Several of the suspensions were kicked pretty hard (600+ mV on some sensors) as a result of this quick vent wind.  All of the suspensions are damped now, so it doesn't look like we suffered any damage to suspensions.

Steve and I managed to access the fuse in the vacuum VME crate, but replacing it did not bring it back up.  We decided to replace the entire crate.

We manually checked that the most important valves, VC1, VM1 and V1, were all closed.  We disconnected their power so that they would automatically close, and we wouldn't have to worry about them accidentally opening when we rebooted the system.

We noted where all the cables were, disconnected everything, and removed the crate.  We noted that one of the values switched when we disconnected one of the IPC cables from a VME card.  We'll note which one it was in a followup post.  We thought that was a little strange, since the VME crate was completely unpowered.

Anyway, we removed the crate, swapped in a spare, replaced all the cards and connections, double checked everything, then powered up the crate.  That's when minor chaos ensued.

When the system came back online after about 20 seconds, we heard a whole bunch of valves switching.  Luckily we were able to get the medm screens back up so that we could see what was going on.

Apparently all of the ION pump valves (VIPEE, VIPEV, VIPSV, VIPSE) opened, which vented the main volume up to 62 mTorr.  All of the annulus valves (VAVSE, VAVSV, VAVBS, VAVEV, VAVEE) also appeared to be open.  One of the roughing pumps was also turned on.  Other stuff we didn't notice?  Bad.

We ran around and manually unplugged all of the ION pump valves, since I couldn't immediately pull up the vacuum control screen.  Once that was done and we could see that the main volume was closed off we went back to figure out what was going on.

We got the medm vacuum control screen back (/cvs/cds/caltech/medm/c0/ve/VacControl_BAK.adl.  really??)  There was a lot of inconsistency between the readback states of the valves and the switch settings.  Toggling the switches seemed to bring things back in line.  At this point it seemed that we had control of the system again.  The epics readings were consistent with what we were seeing in the vacuum rack.

We went through and closed everything that should have been closed.  The line pressure between the big turbo pump TP1 and the rest of the pumps was up at atmosphere, 700 Torr.  We connected the roughing pumps and pumped down the lines so that we could turn the turbos back on.  Once TP2 and TP3 were up to speed, we turned on TP1 and opened V1 to start pumping the main volume back town.   The main volume is at 7e-4 Torr right now.

So there are a couple of problems with the vacuum system.

• Why the hell did valves open when we rebooted the VME crate?  That's very bad.  That should never happen.  If the system is set to come up to an unsafe state that needs to be fixed ASAP.  The ION pump valves should never have opened.  Nor the annulus valves.
• Why were the switches and the readbacks showing different states?
• Apparently there is no control of the turbo pumps through MEDM.  This should be fixed.

I connect belledona, the laptop at the vacuum station to the wired network, so that it's connection would be less flaky.

 That low frequency effect was due to AA board, now it is gone.

VME crate  blew its fuse and the fuse housing is hard to access.

 The PRM damping restored. Oplev  PIT gain  0.15 and YAW  gain   -0.3 turned to zero.

 This could come from AA board, its range is +/- 2.5 V, RMS of the ETMX table motion is a few times higher then ground motion, so ETMX accelerometer signal was corrupted.

As this small AA range has already caused problems before, I decided to increase it. I've looked through the board scheme and found that all its differential line receives and output amplifiers have absolute maximum range of 40V. We used KEPKO power supply for this board with a voltage range up to 6 V. So I've replaced it with a BK PRECISION power supply and set it to +/- 15 V. Now AA board range is 7.5 V.

I'll leave accelerometers near ETMX table. It's interesting to measure table motion in the morning when trucks drive by.

 High frequency (>60 Hz) resonances that are present at the ETMX motion spectrum seem to be understandable. Amplification ETMX/GROUND of a factor of 2 at 1 Hz is interesting. I've monitored ACC DQ channels for a few hours and noticed that usually spectrum looks like in the previous elog. But every ~40 min ETMX motion is much higher then ground motion at low frequencies (<5 Hz). I wonder if this a reaction of a table to outside disturbances or accelerometer issue.

 I've measured ETMX table motion compared to ground motion using accelerometers. Data and settings in the xml file are at the svn directory 40m_seismic/etmx.

 I've added xml file with measurement settings and data to 40m svn at directory 40m_seismic/etmy.

