[Manasa, Jenne]

We think this math is wrong.

If we have P mW through the Faraday, PRM's transmission is 5.5%, BS transmission is 50%, Recycling gain is ~10, pickoff fraction is ~100ppm, we have:

P mW * 5.5e-2 * 0.5 * 10 * 100e-6 = P * 2750e-8 mW = P * 2.7e-5 mW.

So, if P=10 (10mW through the Faraday), we should have 2.7e-4 mW = 2.7e-7 W = 0.27 microwatts = not so many watts.    

If P = 100 (100mW through the Faraday), we should have 2.7 microwatts. Still, not so many watts.

We have the Watec pointed at POY right now, DRMI is flashing, I'm waving the IR card in front of the mirror, and Manasa isn't able to see anything on the monitor.  The power into the vacuum is 100mW (we just measured and adjusted it), so even if we were getting a full 100mW through the Faraday, it would be hard to see.  If we're assuming we get ~half the power through the Faraday, then we should only have 1 microwatt

 QPD at IPPOS has been centered by removing the filter at the QPD.

So we need to remember to check back on AS camera path and the IPPOS as well in addition to the usual MCrefl path.

P.S. We would be happy to have a new laptop in the lab to replace "Belladonna"!

 Hi everyone,

I've been working a bit on neural network code for a controller, and thus far I have code that creates a reference plant neural network. This is necessary to perform a gradient-descent learning algorithm on a controller neural network (one that reads an error signal and outputs actuation force). Because the error signal is read after the previous output of the controller neural network has passed through the plant, in order to calculate the gradient, either the inverse of the plant needs to be calculated, or the plant can be simulated through a neural network, and the error signal can thus be back-propagated through the plant neural network to find the gradient with respect to the output (as opposed to with respect to the plant), and thus back-propagated through the controller network in order to learn. 

I have uploaded to my directory a directory neural_plant. The most important file is reference_plant.c, which compiles with the command

 gcc reference_plant.c -o reference_plant -lfann -lm

The code runs on a file called reference_plant.data, which consists of a series of delayed inputs i_1, i_2, i_3 ... i_{n - 1} of the plant signals and then an output that is i_{n}, the subsequent signal. Parallel streams may also be used, if more than one signal is to be read. The top of the file must contain the number of total training packets (input-output groups), followed by the number of inputs, and the number of outputs. reference_plant.c also has constant variables which specify the number of hidden neurons, which can be changed. 

 Steve!! The light access connector got more ripped during the work last night...we've just taped it back. We might need  to figure out a better way to do this than just cutting through the cover.

[Jenne, Manasa]

Using the alignment of the PZTs and BS from pre-dinner, where the beam was hitting the center of both ETMs, we aligned the DRMI.  The beam was off on the SRM in yaw by ~half a beam diameter, so I undid Koji's movement of SR2 from a week ago.  I loosened the SR2 dog clamps, touched it gently on the base to do a little bit of angle, then re-clamped it.  Once again, Steve's new brass centering target was awesome, since it was on the SRM while I was moving SR2. 

We approximately recentered the beam on the AS camera, although it didn't need much once we got the beam out of the vacuum, by centering it on all of the output AS path mirrors.

We also got IPPOS out of the vacuum.  Manasa was in the process of centering the QPD when the laptop died from too long being unplugged, so we leave that for tomorrow.

Left to do:

REFL path.  REFL is not coming out of the vacuum, and with the light access connector I can't reach any of the REFL steering mirrors, since they're in the center of the IOO table.

IPANG.  Should be easy.

POP, POX, POY.  Need to the the camera-on-a-stick back down to the corner (from ETMY) and point it at the pickoff mirrors to ensure that beam is getting out of the vacuum.

Mannasa and Unni and I looked at the RF driver for the AOM. It was fine.

With the ALC input left unconnected, with the power supply set to +28V, it was drawing 0.56 A.

By adjusting the modulation input we were able to get 1.1 Vrms into the scope (terminated at 50 Ohms) after going through 2 10dB attenuators. 11 Vrms into 50 Ohms is 33.8 dBm ~ 2W.

The RF power trimpot on the front of the driver is now adjusted so that -0.31 to 0.69 V takes the driver output from off to 2W output at 80 MHz.

