I took down the IFO at 717 PST for locking training. The insiral range shows that it was locked for over 12 hours. Awesome!

Based on measurements of the DARM OLTF and the EY PUM/test crossover that Jeff and I took last week, we can estimate the current value of the EY ESD actuation coefficient. It is 1.45×10⁻¹⁰ N/V², with a 20 % uncertainty. This is an 80 % increase compared to the previous value (0.8×10⁻¹⁰ N/V²) which was measured at the end of May.

This number mostly relies on the crossover measurement, since above 10 Hz, the effect in the OLTF of changing the actuation coefficient is largely the same as changing the optical gain. Additionally, this number requires us to assume a number for the PUM actuation coefficient. Here I assume that it has not changed since the May calibration (i.e., I use 7.0×10⁻¹³ m/ct at dc).

Note that the modeled crossover doesn't really agree well with the measurement above 80 Hz. More investigation is required, particularly since during ER7 we already knew there was an issue with the DARM model around 10 Hz. (This discrepancy is why I say the uncertainty in the coefficient is no better than 20 %.)

To generate this number, I took Jeff's ER7 calibration script and made a version (H1DARMXO.m) that models the crossover. Both scripts (and the parameters file for the July 25 measurement) live in the CalSVN under Runs/PreER8/H1/Scripts/DARMOLGTFs. All parameters were left the same as their ER7 values except for the optical gain (I use 1.0×10⁶ ct/m), the DARM pole (I use 330 Hz), and the EY L3 drive strength (I use 11×10⁻¹⁵ m/ct at dc). By tuning the L3 drive strength to match the measured crossover, we can extract the ESD actuation coefficient, assuming the rest of the L3 signal chain has been well-characterized. This is how I get the number quoted above.

The 80 % increase is sort of consistent with the 70 % increase that we saw when retuning the L3 digital gain post-vent. Strictly speaking, that was a measurement of the relative strengths of EX and EY. However, the DARM OLTF (with the retuned EY L3 gain) stayed roughly the constant before and after the vent, indicating that this 70% increase really is a change in EY.

Evan, Stefan

Using a line at 6409.7Hz we found a null in the frequency noise coupling. Thus we took the REFL_9 to OMC_DCPD_A transfer function (plot 1).
This is a nice modelling challenge - thus we spend some time calibrating it in meaningful units.

Calibration:
 - OMC_DCPD_9 (H1:IOP-LSC0_MADC0_TP_CH12):
   - This is just a copy of the filters in the PDA filter module, plus the 13.7kHz 17.8kHz poles from preamp (alog 17647). We did not take into account the AA board, but it should be the same for both channels:
   - G=8.63151e-07
   - P=0.87, 7.689, 7.689
   - Z=10.07, 78.912, 90.642, 13700, 17800
 - REFL9 (H1:IOP-LSC0_MADC1_TP_CH28):
   - The gain is 2900V/WattsRF (see alog 10661) x (12dB whitening) x 1638.4cts/V, plus a z=1,p=10Hz whitening filter (again, no AA board compensation)
   - G=5.2867e-08 = 1/(2900*3.9811*1683.4)
   - P=1
   - Z=10

Plot 2 shows the 9 OMC_DCPD traces during heating, calibrated in A/rtHz

The xml template is in LIGO-T1500414
https://dcc.ligo.org/T1500414

The broadband level below 8kHz didn't change, but the structure around 15kHz disappeared... (Compare to alog 19856)

Attached is a plot of the frequency noise in OMC_PDC_A at six different times during the thermal tuning. There is a significant drop in frequency noise coupling. However the feature that changes most drastically is the carrier 10-mode resonance at 5.1kHz. Thus some of this might be alignment induced effects.

Plot 2 has three more traces with even more heating.

This is a brief update of the current situation of the ISS 2nd loop.

• It successfully lowers intensity noise.
• The resonant peaks of the periscope mirror mount, which would contribute significantly to DARM, can be pushed to more than a factor of 10 below the current DARM floor, in 20-2k Hz band.