Accelerometers were installed on the ETMY table and nearby ground to measure amplification of the seismic noise due to the table. During this experiment ground and table motions were measured simultaneously.

      I  was wrong The instability will be the same when the coils are actuated.

 Rijuparna, Manasa

Today I have checked the optical layout of the MC transmission RFPD table and measured the laser powers at different points. Manasa helped me for that. I found the power entering the RF photodiode is 0.394mW while the transmitted power of the cavity is 2.46mW. (I will give the diagram later).

Could you tell us why? Are you thinking about induced current damping?

 We can only decide the need of pitch damping when the coils are activated.

[Koji / Kiwamu]

 We have realigned the interferometer except the incident beam.

 The REFL beam is not coming out from the chamber and is likely hitting the holder of a mirror in the OMC chamber.

So we need to open the chamber again before trying to lock the recycled interferometers at some point.

--- What we did

•  Ran the MC decenter script to check the spot positions.
  • MC3 YAW gave a - 5mm offset with an error of about the same level.
  • We didn't believe in this dither measurement.
•  Checked the IP-POS and IP-ANG trends.
  • The trends looked stable over 10 days (with a 24 hours drift).
  • So we decided not to touch the MC suspensions.
• Tried aligning PRM
• Found that the beam on the REFL path was a fake beam
  • The position of this beam was not sensitive to the alignment of PRM or ITMs.
  • So certainly this is not the REFL beam.
  • The power of this unknown beam is about 7.8 mW
• Let the PRM reflection beam go through the Faraday
  • This was done by looking at the hole of the Faraday though a view port of the IOO chamber with an IR viewer.
• Aligned the rest of the interferometer (not including ETMs)
  • We used the aligned PRM as the alignment reference
  • Aligned ITMY such that the ITMY reflection overlaps with the PRM beam at the AS port.
  • Aligned the BS and SRM such that their associated beam overlap at the AS port
  • Aligned ITMX in the same way.
  • Note that the beam axis, defined by the BS, ITMX  and SRM, was not determined by this process. So we need to align it using the y-arm as a reference at some point.
  • After the alignment, the beam at the AS port still doesn't look clipped. Which is good.

---- things to be fixed

   - Align the steering mirrors in the faraday rejected beam path (requires vent)

   - SRM oplev (this is out of the QPD range)

   - ITMX oplev (out of the range too)

Just some plots. There is nothing new here except for the fact that I learned how to analyze phase maps myself and how to prepare them for Finesse. In other words, everything is ready for a Finesse simulation.

These phase maps show the raw measurement of ITMY, ITMX and PRC:

Subtracting out the tilt from all phase maps, and the curvature from the PRC (I found the fit 121m consistent with previous fits), the one obtains the following residuals that can be used in Finesse (order is again ITMY, ITMX and PRC):

 I've manually corrected MC1 input matrix by looking at UL, UR, LL, LR transfer functions between each other. This improved pos significantly and slightly yaw.

I rebuilt/install/restarted c1ass.  It came up with no problems.  It's now showing DAC0 with no errors.

After lunch I'll test the outputs.

FSS SLOW control did not drift during the lock at night with MCL path working and AC coupled.

I finally got around to rebuilding, installing, and restarting all the IOP models.  Everything went smoothly.  I had to restart all the models on all the screens, but everything seemed to come back up fine.  We now have many fewer dmesg error messages, and the GDS_TP screens are cleaner and don't have a bunch of needless red.

A frame builder restart was also required, due to name changes in unused (but unfortunately still needed) channels in the IOP.

Tip Tilt pitch adjustment on existing-in vacuum suspension. This can be added by a simple installation of a 1.25" long 2-56 threaded rod with nuts.

I started to create a Finesse model of the PRC cavity. We have the phase maps for the PRC and the two ITMs. I could not find anything for PR2,3 and BS. All files can be found in my SVN folder /janosch/PRC40m. I used the AutoCAD model to determine angles of incidence and distances. These numbers are largely inconsistent with numbers that you can find elsewhere on the 40m wiki, but this certainly depends on what accuracy is required for interferometer alignment and I don't understand anything about alignment.

The phase maps come in a format that needs to be modified before they can be used in Finesse. I have started with this work, but maybe someone else can take over. The phase maps show tilts and the PRC also has the curvature. These have to be subtracted out before the maps can be loaded into Finesse. I asked GariLynn for the code that they use. The Finesse model (MichPRC_40m.kat) does not load the phasemaps yet, and I just wrote some random parameter values for the TEM00 input beam to the PRC. So these Gauss parameters need to be corrected.