The previous distorted signal that Jan and Manasa saw was at a level of ~100 mVrms, which is ~0.5 mW of power. At this tiny drive level, the internal amplifier is not linear and is mostly putting out a signal at ~160 MHz.

We checked by putting a square wave into the modulation input that the RF power from the driver would indeed shut off with a time scale of ~20 ns. Manasa will add a picture to this entry. We are ready now to calibrate the transmitted power of the AOM v. the modulation input voltage and then to measure the step time of the AOM.

Remember: do NOT believe the spec sheet of whatever PD you are using. All commercial PDs are slower than they advertise. In order to measure a <1 us step time you must use a PD with a >50 MHz 'bandwidth'.

[Unni, Manasa, Jenne]

It turned out that the beam was a teeny bit high in the corner, so we touched PZT1 and PZT2 knobs to translate the beam down a bit.

Now the beam is centered on the BS (using the 45 degree non-iris target), centered on ETMY (using Steve's latest target, which worked perfectly), and then BS was aligned a tiny bit (really, it didn't need much) to get the beam centered on ETMX.

After dinner I'll align ITMX and ITMY such that their beams retroreflect and I get MICH fringes.  I'll also align SRM and PRM to retroreflect.  Check no clipping on AS path, get REFL path out, center IPPOS and IPANG, check POX, POY and POP.  Then, I think we might be almost done.

Time domain control using LQR technique is now applied to ETMX sus position. The plan was to do it for oplevs, I'll do it after the vent.

The cost function for state space variables was determined by TF  900 / (s + 30)^2. There was no penulty imposed for velocity, only for position. We can try that configuration as well.

 AOM driver has been removed from the PSL table for testing. However the AOM is still inside; so there should be no problems with the alignment. 

The power buildup in the MC is ~400, so 100mW of incident power would give about 40W circulating in the mode cleaner.

Rana points out that the ATF had a 35W beam running around the table in air, with a much smaller spot size than our MC has, so 40W should be totally fine in terms of coating damage.

I have therefore increased the power into the vacuum envelope to ~75mW.  The MC REFL PD should be totally fine up to ~100mW, so 75mW is plenty low.  The MC transmission is now a little over 1000 counts.  I have changed the low power mcup script to not bring the VCO gain all the way up to 31dB anymore.  Now it seems happy with a VCO gain of 15dB (which is the same as normal power).

 {Jan, Manasa}

We started again to calibrate the RF driver. We connected the ALC to the power supply and observed the output RF power on the scope. The RF power did change with ALC voltage, but the RF signal still seems not to be operating at 80MHz 

There is some kind of additional disturbance to the waveform at 80MHz (the frequency of just the waveform with tall peaks or small peaks alone). We made sure we get a snapshot this time!! I am not sure if it will be safe to feed this RF signal to the AOM as such

SOS alignment tool with ID 9.5 and 6.3 mm

 24" diameter clear acetate access connector is in place. The 0.01" thick plastic is wrapped around twice to insure air and bug tight barrier for the MC to lock overnight. The acetate transmission for 1064 nm is 90 % This was measured at 150 mW   2.5 mm beam size.

I'm making a separate entry to go along with this thread of photos...

Putting the camera and 'bathroom' mirror on any table pretty significantly changes the leveling of any table.  The mirror especially is very heavy, although the camera is not feather light.  We need to come up with a new plan for taking alignment-confirming photos without adding anything to the tables.  That, or we have to level the table between each camera shot.  Anyone who has ever leveled one of our in-vac tables should shudder in horror at idea #2, so we need to put some thought into idea #1 before our next vent.  Vent Czar - can you put this on the list, in addition to the REFL rearrangement stuff?

As a result of this, PZT2 needed to be reverted to the place it was before work began on Saturday (so that the beam goes through the 45 degree target without any extra stuff on the table).  This means, unfortunately, that all of the photos / still captures of optics after PZT2 are invalid.

The DRMI was aligned once again tonight. 

Here's a video: http://youtu.be/Cy8nHL9yMeM  (Can someone please tell me / remind me how to make the elog embed videos?!?)

Description of video:

Video capture of AS camera.
NOTE: The beam is a few centimeters above ETMY with this alignment, so it will not be final.

