Reports 1-1 of 1 Clear search Modify search
AdV-ISC (Commissioning up to first full interferometer lock)
casanueva, valentini - 23:15 Friday 19 March 2021 (51156) Print this report
ISC shift report [10.8 engage the automatic alignment for the needed DOFs, 10.5 study of the 3f signals]

The aim of the shift was to study the relative gain between the 1f and 3f error signals used for the CITF lock and  to tune the quadrants for the CITF alignment.

After the tuning of the B2 PDs and QDs' VGA gains perfomed by Gouaty and Masserot (see entry 51154)  the DRMI CITF  was locked at 16.00 UTC.

Figure 6 shows that the B2_PD1 photodiode is affected by numerous noise lines, suggesting a possible start of saturation of the preamplifier. Figure 7 Shows the the B2_169 MHz signal amplitude when the CITF is locked, Figure 8 shows that the B2_PD1 and B2_PD3 signals did not saturate the ADC during the shift.

The first activity focused then on injecting lines on the citf LSC loops above the UGF frequency to check of the 1f and 3f signals, as follows:

  • 16.16 UTC: PRCL injection. Freq: 45 Hz, amplitude 5e-4. The measured ratio between the B2_6MHz_I and B2_18MHz_I error signals resulted 7e-3 (See figure 1).
  • 16.35 UTC: MICH injection. Freq: 30 Hz, amplitude 5e-1. The  demodulation phase of the 3f signal B2_169MHz required tuning, the new phase is 1.3 rad. The ratio betwen the B2_169MHz_I and B2_56MHz_I signals resulted in -4e-4.(See figure 3)

During the MICH injection, a strong coupling between the B2_56_MHz_I and  B2_56_MHz_Q (used respectively for the MICH and SRCL lock) signals has been noted ( see figure 2). Therefore, the B2_56MHz PD phase was retuned (while injecting a line on SRCL). The phase of the B2_56_MHz_I photodiode has been changed to 1.15 rad. Some attempts to change the MICH driving matrix in order to furthermore reduce the MICH-SRCL coupling (while injecting a line on MICH) has been attempted without obtaining substantial improvement. The MICH-SRCL coupling is very high, this needs to be better understood.

At 17.20 UTC the PR alignment was also retuned to increase the lock stability.

  • 18.05 UTC: SRCL injection. Freq: 25 Hz, amplitude 2e-1. The ratio betwen the B2_169MHz_Q and B2_56MHz_Q signals resulted in -3.6e-4 (see figure 4).

At this point an unexpected oscillation at 11 and 38 Hz  affecting all the CITF error signals (see figure 5) brought the loops close to saturation, requiring us to disengage the CITF lock at 18.13 UTC.

The CITF was relocked at 18.34 UTC.  in order to proceed with the second activity, focused on studying the B2_6MHz QPDs. The study of the B4 QPDs was instead not possible due the impossibility of engaging their corresponding galvo loops.

Figure 9 shows that the B2_6MHz QPDs are affected by an unexpected noise below 20 Hz.

Since the CITF lock was very unstable at this point, a complete realignment of PR and SR mirrors was performed  and the CITF was relockeda gain at 20.40 UTC.

At 20.50 UTC the B2_6MHz QPDs demodulation phases were then tuned by applying offsets to the PR ty and PR tx signals, as shown in figure 10:

  • QD1 V 0.78 rad QD1 H 0.78 rad
  • QD2 V -0.2 rad QD2 H -0.3 rad.

The injection of a 3.5 Hz line  on the PR tx and ty DOFs to calibrate the QPD's response was unsuccesful as it caused unlocks before being noticeable on the quadrant or optical lever signals. We were trying to use the usual line in the DSPs, however, when increasing the amplitude, some movement was visible on the LCs but it was definetly not a line. Probably we missed some logic step in the DSP.

However, we tried to use the 0 of the quadrants to align "by hand" the PR. The working points given by the quadrants demodulated signals actually bring the powers up, and to a good beam superposition in the camera, with a precision around 0.1urad (at least by hand) so at least the DC part seems to be a good error signal. The noise dominating the signals need to be understood if we want to use the signal up to higher frequencies.