I will only go on with this work if Rana tells me that I should do so, otherwise it is on hold until we have a volunteer.

 I've actuated on MC1 with UL, UR, LR, LL coils in turn and measured sensor readings. All coils separately work fine from the first look.

On the plot: black - free mirror, blue - UL coil actuation, green - UR, grey - LR, red - LL.

I've written MC123 input matrixes to the front-end.

MC1 diagonalization is poor, better then before, but still pitch is seen in pos and yaw. Either smth is malfunctioning or flags touch sensors and do not move freely. On the plot mc1_new black lines - before, red - after rediagonalization.

I've added MCL and WFS stop triggers into C1MCS/SUS model. Threshold value of MC_TRANS can be changed in the text entry located in MC2_POSITION medm screen. I tried 2 cases: trigger either blocks signal before MCL filter bank input or after output. Due to filter history in the 1 case MC2 was still slightly disturbed (C1:SUS-MC2_ULPD_VAR ~= 15) right after unlock. In the second case there was no disturbance as we zero output signal, but then I had to add "clear history" command to the mcup script.

WFS triggers block the signal before ASCPIT/YAW filter bank.

I have added some control logic and appropriate output DAC channels for the input tip-tilts (TT1 and TT2) to the c1ass model.

The plan is for all the tip-tilt drive electronics to live in a Eurocrate in 1Y2.  They will then interface with a DAC in c1lsc.

c1ass runs on the c1lsc front-end machine, and therefore seemed like an appropriate place for the control logic to go.

I added and interface to DAC0, and a top_named IOO block, to c1ass:

The IOO block includes two TT_CONTROL library parts, one for each of TT1 and TT2:

This is just a start so that I can start testing the DAC output.

I have not recompiled c1ass yet.  I will do that tomorrow.

 Also, when I walked through the control room later, the WFS were driving the MC crazy.  I turned off / disabled the WFS from the WFS screen.   In my infinite spare time, I need to put in the real-time triggering, so that the WFS turn off as soon as the cavity unlocks. 

 Rijuparna, Jenne

Today I checked the optical lay-out in MC REFL board of the MC REFL path on the AS table (I will put the updated diagram in a few hours), and took a record of the reflected power of unlocked MC and power entering MC REFL PD. The power coming out of MC cavity when unlocked is 1.25W and power entering REFL PD 112mW (Jenne measured these powers for me). 

I also got a description of the MC demodulation board from Jenne.

(Edits by Jenne)

 Aluminum sheet as shown will replace the acetate. Side entries for your arms and "window" on the top will be covered with acetate using double- sided removable-no residue tape 3M 9425

1)  Mirror mount holder for "large silvered mirror" inside of the 8" OD tube vacuum envelope.

 Since we upgraded the CDS system, I guess our ADC ranges have gone up but we never did anything to the OLs to match the ADC ranges. From Liz's daily summary page of the OL, I see this:

So we need a factor of 5-10 increase in the electronics gain (why isn't the BS SUM on there?). This might be accomplished in the head, but for the ones with whitening boards, might be better to do there.

(** add to Jamie's list of long term tasks **)

 Absolutely hokey. What are our requirements for this RFPD? What are the power levels and SNR that we want (I seem to remember that its for 22 as well as 110 MHz)? Perhaps we can test an aLIGO one if Rich has one sitting around, or if the aLIGO idea is to use a broadband PD I guess we can just keep using what we have.

I moved some of the REFL optics on the AS table by a teeny bit to accomodate the new place that the REFL beam exits the chamber (none of this was done while we were at air....we were only dealing with the AS beam at the time, and were happy that REFL came out of the vacuum).

The REFL beam is now on the REFL camera (with PRMI aligned), and the beam is going toward the 4 REFL RF PDs, but it's not aligned to any of them.

I have some questions as to mystery optics on in the REFL path.  There is a 90% BS, and I don't know where the 10% reflection goes....is it going to beat against the AUX Stochino laser?

I have to go, and I didn't fix the videocapture script today, so pix tomorrow, I promise.

Riju hasn't been in the lab in a long time to do any measurements, so I put the signals back to how they should be. 

I turned off / confirmed off the things which were sending signal to the EOM:  the network analyzer, the RF generator box, and the Marconi which supplies the 11MHz. 

I removed the cavity scanning cable, and terminated it, and put the regular 11MHz cable back on the splitter.