Beginning is ITMY only.
ITMX is realigned to form MICH.
PRM is realigned to form PRMI.
SRM is realigned to form DRMI.
PRM is misaligned to form SRMI.
ITMX is misaligned to form SRY.

With this alignment, I opened up the ETMY door to find the beam there.  The beam is ~half on, ~half off of the top of the glass baffle.  Not the top of the hole, but the top of the piece of glass.  This means that it's many centimeters too high at ETMY.  This helps explain why, while swinging PZT2 around the other day, I could not see any beam on the cage.  It did, however, look pretty close (within a centimeter....I didn't look closer than that since it was so off in pitch) to centered yaw-wise.

Tomorrow I'd like a Clean assistant to help tweak PZT2 to hit the center of ETMY.  We'll need to put the 45 degree target back on to make sure that we don't end up pointing funny down the arm.  Then I'll realign the DRMI one more time.

Tonight, I can't check the full AS path, or any of the REFL path once it diverges from the main path.  Steve's new contraption (which is awesome!) doesn't have doors/windows yet, so I can't open it to get an IR card anywhere near any optics in the IOO or OMC chambers.  I waved PRM around a bit, but I can't find the beam on the REFL camera, so I definitely need to check that whole path again before we close up.

So, we're not closing up tomorrow, but progress has been made, and we're getting closer.

Note to self:  These are the ITMX, ITMY, PRM, BS, SRM biases with this DRMI alignment.  The DRMI is good, but the arms aren't, so these won't be final.  The saved alignments are still those with (for the Yarm) the beam bouncing several times between ITMY and ETMY.  BS was aligned at the time to hit the center of ETMX, and PRM and SRM should be retro reflecting in that alignment.  So, it's possible, that aligning PZT2 to hit the center of ETMY and restoring all of the optics will get me close to being back to DRMI aligned, but in a condition that the arms are align-able too.

Results of the Razor Blade Beam Scan

The horizontal blade test measured the beam intensity as a razor blade passed in between it and a power meter from the left side of the beam (negative x values) until blocking it. The resulting function, found through least-squares regression of the error function, calculates a beam height of 3.6 mm +/- 16 mm. However, the function has a chi-squared value of 3.2, so that value may not be accurate.

The vertical blade test measured beam intensity as a razor moved from below the beam (negative x values) until blocking it. This function, found the same way as above, calculates a beam width of 2.8mm +/- 9.6 mm, and has chi-squared value of 0.77.

Both data sets have a y-error of 0.5 micro-Watts, and an x-error of 0.127 mm. The Python code used to analyze the data and plot the results is attached.

 {Jan, Manasa}

We tried towards calibrating the RF driver of the AOM. We decided to use the normal power supply for both the driver control voltage and the ALC voltage.  But we could not figure out the type of the ALC port to find a compatible mating connector...it did not match with SMA, SMB or SMP. Finally I wrote to the company and got to know it is a filtered feed through. Now that we know how to control the ALC voltage, we will try looking at the signal for varying ALC voltage and see how that goes. 

But when we tried to see the 2W RF signal through the RF scope, with ALC open, we found that the RF signal was distorted and did not measure 80MHz.  It was lame that we did not get a snapshot 

P.S. The AOM has been left disconnected from the RF driver. 

I have touched PZT2 such that the beam goes through the 45 degree non-iris target on the beam splitter.  This puts the beam at the center of ITMY, and without moving the BS, at the center of ITMX.  I say "at the center", but what I really mean is I put the target approximately at the center, within what looks like, say, 2 mm, by looking from above.  The target was many (5ish) centimeters away from the optic though, so that's why my side-to-side centering isn't so precise.  Given that, the beam was always more than half going through the hole of the target for both ITMs, so I'm claiming that the spots on the ITMs are within a few mm of center.

With this alignment, the beam was also hitting the center of the SRM (with all the same caveats).

I was able to get the SRM to retroreflect, while I still had Michelson fringing, so I think that I had the SRMI at least close to aligned (I was looking at the SRM retroreflection at the beam splitter, not all the way out to the AS port).  PRM is also pretty easy to align.

We're hitting the top of the AS camera, so I think things are pretty good.  I don't see beam on the REFL camera, but no investigation of that has been done as yet.

There is some scattering going on in the BS / ITMX chambers that's making me kind of unhappy.  I don't know how to get this to embed the youtube video, so here's the embed link, as well as the regular link:

youtube of AS and BS/PRM camera.