After this, the PR and SR mirrors have been put in the MISALIGNED state and the arm cavities have been realigned and locked to the IR signal.



Images attached to this report
Comments to this report:
ruggi - 11:00 Sunday 21 March 2021 (51157) Print this report

The lock of the arms seems engaged without drift control (the switch sent to the DSP is 0). This morning a manual alignment has been performed, but the drift control is still off.

Images attached to this comment
bersanetti - 11:44 Sunday 21 March 2021 (51158) Print this report

The main ASC_ENABLE flag was set to zero after a process restart and was not set back to 1; I did that manually and now the drift control is working again.

mantovani, casanueva - 12:41 Monday 22 March 2021 (51167) Print this report

The DC of the QPD error signals seems to reproduce well the PR misalignment, as it is visible for the vertical signals of QPD1 during a PRtx alignment figure 1.

Images attached to this comment
casaneuva, gouaty, masserot, was - 16:17 Tuesday 23 March 2021 (51183) Print this report

We have a look at the numerous lines present on the B2_PD RF channels for these 2 periods

  • 2023-03-18-21h53
    • ITF in DRMI mode : MICH/SRCL with B2_PD3_56MHz and PRCL with B2_PD3_6MHz
    • B2_DC at 1mW per photodiode
    • no saturationat the ADC level
  • 2023-03-19-16h07
    • ITF in DRMI mode : MICH/SRCL with B2_PD3_56MHz and PRCL with B2_PD3_6MHz
    • B2_DC at 10mW per photodiode:  B2_PD2 has its shutter closed
    • thank to the VGA tuning, there is no saturation at the ADC level

In the following plots, we shows the spectrograms of the SIB2_B2_PD1_sample and SIB2_B2_PD3_sample channels

  • Blue: when the ITF has only its arms locked with B2_DC at 10mW
  • Purple: ITF in DRMI mode  with B2_DC at 10mW the 2023-03-19-16h07
  • Brown:  ITF in DRM mode with  B2_DC at 1mW the 2023-03-18-21h53


In the ITF_DDM.xlsx file for the B2_PD1 photodiode, we tried

  • to identify each line
  • to report it ASD value in mW/sqrt(Hz) .
  • and the ratio between the 10mW case and the 1mW one

One can see that

  • the same lines spaced by 2.01MHZ (F8MHz - F6MHz ) are present whatever the B2_DC power: 1mW  or 10mW
  • for most of the lines, theirs amplitudes increased by a factor close to 10 . If there was any harmonic distortion in the readout chain, one would expected to see the reduction of some line amplitudes : it 's not the case
  • The amplitude of the line at 6 MHz measured when the arms are locked (blue curve) is a factor 10 smaller on the photodiode B2_PD1 than on the photodiode B2_PD3. This is due to the effect of a notch at 6.5 MHz that was placed on the B2_PD1 preamplifier.
  • In the DRMI configuration, the ratio between B2_PD1 and B2_PD3 at 6 MHz changes: the height of the line on the B2_PD3 photodiode is smaller due to the fact the demodulated signal at 6 MHz of this photodiode is used as error signal for the locking.
  • On the B2_PD1 photodiode (out of loop photodiode) the height of the 6 MHz line increases by a factor 40 when the power changes from 1 mW to 10 mW. Thus the 6 MHz line increases faster than the power. In case of electronics non linearity one would expect the opposite, that is to say the 6 MHz line increasing slower than the power. Thus we can exclude the electronics non linearity to be the cause of this behaviour.
  • Looking at the blue spectrum (arms locked) one can see that the line at 700 kHz is not visible on B2_PD1 while it is present on B2_PD3. This coincides with having a larger 6 MHz line and 8 MHz line on B2_PD3 than on B2_PD1 when the photodiodes are both out of loop. It could be an indication that the 700 kHz line is related to a combination of frequencies involving the 6 MHz or the 8 MHz.


Images attached to this comment
Non-image files attached to this comment
Search Help