I then turned on the RF gen box and the Marconi.  The Marconi had been off, so we were not getting any 11MHz or 55MHz out of the RF gen. box.  This is why I couldn't lock any cavities last night (duh). 

On to locking!

----------------- In other news,

While swapping out the EOM cable, I noticed that the DC power supply sitting under the POX table was supplying a weird value, 17 point something volts.  I checked on the table to remind myself why that power supply is there...it's powering an RF amplifier right after the commercial PD that is acting as POP22.  The amplifier wants +15 and GND, so I reset the power supply to 15V.  We should add this to the list of things to fix, because it's dumb.  Either we need to put in the real POP22 (long term goal), or we need to get this guy some rack power, and do the same for any amplifiers for the Beat setup.  It's a little hoakey to have power supplies littering the lab.

What are "LID" sensors?  Do you mean the OSEM shadow sensors?  I'm pretty sure that's what you meant, but I'm curious what "LID" means.

I went inside to align the beam on WFS and noticed that oscillations in yaw are ~10 times stronger then in pitch. I've plot rms of pitch and yaw measured by LID sensors and saw that MC3 yaw rms motion is a few times larger then pitch.

Also MC1 input diag matrix does not diagonalize signals for pitch and yaw. In the spectrums of these signals all 4 resonance are equally seen during the free swinging. I think we should rediagonalize MC1.

Another thing is that if MC1 and MC3 are on the same stack,  pitch and yaw spectrums of these mirrors should be comparable. But MC1 signal is ~2-3 times larger then of MC3. I think we should correct calibration.

[Evan, Jenne]

We applied some volts across both the pitch and yaw pin sets of the ribbon cable that goes to PZT2.  We ended up with ~40V yaw and ~14.5V pitch.  That was the nice happy center of the clipping that we can see on the AS camera.  Once we found the center of the PZT clipping range with the ITMY beam, we recentered the AS camera (actually, this took a few iterations, but now it's good). 

We then aligned MICH, but aren't able to get it to lock.  Before falling asleep, we have decided to align the PRM and SRM, so right now we have a flashing DRMI.  Both the SRMI and PRMI look a little funny the closer you get to 'good' alignment, so I'll investigate a little more tomorrow, and include pictures.  (The video capture script has barfed again, and I'm not in the mood to deal with it today.)

[Evan, Jenne]

We were sitting trying to lock MICH (hooooorraaaayy!!!), and the emergency lights above the control room door came on, and then ~1 minute later turned off.  Steve, can you see what's up?

 The shame:

There is no situation in which it is OK to install a mount like this. Steve had installed this flaky and shaky mount to optimize the beam size on the OL QPD.

Everyone in the lab should know better. Putting in something like this is just like sabotage - it creates extra noise in our interferometer in a sneaky way and just makes locking harder. All mounts for anything useful (including QPDs) must have highly rigid mounts. 

Use the example from the PSL relayout: use the 3/4" steel mounts and the wide aluminum bases from Newport. No more art projects using home made mounting crap, Steve.

Note to self:

The ENV-40 amplifiers that we have supply -10V through +150V .... so don't exceed those limits.

piezojena link

After the fuse-blowing fiasco earlier this afternoon, Koji and I took another look at the PZT controllers.

We put an ammeter in place of the fuse, and watched the current as we turned on the transformer module.  The steady-state current with no other modules plugged in is ~15mA.  However, there is a surge current right when you turn on the box which sometimes goes as high as 330mA.  Since the fuse is 250mA, this explains the fuse blowing, even though Koji had already checked out the low voltage path.

The high voltage line was connected, with +180V to the HV out pin of the backplane connector, and the (-) terminal of the power supply connected to signal ground on the board.

We inserted the PITCH module for PZT2, and we started with ~10V as our "high" voltage, and slowly increased the value (current at this time was ~60mA).  We also had a function generator plugged into the "MOD" input, which is where the epics slider goes, so that we should see a changing output voltage.  We never saw a changing output voltage.  Increasing the HV power supply didn't help. 

When Koji spun the "DC offset" knob really fast and then stopped, sometimes the output voltage as measured on the connector-converter board between the white and red wires would jump up, and then settle back down. It came back to the same value that it always was, but it was bizzarre that it would jump like that.  We suspect that that knob is an offset for use with the closed loop setting, so it isn't relevant for us anyway.  Watching the MON output, the value never changed, even when Koji did his fancy knob twirling.

