It's lossy. 3%  0.16% 

Final correction for MM subtraction. Now subtracted MM off the on resonance dip. 

1.06%  LOSS

[Edited after some input from Terry, Sheila and Daniel]

Calculation:

rloss = 0.99467

Amplitude loss t_L = sqrt(1-rloss) = 0.1031 

Power loss L = 1-rloss^2= 0.0106

Here's what we have:

70% mode matching calculated from transmission (00 peak/all peaks = 0.06/0.0856 = .7) 

Mode mismatch power = (V off resonance)*(1-.7) = 0.108 V

This yields our green dip to be 89.6% of the off-resonance power ( (0.36-0.108) - (0.134-0.108) ) / (0.36-0.108)

Hang, Gabriele, TVo, Amber, Stefan, Danny

This morning we measured, using the AS_A and AS_B DC SUM, MICH dark and estimate 9 counts with 50 mW CO2 central heating on ITMY and 8000 counts for MICH bright (with no CO2) which leaves us with about 0.1% contrast defect.

We followed up the measurement by estimating the amount of sideband power we could be seeing :

Where P_f is the power of the sidebands of frequency f at the Michelson output,  P_o is the input power to the Michelson, Gamma_f is the modulation depth of the sideband of frequency f, and t_f is the amplitude transmission coefficient of the sideband frequency f for the Michelson.

P_9MHz: 3.77*10^(-2) counts (0.4% of what we see when dark)

P_45MHz:  1.3 counts (14.4% of what we see when dark)

Hang and TVo also thought about low frequency movement of the beamsplitter and how this could introduce higher order misalignment modes. We looked at the power spectrum of the beamsplitter oplevs below 0.5 Hz and estimate 10*10^(-9) radians (in both pitch and yaw) is dominant at those frequencies. We estimate the amount of HOM misalignment mode that could be contributing: 

Where P_o is the input power to the Michelson, alpha is the beamsplitter misalignment (in rad), w(z) is the beam size at the misaligned optic, and lambda is the wavelength of the carrier. (The factor of 2 is added to include both pitch and yaw)

P_HG01 = 14.4*10^(-2) counts (1.6% of what we see when dark)

Craig, Sheila

Craig moved pico motors on TMSX M14 and M4 (D1000484) so that the beams are now centered on the QPDs with our current alignment which gives a recycling gain of about 32.  

I attempted to do this for the TMSY picos, but ran into some kind of bug.  

The worst feature of this bug is that you can select a current motor, and it looks like you have selected it if you are looking at one of the single motor screens.  In the screen shot attached I circled in red that the current motor is 3, but you can see from the green dots above motor 1 (these channels are actually readbacks of the LEDs on the controller, so they are reliable about which motor is the current motor).  

However, if you opened the single motor screen and start clicking on the buttons to move the motor you will move the wrong pico.  

I did this, moving M3 which is the steering to the green QPD by mistake.  We unshuttered the green beam for the y arm, I tried to reset the pico, and then Craig and I reset the QPD offsets to zero the green WFS signals so that our green reference is still well set.  After this the camera spot is in the reference location, as expected.  

We have tried many combinations of pushing buttons to try to change which motor is enabled, but we can't change the led lights right now.  We will try power cycling the controller next time we loose lock. 

[LHO Commissioners]

We were able to fully transition to DC readout, and we sat there for about 1 hour, from 22:14:00 to 23:12:00 UTC on 23Aug2018.  Very exciting!  We are on the BNS range board with a small, but non-zero number!

I removed AA chassis S1300104 to repair channels 2, 7, and S1001053 channel 29. In all 3 cases the 8627 op amp at the input had to be replaced. The units are now fully functional and they have been returned to service. WP 7788.

Cheryl requested that the image resolution for cam28 and cam29 be changed from mono12 to mono8 to reduce the size of the resulting tiff image file. The file size was reduced from 1.2M to 0.3M after this was done. Also the tiff image file (and gstreamer annotation) name was changed to replace spaces with underscores and remove brackets (Cheryl is using matlab to process these image files and matlab does not like spaces and brackets in file names).

INI files were changed, loaded and committed to SVN.

