashes, I'll have some output indicating why or at least the line in the code.  I can then look at the raw source code or send the line back to the developer and see if he has any ideas.

I used the free software called 'ABCD' for Mac to construct this mode matching solution for going from the PMC to the IMC.

After getting it close by eye, I plugged the initial guess into Matlab and let it optimize the distances. I then plugged this into 'ABCD'

to get the exact solution. ABCD doesn't actually optimize anything; it just makes a nice table and graphically plots the solution.

• The first waist between the first lens (f = +200 mm) and the second lens (f = -150 mm) is where the triple mod EOM goes. I have not accounted for the index (1.75) of the KTP.
• The third lens we need is a f = +400 mm lens. I have put the lenses in the new layout drawing at the positions indicated in the Omnigraffle drawing. Each grid square corresponds to 1 inch.

The part numbers for these lenses are:

PLCX-25.4-103.0-UV-1064

PLCC-25.4-77.3-UV-1064

PLCX-25.4-206.0-UV-1064

Bertine Robby, Bertin and Jerry completed the installation by pouring concrete yesterday and cleaning up today.

Atm1-2, steel bars, grouted feet and side forms are ready

Atm3,    bounding agent is applied for better bounding

Atm4,    pumping concrete

Atm5,    shaker is used to fill-all

Atm6,    late after noon the forms- sides were removed,     NOTE: actual table to concrete distance is 12.75"

Atm7,    construction people left and the tile man is on the way

The tile patching should complete today so tomorrow we can remove plastic covers.

...was down.  Have restarted it.

Woops, I am sorry about that. I've just cleaned them up.

Bad CDS team. Bad.

 {Joe, Kiwamu}

Today we did the following works in order to get ready for the new CDS test.

 - solved the DAC issue.

 - checked all the channel assignments of the ADC and the DAC.

 - preparation for modification of the AA filter chassis. 

 - checked DAC cable length.

 - connected the power cables of the BO boards to Sorensens.

Although we performed those works, we still couldn't do the actual damping tests.

To do the damping tests, we have to modify the AA chassis to let the SCSIs go in it. Now Joe and Steve are working for this issue.

 Also we found that we should make three more 37pin Dsub - 40pin IDC cables.  

But this is not a critical issue because the cables and the connectors are already in our hands. So we can make them any time later.

(Notes)

 (DAC issue)

  Now all the DAC channels are working correctly.  

There had been a combination of some issues.

  When I posted the elog entry last time, the DAC was not working at all (see here).

 But in the last week Joe found that the IO process didn't correctly run. He modified the IOP file named 'c1x02.mdl' and ran it after compiling and installing it.

 This made the situation better because we then were able to see the most of the signals coming out from the DACs.

     However we never saw any signals associated with SIDE_COILs.

 We checked the DAC cards, their slots and their timing signals. But they all were fine.

At that time we were bit confused and spent a couple of days because the DAC signals appeared at a different slot some time after we rebooted the computer. Actually this issue still remains unsolved...

Finally we found that SIDE_COILs had an input matrix which didn't show up in the medm screen.

We put 1 in the matrix and we successfully got the signal coming out from the DAC.

(channel assignments)

We checked all the channel assignments of the DACs and the ADCs.

All the channels are assigned correctly (i.e. pin number and channel name).

(AA chassis)

 We have been planning to put the SCSI cables into the AA chassis to get the ADC signals.

As Joe said in the past entry (see here) , we need a modification on the AA chassis to let the SCSIs go in it.

Joe and Steve will put an extension component so that we can make the chassis wider and eventually SCSI can go in.

(DAC cable length)

 In the default plan we are going to reuse some DAC cables which are connected to the existing systems.

To reuse them we had to make sure that the length of those cables are long enough for the new CDS.

After stopping the watchdogs, we disconnected those DAC cables and confirmed they were long enough.

Now those cables are connected to the original place where they have been before.

The same test will be performed for the binary outputs.

(power cables to Sorensens)

 Since the binary output boards need +/- 15V power, we hooked up the power cables to Sorensens sitting on the new 1X5 rack.