[Contribution to DARM]

Here are estimation of the contribution of intensity noise to DARM with and without closing the 2nd loop:

The left plot is an estimation without the 2nd loop closed, and the right plot with the 2nd loop closed. The estimation was done by measuring the coupling transfer function from the ISS 2nd loop PD array to the DARM error point and propagating the measured spectra of the ISS array via the measured transfer function to DARM displacement. As you can see, when the 2nd loop is open, it pollutes the DARM spectrum in 100- 400 Hz. Though it looks like I am missing a factor of two or so for some reason in order for the intensity noise to be dominant in the DARM spectrum as observed.

As shown in the right plot, the loop suppresses intensity noise and pushes it to a point more than a factor of 10 apart from the DARM noise floor. This is good. One thing I noticed is that the spectral shape of intensity noise changed depending on whether the 2nd loop was closed or not. For example, the left plot clearly shows structure in 100 - 400 Hz band in the estimated contribution of intensity noise while the right plot does not show the structure. I am still studying what is going on, but it seems like it is from some kind of readout noise of the PD array which covered the real spectral shape.

Also, I attach a measured coupling from the PD arrays to DARM displacement [meters/RIN].

As shown, it is on the order of 3-4 x 10^-14 [meters/cnts]. By the way, meters, I refer here is unsuppressed DARM displacement. The measurement was done by injecting swept sine at  H1:PSL-ISS_TRANSFER1_INJ with the 2nd loop intentionally disabled in order to get high coherence. The interferometer was locked fully in low noise and a PSL power was at 24 W at around 8:40 PT of July 31st.

[Suppression]

As mentioned above, the suppressed noise does not make sense. They look as if they are covered by some kind of readout noise.  Nevertheless, it is evident that the suppression at 300 Hz is approxiamtely a factor of 60 as expected. Good. Later, Stefan and I undid the anti-whitening in the measured spectra in order to check whether the signals were above ADC noise or not -- they were above ADC noise by a factor of 10 or so. Also, I believe that the structure in 10-20 Hz is some scattering induced noise which Gabriele reported before (for example, alog 15198).

[Next to do]

* Modeling of the measured intensity noise coupling. 

* Figure out the read out noise.

Evan, Stefan,

We set up a function generator to get a 3kHz-30kHz broadband frequency noise injection. Details:

 - Function generator, set on 'noise'
 - Followed by an SR560 (AC, band-passed between 3kHz and 30kHz, gain 1)
 - This drives drive a noise floor of about 22uVrms/rtHz at 10kHz.
 - We hooked this up the the CM board input 2 (note to self - need to reconnect SUM_B for the IFO to re-lock).

 - Next we turned on the CM board input 2 with -20dB gain: this should add about 2.2uVrms/rtHz at 10kHz.

Attached is a plot of REFL_9 and OMC_DCPD_A, with annotated features at the carrier 10 and carrier 20 modes.

We left this excitation on for the thermal run tonight.

Kiwamu, Stefan, Evan

We are trying to minimize the coupling of frequency and intensity noise into DARM by tuning the central heating on the IX CP.

The following excitations have been set up:

• Frequency excitation at 6409.7 Hz. 40 ct excitation into ALS-C_REFL_DC_BIAS (FM gain 1, all filters off). This goes into input 1A of the summing node board with a slider gain of 0 dB. We use LSC demodulator #1.1 to demodulate the signal in OMC DCPD sum.
• Intensity excitation at 2202.7 Hz. 40 ct excitation into PSL-ISS_TRANSFER1_INJ (FM gain 1, all filters off). This goes into the error point of the inner ISS loop. The outer ISS loop is off. We use LSC demodulator #2.1 to demodulate the signal in OMC DCPD sum.
• SRCL excitation at 748.7 Hz. 5000 ct excitation into SRM via LSC lockin oscillator #3. We use LSC demodulator #3.1 to demodulate the signal in OMC DCPD sum.
• Additionally, Stefan set up a broadband, analog frequency noise injection from 3 kHz to 30 kHz into input #2 of the common-mode board. This must be reverted in order to relock the interferometer.