<iframe width="420" height="315" src="http://www.youtube.com/embed/QUbnMLXSS5U" frameborder="0" allowfullscreen></iframe>

Manasa watched the camera while I waved an IR card around in the BS chamber, and the only way I was able to get all the scatter spots to go away was to either block the beam incident on the BS (duh), or block the beam reflected off the BS, heading to ITMX.  Manasa said that the scatter spots still looked like they were fringing though, so I'm confused.  I may wave a card around in the ITMX chamber when I come back later tonight, to see what I can see.  Also, I just misaligned the SRM, and the scatter spots moved.  Now there's just some scatter off of what looks like the BS OSEM holders, as seen through the BS optic.

 The photos on the OMC table are particularly tricky, since the camera plus the 'bathroom' mirror add a lot of weight....even if the MC locked, the input beam would be completely different, so all of the beams would be wrong.

During some of the work on the BS table, ITMY was realigned to have its beam retro-reflect, since the weight of the camera plus mirror was shifting all of the suspended optics on the BS table.  ITMY was restored after that, for subsequent photos.

 My hope is that the DRMI flashes will be bright enough to see on the PO beams. IF we get 10 mW through the Faraday, you should get some buildup when the carrier resonates in the DRMI.

If the recycling gain is 10 and the pickoff fraction is 100 ppm you ought to get ~10 uW on PO. How much of the recycling cavity power gets out of POP?

[Manasa, Jenne]

We took the last of the in-vac photos of mirrors today.  I'll post in the morning.

Tomorrow, I'll align the DRMI once more to check, and get IPPOS and IPANG out of the vacuum.  I'll  take a look at POX, POY and POP, but we may just have to cross our fingers and hope for the best on those ones.  They were pretty hard to get out of the vac during their initial alignment, since they're so weak.

Also, tomorrow morning Steve is going to try out our new light access connector!!!!  I'm so excited!

The goal is to put heavy doors on, on Wed, and start pumping Wed afternoon / Thurs evening.

 There is just so much room on this table.

 Have eaten.  Here's a PDF with all the pictures to-date, along with a few notes.

Also, the first thing we did on Saturday was to fix the yaw pointing of MMT1, so that the beam hit the center of MMT2.  Then we had to touch PZT2 to compensate.  We put the iris target on the BS, and adjusted PZT2 until the beam went nicely through there.  The resulting beam looks good on the SRM, and teh beam is still hitting the AS camera.

 This was in response to my suggestion to move the REFL beam path to the table containing the BS/PRM Oplevs. From this seismic data it is clear that the BS table is no worse than the AS table, so we should plan to make the layout change during the next vent.

 This seismic measurement is for BS and AS tables.

I've suspended microphones around the lab

{EricQ, Jenne]

More photos were taken.  Will post Monday, because too hungry now.

I've measured seismic and acoustic noise on BS and AS tables. It seems that horizontal motion of BS table is ~1.5-2 times more then AS table in the frequency range 5-50 Hz.

Edit by Den: this was POI table, not BS!

All the photos so far:

PZT1:

MMT1:

MMT2:

PZT2:

IPPO:

 I've been taking more photos.  Obviously, it gets quicker as I go along and get the hang of it.  Also, I've been taking overhead pictures with the Nikon so we can see what kind of parallax there is for each snapshot.

However, I just took MMT2, and the beam is nearly falling off the side of the optic!  It seemed fine last night when Rana and I were working on it.  The MC spots haven't moved significantly (I had measured yesterday, and again a few hours ago).  WTF?

This means that I need to move the knobs of MMT1, and then redo the whole alignment chain all over again.  Lame.

EDIT:  MC spot positions, last night at 12:33am, and this afternoon at 2:12pm:

                        year month day hour minute       MC1pit         MC2pit          MC3pit            MC1yaw         MC2yaw          MC3yaw

./data_spotMeasurements/MCdecenter201209140033.dat      1.749759        9.744013        1.025681        -0.791683       -1.338786       -1.779958      
./data_spotMeasurements/MCdecenter201209141412.dat      1.702974        7.916438        0.986519        -0.888736       -0.170237       -1.771267

 Looks good. Any way that you can tell in an unambiguous way, where the beam is, is very good. Ideally we want to have1-2 mm accuracy.