After cabling them, we turned on the power and successfully saw the green LEDs shining on the back panel of the boards.

Now watchdogs are back.

The suspensions are well damped.

For the new CDS test, I turned off the watchdogs for PRM, SRM, BS, ITMs and MCs.

I will restore these watchdogs after several hours from now.

 Elog was down, I restarted it.

http://lhocds.ligo-wa.caltech.edu:8000/40m/PSL/LayoutUpgrade

P-pol = purple

S-pol = red

The .graffle file for this is in the 40m SVN's omnigraffle dir/

I finished assembling the frequency generation unit for the upgrade. I tested it through to check that the power levels are as expected at the various connection (see attached png, showing in black the design power values, and in red the measured ones).

Because of some modifications made on the design along the construction, I have to recalculate the SNR along the lines.

I can now start to measure phase noise and distortion harmonics.

A document with a description of the design and the results of the characterization measurements will be available in the end.

C1: PEM-count channels are dead since August 12, 2010

The preparation for pouring concrete is continuing. The grout forms were removed. Half inch ribbed steel bars were epoxied into our 4" thick floor.

The concrete holding sides- forms are cut to size. They will be installed on Monday morning and pouring concrete should follow.

We can be planning for cleaning up on Tuesday

Atm1, epoxy used bars

Atm2, south west tripod feet grouted

Atm3, ribbed bars are getting ready

[Yoichi, Jenne]

Hooray!!! The Tip Tilts are completely done!  The only things remaining are (1) Install 4 TTs in chambers sometime in September and (2) do shake tests / take TFs of the spare.

Today we balanced and characterized #'s 1 and 5.  All 5 TTs are waiting happily on the flow bench in the cleanroom for installation.

I had to restart the elog again.

At this point, I'm going to try to get one of the GC guys to install gdb on nodus, and run the elog in the debugger, that way when it crashes the next time, I have some error output I can send back to the developer and ask why its crashing there.

 script

The 10' x 6' optical table is seating on 6 tripods. Each tripod leg has 3 feet. They are supported by jack screws against copper plates on the concrete floor.                            These set screws allowed us to set the height,  level and distribute the load of the table.

To lock this position of the table grouting is used around the jack screws and between the concrete floor and the tripod feet.

Once the grout set the load will be carried by the hole feet  of 4" x 3.25" x3 = 39 sq inches per tripod leg

These ~4" thick grouted islands of 8-9" OD will be covered by 4" concrete. The area between the legs will be filled with 8" thick concrete.

http://www.fivestarproducts.com/techinfo/GroutHandbook.pdf

Atm1, http://www.fivestarproducts.com/cementgrouts/specsheets/grout_ss.pdf

         Cement based grout, early height change 0 - 4%, hardened height change 0 - 0.3%

Atm2-3,  pouring grout

Atm4,  south west leg is done

Tomorrow: setting up ribbed bars, building frame to hold concrete and pour concrete

 I phoned Phil Ehrens and found out that all these wikis have been moved to a new wiki site

The ATF wiki can now be found here

I have updated the link from the 40m wiki to reflect this

The ATF wiki page doesn't seem to be working any more.  Does anyone know where this is held so we can try to get it back online?  Thanks

 I restarted it using start-elog-nodus and this worked out fine - even though I did it from Pete's on my phone ;-)

I found the elog dead again this morning, and the script didn't kill again. I modified the script to use the following line instead of "pkill elogd":

kill `ps -ef | grep elogd | grep -v grep | grep -v start-elog-nodus | awk '{print $2}'`

Hopefully the script will be a bit more robust in the future.

 Just for added interest, I tried a different method when the restart script broke.  The "start-elog-nodus" script has a line "kill elogd".  This seems not to be actually killing anything anymore, which means the elog can't restart.  So this time I went for "kill <pid number>", and then ran the startup script.  This worked.  So it's the "kill elogd" which isn't working reliably.

WE programmed the MOPA chiller's AC unit to turn off at 6PM each day and automatically turn on at 7 AM each morning. Its now very quiet in the control room at night.

Next, we'll leave it off during the day and see if it makes Steve go crazy or not.