Around 2:30 on August 7, 2012 I attached 5 grounding plugs to the "TEST IN" connections at the Y end station.  At this time, all ITMY and ETMY coil driver boxes (for TOP, UIM, and PUM) have grounding plugs connected to each of the TEST IN connections.  I then took some power spectra measurements of the coil driver NoiseMonitor channels to see what effect the grounding plugs have on the coil driver noise.

Starting with ETMY, the UIM driver board didn't exhibit any significant change between COIL IN and TEST IN configurations.  This is because the noise in the coil drivers is dominated by component noise, and selecting the TEST IN state allows one to see the component noise without the added DAC noise.

The ETMY PUM driver board is dominated by DAC noise, since the PUM driver is not filtered as much as the UIM board in the lowest noise mode.  Therefore when we switch to the TEST IN state, the expectation is that the noise in all coil drivers should drop to the component self-noise level.  This is what we are seeing now, as compared to Jeff K's alog #3111 when the noise levels increased and the frequency dependence was different for those channels.

https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=3111

For the ITMY, the UIM behavior is generally the same as on ETMY coil drivers.  However, the ITMY PUM driver shows an increase in noise after switching to the TEST INPUT with grounding plug attached.

Due to a calibration error, the dust monitor at location 15 in the LVEA (in the clean room over HAM3) was swapped. It was removed around Aug. 8 17:20 UTC and the new one put in around Aug. 8 18:01 UTC.

This is the transfer function between the phase-frequency discriminator and the FET IQ demodulator while the cavity is locked. The blue trace is "initial" conditions, i.e., with a well-aligned cavity and standard modulation sidebands. The green "sb shift" trace was taken with the RF modulation sideband frequency tuned 200 Hz higher than its initial value. The red "alignment shift" trace was taken with the ITM yaw misaligned. 

The FSR is 37.512 kHz, with a higher-order FSR peak at 75.018 kHz. Given this FSR, the cavity length is 3995.95 m.

The peaks due to the RF modulation sidebands are at 54.432 kHz and 58.098 kHz in the initial trace. These peaks shift 200 Hz in the green trace.

Additional structure is observed at 46.301 kHz, 55.325 kHz, 57.252 kHz, and 66.229 kHz. The peaks at 46.301 kHz and 66.229 kHz increase for the case of yaw misalignment and are thus likely to be (1,0) modes. 

The separations between these peaks are:

75.018 - 66.229 = 8.789 kHz

66.229 - 57.252 = 8.977 kHz

46.301 - 37.512 = 8.784 kHz

55.325 - 46.301 = 9.024 kHz

The mean modal spacing is 8.8935 kHz.

From this modal spacing, the g-factor is 0.540532, corresponding to approximate mirror radius of curvature of 2302.86 m. 

The FWHM of the FSR peak is approximately 94 Hz. Finesse is thus 37512/94 = 399.064. This would occur for a reflectivity of about 99.2%.

There's a follow-up of this in the works for characterizing cavity properties during ring heater use.

I turned on the ETMY RH, requesting 630mA from each ring heater segment. This equates to roughly 30W of power input into the ETM. I'm monitoring the Hartmann sensor signal to check for a response.

The ITMY YAW was set to zero (from 717.81 counts) to keep the cavity from resonating or flashing during this measurement.

The time constant of the ring heater is several hours, so we will run this at least until noon.

I locked the arm and did a dither alignment (>8000 on the REFL diode) and engaged the off laod at 23:10h local time, so good times are from 23:15pm.

With the work on the RefCav I will be keen to see the noise at higher frequencies. Vincent reinstalled the boosts on the ISI isolation, so all should be nice.

Should be good for the night.

Attached are plots of dust counts > .5 microns in particles per cubic foot. The dust monitor in the clean room over HAM 3 (location 15) has a calibration failure. I'm not sure when it started, but I think it was sometime today.

I was only able to test ~20 joints today on the Vertex due to leaks.  Aaaargh!  At this rate, I'll need two or more additional days to finish!  A few of the leaks were on "legacy" joints circa 1998 but most were on new aLIGO joints.  I was able to retorque and get most of these to seal but this wastes time and is "not on the schedule" - better to do it right the first time!

During the reboots I took the opportunity to have another look into the OAT RefCav in the optics lab.

I removed the 50/50 beam pslitter infront of the RF LSC diode, so now ~30 mW of power is incident. In addition in optimised the HWP infront of the PM to reduce the AM, 2f_sb = < -80dB.

The few irises in the beam path are closed down, I looked with the viewer and reduced the iris so it was just not clipping.

I didn't find a low enough ND filter to o the paracitic interfernce hunt.

The cavity locks stable with a UGF ~300kHz, the settings are

The Laser temperature on the controller displays: T+40.991, the slow output (T-ed to the laser slow input and the oscilloscope) is at -620 mV.

Once we get the cavity locked again we can see if we gained anything.

I put the boost filters back in FM10 of the ISO filter bank on BSC6 and BSC8. They switch when crossing zero but I have changed the tolerance (300 to 50 counts) and the time out (2s to 20s).

Elli and Bram

We went to the EY and injected a sine wave at 2280 Hz, 200 mVpp into the CMB-B (PDH), Exc A input. Then we used ddt to measure the tranfer coefficients of the SegX_Ix/SegX_Qx. We optimised the demodulation phase and maximised the transfer coeeficient. Although we could not zero the Q-phase, we got a factor of ~30x between the I and Q phase for all segments in both WFS-A and WFS-B.

in between the measurements we aligned the beams onto the WFS but adjusting the picomotor mounts. We actually did this manual and did NOT use the picomotor driver, which would be nice in the future ...