Here is the effective change to cam28's image file:

-rw-r--r--  1 controls     controls    1.2M Aug 23 13:55 H1 IO GIGE1 (h1cam28)_2018-08-23-20-55-25.tiff
-rw-r--r--  1 controls     controls    301K Aug 23 15:26 H1_IO_GIGE1_h1cam28_2018-08-23-22-26-28.tiff

We tried to see how well we could model the DHARD OLTFs, and it seemed that my knowledge on the suspension model was not sufficient to give satisfactory results...

In the first two plots we show the modeled OLTFs for PIT and YAW in blue traces. As a reference, we also show the measurement data in the orange crosses. The measurements can be found at /ligo/svncommon/IscSVN/iscmodeling/trunk/ALIGOH1/ASC_loops/Measurements/DHARD/DAHRD_P_OLG_SS.xml and DHARD_Y_OLG_SS.xml. For PIT I took the REF4 curve and YAW I took the REF3 curve. The third and forth plot showed the modeled M0/L2 crossover.

For the SUS model I directly ran the generate_QUAD_Model_Production function in the SusSVN with default damping filters. The digital suspension filters (LOCK_P or LOCK_Y) we used were L2: FM2, FM3, L1: FM5, M0: FM3, FM4, FM5, FM7 for PIT, and FM3, FM4, FM5 for YAW. The Nm/ct actuator strength were 1.28e-6 [Nm/ct] for M0, and 6.32e-10 [Nm/ct] for L2, based on T1100378.  We don't have a good calibration of the optical gains at this point so we just scale the model to have UGF at ~ 4 Hz. It did not seem the model could match very well to the measured data in the 1-3 Hz band (second and third sus resonances).

Moreover, for PIT, the SUS model predicted the main P2P sus resonance was at 0.525 Hz. Yet this cannot be true because if so our current control filter (with a pair of zeros at 0.65 Hz to invert the sus resonance) could not be stable. Also in the ISIFF measurements the main P2P resonance seemed to be at ~ 0.57 Hz.

If we want to optimize the DHARD loops in the future, we might need to have some good broad-band excitations to measure the OLTF, resolving at least the first and second sus resonances.

Here are measured carrier powers for different DARM offsets while we are locked on AS45Q, these values for DARM offset are entered in DARM1_OFFSET.  

There is a small assymetry. 

Sheila, Craig

- We reset the green QPD offsets and green ITM camera positions while sitting on Full IFO with five ASC loops closed (CHARD, DHARD, MICH, PRCL2, INP1).  Pics 1 through 3 show those new settings.  We accepted them to SDF.

- DHARD_Y was oscillating at about 0.1 Hz, so we took a DHARD_Y TF and found the gain was about a factor of 2 low, with UGF ~ 3 Hz.  DHARD_Y gain was increased from -40 to -80, and we retook the measurement, giving a UGF ~ 4 Hz (Pic 4).  This worked really well for stabilizing the buildups (Pic 5), so we added it to guardian.

We are having trouble with the DC_READOUT transition
- Our RF DARM sensor is ASC-AS_A_RF45_Q_SUM_NORM (hereby called ASQ).  We are trying to transition to LSC-OMC_DC_OUT (now called OMC).  If we use DARM noise to take the TF between the two sensors, we get something that looks like Pic 6, with a 1/f dependence for ASQ/OMC at high frequencies.  We see ASQ WFS sensor noise impressed upon OMC.
- So we have to take a driven DARM TF measurement to get the true TF between ASQ and OMC.  We got a gain of about 58 with no sign flip.
- Since the ASQ sensing matrix element is -1e-6, we expect the OMC matrix element to be -5.8e-5
- We then tried a 1/2th transition, which is where you blend your error signals using 1/2th of the sensing matrix element you'd expect for the new sensor and 1/2ths of the old sensor's matrix element.  So we put in -0.5e-6 for ASQ, and -2.9e-5 for OMC.
- When we did this, the OMC transmission increased, indicating to us that our DARM offset had increased.  Our DARM offset in LSC-DARM1 is 9e-5.
- We measured our DARM TF and found the gain was increased by 1.47.  
- To maintain the same DARM gain, we calculate our OMC matrix element should be -2e-5 when completely controlled using OMC DCPDs.
- So we redid the half transition, yielding Pic 7, which matched our RF DARM OLG very well, but the power on the DCPDs increased to about 30mA from about 11mA before we started to transition.
- However, we are still not sure why our DARM offset should become larger when the DARM OLG remains the same.  We have ensured that RF DARM cts = DC DARM cts.  Sheila suspected the dark offsets of ASQ, but after a bunch of OMC scans at different DARM offsets it seems like DARM1 offset of zero really is zero.
- In our last DC READOUT transition attempt, it seemed like we didn't have any real feedback from the OMC DCPDs to DARM.  We changed the OMC matrix element, but saw no change in the DARM OLG.  This could explain this DARM offset business, but before we were able to maintain the gain, so unclear what's happening.