The amplitudes were chosen so that each line has an SNR of 50 or so in OMC DCPD sum with a 10 s FFT. Each demodulator demodulates OMC DCPD sum at the appropriate frequency, and then lowpasses I and Q with a 100 mHz, 4th-order butterworth.

At 2015-08-03 01:19:45 Z we changed the IX CP heating power from 0.23 W to 0.36 W.

At 2015-08-03 02:57:25 Z we changed the IX CP heating power from 0.36 W to 0.53 W.

At 2015-08-03 04:26:20 Z we changed the IX CP heating power from 0.53 W to 0.41 W.

Additionally:

• At 2015-08-03 03:26:30 Z we turned off the broadband frequency noise injection.
• At 2015-08-03 03:45:50 Z we closed the outer loop of the ISS.

Jamie, Sheila, everyone,

Over the past several days, TJ's verbal alarams have been warning us about ETMX software watchdog trips which aren't really happeneing.  This is interseting though, since we've noticed that sometimes this seems coincident with a huge glitch in DARM that can be seen in the spectrum.  The verbal alarm script is checking the channel H1:IOP-SEI_ETMX_DACKILL_STATE.  It sometimes jumps to a value of 3 for about a second and comes back to 0. 

Three incidents from Friday night happened in the 10 to 20 seconds proceeding these times (UTC):

8/1/2015 7:37:40, 5:32:10, 4:00:00

One of these incidents a huge glitch is visible in the DARM time series before the DACKILL state changed. 

Two questions probably need further investigation, is DACKILL behaving the way we want it to, and are the glitches in DARM cauing the DACKILL state to change or is something else causing both DARM glitches and the change in DACKILL state?

Matt, Lisa, Hang, Evan

Tonight we went to full power, turned on the new boost and cutoff in dHard (along with a small lead filter around 3 Hz), and turned off the oplev damping on the ITMs. Then we took an OLTF. Blue shows the new loop with the lead off, and red is the loop with the lead on. So far we've been at full power for more than 3 hours without any sign of instability.

There is some new, untested code sitting in the ISC_LOCK guardian:

• In COIL_DRIVERS, the ISC_LOCK guardian requests the IMC_LOCK guardian to go to PREPARE_ISS.
• In LSC_FF, the new dHard filters are engaged, the ITM oplevs are ramped off, and the IMC_LOCK is requested to go to ISS_ON.

However, this new, untested code is commented out (search ISC_LOCK.py for 'guardbomb' to find it). We can uncomment it the next time there is someone in the control room to supervise the lock acquisition.

Additionally, I got impatient with damping the roll modes during the acquisition sequence, and so I have set the quad coil drivers to be high range until the COIL_DRIVERS state, at which point they are switched to low noise. This seems to work fine (i.e., I didn't notice any glitching on the cameras).

The IMC_LOCK guardian now has a state PREPARE_ISS which tunes the offset slider to bring the second loop servo board out of saturation before engaging the second loop.

This work was previously being done by the CLOSE_ISS state, but since it can take a few minutes and the offset tuning does not disturb the IFO, it can be done in parallel with other changes as soon as the operating power level is reached.  The ISC_LOCK guardian will be changed accordingly.

A secondary advantage is that the IMC_LOCK guardian can return from PREPARE_ISS to LOCKED without going through ISS_ON, which can be useful for testing.

Drift monitor thresholds updated with 120 seconds averages during lock at 1122488064 GPS,
Aug 01 2015 18:14:07 UTC.

It seems like the rate of epics freezes has increased today, I have seen more than 5 in the last 2 hours. 