After much fussing, we got a picture of MMT1 with the beam.

Using the iris doesn't seem feasible. Since it has to be significantly separated from the optic, it is hard to judge whether it is centered, especially in yaw.

It took ~30 min to get this picture. Comments on whether this kind of picture is good enough are welcomed, since there are many more to be taken.

I was helping Den get started in the cleanroom yesterday, and I noticed that the new active TTs, like the old passive ones, are set to be 4" from the table.  So, like the old ones, we need 1.5" risers to get the center of the mirror up to our in-vac 5.5" beam height.  I didn't see any risers in there when I was looking around. 

Steve says he still has the drawing that he gave to the shop for the old tip tilts, so he'll double check that the dimensions are the same, and then ask the shop to make 4 more.

After looking at the in-vacuum layout I think we should make two changes during the next vent:

1) Reduce the number of mirrors between the FI and its camera. We install a large silvered mirror in the vacuum flange which holds the Faraday cam (in the inside of the viewport). That points directly at the input to the Faraday. We get to remove all of the steering mirror junk on the IO stack.

2) Take the Faraday output (IFO REFL) out onto the little table holding the BS and PRM Oplevs. We then relocate all 4 of the REFL RFPDs as well as the REFL OSA and the REFL camera onto this table. This will reduce the path length from the FI REFL port to the diodes and reduce the beam clutter on the AS table.

 I've installed the mirrors on 4 tip-tilts. I was able to align 3 of them in pitch, the last one has a screw with damaged thread, I'll continue with it tomorrow.

Alignment accuracy in pitch is ~0.1 mrad. Mirrors oscillate a lot probably due to air flow coming from the side wall.

[Rana, Jenne]

We aligned the DRMI, and have concluded that it looks good enough that we should close up and pump down soon. We still need to use the camera to check things, and get all pickoff beams out of the chambers, so don't get too excited yet.

We looked at the mode matching telescope's calculated beam propagation, and since we're using spherical telescope mirrors at non-zero degree incidence angle, we expect an astigmatism about like what we are seeing on the AS camera.  This matches up with the measurements that Mike posted from his and Q's measurements earlier today.  We think that it has 'always' been this way, and someone just picked a camera position such that the beam used to look more round than it does now.

We aren't entirely sure what's up with the SRM - it almost looks like the pitch and yaw are coupled, but it was pretty easy to align the PRMI.  We don't see any evidence of the crazy, crappy beam that we did before the vent.  This means we have fixed most of the bad clipping problems we were seeing over the last ~year.

In the process of aligning the DRMI, we fixed up the input beam alignment - we were not hitting the exact centers of the MMT mirrors (in pitch, mostly), so we fixed that, and propagated the alignment fix through the chain.  In all, we touched the knobs on PZT1, MMT1, MMT2, PZT2.  The beam then went through the SRM, and we touched a few of the output steering mirrors to get the beam centered on all mirrors. 

I remeasured the MC spot positions, and they're a little worse than they have been.  Some of the spots seem to be off by 1.75mm (or less) on MC 1 and 3.  The numbers, MC1,2,3 pitch, then MC1,2,3 yaw are:   1.749759        9.744013        1.025681        -0.791683       -1.338786       -1.779958

A question to consider before doing the final-final alignment checking is: do we need to get the MC spots centered better than this, especially in light of the potential PMC axis having moved? 

The PMC reflection made it seem that the beam going into it was misaligned. I went to the table and aligned the input beam to maximize the PMC transmission. I got ~10% improvement.

Just to check if something was loose, I started tapping things upstream of the Faraday. When I tapped the actual PMC body it seemed to get unseated and the reflected (unlocked) power jumped up by more than a factor of two.

I don't understand how this could be. The attached trend of the PMC channels shows that ever since the PSL upgrade, the PMC refl has been at the low level of ~0.3 V, except for a brief phase soon after the upgrade late in 2010 and then also for a few hours early in May of 2012.

If the PMC body actually moved, it seems that the pointing into the MC would also change and I don't see that. So what else can it be? Is there some clipping or dust or a burn spot on the PMC REFL path?