We have turned the CO2X laser on at 1 W requested power, which resulted in about 1 W of CO2, at 10 UTC.  We also started the templates that Thomas left on ZOTWS10

We had an earthquake at about 3UTC.  We changed the ISI configuration to LARGE_EQ, and brought it back to windy at about 5 UTC. 

We have been sitting here with the green arms locked waiting for things to settle down more.  The strange thing is that the X arm is moving 5 times more at the EQ frequencies than the Y arm is (see attached screenshot, the yellow and brown traces are both calibrated into um and shown on the same scale).  I don't see anything different with the seismic configurations of the two arms. 

During today's lock we did a quick MICH PITCH OLTF measurement after the 'ENGAGE_REFL_POP_WFS' state.

Please see the first image for the OLTF with UGF ~ . The blue trace is predicted OLTF based on the SUS model and the ctrl filter banks. The orange trace is the measured data. There seems to be some small discrepancies at the sus resonances at 0.47 Hz and 1.1 Hz. Other than that the model matched to the measurement pretty well. We might need to compensate for the 1.1 Hz peak better if we want to increase the MICH loop BW.

In the second plot we show the same OLTF model (blue), and the M1 (orange) and M2 (yellow) stages' contribution. The crossover freq ( using the convention that |M1_P2P| = |M2_P2P|) happens at ~ 0.03 Hz.

At our current crystal position (unoptimized), we have a threshold power of 12 mW. Sheila and Terry measured 7.4-8 mW in March (41150). We can do better with the crystal position.

How I took the measurement:

1) Shuttered the pump, scan the seed to get an unamplified power. I also lowered the seed power going into the fiber coupler hoping to mitigate the crystal heating an green up-conversion.

2) Scan and set temperature for co-resonance (pump shutter opened).

3) Lock the OPO and optimize the non linear gain with temperature while scanning the seed PZT. Record the highest power, repeat for various green input power.

To figure out how much is actually going into the cavity I corrected green refl power (off resonance) for 30% mode mismatch and corrected for 47% intracavity loss found in alog43601. I was hoping that I was wrong about 47% loss but when I tried to propagate pump power from ISCT6 to SQZT6 given the known fiber transmission (alog43166), faraday isolator loss (~4%) and known rejected power the number falls in place with 47% loss. Note to self: forget this. They're not the same thing.

Currently the maximum green power can be sent to the coupler on ISCT6 is 28 mW. The SHG power monitor diode reported ~50 mW. ~20mW is lost between SHG and the coupler. I highly suspect that pointing to the EOM might be bad (faraday in the path has quite a large hole).

When locked OPO at 28mW (-6.2% from a pick off BS to SHG launch PD to be exact) of green input to the coupler the (PZT) pointing was off. The loop was locked to the point where green transmission wasn't optimal. This could not be fixed by temperature alone, PZT pointing  offset had to be adjusted. 

Since we have been operating at a gutless nlg region for pump power this might explain why we haven't been seeing squeezing. Once alignment is fixed on the ISCT6 we should test and optimize OPO loop stability with higher green power along with nominal operating CLF power. 

We have seen one of our bounce modes, which we hadn't seen until today since the installation of the bounce mode dampers.  

The mode we see is at 9.726Hz, and seems to be ITMX.  

Jim W suggested that ITMX might be the problem since the suspension was left damped overnight while the ISI was isolated, which we know causes problems.  I tried the same phase as the O2 damping: -90 degrees, gain of 1

Blades were selected for the baffle, best ones, but discovered that the hole on blades needs to be cleaned up (deburr) to be able to complete the assembly.  Tools are in the oven, need them to be "class B".