Evan, Lisa

This entry is to clarify the fact that the impact of  this excess of high frequency noise  is actually bigger than the coherence with the ASC channels suggests, as it can clearly be seen by comparing OMC NULL and SUM.

For example, around 2 kHz, the discrepancy in the noise floor between OMC SUM (total noise) and OMC NULL (shot + dark noise) is about 15%, so corresponding to a noise which is 0.6 times shot + dark.

The attachment shows OMC SUM/NULL in H1 at low noise (left) compared to L1 (right). 

So, the message is that we are looking for something quite big here..

After 17:30 UTC the interferometer was not undisturbed: I was making PEM injections.

Matt, Lisa, Evan

Tonight we looked at the coherences between the OMC DCPD channels and ASC AS C, this time at several different interferometer powers. In the attached plots, green is at 11 W, violet is at 17 W, and apricot is at 24 W.

Evidently, the appearance of excess high-frequency noise in OMC DCPD sum (and the coherence of OMC DCPD sum with ASC AS C) grows as the power is increased. We believe that this behavior rules out the possibility that this is excess noise is caused by RIN in the AS port carrier, assuming that any such RIN is independent of the DARM offset and of the PSL power. Since the DARM offset is adjusted during power-up to maintain a constant dc current on the DCPDs, RIN in the AS carrier should result in an optical power fluctuation whose ASD (in W/rtHz) does not vary during the power-up. This is the behavior that we see in the null stream, where the constant DCPD dc currents ensure that the shot-noise-induced power fluctuation is independent of the PSL power.

We observed broadband coherence of OMC_DC_SUM with ASC_AS_C_LF_SUM and ASC_A_RF36_PIT. We made some numbers and plots, using the 64kHz version of the channels.

First the measurements we made on OCXO oscillator:
- ASC_AS_C sees a RIN of about 5e-7/rtHz above 100Hz (either from H1:ASC-AS_C_SUM_OUT_DQ or from H1:IOP-ASC0_MADC6_TP_CH11). The same is true for its segment 1.
- The calculated shot noise RIN at 20mA (quantum efficiency 0.87) detected is 4.0e-9/rtHz.
- The 4.0e-9/rtHz agrees with DCPD_NULL_OUT_DQ's prediction (8.0e-8 mA/rtHz/20mA).
- DCPD_SUM_OUT_DQ sees a slightly elevated RIN of 4.6e-9/rtHz (9.2e-8 mA/rtHz/20mA).

- The RIN in DCPDA (H1:IOP-LSC0_MADC0_TP_CH12, corrected for the whitening) is about 5.9e-8 mA/rtHz, or RIN = 5.9e-9/rtHz at 20mA/2diodes (~15pm DARM offset)...
- ...or about 3.3e-8 mA/rtHz or 1.2e-8/rtHz at 5.7mA/2diodes (~8pm DARM offset).

- ASC-AS_C_SEG1 (H1:IOP-ASC0_MADC6_TP_CH11) and OMC-DCPD_A (H1:IOP-LSC0_MADC0_TP_CH12) shows a coherence of 0.053 at 20mA, suggesting a white noise floor a factor of 0.23 below shot noise.
- At 5.7mA the same coherence is about 0.13, i.e. the white noise floor is a factor of 0.39 below shot noise.
- These two measurements are in plot 1.

- Taking the last two statements together, we predict a coherent noise of
  - 5.9e-8 mA/rtHz *0.23 = 1.4e-8 mA/rtHz at 20mA/2diodes (~15pm DARM offset)  (RIN of coherent noise = 1.4e-9/rtHz) - The pure shot noise part is thus 5.7e-8 mA/rtHz
  - 3.3e-8 mA/rtHz *0.39 = 1.3e-8 mA/rtHz at 5.7mA/2diodes (~8pm DARM offset)  (RIN of coherent noise = 4.5e-9/rtHz) - The pure shot noise part is thus 3.0e-8 mA/rtHz.