We switched to the other PITCH module, and watched the output voltage on the MON output.  This time, with the function generator unplugged, so no modulation input (so we were expecting a steady DC output voltage) the number on the LCD and the MON output fluctuated wildly.  We plugged in the function generator, and the fluctuations did not change in approximate amplitude or DC offset.  They kind of looked the same. 

So, we have concluded that (a) the PZT drivers don't work, and (b) we don't understand why.  Therefore, we don't know how to fix them.

With that in mind, we are thinking of totally circumventing the PZT drivers. 

I plugged in the PZT1 connector converter board, which has Koji's circuit that he made last time when PZT1 died.  I plugged the ribbon cable which goes to the PZT, and the +\- 30V power supply, and the PZT responded!  Just plugging in the power supply puts the PZTs near the center of their nominal range.  I then put a function generator on the epics inputs for pitch and yaw (one at a time), and saw the spot move around at the ~1Hz that I was applying.  Yay!

What I think I'll do for tonight - modify the other connector converter board so that I can just use 2 HV power supplies (current limited) to steer the PZT.  I set up a TV monitor next to the PZT electronics (1Y3? 1Y4?  I forget), and it's connected to output 20 of the video switch, so I can watch the AS camera and move the PZTs by hand.  Then maybe I can try to align some stuff. (Evan is coming to work tonight, so if I electrocute myself, someone will be here to call 5000)  Koji suggested buying 2 single-channel thorlabs piezo drivers, like we have on the PSL table for the FSS loop.  These take in 0-10V and output either 0-75V, 0-100V or 0-150V (depending on which setting you choose).  These cost $712 each. This would be a more permanent solution than me just sitting out there, since we could once again control PZT2 via epics.

 I connected a thick wire to pin 22 of the backplane connector of the transformer / power supply module of the PZT box.  This is the pin that +180V is supposed to go on, to be distributed to the other boards in the crate.  Last week I had drilled a hole in the front panel so the wire can come out (since no one on campus seems to have HV panel mount connectors in stock). 

While the transformer module was isolated, not touching anything else, I applied (slowly ramping up) 180V DC, and it all looked good.

When I plugged the module back into the crate (first turning off and disconnecting the HV), I blew the 250mA fuse again.  No HV yet, just the low voltage stuff that Koji had fixed last week.  :( 

We're now out of 250mA fuses, we're supposed to get a box of them tomorrow.

 I think state estimation and optimal control are two different techniques that are often used together. Sometimes (as for pendulum) we can use LQG without state estimation as we need only position and velocity. But for more complex systems (like quad suspension) the states of all 4 masses can be reconstructed in some optimal way using information from only one of them if the dynamics is sufficiently well known. When current system states are measured/estimated we can apply control where all our filters are hidden.

 The algorithm needs to know about expected seismic noise transmission from the ground to the optic, but it might be not very precise. I gave it some rough estimate, there are better ways to do it. I think that we'll understand whether we need state estimation or not when we'll move to more complex systems. Brett uses a similar approach for his modal control. Interesting if these methods + seismometer readings will be able to say if one of your sensors is noisier then others.

 For lack of a better idea, I keyed the crate.  The computer came back up just fine, ETMX is happily damped again.

Something bizarre-o was going on with the PMC and FSS over the weekend.  On the striptool, PMC's PZT looks like it was doing a sawtooth pattern for several hours.  I opened up the FSS screen, and the FSS SLOWDC had walked itself up to +10.  It's not supposed to get that far from 0. 

Here are some trends, so you can see what was going on.

10 day trend:  This weird behavior began ~Friday evening (FSS_SLOWDC ramps quickly).

1 day trend:  You can see the sawtooth pattern in PMC_PZT more clearly here.  It's at the same time as the FSS_SLOWDC is ramping rapidly, and the FSS_FAST is saturated.

I think the ETMX slow machine might be dead.  All of the regular FE readbacks are fine, and the c1iscex FE computer looks fine, but the slow readbacks are all whited out. 

I turned off the damping loops for ETMX, since I don't have access to the watchdog disable/enable switch.  I guess checking this out will be task #1 for Monday morning.

 I guess that the estimated state has the same low pass filter, effectively, that we use to low pass the feedback signal in SUSPOS. I wonder if there is an advantage to the state estimation or not. Doesn't the algorithm also need to know about the expected seismic noise transmission from the ground to the optic?