Now we can use the WFS for alignment feedback to the Quads, well at least measure the TFs.

Attached are plots of dust counts > .5 microns in particles per cubic foot.

I attached 5 grounding plugs to the ITMY coil drivers Test Input this afternoon ~ 4:30 pm.  Three were connected at the top stage driver, and one each for UIM and PUM drivers.  I plan on doing the same for ETMY tomorrow before making measurements of the coil driver noise.

(Betsy, Travis, Deepak, Cheryl, Dale on the camera)

All the prep work paid off and the MC2 was installed into the chamber relatively easy this afternoon.  Using the genie, we hoisted it up to the elevator which was attached to the arm, attached the spacer to the bottom of the suspesnion, and then used the arm to swing it into place on the ISI table against the cookie cutter.  We added a few dog clamps to make sure it doesn't walk away overnight.  Tomorrow we continue adding dog clamps, hooking up cables, etc.

Report of site activities:

• Betsy B. Corey G. and others in to the LVEA to move cleanrooms.
• Cleaning crew Christina C. and Karen R. to LVEA to clean HAM03 per WP#3378.
• Corey G. to the LVEA, to do HAM03 accounting of weights.
• Kyle R. and Apollo crew  to the LVEA, H1 output tube to torque GNB flanges.
• Apollo crew to the LVEA, surveying.
• Patrick T. to the LVEA, check dust monitors around HAM03 area.
• Filiberto C. to the LVEA, add some electronics for TCS ring heater.
• Dale I. to the LVEA, film activities around HAM03.
• Hugo and Mitchel to the LVEA, to do inventory of cables around HAM03.
• Jodi F. to Y-End and LVEA, take photos.

During the past few weeks I have been working on calibrating the sensor readouts from the Quad Suspension OSEMs on all four stages: main chain top M0, reaction chain top R0, upper intermediate mass (L1 stage or UIM), and penultimate mass (L2 stage or PUM).  One way to check the OSEM readouts is to measure their response to an excitation of the top mass, and then compare this to the response predicted by the 20120601TMproductionTMrehang mathematical model.  The Quad Suspension has a 0.43 Hertz mostly "Longitudinal" mode, and exciting this mode by shaking the main chain top mass will result in peaks at 0.43 Hz for all OSEM sensor spectra (M0F1 spectrum attached as an example).  Comparing the peak height data of each OSEM at 0.43 Hz to the data predicted by the model should show any dissimilarity between the predicted behavior of the OSEM sensors and what is actually seen.  

The file OSEMsensorplot.pdf shows the magnitude of the OSEM sensor readouts (vertical axis) for each OSEM ID(horizontal axis) scaled so that the magnitude of the Main Chain F1 OSEM is always one.  I tried a few different methods of measuring the power spectra that generate this plot, and they are labeled on the side of each plot.  In each case damping of the top stage reaction and main chains was OFF.

The model prediction is the sensor response at the 0.43 Hz peak for each OSEM derived from the mathematical model of the quadruple suspension.  The measurements labeled "White Noise" are from excitations of white noise into all Main Chain TOP degrees of freedom (Longitudinal, Transverse, Vertical, Roll, Pitch, and Yaw).  These white noise power spectra were collected WHILE the excitation was being performed.  In the 7/3/2012 White Noise measurement the coherence in the L2 stage OSEMs was terrible, and so these data points are not reliable.  The measurements labeled "Offset" are derived from an excitation where each main chain TOP degree of freedom was displaced with a static offset value after a turn-on ramp time of 10 seconds, then this static offset was instantly changed to zero to let the suspension swing freely.  Lastly the measurements labeled "Sine Relax" were taken by exciting the main chain TOP mass Longitudinal DOF with a 0.43 Hz sine wave drive for about five minutes, then the drive was turned off and the suspension allowed to swing freely for about 2 minutes before the power spectra were taken.

The sensor plot shows that the different measurement approaches lead to consistent OSEM data, but we see immediately that the lower stages have some problems.  The L1 stage magnitudes on each suspension appear slightly lower than predicted, but only ITMY OSEM L1LL is way off.  This could be a problem with the L1LL alignment or an electronic issue that fails to correctly communicate the position of the flag on that sensor.

The L2 stage is troubled with multiple symptoms, each OSEM is well below the model prediction.  I am still struggling to understand why the response is so small.  One glaring problem is that the L2UR sensors on both ETMY and ITMY are practically zero compared to the other L2 OSEMs.  I suspect that these specific sensors dead or barely working, but will need to check this somehow.

I replaced the H2GDSAWGL0 model with H2TCSITMY on h2tcsl0 following the install of a second ADC card in this front end. The front end now has 2 ADC and 1 DAC cards. PEM uses the first ADC, TCS uses the second ADC and the DAC card.

The H2 DAQ was restarted with the new configuration.

The h2tcsl0 restart invalidated all the front end data. After the DAQ restart I had to restart the data streamers on each model, so the H2 data gap was longer than anticipated at about 10 minutes.

Jeff K. , Thomas V.

Below are the calibration parameters for the H2 ITMY Optical Lever:
        Slope          Y-intercept
Pitch [ 580.41229822  -11.51774017]
Yaw  [ 689.70532274  -13.3345144 ]


And here are the calibration parameters for the H2 ETMY Optical Lever:
         Slope           Y-intercept 
Pitch [ 1666.80728788   -33.14694715]
Yaw  [ 1727.67131855     23.41350328]

Both sets of calibrations were attained via the same process of moving the QPD along a translation stage and measuring the output signal.  All four sets of slopes are in units of radian*meters.