- AS_C calibration:
 - 200V/W (see alog 15431)
 - quantum efficiency 0.8 (see alog 15431)
 - 0.25% of the HAM 6 light (see alog 15431)
 - We have 39200cts in the AS_C_SUM. Thus we have
   - 39200cts / (1638.4cts/V) * 10^(-36/40) (whitening) / (200V/W) = 1.89mW and AS_C. (shot noi
   - 1.89mW/0.025 = 76mW entering HAM6. I.e. we have slightly more sideband power than carrier power (Carrier: 27mW in OMC transmission).
   - Shot noise level on AS_C_SUM is at 2.0e-8 mA/rtHz, corresponding to a RIN of 1.6e-8/rtHz. I.e. the coherent noise seen at 5e-7/rtHz is high above the shot noise. Dark noise TBD.
   - The light entering HAM 6 has a white noise of 5e-7/rtHz*76mW = 3.8e-5 mW/rtHz 
    

Bottom line:
 -We have ~1.4e-8mA/rtHz, or 1.9e-8mW/rtHz of coherent white noise on each DCPD.
 -It corresponds to 3.8e-5mW/rtHz before the OMC, i.e. the the OMC seems to attenuate this component by 2000.
 -This noise stays at the same level (in mW/rtHz) for different DCPD offsets.


Next, we switched back to the IFR for testing. plot 2 shows the same coherences (all at 5.7mA / 8pm DARM offset), but on the IFR. Interestingly now AS_C and AS_A_RF36 start seeing different noise below 2kHz. We convinced our selfs that the higher excess noise seen in AS_A_RF36 is indeed oscillator phase noise from the IFR - so that is clearly out of the picture once of the OCXO. (Evan will shortly log the oscillator phase noise predictions.)


64k Channel list:
H1:IOP-LSC0_MADC0_TP_CH12:     OMC-DCPD_A  (used in plot)
H1:IOP-LSC0_MADC0_TP_CH13:     OMC-DCPD_B
H1:IOP-LSC0_MADC1_TP_CH20:     REFLAIR_A_RF9_Q
H1:IOP-LSC0_MADC1_TP_CH21:     REFLAIR_A_RF9_I
H1:IOP-LSC0_MADC1_TP_CH22:     REFLAIR_A_RF45_Q
H1:IOP-LSC0_MADC1_TP_CH23:     REFLAIR_A_RF45_I
H1:IOP-LSC0_MADC1_TP_CH28:     REFL_A_RF9_Q
H1:IOP-LSC0_MADC1_TP_CH29:     REFL_A_RF9_I
H1:IOP-LSC0_MADC1_TP_CH30:     REFL_A_RF45_Q
H1:IOP-LSC0_MADC1_TP_CH31:     REFL_A_RF45_I


H1:IOP-ASC0_MADC4_TP_CH8:      ASC-AS_A_RF36_I1
H1:IOP-ASC0_MADC4_TP_CH9:      ASC-AS_A_RF36_Q1
H1:IOP-ASC0_MADC4_TP_CH10:     ASC-AS_A_RF36_I2
H1:IOP-ASC0_MADC4_TP_CH11:     ASC-AS_A_RF36_Q2
H1:IOP-ASC0_MADC4_TP_CH12:     ASC-AS_A_RF36_I3
H1:IOP-ASC0_MADC4_TP_CH13:     ASC-AS_A_RF36_Q3   (used in plot)
H1:IOP-ASC0_MADC4_TP_CH14:     ASC-AS_A_RF36_I4
H1:IOP-ASC0_MADC4_TP_CH15:     ASC-AS_A_RF36_Q4

H1:IOP-ASC0_MADC6_TP_CH11:     ASC-AS_C_SEG1  (used in plot)
H1:IOP-ASC0_MADC6_TP_CH10:     ASC-AS_C_SEG2
H1:IOP-ASC0_MADC6_TP_CH9:      ASC-AS_C_SEG3
H1:IOP-ASC0_MADC6_TP_CH8:      ASC-AS_C_SEG4