The PMC refl image was lost after the body re-settled itself. Jenne and I re-aligned it and added a 0.5 ND filter to the existing ND in order to account for the higher power. We should hide all of the reflective ND filters and just use absorbtive ND for the cameras to prevent reflections.

This image of the past hour shows the event at just before midnight (0650 UTC) where the PMC reflection goes up from 0.28 to 0.85.

We need to do the following things:  Images of optics in DRMI chain, place black glass beam dumps, make sure pickoff beams get out, align IP POS/ANG.

Black glass: behind MMT1, behind IPPOSSM3, forward-going POP beam.

Images and pickoff stuff should happen at the end of each vent.

Images need to be taken of the following optics (with ruler edge at center of optic):

* PZT1

* MMT1

* MMT2

* PZT2

* PRM

* PR2

* PR3

* BS (front and back?)

* ITMX

* ITMY

* SR3

* SR2

* SRM

* OM1

* OM2

* OM3

* OM4=PZT3

* OM5=PZT4

* OMPO

* OM6

* Viewport as AS beam leaves chamber

* POYM1 (check no clipping on edge of mount)

* POXM1 (check no clipping on edge of mount)

Pickoff / aux beams:

* REFL path

* POX

* POY

* POP

* IPPOS

* IPANG

**EDIT:** Mixed up X and Y. Beam is 3.5844 mm tall and 2.7642 mm wide

14.112 hundredths of an inch in the vertical direction

3.5844 millimeters

10.883 hundredths of an inch in the horizontal direction

2.7642 millimeters

Plots and error bars to come soon.

I have given Den 4 G&H R>99.99% mirrors to be installed on the 4 active tip tilts.  He's in there working on things (incl. installing and balancing the pitch of the mirrors) right now.  He'll elog his work later.

 Please, restore condition after you finished and update elog right away! People wasted hours yesterday not knowing the condition of the MC

 [ericq, mikej, some input from zach]

After realigning the MC, the measurement was repeated this afternoon. This time, however, we isolated the beam from ITMY by misaligning ITMX. The beam looked somewhat elliptical to me, and Mike should have fits up tonight. Afterwards, ITMX was returned to the position I found it in, and the PMC shutter and access connector were closed. (Sorry about last night!)

 Summary: Recorded the presence of higher order modes in IMC

What I did: Misaligned the flat mirror MC1 by small amount in both pitch and yaw (it was needed to be done cause at the beginning of the experiment no higher order modes were present)  and scanned the cavity for frequency-range 32MHz to 45MHz.

I found the presence of higher order modes around 36.7MHz (1st order)  and 40.6MHz (2nd order) along with two other strong modes near 35MHz and 42.5MHz.

 Rana is proposing 50 mm hole in the baffle plate that is attached to the tower.  Atm1

Atm2 is showing the back side where the solid line is 40 mm

Eric Quintero and Mike Jenson received 40m specific basic safety training.

 The vacuum envelope must be sealed with light doors on o-rings to insure a bug free IFO.  This was a violation!

We conducted a beam scan on the AP table of the AS beam. We used a lens to focus the beam onto a power meter, and slowly moved a razor blade across the beam using a micrometer, vertically and horizontally both in front of and behind the beam. We also had to block the beam next to the AS beam in order to do this, but is unblocked now. Mike will begin curve fitting the data to try and see if there is a different spot size given by the x-axis vs. the y-axis, and if the lens has any effect.

 On Friday, Koji and I adjusted the beam pointing into the DRMI using the PZT yaw and found that the beam inside the DRMI (as seen on the AS camera) looked OK (not distorted too much).

So it seems that the issue seen before, namely that the DRMI resonant mode is very strange, is no longer true.

The camera image at the AS port still looks elliptical. So Jenne and Mike have started to make this beam round by adjusting the lenses.

Our plan now is:

1) Fix AS camera optics to get a round beam (single bounce off of ITMY).

2) Flash DRMI to make sure the beam at AS is still round.

3) Using the moveable Watec camera and Sensoray, get images of the spot on all DRMI mirrors with DRMI flashing. Use targets and rulers whenever possible to get quantitative measurements of the beam positions. (i.e. just saying "Oh, its pretty much in the center" is the Mickey Mouse approach to science)

4) Align all pickoff beams in this situation. Make sure there is no in vac clipping. Align IP POS and ANG using this input beam pointing.

5) Pump down.